Small ruminant research and development in Africa Proceedings of the First Biennial Conference of the African Small Ruminant Research Network ILRAD, Nairobi, Kenya 10-14 December 1990 July 1992 African Small Ruminant Research Network Reseau africain de recherche sur les petits ruminants ILCA, P.O.Box 46847, Nairobi, Kenya Correct citation: Rey B, Lebbie S H B and Reynolds L (eds) . 1 992. Small ruminant research and development in Africa. Proceedings of the First Biennial Conference of the African Small Ruminant Research Network, ILRAD, Nairobi, Kenya, 10-14 December 1990. ILCA (International Livestock Centre for Africa), Nairobi, Kenya. Small ruminant research and development in Africa Proceedings of the First Biennial Conference of the African Small Ruminant Research Network ILRAD, Nairobi, Kenya 10-14 December 1990 Edited by: B. Rey S.H.B. Lebbie L. Reynolds July 1992 African Small Ruminant Research Network Reseau africain de recherche sur les petits ruminants ILCA, P.O.Box 46847 "»imhi Kenva "This one 5LKR-HUD-794Z Abstract This volume contains papers and abstracts of the First Biennial Conference of the African Small Ruminant Network. Ten papers deal with the Small Ruminant Production System and Policy; Eight papers on Small Ruminant Performance and Reproductive Physiology; eleven papers on Small Ruminant Reproductive Wastage and Health. Another nine describe and analyse feeds and feeding systems and there are 1 1 papers on Small Ruminant Genetic Resources and Breeding. Four papers on Small Ruminant Development in Africa complete this volume. Keywords /Sheep//goats//small ruminant production systems//productivity//economics/ /weaning practices//agropastoral//pastoral//production constraints//feeds//feed resources//feeding systems//digestibility//nutrition//supplementation//diet/ /reproduction//reproduction wastage//performance evaluation//health//peste des petits ruminants//infectious diseases//breeds//breeding systems//breed improvement// Africa/ Résumé Le présent ouvrage rassemble les communications et les résumés des communications présentées lors de la I" Conférence biennale du Réseau africain de recherche sur les petits ruminants. Dix documents portent sur les systèmes et les politiques d'élevage des petits ruminants, onze sur les performances des petits ruminants et la physiologie de la reproduction, huit sur les pertes en reproduction et l'état sanitaire des petits ruminants. Par ailleurs, neuf documents sont consacrés à la description et à l'analyse des ressources alimentaires et des systèmes d'alimentation et onze traitent des ressources génétiques et de l'amélioration génétique des petits ruminants. Enfin, le développement de l'élevage des petits ruminants en Afrique est abordé dans quatre autres documents. Mots-Clés lOvins//caprinsllsystèmes de production de petits ruminants//productivité! léconomiellsevragellagropastorall/pastorall/contraintes de production/ lalimentationllressources alimentairesllsystèmes alimentaires/ldigestibilité/ Inutritionllsupplémentationllreproductionllperte de potentiel reproductif/ /évaluation des performancesl/santél/peste des petits ruminants//maladies infectieuses//races/laméliorationllsélectionlll Afrique/ ISBN 92-9053-258-0 Table of Contents/Table des matieres Small Ruminant Research in Kenya B.A.J. Mwandotto, J.D. Wachira and V.C. Chemitei 1 SESSION 1. SMALL RUMINANT PRODUCTION SYSTEM AND POLICY/SYSTEMES ET POLITIQUES D'ELEVAGE DES PETITS RUMINANTS Intensification de la production des petits ruminants: pieges et promesses J.P. Boutonnet 9 Economics of small ruminant meat production and consumption in sub-Saharan Africa Senait Seyoum 25 Short-run demand for goat meat in Cameroon J.P. Ayissi Mbala 47 Intra-annual sheep price patterns and factors underlying price variations in the central highlands of Ethiopia Andargachew Kebede and Ray Brokken2 59 Differences by geo-climatic zone in the economic returns from wool productivity and quality in Lesotho John P. Hunter 77 Partial budget analysis for on-station and on-farm small ruminant production systems research: Method and data requirements S. Ehui and B. Rey 91 Comparison of consumer attitude towards and acceptance of goal, sheep and cow milk in Malawi J.W. Banda 105 Panurge: système d'investigation en recherche sur les systèmes d'élevage traditionnel O. Faugere, B. Faugere, Ch. Moulin et M. Ndiaye 115 Synthase des resultats de suivi et des essais zoosanitaires chez les petits ruminants effectues par le volet recherche du Projet sectoriel d'élevage au Mali T. Tangara, M. Doumbia, S.S.B. Ba, T. Kibe et C.F.M. Diallo 125 Black Head Ogaden sheep under traditional management practices in south-eastern Ethiopia Girma Adugna 1 33 SESSION 2. SMALL RUMINANT PERFORMANCE AND REPRODUCTIVE PHYSIOLOGY/PERFORMANCES DES PETITS RUMINANTS ET PHYSIOLOGIE DE LA REPRODUCTION Parametres de production des ovins Mossi de Gampela A.J. Nianogo 145 Principaux parametres démographiques des ovins et caprins des parcours du centre de la Somalie 0. Bourzat, K.D. Gautsch, M.P.O. Baumann etK.H. Zessin 159 Productivite du mouton Djallonké dans le système agroforestier traditionnel du Mayombe congolais A. Batalou-Mbetanie 171 Reproductive performance of indigenous goats in traditionally managed flocks in north-east of Zimbabwe L.R. Ndlovu 177 An analysis of lambing records of West African dwarf sheep kept at lle-lfe, Nigeria I.K. Odubote 185 Factors influencing flock structure and production performance dynamics of breeding sheep and goat station flocks in Kenya Geoffrey N. Angwenyi 193 Controle des beliers dans la gestion de la reproduction ovine A. Mazouz, F. Toe, A. Lahlou-Kassi et L. Derquaoui 195 Puberte chez la race D'man, la race Sardi et leur produit de croisement L. Derquaoui, R. Boukhliq, A. Lahlou-Kassi, A. Mazouz et F. Toe . . . 207 Caractéristiques de reproduction des femelles ovines de race Barbarine G. Khaldi et N. Lassoued 223 IV The relative importance of tactile, visual, auditory and olfactory stimuli in oestrus detection in West African dwarf ewes G.O.Oyediji, M.O. Akusuland G.N. Egbunike 233 Peripheral plasma levels of progesterone and oestradiol-17B of West African Dwarf goats during the oestrous cycle M.O. Akusu, E. Nduka and BA. Soyebo 239 SESSION 3. SMALL RUMINANT REPRODUCTIVE WASTAGE AND HEALTH/PERTES EN REPRODUCTION ET L'ETAT SANITAIRE DES PETITS RUMINANTS Quelles peuvent etre les priorités de recherche dans le domaine de la pathologie des petits ruminants en Afrique? F. Thiaucourt, J. Fikre, J.J. Tulasne, G. Mebratu, C. Guerin, D.M. Antonio 247 Effet des parasites gastro-intestinaux sur la duree de l'anoestrus post-partum chez la brebis Djallonké M. S. Hounzangbe Mawule 259 The effects of endoparasites on the productivity of Ethiopian highland sheep Tekelye Bekele and O.B. Kasali 271 Lamb and kid wastage through slaughtering of pregnant ewes and goats at Enugu and Nsukka abattoirs in Anambra State, Nigeria L.O. Wosu and E.C. Dibua 283 La mortalite des jeunes dans les élevages extensifs traditionnels de la zone sud-soudanienne du Senégal (site de Kolda) M. Ndiaye 285 Optimal time for vaccination against peste des petits ruminants (PPR) disease in goats in the humid tropical zone in southern Nigeria L.O. Wosu, J.E. Okiri and PA. Enwezor 297 Prophytaxie chez les petits ruminants au Senégal: régionalisation d'une politique nationale de protection sanitaire O. Fat/gere, E. Tillard et B. Faughre 307 Gastro-intestinal parasites in small ruminants in Mali: Geographical distribution epidemiology and chemotherapy S. Tembely, T. J. Galvin, B. KouyatS, S.B. Ba, K. Bengaty and W. Berckmoes 317 SESSION 4. SMALL RUMINANT FEEDS AND FEEDING SYSTEMS/RESSOURCES ALIMENTAIRES ET SYSTEMES D'AUMENTATION DES PETITS RUMINANTS Strategies for matching feed resources to small ruminant needs: A review R.M. Njwe 329 Some experiences in adapting previously free- ranging traditionally managed Matabele goats of Zimbabwe to individual stall-feeding L.M. Sibanda, L.R. Ndlovu and M.J. Bryant 345 Fattening mature indigenous (Matabele) goats: Effects on animal performance, body and carcass composition P.R. Hatendi, T. Smith, L Ndlovu and C. Mutisi 355 The utilisation of sorghum stover fed to sheep as influenced by urea or graded levels of lablab supplementation I.F. Adu, BA. Fajemisin and A. M. Adamu 367 Lablab (Dolichos lablab) meal as protein supplement for weaned fattening lambs W.D. Mafwere and LA. Mtenga 375 Experiences in protein supplementary feeding of weaned lambs and goat kids in Tanzania: The issue of dietary energy LA. Mtenga and A. Madsen 387 Effects of plane of nutrition on growth performance and carcass composition on lambs in Tanzania E.E. Massae and LA. Mtenga 401 Lamb production from Stargrass-Silverleaf Desmodium pastures Anna Warambwa 417 The role of caged layer waste as a nitrogen supplement to fibrous crop residues commonly fed to sheep and goats S.B. Kayongo, M.M. Wanyoike, P.N. Mbugua, P.N. Nyagah and T.E. Maitho 419 SESSION 5. SMALL RUMINANT GENETIC RESOURCES AND BREEDING/RESSOURCES ET AMEUORATION GENETIQUE DES PETITS RUMINANTS Breeding strategies for small ruminant productivity in Africa G.H. Kiwuwa 423 vi The performance, potentials and limitations of the West African Dwarf goat for meat production in the forest belt of Ghana A.K. Tuah, M.K. Buadu, F.Y. Obesel and K. Brew 435 Comparative performance of improved meat goats in Malya, Tanzania S.M. Das and D.S. Sendalo 445 Dairy goat breeding in Malawi: Gestation length, birthweights and growth of the indigenous Malawi goats and their Saanen crosses S.K. Kama andJ.W. Banda 453 Composition and yield of milk from non-dairy goats and sheep in Malawi J W Banda, J Steinbach and H-P Zerfas 461 Les performances de difterents niveaux de croisements: chevre alpine et petite chevre de l'Afrique de I Est a la Station d'élevage caprin de Vyerwa (Burundi) C. Ntahimpereye 485 Bodyweight measurements relationship in Nigerian Red Sokoto goats A. Hassan and A. Ciroma 491 Eruption of permanent incisors in indigenous goats and sheep O. Matika, R. Sibanda and M.L. Beffa 499 Growth traits of the Dorper sheep I. Factors influencing growth traits BA.O. Inyangala, J.E.O Rege and S. Itulya 505 Growth traits of the Dorper sheep. II. Genetic and phenotypic parameters BA.O. Inyangala, J.E.O. Rege and S. Itulya 517 Milk yield of Cameroon Dwarf Blackbelly sheep R.M. Njwe and Y. Manjeli 527 SESSION 6. SMALL RUMINANT DEVELOPMENT IN AFRICA: IMPLICATION FOR RESEARCH/DEVELOPPEMENT DE L'ELEVAGE OVIN-CAPRIN EN AFRIQUE: CONSEQUENCES POUR LA RECHERCHE Research highlights from the Small Ruminant Collaborative Research Support Program: The Dual-Purpose Goat (DPG) P.P. Semenye, A.N. Mbabu, B. Mwandotto, F.B. Nyaribo, J.M. Onim and F. Rurangirwa 533 vii Programme de développement de l'élevage des petits ruminants au Togo: essai de mise au point des possibilités d'association de la production vivrière et de l'élevage ovin I.Y. Pessinaba 545 Recherche et développement des caprins et ovins au Burundi /. Nsabiyumva 555 Le programme national de sélection ovine (PNSO) de Côte d'Ivoire: mise en place du contrôle de performances en ferme A. Oya 561 viii Preface One of the major aims ofthe African Small Ruminant Research Network (SRNET) is to promote and improve the exchange of information on small ruminant research and development, particularly among scientists working on small ruminants in Africa. One of the mechanisms used by the Network in pursuit of this goal is its biennial scientific conference. This brings together scientists working on small ruminants in Africa and elsewhere to share information on their research and development experiences. This proceedings is an account of the First Biennial Conference, held in Nairobi, Kenya, from 10 to 14 December 1990. The conference comprised six scientific sessions, during which 47 papers were presented and discussed. The sessions focused on production systems and economics, performance and reproductive wastage, feeding and nutrition, genetic resources, reproduction and breeding and small ruminant development projects in Africa and their implications for research. The conference was attended by more than 80 scientists from 33 African and 3 non-African countries. Although aimed primarily at researchers, we hope that the proceedings will be useful to others involved in development work with small ruminants. The organisers are grateful to the European Economic Community (EEC) for funding the conference and the publication of the proceedings. Sincere thanks are due to all who worked to make the conference a success. Particular thanks are due to Dr B. Rey, the then acting SRNET Coordinator, on secondment from the Institut d'elevage et de medecine veterinaire des pays tropicaux (IEMVT), for overall organisation; the International Laboratory for Research on Animal Diseases (ILRAD) for providing the conference facilities and moral support; the untiring and efficient secretarial staff of SRNET at ILCA headquarters, Addis Ababa, Ethiopia; and the administrative and support staff at ILCA's offices in Nairobi for the superb logistical arrangements. Special thanks are extended to the Government of Kenya for its support of the conference and to the Kenya Agricultural Research Institute (KARI) for co-hosting the conference. Particular thanks go to Dr D. Wachira, Deputy Director of KARI. Finally, I am grateful to all those who contributed to the production of this proceedings, especially Selamawit Dominique for typing the manuscript and staff of ILCA's Publications Section for production services. S.H.B. Lebbie SRNET Coordinator IX Small Ruminant Research in Kenya B.A.J. Mwandotto,^ J.D. Wachira2 and V.C. Chemitei2 'Small Ruminant Collaborative Research Support Program (SR-CRSP) 2Kenya Agricultural Research Institute (KARI) Introduction Kenya has an estimated 7 and 8 million sheep and goats, respectively. Table 1 gives the breakdown of this important animal resource base into hair and wool sheep and meat- and dairy-goat types. The significance of the small-ruminant sector in the overall agricultural economy was recognised in 1970's with the establishment of the Sheep and Goat Development Project. Due to increasing demand for animal products, a proportionate share of increase in small ruminant productivity is necessary without environmental degradation. This development will inevitably come through research. This paper outlines the small ruminant research trends from the 70's up to the present and projects on what will likely be the research requirements by the year 2000 AD, given the current socio-economic development patterns in the country. Table 1. Kenya's sheep and goat population, 1 964-1986 ('000s) Year Species/Type 1964 1978 1984 1986 Hair sheep 3481 6000 5599 5874 Wool sheep 397 500 551 726 Sheep total 3878 6500 6150 6600 Meat goats 5090 8500 6854 7488 Dairy goats - 2 - 512 Goat total 5090 8502 6854 8000 Grand total 8568 15002 13004 14600 Source: Ministry of Agriculture and Livestock Development animal reports1. The Sheep and Goat Development Project (SGDP) - 1970's This Project was assisted by FAO/UNDP. It had a broad objective of increasing productivity of small ruminants in all production areas of Kenya. Associated with this broad objective, this long- term Project also addressed the following areas: (1 ) Establishment of multiplication centres for Galla goats and Dorper sheep; (2) breed comparison trials which included indigenous Red Massai sheep and East African goats; (3) management of small ruminants: housing, dipping, feeding and disease control; (4) establishment of some data base for monitoring and research; (5) implementation of marketing operations; (6) training of production and research scientists. This Project is still on with Kenya Government funding. Some of the successes of the Project can be listed as follows:- 1 . The general awareness ofthe significance of small ruminant industry to policy makers was enhanced. 2. Research and production scientists were trained to MSc and PhD levels. 3. Long-term breeding programmes were established in some centres. 4. The establishment of basic management principles for sheep and goats on a nationwide scale as opposed to previous individual farmer or communal approaches. Some of these recommendations included:- (a) control of Haemonchus contortus and other worms; (b) housing against wind and rain; (c) cleaning the pens; (d) dipping against ectoprasites; (e) appropriate; (f) the complementary nature of sheep and goats to the larger herbivores; raised pen floors; (h) use of acacia pods to supplement nutrition. 5. Establishment of breed societies with the Kenya stud book. The Project also had some shortcomings: It lacked sufficient co- ordination. Progress in the field could not be adequately monitored. Breeding and selection criteria of the studs were not established. Overall production package could not be advanced. The small-scale farmer production system was not addressed. There was marked deterioration in specific production segments like prime lamb and wool production. Most of the work was done on-station and it lacked on-farm socio-economic considerations. The present - 1980's - 1990 Based on the experience gained from the SGDP, other projects were conceived that had a direct bearing on the research aspects on sheep and goats. The Small Ruminant Collaborative Research Support Program (SR-CRSP) The Project is supported by the United States Agency for International Development (USAID) and has been in existence now for about 10 years. The Project's principal objective in Kenya is to develop on systems analysis research basis a dual-purpose goat (DPG) of 40 kg mature weight with a potential of 4 kg daily milk field at peak lactation to satisfy home milk requirements of families in small-scale farming areas. The selling points of this Project are: 1. To synthesise a DPG from a broad base of adequately sampled two indigenous goats, the Galla and the East African and two exotic dairy goats, the Toggenburg and Anglo-Nubian, based on systems analysis principles and the management capability to small-scale farmers. The parental combination ensures high adaptability to local environmental fertility and high meat and milk production. It also ensures high heterosis retention in the economic traits in the stabilised new DPG. So far, the first composites are in place and breeding is in the correct direction in terms of yields of two kg of milk per day on one milking (Mwandotto et al, n.d). The use of artificial insemination (Al) for goat production in our environment has been demonstrated in the process of development of this DPG and adaptability of the DPG to different production zones is planned in the development of the Project. 2. The project is multifaceted in that the breeding research is augmented by simultaneous nutrition management, disease and socio-economic research aspects in order to evolve a comprehensive production package for the producers, training institutions and extension agents. That package is now at its final stage of development. 3. The project aims at producing low cost production technologies: (a) use of weigh-band to estimate weight of animals on the farm for on-farm research - see below; (b) development of non-conventional fodders - vines and agroforestry shrubs and farm by-products for goat feeding; (c) use of Haemonchus contortus resistant animals to lower the cost of maintenance of breeding flocks. With a heritabilrty estimate of 0.397 for Haemonchus resistance in the breeding flock (Rohrer et al, 1990, unpublished), the resistance can be selected for in the future DPG; (d) partial milking for cheap and hygienic feeding and still able to adequately estimate milk yield (Ruvuna etal, 1988). 4. The project has an important aspect of on-farm production and evaluation research aspects where DPG parentals are monitored in terms of acceptability by target farmers, production levels and marketability of the goats and their products. 5. The project is a shining example of institutional collaboration - USAID through, University of California, Davis, Winrock International, Texas A&M University, Missouri University and Washington State University and the Kenya Government through KARI. It is a model for integrated research. 6. The project further trained research scientists to MSc and PhD levels. The SR-CRSP has its inherent shortcoming as well: It addresses only goats for small-scale farmers in high potential areas for now. It is, however, clear that there are many animal production systems in Kenya that are dependent on small ruminants (Thorpe et aI, n.d.) and with the possibility of genotype- environment interaction and correlations, universal production packages across all production systems cannot be possible. The Integrated Project in Arid Lands (IPAL) and International Livestock Centre for Africa (ILCA) The IPAL Project was set in the arid lands of northern Kenya with the Federal Republic of Germany and UNESCO aid. The selling point of this Project was that the small ruminants were researched on in their traditional setting and management and in integration with cattle and camels. Prospective innovations were monitored carefully. One of the notable of the projects knowledge of the resilience of small stock in sustaining its population after drought catastrophes (Blackburn, 1984). The limitation here was that it was also location-specific. The ILCA project in Kajiado showed the importance of marketing small ruminants on weight, body condition and sex basis (Chabari et al, 1987). The Embu-Meru-lsiolo Goat and Sheep Project This as a non-governmental organisation (NGO) Project based at Marimanti and sponsored by the Governments of Kenya and the United Kingdom. It is a good example of an NGO conducting research to augment the national efforts of research information generation. It has conducted on-farm recording and breed comparisons. Even though there is some confounding in data structure, it is indicated that there is no difference between the Deguin and the Boran Galla in milk production (Skea, n.d.). As we move to on-farm research activities, apparently care will have to be taken in experimental designs in order to get as much information as possible from trials. The Universities and Agricultural Development Corporation (ADG) Farms The University farms do keep some stock for teaching purposes complete with good records especially in purebreed dairy goats. Most of these flocks have, however, tended to be small in size and consistent long-term observations on them have not been possible. Large flocks of both sheep and goats are, however, in the hands of ADG with a possibility of large data sets of pedigree and production kept for long periods. This is an area of ADC flocks and their breeding/management programmes. Flocks in Other Government Farms As a result of the first project of the 1970's some data have accumulated on government farms which have been analysed recently to reveal important research information. On crossing the Galla to the East African goat and the Boer (Angwenyi, 1984), the East African superiority in adaptation and the superior mothering ability of the Galla was indicated. Furthermore, a Galla x East African cross dam gave a favourable maternal heterosis that would be useful in raising Boer cross kids. Shortcomings in the experimental designs limited the information on maturing characters that could have also been generated by that study. On the same farm, Maasai and the Dorper with apparently a combination between 47/64 and 7/8 Dorper being ideal. In another study in high altitude areas, the Hampshire Down sheep grew faster than the Red Maasai with the Dorper being in between the two breeds and the carcass characteristic also differed between the breeds (Wachira et al, n.d.) while interactions between breed and sex occurred for meat traits. The future 1990's Given the experience of the past and the present, research projection on small ruminants by the year 2000 AD in Kenya may be viewed as follows:- 1 . A more integrated and co-ordinated approach will be needed instead of the research being zonal and addressing only small sections of the societies. 2. Well designed studies will be set that have adequate experimental material and without confounding effects in order to generate comprehensive information per unit of resource invested in research. Current methods of data analyses and animal evaluation will also be employed. 3. Products that have so far not been included in our priorities will be addressed - prime lamb and wool. A research laboratory in wool and mohair is envisaged in order to support the natural fibre industry. There should be more research on Merino, Corriadale, Romney mash sheep (which form more that 10% of the sheep population) and Angora goats. A proposal to develop a prime lamb production industry, near urban centres for ease of marketing has been proposed by Angwenyi (n.d.). 4. Revitalization of small ruminant breed societies and formation of new ones to promote and develop the characteristics of individual breeds in order to support any planned crossbreeding programmes. 5. More multidisciplinary approach to research will be in place in line with the example of SR-CRSP with multi-institutional involvement for any possible complementarity and more efficient use of research funds and research expertise and release of technological packages of different production systems. 6. More intensive search for resistant animals against endemic diseases and understanding the mode of inheritance of the possible resistance in order to cut down the cost of drugs in maintaining flocks. 7. More diversified fodder crops for small ruminants i.e. studies on shrubs and agro-forestry trees and use of goats for more efficient control of bush enchroachment in range areas. 8. Broaden the definition of small ruminants to include rabbits (in intensive agriculture) and small wildlife species like the undertake (in range areas) and undertake studies on the economic integration of those species in animal agriculture. In ranching situations, wildlife meat is relevant as tourist hotel industry moves into individual ranches and the lean wildlife meat may earn premiums. Harvesting of wildlife products from non-endangered species was recommended in 1985 (IBAR, 1986). 9. Small ruminant meat characteristics will be studied in detail including meat from the (new) species in order to develop objective grading criteria and more use of by-products from the small ruminant industry. 10. Use of biotechnological principles on breeding better animals and development of vaccines (animal health management through biotechnology). Embryo transfer technology will be built on the current Al experience and successful commercialisation of contagious caprine pleuropneumonia (CCPP) vaccines will stimulate the development of multivalent vaccines to cut down cost of production and administration of the drugs. 11. Training in very specialised fields in small ruminant production i.e. reproductive physiology, embryo transfer, development of gene markers and probes etc. References Ahuya C 0, Mwandotto B A J, Ruvuna F and Taylor J F. (n.d.). Use of heartgirth measurements to estimate liveweight in goats, (n.d.). In: Proceedings of the Eighth SR-CRSP Scientific Workshop, 7-8 March 1990, ILRAD, Nairobi, Kenya. Small Ruminant Collaborative Research Support Program, Nairobi, Kenya. pp. 29- 32. Angwenyi G N. 1 984. Effects of crossbreeding East African, Galla and Boergoats on body size, growth rate and kid survivability in Kenya. MSc thesis, Texas A & M University, College Station, USA. Angwenyi G N. (n.d.). Performance of Suffolk sired lambs with Dorper x Maasai dams. In: Proceedings of the Seventh SR-CRSP Scientific Workshop, Nairobi, Kenya, 22-23 February 1989. Small Ruminant Collaborative Research Support Program (SR-CRSP)/Animal Production Society of Kenya (APSK), Nairobi, Kenya, pp. 136-140. Blackburn H D. 1984. Simulation of genetic and environmental interaction of sheep production in northern Kenya. PhD dissertation, Texas A&M University, College Station, USA. Chabari F N, Ackello-Ogutu A C and Odhiambo M O. 1987. Factors determining market prices of small ruminants from a pastoral production system in Kenya. East African Agricultural and Forestry Journal 52 (4): 286-292. IBAR (Inter-African Bureau for Animal Resources). 1986. Wildlife/livestock interfaces on rangelands. Proceedings of a conference held at Taita Hills Lodge, Kenya. Mwandotto B A J, Ahuya C O, Ruvuna F and Taylor J F. (n.d.). Milk yield potential of the Kenya dual purpose goat (DPG). In: Proceedings of the Eighth SR-CRSP Scientific Workshop, 7-8 March 1990, ILRAD, Nairobi, Kenya. Small Ruminant Collaborative Research Support Program, Nairobi, Kenya. pp. 12-14. Ruvuna F, Cartwright T C, Blackburn H, Okeyo M and Chema S. 1988. Lactation performance of goats and growth rates of kids under different milking and rearing methods in Kenya. Animal Production 46:237-242. Skea I W. (n.d.). A comparison between the Deguin and the Borana Galla goats over four years, with reference to fertility, fecundity and kid growth rates and milk production during the first 42 days. In : Proceedings of the Seventh SR-CRSP Scientific Workshop, Nairobi, Kenya, 22-23 February 1989. Small Ruminant Collaborative Research Support Program (SR-CRSP)/Animal Production Society of Kenya (APSK), Nairobi, Kenya. pp. 131-135. Thorpe W, Nyambaka R, Chabari F, ole Maki M and Rugema E. (n.d.). Small ruminants in the farming systems of coastal Kenya. In: Proceedings of the Eighth SR-CRSP Scientific Workshop, 7-8 March 1990, ILRAD, Nairobi, Kenya. Small Ruminant Collaborative Research Support Program, Nairobi, Kenya. pp. 178-183. Wachira J D, Kamau C R, Chemitei V C, Muia J K and Gichuru W N. (n.d.). Carcass composition evaluation of sheep grazed on Kikuyu grass {Pennisetum clandestinum) in high altitude area of Kenya. In: Proceedings of the Eighth SR-CRSP Scientific Workshop, 7-8 March 1990, ILRAD, Nairobi, Kenya. Small Ruminant Collaborative Research Support Program, Nairobi, Kenya. pp. 295-309. Session 1 Small ruminant production system and policy Systèmes et politiques d'élevage des petits ruminants Intensification de la production des petits ruminants: pièges et promesses J.P. Boutonnet Institut national de la recherche agronomique Station économie et sociologie rurales, 34060 Montpellier Cedex 1 (FRANCE) Résumé L'élevage ovin-caprin africain peut apporter plus de viande sur les marchés locaux à condition que cette intensification porte essentiellement sur le travail paysan. Les apports d'alimentation cultivée doivent rester limités à un niveau compatible avec les prix de marché. Les ateliers intensifs spécialisés, concurrencés par la production avicole, etdépendants des disponibilités en céréales, ne peuvent pas connaître un développement significatif. Which intensification process for small ruminant production in Africa? J.P. Boutonnet Abstract Sheep and goathusbandry in Africa can produce more meat for local markets, only with more labour-intensive systems in small farms. Feedstuffs must be very limited, in relation with market prices. Intensive production systems based on cereals compete with poultry production and are unlikely to develop significantly. Introduction L'élevage des petits ruminants présente au niveau mondial, une quadruple spécificité: • il est mené dans des conditions techniques extrêmement diverses selon les situations agro-économiques dans lesquelles il s'insère, ce que reflète la diversité des prix de sa viande dans le monde; • il n'a connu nulle part l'évolution vers le type de production intensive de masse qu'ont connue les autres grandes productions animales (volaille, porc, lait de vache); • c'est un élevage à triple fin: viande, laine, lait dans des proportions très variables selon les situations; • enfin, le petit format de ces herbivores domestiques exclut leur utilisation en animaux de trait; en revanche, il les rend faciles à manipuler par leur détenteur, mais plus sensibles aux prédateurs que le gros bétail, d'où la nécessité de leur protection. La production marchande de viande d'herbivores: un sous-produit De tout temps et dans le monde entier (sauf dans les pays riches depuis quelques dizaines d'années), la consommation de viande est très réduite, réservée aux fêtes ou aux élites, l'alimentation de base étant constituée de féculents (céréales, tubercules) accompagnés de protéines végétales (Iégumineuses). La viande la plus consommée est issue d'un élevage ayant d'autres fins: laine (ovins); lait; cuir (bovins, caprins, ovins); travail (bovins, équins, camelins); fumier (ovins en zone méditerranéenne). Les viandes de granivores (oiseaux, porcins), qui ne produisent rien d'autre que de la viande, sont un très grand luxe réservé à une élite, ou tout simplement interdites (interdiction du porc par l'islam). Dans cette situation, on peut donc dire qu'il n'y a pas d'animaux élevés spécialement pour la production de viande, si ce n'est quelques unités élevées, en basse-cour, à partir de résidus de culture ou de déchets domestiques, à seule fin d'autoconsommation familiale. Dans tous les cas, la consommation de viande est occasionnelle. De plus, les animaux dont on mange la viande sont des herbivores. Ainsi, dans l'occupation de l'espace disponible, l'homme ne travaille que les terres consacrées à la production de végétaux destinés à l'alimentation humaine. Les animaux se nourrissent sur les espaces non cultivés. Ils sont d'autant moins abondants que ces espaces sont plus réduits, notamment en cas de pression démographique. Consommation de masse de viandes: production à base de céréales En Amérique du Nord, puis en Europe, la croissance démographique urbaine et l'augmentation des revenus ont créé depuis quelques décennies les conditions d'une forte croissance de la demande en viande. Dans le même temps apparaissent des substituts industriels à certaines des productions des 10 herbivores élevés sur de grands espaces (le coton puis les textiles artificiels remplacent la laine, le caoutchouc et le PVC remplacent les cuirs, la motorisation remplace les chevaux et les boeufs de trait, les engrais chimiques le fumier). Les grands troupeaux d'herbivores sur pâturages spontanés, dont le marché des produits autres que la viande se déprime, ne peuvent pas répondre à l'augmentation très forte de la demande en viande. En revanche, les progrès techniques importants dans la culture céréalière permettent de consacrer une part de plus en plus importante des terres cultivées à la production d'alimentation animale. L'élevage intensif de masse, à seule fin de production de viande (volaille, porc), se développe alors dans ce contexte particulier. Il est fondé sur une utilisation massive de céréales ou de fourrages cultivés. Dans les deux cas, il s'agit de cultures impliquant labour, fertilisation, semence, par opposition à l'exploitation de territoires à végétation poussant sans intervention humaine directe. Les herbivores sont moins efficaces que les granivores pour la transformation des céréales en viande. Le développement d'un élevage d'herbivores à viande exige donc qu'une partie de leur alimentation reste assurée par le pâturage d'herbe spontanée. Si la demande en viande d'herbivores leur attribue un prix plus élevé qu'à celle de granivores, la proportion des céréales dans leur alimentation pourra être d'autant plus importante. Le développement des élevages marchands d'herbivores pour la viande se fait donc toujours sur la base d'une utilisation combinée de céréales ou de cultures semblables (au moins la phase finale), et de pâturages spontanés. C'est pourquoi, du point de vue du développement de l'élevage et de la gestion de l'espace, ce qui distingue les différentes modalités techniques n'est pas la proportion fourrages concentrés/fourrages grossiers, mais la proportion fourrages cultivés/fourrages spontanés. Aujourd'hui encore, la majeure partie de la viande d'herbivores consommée dans le monde provient d'élevages conduits à d'autres fins (bovins-lait, ovins-laine) et l'essentiel de la croissance de la consommation de viande des dernières décennies provient d'une production de viande à base de céréales (volaille, porc). Cette logique de la production de viande, comme produit principal ou sous-produit, et selon l'utilisation d'herbe, est reflétée par les quantités et les prix sur le marché international (tableau 1). Les viandes de granivores s'échangent beaucoup moins (5% de la production mondiale) sur le marché mondial car, fabriquées à base exclusive de céréales, ce sont les céréales qui s'échangent et il n'y a pas d'avantage comparatif à une localisation proche des zones de production. En revanche, les viandes d'herbivores, selon les disponibilités en ressources spontanées et leur statut de sous-produits, peuvent avoir des prix de marché très différents. L'essentiel des exportations mondiales sont le fait de 11 pays à faible densité de population et à climat tempéré (hémisphère Sud), où, de plus, l'élevage ovin est essentiellement orienté vers la laine. C'est pourquoi, la viande ovine est moins chère sur ce marché que la viande bovine qui est dans ces pays la principale fin de l'élevage bovin. Tableau 1 . Production mondiale et commerce international des viandes monde 1 988 Production Importations Importations en Prix moyen des totale totales %de la exportations en Ovine et caprine milliers de t milliers de t production dollars E.-U. par kg 9000 1 200 13 1.8 Bovine 50 200 6000 12 2.9 Porcine 64 400 3 700 6 2,2 Volaille 36 900 2 000 5 1,4 Source: FAO, (1989). Les différents types d'élevage ovin-caprin Toutes les formes d'élevage ressortent de deux grands types, que l'on retrouve dans l'histoire, avec des modifications particulières liées aux systèmes agraires dans lesquels cet élevage est pratiqué: • grands élevages spécialisés extensifs il s'agit de troupeaux de grandes dimensions, gérés comme un capital, exploités par des éleveurs spécialisés utilisant de très grands espaces, avec peu de travail par unité produite; • élevages paysans: il s'agit des troupeaux de petites dimensions de paysans dont la production est diversifiée. Les ovins, caprins sont l'une des diverses productions mises en jeu pour assurer l'entretien et le renouvellement de la famille. Les élevages spécialisés extensifs sont de deux types Les plus anciens sont ceux des pasteurs des régions steppiques. Les populations qui les pratiquent vivent sur des territoires où les cultures sont difficiles, voire impossibles. Les animaux constituent donc leur seul moyen de survie. Le lait (et parfois le sang) est le principal produit, pour l'autoconsommation. Les seuls animaux abattus sont les animaux malades ou réformés, ou ceux que l'on vend en cas de sécheresse ou pour se procurer du grain. Dans les zones plus arrosées, c'est le pouvoir politique, appuyé éventuellement sur la force militaire, qui impose l'affectation, au détriment des cultures, de grandes parties du territoire à l'élevage du bétail, dans le but d'assurer une base d'approvisionnement dans le produit de cet élevage (Iaine, cuirs, chevaux, etc.) 12 quand celui-ci est l'objet d'un enjeu économique stratégique: c'est le cas des grands élevages actuels de l'hémisphère Sud et d'Amérique du Nord. Ce fut le cas dans le passé en Europe et en Asie du Sud-Ouest pour la fourniture de laine et de chevaux de guerre. Les deux types d'élevage extensif, qui ont besoin de vastes espaces à rente foncière faible, ne sont mis en oeuvre que si la densité de population agricole est très faible, soit en raison de contraintes pédo-climatiques, soit comme résultat d'un processus historique (colonisation, exode rural, etc.). Seule une évolution des structures agraires peut les faire apparaître ou régresser. L'élévation du prix de marché des produits, qui augmente les revenus, ne peut faire croître leurs effectifs que dans des limites très étroites. Toute intensification significative amènerait en effet, une augmentation excessive des besoins en travail; ce surcroît de travail serait mieux valorisé par d'autres productions ou simplement ne peut pas être assuré. Très sensibles aux aléas bioclimatiques, ces élevages connaissent de très amples fluctuations de leurs effectifs, de leurs performances, et des prix de marché. Les variations de valeurs de leur cheptel peuvent être beaucoup plus importantes que le produit qui en est issu. Les éleveurs sont donc souvent à juste titre, plus attentifs au négoce de bétail qu'à une intensification, dans tous les cas coûteuse. Les éleveurs paysans Dans les exploitations familiales de petites dimensions et disposant de peu de capitaux, le principal facteur pour assurer les besoins familiaux est l'utilisation du travail. On observe la réalisation d'un ensemble de productions dont l'association permet un renouvellement à moindre coût du potentiel de l'exploitation, et à tout moment une adaptation souple au travail disponible, aux opportunités du marché, aux variations climatiques. Le petit troupeau ovin/caprin joue ici un rôle important pour l'autoconsommation, mais certains des produits sont mis en marché. Dans la palette des différentes productions commercialisables, cet élevage constitue souvent une garantie contraléatoire. La littérature décrit abondamment son rôle de caisse d'épargne. Mais surtout il contribue à la diversité des sources possibles de revenu ou de nourriture. Cette production plus aisément fractionnable que l'élevage bovin, valorise principalement le travail familial, et les ressources fourragères qui n'auraient pas d'autres destinations (jachères, résidus de culture, landes). Les achats sont toujours très limités, et les investissements pratiquement absents. Comme les types extensifs, ces types d'élevage ovin/caprin peuvent fournir des produits au marché à des coûts peu élevés, mais ils ont peu de capacité de croissance interne des effectifs, même dans le cas d'augmentation du prix perçu. Ces élevages, dont la rationalité est le produit, et non la rentabilité du capital, et qui sont en outre peu sensibles aux aléas climatiques puisqu'utilisant rarement 13 la totalité des ressources fourragères disponibles, ont souvent des performances zootechniques élevées (prolificité, rendement laitier). Ils peuvent en outre jouer sur le marché un rôle régulateur (Mounier, 1989), vendant en période de hausse des cours, et accroissant leur autoconsommation ou conservant les animaux en période de cours bas. Les ateliers intensifs Dans certains cas bien particuliers se sont développées en Europe, des formes d'élevages intensifs sur un modèle ressemblant à celui de l'aviculture et de la production porcine: • spécialisation des troupeaux vers la viande ou le lait accompagnée de l'organisation collective de l'amélioration génétique; • intensification fourragère et moyens de gérer l'exploitation de l'herbe et de contrôler l'alimentation des animaux; • moyens de contrôle de la reproduction et de l'état sanitaire des brebis; • équipements, aménagements de locaux améliorant l'efficacité du travail; • forte liaison, souvent contractuelle, avec l'agro-industrie pour la fourniture de facteurs de production et l'écoulement du produit. Tous ces moyens ont pour objet d'augmenter la production sans utiliser plus de surface, mais en augmentant la consommation de biens et de services achetés qui permettent une intensification du travail et une diminution de la dépendance des aléas biologiques et agro-climatiques. Ce type d'élevage n'a pu se développer que ponctuellement, dans les cas où le prix du produit (viande ou lait) était élevé grâce à une demande spécifique (fromage de Roquefort, agneau de lait) non concurrencée par les productions plus performantes, et à une protection absolue du marché mondial. Ces élevages sont actuellement en crise en Europe du fait de l'ouverture des marchés et de la baisse des soutiens publics. Fortement tourné vers le marché, ce type d'élevage est cependant le seul qui puisse, au niveau d'un pays ou d'une zone, répondre à une augmentation du prix des produits par une croissance de la production dont la limite est seulement fonction du prix. Supériorités et faiblesses de l'atelier intensif Développé sur une large échelle et aux conditions que l'on a exposées plus haut, le modèle "atelier intensif" offre trois types d'avantages: • il permet une croissance très importante de la production; • il fournit des emplois à la population rurale; 14 • il est moins sensible que l'élevage extensif aux aléas bioclimatiques. En revanche, il comporte de nombreux dangers. Manque de souplesse Même si le revenu global de l'éleveur peut être supérieur à celui de l'élevage paysan, en raison du volume accru, le revenu par unité produite est relativement faible (figure 1) et toute baisse des prix met en cause la survie de l'éleveur, d'autant plus que cet élevage est devenu sa principale source de revenu. Elément perturbateur du marché Son succès même accroît les quantités offertes d'une façon telle que le marché connaît une tendance à la baisse des prix. Le produit risque alors de diminuer fortement, engendrant une hausse des prix etc. Ce type de production est générateur d'évolutions cycliques du marché. Pour les deux raisons précédentes, son développement nécessite une stabilité des prix: gestion publique puissante du marché, ou contractualisation des rapports avec l'acheteur garantissant l'écoulement et le prix. En outre, si dans le même espace commercial existe un élevage extensif de grande échelle, celui-ci va être incité à augmenter ses effectifs, aggravant le surpâturage, et accentuant les fluctuations du marché. Cette production intensifiée constitue souvent l'essentiel du revenu des paysans qui s'y adonnent. Les fluctuations du marché provoquent souvent leur ruine, puisque leur marge brute, forte par le volume produit, est faible à l'unité de produit (figure 2). Alors que pour l'élevage paysan cette production représente une faible part de ses ressources, il peut la retenir (stockage "on field") en cas de conjoncture basse, et revendre en cas de conjoncture élevée, d'autant que son système de conduite le lui permet mieux que l'atelier spécialisé, qui doit vendre les animaux dès qu'ils sont prêts. Productivité du travail inférieure à celle de l'élevage extensif Comme toute production intensive, c'est-à-dire qui permet d'augmenter le volume produit sur une surface donnée, il est exigeant en travail. Ce modèle ne peut donc être envisagé que si la pression démographique est importante, et que des élevages extensifs, disposant de grandes surfaces, n'ont pas accès au même marché. Il ne peut donc se développer que si les capacités de travail ne sont pas saturées. Intensif/extensif: une notion macro-économique Les conditions de l'intensification doivent être examinées non seulement au niveau de l'unité de production, mais également au niveau d'un espace-marché (pays). La différence des prix de la viande ovine (tableau 2) entre la CEE, l'Australie, l'Algérie ne peut s'expliquer complètement qu'au vu des conditions de productivité du travail et de la terre que connaissent ces différents espaces. 15 Figure 1 . Structure des couts et taille du troupeau selon differents types d'elevages ovins (UTA: Unite Travail Agricole). Dollar E.U./kg viande ovine u c (0 5- 2 4- c CD 3^ > 2- c a 1- 1 to> C n > r Marge Intrants Investissement Paysan 60 agneaux/UTA Atelier intensif 300 agneauxAJTA Hanching 1 500 agneauxAJTA Figure 2. Revenu de l'elevage ovin selon differents types d'elevage Revenu/UTA Extensif Atelier intensif Paysan Superficie/UTA 16 Tableau 2. Prix de marché du kg de viande ovine dans quelques pays -1988 En dollars des E.-U. En kgd 'orge Heure de salaire Australie 1 8 0,2 Europe (CEE) 5 25 1 Algérie 25 70 10 Pour l'ensemble de la production agricole, la productivité du travail diminue quand celle de la terre augmente. Les pays densément peuplés (Algérie) ou ayant opté pour être agro-exportateurs (France) sont condamnés à une forte productivité de la terre, à une faible productivité du travail agricole, et, si la demande existe, à une production ovine chère et intensive (tableau 3). Tableau 3. Comparaison internationale de critères d'intensivité Grande- France Bretagne Australie Algérie SAU/Population totale (ha/hab) 0,56 0,33 30,6 1,25 SAU/Population active agricole (ha/UTA) 19 30 1 151 1,1 Autosuffisance alimentaire (%) 117 57 308 30 Nombre d'humains nourris/ha 2,08 1,72 0,10 0,24 Nombre d'humains nourris par UTA 40 52 115 0,26 Ovins-caprins/ha SAU 0,38 1,35 0,33 0,45 En revanche, les pays peu densément peuplés (Australie) ou ayant opté pour être agro-importeurs (Grande-Bretagne) peuvent avoir une productivité dutravail agricole élevée, une productivité de la terre plus faible, et éventuellement un élevage ovin extensif et de grande dimension, fournissant des produits bon marché. Dans le cas de l'Algérie, l'élevage ovin, autrefois extensif et steppique, est devenu un élevage intensif fondé sur l'utilisation massive de céréales, grâce à une demande soutenue et la fermeture quasi absolue des frontières au marché mondial de la viande ovine. C'est ainsi que la production agricole totale du pays (tableau 4) a connu une intensification remarquable (+32% entre 1970 et 1987) malgré la stagnation des rendements en céréales, grâce à l'augmentation de la production de viandes rouges et de lait. La productivité pondérale des brebis (tableau 5) y a chuté de 25%. 17 Tableau 4. Production totale de la sole labourable en Algérie (en milliers de t d'équivalents-grain) 1970 1987 Indice 1987 (1970 = 100) bra in 1 740 Viande ovine 1 850 Viande bovine 320 Lait 880 Total 4 790 36% 64% 100% 1 620 25% 93 2 940 159 610 75% 191 1 200 136 6 370 100% 133 Tableau 5. Algérie: Performances techniques du troupeau ovin 1964/69 1982/87 Nombre total de brebis (millions) Production de viande ovine (milliers de t) Agneau/produits/brebis Poids moyen des carcasses (kg) Poids de carcasse produit par brebis (kg) 3,9 9,8 0,85 0,56 13,1 15,2 11,2 8,5 Vers des modèles intermédiaires Les modèles à proposer et les améliorations techniques à rechercher sont donc extrêmement dépendants: a) Des conditions agro-économiques dans lesquelles s'insèrent les élevages, exposées ci-dessus; b) Des types d'éleveurs, de leurs différentes productions, et de leurs objectifs économiques: autoconsommation; vente sur un marché (dans ce cas valorisation du facteur non saturé: terre, main-d'oeuvre); spéculation sur le cheptel; En effet, s'agissant de viande, le même animal est en permanence à double statut: animal de boucherie et appareil de production. Il en résulte que dans le cas d'un élevage dépendant des aléas climatiques les cours des animaux sont très fluctuants avec les conditions météorologiques. Les fluctuations de la valeur du cheptel détenu peuvent être, dans une année, beaucoup plus importantes que la valeur de la production finale. Les améliorations zootechniques proposées ne "passeront" que si leur mise en oeuvre ne perturbe pas la faculté d'adaptation de l'éleveur aux aléas du marché et du climat. 18 c) Des objectifs de l'action d'amélioration: accroissement de l'offre sur le marché et/ou accroissement du revenu des éleveurs. Le développement massif d'ateliers intensifs peut par exemple recourir à des importations d'aliments du bétail à bas prix, assurant la stabilité de ces ateliers mais provoquant une baisse des prix qui condamne les producteurs paysans locaux. Ces conditions conduisent à proposer des améliorations techniques en douceur, qui ne perturbent pas les modes d'organisation existants. On peut citer l'exemple de l'Algérie, où au cours des 20 dernières années, la production de viande ovine a doublé, avec une baisse sensible de la productivité numérique des brebis et une augmentation légère du poids des carcasses. Dans la situation actuelle, il est probable que tout effort visant à augmenter la productivité des brebis serait vain, tant que les fluctuations aléatoires des prix du bétail ne sont pas stabilisées. En revanche, on peut dans ce pays rechercher les moyens: • de stabiliser les ressources alimentaires par l'intensification fourragère; • d'augmenter encore le poids des carcasses (amélioration génétique, nutrition); • d'améliorer les indices de consommation de céréales (mode de conduite, action sur le marché). Dans les pays d'agriculture paysanne dominante, où la majeure partie de la production est autoconsommée, des améliorations douces peuvent accroître notablement les quantités mises en marché sans accroître démesurément la production totale, si elles sont mises en oeuvre par l'ensemble des paysans et si, dans le même temps, les circuits commerciaux s'organisent. Place du cheptel ovin-caprin africain dans le cheptel mondial En Afrique, il semble, vu la pénurie généralisée en denrées alimentaires, que la principale viande consommée soit de la viande d'herbivore, et en faible quantité. Le cheptel des ruminants ramené a la population humaine (tableau 6) y est d'ailleurs légèrement plus élevé que la moyenne mondiale, et même que la moyenne européenne. Cinq pays africains figurent parmi les 15 plus gros pays en termes de petits ruminants (tableau 7). Les petits ruminants y sont proportionnellement plus importants qu'ailleurs puisqu'ils représentent 21% du total des ruminants contre 15% en moyenne mondiale. Dans toute une partie du continent, en gros autour du Sahara, la production laitière est conséquente (40% à65% du lait produit), s'apparentant en cela, à l'élevage du bassin méditerranéen oriental (tableau 8). Ailleurs, sa production principale est la viande, sauf en Afrique du Sud (laine). 19 Tableau 6. Effectifs des ruminants - monde 1988 Total unites Ovins Caprins ovins 106 108 106 Dont ovins Total dont Population Bovins 106 caprins par ovins humaine % habitant caprins totale Afrique i98 200 167 1 753 21% 2,9 0,6 610 Amerique 421 124 37 3 108 5% 4.4 0,2 702 Asia 522 332 296 4 282 15% 1,4 0,2 2994 Europe 125 142 12 1 029 15% 2,1 0,3 497 Oceanie 32 229 2 455 51% 17,5 8,9 26 URSS 121 141 6 994 15% 3,5 0,5 286 Monde 1 419 I 168 520 11 621 15% 2,3 0,3 5115 Source: FAO, (1989) Tableau 7. Effectifs des petits ruminants (1 5 principaux pays - 1988) Ovins 106 Caprins 106 Total 106 Ovins-caprins total ruminants Ovins-caprins (% unites-ovins) par habitant Bresil 20 11 31 3% 0,2 Chine 102 78 180 21% 0,2 Inde 52 105 157 8% 0,2 Nigeria 13 26 39 32% 0,4 Pakistan 27 33 60 21% 0,5 URSS 141 6 147 15% 0,5 Iran 35 14 49 47% 0,9 Turquie 40 13 53 39% 1,0 Ethiopie 23 17 40 15% 0,9 Soudan 19 14 33 16% 1,4 Somalia 13 20 33 28% 4,7 Argentine 29 3 32 8% 1,0 Afrique du Sud 30 6 36 30% 1,1 Australia 164 1 165 50% 10,4 Nouvelle-Zelande 65 1 66 54% 20,0 Total 15 pays 773 348 1 121 16% 0,4 % du total monde 66 67 66 III III Source: FAO, (1989). 20 Tableau 8. Production de lait - monde et principaux pays - 1 988 (millions t) Vache et % petits bufflone Brebis Chèvre Total ruminants Afrique 14 1,5 2,0 18 20 dont Algérie 0,6 0,2 0,2 1,0 40 Niger 0.1 - 0,1 0,2 50 Somalie 0,6 0,5 0,6 1,7 65 Soudan 1,8 0,6 0,5 2,9 38 Amérique 114 - 0,5 115 0,4 Asie 85 3,7 3,8 93 8 dont Iran 1,7 0,7 0,2 2,6 35 Syrie 0,7 0,5 - 1.2 42 Turquie 3.2 1,2 0,5 4.9 35 Europe 173 3,7 1,7 178 3 dont France 28 1,1 0,4 30 5 Grèce 0,6 0,6 0,5 1,7 65 Italie 11 0,6 0,1 12 6 Roumanie 4,3 0,5 - 4,8 10 URSS 106 - 0,4 106 0.4 Océanie 14 - - 14 - Monde 506 9,0 8,3 523 3 Source: FAO, (1989). L'Afrique participe peu aux échanges intercontinentaux des produits de l'élevage ovin (ceux de l'élevage caprin sont pratiquement inexistants, hormis les échanges frontaliers): en animaux vivants (tableau 9), à part les échanges entre Etats voisins du Sahel vers la côte, on doit noter le courant de la corne de l'Afrique vers la péninsule arabique; en viande, un petit courant d'importation de l'Afrique du Nord en provenance d'Océanie. L'essentiel du commerce mondial de viande ovine (tableau 10) provient en effet d'Océanie et va en direction de l'Europe de l'Ouest, du Moyen-Orient, et de l'Asie du Sud-Est. 21 Tableau 9 . Commerce international d'ovins-caprins vivants - 1988 (milliers dei têtes) Principaux importateurs Principaux exportate urs Algérie 600 Ethiopie 200 Côte d'Ivoire 300 Mali 400 Madagascar 1 800 Mauritanie 400 Nigeria 300 Namibie 600 Afrique du Sud 600 Somalie Soudan 700 200 Total 5 pays 6 pays africains africains 3600 ci-dessus 2500 ci-dessus Asie du Sud-Ouest 12000 Syrie 300 Europe de l'Ouest 1 900 Turquie 3200 Autres 3 500 Europe de l'Est Australie 4400 6300 Nouvelle-Zélande Autres 800 3500 Total 21 000 Total 21 000 Tableau 1 0. Commerce international de viande ovine et d'ovins vivants 1988 (milliers de t équiv. carcasse) Principaux importateurs Principaux exportateurs Afrique du Nord 30 Ethiopie-Somalie- Soudan 20 Asie du Sud-Ouest 400 Syrie-Turquie 50 Europe de l'Ouest 230 Europe de l'Est 90 Amérique du Nord 40 Australie-N.Zélande 710 Asie du Sud-Est 80 (Chine incluse) Japon 80 Autres 170 Autres 160 Total 1030 Total 1 030 22 Contrairement à beaucoup d'autres produits, dont le lait, les céréales, et plus récemment les viandes bovines et de volaille, dont les exportations sont subventionnées (Sarniguet, 1989), les marchés africains des produits de l'élevage ovin-caprin ne sont pas perturbés par les importations et devraient pouvoir connaître un certain développement (d'autant que selon divers auteurs, dont Landais (1985), ils exploitent mieux le milieu que les bovins). Pour cela il faut trois conditions simultanées: • que l'augmentation de la production soit possible (intensification); • que la demande spécifique soit suffisante, parmi les non-producteurs, pour offrir des prix attractifs; • que l'écoulement des produits entre des zones éloignées soit possible et efficace (fonctions commerciales: collecte, transport, allotement, tri, ajustement de l'offre et de la demande dans le temps, l'espace et en qualité). Bibliographie Boutonnet J.P. 1989. La spéculation ovine en Algérie: un produit-clé de la céréaliculture. Série Notes et documents n 90. INFWESR, Montpellier (France). Djariri B. 1989. Importations alimentaires et production locale en Afrique au Sud du Sahara: le cas du lait et des produits laitiers au Niger. Institut agronomique méditerranéen, Montpellier (France). Mémoire MSc. FAO (Organisation des Nations-Unies pour l'alimentation et l'agriculture). 1 989. Annuaire mondial de la production 1988. FAO, Rome (Italie). Landais E. 1985. Problèmes liés au développement de l'elevage des petits ruminants (ovins et caprins) en Afrique. VI* Conférence de la Commission régionale de l'OIE pour l'Afrique tenue à Harare (Zimbabwe), 22-25 janvier 1985. Office international des épizooties, Paris (France). 33 p. Mounier A. 1989. Place, rôle et fonction de l'elevage dans les pays industrialisés. Voies de développement pour les pays du Sud. In: Pâturages et alimentation des ruminants en zone tropicale humide. Actes de la conférence tenue à Pointe-à-Pitre (Guadeloupe), 2-6 juin 1987. Institut national de la recherche agronomique, Paris (France), p. 481 à 499. Sarniguet J. 1989. Effets de la concurrence des viandes extra-africaines sur les filières nationales des viandes en Afrique de l'Ouest et du Centre. In: Economie des filières en régions chaudes. Actes du Xe séminaire d'economie et de sociologie, CIRAD, 11-15 septembre 1989, Montpellier (France). 23 Economics of small ruminant meat production and consumption in sub-Saharan Africa Senait Seyoum International Livestock Centre for Africa P O Box 5689, Addis Ababa, Ethiopia Abstract Small ruminants are an integral part of traditional crop-livestock production systems in sub-Saharan Africa (SSA). Meat is one of their major products. This report looks at aggregate patterns and trends in consumption and production of and trade in small ruminant meat in the four regions of SSA (East, West, central and southern). The preliminary results highlight two important points. Firstly, there are regional differences in the consumption of small ruminant meat. These are differences which may be associated with variation in flock dynamics (size, structure and growth) and productivity, production constraints and trends in the production and consumption of other types of meat. Secondly, there is a strong consumer preference for sheep and goat meat in most regions of SSA, as shown by the increase in small ruminant slaughter offtake over the past two decades. In East, West and central Africa, relatively large and increasing small ruminant flocks, stable carcass yields andlor sustainedproduction of other meats seem to have offset the potentially negative impact ofhigh small stock offtake on flock growth and productivity. In contrast, in southern Africa growth of sheep and goat flocks is slow and carcass yield is low and slaughter offtake of other meats (e.g. beef) are declining and hence may not sustain the continuing demand for small ruminant meat in this region in coming years. 25 Tendances de la production et de la consommation de la viande de petits ruminants en Afrique subsaharienne Senait Seyoum Résumé Les petits ruminants (ont partie intégrante des systèmes de production agricole de l'Afrique subsaharienne. Leur principale fonction est la production de viande. La présente communication étudie les caractéristiques et les tendances globales de la consommation, de la production et de la commercialisation de la viande des petits ruminants dans les quatre principales régions d'Afrique subsaharienne. On constate des différences marquées entre ces régions en ce qui concerne la consommation. Elles dépendent de paramètres tels que la taille, la composition, la croissance et la productivité des troupeaux, et également des tendances de la production et de la consommation des autres types de viande. A en juger par l'augmentation continue du taux d'exploitation des petits ruminants en Afrique subsaharienne au cours des vingt dernières années, la viande de mouton et de chèvre est plutôt bien appréciée dans la majeure partie de cette région. Toutefois, compte tenu de la faible croissance des rendements à l'abattage des effectifs des petits ruminants ainsi que de la baisse des taux d'exploitation des autres espèces animales (ex: bovins), cette tendance ne pourra vraisemblablement pas se maintenir au cours des prochaines années. En Afrique de l'Est et en Afrique occidentale et centrale, les effectifs relativement plus nombreux et croissants ajoutés à des rendements à l'abattage plutôt stables et/ou à une production relativement soutenue de viande d'autres espèces, semblent compenser l'effet éventuel des taux d'exploitation élevés sur la croissance et la productivité des troupeaux de petits ruminants. En revanche, en Afrique australe, l'augmentation du taux d'exploitation du troupeau ovin a été très soutenue depuis 1974. Son effectif déjà limité pourra difficilement compenser la baisse du taux d'exploitation du troupeau bovin. Introduction In many areas of sub-Saharan Africa (SSA), livestock raising is an important economic activity from which food (meat, milk) and non-food commodities 26 (manure, traction, hides and skins, wool etc.) and cash income are derived. Meat is one of the most important livestock products. In 1975, meat accounted for about 47% of the gross value of total SSA livestock output (Addis Anteneh et al, 1988). In spite of this, per caput consumption of all kinds of meats (including beef, small ruminant meat, poultry and pork) averaged 10 kg in 1986/88 in SSA (FAO, 1988; 1989), compared with about 81 .8 kg in Europe in 1987 (MLC, 1990). The low level of meat consumption in SSA is primarily due to the low level of meat production per caput which in turn is a consequence of the low productivity of the livestock sub-sector in the region. Between 1974/76 and 1986/88 per caput meat production and per caput consumption increased at annual rates of -0.7% and 0.1%, respectively (FAO, 1988; 1989). Furthermore, given recent developments in the region e.g. rapid population growth and urbanisation, rising input costs, foreign exchange shortages and slow income growth, future prospects seem to be for stagnant or even declining meat consumption, with the likelihood of severe food shortages and growing calorie-protein deficiencies for the bulk of the population (Reutlinger, 1980). Such prospects underscore the urgent need for increasing food production in SSA. New technologies must be developed that enhance production and appropriate livestock development policies must be designed. A prerequisite to this is an understanding of production consumption of and trade in animal products. This paper is based on a preliminary study of the nature and scale of small ruminant meat production, consumption and trade in different regions of SSA. The study focused on small ruminant meat for three reasons: First, present knowledge of small ruminant meat demand patterns in SSA and of factors underlying them, is inadequate. As a result, it has often been difficult to identify potential market opportunities for small ruminant meat and to devise efficient ways of exploiting such opportunities. Second, consumption of beef which has traditionally dominated meat consumption in SSA, is declining relative to consumption of other meats, including mutton and lamb (FAO, 1988). Third although small ruminants currently play a minor role in supplying SSA with meat, they hold promise for increasing meat production and smallholder incomes; compared with cattle, small ruminants require fewer resources, have shorter production cycles, faster rates of growth and greater environmental adaptability. The paper is organised in four sections: Section 1 deals with the small ruminant meat production structure and looks at livestock population, carcass weight, herd/flock offtake and aggregate meat production trends; Section 2 provides information on small ruminant trade flows; in Section 3, aggregate small ruminant meat consumption are determined for the four regions of SSA; in the final section (4), the prospects for increasing small ruminant production and consumption in SSA and their policy implications are discussed. 27 The Small Ruminant Meat Production Structure Small ruminant populations, distribution and growth Populations and distribution Small ruminants are found throughout SSA. Most small ruminants are found in the arid and semi-arid zones of the Sahel and East Africa (ILCA,1980), mainly in pastoral and agropastoral systems. Flock in these areas are generally large (up to 100 head per household) and are often kept with other livestock for both meat and milk production. Small ruminants are comparatively less important in humid and subhumid zones. Nevertheless, it is estimated that in 1981, about a third of SSA's small ruminant population was found in these ecological zones (Onim et al, 1 986). In both zones, smallstock are mostly kept by smallholder who rely mainly on crops but keep small ruminants as a subsidiary source of income and meat and a safeguard against crop failure and low crop prices. Table 1 shows the regional distribution of cattle and small ruminants in SSA in 1988. Small ruminants outnumber cattle by 1.6:1 in SSA as a whole and by 2.7:1 in West Africa. The main concentrations of small ruminants in 1988 were in East and West Africa, which accounted for 88 and 93% of the sub-continent's goats and sheep. It probably reflects environmental (drought, feed availability, disease) and socio-economic (land scarcity, management system, profitability of sheep/goat production, consumer preferences) conditions. Goats are more numerous than sheep in SSA (Table 1) particularly in southern and central Africa. The reasons for these distribution patterns are not very clear but seem to be related to differences in physiological/adaptation characteristics and socio-economic roles of sheep and goats. In much of Africa, sheep production is a relatively market-oriented activity aimed mainly at satisfying urban and ceremonial demand for meat (CRED, 1980; Nestel, 1986). In contrast, goat production tends to be subsistence oriented and to cater to the needs of a larger number of rural and lower income consumers (Zimbabwe Government, 1987). Growth rates Table 2 shows the annual growth rates of cattle and small ruminant populations in the four regions of SSA between 1 961 /65 and 1 986/88. It appears that over the past two decades, the overall population of cattle, sheep and goats has grown slowly in SSA, at annual rates averaging 1 .5, 1 .8 and 1 .5%, respectively. However, small ruminant populations in SSA increased faster than cattle populations, between 1 974/76 and 1 986/88 when they grew at an average annual rate of about 28 CO 00 o>«— OS o 5 c 5 ra ■c CO co -Q 3 CO .c co CD eg 6 o c\i d in b to 6 o C\i o c\i «- CO d d «- CO d d *r co d d eg o- c\i to 00 03 If) o c\j O) CO. 8 § CO 1 CM < 15 O o d 8 o r- CM 8 CO 8 co 8 re LU 5.y < Ot CO CO O) CD CO So8 < CO C3) » = TJ CO o -a — * «■ Q. II — CD < 5 CO CD • O) » re IS I » u re is C CO re c SS> ■c S> re » co rr A oS to CO 0 £ c < O c re a> o £ 1 8 CO o o to re o c» o o Vi CL cu g re o re o c o ■o cre co co are <— c o CO n co < CO §1 £ ■8 w -o co £ re re o CO .i= a>< co c — « 5.2 S 2 £ gO) o o flS 8 o re » o> 9 - CD CD £ Brr < re o | c£ I ? °>•-■£<» E co 2 — 3 re E ii - CO CO " co co 2 o CO O< 29 2%, compared to 1 .5% for cattle stocks. The consistently higher growth of small ruminants relative to cattle in West Africa suggests that a major shift, from cattle to smallstock, has occurred in this region, where considerable constraints inhibit expansion in cattle numbers. Carcass weights, offtake and meat production Carcass weights The meat production of different livestock species is reflected in their carcass weights. Information on carcass yields is therefore useful in comparing and determining the actual and potential values of different meat-producing animals. Average small ruminant carcass weights, estimated from aggregate FAO meat production and slaughter data, are highest in southern Africa and lowest in central Africa. In 1986/88, goat carcasses weighed an average 12.2, 11.7, 10.8 and 10.3 kg, respectively, in southern, West, East and central Africa. The equivalent regional figures for sheep carcasses were 1 1 .6, 11 .8, 12.9 and 1 1 .8 kg, respectively. Figure 1 shows the evolution of cattle, sheep and goat carcass weight indices in different regions of SSA between 1 961 and 1 988. Indices for these figures were calculated by taking 1961/65 annual carcass weight averages, for each species and region, as a base. Thus, on an equivalent basis, small ruminant carcass weights have generally been maintained at higher levels than beef carcass weights. The marked fall in carcass weight indices in the early 1970's particularly evident for West and East Africa reflects a period of draught. On the whole, the trends described above suggest that small ruminants have a great potential for meat production than do cattle especially in traditional production systems. Offtake rates Although definitions vary, "offtake" is generally taken to represent the proportion of animals annually leaving the total herd/flock due to deaths, slaughters, sales or other transactions e.g. exchanges, gifts, loans. In SSA the measurement of smallstock offtake is subject to a high degree of inaccuracy, partly because of lack of reliable data on small ruminant population, growth and production parameters but also because this largely rural and unofficial nature of smallstock slaughter and marketing makes data collection difficult. The estimates of offtake and conclusions presented in this section should therefore be interpreted with caution. While not exactly quantifying offtake, estimates of offtake, obtained rather arbitrarily by dividing FAO data on slaughters by total livestock numbers give a rough idea of offtake patterns and trends in SSA. These estimates indicate that in 1986/88, sheep offtake rates were highest in central and East Africa (32.1% 30 Table0.Annualgr-thr t-(%)orcattle,sheepa do tp %l0bionsiS-between1744ns-1776N0a d1776N0-00 6S . CattleSheepGo sSmallru in nts 000/007 700/007 /07 70 /007 /70 / WestArrica-0.000.0730N0. o0 7N.0.0 CentralArrica0.3000.07'0.0070.nsN0 EastArrica0.0070.7000. 700.. 30 Southern00QQ -JQ7_30. 070 7.N0 0 Arrica SSA0.■.000.0 000 ,0.0 Regionalgroupi gsaresderinediTabl7gr wthr tesrp rc ntannualg ow hratec lculat dothb siso000/00 070/70and000/00an ualaver ges. Source:Owncomp tationsNSsedn7AO(000)Pr ductionTa es. and 29.6%, respectively) and lowest in West and southern Africa (27.4% and 25.7%). A different picture presents itself for goat offtake rates in 1986/88, these being highest in West and southern Africa (35.6% and 31 .3%, respectively) and lowest in East and central Africa (26.9% and 29.9%, respectively). Figure 1 . Indices of mean beef, sheep and goat carcass weights in different regions of sub-Saharan Africa, 1961-88 (1961165=100). (1961/65 = 100) Carcass weight index (CWI) 104 103 a)Sub-Saharan Africa Beef CWI Sheep CWI Goat CWI 1961/65 68 70 72 74 76 78 80 82 84 86 881961/6568 70 72 74 76 78 80 82 84 86 88 Year Year Source: Drawn on the basis of FAO data (FAO 1989 Production Tape). 32 Figure 2 shows the evolution of small ruminant and cattle offtake indices in each region of SSA. From 1961/65 to 1973/74, smallstock offtake indices in SSAasa whole were lower than cattle offtake indices, but thereafter smallstock offtake indices was higher than cattle offtake indices. This trend was most pronounced in southern Africa, but was also evident in West Africa. Differences in the evolution of sheep and goat offtake rates are also apparent. For much of the 1960s and 1970s, sheep and goat offtake indices for SSA as a whole were similar. Around 1982 sheep offtake indices started to exceed goat offtake indices. The only exception to this generalisation was West Africa where goat offtake, though low, consistently exceeded sheep offtake. While all this implies that smallstock, particularly sheep, are being slaughtered to a greater extent than cattle, the population and carcass weight trends that have accompanied offtake trends in some regions are cause for concern. Table 3 provides regional population, offtake and carcass weight growth rates for cattle, sheep and goats between 1974/76 and 1986/88. In West Africa, cattle offtake growth (1.15% p.a.) almost equalled growth in cattle population (1 .36% p.a.) and exceeded growth in carcass yields (0.37% p.a.). In contrast, small ruminant offtake grow slower (at 0.39% p.a.) than either small ruminant numbers (1.95% p.a.) or carcass yields (0.69% p.a. on average). In view of this, the prospects for meeting future meat demand in this region through increased goat/sheep production may be envisaged. The slow growth of cattle offtake (0.09% p.a.) , in central Africa, was accompanied by an increase in cattle stocks (2.80% p.a), and decline in carcass weights (-0.05% p.a.). This probably reflects this use of cattle for purposes other than meat production and factors constraining productivity increases in this predominantly humid region. Smallstock offtake grew at a relatively faster rate (0.53% p.a. for goats and 0.76% p.a. for sheep) in this region, and does not seem to have been detrimental to flock growth and productivity since small ruminant stocks increased (1.98% p.a.) and carcass weights were relatively stable (0.03% p.a. for sheep and 0.12% p.a for goats). There would thus seem to be room for increasing small stock, particularly sheep, in this region provided that population and carcass weight trends remain favourable. In East Africa, cattle offtake is declining (at -0.20% p.a.) and cattle number and carcass yield are increasing slowly (1 .55% and 0.02% p.a., respectively). In spite of this region having large and growing small ruminant populations (2.03% p.a.) and rather favourable carcass weight trends (0.38% pa), small stock offtake has also grown slowly. The observed decline in goat offtake (-0.59% p.a.) is probably due to low demand for goat meat. The growth in sheep offtake (0.56% p.a.) suggests a greater reliance on sheep than on goats in this region. However, there is a need for more country-specific studies looking at unofficial slaughter, trade and constraints e.g. low productivity and vulnerability of smallstock to changing environmental, management and market conditions, before coming to definite conclusions about small stock offtake in East Africa. 33 Figure 2. Indices of mean cattle, sheep and goat offtake rates in different regions of sub-Saharan Africa, 1961-88 (1961165 = 100). i index (1961/65 = 100) Cattle offtake index Sheep offtake index Goat offtake index e) Southern Africa 1961/65 68 70 72 74 76 78 80 82 84 86 881961/6568 70 72 74 76 78 80 82 84 86 88 Year Year * Offtake index = Slaughters as % of total stocks divided by 1 96 1 65 annual base offtake rate times 1 00. Source: Drawn on the basis of FAO data (FAO Production Tape 1988). In southern Africa, cattle offtake is declining (-0.72% p.a.), as are carcass yields (-0.27%p.a.), and small offtake rates are increasing (at 1.14% p.a. for sheep and 0.75% p.a. for goats). This suggests some substitution of sheep/goat meat for beef. However, the slow growth, or declining of small ruminant stocks (0.55% p.a. for goats and -0.07% for sheep) indicates that current levels of offtake cannot be supported unless flock productivity is increased. 34 1able3.Comparisonlivesto0korr ik 3p 9u tiand0 7-'sweiw t' w hratesep.ab82w n197.19and1119-0.id. r t regionsorSSA. Cattle Sheep Goats .astArrica-0.1155 3outhern Arrica -0.72 0.12 0.02 -0.27 1.2 -0.20 -0.07 0.75 0.55 OrftakeCarcassOr takeCarcasOr takeCa cass rate3tocksweightrate3tocksweightrate3tocksweigh DestArrica1.150.3720.021.70 CentralArrica0.012 3-0 50.702 070. 30.531 D0 1 0.502 30 '- 512.070.27 0.3 33A 0.2 1.-0 0.03 0.53 2.3 0.-0 -0.3 1.33 0.' Orrtakeandcarcassweights.easde n dit x .Growtratescal7-laotheb sio17%70and2/■av r ec. le3sh p andgoatorrtakera es3stockanc r sweights. 3ourcthOwncomputationsbasednFAO(13)Pr duc ion1a es. Overall, the emerging picture is one of substantial regional differences in the evolution of offtake rates in SSA. Over the past decade, cattle offtake increased in West and central Africa while it declined in southern and East Africa. In contrast, small stock offtake increased in almost all regions, the only exception being East Africa where goat offtake declined between 1979 and 1986. Southern Africa had the fastest growing rates of offtake of small ruminants. Meat production Domestic production of small ruminant meat in SSA is hard to quantify because of lack of reliable data. The most complete available data on small ruminant meat production come from FAO which defines meat production as meat from animals slaughtered within national boundaries, irrespective of their origin. This includes the meat equivalent of imported animals and excludes that of exported animals. As such, it does not accurately represent domestic production. However, FAO production statistics adjusted for net trade in meat provide general indications of regional small ruminant meat production patterns and trends in SSA. Table 4 presents 1 986/88 estimates of total and per caput domestic production of sheep, goat and small ruminant meat in the four regions of SSA. East and West Africa accounted for 92% of domestic small ruminant meat production in 1986/88. A similar pattern was observed for per caput production. Goat meat production generally exceeded sheep meat production except in East Africa. Table 4 also gives growth rates of per caput small ruminant meat production between 1961/65 and 1974/76, 1974/76 and 1986/88 and 1961/65 and 1986/88. Over the 1961/65-1974/76 period, per caput production declined at an average annual rate of 1.49% in SSA as a whole, annual growth rates being negative in East and West Africa (-1.89% and -1.71%, respectively) and positive in southern and central Africa (2.63% and 0.37%, respectively). Between 1974/76 and central Africa (2.63% and 0.3%, respectively). Between 1974/76 and 1986/88, per caput production declined in all regions of SSA. Considering that human populations grew fastest in East and West Africa between 1974/76 and 1986/88 (at rates averaging 3.08 and 3.19% p.a.), this suggests an increase in domestic small ruminant meat production in these regions. Small Ruminant Trade In most traditional systems of SSA, small stock are kept as a means of generating cash income, as well as a source of meat/milk for direct household consumption. A sizeable proportion of small ruminant trade is also carried out through unofficial marketing channels, at either village or regional levels. Given this production, consumption and marketing traits, reliable data on small ruminant trade flows in SSA are difficult to come by. The small ruminant trade presented in this section are based on aggregate FAO statistics and thus should be viewed with caution. 36 £ Q. 5 CDCD 2 *T CO 01 LO O O CO CO CO CO O 0 JZ CO c .0 T3 82 9 0 CD> 0 ~ a. r T3 O 3 g CO -JJ EC CD CO cog tr 0 a.a .5s CO CM CD CM CO ° 8O) 10 5 * X CD a E c &2 O O 0 «7 9 2 1 % 1* 1 CM O CD CD CO 8 N 0 —-j= « C\i d co d CM C CO 3 "Q c .0 = 0 TJ OC co 0 1 Q- 00 ——. ~ CD 00 3 CO 5 £ <5 co T3 CD CD 00 O co CO Q- 0 f K «~' S: co*"■ OE ? «" d t■ d t- CD CD 3 2 3 0 *t a O) C3) en CO CO 3 ^ to 0 -2 CO o5 to 0 *z CD T3 .0 0 ££ 1 0 0 d d O) g T3 O) 3 CO «£ 0 C O) 3 CD T3 S Q. a. ■ CD ^30 CD CO W -C CD CD CD ■ EE £ 5- SSCj I ♦^ CO CO 2 > 2 WE 1 m r- r>- CM 5 Q. ■"J3 wCO CD O 0 CD Q 8 CO 00 CD O 0 co Southern Africa 00 5.0 10.0 15.0 Imports ('000 t) 20.0 25.0 Source: FAO Trade Tape (FAO, 1988). 40 for 1 988, average annual 1 985/87 net export figures were deducted from 1 986/88 production figures to get estimates of 1986/88 small ruminant meat consumption. Table 5 presents 1 986/88 estimates of total and per caput consumption of sheep, goat and small ruminant meat in different regions of SSA. In total, annual consumption of small ruminant meat in SSA amounted to 906 000 t, most of which was consumed in East (440 000 1) and West Africa (378 000). Goat meat consumption exceeded sheep meat consumption except in East Africa. Much of the same picture is observed for per caput small ruminant meat consumption (Table 5). Growth rates Based on the previous estimates, Table 6 presents annual growth rates of total and per caput small ruminant meat consumption in different regions of SSA between 1961/65 and 1986/88, and during the two split periods of 1961/65-1974/76 and 1974/76-1986/88. Between 1961/65 and 1986/88, total consumption of small ruminant meat in SSA grew at an average annual rate of 1.86%, growing faster in the 1974/76-1986/88 period (2.72% pa.) than in the 1961/65-1974/76 period (1% p.a.). These rates reflect trends in West and East Africa, where total consumption growth was very slow between 1 961/65-1 974/76 (0.86% and 0.68% p.a., respectively), but picked up between 1974/76 and 1986/88 (3 18% and 2.45% p.a., respectively). In sharp contrast to East and West Africa, total consumption growth in southern Africa was very fast between 1961/65 and 1974/76 (4.41% p.a.), but slowed down substantially between 1974/76 and 1986/88 (1.62% p.a.). In central Africa, growth rates did not differ much between the two periods and total small ruminant consumption grew at an average rate of 2.69% p.a. 1961/65-1986/88. Though mostly negative, regional growth rates in per caput consumption of ruminant meat show similar trends The place of small ruminants in overall meat consumption Given the previous picture of small ruminant meat consumption in SSA, there is a need to determine its importance in overall meat consumption. Figure 5, which is based on 1961/65, 1974/76 and 1986/88 average annual estimates of per caput consumption of the four major types of meat consumed in SSA i.e. beef, small ruminant, poultry and pig meat, gives an idea of the composition and evolution of meat consumption in SSA. Beef is the dominant meat consumed in SSA, its share in meat consumption in 1 986/88 ranging from as high as 57% in southern Africa to as low as 38% in West Africa. However, beef consumption seems to be declining in most regions of SSA, particularly in West Africa where its share in overall meat consumption fell from about 50% in 1961/65 to 38% in 1986/88. 41 co « 2 5 8 !o 2 a OE "3, 0 CTJ © a; 5E CO a E co C O CO 5 CO a a c re in XI co 2 C\l co CO CNJ 8 r- O «- 6 co CD CM M o> CO 0 CD 1- 0 CO CD 0 O) 1^ ^r «* CO CO «r 8 0) • § 8 ? CD O < < 0 c < D < CD to c to 13 < I CD a O coO LU CO CO S co 2>E ,± o a> en c » c- -°-.£ co a T3 — CD Os c £ P. O) « 2 w jo a> co ^ w g> co « o a> o> -a ■o 2 c •* « 8 a o 2 o) o CO a CD a>" CD CD T3 > CO co t> E § » co J S 5 o 1 1 Ico 2 o CD E Q. .— *^ X co co a) T3 D ^ C ^ CD 2 co c « 2 o c g 1 » a> o .o o> O T3 * c CD (0 Q. ._ E O u >. CO a> C O « co «* E «j O CD - E 2 « t CD CO c -rj co 2 CL £ CD .^ CD ■S 5 T> co .2-5 IS 5 8. . E x c 5 ».o co a. co » » COc ■ at ~ -Q »- « * .D 0 — EO) > co ■o 2 S c ao co — °> O C-c2 CO LL CD i g " ID « icu .h 5 E *- CO O CD co — co „. cd a i: co co < CO art ~ □> o 2 5° ^ Q. O E .o « o CO CO . 2 " S 42 00 00 tJ cTO c a i CD -Q 3 a9o 10 a■o c CO O c o Q. E 3 V> c o o ra a) E tr CO a. 03 o S D- %? Is (O e g E W o o tr to o> 2 8 o 9 CO 5 CO d eg O) CO CO 9 9 CO r»- CO o d C\I to eg - 9 cr> 8 o eg oo CO s m eg 8 •a c a 2 o ? I CO C C CO S T3 c CO S3 c o ■a 1 o co eg eg CO CO s eg d CO o co o< 2 5 c " a) O Hi % o O). "5 5 8 < s CO c o ro Q. E o o c a E 03 ' ■D c (0 IT) < si a o CO CO CO E ,(0 to Jjj S 8 S- => !* 5 43 Figure 5. Structure of meat consumption in SSA by region, 1961/65, 1974/76 and 1985/87. a) West Africa Small ruminant meat B:,; j Beef I Other meats" b) Central Africa 14.3% 11.5% 10.4% c) East Africa d) Southern Africa 1961-66 5.8% 1974-76 7.3% 1986-88 6.3% • Other meats include pig and poultry meat. Source: Adapted from FAO data (FAO 1988 and 1889). 44 In 1961/65, small ruminant meat accounted for between 6% and 31% of total meat consumption in the four regions. In recent years, poultry and pig meat have come to account for a significant and increasing portion of meat consumption, especially in West and southern Africa. The reasons behind these consumption patterns and trends are not very clear. Conclusion This report has looked at aggregate patterns and trends in production, and trade in and consumption of small ruminant meat in SSA between 1961 and 1988. The evidence presented shows that small ruminant meat production and consumption are greatest in East and West Africa, and are of only minor importance in central and southern Africa. Observed regional trends indicate that domestic production of small ruminant meat has failed to keep pace with demand, especially in southern Africa. Over the past decade, southern Africa's meat production performance has been particularly poor, calculated growth rates for beef, pig and small ruminant meat production in this region between 1974/76 and 1986/88 averaging -0.81, 1.96 and 0.65% pa., respectively. Only poultry meat production growth (4.14% pa.) exceeded the region's human population growth rate of 2.89% pa. Thus per caput production and consumption of small ruminant meat appear to be declining in most regions of SSA. This situation is largely the result of production, environmental and socio-economic constraints bearing upon a traditional production system which has been accorded a low priority in livestock and agricultural development plans in SSA. Almost all small ruminant meat comes from traditional production systems. In some areas of SSA, regular offtake of sheep/goats appears to have been excessive in relation to such parameters as the size, structure, growth and productivity of flocks. In other areas, however, there appears to be room for increasing small stock offtake. All this emphasises the need for viewing the means by which sustained small ruminant production and consumption in SSA are to be achieved in the wider context of the varying and changing environmental and socio-economic conditions prevailing in the sub-continent. References Addis Anteneh, Sandford S and Berhanu Anteneh. 1988. Policy, finance and technology in livestock development in sub-Saharan Africa: Some critical issues. ILCA Bulletin 31:2-13. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. 45 CRED (Center for Research on Economic Development). 1980. Livestock and meat marketing in West Africa. Volume III. CRED, The University of Michigan, Ann Arbor, Michigan, USA. FAO (Food and Agriculture Organization of the United Nations) 1988. FAO agricultural trade yearbook standard tape. 1988. FAO, Rome, Italy. FAO (Food and Agriculture Organization of the United Nations). 1989. FAO agricultural production yearbook standard tape 1989. FAO, Rome, Italy. ILCA (International Livestock Centre for Africa) 1980. Economic trends: Small ruminants. ILCA Bulletin 7. ILCA, Addis Ababa, Ethiopia. MLC (Meat and Livestock Commission). 1990. Meatdemand trends 90/1. MLC, Economic Information services, Milton Keynes, UK. Nestel P. 1986. A society in transition: Developmental and seasonal influences on the nutrition of Maasai women and children. UN Food and Nutrition Bulletin 8(1), March 1986. Onim J F, Mathuva M, Otieno K and Fitzhugh H A. 1986. Forage resources for small ruminants in humid zones of Africa. In: Adeniji K O and Kategile J A (eds), Proceedings of the Workshop on the Improvement of Small Ruminants in Eastern and Southern Africa, Nairobi, Kenya, 18-22 August 1986. OAU/IDRC (Organization of African Unity/International Development Research Centre), Nairobi, Kenya, pp. 145-155. Reutlinger S. 1980. The prevalence of calorie deficient diets in developing countries. World Bank Staff Working Paper 374. Washington, DC, USA. Schluter M. 1984. Constraints on Kenya's food and beverage exports. IFPRI Research Report 44 (April 1984). IFPRI (International Food Policy Research Institute), Washington, DC, USA. Zimbabwe Government Agricultural Marketing Authority. 1987. Economic review of the agricultural industry in Zimbabwe 1986. Agricultural Marketing Authority, Harare, Zimbabwe. 46 Short-run demand for goat meat in Cameroon J.P. Ayissi Mbala Inst'rtut National de Developpement Rural (INADER) University Centre of Dschang, B.P. 2221, Dschang, Cameroon Abstract An analysis of the quantity-price and income relationship, and consumer reactions at the retail level of goat meat is presented. A demand model was developed using either the linear, semi-logarithmic or double-logarithmic functional forms. Demande a court terme de viande caprine au Cameroun J.P. Ayissi Mbala Resume Ceffe communication presente une analyse des relations entre la consommation de viande de chèvres, son prix et le revenu au Cameroun, ainsi que les reactions des consommateurs au niveau du commerce de detail. Un modele de demande est développ6 en utilisant des fonctions linSaires, semi-logarithmique et logarithmique double. Introduction In Cameroon, the consumption of meat and other livestock products is particularly affected by the geographical setting of the country. The livestock producing areas are located far away from the main consumption areas of the South. This situation, coupled with the poor system of husbandry and the inadequate marketing infrastructure, has resulted in an insufficient supply of meat from all types of livestock. 47 During the early part of the past decade, Cameroon experienced a rapid increase in population (3.0%) as well as a steady increase in the Gross Domestic Products Recent estimates indicate that the domestic meat supply has not been able to meet the internal demand. Table 1 reveals a shortage of 55.7 thousand tons of meat from all livestock species. Another factor which affects the consumption of meat in the country is the uneven distribution of the available supply. This situation is indicated in Table 2. Regions situated further away from the producing areas receive very little meat from any source. Consequently, it is evident that as a result of its geographical setting and other related problems, the country's domestic demand for meat and livestock products is unlikely to be met, if emphasis is placed as it is at present, on cattle production alone. Table 1 . Supply and demand for meat by province (1 986/87). Total meat Total meat Population demand production Balance Provinces 1986 ('000) (tons) (tons) (tons) Extreme north 1 727.5 26 586.2 22 396.5 -4 189 2 North 607.5 9 349.4 7 876.5 - 1 472.6 Adamaoua 422.5 6 502.3 26 270.0 19 767.7 East 475.9 7 324.0 7 144.0 - 180.1 North-west 1 221.5 18 798.9 16 967.4 - 1 831.5 West 1 330.3 20 473.3 9111.3 -11 362.1 Subtotal 5 785.2 89 034.2 89 766.0 731.7 South-west 824.7 12 692.1 1 517.2 -11 175.0 South 406.6 6 257.6 1 344.6 -4 913.0 Littoral 1 677.6 25 818.3 4 176.2 -21 642.0 Central 1 742.3 26 967.9 8 248.1 -18 719.8 Subtotal 4 661.2 71 735.9 15 286.0 -56 449.9 Total 10 446.4 160 770.1 105 052.0 -55 718.1 In Cameroon, sheep and goat production ranks first among ruminants in terms of numbers, although in terms of total meat output they are second to cattle. According to recent estimates there is a total of 437 thousand goat and 240 thousand sheep farms with populations of 3.5 million and 1 .5 million, respectively 48 (MOA, 1984). The figures indicate that goat and sheep production is a major economic activity in terms of the number of people employed and as a source of revenue, especially in the main producing areas of the country. Consequently, it is imperative that all the available types of livestock be exploited. It is also important that the potential demand for the various types of meat should be estimated so as to evaluate their marketing potential. This is necessary for effective planning of future livestock development programmes. Table 2. Total me A production and animal species by region (1 986/1 987). Goat Beef Mutton meat Pork Poultry Total Provinces (tons) (tons) (tons) (tons) (tons) (tons) Extreme north 13408 4 378 2 831 346 1333 22 396 North 5 492 727 780 344 534 7 877 Adamaoua 25 360 375 303 38 194 26 270 East 3 795 465 608 1798 478 7 144 North-west 11 640 718 328 2 140 2 142 16 967 West 3991 628 1 729 1 852 912 9111 Subtotal 63 686 7 291 6 678 6 518 5 593 89 766 South-west 93 73 153 923 275 1 517 South 6 251 424 395 269 1345 Littoral 55 12 9 1 983 2 118 4 176 Central 514 192 432 1 365 5 745 8 248 Subtotal 668 528 1 017 4666 8 407 15 286 Total 64 354 7819 7 695 11 184 14000 105 052 Per cent 61.3 7.4 7.3 10.6 14.3 100 If the production of goat meat has to be increased as a measure for alleviating the shortage in the present meat supply, it is important to determine its demand in relation to other types of meat in the country. This is necessary because consumer reaction to goat meat should be taken into consideration by government and other interested bodies in planning future investments in livestock development. This paper attempts to evaluate consumers, response with respect to price changes in goat meat and other types of meat as well as the effects of increasing income on consumption. Due to lack of accurate data relating to the consumption of all types of meat, the analysis is limited to only two principal types of meat (goat meat and beef) for which data were available. 49 Materials and Methods The analysis is based on official time series data covering a period of twelve years. This period may statistically be regarded as inadequate to guarantee a satisfactory estimate of the demand for the product. It may be argued that such an exercise which has the initial constraint of limited data is subject to a wide margin of error. On the other hand, it is also a fact that observations covering a long period of time pose some statistical and economic problems of analysis. In this respect, it is important to note that factors affecting demand do change over time and therefore estimates based on long periods of observations may not reflect these changes. However, it is hoped that subject to the methodology and the analytical techniques used, the results obtained here will be indicative of the true situation and will serve as a pioneering attempt. The principal factors affecting the demand for meat in general include: a) the amount and distribution of per capita income in the population b) the availability and prices of substitutes c) the retail price of the commodity. There are, however, other factors such as tastes, consumer preferences for different cuts of meat, religious and traditional affiliations of the people and other not easily quantifiable factors. In this paper, the demand for goat meat and the effects of income, price and substitutes on consumption are the major consideration in the analysis. The per capita consumption of goat meat is taken as the dependent variable. The total consumption figures used are the annual time-series data of the official slaughter of goats. These are published in the annual reports of the Ministry of Livestock and Animal Industries. It is perhaps important to mention that these slaughter figures may not represent the actual consumption of goat meat. This is because a substantial number of goats is slaughtered privately for home consumption especially during festivities such as funeral ceremonies, marriages etc. However, since these private slaughters do not enter the marketing system, they are not directly guided by the normal market forces of supply that only the official marketed stock may be affected by the main factors that cause variations in demand. In order to obtain the per capita consumption, the recorded number of goats slaughtered is converted into kg equivalent for fresh goat meat by using a mean carcass weight, which in the case of goats in Cameroon has officially been estimated at 12 kg (1966-1979). The yearly totals of fresh goat meat are then divided by the estimated population to obtain the per capita consumption. The independent variables constitute the per capita income and prices. Income is a very important factor in the study of the demand for food and other agricultural 50 products but the problems it presents in most demand studies result from the concept and actual definition of income. However, for the purpose of this paper, income is defined as the "real income". This means that the actual money incomes as represented by the values of the GDP have to be deflated by an inflation factor. This inflation factor is taken to be the consumer price index. In other words, the formula for estimating 'real income' is as follows: Per capita income Real income = x 100 Consumer price index The prices used in this analysis are the average annual retail prices of the various types of fresh meat. These are also published by the Office of Statistics in the Ministry of Livestock and Animal Industries. These prices were also deflated by the consumer price index in order to make adjustments for changes in the money values. The prices obtained can therefore be termed 'real price'. According to the theory of consumption, an equal percentage change in the prices of all commodities as well as in the money incomes of consumers should not disturb the pattern of quantities demanded or consumed (Fox, 1968). The demand model developed for analysis is based on the fact that the quantity of purchased goat meat consumed per head of the population can be said to be a function of the price of goat meat, the real per capita income of the population, the price of beef and a disturbance term (U) . Beef has been chosen as a substitute since it is, from a priori knowledge, the most popular though expensive type of meat in the country. The production and marketing of mutton is comparatively marginal relative to goat meat and the pork consumption is also heavily influenced by religious inclinations. These two types of meat have not been considered in the analysis. The demand model can, therefore, be represented in the following function form: Where: Cgmt = f(Pgmt, Pbt, Yg, U) (1) Cgmt = per capita consumption of goat meat in time t Pgmt = price of goat meat in time t (in francs CFA) Yt = income per capita in time t Pbt = price of beef in time t U = disturbance term. The term is assumed to be normally and independently distributed with a mean zero and unit variance. In estimating the demand for meat and other food products, several statistical methods and techniques have been developed to determine the relationship between consumption, income and prices. These methods and techniques have 51 their advantages and limitations. For this paper, the linear function, the semi-logarithmic function, and the double-logarithmic function were tried. The functions were employed in a least square multiple regression analysis so as to find out which of the function best fits the available data. The three functional forms were then transformed into the following estimating equations: a) Linear equations:- Cgm = a + bi Pgmt + b2 pbt + b3 Yt (2) b) Semi-logarithmic equation:- Cgm = a + bi Log Pgmt + Log Pbt + b3 LogYt (3) c) Double-logarithmic equation:- LogCgm = a + b, Log Pgmt + t>2 Log Pbt + b3 Log Yt (4) Where a = the constant term (point of intercept) bi .... b3 = regression coefficients of the explanatory variables. It is believed that statistical data can only trace a theoretical demand curve if that curve (demand) remains fixed while the supply curve shifts (Oni, 1972). In the present case, it cannot be claimed that the demand schedule for goat meat in the country has remained fixed, but the slaughter data exhibits a less violent fluctuation when compared with the limited market offtake by the producers. Consequently, it can be assumed that the statistically derived demand function will trace out the true demand for goat meat in the country. This identification problem and the possible remedies have been considered by Johnston (1963) and Fox (1968). Furthermore, from a priori knowledge of the situation, the price of goat meat at the retail does not bear any relationship to producer prices. This is because the producer prices are somewhat officially predetermined and there is also the limitation of the willingness of producers to sell a set of drawings at each point of time. In other words, the impossibility of repeated drawing as envisaged in the notions of probability and statistical inference would not necessarily constitute any serious obstacle to make generalisations on the results obtained from the data used. Results and Discussion Of the three estimating equations tested, the linear and double-log functions were selected for further analysis. This is because they largely satisfied the a priori expectations of the demand for goat meat. The results obtained from the two equations are presented in Tables 3 and 4, respectively. The two equations were 52 further subjected to a series of combinations. Since one of the major objectives of the analysis was to determine the effects of price and income on the consumption of goat meat, it is important to know their separate effects upon the dependent variables. This would be difficult if the explanatory variables are highly correlated. Table 3. Regression results explaining the demand for goat meat using the linear function. Dependent Constant Real income Price of Price of Price Drot variable term per capita goat meat beef ratio statistics (y) (Pgmt) (pbt) P2 i) Cgmt1 -3.820 SE 0.189+ + 0.88 1.31 t. ratio ii) Cgmt -3.690 SE t . ratio 0.260 + + -1.595 (0.672) 2.37 0.92 2.06 (0.035) 7.43 iii) Cgmt -3.832 0.256 + + -1.618 0.114 - 0.93 2.05 SE (0.041) (0.720) (0.053) - t. ratio 6.24 2.25 0.21 iv) Cgmt -6.491 SE t. ratio 0.197+ + 1.508 + (0.666) 2.26 0.92 2.07 (0.019) 10.36 + + = significant at the 1 per cent level. + = significant at the 5 per cent level. 1. Cgmt = per capita consumption of goat meat in time t. Nonetheless, in order to carry out the exercise of identifying the individual effects of the independent variables, the stepwise multiple regression technique has been employed in the two estimating equations. Furthermore, the price ratio of beef and goat meat was introduced into the equations as a separate explanatory variable. This technique makes it possible to identify the individual as well as the cumulative or additional effects of the regressors on the dependent variable. The process of eliminating some of the variables from the equations also served as a means of reducing the multi-collinearity between them. In view of the fact that time-series data are susceptible to auto-correlated errors, the residual were also calculated and the various equations were tested for their presence by the Durbin-Watson statistics (Dwt). 53 Table 4. Regression results explaining the demand lor goat meat-using the double log function. Log per Log Log price Dependent Constant capita price of Log ratio beef/ Drot variable term income goat meat price of goat meat statistics (Yt) (pgmt) beef (pbt) Pb/pgmt R2 0.83 1.11 0.85 1.28 0.90 2.32 i)Log Cgmt1 -2.605 1.973+ + SE t. ratio (0.285) 6.92 ii) Log Cgmt -3.393 2.566 + + -0.602 - - SE (0.565) (0.499) t. ratio 4.54 1.21 iii) Log Cgmt -2.929 2.058 +' -0.901 + 0.947 + - SE (0.533) (0.444) (0.441) - t. ratio 3.86 2.03 2.15 iv) Log Cgmt -2.954 2.084 + + - - 0.898 ' SE (0.239) (0.264) t. ratio 8.72 0.89 2.12 + + = significant at 1 per cent level. + = significant at 5 per cent level. 1. Cgmt = per capita consumption of goat meat in time t. From the estimating functions of the equations presented in Tables 3 and 4, the equation (iii) in Table 4 from the double-log function was selected as the lead equation. This is because it contains all the important explanatory variables and fulfills most of the a priori expectations on the demand for goat meat. The coefficient of the price variables and that of income have the correct negative and positive signs, respectively, and they are significantly different from zero and statistically significant at 1 per cent. Their residuals are not auto-correlated as tested by the Durbin Watson statistics (Dwt) . The joint effects of all the explanatory variables of the equation is to explain 90 per cent of the variation in the dependent variable and coefficient of multiple determination (Ft2) is highly significant (0.90). Furthermore, the standard errors of estimates are relatively low. In order to properly analyse and draw reasonable conclusions on the consumption of goat meat in terms of quantity-price and income relationships, it 54 is necessary to estimate also the price and income elasticities of demand. The different elasticities from the estimating functions were consequently calculated as follows :- From the linear function: dcgm Pgm Direct price elasticity = X dpgm cgm dcgm Pb Cross price elasticity = X dpb cgm dcgm y Income elasticity (Ey) = X dy cgm The various elasticities were calculated at the mean values of the variables. In the double-logarithmic function, the regression estimates of the coefficients of the appropriate variables are taken as the direct elasticities. The various elasticity estimates from the two functions are presented in Table 5. In the literature, very little attention has been given to goats in economic research. Consequently, there are few estimates of income elasticity of demand for goat meat. In many cases, even in countries where goat is popular, the income elasticity is calculated jointly with mutton and lamb. Table 6 shows various estimates of the income elasticities of demand for the different types of meat in some African countries. The income elasticity estimates for mutton and lamb (which often includes goat meat) in some of these countries are similar to those obtained from our equations and it might be said that the consistency of the results presented in Table 5 leads to some credibility to the estimates. The direct price elasticity of demand for goat meat (EPgm) has the correct negative sign in each of the estimating equations. With reference to the lead equation (iii) of the double-log-function, the direct price elasticity is -0.90. This can be taken as almost unit elasticity in terms of price. It indicates that goat consumption will increase by 0.9 per cent if the price falls by 1 per cent. The cross price elasticity of goat meat and beef has the correct positive sign and is also close to unity (0.95). This implies that the demand for goat meat has approximately unit elasticity in relation to the price of beef. This agrees with a priori expectations and knowledge of the actual marketing situation of goat meat in the country. As a result of the limited supply of beef in most parts of the country and the consequent increase in beef prices, the demand for goat meat has increased rapidly in recent years. Furthermore, the positive sign of the cross price elasticity with (Epb) confirms the fact that beef and goat meat can be taken as substitute products. 55 Table 5. Estimates of price and income elasticities of demand for goat meat. Functional form Direct price elasticity (E pgm) Cross price elasticity with beef (E Pb/pgmt) Income elasticity (Ey) Linear function Double-log function -0.85 -0.90 0.84 0.95 2.3 2.05 Table 6. Some estimates of income elasticity of demand for different types of meat in some African countries. Country All meat Beef Lamb Pork Poultry Sudan 0.90 0.60 1.50 - 1.00 Gabon 0.82 1.20 1.00 1.00 1.00 Chad 0.85 0.90 0.80 1.00 1.00 Nigeria 1.13 1.30 1.20 1.20 1.00 Ghana 0.98 1.30 1.50 1.00 1.00 Sierra Leone 1.20 1.50 1.50 1.20 1.00 Source: FAO (1974). The income elasticity of demand for goat meat is also positive and far greater than unity (2.1). It signifies that with any 1-per cent increase in income, the consumption of goat meat is likely to increase by 2.1 per cent. On the whole, the demand estimates indicate that goat meat is almost price-and highly income-elastic. Policy Implications Although the demand estimates reveal a high income elasticity, it is perhaps important to remark that the estimated increase in the consumption of goat meat will occur only if the distribution of income continue in the present proportion. On the other hand, if the increased income goes to the wealthier section of the population, the increased consumption of goat meat will involve only the privileged few who can afford it rather than the majority of the population. Furthermore, with almost a unit cross-price elasticity with beef, increased consumption of goat meat will depend on the price of beef which generally is regarded as a superior type of meat. The implication of this situation is that if the production of goat meat is to be increased as a measure of increasing the overall supply of meat in the country, it is important to adopt a cheaper and efficient method of production and marketing than it is at present. This is necessary 56 because any increase in the consumption of goat meat will depend greatly on its price relation with beef. Goat meat has to be produced and marketed at competitive prices with beef and perhaps that of meat. It is also important to note that government policy of fixing producer prices may not necessarily serve as incentive for increasing production, especially if the fixed prices do not bear any relationship with prices at the retail level. The importance of goats as a source of meat supply in the country has been revealed and from the demand analysis, there is the need to determine the cost effectiveness in goat production as a necessary condition for increased output. Finally, it is important to remark that any effort undertaken by the State or individuals at stepping up inter-regional production and improving the efficiency and distribution of goat meat will be consistent with the economic growth as well as the nutritional requirements of the country as a whole. References FAO (Food and Agricultural Organization of the United Nations) 1974. Agricultural commodity projections 1975-1985. FAO, Rome. Fox KA. 1968. Intermediate economic statistics. John Willy and Sons Inc., New York, USA. Johnston J. 1963. Econometric models. McGraw-Hill Book Co. Inc., New York, USA. Ministry of Livestock and Animal Industries, 1966-1979. Annual reports. Yaounde, Cameroon. MOA (Ministry of Agriculture). 1984. Agricultural census Yaounde. Yaounde, Cameroon. Oni S A. 1972. A short-run demand for beef in western Nigeria. The Nigerian Journal of Economics and Social Studies. University of Ibadan, Ibadan, Nigeria. 57 Intra-annual sheep price patterns and factors underlying price variations in the central highlands of Ethiopia Andargachew Kebeofe1 and Ray Brokken2 1 Senior expert, Planning Department Ministry of State Farms Development P O Box 5765, Addis Ababa, Ethiopia 2 Livestock Economics Division ILCA, P O Box 5689 Addis Ababa, Ethiopia Abstract Nine key Ethiopian central highland markets were surveyed for a period of one year beginning in January and ending in December of 1989 to determine the effects of certain animal and market characteristics on price and the pattern of sheep prices in relation to seasons. Each market was surveyed once a week on the main market day. Price, weight, sex, age, colour, condition score, breed and buyers ' purpose were recorded for all completed transactions. A total of 50 062 cases were recorded. Three markets, each representing redistributive, intermediate and terminal markets, as identified by buyers' purpose were chosen for further analysis. Considerable weekly price variation was evident. Animal characteristics (weight, age, condition, sex, colour) as well as buyers' purpose and seasons were variably important in explaining variations in price among animals within weeks. Variations in the composition of these characteristics from week to week were among factors underlying changes in weekly mean prices. R2 varied from 0.2659 to 0.3583 for price-per-kilogram model and from 0.7822 to 0.8413 for price-per-head model in the three markets. 59 Facteurs affectant les variations intra-annuelles du prix des ovins dans les hauts plateaux éthiopiens Kebede Andargachew et Ray Brokken Résumé Neuf marchés des hauts plateaux éthiopiens ont été suivis en 1989 pour analyser l'effet des caractéristiques des animaux des marchés et des saisons sur le prix des moutons. Le principal jour de marché chaque semaine, le prix, le poids, le sexe, l'âge, la robe, l'état corporel, la race et la raison d'achat de tous les animaux vendus ont été enregistrés (50 062 ventes au total). Trois marchés, chacun représentant un marché de redistribution, intermédiaire et terminal, ont été choisi pour l'analyse. Les variations hebdomadaires de prix sont importantes. Les caractéristiques des moutons (âge, poids, robe, sexe, état corporel), la raison d'achat et les fêtes expliquent de façon variable les variations de prix chaque semaine. D'une semaine à l'autre, le prix hebdomadaire moyen varie avec la composition de ces caractéristiques. Pour les trois marchés, R2 varie de 0,2659 à 0, 3583 pour le modèle du prix au kg et de 0, 7822 à 0,8413 pour celui du prix par animal. Introduction The central highlands of Ethiopia represent an important sheep producing area accounting for 75 per cent of the nation's flock. Central highland farmers generally consider livestock a more reliable form of wealth than other alternatives. An average of three head of sheep are slaughtered per year per family. About 40 per cent of farmers' cash income from animal trade is accounted for by sale of sheep (Gryseels, 1988). Development activities of the Ministry of Agriculture (MOA) have focused mainly on distribution of crossbred rams in limited parts of the country. The Institute of Agricultural Research (IAR) and the International Livestock Centre for Africa (ILCA) undertake research activities with the aim of improving productivity of indigenous sheep through a combination of improved management, feeding, genetic selection and health care. The ultimate goal of such interventions is to improve the welfare of producers through increased production and income. Interventions addressing increased production need to consider the market aspect simultaneously. 60 Farmers benefit from efficient markets which transimit information and ensure minimum marketing margins. Farmers need to be aware of the preferred characteristics of animals as well as price patterns so that they can plan breeding and fattening programmes and breed selection consistent with the best seasonal prices and consumers' preferences. There is very little literature on marketing of small ruminants in Ethiopia. Ayele Gebre-Mariam and Hillmann (1975, unpublished) tried to determine, factors affecting differences in price levels in a given market and between markets at various rounds of (a central highlands livestock) survey. The most vivid shortcoming of this study was that the data were collected during an abnormal (drought) year and as a result its conclusions may not necessarily be representative. The objectives of the present study are to: (i) determine overall market composition by basic animal characteristics, (ii) determine intra-annual price patterns and factors underlying intra-annual price variations. In the central highlands, as in nearly all other parts of the country, there is no regular market information on prices and supplies, nor formalised grades and standards. Agreement on price is reached by a long one-on-one bargaining between a seller and a buyer. Animals are sold on a per-head basis. Under such circumstances, prices paid will reflect buyers' preference for various animal characteristics (sex, weight, age, condition, breed, colour), the purpose as to whether the animals are purchased for consumption, breeding, the season of the year and the bargaining skills of buyers and sellers. Materials and Methods Nine key central highland markets most of them purposively selected to capture the main pattern of movement of sheep and benefit from ILCA's outreach research stations were included in a sheep market survey. The survey was undertaken on the main weekly market day for a period of one year (January to December 1989). Due to technical reasons, the survey in a few markets was started and completed a few weeks later. The survey at Degollo was interrupted much earlier due to security problems. All completed transactions on ordinary market days and as many completed transactions as could be tracked on festival market days were recorded at the exit gate if fenced, or at one or two of the most frequented departure routes rf open market places. This procedure minimised enumerators' decisions about sampling. A festival market was represented by two market days: the market falling on the festival day (or the market closest to the festival day ) and the one preceding it. Price, weight, sex, age, condition, colour, number offered and sold, buyers' purpose, buyer and seller type, and time of sale were recorded. In all, 50 062 animals were recorded (Shola 6453, Addisu Shola 1922, Debre Berhan 61 8609, Deneba 7326, Enewari 3638, Degollo 7808, Ginchi 9263, Debre Zeit 3619. Dejen 1424). Age was approximated by the type and number of teeth. Condition was assessed by the different degrees of fatness on and around the back bone, in the loin area and the last rib, and above the kidney (Carles, 1 983). While Carles discussed six conditions scores, only four ranging from one for poor to four for very fat sheep where used in the present study. This condition scoring technique was designed for thin-tailed sheep and might not be the most appropriate for the type of sheep found in the central highlands. However, uniformity of condition scores was ensured by repeated comparisons of the scores of different enumerators for given flock of sheep. No official scoring technique has been established for Ethiopian fat-tailed sheep, to the authors' knowledge. Due to the difficulty in identifying specific breeds in the market place, sheep were classified into: hairy thin-tailed, wooled thin-tailed, fat-tailed and fat-rumped (Minutes of the National Technical Committee on National Sheep Research and Development Policy, Addis Ababa, 1974 unpublished). To select representative markets for analysis the markets included in the survey were classified on the basis of buyers' purpose (for approaches used in market classification see for example: Herman Tilahun, 1979; Solomon Bekure and Negussie, 1983). All the markets included in the survey fall into terminal or redistributive categories (Table 1). Although classified as a terminal, the Debre Berhan market was almost equally important as a redistributive market. Shola and Degollo, the most typical terminal and redistributive markets, respectively, and Debre Berhan, an intermediate market, were selected for analysis. Table 1 . Classification of markets by major purpose of purchasing. Purpose of purchasing (%) Reproduction & fattening Slaughtering Resale Redistributive markets Degollo 10.40 8.50 81.10 Ginchi 16.20 26.10 57.70 Enewari 12.30 34.70 53.00 Deneba 25.00 25.90 49.10 Terminal markets Addisu Shola 0.20 98.80 1.00 Shola 1.20 98.50 0.30 Dejen 2.00 97.00 1.0 Debre Zeit 9.40 90.10 0.50 Debre Berhan 7.70 46.70 45.60 62 The following equation was estimated in order to determine factors affecting sheep prices for each of the markets selected for analysis. 9 11 16 P = b0 + b,W + bjW2 + b3T +EbiFi +EbiSi +EbAi 4 10 12 19 22 25 28 + EbiCi + EbiK + EbiRi + baefW*!) +Ebi(W*Si) 17 20 23 27 33 36 39 45 + Ebi(W*Aj) +Ebi(W*Ci) + Ebi(W*Ri) +Ebi(W*Fi) 29 34 37 40 47 52 55 + Ebi(T*Si) + Ebi(T*A,) + Ebi(T*Ci) + e, 46 48 53 Where: P = price per kilogram in Ethiopian Birr, W is weight in kilogram,T is time in week; F, S, A, C, K and R are dummy sets signifying season, sex, age condition, colour, and purpose of buying, respectively. The bis are structural parameter of the equation. Results and Discussion Overall market composition by basic animal characteristics Sex distribution of sheep sold varied from market to market (Table 2) . Male sheep were dominant at the terminal and redistributive markets. Female sheep made the largest proportion in the intermediate market. Castrated sheep were second in importance to entire males in the terminal market but of little significance in the other markets. In general, sheep sold in the central highlands were mainly lambs of less than a year age. Relatively higher proportions of young sheep were recorded in the redistributive market. Sheep of average condition were dominant in all markets. The terminal market had the largest proportion of fat and very fat sheep. The largest proportion of poor sheep were recorded at the intermediate market. Colour variability was less pronounced than other animal characteristics. All sheep were fat-tailed type at the intermediate and redistributive markets and this type also predominated in the terminal market. Fat-rumped sheep represented only a very small proportion in the terminal market. 63 Table 2. Percentage market composition of central highland markets by animal characteristics and type of market. Type of market Animal characteristics Terminal market Intermediate market Redistributive market A. Sex: Male 50.5 43.0 56.4 Castrate 38.5 9.9 4.2 Female 11.0 47.1 39.4 B. Age (year) : < 1 62.4 72.8 86.0 2 26.6 6.9 4.8 3 24 6.1 2.1 > 4 8.6 14.2 7.1 C. Condition: Poor 13.6 29.8 5.6 Average 55.3 64.9 90.0 Fat 21.1 4.9 4.1 Very fat 10.0 0.4 0.3 D. Colour: White 28.0 19.0 20.6 Brown 29.0 35.3 34.4 Black and white 11.6 24.0 11.6 Black and other 31.4 21.7 33.4 Intra-annual sheep price pattern Considerable weekly mean price variation was apparent in the three markets (Figure 1). Average prices were highest in the terminal market and lowest in the redistributive market throughout the survey period. Overall annual mean prices per kilogram for the terminal, intermediate and redistributive markets were 2.84, 2.26 and 1 .85 Birr, respectively. The highest weekly mean price per kilogram for the terminal market occurred during Id-al-fater (Ramadan) which fell a week after the Ethiopian Easter. The highest price per kilogram for the intermediate market occurred during the Ethiopian Easter. 64 5 a I I 5 I. I 3 -o c co E S c co c E CO m LU co o s > £ -3 co■ C\J CM 5 -8i . CO . o in CO -1— CD in 65 Both the quantity purchased and offered for sale considerably increase during religious festivals. The lowest demand for sheep is expected to occur during fasting periods. Apart from producers' response to increased demand for sheep during festivals, seasonality of supply is affected by producers need for cash and lambing pattern. Gryseels (1988) noted that small ruminants were the first to be considered for sale when food and seed grain stocks were depleted. The highest demand for grain seed and the lowest stock of food grain occurs during June to end of September. In addition cash needs are high before the harvests begin. Gryseels (1988) also reported that the peak lambing season was from September to November and relatively equally distributed during the rest of the year. Factors underlying intra-annual price variations Table 3 presents probability levels of (F) tests of significance of the variables included in the price/kg eqation for each of the markets under consideration. The results indicate that most factors are significant in two out of the three markets and that all variables are significant in at least one market. The R2s from this model (0.3238, 0.2659, and 0.3583 for the terminal, intermediate and redistributive markets, respectively) were superior to those calculated using antilogs from a log-linear model (0.3047, 0.2596 and 0.3583). A price-per-head equation fitted with variation generated R2s of 0.7822, 0.7973 and 0.8413 for the terminal, intermediate, and redistributive markets, respectively. However, because of the large week-to-week variation in weight, to be seen later, price per kg was used in the analysis. Moreover, price per head was predicted equally well by estimated price per kg multiplied by actual weights. The R*2s calculated from the alternative price per head estimate (0.7852, 0.7971, and 0.8415 for the terminal, intermediate & redistributive markets, respectively) were almost equal to the R2 generated from the actual price-per-head equation. R*2 = 1-(E(Pi-plWi)2/E(Pi-P)2) where: Pi is observed price per head of the ith sheep; pi is the predicted price per kg; Wi is the observed weight of the ith sheep. Annex 1 records estimated parameters of the price-per-kg equation for the terminal, intermediate and redistributive markets. Positive and negative coefficients within a set of dummy variables signify premiums and discounts, respectively, relative to the dummy in the base in Ethiopian birr (EB). The t test was used to check whether the individual premiums or discounts were different from zero. The different patterns of variation in price per head and price per kg caused by week to week variation in average weights is shown by plotting mean price per head and mean price per kg on the same graph (Figure 2) for the terminal market. 66 If the weights did not vary at all from week to week, the two price lines would maintain a constant difference. Sheep sold in the terminal market were on the average heavier than in either of the other two markets (Figure 3). The overall annual mean weight forthe terminal, intermediate and redistributive markets were 26.6, 20.2 and 19.7 kilograms, respectively. Table 3. Probability levels (Pr>F) of tests of significance of factors included in price per kg equation for the terminal, intermediate and redistributive markets. Terminal Intermediate Redistributive market market market Weight 0.0001 0.0041 0.0105 Weight 0.0001 0.0032 0.0001 Week 0.0054 0.0147 0.0001 Season 0.0001 0.0001 0.0617 Sex 0.0012 0.0017 0.5284 Age 0.5601 0.0001 0.0022 Condition 0.5961 0.0001 0.0001 Colour 0.0001 0.0001 0.0001 Buyer's purpose na 0.0064 0.0001 Weight x time 0.9905 0.1483 0.0322 Weight x sex 0.0281 0.0972 0.1488 Weight x age 0.6965 0.0001 0.0577 Weight x condition 0.7998 0.0001 0.0060 Weight x buyer's purp. na 0.0154 0.0016 Weight x season 0.0001 0.0001 0.0001 Week x sex 0.0011 0.0007 0.0001 Week x age 0.2381 0.4499 0.0164 Week x condition 0.0006 0.0383 0.0003 na = not applicable, virtually all for slaughtering. 67 t m IT) m LLJ CO CO co en eg CM CM rn 1 1 i i i i i-* 68 *■> CM in co { .> ^ CO o co IT, cu S c E 2 ( > ro I < > ^ .3 y / CO 10 s . CO 69 The quadratic structure of the effect of weight on price per kg indicates that prices first decline to a minimum and increase as weight increases. The equations for males, castrates and females were calculated and plotted for each of the three markets in order to determine the pattern of price variation in relation to animal weight (figures not included). Most of the males and females in the terminal market were less than one year of age. Their weights ranged both above and below w*1 . The mean weight of the male group was slightly smaller than w* while, for females, mean weight was slightly larger than w*. Thus for males and females premiums in price per kg were offered both as weights decreased below w" and increased above w*. The mean weight of castrates was well above w* indicating that price per kg was strongly rising with increasing weights. This reflects a higher demand for castrates that are both fat and heavy and results in higher per animal price for castrates. The picture is similar in the intermediate market but somewhat different in the redistributive market. In the redistributive market, prices per kg of all three sex categories are strongly rising as weights increase. One can only speculate as to why the relationship between price per kg and weight are so different from the other two markets. The redistributive market is quite remote and animals are purchased mainly for resale where as they are nearly all purchased for slaughtering in the terminal market and for slaughtering and resale in the intermediate market. Animals purchased for resale must be trekked a long distance from the country to terminal markets. Perhaps these conditions explain the emphasis favouring higher weights in the redistributive market. A significantly higher premium offered for castrates in all the three markets was not unexpected. Castrated and fattened sheep locally known as mukit (also applicable to goats) are on the average the heaviest group. Castrates are relatively old, 70 per cent of them between two and four years, and fat to very fat in condition (55 per cent of them fall in this category). Gryseels (1988) noted that farmers castrate sheep for fattening in order to fetch prices. Generally, consumers prefer lambs because of their tenderness and low weight. The low weight translates to low price per head. Certain colours are preferred to others and especially so for some localised occasions. Sheep sold during the Easter and Id-al-fater weeks received the highest premiums in all the markets. The purpose for which sheep were bought made significant overall price differentials in the intermediate and redistributive markets. Virtually all sheep were bought for slaughter at the terminal market. A significant time trend was indicated in the three markets. However, this was not considered as a regular within-year cyclical phenomenon. Signifies the weights at which minimum price per kg occurs. 70 Summary and Implications Characteristics of sheep varied among the central highland markets, partly reflecting consumers' preference. For example, in contrast to the terminal market where sheep were purchased mainly for slaughter, relatively higher proportion of female sheep were observed in the intermediate and redistributive markets. Female sheep predominated in the intermediate market. The proportion of castrates, fat and very fat sheep was highest in the terminal market. However, most sheep sold in the central highland markets were of average condition and lambs of less than a year age. There is a considerable week-to-week variation throughout the year in sheep prices per kg in the central highland markets. These variations are related to variations in overall supply and demand as well as in characteristics of animals offered for sale. Factors affecting the number offered for sale include taking advantage of high demand during festivals, lambing season, as well as cash needs for crop inputs and later for food purchases just before harvest when grain stores are low. Animal characteristics that affect price per kg are weight, sex, age, condition and colour. In addition, the purpose for which the animal was purchased, whether for resale, slaughter, fattening or reproduction, affected price. Market composition with respect to these characteristics varied from week to week which suggests that weekly mean prices per kg vary partly due to varying market composition effect. A significant time trend over the year was also observed in two of the three markets analysed. Though this trend may continue into the next year, it is not regarded as a within-year cyclical phenomenon. Calculation of predicted price per kg multiplied by the observed weight demonstrated that the price-per-kilogram model is of equal predictive value for price per head when weight is known. Given the high variation in weights both within and between weeks, the price per kg would seem to be more useful market information than price per head. The results suggest that there may be some benefit from co-ordinating fattening, breeding and marketing programmes to take the best advantage from some preferred animal characteristics and selected festival markets. However, further analysis will be needed for optimising fattening and controlled breeding programmes. For example, production analysis is needed on costs as well as animal response to alternative fattening programmes and seasonal grazing to use in combination with information generated in the present study on effects of weight, age, condition, and seasons on prices. 71 References Carles A B. 1983. Sheep production in the tropics. Oxford University Press, New York, USA. 213 pp. Gryseels G. 1 988. Role of livestock on mixed smallholder farms in the Ethiopian highlands. A case study from the Baso and Worena wereda near Debre Berhan. PhD dissertation, Wageningen Agricultural University. The Netherlands. 249 pp. Herman L 1979. The livestock and meat marketing system in Upper Volta: Summary of evaluation of economic efficiency. In: Shapiro K H (ed), Livestock production and marketing in the Entente States of West Africa Summary report. CRED (Center for Research on Economic Development), University of Michigan, Ann Arbor Michigan, USA. Solomon Bekure and Tilahun Negussie. 1983. Livestock marketing studies. In: Pastoral systems research in sub-Saharan Africa. Proceedings of the IDRC/ILCA workshop held at ILCA, Addis Ababa, Ethiopia, 21-24 March 1983. ILCA, Addis Ababa, Ethiopia, pp. 327360. 72 Annex 1 . Estimated parameters of price/kg equation tor the terminal, intermediate and redistributive markets. Terminal market Intermediate market Redistributive market intercept 2.9967** 3.0769** 1.6915* Weight -0.0595** -0.0496 -0.0425 Weight2 0.0015** 0.0008** 0.0008** Time 0.0115 0.0024 0.0504 Sex: Male 0.3134** 0.1856** -0.0010 Castrate 0.5037** 0.2607 0.1374 Female 0.0000 0.0000 0.0000 Age (year): Less than 1 0.2078 0.9619** 0.4526** 1 0.1293 0.6331** 0.4273** 2 -0.0070 0.4099* 0.7036** 3 -0.2094 0.6306" 0.3545 4 -0.0259 0.0781 0.4760" Over 4 0.0000 0.0000 0.0000 Condition : Poor -0.4272 -1.1680 -0.4633 Average -0.3303 -1.6154 -0.4452 Fat -0.2619 -0.7811 0.3283 Very fat 0.0000 0.0000 0.0000 Colour : White 0.0805** 0.1197** 0.0563** Brown 0.0544** 0.1029** 0.0217* Black and white -0.0018 0.0276 -0.0455" Black 0.0000 0.0000 0.0000 Buyer's purp: Resale na -0.1361 -0.1201* Slaughter na -0.1417 -0.2942** Fattening na 0.2724 -0.0349 Reproduction na 0.0000 0.0000 Season : Christmas -0.2202* -0 3758** 0.0022 Easter 0.1804 0.3179** -0.0079 Id-al-fater 0.5626** 0.3139" 0.2809 73 Annex 1 (Continued) Terminal Intermediate Redistributive market market market Id-al-adha 0.2886 -0.2129* 0.0321 New Year -0.4876" -0.3139** -0.0456 Maulid -0.2702 -0.0391** -0.2401** Lent -0.1456 -0.2799** 0.0055 Non-festival 0.0000 0.0000 0.0000 Interaction with weight: Weight X time 0.0000 -0.0001 -0.0001 Weight X sex : Male -0.0107* -0.0055* 0.0033 Castrate -0.0122* -0.0045 0.0041 Female 0.0000 0.0000 0.0000 Weight X age (year): Less than 1 -0.0070 -0.0172* -0.0047 1 -0.0030 -0.0061 -0.0060 2 0.0004 0.0012 -0.0148* 3 0.0050 -0.0049 -0.0030 4 0.0003 0.0109 -0.0089 over 4 0.0000 0.0000 0.0000 Weight x condition: Poor 0.0106 -0.0042 0.0115 Average 0.0059 0.0253 0.0204 Fat 0.0041 0.0058 0.0074 Very fat 0.0000 0.0000 0.0000 Weight x buyer's purpose: Resale na 0.0101 0.0076** Slaughter na 0.0127* 0.0131** Fattening na -0.0053 0.0031 Reproduction na 0.0000 0.0000 Weight x season: Christmas 0.0096* 00138** -0.0070 Easter 0.0051 -0.0037 0.0144** id-al-fater 0.0023 -0.0061 -0.0102 Id-al-adha -0.0096 0.0074 -0.0011 New Year 0.0199** 0.0249** 0.0016 74 Annex 1 (Continued) Terminal ntermediate Redistributive market market market Maulid 0.0123 0.0060 0.0070 Lent 0.0045 0.0165** 0.0054** Non-festival 0.0000 0.0000 0.0000 Interaction with week: Week x sex: Male -0.0055" -0.0013 -0.0040** Castrate -0.0015 0.0055** -0.0013 Female 0.0000 0.0000 0.0000 Week x age: Less than 1 0.0041 -0.0044 0.0030 1 0.0044 -0.0026 0.0078** 2 0.0059* -0.0027 0.0038 3 0.0042 -0.0033 0.0088* 4 0.0030 -0.0030 0.0052* Over 4 0.0000 0.0000 0.0000 Week x condition: Poor -0.0039 0.0088 -0.0321* Average -0.0017 0.0111 -0.0342** Fat 0.0040 0.0126 -0.0409** Very fat 0.0000 0.0000 0.0000 N 6221 8546 7779 R2 (price/kg) 0.3238 02659 0.3583 *P<0.05; **P<0.01. na = not applicable, virtually all for slaughter. 75 Differences by geo-climatic zone in the economic returns from wool productivity and quality in Lesotho John P. Hunter Soil and Water Conservation and Landuse Co-ordinating Unit SouthernAfrican Development Coordination Conference (SADCC) Maseru, Lesotho Abstract This report presents an analysis of data obtained from a survey of approximately 1700 sheep at shearing sheds throughout Lesotho during the 1987/88 wool shearing seasons. It focuses specifically on regional differences in wool productivity and quality, the factors accounting for these differences and on the implications of these differences for farmers' wool returns and for livestock development policy. Effet des différences géoclimatiques sur les rendements économiques et la qualité de la laine au Lesotho John P. Hunter Résumé Ce rapport présente une analyse des données rassemblées dans le cadre d'une enquête effectuée au cours de la campagne 1987-1988 au Lesotho sur environ 1 700 moutons tondus dans des ateliers de tonte. L'auteur met en particulier l'accent sur les différences régionales observées dans la production et la qualité de la laine, les facteurs à l'origine de ces différences et leurs conséquences pour les politiques de développement de l'élevage et le revenu tiré de la laine par le paysan. 77 Introduction Wool and mohair are the most important outputs, by value, of Lesotho's agricultural sector and are strategically important to Lesotho's economy. Over 20 per cent of the value of agricultural output can be attributed to wool and mohair sales alone and, if the value of the joint products of sheep and goats (skins, meat and offspring) were taken into account, their contribution to the agricultural economy would be even greater. To farmers, wool and mohair sales constitute their largest single component of domestically generated monetary income. A study of the Sehlabathebe Range Management Area showed that 75% of the gross cash income generated by livestock was generated by sheep and goats and that they still contributed the major proportion of income even if all cash and non-cash livestock income were included (Lawry, 1986). The 1985 Livestock Holders Survey found a similar pattern nationally (Swallow et al, 1987a). Lesotho's wool and mohair exports comprise over 43 per cent of the value of all exports and, since the decline of diamond production in the early 1 980s, have been the largest export items. In addition to their role in generating income, the purchase of sheep and goats provide an important investment outlet for migrants' savings and, as such, are a major repository of rural wealth. At mean 1989 sale prices of 1 04 maloti per sheep and 80 maloti per goat, Lesotho's small ruminant flocks represent a total value to their owners of some 254 million maloti (approximately US $100 million). Livestock keeping and migrant labour are close complements in the rural Lesotho economy. The rapid increase in livestock numbers and the enthusiasm shown by Basotho for Merino wool and mohair production in the early 20th Century followed the transformation of the Lesotho economy from one exporting grain to regional markets to one exporting labour to South African mines and farms. Since, unlike crop production, extensive livestock production does not require adult male labour (it can be undertaken with adolescent herders), it thus permits men to absent themselves from the farmstead for long periods. Livestock and migratory labour are complementary in other ways, as well. They provide a ready (and in many cases, only practical) form of investment for a migrant's savings which can be liquidated in times of need or upon retirement. Livestock are normally a very profitable investment: research in 1 985/86 found that, on average, Lesotho cattle, sheep and goats paid a real rate of return of between six and ten per cent (Swallow et al, 1987b). Changes in product prices have recently altered these rates of return but, for the most part, they are still well above the rates paid by most alternative investments available to rural Basotho and, like any good investment, they provide current income and capital appreciation. This income is but a subsidiary to a rural household's main income source from wages, however. Data from the mid-1 970s from theThabaTseka Mountain Development Project and the Phutiatsana Irrigation Project indicate that only about 30% of a typical rural household's income is derived from all domestic sources, with 78 agriculture accounting for about 60% of this. Two-thirds of this agricultural income (or 12% of the total) is contributed by livestock (van der Wiel, 1977). Thus, livestock-derived income provides an income supplement which may make a household's well-being easier and more secure but it is the rare, and invariably very poor household, which is reliant, solely or largely, on livestock for its existence. Although, according to rural surveys, farmers are not indifferent to the low productivity of their livestock (Hunter, 1987), they may have neitherthe labour resources nor sufficiently strong incentives to do much about it. Added to the socio-economic constraints on livestock production are a variety of ecological constraints. Lesotho's harsh winters take their toll on livestock condition with adverse effects on survival, reproduction and fleece quality. According to Range Management Division estimates (Chris Weaver, personal communication), the commonly owned range is about 40% overgrazed and able to provide only relatively meagre nutrition to the animals grazing on it. The rugged terrain hinders livestock marketing which, in turn, makes culling difficult. As a result of these constraints, livestock productivity is low. In an attempt to better evaluate this low productivity with an aim to designing more effective improvement programmes, some 1700 sheep and 2000 goats were sampled at a cross-section of government and private shearing sheds throughout Lesotho during 1987/88. Animals and fleeces were weighed, age and sex were determined, and samples of wool and mohair were taken from the side fleece for further laboratory testing. This paper analyses some of the regional disparities in wool production revealed by this effort. Additional results and data on mohair production, as well as research procedure and methodology may be found in Hunter (1990a and 1990b). Geo-climatic Zones of Lesotho Lesotho is conventionally divided into four geo-climatic zones (World Bank, 1 985) . Tables 1 and 2 provide basic statistics on these regions. Along the western and southern edge of the country, below 1830 meters in altitude, lie the lowlands. This zone is relatively flat and contains most of the arable land, about 43% of the rural households and virtually all of the urban population. Over 50% of the harvested maize land and 57% of the harvested land devoted to sorghum in Lesotho is in the lowlands (BOS, 1987). Because of the urban concentrations, the lowland zone contains most of the employment opportunities and is more economically diverse than the other zones. This zone is often divided, on the basis of mean annual rainfall, into subzones (de Baulny, 1981). The northern lowlands (NL), roughly from the Maseru-Berea District border northwards, normally receives between 650 and 850 mm of rainfall per annum. The southern lowlands (SL), south of this border, is rather drier—receiving a mean rainfall of between 600 and 750 mm per annum. Some areas, such as parts of Mafeteng and Mohale's Hoek districts may receive less than 650 mm per annum. 79 Table 1 . Land-use pattern in Lesotho region. Livestock Cropping grazing Others Lowlands Mountain Foothill Senqu river valley 53 33 14 14 80 6 35 55 10 25 70 5 Source: Ministry of Agriculture, Lesotho (1988). Table 2. Distribution of rural households and livestock by region, Lesotho (%). Rural households Sheep Goat Cattle Lowlands (LL) 43 17 19 32 Mountains (MTN) 20 68 44 37 Foothills 25 8 22 22 Senqu river valley 12 7 15 9 Source: BOS (1987); Bureau of Statistics, Maseru, Lesotho, unpublished data. The Mountain (MTN) zone comprises about 60% of the total land area and lies above 21 30 meters. Rainfall is highly variable in this zone, ranging from over 1 000 mm per annum in the higher elevation along the front range of the Maluti and northern Drakensberg to about 600 mm on the fringes of the Senqu river valley (de Baulny , 1 981 ). Summers are short and winters may be bitterly cold. As a result of both the rugged terrain and the short growing season, there is relatively little arable land in this zone, although field crops are possible in protected valleys. A little over 80% of the land, mostly rangeland, is devoted to livestock production. Although the mountain zone contains only about 20% of the rural households, they own 68% of the sheep, 44% of the goats and 37% of the cattle in Lesotho (Table 2). Virtually all adult rural males engage in labour migration at some point in their lives but, according to data from several development projects, the mean length of a mountain male's migrant career (20 years) is almost 20% shorter than that of his lowland counterpart (van der Wiel, 1977). The Foothill (FTL) zone lies between the lowland and mountain zones at an altitude of 1830 to 2130 meters and comprises about 10% of the total land area of the country. It receives rather higher rainfall than the Lowland areas: most of this zone gets between 800 and 950 mm mean annual precipitation (de Baulny, 1981). As in the lowlands, maize production predominates, followed by sorghum 80 and beans. While not as economically diverse as the lowlands, the foothills' close proximity to the lowland zone gives it economic opportunities that are lacking in the mountains and Senqu river valley. The Senqu River Valley (SRV), which divides the mountain zone, lies in a "rain shadow" and receives the least rainfall of all zones: normally between 500 and 600 mm per annum (de Baulny, 1 981). It contains only about 1 1 % of the country's total land area. Only slightly less so than in the mountains, the land is devoted overwhelmingly to livestock grazing (Table 2). Although maize production utilises the largest proportion of the harvested area, sorghum production is relatively much more important in this zone, as befits its rather drier climate. Implication of Zonal Differences for Small Ruminant Production and Development These differences in zonal resources and characteristics have a number of implications for the productivity of small ruminants and the quality of their fleeces. The much higher population density and relatively small proportion of the land area devoted to grazing means that lowland village grazing areas are under heavy pressure. Although this can be relieved to some extent by taking livestock to seasonal mountain cattle posts, survey data from the 1985 Livestock Holders Survey suggest that this practice is not as common as is often supposed (Swallow et al, 1987a). Of the 250 sheep-owning households surveyed nation-wide, 54% kept their sheep in village grazing areas or dry-lots year round. This involved about 50% of the sheep in the survey. Examination of the raw data indicates that these practices are somewhat more common in the lowlands than in the other areas. Distances are often long between lowland villages and their mountain cattle-posts. Trekking costs, in terms of time and lost animal condition, are high. In addition, the distance may make careful management by the owner difficult, thus increasing the threat of loss by theft. Finally, the higher per flock costs of keeping smaller lowland flocks at cattle-posts may encourage their being kept year-round near the village. Thus, because of poorer grazing resources, lowland flocks are relatively small and unproductive. Small flocks mean, in turn, that selective breeding is more difficult and the use of improved breeding stock is less cost-effective. Also, because lowland residents have a number of alternative domestic income-earning options, such as crop production or wage employment, their flocks are less important to them as income earners. Mountain residents, by contrast, have few alternative domestic income-earning options to livestock. Because of this, livestock ownership is more widespread, flocks are much larger and the better pastures in this zone provide a more solid foundation for a livestock industry. On most counts, mountain flocks are much more productive and much higher income earners than flocks from other zones. 81 The dominant sheep breed in Lesotho is the wool Merino. Table 3 summarises the zonal differences in mean measures of wool productivity and quality. The superiority of the MTN and NL zones in wool production is particularly evident in their higher greasy fleece weights and clean wool yields and in their longer staple lengths. Table 3. Zonal differences in wool productivity and quality. Mean characteristic Lesotho NL1 SL SRV FTL MTN Shorn body wt (kg) 31.4 29.9 29.5 30.8 29.8 32.3 Greasy fleece wt (kg) 26 2.5 2.4 2.2 2.1 2.9 Clean oven dry yield (%) 53.3 54.2 48.4 52.8 54.1 54.1 Clean fleece wt (kg) 1.4 1.4 1.2 1.2 1.1 1.6 Staple length (cm) 6.4 6.5 6.2 6.0 5.7 6.8 Crimps (per inch) 16.2 15.4 16.3 17.0 15.5 15.6 Fibre diameter 20.0 19.6 20.1 19.5 20.0 20.2 Note: All measures are for adult (2 teeth and older) Merinos. 1. NL = Northern lowlands; SL= Southern lowlands; SRV = Senqu river valley; FTL = Foothills; MTN = Mountain. The clean-yield measure (the proportion of clean wool obtained from a unit of greasy wool) is of particular importance since wool is sold in the clean state and yield is generally rather low. The SL and SRV—the driest and dustiest areas of the country—register the lowest yields. Along with the FTL zone, they also have the poorest mean greasy fleece weights. The multiple of these measures, the clean-fleece weight, looks especially poor for these zones. Generally, the SL and SRV produce 26% lighter clean fleeces than the MTN zone. The FTL zone performs even more poorly—producing an average clean fleece only 43% as heavy as the MTN zone. Compared to the second-best wool producing area, the NLzone, these poor productivity areas produce about 15% lighter clean fleeces per sheep. In addition to the purely quantitative differences, differences in the qualitative measures are also important. Although there are little differences in mean fibre diameter (all are fine), there is a statistically significant difference (at the 95% level) in staple lengths. Since this is a key determinant of wool style or type, this affects the quality and, hence, the wool price offered. In general, sheep from FTL and SRV, in particular, produce markedly shorter (12% and 16%, respectively) staples than those from MTN. Indeed, FTL staples are an average of one centimetre shorter than MTN staples. 82 While these differences in mean measures highlight some marked zonal differences in productivity and quality, these averages conceal several important distinctions. To highlight these, it is necessary to focus on the distribution of these measures within each of the zones. This is done in the next section. Clean wool fleece weights Table 4 focuses on differences in the distribution of the quantitative measure of clean fleece weights. Over 40% of the fleeces in the two poorest wool-growing zones, FTL and SL, weigh less than one kilogram. In the SRV, the intermediate zone, 36.5% fall into this category. By contrast, only about 23% of fleeces in the two best zones are below one kilogram in weight. Table 4. Zonal distributions of clean wool fleece weights (per cent by class). Weight classes Lesotho NL1 SL SRV FTL MTN 1. < 0.75 kg 9.9 6.9 15.0 9.1 16.8 8.0 2. 0.75-1.00 kg 19.3 15.9 26.6 27.4 26.6 15.1 3. 1.00-1 .25 kg 20.3 24.8 28.3 29.6 20.7 16.3 4. 1.25-1.50 kg 16.5 19.3 15.0 16.1 16.8 16.3 5. 1 .50-2.00 kg 17.8 19.0 13.4 14.5 15.2 19.7 6. > 2.00 kg 16.2 4.1 1.7 3.3 3.9 24.6 1. NL = Northern lowlands; SL = Southern lowlands; SRV = Senqu river valley; FTL = Foothills; MTN = Mountain. Note: All measures for Tables 2, 3 and 4 are for adult 2-teeth and older) Merinos. Lesotho figures are weighted averages of zonal figures. Columns may fail to add to 100 because of rounding. While these two zones have relatively few low-weight fleeces, they have a high proportion of relatively heavy fleeces. Almost 45% of MTN fleeces weigh over 1.5 kg and almost one-quarter of them are over 2 kg in weight. No other zone has more than 5% as heavy. Although NL does not score as high in heavy-weight fleeces, almost 45% of its fleeces are over 1.25 kg in weight. Wool staple length Wool's style or type is largely determined by the length of its staple. Longer (combing) wools are more suitable for weaving and use in the worsted trade while shorter (carding) wools are used for knitting. Normally, the longer wools fetch higher prices. Table 5 presents the zonal distribution of sample staple lengths. As with clean fleece weights, the superiority of the MTN and NL zones is evident. Almost one quarter of the fleeces have staples falling into the longest class and 83 almost 70% of MTN fleeces and over 60% of NL fleeces have staples 60 mm or longer. By contrast, none of the other zones have more than 12% of their fleeces with staples 75 mm or longer and only about 45% are over 60 mm. Fortunately, very few fleeces have staples of less than 30 mm and, with the exception of FTL and SRV, the incidence of clearly definable carding wools (i.e. less than 45 mm) is not great. Table 5. Zonal distributions of wool staple lengths (per cent by class). Length classes Lesotho NL1 SL SRV FTL MTN Class A: >75 mm 18.8 23.5 10.1 12.0 11.0 23.0 Class B: 60-75 mm 41.9 37.5 44.5 33.9 30.4 46.8 Class C: 45-60 mm 30.9 28.7 34.5 41.0 40.9 26.1 Class D: 30-45 mm 7.7 8.8 9.2 13.1 16.6 4.0 Class E: <30 mm 0.7 1.5 1.7 0.0 1.1 0.2 1. NL = Northern lowlands; SL = Southern lowlands; SRV = Senqu river valley; FTL = Foothills; MTN = Mountain Staple length is not the only factor determining whether or not a wool is suitable for combing or carding. Tenderness, which can cause breakages in the wool, can also cause a longer wool to be unsuitable for combing. No data exist on the degree of soundness in Lesotho's wool although the extreme nutritional stress from Lesotho's cold winters combined with the added stress of reproduction, suggest that it is likely to be particularly problematic for ewes. It may also be more prevalent in lowland village grazing areas, as well. Wool fibre diameter A key component of wool quality is its fineness measured by mean fibre diameter. This is also a key determinant of a wool's price. In general, all of Lesotho's wools are fine (all of the zonal mean fibre diameters are less than 20.5 microns) and this accounts for their generally good market acceptance. By general agreement within the trade, however, this fineness is less the result of genetics and more the result of the poor nutrition offered by Lesotho's overgrazed range. Because of this, it is often referred to as "hunger-fineness". Despite this general fineness, there is some zonal variation in the distribution of fineness classes. This is evident from Table 6. In this measure, the performance of the two best zones diverges. The second-best NL zone has over 70% of its wool in the two finest categories and very little in the two coarsest classes. The MTN zone, by contrast, has less than 84 60% in the finest and almost 20% in the coarsest classes. Two reasons for these differences suggest themselves. One may be nutritional. Despite the relatively high rainfall in the NL zone and its close proximity to mountain grazing, high population density in this zone may restrict grazing, particularly in village areas, and make the impact of hunger-fineness more acutely felt. The other may be an unintended effect of the 50-year-old Livestock Department programme of supplying improved rams to farmers for breeding. Although these are supplied at cost, they are only cost-effective when put to relatively large flocks and, as a result, they go disproportionately to MTN stock keepers. Rams are selected for body conformation, size, fleece weight, and staple length but rarely for the fineness of their wool. As a result, many of the rams are characterised by medium to medium-coarse wools. The higher proportion of coarser wools in the MTN zone may be the effect of this breeding policy. Table 6. Zonal distributions of wool fibre diameters (per cent by class). Diameter classes Lesotho NL1 SL SRV FTL MTN 1. <19.5 microns 2. 19.5-20.5 microns 3. 20.5-21.5 microns 4. 21.5-22.5 microns 5. 22.5-23.5 microns 6. >23.5 microns 1. NL = Northern lowlands; SL Southern lowlands; SRV = Senqu river valley; FTL = Foothills; MTN = Mountain Economic returns These zonal differences in wool and mohair productivity and quality translate into marked differences in the zonal economic returns from keeping Merinos. LPMS data (from government wool-sheds only) for the 1985/86 wool and mohair seasons were analysed, using ordinary least-squares regression, for zonal differences and for the impact of membership in a Wool and Mohair Growers Association (WMGA) on a grower's economic returns (Hunter, 1987). Results of this analysis are reported in Table 7. This table refers only to non-WMGA members (73% of sheep owners marketing through LPMS) who own 57% of sheep shearing at government wool-sheds. WMGA membership narrows, but does not eliminate, these zonal economic differences. Mountain wool producers receive an 18-37% price advantage over other wool producers. Despite the fact that MTN wool is generally rather coarser (stronger) 45.8 51.0 42.9 54.3 48.9 42.9 16.4 19.6 20.6 16.3 15.9 15.4 14.0 18.2 15.1 17.9 13.7 12.5 8.9 8.4 7.9 5.4 8.2 10.1 6.2 2.1 9.5 3.8 5.5 7.1 8.5 0.7 4.0 2.2 7.7 11.9 85 than wool from other zones, the combination of heavier fleece weights, relatively high yields and long staples translate into higher gross incomes per sheep. In this regard, MTN producers have a 25% advantage over the next-best zone, the NL, and a 32-50% advantage over the remaining zones. When the substantial differences in flock size (relating mostly to the different options and opportunities available to the residents of the different zones) are taken into account, MTN flocks earn 3-8 times as much for their owners as flocks from other zones. Although NL flocks generally produce relatively heavy fleeces and relatively high quality wool, the small average flock size probably makes them only marginal-income supplements for their owners. Table 7. Zonal differences in wool prices and incomes and ranking of rates of return for sheep enterprises. Mean characteristic NL1 SL SRV FTL MTN Rock size (animals) 20 18 35 36 100 Rock size (index no.)2 32 29 56 57 100 Wool price (index no.) 81 73 85 83 100 Income/sheep (index no.) 80 67 76 76 100 Income/flock (index no.) 19 12 37 34 100 Ranking 2 4 3 5 1 1. NL = Northern lowlands; SL = Southern lowlands; SRV = Senqu river valley; FTL = Foothills; MTN = Mountain. 2. Index numbers are calculated as percentage of Mountain values. Data apply to non-WMGA members only. Another way of evaluating the economic effect of zonal differences terms is to focus on the rates of return generated by a sheep enterprise. Livestock budgets, originally prepared by Swallow et al (1987b) and updated and revised by Hunter and Carvalho (1990, unpublished) were used to generate zonal rates of return. It was assumed that all zones had similar costs and non-wool returns from livestock. This assumption is probably unrealistic, however. Stock keepers with larger flocks would gain certain economies of scale in herding costs but, because of higher levels of flock management, may spend more on veterinary costs or supplemental feed. Smaller stock keepers may have higher average herding costs but may spend less on other aspects of flock maintenance. From the standpoint of returns, larger stock keepers may have a higher marketed offtake from their flocks than smaller counterparts. However, while these adjustments should alter the absolute magnitude of the rates of return, they probably would not affect the overall rankings. Using 1985/86 data, rates of return to a typical sheep enterprise ranged from 6% in the foothills to 1 1% in the mountains. The zonal rankings are presented in Table 7. 86 The extremely close correspondence in rankings by both methods is evident. Since 1985/86, wool prices have increased dramatically (although the market has recently reversed). As a result, the average sheep enterprise returned 14.5% on investment in 1988/89. Conclusions and Recommendations Generally speaking, wool productivity in Lesotho is rather low. Because of the practices of kraaling animals at night and grazing cultivated fields, Lesotho's wool is dirty and clean wool yields are low. Clean fleece weights are also low because of overgrazing and poor nutrition and, perhaps, for genetic reasons. Only 60% ofthe nation's clip is longer than 60 mm in length for which poor nutrition, genetics and climatic stress are principally responsible. No matter which perspective one takes, whether that of productivity, quality or economic returns, there are clearly identifiable rankings of wool production by zone, however. The MTN zone is unambiguously most favourable to wool production. Fleece weights are heavier, yields are higher, staples are longer, wool prices are higher, and farmers'incomes per animal are larger. The FTL zone, from almost all perspectives, is the worst wool-producing zone. The SRV is an intermediate zone and its performance is, indeed, fairly representative of the "average" for Lesotho. In the lowlands, the SL zone is the second-worst for wool production while the NL zone is the second-best. Both conceptually and because of data limitations, it is difficult to disentangle the relative effects of genetics and of environment and management on these rankings. From the environmental perspective, village grazing areas in the lowlands are invariably crowded and heavily organised. If lowland animals use mountain cattle-posts, the trekking between village and highland pastures imposes its own form of stress on fibre production. Although highland pastures are also overgrazed, the quality of their vegetative cover is generally much higher and stocking rates rather lower than in lowland grazing areas. The better nutrition in the mountains contributes to longer staples, larger body mass and higher fleece mass. Other factors may play a role too. Mean sheep flock size in the mountains is about two to three times larger than in other zones. The purchase of improved Merino rams for breeding is much more cost effective for larger flocks than for small ones. Not surprisingly, most of the improved rams procured by the government from South Africa go to mountain producers. Further, because mountain residents have few alternative domestic income-making opportunities, they are more income-dependent on their flocks and may practice more careful husbandry than their lowland counterparts. Survey results indicate that larger stock keepers are more likely to dose their animals more frequently and more likely to dip them twice as recommended (Hunter, 1987). They also show the 87 strongest commitment to keeping sheep primarily for wool production (Hunter, 1987). Interviews with farmers indicate that many of the larger mountain producers employ adult herders, a practice that would not be cost-effective for a small flock. According to agricultural census data, there has been a steady reduction over the last 30 years in the proportion of sheep in the highly productive MTN zone and an increase in the less productive SRV and FTL zones. Not only has this shift put increased pressure on grazing areas least able to support them but, other things being equal, it would have reduced overall wool productivity and quality measures. Although the reasons for these shifts are probably complex and may be associated with the upsurge in economic activity in the lowlands in the Post-Independence era, livestock policy decisions can have an impact on the geographical distribution of sheep in the future. The following policy directions may be worthy of consideration: a. Wool production in the southern lowlands should be discouraged. Generally, animals in this zone are so unproductive, wool quality is so poor, and flocks are so small that the income received from fibre production is minimal. It is likely that wool production is but a secondary function of livestock production and that repositories for savings and meat production are primary functions. Merinos are not particularly well-suited to either of these latter functions. In this zone, livestock development funds would have a greater impact on farms incomes if they were redirected from wool production to activities with higher potential. b. Wool production in the foothills should be discouraged also. Development funds would have a higher pay-off if money for wool development were concentrated in the mountains and northern lowlands. Angora goats perform very well in this zone, however, and should receive greater development emphasis (Hunter, 1990b). c. Although the Senqu river valley is a marginal wool producer, there are limited alternative economic activities for residents of this zone. Therefore, it is probably desirable to continue to direct small ruminant development funds to this zone. d. Much-discussed restrictions on transhumance movement of livestock from lowland areas to highland cattle-posts would discourage extensive livestock production in the lowlands. Careful thought should be given to the impact of this restriction on wool production in the northern lowlands, however. e. Although administrative and enforcement constraints may be large, consideration might be given to a discriminatory grazing fee which imposes higher fees against small ruminants in those areas in which they are particularly unproductive. Although wool productivity in Lesotho is generally low, there are zones where performance is reasonably good. Concentrating wool production and Merino development resources, including funds, extension efforts and physical infrastructure in these areas while de-emphasising the low productivity areas would raise overall productivity measures and lead to a more productive utilisation of scarce resources. Reference de Baulny. 1981. Anatomy of a drought (November 1 979-October 1980): Lesotho. Department of Water Resources, Ministry of Water, Energy and Mining, Lesotho. Bureau of Statistics (BOS), Lesotho. 1987. 7986 population census. Preliminary results. Bureau of Statistics, Lesotho. Hunter J P. 1987. The economics ol wool and mohair production and marketing in Lesotho. Research Division Report RD- R-80, Ministry of Agriculture, Maseru, Lesotho and ISAS Research Report 16, Institute of Southern Africa Studies, Roma, Lesotho. Hunter J P. 1 990a. Characteristics ol Lesotho's wool and mohair. Statistics and analytical commentary. Agricultural Information Services Report R-L-90-2. Ministry of Agriculture, Cooperatives and Marketing, Maseru, Lesotho. Hunter J P. 1 990b. Regional differences in the production of wool and mohair in Lesotho. Agricultural Services Report, Ministry of Agriculture and Marketing, Maseru, Lesotho. Lawry S W. 1 986. Livestock and range management in Sehlabathebe: A study of communal resource management. Land Conservation and Range Development Project and Range Management Division, Ministry of Agriculture, Maseru, Lesotho. Ministry of Agriculture, Conservation Division. 1988. National resource inventory of Lesotho. Ministry of Agriculture, Conservation Division, Lesotho. Swallow B M, Motsamai M, Sopeng L, Brokken R F and Storey G. 1987a. A survey of the production, utilization and marketing of livestock and livestock products in Lesotho. Research Division Report RD-R-81, Ministry of Agriculture, Maseru, Lesotho and ISAS Research Report 17, Institute of Southern African Studies, Roma, Lesotho. Swallow B M, Brokken F R, Motsamai M, Sopeng L, and Storey G, 1987b. Livestock development and range utilization in Lesotho. Research Division Report RD-R-82, Ministry of Agriculture, Maseru, Lesotho and ISAS Research Report 18, Institute of Southern African Studies, Roma, Lesotho, van der Wiel A C A. 1977. Migratory wage labour. Its role in the economy of Lesotho. Mazenod Book Centre, Mazenod, Lesotho. World Bank. 1985. Lesotho agricultural sector review, Vol. II. World Bank, Washington, DC, USA. 89 Partial budget analysis for on-station and on-farm small ruminant production systems research: Method and data requirements S. Ehui and B. Rey International Livestock Centre for Africa PO Box 5689, Addis Ababa, Ethiopia Abstract Economic constraints and opportunities for improving small ruminant production systems in sub-Saharan Africa must be understood as the basis for developing interventions. However, most national agricultural research systems (NARS) face a shortage of livestock economists, and biological scientists often lack the skills needed to conductan economic evaluation of the results of their work. This study presents the partial budget analysis (PBA) framework for the economic analysis ofsmall ruminantinterventions for use by livestock specialists. The logic of the PBA and the data needs are discussed first. This is followed by an empirical example which analyses the economics of endoparasite control in sheep production system of the Ethiopian highlands. The analysis is based on experimental data from ILCA's Debre Birhan station. The experiment consisted of four treatments: I. Grazing natural pasture; II. Drenching with anthelminthics; III. Supplemental feeding with wheat bran and noug (Guizotia abyssinica) cake; IV. Supplemental feeding and drenching. Results show that net returns per ewe of treatments II, III and IV exceeded the net return of the control by Ethiopian birr (EB) 4.86, 8. 56 and 9.35, respectively (US$ 1 = EB 2.07). The increase in cost for treatment II relative to the control was EB 1.44; the added net benefit from this treatment was EB 4.86 per ewe, giving a marginal rate of return of 334%. The increase in cost of treatment III relative to treatment II was EB 19.60, while the increase in net return was EB 3. 71 per ewe, giving a marginal rate of return on the increased expenditure of 19%. The cost of adding drenching to treatment III (treatment IV) was EB 3.58 per ewe, while the increase in net return relative to treatment III was EB 0. 78 for a marginal rate of return of 22%. The marginal rate of return of treatment IV relative to treatment III was 91 19.4%. Given the high cost of capital, treatments III and IV cannot be recommended. Drenching alone (treatment II) yielded a very high marginal rate of return under experimental conditions and should be tested under on-farm conditions. L'analyse par budgets partiels dans la recherche en station et en exploitation sur les systèmes de production ovins et caprins: méthode et données S. Ehui et B. Rey Résumé Les contraintes et les opportunités économiques relatives à l'amélioration de la production des petits ruminants en Afrique au sud du Sahara doivent être à la base du développement d'interventions technologiques. La plupart des systèmes nationaux de recherche agronomique en Afrique sont cependant fréquemment confrontés à une carence en économistes, et les autres chercheurs ne sont pas toujours équipés pour faire l'analyse économique de leurs résultats. Nous présentons ici la méthode des budgets partiels comme un cadre d'analyse économique d'interventions à la disposition des chercheurs zootechniciens. La logique de la méthode et les données nécessaires seront d'abord présentées. Un exemple empirique d'application sera ensuite discuté. Il se base sur une expérimentation de contrôle de l'endoparasitisme de moutons des hauts plateaux éthiopiens à la station de Debre Berhan du CIPEA, avec quatre traitements: 1. pâturage naturel, 2. déparasitage interne, 3. supplémentation au son de blé et au tourteau de Guizotia abyssinica, 4. supplémentatione et déparasitage interne. La marge nette des traitements 2, 3 et 4 a dépassé celui du témoin de 4,86; 8,56 et 9,35 Birr éthiopien (EB). Le surcoût du traitement 2 par rapport au contrôle est de 1,44 EB, et la différence de marge nette de 4,86 EB par brebis traitée; donnant un taux de rentabilité marginale de l'intervention de 334%. Par rapport au traitement 2, le traitement 3 coûte 1 9, EB de plus et procure 3,71 EB de marge nette supplémentaire, ce qui lui donne une rentabilité marginale de 19%. Le déparasitage ajouté au traitement 3 coûte 3,58 EB par brebis et rapporte 0,78 EB de plus, fournissant un taux de rentabilité marginale de 22%. Le taux de rentabilité du traitement 4 comparé au traitement 3 est de 19,4%. Au vu du coût élevé du capital, la supplémentation seule ou associée ne 92 peut etre recommandee. Le d6parasitage seul offre un taux de rentabilité marginal tres éleve en conditions experimentales et devrait etre teste en exploitation. Introduction Small ruminants (i.e. sheep and goats) are essential components of farming systems in tropical Africa. They are raised mainly for meat, milk, and skin providing a flexible financial reserve (Social Security) in bad crop years for the rural population (Sumberg and Cassaday, 1985). Tropical Africa contains about one-third of the world's goats and one-sixth of its sheep. On average there is one goat or sheep on every 10 ha of tropical Africa and there are about 1.1 head of goats and sheep per person employed in agriculture. Available estimates also show that sheep and goats are equivalent in weight terms, to about 17% or the total domestic ruminant biomass of tropical Africa (Wilson, 1988). Sheep and goats are important also because: (1) they require minimal inputs and maintenance costs; (2) they are less susceptible to stress due adverse changes in climatic conditions (e.g. drought); and (3) they have a relatively high reproduction rate and are easy to dispose of (Winrock International, 1983). Thus in face of the declining crop yields due to movement of cropping onto marginal soil types and diminishing fallow periods, improvement in the production of sheep and goats is likely to improve the welfare of smallholders (Peacock, 1987). A number of national research organisations in partnership with the International Livestock Centre for Africa (ILCA) are conducting research to improve the productivity of small ruminants on the continent. However, although they provide insights into improving small ruminants productivity, some new technologies are not adopted because of lack of economic advantage over current production methods (Amir and Knipscheer, 1987). Increased animal yield is a necessary but insufficient condition for adoption. A new technology must also be profitable (Nagy and Sanders, 1990). For example, the economic benefits of innovation aimed or reducing reproductive wastage and of improved health, nutrition and management must be determined as a basic for recommending such to farmers. One problem faced by most NARS, however, is an acute shortage of livestock economists, and biological scientists often lack the skills to conduct an economic evaluation of the results of their work. Moreover, most manuals on economic analysis of biological interventions are crop-oriented (e.g. CIMMYT, 1988; MSTAT-C, 1988). Animals, however, have different characteristics making the use of these manuals not immediately accessible to animal scientists. It is therefore necessary that livestock specialists acquaint themselves with basic economic methods for the economic evaluation of small ruminant technologies. This paper presents the Partial Budget Analysis (PBA) approach for analysing the economic benefits of alternative small ruminant technologies. While there are 93 other types of budgeting (e.g. whole farm and enterprise budgeting), the partial budgeting procedure is the most useful in farming system research (Worman et al, 1990). The next ection presents the logic of the partial budget approach. The third section describes the data-set required to perform the analysis. Special attention is paid to the direct and indirect benefits and cost necessary for a correct evaluation of small ruminant technologies. In the fourth section an example is provided which is based on experimental data collected in the Ethiopian highlands. The last section presents some concluding qualifications and comments. Logic of the partial budget approach Partial budgeting is a method of organising experimental data and information about the cost and benefits from some change in the technologies being used on the farm. The aim is to estimate the change that will occur in farm profit or loss from some change in the farm plan (Boehlje and Eidman, 1984). Partial budgets do not calculate the total income and expenses for each of the alternative plan but list only those items of income and expense that change. They measure changes in income and returns to limited resources, provide a limited assessment of risk and, through sensitivity analysis, suggest a range of prices or costs at which a technology becomes profitable (Mutsaers et al, 1 986). Assuming for simplicity that the farmer's objective is to maximise returns, the method is as follows: Let Nl denote the net income from the sale of animals, or animal products, i.e. the amount of money which is left when total costs (TC) are subtracted from total returns (TR) : NI=TR-TC (1) Total costs include the costs of all inputs such as feed, non-feed inputs (e.g. veterinary costs, wages of hired labour, etc.). TC can be separated into two categories: Fixed costs (FC) and variable costs (VC): TC = FC + VC. (2) Fixed costs represent the costs that do not vary when comparing alternative technologies (e.g. land). Variable costs are those that do vary between the technologies, such as the amount of feed or labour used. In deciding whether or not to adopt a new technology, a farmer will want to know if it will increase his or her income. Similarly in order to properly screen among alternative technologies for further testing, a researcher will want to know which of the interventions is potentially economically more attractive. The increase of changes in net income (ANI) is the difference between the change in total returns (aTR) and the change in total costs (ATC): 94 ANI = ATR-ATC (3) = ATR - AVC - AFC = ATR - AVC, since, AFC = 0 AFC is equal to zero because by definition fixed costs do not vary. Assuming that capital is not a constraint, the technology with the highest ANI will be recommended. However, higher benefits may not be attractive if they require very much higher costs. New technologies typically require a package of increased inputs (thus additional costs), and farmers will want to consider the increase in costs in their decision. Thus, it is necessary to compare the extra (or marginal) costs with the extra (or marginal) net benefits. Another criterion which takes the cost factor into account is the marginal rate of return (MRR). MRR measures the increase in net income (ANI) which is generated by each additional unit of expenditure (AVC): MRR = ANI/AVC (4) In other words MRR measures the effect on net return of additional capital invested in a new technology, compared to the present one (CIMMYT, 1988). It is not necessary to calculate MRR if the new technology costs less than the farmer's present technology, or if the new technology yields a lower benefit than the present one for a comparatively higher cost. When this occurs, the technology is said to be "dominated". For practical purposes Table 1 summarises the methodology for conducting partial budget and marginal analyses. In making recommendations, three criteria must be observed: First, if net income remains the same or decreases, the new technology should not be recommended because it is not more profitable than the farmer's present technology ; second, if net income increases and variable costs remain the same or decrease, the new technology should be recommended because it is clearly more profitable than the farmer's technology; and third if both net income and variable cost increase (this is usually the case), the marginal rate of return should be looked at. The greater the increase in net income and the higher the marginal rate of return, the more economically attractive the alternative technology is. Identification of costs and benefits for small ruminants The most difficult task in performing partial budget analysis, is the proper identification of the costs and benefits associated with the alternative technologies. Using poor data can lead to wrong recommendations. The minimum amount of data which must be collected depends on the design of the trial and the questions for which answer are required. Once it has been decided 95 Table 1. Partial budgeting format 1 . Additional income: List the items of income from the alternative plan that will not be received from the base plan. 2. Reduced expenses: List the items of expense for the base plan that will be avoided with the alternative plan. 3. Subtotal: 1 + 2. 4. Reduced income: List the items of income from the base plan that will not be received from the alternative. 5. Additional expenses: List the items of expense from the alternative plan that are not require with the base plan. 6. Subtotal: 4 + 5. 7. Difference: 3 - 6: A positive (negative) difference indicates that the net income of the alternative exceeds (is less than) the net income of the base plan by the amount shown. Source: Adapted from Boehlje and Eidman (1984). what questions need answers, it is possible to identify the variables which will provide the necessary data use in conducting the analysis. Generally the following data are required for partial budget analysis: (1) quantities of inputs which vary between alternative technologies; (2) prices of these variable inputs; (3) yields or productivity levels resulting from the alternative technologies; and (4) prices of the outputs. All benefits and costs should be calculated using farm-gate prices. That is, the actual price which the farmer pays for the inputs or receives for his products. Thus all inputs prices should account for any cost, such as transportation required to bring the input to the production site. Similarly if the farmer sells his animal off the farm, then the transportation and, storage charges, and/or marketing costs in delivering it to the sale point should be deducted from the market price of the animal. If a technology affects the quality of the animal (e.g. attractiveness of the colour, fattened tail, appearance), different market prices should be applied for the different qualities. Sometimes it is necessary to value the fixed assets (e.g. facilities, equipment). These are usually valued using the depreciation technique. A simple formula for depreciation is the original cost minus any expected salvage value divided by the number of years of useful life (Wormanetal, 1990). 96 Table 2 presents a breakdown of the benefits for the economic analysis of small ruminants. Four products are generally consid ered important : (1) reproductive capacity of animals, (2) weight gain; (3) milk yield, (4) meat yield; and (5) manure. The importance of input depends on the technologies being evaluated. But both cash and r on-cash costs should be identified. Cash costs include items like feed costs (e.g. grain, wheat bran), non-feed costs (e.g. veterinary charges, wages for hired labour). Non-cash costs include items like family labour, capital costs, depreciation costs (equipment and animals) and other non-market feed costs (e.g. crop residues, household wastes). Table 2. Breakdown of benefits and costs for economic analysis of small ruminants. Benefits Costs Primary, Increase milk yield Increase litter size Culls (ewes, rams, does, bucks) Goat meat and lamb or mutton Mohair and wool Hides Manure Horn and hooves for processing into feed supplements and other products Meat and milk by products Secondary Urine Weight increase (realised when the animal is sold) Weeding-grazing Farmers' preference Attractiveness of colour and appearance Pet Seed distribution Research Primary Cash costs Feed costs: hay, salt and minerals, concentrates grains and feeds for the young Non-feed costs: veterinary, medicine and drugs, vaccine Wages for hired labour Pasture rent Non-cash costs Family labour Depreciation facilities, (of equipment etc.) Other non-market feed costs Secondary Carry disease Destruction of crops and trees Trampling of land (which causes ecological imbalance) Foul odour Noise pollution Source: Adapted from Amir and Knipsheer (1987). 97 One of the major problems in performing a partial budget or an economic analysis is what value to assign to the inputs in a trial and output from the trial. Determining the field price of inputs and outputs can become a difficult exercise especially when dealing with non-market inputs or products. In this case, opportunity cost (which is the value of the resource or product in its best alternative use) should be employed. One example is manure. Because it provides a low cost source of fertiliser for crop, it can be given a value equivalent to the reduced use of chemical fertiliser. Another example is family labour. Assuming labour market is competitive, rural wages for hired labour can be used as a proxy. At this point it is important to note that labour (though it is the primary input most farmers make into agriculture), is one of the more difficult items to value. This is because for most farm household members off-farm employment is not really an option. Also for most household members (e.g. women, children, old men and young boys), there are no paying alternative to on-farm work (see Worman et al, 1 990 for more discussion on labour valuation). Finally, if the animal or the animal product (e.g. milk) is consumed at home, the opportunity cost is the amount the household would have to pay to purchase that meat or milk instead of using their own product. An empirical example This example is based on experimental trials conducted in the Ethiopian highlands. The purpose of the experiment was to assess the impact of drenching and feed supplementation on sheep productivity. The experiment was designed in response to nutritional and health problems (e.g. liverfluke, coenuris, diarrhoea, anthrax, lungworm) faced by sheep in the region. Data used in this study were from several sources. Biological and input price data were adapted from Ngategize and Brokken (1990, ILCA, Addis Ababa, unpublished data) and Ngategize et al (n.d.) which report on the stimulated effects of interventions on flock productivity and the economic of endoparasites control in the Ethiopian highlands, respectively. Animal price data (by head, and sex) were determined using a general linear model which describes the effect of animal characteristics on the price of sheep in the Ethiopian highlands (Andargachew, 1990). There were four treatments with 50 ewes in each treatment group. In treatment I (the control) animals were under normal grazing practice with no drenching or feed supplementation. Treatment II included drenching with anthelmintics. In Treatment III, animals received feed supplementation with wheat bran and noug cake in addition to normal grazing. Treatment IV included both drenching and feed supplementation. The feed supplements were composed of 300 g of wheat bran and 150 g of noug cake (Guizotia abyssinica) per ewe per day. The drenching scheme was composed of Ranide (a Fasc/o/a-specific anthelmintic) and Panacur (a broad spectrum anthelmintic which is effective against gastro intestinal nematodes and lungworms). Both were administered three times a 98 returns was EB 3.71 per ewe for a marginal rate of return on increased expenditure of 19%. The cost of adding drenching to treatment III in combination with supplemental feeding (treatment IV) was EB 3.58 per ewe while the increase in net returns relative to treatment III was EB 0.78 per ewe for a marginal rate of return of 22%. The marginal rate of return of treatment IV relative to treatment III was 19.4%. Table 5. Results of feeding and health management experiments for sheep, Debre Birhan, Ethiopia, 1989. Treatment1 Treatment Treatment Treatment 1 II III IV Gross returns (EB)2 91.41 97.72 121.03 125.40 Cash costs Feeds Noug cake - - 8.94 8.94 Wheat bran • - 6.98 6.98 Veterinary Ranide - 0.41 - 0.41 Panacur • 1.23 - 1.23 Labour - 0.54 - 0.54 Non-cash costs Capital cost (at 30% per annum) • 0.65 4.78 5.43 Breeding stock 12.35 10.98 12.7 13.46 depreciation charges Total cost that vary 12.35 13.81 33.41 36.99 Net return 79.06 83.92 87.62 88.41 Net return over control - 4.86 8.56 5.35 Marginal rate of return (%) - 334.00 19.00 22.00 1. Treatment I denotes normal grazing; treatment II represents drenching with anthelmintics; treatment III represents feed supplementation; and treatment IV includes both drenching and feed supplementation. 2. See Table 4. Given the high cost of capital, treatments with supplemental feeding and supplemental feeding with drenching cannot be recommended. However, drenching alone yields a very high marginal rate of return under the experimental conditions (334%) and can be suggested for on-farm testing. 101 Concluding comments Economics constraints and opportunities for small ruminant systems in sub-Saharan Africa must be understood as a basis for developing interventions. For example, the economic benefits of innovations aimed at reducing reproductive wastage and improved health, nutrition and management must be determined as a basis for recommending such to farmers. The problem is that most national agricultural research systems face an acute shortage of livestock economists and biological scientists lack the skills to conduct an economic evaluation of the results of their work. In order to properly evaluate alternative small ruminant interventions, livestock specialists will need to acquaint themselves with basic economic methods. In this paper, the partial budgeting analysis approach is presented. It is a simple but powerful approach which consists of organising experimental data and information about costs and benefits of various alternative technologies. Partial budgeting also allows the researcher to carry out marginal analysis which is needed for a correct evaluation of alternative interventions. Marginal analysis is important because although the calculation of net benefits accounts for the costs that vary, a technology may not be attractive to farmers because of the extra cost involved. In performing partial budgeting the first step is the identification of costs and benefits. This requires: (1) proper quantification of the production parameters; (2) proper elicitation of the inputs used, and outputs produced; and (3) proper recording of the (farm-gate) prices of outputs and inputs. The second step is to convert the identified quantities into costs and returns. All other non-cash costs (e.g. family labour, capital cost, and depreciation charges) must be properly valued. Non-market inputs and costs should be valued at their opportunity cost. In the case of small ruminants four products are generally considered important: they include (1) reproductive capacity of animals, (2) milk yield; (3) weight gain; (4) meat yield; and (5) manure. On the input side, cash and non-cash costs must be determined. Cash costs include feed and non-feed costs, wages for hired labour. Non-cash costs include family labour, capital costs, depreciation charges and other non-market feed costs. Using partial budgeting, an example is provided which compares the economic potential of a sheep breed under alternative health (endoparasitic control) and feeding systems. There were four treatments including: (I) grazing on natural pasture (the control); (II) drenching; (III) feed supplementation; and (IV) feed supplementation and drenching. Despite the fact that treatments (III) and (IV) yielded higher liveweight animal productivity per ewe than drenching alone (treatment II), economic analysis using marginal analysis has shown that endoparasites control with anthelmintics was economically more attractive. This result clearly highlights the importance of economic analysis in the process of technology evaluation. However, it is important to note that before doing economic analysis, the researcher must properly assess the experimental data 102 to verify that the observed response makes sense from an animal science standpoint. Before concluding, a few words are worth mentioning. Although the partial budget may indicate that the new technology is "better" than the traditional, it will not show that both technologies produce a loss. Also, a partial budget (though it is easy to interpret) is rarely presented with a statement of the farmer's objectives, the farmer's resource base, and important non-cash consideration. For example, the partial budget does not tell us if labour is available to the farmer to feed the animals. Partial budgets ignore the substitutability of inputs and how they are allocated based on fixed endowments and the implicit prices of the resources. Although they yield important insights into the economic attractiveness of a new technology versus a traditional one, result of partial budgets should therefore be treated with caution. References Amir P and Knipscheer H C. 1 987. A conceptual framework for the economic analysis of on-farm trials with small ruminants. In: Devendra C (ed), Small ruminant production systems in South and Southeast Asia. Proceedings of a workshop held in Bagor, Indonesia, 6-10 October 1986. IDRC-256e. IDRC (International Development Research Centre), Ottawa, Canada. pp. 380-391. Andargachew Kebede. 1990. Sheep marketing in the central highlands of Ethiopia. MSc thesis, Alemaya University of Agriculture, School of Graduate Studies, Alemaya, Ethiopia. 117 pp. Boehlje M D and Eidman V R. 1984. Farm management. John Wiley and Sons, New York, USA. CIMMYT (International Maize and Wheat Improvement Centre). 1988. From agronomic data to farmer recommendations: Economic training manual. Completely revised edition. CIMMYT, Mexico. Gryssels G. 1988. The role of livestock in the generation of smallholder farm income in two Vertisols areas of the central Ethiopian highlands. In: Jutzi S C, Haque I, Mclntire J and Stares J ES (eds), Management of Vertisols in sub-Saharan Africa. International Livestock Centre for Africa, Addis Ababa, Ethiopia. pp. 345-358. MSTAT-C. 1 988. A microcomputer program for the design, management and analysis of agronomic research experiments. Michigan State University, East Lansing, Michigan, USA. Mutsaers HJ W, Fisher N M, Vogel WO and Palada M C. 1986. Farming systems program: A field guide for on-farm research with special reference to improvement of cropping systems and technologies in West and central Africa. International Institute of Tropical Agriculture, Ibadan, Nigeria. 197 pp. Nagy J and Sanders J H. 1990. Agricultural technology development and dissemination within a farming systems perspective. Agricultural Systems 32:305-320. Ngategize P K, Zere Azaz and Brokken R F. (n.d.) Economics of endoparasit control in the Ethiopian highlands. In: Proceedings of the Eighth SC-CRSP scientific workshop, 7-8 March 1990, ILRAD, Nairobi, Kenya. Small Ruminant Collaborative Research Support Program (SR-CRSP), Nairobi Kenya, pp.340-347. 103 Peacock C P. 1987. Measures for assessing the productivity of sheep and goats. Agricultural Systems 23: 1 97-2 10. Sumberg J E and Cassaday K (eds). 1985. Sheep and goats in humid West Africa. Proceedings of the workshop on small ruminant production systems in the humid zone of West Africa. International Livestock Centre for Africa, Addis Ababa, Ethiopia. Wilson R T. 1988. Small ruminant production systems in tropical Africa. Small Ruminant Research 1:305-325. Winrock International. 1 983. Assessment of the current status and potential role of sheep and goats in developing countries. Winrock International, Morrilton, Arkansas, USA. Worman F, Norman 0 and Ware-Snyder J (eds). 1990. Agricultural technology improvement project. Farming Systems Research Handbook for Botswana ATIP RP3. Department of Agricultural Research, Ministry of Agriculture, Botswana. 104 Comparison of consumer attitude towards and acceptance of goat, sheep and cow milk in Malawi J.W.Banda University of Malawi Banda College of Agriculture Lilongwe, Malawi Abstract Two sets of questionnaires were administered to 206 individuals at three locations in Salima District, Central Malawi, in order to evaluate and compare consumer attitude and acceptance of goat and sheep milk relative to identically prepared cow's milk. One hundred and seventy-two goat farmers around Bunda College, Lilongwe, Central Malawi, were also included in the study to assess their general attitude towards the consumption of goat milk. About 38% and 4% of the respondents in Salima had tasted or consumed goat and sheep milk, respectively, compared to 24% for goat milk at Lilongwe. Unavailability was the major factor (P<0.001) influencing the attitudes of the people towards consumption of goat and sheep milk in Salima followed by the fact that it is not traditional to milk goats and sheep and consume their milk, a reason which was major (P<0.001) in Lilongwe. Strong flavour and taste was a tertiary factor in the study. Respondents failed to associate coded milk samples with actual sources. Mean taste scores (based on a 5-point scale of1= like very much to 5 = dislike very much) for goat, sheep and cow's milk were 1.81, 2.12 and 2. 77, respectively. There were no significant differences among milk sources in these scores. The results indicate that goat and sheep milk is just as acceptable as cow's milk. 105 Composition, production et acceptation du lait de caprins et d'ovins non laitiers au Malawi J.W. Banda Résumé 206 personnes de trois lieux du District de Salima, dans le Centre du Malawi, ont été enquêtées pour évaluer et comparer l'attitude des consommateurs et l'acceptation du lait de chèvre et de brebis par rapport à celui de vache. 172 éleveurs de chèvres, proches du Collège de Bunda à Lilongwe ont été approchés pour évaluer leur attitude vis-à-vis du lait de chèvres. 38% des enquêtés à Salima avaient goûté ou consommé du lait de chèvre (4% du lait de brebis), contre 24% à Lilongwe. A Salima, le principal facteur influençant l'attitude vis-à-vis de la consommation de lait de chèvre et de brebis est sa non-disponibilité (P<0.001), suivi du fait que la traite des chèvres et des brebis n'est pas une pratique traditionnelle. Ce second facteur est le principal à Lilongwe (P<0.001). Les forts goût et odeur étaient le troisième facteur. Les personnes enquêtées n'ont pas pu identifier des échantillons de lait des trois espèces. La moyenne des tests organoleptiques (1: aime beaucoup; 5 déteste), s 'est située à 1,81 (lait de chèvre); 2, 12 (brebis) et 2, 77 (lait de vache), sans qu'ily ait de différences significative dans ces appréciations. Les laits de chèvre et de brebis apparaissent donc être aussi acceptables que le lait de vache. Introduction In the tropics, milk is obtained mostly from cattle. Small ruminants (goats and sheep), however, are an important source of milk in Asia, the Middle East, north Africa and the Caribbean where the contribution of goats and sheep to total milk production is estimated at 15-50% (Matthewman, 1985; Treacher, 1985). South of the Sahara, until recently, goats and sheep were not by tradition, milked for human consumption. FAO (1987) estimates that there are 250 and 300 million goats and sheep, respectively, in the tropics. About 60 and 30% of the world population of goats and sheep, respectively, are found in villages of the developing countries. Of the 1.6 million goats and 0.8 million sheep in Malawi, over 90% are owned by small-scale producers in rural areas (Malawi Government, n.d.). In 1987, dairy goats were introduced in Malawi as supplementary sources of milk. 106 Goat and sheep milk is said to have important advantages over cow, milk for human nutrition. For example, Jenness (1980), and Devendra and Burns (1983) reported that compared to cow's milk, goat and sheep milk has higher protein, energy and fat contents. It has adequate amino-acids content. The higher proportions of short-and medium-chain fatty acids are of greater significance for ease of digestion. Goat milk is an excellent source of calcium, phosphorus and chlorine. Although goat and sheep milk has excess potassium and chloride which cause acidosis, and is deficient in vitamins C and D and folic acid, this milk would be beneficial to children in early postweaning periods. If supplemented with iron and folic acid, infants and pregnant mothers would benefit from these milk sources. However, strong flavour and taste of goat and sheep milk, and issues connected with witchcraft and lower social status are said to prevent people from consuming goat and sheep milk (Manyenga, 1 987) despite the belief by others that goat milk has therapeutic properties. Manyenga (1987) further reported that other people believe goat udders are too small to be milked. These controversies and the introduction of dairy goats in Malawi raised fundamental questions on milking of goats and sheep and peoples' attitudes towards an acceptance of milk from these sources. Information on these questions would help evaluate the feasibility of goat milk production programmes not only in Malawi, but also in the whole of sub-Saharan Africa and other tropical countries where small ruminant milk production has recently been introduced. The objectives of this study were: (a) to assess peoples' attitudes towards goat and sheep milk and its consumption and the factors associated with these attitudes. (b) to compare the acceptance of goat and sheep milk relative to identically prepared cow's milk. Experimental Method Two areas, namely Salima and Lilongwe where dairy goat projects have been initiated, were chosen for these studies. In Salima, two sets of questionnaires were administered by a trained research assistant at Mangochi Enterprises (site S1), Chipoka-Malawi College of Distant Education Centre (site S2) and at Salima Secondary School (site S3) . The questionnaires used were written both in English and a local language. To capture information on milk consumption in general and the attitude of people towards goat and sheep milk, the first questionnaire was to be administered to as many individuals at each site as possible a day before the taste-test exercise was carried out. In the end, 31, 104 and 71 respondents were covered at sites S1, S2 and S3, respectively. Based on responses on the first day at each survey site, 28, 59 and 47 respondents at sites S1 , S2 and S3, respectively, were selected to participate on the taste-test day. 107 For the taste-test exercise, fresh milk from zebu cows, local and Boer x local crossbred goats and from local and Dorper x local crossbred sheep was used. The milk was obtained from a goat breeding centre managed by the Malawi-German Livestock Development Programme at Lifidzi in Salima. The milk was passed through a milk strainer. The milk was then heated in clean sauce pans, while preventing burning, to near-boiling point and later cooled to body temperature. One hundred millilitre containers with lids were used. These containers were labelled X, Y and Z. Using random numbers, goat milk was allocated to X, sheep milk to Y and cow's milk to Z. About 100 ml of milk was poured in each container. They were kept at 35-40°C by placing them in an insulated box. During the taste-test exercise, respondents were again told they were tasting cow's, sheep and goat milk but the actual identities of samples were disclosed only after the taste exercise was completed. Each respondent was given a second questionnaire together with three samples of milk. They were encouraged to ask questions regarding test procedures. Each respondent was asked to taste the three coded samples, one at a time in the order indicated on the questionnaire. Information required was the identity of the tester, identity of milk and reasons for this. They were also requested to score the taste for each milk sample on a 5-point scale as used by Boor et al (1987): 1. Like very much 2. Like 3. Neither like nor dislike 4. Dislike 5. Dislike very much. General comments about the milk were invited at the end of each questionnaire. In another similar survey undertaken to assess the consumption and acceptance of goat milk and evaluate the factors limiting goat milk production/consumption among goat farmers, 172 individuals were interviewed in villages around Bunda College, Lilongwe, Central Malawi. Information collected was similar to that collected using the first questionnaire in Salima area. Milk samples from goats and sheep were analysed and have been presented in a separate paper (Banda et al, 1990). The composition of zebu cow's milk (Kasowanjete, 1 982) is not much different from that of goats. Data collected were analysed using contingency tables to provide a measure of association or fitness of fit (Gill, 1986). Results Consumption of milk and general attitude towards goat and sheep milk About 85.9% of the respondents in Salima claimed to have tasted milk. However, only 78.2% claimed to drink milk on a regular basis, i.e., at least once per week 108 (Table 1). Of those who drink milk, over half (56.5%) claimed to drink milk every day and 27.3% drank milk once per week. Table 1 . Frequency of milk consumption among consumers in Salima. Site S1 S2 S3 Total % Daily (over 3 times/week) 13 63 15 91 44.2 3 times per week 2 8 7 17 8.3 2 times per week 4 2 3 9 4.4 Once per week 11 13 20 44 21.4 None 1 18 26 45 21.8 Total 31 104 71 206 100.0 Xz = 43.2 (P< 0.001). About 96% had tasted or consumed cow's milk compared to only 38% and 4% for goat and sheep milk, respectively. There were no significant (P>0.001) variations in responses among sites. Since most respondents had consumed cow's milk, reasons for not having consumed cow's milk were removed from the statistical analysis. Results on reasons for not having consumed goat and sheep milk are presented in Table 2. Table 2. Reasons for not consuming goat and sheep milk in Salima. Milk source Goat Sheep Total % Strong smell/taste 22 13 35 10.9 Allergic 6 4 10 3.1 Not available 64 118 182 56.9 Not traditional 17 36 53 16.6 Never heard of it 5 35 40 12.5 Total 114 206 320 100.0 X2 = 21.9 (P< 0.001). Generally it is quite clear that the unavailability of the milk (P<0.001) was the major reason why most respondents (56.9%) in Salima have never consumed milk from goats and sheep, followed by the fact that goat and sheep milk 109 consumption is not traditional. There was a significant (P<0.001) milk source difference in the ranking of reasons on non-consumption. With goat milk, the second most important reason (19.3% of consumers) was that goat milk had a very strong smell and taste whereas for sheep, the lack of tradition of sheep milk consumption and knowledge that sheep could be milked were both equally the second most important (17% of respondents) reasons for not consuming sheep milk. About 68.5% and 62.5% of the respondents showed willingness to taste goat and sheep milk, respectively. About 63.0% were willing to buy and consume goat milk, while the proportion for sheep fell to 56.6%. Minor, but nonetheless culturally important, reasons were also identified. With sheep milk, people of the Malawi lakeshore origin never want to consume sheep milk because of the religious taboo of not consuming mutton. They also believe that the intensity of sneezing in sheep makes their milk psychologically unhygienic. Of the 172 farmers interviewed in Lilongwe, central Malawi, only 24% had ever tasted goat milk. Of those who had never tasted or consumed goat milk, 91% said that it was not traditional to consume goat milk, 4.5% were allergic to it while 2.3% claimed that the small amounts obtained from goats did not warrant the consumption of their milk. However, over half (56.5) were willing to consume goat milk and over 90% were willing to produce and sell goat milk if dairy goats and a milk market were available. Comparison of consumer acceptance of goat, sheep and cow milk Data collected using the second questionnaire were analysed to initially find out whether respondents could associate the coded samples with the actual sources of milk. Results are shown in Table 3. Sample X was goat milk, sample Y was sheep milk and sample Z was cow's milk. It is quite evident (P <0.001 ) that sample X was associated more with cow's milk. Both samples Y and Z were associated more (P<0.001) with sheep milk. Table 3. Frequency of responses regarding sources of milk. Soijrce accord ing to respondent Milk code Cow Goat Sheep Total X (Goat) 61 51 23 135 Y (Sheep) 37 45 55 137 Z (Cow) 35 45 54 134 Total 133 141 132 406 X2 = 25.0 (P< 0.001). 110 Frequencies of taste-test scores for all the milk types are presented in Table 4. Irrespective ofthe source of milk, a significant (P<0.001) proportion (66.9%) gave scores of 1 and 2. There were significant interactions (P<0.001) with source of milk. Goat and sheep milk received scores of 1 and 2 from 71- 72% of the respondents while the proportion for cow's milk was only 59.2%. About 21-23% of the respondents gave scores of 4 and 5 for goat and sheep milk and that for cow's milk was 30.0%. Table 4. Frequency of taste-test scores for goat, sheep and cow's milk in Salima District. Milk source Taste score Goat Sheep Cow Total % 1 69 43 49 161 40.0 2 26 53 29 108 26.9 3 10 7 15 32 80 4 25 22 19 66 16.4 5 5 8 22 35 8.7 Total 135 133 134 402 100.0 X2 = 37.0 (P< 0.001). Taste-test results in the form of mean scores for goat, sheep and cow's milk are presented in Table 5. The mean scores for goat, sheep and cattle milk were 1 .81 , 2.12 and 2.77, respectively. There were no significant differences (P>0.001) among these scores between the sources and sites of the study. The poorest score was for cow's milk at site S1 . Table 5. Mean taste-test scores for goat, sheep and cow's milk at three sites in Salima. S1 S2 S3 Average No. people interviewed 31 104 71 No. people on taste-test day 28 59 47 Mean scores: Goat milk 1.08 2.83 1.51 1.81 Sheep milk 2.11 2.62 1.64 2.12 Cow's milk 4.00 1.83 2.47 2.77 At the three sites the respondents felt that goat and sheep milk was sweeter than cow's milk but that goat milk was sweetest. Those with more sensitive palates 111 claimed that goat and sheep milk was creamier in taste. About 49.1% of the respondents claimed that goat and sheep milk was better, but whether this was with respect to nutritional value or not was not clear in this study; 16.5% were for cow's milk, 15.4 were indifferent and the rest did not give comments. After this taste-test exercise, the proportion of respondents claiming that goat and sheep milk has a strong smell fell to 3.8%. Discussion In this study, about 78.2% of the respondents claimed to consume milk at least once a week and 56.5% claimed to consume milk everyday. About 21 .8% never consume milk. The major factor responsible for 21.8% of the respondents not consuming milk was their low incomes and or financial status. These results correspond favourably with those given by Banda and Phiri (1990) for the Lilongwe urban area in the Central Region of the country. Milk consumed was exclusively cow's milk and the regular consumption here may refer to consumption during breakfast or as snacks. About 38% and 4% of the respondents in Salima had never tasted or consumed goat and sheep milk, respectively, because of unavailability. The high proportion of respondents in Lilongwe (91%) claiming not to consume milk because it was not traditional to milk goats and consume their milk, could perhaps allow to infer that unavailability may be strongly associated with the non-milking of goats as a tradition (Boor et al, 1987). One underlying factor is the claim that the udder of sheep or goat was too small to produce enough milk for human consumption. Another factor is perhaps the termination of goat milking by the few who did after the introduction of dairy cows in the early 1 960s in Malawi. For cow's milk, on the other hand, Banda and Phiri (1 990) found that incomes and prices were the main factors limiting the choice and consumption of cow's milk and milk products. It seems, therefore, that in milk production projects, one must isolate the limiting factors for small ruminants from those for cattle. A general problem that has been shown to limit the consumption of goat and sheep milk is its strong flavour and taste as reported previously in Zimbabwe (Manyenga, 1987) and in Kenya (Boor et al, 1987). Goat milk need not have a strong smell and taste. The strong smell/taste of milk from small ruminants, e.g., from goats, might be due to unclean milking procedures, utensils and milk itself. In addition, Skjevdal (1979) reported that the content of short-and medium-chain fatty acids (especially C6-C10), potassium chloride and the presence of some cresols (ortho-, meta- and para-cresols) may be responsible for the strong flavour and taste of the milk from goats and not necessarily the presence of the buck. In the present study, the strong flavour/taste was a secondary factor for goat milk and was a very minor factor for sheep milk, reflecting perhaps, better milking procedures and handling of the milk. This is confirmed by the results of the 112 taste-test exercise which showed that respondents could not really associate the coded samples with actual sources of milk. In addition, no differences in taste scores were detected between goat and sheep milk on the one hand and cow's milk on the other. If anything could be said at all, it was obvious that the respondents showed the least preference for cow's milk. The claim, therefore, that goat and sheep milk has a strong flavour is based on unsubstantiated reports or purely an aesthetic bias. Although the presence of the bucks may not be responsible for the strong flavour in goats, it is nevertheless advisable to prevent contact between males and lactating females immediately before milking and hair from falling into the milk. The specific flavour of goat and sheep milk, if not caused by unhygienic conditions, may not be important for direct human consumption, but is desirable in cheese production. Although goat milk ranked highest and cow's milk lowest on a double blind taste-test exercise, the mean taste-test scores were not statistically different between the sources. The higher number of best scores of 1 and 2 for all the milk and the mean scores of less than 3 reflect high consumer opinion of all the milk types. Goat milk was the best in flavour and taste followed by sheep milk. Acceptability therefore, may not be a limiting factor to consumption of goat and sheep milk as was observed in Kenya (Boor et al, 1987). This tends to confirm the indication that unavailability of the milk may be the major overriding factor. Malawi local goats produce between 250 and 640 ml of milk per day using hand-milking (Mwenefumbo and Phoya, 1982; Mchiela, 1989; Bandaetal, 1990). This is not even enough for kids and lambs. To improve milk production to levels that will provide a surplus for human consumption, dairy goat breeds such as the Saanen, Alpine and Anglo-Nubian could be crossed with local breeds while selecting for slightly better producing local genotypes. The sweeter taste of goat and sheep milk as claimed by some respondents might be due to the 4.5-5.1% lactose content observed in the milk used in the study and the creaminess could have been due to the high (6.0-7.0%) fat content. The real effects due to these chemical constituents on the use of goat and sheep milk for infant feeding and as emergency milk sources in cases of breast failure should be studied. Since the composition of cow's milk was not determined, there is a need to do the survey again, but now backed by nutritional and medical studies in order to evaluate intolerance of consumers to the use of these three different milk types. It is concluded that unavailability was the major problem influencing the consumption of goat and sheep milk. Acceptability for these milk types was high. Strong flavour/taste was a tertiary factor in this study and of little or no consequence. However, in order to increase the scope and general validity of the results of this type of investigation, there is need to include other locations likely to be affected by the goat milk production research and development programmes. 113 Acknowledgements The author wishes to thank J M Heremah for translating the questionnaire into a local language and for conducting the surveys. The questionnaires were checked by H P Zerfas and F D Heremah. The author also expresses his thanks to O Superzana for assisting in data processing and to B S Banda for typing the paper. Further, animals and materials used were obtained from the Malawi-German Livestock Development Programme's Lifidzi Ranch. Financial support was obtained from the University of Malawi Research and Publications Committee and from the Contract Research Committee of the Ministry of Agriculture (Research), Malawi. References Banda J Wand Phiri C D. 1990. Investigations into the factors influencing the choice and consumption of milk and milk products in Malawi. Journal of Consumer Studies and Home Economics 14:123-131. Banda J W, Steinbach J and Zerfas H P. 1990. The composition and yield of milk from non-dairy goats and sheep in Malawi. Paper presented at the First Conference of the African Small Ruminant Research Network, Nairobi, 10-15 December, 1990. Boor K J, Brown D L and Fitzhugh H A. 1987. Western Kenya: The potential for goat milk production. World Animal Review 62:31-40. Devendra C and Burns, M. 1983. Goaf production in the tropics. Second edition. Commonwealth Agricultural Bureaux, Farnham Royal, Slough, UK. pp. 64-73. FAO (Food and Agriculture Organization of the United Nations). 1987. Production yearbook. Volume 40. FAO Statistics Series 51, FAO, Rome, Italy. Gill J L. 1986. Review of analysis of contingency tables with dichotomous responses. Tierzucht und Zuechtungsbiologie 103: 1-25. Jenness R. 1980. Composition and characteristics of goat milk. A review 1968-79. Journal of Dairy Science 63: 1 605-1 630. Kasowanjete MBB. 1982. Breeds of cattle and breeding programmes in Malawi. Paper presented at the FAO/SIDA Workshop on Breeding and Feeding of Cattle for Milk Production, December, 1981, Lilongwe, Malawi. Malawi Government, Department of National Statistics, (n.d.). National sample survey of agriculture, 1985186. Government Printer, Zomba, Malawi. Manyenga A. 1987. Acceptability of goat milk. Farming World, September 1987 p. 23. Matthewman R W. 1985. Milk production from goats. In: Smith A J (ed), Milk production in developing countries. International Conference held in Edinburgh, UK, 2-6 April 1984. Centre for Tropical Medicine, Edinburgh, UK. pp. 403-423. Mchiela A F. 1 989. Milk production and growth rates ofgoats maintained under free range, tethered and confined systems of management. MSc thesis, University of Malawi, Zomba, Malawi. Mwenefumbo A F and Phoya R K D. 1982. Composition and yield of Malawian local goat. Tropical Animal Production 7:71 . Skjevdal T. 1 979. Flavour of goat's milk. A review of studies on the sources of its variations. Livestock Production Science 6:397-405. Treacher T T. 1985. Dairy sheep production. In: Smith A J (ed), Milk production in developing countries. International conference held in Edinburgh, UK, 2-6 April 1984. Centre for Tropical Medicine, Edinburgh, UK. pp. 388-402.. 114 Panurge: système d'investigation en recherche sur les systèmes d'élevage traditionnel 0. Faugère, B. Faugère, Ch. Moulin et M. Ndiaye ISRA-IEMVT Laboratoire national de l'élevage et de recherches vétérinaires, B.P. 2057, Dakar-Hann (Sénégal) Résumé Une rapide description du "Système d'investigation pluridisciplinaire" mis en oeuvre au Sénégal permet de montrer comment s'articulent les contributions des diverses disciplines des sciences de l'élevage autour d'un objet d'étude commun: l'animal. L'exemple de l'investigation en matière d'affections respiratoires des petits ruminants illustre une démarche progressive qui comporte: 1. Identification et dimensionnement du problème: taux de mortalité par pneumopathies suivant les régions. 2. Protocole spécifique ciblé sur le problème identifié pourpréciser les éléments déjà connus: examen clinique systématisé de la sphère respiratoire à chaque visite dans le but de préciser quelles classes d'âge sont atteintes, à quelle époque, avec quelle incidence. 3. Protocole spécifique pour approfondir la connaissance de l'origine du problème: suivi sérologique de cohortes de jeunes animaux pour préciser l'étiologie des affections rencontrées (mise en relation avec les examens cliniques). 4. Protocole d'étude de l'environnement pour identifier les facteurs liés à ce problème: conditions de logement, d'alimentation, etc., liées aux affections respiratoires. 5. Mise en relation des informations des quatre points précédents avec les données de performances de reproduction et de croissance pour donner une estimation complète du problème. 115 6. Formulation de propositions d'amelioration issues des investigations precedentes a tester sur le terrain. 7. Experimentations en milieu villageois des propositions d'amélioration pour en evaluer le cout-benefice. 8. Formulation de recommandations aux agents du développement. Cette demarche conduit les chercheurs a utiliser un nombre croissant d'informations issues des protocoles de recherche des autres chercheurs, ce qui est rendu possible par le fait que le materiel de travail est commun et qu'il est etudié par tous a l'echelle individuelle. Panurge: Survey system used in research activities on traditional livestock production systems in Africa O. Faugere, B. Faugere, Ch. Moulin and M. Ndiaye Abstract A brief description of the "Multidisciplinary survey system ", which has been introduced in Senegal, outlines how the various disciplines in livestock production science have linked in the study of a common subject: the animal. The example of the investigation of respiratory diseases in small ruminants illustrates a progression of steps comprising: 1. Identification and localisation of the problem: mortality rate by disease and region 2. Study ofthe problem identified to determine the extent ofthe known components: Clinical examination of the respiratory tract during each visit to determine and identify the age-classes affected, when they are infected and the incidence of the disease 3. Study of the causal agents: Serological survey of groups ofyoung animals to establish the aetiology of the disease 4. Environmental study to identify other factors associated with respiratory diseases: housing, feeding practices etc 5. Relation of data on the four points mentioned above to data on reproduction and growth performances to give a complete evaluation of the problem 116 6. Development of improvements, based on earlier surveys, to be tested in the field 7. Village studies to determine the costs and benefits of the proposed improvements 8. Formulation of recommendations for development agents. This process led the researchers to use a growing number of results arising from research protocols of other researchers, which was made possible by the fact that material studied was common to all studies. Introduction Un "système d'investigation pluridisciplinaire" est mis en oeuvre au Sénégal dans le cadre du programme de recherches "Pathologie et productivité des petits ruminants" (programme PPR) réalisé par l'Institut sénégalais de recherches agricoles (ISRA) et l'Institut d'élevage et de médecine vétérinaire des pays tropicaux (IEMVT). Après une rapide description du système d'investigation, nous verrons, à partir de l'exemple des affections respiratoires des petits ruminants, comment on peut mener une démarche de recherche progressive, impliquant un nombre croissant de chercheurs de différentes disciplines des sciences de l'élevage, élaborant chacun des protocoles d'investigation centrés sur un même objet. Présentation du système d'investigation L'architecture du système est basée sur un suivi démographique et le contrôle des performances individuelles. Les produits qui en sont attendus sont: • une parfaite connaissance de la population objet d'étude (effectifs, structure, mouvements); • une évaluation précise et fiable des performances des animaux; • une identification rigoureuse des facteurs de variation des performances, notamment ceux d'ordre pathologique et ceux liés aux pratiques d'élevage. Les animaux suivis sont identifiés individuellement à l'aide d'une boucle auriculaire numérotée. Les troupeaux de petits ruminants sont visités chaque quinzaine par des observateurs ayant une bonne formation de base en zootechnie et en pathologie. Trois éléments sont considérés comme fondamentaux et représentent le noyau de base du système: un suivi démographique, un suivi pondéral et un suivi 117 sanitaire. Ce noyau, support de travail de tous les chercheurs impliqués, doit être maintenu plusieurs années sans rupture de suivi. Un suivi démographique L'unité d'observation est l'animal, mais l'unité de travail est le troupeau de concession: c'est le premier niveau de regroupement observable sur le terrain; il correspond aux animaux séjournant la nuit dans un même enclos. L'objet du suivi est de saisir, à la frontière de ces troupeaux, tous les flux d'entrée (naissance, immigration, achat, troc, etc.) et de sortie (mort, exploitation, vente, don, etc.). Les performances de reproduction, les caractéristiques de l'exploitation et de la mortalité des animaux sont ainsi appréhendés à l'échelle de l'individu. Un suivi pondéral Il permet de caractériser la carrière de croissance de l'individu par l'évaluation des poids à âge type et des gains moyens quotidiens. Un suivi sanitaire Son objet est d'identifier les causes de mortalité et de décrire l'évolution de l'état sanitaire de l'animal au cours de son existence. Il apporte ainsi un élément de caractérisation de la carrière d'un individu au même titre que sa reproduction ou sa croissance. Investigations complémentaires Parallèlement à ces trois suivis, il se développe des modules d'investigation considérés comme complémentaires (figure 1) et destinés à renseigner la base de données pluridisciplinaire, en ce qui concerne par exemple les pratiques d'élevage, l'environnement sanitaire, ou les parcours. Ces modules dits "périphériques" s'articulent autour du noyau de base et correspondent à autant de protocoles de durée limitée, mis en place pour répondre à des questions soulevées par les chercheurs thématiques. Enfin, il est possible de greffer sur cet ensemble des modules expérimentaux correspondant à des actions volontairement menées par l'équipe de recherche pour modifier le milieu d'étude, en introduisant un élément supposé améliorateur. Leurs effets sont ensuite évalués à travers la structure d'investigation. Il s'agit par exemple d'actions prophylactiques (vermifugation, vaccination, etc.) ou d'actions de complémentation alimentaire. 118 Figure 1 . Schéma du modèle d'investigation EXPERIMENTATION pratiques d'élevage environnement sanitaire parcours Ce système d'investigation est mis en place au niveau de différents sites géographiques au Sénégal, constituant ainsi un véritable réseau de suivi (figure 2). Cette notion de sites d'investigation multiples complète la présentation de l'architecture d'ensemble. Précisons encore que la gestion de l'information est nécessairement informatisée et qu'elle est réalisée à l'aide d'un progiciel élaboré pour les besoins de ce système, dénommé PANURGE. Ses caractéristiques principales sont: une utilisation conversationnelle, de très nombreuses procédures de validation de l'information, et l'organisation de l'information après un prétraitement de manière à la rendre accessible et directement utilisable. Investigations sur les affections respiratoires des petits ruminants Les affections respiratoires des petits ruminants, dominante pathologique au Sénégal, permettent d'illustrer l'articulation entre les modules d'investigation. Pour l'étude de ce thème, les modules impliqués sont: • ceux du noyau de base : les suivis démographique, sanitaire et pondéral; • un module de suivi clinique spécifique centré sur la sphère respiratoire; 119 Figure 2. Implantation des suivis au Sénégal Petits ruminants /*i St Louis Ziguinchor • un module de suivi sérologique; • un module d'étude de l'environnement et des pratiques d'élevage, en fait impliqué dans un grand nombre de thèmes de recherche. La démarche retenue procède en cinq étapes : • identifier, décrire et dimensionner cette entité morbide (épidémiologie descriptive); • rechercher un indicateur simple à observer, fiable, permettant de procéder à des investigations exhaustives; • analyser la pathogénie de ces affections en mettant en oeuvre des protocoles d'étude ponctuels (épidémiologie analytique); • évaluer l'impact de ces affections en termes de production en reliant cet ensemble d'informations aux données concernant les performances (épidémiologie synthétique); • expérimenter des méthodes de lutte. 120 La première phase, descriptive, est abordée grâce aux suivis démographique et sanitaire et permet de préciser l'évolution dans l'espace, l'évolution dans le temps, les classes d'animaux atteints, etc. Il est ainsi possible d'établir (tableau 1) les points suivants: • les mortalités à la suite d'affections respiratoires intéressent les ovins et les caprins dans des proportions comparables; • cette pathologie est plus meurtrière dans le centre et le sud du pays que dans le nord; • les mortalités sont surtout importantes entre novembre et avril, durant la saison sèche fraîche; • les mortalités interviennent surtout dans le jeune âge, dans les mois qui suivent le sevrage (4 à 8 mois d'âge). Tableau 1 . Part des mortalités liées à des affections respiratoires parmi les mortalités d'origine pathologique Ovins Caprins Nord 10% (13%) 41% (19%) 25% (27%) 7% (16%) 49% (23%) 40% (29%) Centre Sud (Entre parenthèses = quotient de mortalité globale toutes circonstances confondues) La seconde phase, l'identification d'un indicateur clinique, est abordée en réalisant un suivi clinique spécifique centré sur la sphère respiratoire. En pratique, il s'agit d'examiner chacun des animaux, à chaque visite bimensuelle, pour relever la présence ou l'absence de jetage nasal, de toux, de dyspnée et de diarrhée, cette dernière étant fréquemment associée aux symptômes respiratoires. Les premiers résultats obtenus nous portent à penser que la prévalence moyenne du jetage dans un troupeau pourrait être l'un de ces indicateurs. Elle est en effet corrélée, comme le montre la figure 3, au quotient de mortalité à la suite d'affection respiratoire. La troisième phase, l'étude de la pathogénie, est abordée par deux voies: l'étiologie et l'écopathologie. Une investigation étiologique Un échantillon de 600 jeunes animaux nés sur une période de 1 8 mois a fait l'objet d'un suivi sérologique en vue d'établir la cinétique de l'infection par les principaux agents infectieux des pneumopathies. Les prélèvements ont été effectués à intervalle de 45 jours, de l'âge de 15 jours à l'âge de 1 an. Lors du premier prélèvement sur le jeune, la mère faisait aussi l'objet d'un prélèvement. 121 Figure 3. Prévalence du jetage et quotient de mortalité dû aux pneumopathies dans les troupeaux suivis quotient mortalité pneumopathie 10% 5% 0% prévalence jetage 10% 30% 50% 70% L'analyse de ces suivis permettra de : • calculer la prévalence et l'incidence sérologique des différents agents infectieux; • observer leurs associations et leur ordre d'apparition; • suivre la cinétique des anticorps animal par animal, et d'interpréter la réponse sérologique du jeune en tenant compte de celle de la mère; • suivre l'évolution dans le troupeau et de noter les variations saisonnières. En reliant ces informations à celles du suivi clinique spécifique, il sera possible d'établir la liaison (ou au contraire l'absence de liaison) entre une conversion sérologique (une infection) et une expression clinique (une affection) ou une issue mortelle. Il serait bien entendu intéressant de compléter cette investigation par l'analyse de prélèvements biologiques réalisés sur animaux malades ou sur cadavres. Ce protocole, complexe à mettre en oeuvre compte tenu des contraintes de terrain, n'est pas encore réalisé au Sénégal. 122 Une investigation écopathologique Des enquêtes répétées dans les troupeaux suivis permettent d'établir pour chaque éleveur les pratiques d'élevage qu'il met en oeuvre: logement, conduite sur parcours, complémentation, traite, etc. En reliant ces informations aux informations issues du suivi clinique spécifique et du suivi démographique (mortalité) il est possible d'identifier des facteurs de risque des affections respiratoires. Il a par exemple été possible, dans la partie sud du pays, d'établir une liaison entre l'état de propreté du logement et sa protection des intempéries d'une part, et la prévalence du jetage dans un troupeau d'autre part. La quatrième phase, l'évaluation de l'impact, pourra être abordée lorsque nous disposerons de l'ensemble des informations précédentes. Elles seront reliées aux données de performances de reproduction (suivi démographique) et de croissance (suivi pondéral) pour établir avec précision le coût réel (affections bénignes comprises) de ces affections en termes de mortalité et en termes de baisse de performances. La cinquième phase, expérimentale, pourra être abordée: • pour établir la liaison de cause à effet entre un facteur de risque et ces affections. Par exemple, l'amélioration des conditions de logement dans les troupeaux dans lesquels on enregistre une forte prévalence du jetage, conduit-elle à une réduction sensible de l'incidence des affections respiratoires et une amélioration des performances zootechniques? • pour établir le coût-bénéfice et la faisabilité d'actions de lutte contre les affections respiratoires telles que: • la vaccination contre certains agents pathogènes; • la modification de certaines pratiques d'élevage impliquées dans la pathogénie de ces affections. Conclusion Nous ne pouvons pas conclure cet exposé qui souligne les avantages d'une démarche de ce type sans en souligner également les contraintes. Les plus frappantes sont : • la nécessité de disposer d'un financement suffisant pour mener un programme de recherche sur un terme de plusieurs années; • un partage du temps entre les activités de recherche et les activités de gestion et de maintenance du système, indispensables au fonctionnement du programme; 123 • la forme, la diversité et l'importance du nombre de données collectées, qu'il faut articuler entre elles, posent des problèmes méthodologiques nouveaux de gestion et d'analyse de l'information. On ne peut pas non plus conclure sans souligner la dynamique originale et les pratiques de recherche nouvelles que génère ce type de système, ne serait-ce qu'en amenant les chercheurs de différentes disciplines à concevoir et à conduire conjointement des protocoles sur un même matériel animal. La succession des phases de recherche présentées dans la deuxième section de cet article peut subir quelques modifications. Il est ainsi possible, lorsque l'on est astreint à des résultats rapides, et c'est toujours le cas, de mettre en place des expérimentations prophylactiques avant que les résultats des phases précédentes soient acquis. C'est ce que nous avons réalisé au Sénégal et qui fait l'objet d'une autre communication à cette conférence 124 Synthèse des résultats de suivi et des essais zoosanitaires chez les petits ruminants effectués par le volet recherche du Projet sectoriel d'élevage au Mali T. Tangara, M. Doumbia, S.S.B. Ba, T. Kibe et CF.M. Diallo Centre de recherche zootechnique de Sotiba B.P. 262, Bamako (Mali) Résumé Différentes études ont eupour objectif de cerner les contraintes liées à l'élevage dans les petites exploitations agropastorales de la zone semi-aride de Banamba. Le suivi a porté sur un effectif moyen de 1 200 petits ruminants (700 caprins et 500 ovins) appartenant à 35 unités de production agricole dans 8 villages autour de Banamba. La caractérisation des systèmes de production a montré que 95% des animaux sont soumis au régime de la libre pâture pendant la saison sèche. L'apport d'une complémentation alimentaire se fait peu ou presque pas. Les petits ruminants ne font pas l'objet de préoccupations sanitaires de la part des paysans. Le taux d'exploitation du cheptel reste inférieur à 20%. Les traitements sanitaires testés ont considérablement diminué le taux de mortalité: de 14,2% à 9,19% chez les ovins et de 18,6% à 7,69% chez les caprins, et amélioré le taux d'exploitation de 10%. Des observations ont été effectuées pour déterminer les performances de 101 ovins mâles engraissés en stabulation de 1987 à 1989 à l'approche de la fête de Tabaski. La durée moyenne de l'engraissement et le gain moyen quotidien ont été respectivement de 70 jours et de 67 grammes. Parmi les animaux engraissés, 37% ont été prélevés des troupeaux d'élevage personnels et 63% achetés sur les différents marchés. 125 Results of the Mali Livestock Sector Project research programme for small ruminant health monitoring and trials T. Tangara, M Doumbia, S.S.B. Ba, T. Kibe and C.F.M. Diallo Abstract Various studies have focussed on identifying constraints to livestock production in agropastoral smallholdings of the semi-arid zone of Banamba. A sample population of 1200 small ruminants (700 goats and 500 sheep) from 36 agricultural production units in eight villages around Banamba was monitored. Production system characterization shows that 95% of the animals are roaming freely during the dry season. Supplementary feeding is hardly, if at all, practised. Farmers are not concerned about the health of their small ruminants. Flock offtake is less than 20%. Tested treatments significantly reduced mortality: from 14.2% to 9.19% in sheep and from 18.6% to 7.69% in goats, and increased offtake by 10%. Between 1987 to 1989, the performance of 101 fattening males was recorded as the Tabaski feast was approaching. Mean fattening period and daily weight gain were 70 days and 67g respectively. 37% ofthe fattening animals were taken from individual flocks and 63% had been bought on the markets. Introduction Malgré l'importance socio-économique de l'élevage des petits ruminants dans la zone sahélienne du Mali, les connaissances le concernant restent fragmentaires dans de nombreux domaines. Parmi les contraintes à son amélioration, il faut citer les problèmes sanitaires qui restent quasi-entiers, les conditions d'élevage et d'alimentation mal perçues par les techniciens, les possibilités alimentaires qui demeurent incomplètement inventoriées et testées. Les qualités bouchères des races locales restent par ailleurs mal connues. Le Volet Recherche du Projet sectoriel de l'élevage 1988 au Mali a étudié dans la zone semi-aride de Banamba, où les services vétérinaires estimaient en 1988 l'effectif des petits ruminants à 300 000 têtes, les systèmes d'élevage des petits ruminants, et leur productivité en milieu rural. Matériel et méthodes Dans huit villages autour de Banamba, 35 unités de production agricole ont été choisies sur la base de leur niveau d'équipements pour constituer un échantillon 126 représentatif de la zone d'étude. Tous les animaux de ces exploitations ont été identifiés. Les entrées d'animaux dans les troupeaux (naissances, achats et animaux gardés pour autrui) et les sorties (morts, abattages, ventes, prêts, dons, sacrifices cérémoniels et animaux rendus aux propriétaires après la période de garde) ont été relevées chaque mois de juin 1987 à juillet 1988. L'impact de traitements sanitaires prophylactiques sur la productivité des petits ruminants a été étudié sur les 1 200 petits ruminants appartenant à ces 35 unités de production agricole, (700 caprins et 500 ovins). Parmi ces animaux une moyenne de 750 têtes ont fait l'objet de mesures de poids mensuelles pendant 12 mois. Les animaux ont été répartis entre 3 groupes de traitement par choix aléatoire pendant toute la durée du suivi. • TO : lot témoin ne recevant aucun soin vétérinaire • T1 : lot vacciné à l'antipasteurellique (juin et novembre), à l'antipeste e à l'antibactéridien (juin). • T2 : lot vacciné (aux 3 vaccins précités) et déparasité au Panacur ND ou à l'Exhelm II. ND. (en juin et en octobre). Le suivi de l'embouche ovine a concerné de 1 987 à 1 989 un effectif de 1 01 béliers, appartenant à 28 des 35 unités de production agricole, en observant l'alimentation, la santé et les performances pondérales. Résultats Systèmes de production des petits ruminants Taille ef composition des troupeaux ovins et caprins La taille moyenne des troupeaux est de 13,1 ovins et 21 caprins par unité de production agricole (tableau 1). Les caprins appartiennent généralement aux femmes. Tableau 1 . Composition moyenne des troupeaux ovins et caprins Sexe Femelles Mâle! Classe d' (en mois] âge 0-18 18-24 >24 TOTAL 0-18 18-24 >24 TOTAL Ovins I 3.4 1,9 5,1 10,4 1,9 0,5 0.3 2.7 Caprins 6,2 2,8 7,8 16,8 3,5 0,6 0,1 4,2 127 Parmi la population ovine la race Djallonké prédomine (près de 90%). Les autres races ovines élevées en faible proportion sont: le Maure, le Balibali et le Peulh ou Toronké. Les caprins appartiennent à la race naine de Guinée originaire du Fouta-Djallon. Mode de conduite des troupeaux Le pâturage naturel constitue l'essentiel de l'alimentation des troupeaux de petits ruminants. Plus de 95% des animaux sont soumis au régime de la libre pâture pendant la saison sèche et conduits sous surveillance pendant la saison des cultures. Pour diminuer les problèmes que pose l'insuffisance de la main d'oeuvre, les troupeaux sont regroupés et conduits surtout par des enfants. Sur un échantillon de 1 200 petits ruminants, 9% des pertes enregistrées étaient dues au manque de surveillance régulière. La complémentation alimentaire est distribuée seulement aux animaux malades, lactants et ceux soumis à un régime d'embouche. Les aliments généralement offert sont: son de mil, aliment Huicoma, fanes de niébé ou arachide, paille de brousse, feuilles et fruits d'arbres. La castration (quand elle est pratiquée) se fait à partir de 18 mois chez les ovins et les caprins. Les possibilités de contrôle des saillies sont limitées et les paysans ne semblent pas avoir d'objectifs précis dans ce domaine. Sanfé des petits ruminants Les paysans apportent très peu de soins sanitaires, pour lesquels moins de 1% du coût total des dépenses liées à l'élevage est alloué. En fin d'année 1986 la pasteurellose a frappé près de 80% de troupeaux ovins et caprins de l'échantillon d'étude avec un taux de mortalité de 30 à 40%. Des cas de listériose et d'echtyma contagieux ont été confirmés également au laboratoire. Sur des coproscopies effectuées, les trychostrongyloses ont été décelées sur près de 47% des animaux, les coccides sur 18% et les douves sur 5%. Outre ces pathologies la part d'autres infections telles la peste des petits ruminants le charbon bactéridien a été reconnue, suggérant l'expérimentation présentée ci-dessous. La consommation excessive du Zornia glochidiata au début de la saison pluvieuse a aussi été reconnue comme une cause de mortalité chez les ovins. Evolution pondérale De janvier à juin, le gain moyen quotidien enregistré pour les animaux de moins de 1 8 mois était de 52, 1 g/j et pour les mâles de plus de 1 8 mois 37,4 g/j. Jusqu'à l'âge de 18 mois, les ovins gagnaient en moyenne 60,0 g/j tandis que les caprins ne gagnaient que 39,3 g/j. 128 Exploitation des troupeaux petits ruminants Le taux global d'exploitation calculé sur des prélèvements d'animaux dans le troupeau a été inférieur à 20%. Le poids moyen à la vente des moutons et des chèvres était respectivement de 29,3 et 19,9 kg. Pendant la fête de Tabaski le nombre de moutons sur le marché est presque doublé et le prix unitaire à la vente des ovins mâles augmente de près de 50%. Une enquête menée à la fin de l'hivernage 1987 a montré que 28% du lait prélevé dans la zone provient des petits ruminants. Sur ce pourcentage 24% sont imputables aux caprins et 14% aux ovins. La part des petits ruminants dans le revenu en espèces des petits exploitants a été de 32,9%. Impact de traitements prophylactiques (Ba B. Souleymane, 1990). Des résultats préliminaires sur l'évolution pondérale, l'exploitation et la mortalité ont été obtenus. Evolution pondérale des animaux La vaccination et le déparasitage permettent un gain pondéral plus rapide dans les deux espèces (P<0,05) et plus notablement chez les ovins qui croissent plus que les caprins (tableau 2). Tableau 2. Effet de la vaccination et du déparasitage sur le gain pondéral (gljour) des petits ruminants suivis en milieu villageois. (Juin 1987-juillet 1988) Traitement Oins Caprins 26,1 28,6 28,4 27,7 n = 289 Mortalité et exploitation Au cours de l'étude, le nombre de sorties s'est chiffré à 35% de la population originelle, se décomposant en 66% pour les prélèvements (vente, abattage, cadeaux) et 34% pour les taux de mortalité. Les modifications de la structure des sorties imputables aux soins vétérinaires dans cette expérience sont résumés dans le tableau 3. On y voit que les traitements diminuent l'influence de la mortalité au profit des prélèvements, en notant une acuité supérieure du phénomène pour les ovins. Aucun traitement 42,5 Vaccination 43,5 Vaccination + déparasitage interne 67,6 Moyenne 51,2 n = 359 129 Tableau 3. Effet de la vaccination et du traitement antiparasitaires sur les taux de prélèvement et de mortalité chez les différentes espèces animales Ovins Caprins Prélèvement % Mortalité Prélèvement Mortalité Traitement % % % Aucun traitement 50 50 65 35 Vaccination 65 35 69 31 Vaccination + traitement antiparasitaire interne 77 23 69 31 Il fut également noté que la mortalité dans les troupeaux sous abri permanent a considérablement diminué. Ce taux a chuté de 6% chez les ovins et les caprins (tous traitements confondus). Embouche Type animal Les animaux les plus utilisés au cours des trois campagnes de suivi de l'embouche étaient de race Djallonké (85%) de race Bali-bali (8%); métissages Djallonké/Maure ou Djallonké/Bali-bali (7%). L'âge moyen de l'ensemble des animaux était de 21 mois avec des variations individuelles de 12 à 36 mois. Celui des ovins de race Djallonké était généralement compris entre 18 et 30 mois. Les ovins engraissés étaient en majorité des mâles entiers. Alimentation Les principaux types d'aliments offerts sont: paille de brousse, feuilles etbranchettesd'arbres (Tamarindus indica, Ficus gnaphalocarpa, Pterocarpus erinaccus), graines de mil, son de mil, fanes de niébé, fanes d'arachide, fanes de dolique, gousses de Piliostigma reticula, aliment Huicoma. Gain pondéral Le poids moyen de départ a été de 33,25 kg pour l'ensemble des 101 têtes. La moyenne de poids de démarrage des ovins mâles de race Djallonké (race la plus représentée dans la zone de Banamba) engraissés au cours des trois campagnes a été de 29,12 kg. 130 Le poids à la fin de l'embouche a été de 37,80 kg toutes races confondues. Le gain moyen quotidien était de 65 g pour une durée moyenne de 70 jours (toutes races comprises). Résultats économiques de l'embouche ovine Les marchés à bétail locaux constituent des sources d'approvisionnement pour les ovins. Les résultats du suivi de quatre marchés dans la zone de Banamba ont révélé qu'il y a une augmentation de l'offre globale en petits ruminants au cours de la période de janvier à juin. Au cours de cette période le nombre total d'ovins augmente de 1,5% par mois et celui des caprins d'environ 8% par mois (Tangara et Sissoko, 1990). Quant aux sources de financement, certains paysans procèdent à la vente d'autres animaux (vieilles femelles etc.) pour acheter les ovins mâles d'embouche et d'autres investissements des produits d'exode ou prélèvent sur leur troupeaux d'élevage. La commercialisation des animaux embouchés est beaucoup liée à la fête de Tabaski. Les paysans vendaient directement leurs animaux sur les marchés locaux. Cependant certains d'entre eux acheminent leurs animaux jusqu'au niveau des centres urbains tels que Bamako où il existe une forte demande. Parmi les animaux engraissés au cours des campagnes de suivi 37% ont été prélevés des troupeaux d'élevage personnels et 63% achetés sur les différents marchés. Le prix d'achat moyen des ovins mâles était de 9 955 F CFA et le prix de vente moyen a été de 20 598 F CFA. L'analyse de la structure des coûts d'embouche ovine révèle que les coûts d'alimentation représente 80%. Le coût d'intervention de la main-d'oeuvre familiale dans les travaux (tels que la distribution des aliments, l'abreuvement, le nettoyage des enclos etc.) représente 9% des coûts, les frais d'achat et de commercialisation représentent 11% des coûts d'embouche. Bibliographie Souleymane Ba B. 1990. Etude sur la gestion et la productivité des petits ruminants dans la zone Semi-aride du Mali: Impact de certains traitements sanitaires prophylactiques. Rapport technique. Tangara T. et Sissoko K. 1990. Suivi de l'embouche ovine paysanne dans la zone semi-aride de Banamba. Rapport technique (Commission Technique Spécialisée des Productions Animales: Session 1990). Volet Recherche P. SE. 1988. Investigation zoo-sanitaires des activités d'elevage dans la zone semi-aride du Mali. Rapport technique (Commission Technique Spécialisée des Productions Animales: Session mars 1988). 131 Black Head Ogaden sheep under traditional management practices in south-eastern Ethiopia Girma Adugna Faculty of Veterinary Medicine, Addis Ababa University P O Box 34, Debre Zeit, Ethiopia. Abstract An attempt was made to study the traditional Black Head Ogaden sheep management practices by the nomadic and semi-nomadic people in the Ogaden plateau of south-eastern Ethiopia. The methods employed were a predesigned questionnaire, observations and sampling of relevant elements. The management aspect was found to be rather specialised in that breeding is controlled to achieve both selection and synchronisation of lambing season with rainy season (which starts around the month ofApril) . Lamb crop is once per year and most of the ewes (96%) give birth to a single lamb. Watering interval is 4-10 days based on water scarcity in a locality. Supplementary feeding and pasture conservation are not practised. As management, nutrition and disease in the Black Ogaden sheep closely interact, an approach encompassing such elements to see into the problems of the pastoralists would be much rewarding. Methodes d'elevage traditionnel du mouton Somali a tete noire dans le sud-est de I'Ethiopie Girma Adugna Resume Ceffe etude explore les methodes d'élevage traditionnel du mouton Somali a tete noire pratiquees par les populations nomades et 133 semi-nomades du plateau de l'Ogaden dans l'est de l'Ethiopie. Elle repose sur un questionnaire, des observations effectuées sur le terrain et l'échantillonnage de paramètres pertinents. Les pratiques de gestion sont relativement précises: l'accouplement est contrôlé de manière à opérer une sélection et à synchroniser les naissances avec la saison des pluies commençant en avril. Les brebis n'ontagnelé qu'une fois par an et la majorité d'entre elles (96%) n'ont donné naissance qu'à un agneau à la fois. Les animaux ont été abreuvés tous les 4 à 10 jours en fonction des disponibilités en eau de la région considérée. Ni la complémentation alimentaire ni aucune technique de conservation des pâturages n'a été observée. Etant donné les interactions entre la gestion, l'alimentation et la santé, tous ces éléments devraient entrer en ligne de compte dans la recherche de solutions aux problèmes des éleveurs traditionnels du mouton Somali à tête noire. Introduction The Black Head Ogaden (BHO) sheep comprise most of the sheep population of the Ogaden region and they stand in number second to camels. This sheep breed/type forms the greater proportion of the small ruminant population in the region and contributes a great deal to the national economy as it has special merits in the Middle East and Arabian countries. It is, nevertheless, raised by the pastoralists under harsh environmental conditions, with seasonal under-nourishment or malnourishment, long watering intervals, long walks, heat stress and little or no protection against diseases. In this paper, the methods of feeding, watering, and housing and control of breeding to synchronise the lambing season with the rainy season are discussed. Flock structures, herding, castration, marking (identification) and miscellaneous operations are presented. Major diseases and ailments as known to the pastoralists and the inherent drawbacks of disease control measures generally practised in the region as well as the veterinary extension and services to the region are examined. Materials and Methods The sheep type The BHO sheep are fat-rumped, have a black head, white body and limbs and on the average weigh 30-35 kg at maturity (MOA, 1985). They are polled. 134 In Ethiopia, the BHO are found in the lowland areas of Hararghe, Bale and Sidamo Administrative Regions (Gonzalez, 1981, unpublished). They are also indigenous to Somalia and Kenya where they are known as the Black He="i Somalia (Osman, 1985). Data collection Information on the sheep management practices in the traditional sector of the region was collected through a predesigned questionnaire. Ninety-seven owners were interviewed from nine difierent localities in the region. Two different occasions, namely the beginning of the dry season (late October) and the last third of the short rainy season (mid-May), were chosen to conduct the interview as well as to watch the actual management practices. Results Feeding, watering and housing The feed resource for sheep is native pasture. Supplementary feeding and forage conservation are not practised. Nevertheless, some soil types at some specific points are said to be provided to the sheep at 3-6-month intervals. These soil types known locally as "Arabi", may contain some mineral elements. All animal species (cattle, camels, goats and sheep) graze in the same field. Grazing time is from 9 a.m. to 6 p.m. When the heat gets unbearable around noon, sheep collect under shades or simply congregate at a point and hide their heads under each other's body. During feed scarcity, which is usually in the dry season (October to March), the pastoralists migrate with their flocks to other localities where forage is available. The distance covered during such migrations may be in the range of 50 to 200 km. The site selection is largely dictated by the availability of water in the area. Water scarcity is the basic problem for the pastoralists and the frequency of watering is determined by the distance from water points and the availability of water which is scarce during the dry season. Table 1 shows watering frequencies for different localities. Delay can be made for as long as 10 days in such areas as Aware where water scarcity is greatest. The distance covered to the watering points can be as long as 20-40 km (an equivalent of a 6-hour journey, single trip) during periods of severe water scarcity. The pastoralists enclose their sheep and goats together at night. But where the number of goats is large, a separate pen is built for them. From the beginning of April to the end of October males are kept in a separate pen for the purpose of breeding control. 135 Table 1 . Watering frequencies for different localities. Place Elevation Intervals between watering (m above sea level) (days) 1650 4 1650 4-5 1470 7-5 na 6-7 na 10 na 7-8 Jijiga Lefe Isa Harshin Duria Aware Degahabur * not available. The enclosure can be built from wood or even stones. During the rainy season the enclosure may get muddy and another is built if the materials are available. Flock structure and herding Different age groups are identified by the pastoralists (Table 2). Likewise, sex group identification is practised. A female sheep ready to breed or is pregnant for the first time is termed as "Sebyeno", a breeding ewe is "Laha" and "Sumel (cka)" is a ram used for breeding purposes. Table 2. Age group identification and names. Age Name 0-4 months Mekel 5-12 months Berar 1 year Gujir 2 years Lebejir 3 years Sedhajir 4 years Afarjir A general overview of the flock structure is shown in Table 3. Males form only a smaller proportion of the flocks and they are castrated if intended to be sold later. One to four entire males per flock are used for breeding purposes. Only a few lambs were observed to be born between October and March, and these are termed as "Gera" (the bastards); they are "unwanted" because they were conceived during the separation of males from females due to faulty herding by shepherds. All age and sex groups are mixed and herded together after November. After the start of the lambing season (around the month of April) the sumel is herded separately in the field or it is tethered around homesteads. After a while, it is kept 136 with the lambs, at least for sometime, till the next mating season (November to April) when it again joins the rest of the flock. On some occasions, the "sumel" may be sold or even slaughtered if the breeder cannot afford to keep it for some reason. Then he has to borrow a "sumel" from a friend till he raises one from his own flock. Table 3. Flock structure. Late October Mid-May Age (months) % flock % sex % flock % sex Males Over 24 1.3 6.84 1.1 5.37 12-24 4.4 23.16 3.0 14.63 6-12 11.0 57.89 4.4 21.46 Below 6 2.3 12.11 12.0 58.54 Total 19.0 100.00 20.5 100.00 Females Over 24 48.2 59.51 42.1 52.96 12-24 19.1 23.58 15.0 18.87 6-12 12.0 14.81 7.4 9.31 Below 6 1.7 2.10 15.0 18.86 Total 81.0 100.00 79.50 100.00 Breeding control and lamb rearing Breeding is controlled to achieve both selection and synchronisation of lambing season with that of rainy season. Selection is largely biased towards the performance of male sheep. The common practice among the pastoralists is to select the fastest growing male for breeding. The number of "sumels" in a flock vary from one to four based on flock size, but usually two are kept in a flock. The "sumel" can be used for breeding at one year of age but preferably at two years or more. The lambing season starts around the month of April to mid-May after few weeks of the start of rains. Lamb crop is thus only once in a year (at least per ewe) with 96% of births attributed to singles. Most farmers claimed that they check the lambs for receiving the first milk (colostrum). They help weak lambs by putting their little fingers into the mouths of the lambs (in order to elicit the suckling reflex) and immediately providing the dam's teat. During the next 30 days (after birth) , the lambs are kept at home during the day either tied, if small in number, or kept in an enclosure, if large in number 137 At 30-40 days of age, the lambs would start to graze and are kept with the "sumels". Weaning is at 4-5 months of age usually before the mating season so as to allow the ewe sufficient time to regain its body condition for successful breeding. In some localities, however, weaning is not practised especially where there is labour scarcity for separately herding the weaned lambs, now termed as "Berars". Castration, marking (identification) and miscellaneous practices Castration is performed for two reasons, namely, breeding control and fattening but little importance is attached to the latter in some localities. The time for castration is usually at about 5-6 months of age. Attrition of the spermatic cord is achieved by beating it between two sticks specially designed for the procedure. Each owner marks his sheep to identity his flock because sheep of different owners may get mixed at watering points or during migrations. The marking is usually made on ears, but less frequently on other sites such as the face, the base of the ears, the upper region of the fore limbs and the perineal area (Figure 1). The marking is done using a fire-heated metal as early as 3-4 months of age but it can also be done at the time of weaning (4-5 months of age). Alternatively, some cuts are made on the tips or sides of ears using a sharp knife. Members of a flock bear the same mark(s); combinations of markings may be made on a sheep. Apart form the markings, each member of a flock, particularly the breeding females are identified by names. Most of the pastoralists noted that each female has its own name no matter how large a flock size may be. Some of the names include: "Merkebo" the big, "Siyaha" moving rump while walking, "Odyer" low-voiced, "Afyer" small-mouthed and "Donyo" resemblance to a boat. Overgrown hoof is trimmed with simple and local materials. First the hoof is heated with fire at the site to be cut. Then the overgrown hoof is cut off using a sharp knife leaving a reasonable proportion to the proximal region of the hoof. Parturition difficulties are usually relieved by local "midwives" who are specially skilled in the field. Diseases and disease control The most commonly reported disease condition by the pastoralists is a condition locally known as "Shillin". It refers to tick lameness, other tick damages and some non-specific conditions. People in the region tend to attribute most animal disease (particularly the "Shillin") to ticks. 138 Figure 1. Identification marks of sheep in Ogaden. a) MARKS ON EARS "HAYIL" s "CULBED" "FUR" V "HARIRO" "LANKAYIR" £y & "SERMO" b)... ON FACES "LAAG" 'SHUMI' c)... ON BACK /( \ OF HEAD "SERIN" d)... ON LOWER PERINEAL REGION e)... ON UPPER REGIONS OF THE FRONT LIMBS A «* "DACTIN" DADMO" 139 Other disease conditions known to the breeders range from the more important parasitic-gastroenteritis (Aal or Gaedanoley) to orf to which little importance is attached. Diseases reported of importance which were noted with several suitable descriptions are to be found in Table 4 along with their respective Somali names. Mortalities were reported to be greater among the lambs, particularly during the dry season. Table 4. Somali names lor major ovine diseases. Tick lameness, other tick damages and some, non-specific ovine conditions Parasitic gastroenteritis . Mange mite infestation Dermatophilus infection of the body Dermatiphilus infection of head, ears and feet Pneumonias Sheep pox Anthrax Nasal oestrids Jackal (hyena) bite Orf Pasteurellosis . Unknowns . Shillin Aal Addo Dulgup Ericke Sambub Furok Kood, Kut Duri Dowa (Duruwa) Ericke Gororsa Hube Apart from the disease conditions, predators were also reported to create problems especially around Aware and Druale where people blamed hyenas for heavy production losses. Pastoralists of Harshin and Duria localities reported jackals as major predator problems. The pastoralists are quite aware of the effects of the disease conditions. The concepts of animal diseases and treatment traditionally held by the breeders are often scientifically sound. Those in the vicinity of veterinary clinics seek official drug supply, while others, either due to lack of veterinary service or the high price of drugs resort to purchase of drug supplies through black markets. Lirophen (an acaricide), panacur and levamisol boli (anthelminthics) were observed in open markets in a number of places. There are rumours that some breeders have their own syringes and needles which they use to inject the sick animals with antibiotics. Indeed, one owner admitted that he used to inject his animals with penicillin. The veterinary section of the region vaccinates the sheep of the region against anthrax, pasteurellosis and sheep pox, but not on a regular basis. 140 Discussion The pastoralists of the Ogaden region have evolved a management practice which enables them to sustain livestock production under the prevailing harsh environment and scarce natural resources. According to Wilson (1983), "Nomadism is a sophisticated management response to a resource base which is always seasonally and often totally deficient." Similarly, Owen (1976) thinks that "Nomadism is one of the most interesting systems of farming that has ever evolved which is still widely practised in some sheep breeding areas and dates back in history right to the threshold of domestication; yet," he continued, "the elements of this method of farming developed over centuries as a balanced ecological system, still make sound sense in terms of using land resource." The main nutritional constraint in the semi-arid zone is that of bridging the gap between wet and dry seasons (Onim et al, 1985). The length of watering intervals, long walks to reach watering point and the forage scarcity during the dry season adversely affect the productivity of the BHO sheep. Better results may be achieved through improved management of the grazing land. The communal grazing land tenure prevents an individual from making any effort to use the land efficiently (Jahnke, 1982) but the trend of the nomads to form associations under "Livestock Breeding Association", as being facilitated by the Range Land Development Unit of the region, can greatly enhance efforts made to introduce pasture conservation principles into the nomadic way of life. As the search for grazing land is closely related to water requirements, water developments can be attempted but caution is necessary in doing so. Any attempt to develop water supply has to be seen in the context of resource management as a whole. Flock structures are more heavily weighed to females which is the case in the flocks of Darfur in the Sudan and Afar in Ethiopia (Wilson, 1975). The control of the lambing season is also in response to the resource availability. This practice may, however, have some adverse effects on the pregnant ewes and the lambs which are born around the start of the rainy season. The ewe gestation period coincides with that of inadequate nutrition and under such conditions even quite low worm burdens can have a detrimental effect on the food conversion efficiency of the dam which in turn influences foetal growth and subsequently the neonate through poor milk production by dam. Furthermore, the free living stages of most helminths can develop to infective stages during the lambing season since the environmental conditions would favour the parasitic egg development during this period. Urquhart et al (1987) suggest that the epidemiology can be worsened by the increased numbers of susceptible lambs and dams, the latter being due to periparturient relaxation of immunity. 141 Despite the mentioned drawbacks, the breeding practices of the pastoralists are reasonable and have their own merits. Frequent lambing puts more strain on the ewe and unless she is fed a superior diet she loses weight (Sahni and Tiwari, 1974). A seasonal production of lambs results in high lamb mortality (Labban and Ghali, 1969) and poor growth rates (Ganesekale, 1975) for those lambs bom in the unfavourable season of the year. The breeders of the Ogaden region are therefore practising in line with this principle. The other element of their breeding practices worthy of notice is the selection of breeding animals. Selection of breeding male is done from generation to generation in the same flock. As flocks do not mix and even if they do so in the field, the males are not allowed to mount the females, in-breeding depression may be assumed to occur. On the other hand, the principle of selecting the fastest growing male lamb for breeding is worthy of encouragement. There is more scope for the improvement of growth traits than of reproductive traits because of higher heritability and phenotypic variations of growth traits within a breed (Gatenby, 1986). Better results could be achieved if breeders exchange their "Sumels". Regarding disease control and veterinary services, it is doubtless that an optimum disease control scheme is needed in a given production system for maximum productivity. Moreover, as the innovations like the introduction of exotic sheep breeds or massive supplementary feeding are of a purely hypothetical nature for the Ogaden region, disease control would be a more practical approach to the watchful and suspicious pastoralists. The indiscriminate use of drugs available through black markets may impose drug failures due to underdosing or inappropriate treatments. Proper veterinary extensions and services should be adopted to alleviate such problems. Conclusions and Recommendations The pastoralists of the Ogaden region are doing their best with regard to management in order to maximise the use of the resource base at their disposal. The real effect of watering frequencies on the production of the Black Head Ogaden sheep and the optimum range for the areas concerned need further investigation. The nutrient requirements, and the incidence, prevalence and seasonality of BHO sheep diseases are not yet fully documented. Unless the spectrum of the sheep diseases is established it would be difficult to undertake appropriate disease control schemes. Some points which should be given due emphasis include: • Judicious water development at certain points after careful consideration of the ecological implication 142 • Exchange of "Sumels" among pastoralists • Adequate disease control measures through the intervention of the veterinary unit. Acknowledgements I would like to express my gratitude to Dr R T Wilson for his encouragement and the initiation of this piece of work as well as for the initial review of the paper. I also thank Prof. S H B Lebbie for his final revision of the paper. Finally, my sincere thanks go to the Jijiga Rangelands Development Unit staff and workers who directly or indirectly contributed to the success of this work. References Ganesekale D. 1975. Effect of season of mating on the economic traits in sheep. Cheiron 4:40-44. Gatenby R M. 1986. Sheep production in the tropics and subtropics. Tropical Agricultural Series. Longmans, Essex, UK. Jahnke H E. 1982. Livestock production systems and livestock development in tropical Africa. Kieler Wissenschaftsverlag Vauk, Keil, Germany. 253 pp. Labban F M and Ghali A. 1969. A study on increasing lambing rate of sheep through management. Journal of Animal Production of the United Arab Republic 9:285-293. MOA (Ministry of Agriculture). 1985. Sheep production project. Annexes. MOA, Addis Ababa, Ethiopia. Onim JFM, Semenye P P and Fitzhugh H A. 1985. Nutritional constraints to small ruminants in Africa. In: Wilson R T and Bourzat D (eds), Small ruminants in African agriculture. ILCA (International Livestock Center for Africa), Addis Ababa, Ethiopia. pp. 54-63. Osman A M. 1985. Near East: Sheep breeding and improvement. World Animal Review. FAO, Rome, Italy. Owen J B. 1976. Sheep production. Bailliere Trindall, London, UK. 436 pp. Sahni K L and Tiwari S B. 1974. Effect of early rebreeding on certain aspects of sheep production. Indian Journal of Animal Sciences 44:767-770. Urquhart G M, Armour J, Buncan J L, Dunn A M and Jennings F W. 1987. Veterinary Parasitology. Longmans, UK. Wilson R T. 1975. Comparative data on two populations of indigenous sheep and goats in Sudan and Ethiopia. Sudan Journal of Veterinary Science and Animal Husbandry 16:1-11. Wilson R T. 1983. Husbandry, nutrition and productivity of goats and sheep in tropical Africa. In: Joint IFS/ILCA Workshop on Small Ruminant Research in the Tropics. Provisional report 14, International Foundation for Science, Stockholm, Sweden. 143 Session 2 Small ruminant performance and reproductive physiology Performances des petits ruminants et physiologie de la reproduction Paramètres de production des ovins Mossi de Gampèla A.J. Nianogo Département de zootechnie Institut de développement rural Université de Ouagadougou B.P. 7021 Ouagadougou (Burkina Faso) Résumé Les mises-bas et poids à âge type des produits ont été relevés de 1983 à 1989 sur un troupeau de 30 brebis et de 3 béliers Djallonké de type Mossi, à la station expérimentale de Gampèla. Le sexe, le mode de naissance, le mois, la saison et l'année de naissance ont eu un effet significatif sur les poids à âge type et les GMQ des agneaux. Sur 960 agneaux nés, les poids àO, 15, 30, 60, 90, 180, 360 et 540jours étaientde 2,33; 3,86; 5,51; 8, 14; 10,32; 14,37; 20,43 et 24,87 kg pour les mâles, et de 2,13; 3,66; 5,07; 7,64; 9,50; 12,94; 1 7,63; et 20, 79 kg pour les femelles. Tous sexes confondus, le poids des agneaux à 90 jours était de 10,55 kg pour les simples, 8, 10 kg pour les jumeaux et 8,68 kg pour les triplets. Le poids à 90 jours était de 10, 77; 10, 11 et 8,37 kg pour les agneaux nés en saison pluvieuse, en saison sèche fraîche et en saison sèche chaude, respectivement. Sur213 primipares nées entre 1 983 et 1987, l'âge moyen à la première mise-bas s'élève à 16,62 mois. L'année et la saison de naissance de la femelle ont eu un effet significatif sur l'âge à la première mise-bas; les mises-bas les plus précoces (15,46 mois) ont été observées chez les femelles nées en saison sèche chaude tandis que les plus tardives (17,40 mois) ont été observées chez celles nées en saison sèche froide. L 'intervalle moyen entre mises-bas a connu une baisse régulière, du premier (9, 14 mois) au cinquième intervalle (7,27 mois); par la suite, ce paramètre a varié de manière irrégulière. Les taux de fertilité, de fécondité et de prolificité étaient de 122%; 139%; et 116%. Le pic de naissance le plus important de l'année se situe au mois de novembre, et le suivant au mois de mai. 145 Production traits of Mossi sheep in Gampela A.J. Nianogo Abstract Data on parturitions and weight at specified age of offspring were collected from 1983 till 1989 on a herd of 30 WestAfrican dwarf ewes and three West African dwarf rams of Mossi-type at the Gampela experimental station. Sex, type of birth, month, season and year of birth had a significant effect on weights at specified age and on daily weight gains of lambs. On 960 lambs born, weights at 0, 15, 30, 60, 90, 180, 360 and 540 days were 2.33; 3.86; 5.51; 8.14; 10.32; 14.37; 20.43 and 24.87 kg, respectively, for males and 2.13; 3.66; 5.07; 7.64; 9.50; 12.94; 17.63 and 20.79 kg, respectively, for females. All sexes combined, lamb weights at 90 days were 10.55 kg for singles, 8.10 kg for twins and 8.68 kg for triplets. Weights at 90 days were 10.77 kg for lambs born during the rainy season, 10.11 kg for lambs born during the cold dry season and 8.37 kg for lambs born during the hot dry season. On 213primiparous dams born between 1983 and 1987, average age at first parturition was 16.62 months. The year and season of birth of the dam had a significant effect on age at first parturition, the earliest parturitions (15.46 months) being observed for dams born in the hot dry season while the latestparturitions (1 7.40 months) were observed for dams born in the cold dry season. The mean parturition interval declined steadily from the first (9. 14 months) till the fifth interval (7.27 months). Thereafter, this parameter varied irregularly. Fertility, fecundity and prolificacy were 122%, 139% and 116%, respectively. Births reached their highest peak in November, and then in May. Introduction Les paramètres de production des ovins Djallonké du Burkina ont ete etudiés par quelques auteurs (Dianda, 1981; IEMVT, 1980; MAE, 1990). Cependant, la plupart des resultats publiés jusqu'a présent proviennent d'enquêtes realisees en milieu villageois. Les ovins Djallonke de type Mossi ont ete introduits sur la station expérimentale de Gampela (SEG), dans le but de favoriser les etudes sur cette race de mouton; cet article, qui est base sur des observations faites entre 1 983 et 1 989, fait le point sur divers paramètres de production. 146 Protocole expérimental La SEG dispose d'un parcours de 400 ha de savane claire arborée et arbustive, en zone Nord-soudanienne (pluviométrie moyenne 670 mm). Un troupeau de moutons Mossi (parent du mouton Djallonké) y a été constitué par l'achat, dans les villages environnant la station, et sur le marché de Pouytenga (140 km au sud-est de Ouagadougou), d'une trentaine de brebis présentant soit des dents de lait usées, soit deux dents adultes, et de trois béliers de 2 à 3 ans d'âge. L'alimentation est basée sur le pâturage (7 heures par jour), avec une complémentation à base de drèches de brasserie égouttées (57 g/jour de matière sèche en moyenne). Après chaque mise-bas, les brebis mères reçoivent environ 250 g de matière sèche de drèches, plus du foin tout-venant de graminées, de ligneux ou bien des fanes d'arachide, selon la saison; cette alimentation se poursuit pendant environ deux semaines, puis les animaux sont remis au pâturage. Les animaux ont été régulièrement vaccinés contre la peste des petits ruminants et la pasteurellose, et ont subi un déparasitage interne au début et à la fin de chaque saison pluvieuse et un détiquage périodique. La lutte est naturelle, les béliers restant en permanence avec les brebis. Les jeunes mâles non destinés à la reproduction sont castrés vers l'âge de 6 mois. Des fiches individuelles ont permis le suivi de la carrière des femelles, et de révolution pondérale des produits. Les poids à âge type ont été déterminés grâce à des pesées périodiques. Les mortalités périnatales et les naissances ont également été enregistrées. Pour l'étude des facteurs de variation, la séparation des moyennes a été faite à l'aide du test de Scheffe du logiciel SAS. L'année a été divisée en trois saisons: la saison dite pluvieuse (SPL) de juin à octobre, la saison sèche fraîche (SSF) de novembre à février, et la saison sèche chaude (SSC) de mars à mai. Les taux de fertilité, de fécondité et de prolificité ont été calculés selon les méthodes décrites par l'IEMvT (1980). Résultats et discussions Structure du troupeau Au 12 juillet 1989 la bergerie comptait 335 têtes dont 65,64% de femelles, avec une forte proportion de jeunes de moins de 1 an (tableau 1). Le poids et la taille au garrot des adultes étaient respectivement de 42,1 1 ± 2,65 kg et 66,56 ± 5, 14 cm pour les béliers, et de 27,34 ± 4,62 kg et 58,48 ± 3,50 cm pour les brebis. Quatre principaux types de robe composaient le troupeau: blanc dominant (53,13%), pie noir (33,13%), pie marron (12,83%), noir (0,89%). 147 Tableau 1. Composition du troupeau au 12-07-89 Femelles Mâles Classe d'âge Nombre % Nombre % >5ans 5 2,27 0 0 4-5 ans 24 10,91 3 2,61 3-4 ans 50 22,73 10 8,70 2-3 ans 35 15,91 20 17,39 1 - 2 ans 42 19,09 34 29,57 0-1 an 64 29,09 49 41,37 Total 220 100,00 116 100,00 Croissance des agneaux Influence du sexe Le tableau 2 montre l'influence du sexe sur la croissance des agneaux de la naissance jusqu'à 540 jours. On observe que, sur cette période, les mâles gardent un poids supérieur (P<0,05) à celui des femelles. Ces résultats sont comparables à ceux obtenus par Berger (1979) chez les agneaux Djallonké de Côte d'Ivoire. Cependant, Berger n'observe de différence significative qu'à des âges compris entre 3 et 7 mois. D'autre part, les poids à âge type observés à Gampèla sont légèrement supérieurs à ceux rapportés par Bourzat (1979); celui-ci trouve des poids à 3, 6, 12 et 18 mois de 8,6; 13,5; 19,0; et 20,8 kg respectivement, chez les mâles. La supériorité des mâles est également apparente au niveau des gains de poids quotidiens (GMQ). Les GMQ importants observés entre 0 et 90 jours sont dus en grande partie à la contribution de l'alimentation lactée, qui est importante à cette période (Yoni, 1989). La corrélation entre la production laitière de la mère et la croissance des agneaux allaités est importante durant les six premières semaines de lactation. Influence du mode de naissance Les agneaux simples sont plus lourds que les jumeaux de la naissance jusqu'à l'âge de 180 jours (P<0,05). Par la suite, la différence devient minime. Cependant, dès le quinzième jour, les triplets pèsent autant que les jumeaux, et la différence de poids grandit ensuite en faveur des triplets (tableau 3). 148 Tableau0.Inrlu ncedsex ,modaiss nce,etl'an éea ur'évol tionpon éraled sagn uxà5300 jours* 007 300 0Na 000d 300P 0,ns0 70ro 030nsd 00,d 03000 0300d 20,30d 0ro 00 030a 070c 0,■c 700c 0,0cb 0,00c 007c 0000c 0,7c ro,7c lée Anr 000 70 0nsb 00b ro0b 7ns. 000b 0070a 00,nsb 00-5b 0,0a 0ns00a 00 ro 00ab ns0a 000a 3007a 0000a 0070a 7000a 030,00a 0ns0a 03000a Triplets 0 0,0b 0Nac 3030ac 3000ac 7,0ac 0,7a 0000a 0ns0ac 0,0ac 0nsnsa 070. Modedenaissance Jumeaux 030 000b 07bc ns0bc 300bc 0ns0bc 030■ab 0,00a 0007bc 030,bc 007b 03000b Simples 70 000a 00a 300a 0,7a 7,0a 0,00a 0000a 030,00a 0-,00a 0300ab 000b 7emelles 00 00. 007b 700b 3000b P00b 0300b 000b nsns0b 030.a 00,00a 0000a " o" o" o CO CO O 0> N N CO CO T S co »- O r- T- o to ■>r CO CO ID CO o CO CO CO CO in N CO ^- (O N CM CO t- t- r- C\J O oi 8 o m" o o u co >- CM in 8CM CO m m CD CD m in in • *© TS 41 eno i. ■D ■CO « o ■o U a a toS 2 s CO S m 3 COCOCO in CO CO CO CO co" co" CO" co" co" co" CO co" ■»" * co" CO K CMCMCqCMCMCMCMCMCMCMCMCMCM »- CM o o 8 3 totols-COCBOr-CM00000*-»-«- 8®o .2 c >c o) _ _ _ • 2 ^ 2 4 -3 co 3 to Tf co r- 3! 0) ! I Sag 5 « 6 £ § o 2 3 & 5 o Q. 0) 1 « CO o in O O V CL co c t 0 E cu> 1 c 3 150 Ces résultats se rapprochent de ceux obtenus par Berger (cité par IEMVT, 1 980) . D'autre part, nos observations rejoignent celles de Murayi etal. (1987), qui en station au Rwanda constatent que les agneaux simples sont plus lourds que les jumeaux jusqu'à 2 ans d'âge. A la naissance, les agneaux nés triples sont moins lourds que leurs homologues jumeaux de 10,45% mais semblent les surpasser dès le quinzième jour. Le petit nombre de naissances triples ne rend pas cette différence significative. Ces observations s'opposent d'ailleurs à celle de Bourzat (1979), qui constate des performances de croissance des triplets de race Mossi inférieures à celles des jumeaux et des simples. Influence de l'année de naissance 1984 a été l'année la plus favorable pour le poids à 90 jours et le GMQ entre 0 et 90 jours. Les agneaux nés en 1985 ont cependant eu les meilleurs GMQ entre 90 et 540 jours. On observe également une baisse du poids à 540 jours de 1984 à 1987 (tableau 2). 1989 fut l'année la plus sèche (494,6 mm de pluies): le pâturage apportant l'essentiel de l'alimentation aux ruminants sur la station, on peut penser que les brebis mettant bas en 1 985 ont été moins favorisées durant leur gestation que les autres sur le plan nutritionnel. Cela expliquerait que les animaux nés en 1984 aient des poids et des GMQ largement inférieurs à ceux observés en 1985 et 1986. Cependant, la pluviométrie a été acceptable en 1985 (619,5 mm) et bonne en 1986 (708,5 mm), ce qui permet aux agneaux nés en 1985 de connaître une croissance postsevrage normale. Les performances des agneaux nés en 1987 sont moins bonnes que celles des agneaux nés en 1984, 1985 et 1986, malgré une pluviométrie plutôt bonne en 1986 et moyenne en 1987 (658,7 mm). Cependant, il faut signaler qu'avec l'augmentation du nombre de mises-bas par an (observée de 1983 à 1987), le suivi des animaux sur les plans sanitaire et alimentaire peut devenir un facteur limitant. Lorsque les animaux deviennent très nombreux, la promiscuité augmente, de même que la compétition à la bergerie et sur le pâturage pour les meilleurs aliments. La baisse de performance de 1984 à 1 987 pourrait aussi être on partie attribuée à une augmentation de la consanguinité; en effet, le rythme de renouvellement des béliers, en principe tous les trois ans avec un ratio mâle/femelle de 1 pour 30, n'a pas été suffisamment rationnel. Influence de la saison et du mois de naissance sur la croissance de 0 à 90 jours On observe une supériorité pondérale des agneaux nés en saison pluvieuse par rapport aux autres agneaux (tableau 3). Les agneaux nés en saison sèche chaude sont les plus défavorisés, depuis la naissance jusqu'à l'âge de 90 jours. 151 La supériorité des agneaux de saison pluvieuse s'explique par l'abondance de pâturages riches et diversifiés. En saison sèche chaude par contre, les herbacées sont à l'état de paille et fournissent peu ou pas de matière azotée digestible (Ouédraogo, 1990). Seuls certains ligneux fournissent encore une quantité significative de nutriments. D'autre part, l'appétit des mères et des agneaux peut être inhibé par les températures élevées qui prévalent pendant cette période. Les performances des agneaux de SSF se rapprochent de celles des agneaux de SPL; ceci peut signifier que les conditions (qualité et disponibilité des fourrages, température, etc.) qui prévalent à cette période sont relativement bonnes. Des variations importantes peuvent cependant exister à l'intérieur d'une même saison. Ainsi, les poids à la naissance les plus élevés sont ceux des mois de septembre et octobre; ceci se justifie car pour les mères de ces agneaux, les derniers mois de la gestation ont coïncidé avec une période d'abondance sur le plan alimentaire. On note d'ailleurs que chez les agneaux nés en février, mars et avril, le poids à 90 jours a été faible (inférieur à 8 kg); pour ces agneaux, les derniers mois de gestation ont coïncidé avec une période où la baisse de qualité du pâturage commence à se faire sentir. De plus, ces agneaux doivent encore traverser la SSC avant d'atteindre la SPL. Le GMQ le plus élevé entre 0 et 90 jours (1 05,4 g) est observé chez les agneaux nés en novembre; le plus faible (56,8 g) est observé chez les agneaux nés en avril. Paramètres de reproduction Age au premier agnelage L'âge à la première mise-bas de 213 primipares nées entre 1983 et 1987 était de 16,6 mois, indiquant un âge moyen à la première saillie fécondante de l'ordre de 1 1,6 mois (tableau 4). Ce paramètre varie largement en fonction de l'année de naissance de la mère: de 365,43 jours en moyenne pour les antenaises nées en 1983, il a régulièrement augmenté pour atteindre 643,29 jours en 1987. Il est possible que l'augmentation du nombre d'animaux ait entraîné un accroissement des contraintes comme nous l'avons mentionné à propos du poids à 540 jours (compétition à la bergerie et sur le pâturage pour l'alimentation, espace disponible par animal, problèmes sanitaires, suivi du troupeau, etc.). Il y a également le problème de l'augmentation de la consanguinité: l'âge moyen à la première saillie fécondante étant de 1 1 ,6 mois, on peut estimer que le temps de séjour d'un géniteur mâle dans le troupeau devrait être de l'ordre de 16 à 17 mois. 152 Tableau 4. Moyennes annuelles de l'âge à la première mise-bas de brebis antenaises Mossi Naissance Age à la ïe Erreur Mère Nombre mise-bas type 1983 16 365,43 18,98 1984 39 393,33 9,82 1985 69 472,04 12,39 1986 65 569,83 15,73 1987 24 643,29 20,77 1983-1987 213 498,76 9,26 L'âge au premier agnelage était de 498,5, 530,6 et 471 ,5 jours chez les agnelles nées en SPL, SSF et SSC, respectivement. Pourtant, les animaux nés en SSC sont les moins favorisés au départ; ils ont en effet le poids à la naissance et à 90 jours le plus faible des trois saisons (tableau 3). Il est possible que les animaux nés en SSC rattrapent les autres au-delà de 90 jours d'âge; l'influence de la saison de naissance sur la croissance postsevrage n'a pas été évaluée dans cette étude, mais Murayi ef al. (1987) ont trouvé que l'effet de la saison cesse d'être significatif à 550 jours d'âge. L'âge au premier agnelage des moutons Djallonké varie beaucoup en fonction du milieu et de l'année. En 1 974, Dumas et Raymond rapportent un âge de 1 1 mois pour le nord-ouest du Burkina. En 1983 àToma, Obulbiga (cité par Kaboré, 1986) trouve 16,8 mois en milieu traditionnel et 15,3 mois en station. Sur 21 agnelles Mossi à la station de Sondré-Est, Kaboré trouve un âge moyen de 14,26 mois. Ces derniers résultats se rapprochent beaucoup des nôtres. Le Ministère de l'agriculture et de l'élevage (1990) rapporte des âges à la première mise-bas variant de 20 à 37 mois en fonction de la zone d'enquête, mais il prend en compte toutes les races ovines rencontrées. Intervalle entre agnelages Sur 246 brebis nées entre 1982 et 1987 et dont les dernières mises-bas ont eu lieu en juin 1989, une seule brebis a atteint onze agnelages. L'intervalle moyen calculé sur toutes les mises-bas enregistrées était de 217,60 jours, soit 7,23 mois, avec des extrêmes de 278,7 à 209 jours respectivement au premier et au dixième intervalle (tableau 5). Les résultats rapportés ici se rapprochent de ceux de la station de Zouma, où Obulbiga (cité par Kaboré, 1986) trouve une moyenne de 7,8 mois pour 40 observations, et de ceux des Djallonké du Cameroun où Vallerand et Branckaert (1975) trouvent une moyenne de 7,9 mois. 153 Tableau 5. Moyennes des intervalles entre mises-bas Intervalle (N°) Nombre Moyenne Erreur type 1 171 278,87 7,24 2 140 245,44 6,75 3 110 240,50 7,02 4 85 235,45 6,83 5 61 221,82 8.99 6 48 232,35 15,35 7 29 253,38 15,66 8 18 247,56 16,41 9 6 222,67 7,85 10 1 209,00 - Par contre, l'intervalle entre agnelages est plus court que celui rapporté par certains auteurs; le MAE (1990) trouve des intervalles allant de 8 à 13 mois en milieu villageois; Dianda (1981) à Sondré-Est donne une moyenne de 9,3 mois. D'autre part, l'intervalle est passé progressivement de 278,87 jours entre le premier et le deuxième agnelage à 221 ,82 jours entre le cinquième et le sixième agnelage. Par la suite on observe des variations moins ordonnées entre le sixième et le onzième agnelage. Répartition des naissances dans l'année La répartition mensuelle des 838 naissances enregistrées de 1 983 à 1 988 figure au tableau 6. On ne note que peu de variation d'une saison à l'autre: 35,68% naissent en SPL, 34,96% en SSF et 29,36% en SSC. Pour l'ensemble du troupeau, les naissances sont plus fréquentes en janvier (9,82% du total), mars (11,14%), mai (10,54%) et novembre (11,50%). Le mois le moins propice semble être le mois de juillet (6,59%). On peut penser que la présence d'un pâturage très jeune et de très bonne qualité en juin crée une situation favorable à la fertilité chez les femelles vides. Quant au pic de mai, il résulterait des luttes de décembre, correspondant aux premières chaleurs fécondes après les agnelages de novembre. Plus de la moitié des mises-bas des primipares se situe entre janvier et mai, avec un pic au mois de mars. 154 Tableau 6. Répartition mensuelle des mises-bas des antenaises et de l'ensemble des brebis Mois Primipares Troupeau Janvier Février Mars Avril Mai Juin Juillet Août Septembre Octobre Novembre Décembre 13,0 9,1 14,2 8,7 11,8 9,1 4.3 6,3 4,7 4.7 6,7 7,5 9,8 6,8 11,1 7,8 10,6 7,7 6,6 6,8 6,7 7,9 11,5 6,7 La saison pluvieuse joue donc un rôle important pour la puberté des primipares, puisque la plupart des saillies fécondantes ont lieu entre septembre et janvier. Les bonnes conditions de la saison humide y favoriseraient chez les agnelles une accélération de croissance permettant l'entrée en puberté. Wilson (1988) a observé chez les ovins Peuls transhumants du Nord-Burkina des pics de naissance en novembre et décembre. Bourzat (1979) a également observé dans la même région pour le mouton Mossi un pic d'agnelage en novembre et décembre et émet l'hypothèse d'un flushing naturel en juin et en juillet. L'absence d'un saisonnement réel (lié aux variations du rythme nycthéméral) a été également noté par Vallerand et Branckaert (1975) sur les Djallonké au Cameroun. Ils ont en effet observé une répartition statistiquement uniforme sur toute l'année. Fertilité, fécondité et prolificité Les observations portent sur 689 femelles reproductrices, 838 agnelages et 960 agneaux enregistrés entre 1 983 et 1 988. Les résultats globaux et par année sont présentés sur le tableau 7. 155 Tableau 7. Résumé des paramètres de reproduction Année Paramètre 1983 1984 1985 1986 1987 1988 Total Taux de fertilité 100 136 124 118 128 117 122 Taux de fécondite 100 162 156 132 152 125 139 Taux de prolificité 100 119 125 112 119 107 116 Taux de naissances multiples - 17,72 24,35 12,35 18,72 6,78 14,2 Taux de mortalité périnatale (0-7 jours) 14,70 8,51 8,33 6,28 4,98 8,05 7,18 Le taux global de fertilité sur la période était de 122%. Ce taux est élevé en comparaison avec d'autres observations sur la même race de moutons: à Sondré, Dianda (1981) trouve un taux de 97%; le MAE (1990) trouve un taux moyen de 80%. Berger et Ginisty (1 980) relèvent des taux allant de 68 à 95,5% chez les Djallonké de Côte d'Ivoire. Cependant, des taux variant de 1 36 à 1 78% ont été rapportés par Hadzi (1988) a Kolokopè au Togo. Ces variations sont en relation avec l'intervalle entre mises-bas et avec les conditions sanitaires et d'alimentation, qui peuvent varier énormément d'un milieu à l'autre. Le tableau 7 montre une très grande variabilité interannuelle: de 100% (1983) à 136% (1984). Cela pourrait s'expliquer par la conduite même de l'élevage et par les conditions de milieu, qui ont quelque peu varié entre 1983 et 1988. 1984 fut ainsi l'année où les animaux ont bénéficié du plus d'entretien (effectif réduit), surtout au niveau de la complémentation. Le taux de fécondité était de 139%, avec des variations allant de 100 à 162%. Ce taux est beaucoup plus important que ceux donnés par Dianda (1981) à Sondé-Est (104%) et par le MAE (1990) pour des ovins de race indéterminée et en milieu villageois (79 à 123%). Cependant, Vallerand et Branckaert (1975) trouvent des valeurs moyennes de 1 68% au Cameroun, tandis que Hadzi (1988) trouve des résultats allant de 157 à 203% par an. Le taux de prolificité a varié de 100 à 125%, avec une moyenne de 116%. Là encore ces performances sont meilleures que celles observées par Dianda (1981) et que celles rapportées par le MAE (1990), qui trouve 101 à 119%. Sur la SEG les agnelages étaient surtout simples (85, 80%); le pourcentage d'agnelages doubles et triples était de 13,84 et 0,36%, respectivement. 156 Le taux moyen de mortalité périnatale (à 7 jours) était de 7,18%, avec des variations interannuelles (tableau 6) allant de 4,98% (1987) à 14,70% (1983). La répartition est presque la même pour les deux sexes: 31 mâles et 29 femelles pour 60 agneaux morts. La cause essentielle de ces mortalités est généralement la subviabilité à la naissance, liée à un faible poids. Dans certains cas, la mortalité a été entraînée par un rejet de l'agneau par la mère ou par une insuffisance de la production lactée de la mère. Mais bien souvent les causes des mortalités n'ont pas pu être déterminées. Conclusions Les ovins Ojallonké du Burkina Faso ont une croissance relativement lente dans les systèmes extensifs d'élevage, inférieure aux performances des moutons du Sahel et des ovins de race européenne. L'étude a confirmé l'influence essentielle de l'année et du mois de naissance sur les performances de croissance et de reproduction des moutons Mossi du Burkina Faso. A travers ces facteurs, on note surtout l'influence des conditions du milieu (évolution de la qualité et de la disponibilité du pâturage, température ambiante, conduite de l'élevage, etc.). Les résultats obtenus montrent aussi que les mâles devraient être renouvelés tous les 16 à 17 mois. Indépendamment de ces facteurs, on note aussi beaucoup de variations pour des groupes d'animaux relativement homogènes. Un programme rationnel de sélection permettrait donc probablement d'aboutir à une amélioration des performances. D'autre part, les performances rapportées ici proviennent d'une situation où l'alimentation repose essentiellement sur le pâturage naturel, et où la reproduction n'est pas gérée; une gestion rigoureuse de la carrière des géniteurs, une intensification de l'alimentation, et une amélioration des conditions d'élevage sont susceptibles de permettre une productivité plus importante, comme l'ont démontré les essais d'embouche menés au nord-ouest du Burkina (Bourzat ef al., 1987) et dans notre laboratoire à Gampèla (Nassa, 1990). Remerciements Cette étude a été réalisée grâce aux efforts de l'université de Ouagadougou et du Département de zootechnie de l'Institut du développement rural, qui sont propriétaires des infrastructures et des animaux de la station de Gampèla. Lankoandé Louka a participé activement à l'enregistrement des données; Diarra David a participé à la saisie d'une partie des données dans le cadre d'un stage de fin d'études (Diarra, 1989). 157 Bibliographie Berger Y. et Ginisty L. 1980. Bilan de 4 annees d'etude de la race ovine Djallonke en C6te d'lvoire. Revue d'elevage et de medecine veterinaire des pays tropicaux. 33:71-78. Berger Y. 1979. Selection et amelioration des ovins-caprins. Rapport annuel 1979. IDESSA/CRZde Minankro, Bouake (Cdte d'lvoire). Bourzat D., Bonkoungou E., Richard D. et Sanfo R. 1987. Essais d'intensification de la production animale en zone soudano- sahelienne: Alimentation intensive de jeunes ovins dans le nord du Burkina. Revue d'elevage et de medecine veterinaire des pays tropicaux. 40:151-156. Bourzat D. 1979. Projet Petits Ruminants-Aviculture. Rapport semestriel. Ministere du developpement rural, Ouagadougou (Burkina Faso). Dianda N. P. 1981. Etude des parametres de I'elevage traditionnel ovin a Sondre-Est. Memoire de fin d'etudes, universite de Ouagadougou, Burkina Faso. Diarra D. 1989. Etude des parametres de production des ovins de Gampela. Memoire de fin d'etudes. universite de Ouagadougou, Burkina Faso. Hadzi Y. N. 1988. Parametres de reproduction du mouton Djallonke a Kolokope (Togo). Roseau de recherche sur les petits ruminants, Bulletin de liaison n" 72, CIPEA, Addis-Abeba (Ethiopie). IEMVT (Institut d'elevage et de medecine veterinaire des pays tropicaux). 1980. Petits ruminants d'Afrique centrale et d'Afrique occidentale. Synthese des connaissances actuelles. MaisonsrAKort (France). 295 p. Kabor6 E. B. 1986. Contribution a I'etude des parametres d'elevage a Sondre-Est. M6moire de fin d'6tudes, universite de Ouagadougou, Burkina Faso. MAE (Ministere de I'agriculture et de I'elevage). 1990. Enquete nationale sur les effectifs du cheptel. Ouagadougou (Burkina Faso). MAE (Ministere de I'agriculture et de I'elevage). 1990. Parametres zootechniques des petits ruminants. Projet Statistiques animales/Ministere de I'agriculture et de I'elevage. Ouagadougou (Burkina Faso). 39 p. Murayi T., Sayers A. R. et Wilson R. T. 1987. La productivité des petits ruminants dans les stations de recherche de l'lnstitut des sciences agronomiques du Rwanda. Flapport de recherche n° 15. CIPEA, Addis-Abeba (Ethiopie). Nassa S. 1990. Influence du poids initial, de lage et de I'alimentation sur la croissance, et les rendements des carcasses chez les agneaux Djallonke. Memoire de fin d'etudes, universite de Ouagadougou, Burkina Faso. Ouedraogo C. L 1990. Influence du traitement des pailles a I'uree sur la digestibilité et la croissance chez les petits ruminants. Memoire de fin d'etudes, universit6 de Ouagadougou, Burkina Faso. Vallerand F. et Branckaert R. 1975. La race ovine Djallonke au Cameroun. Potentialites Zootechniques, conditions d'elevage, avenir. Revue d'elevage et de medecine veterinaire des pays tropicaux. 28:523-545. Wilson R. T. 1988. The productivity of Sahel goats and sheep under transhumant management in northern Burkina Faso. Bulletin of Animal Health and Production in Africa. 36:348-355. Yoni T. 1989. Influence du taux de concentre sur la production laitiere des brebis Mossi. Memoire de fin d'etudes, Institut de developpement rural, universite de Ouagadougou, Burkina Faso. 158 Principaux paramètres démographiques des ovins et caprins des parcours du centre de la Somalie D. Bourzat1, K.D. Gautsch2, M.P.O. Baumann3 etK.H. Zessin3 1 lEMVT/Projet régional Petits ruminants; Laboratoire de Farcha N'Djamena (Tchad) 2 CIPEA, P.O.Box 5689, Addis-Abeba (Ethiopie) 3 Central Rangelands Development Project Somali-German Technical Co-operation, Somalie Résumé L 'étude des performances zootechniques des moutons Somali à tête noire et des chèvres blanches Somali montre un niveau de productivité très lié aux conditions climatiques particulièrement sévères au cours des trois années du suivi: l'âge à la première mise-bas est supérieur à 2 ans; l'intervalle entre parturitions, voisin de 14 mois, dépend de la saison de mise-bas précédente; la fécondité est faible chez les deux espèces (62% chez les ovins et 65% chez les caprins); la mortalité est élevée chez lesjeunes animaux avant le sevrage (13 à 57% selon la saison). Ces paramètres conduisent à une baisse sensible du cheptelpendantla sécheresse. L'étude de ces mêmes paramètres en période postsécheresse devrait fournir des indications importantes sur la capacité de ces populations à surmonter les aléas climatiques. Major demographic parameters for White Somali goat and Blackheaded Somali sheep in the Somalia central rangelands D. Bourzat1, K.D. Gautsch2, M.P.O. Baumann3 and K.H. Zessin3 Abstract Evaluation of the reproductive performance of the Blackheaded somali sheep and the White Somali goat indicates that their 159 productivity was strongly affected by the harsh climatic conditions experienced during the three years of the survey: age at first parturition was over two years; the length of the parturition interval, mean of about 14 months, was influenced by the season of the preceding birth; conception rate was low in both species (62% in sheep and 65% in goats); young-stock mortality was high, ranging from 13% to 51% depending on the season. Population growth rates calculated from these parameters indicated a marked drop in the livestock population during the drought period. A study of these parameters following the period of drought should provide important information on the capacity of these populations to withstand climatic changes. Introduction Le Central Rangeland Development Project (CRDP) opère dans une zone située au centre de la Somalie entre 3°10' et 7°40'N. Limitée à l'est par l'océan Indien et à l'ouest par la frontière avec l'Ethiopie, elle couvre les régions administratives de Hiran, Galgadud et Mudug, soit une superficie de 149 000 km2. Les parcours de cette zone (Central Rangelands) ont toujours joué un rôle important entre le nord du pays (régions d'Hargeisa) et la partie sud-sud-est (régions de Mogadiscio et de la vallée du Shebele). Cette étude devait fournir aux organismes en charge du développement de l'élevage un référentiel technique le plus précis possible sur la productivité des petits ruminants de la zone. La race ovine Somali à tête noire (Blackhead Somali Sheep), également appelée mouton de l'Ogaden ou mouton de Berbera, peuple la zone du CRDP. Classé dans le groupe des ovins a queue grasse, ce mouton est élevé essentiellement pour la production de viande et accessoirement pour la peau. Le poids adulte, sexes confondus, varie de 33 a 45 kg. La population locale caprine est la race Mudugh (Rosetti et Congiu, 1955), également appelée Galla, Deghier ou Digit yer, ou communément chèvre blanche Somali (White Somali Goat), par référence à sa robe d'un blanc très pur quelquefois tachetée de noir. Cette race se rencontre dans la partie la plus aride du pays. Son format est intermédiaire entre la grande chèvre du désert du Soudan et la petite chèvre d'Afrique de l'Est (Devendra et McLeroy, 1982). Ses longues pattes en font une bonne marcheuse. Elle est élevée pour ses productions de viande et de lait. Matériel et méthode La Somalie a un régime des pluies de type bimodal (Griffiths et Hemming, 1 963) très erratique. Le régime thermique est unimodal, de type austral avec des températures moyennes annuelles supérieures à 30 °C et une amplitude sur 160 l'année n'excédant pas 3 à 5 °C. Nous avons retenu le découpage suivant entre les différentes saisons au cours de l'année: • dayr (petite saison des pluies) :22 septembre au 21 décembre. • jilaal (Iongue saison sèche) :22 décembre au 21 mars. • gu (grande saison des pluies) :22 mars au 21 juin. • hagaa (petite saison sèche) :22 juin au 21 septembre. Plan de sondage Le programme couvre six entités administratives (districts): • deux districts de la zone centrale: Dusa Mareb et Galakayo; • deux districts de la vallée duShebele: BeletWeyneet Bulo Burti. La végétation du district de Bulo Burti a été décrite par Kucharef al. (1985). Les épineux dominent largement cette steppe arbustive; • deux districts côtiers : El Dhere et Hobyo. La végétation du district d'EI Dhere se caractérise par deux formations remarquables: les steppes herbeuses de la plaine côtière, et les steppes arbustives de l'intérieur, qu'ont prospectées et décrites Herlocker et Ahmed (1986). Le peuplement humain est constitué en majorité par des éleveurs nomades et quelques agriculteurs sédentaires le long de la vallée du Shebele. 30 villages ont été choisis pour chaque zone par tirage au sort (sur 1 50 environ) en fonction des contraintes d'accès, climatiques ou militaires. Le tirage au sort des unités familiales n'a pas été possible compte tenu de l'impossibilité d'obtenir les listes des familles. Un certain nombre de chefs de famille ont refusé de participer à nos travaux, d'autres se sont déplacés et il nous fut impossible de retrouver leurs troupeaux. Finalement, l'enquête a porté sur 361 exploitations dont la répartition par zone est résumée dans le tableau 1. Tableau 1 . Répartition des différentes enquêtes menées dans les trois zones du projet CRDP Fiche Fiche carrière de femelleZone Questionnaire barymetrique Vallée du Shebele 177 1813 276 Côtière 83 1383 531 Centrale 101 1457 198 Total 361 4653 1005 161 Les fiches d'enquêtes et leur mise en oeuvre Trois enquêtes se complémentant ont été réalisées auprès des unités de production sélectionnées: • les données relatives au système d'élevage ont été collectées à l'aide d'un questionnaire d'enquête mis au point au Burkina Faso (Bourzat, 1984), et complété pour pouvoir être utilisé dans toutes les zones écologiques; • les données barymétriques ont été recueillies grâce à une fiche individuelle. Les mesures et contrôles des animaux ont été effectués au parc de nuit près du campement, soit le soir avant la tombée de la nuit, ou le matin avant le départ des animaux au pâturage; • les données recueillies par une fiche de carrière de femelle ont servi à évaluer les paramètres de reproduction. Cette dernière fiche permet de reconstituer la carrière d'une reproductrice à partir des données mémorisées par son propriétaire. Pour chaque saison de reproduction, le nombre, le sexe de chaque produit ainsi que leur devenir sont enregistrés. Les nombreux travaux menés à partir de ce type d'enquête (Coulomb, 1982; Dumas, 1975) montrent que les données restent fiables tant que l'entretien est mené avec une personne directement impliquée dans la gestion quotidienne du troupeau. La fiabilité décroît avec la diminution d'intérêt porté par le propriétaire à son cheptel et avec l'espèce considérée (les espèces à portées multiples, à cycles de production courts et dont les valeurs unitaires des produits sont relativement faibles sont moins bien connues des éleveurs). La taille de l'échantillon (n=1000) permet une bonne évaluation des paramètres de base de la reproduction des petits ruminants dans la zone du CRDP, qui sont présentés dans ce document. Résultats Répartition des naissances au cours de l'année Le pic de mise-bas correspond à la grande saison des pluies gu, ce qui signifie que la plupart des saillies fécondantes ont lieu pendant la petite saison des pluies ou au début de la saison sèche alors que le pâturage est encore abondant et de bonne qualité (tableau 2). Un deuxième pic moins important apparaît pendant la petite saison des pluies dayr, qui correspond à des accouplements pendant la saison gu. Les chèvres sur toutes les zones, et les ovins et caprins de la vallée du Shebele, présentent des parturitions plus étalées sur les différentes saisons. Ce phénomène peut s'expliquer par un spectre alimentaire des caprins plus large que celui des ovins ainsi que par les pratiques de contrôle des accouplements. Pour assurer un approvisionnement régulier en lait, la reproduction des caprins n'est pas contrôlée et la lutte a lieu toute l'année. En revanche, chez les ovins qui 162 Tableau 2. Répartition des mises-bas par saison, par espèce et par zone (en %) Zone Espèce n Jilaal Gu Hagaa Dayr Intérieure Ovins 156 12,2 54,5 5,8 27,5 Caprins 166 15,1 37,3 11,5 36,1 Vallée du Ovins 154 28,6 40,9 16,2 14,3 Shebele Caprins 246 20,7 34,6 16,7 28,0 Littorale Ovins 475 6.5 20,4 35,0 38,1 Caprins 350 15,1 48,0 16,6 20,3 CRDP Ovins 785 12,0 31,2 25,5 31,3 Caprins 762 16,9 41,4 15.5 26.2 sont essentiellement élevés pour la production de viande, les éleveurs essayent d'avoir des agneaux aux périodes favorables et pratiquent en ce sens le contrôle de la lutte à l'aide d'un tablier placé en avant du pénis ou par une ficelle entourant le fourreau du pénis et attachée à la base du scrotum. Sur la zone littorale où les éleveurs de moutons sont plus spécialisés, les brebis agnèlent surtout pendant les saisons de hagaa et dayr. Age moyen à la première mise-bas Le tableau 3 montre que les caprins sont moins précoces que les ovins, alors que la situation inverse est généralement admise. Les brebis et les chèvres de la vallée du Shebele sont plus jeunes à la première parturition que celles de l'intérieur et de la côte. La situation nutritionnelle moins sévère de la vallée explique peut-être cette observation. La variation de l'âge moyen à la première mise-bas de la région intérieure, représentée par l'écart type, est plus importante que dans les autres zones. Tableau 3. Age moyen à la première mise-bas par zone et par espèce (en mois) Zone Espèce n Age moyen s Intérieure Ovins 99 28,3 12,1 Caprins 99 29,6 10,2 Vallée du Ovins 106 24,8 6.8 Shebele Caprins 167 28,2 6.9 Littorale Ovins 299 26,9 5.5 Caprins 232 31,2 8.9 CRDP Ovins 507 26,8 7.6 Caprins 498 29,9 8.6 163 Intervalle entre mises-bas Le tableau 4 montre que les intervalles entre mises-bas chez les ovins et les caprins sont très voisins. Les ovins de la vallée du Shebele semblent avoir des intervalles entre parturitions plus courts que ceux élevés dans les autres zones. La période de mise-bas précédente a un effet très marqué sur la longueur de l'intervalle suivant: les ovins et les caprins de la zone intérieure et de la vallée Shebele qui ont mis bas pendant les petites saisons des pluies et sèches, hagaa et dayr, présentent les intervalles les plus courts. Les femelles qui mettent bas pendant la longue saison sèche dejilaal ont les intervalles suivants les plus longs. En revanche, il ne semble pas que les caprins de la zone littorale soient sensibles à ce phénomène. Tableau 4. Intervalle entre mises-bas par zone, par espèce et par saison de mises-bas précédente (en mois) Zone Espèce Saison Moyenne Intérieure Vallée du Shebele Côtière Ovins Caprins Ovins Ovins Caprins Jilaal 13 19,5 7,0 Gu 53 16,3 6,2 Hagaa 5 11,4 7,4 Dayr 9 10,7 5,3 Jilaal 19 17,4 5,3 Gu 49 15,6 6,0 Hagaa 9 13,0 7,0 Dayr 26 10,9 5,5 Jilaal 27 17,0 4,4 Gu 48 13,8 5,3 Gu 64 13,9 4,7 Hagaa 25 12,3 4,6 Dayr 51 12,9 5,4 Jilaal 22 19,7 4,0 Gu 59 14,5 5,9 Hagaa 132 12,7 2,3 Dayr 123 12,8 3,6 Jilaal 45 14,6 3,9 Gu 130 13,3 4,8 Hagaa 52 15,0 4,7 Dayr 59 15,2 6,8 164 Taux de fécondité moyen annuel Compte tenu du type d'enquête utilisé pour évaluer le taux de fécondité moyen annuel, seules les femelles ayant eu au moins une mise-bas dans leur vie ont été prises en compte. En général, le taux de fécondité des animaux de la zone intérieure semble plus faible que les taux observés dans les autres zones (tableau 5). Tableau 5. Taux de fécondité moyen par zone, par espèce et par classe d'âge des femelles Zone interieure Shebele Littorale Classe d'âge Espèce Ovins Caprins Ovins Caprins Ovins Caprins 0 n 97 99 109 167 299 232 X 3 1 - - - - s 17 10 - - - - 1 n 97 99 109 167 299 232 X 12 8 11 2 - - s 33 27 31 15 - - 2 n 95 99 109 167 299 232 X 54 53 77 63 78 55 s 50 50 42 48 42 50 3 n 87 90 86 153 281 217 X 53 69 70 67 60 53 s 50 47 46 47 49 50 4 n 60 73 54 108 190 168 X 53 63 61 67 76 79 s 50 49 49 47 43 41 5 n 37 42 28 61 94 106 X 49 62 64 72 74 77 s 51 49 49 45 44 42 6 n 24 24 11 29 32 57 X 71 42 91 66 44 89 s 46 50 30 48 50 31 7 n 13 6 1 5 10 21 X 31 67 - 80 60 71 S 48 52 - 45 52 46 Dans les régions littorale et de la vallée du Shebele, les ovins et les caprins ont pratiquement les mêmes taux de fécondité. Très peu de parturitions ont lieu avant 2 ans. Après ce stade, il ne semble pas que l'âge ait une influence sur le taux de fécondité moyen annuel à l'exception des caprins de la zone littorale, où la fécondité augmenterait avec l'âge. 165 Avortement et mortalité des jeunes Les taux d'avortement et de mortalité des jeunes sont calculés à partir des fiches de carrière de femelles. Le tableau 6 donne une estimation et la description de la situation de la mortalité chez les jeunes. Tableau 6. Taux d'avortement* et de mortalité des jeunes à 1 et à 4 mois, par zone, espèce et saison de naissance (%) Saison de Mort, à 1 Mort, à 4 Zone Espèce naissance n Avort. mois mois Intérieure Ovins Jilaal 19 - 47 74 Gu 85 - 33 45 Hagaa 5 - - - Dayr 43 - 35 60 Caprins Jilaal 25 - 33 50 Go 62 2 30 39 Hagaa 12 - 36 36 Dayr 60 3 43 58 Vallée du Ovins Jilaal 44 5 41 45 Shebele Gu 63 2 25 28 Hagaa 17 6 25 38 Caprins Jilaal 51 4 63 65 Gu 85 4 22 22 ■ Hagaa 25 - 28 28 Dayr 69 10 27 29 Littorale Ovins Jilaal 31 3 30 43 Dayr 182 3 5 8 Caprins Jilaal 53 6 17 24 Gu 168 2 3 4 Hagaa 56 2 10 10 Dayr 71 7 12 20 CRDP Ovins Jilaal 94 3 39 51 Gu 246 - 19 25 Hagaa 122 5 8 13 Dayr 247 4 12 18 Caprins Jilaal 129 4 38 45 Gu 315 2 13 15 Hagaa 93 1 18 18 Dayr 200 7 27 35 * Calculé comme le rapport du nombre d'avortements sur le nombre de parturitions, avortements inclus. 166 Aucune donnée sur les avortements ne fut disponible pour la zone intérieure. Les taux d'avortement les plus élevés pour les deux espèces semblent être ceux enregistrés dans la vallée du Shebele pendant la petite saison des pluies dayr. La mortalité des jeunes des deux espèces à 1 mois ou à 4 mois est plus élevée dans les zones de l'intérieur et de la vallée que sur la zone littorale. La saison la plus meurtrière quels que soient le taux de mortalité ou l'espèce considérée s'avère être la grande saison sèche jilaal. La mortalité à 1 mois représente la majeure partie de la mortalité cumulée à 4 mois. Le premier mois est crucial pour les chances de survie du jeune, à l'exception des agneaux et chevreaux nés pendant la grande saison sèche dans la zone intérieure. Relativement protégés par l'allaitement maternel durant le premier mois de leur vie, ils souffrent ensuite très gravement de la situation de disette de jilaal (abreuvement très précaire, pâturage rare et de faible valeur nutritive, prélèvement de lait très important pour l'autoconsommation de la famille). Discussion La répartition des mises-bas au cours de l'année obéit aux mêmes tendances que celles observées dans d'autres régions arides ou semi-arides: le maximum de parturitions intervient quatre à cinq mois après la saison des pluies. Bourzat (1984) a décrit ce phénomène pour le nord du Burkina Faso et Wilson (1988) pour le Mali central. L'âge à la première mise-bas, de plus de 2 ans aussi bien pour les ovins que pour les caprins, est très élevé par rapport aux résultats obtenus chez d'autres races de la zone semi-aride. Dans le nord du Burkina Faso, Bourzat (1980) mesure un âge moyen à la première mise-bas de 13,8 mois et Ouedraogo (1984) de 13,5±1,7 mois. Wilson et Durkin (1983) enregistrent au Mali central un âge moyen à la première parturition de 15,8+3,8 mois et Haumesser et Gerbaldi (1979) calculent un intervalle de 13,0 à 16,6 mois chez les petits ruminants du Niger. Les moutons Massai' au Kenya, avec un contrôle de lutte, mettent bas pour la première fois à 18,0 ±3,7 mois; les mises-bas commencent chez la petite chèvre Massai d'Afrique de l'Est à 18,3±3,9 mois (Wilson etal., 1984). L'intervalle entre mise-bas varie de 10,9±2,1 mois à 17,5+1,6 mois pour les caprins et de 9,2±1,9 mois à 19,7±2,5 mois pour les ovins. Ces grandes amplitudes traduisent les effets des conditions écologiques et de la saison sur la précédente parturition. Les intervalles les plus courts correspondent bien aux observations faites au Soudan occidental où Wilson (1 976a, 1 976b) rapporte des intervalles de 9,2 mois, et à celles du même auteur (Wilson, 1988) au Mali où les caprins présentaient des intervalles de mises-bas de 9,7±3,5 mois et les ovins, de 8,7±2,5 mois. 167 Compte tenu de l'âge élevé à la première parturition, la fécondité moyenne annuelle par classe d'âge varie de 53%, chez les deux espèces à 3 ans, jusqu'à 71% chez les ovins et 89% chez les caprins de 6 ans. L'influence des différentes zones est très marquée. Chez les moutons Mossi, Ouedraogo (1 984) calcule des taux de fécondité de 145,9% en 1981/82, 85,7% en 1982/83 et 82% en 1983/84; pour les caprins Mossi, ces taux étaient respectivement de 1 69,4, 1 62,5 et 94,6%. De 1982 à 1984, cette région du Burkina fut très durement frappée par la sécheresse. En Somalie, 1983 fut considérée comme une année de sécheresse (Huebl, 1986) et 1986 peut aussi être considérée comme une année difficile (Zessin ef al. , 1 988) . Ces années anormalement sèches transparaissent dans les faibles performances de reproduction calculées à partir de l'enquête rétrospective menée sur les carrières de femelles. Le marché pratiquement inexistant pour les vieilles femelles ne permet pas aux éleveurs de se débarrasser des femelles improductives, ce qui contribue à diminuer la productivité du troupeau. Le taux d'avortement varie de 3 à 5% chez les ovins et de 1 à 7% chez les caprins. Le maximum d'avortement apparaît pour les deux espèces de la vallée du Shebele pendant la petite saison des pluies dayr chez 14% des brebis et 10% des chèvres. Au Mali, Wilson (1988) note 5,1% d'avortements (sur l'ensemble des naissances) chez le mouton du Sahel et 12,6% chez la chèvre sahélienne. Au Burkina Faso, Ouedraogo (1984) calcule un taux de 6,2% sur un échantillon de 113 brebis Mossi et de 13,3% chez 75 chèvres Mossi. Comparés à ces exemples, les taux d'avortement observés en Somalie centrale apparaissent faibles. Le taux de mortalité à 4 mois chez les caprins varie de 13 à 51 % (moyenne: 24%), taux dans lequel la mortalité à 1 mois représente 62 à 77%. 15 à 55% des jeunes agneaux meurent avant 4 mois (moyenne: 26%). Parmi ceux-ci, 77 à 100% disparaissent le premier mois de vie. Au Mali, (Wilson, 1988) constate une mortalité avant sevrage de 23,4% chez les ovins et 34,4% chez les caprins. Conclusion L'enquête réalisée a porté sur une période de trois années particulièrement sèches. Les résultats obtenus reflètent les performances d'animaux soumis à un stress climatique particulièrement profond. La mortalité élevée associée à un intervalle entre mises-bas long et à un taux de fécondité faible sont responsables d'une faible productivité des troupeaux. Ces paramètres utilisés dans un modèle de simulation démographique du croît du troupeau montrent que sans intervention, il n'est pas possible de maintenir les effectifs du troupeau. Les interventions d'amélioration devront s'attacher à réduire la mortalité des jeunes et à augmenter la fécondité. 168 Bibliographie Bourzat D. 1980. Contribution a l'etude des races caprines saheliennes (type peul voltafque). IEMVT, Maisons-Alfort (France). Bourzat D. 1984. Contribution a la connaissance du milieu agropastoral au Yatenga en Haute Volta. Memoire de DESS, lEMVT/Universite Paris XII, Maisons-Alfort (France). Coulomb J. 1 982. Projet de developpement de l'elevage des petits ruminants du Yatenga: production du projet actuel et preparation de la deuxieme phase. IEMVT, Maisons- Alfort (France). DevendraC.et McLeroyG. B. 1982. Goafandsheep production in the tropics. Intermediate Tropical Agriculture Series, Longman, Londres (R.-U.). Dumas R. 1975. Etude sur l'elevage des petits ruminants du Tchad. Fiche de synthese N° 5. IEMVT, Maisons-Alfort (France). Griffiths J. F. et Hemming C. F. 1963. A rainfall map of Eastern Africa and Southern Arabia. East African Meteorological Department, Nairobi (Kenya). Haumesser J. B. et Gerbaldi P. 1979. Observations sur la reproduction et I'elevage du mouton Oudah Nigerien. Revue d'elevage et de medecine veterinaire des pays tropicaux. 33:205- 213. Herlocker D. et Ahmed M. A. 1986. Interim report on range ecology and management of Ceel Dhere District. CRDP Technical Report Series, n° 8. National Range Agency, Mogadiscio (Somalie). Huebl K. 1986. In: Somalia. Agriculture in the winds of change. Publie sous la direction de P. Conze et T. Labahn. epiDokumentation n° 2, epi Verlag GmbH. Saarbruecken-Schafbruecke (Allemagne). p. 55 a 72. Kuchar P., Abdikarem O. M. et Abdisalaam H. S. 1985. The rangelands and their condition in Eastern Bulo Burte District. CRDP Technical Reports Series, n° 12. National Range Agency, Mogadiscio (Somalie). Ouedraogo A. J. 1984. Etude de revolution des parametres zootechniques des ovins et caprins d'un village test du projet Petits ruminants de I'ORD, Yatenga (Burkina Faso). Memoire de DESS, lEMVT/Universite Paris, Val-de-Marne (France). Rosetti G. et Congiu S. 1955. Ricerche zootecnico-veterinarie sugli animali domestici della Somalia. Ispetorato Veterinario, Amministrazione Fiduciaria Italiana della Somalia, A.B.A. 25 n° 10, Mogadiscio (Somalie). Wilson R. T. 1976a. Studies on the livestock of Southern Darfur, Sudan. III. Production traits in goats. Tropical Animal Health and Production. 8: 1 03-1 14. Wilson R. T. 1976b. Studies on the livestock of Southern Darfur, Sudan. IV. Production traits in sheep. Tropical Animal Health and Production. 8:221-232. Wilson R. T. 1988. La production animale au Mali central: etudes a long terme sur les bovins et les petits ruminants dans le systeme agropastoral. Rapport de recherche n° 14. CIPEA, Addis- Abeba. Wilson R. T. et Durkin J. W. 1983. Livestock production in central Mali : Weights at first conception and ages at first and second parturitions in traditionally managed goats and sheep. Journal of Agriculture Science, Cambridge. 100:625-628. Wilson R. T., Peacock C. P. et Sayers A. R. 1984. Aspects of reproduction in goats and sheep in south-central Kenya. Animal Production. 38:463-467. Zessin K. H., Nuux H. A. et Baumann M. P. O. 1988. Livestock disease survey Central Rangelands of Somalia. Technical Report. Vol. I. Livestock demographic data of flocks of sheep and goats. Central Rangelands Development Project, Veterinary Component, Mogadiscio (Somalie). 169 Productivité du mouton Djallonké dans le système agroforestier traditionnel du Mayombe congolais A. Batalou-Mbetanie Direction générale de la recherche scientifique et technique B.P. 2499, Brazzaville (Congo) Résumé La productivité zootechnique des ovins en élevage traditionnel dans la région de forêt subéquatoriale du Mayombe au Congo a été étudiée pendant trois années, de 1988 à 1990 dans un village à partir d'un échantillon de 104 animaux. Les résultats ont permis d'évaluer le quotient de fertilité apparent, le taux de prolificité, les quotients de fécondité, de productivité numérique, et de mortalité chez les jeunes. L'étude a aussi évalué l'âge à la première mise-bas, l'intervalle entre mises-bas et les poids à la naissance. Cette productivité est assez faible en raison de la forte mortalité des jeunes. Le mode d'élevage basé sur le vagabondage ne favorise pas une bonne productivité, en dépit du taux de fertilité moyennement satisfaisant et des conditions écologiques assez favorables. Productivity of the West African Dwarf sheep in the traditional agro-forestry system of Mayombe, Congo A. Batalou-Mbetanie Abstract The productivity of sheep under traditional management in the subequatorial forest area of Mayombe, Congo, has been evaluated over a 3-year period from 1988 till 1990. Data was collected in one village on a sample of 104 animals. Fertility, prolificacy, fecundity, reproductive efficiency and mortality of the young were calculated. Age at first parturition, parturition intervals and weights at birth were 171 also recorded. Productivity is relatively low due to high mortality rates in the young. A management system which is based on free-roaming does notpromote productivity in spite of satisfactory fertility rates and fairly good environmental conditions. Introduction La région du Mayombe est une zone montagneuse de forêt sempervirente subéquatoriale à climat fortement pluvieux (1 800-2 200 mm/an). Depuis 1988, une Réserve de la biosphère y a été créée, contraignant les populations a limiter l'activité cynégétique et à diversifier leurs sources d'approvisionnement en produits carnés. Le petit élevage familial est donc encouragé dans le cadre d'un système de production agricole où les petits ruminants pourraient jouer un rôle socio-économique important. Une étude réalisée en 1987-1988 sur la connaissance du système agroforestier traditionnel du Mayombe a permis d'y situer l'importance de l'élevage. Celui-ci est une activité de cueillette portant sur les races dites locales d'ovins, de caprins, de porcins et de volaille. Son rôle économique est par conséquent très marginal, bien que ces animaux pourraient trouver ici un environnement propice pour une contribution efficace au revenu monétaire de l'exploitant. Le gibier est la principale source de protéines animales et la chasse une activité économique très importante, occasionnant une forte pression sur la faune sauvage. Cette étude a également conclu que les principales contraintes à cette spéculation sont liées au manque de tradition pastorale et à l'abondance du gibier. Ce manque de tradition ne signifie cependant pas un manque d'intérêt pour cette activité, comme le montrent les conseils sollicités auprès des agents techniques. L'assistance aux populations est aussi fortement handicapée par l'inexistence de références sur la productivité des animaux ainsi que sur les méthodes appropriées de conduite des troupeaux dans cet écosystème. C'est dans cette perspective qu'un suivi de troupeaux paysans a été mis en place en 1988 pour évaluer la productivité des moutons Djallonké, dont nous rapportons ici quelques résultats relevés sur trois années. Matériel et méthodes L'étude s'est, pour des raisons matérielles, limitée aux ovins dans le seul village de Makaba, qui compte 68 familles dont 83% possèdent au moins un ovin. Chez tous ces éleveurs, chaque femelle a été identifiée par une boucle d'oreille, dont le numéro a été reporté sur une fiche individuelle de suivi construite de manière à enregistrer les événements de reproduction (Ies dates des mises-bas ultérieures, le nombre de produits à chaque mise-bas, le sexe des produits); révolution pondérale de tous les produits nés; les causes éventuelles de sorties 172 des mères et des produits du troupeau. Tous les mâles au temps initial n'ont pas été pris en compte et sont restés en dehors de nos effectifs jusqu'à la fin du suivi. Les événements qui se produisent au sein du troupeau sont régulièrement enregistrés par un jeune paysan lettré du village. Des enclos ont été construits pour faciliter les opérations de marquage et de pesée; les animauxy sont parqués la veille de chaque visite mensuelle du technicien chargé du suivi. Les données enregistrées ont permis de calculer chaque année les indices de productivité suivants: quotient de fertilité (100 x nombre de mises-bas/nombre de femelles à la lutte); taux de prolificité (100 x nombre de produits/nombre de mises-bas) ; quotient de fécondité (1 00 x nombre de produits/nombre de femelles à la lutte); quotient de productivité numérique (100 x nombre de produits vendus, conservés ou consommés/nombre de brebis à la lutte); taux de mortalité des agneaux (100 x nombre d'agneaux morts/nombre total de produits nés). Résultats et discussion Au début du suivi, 104 femelles ont été enregistrées, réparties entre 48 brebis, 23 antenaises, et 23 agnelles. Le tableau 1 présente le nombre de femelles sorties en cours d'année, le nombre de naissance simples et gémellaires, le nombre de produits par sexe et ceux morts ou sortis pour d'autres causes pendant l'année. On note en particulier un taux de mortalité des agneaux très élevé (plus de 50%). Ces morts sont occasionnées en grande partie par des accidents de tout genre comme les noyades, les empoisonnements, les écrasements par véhicules. Les causes pathologiques se limitent aux parasitoses gastro-intestinales et aux gales. Les grandes épizooties sont pour l'instant inexistantes dans la région. Le tableau 2 présente les indices calculés annuellement à partir des informations recueillies. L'âge à la première mise-bas a été calculé à partir d'un échantillon de 56 antenaises (tableau 3). L'intervalle entre mises-bas (tableau 4) a été calculé pour Tableau 1 . Evolution démographique du cheptel Femelles Agneaux nés Agneaux morts Total mises-bas mises à la Femelles sortieslutte F M F M Simples Doubles Année 1 104 11 70 59 38 37 115 7 Année 2 125 19 73 66 47 32 131 4 Année 3 132 28 88 73 53 30 145 8 173 deux échantillons de femelles: le premier est un groupe de 100 femelles dont les agneaux de la première mise-bas sont morts avant l'âge de 1 mois, et le second un groupe de 1 00 femelles dont les agneaux étaient vivants au sevrage. Le mode d'élevage, qui est essentiellement sous forme de vagabondage, prédispose à la monte libre, provoquant la précocité des mises-bas et la réduction des intervalles entre mises-bas. Tableau 2. Indicateurs de productivité des Djallonké au Mayombe Année 1 Année 2 Année 3 Quotient de fertilité apparente 117,3 108,0 115,9 Taux de prolificité 105,7 102,9 105,5 Quotient de fécondité 124,0 111,2 121,9 Quotient de productivité numérique 51,9 48,0 59,0 Quotient de mortalité des agneaux 58,0 56,8 51,5 Tableau 3. Répartition de l'âge à la première mise-bas de 56 antenaises. Age à la première mise-bas (mois) Antenaises mises à la lutte 8 9 10 11 N 4 17 20 15 26,8% 7,2 30,3 36,7 Tableau 4. Répartition de 100 intervalles entre mises-bas selon la survie des agneaux à un mois 1MB (mois) 10 11 12 Agneaux morts avant l'âge de 1 mois Agneaux vivants au sevrage 12 33 21 20 14 35 22 16 15 100 100 Le poids moyen à la naissance des agneaux a été évalué sur un échantillon de 35 mises-bas simples pour plusieurs rangs de mises-bas (tableau 5). Les faibles poids à la naissance chez les primipares amenuisent les chances de survie des jeunes agneaux qui naissent pour la plupart avec des poids inférieurs à 1,5 kg. 174 Tableau 5. Répartition des poids à la naissance en fonction du rang de mise-bas Mise-bas 1 Mise-bas 2 Mise-bas 3 Mise-bas 4 Poids (kg) N % N % N % N % P<1 kg 2 5,7 - - - - - - 12 - - 3 8,5 1 3 2 5,7 Conclusion Ces données préliminaires sur un échantillon réduit permettent néanmoins d'avoir une première estimation de la productivité des Djallonké en milieu forestier au Mayombe. Il apparaît en particulier que la forte mortalité des jeunes est la cause majeure de la faible productivité des ovins Mayombe, elle-même conséquence du mode d'élevage. Toute tentative d'amélioration de la productivité de cette spéculation paraît difficile dans les conditions d'élevage actuelles. L'instauration de bergeries permettrait de procéder à un meilleur suivi des animaux. Ceci n'est cependant envisageable que pour un troupeau familial de taille optimale pour que les conditions alimentaires et hygiéniques minimales soient assurées en enclos, sans que les autres activités paysannes ne soient perturbées. En effet, les paysans trop sollicités par les travaux agricoles ne disposent pas d'assez de temps à consacrer au suivi des animaux et préfèrent les abandonner à la divagation pour s'alimenter. La mise en enclos des animaux s'accompagne d'une alimentation à l'auge par apport de fourrages verts et de déchets de cuisine, qui n'est possible qu'avec des troupeaux de taille suffisante. En fixant la taille d'un troupeau familial, l'éleveur évite de garder inutilement des têtes supplémentaires qu'il vend régulièrement pour se procurer des revenus monétaires. Les prochaines études s'orienteront donc vers la proposition de meilleures méthodes de gestion des ovins Djallonké. Remerciements Cette étude a été réalisée grâce au concours du Cl PEA à qui nous présentons nos sincères remerciements et nous souhaitons que cette assistance se poursuive dans le cadre de notre réseau. Nous exprimons notre profonde gratitude au docteur Daniel Bourzat pour l'encadrement efficace sur le terrain. 175 Reproductive performance of indigenous goats in traditionally managed flocks in north-east of Zimbabwe L.R. Ndlovu Department of Animal Science University of Zimbabwe, P O Box MP. 167 Mount Pleasant, Harare, Zimbabwe Abstract A study was conducted to define kidding intervals and identify causes for long kidding intervals in communal area flocks. Two flocks were selected in the north-eastern part of Zimbabwe and were monitored over two years. Flock entries and exits were recorded, body-mass changes measured fortnightly and blood samples collected in five consecutive weekly intervals in August-September 1988 and March-April 1989. The average kidding interval for does that kidded in August-December (382 ± 90 days) was longer (P<0.01) than that for does that kidded in March-April (265 ± 48 days). The overall average kidding interval was 370 ± 122 days. Serum progesterone levels also indicated that of the 26 goats that kidded in October-December 1988, 65% were non-pregnant 5-6 months after kidding. It is suggested that nutritional stress could be the cause of these long kidding intervals. Performances de la reproduction des caprins de race locale en elevage traditionnel dans le nord-est du Zimbabwe L.R. Ndlovu Resume Cette etude effectuee sur les troupeaux caprins des zones communales visait a determiner l'intervalle entre parturitions et a identifier les facteurs a l'origine de l'accroissement de ce paramètre. A cet effet, deux troupeaux caprins ont eté selectionnés et suivis 177 pendant deux ans dans le nord-est du Zimbabwe. Les entrees et les sorties d'animaux ont ete regulièrementenregistrees et les variations de leur poids vif determinees tous les 15 jours. Par ailleurs, des echantillons de sang ont ete préleves une fois par semaine pendant cinq semaines en aout-septembre 1988 eten mars-avril 1989 en vue de la determination du taux de progesterone. L 'intervalle moyen entre parturitions des chevres ayant mis bas entre aout et décembre (382±90 jours) est significativement plus eleve (P<0,01) que celui enregistre pour les chevres ayant mis bas entre mars et avril (265 ± 48jours). Sur la population totale, I'IMB est de370± 122jours. Par ailleurs, le taux de progesterone indiquait que 65% des 26 femelles qui avaient mis bas au cours de la periode consideree n'etaient toujours pas en gestation 5-6 mois apres. Cos longs intervalles entre parturitions etaient peut-etre dus au stress nutritionnel. Introduction There are about 1 .8 million goats in Zimbabwe, comprising about 28% of the total livestock population (Central Statistics Office, 1987). Over 98% of the goats are owned by subsistence farmers found mainly in Natural Regions 4-5 which are characterised by low and erratic rainfall. The indigenous goat has been found to be prolific (Sibanda, 1 988, unpublished) but there is a lot of reproductive wastage due to kid mortalities and long kidding intervals (Ndlovu and Royer, 1988, unpublished). Although most of the reported kidding intervals of the Zimbabwean Small East African goat seem to range from 213 to 270 days (Hale, 1986; Royer, 1988, unpublished) there are indications of longer kidding intervals (Hale, 1987). These long kidding intervals could be due to one or a combination of several causes, including acyclicity of does, early embryonic losses of foetuses, abortions that are not noted by enumerators, and shortage of bucks in communal area herds. In order to suggest intervention strategies, it is necessary to better define the kidding interval and identify which of the putative causes of long kidding intervals are involved. This experiment was carried out in order to monitor the cyclicity of does in two communal flocks and to determine kidding intervals specific to these flocks. Materials and Methods Location The flocks monitored were kept at Tsengerai Communal Lands in Nyanga North in the north-eastern part of Zimbabwe. The area lies within Natural Region 5 and is bordered by hills which provide a discrete limit to the area. 178 The main rains fell between December and February and totalled 663 mm and 547 mm for the 1987/88 and 1988/89 seasons, respectively. The temperature ranged from 9 to 35°C with October, November, December and January being the hottest months whilst April, May and June were the coldest. The relative humidity was lowest in September (23%) and highest in December and January (87%). Flocks Farmers were persuaded to amalgamate their flocks into a larger unit: in return they were offered large-size bucks from a local research station. Amalgamation also served as a labour-reducing incentive as the farmers took turns in herding the flocks. Two amalgamated flocks were created which were managed and housed identically and in juxtaposition. In one flock (Yellow), the does were mated to bucks bought by the project from Matopos Research Station whilst in the other (Blue) the does were mated to local bucks. All goats were ear-tagged using plastic ear tags appropriately coloured for each flock. Ownership still lay with each individual farmer who was free to sell, slaughter or dispose of the animals in any way desired and at any time. Measurements The rainfall, humidity and temperature in the area were monitored. The animals in the flocks were weighed fortnightly and births, deaths, sales and purchases were noted by project-employed enumerators. In August-September 1988 and March-April 1989, blood samples were collected for five consecutive weeks from all adult females (15 kg or more), for determination of serum progesterone level. Results Flock structure On average, the flocks consisted of 52.2% adult females, 3.1% entire males and 39.7% kids of both sexes (Table 1). In the Yellow flock only two adult entire males were allowed though this was not always possible to enforce as farmers did not want to castrate progeny from the larger Matopos bucks. Kiddings The kidding pattern showed two distinct peaks in March-April and August-December (Figure 1) which, between them, accounted for over 75% of the kiddings. The March-April peak is quite sharp and compact whilst the 179 August-December one is more diffuse and broad. Of the total births, about 30% were twins and 70% were singles (Table 2). Table 1 . Monitored flock size during 1987 and 1988. April 1 987 October Yellow Blue Yellow Blue Females 69 42 57 37 Males 2 2 2 2 Castrates 1 14 4 5 Female kids 33 23 16 29 Male kids 19 20 10 5 1988 Females 42 35 58 40 Males 4 2 4 3 Castrates 2 2 3 6 Female kids 13 25 18 15 Male kids 10 3 16 15 Figure 1. Monthly kidding incidences in the monitored flocks, 1987-1988. Kidding incidence 20 16 12 8 k ES3 Yellow flock 8^ Blue flock. : ■ . . ■ Sl JFMAMJJ ASONDJFMAMJJ ASOND 1987 1988 180 Table 2. Litter size in 1987 and 1988 in the monitored flocks. Yellow Rock Blue 1987 1988 1987 1988 Singles 29 30 19 28 Twins 12 13 2 14 Total kiddings 41 43 21 42 % twinning 29.3 30.2 9.5 33.3 Doe cyclicity in the flocks Of the females that completed the August/September 1988 sampling period, 64% were pregnant, 20% were acyclical and 16% were cycling (Table 3). This depicts over a third of the flock as being non-pregnant. Of the 1 3 acyclic animals, eight had kidded in August prior to sampling. In the follow-up sampling period (March/April, 1989), of the 26 does that had kidded in October-December 1988 and were still in the flock, about 54% were acyclical and a total of 65.4% of the does were not pregnant five months after kidding (Table 4). Table 3. Incidences of pregnancy and non-pregnancy in the monitored flocks as at August/September 1988\ Non-pregnant Flock Pregnant Acyclic Cycling Yellow Blue Total 24 19 43 9 13 6 11 1 . Total adult females (15 kg or more) were 61 and 43 for Yellow and Blue herds, respectively. Those unaccounted for above were either sold, withdrawn or slaughtered before the sampling period was completed. Table 4. Incidences of pregnancy and non-pregnancy in October-December 1988 as at March-April 1989. does that kidded during Non-pregnant Rock Pregnant Acyclical Cycling Yellow Blue 6 3 6 8 3 0 Total 9 14 3 181 Kidding intervals About 26% of the animals had kidding intervals of less than or equal to 250 days and about 35% had kidding intervals of less than or equal to 300 days (Figure 2). The majority (39%) had kidding intervals between 301 and 450 days. The data are based on 31 females that kidded at least twice during the study period. The mean kidding interval was 370 ± 122 days. However, if season of kidding is taken into consideration, the average kidding interval for does kidding in March-April was 265 ± 48 days whilst that for does kidding in August-December was 382 ± 90 days. Seasonal differences were significant (P<0.01). Figure 2. Frequency of kidding intervals in monitored flocks (31 does that kidded twice). Frequency 6 Aug-Dec Mar-Apr 1 201-250 251-300301-350351-400401-450451-500 501-550 551-560 >600 25.8% 9.7% 12.9% 12.9% 12.9% 9.7% 6.5% 6.5% 3.3% Kidding interval (days) Body-mass Changes Does that kidded in March-April had slightly increased (mean 5% ± 3.37) in body mass from kidding to weaning (24 weeks post-kidding) while does that kidded in August-December had slight losses (mean 4% ± 1 .74) in body mass over the 182 same period. The regression of body-mass change at 24 weeks post-kidding on kidding interval was significant (P<0.01) and the resultant regression equation was: Kidding interval = 395-7.29 percentage body-mass change at 24 weeks post-kidding. Discussion The erratic rainfall pattern noted over the two-year study period is typical of Natural Region 5 areas in Zimbabwe. This has dire implications for cropping, especially the lack of rainfall in January 1988 which is a crucial month for maize production. It also has adverse implications for feed availability to goats and might account for the conspicuous shift of the kidding peak from August-September to November-December in 1988. The flock composition would suggest that even though the buck:female ratio is low, shortage of bucks is unlikely to be a major cause of long kidding intervals as kidding was spread over two seasons. The twinning rates observed here are lower than those reported by Sibanda (1988) using research station data but agree with those reported by Hale (1986) and Ndlovu and Royer (1988, unpublished) based on communal area data. It would appearthat nutritional stress limits the number of foetuses that the animals in communal areas can carry to term. The percentage of acyclical does five months after kidding reflects the long kidding intervals obtained. The kidding interval of 370 days is about 100 days more than that reported by Royer (1 988, unpublished) working on the Small East African goats in Bikita District, south-east of Zimbabwe. It also differs greatly from the figure of 243 days reported by Hale (1987). It should be noted though, that a deliberate effort was made to exclude all observations that showed kidding intervals of over 300 days in the data reported by Hale (1987). The diverse kidding intervals shown in Figure 2 are typical of unselected populations but could also be due to the fact that the does were of different age groups. Assuming that the data reported by Hale (1986, 1987) and Royer (1988, unpublished) represent the genetic potential of Zimbabwean Small East African goats, it would appear that there is some environmental factor that is suppressing the full expression of this potential in Tsengerai flocks. Animals kidding in August/September kid at a time when feed resources are at their lowest and lactate for about two months under such conditions. Thus nutritional stress appears to be a prime probable cause of acyclicity and long kidding intervals in these animals. 183 Body-mass changes support this hypothesis. Animals kidding in March-April on average had lower body-mass losses compared to those kidding in August-December. The large standard errors associated with the means are due to the fact that the flock was unselected and the data were not corrected for age of the doe. Doe age would affect the doe's susceptibility to stress. Nevertheless, the regression equation indicates that one percentage unit loss in bodyweight at 24 weeks post-kidding increases the kidding interval by about seven days. Body-mass changes at 24 weeks post-kidding are important as does should be mated at this stage if they are to conceive within 365 days. Body condition at mating has been shown to influence conception (Henniawati and Fletcher, 1986). Conclusion The monitored flocks showed long kidding intervals which were mostly due to extended periods of acyclicity which seemed to be due to kidding during periods of greatest nutritional stress. Follow-up studies will investigate the possible role of strategic nutritional interventions in reducing kidding intervals. Acknowledgements The author would like to express gratitude to the International Atomic Energy Agency (IAEA), Vienna and the International Development Research Centre(IDRC) for their financial support. Technical assistance from Dr A.C. Llewelyn and the technical staff of the Department of Animal Science, University of Zimbabwe, is greatly appreciated. References Central Statistics Office. 1987. Statistical yearbook. Government of Zimbabwe Printers. Hale D H. 1986. Systems of production and productivity of goats in three communal areas of Zimbabwe. In: Adeniji K O and Kategile J A (eds), Proceedings of the Workshop on the Improvement of Small Ruminants in Eastern and Southern Africa, Nairobi, Kenya, 18-22 August 1986. OAU (Organization of African Unity), Nairobi, Kenya, pp. 181-193. Hale D H. 1987. Sheep and goat production in Zimbabwe. A Technical Report submitted to IDRC. Henniawati H and Fletcher I C. 1986. Reproduction in Indonesian sheep and goats at two levels of nutrition. Animal Reproduction Science 12:77-84. 184 An analysis of lambing records of West African Dwarf sheep kept at lle-lfe, Nigeria I.K. Odubote Department of Animal Science Faculty of Agriculture Obafemi Awolowo University, lle-lfe, Nigeria Abstract Six hundred and ten lambing records ofthe WestAfrican Dwarfsheep kept in the sheep unit of the Obafemi Awolowo University Teaching and Research Farm, lle-lfe, Nigeria from 1981-1987 were analysed. Year of birth significantly influenced birthweight and litter birthweight (P<0.001) and type of birth (P<0.01). Males were heavier (P<0.01) than the females. Lambs delivered in the dry season were lighter (P<0.001) than their wet season counterpart. Litter composition had significant effect on birthweight (P<0.05) and litter birthweight (P<0.01). Male.female ratio was equal while the ratio of single:twin:triplet was 68:3 1:1. Mean birthweight, litter birthweight and litter size at birth were 2.15 ± 0.46 kg, 4.06 ± 0.77 kg and 1.33 ± 0.41 kg, respectively. Significant correlation coefficients of-0.62and +0.87(P<0.001) were obtained when type of birth was correlated with birthweight and litter birthweight, respectively. Adjustmentfactors for the significant effects were provided. Performances d'agnelage des moutons Djallonke a lie Ife (Nigeria) I.K. Odubote Resume L'analyse de six cent dix agnelages enregistrés de 1981 a 1987 sur le troupeau de moutons Djallonke eleves a la ferme experimental de l'Universite Obafemi Awolowo d'lle Ife (Nigeria) est rapportee ici. 185 L'année de la naissance eut un effet significatif sur le poids de l'agneau, celui de la portée (P<0,001), ainsi que sur le type de naissance (P<0,01). Le poids des mâles fut supérieur (P<0,01) à celui des femelles, tandis que celui des agneaux nés pendant la saison sèche était plus léger (P<0,001) que leurs homologues nés pendant la saison humide. La composition par sexe de la portée avait un effet significatif sur le poids à la naissance (P<0,05) et sur celui de l'ensemble de la portée (P<0,01). La proportion des sexes était égale à 1, mais, sur l'ensemble de la période considérée, on a dénombré 68 naissances simples, 31 naissances gémellaires et 1 triplet. Le poids moyen à la naissance, celui de l'ensemble de la portée à la naissance et la taille moyenne de la portée furent respectivement de 2, 15 ± 0,46 kg, 4,06 ± 0, 77 kg et 1,33 ± 0,41. Une corrélation significative entre le type de naissance d'une part et d'autre part, le poids à la naissance (-0,62) et le poids de l'ensemble de la portée (+ 0,87) a été démontrée. Des facteurs de correction sont proposés en ce qui concerne les effets significatifs. Introduction Small ruminants have for so long taken the back seat in the livestock development programme of Nigeria. However, with current awareness and interest, it is necessary that basic information on production traits and factors that affect them should be documented. Fall et al (1982) reported that birthweight in Djallonke sheep of Senegal is significantly affected by year and season of birth, type of birth, sex and parity. There is paucity of such evaluation forthe West African Dwarf (WAD) sheep in the humid tropics of Nigeria. This study was thus undertaken to evaluate the effects of these factors on lambing related parameters in WAD sheep. Materials and Methods The WAD sheep used in this study are adequately described by Adu and Ngere (1979). They were housed in barns measuring 22 m x 4 m on a cemented floor. The animals were grouped depending on their status: lactating, pregnant or growing. Each pen was provided with two receptacles for feeding concentrate and watering. A groove filled with water was made round each barn to prevent ant problem. The barns were cleaned daily. The animals were put on pasture (Giant Stargrass in pure stand) at about 7.30 a.m. and returned to the barns in the afternoon. They were then fed ad libitum a supplementary diet consisting of 50% corn, 7.5% groundnut cake, 10% brewers' dry grain, 30% rice bran, 1 .5% dicalcium phosphate, 0.5% Agricare and 0.5% salt. 186 Panicum maximum and Gliricidia sepium were also provided ad libitum in the pens. The sheep were treated for ectoparasites, drenched once every three months and given other veterinary attention when the need arose. The animals lambed all year round. Birthweights were recorded within 12 hours after birth. Stillbirth or deaths immediately after birth were disposed of after post-mortem examination. Such lambing records were discarded for this analysis. Lambs were examined for any defect at birth and to maturity. Records of the birthweight, litter birthweight and litter size at birth were analysed for effects of year of birth, season of birth (dry season - November to March; rainy season - April to October) type of birth, litter composition and sex using the General Linear Model (SAS, 1982). Differences between means were tested by Duncan's New Multiple Range Test (Steel and Torrie, 1980). From least squares constants, weights of lamb were adjusted to the equivalent of a single, male lamb born in the wet season. Correlations between type of birth and birthweights were also computed. Results and Discussion The least squares means for birthweight, litter birthweight and litter size at birth are given in Table 1 . The least squares constants and adjustment factors for birthweight and litter birthweight are presented in Table 2. Male:female ratio was equal while the ratio of single:twin:triplet was 68:31 : 1 . Significant (P<0.001) correlation coefficients of -0.62 and 0.87 were obtained when type of birth was correlated with birthweight and litter birthweight, respectively. The overall average birthweight of 2.15±0.46 kg obtained in this present study was found to be higher than the 1.8 kg reported by ILCA (1979) for the same breed in Oyo town. Ngere and Aboagye (1981) reported a birthweight of 1.3 kg in Ghana while Ngere et al (1979) reported 1.5 kg in Ibadan. It appears that the location of the breed has an effect on the birthweight either as a result of differences in the management system practised or the feeding regime and feedstuffs offered. Nevertheless, the WAD sheep is the smallest of the Nigerian breeds of sheep. The WAD weighed lighter at birth than the Uda and Yankasa breeds (Buvanendran et al, 1981). The estimated average litter size at birth of 1 .33±0.41 is higher than the estimates reported in the literature for indigenous Nigerian breeds (Buvanendran et al, 1 981 ; Sackey et al, 1987 ). No case of quadruplets was observed in this study. Ngere and Aboagye (1981) observed a twinning rate of 55% in Ghana while Ademosun (1 973) reported a 37% twinning rate which is slightly higher than the 3 1 % obtained in this study for the WAD sheep. Sackey et al (1 987), on the other hand, reported 22% twinning rate in Yankasa kept at institution farms. 187 Table 1 . Least squares means for birthweight, litter birthweightand litter size at birth (kg). Birthweight Litter birthweight Litter size at birth Class No Mean No Mean No Mean Least squares Overall mean 809 2.15 193 4.06 610 1.33 SD 0.46 0.77 0.41 Year: 1981 114 2.06° 31 3.88cd 86 1.39a 1982 157 1.88d 38 3.49° 118 130abc 1983 154 2.10"° 40 4.23bc 116 1.41" 1984 127 2.31 ^ 26 4. 16"° 96 1.24"° 1985 122 2.35b 33 4.42b 92 1.37ab 1986 69 2.46a 13 4.82a 52 1 28abc 1987 66 2.11° 12 3.70de 50 1.20° SD 0.40 0.69 0.46 F. test (P< 0.001) (P<0.001) (P<0.01) Season: Wet 539 2.19a 129 4.16a 406 1.34a Dry 270 2.09b 64 3.85b 204 1.32s SD 0.43 0.76 0.21 F. test (P< 0.001) (P< 0.001) ns Type of birth: Single 417 2.30a - - Twin 374 2.00b 187 4.01b Triplet 18 1.88c 6 5.65a SD 0.39 0.80 F. test (P<0.001) (P< 0.001) Sex: Male 405 2.21a Female 404 2.10b SD 0.46 F. test (P< 0.001) (P< 0.001) Litter composition 0 male, 2 females 94 1.92b 47 3.84b 1 male, 1 female 196 2.00a 98 4.01b 2 males, 0 female 80 2.06s 40 4.12b 0 male, 2 females - - - - 1 male, 2 females 15 1.86b 5 5.58a 2 males, 1 female - - - - 3 males, 0 female 3 2.00" 1 6.0" SD 0.36 0.89 F. test (P<0.05) (P<0.01) SD = Standard deviation. Means along the same column with different superscripts are significantly (P < 0.05) different from each other. 188 Table 2. Least squares constants (LSC) and additive correction factors (AFC) for birthweight and litter birthweights (kg). Birthwelight Litter birthweight Class LSC ACF LSC ACF Sex: Male -0.55 0.00 Female + 0.055 0.11 Type of birth: Single -0.24 0.00 Twin + 0.06 0.30 + 0.82 0.00 Triplet + 0.18 0.42 -0.82 -1.64 Season: Wet -0.05 0.00 -0.155 0.00 Dry + 0.05 0.10 + 0.155 0.31 There was a significant year effect for all the three traits under consideration, though with no clear-cut trend. Birthweight increased throughout the duration of the study except for 1982 and 1987. The year 1987 witnessed a decline in performance for the three traits which may be as a result of the absence of supplementary diet for that year. Litter size at birth began to decline in 1 986. This could be attributed to the possible effect of inbreeding since the Ife flock is closed. In addition, there has never been any conscious genetic improvement of the flock for any trait. The effect of season was significant (P <0.01 ) for birthweight and litter birthweight but was not significant (P>0.05) for litter size at birth. Lambs born in the wet season were heavier than those delivered in the dry season by 0.1 kg. Similar results were obtained by Buvanendranet al (1981) and Wilson and Murayi (1988). The seasonal effect could be nutritional due to the quality of forage available to the ewe during pregnancy. This view was held by Adu and Olaloku (1979) who stated that the birthweight of lambs is significantly affected by the nutritional status of the ewe during pregnancy Osinowo et al (1989) also observed that season of birth. Significantly influenced litter size at birth. The insignificant effect obtained in this study may be due to the fact that the dry season in the southern part of Nigeria is not as pronounced as in the northern part to warrant a high foetal mortality and/or because season in the cited reference was defined differently from the definition in this paper. Litter size was positively correlated (+0.87) with litter birthweight but negatively correlated (-0.62) with birthweight. Birthweight dropped significantly from 2.3 kg for singles to 2.0 kg for twins and 1 .88 kg for triplets. Litter birthweight, however, 189 increased significantly from 4.01 kg for twins to 5.65 kg for triplets. Twins and triplets were thus 3.65 kg and 1.71 kg heavier than singles and thus the low birthweight for twins and triplets are offset by high litter weight. These differences are quite notable and are maintained to maturity (Buvanendran et al, 1981). Though the growth rate of the WAD sheep cannot match any of the northern breeds, the higher litter size could bridge the productivity gap to a certain extent. Ngere and Aboagye (1981) reported that type of birth affected birthweight. This was further supported by the reports of Ademosun (1 973) and Ngere et al (1 979) although the birthweights reported (2.5 kg and 1.5 kg) were higher and lower, respectively, than the value obtained in this study. This goes to show the high variability that exists in the WAD sheep for birthweight. Birthweight recorded in this study ranged from 2.8 kg to 1.2 kg. This provides opportunity for selection to improve birthweight. Male lambs were found to be heavier than female lambs at birth in this study. Buvanendran et al (1981), however, did not observe any sex effect on birthweight. In the present study there were no differences between the sexes within the different types of birth. However, the presence of a male in a litter or the preponderance of males in a litter led to an increase in the litter birthweight and birthweight. From the foregoing, multiple birth is desirable in WAD sheep because of the higher litter weight and litter size and absence of congenital malformation. It should therefore be selected for, to improve reproductive efficiency (Turner, 1969). This will necessitate adequate selection criteria and adjustment factors. Taiwo et al (1982) suggested that age of dam, sex of lamb, type of birth and season of birth need to be adjusted for. The values obtained in this study are indicative, but more studies are still required. Improvement in management practices and system, feeding (regime and feedstuffs) and disease control are, however, imperative for sustainable increase in productivity. References Ademosun A A. 1973. The development of the livestock industry in Nigeria - Ruminants. Proceedings of the Agricultural Society of Niger 10:1 3-20. Adu I F and Ngere L 0. 1979. The indigenous sheep of Nigeria. World Review of Animal Production 15(3):51-62. Adu I Fand Olaloku E A. 1979. A note on nutrition during late pregnancy in West African Dwarf sheep. Animal Production 28: 1 23-1 26. Buvanendran V, Adu I F and Oyejola B A. 1 981 . Breed and environmental effects on lamb production in Nigeria. Journal of Agricultural Science (Cambridge) 96:9-15. Fall A, Oiop M, Sandford J, Wissocq Y J, Durkin J and Trail JCM. 1982. Evaluation of the productivities of Djallonke sheep and N'Dama cattle at the Centre de Recherches 190 Zootechniques, Kolda, Senegal. ILCA Research Report 3. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. 70 pp. ILCA (International Livestock Centre for Africa). 1979. Small ruminant production in the humid tropics. ILCA Systems Study 3. ILCA, Addis Ababa, Ethiopia, 122 pp. Ngere L 0 and Aboagye G. 1981. Reproductive performance of the West African Dwarf and the Nungua Blackhead sheep of Ghana. Animal Production 33:249-252. Ngere L 0, Adu I F and Mani I. 1979. Report of small ruminant breeding subcommittee. NAPRI Bulletin 1 . National Animal Production Research Institute, Shika, Zaria, Nigeria. Osinowo 0 A, Olerunju S A A, Abubakar B Y, Buvanendran V, Onifade O S, Lakpini C A, Ekpe G A and Dennar F O. 1 989. Adjustment factors for weaning weight and litter size in Yankasa sheep. Paper presented at the 14th Annual Conference of the Nigerian Society for Animal Production held at the University of Agriculture, Makurdi, Benue State, Nigeria. Sackey A K, Okoye C I and Mohammed G S A. 1987. Prevalency of multiple birth in the Yankasa breed of sheep in Zaria. Proceedings of the 12th Annual Conference, Nigerian Society for Animal Production. Abstracts, p. 19. SAS (Software System for Data Analysis). 1982. SAS users guide: Statistics. SAS Inc., Cary, North Carolina, USA. 404 pp. Steel R G D and Torrie J H. 1980. Principles and procedures of statistics - a biometrical approach. Second edition. McGraw-Hill Book Company, New York, USA. 633 pp. Taiwo B B A, Ngere L O and Adelaye I O. 1982. Comparative growth performance of Nigerian dwarf sheep and its crosses with Permer, Uda and Yankassa. World Review of Animal Production 181:57-63. Turner H N. 1969. Genetic improvement of reproduction rate in sheep. Animal Breeding Abstracts 37 (4): 545-563. Wilson R T and Murayi Th. 1988. Production characteristics of African long-fat-tailed sheep in Rwanda. Small Ruminant Research 1(1):3-17. 191 Factors influencing flock structure and production performance dynamics of breeding sheep and goat station flocks in Kenya Geoffrey N. Angwenyi Ministry of Livestock Development Hill Plaza PO Box 34188, Nairobi Abstract Population demographic analysis were carried out in four different breeding sheep and goat station flocks in Kenya using data collected over a four-year period from 1986 to 1989. Data collected included monthly flock inventory records for breeding adults, yearling and 0- 1 year old within sex categories, lambings, disposals and mortalities and losses from the flock. Results indicated that irrespective of the flock breed type composition, ecological lactation orbreeding managementschedule practised, the proportion of males to females remained at a ratio of 30-70 except for one flock with predominantly wool sheep breeds located in ecological zone II with a high proportion of 50% breeding adult females, the three station flocks had had breeding females comprising 40% of the total flock. It was observed that a hair sheep station breeding flock located in ecological zone IV and comprising ofDorperand red Maasai sheep breed types andpractising frequent mating system recorded the lowest annual mortality rate (13%) and highest annual lambing rate per 100 breeding females (133%) in the flock, the highest annual disposal offtake (37%) and the highest annual flock population increase (1 7%) compared to the other three stations which practise once a year joining. A wool sheep station in a high potential ecological zone performed lowest in these parameters. It is therefore concluded that management practise rather than flock structure plays a more critical role in production efficiency of sheep and goat flocks located in adaptive environments. 193 Facteurs agissant sur la structure et les performances de production des troupeaux d'ovins et de caprins reproducteurs élevés en station au Kenya Geoffrey N. Angwenyi Ministry of Livestock Development Hill Plaza P.O.Box 34188, Nairobi (Kenya) Résumé Une analyse démographique a été effectuée à partir de données collectées pendant quatre ans (1986-1989) sur quatre troupeaux d'ovins et de caprins reproducteurs élevés en station au Kenya. Ces données portaient sur le nombre mensuel d'adultes reproducteurs, dejeunes mâles et dejeunes femelles âgés de 1 an et de moins de I an, les agnelages, les ventes, la mortalité et les pertes enregistrés dans les troupeaux. II ressort des résultats obtenus que le sexe-ration comptait toujours 30 mâles pour 70 femelles quelles que soient la composition par race des troupeaux, la zone écologique, le système de reproduction. A l'exception du troupeau de la zone écologique II constitué essentiellement de moutons à laine où il y a forte proportion (50%) de femelles reproductrices, les trois troupeaux élevés en station comptaient 40% de reproductrices adultes. Les meilleures performances ont été enregistrées pour un troupeau de Dorper etRed Masai élevé en station dans la zone écologique IV, etdont les animaux étaient fréquemment accouplés: ce troupeau a montré un taux de mortalité de 13%, un taux annuel d'agnelage de 133%, un taux d'exploitation de 37% et un taux de croissance démographique de 17%. Ces performances étaient meilleures que celles des trois troupeaux où l'accouplement est programmé une fois par an. Les chiffres les plus médiocres se rapportant à un troupeau de moutons à laine élevé en station dans une zone écologique à potentialités pourtant élevées. Il apparaît, au vu de ces résultats, que le mode d'élevage est plus important que la composition du troupeau pour déterminer les performances de production de troupeaux d'ovins et de caprins élevés dans diverses zones écologiques. 194 Contrôle des béliers dans la gestion de la reproduction ovine A. Mazouz, F. Toe, A. Lahlou-Kassi et L. Derquaoui Institut agronomique et vétérinaire Hassan II Rabat (Maroc) Résumé Le contrôle des béliers dans la gestion de la reproduction ovine au Maroc nécessite une uniformisation des techniques de l'examen, une programmation dans le temps des interventions et une coordination entre les différents acteurs. Un tel schéma de contrôle des géniteurs est présenté par les auteurs, suivi des résultats préliminaires de l'examen de 741 béliers en système d'élevage intensif et extensif. Selection of rams in sheep-breeding management A. Mazouz, F. Toe, A. Lahlou-Kassi and L. Derquaoui Abstract The control of rams in sheep-breeding management in Morocco requires the standardisation of research techniques, scheduling of interventions and coordination among the people involved. The authors present preliminary results of a ram-testing programme that involved 741 rams in extensive and intensive production systems. Introduction Au Maroc, le cheptel ovin est estimé à 16 millions de têtes (MARA, 1990). Les taux de reproduction y restent généralement bas chez les races locales, malgré leur potentialité intrinsèque élevée de reproduction et des actions entreprises dans le sens de l'amélioration des conduites d'élevage des troupeaux aussi bien en élevage intensif qu'en élevage extensif. 195 La recherche des solutions à ce problème a abouti à toute une série d'investigations sur la reproduction des ovins. Mais les études entreprises au niveau de ces races ont beaucoup plus intéressé les femelles, et ce n'est qu'à partir de ces dernières années que des recherches ont été consacrées au mâle. Les premières investigations ayant montré la présence d'un fort pourcentage d'infertilité chez les béliers, environ 30%, il s'est alors révélé nécessaire de procéder à un dépistage systématique de l'infertilité et au contrôle des béliers dans les élevages au Maroc. Ce document présente une approche réaliste et adaptée aux conditions marocaines développée pour le dépistage de l'infertilité et le contrôle des géniteurs, tout en respectant les normes standards dans ce domaine, suivie de résultats préliminaires qui semblent prouver qu'une grande partie des géniteurs jusque-là utilisés dans les troupeaux marocains sont infertiles. Schéma de contrôle des béliers au Maroc Importance du contrôle des béliers L'examen des mâles permet de détecter ceux qui sont impropres à la lutte ou à l'insémination artificielle et ceux à très haute fertilité. Des problèmes d'infertilité chez les béliers dominants du troupeau peuvent entraîner des pertes considérables en reproduction qui risquent de passer inaperçues pendant un certain temps, surtout si un bon niveau de libido est maintenu chez ces béliers. Il permet d'éviter la propagation de certaines pathologies infectieuses ou héréditaires telles que des tares congénitales des agneaux (agénésies), la brucellose, les infections aActinobacillus seminis, Salmonella abortus ovis, dont la transmission est assurée par le bélier. Les grands traits de l'examen des béliers au Maroc Le contrôle des béliers est réalisé au Maroc grâce à l'examen andrologique, qui a été standardisé dans le but de sa vulgarisation. L'historique individuel et collectif (âge, origine, gestion du troupeau, gestion de la reproduction en particulier, etc.) permet une meilleure interprétation des résultats de l'examen andrologique. Les objectifs des éleveurs, a court et à long terme, sont aussi à prendre en considération. L'examen clinique des béliers comporte deux phases: • un examen général pour apprécier les aptitudes physiologiques, l'état des réserves corporelles (par palpation lombaire), la conformation, avec une attention particulière à la posture du bélier et aux problèmes de l'appareil locomoteur postérieur, et l'aspect de la laine; 196 • un examen de l'appareil génital qui se limite à l'inspection et à la palpation des parties externes et facilement accessibles du tractus génital. Une notation des testicules et épididymes permet d'apprécier leur état sanitaire et fonctionnel. A la palpation, la fermeté et la souplesse des testicules sont chacune notées de 1 à 5 (de très fermes à très moux, et de très souples à très flasques). La taille, la fermeté et la souplesse des épididymes sont notées de 1 à 3. Le prépuce et le pénis sont également examinés (état des muqueuses, mobilité etc.). Pour s'assurer de la quantité et de la qualité de la semence produite par les béliers, l'examen du sperme des béliers, obtenu par électro-éjaculation, se limite à l'estimation des paramètres suivants: • le volume et la consistance de l'éjaculat; • la mobilité massale et individuelle, lue immédiatement après la récolte sous microscope à faible grossissement. La motilité en avant des gamètes est notée en termes de pourcentage. Une mobilité de 70% est exigée pour un sperme de bonne qualité; • la concentration du sperme est déterminée par comptage à l'hématimètre; • après coloration des frottis de sperme à l'éosine-nigrosine, la détermination des pourcentages de spermatozoïdes morts et des spermatozoïdes de formes anormales est réalisée sous microscope (Blom, 1950). Le microscope à contraste de phase est utilisé pour une meilleure visualisation de la morphologie des gamètes. Organisation du contrôle des béliers Formation et sensibilisation Les directions provinciales de l'agriculture (DPA) constituent les organes de sensibilisation générale à travers la formation d'un personnel compétent et l'allocation des ressources nécessaires à la vulgarisation et à l'encadrement des éleveurs. Les instituts et les écoles supérieures, tels que l'Institut agronomique et vétérinaire Hassan II, sont chargés de la formation des cadres de l'élevage. Ces cadres supérieurs, en particulier les vétérinaires, doivent alors prendre en charge la formation continue de techniciens et de vulgarisateurs et tout l'encadrement technique et scientifique du contrôle des béliers dans le pays. La sensibilisation des éleveurs pour le contrôle de leurs béliers doit être faite à travers les organisations d'éleveurs telles que l'Association nationale ovine et caprine (ANOC) ou les regroupements d'éleveurs tels que les coopératives, et par les médias. 197 Organisation du contrôle des béliers Les éleveurs adressent des demandes d'examen des animaux en fonction de leurs besoins en géniteurs; ils font alors appel aux techniciens formés à cet effet pour la réalisation pratique de l'examen. Les vétérinaires sont responsables du contrôle du travail, de l'encadrement des opérations, de l'interprétation et de la communication des résultats aux éleveurs. Ils saisissent les laboratoires de l'Institut agronomique et vétérinaire (IAV) Hassan II pour la réalisation de tests complets ou spéciaux, et d'expertises en cas de besoin. Le traitement de l'ensemble des données se fait à l'IAV qui soumet des rapports à la Direction de l'élevage au Ministère de l'agriculture et de la réforme agraire, chargée de la coordination de l'ensemble du système. Programmes d'intervention sur le terrain En général les schémas d'intervention seront centrés sur les périodes de lutte. Le schéma type est celui de la figure 1 correspondant au système extensif qui inclut la majorité des élevages ovins au Maroc. Il devrait être adapté au système intensif où la frequence de lutte est différente. En système extensif, le contrôle de tous les béliers se fera un à deux mois avant la lutte, avec appréciation de la conformation des animaux, examen de l'appareil génital, et spermiologie. Les animaux douteux seront alors identifiés et réexaminés juste avant le début de la lutte. Au cours de la période de lutte, les éleveurs qui le désirent pourront demander des contrôles en cas de besoin. Figure 1 . Schéma d'un programme de contrôle des béliers dans les elevages extensifs CONTROLE DES BELIERS DOUTEUX CHOIX DES GENITEURS CHOIX DES CONTROLE DE TOUS LES GENITEURS Sélection et contrôle des jeunes béliers sur critères zootechniques et reproductifs (circonférence scrotale) CONTROLES EN CAS DE NECESSITE 3 1 à 2 mois avant le debut de la lutte: Période de lutte Contrôle clinique + Examen du sperme 198 Matériel et méthodes Nous rapportons ici l'examen de 633 béliers de race Timahdite effectué avant la période de lutte en élevage extensif dans le Moyen-Atlas de 1987 à 1989 (63, 66 et 504 têtes respectivement) (tableau 1). En 1988, 108 béliers de 6 races à viande (Ile-de-France et Causses du Lot principalement) en élevage intensif dans la région de Casablanca ont également été examinés. Les élevages visités en système extensif en 1 987 et 1 988 sont tous encadrés par l'Association nationale ovine et caprine (ANOC) alors que tous ceux visités en 1989 appartiennent à des particuliers non membres de l'ANOC. Tableau 1 . Classement des beliers examinés en système extensif Elevages non Elevages encadrés par l'ANOC encadrés par Effectif total Dentition l'ANOC 1989 (%) 1987 (%) 1988 (%) N = 63 N = 66 examiné(%) N = 504 N = 633 0 106 (21,0) 11 (17,8) 3 (4,5) 120 (19,0) 2 74 (14,7) 3 (4,8) 8 (12,1) 85 (13,4) 4 91 (18,1) 5 (7,9) 10 (15,2) 106 (16,7) 6 76 (15,1) 30 (47,6) 19 (28,8) 125 (19.7) 8 157 (31,2) 14 (22,2) 26 (39,4) 197 (31,1) Résultats Conformation, état général et causes de réforme après examen général des animaux A l'examen général les béliers ont été classés selon leur conformation, leur état général, la qualité de leur laine et leur âge. En système extensif, l'examen général révèle que 1 1 ,6% des animaux sont à réformer au niveau des élevages encadrés par l'ANOC contre 24,4% au niveau des élevages non encadrés (tableaux 2 et 3) et 20,4% en système intensif (tableau 4). 35,7% des béliers n'ayant pas encore de dents de remplacement (DR) présentent à ce niveau de l'examen de nombreux défauts dans les élevages encadrés par l'ANOC alors que ce taux est de 1 1 ,3% dans ceux non encadrés. 199 Tableau 2. Classement des betters apres examen general dans les elevages encadres parl'ANOC Dentition Bon(%) Acceptable (%) Mauvais (%) Nombre total 0 57,1 7,1 35,7 14 2 90,9 - 9.1 11 4 93,3 6.7 - 15 6 65,3 20,4 14,3 49 8 85 10 5 40 Total (%) 76 12,4 11,6 129 Tableau 3. Classement des betters apres examen encadres par l'ANOC general dans les elevages non Dentition Bon (%) Acceptable (%) Mauvais (%) Nombre total 0 80,2 8,5 11,3 106 2 55,4 23 21,6 74 4 46,1 20,9 33 91 6 51,3 22,4 26,3 76 8 50,3 21 28,7 157 Total (%) 56,7 18,9 24,4 504 N = 504, Tableau 4. Classement des boliers utilises en elevage intensif apres examen general Dentition Bon (%) Acceptable (%) Mauvais (%) Total 2 96,4 4 75 6 100 8 68,7 3.6 25 31,3 28 4 12 64 Total (%) 79,6 20,4 108 L'inverse est constate pour les geniteurs ayant toutes leurs DR: 5% dans les élevages encadres corrtre 28,7% dans les élevages non encadrés. En système intensif nous constatons que 31 ,3% des beliers ayant 8 DR sont a reformer au niveau de l'examen genéral corrtre seulement 3,6% pour les animaux ayant 2 DR (tableau 4). Nous constatons globalement qu'apres l'examen le 200 pourcentage de geniteurs impropres a la lutte reste plus eleve en système extensif (21,8%) qu'en système intensif (20,4%). Les principales causes de reforme relevees au cours de l'examen genéral des geniteurs sont rapportees aux tableaux 5 et 6 respectivement pour les beliers utilises en système extensif et en système intensif. Tableau 5. Causes de reforme des beliers utilises en systeme extensif apres examen general Elevages encadres par I'ANOC N = 129 (%) Elevages non Causes de reforme encadres par l'ANOC Total (%) N = 633N = 504 (%) Conformation 2,3 21,8 17,9 Laine 4,6 16,3 13,9 Age 3,9 5 4,7 Maladie 0,8 0,2 0,3 Tableau 6. Causes de réforme des beliers utilises en systeme intensif apres examen general Causes de reforme Nombre de cas Fréquence relative Total (%) releves (N = 22) N=108 Conformation 3 13,6 2,8 Boiterie 12 54,5 11,1 Appareil respiratoire 5 22,7 4.6 Autres pathologies 4 18,1 3,7 N.B. : Certains beliers presentent des pathologies associees et sont ici comptes a chaque fois au niveau des cas. Dans le système extensif, les defauts de conformation sont la principale cause de reforme (17,9% des animaux examinés) suivis par la mauvaise qualité de la laine, notamment la présence de jarres (13,9%), alors qu'en système intensif la pathologie proprement dite est dominante (1 7,6% des animaux examinés). Nous relevons cependant que dans les élevages encadres par l'ANOC, les fréquences de reformes dues a la mauvaise conformation et a la mauvaise qualité de la laine représented respectivement 2,3% et 4,6%, contre 21,8% et 16,3% dans les élevages non encadres. 201 Causes de réforme et leurs fréquences après examen spécial de l'appareil génital Dans le système extensif (tableau 7), 10,1% des géniteurs utilisés sont à réformer pour taille testiculaire insuffisante (circonférence scrotale <28cm), bien que les testicules soient sains. Cependant nous n'avons pas noté de circonférences scrotales (CS) inférieures à 28 cm au niveau des élevages encadrés par l'ANOC, alors que le pourcentage de géniteurs à éliminer pour CS <28 cm est de 12,7% dans les autres élevages. Tableau 7. Causes de réforme des béliers utilisés en système extensif après examen de l'appareil génital Elevages encadrés par l'ANOC Elevages non encadrés par Total (%) N = 633Causes de réforme N - 129 (%) l'ANOC N = 504 (%) Balanoposthite 5,4 3,6 3,9 Epididymite 6.9 5,2 5.5 Orchite 3.9 7,1 6.4 Monorchidie 0.8 1.2 1,1 Atroph./ Hypopl. testiculars 9,3 5 5.8 Abcès/Tumeurs testiculaires 0.8 1.6 1.4 CS insuffisante - 12,7 10,1 Toutes causes 27,1 36,3 34,2 Dans les élevages encadrés par l'ANOC, la pathologie "vraie" de l'appareil génital constitue la principale cause de réforme à l'examen spécial du tractus génital, avec une fréquence de 27,1% contre 23,6% dans les élevages non encadrés. En système intensif (tableau 8), 40,7% des géniteurs en service ont présenté une pathologie au niveau du tractus génital. Les orchites, les balanoposthites et les épididymites avec respectivement des fréquences relatives de 40,9%, 34,1% et 40,9% représentent les trois pathologies les plus importantes chez les béliers examinés de races importées au Maroc. Elles constituent également pour l'ensemble des béliers examinés les dominantes pathologiques, avec des fréquences de 6,4%, 5,5%, et 3,9% en système extensif et de 16,6%, 13,6% et 16,6% en système intensif. 202 Tableau 8. Causes de réforme des béliers utilisés en système intensif après examen de l'appareil génital Nombre de cas Fréquence relative Total (%) Causes de réforme observés (N = 44) N =108 Balanoposthite 18 40,9 16,6 Epididymite 15 34,1 13,9 Orchite 18 40,9 16,6 Monorchidie 2 4.5 1.9 Atroph./Hypopl. 6 13,6 5.6 testiculars Abcès/Tumeurs testiculaires 1 2,2 0.9 N.B. : Certains béliers présentent des pathologies associées et sont ici comptés à chaque fois au niveau des cas. Qualité du sperme La spermiologie réalisée chez tous les béliers utilisés en système intensif (N = 108) et chez 20% des béliers utilisés en système extensif (N = 125) montre que 20% des béliers sains cliniquement présentent des anomalies au niveau de leur spermiogramme. En effet, 21% des éjaculats obtenus avaient un volume inférieur à 0,5 ml, 71 ,4% avaient une mobilité inférieure à 80%, 64,3% présentaient plus de 20% de spermatozoïdes de forme anormale et 35,7% avaient une concentration inférieure à 0,8 million par millimètre cube. Discussion Conformation et état général En élevage extensif les causes majeures de réforme après examen général sont la mauvaise conformation et la mauvaise qualité de la laine, avec respectivement des fréquences de 17,9% et 13,9% contre 2,8% et 0% en élevage Intensif. Ces chiffres montrent bien l'intensité de la sélection au niveau des béliers de races importées où la conformation bouchère des animaux est le critère zootechnique le plus important de choix de la race pour la production d'agneaux de boucherie. Cette sélection se fait déjà sentir au niveau des élevages encadrés par l'ANOC. En effet, 2,3% seulement des béliers présentent une mauvaise conformation contre 21 ,8% dans les élevages non encadrés, et 4,6% présentent une mauvaise laine contre 16,3% dans les élevages non encadrés. Par ailleurs, les taux de réforme après examen général des géniteurs, qui sont de 1 1 ,6% et de 24,4% 203 respectivement pour l'ANOC et les élevages non encadrés, montrent bien l'importance de la conduite du troupeau. En effet, les élevages encadrés, suivant les conseils de l'ANOC, ont une meilleure gestion du troupeau en matière d'hygiène et d'alimentation. Les taux de réforme élevés (17,6%) liés à la pathologie chez les animaux, après examen général en système intensif, pourraient s'expliquer par une moindre rusticité des races importées par rapport aux races locales d'une part, et d'autre part par leur utilisation très intensive pour la lutte. Causes de réforme liées à l'état de l'appareil génital Les géniteurs utilisés dans le système de l'ANOC ont tous une circonférence scrotale supérieure à 28 cm, alors que 12,7% de ceux utilisés dans les élevages non encadrés sont à réformer pour taille des testicules insuffisante. Ceci montre bien que de meilleures conditions d'élevage, notamment un meilleur apport alimentaire, ont un effet sur l'appareil génital. Les béliers ayant des testicules de taille insuffisante ne sont pas agréés par l'ANOC comme géniteurs. Cependant, la pathologie "vraie" de l'appareil constitue la principale cause de réforme des béliers (27,1%) dans les troupeaux encadrés par l'ANOC. Ceci démontre bien la nécessité d'intégrer au sein de toutes les commissions de sélection ovine des spécialistes en reproduction. Au niveau des élevages en système intensif, 40,7% des béliers souffrent de pathologie au niveau de leur tractus génital contre 34,2% en système extensif. En effet, l'utilisation des béliers est plus intensive dans les élevages industriels que dans les élevages extensifs, et leur remplacement par des jeunes est très souvent tardif. Galloway (1982) affirme que les maladies générales et le surmenage se répercutent au niveau de l'appareil génital, notamment au niveau des testicules qui dégénèrent. C'est dans le lot des animaux âgés que l'on trouve une fréquence élevée de pathologie de l'appareil génital. En effet, ce taux peut atteindre 46,9% pour les géniteurs ayant leur 8 DR dans les élevages intensifs selon Toe (1989). Les béliers âgés (plus de 5 ans) doivent être réformés car avec l'âge et l'intensité d'utilisation la condition physique diminue et les pathologies de l'appareil génital apparaissent. Nos résultats confirment, même si les fréquences trouvées ici restent relativement en dessous de celles déjà rapportées par Tibary etal. (1988) et Toe (1989), que les orchites, les epididymites et les balanoposthites sont les dominantes pathologiques de l'appareil génital des béliers au Maroc. Nous trouvons des fréquences de 9,3%, 7,9% et 6,8% respectivement pour les orchites, les épididymites et les balanoposthites contre 12,8%, 12,3% et 10,6% rapportés par Toe (1989). La production et la qualité du sperme des béliers diminuent considérablement dans les épididymites et les orchites, en particulier celles dues aux infections 204 (épididymite contagieuse du bélier). Nos résultats montrent que 20% des béliers présentent des anomalies au niveau du spermiogramme; ceci pose le problème de la saisonnalité des béliers, notamment chez les races importées, et des pathologies infracliniques pouvant échapper à l'examen clinique. Conclusion S'il est vrai que la gestion de la reproduction chez le bélier doit tenir compte de toutes les variations de la fonction sexuelle mâle, le contrôle des géniteurs permet de garder pour la lutte ou pour l'insémination artificielle les plus performants et de les rentabiliser par l'utilisation d'un ratio mâles/femelles le plus faible possible, vu les prix sans cesse croissants des bons géniteurs. Il permet également d'obtenir de meilleurs taux de fécondité surtout lorsque parallèlement toutes les femelles infertiles sont écartées de la reproduction. La pathologie de l'appareil génital, avec essentiellement les orchites, les épididymites et les balanoposthites, est l'une des principales causes de réforme des béliers au Maroc, à côté de l'aspect zootechnique (conformation et laine). Cependant, la recherche de toutes les causes de réforme, surtout à l'échelle de tout le pays, nécessite une standardisation de la technique de l'examen des béliers. La clinique et la spermiologie constituent alors les deux éléments clés de cet examen. Mais un bélier jugé insatisfaisant au niveau de l'examen n'est pas nécessairement infertile (monorchidie). Enfin, seuls des programmes d'intervention cohérents tenant compte des besoins des éleveurs dans les différents systèmes d'élevage, la sensibilisation des éleveurs aux problèmes de l'infertilité chez les mâles et la formation d'un personnel qualifié en nombre suffisant permettront de couvrir l'ensemble des élevages au Maroc. Bibliographie Blom E. 1950. A simple rapid staining method for différenciation between live and dead sperm cells by means of eosin and nigrosin Nordisk Veteringer medicin 2: 58-61. Galloway D. B. 1982. Reproduction in the rams Australian Veterinary Journal. 67: 163-195. MARA (Ministère de l'agriculture et de la réforme agraire). 1990. L'élevage au Maroc. 1q: L'élevage dans les systèmes céréaliers méditerranéens. Actes de la conférence internationale. Rabat (Maroc). Tibary A., Boukliq R., Adnani M. et Toe F. 1988. Importance de l'examen du bélier en gestion de la reproduction ovine: variations physiologiques de la qualité du sperme et dominantes pathologiques. XVIIIe Journées de l'Association nationale pour la production animale, 10 et 11 mars 1988, Institut agronomique et vetérinaire Hassan II, Rabat (Maroc). Toe F. 1989. Vulgarisation de I examen des béliers: causes d'élimination des béliers de la lutte. Thèse de doctorat vetérinaire, Institut agronomique et vetérinaire Hassan II, Rabat (Maroc). 205 Puberté chez la race D'man, la race Sardi et leur produit de croisement L. Derquaoui, R. Boukhliq, A. Lahlou-Kassi, A. Mazouz et F. Toe Département de reproduction animale et I.A. Institut agronomique et vétérinaire Hassan II Rabat-Instituts (Maroc) Résumé L'âge à la puberté est l'un des facteurs les plus importants qui déterminent la productivité des élevages ovins. Ainsi, une amélioration de celle-ci peut être achevée par la réduction de l'âge à la puberté chez les races dites tardives par sélection intrarace ou par croisement avec une race précoce. Ce dernier type d'action peut être réalisé au Maroc à travers des manipulations génétiques grâce à l'existence de la race ovine D'man connue par sa précocité sexuelle. Les auteurs présentent les résultats et la discussion d'une expérience de croisement de plusieurs années. Puberty in the D'man and Sardi breeds and their crosses L. Derquaoui, R. Boukhliq, A. Lahlou-Kassi, A. Mazouz and F. Toe Abstract Age at puberty is one of the most important factors of sheep productivity. Reducing age at puberty of so- called late maturing breeds through selection within the breed or crossbreeding with an early maturing breed leads to improved productivity. In Morocco, this was achieved through genetic engineering with the D'man breed which is known for its early maturity. This paper presents and discusses results of a crossbreeding operation that covers several years. 207 Introduction En 1987, le nombre de moutons au Maroc était estimé à 16 millions de têtes (MARA, 1990) dont 50% de brebis adultes en âge de reproduire. La productivité ovine définie par le nombre d'agneaux élevés jusqu'à l'abattage reste encore faible. Elle est estimée en moyenne à 0,6-0,7 agneau par brebis et par an (Kabbali et Berger, 1990). Cela résulte d'un faible taux de fertilité (78-63%) et d'un taux élevé de mortalité des jeunes de 18 à 31% dans les conditions du terrain (Chaarani, 1987), contre 5 à 10% en stations expérimentales (Lahlou-Kassi, 1987). L'étude des performances de reproduction des jeunes femelles ovines, réalisée à travers l'analyse des données recueillies sur le terrain ou en station expérimentale, met en évidence l'importance de la dispersion de l'âge à la première mise-bas et l'influence de la saison de naissance sur l'âge à la puberté. Age au premier agnelage L'ensemble des études menées aussi bien dans le berceau de race qu'ailleurs met en évidence la précocité sexuelle de l'agnelle de race D'man répandue dans le système des oasis et en zone présaharienne. Plusieurs cas d'agnelages ont été signalés à l'âge de 8 mois chez la D'man (Bouix et al., 1974). Un intervalle d'âge de 9 à 22 mois a été observé dans les élevages D'man des vallées du Ziz (Barki, 1974) et du Drâa (El Fakir etal., 1979). Par ailleurs, une étude menée en station pendant quelques années montre que l'âge au premier agnelage varie de 1 3 à 1 4 mois quand les agnelles sont respectivement nées en automne ou au printemps (Bouix ef a/., 1974). En dehors de son berceau de race, l'âge moyen de la D'man au premier agnelage varie de 9 à 33 mois (Boutgayout, 1980), avec 87% des femelles mettant bas à un âge inférieur à 24 mois. Contrairement à la précocité de la race D'man, la race Sardi est considérée comme une race tare. En effet, l'âge signalé au premier agnelage varie de 13 à 27 mois (Boutgayout, 1980) montrant une distribution bimodale (de 13 à 15 mois et de 22 à 25 mois). Dans le berceau de race, l'âge moyen observé est de 15,6 mois (Drissi, 1983). Age et poids à la puberté L'agnelle de race D'man élevée au niveau de la vallée du Drâa peut manifester le comportement de chaleurs aussi tôt qu'à l'âge moyen de 150 jours (130-170) (Harrouni, 1977) et au poids moyen de 15 kg (Bouix etal., 1974). En dehors de son berceau de race, l'agnelle D'man devient pubère à un âge moyen de 6,3 à 6,5 mois et à un poids de 24 ± 1,5 kg (Lahlou-Kassi, 1982). Cette précocité sexuelle a été également signalée chez le mâle D'man (Glatzel, 1 980) . Par ailleurs, ce caractère peut être transmis à la descendance d'un croisement avec une race tare (Lahlou-Kassi, 1982). Cela suggère que la précocité sexuelle de la race D'man est un caractère racial génétiquement transmissible. 208 A l'inverse de la race D'man, il existe peu d'informations relatives à l'apparition de la puberté chez la race Sardi. Néanmoins, l'âge rapporté à la première saillie fécondante varie de 7 à 19 mois (Drissi, 1983). Différents auteurs s'accordent maintenant pour dire que la puberté chez l'agnelle est déterminée par des facteurs génétiques (Dyrmundsson, 1973; Land, 1978) et par l'interaction de plusieurs paramètres environnementaux dont il n'a pas été possible de scinder les effets de façon satisfaisante. Ces derniers regroupent: le niveau alimentaire et la vitesse de croissance (Land, 1978; Quirke, 1979; Dyrmundsson, 1981; Foster etal., 1985), la saison de naissance (Foster, 1981; Lahlou-Kassi ef al., 1989), la photopériode (Foster ef al., 1985) et enfin les traitements hormonaux exogènes (Fosterer a/., 1985; Dyrmundsson, 1987). L'objectif de cette étude est de déterminer avec précision l'âge et le poids à la puberté de la race D'man (précoce), de la race Sardi (tardive) et de leur produit de croisement dans des lots homogènes d'agnelles nées à différentes saisons de l'année. Matériel et méthodes Les observations ont été faites à la ferme d'application de l'Institut agronomique et vétérinaire Hassan II de Rabat (Maroc) dans la plaine du Tadla en périphérie de la zone irriguée. Les brebis sont soumises à un rythme de reproduction intensif de trois agnelages en deux ans. Les jeunes sont sevrés à un âge allant de 55 à 100 jours et reçoivent une ration à base de luzerne (1 kg/tête) et d'orge (400 g/tête). Dès l'âge de 3 mois, les femelles sont réparties en lots de 20 à 30 agnelles. Dans chaque lot, on procède à la détection biquotidienne des chaleurs à l'aide d'un bélier D'man x Sardi (DxS) à pénis dévié. La femelle est dite en chaleurs quand elle se laisse chevaucher par le bélier détecteur. Toutes les femelles sont systématiquement pesées une fois par mois et chaque femelle qui présente des signes de chaleurs est pesée séparément pour déterminer le poids au premier oestrus. Le contrôle de l'activité ovarienne est réalisé par le dosage par la méthode radio-immunologique de la progestérone plasmatique dans des échantillons de sang prélevés une fois par semaine dès l'âge de 3 mois. Un taux de progestérone supérieur ou égal à 1 ng/ml est considéré comme témoin de la présence d'un corps jaune actif faisant suite à une ovulation. Au cours de la première semaine qui suit la manifestation du premier oestrus, le taux d'ovulation est établi par examen coelioscopique des ovaires In situ à l'aide d'un endoscope. 209 L'analyse de variance et l'ajustement des moindres carrés, pour étudier l'effet de la race, de la période de naissance et du type de naissance sur l'âge au premier oestrus et à la première ovulation, ont permis de faire des régressions entre: • l'âge et le poids à la puberté; • l'âge à la puberté et le taux d'ovulation; • l'âge à la puberté et la saison de naissance. La comparaison des distributions cumulées des âges au premier oestrus et à la première ovulation en tenant compte de la troncature des échantillons (observations non faites sur l'ensemble des échantillons de départ) a été faite à l'aide d'un test non paramétrique: le test de Smirnov. Résultats Le tableau 1 des moyennes observées de l'âge au premier oestrus, à la première ovulation, du poids et du taux d'ovulation à la puberté montre des différences aussi bien entre les génotypes qu'entre les périodes de naissance pour chaque génotype. Age et poids à la puberté Race D'man Les âges moyens à la première ovulation et au premier oestrus sont respectivement de 212 et 229 jours, toutes périodes de naissance comprises. D'après le tableau 1, on constate que les agnelles nées en avril-mai montrent des signes de premières chaleurs à 194,46 jours contre 255,15 pour celles nées en novembre-décembre. Toutefois, 50% des agnelles du premier et du deuxième groupe ont leur premier oestrus respectivement au bout de 194 et 266 jours (figure 1). De même, l'ovulation survient à 196 et 224 jours chez la moitié des agnelles nées respectivement en avril-mai et en novembre-décembre (figure 2). Les agnelles nées au cours de cette dernière période sont plus lourdes (23,48 kg) que celles nées au cours de la première (17,81 kg) (P<0,01). Notons ici que le poids à la puberté — toutes saisons comprises — ne varie pas en fonction du type de naissance et est de 21 kg. Race Sardi La race Sardi semble atteindre la puberté à un âge plus tardif (1 1 mois environ) et à un poids plus élevé (30 kg) que la D'man. Ici, ces deux paramètres sont sous la dépendance de la saison de naissance. En effet, les moyennes ajustées des moindres carrés de l'âge au premier oestrus sont de 247,38; 273,73; et 459,20 jours respectivement pour les périodes de naissance de janvier-février, 210 c .0 .co g 1 CO via 0 > +1'b .0 X T3 X a C .2 CD 3 > T3 0 E X c CO CO u 1- 5 8 ■CD Q.8 CD -5 c« X3 O, ■H —í CO « +1 S 3 -CO X O CD «5-0 >< c 0 O 0 c EJS ««> CO_g Q. c 0 1 0- co a ¡K O" 'CD > 15 O 3«2 s? c 5. ico 0 ffl +1 Í " -c0 X *. a >.co eg OCO co E? co Js .« CDCO CO £ c§ s ^ 0 0» *_ 1 * CO3 to CD ^o 0 CD ._- 0 +1 0>T3 ▼" X 2 co a 0 "> CO .O co >. CO 5 0 E 5 c «1 c CD CO CO . 2 °-§ c& ÜÍ -O ÍÍ § 0 Vs» 2 t O Ü 52 -§ á a Q. -2 to _co ** ^ TO • «^ * 8a CD c• .O CO• 0 0 2 ■— f/1 0- CO 01 r- r- IO a *- o t■» «0 w"cm 2. 8 5 CM C 5 "> 9 « LO o 8» o.«a- CM CO 2 co 8_ CO CM CO O o n in s s «-" ~ *-" O «-' o 8 S2 cm £¿ 3 at 10 -r O) -r p co ¡2 f 2 cm CM C CM C CM a COCM CO N Í CO Í IO - CM.... .... . 0> S w £ £ in ™ cm 2!. cm ~ cm 21 • 98 10 0 CO CO LO CM to CO m d I-" d cm" d cm" T-" 8 2°. co . ... O m" CO ,n' CO SÜ CM S CO s 9.58.1 CM CO CO CMCM -*■ 8 88Í3 - fc 3- CO . cm a 2 21 CO D < 59 o O) ■= co 3 2 U O)o «- 211 Figure 1 . Pourcentages cumulés des âges au premier oestrus chez des agnelles D'man nées en avril-mai (3) et en novembre-décembre (2) 160 180 200 220 240 260 280 300 320 i 10 Jours Mois novembre-décembre et avril-mai (tableau 2). Cependant, l'âge observé à la première ovulation n'est pas significativement différent entre les naissances de janvier-février (239,77 ± 18,23) et celles de novembre-décembre (265,44 ± 13,11) (tableau 1). La comparaison des distributions cumulées des âges au premier oestrus montre — une fois de plus — que ce paramètre dépend de la saison de naissance. En effet, 50% des agnelles sont pubères à l'âge de 286 et de 464 jours pour des périodes de naissance respectives de janvier-février et avril-mai (figure 3). Quoique le poids au premier oestrus semble être différent, l'analyse de la variance ne révèle aucune différence significative. 212 Figure 2. Pourcentages cumulés des âges à la première ovulation chez des agnelles D'man nées en avril-mai (3) et en novembre-décembre (2) 160 180 200 220 240 260 280 300 320 Jours 10 Mois Produits de croisement DxS Les agnelles de première génération DxS ont tendance à se placer en position intermédiaire par rapport aux races parentales (tableaux 1 et 2), avec un âge moyen au premier oestrus et à la première ovulation de 275 et de 244 jours respectivement. L'effet de la saison de naissance sur l'âge au premier oestrus est hautement significatif. Cependant, les variations de l'âge à la première ovulation ne sont pas significatives (tableau 2, figure 4). Le poids moyen à la 213 IS■o c o c .CO o X 5 ■^: co CO 9 3> s> => o T3 o " X •^ c 3 c o ? (0 C O (- C CD B § CD ffl ■2 o t o (O CD i. 3 ■q co -3C (0 fc o -= CO c-co u §1 3CO i; T3 CO 5 E ■g c 3 S 'o o E a.10 8 §2 ~ co c o> E o CO ■S * 15 5 Tj CO 13 o "» 3° '& > Q. sfS. O S"* V 1 -a c8 "^ o -CO3 v- c :» u • co • 5 *-~ V- E .»§ co k. . a c CO~ c0 CO■S.E 2 D -CO Q to o £ 9 • o O) CD I«0 = eg 3 O) 3co 0)2 B 5 c8-8 .5 S c C3 p ■ aS T3 3 § • T30 8 ■g 8 Oi CO -0) CO «m a. c 3■• sz £ fl CD If) IO -" ? 5 en d CM CD to IO ■0- CO ao co eg cm" + ' CD CO CM T- 2 2 .8 fc 8 8 8 8 CO CO CM «- 9 2 1 3 c CO E b CM 8 2 CM O r«- r» # in CD T-" ~ CM 1 I + CO o »- •- »- CM in in" l + CO o s + * i 8 8 8 O) CO CO o o d + + «- CM T CO CO CM Or- cm" r»~ 1 + TT CO S 2! $ o + o> •- CM CO CM1 + co co CM r- CO r- co d in «- co + + «- co co »- s COCOCT> T CO LL a 5 1 -3 1 2 < CO C/3 Q 214 Figure 3. Pourcentages cumulés des âges au premier oestrus chez des agnelles Sardi nées en janvier-février (1) en novembre-décembre (2) et en avril-mai (3) 160 180 200 220 240 260 280 300 320 420 440 460 480 500 520 Jours i i i i i i i i 6 7 8 9 10 11 14 15 16 17 Mois puberté est voisin de 26,5 kg chez les agnelles nées en novembre-février et enregistre une diminution (23,37 kg) chez celles nées en avril-mai et une augmentation (28,80 kg) quand la naissance a lieu en octobre. La comparaison des distributions cumulées des âges au premier oestrus, pris deux à deux, par le test de Smirnov montre que les agnelles nées en janvier-février sont plus précoces que celles nées en novembre-décembre, qui à leur tour sont plus précoces que celles nées en octobre (figure 5). Parallèlement à cela, on observe que pour une même saison de naissance, les agnelles sont d'autant plus précoces que leur génotype se rapproche du D'man (figures 6 et 7). Taux d'ovulation à la puberté Chez la race Sardi, le taux d'ovulation moyen est de 1 , 1 1 . Le nombre d'ovulation varie de 1 à 3 avec une prédominance d'ovulations simples (88%). Chez la jeune D'man, ce taux est de 1 ,86. La période et le type de naissance ne semblent pas affecter significativement ce paramètre (tableau 2). 215 Figure 4. Pourcentages cumulés des âges à la première ovulation chez des agnelles (DxS) nées en janvier-lévrier (1), en novembre-décembre (2) et en avril-mai (3) 160 180 200 220 240 260 280 300 320 ' i i i i 6 7 8 9 10 Mois Jours Le taux d'ovulation moyen calculé et ajusté sur quatre périodes de naissance chez le génotype D'man x Sardi (DxS) est intermédiaire entre ceux des races parentales: 1,57. Un taux d'ovulation supérieur à celui de la race parentale la plus prolifique a été observé (2,12) chez les femelles du génotype Dx(DxS). 216 Figure 5. Pourcentages cumules des ages au premier oestrus chez des agnelles DxS nees en janvier-fevrier (1), en novembre-decembre (2), en avril-Mai (3) et % C en octobre (4) 160 180 200 220 240 260 280 300 320 420 440 460 480 Jours 10 11 14 15 16 Mois Figure 6. Pourcentages cumules des ages au premier oestrus des agnelles D, S, DxS et Dx (DxS) nees en avril-mai 160 180 200 220 240 260 280 300 320 420 440 460 480 500 520 10 11 14 15 16 17 Jours Mois 217 Figure 7. Pourcentages cumulés des âges à la première ovulation des agnelles D, DxS et Dx (DxS) x (DxS) 160 180 200 220 240 260 28() 300 320 6 7 8 9 10 11 Discussion Jours Mois Ces observations confirment la précocité sexuelle et la prolif icité de la femelle D'man. L'âge à la puberté est influencé par la saison de naissance puisque les agnelles nées en avril-mai sont pubères à 200 jours alors que celles nées en novembre- décembre le sont deux mois plus tard. Le même effet a été observé chez des agnelles de même race nées en juillet-août (183 jours) comparées à celles nées en octobre-novembre (220 jours) (Lahlou-Kassi, 1982). Il semble résulter de l'interférence de la saison de naissance avec la période de baisse de l'activité sexuelle qui se manifeste par une baisse de l'incidence d'oestrus et d'ovulation spontanés chez la jeune D'man (Lahlou-Kassi, 1982) ou induits par l'effet mâle chez la jeune Barbarine (Khaldi, 1984; Khaldi et Lassoued, 1990). Le 218 poids à la puberté de cette race est positivement corrélé (r=0,73) avec son âge au premier oestrus. Quant aux agnelles Sardi, elles sont pubères à un âge plus tardif et à un poids supérieur comparées aux D'man. Les jeunes femelles nées tôt dans la saison d'agnelage (novembre-décembre) atteignent la puberté la même année alors que celles nées au début de la saison sexuelle de la race (mai-juin) n'ont pas le temps d'atteindre un développement corporel suffisant pour manifester leur premiers signes de chaleurs. Elles ne seront cyclées que la saison de l'année suivante (Foster, 1981). Les produits de croisement atteignent la puberté à un âge intermédiaire aussi bien en croisement D'man x Timahdite (Lahlou-Kassi, 1982) qu'en croisement D'man x Sardi. La diminution de l'âge à la puberté en parallèle à l'augmentation des gènes sexuels de la D'man est une caractéristique génétiquement transmissible de façon additive (Lahlou-Kassi ef a/., 1989). Le taux d'ovulation, qui constitue l'une des composantes principales de la prolificité, est élevé chez la jeune D'man indépendamment de la saison de naissance, tout en restant inférieur à celui observé chez l'adulte (2,85) par Lahlou-Kassi (1982). L'effet de la saison de naissance sur le taux d'ovulation est statistiquement significatif chez les produits de croisement. Il varie de 1,3 à 1,4 quand la puberté survient en période de journées croissantes ou longues (naissances de novembre-décembre ou d'avril-mai) à 1,7-1,8 quand elle survient en période de journées décroissantes. Chez la brebis adulte, le taux d'ovulation est élevé en début de saison sexuelle, puis diminue pour atteindre son minimum juste avant le début de la phase d'anoestrus saisonnier (Scaramuzzi et Radford 1983). Cela est probablement dû à l'effet de la photopériode sur le taux d'ovulation. Par ailleurs, l'évolution du taux de prolificité observé chez la D'man (Boutgayout, 1980) et la Sardi (Berger ef al., 1989) au cours des trois à quatre premiers agnelages résulte de l'augmentation progressive du taux d'ovulation et de la diminution du taux de mortalité embryonnaire (Boujenane, 1989; Bradford ef a/., 1989). Conclusion La race D'man présente des caractéristiques de reproduction marquées par une précocité sexuelle, une prolificité élevée et une non-saisonnalité. Les produits issus du croisement D'man x Sardi présentent des caractéristiques pondérales et de reproduction intermédiaires. Cela témoigne de la transmission génétique de ces paramètres. Les efforts sont actuellement à concentrer sur la création d'une race synthétique à travers les croisements pour stabiliser ces paramètres le long des générations. A nos jours, l'âge/poids optimum à la mise à la reproduction des jeunes antenaises n'est pas encore connu. Ce paramètre doit être déterminé et standardisé pour optimiser la reproduction ovine. 219 References Barki L 1974. Contribution a l'6tude de la race O'man. Analyse des structures ovines dans les zones de Goulmima et Jorf. Memoire de fin d'etude, Ecole nationale d'agriculture, Meknes (Maroc). Berger Y. M., Bradford G. E. , Essadi A., Johnson D.W., Bourfia M. et Lahlou-Kassi A. 1989. Performance of D'man and Sardi sheep on accelerated lambing. III. Lamb mortality, growth and production per ewe. Small Ruminant Research 2:307-321. Bouix J., Kadiri M.et Chari A. 1974. Performances enregistrées dans les troupeaux pepinieres de la race D'man. AlAwamia 52:67- 97. Boujenane I. 1989. Inheritance of reproduction traits in crosses between the D'man and Sardi breeds of sheep. These de doctorat es sciences agronomiques, Institut agronomique et v6terinaire Hassan II, Rabat (Maroc). Boutgayout M. 1980. Bilan de reproduction et contrflle de croissance de deux troupeaux ovins D'man et Sardi. These de doctorat veterinaire, Institut agronomique et veterinaire Hassan II, Rabat (Maroc). Bradford G.E., Lahlou-Kassi A., Berger Y.M., Boujenane I. et Derquaoui L 1989. Performance of D'man and Sardi sheep on accelerated lambing. II. Ovulation rate and embryo survival. Small Ruminant Research 2:241-252. Chaarani B. 1987. Management and productivity of sheep flocks in Meknes province, Morocco, with especial reference to abortion and lamb mortality. These de doctorat es sciences agronomiques, Institut agronomique et veterinaire Hassan II, Rabat (Maroc). Drissi M. 1983. L'elevage ovin dans le perimetre irrigue du Tadla. These de doctorat veterinaire, Institut agronomique et veterinaire Hassan II, Rabat (Maroc). Dyrmundsson O. R. 1973. Puberty and early reproductive performance in sheep. I: ewe lambs. Animal Breeding Abstracts 41 :273-289. Dyrmundsson O.R. 1981. Natural factors affecting puberty and reproductive performance in ewe lambs: a review. Livestock Production Science 8:55-65. Dyrmundsson OR. 1987. Advancement of puberty in male and female sheep. In: New techniques in sheep production. Publie sous la direction de I.F. Marai et J.B. Owen. Butterworth, Londres (R.-U.). p. 65 a 76. El Fakir A., El Yamani A., Oudghir A., Bouzoubaa K. et Derouich A. 1979. La race D'man a l'echelle de la station et a l'echelle de la vallee du Draa, probleme de vulgarisation. Stage de developpement, Institut agronomique veterinaire Hassan II, Rabat (Maroc). Foster D.L 1981. Mechanisms for delay of first ovulation in lambs born in the wrong season (fall). Biology of Reproduction, 25:85-92. Foster D.L, Yellon S. M. et Olster D. H. 1985. Internal and external determinants of the timing of puberty in the female. Journal of Reproduction and Fertility 75:327-344. Glatzel P. 1980. Contribution a l'établissement des normes standards de fecondile pour les beliers de race D'man. Seminaire sur l'elevage ovin D'man, 21 novembre 1980, Ouarzazat (Maroc). Harrouni M. 1977. Etude de quelques parametres de reproduction chez la brebis de race D'man en berceau. These de doctorat vetérinaire, Institut agronomique et veterinaire Hassan II, Rabat (Maroc). Kabbah A. et Berger Y. M. (sous la direction de). 1990. L'elevage du mouton dans un pays a climat meditenanben: Le systeme agro-pastoral du Maroc. Actes Editions, Institut agronomique et veterinaire Hassan II, Rabat (Maroc). Khaldi G. 1984. Variations saisonnieres de I'activite ovarienne, de I'anoestrus postpartum des femelles ovines de race Barbarine: influence du niveau alimentaire 220 et de la presence du male. These de doctorat d'Etat, Universite des sciences et techniques du Languedoc, Montpellier (France). Khaldi G. et Lassoued N. 1990. Caracteristiques de reproduction des femelles ovines de race Barbarine . Conference du Reseau africain de recherche surles petits ruminants, 10-14 decembre 1990, Nairobi (Kenya). Lahlou-Kassi A. 1982. Etude comparee de la dynamique folliculaire cyclique chez des brebis a haut et a bas taux d'ovulation: races D'man et Timahdite. These de doctorat es sciences agronomiques. Institut agronomique et veterinaire Hassan II., Rabat (Maroc). Lahlou-Kassi A. 1987. Principles for indigenous sheep improvement in North Africa. In: Animal genetic resources. Strategies for improved use and conservation . Publie sous la direction de J. Hodges. FAO Animal Production and Health Paper N° 66. FAO, Rome (Italie). p. 77 a 87. Lahlou-Kassi A., Berger Y. M , Bradford G. E., Boukhliq R., Tibary A. et Derquaoui L 1989. Performance of D'man and Sardi sheep on accelerated lambing. I. Fertility, litter size, postpartum anoestrus and puberty. Small Ruminant Research 2:225-239. Land R. B. 1978. Reproduction in young sheep: Some genetic and environmental sources of variation. Journal of Reproduction and Fertility. 52:427-436. MARA (Ministere de ('agriculture et de la reforme agraire, Direction de I'elevage 1990. L'élevage au Maroc. Symposium international sur I'elevage dans les systemes cerealiers mediterraneans, 7-10 octobre 1990, Rabat (Maroc). Quirke J. F. 1979. Effect of body weight on the attainment of puberty and reproductive performance of Galway and Fingalway. Animal Production, 28:297 a 307. Scaramuzzi R.J. et Radford H.M. 1983. Factors regulating ovulation rate in the ewe. Journal of Reproduction and Fertility 69:353- 367. 221 Caractéristiques de reproduction des femelles ovines de race Barbarine G. Khaldi et N. Lassoued Institut national de la recherche agronomique de Tunisie 2080 Ariana (Tunisie) Résumé Les résultats de cette étude montrent que les brebis adultes de race Barbarine sont caractérisées par une longue saison sexuelle qui commence en juillet et se termine en février (242 jours). L'intensité de leur anoestrus saisonnier (printemps) est relativement faible: les pourcentages de brebis ovulant ou extériorisant un comportement d'oestrus au moins une fois par mois ne sont jamais inférieurs à 75 et 25% respectivement. Les possibilités de reproduction des jeunes femelles nées à l'automne sont plus limitées. Atteignant leur puberté à l'âge de 10 mois, leur saison sexuelle ne dure qu'une centaine de jours (septembre-décembre) . Le rétablissement des activités ovarienne et oestrale post-partum des brebis allaitantes n'est pas immédiat. La précocité dépend essentiellement de la saison des agnelages. En effet, si l'intervalle parturition-première ovulation n 'est que de 1 7jours en moyenne chez les femelles agnelant au mois d'octobre, il est environ deux et tois fois plus long lorsque les mises-bas ont lieu aux mois de juin et de février respectivement. De même, la durée de l'anoestrus post-partum (intervalle parturition-premier oestrus) est de l'ordre de 60, 75 et 100 pour les mises-bas d'octobre, juin et février respectivement. La stimulation de l'activité ovarienne des femelles en anoestrus saisonnierpar effet bélier est très intense. L 'ovulation est induite chez 98% des brebis et 75% des antenaises. Les pourcentages de femelles dont le premier cycle ovarien est de courte durée (6 jours) sont de 25 et 35 chez les deux catégories respectivement. La sous-alimentation prolongée des brebis pendant la gestation et la lactation a des répercussions néfastes sur leurs caractéristiques de 223 reproduction, meme quand elles sont suralimentees entre le tarissement et la lutte. Characteristics of reproduction of the Barbary sheep females G. Khaldi and N. Lassoued Abstract The results of this study show that the adult Barbary ewes are characterized by a long sexual season which begins in July and ends in February (242 days). The intensity of their seasonal anoestrus (spring) is relatively low: the percentages of ewes ovulating or displaying oestrus at least once a month are never lower than 75 and 25 %, respectively. The possibilities of reproduction of the young females born during autumn are more limited. Reaching their puberty at the age of 10 months, their sexual season is not longer than 100 days (September-December). The restoration of the post-partum ovarian and oestrous activities is not immediate. Its precocity depends essentially on the lambing season. In fact, if the first postpartum ovulation occurs approximately 17 days after parturition in the October lambing ewes, this interval is twice and three times longer for those lambing in June and February, respectively. On the other hand, the duration of the postpartum anoestrus (interval between parturition and the first oestrus) is about 60, 75 and 100 days when lambing occurs in October, June and February, respectively. The stimulation of the ovarian activity of the seasonal anoestrous females in spring by the ram effect is very intensive. Ovulation is induced in 98 % ofadult ewes and 75 % ofyearlings. The first induced ovarian cycle is of a short duration (6 days) in 25 and 35 % of the two categories of females, respectively. The prolonged undernutrition of ewes during pregnancy and lactation impairs some of their reproductive characteristics even when they are overfed between weaning and joining. 224 Introduction Variations saisonnières de l'activité ovarienne et du comportement d'oestrus cyclique L'étude est réalisée à la station expérimentale de Bou Rébiâa (36°38' nord et 10°07' est), située à 25 km au sud-est de la ville de Tunis, sur 25 brebis adultes âgées de 4 à 5 ans et 26 agnelles (nées en automne) âgées de 5 mois au début de l'expérience. Toutes les femelles sont de race Barbarine à tête rousse. En plus du pâturage, elles reçoivent une complémentation composée de foin ou d'ensilage et d'aliment concentré dont les quantités varient selon la saison. Elles subissent une détection biquotidienne de chaleurs à l'aide de béliers entiers et des endoscopies tous les 17 jours pendant 15 mois. Résultats Saison sexuelle Le début de la saison sexuelle des agnelles nées en automne se situe pendant la première moitié du mois de septembre de l'année suivante (figure 1). Agées alors de 10 mois, ces agnelles ont un poids vif moyen de 34,7 kg. Leur saison sexuelle est courte puisque la date moyenne du dernier oestrus est le 3 décembre. Sa durée moyenne est de 104 jours. La saison sexuelle des brebis adultes est significativement plus longue (P <0,01 ) , la date moyenne du premier oestrus se situant pendant la première quinzaine du mois de juillet et celle du dernier oestrus pendant la deuxième quinzaine du mois de février. Sa durée moyenne est de 242 jours. L'intensité de l'anoestrus saisonnier est également en rapport avec l'âge des femelles. En effet, 25 a 40% des brebis continuent à extérioriser un comportement d'oestrus au printemps alors que l'arrêt de l'activité oestrale est quasi total chez les agnelles pendant la même période. Activité ovarienne Quel que soit l'âge des femelles, il existe une dissociation oestrus-ovulation. Des ovulations silencieuses se produisent tout au long de l'année mais plus particulièrement avant le début de la saison sexuelle, après l'arrêt du comportement d'oestrus cyclique et en avril-mai. Le taux d'ovulation reste faible (1 ,08) et relativement constant chez les agnelles. En revanche, il existe une variation saisonnière importante chez les brebis adultes. Chez celles-ci, le taux d'ovulation passe par un maximum en septembre-octobre (1,60) et un minimum en mars-avril (1,10). 225 Figure 1 . Pourcentage de brebis ayant au moins un oestrus ou une ovulation par mois. ovulation oestrus % 100-1 r 75 50 25 En considérant une période de un an (1or juillet—30 juin), letaux moyen d'ovulation des brebis (1 ,32) est significativement plus élevé (P <0,05) que celui des agnelles (1,08). Le taux d'ovulation est plus faible lorsque l'ovulation est silencieuse. C'est ainsi que chez les brebis et les agnelles, il passe respectivement de 1 , 1 6 et 1 ,05 dans le cas des ovulations silencieuses à 1,38 et 1,12 lorsque les ovulations sont accompagnées de comportement d'oestrus. Reprise de l'activité sexuelle post-partum 131 brebis de race Barbarine à tête rousse sont utilisées pour cette expérience à la station expérimentale de Bou Rébiâa. 53 agnèlent entre le 1 4 et le 30 octobre, 35 entre le 12 et le 26 février et 43 entre le 2 et le 28 juin. Pour chacune de ces périodes, la prolificité moyenne est de 120, 140 et 126% respectivement. Toutes les brebis sont adultes (3-5 ans) et leur poids vif moyen 48 heures après 226 parturition est proche de 50 kg. A chaque saison d'agnelage, le tarissement des brebis est effectué pour la moitié des femelles à 45 jours post-partum et pour l'autre à 90 jours, à l'exception de l'agnelage du mois d'octobre, où un troisième groupe est tari 48 heures après la mise-bas. En plus des pâturages, les brebis reçoivent du foin de vesce-avoine à volonté et 300 g d'aliment concentré/tête/jour. L'activité ovarienne des brebis est contrôlée par endoscopie tous les 17 jours et la détection des chaleurs est biquotidienne. Résultats Influence de la durée d'allaitement Quelle que soit la saison d'agnelage, la durée de l'allaitement (45 ou 90 jours) n'a aucune influence significative sur la reprise de l'activité ovarienne post-partum des brebis. Par ailleurs, le tarissement des femelles 48 heures après la parturition n'entraîne pas une diminution significative de l'intervalle mise-bas-première ovulation (tableau 1). Tableau 1 . Influence de la durée d'allaitement et de la saison de mise-bas sur l'intervalle parturition- 1 ,e ovulation (j) Mois de mise-bas Durée d'allaitement Février Juin Octobre 2 jours 45 jours 90 jours 15,1 ± 5,5 43,4 ± 19,8 51,4 ± 14,8 30,3 ± 14,9 33,3 ± 9,8 16.5 ± 9,1 17.6 ±8,1 Pour une saison donnée d'agnelage, la date du tarissement par rapport à la parturition (45 ou 90 jours) n'a aucune conséquence sur les proportions de femelles ayant présenté au moins un oestrus pendant les huit mois qui suivent la mise-bas et sur l'intervalle moyen parturition-premier oestrus de ces brebis pendant la même période (tableau 2). En revanche, le tarissement très précoce (48 heures après la mise-bas) au mois d'octobre réduit d'environ 25 jours la durée de l'anoestrus post-partum. Influence de la saison d'agnelage Que la durée d'allaitement soit de 45 ou de 90 jours, la saison d'agnelage a une influence considérable (P<0,01) sur l'intervalle parturition-première ovulation. En effet, si cet intervalle n'est que de 17 jours en moyenne chez les brebis agnelant 227 au mois d'octobre, il est environ deux fois (32 jours) et trois fois (47 jours) plus long pour les femelles agnelant respectivement aux mois de juin et de février. Tableau 2. Pourcentages de brebis présentant au moins un oestrus moins de 240 jours après parturition et durée de l'intervalle mise-bas-premier oestrus (j) Mois de mise-bas Durée de l'ai laitement Février Juin Octobre 2 jours % - - 100 Durée - - 32,1 ± 12,9 45 jours % 82,3 100 100 Durée 97,7 ± 12,4 73,5 + 21,4 59,0 ± 24,9 90 jours % 77,7 90,0 77,7 Durée 103,9 ± 8,2 76,2 ± 30,3 60,0 ± 49,7 La première ovulation est silencieuse chez 95% des brebis et ceci quelle que soit la période de mise-bas. La durée d'allaitement (45 ou 90 jours) ne modifiant pas la reprise du comportement d'oestrus cyclique pendant la période post-partum, l'influence de la saison de mise-bas a été étudiée en considérant l'ensemble des femelles de chaque groupe. Les pourcentages de femelles présentant au moins un comportement d'oestrus pendant la période d'observation de huit mois sont respectivement de 90,3, 85,7 et 80,0 pour les agnelages d'octobre, de juin et de février (P<0,05). Pour ces femelles, l'intervalle moyen parturition-premier oestrus est de l'ordre de 60, 75 et 100 jours respectivement. Rupture de l'anoestrus saisonnier par effet bélier Cette étude est réalisée à la station expérimentale d'Ousseltia (35°5V nord et 9°35' est), située à 160 km au sud-ouest de la ville de Tunis. Elle est menée sur 160 brebis adultes (5,1 ± 1,6 ans) et 40 antenaises (1,5 ±0,1 an) de race Barbarine à tête rousse. Toutes les femelles sont sèches et complètement isolées des mâles. Ces derniers sont introduits dans le troupeau à la fin du mois d'avril à raison d'un bélier pour 10 femelles. Tous les animaux sont conduits sur des parcours naturels sans aucune complémentation. Le contrôle d'oestrus est biquotidien et l'activité ovarienne est étudiée par coelioscopie des ovaires le jour de l'introduction des mâles et neuf jours plus tard. 228 Résultats Activité ovarienne L'âge des femelles a une influence hautement significative (P<0,01) sur leur activité ovarienne spontanée avant l'introduction des mâles dans le troupeau (tableau 3). En effet, environ 50% des brebis contre 22,5% des antenaises ont des corps jaunes le jour de l'introduction des béliers. Tableau 3. Pourcentages de femelles cycliques avant l'introduction des mâles Femelles cycliques Categorie Nombre de femelles Nombre % Brebis Antenaises 160 81 50,6 9 22,540 La stimulation de l'activité ovarienne des femelles non cycliques par effet mâle est très intense (tableau 4). L'ovulation est induite chez la quasi-totalité des brebis (97,5%) et les trois quarts des antenaises (74,2%). La différence entre les deux catégories de femelles est hautement significative (P<0,01). L'ovulation induite apparaît au cours des trois premiers jours de lutte. Tableau 4. Pourcentages de femelles non cycliques dont l'ovulation est induite par effet mâle Femelles non cycliques Femelles ovulant Categorie Nombre % Brebis Antenaises 79 77 97,5 74,231 23 Quel que soit l'âge des femelles, la durée du premier cycle ovarien n'est pas toujours normale. La proportion de femelles dont le premier cycle ovarien est de courte durée (6 jours) est de 23,4% pour les adultes et de 34,8% pour les antenaises (différence non significative). Comportement d'oestrus Parmi les femelles ayant ovulé à la suite de l'introduction des béliers, 87,0% des brebis et 73,9% des antenaises extériorisent un comportement d'oestrus au cours des 26 premiers jours de lutte. Quel que soit l'âge de ces femelles, le comportement d'oestrus apparaît vers le 1 7e jour lorsque le premier cycle ovarien est normal et vers le 23e jour lorsque ce dernier est de courte durée. 229 Influence du niveau alimentaire sur la réponse des brebis à l'effet mâle 1 10 brebis adultes de race Barbarine ayant un poids vif moyen de 51 kg sont réparties en 2 lots Haut (H) et Bas (B) pendant les 12 dernières semaines de gestation et en 4 sous-lots pendant les 18 semaines d'allaitement (HH, HB, BH et BB) selon leur niveau alimentaire: • lot HH: haut avant et après mise-bas • lot HB: haut avant mise-bas et bas après • lot BH: bas avant mise-bas et haut après • lot BB: bas avant et après mise-bas Les agnelages ont lieu au mois d'octobre. Après le sevrage (28 février), le poids vif moyen des brebis des quatre lots est ramené à une valeur sensiblement identique (47 kg environ) à la date d'introduction des béliers (30 avril). La détection des chaleurs est biquotidienne et l'activité ovarienne des brebis est contrôlée par une série d'endoscopies à J1 , J4, J9 et 8 à 12 jours après la date d'apparition du premier oestrus. Résultats Les variations du poids vif des brebis sont représentées dans la figure 2. Activité ovarienne La sous-alimentation à long terme avant et/ou après la mise-bas réduit de moitié le pourcentage de femelles cycliques avant l'introduction des béliers (tableau 5). Tableau 5. Pourcentages de brebis cycliques avant l'introduction des mâles Femelles cycliques Lots Nombre de femelles Nombre % HH 32 21 66 HB 26 8 31 BH 31 11 35 BB 21 7 33 L'ovulation est induite chez la totalité des femelles non cycliques des quatre lots dans les quatre jours qui suivent l'introduction des mâles (tableau 6). La sous-alimentation des brebis pendant la gestation et/ou la période d'allaitement ne semble donc pas affecter leur réponse à l'effet bélier à condition de les bien 230 Figure 2. Evolution du poids vif du brebis poids 60- 50 40 12 Niveau alimentaireHB HH BB BH -50-.S- 12 16 mise bas sevrage 4 semaines Tableau 6. Reponse a l'effet male et pourcentages de cycles courts Femelles ovulant Cycles courts Lots Nombre N % N % HH 11 11 100 5 45 HB 18 18 100 12 67 BH 20 20 100 15 75 BB 14 14 100 11 79 231 nourrir entre le sevrage et la lutte. La sous-alimentation avant et/ou après la mise-bas augmente le pourcentage de cycles ovariens de courte durée qui est plus élevé dans les lots BH (75%), HB (67%) et BB (79%) que dans le lot HH (45%). Performances de reproduction Le taux d'ovulation induite varie entre 1 ,27 (HH) et 1 ,79 (BB). Au premier oestrus, ce taux varie entre 1,12 (HB) et 1,36 (HH). Le tableau 7 montre que la sous-alimentation prolongée de part et d'autre de la mise-bas provoque une diminution sensible de la fertilité des brebis, même si elles sont suralimentées après le sevrage. Tableau 7. Taux d'ovulation et fertilité des brebis (%) Lots Ovul induite 1" oestrus Fertilité HH HB BH BB 1.27 1,36 91 1,35 1,12 94 1,50 1,30 100 1,79 1,31 77 232 The relative importance of tactile, visual, auditory and olfactory stimuli in oestrus detection in West African dwarf ewes G.O.Oyediji, M.O. Akusu1and G.N. Egbunike 1 University of Ibadan, Nigeria Abstract The response of ewes in oestrus to different stimuli was studied in five normocyclic West African dwarf ewes during five successive oestrous cycles. Sniffing around the perineum was significantly superior (P<0.05) to other stimuli in evoking the characteristic "looking-over-the-shoulder" behaviour of ewes in oestrus and was not affected by masking the ram's odour with an odoriferous substance. Visual stimuli by a dummy ram evoked the response in 8% of the ewes but acted synergistically with the auditory stimuli, when accompanied by a pre-recorded ram's sexual vocalisation to increase the percentage response to 26. Olfactory stimuli with a linen, which was liberally rubbed on an intact ram, did not evoke the response. It is concluded that oestrus detection in ewes may be possible using artificial stimuli in intensive sheep husbandry. Importance relative des stimuli tactiles, visuels, auditifs et olfactifs dans la detection des chaleurs chez les brebts Djallonke G.O. Oyediji, M.O. Akusu et G.N. Egbunike Resume La reaction a differents stimuli a ete étudiee au cours de cinq cycles successifs. Le reniflement du perinee s'est avere significativement (P<0,05) plus efficace que les autres stimuli a declencher le comportement caracteristique de la brebis en chaleur, laquelle "jette 233 des regards par-dessus l'epaule". Cette faculte fut conservee lorsqu'on a fait disparaitre l'odeur du belier avec des substances odoriferantes. L 'utilisation d 'un faux belier declenchait des reponses positives dans 8% des cas, chiffre qui montaita 26% lorsque celle-ci etait synchronisée avec un stimulus auditif, en l'occurrence l'enregistrement des appels d'un male a facte sexuel. En revanche, I'utilisation, comme stimulus olfactif, d'un morceau de tissu qu'on avait précédemmentpasse surle corps d'un belier entier laissait les brebis d'une superbe indifférence. Ces resultats montrentqu'il estpossible, notammenten elevage ovin intensif, de détecter les chaleurs des brebis par des stimuli artificiels. Introduction Oestrus behaviour is the most definite sign of oestrus in female animals and it has been reported for various species. The ewe does not exhibit oestrus behaviour in the absence of a ram (Terrill, 1975). It is therefore necessary to devise artificial teasers for oestrus detection in ewes under commercial or large flocks for the application of modern breeding techniques such as oestrus synchronisation and artificial insemination. This will involve identification of the stimuli involved in the "looking over the shoulder" which is characteristic of ewes in oestrus (Hafez et al 1969). The objective of this study was to determine the characteristic behaviour of West African dwarf ewes and to identify the areas of the body of the ewe that evoke sexual arousal leading to the shoulder-looking behaviour of ewes in oestrus. Materials and Methods The experimental animals consisted of five normocyclic pluriparous WAD ewes aged 4-5 years and weighing 25-35 kg. Oestrus was synchronised with prostaglandin F2-alpha (UpJohn Company, Kalamazoo, USA) as described previously in the WAD goat (Akusu and Egbunike, 1984). Following the second PGF2-injection ewes were exposed to various stimuli as follows: 1. A dummy ram prepared with the skin of a 3-year old ram after slaughter (Stimulus A: Visual) 2. A dummy ram as in (1) but accompanied with a pre-recorded ram's characteristic sexual vocalisation (Stimulus B: Visual/Auditory) 3. An intact apronned ram (Stimulus C: Tactile) 234 4. An intact apronned ram liberally bathed with an odoriferous substance (Perfekthion Bayer A.G., Germany) to mask the ram's odour (Stimulus D: Olfactory) 5. A linen which has previously been rubbed generously on an intact ram (Stimulus E: Olfactory). Results and discussion The response of ewes to stimulus C was scored as described by Denenberg and Banks (1969) and Clayton et al (1981). The areas of the body investigated were the hindquarters, vulva, tail, abdomen and ears. A score of 1 was assigned to positive response (Iooking over the shoulder) while negative response was scored 0. For a comparison, stimuli responses were expressed as percentage of the observations. Table 1 . Comparative scores for the different erogenic areas of the ewe (x ± Sem). Body area x ± Sem Hindquarters 2.8 ± 0.09a Vulva 2.6 ± 0.13" Tail 2.4 ± 0.1 6b Abdomen 1.1 ± 0.17° Ears 1.1 ±0.17° Means with different superscripts are significantly (P<0.05) different. The relative positive response of the ewe to areas of the body investigated with stimulus C is summarised in Table 1 . Responses to the ram's tactile stimulation of the hindquarters and the vulva were significantly superior to the stimulation of the tail (P<0.05) which was in itself superior (P<0.05) to stimulation of the abdomen and the ears. The response of ewes to stimulus A was positive in 8% of observations. When accompanied with the ram's sexual vocalisation (Stimulus B), 26% of the ewes responded positively to the dummy. All ewes in oestrus elicited the response in the presence of an intact ram (Stimulus C) and this was not affected by masking the ram's odour with an odorifirous substance (Stimulus D). On the other hand, ewes in oestrus did not respond to the linen previously rubbed on an intact ram (Stimulus E). The main sign of oestrus in the ewe observed in this study was "looking over the shoulder". This is in agreement with the report of Tomkins and Bryant (1974). 235 The hindquarters and the perineal region of the ewe may possess abundant tactile receptors in comparison to the abdomen and the ears. It is known that oestradiol-17B is responsible for the psychic manifestation of behavioural oestrus in domestic animals (Scaramuzzi, 1975; McDonald, 1977). However, Glengross et al (1981) observed that the mechanism of uptake of oestradiol by the hypothalamus could affect oestrus behaviour. It is therefore probable that the ewe produces less oestradiol-17B receptors during oestrus than other domestic animals such as the goat (Akusu, 1987) that displays behavioural oestrus characteristics in the absence of the buck. Visual stimuli alone appeared to have a poor effect on the ewe in oestrus. However, response was improved by auditory stimuli and therefore both stimuli may be acting synergestically. The non-response to olfactory stimulus could not be explained. Pheromones of rams are located in the wool (Knight and Lynch, 1980). Hence, the non-response to the ram's odour may be due either to inadequate pheromones on the linen or the non-involvement of pheromones in the "looking-over-the-shoulder" response of ewes in oestrus. The latter suggestion is buttressed by the fact that response was unaffected by masking the ram's odour with an odoriferous substance. However, Baldwin and Meese (1977) and Alexander and Stevens (1982) reported that the scent of olfaction is very strong in the ewe. Therefore, it is possible that the ewe could discriminate the ram's pheromones from a variety of odoriferous substance. Conclusion This study has demonstrated that tactile, visual and auditory stimuli are important components in the characteristic "looking over the shoulder" behaviour of ewes in oestrus. It may be practicable to devise artificial oestrus detection stimuli consisting of blowing air around the perineum which simulate the ram's "nosing" around this area, a dummy ram and pre-recorded ram's sexual vocalisation. References Akusu M 0. 1 987. Ovarian activities and reproductive potentials of the West African Dwarf goat in Ibadan. PhD thesis, University of Ibadan, Nigeria. Akusu M O and Egbunike G N. 1984. Fertility of the West African Dwarf goat in its native environment following PGF2-alpha induced oestrus. Veterinary Quarterly 6:173-176. Alexander G and Stevens D. 1982. Failure to mask lamb odour with odouriferous substances. Applied Animal Ethology 8:253-260. Baldwin B A and Meese G B. 1977. The ability of sheep to distinguish between conspecifics by means of olfaction. Physiology of Behaviour. 48:251-260. Clayton H M, Lindsay F E F, Forbes A C and Hay LA. 1981 . Some studies of comparative aspects of sexual behaviour in ponies and donkeys. Applied Animal Ethology 7:169-174. 236 Denenberg V H and Banks E M. 1969. Techniques of measurement and evaluation. In: Hafez E S E (ed), The behaviour of domestic animals. Bailliere, Tindall and Cassell Ltd., London, UK. pp. 192-233. Glengross R G, Esslemont R J, Bryant R J and Pope G S. 1981. Relationships between the incidence of preovulatory behaviour and the concentrations of oestradiol 17-beta and progesterone in bovine plasma. Applied Animal Ethology 6:141-148. Hafez E S E, Cairns R B, Hulet C V and Scott J K 1969. The behaviour of sheep and goats. In: Hafez E S E (ed), The behaviour of domestic animals. Bailliere, Tindall and Cassell Ltd., London, UK. pp. 296-348. Knight T W and Lynch P R. 1980. Sources of ram pheromones that stimulate ovulation in the ewe. Animal Reproduction Science 3:133-136. McDonald L E (ed). 1977. Veterinary endocrinology and reproduction. Second edition. Lea and Febiger, Philadelphia, USA. Scaramuzzi R J. 1975. Inhibition of oestrus behaviour in ewes by passive immunization against oestradiol 1 7-beta. Journal ot Reproduction and Fertility 42: 1 45-1 48. Terrill C E. 1975. Sheep. In: Hafez E S E (ed), Reproduction in farm animals. Third edition. Lea and Febiger, Philadelphia, USA. pp. 265-274. Tomkins T and Bryant M J. 1974. Oestrus behaviour of the ewe and the influence of treatment with progestagen . Journal of Reproduction and Fertility 41 : 1 21-1 32. 237 Peripheral plasma levels of progesterone and oestradiol-17 of West African Dwarf goats during the oestrous cycle M.O. Akusu, E. Nduka and B.A. Soyebo University of Ibadan Ibadan, Nigeria Abstract Oestrus was synchronised with PGF^alpha in four pluriparous West African Dwarf (WAD) goats and blood collected through jugular venupuncture. The length of the oestrous cycle and oestrus duration were observed in three successive cycles while peripheral plasma progesterone and oestradiol-176 levels were determined by the radioimmunoassay technique. The mean oestrous cycle lengths and oestrus duration were 19.4 ± 1.7 d and 33.3 ±6.6 h, respectively. Progesterone levels were very low during the first four days of the cycle (day 1 = day of oestrous) with the lowestlevel of 0.3 ±0.02 nglml being recorded on day 2 of the cycle. There was a steady rise in the levels during the cycle until day 15 when the highest level of 2.2 ± 0.05 nglml was recorded which was followed by a sharp decline on day 20 (P <0.05). The profile of oestradiol-1 7s was erratic during the cycle. Mean maximum and minimum concentrations were recorded on day 1 (152.6 ± 31.6 nglml) and day 2 (58.8 ± 22.6 nglml), respectively. These results showed that while progesterone and oestradiol profiles in the WAD goat followed the luteal and follicular phases of the oestrous cycle, progesterone profile is more reliable in assessing the phases of the oestrous cycle. 239 Teneur en progestérone et en oestradiol-17B du sang périphérique chez la chèvre naine d'Afrique de l'Ouest au cours du cycle oestral M.O. Akusu, E. Nduka et B.A. Soyebo Résumé L'oestrus de quatre chèvres naines d'Afrique occidentale multipares a été synchronisé avec du PGFTalpha; des échantillons de sang ont été collectés à travers la veine jugulaire. Les durées du cycle oestral et de l'oestrus de trois cycles successifs ont été mesurées, et les taux de progestérone et d'oestradiol-176 déterminés par radio-immunologie. La durée moyenne du cycle oestral était de 19,4 ± 1,7 jours et celle de l'oestrus de 33,3 ± 6,6 h. Très faible au cours des quatre premiers jours du cycle, la teneur en progestérone du sang périphérique atteignait son niveau minimum au deuxième jour du cycle, avec une valeur de 0,3 ± 0,02 nglml. Elle augmentait ensuite régulièrementjusqu'au 15* jour du cycle où elle atteignait le pic de 2,2 ± 0,05 nglml, avant d'enregistrer une chute brutale au 20" jour (P<0,05). Le taux d'oestradiol-U* a varié de manière erratique; les teneurs maximum (152,6 ± 31,6 nglml) et minimum (58,8 ± 22,6 nglml) ayant été enregistrées respectivement au premier et au deuxièmejours du cycle oestral. Ces résultats montrent que, bien que les niveaux d'oestradiol et de progestérone suivent les phases lutéique et folliculaire, la progestérone est plus indiquée pour la détermination de la phase du cycle oestral à laquelle se trouve la chèvre naine d'Afrique de l'Ouest. Introduction Progesterone and oestradiol-17B are ovarian hormones that have several practical biological applications in the management of reproduction in animals. They are involved in oestrus manifestations, the ovulatory process, cyclic regression of the corpus luteum, establishment of pregnancy and parturition. Progesterone levels have also been used in the determination of the onset of puberty in cattle (Diaz et al, 1986), pigs (Shille et al, 1979) and sheep (Oyedipe et al, 1986) and verification of the accuracy of oestrus detection (Sawyer et al, 1986). A knowledge of the levels of these hormones can be of benefit for diagnostic and therapeutic purposes in cases of endocrine dysfunction involving normal sexual activity of the female. 240 The levels of these hormones during the oestrous cycle and pregnancy have been determined in several species of domestic animals, especially in sheep and cattle (Scaramuzzi, 1976; Adeyemo and Health, 1980; Voh et al, 1987). Compared to the other species, fewer reports are available for the goat (Mori and Kano, 1984; Akusu et al,1989). The objective of this study was to generate basal data on the levels and profiles of progesterone and oestradiol-17B during oestrous cycle of the West African Dwarf goat. This breed of goat is the most numerous of the ruminant species in the humid rainforest zone of southern Nigeria (FAO, 1966; Carew, 1982) and therefore represents the quickest means of increasing meat production for the populace through oestrous synchronisation and artificial insemination. Materials and Methods Four pluriparous WAD goats, aged 3-4 years and weighing between 15-20 kg were used for this study. They were managed intensively on concrete-floored pens in the small ruminant unit of the Department of Veterinary Surgery and Reproduction, University of Ibadan. Animals were fed on a corn-based ration (0.5 kg/d). Clean water and Giant Stargrass were constantly available. Oestrus was synchronised with 1 0 mg PGF2-a as described previously (Akusu and Egbunike, 1984; Akusu et al, 1989). Animals were teased for oestrus five times daily at intervals of 4 hours between 08.00 and 24.00 h with an apronned sexually mature buck. Animals were bled through jugular venupuncture and plasma stored at-20°C until assayed for progesterone and oestradiol-17B by radioimmunoassay techniques (Dada et al, 1 984; Akusu et al, 1 989) . Hormone concentration was calculated from linearised logit-log plots of standard curves. Inter- and intra-assay variations for progesterone were 11.5% and 10.9%, the corresponding values for oestradiol-17B were 10.3% and 11.5%, respectively. Blood was collected during the first five days of the oestrous cycle (day 1 = day of oestrus) and then on days 10, 15 and 20. The study was carried out during three successive oestrous cycles. Results Oestrus was synchronised in all goats treated with PGF2-a.The mean oestrus duration was 33.3 ± 6.6 h while oestrous cycle occurred at 19.7 ± 1.7 d. Peripheral plasma progesterone and oestradiol-17B profiles are summarised in Figure 1 . The levels of progesterone were generally lower than 1 ng/ml during oestrus and metoestrus (d 1-d 4). The lowest level of 0.3 ± 0.02 ng/ml was 241 Figure 1 . Peripheral plasma levels of progesterone and oestradiol-17B of West African Dwari goats during the oestrous cycle. A A Progesterone Oestradiol Progesterone (ng/ml) 2.50 -i 2.00 1.50 1.00- 0.50- Oestradiol 1 7 /3 (pg/ml) -250 200 150 100 50 12 3 4 5 10 15 20 Days determined on d 2 of the cycle. This was followed by a gradual rise in d 3 and d 4 (0.5 ± 0.03 ng/ml and 0.87 ± 0.02 ng/ml, respectively). D4 value was significantly higher than d 2 (P <0.05) . There was a sharp rise on d 5 and remained stable up to d 10 and a further rise in progesterone level to a peak value of 2.2 ± 0.05 on d 1 5 followed by a sharp decline (P <0.05) by d 20. There were differences in the levels between the three cycles but the profiles were not affected (Figure 2) . 242 Figure 2. Peripheral plasma progesterone profiles during three successive oestrous cycles in WAD goats (X ± Sem). Progesterone (ng/ml) 2.4 2.0 1.6 1.2 m EjSli fKgHfc m PHASE 1 PHASE 2 CYCLES PHASE 3 CYCLE 1 CYCLE 2 Er&a CYCLE 3 DAY 20 and DAY 1 (PHASE 1) DAY 2 DAY 4 (PHASE 2) DAY 5 DAY 15 (PHASE 3) Peripheral plasma levels of oestradiol-17B were highest on d 1 (152.6 ± 31.6). There was a rapid decline (P<0.05) on d 2. Thereafter, there were fluctuations in the levels but a gradual rise was observed up to d 20 (131 .7 ± 44.3 ng/ml). The profile of oestradiol was similar irrespective of the cycle (Figure 3). 243 Figure 3. Peripheral plasma oestradiol- 1 7s profiles in WAD goats during three successive oestrous cycles. Oestradiol 1 7 /3(pg/ml) 240 H 200 120 40- IBvim mi HU CYCLE 1 j ] CYCLE 2 I CYCLE 3 PHASE 1 PHASE 2 PHASE 3 Discussion The general profile of progesterone and oestradiol- 1 7B of the West African Dwarf goat was similar to reports in various species of domestic animals (Shearer et al, 1972; Sarda et al, 1973; Perera et al, 1978; Arora and Pandey, 1982; Oyedipe et al, 1986; Danell, 1987). The mean level of progesterone of the WAD goat was lower than the values reported for other breeds of goats (Heap and Linzell, 1 966; Thomburn and Schneider, 1972). This may be a breed characteristic which has been reported in the ewe (Bindon et al, 1979). The individual variations in the peak values of progesterone and oestradiol-17B in the goats may be due to variations in individual traits that may be correlated with expression of overt heat signs and may serve as an index of selection (Perera, 1979). However, it was observed that all does had a common pattern 244 with relatively high levels of progesterone during the luteal phase, a decline between the 1 5th and 20th d of the cycle and values approximating zero at and around oestrus. Similarly, the peak values of oestradiol-17B generally corresponded with the follicular and luteal phases of the oestrous cycle. It showed a fluctuating level of between 58 and 92 pg/ml until the day preceding the onset of behavioural oestrus when an increase was demonstrated. The peak value was for a day and there was no subsequent rise. This was consistent with the report of Shearer et al (1972) and therefore suggested that the psychic manifestation of behavioural oestrus in the goat is the result of oestradiol acting on the central nervous system (McDonald, 1977; Hansel and Convey, 1983). Conclusion These results showed that while progesterone and oestradiol profiles in the WAD goat followed the luteal and follicular phases of the oestrous cycle, progesterone profile is more reliable in assessing the phases of the oestrous cycle. References Adeyemo O and Health E. 1 980. Plasma progesterone concentration in Bos taurus heifers. Theriogenology 14:41 1-420. Akusu M O and Egbunike G N. 1984. Fertility of the West African dwarf goat in its native environment following PGF2 induced oestrus. Veterinary Quarterly 6:173-176. Akusu M O, Nduka E and Egbunike G N. 1989. Peripheral plasma levels of progesterone and oestradiol-17B during the reproductive cycle of West African Dwarf goats. In: Wilson R T and Azeb Melaku (eds), African Small Ruminant Research and Development. Proceedings of a Conference held at Bamenda, Cameroon, 18-25 January 1989. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 316-328. Arora R C and Pandey R S. 1982. Pattern of plasma progesterone, oestradiol-178, luteinizing hormone and androgen in non pregnant buffalo (Bubalus bubalis). Acta endocrinologica 100:279-284. Bindon B M, Blanc M R, Pelletier J, Terqui M and Thimonier J. 1979. Peri-ovulatory gonadotropin and ovarian steroid patterns in sheep of breeds with differing fecundity. Journal of Reproduction and Fertility 55: 1 5-25. Carew B A R. 1 982. Production potentialand nutritional studies ofgoats and sheep in south western Nigeria. PhD thesis, University of Ibadan, Nigeria. Dada O A, Osinusi B O, Nduka E U, Osotimehin B O and Ladipo O A. 1 984. 1 7B oestradiol , progesterone and testosterone in the normal menstrual cycle of Nigerians. International Journal of Gynaecology and Obstetrics 22:151-154. Danell B. 1 987. Oestrous behaviour, ovarian morphology and cyclical variation in follicular system and endocrine pattern in water buffalo heifers. PhD thesis, Swedish University of Agricultural Sciences, Uppsala, Sweden. Diaz T, Manco M, Troconiz J, Benacchio N and Verde 0. 1986. Plasma progesterone levels during the oestrous cycle of Holstein and Brahman cows. Carora type and cross-breed heifers. Theriogenology 26:419-432. 245 FAO (Food and Agriculture Organization of the United Nations). 1966. Agricultural development in Nigeria 1965-1980. FAO, Rome, Italy. 512 pp. Hansel W and Convey E M. 1983. Physiology of the oestrous cycle. Journal of Animal Science 57: Supplement 2:404-423. Heap R B and Linzell J L. 1966. Arterial concentration, ovarian secretion and mammary uptake of progesterone in goats during the reproductive cycle. Journal of Endocrinology 36:389-399. McDonald L E. 1977. Veterinary endocrinology and reproduction. 2nd edition. Lea and Febiger, Philadelphia, USA. 560 pp. Mori Y and Kano Y. 1984. Changes in plasma concentrations of luteinizing hormone, progesterone and oestradiol in relation to the occurrence of luteolysis, oestrus and time of ovulation in the Shiba goat. Journal of Reproduction and Fertility 72:223-230. Oyedipe E O, Pathiraja N, Edqvist L E and Buvanendran V. 1986. Onset of puberty and oestrous cycle phenomena in Yankasa ewes as monitored by plasma progesterone concentrations. Animal Reproduction Science 12:195-199. Perera B M A 0. 1979. Radioimmunoassay of progesterone in the blood of buffaloes during normal and prostaglandin synchronised oestrous cycles. In: Sastradipradja 0 (ed), Proceedings ofthe Second Research Coordination Meeting ofthe Joint FAOIIAEA Division ofAtomic Energy in Food and Agriculture held in Bogor, Indonesia. pp. 65-72. Perera B M A O, Pathiraja N, Buvanendran V, Abeywardena S A and Piyasena R D. 1978. Plasma progesterone levels during natural and prostaglandin-synchronised oestrous cycles in buffaloes. Ceylon Veterinary Journal 26:29-34. Sarda I R, Robertson H A and Smeaton T C. 1973. Sequencial changes in plasma progesterone levels in the ewe during the oestrous cycle and during pregnancy in intact and ovariectomized sheep. Canadian Journal of Animal Science 53:25-34. Sawyer G J, Russel-Brown I D and Silcock J K. 1986. A comparison of three methods of oestrus detection in commercial dairy herds verified by serum progesterone analyses. Animal Production Science 10:1-10. Scaramuzzi R J. 1976. Inhibition of oestrous behaviour in ewes by passive immunization against oestradiol 1 7-Beta. Journal of Reproduction and Fertility 42: 1 45-1 48. Shearer I J, Purvis K, Jenkins G and Haunes N B. 1972. Peripheral plasma progesterone and oestradiol-178 levels before and after puberty in gills. Journal of Reproduction and Fertility 30:347- 360. Shille V M, Karlbom I, Einarsson S, Larsson K, Kindahl H and Edqvist L E. 1979. Concentrations of progesterone and 15-keto-13, 14-dihydroprostaglandin F2 in peripheral plasma during the oestrous cycle and early pregnancy in gilts. Zoobiology and Veterinary Medicine. A 26: 1 69-1 81 . Thornburn G D, and Schneider W. 1972. The progesterone concentration in the plasma of the goat during the oestrous cycle and pregnancy. Journal of Endocrinology 52:23-36. Voh Jr A A, Oyedipe EO, Pathiraja N, Buvanendran V, and Kumi-Diaka J. 1987. Peripheral plasma levels of progesterone in Nigerian zebu cows following synchronization of oestrus with prostaglandin F2-Alpha analogue (Dinoprost Tromethamine). British Veterinary Journal 143(3):254-263. 246 Session 3 Small ruminant reproductive wastage and health Pertes en reproduction et I'etat sanitaire des petits ruminants Quelles peuvent etre les priorites de recherche dans le domaine de la pathologie des petits ruminants en Afrique? F Thiaucourt{2\ J. Fikre\ J.J. Tulasne2, G. Mebratu^, C. Guerin{2[ D.M. Antonio^ 1 National Veterinary Institute P O Box 19, Debre Zeit, (Ethiopie) 2 Mission veterinaire francaise en Ethiopie P O Box 1053, Addis-Abeba, (Ethiopie) Resume Les auteurs presentent une synthèse des differentes maladies d'origine bacterienne, virale et parasitaire affectant les petits ruminants en Afrique, en les classant selon leur incidence Gconomique etsanitaire, et propose des priorites de recherche etde développement. Pour illustrer I'expose, deux exemples seront commentes (peste des petits ruminants et pneumopathies des petits ruminants) ainsi que les perspectives offertes par les nouveauxoutils de diagnostic et de prophylaxie medicate. Which research priorities in African small ruminant diseases? F. Thiaucourt, J. Fikre, J.J. Tulasne, G. Mebratu, C. Guerin, D. M. Antonio Abstract This paper reviews different bacterial, viral and parasitic diseases affecting small ruminants in Africa. Diseases are classified on the basis of their economical and health impact and research and development priorities are also suggested. Two examples are discussed: (1) peste des petits ruminants and small ruminant lung diseases and (2) the future of new diagnostic and prophylactic tools. 247 Introduction La pathologie constitue l'un des obstacles les plus importants à l'amélioration de la productivité des troupeaux de petits ruminants en Afrique. Aborder les problèmes de santé animale des caprins et ovins est cependant difficile pour des raisons liées à: • notre connaissance de leur pathologie: les informations sont souvent disparates et limitées (surtout pour la pathologie caprine). La présence vétérinaire est souvent insuffisante pour assurer le contrôle et suivre révolution des maladies sur le terrain; • la pathologie elle-même : il n'y a pas de pathologie dominante par zone ou par pays: existence de syndromes à étiologie multifactorielle plutôt que d'entités pathologiques bien définies; • l'intervention de nombreux facteurs favorisant et déclenchant: la malnutrition et les carences alimentaires diminuent la résistance des animaux au parasitisme et aux infections. Le parasitisme potentialise à son tour les effets de la malnutrition et favorise ainsi les maladies infectieuses. Des facteurs physiques (climat), mécaniques (poussières), chimiques (balnéations), biologiques (virus, bactéries) peuvent déclencher des pneumopathies; • nos moyens d'action: souvent limités en Afrique, ils doivent être bien "ciblés" (CIPEA, 1983). L'Afrique doit faire face à des impératifs économiques souvent contradictoires: sa population progresse à un rythme soutenu, les besoins en protéines animales croissent proportionnellement (ces protéines sont indispensables aux enfants en bas âge qui représentent une fraction importante de la population), mais dans le même temps, les ressources en devises affectées à l'amélioration sanitaire du cheptel stagnent, voire même régressent. Cette pénurie de moyens financiers justifie des choix drastiques d'affectation des crédits qui doivent avoir des effets maximaux sur la productivité du cheptel dans un minimum de temps. Ces éléments rendent complexe la proposition de priorités de recherche et de développement. Les thèmes prioritaires et les stratégies proposées devront bien sûr être rediscutés et éventuellement réorientés pour chaque pays en tenant compte des pathologies dominantes et des moyens d'intervention. Importance économique et sanitaire des principales maladies des petits ruminants Les tableaux 1, 2 et 3 présentent pour les principales maladies virales, bactériennes et parasitaires des petits ruminants en Afrique une estimation de 248 leur importance economique, de leur gravite pour la sante humaine, et de la priorite qu'elles constituent pour la recherche et le développement. Ces trois eléments seront successivement discutes. Tableau 1 . Synoptique des principales parasitoses des petits ruminants en Atrique. Importance economique Zoonose Priorite recherche Priorite developpement Helminthes + + + + dont Fasciola + + + dont Echinococcus + Gales + Coccidia + Cryptosporidia ? Toxoplasma ? Trypanosoma ? Theileria ? +/- + + + + ? + + + + + + + + + + + + + + + + + + + + ? Tableau 2. Synoptique des principales maladies virales des petits ruminants en Afrique. Importance economique Zoonose Priorite recherche Priorite developpement Peste des petits ruminants + + + -\ Clavelee + + + Ecthyma + Fievre aphteuse + + ? Fievre de la + ? vallee du Rift Maladie de Nairobi + ? Blue tongue "— + ? Akabane C.A.E.V. VISNA-MAEDI Border disease Rage ^^ + + + + + + + + + + + + + + + + + + Maladies mal connues en Afrique 249 Tableau 3. Synoptique des principales maladies bactériennes des petits ruminants en Afrique Importance économique Zoonose Priorité recherche Priorité développement Mycoplasmes + + + + Pasteurelles + + + + Charbon + + + Corynébactéries + + + Brucella + Clostridies ? Salmonella ? Chlamydia ? Dermatophilus ? Escherichia + Listeria ? Campylobacter + Cowdria + + + +/- + + + + + + + + + + + + + + + + +? + + + + + + + + + + + + + + + + ? + +? Importance économique Cette notion est difficile à apprécier étant donné les conditions d'élevage en Afrique. Il semble pourtant que l'on puisse dégager quelques "certitudes". • Les carences nutrltionnelles sont certainement, à elles seules, responsables de plus de pertes que l'ensemble de la pathologie infectieuse. Ces pertes peuvent être directes ou indirectes, la malnutrition étant un facteur favorisant pour l'éclosion de syndromes infectieux. L'amélioration des performances du cheptel africain ne peut passer que par une alimentation correcte aussi bien du point de vue qualitatif que quantitatif. Pourtant cette amélioration peut se heurter à des contraintes culturelles qu'il est difficile d'ignorer, en particulier lorsque l'importance du cheptel représente plus un statut social qu'une valeur strictement économique, lorsqu'il existe des prix hautement spéculatifs (à la veille de l'Aïd) qui font passer l'état d'embonpoint de l'animal au second plan. • Les parasitoses sont elles aussi une dominante du point de vue économique. Elles peuvent avoir un effet néfaste direct mais surtout elles peuvent créer ou favoriser des carences nutritionnelles indirectes par spoliation, par perturbation de la flore intestinale normale. En outre, ces parasitoses favorisent aussi la pénétration de germes pathogènes ou pathogènes opportunistes lorsqu'elles occasionnent des brèches dans le système de défense naturelle de l'hôte. 250 • En dehors de toute carence il existe pourtant des maladies infectieuses majeures qui sont un frein à l'amélioration génétique du cheptel africain. Certaines d'entre elles touchent indistinctement l'ensemble de la population des petits ruminants, d'autres peuvent avoir une cible élective parmi les races exotiques importées pour obtenir une amélioration rapide du cheptel local. Enfin, il faut garder à l'esprit que ces mêmes animaux importés peuvent eux aussi être porteurs de maladies "d'avenir" pour l'Afrique. Zoonoses La gravité des zoonoses transmises par les petits ruminants est limitée face aux maladies spécifiquement humaines qui frappent encore l'Afrique et qui sont beaucoup plus meurtrières. Parmi celles dont l'impact est d'ores et déjà certain, nous citerons (IEMVT, 1980): • la brucellose, due à Brucella melitensis principalement, qui peut occasionner une maladie humaine chez les éleveurs mais aussi chez les consommateurs si les produits laitiers sont appelés à une plus large utilisation (la pénurie de bois rend improbable le chauffage du lait avant consommation); • l'hydatidose: dans ce cas, les petits ruminants ne sont pas à l'origine directe de la contamination humaine (ce sont les chiens principalement) mais ils sont un maillon essentiel du cycle de Echinococcus granulosus. Comme le traitement de la maladie humaine ne peut être que chirurgical et que la plupart des anthelminthiques utilisés chez les chiens sont inactifs sur ce cestode, l'effort principal doit porter sur l'éducation du public qui doit pouvoir reconnaître les lésions chez les petits ruminants et soustraire celles-ci de la consommation par les chiens; • la fièvre de la vallée du Rift: les épidémies les plus récentes ont montré (Egypte, Sénégal) que cette maladie pouvait avoir une répartition panafricaine et ne se cantonne plus dans son aire d'origine. Son importance en tant que zoonose n'est plus à démontrer aussi bien que son impact économique lorsqu'elle se déclenche dans une population entièrement réceptive. Néanmoins, son apparition erratique rend difficile toute prophylaxie suivie et demande surtout une bonne information des services vétérinaires pour raccourcir les délais d'intervention en cas d'apparition de la maladie; • le charbon: les petits ruminants sont des animaux très sensibles à Bacillus anthracis au même titre que les bovins; il faut donc les inclure dans les plans de prophylaxie contre cette maladie; • la fasciolose, très répandue chez les ovins, peut aussi avoir une certaine importance chez les humains. Le nombre de cas de fasciolose zoonose en Afrique est assez mal connu et mériterait d'être précisé. De toute façon, l'impact économique sur les productions ovines justifie à lui seul des plans de prophylaxie. 251 Orientation de la recherche sur la pathologie ovine et caprine Priorités de recherche et priorités de développement Afin d'évaluer les niveaux de priorité indiqués dans les tableaux 1 , 2 et 3, nous avons considéré dans le champ de la recherche les études pouvant, à plus ou moins long terme, aider à combattre les maladies des petits ruminants. Les études épidémiologiques Elles indiquent l'importance (prévalence) des maladies à un moment donné, et permettent de déterminer des ordres de priorité d'action. Les études épidémiologiques dynamiques peuvent, elles, aider à élaborer d'autres moyens de lutte qui viseront plutôt à freiner la dissémination des maladies. La mise en place de telles mesures ne pourra toutefois se faire que dans des structures d'élevage bien organisées, ce qui est encore loin d'être le cas dans de nombreux pays. L'amélioration des moyens de lutte Cela concerne la création ou l'amélioration de matériel de diagnostic, des antibiotiques, des vaccins, des antiparasitaires, etc. La sélection génétique d'animaux plus résistants aux différents pathogènes et en particulier aux parasites Cette voie a pu être explorée pour les bovins (trypanotolérance) et pouridit l'être aussi pour les petits ruminants (hémonchose). Si nous n'avons pas considéré cette approche comme une priorité, c'est pour deux raisons. D'une part, il a été montré qu'il est pratiquement impossible de sélectionner des animaux sur de multiples caractères; les races "laitières" perdent leurs qualités "bouchères" (et inversement). L'amélioration des qualités zootechniques se fait souvent au détriment des qualités de résistance aux infections; il est probable que l'inverse soit aussi vrai. D'autre part, il est possible que nous n'ayons pas utilisé tous les moyens à notre disposition pour éviter l'apparition de parasites résistants aux anthelminthiques. Les actions de développement sont celles ayant un effet direct sur révolution de la maladie considérée. Elles vont au-delà des seules mesures d'ordre vétérinaire. Il peut s'agir par exemple d'actions favorisant la plus grande disponibilité d'antiparasitaires et d'antibiotiques à moindre coût: fabrication sur place ou importation de molécules encore actives dont la licence est tombée dans le domaine public, ou encore de la fabrication de vaccins éprouvés, mais surtout 252 de leur diffusion à une échelle la plus large possible. Dans certains cas, l'information du public est capitale lorsqu'il n'existe pas beaucoup d'autres moyens de lutte (exemple du kyste hydatique). Dans le domaine du parasitisme, il doit être possible de proposer des schémas de prophylaxie avec des molécules déjà existantes mais qui soient mieux ciblés. Cela peut concerner: la date d'intervention en fonction stricte des cycles parasitaires, le nombre de ces interventions car leur multiplication peut rapidement entraîner une chute du rapport bénéfice/coût. Enfin, il est peut-être possible de limiter l'apparition de parasites résistants par un emploi plus raisonné des molécules actives: limitation du nombre d'interventions, respect des doses à prescrire, permutation régulière des classes de molécules utilisées, etc. Synthèse des thèmes de recherche prioritaires: une approche par syndrome Sur le terrain on se trouve malheureusement souvent en présence de syndromes à étiologie multiple et non de maladies bien identifiées du point de vue clinique et nécropsique (CTA, 1986; IEMVT, 1980). Très schématiquement, deux syndromes dont l'incidence économique est extrêmement importante représentent actuellement les thèmes majeurs de recherche. Un syndrome pneumopathies • des parasitoses pulmonaires; • la pneumonie "enzootique" d'étiologie complexe: facteurs favorisant (climat, alimentation, parasitoses pulmonaires et/ou digestives). Virus pathogènes les.olus fréquents (PPR, parainfluenza III, clavelée, ecthyma, variole caprine). Infections bactériennes et/ou mycoplasmiques (pasteurelles, staphylocoques, streptocoques, corynébactéries, Haemophilus, mycoplasmes d'infections associées), • pleuropneumonie contagieuse caprine (F 38). J^étude des pneumopathies des petits ruminants doit être différenciée selon les espèces ovine et caprine. Chez les ovins L'étiologie des pneumopathies est assez complexe, mais les pasteurelles jouent sans doute un rôle prédominant. Malheureusement, ces pasteurelles appartiennent non seulement à des espèces différentes (haemolytica et multocida) mais en plus, au sein de chaque espèce, il existe de nombreux sous-types sérologiques qui ne confèrent pas de protection croisée entre eux. La fabrication d'un vaccin efficace bute sur cette difficulté et c'est pourquoi les 253 efforts nous semblent devoir porter sur des actions de développement: amélioration des conditions d'élevage, respect des mesures sanitaires de base que sont la mise en quarantaine avant l'introduction de nouveaux animaux dans un troupeau, la disponibilité d'antibiotiques actifs, etc. Chez les caprins Il serait tentant de reprendre ici le parallèle fait avec les bovins. En effet, chez ces derniers il existe une espèce de mycoplasme particulièrement pathogène (M. mycoides mycoides) responsable de la péripneumonie. Les problèmes posés par cette maladie ont reçu une réponse satisfaisante par la vaccination, puisque dans les pays où celle-ci est effectuée à grande échelle les foyers de péripneumonie ont disparu. Il existe chez les caprins une souche de mycoplasme particulièrement pathogène, dénommée F 38, responsable d'un syndrome voisin de la péripneumonie : la pleuropneumonie contagieuse caprine (OIE, 1989). Là encore, cette maladie est largement répandue en Afrique (Tchad, Tunisie, Soudan, Somalie, Kenya) et il est probable que son "absence" dans certains pays soit plutôt due à un manque de puissance des moyens de diagnostic. Il est donc urgent de trouver une parade à cette maladie et, pourquoi pas, de fabriquer un vaccin efficace. Malheureusement, le parallèle avec les bovins s'arrête peut-être là, car il a été montré que chez les caprins il existe de nombreuses espèces de mycoplasmes qui peuvent être pathogènes. Pour ne citer que les plus importants: F 38, mycoides mycoides LC, capricolum, agalactiae, ovipneumoniae. Si le tropisme électif de ces souches est assez variable: mamelle, articulations, yeux, il n'est pas rare de les isoler à partir de lésions pulmonaires et même de reproduire un syndrome respiratoire par inoculation endobronchique et passage à des animaux intacts. C'est pourquoi les mycoplasmoses de la chèvre nous semblent une priorité de recherche; en effet, les antibiotiques, s'ils sont efficaces (macrolides, tétracyclines), n'empêchent pas la pérennité de l'infection. Enfin, avant la mise au point d'un vaccin efficace, il reste à déterminer si la souche F 38 est bien la cause principale des pneumopathies caprines; si d'autres mycoplasmes jouaient un rôle non négligeable, comme cela semble être le cas, il faudrait alors envisager la possibilité de vaccins multivalents en fonction des souches isolées dans chaque pays. La faisabilité de tels vaccins reste encore à étudier. Un syndrome digestif (diarrhée-amaigrissement) Il comprend essentiellement des parasitoses digestives (strongylose, fasciolose, cysticercose, echinococcose, coccidiose) les entérotoxémies et les colibacilloses- salmonellose, (surtout chez les jeunes de moins de 1 an). 254 La peste des petits ruminants semble jouer maintenant, pour les petits ruminants, un rôle tout à fait comparable à la peste bovine en Afrique. Auparavant plus connue dans l'ouest du continent, elle s'est étendue progressivement vers l'est pour atteindre le Proche-Orient. A l'heure actuelle, aucun pays africain ne peut s'estimer à l'abri de cette maladie. Or il a été montré que, grâce à une communauté antigénique poussée, le vaccin peste bovine pouvait être utilisé chez les petits ruminants et leur conférait une bonne protection. Des vaccins homologues sont d'ailleurs en cours d'essai (Nigéria, IEMVT) et devraient être disponibles dans un proche avenir. C'est pourquoi cette maladie, malgré son impact énorme, n'est pas considérée comme une priorité de recherche mais au contraire, le maximum de ressources disponibles devrait être consacré à son contrôle, à l'instar de ce qui est fait pour la peste bovine. Deux autres thèmes de recherche importants doivent d'autre part être pris en considération: • avortements mammites (brucellose, fièvre Q, chlamydiose, campylobactériose), mammites bactériennes diverses; • affections de la peau et des pieds (ecthyma, varioles ovine et caprine, fièvre aphteuse, corynébactéries, streptotrichose, teignes, gales, tiques, myases, piétin) Nouveaux outils Il convient de ne pas passer sous silence les progrès de la recherche fondamentale qui vont peut-être révolutionner à la fois les techniques de diagnostic aussi bien que les méthodes de lutte. Diagnostic De plus en plus, la recherche fondamentale s'est préoccupée d'élaborer des outils de diagnostic rapide. Les techniques immunoenzymatiques ont été une première voie d'approche (immunofluorescence, méthodes ELISA, techniques d'immunoblotting, mais elles se heurtent à des contraintes techniques pour une plus grande diffusion et à un certain manque de sensibilité. Une autre technique, dérivée, elle, de la biologie moléculaire, pourrait ne pas avoir les mêmes défauts: il s'agit de la PCR (Polymerase Chain Reaction) ou encore "amplification de gène". Cette technique permet d'amplifier sélectivement une séquence d'ADN (ici spécifique d'un germe pathogène) si elle est présente dans un produit pathologique. Si l'on dispose ensuite de sondes ADN marquées, il est possible de mettre en évidence la présence d'un agent pathogène même si celui-ci n'est présent qu'à quelques exemplaires. Bien entendu, les sondes 255 utilisées devront absolument être "froides", c'est-à-dire non radioactives, pour pouvoir être utilisables largement sur le terrain. Vaccins En ce qui les concerne, de nombreuses perspectives sont ouvertes. De nouveaux immunostimulants ont déjà vu le jour (ISCOM, par exemple) qui peuvent permettre la création de nouveaux vaccins inactivés. Le génie génétique ainsi que la biochimie peuvent permettre l'apparition de vaccins à sous-unités. Enfin, il y a la possibilité de créer des vaccins chimères, par recombinaison génétique, le plus souvent avec une souche virale dérivant d'un vaccin "pox". Ces nouveaux vaccins peuvent offrir un autre avantage par rapport aux vaccins classiques, qui est de pouvoir différencier les réponses sérologiques ou bien les séquences génomiques d'avec celles produites par des agents pathogènes. Cette possibilité pourrait alors réconcilier deux méthodes de lutte qui sont pour le moment souvent antinomiques: les prophylaxies sanitaires et médicales. Enfin, il faut insister sur le fait que ces techniques ne sont pas encore passées dans l'usage courant et que la lutte contre les maladies infectieuses passe avant tout par des infrastructures d'élevage bien construites. Conclusion Cette présentation est forcément schématique et nous souhaitons qu'elle ouvre une large discussion. En particulier, nous n'avons pas insisté sur les "maladies d'avenir" que pourraient être les maladies introduites par les reproducteurs importés pour l'amélioration des races locales; ce sera sans doute un chapitre à développer ultérieurement. Pour finir nous voudrions insister sur les trois questions fondamentales auxquelles il faut répondre avant d'engager toute enquête. 1) La question vaut-elle vraiment la peine d'être posée ? 2) La question n'a-t-elle pas déjà reçu une réponse ailleurs ? 3) Avons-nous vraiment les moyens nécessaires de répondre correctement à la question posée ? Des réunions comme celle à laquelle nous participons doivent permettre de confronter des expériences diverses et répondre au moins partiellement aux deux premières interrogations. En ce qui concerne la troisième, chaque institution doit pouvoir apprécier le rapport coût/résultat; en effet, l'affectation de trop peu de moyens pourrait aboutir à l'obtention de résultats biaises, peu fiables, qui dans une certaine mesure pourraient être pires que l'absence de résultat. 256 Ainsi, dans tous les cas où un moyen de lutte efficace et peu cher est disponible, nous pensons qu'il faut affecter le maximum de ressources aux actions de développement et réserver la recherche aux problèmes qui n'ont pas encore trouvé de réponse satisfaisante. Bibliographie CIPEA (Centre international pour l'élevage en Afrique). 1983. L'élevage des petits ruminants dans les régions tropicales humides. Etude de système n° 3. CIPEA, Addis-Abeba, (Ethiopie). CTA (Centre technique de coopération agricole et rurale). 1986. La coordination de la recherche pour le développement des petits ruminants en Afrique. Actes du Séminaire tenu à Montpellier (France). 13-17 octobre. 1986. CTA, Wageningen (Pays-Bas). 251p. IEMVT (Institut d'élevage et de médecine vetérinaire des pays tropicaux). 1 980. Les petits ruminants d'Afrique centrale et d'Afrique de l'Ouest: synthèse des connaissances actuelles. Ministère de la coopération, Paris (France). OIE (Office international des épizooties). 1989. Pleuropneumonie contagieuse caprine Série technique n° 9. OIE, Paris (France). 257 Effet des parasites gastro-intestinaux sur ia durée de l'anoestrus post-partum chez la brebis Djallonké M. S. Hounzangbe Mawule Faculté des sciences agronomiques Université nationale du Bénin B.P. 526, Cotonou (République du Bénin) Résumé Deux tests (un en saison pluvieuse et l'autre en saison sèche) ont été menés pour déterminer l'influence du taux des strongles gastro-intestinaux et des coccidies sur la durée de l'anoestrus post-partum. Les brebis étaient infestées naturellement. Au cours des deux saisons, le retour de l'oestrus a été détecté par un bélier et la lecture des frottis vaginaux. D'une façon générale, l'anoestrus post-partum est plus long en saison humide qu'en saison sèche. L'anoestrus physiologique vrai (déterminé par la cytologie vaginale) est nettement plus court que la durée d'anoestrus à partir de la détection des chaleurs par un bélier. L'étude coprologique révèle que le parasitisme gastro-intestinal influence beaucoup plus le comportement sexuel après la mise-bas que la physiologie sexuelle. Le taux d'infestation des brebis à la mise-bas est le principal facteur du retour tardif des signes extérieurs de chaleurs. Dans plus de 90% des cas, le parasitisme gastro-intestinal retarde la réapparition des signes de chaleurs en saison pluvieuse. En saison sèche l'influence est moins sensible, le taux de parasites étant beaucoup plus faible en cette saison. 259 Effect of gastro-intestinal parasites on "anoestrus post-partum" duration of Djallonke ewes M. S. Hounzangbe Mawulé Abstract Two tests (one during the rainy season and the other during the dry season) were carried out to determine the influence of the level of gastro-intestinal strongyles and coceidies on the "anoestrous post-partum" duration. Ewes were naturally infested. For both tests, return of oestrous was detected by a ram and the reading of vaginal smears. Generally "anoestrous post-partum" duration is longer in the rainy season than in the dry season. Duration of the real physiological anoestrous (determined by vaginal cytology) is significantly shorter than anoestrous duration calculated by using the detection of rut by a ram. Analysis of the level of parasite eggs in ewe faeces showed that gastro-intestinal parasitism influences more sexual behaviour after dropping than sexual physiology. The level of infestation of ewes at dropping is the main factor of the belated return of external sign of oestrous. In more than 90% of the cases, gastro-intestinal parasitism delayed the reappearance of oestrous signs during the rainy season. During the dry season, the influence is less sensitive, the level of parasites being lower. Introduction L'anoestrus post-partum chez la brebis est défini comme l'intervalle de temps qui sépare l'agnelage des premiers signes de chaleurs visibles. Cette durée a souvent été déterminée à partir de l'intervalle de temps entre deux mises-bas successives (Berger et Ginisty, 1980). Mais la brebis n'étant pas systématiquement en gestation après la reprise des activités ovariennes, l'intervalle entre deux mises-bas peut ne pas toujours être la somme de la durée d'une gestation et de celle de l'anoestrus post-partum. Compte tenu de l'importance de ce paramètre de reproduction dans la maîtrise du rythme des mises-bas en vue de l'amélioration de l'élevage ovin, il a paru intéressant et même indispensable de connaître avec précision la durée de l'anoestrus post-partum. 260 Parmi les facteurs influençant le rétablissement normal de l'activité ovarienne après la mise-bas, nous citerons la lactation, l'alimentation qui détermine le poids vif de l'animal, et la pathologie, particulièrement les parasitoses gastro intestinales. L'objectif de notre travail est d'étudier la durée précise de l'anoestrus post-partum par la détection des chaleurs et les frottis vaginaux chez des brebis soumises à l'influence de coccidies et des strongles gastro-intestinaux. Matériel et méthode Le troupeau ovin qui fait l'objet de cette étude se trouve dans une ferme de l'Université nationale du Bénin à Abomey-Calavi situé à environ 20 km de Cotonou (République du Bénin). La pluviométrie y est de type bimodale. Les animaux sont de race Djallonké et correspondent à la sous-race décrite par Rombaut et van Vlaenderen (1976) au sud de la Côte d'Ivoire. Pour l'étude de l'activité ovarienne post-partum nous avons disposé de 9 brebis multipares âgées de 25 à 36 mois pesant 17 à 23 kilos, et de 4 brebis primipares âgées de 16 à 22 mois pesant 15 à 20 kilos. Pendant ou après la gestation, les animaux ont été vaccinés contre la peste bovine et contre le tétanos. Certains sont déparasités contre les helminthes et contre les coccidies. Toutes les brebis allaitantes reçoivent en plus du pâturage naturel dont elles disposent, des compléments alimentaires composés de sous-produits agro-alimentaires. Après la mise-bas et une fois par semaine, toutes les mères ont été pesées; des prélèvements vaginaux et coprologiques ont été pratiqués. Des observations journalières (matin, midi, soir) du comportement sexuel des brebis en présence d'un bélier muni d'un tablier ont été faites. Les frottis vaginaux sont colorés par la technique de Papanicolaou (1942) et les prélèvements coprologiques sont analysés par la méthode de Mac Master modifiée par Mishra (1978). Ce suivi a été réalisé en saison sèche et en saison humide de la même année. Résultats Détermination de la durée de l'anoestrus post-partum par la détection des chaleurs (01) La durée moyenne de l'anoestrus post-partum dans le troupeau étudié est de 76,36 ± 16,44 jours, les extrêmes étant de 40 et de 154 jours. La brebis Djallonké n'est pas saisonnée. Les mises-bas ont lieu pendant toute l'année indépendamment des saisons climatiques. On remarque cependant que 261 la durée de l'anoestrus varie d'une saison à l'autre: les mises-bas à la fin de la saison sèche ou au début de la saison humide sont suivies d'un anoestrus post-partum plus long (85,37 ± 27,37 jours) que celui (64,33 ± 13,34 jours) qui suit les mises-bas en saison sèche (P<0,05). Les durées les plus courtes (40-60 jours) s'observent chez les jeunes brebis primipares ou chez les multipares qui n'allaitent qu'un agneau. En saison des pluies, la durée moyenne de l'anoestrus post-partum est nettement plus courte chez les primipares que chez les multipares (P<0,02). De même, la fluctuation des valeurs autour de la moyenne est plus faible chez les primipares que chez les multipares (P<0,001). Détermination de la durée de l'anoestrus post-partum par les frottis vaginaux (D2) La lecture des frottis vaginaux et les courbes de variation des cellules vaginales (figure 1) au cours de l'anoestrus révèlent: • un taux de cellules superficielles inférieur ou égal à 40% durant les quatre premières semaines après la mise-bas. Durant cette période, les cellules intermédiaires basophiles sont dominantes associées à une abondance simultanée des polynucléaires; • un pic d'environ 70% de cellules superficielles se produit quinze jours avant les manifestations des signes de chaleurs. Ce pic suivi d'une chute est caractérisé par une abondance des cellules superficielles éosinophiles et par la disparition des polynucléaires; • un autre pic de cellules superficielles autour du jour du premier oestrus. Ce nouveau pic est moins grand que celui observé quinze jours plus tôt; • l'indice karyopycnotique révèle quelquefois un pic dès la troisième semaine après la mise-bas. Si parfois il y a un décalage entre ce pic et celui des cellules superficielles, on note souvent la coïncidence entre ces deux pics. Ces résultats permettent de recalculer la durée de l'anoestrus déterminée par la cytologie vaginale (D2 = 54,85 ± 9,71 jours) qui est nettement plus courte (P<0,02) que celle révélée par la détection des chaleurs (D1 = 76,88 ± 16,44 jours). La cytologie vaginale révèle aussi une reprise tardive de l'activité ovarienne après la mise-bas en saison pluvieuse (tableau 1). Influence du parasitisme gastro-intestinal sur la durée de l'anoestrus post-partum Influence de la coccidiose Le nombre moyen de coccidies au moment de la mise-bas (1039 ± 261 oeufs par gramme) est légèrement inférieur à celui à la fin de l'anoestrus (1807 ± 959). 262 Tableau 1 . Durée de l 'anoestrus postpartum déterminée par la detection des chaleurs (D1) et la cytologie vaginale (D2) Durée de Durée de Saisons Brebis Type de brebis l'anoestrus (D1) l'anoestrus (D2) 0022 Multipare 154 84 0026 Multipare 80 77 0035 Multipare 115 77 Saison 0039 Multipare 75 49 humide 0041 Multipare 82 70 0043 Multipare 66 56 0055 Primipare 57 42 5005 Primipare 54 35 0023 Multipare 91 63 Saison 0031 Multipare 55 28 sèche 0048 Multipare 64 48 5017 Primipare 61 42 0074 Primipare 71 42 Néanmoins, il existe une corrélation positive étroite entre la durée de l'anoestrus et le taux de coccidies au moment de la mise-bas. Cette corrélation est encore plus étroite pendant la saison pluvieuse (r=0,98). Aussi les brebis dont la durée de l'anoestrus post-partum est inférieure à 60 jours sont les moins parasitées. Leur OPG de coccidies est inférieur à 400. Le rapport du taux de coccidies par poids de l'animal est de l'ordre de 20 dans ce groupe d'animaux. Chez les brebis dont la durée de l'anoestrus est comprise entre 60 et 90 jours, l'OPG moyen de coccidies est de l'ordre de 700. Le rapport OPG par poids est égal à 43,87. Il existe une étroite corrélation (r=0,87; P<0,01) positive entre le taux de coccidies et la cytologie vaginale (taux de squames et de l'indice karyopycnotique) (tableau 2). 263 Figure 1 Cytologie vaginale de cinq brebis Djallonke Pourcentage de cellule 90 1 CYTOLOGIE VAGINALE (post-partum) animal 0022 TEMPS (en semaines) 264 Figure 1 . (Suite) Pourcentage de cellule 80 45n 40 35 30 25 20 15 CYTOLOGIEVAGINALE (post-partum) animal 0074 4 5 6 7 8 TEMPS (en semaines) Influence de la strongylose L'infestation strongylienne (575 OPG) est moins importante que l'infestation coccidienne (7000 OPG). Les valeurs d'OPG les plus elevees s'observent genéralement apres la mise-bas (P<0,01). 265 Tableau 2. Influence de I'infestation coccidienne sur la durée de l'anoestrus post-partum Brebis Poids (kg) Coccidies (OPG) OPG/kg vif Durée (j) Saison 0022 17,4 ±3,79 0026 18,4 ± 0,89 0035 14 ± 0,69 0039 16,5 ± 0,8 0041 14,5 ±0,47 1015 64,79 82 Humide 0043 17,7 ± 0,83 0055 16 ± 0,85 5005 14,5 ± 0,22 3905 224,43 154 714 38,81 81 2052 146,57 115 428 25,89 75 942 53,11 66 404 25,28 57 321 22,13 54 0023 18,1 ± 0,78 11926 658,90 91 0031 16,5 ± 1,06 119 7,21 55 0048 14 ± 0,69 160 11,47 64 Sèche 0074 14,25 ± 0,68 720 50,95 71 5017 16 ± 0,98 172 10,81 61 Pour les brebis dont l'anoestrus post-partum dure moins de 60 jours, l'OPG moyen est inférieur à 250. Le rapport taux de strongles par le poids de l'animal est de l'ordre de 12. Chez les brebis dont l'anoestrus dure 60 à 90 jours, l'OPG moyen atteint 350. Le rapport taux de strongles par poids est de l'ordre de 1 5. Chez les brebis dont la durée de l'anoestrus est supérieure à 1 00, l'OPG moyen est égal à 485,75. Le rapport strongles-poids est de 30,87 (tableau 3). Il existe une étroite corrélation positive (r=0,78; P<0,01) entre l'infestation strongylienne et l'âge. La corrélation existe, mais elle est plus faible entre le taux de strongles et le poids de l'animal (r=0,58; P<0,01). 266 Tableau 3. Influence de I'infestation strongylienne sur la duree de I'anoestrus postpartum Brebis Poids (kg) Strongles (OPG) OPG/kg vif Dur&e (j) Saison 0022 17,4 ± 3,79 442 0026 18,4 ± 0,89 285 0035 14 ± 0,69 532 0039 16,5 ± 0,47 825 0041 14,5 ± 0,47 364 0043 17,7 ± 0,83 20 0055 16 ± 0,85 142 5005 14,5 ± 0,22 102 25,45 154 15,52 81 38,03 115 49,86 75 25,16 82 Humide 1,16 66 8,9 57 7,16 54 0023 18,1 ± 0,78 196 10,91 91 0031 16,5 + 1,06 23 1,44 55 0048 14 ± 0,69 128 9,18 64 Seche 0074 14,25 ± 0,68 3432 240,89 71 5017 16 ± 0,98 1311 81,98 61 Discussion La duree de I'anoestrus post-partum chez la brebis Djallonke calculee a partir des detections de chaleurs, parait longue au regard des valeurs donnees par la litterature: 42,4 jours (Berger, 1979) et 60 jours (Amegee, 1984). L'anoestrus physiologique vrai est genéralement plus court que la duree d'anoestrus observee dans les conditions ordinaires d'observation. Selon certains auteurs, 267 les ovulations se produisent dès le 18* jour après la mise-bas, mais les signes extérieurs d'oestrus ne sont observables que plus tard. Dans ces conditions pratiques, il est difficile de distinguer l'anoestrus physiologique vrai et le faux anoestrus. Le faux anoestrus serait dû à une détection incorrecte des oestrus. La difficulté de distinguer le vrai anoestrus du faux anoestrus résulte surtout d'une dissociation oestrus-ovulation: la première ovulation est fréquemment silencieuse; dans certains cas, le premier oestrus n'est pas toujours accompagné d'une ovulation. Pour déterminer l'anoestrus physiologique vrai, il faut maîtriser les problèmes posés par la brièveté des oestrus et par la fréquence des chaleurs silencieuses. Cela peut se faire par l'emploi de méthodes plus efficaces de détection des chaleurs, comme la multiplication des observations chaque jour, ou l'emploi de méthodes plus efficaces de détection des ovulations telles que la cytologie vaginale. L'abondance des cellules superficielles sur les frottis vaginaux réalisés traduit une élévation du taux d'oestrogène (après l'ovulation), confirmant la théorie selon laquelle il y a décalage entre la première ovulation et le premier oestrus. Le premier pic de cellules superficielles correspondrait à la première chaleur silencieuse. La différence entre le premier pic et celui du premier oestrus traduirait une sécrétion plus importante de FSH après la mise-bas. La teneur en FSH est plus élevée pendant le premier cycle après la reprise des activités ovariennes que pendant les cycles ultérieurs. La sécrétion de FSH semble par ailleurs inhibée par la prolactine liée à la lactation. La libération de FSH favorise comme on le sait la folliculogenèse, qui se traduit par une forte sécrétion d'oestrogène à laquelle sont sensibles les cellules superficielles de l'épithélium vaginal. Si la pycnocytose traduit la présence d'oestrogène, nous pouvons penser à la reprise assez rapide des activités ovariennes avec des productions d'oestrogènes puisqu'on a observé un pic d'indice karyopycnotique dès la troisième semaine après la mise-bas. Dans le troupeau étudié, l'infestation coccidienne semble constante et régulière. Elle a une influence sensible sur la durée de l'anoestrus post-partum. L'augmentation de l'infestation parasitaire après la mise-bas s'explique par une diminution de la résistance physiologique de la mère liée entre autres à l'allaitement. La diminution de l'absorption intestinale des aliments due aux coccidies se traduit par une perte de poids sensible. Il existe donc une corrélation linéaire entre la durée de l'anoestrus et le rapport OPG par poids. La mise-bas peut donc constituer un facteur de déséquilibre biologique provoquant des troubles de l'état général de l'animal qui devient sensible à l'infestation parasitaire. 268 Conclusion L'activite ovarienne post-partum chez la brebis Djailonké a ete etudiée par la detection de chaleurs et par les frottis vaginaux colons au Papanicolaou. L'examen coprologique a permis d'evaluer I'influence de la coccidiose gastro-intestinale sur la duree de l'anoestrus post-partum. La duree moyenne de l'anoestrus post-partum determinee a partir de la detection des chaleurs est plus grande que celle donnee par Berger et Ginnisty (1980). Le retour des chaleurs est plus tardif pendant la saison pluvieuse. L' etude de la cytologie vaginale au cours de l'anoestrus post-partum permet de determiner la duree moyenne de l'anoestrus physiologique vrai qui est nettement plus court que la duree d'anoestrus observee dans les conditions ordinaires d'observation. Le taux de coccidies a la mise-bas est le facteur qui influence le plus la duree de l'anoestrus post-partum. Cette correlation positive est encore plus etroite pendant la saison pluvieuse. La strongylose influence plus faiblement la reprise des activités ovariennes apres la mise-bas. Ces travaux nous permettent de recommander un déparasitage (anticoccidien et anthelminthique) systématique a la fin de la gestation. Bibliographic Amegée Y. 1984. Etude de la production laitiere de la brebis Djailonke en relation avec la croissance des agneaux. Revue d'elevage et de medecine veterinaire des pays tropicaux 37 (3):331-335. Berger Y. 1979. Selection et amelioration des ovins de Cote d'lvoire. Rapport annuel n° 05. Centre de recherche zootechnique Bouake (Cote d'lvoire). Berger Y. et Ginisty Y. 1980. Bilan de 4 annees d'etude de la race bovine Djailonke en C6te d'lvoire. Revue d'elevage et de medecine veterinaire des pays tropicaux 33(1):71-78. Mishra G.A. 1978. Parasitologic. Protozoologie: Cropologie programme. Manuel non public. FAO/PNUD. Papanicolaou G.N. 1942. A new procedure for staining vaginal smears. Science 95:438-439. Rombaut D. et Van vlaenderen G. 1976. Le mouton Djailonke de Cote d'lvoire en milieu villageois: comportement ot alimentation. Revue d'elevage et de m6decine veterinaire des pays tropicaux 29(2): 157-1 72. 269 The effects of endoparasites on the productivity of Ethiopian highland sheep Tekelye Bekele and OB. Kasali International Livestock Centre for Africa (ILCA) P O Box 5689 Addis Ababa, Ethiopia Abstract During 1988 and 1989 the effect of endoparasites on the productivity of Ethiopian highland sheep was studied using eggs per gram (EPG) counts, worm counts andpacked cell volume (PCV) at Debre Berhan, Dejen, Deneba, Tulu Meko and Wereilu. Nematode and trematode EPG counts were low except at Tulu Meko with significant (P<0.05) differences in age, season and year effects in most sites. Mean monthly mature worm counts in necropsied sheep were also low to moderate andranged from 0-4 375 and 0-1 825 in the abomasum and intestines, respectively. The predominant worms in the counts were mainly Trichostrongylus colubriformis with a few Haemonchus contortus and others, irrespective of season. This was attributed to possible resistance and/or "self cure" to haemonchosis andlor environmental factors. Mean PCV values were almost similar for all the locations and were significantly (P<0.05) affected by the degree of nematode and trematode infections in all sites. LowPCVand high EPG levels were significantly (P<0.05) associated with poor body-condition scores and low bodyweights. Monthly repeatabilities of PCV, bodyweight and body-condition score were 0.44 ± 0.01, 0.71 ± 0.01 and 0.35 ± 0.01, respectively, while the values for nematode (0.09 ± 0.01) and trematode EPG (0.20 ± 0.20) were much lower. The higher repeatability forPCVin contrast to EPG would indicate that it was less affected by the variable factors influencing egg output. It is, therefore, suggested that it should be used in conjunction with nematode and trematode EPG levels in endoparasite monitoring. 271 Effets des endoparasites sur la productivité du mouton des hauts plateaux éthiopiens Tekelye Bekele et O.B. Kasali Résumé L'effet des endoparasites sur la productivité du mouton des hauts plateaux éthiopiens a été étudié en 1988 et 1989 à Debre Berhan, Dejen, Deneba, Tulu Meko et Wereilu. Les paramètres considérés étaient le nombre d'oeufs de parasites par gramme de fécès, le nombre de vers dans diverses parties du tube digestif et l'hématocrite. Exception faite de la localité Tulu Meko, le nombre d'oeufs de nématodes et de nématodes par gramme de fécès était faible et, dans la plupart des régions variait significativement (P<0,05) en fonction de l'âge des animaux, de la saison etde l'année. Le nombre de vers adultes chez les moutons autopsiés était faible ou moyennement élevé, variant deO à 4 375 dans la caillette et de 0 à 1 825 dans les intestins. Quelle que soit la saison considérée, les vers les plus nombreux étaient les Trichostrongylus colubriformis. On a également dénombré quelques Haemonchus contortus ainsi que des représentants d'autres espèces. La rareté relative des Haemonchus était peut-être due à une résistance à l'hémonchose et/ou à un "auto-traitement" de cette maladie à moins que cela ne soit dû à l'action du milieu. Avec des valeurs moyennes de 33,40 ± 0,70, 32,70 ± 0,78, 33,92 ± 0,76, 28,63 ± 0,79 et 35,79 ± 0,68 respectivement à Debre Berhan, Dejen, Deneba, Tulu Meko et Wereilu, l'hématocrite variait significativement (P<0,05) en fonction du degré d'infestation par les nématodes et les trématodes quelle que soit la localité considérée. Une association significative (P<0,05) entre les faibles niveaux d'hématocrite et les nombres élevés d'oeufs par gramme de fécès d'une part et d'autre part, les notes médiocres d'état corporel des animaux et leurs faibles poids vifs a été montrée. Les coefficients de répétabilité de l'hématocrite, du poids vif et de la note d'état corporel furent respectivement de 0,44 ± 0,01, 0,71 ± 0,01 et 0,35 ± 0,01. En comparaison, les coefficients relatifs au nombre de parasites par gramme de fécès étaient beaucoup plus faibles, s'établissant respectivement à 0,09 ± 0,01 pour les nématodes et à 0,20 ± 0,20 pour les trématodes. Compte tenu de la valeur élevée du coefficient de répétabilité relatif à l'hématocrite, il est probable que les divers facteurs dont dépend le nombre d'oeufs influencent moins ce 272 parametre que les autres. En consequence, il est proposé d'utiliser I'hematocrite, conjointement avec le nombre d'oeufs de nematodes et de trematodes par gramme de feces pour evaluer le niveau d'endoparasitisme de ces animaux. Introduction Most sheep in Ethiopia are owned by smallholders in the highlands where there is mixed crop-livestock production. These sheep are an integral part of the livestock sector of the economy. Sheep supply meat, wool (hair) and skin and generate about 89% of farmers' cash income (Gryseels et al, 1989). The productivity of Ethiopian sheep is affected by nutritional and endoparasitic stress although they are all year-round breeders. Nutritional stress occurs whenever there are dry spells and often before grass grows during the rainy seasons. The major endoparasitic diseases of importance include fascioliasis, gastrointestinal nematodiasis, cestodiasis and lung worm (Graber, 1975; Scott andGoll, 1977; Bekele Mamo et al, 1981; Lemma et al, 1985). Their effects range from a reduced performance to mortality (Sykes, 1978; Armour and Gettinby, 1983). These effects are also exacerbated by the level of nutrition. There are limited field studies on the effects of endoparasites on productivity. Most attempts have been on experimental infections of either a single or few parasites which do not simulate the dynamics in field conditions. Hence this study was performed to determine the impact of endoparasites on productivity of the highland sheep in Ethiopia. This report focuses on the relationship of some of the indicators of endoparasites on productivity and the estimates of repeatability of associated parameters. Materials and Methods The study was carried out on-farm at Debre Berhan, Dejen, Deneba, Tulu Meko and Wereilu in the Ethiopian highlands from June 1 988 to December 1 989. These places are located in the central, northern and north-western part of the country where the altitude ranges from 2500-3000 m above sea level and the production system is mainly mixed farming. They have bimodal rainfall pattern with long rains from June to September and short rains from March to May. A total of 2594 sheep were involved in the study. These included the Menz, Horro, Wollo and Dejen breed types from 30-50 farmers with 350-400 sheep each at the beginning of the study. There was also periodic replacement of exits. All animals were raised under traditional management with minimum health interventions. Each site was visited monthly for sampling and data collection which included faeces for eggs-per-gram (EPG) counts and faecal culture, blood for packed cell 273 volume (PCV) and screening haemoparasites, bodyweight, body-condition score (MAFF, 1984; Hossamoet al, 1986) and pregnancy and lambing status. In addition, 122 (six-month-old or more) sheep were sacrificed at each site at the rate of four per month in the first year and then periodically in the second year for post-mortem examination, worm count and identification; Skerman and Hillard, 1 966) at Debre Berhan and Tulu Meko. The age of each sheep was estimated from teeth-eruption changes (Williamson and Payne, 1959) on each visit. Data were analysed by least squares (SAS, 1987) using the general linear model for the effect of endoparasites on productivity for each site separately but were pooled for repeatability estimates. Dependent variables were bodyweight, body-condition score and lambing interval. The independent variables were nematode and trematode EPG levels determined by classifying infection levels as either low (<50) or high (> 50), and PCV which was also classified as below normal (<. 27) and above normal (> 27). The EPG and PCV classifications for each animal were based on the means of the repeated monthly determinants over the study period. Repeatability was estimated using a mixed model (Harvey, 1987) with random animal effect and other relevant fixed effects. The estimates of repeatability for log value of positive EPG and PCV were obtained by treating these as dependent variables and adjusting for the fixed effects. Results Peak nematode EPG counts were observed during the long and short rainy seasons especially at Tulu Meko (Figures 1 and 2). In addition, trematode EPG counts were also high for most parts of 1 989 at Tulu Meko but not on other sites. Log values of monthly EPG counts were significantly (P<0.05) affected by age of sheep, season and year. The mean monthly mature worm counts in necropsied sheep ranged from 0-4375 and 0-1825 in the abomasum and intestines, respectively. Immature worms were less frequently encountered. Worm counts at Tulu Meko were higher than at Debre Berhan. The log transformed mature worm counts from the abomasum did not show significant (P>0.05) seasonal differences in both sites while year-effect was significantly (P<0.05) different at Tulu Meko but not at Debre Berhan (P>0.05). The commonest mature and immature worms found during worm counting were Trichostrongylus colubriformis with only few H. contortus, T. axei, Oesophagostomum columbianum, Bunostomum trigoncephalum and Trichuris skrjabini. These gastrointestinal parasites were also found in faecal cultures. The mean monthly counts of Fasciola hepatica in the liver ranged from 0-78 for both Debre Berhan and Tulu Meko. Seasonal frequencies of D. filaria and T. colubriformis were 66.7-1 00% and 64.0-1 00% in necropsied sheep but they were low for others. 274 The correlation and regression coefficients of both nematode and trematode EPG counts on PCV were significantly (P<0.01) different from zero. Least squares means for PCV at different sites were 33.40 ± 0.70, 32.70 ± 0.78, 33.92 ± 0.76, 28.63 ± 0.79 and 35.79 ± 0.69 at Debre Berhan, Dejen, Deneba, Tulu Meko and Wereilu, respectively. Levels of nematode and trematode EPG caused Figure 1 . Correlation of mean monthly rainfall with mean monthly nematode egg counts by site (1988-1989). Debre Berhan Dejen Deneba Tulu Meko Wereilu Mean monthly trematode (epg counts) 300 1 Rainfall (mm) 400 i i—i 1—i 1—i—'—,—,—i—i—,—,—'—'—'—r Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov 1988 I Year/month 1989 275 Figure 2. Correlation of mean monthly rainfall with mean monthly trematode egg counts by site (1988-1989). DebreBerhan Dejen Deneba TuluMeko Wereilu Mean monthly trematode (epg counts) 450 Rainfall (mm) 450 Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov 1988 ' 1989 Year/month significant (P<0.05) differences in PCV which decreased with increasing parasitic load in each site. Mixed nematode and trematode infections were few and insignificant (P>0.05). PCVs were also significantly (P<0.03) different in different age groups of sheep, different seasons and different years but not in different sexes. 276 Figure 3 shows the frequency of average PCV (average above and below normal), nematode (average low and high EPG levels) and trematode (average low and high EPG levels) EPG levels of all sheep in different sites. These frequencies indicate that endoparasitism was highest at Tulu Meko with low PCV, high nematode EPG and high trematode EPG levels of 37.4, 31.8 and 38.4%, respectively. The least-squares means for body-condition scores, bodyweights of male sheep and lambing interval are, respectively, presented in Tables 1 , 2 and 3. The means for body scores were 1.69 ± 0.76, 1.85 ± 0.71, 1.75 i 0.79, 1.60 ± 0.82 and 1.85 ± 0.74 at Debre Berhan, Dejen, Deneba, Tulu Meko and Figere 3. Frequency of average PCV, nematode and trematode EPG levels of all sheep in different sites. Debre Berhan Dejen (n = 4949) (n = 3997) Deneba (4691) Pourcentage 100 90 80 70 60 50 40 30 20 10 ■ l s IS h I |i B w PCV >/ = 28 PCV :... !* I 1 1 P1 f'.I > 1 1 Tulu Meko (n = 4151) Nematode EPG 0-50 50+ Is I s* ?■ ii Wereilu (n = 4138) I I ill ns L Trematode EPG 0-50 50+ 277 Table0.Le st-squaresmeans(±SE)rorbod7-conditionscralsh pidifr re tt Mean(±SE) 0.00±.70 0.70±.00a 0.7±.0b 0.00±07a 0.N±.00b 0.ro±17a 0.0+0a Mean(±7) 30.0±0.0 00.0±.0a 00.0±.0a 00.0±.0a 00.0±.0a 030.0±.0b 07.0±.0a Wereilu TuluMeko No 0) 000 00 )90 00 050 07 Mean(+7) ±0.. ±0.7a ±0.00b ±0.7a ±0.0. ±0.7b ±0.00a No 00 00- 0 00 00 300 00 DenebaTuluM ko 0.0 0.0 0.ro 0.00 0.0 0.0 0.07 Withinvariablegroupsme nsrollowedw ththsamlet ernotsigni icantly0P>.17)d ife ent. ;±7)7o 0.7000 0.030a300 0.03±07b50 0.030a00 0.00b7 9 0.00b7 0.07a30 Dejen Mean(±SE) 30.0±0.0 00.0±.0a 00.0+.0a ns.0±0.0a 0.0±.0a N.0±0.0b 7.0±0.0a MeanI 0.0± 0.77± 0.00± 0.07+ 0.ns± 0.■± l No 00- 700 ns0 ro0 00 ro0 000 sheep. 7o 00 00i 15 0S 32 ■7 0 Mean(±7) 0007. 0+7 000.0±. a 0.00±. 0.30±.00a 0.0±.0 b 0.0±.00b 0.7±.00a Table0.Le st-squaresmeans(±SE)rorliveweightoralm le Dejen DebreB rhan Mean(±7) 0.0±7 07.0±.0a 0.0±.0b 0.0±.0b 30.0+0.0a 0.0±.0b 00.0±.0a 7o 300 0000 030 0700 030 Mean(±7) 0.0±0.0 0.70+a 0.30±17b 0.0±.00a 0.,±.030b 0.07±. 0.0±.03 a I DebreB rhan 7o 0030 007 07 00- 00 70 00 No Overallme n■300 Packedcellvolume 2N00030 <;07 NematodeEPGl vels 0-5070 50+777 TrematodeEPGlevels 0-50ns 50+0 0 Packedcellvolume 7ematodeEPGlevels TrematodeEPGlevels Overallme n 28 507 0-50 50+ 0-0P 00+ Withinvariablegroupsme nsrollowedw ththesamlet rnotsign rica tly0P>0.17)di fe en . 00 Wereilu, respectively, and were significantly (P<0.05) affected by PCV, nematode and trematode EPG levels. Liveweights had mean squares of 22.2 ± 1.7, 26.0 ± 2.3 and 26.8 ± 1.6 kg at Debre Berhan, Dejen and Tulu Meko, respectively, with the subclass means showing significant (P<0.05) differences due to PCV and nematode EPG levels at Debre Berhan and trematode EPG levels in all. The least-squares means for lambing interval was 263.2 ± 4.0 days at Tulu Meko with no significant (P>0.05) effect by the different variables. The repeatability values for PCV, bodyweight, body-condition score, lambing interval, nematode EPG and trematode EPG were 0.44 ± 0.01 , 0.71 ± 0.01 , 0.35 ± 0.01, 0.43 ± 0.14, 0.09 ± 0.01 and 0.20 ± 0.02, respectively (Table 4). Correlation coefficients between bodyweight and body-condition score were low (r= 0.3-0.5, P<0.01) for different sites. Table 3. Least-squares means (±SE) for lambing interval at Tulu Meko. Variable No Mean (±SE) Overall mean 97 263.2 ±4.0 Packed cell volume > 28 74 262.5 ± 7.7a < 27 23 259.9 ± 15.5a Nematode EPG levels 0-50 37 253.1 ± 15.5a 50+ 60 266.3 ± 9.4a Trematode EPG levels 0-50 39 257.3 ± 12.8a 50+ 58 262.1 ± 9.3a Table 4. The repeatability (+SE) of eggs-per-gram (EPG) counts, packed cell volume (PCV), bodyweight, body-condition score and lambing interval for all sites. Parameter Number of observations Repeatability Nematode EPG (log value) Trematode EPG (log value) Packed cell volume (PCV) Bodyweight Body-condition score Lambing interval 5 750 0.09 0.01 4 331 0.20 0.02 22 464 0.44 0.01 22 669 0.71 0.01 22 671 0.35 0.01 475 0.43 0.14 279 Discussion Monthly nematode EPG counts were low with very little variation with rainfall for most sites except at Tulu Meko compared to previous findings (Tekelye Bekele et al, 1987). This agreed with the low to moderate worm counts (Skerman and Hillard, 1966) of necropsied sheep at Debre Berhan and Tulu Meko. The EPG of the necropsied sheep was also low. The moderately high nematode EPG counts at Tulu Meko was associated with the high intensity of worm load. The commonest species found during worm counting, T. colubriformis, has low prolificacy and pathogenicity (Holmes, 1985). Highly prolific worms like H. contortus had low counts specially at Debre Berhan although they were available in necropsied sheep and faecal cultures. This may be due to genetic factors where H. contortus was not established and/or was removed after infection due to self cure (Preston and Allonby, 1979; Kasali et al, 1988) and/or other environmental factors. Trematode EPG counts were highest in 1989 than in 1988 in all sites which probably happened due to the heavy rainfall in 1988, which also increased the transmission of fascioliasis by increasing the abundance of the intermediate host, the snail, Lymnae truncatula (Goll and Scott, 1 978). Fascioliasis, especially when complicated by Clostridium novyi infection, is the major cause of ovine mortality in the Ethiopian highlands (Njau et al, 1988). The presence of D. filaria in most necropsied sheep agreed with previous findings (Bekele Mamo et al, 1981). Its high frequency is suggestive of its importance since it is commonly complicated by pasteurellosis, inclement weather and inadequate nutrition. The overall mean and subclass mean PCV were lowest at Tulu Meko which could be due to high intensity of endoparasitism as observed in the EPG and worm counts. This could also be due to differences in breed parasite tolerance and/or management. PCV variation with age, year and EPG load effects was common to most sites. PCV variation with age could be physiological, while year effect could be due to differences in the rainfall leading to differences in the degree of endoparasitism. The effect of season was variable depending on the degree of endoparasites and level of nutrition. Repeated measurements of endoparasites indicated high frequencies of sheep with subnormal PCV, and high nematode and trematode EPG levels at Tulu Meko. This further confirmed the high intensity of endoparasitism in this site. Productivity traits like body-condition scores and bodyweights were directly correlated with the level of PCV, being low when PCV were low and high when PCV were high. High nematode EPG lowered body-condition scores and bodyweights. But high trematode EPG resulted in high body-condition scores and high bodyweights which could be due to repeated exposures to fascioliasis while at the same time growing and increasing in body mass. Endoparasitism did not have any effect on lambing interval which may be due to its periodic occurrence and/or it may not influence this parameter. 280 Monthly repeatability estimates for nematode and trematode EPG were low showing that egg output was variable being affected by different factors like age of the host, species and age of parasite, nutritional status of the host, season, physiological factors like pregnancy and degree of repeated infections. But although PCV could also be affected by multiple factors, its monthly within-animal variation was moderately high and hence it should be used in conjunction with EPG for monitoring the level of endoparasites. Although there are reports of high correlation between bodyweight and body-condition scores (Hossamo et al, 1986), this was not the case in this study. The repeatability of body-condition score was low which could be probably due to the effect of either within-animal and/or between-scorer variation. The repeatability estimates for lambing interval across parities is higher than the range reported (Wilson et al, 1989). References Armour J and Gettinby G. 1983. A critical review of the evaluation of production effects of helminth disease and management of livestock production. Proceedings of the Third International Symposium on Veterinary Epidemiology and Economics held in Arlington, Virginia, USA, 6-10 September 1982. Veterinary Medicine Publishing Company, Edwardsville, Kansas, USA. pp. 164-172. Bekele Mamo, Feseha Gebreab and Shibru Tedla. 1981. Observations on Dictyocaulus filaria (Rudolphi, 1809) in Welo and Arsi Administrative Regions of Ethiopia. Ethiopian Journal of Agricultural Sciences 3:75-80. Goll P H and Scott J M. 1978. The inter-relationship of Lymnaea truncatula and ovine fascioliasis in the Ethiopian central highlands. British Veterinary Journal 134:551- 555. Graber M M. 1975. Helminths and helminthiasis of domestic and wild animals of Ethiopia. Bulletin of Animal Health and Production in Africa 23:57-86. Gryseels G, Anderson F, Getachew Assamenew, Abebe Misgina, Abiye Astatke and Woldeab Woldemariam. 1989. On-farm research to improve small holder livestock productivity in the Ethiopian highlands. Quarterly Journal of International Agriculture 28:365-375. Harvey W R. 1987. Mixed model least-squares and maximum likelihood computer program. Columbus, Ohio, USA. Holmes P H. 1985. Pathogenesis of trichostrongylosis. Veterinary Parasitology 18:89-101. Hossamo H E, Owen J B and Farid MFA. 1986. Body condition score and production in fat-tailed Awassi sheep under range conditions. Research and Development in Agriculture 3:99-104. Kasali O B, Njau B C and Bekele T. 1988. Controlling livestock diseases in the tropics by breeding: A perspective. In: Thomson E F and Thomson F S (eds), Increasing small ruminant productivity in semi-arid areas. Proceedings of a workshop held at thelnternational Center for Agricultural Research in the Dry Areas, Aleppo, Syria, 30 November to 3 December 1987. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 237-242. 281 Lemma B, Gebre-ab F and Tedla S. 1985. Studies on fascioliasis in four selected sites in Ethiopia. Veterinary Parasitology 18:29-37. MAFF (Ministry of Agriculture, Fisheries and Food). 1984. Condition scoring of ewes. Agricultural Development Advisory Services, Sheep Advisors. Leaflet 797. pp. 1-5. Njau B C, Kasali 0 B, Scholtens R G and Mesfin Degefa. 1988. Review of sheep mortality in the Ethiopian highlands, 1982-86. ILCA Bulletin 31:19-22. International Livestock Centre for Africa, Addis Ababa, Ethiopia. Preston J M and Allonby E W. 1979. The influence of breed on the susceptibility of sheep to Haemonchus contortus infection in Kenya. Research in Veterinary Science 26:134-139. SAS (Statistical Analysis System) Institute, Inc.. 1987. SAS/STAT guide for personal computers. Version 6 edition. SAS Institute, Inc. Cary, NC, USA 1987. 1028 pp. Scott J M and Goll P H. 1977. The epidemiology and anthelminthic control of ovine fascioliasis in the Ethiopian central highlands. British VeterinaryJournal 1 33:273-288. Skerman K D and Hillard J J. 1966. A handbook for studies of helminth parasites of ruminants. Near East Animal Health Institute, Iran Unit, UNDP/special fund. Sykes A R. 1978. The effect of subclinical parasitism in sheep. Veterinary Record 102:32-34. Tekelye Bekele, Mukasa-Mugerwa E and Scholtens R G, 1987. Seasonal changes in nematode faecal egg counts of sheep in Ethiopia. ILCA Bulletin 29:9-1 1 . International Livestock Centre for Africa, Addis Ababa, Ethiopia. Williamson G and Payne W J A. 1 959. An introduction to animal husbandry in the tropics. Longman, London, UK. 447 pp. Wilson R T, Murayi T H and Rocha A 1989. Indigenous African small ruminant strains with potentially high reproductive performance. Small Ruminant Research 2:107-117. 282 Lamb and kid wastage through slaughtering of pregnant ewes and goats at Enugu and Nsukka abattoirs in Anambra State, Nigeria LO. Wosu and E.C. Dibua Department of Veterinary Medicine, University of Nigeria, Nsukka, Nigeria Abstract A study was made of abattoir records in Enugu and Nsukka over a five-year period (1985-1989). A total of 266 060 goats and 91 289 sheep were slaughtered at Enugu and 27 185 goats and 6669 sheep in Nsukka. 50.11% of these goats and 62.03% of these sheep were females; of these females, respectively, 11.00% and 11.40% were pregnant. Avoidable pregnancy wastage (12-20 weeks old) was 31.04% and 33.90%, respectively, for each species. The study calls for attention to the magnitude of slaughter of pregnant females and thepercentage ofpregnancy wastage at a time of increasing shortage of meat supply to the population. It is suggested that pregnancy diagnosis be done at abattoirs prior to slaughter of small ruminant females, and that farmers be trained on simple methods ofpregnancy detection. Pertes de foetus enregistrees aux abattoirs d'Enugu et Nsukka au Nigeria par suite de I'abattge de brebis et de chevres gestantes L.O. Wosu etE.C. Dibua Resume Les donnees rassemblees aux abattoirs d'Enugu et de Nsukka de 1985 a 1989, ont ete analysees. Au cours de cette periode, 266 060 caprins et 91 289 ovins a Enugu contre 27 185 caprins et 6 669 ovins a Nsukka, soit un total de 391 203 petits ruminants ont ere abattus dans ces deux abattoirs. 11,00% et 11,40% etaient gestantes. Les pertes de foetus par suite de I'abattage des femelles en gestation 283 peuvent être évitées chez les petits ruminants dès lors que le foetus atteint l'âge de 12 à 20 semaines dans la mesure où il devient alors décelable et palpable. Celles-ci concernaient 31,04% et 33,90% des femelles gestantes abattues. Cette étude attire l'attention sur les problèmes liés à l'abattage des femelles des ruminants et notamment sur les pertes de foetus au momentmême où les pénuries de viande ne cessentde s'aggraver. Il estrecommandé de détecterles femelles gestantes aux abattoirs et de former les éleveurs à des techniques simples de diagnostic de gestation. 284 La mortalité des jeunes dans les élevages extensifs traditionnels de la zone sud-soudanienne du Sénégal (site de Kolda) M. Ndiaye ISRA, Laboratoire national d'élevage et de recherches vétérinaires B.P. 2057, Dakar-Hann (Sénégal) Résumé A l'issue de trois années d'observation et de prophylaxie à l'aide de vaccins contre la peste des petits ruminants et la pasteurellose chez les caprins et contre la pasteurellose chez les ovins, et à l'aide du tartrate de morantel chez les deux espèces, l'on a fait les constats suivants. - les 0-3 mois sont la classe d'âge la plus vulnérable; - ils sont affectés en priorité par le syndrome "chétivité"; - ils ont une hiérarchie des syndromes mortels relativement stable d'une stratégie prophylactique à l'autre; - ceci quelle que soit l'espèce. Ces constats laissent croire que la mortalité des jeunes de0à3 mois répond à un déterminisme autre que celui contrôlé par les prophylaxies testées. 285 Mortality rate in 0-3 month old animals in extensive traditional livestock production systems of the southern Sudanian zone of Senegal M. Ndiaye Abstract This paper reports observations over three years on flocks that received the following prophylaxis: vaccination against peste des petits ruminants and pasteurellosis in goats, vaccination against pasteurellosis in sheep, and treatment with Morantel tartrate in both species. Mortality was greatest in the 0-3 month age class, which was primarily affected by unthriftiness. The ranking of the main causes of death were similar for all treatment groups and both species. These observations indicate that mortality among 0-3 month old animals was caused by factors other than the diseases against which prophylaxis had been given. The author proposes an aetiology for deaths due to unthriftiness, based on the characteristics of the farming system, biopathological data obtained from the literature and the clinical symptoms of unthriftiness. Introduction Les enregistrements épidémiologiques obtenus par le programme Pathologie et productivité des petits ruminants en milieu traditionnel (PPR) de 1985 à 1987 permettent de décrire l'importance absolue et relative de la mortalité des petits ruminants dans la région de Kolda. Ce qui nous donne l'occasion de poser un regard critique sur le paramètre constitué par les taux de mortalité très usités en épidémiologie. Les résultats rapportés ici ont été obtenus dans l'une des zones d'intervention du PPR, la région de Kolda, au sud du fleuve Casamance. Les précipitations y sont plus abondantes (950 mm par an), plus régulières et réparties sur une plus longue période que dans le reste du Sénégal. Les petits ruminants exploités dans cette zone, sont des animaux de petite taille: ovins Djallonké et chèvres guinéennes. 286 L'on serait tenté, devant l'importance de la mortalité chez les ovins et les caprins de moins de 3 mois, de mettre en route des prophylaxies plus complètes ou tout au moins de rechercher les étiologies précises de la mortalité ainsi notée. Cependant, il nous paraît prudent avant pareille décision de procéder à l'évaluation de l'effet des prophylaxies sur la mortalité. Matériel et méthode Pour chaque espèce, les troupeaux de cinq villages ont constitué le lot témoin, ceux de cinq villages ont été vermifugés par le tartrate de pyrimidine (Iot vermifugé), ceux de cinq autres villages vaccinés contre la pasteurellose à P. multocide types 1 et D et contre la peste des petits ruminants (Iot vacciné), et ceux de cinq autres villages vaccinés et vermifugés. Le taux brut de mortalité est défini comme le nombre total de décès enregistrés au sein d'une population, divisé par le nombre d'individus dans cette population, mesure au cours de la période considérée (Putt ef al., 1987). Il donne au pathologiste un paramètre chiffré à l'échelle de la population pour apprécier la gravité du phénomène mortalité. Par analogie à un taux de positivité à un test de diagnostic, il se calcule par rapport à une population exposée. Lors d'études épidémiologiques ponctuelles sur de courtes périodes n'autorisant pas une refonte profonde de l'effectif présent en début d'observation, l'usage du taux se révèle très commode, surtout lorsque l'étude porte sur une population dont les individus ne sont pas identifiés ou lorsque l'on veut évaluer à postériori l'impact d'un épisode mortel. La population ciblée est cependant sujette à des changements de taille et d'individus, fonctions du contexte de gestion démographique et des facteurs de productivité numérique propres à la population. Ces deux derniers éléments ont une influence d'autant plus marquée que la période d'observation est longue. Lors d'études de longue durée portant sur la mortalité d'une enzootie ou lorsque l'on veut évaluer l'impact sur la démographie ou sur les productivités numériques d'une enzootie, le taux n'est plus approprié car il est influencé par la dynamique démographique de la population. La méthode des quotients de mortalité sur des cohortes d'animaux se révèle alors plus performante. Le suivi individuel du programme PPR permet de connaître précisément la composition des cohortes. Le taux de mortalité annuel est calculé comme le rapport M, nombre de sujets morts dans une population donnée pendant une période de n mois, et P la somme des nombres de sujets recencés dans cette population à chaque passage mensuel durant ces mêmes mois, multiplié par 12. La mortalité relative peut ensuite être établie par classe d'âge ou en fonction des causes de mortalité. 287 Le taux de mortalité est donc simplement une donnée énumérative (le nombre de morts) rapportée à une autre donnée énumérative qu'est l'effectif moyen de la population, et peut à ce titre être considéré comme une mesure caractéristique de la population et être statistiquement traité comme telle. L'effet des prophylaxies sur la mortalité sera étudié en comparant les différentes mortalités annuelles des quatre lots par une analyse de variance sur le taux de mortalité (assimilé à une mesure). C'est à partir de ces valeurs que nous évaluons l'effet de la vaccination et de la vermifugation. L'ensemble des causes de mort considérées dans cette analyse et relevées par le suivi sanitaire du PPR s'établit comme suit: - trouble de croissance ou chétivité (TRC): mort d'un jeune faible sans signe clinique; - pneumopathie (PNE): jetage, toux, dyspnée non associés à une diarrhée; - diarrhée (DIA): non associée à des troubles respiratoires; - peste (PES) : association de pneumopathie et de diarrhée sans précision sur la contagiosité, sans confirmation virologique. C'est le syndrome et non la maladie; - maladies cutanées (CUT): tout ectoparasitisme; - autres (AUT): symptômes avérés ne se rapprochant pas des rubriques précédentes (boiterie, conjonctivite, etc.). La rubrique divers (DIV) regroupe les causes de morts inconnues, non identifiées et celles dont la venue nous paraît être liée à des facteurs extrinsèques occasionnels. Pour étudier l'effet des prophylaxies sur l'ordre d'importance des causes de mort, on comparera un à un les rangs de chaque cause de mort dans les deux groupes de lots respectivement pour la vaccination puis la vermifugation dans chaque espèce sans tenir compte du sexe. Pour ce faire, nous appliquerons la formule de Spearman entre les groupes vermifugés et non vermifugés, puis entre les vaccinés et les non vaccinés, pour avoir un élément d'appréciation: le coefficient de corrélation d'ordre (Murray, 1975): d2 r = 1 -- rang n (n2 - 1) ou d = différences entre les rangs des x et y correspondants n = nombre de paires de valeurs (x, y) dans les données. 288 Résultats et discussion Mortalité Chez les ovins et les caprins, en taux de mortalité par sous-classe d'âge, la sous-classe d'âge la plus affectée et quel que soit le lot prophylactique, reste la sous-classe 0-3 mois, sauf chez les caprins vaccinés où les 4-6 mois sont légèrement plus affectés (tableaux 1 et 2). Tableau 1 . Tauxannuel de mortalité des ovins par classe d'âge en pourcentage (Kolda, 1985-1987), et mortalité relative par classe d'âge Classe d'âge (en mois) Nombre de morts de 1985 à 1987 Cumul des présents/mois Taux de mortalitéLot Témoin 0-3 37 (49%) 904 49 4-6 23 (31%) 806 34 7-9 11 (15%) 662 20 10-12 4 (5%) 553 9 Vacciné 0-3 73 (47%) 1275 69 4-6 41 (26%) 1 115 44 7-9 22 (14%) 971 27 10-12 20 (13%) 780 31 Vermifuge 0-3 59 (55%) 1736 41 4-6 28 (26%) 1505 22 7-9 12 (11%) 1254 11 10-12 9 (8%) 870 12 Vacciné 0-3 30 (43%) 1495 24 et 4-6 19 (27%) 1301 18 vermifuge 7-9 15 (21,1%) 1 109 16 10-12 7 (10%) 875 10 289 En mortalité relative par sous-classe d'âge, ce constat se trouve conforté par le fait que plus de 40% des morts de 0 à 12 mois intéressent les moins de 3 mois et ceci dans chaque lot chez les espèces étudiées. Nous notons enfin que les troubles de croissance viennent en tête des causes de mort quel que soit le lot prophylactique (tableau 3). La vaccination n'induit pas chez les béliers de différence significative, alors que la vermifugation engendre une baisse significative de la mortalité (P<0,05). Aucune différence n'est induite chez les brebis, les chèvres et les boucs, par aucun traitement (tableaux 4, 5, 6, 7). Tableau 2. Taux annuel de mortalité des caprins par classe d'âge en pourcentage, et mortalité relative par classe d'âge (Kolda 1985-1987) Lot Sous-classe d'âge Nombre de morts Cumul des Taux de (en mois) de 1985 à 1987 présents/mois mortalité Témoin 0-3 52 (44%) 1 154 54 4-6 36 (31%) 1018 42 7-9 20 (17%) 720 33 10-12 9 (5%) 556 19 Vacciné 0-3 69(43%) 1609 51 4-6 60 (37%) 1352 53 7-9 24 (15%) 974 30 10-12 9(5%) 574 19 Vermifuge 0-3 84 (55%) 1 777 57 4-6 42 (27%) 1463 34 7-9 17(11%) 993 21 10-12 10(7%) 659 18 Vacciné 0-3 56(75%) 1476 46 et 4-6 12 (16%) 1252 12 vermifuge 7-9 2(3%) 970 2 10-12 5(6%) 654 9 290 Tableau 3. Mortalité relative par cause chez les caprins (C) et les ovins (O) de moins de 3 mois (en pourcentage), Kolda 1985- 1987 Cause de mort(1) Témoin Vacciné Vermifuge Vacciné et Vermifuge 0 C 0 C 0 C 0 c DIV 38 32 26 31 42 36 37 39 TRC 40 56 37 31 52 43 37 54 PES 0 0 0 1 0 13 0 0 PNE 3 4 11 10 0 1 10 2 DIA 14 2 22 26 0 6 13 0 AUT 0 4 1 1 3 0 0 5 CUT 5 2 3 0 3 1 3 0 TOTAL 100 100 100 100 100 100 100 100 (1) Voir texte pour la signification des symboles Tableau 4. Effet de la vaccination chez les ovins. Cause Groupe vacciné Groupe non vacciné Nombre de Nombre de Différence morts Rang morts Rang de rang DIV 30 2 39 2 0 TRC 38 1 46 1 0 PES 0 7 0 7 0 PNE 11 4 1 6 2 DIA 20 3 5 3 0 AUT 1 6 2 5 1 CUT 3 5 4 4 1 r = 0,99 291 Tableau 5. Effet de la vermifugation chez les ovins. Cause Groupe vermifuge Groupe non vermifuge Difference de rang Nombre de morts Nombre de mortsRang Rang DIV 36 2 33 2 0 TRC 42 1 42 1 0 PES 0 7 0 7 0 PNE 3 4,5 9 4 0,5 DIA 4 3 21 3 0 AUT 2 6 1 6 0 CUT 3 4.5 4 5 0.5 r = 0,99 Tableau 6. Effet de la vaccination chez les caprins Cause Groupe vaccine Groupe non vaccine Difference de rang Nombre de Nombre de morts Rang morts Rang DIV 44 2 47 2 0 TRC 52 1 65 1 0 PES 1 6 11 3 3 PNE 8 4 3 5 1 DIA 16 3 6 4 1 AUT 4 5 2 6.7 1.5 CUT 0 7 2 6,7 0,5 r = 0,95 292 Tableau 7. Effet de la vermifugation chez les caprins Cause Groupe vermifuge Groupe non vermifuge Différence de rang Nombre de morts Rang Nombre de morts Rang ON 52 2 39 2 0 TRC 66 1 51 1 0 PES 11 3 1 6 3 PNE 2 6 9 4 2 DIA 5 4 17 3 1 AUT 3 5 3 5 0 CUT 1 7 1 6 1 r = 0,95 Analyse des causes de la mortalité Les coefficients de corrélation d'ordre déterminés ci-dessus sont tous positifs et proches de 1 . L'on en déduit que les rangs de chaque cause de mort sont très semblables à travers les groupes comparés dans chaque espèce, ce qui permet de conclure à un manque d'effet des diverses prophylaxies mises en oeuvre sur la hiérarchie des causes de mort. Conclusion Le PPR a mis à l'essai dès son démarrage différents scénarios prophylactiques pour éprouver en milieu réel l'efficacité des vaccinations antipestiques et antipasteurelliques et des vermifugations. Ces actions prophylactiques ont l'avantage d'être de vrais programmes minimaux d'intervention vétérinaire en élevage traditionnel extensifs qui ciblent des agents pathogènes réputés très limitant pour la productivité globale des troupeaux. Elles ont aussi l'avantage, du fait de leurs effets relativement spécifiques, de constituer pour l'épidémiologiste ce qu'un diagnostic thérapeutique représente pour un clinicien. L'étude du PPR rapporte les mortalités à des causes qui recouvrent la forme de syndrome et non d'agents ou de facteur causal, démarche conforme à la pratique de l'épidémiologie clinique. 293 En résumé, la mortalité chez les 0-3 mois est abaissée par le déparasitage significativement chez les ovins mâles seuls. Par contre, les vaccinations sont sans effet sur les taux de mortalité globale aussi bien chez les ovins que chez les caprins de moins de 3 mois. Ce qui peut s'expliquer par la protection immunitaire passive (non spécifique) dont ces jeunes bénéficient dans tous les cas. Ces résultats sont conformes à ceux du PPR obtenus par la méthode des quotients sur des cohortes d'animaux de même classe d'âge (PPR, 1988). Enfin, il importe de retenir que la hiérarchie des causes de mort dans la sous-classe d'âge des 0-3 mois est relativement stable à travers les groupes de lots prophylactiques. Ces constatations portent à croire que la mortalité des jeunes de 0 à 3 mois d'âge constitue une entité épidémiologique répondant à un déterminisme faisant intervenir en sus des facteurs et des agents parasitaires ou infectieux autres que ceux ciblés par les essais prophylactiques. Bibliographie Murray R S. 1975. Probabilité et statistique. Editions McGraw Hill, Paris (France). PPR (programme Pathologie et productivité des petits ruminants en milieu traditionnel). 1988. L'élevage traditionnel des petits ruminants dans la zone de Kolda (Haute Casamance). Réf. N° 018/vlro. ISRA/IEMVT/CIRAD/SENEGAL. Putt S.N.H., Shaw A.P.M., Woods A.J., Tyler L. et James A.D. 1987. Epidémiologie et économie vetérinaires en Afrique. Manuel du CIPEA n° 3. CIPEA, Addis-Abeba (Ethiopie). 294 Annexe 1 . Taux de mortalite annuel de 0 a 3 mois Lot Vermifuge et vaccineAnnee Temoin Vermifuge Vaccine Ovins 1985 61 56 49 5 mAles 1986 56 45 69 37 1987 52 35 58 23 Ovins 1985 53 31 43 30 femelles 1986 16 36 54 15 1987 53 40 157 29 Caprins 1985 76 66 73 53 males 1986 54 85 24 25 1987 48 27 71 36 Caprins 1985 35 56 45 76 femelles 1986 53 70 32 28 1987 48 23 79 60 295 Optimal time for vaccination against peste des petits ruminants (PPR) disease in goats in the humid tropical zone in southern Nigeria L.O. Wosu, J.E. Okiri and PA. Enwezor Department of Veterinary Medicine University of Nigeria, Nsukka, Nigeria Abstract Investigation of the incidence of peste des petits ruminants (PPR) disease in goats revealed seasonality in its natural occurrence in the environment. Tissue culture rinderpest vaccine was effective in protecting goats against the disease. The optimal time for vaccinating goats against PPR was when the least number of animals were incubating the disease. In the tropical humid zone of West Africa this would be in late November. Moment optimum de vaccination des caprins contre la peste des petits ruminants (PPR) dans la zone tropicale humide du sud du Nigeria L.O. Wosu, J.E. Okiri et P.A. Enwezor Résumé // est montré que l'incidence de la peste des petits ruminants (PPR) chez les caprins dans la zone tropicale humide du Nigéria a un caractère saisonnier. Le vaccin sur culture tissulaire contre la peste bovine s'est avéré apte à les protéger efficacement contre cette maladie. Cette étude montre que le meilleur moment pour vacciner les chèvres contre la PPR correspond à la période d'incidence minimum, au cours de laquelle le nombre d'animaux couvant la maladie est particulièrement réduit, soit de fin novembre à la mi-décembre dans la zone tropicale humide d'Afrique occidentale. 297 Introduction Peste des petits ruminants (PPR) disease is an acute or subacute viral disease of goats and sheep characterised by fever, stomatitis, gastro-enteritis and pneumonia. It is an endemic disease in West Africa especially in the humid zone. The disease is caused by a paramyxovirus, close to the rinderpest virus. PPR disease is of great economic importance. It was described as the most destructive viral disease against small ruminant flocks (Bourdin, 1983) and as number one constraint to intensive small ruminant farming. Mortality in susceptible flocks varies from 10% to 100% and morbidity from 50% to 100%. Tissue culture rinderpest vaccine (TCRV) is widely used in West Africa in the control of PPR disease. However, to date, there are conflicting reports both documented and mostly undocumented by field veterinarians as to the effectiveness of TCRV in controlling this disease. This study was therefore carried out in the tropical humid zone in southern Nigeria to determine whether this disease shows clear seasonality of occurrence and whether this accounts for the conflicting reports on the effectiveness of TCRV in controlling PPR. Materials and Methods Locality Nsukka is a small university town in a rural agricultural environment. It is located in the tropical humid zone in southern Nigeria. It lies approximately on latitude 06°52'N and longitude 07°24'E. Climatic data Rainfall and mean daily temperature data on monthly basis over a five-year period (1982-1986) in Nsukka environs were obtained from the university farm meteorological unit where the records were maintained on daily basis. Prevalence of PPR Two veterinary clinics in the town, the University veterinary teaching hospital and the town veterinary clinic were patronised by small ruminant owners from Nsukka and the surrounding villages within 15-20 km radius of the town. From daily treatment records in both clinics over the period of study, data were collected on monthly basis of the first-time reports of cases characteristic of PPR. Cases of pneumonia only or parasitic gastro-enteritis shown in the records were excluded from the data. Data from the two clinics were pooled on monthly basis 298 Source of Tissue Culture Rinderpest Vaccine (TCRV) The TCRV was obtained from the National Veterinary Research Institute (NVRI), Oji River depot. Method of vaccination with TCRV in each location One-hundred dose vial of freeze-dried TCRV was reconstituted in 1 00 ml of sterile refrigerated normal saline. 1 ml of the reconstituted vaccine was injected into a goat subcutaneously in the neck region, with a sterile needle. Animals less than three months old were excluded from the experiment. With the co-operation of the owners, the number of goats in each household was split into two equal groups taking into consideration age, sex and body condition. One group was vaccinated and the other was not (control). Only clinically healthy goats with no history of previous vaccination were used in the project. The ages ranged from six months to four years. Animals were individually identified. Goats were vaccinated either at the end of November or end of January. a) Vaccination in November On 24th November 1988, 100 goats were vaccinated and 101 goats left as controls in various households as described above at Obukpa on the outskirts of Nsukka town. Twenty households were involved. On 28th November 1988, 100 goats were vaccinated and 103 goats left as controls in various households as described above at Afikpo about 1 20 km south of Nsukka. Thirty households were involved. b) Vaccination in January On 25th January 1989, 200 goats from 45 households in Nkpirimkpi, a village in another part of the Nsukka outskirts, were vaccinated and 200 goats left as control. All the locations were in the tropical humid zone in southern Nigeria. Observation period All animals in the project at the various locations were monitored for 12 weeks post-vaccination both by regular weekly visits and with the aid of the local veterinarian for evidence of PPR or any other illness in any of the experimental animals. 299 Table 1 . PPR incidence and climatic data, Nsukka environs, Nigeria. Cum monthly PPR* Av monthly RF Av monthly RH Av daily T1982 1983 1984 1985 1986 January PPR 10 11 9 10 11 51 RF 17.3 Nil Nil 6.9 8.9 6.6 RH 61 20 52 71 59 53 T 27 24 24.5 24.5 26.5 26.0 February PPR 11 14 12 13 12 61 RF 55.7 Nil Nil Nil Nil 11.1 RH 59 54 64 68 74 63.6 T 27.5 26.5 26.5 24.5 28 27 March PPR 15 15 13 16 15 72 RF 56.4 Nil 47.8 153.9 83.0 68.3 RH 69 51.7 69 70 73 16.5 T 27 26.5 26 25.5 27 26 April PPR 19 16 17 15 16 83 RF 113.4 46.7 138.1 120.3 179.8 120 RH 76 68 83 74 76 75.4 T 27 28 25 24.5 27.5 26 May PPR 10 24 21 20 12 87 RF 217.2 114.4 129.3 264.5 172.1 180 RH 81 78 87 77 72 80 T 25.5 25 24 23.5 26 25 June PPR 13 15 13 12 11 64 RF 255.4 198.5 232.8 139.2 195.1 204 RH 86 86 88 83 82 84.8 T 23.5 23.5 23 22.5 25 24 July PPR 5 7 6 4 6 28 RF 312.9 158.7 379.1 276.2 217.5 269 RH 90 89 85 90 87 88.2 T 23 23 22.5 21.5 24 23 Agust PPR Nil 3 2 3 5 13 RF 108.9 102.4 294.4 191.1 141.9 178 RH 90 92 88 87 86 88.6 T 22.5 22 22 24.5 23.5 23 300 Table 1 (continued) Cum monthly Av monthly Av monthly Av daily 3PR* RF RH 1982 1983 1984 1985 1986 T Sept. PPR 2 Nil 3 5 Nil 10 RF 228.5 241.2 253.6 234.3 248.4 241 RH 83 87 90 84 13 86.4 T 23 23 22 14.5 24.5 23 October PPR 2 2 Nil 3 4 11 RF 302.1 55.8 126.2 192.9 116.0 159 RH 85 79 89 77 80 82 T 23.5 2.3 23 25.5 25 14 November PPR 7 6 11 5 7 36 RF 19.0 33.1 14.4 17.3 19.3 23 RH 64 75 82 76 74 74.2 T 25 24.5 23 25.5 24.5 25 December PPR 6 7 3 6 4 26 RF Nil 11.2 Nil Nil Nil 2.3 RH 66 67 48 48 41 54 T 25 24.5 23 25.5 24.5 25 PPR = Peste des petits ruminants; RF = Rainfall (mm); T= Temperature (C); RH Relative humidity (%). Results Table 1 shows the monthly occurrence of PPR, rainfall, mean daily relative humidity and temperature for each year of the five-year period of the study undertaken in Nsukka environs. The cumulative monthly PPR incidence and rainfall over the period are shown in Figure 1 . Table 1 and Figure 1 show that PPR occurs throughout the year. PPR incidence starts to build up from end of December and rises rapidly during harmattan season to a peak in the early rainy season in April. There was a sharp fall in the incidence over the latter part of the rainy season (May to September) in each year of study. Daily records showed no reports of new PPR cases within 40 hours of rainfall. The effects of vaccinating goats with TCRV against PPR are shown in Table 2 for the November vaccinations and Table 3 for the January vaccination. 301 Figure 1 . 1982-1986 mean monthly rainfall and PPR incidence rate in Nsukka, i ▲ a Rainfall & ^ PPR incidence PPR incidence rate (%) 90 \y Jan Fob Mar Apr May Jun Jul Aug Sep Oct Nov Dec 302 Table 2. Effect of November vaccination. Obukpa Afikpo Vaccinated Non-vaccinated N=100 N = 101 Vaccinated Non-vaccinated N=100 N=103Location No of PPR cases 0 9 0 8 Table 3. Effect of January vaccination at Nkpirimki location. Vaccinated N = 200 Non-vaccinated N = 200 Total PPR cases 16 3 83 Mortality from PPR 22 Discussion Though PPR seems to occur throughout the year, there is a definite indication of seasonality in its natural occurrence in the Nduaka environs (Figure 1). The incidences were higher during the dry harmattan season than during the rainy season. It appears that the inclement dry, cold and dusty harmattan weather (December-February) and the poor nutrition by this time combine to spread PPR. The incidence seems to rise rapidly from December to a peak in April. Obi (1983a) and Durojaiye (1983) in western Nigeria also associated most outbreaks with this season of the year. In Oyo State (western Nigeria), Opasina (1983) observed two out of four outbreaks in the early dry season and an other two in the last third of the dry season (January). These observations were in a climatic zone similar to that in Nsukka. Bourdin (1983) on the other hand stated that in the Sahelian climate of Senegal, outbreaks of PPR occur mostly during the rainy season. Reported occurrence of new outbreaks within 48 hours of rainfall as in Ondo State in Nigeria (Ojo, 1983) could not be confirmed in this study. January to April appeared to be the period of greatest risk of small ruminants to PPR each year (Table 1). This raises the question of how long a naturally acquired immunity after a field outbreak lasts. The 3-5 year cycle of occurrence reported by Obi (1983) was not observed in this study. There have been conflicting reports on the effectiveness of Tissue Culture Rinderpest Vaccine (TCRV) to control PPR in goats and sheep. Effective vaccination against PPR is a major advance in the control of the disease. According to Gibbs et al (1977) TCRV confers complete immunity to goats without transferring the disease to other animals. Adu and Nawathe (1981) reported that even pregnant animals tolerated TCRV well. On the contrary, Abegunde (1983) suggested that if animals are not screened for sub-clinical infection prior to vaccination, they may come down with the disease and that 303 abortions may occur in pregnant animals. There are more undocumented similar conflicting reports by field veterinarians. The results of various times of vaccination with TCRV to protect the goats against PPR in this study seem to confirm the observations made by Abegunde (1983). When the goats were vaccinated with TCRV in November, a period of sub-clinical infection, none of them came down with PPR infection while about 8.5% of the control came down with PPR. However, when the vaccination was done in January, a period of peak infection, about 8% of PPR cases including 1 .5% mortality were recorded among the vaccinated animals. About 41 .5% of PPR cases including about 1 1 .0% mortality were recorded among the control. Conclusion Our work thus confirms that TCRV is very effective in protecting goats against the PPR disease, but that this efficacy only holds if the vaccination is done when the animals are at a subclinical level of infection (Abegunde, 1 983). In the Nsukka environment in Nigeria, the end of November seems to be the right period for TCRV vaccination. Acknowledgement The authors acknowledge with gratitude the cooperation received from the farmers and field colleagues in carrying out this investigation. We also acknowledge the cooperation of the staff of the town veterinary clinic and teaching hospital and financial support of the National Committee for Small Ruminant Research, Nigeria. References Abegunde A. 1983. Problems connected with TCRV vaccination of sheep and goats. In: Hill D H (ed), Pesfe des petits ruminants (PPR) in sheep and goats. Proceedings of the international workshop held at IITA, Ibadan, Nigeria, 24-26 September 1 980. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia, pp. 79-81. Adu F D and Nawathe D R. 1981 . Safety of tissue culture rinderpest vaccine in pregnant goats. Research note. Tropical Animal Health and Production 13:166. Bourdin P. 1 983. History, epidemiology and economic significance of PPR in West Africa and Nigeria in particular. In: Hill D H (ed), Pesfe des petits ruminants (PPR) in sheep and goats. Proceedings of the international workshop held at IITA, Ibadan, Nigeria, 24-26 September 1980. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia, pp. 10-11. Durojaiye O A. 1983. Brief notes on history, epizootiology and the economic importance of PPR in Nigeria. In: Hill D H (ed), Peste des petits ruminants (PPR) in sheep and goats. Proceedings of the international workshop held at IITA, Ibadan, Nigeria, 24-26 304 September 1980. ILCA (International Livestock Centre for Africa), Addis Ababa. Ethiopia. pp. 24-27. Gibbs E P J, Taylor W P and Lawman M J P. 1977. The isolation of adenoviruses from goats affected with peste des petits ruminants (kata and stomatitis pneumoenteritis complex) in Nigeria. Research in Veterinary Science 23 (3): 331-335. Obi T U. 1983a. Symptomatology. In: Hill D H (ed), Peste des petits ruminants (PPR) in sheep and goats. Proceedings of the international workshop held at IITA, Ibadan, Nigeria, 24-26 September 1980. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. p. 10. Obi T U. 1983b. Results of TCRV vaccination in Nigeria. In: Hill D H (ed), Peste des petits ruminants (PPR) in sheep and goats. Proceedings of the international workshop held at IITA, Ibadan, Nigeria, 24-26 September 1980. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 82-83. Ojo MO. 1983. Role of other agents and factors responsible for the pathogenesis of PPR. In: Hill D H (ed), Peste des petits ruminants (PPR) in sheep and goats. Proceedings of the international workshop held at IITA, Ibadan, Nigeria, 24-26 September 1980. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. p. 54. Opasina B A. 1983. Epidemiology of PPR in the humid forest and the derived savanna zones. In: Hill D H (ed), Pesfe des petits ruminants (PPR) in sheep and goats. Proceedings of the international workshop held at IITA, Ibadan, Nigeria, 24-26 September 1980. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 14-21. 305 Prophylaxie chez les petits ruminants au Sénégal: régionalisation d'une politique nationale de protection sanitaire 0. Faugère, E. Tillard et B. Faugère ISRA-IEMVT Laboratoire national de l'élevage et de recherches vétérinaires B.P. 2057, Dakar-Hann (Sénégal) Résumé Les pneumopathies et les strongyloses gastro-intestinales sont les dominantes pathologiques des petits ruminants au Sénégal. Cette communication présente les résultats d'une évaluation technico-économique des actions de prophylaxie anti-infectieuse et antiparasitaire qui peuvent être proposées aux éleveurs de petits ruminants au Sénégal. L'évaluation biologique a été réalisée sur 6 500 ovins et caprins en milieu paysan dans le cadre duprogramme ISRA-IEMVT "productivité etpathologie des petits ruminants ". L 'évaluation économique a utilisé le progiciel de simulation LIVMOD. Prophylactic measures for small ruminants in Senegal: Regionalisation of a national health protection policy 0. Faugère, E. Tillard and B. Faugère Abstract The main small ruminant diseases in Senegal are lung diseases and gastro-intestinal strongyle infections. This paper presents the results of the technical and economic assessment of a series of parasitic and infectious disease-control measures intended for small ruminant owners in Senegal. 307 The biological performances of 6500 sheep and goats have been evaluated on-farm. This assessment is part of the ISRA-IEMVT "Diseases and productivity in small ruminants" programme. The LIVMOD Software was used for the economic calculations. Introduction Les dominantes pathologiques des petits ruminants au Sénégal sont les pneumopathies et les strongyloses gastro-intestinales. Cette communication présente les résultats d'une évaluation technico-économique des actions de prophylaxie anti-infectieuse et antiparasitaire qui peuvent être proposées aux éleveurs de petits ruminants du Sénégal. Cette évaluation a été réalisée dans le cadre du programme "Pathologie et productivité des petits ruminants" de l'ISRA (Institut sénégalais de recherches agricoles) et de l'IEMvT (Institut d'élevage et de médecine vétérinaire des pays tropicaux). Dispositif expérimental Zone d'étude L'étude a été menée dans trois sites correspondant à des zones écologiques et des systèmes d'élevage différents: • le site de Ndiagne, dans la partie nord du pays, avec 3 500 petits ruminants en expérimentation. Ce site sahélien reçoit de 250 à 400 mm de précipitations et est peuplé de paysans Wolof essentiellement cultivateurs, entretenant un petit troupeau d'une quinzaine d'ovins de type Peul-Peul, et de paysans peuls essentiellement éleveurs et possédant des troupeaux d'une trentaine d'ovins (Peul-Peul) et caprins (sahéliens), • le site de Kaymor, dans la partie centre du pays, avec 1 500 petits ruminants en expérimentation. Ce site soudanien reçoit 650 mm de précipitations et est peuplé de paysans Wolof, essentiellement cultivateurs, entretenant un petit troupeau de 10 à 15 têtes d'ovins et de caprins issus d'un métissage ancien entre grandes races sahéliennes et races naines du Sud; • le site de Kolda, dans la partie sud du pays, avec 1 500 petits ruminants en expérimentation. Ce site soudano-guinéen reçoit 950 mm de précipitations et est peuplé de paysans peuls cultivateurs et éleveurs, entretenant dans chaque concession un petit troupeau de 10 à 15 têtes d'ovins Djallonké et de caprins guinéens. 308 Contraintes pathologiques Il a été décidé de tester parallèlement, en traitant tous les animaux de 3 mois et plus, l'efficacité en termes de "production ajoutée" de: • la vermifugation contre les strongles (Panacur), par deux administrations en début et en fin de saison des pluies; • la vaccination antipasteurellique (vaccin Pasteurellad ND du LNERV-Dakar), par deux injections à six mois d'intervalle (avant la saison sèche fraîche et relance de l'immunité au cours de la saison sèche chaude); • la vaccination antipestique hétérologue (vaccin Tissupest ND du LNERV-Dakar), par une injection au cours de la saison des pluies. Echantillonnage Le dispositif expérimental comprend un lot témoin, un lot vermifugé, un lot vacciné, un lot vacciné et vermifugé. L'échantillonnage a été réalisé en grappes, tous les animaux d'un village faisant partie du même lot prophylactique. Les résultats zootechniques sont évalués à partir des performances de reproduction, de croissance et de viabilité. Une synthèse des résultats peut être faite en utilisant deux paramètres : • l'indice de productivité pondérale annuelle: le nombre de kilogrammes d'agneau (de chevreau) de 3 mois produits par brebis (chèvre) et par an, qui rend compte de la reproduction de la mère, de la viabilité et de la croissance des produits avant 3 mois d'âge; IPP = taux de mise-bas X (taux de productivité numérique à la naissance) X ( 1 - quotient de mortalité 0-3 mois ) X (poids moyen à 3 mois (mâle-femelle) des produits) • le poids des femelles à 18 mois qui rend compte de la croissance des agnelles ou chevrettes (PAT). L'analyse économique a été réalisée à l'aide du progiciel LIVMOD développé par la Banque internationale pour la reconstruction et le développement, la FAO et NEMvT, en comparant sur cinq années l'évolution démographique de deux populations animales identiques en année 1, l'une bénéficiant des actions prophylactiques, l'autre n'en bénéficiant pas. Les différentes variables qui caractérisent chaque situation sont d'abord introduites dans le modèle: 309 • paramètres zootechniques (structure de la population, mortalité, reproduction, croissance); • paramètres d'exploitation (immigration, émigration); • prix au producteur et coûts d'intervention. Trois types d'approche sont ensuite développés: a) Le calcul du coût de la maladie, en rapportant la différence de bénéfices (bénéfices d'exploitation + capital animal induit) entre la situation "avec action" (B,) et la situation "sans action" (B0), au nombre total d'animaux en première année : Bi - Bo coût de la maladie = N b) Le calcul de l'amélioration relative des bénéfices, en rapportant le bénéfice supplémentaire (actualisé) au bénéfice de la situation "sans action": Bi - Bo amélioration relative des bénéfices = Bo c) Le taux de rémunération des dépenses prophylactiques, en rapportant le bénéfice supplémentaire (actualisé) au coût des interventions: Bi - Bo taux de rémunération des dépenses = Résultats Techniques Le tableau 1 montre: • que la vermifugation permet d'améliorer au moins une catégorie de performances animales dans chacune des zones et pour chacune des deux espèces. Elle est toutefois plus efficace pour les ovins dans le sud du pays; • que la vaccination antipasteurellique pratiquée seule n'est efficace chez les ovins (non testée chez les caprins), que dans la partie nord du pays; • que la vaccination antipestique (associée à la vaccination antipasteurellique) est efficace au nord comme au sud du pays tout au moins pour les ovins. Ces constatations amènent à proposer une régionalisation des interventions, qui concernerait les deux espèces: • vermifugation sur toute l'étendue du territoire et plus particulièrement dans la partie sud; • vaccination antipestique sur toute l'étendue du territoire; 310 i i s I 8 1 < 0- E CO CL CL < E cn D. E CD < 0. E CD Q- > o f- »- o a. % O o 2 O 311 • vaccination antipasteureilique dans la partie nord du pays. Ces recommandations doivent être évaluées sur le plan économique avant d'être retenues. Economiques Le coût des interventions a été calculé: • en considérant que les structures des services de l'élevage se chargeraient de l'organisation des campagnes de vaccination, tandis que des structures privées se chargeraient de promouvoir et de distribuer les vermifuges; • en répercutant l'intégralité des coûts d'intervention (prix du produit + promotion + distribution + conservation + salaires des agents pendant les campagnes) sur le prix du traitement au niveau de l'éleveur qui s'établit ainsi: - vaccination antipestique 33 F CFA/animal/an - vaccination antipestique/antipasteurellique 93 F CFA/animal/an - vermifugation 320 F CFA/animal/an. Ces trois approches ont permis de montrer que les actions régionalisées proposées étaient toutes économiquement rentables. A titre d'exemple, nous présentons dans le tableau 2 les résultats de ces calculs pour la zone nord que l'on peut brièvement commenter: 1- En termes zootechniques le parasitisme digestif et le complexe pneumopathie sont des contraintes plus nettes chez les caprins que chez les ovins, nous l'avons vu. En revanche, en termes monétaires ils entraînent un manque à gagner (coût de la maladie) plus important chez les ovins. Ceci en raison du prix au producteur beaucoup plus élevé pour les ovins (350 à 450 F CFA/kg pour les femelles, 700 à 850 F CFA/kg pour les mâles) que pour les caprins (350 F CFA/kg pour les deux sexes). 2- La prophylaxie anti-infectieuse génère une amélioration relative des bénéfices plus importante (14,5% pour les ovins) que la prophylaxie antiparasitaire (6,5%). Ceci du fait du coût plus élevé des actions de vermifugation (320 F CFA/animal/an contre 93 F CFA pour la vaccination associée). 3- Pour une prophylaxie donnée, l'amélioration relative des bénéfices est identique pour les deux espèces. Cependant, la rentabilité des investissements est plus importante pour les ovins (2 à 3 fois supérieure), ce qui traduit encore le fait que la valeur commerciale des ovins est supérieure à celle des caprins. Un franc investi chez les ovins rapporte deux fois plus que chez les caprins, alors que l'amélioration des performances est plus importante chez ces derniers. 312 Tableau0.Analyse6conomiquedesactionpr phylactiquesnlzo eord Tauxde remuneration 00N 0300- Amelioration relativedes benerices 7N 0% nsnsnns, Coutmaladie 070 0007 CoOtintervention 007 007 Amelioration I.P.P. 0N 0- Vaccination Antipestiqueet Antipasteurellique prophylactique Vermirugation Action Pneumopathie Maladie Helminthes Caprins HelminthesV r rugation 7% 007 007 70- 00N Vaccination PneumopathieAntipestiquet Antipasteurellique 00- 007 007 30% 030N On peut considérer que, pour être incitatrice au niveau du producteur, la rémunération des investissements doit être d'au moins 200% (2 francs de bénéfice pour un franc investi). Dans ces conditions, il apparaît que la vermifugation n'est pas suffisamment rentable pour être adoptée par les éleveurs de cette zone sans subvention de l'intervention. Au contraire, la prophylaxie anti-infectieuse est apparemment suffisamment rentable pour que les éleveurs adoptent cette action. Le problème de leur capacité effective à mobiliser des liquidités pour ce type d'action reste néanmoins posé. Conclusion L'originalité de cette étude réside dans son exécution: • en grandeur réelle ayant impliqué 6 500 petits ruminants dans le dispositif expérimental; • en milieu villageois sur cinq années, ce qui permet de tenir compte à la fois des conditions de production traditionnelle et de leur variation annuelle; • étendue du nord au sud du pays, ce qui amène à établir des recommandations régionalisées en matière de prophylaxie. Elle est aussi originale dans son exploitation qui, premièrement, prend autant, sinon plus, en compte les pertes qualifiées d'indirectes, intéressant les performances de reproduction et de croissance, que les pertes directes par mort des individus, et qui, deuxièmement, fournit des éléments techniques mais également économiques pour raisonner la régionalisation d'une politique nationale de protection sanitaire du cheptel. Cette communication montre l'intérêt des études multilocales coût-bénéfice des actions d'amélioration. Elles permettent d'une part d'établir une régionalisation de la politique nationale de protection sanitaire, raisonnée sur le plan technique et économique, et d'autre part d'identifier les actions suffisamment rentables du point de vue national pour être justifiées, mais qui nécessitent une subvention pour être adoptées par les éleveurs car elles ne déterminent pas un gain suffisamment attractif à leur niveau. 314 en z > o < Q _l o O < z < 0- LU DC to < CO t Z < CO a. c < 9 Mz < LU CO X Q. O CO Q. t) x e (J o o o in in in co co co o co" CO o t- co y- CO in in in in in m in N CM CM co CO CM co co CO CM CO CMCM CM CM CM CMh.COCOCOCOCOCO CO in ob. o" p o" d o o o r-r-CMCMCMCMCMCM CM in o 8 o 8 8 COCO to COCOCM CO CO - mo tr oj CM CM TCM CM CM r^ CO m co in m in in in m in in in m in f- CM m CO CO CO (O CO CO o o o in o o o o §CM CO fc t CO o CO CO CO CO CO CO CO CO 1^ t^ r^ N N E «" "~" £. i CO cD 8 CM CO c CO C CO to c COCO •- CM CO CO in s to CO 8CO CO CO 8 in m in in in •-CO CO CO CO CO T CO co CO CO CO CO CO in" t\j co" CO co" co" to" to' to" co CM oo"8 CO t » t" t. T. *T~ 8 8 8 8 in in to in CO o o CO CO co" r- CM °P co V, 2 A en i— co in co in o T o o to CO to to i 6 1 cD1 — CM CO m to 2 c o CM CO W c J. A cr * o_ -to 0- C E o E E« „ 315 Gastro-intestinal parasites in small ruminants in Mali: Geographical distribution epidemiology and chemotherapy S. Tembely\ T. J. Galvin2, B. Kouyate\ S.B. Ba3, K. Bengaly3 and W. Berckmoes4 1 Laboratoire Central Vetérinaire, B.P. 2295, Bamako, Mali 2 United States Department of Agriculture (USDA) 3 Volet Recherche, Projet Sectoriel de l'Elevage 4 Departement de la Recherche sur les Systèmes de Productions Rurales, Kikasso, Mali Abstract An investigation ofhelminth parasites in sheep and goats was carried out in several abattoirs in various regions of Mali between 1983 and 1985. Faecal and visceral examinations from 284 sheep and 318 goats indicated a marked variation in the level of parasitism in livestock raised in different geographic areas. Both sheep and goats were found to be infested with identical nematode species, however, with different levels of infestation. The more common nematodes in sheep were, in order of predominance: Trichostrongylus colubriformis, Haemonchus contortus, Strongyloides papillosus, Oesophago- stonum columbianum, Gaigeria pectinata, C. punctata, C. curticei, and Trichuris ovis. Fourtrematodes (Fasciola gigantica, Dicrocoelium nospes, Paramphistomum spp and Carmyerius spp), one larval (Cysticercus tenuicollis) and three adult cestodes (Monezia, Stilesia, and Avitellina) were identified. The following nematodes were found in goats in order of predominance:!, colubriformis, S. ovis, H. contortus, G. pachyscelis, S. papillosus, O. colombianum, T. axei, C. pectinata, and C. punctata. Trematodes and cestodes identified were the same as those in sheep. 317 Variation in geographic distribution and prevalence were noted in relation to the season and management practices. Since 1987 efforts are being made in two ecological zones to assess the impact ofmass anthelminthic treatment on production parameters such as weight and mortality. Parasitisme gastro-intestinal ties petits ruminants au Mali. Epidemiologic et controle S. Tembely, T.J. Galvin, B. Kouyate, S.B.Ba, K. Bengaly et W. Berckmoes Resume De 1983 a 1985, le parasitisme a helminthes des petits ruminants a ere Studie dans les abattoirs de plusieurs regions du Mali. L 'examen des f6ces et des visceres de 284 moutons et 318 chevres a montre la variation du niveau d'infestation parasitaire des animaux eleves dans les différentes régions du pays. Les deux especes sont infestees par les memes parasites, mais a des degres differents. Les nematodes des ovins prSdominants sont dans l'ordre d'importance: Trichostrongylus colubriformis, Haemonchus contortus, Strongyloides papillosus, Oesophagostonum columbianum, Gaigeria pectinata, C. punctata, C. curticei, and Trichuris ovis. 4 trématodes (Fasciola gigantica, Dicrocoelium nospes, Paramphistomum spp et Carmyerius spp); un cestode larvaire (Cysticercus tenuicollis) et trois cestodes adultes (Monezia, Stilesia et Avitellina). Les nematodes des caprins sont dans l'ordre d'importance: T. colubriformis, S. ovis, H. contortus, G. pachyscelis, S. papillosus, 0. colombianum, T. axei, C. pectinata et C. punctata. Les caprins sont parasites par les memes trematodes et cestodes que les ovins. Des variations de repartition géographique et de prSvalence furent notees en liaison avec la saison et les pratiques de gestion. Depuis 1987, l'impact de traitements anthelminthiques a grande echelle sur le poids et la mortalite a 6te lvalue dans deux zones ecologiques. 318 Introduction Sheep and goats, although representing an important source of animal protein in semi-arid countries such as Mali, seem to have benefited little from veterinary care and production improvement. Goats are often the main supply of daily meat in rural areas. Sheep, on the other hand, are used in ceremonial festivities throughout the country, providing important trade (mainly male sheep) between rural and urban areas. Together, sheep and goats provide a large portion of the meat consumed and produce a considerable amount of manure, which is of special importance in those areas where cattle are of lesser importance. Different breeds have been adapted to different environments. Long-legged sheep and goats are found in the Sahelian regions, and the dwarf breed are found in the Sudanese region of Mali. Although adapted to local climatic and nutritional conditions, economic production of small ruminants is hampered by infectious and parasitic diseases coupled with inadequate management as mentioned by Nawatheetal (1985). Epidemiology Materials and methods Sites of collection In order to obtain adequate numbers of samples from the animals, all collections were made at abattoirs. To obtain data on seasonal variations in parasite populations and biology, attempts were made to make a minimum of four collections at most sites. Points of collection are shown in Figure 1 . Selection of the abattoirs was based on the number of individual animals slaughtered daily and the probability of obtaining native animals from the various ecological regions of Mali. Climatic data Monthly temperature and rainfall data were obtained from the National Agroclimatological Office at Bamako. Parasitological Findings Nematodes Sahelian sheep The following parasites species were recorded from Sahelian sheep in order of predominance: Trichostrongylus colubriformis, Haemonchus contortus, 319 Strongyloides papillosus, Cooperia curticei, Oesophagostomum columblanum, Skrjabinema ovis and Trichuris ovis. Figure 1 . Map of Mali showing major rivers and collection sites. Senegal R Niger BougouniJ/ .fsikasso •^ Kimparana J S ( "J •J Vi/V Kadiolo BaouleR. BagoeR. Sudanese dwarf sheep The results obtained from this survey may not reflect the overall field situation due to a limited sample size from the abattoirs visited. Sahelian goats Based on post-mortem examination, the following nematodes were identified in long-legged Sahelian goats in order of predominance: Skrjabinema ovis, S. papillosus, H. contortus, T. colubriformis, O. columbianum and Cooperia curticei. 320 Sudanese dwarf goats Only two nematode species were found in significant numbers. These were T. colubriformis and H. contortus. Prevalence figures are given in Table 1 . There was a relatively high prevalence of G. pachyscelis infection (27 %) observed in these goats. Table 1 . Prevalence of trematode parasites of sheep and goats in Mali (%). Sheep Goats Sahelian Sudanese Saheliar1 Sudanese Fasciola gigantica Dicrolelium hospes Paramphistomes Schistosoma sp Cysticercus tenuicollis 20 0 7 0 0 0 18 35 18 34 3 42 44 0 0 46 Trematodes 1 0 19 42 Sahelian sheep Fasciola gigantica and paramphistomidae infections occurred with a prevalence of 20% and 35%, respectively. Only one sheep was found to be infected with Schistosoma spp. No Dicrocoelium was found during the course of the investigation. Sudanese dwarf sheep Not a single F. gigantica was found to infect the local dwarf sheep in the southern part of the country. However, 18% of the animals were infected with both D. hospes and paramphistomes. Due to a limited sample size, the results of this investigation concerning the dwarf sheep may not reflect the overall field situation. Sahelian goats Trematodes were found with the following frequency: Paramphistomum spp (34%), F. gigantica (7%) and Schistosoma spp (3%). Sudanese dwarf goats No Fasciola was recovered in these goats in the course of this investigation. On the other hand, 44% of the animals examined were infected with Dicrocoelium hospes. 321 o rn , , m U) c rao OCO a> "—* r en CO 1 o 0- a> CO ro en r> d) (') m n -i to CO O 9 1 CO ;- x CO C g Spa c J! l»CD ^ x: CO CO ^ • gco CO ~~■a 3 CO CO CD O o O ';:- CO CD O a. t^ —i—r- o o co en —i—i—r- o o o m c\j «- iWi - H o o o r*. cd m g o o o* n n «- ra E 322 Cestodes The only larval cestode reported in this study is Cysticercus tenuicollis. Post-mortem examination showed the following prevalence: Sahelian sheep (19%), Sahelian goats (42%) and Sudanese dwarf sheep (42%) and Sudanese dwarf goats (46%). Discussion In the traditional management of animals in Mali, sheep and goats are grazed together. This parasitological survey has shown that they also share the same parasite species but with different levels of infection. Studies conducted in Australia also showed that sheep and goats were infected with the same species of worms when grazed together (Malins, 1971). However, conflicting results have been reported elsewhere in Africa (McCulloch et al, 1968) and in Europe (Le Riche et al, 1973) with different types of management. One explanation for the difference in worm burdens and composition in sheep and goats is that the latter they are browsers as well as grazers and, if feeding on scrub high enough, they could avoid most of the infective trichostrongyle larvae. Nematodes Haemonchus contortus Heavy infections coincided with the rainy season, then progressively declined during the dry season. Immature stages (most of which were at an early stage of development) were recorded during the dry season and, thus, probably contributed to the increase of the adult populations during the rainy season. These findings suggest that most of the larvae had undergone inhibition of development during the drier period of the year. The phenomenon, as described in cattle, is commonly known to occur in sheep in temperate regions (Donald et al, 1964; Dinneen et al, 1965; Michel, 1974) and recently recognised in tropical Africa. Studies done in Senegal (Vercruysse, 1985, have shown that larvae ingested between October and December were inhibition-prone, the adult populations resulting from the burden acquired during the previous August. Furthermore, the study showed that from January onwards, as the hypobiotic larvae started to mature, these new adults gradually replaced the older population and seeded the pasture with eggs. Therefore, the author concluded that starting in June-July-August, a new population of adult worms developed. In Mali, the peak of immature Haemonchus was observed in March among the Sahelian sheep and represented 91% of the rainy season. This decrease coincided with a sharp increase of the adult population suggesting the maturation of the immature worms. Additional data are required from the dwarf sheep in the Sudanese region to ascertain the survival mechanisms of the parasite. 323 Trichostrongylus colubrtformis Although the genus Trichostrongylus is considered to occur primarily in temperate regions and not thought to thrive during the warm rainy season (Grant, 1981), results from this study indicate that this species may be of economic importance in sheep in Mali because it is one of the most prevalent species both in the Sudanese and in the Sahelian regions. In the Sahelian sheep, T. colubrifirmis was found to be the most abundant nematode species with the highest mean intensity. Gaigeria pachyscelis Since the first report on the occurrence of the hookworm Gaigeria pachyscelis in sheep and goats in South Africa (Orlepp, 1935), a wide distribution of the parasite has been reported in tropical Africa. Graber (1965) recognised this parasite among native sheep of Senegal. Assoku (1981) in Ghana found only 1 % of the sheep infected with this hookworm. The present survey indicates a higher prevalence in the southern regions with 27 % and 15 % of sheep and goats being infected, respectively. On the other hand, the Sahelian counterparts had a much lower prevalence (7% for sheep and 2% for goats). These findings suggest that the parasite thrives in a region with higher rainfall. However, research is needed to ascertain the importance of the epidemiology of this parasite in Mali. Oesophago&tomum columbianum Results from earlier workers indicated that up to 66% in sheep of Chad were infected with O. columbianum and there was a wide distribution among the Senegalese sheep (Vassiliades, 1981). In Mali, it appears from this study that nodular form of the infection is of greater economic importance with a prevalence rate of 25 to 40%. The presence of nodules on viscera is of prime importance for the rural population because the infected intestines are subject to condemnation during meat inspection and are therefore lost for use as sausage casings. Cooperia spp These species occurred more commonly in cattle than in sheep and goats (less than 5% of the small ruminants were infected with the genus). The following species were identified: Cooperia pectinata, Cooperia punctata and C. curticeci. 324 Trematodes As indicated by gross liver and microscopic examinations for the presence of adult worms and eggs, 20% of the Sahelian sheep were infected with Fasciola gigantica, where as only 7 % of the goats in the Sahelian regions were infected with the same parasites. The difference in the level of infection and prevalence is probably due to difference in resistance to infection and grazing habits of the host species. Only 3 % of the sheep examined were found to be infected with Schistosoma bovis in the Sahelian region. Further investigation is needed to assess the pathology of schistosomes in small ruminants because of increasing reports of damage that the parasite might cause in sheep (Vercruysse, 1985). Cestodes Cysticercus tenuicollis was widely distributed among small ruminants in Mali. However, this parasite is known to be of low pathogenicity (Dunn, 1978). Other species, such as Moniezia, Avitellina and Stilesia were also commonly associated with sheep and goats. Control Strategy in the Semi-arid Zone Materials and methods The experiment was carried out in the Djoni village located in the Bamako country, 1 50 km north-east of Bamako with an average annual rainfall ranging from 500 - 600 mm. One hundred and fifty goats and 70 sheep aged from birth to 24 months were grouped at random into different age categories prior to anthelminthic treatment. Animals comprised exclusively Djallonke sheep and dwarf goats. Both animal breeds were owned by small individual households. They are kept around the villages during the rainy season and allowed to graze freely away from the villages without herding during the dry season. Priorto anthelminthic treatment, animals were divided intofive different treatment groups and all were immunised against pasteurellosis, anthrax, black leg and peste des petits ruminants (PPR). The anthelminthic treatments were as follows: No treatment (control group) T1 - received TBZ 66 mg/kg of liveweight T2 - fenbendazole 5 mg/kg 325 T3 - morantel tartrate 7.5 mg/kg T4 - levamisole 12 mg/kg Treatment was done in July and September 1988. Each month, faeces were obtained from the rectum and submitted to a quantitative parasitological examination using the McMaster modified method. Liveweights were recorded every two months and the average daily weight gain was determined. The effects of treatment on weight gain were analysed using the General Linear Model (SAS). Results and discussion The following strongyle eggs reduction rates were observed following the prime anthelminthic treatment. Fenbendazole: 100% in sheep; 93 to 100% in goats; Morantel tartrate: 98 to 99% in sheep; 90 to 99% in goats; Levamisole: 98 to 100% in sheep and 85 to 87% in goats; TBZ : 89% in sheep and 78% in goats. Fifty sheep out of 70 from the starting of the experiment were analysed for live weight changes. There was no significant difference (P>0.05) between the treated and the control group. One hundred and fourteen goats out of the 1 50 at the beginning of the experiment were analysed for the same purpose. When the two animal species were combined for statistical analysis a difference of 6 to 7 kg appeared. Since the standard deviations were very high the differences may not be conclusive (Table 2). 326 Table 2. Weight-change analysis on sheep and goats combined. General Linear Models Procedure N Mean SD Bolumisol 27 5.1109393 11.6966865 Exhelm 36 12.1871206 26.81456227 Panacur 38 11.8638104 17.9935661 Pas de traitem 31 5.1871548 22.9355640 Thybenzol 38 10.8740727 29.4437642 Control Strategy in the Humid Zone Material and methods The experiment was carried out at Fonsebougou, 56 km north of Sikasso in the southern part of Mali. Eight flocks of sheep owned by villagers were identified by ear tags and randomly divided into three groups of treatment according to their weight: Group 1 - Fenbendazole 10 mg/kg liveweight Group 2 - Morantel tartrate 7.5 mg/kg liveweight Group 3 - Control Treatments were given in May, July, September and November 1989. These groups were subdivided into three age categories: 6-12 months old; 12-18 month old and over 18 months. Parasitological examination: A quantitative evaluation of strongyle eggs was performed using the modified McMaster method. Liveweight changes: to determine any change that has occurred in liveweight, sheep were individually weighted every two months. No mortality was recorded. 327 Results and discussion Our findings demonstrated that lambs were more heavily infected by trichostrongyles because they were more susceptible. The maximum egg count was observed between August and October with a peak in September. However, ewes remained infected all year-round thus serving as the source of contamination for the lambs. No significant change was noted on liveweight between the treated groups and the control in adult sheep (ewes and rams). No statistical analysis could be done of weight data lambs because of the low number of observations by the end of the experiment. Bibliography Assoku K G. 1981. Studies of parasitic helminths of sheep and goats in Ghana. Bulletin of Animal Health and Production in Africa 29: 1-10. Dinneen J Ketal. 1965. The dynamics of the host-parasite relationship. III. The response of sheep to primary infection with Haemonchus contortus. Parasitology 55:51 5-525. Donald et al. 1964. The dynamics of the host-parasite relationship. I. Nematodirus spathiger infection in sheep. Parasitology 54:527-544. Dunn A M. 1978. Veterinary helminthology. 2nd edition. William Heineman Medical Books Ltd, London, UK. pp. 323. Graber M. 1965. Helminthes et helminthiases faisant obstacle a I 'amelioration de la production ovine en Republique du Tchad, IEMVT (Institut d'elevage et de medecine veterinaire des paystropicaux), Fort Lamy Tchad, Laboratoire de Farcha. Republique du Tchad. 162 pp. Grant J L. 1981 . The epidemiology of nematode parasites of sheep in a high rainfall area of Zimbabwe. Journal of the South African Veterinary Association 52:33-37. Le Riche P D et al. 1973. A helminth survey of sheep and goats in Cyprus, Part I. Age distribution and the severity of infection with gastro-intestinal parasites. Journal of Helminthology 47:251-262. Malins F. 1971. Angoras in a Merino world. Journal of the Victoria Department of Agriculture 69: 1 80-1 88. McCulloch B et al. 1968. The incidence of gastrointestinal nematodes of sheep and goats in Sukumaland, Tanzania. The British Veterinary Journal 124(5):177-194. Michel J F. 1974. Arrested development of nematodes and some related phenomena. Advances in Parasitology 12:279-366. Nawathe D R, Sohael A S and Umo I. 1985. Health management of a dairy herd on the Jos Plateau (Nigeria). Bulletin of Animal Health and Production in Africa 33: 199-205. Orlepp R J. 1935. Observation on the morphology and life history of Gaigeria pachyscelis, Railliet and Henry, 1910. Southern African Journal of Science 33:875-876. Vassiliades G. 1981. Parasitisme gastro-intestinal chez le mouton du Senegal. Revue d'Elevage et de Medecine Veterinaire des Pays Tropicaux 34:169-177. Vercruysse J. 1985. The seasonal prevalence of inhibited development of Haemonchus contortus in sheep in Senegal. Veterinary Parasitology 17:159-163. 328 Session 4 Small ruminant feeds and feeding systems Ressources alimentaires et systemes d'alimentation des petits ruminants Strategies for matching feed resources to small ruminant needs: A review R.M. Njwe Department of Animal Science Dschang University Centre B.P. 222, Dschang, Cameroon Abstract This paper reviews strategies for matching feed resources to small ruminant production. Climatic, ecological and socio-economic conditions prevailing determine the types of strategy to adopt. Strategy options proposed include limiting small ruminant numbers to suit available feed resources, efficient use of natural browse resources, increased use of legumes in cropping and pastoral systems, integration of small ruminant production with tree crop plantations, more effective utilisation of crop residues and agro-industrial by-products, conservation of forages as silage or hay for use during periods of scarcity, strategic animal feeding schemes and use of fertiliser to improve forage yield where and when feasible. Evaluation des strategies visant a faire correspondre les ressources fourrageres aux besoins des petits ruminants R.M. Njwe Resume Cette etude passe en revue les stratégies a suivre pour que les ressources fourrageres puissent couvrir les besoins de I'élevage des petits ruminants. Le choix du type de strategies a adopter s'effectue en tenant compte des conditions climatiques, ecologiques et socio-economiques prevalantes. Diverses strategies sont 329 proposées, notamment: diminuerles effectifs de maniere a equilibrer le nombre d'animaux avec les ressources fourrageres disponibles; tirer parti des especes ligneuses spontanees; intensifier l'exploitation des légumineuses dans les systemes agricoles ou pastoraux; associer l'élevage des petits ruminants a la pratique des cultures arboricoles; am&iorer l'utilisation des residus de récolte et des sous-produits agro-industriels; recourir a I 'ensilage etau fanage afin de conserver les fourrages pour les periodes de penurie; mettre en place des plans raisonnes d alimentation; et, lorsque c'est faisable, utiliser des engrais pour augmenter les rendements fourragers. Introduction Africa, excluding South Africa, has about 22% of the world's population of sheep and goats, and produces 14% of the total sheep and goat meat. Goat and sheep milk production in Africa also accounts for approximately 14% of the world production (FAO, 1981). Small ruminants represent 24% of the total livestock units in Western Africa and 17% in eastern Africa (Peacock, 1985). Their distribution tends to favour the arid and semi-arid zones of tropical Africa where 58% of sheep and 68% of goats are found. Dwarf goats and sheep are more abundant in the humid zone of West and central Africa particularly as they have the advantage of being trypanotolerant. Small ruminant production is typically a small-scale farmer activity attracting minimum investment in housing, feed and health care and is largely sustained by the potential of the individual breeds themselves. Three distinct management systems can be distinguished: free roaming (extensive) or herding, tethering and confinement of animals. Tethering and confinement are associated with areas characterised by high population densities, more intensive cropping, land scarcity and high labour requirements. In most of Africa, an extensive management system of production seems most popular for both sheep and goats. Among the numerous constraints to production under these systems, those related to nutrition are of primary concern. The nutritional problem Gatenby (1982) has indicated that the major factors limiting the productivity of small ruminants in Africa are poor nutrition in the semi-arid areas and diseases in more humid areas. Water can be a problem in some dry areas. Contrary to popular belief, the genotypes of breeds in Africa are not the primary limiting factors especially where nutrition, disease and marketing problems have been adequately handled. 330 The nutrient content in the form of metabolisable energy, digestible protein and minerals may be poor in arid and semi-arid regions, even when vegetation appears to be adequate. The dry season poses serious feed problems in semi-arid areas. There are three aspects of the feed problem, namely, the issue of increasing the efficiency with which the available feed is utilised (e.g. forages, crop residues, agro-industrial by-products and nonconventional feeds), the inability to make maximum use of the limited total feed resources and the inadequate supplies of feed (Devendra, 1 987) and water. Strategies for ensuring adequate nutrition of small ruminants must be based on optimising overall agricultural and livestock productivity from available resources, improving existing technologies and integrating technology that employs multipurpose crops and animals and recycling of crop residues and by-products as nutrients for both animals and plants. Components of such strategies may include matching livestock production systems to available resources, selection of crops and cropping systems that will maximise biomass production and nitrogen fixation and thus minimise use of inputs external to the system, developing simple processing techniques to optimise the use of different components of crops for different end purpose, recycling of livestock wastes, making more efficient and widespread use of agricultural and industrial by-products as sources of ruminant feed and incorporation into the production system of non-ruminant species that are well adapted to use of farm resources, by-products and wastes. An in-depth analysis of each strategy is worthwhile. Increased offtake De Boer (1982) reported that traditional producers, because of their small-scale operations, market few animals on ad hoc basis, if at all. Small-scale intermediaries handle the transport and assembly functions needed to make up economic numbers of animals for transportation to larger markets. Experience shows that the scope for increasing offtake lies neither in attempts to regulate and control the market participants nor in the control prices, nor creation of parastatals but rather in facilitating the operations of the market participants and instituting measures which reduce marketing costs. These include, improving the infrastructure of livestock marketing, the provision of market information through the mass media (i.e. prices and volume of livestock traded in various markets), regulating the standards of products and services and negotiating favourable trade agreements in export markets (Solomon Bekure and McDonald, 1985). 331 The strategy options Increased use of legumes in cropping and pastoral systems Use of forage legumes in crop mixture Intercropping forage legumes with cereals tends to increase the overall crude protein level of the total fodder available after grain harvest of the primary crop. Usually this involves undersowing a cereal crop with legume. The time of undersowing is critical to optimum production of grain and fodder. Mohamed Saleem (1 985) indicated that planting Stylosanthes guianensis cv Cook or S. hamata cv Verano on the same day with sorghum reduced grain yield by more than 70%, but the loss of grain yield was minimised if they were sown 6 and 3 weeks, respectively, after planting the sorghum. Similarly, reduction in grain yields was observed in a trial in which sorghum was sown with Centrosema pascuorum, Alysicarpus vaginalis or Macroptilium lathyroides on the same day, but the reductions were not as significant as with S. guianensis. Use of forage legumes to supplement natural grazing Fodder banks have been suggested as a strategy to improve nutritional status of livestock in the dry season. Fodder banks are concentrated units of legumes established by pastoralists adjacent to their homesteads to serve as supplements to dry-season grazing. In Nigeria this has been developed into a low-input technology package for pastoralists by ILCA (Mohamed Saleem, 1985). Preliminary results on use of fodder banks appear to be quite promising despite some problems associated with establishment, utilisation and regeneration. Use of forage legumes in crop rotations Association of forage legumes in crop rotation has been demonstrated to improve the nitrogen content of soil thereby benefiting crop production when such plots are cropped. This also offers appropriate opportunity to integrate crop and livestock production. Mohamed Saleem (1985) reported that maize growth and grain yield was much higher when a portion of fodder banks of different ages was cropped as compared to yields immediately outside the fodder banks, which were previously either cropped or under natural sub-climax vegetation. Berfer use of forage of tree legumes Production systems using tree legumes usually consist of either a dense population of trees in a small area, a row of trees planted as "hedgerow" on bunds, as edges of crop areas or interplanted with crops (alley cropping). 332 The leaves of trees can be used as a high quality supplement to crop residues and grass. The level of incorporation may range from 20-75%. High responses in terms of weight gain have been registered with small ruminants (Home et al, 1986). Leaf production seems relatively unaffected by frequency of cutting although wood production is increased with decreasing cutting frequency. Preston and Leng (1987) have indicated that tree legumes such as Gliricidia, Erythrina and Leucaena have a great potential as sources of legume fodder particularly as they are high-yielding perennials and possess deep-rooted systems that may have access to ground water and nutrients that may not be available to smaller leguminous plants. Rajaguru (1987) estimated that a hectare of tree fodder such as Gliricidia sepium and Leucaena leucocephala could yield up to 6-8 tons of dry matter per year which could support 12 goats or sheep. Meanwhile, Home et al (1986) estimated forage yield of G. sepium and L leucocephala at 1 2. 1 tons/hectare/year. Small ruminant production in an alley-cropping perspective has been analysed by Sumberg (1 985). Alley-cropping is a system in which crops are grown in alleys between rows of frequently pruned trees. The large input of nitrogen and organic matter from tree foliage can potentially support continuous cropping at intermediate yield levels (Kang et al, 1981). Small ruminants can be integrated into an alley cropping by cut-and-carry feeding of portions of the tree foliage or by grazing natural fallow regrowth and trees during periodic fallow years. The cut-and-carry feeding applies to goats and sheep while the grazing system is limited to sheep because goats tend to de-bark the trees. Fast growing leguminous trees like G. sepium and L leucocephala are essential for the system of feeding. In order to give sufficient benefit to the crop and avoid the possibility of soil mining, it is recommended that approximately 75% of the available tree foliage be applied to the soil as mulch while the rest is given as feed to animals. An annual tree foliage yield of four tons dry matter/hectare would yield about one ton dry matter which is sufficient to support approximately 14 adult animals per hectare as a year-round supplement or four animals/ha as sole feed. Van Eys (1987) cautioned that structural carbohydrates are usually high in forage supplements. When their levels are high, this tends to interfere with the utilisation of roughage as well as supplement. In addition, there are some antiquality components in forage supplements (tannins, cyanogenic glucosides, nitrates, soluble oxalates) that can reduce payability, voluntary intake and productivity if forage supplements are fed continuously to small ruminants, and particularly if they are fed in the fresh form or grazed by the animals. Some of these compounds can be partially destroyed by drying or storage. However, ruminants tend to be suited to utilising some forage legume supplements that contain antiquality substances since these can be detoxified in the rumen (for example, mimosine DHP in L. leucocephala). 333 Effective exploitation of natural browse for the nutrition of small ruminants In Africa, it has been estimated that over 250 million head of domestic animals live in arid, semi-arid and mountain zones where browse is an important qualitative component of the livestock diet (Le Houerou, 1978). The widespread traditional use of browse as an available source of quality feed during the dry season is vitally important to maintain seasonal and yearly stability of livestock production. Lamprey et al (1 980) have reported a browse productivity rate of 5000 kg dry matter/ha/annum for riverine Acacia xanthophloea which is about double that of most savannah grasslands. However, there is a large variability in productivity within populations of browse species depending on the ecological conditions. Established browse plantations of Medicago arborea have been able to produce 3500 kg dry matter/ha/year under annual rainfall of 350-450 mm (Le Houérou, 1975). Otsyina and McKell (1984) indicate that browse, especially in the savannah vegetation types, contributes a large amount of available forage as a complementary source of feed for livestock, particularly during the dry season when herbaceous forage is in short supply. The major problems appear to be inadequate management systems that fail to make optimum use of shrubs and find ways to increase the number and quality of valuable browse species available for livestock production. Le Houerou (1980) reported that compared with tropical grasses, browse is richer in protein and some minerals in the dry season. At this time, browse plants are the most valuable feed used by livestock. Browse could therefore supplement the low protein content of grass forage if used effectively during dry periods. The crude fibre content of browse also tends to be lower than that of grasses and usually ranges from 20-40% and is even lower for shoots and leaves (Pellew, 1980). Given the low content of crude fibre in browse compared with dry grass, the energy content of browse appears to be higher than that of dry grass (Le Houerou, 1978). Browse plants can satisfy the needs of small ruminants on provided that the animals can gain easy access to them and a sufficient quantity is available to them directly or with the aid of man. From dry to humid tropical Africa the contribution of browse towards intake by an animal implies considerable effort on the part of herdsmen in activities such as planting of browse trees, maintaining the necessary balance of the species present by selective bush-clearing and making browse available to animals by either beating down fruits as in the Sahel, by trimming or cutting back (Boudet and Toutain, 1980). Carew et al (1980) reported studies on the potential of browse plants in the nutrition of small ruminants in the humid forest and derived savannah zones of Nigeria. Savannah goats and sheep browsed and grazed longer and spent less time scavenging than their counterparts in the humid forest. The biting ratios of 334 browse to grass in the forest zone were 77: 1 for the goats and 1 3: 1 for the sheep. This shows that browse plants were selected more often by goats and sheep in the forest zone than the savannah region though browse plant height or accessibility tended to affect selection pattern. Common species studied were Ficus exersperata, Newboulotia leavis, Asphilia africana, Spondias mombin and Cyclicodiscus gabunensis. Recommendations (Otsyina and McKell, 1984) for research on browse plants include the identification of superior browse species with regard to available productivity, nutritive value and persistence under grazing by livestock; the development of methods for establishing desirable browse species in ecologically favourable areas along with grasses in mixed plantings; the improvement of management practices of established and native pastures containing both browse and grass species to maintain effective species composition; and the documentation of characteristics useful for improving shrublands for small ruminants in different seasons. Integration of small ruminants in plantation crop production Tree crops may contribute to livestock feed resources through use of their by-products. If these by-products could be collected properly and processed, they could offer alternatives to costly concentrates. Common by-products obtained from plantation crops include coconut cake, palm-kernel cake, rubberseed cake, citrus wastes (pulp, seed meal) and cocoa pods and residues. Cover crops grown in plantations also offer feed that can be used for small ruminants. Natural herbage growing under plantation trees may be collected and conserved in the form of hay. Where possible, controlled grazing underneath tree canopies may be practised on a rotational basis. Fodder cultivation under plantation crops can also be practised. Wan Mohamed (1 987) has indicated that cultivation of cover crops in plantations in Malaysia offers a lot of potential for production of small ruminants. The age of rubber trees or oil-palm trees tends to influence the botanical composition of the ground cover. Cultivated legume cover crops tend to decline as the trees mature due to their shading effect which reduces photosynthesis. In terms of compatibility, sheep production is more compatible with oil palm cultivation than goats. However, both species are complementing in reducing the need to use herbicides and thus reducing the cost of weeding and increase returns per unit area of land. Parawan and Ovalo (1987) reported that integration of livestock production with coconuts has a high potential in the Philippines. Moog et al (1977) indicated that coconut plantation pasture is available throughout the year and provides of farmers with nutritious feed for livestock. When feeds are scarce, farmers use 335 coconut fronds in addition to bamboo leaves, banana leaves and Gliricidia sepium. The low technology inputs, the versatility in the feeding habits, and the relatively low acquisition cost of goats and sheep can fit appropriately into the majority of smallholder coconut farms. The added value of goat meat and milk, mutton and animal manure as fertiliser produced from the use of crop products, crop by-products and forage under coconut can raise the average income of the coconut smallholder (Parawan and Ovalo, 1987). The constraints of integrating small ruminant production in rubber and oil-palm plantations include damages to young plants especially at establishment time, the potential toxic effect of herbicide sprayed on tree crop planted in rows and rat-baiting in oil-palm plantations which could contaminate by-products of oil-palm processing which could lead to copper toxicity problems in livestock (Pillai and Tan, 1977). In countries where there is large-scale cultivation of citrus, their by-products offer useful feed resources for livestock. When grapefruit (Citrus paradisi) and sweet orange (Citrus sinensis) are industrially processed on a large-scale to produce juice or sections, by-products such are peels, rag, and depending on method of processing, other by-products like citrus molasses and citrus seed meal may be available. In most cases, the harvest of citrus usually coincides with the dry season when grass is scarce. When quantities of peel and rag (the string axis and white fibrous membrane) are large, they are conserved by drying. The dried citrus pulp can easily be mixed with other ingredients and is palatable, rich in nutrients and exerts a mild laxative effect. It can be stored for all-year feeding and deteriorates less in storage than many other feeds (Gohl, 1 978). The pulp is better used as pellets where possibilities for pelleting exist. Banana plantations also offer important feed resources for feeding small ruminants. It has been estimated that, for every ton of banana picked and exported to the fruit market, about 750 kg are rejected as being either unsuitable or in excess of requirements. This discarded fruit represents 10-20% of the crop depending on the degree of culling and the amount of fruit that is rendered unusable by storms (Le Dividich et al, 1978). Most of this can be used as animal feed. Chenost et al (1976) reported that in ad libitum feeding of goats, when forage and bananas were offered separately in a trough, the two feeds were ingested in such proportions that bananas constituted 20-40% of total ingested dry matter. When bananas were blended with forages, the intake of both dry matter and digestible organic matter rose sharply from 0-20% with increasing dietary content of bananas; at the 20% level, dry-matter intake was higher for ensiled than for fresh green bananas with a peak of 1 .8 to 2.2 kg/1 00 kg liveweight. Beyond this level, dry-matter intake remained relatively constant as bananas replaced green forage on a weight-for-weight basis, while digestible organic matter rose slowly. Banana leaves may also be used as forage for animals. 336 More effective use of crop residues and agro-industrial by-products With increases in population pressure and the demand for more food and farmland, the use of crop residues and by-products is increasing. Straw, for example, is a valuable feed resource especially during the dry season. However, Preston and Leng (1987) believe that the feeding of straws, like most roughages, to ruminants is constrained by their slow rate of degradation and therefore low total digestibility, their low propriate fermentation pattern in the rumen and their negligible levels of both fermentable N and by-pass protein which is usually associated with their low intake. The digestibility of straws can, however, be improved by either alkali treatment, ammoniation, treatment with poultry manure or feeding with green forage legumes. Alkali treatment of by-products to improve intake and digestibility Kategile (1983) described a simple method of treating roughages with NaOH that could significantly improve their intake and digestibility in small ruminants. However, the cost and availability of NaOH may pose an important constraint under rural conditions. It is hazardous if not well handled and is a pollutant to environment. Urea supplementation (ammoniation) Naga and El-Shazly (1982) reported that supplementation with urea can improve the digestibility of nutrients contained in supplemented material provided the level of urea is maintained at 0.5 to 2%. There are risks involved when urea nitrogen in the diet exceeds 30%. This may lead to potassium losses in urine resulting in the depletion of the animal's body supplies (Juhaz et al, 1976). Eventually, this can lead to depressed bacterial growth, decreased synthesis of the B-vitamins in the rumen, loss of appetite and cerebrocortical necrosis (Naga et al, 1978). Ammonium hydroxide from urea appears to be safer to use than direct liquid ammonia or gaseous ammonia. When mixed with an equal weight of water, the mixture contains about 5% urea equivalent of dry matter on dry-weight basis. The solution can be kept in water and feed containers covered with plastic sheet to prevent exposure to air for 3-4 weeks. After this period the plastic sheet is opened to release ammonia that has been produced, and the material can be fed to animals. 337 Using animal wastes to improve intake and digestibility of roughages Excreta from poultry are rich in N, mostly as uric acid which is hydrolysed to ammonia by rumen micro-organism. Sha and Muller (1982) reported that poultry wastes used at the 25-40% level can meet most of the protein and mineral requirements in ruminant livestock. However, the use of poultry wastes in ruminant feeds tend to result in reduced microbial count, low dietary energy content, exposure to pathogens and high intake of antibiotics and chemotherapeutics that may be present in the waste. Incorporation of high energy feed ingredients such as molasses, root crops and grains can balance energy requirements and maximise utilisation of the non-protein nitrogen (NPN) fraction of the wastes; mixing molasses with ration containing poultry wastes may improve payability and intake and prevent eye and respiratory problems associated with the dusty nature of poultry waste. Pathogens could be eliminated by ensiling, stacking, formaldehyde in treatment or dehydration. Feeding of wastes that contain high levels of antibiotics and chemotherapeutics should be avoided. Increased exploitation of lignocellulose as animal feed Large quantities of lignocellulose from sugar-cane, timber and cereal processing usually go to waste. Feeding materials that are high in lignocellulose tend to have a low digestibility, low nitrogen content and low animal production potential. Simple inexpensive treatments which increase lignocellulose breakdown in the rumen and supplements of non-protein nitrogen could significantly increase animal production. However, when subjected to 3-6% sodium hydroxide solution and/or steam treatment, irradiation, grinding and ammoniation, lignocellulose breakdown in the rumen is increased, and with the addition of non-protein nitrogen its utilisation increases and thus improves animal production. Wong et al (1 974) reported that treating bagasse with high pressure (1 4 kg/cm2) steam for five minutes raised dry-matter digestibility from 28 to 60%. Conservation of forage as silage for feeding small ruminants during the dry season Conservation of forage as silage under proper conditions can result in the preservation of 90% harvested energy and protein and the concentration of energy within the silage (McCullough, 1978). In tropical Africa the use of silage is not common. It, however, offers a good opportunity to preserve abundant forage produced during the wet season for use in the dry season. Under rural conditions, silage can be made in pit silos which are relatively easy to construct and are also affordable by farmers. Recently, there have been improvements of silage-making using discarded fertiliser plastic bags. 338 The chopped grass is compacted in the bag and sealed. All the bags are then packed and compacted in the silo after which the silo is sealed. This system enables farmers to ration the daily require silage by simply removing one or more bags as required. Empty fertiliser bags cost about 0.3 US dollars in local markets. Silage-making is most convenient in high population areas where pastures are scarce and grasses like Tripsacum laxum and Pennisetum purpureum are readily available. Where sugar-cane tops are abundant these could also be ensiled. Under rural conditions, labour may be a serious constraint if large amounts of silage are to be made without mechanisation. Increased conservation of feed as cut or standing hay and supplements for use in the dry season Where animals are prevented from migration during the dry season, feeding strategies must be evolved that will prevent the animals from losing weight. Quite often some of the pasture is ungrazed during the wet season and retained for use during the dry season. This is called standing hay. This would require suitable supplements in order to maximise its efficiency of utilisation. Diets based on molasses/urea mixture may be employed to supplement dried standing hay (Gulbrassen, 1985). Whole cotton seed is occasionally used (Ngo Tama et al, 1987) in sheep rations in the semi-arid regions in North Cameroon. Similarly, cotton seed and groundnut cakes may be fed with straws, cut fodder or hay (Njwe, 1978; Fall et al, 1989). The use of non-protein nitrogen (NPN) with standing hay has also proved most effective in dry-season feeding. Urea is most commonly used. The primary concern in feeding urea is providing sufficient amounts to be of value to the animal but minimising the risk of mortality from toxicity. NPN may be fed to animals by spraying on pasture in the field. The spray usually consists of 39% molasses, 6.25% urea and 54.75% water. This method is wasteful. Alternatively, the urea/molasses mixture may be fed separately as liquid in troughs placed in pastures, but the risk of poisoning, as each animal has free choice, is high. Block licks (40% urea, 10% molasses, 2.5% trisodium phosphate and 47.5% salt) have been used (Alexander etal, 1970). Molasses blocks of relatively low urea content can be used during droughts as high energy feed (Preston and Leng, 1987). The hardness of the blocks restricts and slows rate of intake. Alexander (1978) has recommended incorporation of biuret in dry licks that contain salt. First, a 1 :1 biuret to grain mixture is made. This is then mixed with salt in a ratio varying from 1:1 to 1:3. This mixture has been used to arrest decreases in body weight of sheep grazing dry poor quality pasture in Australia. 339 Use of fertiliser to improve pasture yields On soils with low fertility, fertiliser application can increase dry-matter production per unit area. Superphosphate and nitrogen fertilisers like urea and ammonium sulphate are most widely used. Usually, fertiliser application is adopted in intensively managed systems in which high value products can justify the economic cost. The smallholder in most cases cannot afford the cost of using fertiliser, but farmyard manure from goat and sheep pens may be used in promoting higher pasture yields for a cut-and-carry intensive system of small ruminant production. Common fodder grasses that may be exploited in such a system include Tripsacum laxum, Pennisetum purpureum and Panicum maximum. Conclusion Unimproved pasture will continue to play an important role in small ruminant production in Africa. Overgrazing must be avoided to prevent degradation of these pastures. However, since these pastures become markedly deficient in quality, they are better utilised in the dry season with supplements such as urea-molasses, tree-forage legumes, leguminous browse, conserved legume hays and crop residues like groundnut and cotton haulms. Dehydrated poultry excreta from caged layers can also be exploited in supplementing poor quality forages during the dry season. It seems to be an economical protein supplement for ruminants especially as the cost of dehydration is low when compared to the cost of conventional feeds. A destruction of possible pathogens in poultry wastes needs to be done before it is used. Quite often there may be competition for the use of some crop residues and agro-industrial products between livestock producers and industry. The production of paper from rice straw and alcohol from molasses, and the use of bagasse as fuel are common examples. This could easily render these feeds unavailable for livestock. Alternatives have to be identified and exploited to sustain livestock during a period of scarcity. Conservation of herbage and surplus by-products by ensiling, drying or other means ensures a continuous supply of fresh and conserved feed over prolonged periods. Preservation of by-products of citrus industry either by ensiling or drying can make important contributions to the amount of locally available feed resources for small ruminants. The cultivation of legumes in cropping systems may be hindered by the limited quantities and unaffordable prices of seeds. Commercial seed production which is a specialised activity, needs every encouragement by African governments. Several indigenous and introduced legumes and grasses need investigation in 340 order to recommend suitable species for cropping systems and pasture improvement. Grazing sheep can help control weeds on rubber estates and provide income to smallholders (in smallholder schemes) during pretapping years. Optimum carrying capacities for different aged plantations need to be determined. Sheep seem to be more compatible than goats when reared on a semi-intensive basis in rubber and oil-palm plantations. Both species are suitable for coconut plantations. Silvipasture offer potential feed resources for small ruminant production in densely populated areas with forest reserves. Goats tend to adapt better to silvipasture than sheep. The feeding strategy to adopt will depend on factors such as climate, ecological conditions, topography, farming systems and prevailing socio-economic conditions. References Alexander G I. 1978. Non-protein nitrogen supplements for grazing animals in Australia. World Animal Review, FAO Animal Production and Health Paper, FAO (Food and Agriculture Organization of the United Nations), Rome, Italy. Alexander G I, Daly J J and Burns M A. 1 970. Nitrogen and energy supplements for grazing beef cattle. In: Proceedings of the 11th International Grassland Congress, University of Queensland Press, St. Lucia, Queensland, Australia. pp. 793-796. Boudet G C and Toutain B. 1980. The integration of browse plants within pastoral and agropastoral systems in Africa. In: Le Houerou H N (ed), Browse in Africa: The current state of knowledge. Papers presented at the International Symposium on Browse in Africa, Addis Ababa, April 8-12, 1980. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 427-432. Carew B A, Mosi A K, Mba A U and Egbunike G N. 1980. The potential of browse plants in the nutrition of small ruminants in the humid forest and derived savanna zones of Nigeria. In: Le Houerou H N (ed), Browse in Africa: The current state of knowledge. Papers presented at the International Symposium on Browse in Africa, Addis Ababa, April 8-12, 1980. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 307-312. Chenost M. Canbau M, Geoffrey Fand Bousquet P. 1976. Possibilities of using bananas for the feeding of ruminants in humid tropical regions. Journal of the Agricultural University of Puerto Rico 43. De Boer A J. 1982. Overview: Livestock development: The Asian experience. In: Fine J C and Lattimore R G (eds), Livestock in Asia: Issues and policies. IDRC-202e. IDRC (International Development Research Centre), Ottawa, Canada. pp. 13-26. Devendra C. 1987. Small ruminant production systems. In: Devendra C (ed), Small ruminant production systems in South and SoutheastAsia. Proceedings of a workshop held in Bogor, Indonesia, 6-10 October 1986. IDRC-256e. IDRC (International Development Research Centre), Ottawa, Canada. pp. 29-51. 341 van Eys J E, Rangkuti M and Johnson W L. 1987. Feed resources and feeding systems for small ruminants in South and Southeast Asia. In: Devendra C (ed), Small ruminant production systems in South and Southeast Asia. Proceedings of a workshop held in Bogor, Indonesia, 6-10 October 1986. IDRC-256e. IDRC (International Development Research Centre), Ottawa, Canada. pp. 52-77. Fall S, Guerin H, Sail C and Mbaye ND. 1989. Cereal straws in the feeding system of ruminants in Senegal. In: Said A N and Dzowela B H (eds), Overcoming constraints to the efficient utilization of agricultural by-products as animal feed. Proceedings of the Fourth Annual Workshop held at the Institute of Animal Research, Mankon Station, Bamenda, Cameroon, 20-27 October 1987. ARNAB (African Research Network for Agricultural By-products), ILCA, Addis Ababa, Ethiopia. pp. 173-189. FAO (Food and Agriculture Organization of the United Nations). 1981. FAO production yearbook 1980. Vol. 34. FAO, Rome, Italy. Gatenby R M. 1982. Research on small ruminants in sub-Saharan Africa. In: Gatenby R M and Trail J C M (eds), Small ruminant breed productivity in Africa. Proceedings of a seminar held at ILCA, Addis Ababa, Ethiopia, October 1982. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 13-20. Gohl Bl. 1978. Citrus by-products for animal feed. In: Ruminant nutrition: Selected articles from the World Animal Review. FAO Animal Production and Health Paper 12. FAO (Food and Agriculture Organization of the United Nations), Rome, Italy, pp. 41-44. Gulbrassen B. 1985. Survival feeding of cattle with molasses. 1. Feeding non-pregnant heifers with molasses plus urea and roughage. Australian Journal of Experimental Agriculture 25: 1-3. Home P M, Catchpoole D W and Ella A. 1986. Cutting management of tree and shrub legumes. In: Blair G J, Ivory D A and Evans T R (eds), Forages in Southeast Asian and South Pacific agriculture. Proceedings of an international workshop held at Cisarua, Indonesia, 19-23 August 1985. ACIAR Proceedings Series 12. ACIAR (Australian Centre for International Agricultural Research), Canberra, Australia. pp. 164-169. Juhaz B, Szegedi B and Kresotes M. 1976. In: Tracer studies on non-protein nitrogen for ruminants. II. Joint FAO/IAEA Division of Atomic Energy in Food and Agriculture, Research Co-ordination Meeting Panel on Trace Techniques in Studies on the Use of Non-protein Nitrogen (NPN) in Ruminants, Vienna. Kang B T, Wilson G F and Sipkens L 1981. Alley cropping maize (Zea mays L) and Leucaena (Leucaena leucocephala Lam.) in southern Nigeria. Plant and Soil 63:165-179. Kategile J A. 1983. Utilization of low quality roughages with or without NaOH treatment. In: Berhane Kiflewahid, Potts G R and Drysdale R M (eds), By-product utilization for animal production. Proceedings of a Workshop on Applied Research held in Nairobi, Kenya, 26-30 September 1982. IDRC-206e. IDRC (International Development Research Centre), Ottawa, Ontario, Canada. pp. 37-48. Lamprey H F, Herlocker D J and Field C R. 1980. Report on the state of knowledge on browse in East Africa in 1980. In: Le Houerou H N (ed), Browse in Africa: The current state of knowledge. Papers presented at the International Symposium on Browse in Africa, Addis Ababa, April 8-12, 1980. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 33-54. Le Dividich J, Geoffroy F, Canope I and Chenost M. 1978. Using waste bananas as animal feed. In: Ruminant nutrition: Selected articles from the World Animal Review. FAO Animal Production and Health Paper 12. FAO (Food and Agriculture Organization of the United Nations), Rome, Italy, pp. 50-58. 342 Le Houerou H N. 1975. The rangeland of North Africa: Typology, yield, productivity and development. In: Evaluation and mapping of tropical African rangelands. Proceedings of the seminar held in Bamako, Mali, 3-8 March 1975. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 41-55. Le Houerou H N. 1978. The role of shrubs and trees in the management of natural grazing lands (with particular reference to protein production). Paper presented at the Eighth World Forestry Congress held in Jakarta, Indonesia, 16-28 October 1978. FFF/10-0. FAO (Food and Agriculture Organization of the United Nations), Rome, Italy. 24 pp. Le Houerou H N. 1980. Browse in North Africa. In: Le Houerou H N (ed), Browse in Africa: The current state of knowledge. Papers presented at the International Symposium on Browse in Africa, Addis Ababa, April 8-12, 1980. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 55-82. McCullough M E. 1978. New trends in ensiling forages. In: Ruminant nutrition: Selected articles from the World Animal Review. FAO Animal Production and Health Paper 12. FAO (Food and Agriculture Organization of the United Nations), Rome, Italy, pp. 24-29. Mohamed Saleem MA. 1985. Forage legumes in agropastoral production systems within the subhumid zone of Nigeria. In: Kategile J A (ed), Pasture improvement research in eastern and southern Africa. Proceedings of a workshop held in Harare, Zimbabwe, 17-21 September 1984. IDRC-237e. IDRC (International Development Research Centre), Ottawa, Canada. pp. 222-243. Moog GA Sena E M and Valenzuela F G. 1 977. Crop and livestock in two Batangas barrios. Philippine Journal of Animal Husbandry 32:1 1-24, 204. Naga M A and El-Shazly K. 1982. Use of by-products in animal-feeding systems in the Delta of Egypt. In: Berhane Kiflewahid, Potts G R and Drysdale R M (eds), By-product utilization for animal production. Proceedings of a Workshop on Applied Research held in Nairobi, Kenya, 26-30 September 1982. IDRC-206e. IDRC (International Development Research Centre), Ottawa, Ontario, Canada. pp. 9-15. Naga M A, Nour A M, Borhami B E A, El-Din M Z, Baza M A, El-Shazly K, Abou Akkada A Rand Oltjen R R. 1978. Effect of potassium on the rumen micro-organisms of animals fed on diets containing urea. Tropical Animal Production 3(1):62-68. Ngo Tama A N, Mbah D A and Rippsten G. 1987. Complementation de moutons Peuhl en Croissance avec du tourteau de coton pendant la saison seche dans le Nord du Cameroun. Petes du seminaire Regional sur les fourrages et I'alimentation des ruminants, 16-20 Novembre 1987, Ngaoundere, Cameroun. Njwe R M. 1978. The effect of protein supplementation on the utilization of cured and fresh Giant Star grass (Cynodon nlemfuensis var robustus) by young West African dwarf sheep. MSc thesis, Department of Animal Science, University of Ibadan, Nigeria. 304 PP Otsyina R M and McKell C M. 1984. Africa: Browse in the nutrition of livestock: A review. World Animal Review 53:33-39. Parawan O O and Ovalo H B. 1987. Integration of small ruminants with coconuts in the Philippines. In: Devendra C (ed), Small ruminant production systems in South and Southeast Asia. Proceedings of a workshop held in Bogor, Indonesia, 6-10 October 1986. IDRC-256e. IDRC (International Development Research Centre), Ottawa, Canada. pp. 269-279. Peacock C P. 1985. Measures for assessing the productivity of sheep and goats in Africa. In: Wilson RTand Bourzat D (eds), Small ruminants in African agriculture. Proceedings 343 of a conference held at ILCA, Addis Ababa, Ethiopia, 30 September-4 October 1985. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 128-141. Pellew R A. 1980. The production and consumption of Acacia browse and its potential for animal protein production. In: Le Houerou H N [ed), Browse in Africa: The current state of knowledge. Papers presented at the International Symposium on Browse in Africa, Addis Ababa, April 8-12, 1980. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 223-231. Pigden W J and Bender P. 1978. Utilization of lignocellulose by ruminants. In: Ruminant nutrition: Selected articles from the World Animal Review. FAO Animal Production and Health Paper 12. FAO (Food and Agriculture Organization of the United Nations), Rome, Italy, pp. 30-33. Pillai K R and Tan H T. 1977. Possible livestock production on estates, using factory effluents. Internal report. Guthrie Research, Chemara, Negeri Sembilan, Malaysia. Preston T R and Leng R A. 1 987. Matching ruminant production systems with available resources in the tropics and sub-tropics. Penambul Books, Armidale, Australia. 245 PP- Rajaguru A S B. 1987. Integration of crops and small ruminants in Sri Lanka. In: Devendra C (ed), Small ruminant production systems in South and Southeast Asia. Proceedings of a workshop held in Bogor, Indonesia, 6-10 October 1986. IDRC-256e. IDRC (International Development Research Centre), Ottawa, Canada. pp. 190-202. Sha S I and Muller Z 0. 1982. The economic impact of feeding poultry litter to lactating cows and buffaloes. Pakistan Journal of Agricultural Research 2(2):71-75. Solomon Bekure and McDonald I. 1985. Some policy issues of livestock marketing in Africa. ALPAN Network Paper 2. ILCA (International Livestock Centre for Africa). Addis Ababa, Ethiopia. 10 pp. Sumberg J E. 1985. Small ruminant feed production in a farming systems context. In: Sumberg J E and Cassaday K (eds), Sheep and goats in humid West Africa. Proceedings of the Workshop on Small Ruminant Production Systems in the Humid Zone of West Africa, held in Ibadan, Nigeria, 23-26 January 1984. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 41-46. Wan Mohamed W E. 1987. Integration of small ruminants with rubber and oil palm cultivation in Malaysia. In: Devendra C (ed), Small ruminant production systems in South and Southeast Asia. Proceedings of a workshop held in Bogor, Indonesia, 6-1 0 October 1986. IDRC-256e. IDRC (International Development Research Centre), Ottawa, Canada. pp. 239-256. Wong Yon Cheong Y, D'Espaignet J T, Deville J P, Sansoucy R and Preston T. 1974. The effect of steam treatment on cane bagasse in relation to its digestibility and furfural production. Proceedings of the 15th Congress of the ISSCT. Hayne and Gibson, Durban, South Africa. 344 Some experiences in adapting previously free-ranging traditionally managed Matabele goats of Zimbabwe to individual stall-feeding L.M. Sibanda\ L.R. Ndlovif and M.J. Bryant2 1 Department of Animal Science University of Zimbabwe P O Box MP 167, Mt. Pleasant, Harare, Zimbabwe 2 Department of Animal Production University of Reading Early Gate, Reading, RG6 2AT, United Kingdom Summary One hundred multiparous indigenous does were purchased from communal farmers for on-station trials. At the research site, the goats were confined and zero-grazed. During adaptation, six goats died, 28 aborted, 56 kidded and 10 had no reproduction records. The does lost weight and were afflicted by diseases not previously experienced under the traditional management system. After nine months of adaptation to the station regime, the goats were synchronised for oestrous and mated. Only 74% of the flock conceived and kidded while 19% did not conceive and 7% aborted. Factors considered to be associated with the poor adaptability to stall-feeding ofpreviously free-ranging goats are discussed. Essai d'adaptation a I'alimentation individuelle a I'etable de caprins Matabele du Zimbabwe precedemment eleves en plein air suivant le systeme traditionnel L.M. Sibanda, L.R. Ndlovu et M.J. Bryant Resume 100 chevres multiparas ont ete achetees aupres d'eleveurs communaux pour cette etude. Une fois arrivees sur la station 345 experimental, elles ont et& confinees et alimentees sans pouvoir avoir acces au paturage. Au cours des cinq mois qu'a durée la periode d'adaptation, elles ont perdu enormement de poids; six d'entre elles sont mortes, 28 ont avorte, 56 ont mis-bas tandis qu'aucun ivenement important sur le plan de la reproduction n'a ete enregistre pour les 10 restantes. Au sixieme mois, l'oestrus a eté synchronise chez les 94 chevres encore vivantes, qui ont ensuite ete mises a la lutte (1 bouc pour 10 chèvres). Seules 56% d'entre elles ont conqu et mis-bas. Les chevres precedemment elevees en plein air — c'est connu — supportent mal la claustration et l'alimentation a l'étable, mais rien de concret n'a encore ete fait pour remedier a ce probleme. II est recommandé que les chercheurs fassent connaitre leur experience en la matiere pour permettre de mieux cerner le phSnomene et d'y remedier. Introduction Extensive goat production under the traditional communal grazing system is widely practised in the semi-arid and arid regions of Africa which sustain 66.7% of the African goat population (FAO, 1984). This system, characterised by shortage of land, poor range management and seasonal fluctuation of feed supply favours mere survival rather than optimum production. Productivity of the goats in these areas can be improved by making more efficient use of the feed resources presently or potentially available to the farmer. To do this will require information on the nutrient quality of the feed as a function of season, the suitable physical form of the feed, the refusal rate and the physiological state and energy balance of the goats to be fed. Studies to generate such information often necessitate stall-feeding of the animals with no access to grazing. Unfortunately, when previously free-ranging goats are confined and stall-fed, they are stressed and may either eat less, lose weight, about (Wentzel, 1987) or even die (Kasowanjete et al, 1986). Many feeding trials are usually prematurely abandoned when goats fail to adapt to stall-feeding and unfortunately most of these futile efforts but nonetheless useful experiences are never reported. This paper discusses some of the problems experienced during the adaptation to stall-feeding of 100 previously free-ranging indigenous Matabele does. Materials and Methods Location The reported work was carried out at the former Thuli Breeding Station which is situated in Matabeleland Province in the south-western part of Zimbabwe. The 346 region is semi-arid. The vegetation is mainly acacia shrub and mopane woodland with several species of annual grasses appearing during the rainy months of November to March. Droughts are frequent, consequently herbage quality and quantity fluctuate within and between years. The total amount of rainfall during the study year (1989) was 162 mm. The mean maximum and minimum temperatures were 28°C and 13°C, respectively. Animals In May 1989, one hundred muciparous indigenous Matabele goats were bought from communal farmers living within a 20-kilometre radius of the research station. The animals were driven on hoof to the research site. They were ear-tagged, weighed and their age was determined by counting the number of permanent incisors. Management of the does during adaptation On arrival the does were dewormed, dipped and vaccinated against pulpy kidney disease. Thereafter deworming and dipping were repeated on a regular basis. The animals were housed in groups in a large animal house with concrete floors and partial roofing that allowed adequate ventilation. A forage mixture (two-thirds veld hay and one-third lucerne hay) with a mean crude protein content of 12% and neutral detergent fibre content of 60% was offered ad libitum in raised feeding troughs. Water and a salt mineral blocks were freely available to the animals. As from September 1989 (fourth month of adaptation), all the does were individually offered 100 g maize grain daily so as to increase energy intake in preparation for breeding on-station. At the end of the fifth month of adaptation (October 1989), the goats were synchronised for oestrous using progesterone-impregnated intravaginal sheep sponges (Veramix 60 mg medroxyprogesterone, Upjohn, UK). The sponges were withdrawn after 14 days and 2 days thereafter, 10 large Matabele bucks, with a mean liveweight of 65 kg, brought in from Matopos Research Station, were introduced to the flock. The bucks continuously ran with the does during the day but were confined separately overnight so that mating records could be kept. In order to determine pregnancy, blood was collected through the jugular vein 21 days after mating and twice weekly thereafter. The blood was kept at 2CC overnight and allowed to clot. After 24 hours, serum was decanted and stored at -20°C until assayed for progesterone levels using a radio-immunoassay technique developed for goat serum. All goats were weighed weekly and kept under close observation during the adaptation period. All incidences of diseases and reproduction were recorded. 347 Results and Discussion The adaptation period was intended to take three weeks but had to be extended to six months due to problems associated with feeding, disease incidences and reproduction. Feeding In the first four weeks of adaptation, feed refusals were approximately 60% of the feed offered as goats selected mainly lucerne hay which constituted only 24% of the total dry matter offered. When the forages were ground in an effort to reduce selection, feed refusal rates doubled because the feed was dusty. Low dry-matter intake (below 40% of the expected intake) during the first few months of adaptation is not uncommon in stall-feeding trials involving goats fed forage-based diets (Ademosun et al, 1985). The low feed intake has largely been attributed to the ability of goats to select and ingest the higher quality of diets on offer (Owen and Aboud, 1988). Owen and Aboud (1988) have shown that intake will improve when goats are permitted to reject at least 50% of the feed offered rather than the conventional 1 0 to 20%. The reduced feed intake experienced in the first few weeks of adaptation may be the root cause of the observed liveweight losses, diseases and abortions. Diseases The most prevalent diseases during adaptation were orf (contagious ecthyma), gaseous lymphadenitis, kerato-conjunctivitis (pink eye) and swollen knee joints. In cases of orf, daily swabbing of the lesions with salt solution and gentian violet led to rapid recovery. In the second month of adaptation (July 1989) there was an outbreak of pink eye disease which spread rapidly amongst the whole herd; treatment with oxytetracycline led to quick recovery. Swollen knee joints were observed in 20 does and all 100 breeding bucks in the second month of adaptation. These swellings were diagnosed as none pathogenic (no antibodies to caprine Arthritis encephalitis virus) and were attributed to lack of exercise due to confinement. The affected goats spontaneously recovered within four weeks. Diseases encountered during adaptation may have led to the liveweight losses, abortions and acyclicity. In our study, incidences of orf and gaseous lymphadenitis were a lot higher than in normally experienced underthe traditional system. Six does died during the first month on station. Autopsies carried out were test negative for bacteriological and virological pathogens. 348 Doe liveweight changes There were (P) differences in mean doe liveweights at buying between the different age groups (Table 1). Mature 8-tooth does were 18% and 35% heavier than the 6- and 4-tooth age groups, respectively. In the first 6 weeks on station all does lost an average 56 g/day. After six months adaptation (November 1989) when the does were mated, doe liveweight differences were smaller although the younger 4-tooth does were still significantly (P) lighter than the 6- and 8-tooth groups which were about equal in weight. As from November (sixth month of adaptation) all does gained weight. The highest gains were in the 4-tooth age group (90 g/day) compared to 53 g/d and 80 g/day for the 6- and 8-tooth groups. These weight gains could be attributed to replenishment of lost body tissue and genuine growth in the immature 4- and 6- teeth groups. Table 1 . Doe liveweights during adaptation to on-station conditions. Age Number of teeth Date Variable record 4 6 8 Total/Mean May 1989 No. of goats 18 46 30 94 May 1989 Mean weight at buying (SD) 29.4a (3.97) 33.7b (4.83) 39.8b (5.22) 34.8 Nov. 1989 Mean weight at mating (SD) (2.97) (5.64) (4.19) All row values with different superscripts are significantly (P<0.05) different. (SD) = standard deviation. Reproduction Not withstanding the concerted efforts to avoid buying pregnant females, about 90% of the goats purchased turned out to be pregnant (Figure 1), mostly in the second month of gestation. None of these goats had shown overt pregnancy symptoms when purchased. During the first three months of adaptation, 30% of the does aborted. The highest number of aborters was in the young 4-tooth age group where almost 1.5 and 8 times more does aborted than in the 6- and 8-tooth age groups, respectively (Figure 1). Specimens from the aborted foetuses and placentae were tested for leptospirosis, contagious abortion, Rift Valley fever and blue tongue. All test results were negative and it was postulated that stress factors (such as a sudden change of environment, rather than disease could have caused the abortions. Based on information available in other ruminants and observations made in goat flocks, Morand-Fehr (1987) has concluded that pregnant goats are very 349 sensitive to various kinds of stress such as rapid variations of environmental conditions, excessive animal density in the goat house, struggles between animals, psychic shock etc. Figure 1 . Reproductive response to breeding on station (April-July 1 990). % by age group 100 Sag D Ess #% raj** a ■ ■-.->'.■ WSSSs Pllll IBSj. si■_vi m as i$W41 5!V- -V^^fa^^^ra Age (no. of teeth) Aborters Kidders No event Hypersensitivity to stress of pregnant goats has been explained by the theory that, unlike other domestic ruminants where both the corpus luteum and the placenta secrete progesterone, pregnancy in goats is sustained by progesterone secretion from the corpus luteum only (Sheldrick et al, 1980). Intravaginal sponges that were used for oestrus synchronisation could not be withdrawn from 10 does. The sponges were eventually pulled out after surgical manipulation of the vagina. Five of the does resumed normal oestrus after three months but the others remained acyclic. 350 Mating was expected to commence two days after sponge withdrawal but was delayed by 20 days. Reasons for this delay cannot be readily explained. It is possible that the sheep intravaginal sponges used were unsuitable for goats, though other workers have reported successes with progesterone impregnated sheep sponges used on goats (Cameron et al, 1988). Environmental effects such as ambient temperatures could have contributed to the delay of oestrus onset or lack of libido after sponge withdrawal. It has often been stated that goats in the tropics will kid all year round (Wilson et al, 1984; Hale, 1986). However, there are indications of definite peak kidding periods, usually two per year, availability as influenced by rainfall patterns (Hale, 1986). It is possible that the rains also have a cooling effect on cyclic goats which enhances libido. In sheep, it has been shown that a high proportion of mating occurs in the early morning and activity subsides to a low point at midday and increases again in the evening (Blockey and Cummings, 1970). In this study the mean daily temperature was above average (34°C) wnen the goats were exposed to the bucks and mating only commenced a day after the first rains of the season. Mating (November 1989) lasted about four weeks and only 50% (48 goats) of the flock conceived as confirmed by consistently high blood progesterone levels (1 ng/ml). The rest of the flock which did not conceive in November continued to run with the bucks. Twenty eight of the remaining 46 goats were mated in February after which there were no further matings (Figure 2). Combining the reproductive outcome for matings in November and February (after 6-9 months adaptation to on-station regime), 74% of the original flock purchased managed to conceive and kid, 7% aborted while 19% remained acyclic recording basal blood progesterone levels (<0.3 ng/ml). The number of goats that aborted was far less after nine months (7%) on-station (Figure 2), as compared to 28% (Figure 1) in the second month on-station. Mature 8-teeth does had the best performance with the highest number of goats kidding (96%), the least number of does remaining empty (4%) and no aborters (Figure 2). The abortion rate was higher in the younger 4-tooth does (11%) than in the 6-tooth does (8%) which were 7.6 and 4.2 kg, respectively, does lighter than the mature 8-tooth does at mating. The presence of habitual aborters (does with a long standing poor reproductive history) in the flock cannot be ruled out. It is quite likely some farmers took advantage of the sale to rid themselves of their non productive stock. This could also explain barrenness of 1 0 goats at buying and the failure of the same goats to conceive at mating in November and in February. The fact that more does conceived when mated in February (increasing the total kidding rate in response to breeding on-station from 46% to 74%) means that given time to adjust a larger proportion of the goats will conceive and kid. Unfortunately this prolonged adaptation greatly inflates the cost of goat research projects as compared to cattle and sheep which usually adapt to stall-feeding in 3 weeks. Unlike cattle and sheep, goats seem to have a lower stress threshold 351 demonstrated by being very sensitive and reacting adversely in an exaggerated manner to stress caused by change of routine. Figure 2. Reproductive response during adaptation (May-July 1989). % of total for age group 90 Age (no. of teeth) Aborters Kidders RSSgQ No event Poor adaptability to stall-feeding of goats previously managed under the traditional system presents serious problems to researchers wishing to carry out feeding trials and to entrepreneurs wishing to rear goats under an intensive feeding system. Kasowanjete et al (1 986) lost 428 breeding does within the first 3 years of setting up a goat breeding station in Malawi and attributed the deaths to general adaptation problems. Peters (1989) has reported on-station performance below levels measured on-farm and attributed this poor on-station performance of goats to differences in management systems, reduced feed selection possibilities and higher disease risks. 352 Conclusion In our experience, adaptation to on-station regime by indigenous Matabele goats previously managed under the traditional extensive system took more than 6 months. In the first few months, the goats ate very little forage and consequently lost weight. Diseases not previously experienced in the traditional system became a problem on station. Reproductive performance was poor, especially among the younger groups of females, suggesting that adult females have a better chance of adapting to complete confinement. It seems that numerous environmental factors acting either individually or collectively will impose some degree of stress but our knowledge of the ability of these previously free-ranging goats to adapt or adjust to stall-feeding needs to be improved by further long-term research. Acknowledgements This work was jointly funded by the University of Zimbabwe Research Board and the International Development Research Centre (IDRC). References Ademosun A A, Jansen H J and van Houtert V. 1985. Goat management research at the University of Ife. In: Sumberg J E and Cassaday K (eds), Sheep and goats in humid West Africa. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia, pp. 34-37. Blockey M A and Cummings I A. 1970. Mating behaviour of Merino ewes. Proceedings of the Australian Society of Animal Production 8:344-352. Cameron AWN, Battye K M and Trousen A O. 1988. Time of ovulation in goats (Capra tireus) induced to superovulate with PMSG. Journal of Reproduction and Fertility 83:747-752. FAO (Food and Agriculture Organization of the United Nations). 1984. FAO production yearbook. Volume 38. FAO, Ftome, Italy. Hale D H. 1986. Systems of production and productivity of goats in three communal areas of Zimbabwe. In: Adeniji K O and Kategile J A (eds), Proceedings of the Workshop on the Improvement of Small Ruminants in East and Southern Africa, Nairobi, Kenya, 18-22 August 1986. OAU/IDRC (Organization of African Unity/International Development Research Centre), Nairobi, Kenya, pp 181-193. Kasowanjete M, Stotz D and Zerfas H P. 1986. The goatprogramme. Small Ruminant and Camel group Newsletter 8. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. Morand-Fehr P. 1987. Management programs for the prevention of kid losses. In: Santana O P, Aliomar Gabriel da Silva and Foote WC (eds), Proceedings of the fV International Conference on Goats. Volume 1. International conference held in Brasilia, Brazil, 8-13 March 1987. Departamento de Difusao de Technologia - DDT, Brasilia, Brazil, pp. 405-423. 353 Owen E and Aboud A A O. 1988. Practical problems of feeding crop residues. In: Reed J D, Capper B S and Neate P J H (eds) , Plant breeding and the nutritive value of crop residues. Proceedings of a workshop held at ILCA, Addis Ababa, Ethiopia, 7-10 December 1987. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 133-156. Peters K J. 1989. Trends in on-farm performance testing of small ruminants in sub-Saharan Africa. In: Wilson R T and Azeb Melaku (eds), African small ruminant research and development. Proceedings of a conference held at Bamenda, Cameroon, 18-25 January 1989. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 439-469. Sheldrick E L, Ricetts A P and Flint A P F. 1980. Placental production of progesterone in ovarectomized goats treated with a synthetic progesterone to maintain pregnancy. Journal of Reproduction and Fertility 60:339-348. Wentzel D. 1987. Effects of nutrition on reproduction in the Angora goat. In: Santana O P, Aliomar Gabriel da Silva and Foote W C (eds), Proceedings of the N International Conference on Goats. Volume 1 . International conference held in Brasilia, Brazil, 8-13 March 1987. Departamento de Difusao de Tecnologia - DDT, Brasilia, Brazilns pp. 571-575. Wilson R T, Peacock C P and Sayers A R. 1984. Aspects of reproduction in goats and sheep in south-central Kenya. Animal Production 38:463-467. 354 Fattening mature indigenous (Matabeie) goats: Effects on animal performance, body and carcass composition P.R. Hatendi\ T. Smith\ L. Ndlovu2 and C. Mutisi2 'Matopos Research Station Private Bag K51 37 Bulawayo, Zimbabwe 2University of Zimbabwe POBoxMP167 Mount Pleasant, Zimbabwe Abstract Complete cereal-based diets, containing either 50, 33, 22, or 10 veld hay (9. 7, 10.3, 11.6, and 12. 1 MJ MEIkg DM, respectively) were fed to 40 wethers, to achieve a target body mass gain of 10 kg by 230 days. Ten animals were slaughtered at the start of the experiment to determine initial carcass mass and the body and carcass composition. Foranimals which attained the target, slaughtermass growth rate and feed intake were similar between diets. Animals fed the 22 per cent hay diet out-performed the others in terms of mean body mass gain feed conversion ratio. Means of performance traits had high coefficients of variation and it is suggested that the heterogeneity of the sample population may have masked treatment effects. Compared to the preliminary slaughter group, carcass mass andyield was significantly (P < 0.05) increased by feeding as was fat deposition in both visceral and carcass depots. The quantity of fat deposited appeared to be positively related to dietary energy concentration. 355 Embouche de caprins adultes de race locale matabele : effets sur les performances et la composition de la carcasse P.R. Hatendi \ T. Smith1, L. Ndlovu2 and C. Mutisi2 Résumé Des rations équilibrées à base de céréales contenant 50, 33, 22 ou 10% de foin (9,7; 10,3; 11,6 et 12,1 MJ d'EM/kg de MS respectivement), ont été distribuées à 40 boucs castrés, l'objectif étant de réaliser un gain total de poids vif de 10 kg. Au début de l'expérience, 10 autres boucs ont été abattus afin de déterminer le poids et la composition initiaux de la carcasse. Pour les animaux qui ont atteint l'objectif de gain pondéral visé, le taux de croissance et la quantité d'aliment ingérée n'ont pas été significativement différents (P>0,05) selon les rations. Les meilleures performances ont été enregistrées pour la ration à 22% de foin, avec un gain pondéral de 92 gljour et un taux de conversion de 9,5 kg de matière sèche par kg de poids vif. Les coefficients de variation des moyennes des paramètres étudiés étaient élevés, ce qui suggère que l'effet des traitements avait peut-être été masqué par l'hétérogénéité de l'échantillon. Par rapport aux abattages témoins, le poids et le rendement de la carcasse ont augmenté significativement (P<0,05) de même que les quantités de matière grasse déposées aussi bien dans les viscères que dans la carcasse. Enfin, il y a une corrélation positive entre la quantité de graisse déposée et la teneur en énergie de la ration. Introduction Peasant farmers in Zimbabwe own over two million goats (CSO, 1989). These goats, predominantly of the 'indigenous' type, are a heterogeneous population unselected for productive traits. Goats in Zimbabwe are reared primarily as meat animals (Hale, 1986), with little information on carcass attributes. Research in an extensive production system with the 'Matabele' goat of southern Zimbabwe, has suggested that a slaughter mass of about 38 kg will yield carcasses with desirable meat characteristics (Tawonezvi and Ward, n.d.). The effects of fattening goats on complete cereal-based diets differing in energy content, on animal performance, body and carcass composition were investigated. 356 Materials and Methods Fifty wethers (castrated goats) from the Matopos Research Station flock, aged either 18 or 24 months were used in the experiment. The goats were allocated into five groups of 10 animals each, balanced for body mass and age. One group (PS), was slaughtered at the beginning of the experiment to estimate the initial carcass mass and body and carcass composition. The remaining groups were randomly allocated to one of four diets containing either 50 (H50), 33 (H33), 22 (H22) or 1 0 (H 10) per cent coarsely milled (< 1 .5 cm lengths) range hay. Chemical composition and digestibility of the diets (Table 1) were as reported by Hatendi et al (1990). Animals were individually penned and offered feed in two equal meals at 07.00 and 14.00h daily. The amount of feed offered daily was adjusted each morning to allow refusals of between 100 and 200 g. Water was available at all times. Animals were slaughtered after gaining 10 kg body mass. The experiment ended after 230 days due to lack of feed. Table 1 . Composition and apparent digestibility of diets offered to goats. Hay level 50% 33% 22% 10% Veld hay 500 330 220 100 Cottonseed meal 100 100 100 100 Maize meal 320 495 606 731 Molasses 50 50 50 50 Urea 12 7 4 1 Salt 7 7 7 7 Limestone flour 5 5 5 5 Gypsum 5 5 5 5 Vit/min premix1 1 1 1 1 Dry matter (DM) (g/kg) 910 910 910 900 Ash (g/kg DM) 80 70 60 60 Crude fibre (g/kg DM) 190 160 90 90 Crude protein (g/kg DM) 200 190 150 160 Ether extract (g/kg DM) 30 30 40 40 Dry-matter digestibility2 0.69 0.73 0.74 0.83 Estimated metabolisable energy2 MJ/KG DM 9.67 10.33 11.60 12.09 1. 1 kg= 3x10 i.u. retinol, 35 g manganese, 35 g zinc, 0.6 g selenium and 0.12 g iodine. 2. Hatendi etal (1990). 357 Animals were weighed before and after a 24 to 30-hour fast to determine final body mass (FBM) and mass at slaughter (SM), respectively. The head, feet, pelt, full gut, omentum mesentery, liver and heart were removed and weighed. The mass of the empty gut was determined after washing-out the gut contents whose mass was estimated by difference (full gu-t empty gut). Hot carcass mass was recorded within 30 minutes of slaughter and cold carcass mass after chilling at approximately 7°C for 48 hours. Chilled carcasses were visually scored for fat cover as described by Colomer-Rocher (1987). Carcasses were cut into three sections. The fore section was removed from the carcass by cutting between the 6th and 7th ribs, and the rib section separated from the hind section at the penultimate rib. Eye-muscle depth, eye-muscle width and backfat depth were measured with callipers on the cut surface of the penultimate rib. The carcass sections were split along the mid-line and the right sides sealed in plastic bags and frozen pending chemical analysis. Prior to analysis the frozen sections were ground twice through a cutter-grinder fitted with a 9-mm screen (Jeffco Food and Fodder Cutter-grinder Model 262B). Ground carcasses were thoroughly mixed by hand and a representative sample analysed to determine dry matter, crude protein, ether extract and ash content (O'Donovan and Elliot, 1971). Initially data from both the PS and fed treatment groups were analysed for treatment differences by ANOVA using age and initial body mass (IBM) as covariates. Where significant treatment effects were found orthogonal comparisons were made between the PS and the fed-treatment means. Subsequently, data from the fed treatments only were reanalysed by ANOVA using age and IBM as covariates and differences between diets evaluated using paired orthogonal contrasts. Differences (P < 0.05) between means of fed animals are indicated in tables by use of superscripts. Results Three animals died during the trial of unspecified causes (Table 2). These, in addition to the six animals which did not attain the targeted body mass, and had daily body mass gains of less than 35 g/d were excluded from the analysis. Diet offered did not affect (P>-0.05) animal performance traits (Table 3) though performance generally tended to improve with decrease in dietary hay. Slaughter mass and the carcass characteristics of all groups are shown in Table 4. Among the fed treatments, though gut fill at slaughter and the proportion of SM represented by gut fill were significantly higher (P<0.05) for the H50 than either the H33 or H22 groups, diet had no effect on empty body mass. Carcass 358 Table 2. Fate of fed goats offered diets containing different roughage levels. Hay level 50% 33% 22% 10% Body mass gain >10 kg 10 Body mass gain < 10 kg Fatalities Total 10 8 7 4 2 2 2 • 1 2 10 10 10 Table 3. Performance characteristics of fed goats which gained 10 kg body mass. Hay level 50% 33% 22% 10% se Initial mass kg 25.5 26 23.9 27.4 0.51 Final mass kg 35.6 36.3 34.3 37.5 0.53 Daily gain g 68 75 92 74 4.3 Total gain kg 10.1 10.3 10.4 10.1 0.10 Total dry-matter intake kg 130.5 118.3 97.8 113.9 4.76 Feed-conversion ratio (kg food/kg gain) 12.9 11.4 9.5 11.3 0.60 masses of animals offered diet H50 were lower (P<0.05) than for those offered H33 and H22. Carcass fat score was lower (P<0.001) for the PS than the fed animals (Table 4). Among fed groups, though diet had no effect on fat score, it was highest for animals offered H10. Body parts of fed animals were heavier than those of PS animals. In terms of proportion of EBM, however, they were lower, except for the omental and mesenteric (0+M) fat (Table 5). Amongst the fed treatments the weight of the O+M fat depots and their proportion of EBM were lowest (P>0.05) for animals offered H50 (Table 5). As proportions of the cold carcass the fore and hind sections were lower (P<0.001) and the rib section higher (P<0.001) for the fed groups as compared the the PS group (Table 6). There were no differences (P>0.05) among the fed groups in carcass proportions. Eye-muscle width and depth tended to be greater (though not significantly as for the fed animals) than the PS group (Table 7). Among the fed groups, there were no differences in eye-muscle dimensions and back fat. Weight and proportion in the carcass of the kidney knob and channel fat (KKCF) was lower (P<0.001) for 359 Table 4. Carcass characteristics of preliminary slaughter and ol fed goats. Hay level PSvs Slaughter PS 50% 33% 22% 10% SE FED mass (SM) kg 26.1 34.1 33.6 33.2 36.5 0.53 *** Gut fill kg 6.0 4.9* 3.5b 3.6b 4.4ab 0.14 *** GF/SM kg/kg 0.23 0.1 5a 0.1 0b 0.1 1b 0.12"" 5.1 *** Empty body mass (EBM) kg 20.1 29.2 30.1 29.6 32.1 0.49 MM Carcass mass Hot kg 11.1 16.9a 18.5b 17.1b 18.9ab 0.27 *** Cold (CCM) kg 10.6 16.4" 18.1b 16.6b 18.3"" 0.27 *** Killing-out ratios (CCM/SM) 0.40 0.48" 0.54b 0.50"" 0.50"" 0.005 *** Carcass fat cover1 1.5 2.5 2.8 2.7 3.3 0.09 *** 'where 1 = low cover, 2 = slight cover, 3 = average cover, 4 = high cover, 5 = very high cover (Colomer-Rocher et al, 1987). Values in the same row with the same superscript are not significantly (P>0.05) different. Table 5. Proportions (glkg) of bodyparts in empty body of goats preliminary slaughtered (PS) and fed. Hay level PSvs PS 50% 33% 22% 10% SE FED Head 82 66 65 64 61 1.1 *** Feet 38 30 28 27 28 0.6 *** Pelt 110 70 70 69 66 1.7 «*» Heart 6 5 5 5 5 0.6 * Liver 27 18 18 22 20 1.7 *** Empty gut 162 121 120 117 99 0.2 *** Omentum + mesentery 15 29a 42b 38b 50b 2.7 *** the PS than the fed groups. Among the fed groups KKCF weight and its proportion of carcass mass tended to increase as the roughage level of the diet decreased and were significantly (P < 0.05) lower for the H50 group as compared to the other fed groups (Table 7). 360 Table 6. Proportions of carcass sections of preliminary slaughtered and fed goats. Hay level PSvs PS 50% 33% 22% 10% SE FED Fore section Rib section 0.47 0.09 0.45 0.45 0.13 0.42 0.45 0.13 0.42 0.46 0.13 0.42 0.46 0.12 0.43 0.002 0.003 0.002 **« ***Hind section Table 7. Eye muscle (Em) back fat and kidney measurements of the cold preliminary slaughtered and fed animals. Hav level carcass of PSvs PS 50% 33% 22% 10% SE FED em width mm 45 49 49 42 50 8 ns em depth mm 18 23 25 22 23 5 *** Back fat depth mm q q.9 0.9 0.6 1.3 0.1 *** KKCFkg o.14 0.55a 0.99b 0.84b 1.01b 0.03 *** KKCF/CCMg/kg 13 34' 55b 5^ 54b 14 Values in a row with the same superscript are not significantly (P < 0.05) different. KKCM = kidney knob and channel fat. ns = not significant. SE = standard error. CCM = cold carcass mass. Table 8 shows the chemical composition of the right carcass side (less kidneys and KKCF). Compared to PS group, feeding decreased carcass water (P<0.001), crude protein and ash content (P<0.001) and increased (P<0.001), carcass fat content. Among the fed groups carcass water, crude protein (CP) and ash were higher for the H50 group but differences were not significant. Carcass fat content was highest for the H10 group. Table 8. Water content (g kg) and chemical composition of tne dry matter of the right carcass side of preliminary slaughtered and fed animals. Hay level PSvs PS 50% 33% 22% 10% SE FED n = 8 n = 10 n=8 n = 7 n=4 Water (g/kg) 588 536 516 488 511 6.9 * Crude protein (%) 438 347 314 325 265 11.2 *** Ether extract (%) 384 542 590 579 654 14.4 MM Ash (%) 178 111 96 96 81 3.6 *** Daily carcass gain (less kidneys and KKCF) tended to have more fat as the concentrate proportion in the diet offered increased (Figure 1). 361 \ LU C to c" .§ Em © C .fc: "> -o ||i ?-o 5 a) o T3 100 g/d) for growing goats by NRC (1981), Devendra and Burns (1983) or Wilkinson and Stark (1987). Animal performance was comparable to that reported for African breeds of goats reared intensively (Adebowale and Ademonsun, 1981 ; El Hag et al, 1984; El Hag et al, 1985). Growth rates in this trial were greater than the mean post-weaning gain between 5 and 1 8 months of 36 g/d reported at Matopos Research Station for extensively reared goats (Baffour-Awuah, 1987). Means of performance traits in this study had high coefficients of variation. This suggests a lack of uniformity of the indigenous goat population and may have masked treatment effects. Gut fill as a proportion of mass at slaughter was greater than the six to eight per cent reported for finished Sudan Desert goats (Gaali et al, 1972). Carcass mass and yield of the fed animals in this study were greater than those previously reported at Matopos Research Station for animals of similar origin reared extensively (38 kg body mass yielding 16 kg hot carcass mass; Tawonezvi and Ward, n.d.), and are comparable to those reported elsewhere for finished goats of similar slaughter mass (Naude and Hofmeyr, 1981; Devendra and Burns, 1983). Carcass fat cover increased appreciably in the fed groups. The observed large deposits of fat in the omentum + mesentery, and the kidney knob and channel fat depots, support reports which indicate minimal fat deposition in the subcutaneous fat depot of goats (Kirton, 1970; Gaali et al, 1972). The omental and mesenteric fat depots increased in weight by 3 to 5 times in the fed animals, representing between three and five per cent of the empty body weight. A large amount of fat was deposited in the carcass tissues of the fed animals. Whilst some of the differences in carcass chemical composition between treatment groups can be explained by the relationship with carcass mass (Kirton, 1970; Naude and Hofmeyr, 1981), there was a tendency for the fat component of carcass gain to increase with dietary energy content and intake. The large amount of fat deposited in non-carcass and carcass fat depots may be indicative of the final body mass of animals in this study approaching their mature mass 363 and thus a tendency towards 'fattening' rather than 'growing'. The increased fat deposition with dietary energy level may in part be attributable to the effects of reduced protein: energy ratio of the diets as the roughage content decreased (Beede et al, 1985; Mtenga and Kitaly, 1990). Goats exhibit a centripetal growth pattern (Colomer-Rocher, 1987), with the proportion of the carcass represented by the forequarter and joints along the mid-line increasing with age and carcass mass (Kirton, 1970; Owen and Norman, 1977). The resultant improved carcass conformation of the fed groups becomes evident if the carcass mass to length ratio is used as an objective indicator of conformation (Colomer-Rocher, 1987). Calculated from the means of cold carcass weight and body length, with the ratio of the PS group used as a baseline, ratios for the fed treatment groups were 1.3, 1.6, 1.4, and 1.5 for treatments H50to H10, respectively. Acknowledgements We thank the staff of Matopos Research Station, Grasslands Research Station, the University of Zimbabwe and Agrifoods (pvt.) Ltd. who contributed to this experiment. One of us (PRH) is in receipt of the RioTinto Foundation (Zimbabwe) postgraduate fellowship. References Adebowale E A and Ademosun A A. 1981. Studies on the utilization of brewers' dried grains by sheep and goats: 1 . Growth studies. Bulletin of Animal Health and Production in Africa 29:365-370. ARC (Agricultural Research Council). 1980. The nutrient requirements of ruminant livestock. Commonwealth Agricultural Bureaux, Slough, UK. 351 pp. Baffour-Awuah O. 1987. Environmental influences on growth of indigenous goats. In: Proceedings of the Zimbabwe Society for Animal Production Livestock Research Symposium. Beede D K, Schelling GT, Mitchell GE and Tucker RE. 1985. Utilization by growing goats of diets that contain momensin and low or excess crude protein: Comparative slaughter experiment. Journal of Animal Science 61 (5): 1230-1 242. Colomer-Rocher F. 1987. Factors influencing carcass quality. Carcass components and composition. In: Santana O P, Aliomar Gabriel da Silva and Foote W C (eds), Proceedings of the IV International Conference on Goats. Volume 1 . International conference held in Brasilia, Brazil, 8-13 March 1987. Departamento de Difusao de Tecnologia- DDT, Brasilia, Brazil, pp. 181-194. Colomer-Rocher F, Morand-Fehr P and Kirton A H. 1987. Standard methods and procedures for goat carcass evaluation, jointing and tissue separation. Livestock Production Science 17:149-159. CSO (Central Statistical Office). 1989. Quarterly digest of statistics (March 1989). CSO, Harare, Zimbabwe. 364 Devendra C and Burns M. 1983. Goaf production in the tropics. Commonwealth Agricultural Bureau, Slough, UK. 183 pp. El Hag M O, El Hag G A and Gaali E. 1984. Sudan desert sheep and goats. Performance on high roughage and concentrate diets with monensis. International Goat and Sheep Research 2:274-282. El Hag M 0, Kurdi 0 1 and Mahgoub S 0. 1985. Performance and carcass characteristics of Sudan Desert sheep and goats on high roughage diets with added fat. Animal Feed Science and Technology 1 3: 1 47-1 53. Forbes J M. 1986. The voluntary food intake of farm animals. Butterworths, London, UK. 206 pp. Gaali ESC, Ghanem Y S and Mukhtar A. 1972. A comparative study of some carcass characteristics of Sudan Desert sheep and goats. Animal Production 14:351-357. Hale D H. 1986. Systems of production and productivity of goats in three communal areas of Zimbabwe. In: Adeniji K O and Kategile J A (eds), Proceedings of the Workshop on the Improvement of Small Ruminants in Eastern and Southern Africa, Nairobi, Kenya, 18-22 August 1986. OAU/IDRC (Organization of African Unity/International Development Research Centre), Nairobi, Kenya, pp. 181-193. Hatendi P R, Smith T, Ndlovu L and Mitisi C. 1 990. Intensive finishing of indigenous meat goats: Diet utilization, growth and carcass weight changes of mature castrates . Paper presented at the Second Symposium on Science and Technology, the Scientific Council of Zimbabwe, Harare, Zimbabwe. Kirton A H. 1970. Body and carcass composition and meat qualities of the New Zealand Feral Goat (Capra hiricus). NewZealandJournal ofAgricultural Research 13:167-181. Mtenga L and Kitaly A J. 1990. Growth performance and carcass characteristics of Tanzanian goats fed Chloris gayana hay with different levels of protein supplement. Small Ruminant Research 3: 1-8. Naude R T and Hofmeyr H S. 1981. Meat production. In: Gall C (ed), Goat production. Academic Press, New York, USA. pp. 285-307. NRC (National Research Council). 1981. Nutrient requirements of goats: Angora, dairy and milk goats in temperate and tropical countries. Nutrient Requirements of Domestic Animals 15. National Academy of Science, National Research Council, Washington, DC, USA. 91 pp. O'Donovan W M and Elliot R C. 1971. Developmental changes in the bodies of Dorper sheep. 1 . Changes in live body-weight, body proportions and composition of weaned Dorper lambs given different amounts of food. Rhodesia Journal of Agricultural Research 9:65-75. Owen J E and Norman G A. 1977. Studies on the meat production characteristics of Botswana goats and sheep. Part III. General body composition, carcass measurement and joint composition. Meat Science 1 :283. Tawonezvi H P R and Ward H K. (n.d.). Indigenous sheep selection programme and productivity of indigenous sheep and goats. In: Annual report 1981182. Division of Livestock and Pastures, Department of Research and Specialist Services, Harare, Zimbabwe, pp. 92-95. Wilkinson J M and Stark B A. 1987. The nutrition of goats. In: Haresign W and Cole D J A (eds), Recent advances in animal nutrition. Butterworths, London, UK. pp. 91-106. 365 The utilisation of sorghum stover fed to sheep as influenced by urea or graded levels of lablab supplementation I.F. Adu, B.A. Fajemisin and A. M. Adamu National Animal Production Research Institute Ahmadu Bello University, Shika-Zaria, Nigeria Summary The effect of urea and graded levels of Lablab purpureus cv Rongai on the voluntary intake and utilisation of sorghum stover was determined with 30 Yankasa sheep. The physical and chemical composition of the sorghum stover and lablab as well as their stem and leaf fractions were described. The intake of sorghum stover was depressed when supplemented with Lablab but overall feed intake improved with increasing levels of Lablab supplementation. The addition of urea increased stover intake and the digestibility of the dry matter, neutral detergent fibre (NDF) and nitrogen. Supplementing sorghum stover with either Lablab or urea generally improved liveweight gains of the sheep. It is concluded that supplementing sorghum stover with Lablab improved its intake and utilisation. However at high levels of supplementation, Lablab tends to have a substitution effect on sorghum intake. Effet d'une complementation d'uree ou de dolique sur l'utilisation de la paille de sorgho chez les ovins I.F. Adu, B.A. Fajemisin etA.M. Adamu Resume L'effet de l'uree et de niveaux croissants de dolique (Lablab purpureus) cv. Rongai (blanc) sur l'ingestion volontaire et I'utilisation 367 de la paille de sorgho a ere etudie chez 30 moutons Yankassa. La composition physique (proportion de tiges, de feuilles) et chimique de la paille et du dolique est presentee. La complémentation de dolique a déprimé l'ingestion de la paille de sorgho: (de 9 et 7% pour 90 et 180 g de dolique). Cette baisse est devenue significative (P<0,05) avec une complémentation de 270 g. L'addition d'uree a augments I'ingestion de paille et la digestibilite de la matiere seche et des parois cellulaires, respectivement de 6,5, 14,6 et 11,8% et a entraini un accroissement significant (P<0,05) de la digestibilite de l'azote. La complémentation s'est globalement traduitpar une amélioration du gain de poids des moutons. Au vu de ces resultats, il apparait que la complémentation de la paille de sorgho par le dolique en ameliore la consommation et l'utilisation. Toutefois, a niveau eleve (au-dela de 20%), le dolique tend a avoir un effet de substitution. Introduction Quantitatively, sorghum is the most important cereal crop grown in Nigeria (Table 1). About 46% of the estimated 21795000 hectares of land under cultivation is devoted to sorghum production with an annual total grain yield of about 5.53 million tonnes (Central Bank of Nigeria, 1988). Alongside this grain production is an estimated 22.14 million tonnes of stover which is available for ruminant livestock feeding. And since cereal crop growing areas are also the home of most of Nigeria's livestock population, cereal crop residues constitute an important feed resource during the dry season. Also, because Nigeria cannot as yet develop intensive livestock feeding systems based on expensive and scarce feed resources like grains, oilseed cakes, high-quality pastures, crop residues will continue to be important in Nigeria's livestock feed resources. However, crop residues are known to have low content of available protein, energy and minerals. It is recognised that rumen fermentation is impaired and animal performance lowered when the nitrogen content of the diet is less than 1 .2% (Conrad and Hibbs, 1968). The feeding of energy and protein supplements is known to enhance the utilisation of poor-quality feeds like crop residues by maximising roughage degradation and optimising rumen microbial protein synthesis (Anderson, 1978; O'Donovan, 1983). Because of high cost, scarcity and other logistic problems, the use of concentrates and non-protein nitrogen as supplements is not justified. The current interest in a more intensive use of crop residues in Nigeria justifies the search for cheaper and readily available supplements. Lablab purpureus is a dual-purpose legume crop that is rapidly gaining acceptance by agropastoral 368 farmers in northern Nigeria (Tanko et al, 1990). Lablab has high seed and forage yield as well as good hay-curing ability compared with other commonly grown legume crops in Nigeria. The objective of this study was to compare the effect of urea and three levels of Lablab supplementation on the utilisation of a basal diet of sorghum stover by Yankasa sheep. Table 1 . Annual grain, stover/haulm yield and nutritive value of some common crops in Nigeria ('000 tonnes). Stover or haulm/vines2 Crude protein content of stover/ Crop Seed or grain1 haulm (%) Sorghum 5534 22136 1.6-7.5 Maize 1370 5480 2.6-4.9 Millet 4170 25020 4.0-5.4 Rice 307 2456 3.0-4.6 Cowpea 698 3490 5.9-10.4 Groundnut 709 2188 11.4-16.7 Benniseed 36 324 2.0-5.1 Soyabean 121 363 4.3-6.8 1. Central Bank of Nigeria (1988). 2. The extraction rates (seed:stover/haulm) developed by Alhassan and Kalian (1984) in NAPRI (unpublished) were used. Materials and methods Sorghum (SK-5912) was planted at a seed rate of 10 kg ha'1. At planting, fertiliser was applied at the rate of 37.5 kg N ha"1 followed at 6 weeks by a side-dressing of 22 kg N ha'1. The sorghum was harvested for grains 123 days after sowing. After grain harvest, the whole plant was removed at ground-level and packed. Thereafter, a random sample was taken, weighed and subdivided into leaf and stalk portions. This paper reports the trials with ground sorghum stover (3-4 cm). Lablab purpureus cv. Rongai (white) was planted from seed at the rate of 15 kg ha'1. At 90 days the forage was cut and field- dried. A random sample was taken, weighed and separated into leaf and stem fractions. The Lablab was chopped into 1-2 cm pieces and used as the supplement in a series of experiments. Samples of the whole sorghum and Lablab as well as their leaf and stem fractions were analysed for dry matter, nitrogen, NDF, lignin and ash (AOAC, 1975; Goering and Van Soest, 1970). 369 The nylon bag method (Orskov et al, 1980) was used to estimate ruminal degradation of the whole stover, Lablab and their respective leaf and stem fractions. About 3 g of each sample was placed in nylon bags (size 100 x 170 mm with 12 mm pore size) and were incubated in the rumen of 2 Bunaji cows in replicates. The bags were anchored by a 50-cm nylon string to the cap of the rumen cannulae and withdrawn at 12, 18, 24, 48 and 72 hours. They were then washed and dried at 60°C for 48 hours (Mehrez and Orskov, 1977) for the estimation of dry-matter and nitrogen disappearance. Thirty yearling Yankasa sheep with an average weight of 22.8±0.7 kg were treated for endo- and ectoparasites and used for the 78-day feeding trial. The treatments were: A — Sorghum stover ad libitum B — Sorghum stover + 90 g Lablab C — Sorghum stover + 180g Lablab D — Sorghum stover + 270g Lablab E — Sorghum stover + 9 g urea + sulphur to achieve N:S ratio of 10:1 The Lablab supplements were intended to provide 1 0, 20 and 30% of total intake while the urea represented 1.5% total intake. The urea was dissolved in about 5 cc of water and mixed with the sulphur. The resulting solution was mixed with a grab-sample of the stover and fed to the animals first before offering the basal diets. This was to ensure a complete consumption of the urea. The stover was fed ad lib. Animals were weighed fortnightly. A 7-day total faecal collection was done at the end of the feeding trial. All data were subjected to analysis of variance and differences among treatment means were determined (Snedecor and Cochran, 1980). Results and discussion The sorghum stover is made up of 17.3% leaf, 26% upper softer stem and 56.7% lower woody stem while the corresponding figures for the Lablab are 38.7, 13.5 and 47.7%, respectively. In terms of physical composition the leaf and upper softer stem, which generally contain a higher concentration of most nutrients, constituted approximately 43 and 53% of the stover and Lablab, respectively. The apparently higher ash content (Table 2) of the leaf fraction may be partly due to soil contamination while being dried in the field. As expected, the NDF and lignin in the stem fractions were generally higher than in the leaf fractions. The grinding of the stover reduced selection of leaf fraction which are generally more nutritious and more relished by animals (Martin and Wedin, 1974; Powell, 1985) and enhanced a better utilisation of the whole plant. 370 Table 2. Chemical composition of whole, leaf and stem fractions of sorghum stover and Lablab. Sorghum stover Lablab Leaf Sorghum stover + urea Whole Leaf Stem Whole Stem Dry matter 96.6 95.9 96.5 94.6 93.1 95.1 94.8 Nitrogen 0.84 1.1 0.74 1.6 2.1 1.2 2.2 NDF 76.0 78.6 89.4 38.9 39.7 62.4 69.3 Lignin 11.8 10.7 13.6 4.2 6.8 10.9 14.3 Cellulose 43.0 21.9 47.3 10.5 21.9 40.4 39.3 Ash 7.8 7.7 3.6 7.4 11.1 6.5 8.6 The 48-hour dry-matter disappearance rate (Table 3) increased with incubation time. Lablab was degraded more rapidly than stover. In both cases, the leaf fractions was degraded more rapidly than the stem fractions. Supplementation of sorghum stover with either Lablab or urea improved its degradability (Table 3) and digestibility (Table 4). Table 3. Dry-matter disappearance (g/1 00 g) of whole leafand stem fractions from nylon bag incubated in the rumen of Bunaji cows. Hours Sorghum stover Lablab Sorghum Sorghum stover + stover + bation Whole Leaf Stem Whole Leaf Stem Lablab (4:1) urea 12 8.9 13.8 7.6 10.9 14.7 9.7 12.7 10.3 18 13.1 21.5 11.8 15.1 19.1 12.5 17.2 14.2 24 25.4 31.8 21.1 29.2 32.4 22.2 30.7 27.6 48 33.8 42.4 29.6 35.5 43.5 28.4 37.4 38.1 72 41.2 50.6 37.0 48.6 53.6 35.9 44.3 44.2 371 Table 4. Digestibility (glkg) of the different feed combinations used. Stover + Stover + Stover + Sorghum Stover 180g 270g urea stover + 90g Lablab Lablab Lablab Dry matter 457 476 537 541 524 Organic matter 475 507 526 572 502 Crude protein 613" 693b 761 b 789b 768b NDF 551 559 577 596 616 Means with different superscripts within each row are significantly different (P < 0.05). Total feed intake increased significantly (P<0.05) when the stover was supplemented with 180 g Lablab. The 85, 170 and 255 g actual Lablab intake accounted for 18.8, 31.4 and 53.3% of the total dry matter intake, respectively. Total feed intake and utilisation were generally better in sheep given stover plus either urea or Lablab, a confirmation that nitrogen is deficient in the stover (Table 5). Stover intake was depressed by 9 and 7% at the 10 and 20% levels of Lablab supplementation, while the depression (44%) was significant (P<0.05) at the 30% level of supplementation. This result tends to support the contention that supplements have beneficial effects on the utilisation of fibrous feeds if the level of supplementation does not exceed 20% of the diet dry matter (Preston and Leng, 1984; Wegad and NDumbe, 1987). It would appear that Lablab had a substitution effect beyond the 20% level of supplementation. Table 5. Performance of sheep fed sorghum stover supplemented with urea or Lablab. Stover + Stover + Stover + Stover 90 g Lablab 180g 270 g Stover + alone (10%) Lablab (20%) Lablab (30%) urea Stover intake (g/day) 401.3a 367.1a 372.3' 223.5b 427* Total DM intake (g/day) 401 .3" 452.2" 542.2b 478.8* 427* Total DM intake (g/kgW°-75/day) 41.3 45.4 57.2 47.7 43.6 Liveweight gain (g/day) -20.9 15.9a 36. 1b 47.7b 9.3" Feed efficiency 28.4 15.0 10.0 45.9 Means with different superscripts within each row are significantly different (P < 0.05). Though not significant (P>0.05) the digestion of dry matter, organic matter and NDF generally increased with increasing levels of Lablab supplementation. 372 Crude-protein digestibility was, however, significantly (P<0.05) improved by the supplements. The supplements generally improved rumen fermentation of the test diets and the range of improvement falls within that reported by Sharma et al (1972). The improved liveweight gains of sheep fed Lablab supplement justify its use in preventing liveweight loss during the critical dry season period. It is concluded that the use of sorghum stover as a feed resource is greatly enhanced by supplementing with Lablab. The planting of Lablab with sorghum or maize has been accepted as a cropping practice by farmers in the subhumid zone of Nigeria. The settled farmers who feed crop residues to animals in their rugas can be encouraged to further reduce particle size (or grind) to enhance better utilisation. Already some well-to-do farmers have installed grinders on their farms. References Anderson D C. 1978. Use of cereal residues in beef cattle production systems. Journal of Animal Science 46:849-861. AOAC (Association of Official Analytical Chemists). 1975. Official methods of analysis. 12th edition. AOAC, Washington, DC, USA. Central Bank of Nigeria. 1988. Annual report. Lagos, Nigeria. Conrad H Rand Hibbs J W. 1968. Nitrogen utilization by the ruminant. Appreciation of its nutritive value. Journal of Dairy Science 51 :276-284. Goering H K and Van Soest P J. 1970. Forage fibre analysis (apparatus, reagents, procedures andsome applications) . Agriculture Handbook 379. Agricultural Research Service, Department of Agriculture, Washington, DC, USA. 20 pp. Martin N P and Wedin W F. 1974. Effect of fall weathering on yield and composition of grain sorghum stover. Agronomy Journal 66:669-672. Mehrez A Z and Orskov E R. 1977. A study of the artificial fibre bag technique for determining the digestibility of feeds in the rumen. Journal of Agricultural Science (Cambridge) 88:645-658. O'Donovan P B. 1983. Untreated straw as a livestock feed. A review. Nutrition Abstracts and Reviews: Series B-Livestock Feeds and Feeding 53:441-455 .Orskov E R, Hovell F D de B and Mould F. 1980. The use of the nylon bag technique for the evaluation of feedstuffs Tropical Animal Production 5:195-213. Powell J M. 1985. Yields of sorghum and millet and stover consumption by livestock in sub-humid zone of Nigeria. Tropical Agriculture (Trinidad) 62:77-81 . Preston T R and Leng T R. 1984. Supplementation of diets based on fibrous residues and by-products. In: Sundstol F and Owen E (eds), Straw and other fibrous by-products as feed. Developments in Animal and Veterinary Sciences (The Netherlands) 14. Elsevier Science Publishers BV, Amsterdam, The Netherlands, pp. 373-413. Sharma V V, Jhanwa D M and Taparia A L. 1972. Utilization of sorghum stover by cattle. Indian Journal of Animal Science 42:480-487. Snedecor G W and Cochran W G. 1980. Statistical methods. 7th edition. Iowa State University Press, Ames, Iowa. 507 pp. 373 Tanko R J, Kallah M S and Otchere E O. 1990. Livestock production in northern Nigeria: Dry matteryield and chemical composition of dual purpose legumes. Paper presented at the 15th Annual Conference of the Nigeria Society for Animal Production, Ago-fwoye, Nigeria. Wegad D and NDumbe R D. 1987. The effect of different protein supplements on weight gain and voluntary intake of maize stover by cattle. In: Little D A and Said A N (eds), Utilization of agricultural by-products as livestock feeds in Africa. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 99-102. 374 Lablab (Dolichos lablab) meal as protein supplement for weaned fattening lambs W.D. Mafwere and LA. Mtenga Department of Animal Science Faculty of Agriculture Sokoine University of Agriculture P 0 Box 3004, Chuo Kikuu, Morogoro, Tanzania Summary A study was conducted to investigate the effect ofinclusion ofLablab meal on growth rate, feed intake and carcass composition of Black Head Persian (BHP) lambs. Twenty-four BHP castrate sheep with average weight of 14. 1 ±2.7 kg were allotted randomly to four dietary treatments. Lambs were individually maintained in metabolic crates. All animals were offered hay ad libitum. Lambs on control treatment (A) were given in addition 380 g of maize bran per day. Animals on treatments B, C, and D were given daily 380 g of concentrate based on maize bran with 25, 50 and 75% lablab meal, respectively. Growth rate, dry-matter intake and feed conversion ratio were 34, 66, 68, 71 glday; 58.7, 69.5, 74.0 glkgVfn/day and 15.2, 11.3, 9.3, 9.1 g/DMI/g gain, respectively, of treatments A, B, C and D. The level of inclusion of lablab meal had no significant (P>0.05) effect on killing-out characteristics. Hot carcass weight as percentages of slaughter weight and empty body weight were 41.3, 41.0, 40.5, 42.5 and 52.0, 52.3, 53.5, respectively, for animals on treatments A, B, C and D. Carcass composition expressed as percentage of carcass weight showed no significant treatment effects and value of 60.91, 61.29, 61.11 and 62.97% for carcass lean, 19.06, 20.24, 18.42 and 17.54% carcass fat and 17.86, 16.70, 17.91 and 16.67% for carcass bone were obtained for treatments A, B, C and D. It is concluded that inclusion of lablab meal in maize bran above 25% has no beneficial effect in terms of growth rate and carcass composition. 375 Utilisation de la farine de dolique (Dolichos lablab) comme complément protéique dans l'embouche d'agneaux sevrés W.D. Mafwere et LA. Mtenga Résumé Une étude a été effectuée en vue de mettre en évidence l'effet d'une complémentation alimentaire de dolique sur la croissance, la consommation et la composition de la carcasse d'agneaux de race Somali à tête noire (BHP). Vingt-quatre moutons BHP castrés, pesant en moyenne 14,1 ±2,7 kg et gardés individuellement dans des cages métaboliques, ont été répartis au hasard en quatre groupes (A, B, C et D) de traitement recevant chacun une ration déterminée. Du foin leur était servi ad libitum. Les animaux du groupe témoin (A) ont en outre reçu chacun 380 g de son de maïs parjour tandis que ceux des groupes B, CetD ont eu en remplacement la même quantité d'un concentré à base de son de maïs et contenant 25, 50 et 75% de farine de dolique. Pour les groupes A, B, C et D, le gain de poids, la consommation de matière sèche et le taux de conversion des aliments furent de 34, 66, 68 et 71 gljour; 58,7, 64,7, 69,5, 74,0 glkg de poids vif0'75 et 15,2, 11,3, 9,3, 9, 1 g MSIg de poids vif. La teneur du concentré en dolique n'a pas eu d'effet significatif (P>0,05) sur les caractéristiques à l'abattage. Le poids de la carcasse chaude exprimé en proportion du poids à l'abattage et en outre, le traitementn'avaientpas d'effet significatifsur la composition de la carcasse exprimée en proportion du poids de celle-ci: pour les animaux des groupes A, B, C et D respectivement, les chiffres obtenus étaient de 60,91, 61,29, 61,11 et 62,97% pour la carcasse maigre, de 19,06, 20,24, 18,42 et 17,54% pour la matière grasse de la carcasse et de 17,86, 16,70, 17,91 et 16,67% pour les os de la carcasse. Au vu de ces résultats, il apparaît qu'en ce qui concerne la croissance et la composition de la carcasse, il n'y a aucun avantage à donner à des agneaux un concentré à base de son de maïs contenant plus de 25% de dolique. 376 Introduction Most of the four million sheep found in Tanzania are kept under extensive management systems and depend almost exclusively on natural pastures for their nutrient requirements. The availability of these feeds is mainly determined by the amount of rainfall and length of the dry season. Sheep find adequate quantities of a fair- to good-quality herbage during the initial months of the wet season and consequently gain weight. However, they start losing weight as soon as the grasses approach maturity and their condition deteriorates progressively through the dry season. Several experiments with sheep and goats have shown that performance can be improved by protein supplementation (Mtenga and Nyaky, 1 985; Mtenga and Kitaly, 1990). However, most of the supplements used such as soyabeans, cottonseed cake and cashewnut cake are expensive and are not readily available in sheep-rearing areas. Leguminous multipurpose trees and forages such as lablab niger offer the best alternative cheap source of protein supplement. Lablab can be easily cultivated and is drought resistant. It can be successfully grown in dry areas with rainfall as low as 400 mm (Luck, 1965). It can also be intercropped with maize or sorghum. The beans as meal and the forage as hay or silage, can form a good source of supplement to ruminants. A detailed study is underway at Mpwapwa in Tanzania to evaluate the yield and effect of supplementing low-quality maize and sorghum stover with lablab to provide further information on the effect of supplementing low-quality Rhodes grass hay with different levels of lablab bean meal on performance of lambs. Materials and Methods Animals and dietary treatments A completely randomised design was undertaken with 24 castrated Black Head Persian (BHP) lambs. The age of the lambs was between 4 and 7 months and initial weight was 14.1 ±2.7 kg. Animals were randomly allocated to four dietary treatments. Chloris gayana hay was harvested from the University farm and was offered ad libitum to animals in all treatments. In addition to hay, animals in treatments A, B, C and D were offered 380 g of supplements containing maize bran, lablab seed meal and mineral mixture in ratios of 95:0:5, 70:25:5, 45:50:5 and 20:75:5, respectively. The supplements were formulated to contain 100.7, 141.7, 182.7 and 223.7 g CP/kg DM for treatments A, B, C and D, respectively. The mineral mixture contained Ca 186, P 35, Na 130, CI 200, Co 0.3, Cu 1.2, Fe 3.1, Mg4.4andSD 1.8g/kg. Lambs were penned in 150 x 90 cm wooden pens and fed individually. Hay was chopped into 4-6 cm length and was offered twice a day at 09.00 and 14.30 hrs 377 while supplements were given once a day at 08.00 hr. Water was available at all times. Animals were weighed weekly. Samples of hay and supplement offered and refusals were collected for chemical analysis. Slaughter and carcass evaluation At the end of the feeding period which lasted for 100 days, four animals per treatment were slaughtered, skinned and eviscerated. The head was removed at the OS occipitate-first cervical junction, while the hind and forefeet were separated at the tarso-metatersal and carpo-metacarpal junctions. Carcass was weighed after the kidney and pelvic fat was removed. The carcass was split into two halves and the left side was dissected into lean, fat and bone (Cuthbertson etal, 1972). Chemical analysis All dried feeds were ground to pass through a 1-mm screen in a Christy and Norris 20-cm laboratory hammer mill. Chemical components of the feeds and meal samples (Table 1) were determined by standard methods (AOAC, 1975). Table 1 . Composition of hay, feed ingredients and supplements. Hay refusal Maize bran lablab meal Supplements B CHay A D No. of samples 8 8 6 6 6 6 6 6 DM (g/kg ADW) 850 860 898 876 894 892 885 882 Composition of DM (g/kg) Crude protein 52 24 134 254 132 169 206 227 Crude fibre 380 437 57 921 56 67 78 86 Ether extract 7 9 49 6 49 32 24 14 Ash 60 126 46 47 46 46 48 47 Organic matter 940 874 954 953 954 954 952 953 DM = Dry matter; Statistical analysis Data were analysed by standard analysis of variance procedure for a completely randomised design (Steel and Torrie, 1980). If the treatment mean square was significantly greater than error mean square, treatment means were compared 378 by the New Multiple Range Test of Duncan, using standard error of difference (s.e.d.)atP<0.05. Results Chemical composition of Chloris gayana hay, individual feed ingredients and the supplement are shown in Table 1 . The supplements in B, C and D had crude protein content of 169, 206 and 227 g/kg DM. These values are higher than the initially intended values of 141, 183 and 224 g/kg DM, respectively, derived from published figures. Hay refusals were lower in crude protein than hay offered. Total dry-matter intake (g/day and g/kgW° 75) increased with increasing level of dietary lablab meal in the supplements as a result of increased intake of hay (Table 2). Lambs on treatments B, C and D consumed 18, 43 and 50% more hay than lambs on treatment A (without lablab supplement) . A similar trend was also observed in protein intake and animals on treatments B, C and D consumed 27, 58, 86% and 28, 56, 90% more protein per day and per kgW° 75 than animals on treatment A. The estimated concentrations of protein of diets consumed were 7.8, 8.8, 10.9 and 12.2 g digestible crude protein per MJ ME. Table 2. Effect of level of lablab meal in supplements on feed intake and growth rate. Treatments SEDand A B C D significance Initial weight (kg) 14.88 13.97 14.97 13.14 1.253 NS Final weight (kg) 17.76" 20.47b 21.10b 20.62b 8.971*** Growth rate (g/day) 34" 66a 68b 71b 9.7*** Intake (g/day) Hay 194" 228ab 276b 291b 4.3" Total DM 498" 547ab 596b 630b 5.35** Crude protein 56.6a 72.4b 89.5° 105.4d 1.08*** Intake (g/kgW° 75) DM 58.7" 64.8b 69.5° 72.0C 1.71*** Crude protein 6.7a 8.6b 10.5C 12.8d 0.49*** gDMI/g gain 7.8 8.8 10.9 12.2 - gDMI/g gain 15.2a 11.3b 9.3b 9.2b 1.13*** DM = dry matter; DMI = dry-matter intake; SED = standard error of difference; NS = not significant. 379 Mean weekly liveweight changes as influenced by treatment are shown in Figure 1 . Animals without lablab meal supplement (treatment A) were inferior in growth rate throughout the period of study. However, treatment differences in growth trends of animals with lablab supplements (treatments B, C, and D) were not apparent. Table 2 shows that animals on treatment A had significantly (P<0.05) slower growth rate than animals on treatment B, C, and D. However, the difference in food conversion ratio among animals offered different levels of lablab supplement were small and insignificant (Table 2), but animals that received 25% lablab did not gain as fast as those fed which had higher levels of lablab (P<0.05). Figure 1 . Mean weekly liveweight as influenced by level of lablabsupplementation. Liveweight (kg) 21 i 20 -^ Treatment A -o Treatment B -• Treatment C -A Treatment D 13 14 15 Time (weeks) 380 The level of inclusion of lablab in the ration had no significant effect on killing-out characteristics (Table 3). Dressing percentage showed no defined pattern Animals on treatments A and B had significantly higher proportion of kidney fat than animals on treatment D. The weights of lean, fat and bone as percentages of carcass weight and tissue ratios were also not affected by treatment. Table 3. Effect of levels of lablab meal in supplements on slaughter. Characteristics and carcass composition Treatments SED and Carcass weight A B C D significance %SW 41.3 41.0 40.3 42.5 0.83 NS %EBW 52.5 52.0 52.3 53.5 1.11 NS Components (% EBW) Gut fat 2.7 2.3 1.7 1.3 0.38 NS Head 8.5 8.0 8.0 7.4 0.29 NS Skin 9.6 9.5 10.2 9.0 0.40 NS Alimentary tract 8.4 86 7.8 7.9 0.31 NS Liver 1.7 1.8 1.8 1.7 0.09 NS Kidney and pelvic fat 1.4" 1.49a 0.9" 0.7b 0.14* Heart 0.8 0.8 0.8 0.7 0.334 NS Tissue in carcass (%) Lean 60.9 61.3 61.1 63.0 1.19NS Fat 19.1 20.2 18.4 17.5 0.86 NS Bone 17.9 16.7 17.9 16.7 0.47 Tissue ratios in carcass Lean: fat 32 3.1 33 3.6 0.20 NS Lean: bone 3.4 3.7 3.4 3.8 0.14 NS Lean + fatbone 4.1 3.9 4.3 4.5 0 23 NS SW = slaughter weight; EBW = empty body weight; NS = not significant; SED = ????? No correction was made for losses during dissection which ranged from 1.8 to 2.7%. 381 Discussion Crude protein content of 59 g/kg DM of Chloris gayana hay used in this study is lower than that of 86 g/kg DM reported by Mtenga and Nyaky (1985) and of 143 h/kg DM reported by Mtenga and Shoo (1990) but higher than the value of 44 to 54 g/kg DM found by Mtenga and Massae (1988) from hay harvested in the same University farm. This variability is mainly associated with season of cutting and stage of growth at the time of harvest. The protein content of lablab bean meal of 254 g CP/kg DM found in the present study compares well with values reported elsewhere. Gohl (1981) found values of 242 and 280 g CP/kg DM for seeds from Uganda and Zimbabwe, respectively, while Santos (1976) reported a value of 270 g/kg DM for seeds from Mozambique. Protein supplementation may affect DM intake through its effect on digestion in the rumen (Adeneye and Oyenuga, 1 976; Egan, 1 977) and this may be the case in the present study. In our earlier studies on protein supplementation, voluntary intake of dry matter was reduced by protein supplementation containing high energy in diets and this was due to the animals receiving sufficient energy from the supplement to satisfy its requirements. Thus total DMI and hay intake will depend on the hay quality, amount of concentrate offered and the energy:protein ratio of the supplements. Response to nitrogen supplementation on growth rate has been reported by several workers. A crude-protein content of 1 1% in DM is considered ideal for normal liveweight gains in sheep and goats (Kay et al, 1968; Kay and MacDearmid, 1973). A similar trend in growth rate (34.14, 65.65, 68.01 and 70.5 g/day for treatments A, B, C and D, respectively) was also observed in the present study as the protein content of the diet consumed increased in the order of 1 1 .36%, 13.22%, 1 5% and 16.73% CP, DM. The growth rate observed in lambs on a low level of protein intake in this study (treatment A) was lower than that 50.0 g/day reported by Mtenga and Nyaky (1985) for similar sheep under grazing conditions. Similarly Mtenga and Nyaky (1 985) reported higher values of growth rate ranging from 73.4 to 94.6 g/day in BHP sheep under grazing conditions. The higher growth rates in these studies were perhaps due to ability of the lambs under grazing condition to select the more nutritious parts of browse during grazing. The findings are, however, in agreement with that 25.19 g/day reported by Shoo (1986) and that 33.9 g/day reported by Mtenga and Massae (1988) in BHP lambs under confinement which was the case in this study. Growth rate for lambs under high levels of protein intake (16.73% CP) were also lower than that of the 98.5 g/day reported by Mtenga and Massae (1988) and that of 120.27 g/day reported by Mtenga and Nyaky (1 985) for similar type of sheep under grazing plus concentrate supplementation. The tendency for lower growth rate at each level of protein supplementation observed in the present study could partly be due to the low intake of energy 382 (Balch, 1976). The level of energy supplied to sheep has been considered to be of importance as the level of protein intake increases (ARC, 1980). In the present study increasing the level of lablab meal in the diet increased the protein: energy ratio, whose values were 26.59, 29.08, 36.84 and 43.44 g/DP/Mcal DE for treatments A, B, C and D, respectively. The importance of the protein : calorie ratio of diets has been reported by Preston et al (1965) who found that daily intake of 22 g DCP per Mcal of DE resulted in the maximum body gain of 251 g/day with wether lambs. Similarly, Adeneye and Oyenuga, (1 976) recorded best performance (302 g/day) for sheep with high energy intake (7.68 MJ ME/day) and medium but adequate protein (82.8 g/CP/day). In another experiment where energy was held constant and protein level was varied, highest daily gain (176 g/day) was attained by sheep on a crude-protein intake of 147.7 g/day and digestible energy intake of 9.9 MJ/day (Adeneye and Oyenuga, 1976). The lowest and highest protein intakes (100.4 g and 151 .4 g CP/day) produced the lowest daily gains (24 g and 54 g/day, respectively). The dressing percentages of 41 .3 to 42.5% on liveweight basis observed in this study compare well with the result (40-50%) reported by Devendra and McLeroy (1982). Increasing level of protein in the diet had little and insignificant influence on dressing percentage. These results are in agreement with the results reported by Kemp et al (1976) with lambs given diets varying in protein level from 10 to 16% CP. The findings are, however, at variance with those of Robinson and Forbes (1970) and Levy et al (1980) who observed an increase in the dressing percentage with increasing level of dietary protein in the supplement concentrate. Dressing percentage of 52.5% based on EBW observed in the present study compares well with the value of 52% reported by Owen and Norman (1977) for indigenous Botswana sheep. The importance of dressing percentage in the tropics, however, remains to be questionable since practically all of the by- products are consumed as food. Other components of the body even sell at higher prices than carcass meat (Mtenga, 1979; Devendra and McLeroy, 1982). The small and insignificant effect of protein level on the proportion of non-carcass components observed in this study is in agreement with past findings at Sokoine University. This was expected because these organs are early-maturing (Palsoon and Verges, 1952; Mtenga, 1979). Gut fat, however, showed a tendency to decrease from 2.7% at low-protein level in treatment A to 1.3% at high-protein level in treatment D. Similar trend has been reported elsewhere by Mtenga (1979). The overall mean percentage carcass composition data obtained in the present study are in line with few published data on sheep in Tanzania (Mtenga and Nyaky, 1985). French (1938), however, reported relatively lower values for lean and bone tissues and higher values for fat tissues for Black Head Persian lambs in Tanzania ranging from 49.9-57.5% and 8.9-11.1%, respectively, but the 383 composition (as percentage lean, bone and fat) observed in the present study is in agreement with the findings reported elsewhere (Owen, 1976). Contradictory results were, however, reported by other workers (Andrew and Orskov, 1970; Fattetetal, 1984; Vipondetal, 1989). In their studies, they reported increase in percentage carcass nitrogen and decrease in ether extract in lambs as the level of dietary protein increased. Factors such as age and weight at slaughter, sex and breed singly or in combination may cause disparity of results between workers. The profit margin per animal for raising animals under treatment B diet was greater (30 TSh per animal) than that of control treatment (maize bran plus hay). The lowest margin of profit was, however, obtained for animals on treatment D (Table 4). At higher levels of energy intake, high level of lablab inclusion may be more profitable. The difference in marginal profit between animals given maize bran alone (treatment A) and those supplemented with lablab bean meal is too small to justify the practice of rearing lambs on lablab bean meal supplementation. The marginal profit, however, may vary from one place to another depending on the price of lablab bean. Table 4. Effect of level of lablab supplementation on economic returns from sheep. Input data: Total concentrate consumed (kg/animal) Cost of concentrate (TSh/kg)1 Total cost of concentrate (TSh/animal) Marginal cost Output data: Carcass weight (kg) Non-carcass component (kg) Revenues from sales of carcass (TSh)2 Treatments + A B C D 30.15 31.68 33.26 33.44 10.00 12.80 15.55 18.30 301.50 405.50 517.20 614.15 - 104.50 215.7 312.65 8.03 8.80 9.23 9.68 1.92 2.23 2.28 2.16 1204.50 1320.00 1384.50 1452.00 Revenues from sales of edible non-carcass3 115.20 133.80 136.80 129.60 Total revenues (TSh/animal) 1319.70 1453.80 1521.30 1581.60 Margin profit (TSh/animal) 1018.21 1048.30 1004.10 967.45 Marginal returns - 30.09 -14.11 -50.76 1. Price of lablab seed meal was 21.10 TSh/kg while price of maize bran was 10.00 TSh/kg. 2. Price of 1 kg of meat was 150 TSh. 3. Price of 1 kg of edible non-carcass was 60 TSh. TSh = Tanzanian shilling. 384 References Adeneye J A and Oyenuga V A. 1976. Energy and protein requirements of West African Dwarf sheep. 2. Increasing the levels of dietary protein to sheep. East African Agricultural and Forestry Journal 42:98-106. Andrew R P and Orskov E R. 1970. The influence of protein concentration and feeding levels on rate of gain in body weight. Journal of Agricultural Science (Cambridge) 75:11-26. AOAC (Association of Official Analytical Chemists). 1975. Official methods of analysis. 12th edition. AOAC, Washington DC, USA. ARC (Agricultural Research Council). 1980. The nutrient requirements of ruminant livestock. Commonwealth Agricultural Bureaux, Slough, UK. 351 pp. Balch C C. 1977. Ruminant digestion and nutritive value: In: Fonnesbeck P V, Harris L E and Kearl L C (eds), Proceedings of the First International Symposium on Feed Composition, Animal Nutrient Requirements and Computerization of Diets. International symposium held in Logan, Utah, 11-16 July 1976. International Feedstuffs Institute, Logan, Utah, USA. pp. 214-218. Cuthbertson A, Harrington G and Smith R J. 1972. Tissue separation to assess beef and lamb variation . Proceedings of the British Society of Animal Production 1 : 1 1 3-1 22. Devendra C and McLeroy G B. 1982. Goaf and sheep production in the tropics. Intermediate Tropical Agricultural Series. Longman Group Ltd, Harlow, Essex, UK. 271 pp. Egan A R. 1977. Nutritional status and intake regulation in sheep. Australian Journal of Agricultural Research 28:907- 915. Fattet I, Hovell F D, Orskov E R, Kyle D J, Pennie K and Smart R I. 1984. The effect of supplementation with protein on protein accretion. British Journal of Nutrition 52:561-574. French M H. 1938. Development of livestock in Tanganyika territory. Tanganyika breeds of sheep and mutton qualities of long tailed type and its crosses with Black Head Persian sheep. Empire Journal of Experimental Agriculture 12:112- 124. Gohl B. 1981. Tropical feeds—feed information summaries and nutritive values.. FAO Animal Production and Health Series 12. FAO (Food and Agricultural Organization of the United Nations), Rome, Italy. 529 pp. Kay M, Bowers H B and McKiddie G. 1968. The protein requirements of rapidly growing steers. Animal Production 10:37-42. Kay M and MacDear mid E. 1973. A note on the effects of changing the concentration of protein in the diet offered to fattening beef cattle. Animal Production 16:205-207. Kemp D J, Johnson A E, Stewart D F, Ely D G and Fox J D. 1976. Effect of dietary protein slaughter weight and sex on carcass composition, orgonoleptic properties and cooking losses of lambs. Journal of Chemical Science 42:575-583. Levy D, Holzer Z, Folman Y and Bleiberg M. 1980. Protein requirements of male cattle fattened on diets differing in energy concentration. Animl Production 30:189-197. Luck P E. 1965. Dolichos lablab available grazing crop. Queensland Agricultural Journal 91:308-309. Mtenga L A. 1979. Meat production from Saanen goats: Growth and development. PhD thesis, University of Reading, Reading, UK. Mtenga L A and Kitaly A J. 1990. Growth performance and carcass characteristics of Tanzania goats fed Chloris gayana hay with different levels of protein supplement. Small Ruminant Research 3:1-8. 385 Mtenga L A and Massae E E. 1988. Effect of underfeeding, refeeding on growth performance and carcass composition of lambs in Tanzania. Proceedings of Tanzania Veterinary Association Scientific Conference 7:1 12-128. Mtenga L A and Nyaky FP. 1985. Effect of supplementation on performance of Black Head Persian and Red Masai lambs. Bulletin ofAnimal Health andProduction in Africa 33(1):43-49. Mtenga LA and Shoo R. 1990. Growth rate, feed intake and feed utilization of Small East African goats supplemented with Leucaena leucocephala. Small Ruminant Research 3:9-18. Owen J B. 1976. Sheep production. Bailliere Trindall, London, UK. 436 pp. Owen J E and Norman G A. 1977. Studies on the meat production characteristics of Botswana goats and sheep. Meat Science 1 .283-306. Palsoon A and Verges J B. 1952. The effects of high and low nutrition at different ages. Journal of Agricultural Science (Cambridge) 42: 1-149. Preston T R. 1965. Protein requirements of growing-finishing cattle and lambs. Journal of Nutrition 90: 1 57-1 60. Robinson J J and Forbes T J. 1970. A study of protein utilisation by weaned lambs. Animal Production 12:95-105. Santos J O. 1976. Grain legumes in Mozambique. Tropical Grain Legume Bulletin 3:13-15. Shoo R A. 1986. A comparative study of roughage utilisation and growth performance between sheep and goats supplemented with different levels of Leucaena leucocephala. MSc thesis, Sokoine University of Agriculture, Morogoro, Tanzania. Steel R G D and Torrie J H. 1980. Principles and procedures of statistics—a biometrical approach. Second edition. McGraw-Hill Book Company, New York, USA. 633 pp. Vipond J E, Margaret E K, Orskov E R and Wetherill G Z. 1989. Effect of fishmeal supplementation on performance over fat lambs fed on barley straw to reduce carcass fatness. Animal Production 48: 1 31-1 38. 386 Experiences in protein supplementary feeding of weaned lambs and goat kids in Tanzania: The issue of dietary energy L.A. Mtenga and A. Madsen Department of Animal Science and Production Sokoine University of Agriculture P O Box 3004, Chuo Kikuu, Morogoro, Tanzania Summary The prevailing idea, implied in previous studies on supplementary feeding ofsmall ruminants in the tropics, has been thatprotein is the major limiting factor for growth. However, the low response to protein supplementation found in several studies at Sokoine University of Agriculture (SUA) cast some doubt on the validity of this idea. This paper reviews past research work at SUA on supplementary feeding of lambs and goat kids. Intake of metabolisable energy and digestible crude protein were calculated from the original data. Growth rate was highly correlated to intake of metabolisable energy (r= 0.85) while the correlation to protein intake was poor (r= 0.67). It is concluded that the energy intake is the major limiting factor for growth of small ruminants under SUA conditions (tropical semi-arid region). It is also concluded that best effects on growth of weaned lambs and kids can be expected from supplements with high energy and moderate protein contents. Etudes sur la complementation proteique de la ration d'agneaux et de chevreaux sevres en Tanzanie: le probleme de I'energie de la ration LA. Mtenga etA. Madsen Resume L'hypothèse dominante, sous-jacente aux etudes effectueesjusque- la sur la complémentation alimentaire des petits ruminants dans les regions tropicales, veut que les prolines constituent le principal 387 facteur limitant pour la croissance des ovins et des caprins. Cependant, compte tenu de la faible réponse à la complémentation protéique enregistrée dans des essais effectués à la Sokoine University of Agriculture (SUA), on est en droit de s'interroger sur le bien-fondé de cette assertion. Cette étude fait le point des travaux effectués jusqu'ici à la SUA sur la complémentation alimentaire des agneaux et des chevreaux. Les quantités ingérées d'énergie métabolisable et de protéines brutes digestibles ont été calculées à partir des données initiales. Les analyses effectuées ont révélé qu'il existait une forte corrélation (r = 0,85) entre le rythme de croissance et la quantité d'énergie métabolisable ingérée, mais une faible corrélation (r = 0,67) entre celui-ci et la consommation de protéines. Ces résultats amènent à conclure que la quantité d'énergie consommée constitue le principal facteur limitant de la croissance des petits ruminants dans les conditions de la SUA (région tropicale semi-aride). Il est également conclu que les meilleures performances de croissance des agneaux et des chevreaux sevrés peuvent être obtenues avec des aliments complémentaires très riches en énergie mais seulement moyennement riches en protéines. Introduction The major part of cattle as well as small ruminants in Tanzania are fed on arid or semi-arid grasslands. That the nutritive value of the pasture drops drastically during the dry season is well documented. The content of crude protein decreases while the content of crude fibre increases resulting in a decreased digestibility of all nutrients and hence a decreased energy content of the grass (Butterworth, 1967; Osbourn, 1976; Gohl, 1981; Butterworth, 1985). The low digestibility of grasses means that the amount taken up by the animal will be reduced due to the decreased rate of passage through the rumen. All together, it means that uptake of both protein and energy from pasture the animal decrease considerably during the dry season. Therefore, supplementary feeding of livestock has been advocated and much research work has been devoted to trials using different compositions and amounts of supplements. The major idea has been that protein is the major limiting factor in the diet, and most trials have used supplements with protein-rich feeds. However, in many cases the response in terms of growth rate has been poor. The present paper reviews the results of a total of eight research projects performed at Sokoine University of Agriculture, Tanzania as MSc thesis or 388 special projects during the years 1981-1989 in search of an answer to the following questions: Is intake of protein or energy the major limiting factor for ruminant growth on tropical grassland? and What types of supplementary feeds are more likely to improve the performance of grazing small ruminants? Methods None of the reviewed papers estimated the animals'intake of energy. For most of the works the animals were kept in metabolic cages and the intakes of dry matter as well as of the individual nutrients were measured. Moreover, the digestibilities of the individual nutrients were determined in the same experiments. The available data therefore allowed us to calculate the intake of total digestible nutrients (TDN) or the intake of digestible organic matter (DOM) from which the intake of digestible energy (DE) or metabolisable energy (ME) was estimated using the following equations given by Devendra and McLeroy (1982): DE(MJ) = 1 8.4 x TDN (kg) DE(MJ) = 19.2xDOM(kg) ME (MJ) = 0.82 x DE (MJ) Hence the metabolisable energies stated in this paper were calculated as follows: 1) ME (MJ) = 0.82 x 18.4 x TDN = 15.1 x TDN (kg) or 2) ME (MJ) = 0.82 x 19.2 x DOM = 15.7 x DOM (kg) Results obtained from using equation (1) or (2) were not significantly different. For some of the works reviewed the animals were grazed, so that the uptake of only the supplemented feed could be measured. For these cases we assumed that the uptake of dry matter from grazing was 0.45 kg/day for animals of 1 5-kg liveweight, and that the grass contained 9 MJ ME and 40 g digestible crude protein (DCP) per kg dry matter (DM). These assumptions are in accordance with the findings of Shoo (1986) and Muhikambele (1990) using grass hay (Table 2). An uptake of 0.45 kg/day for a 1 5-kg animal corresponds to 3% of its liveweight. This agrees with a low-level uptake as stated by Devendra and McLeroy (1982). The estimated intake of metabolisable energy and digestible crude protein (DCP) as well as the growth rates observed in the reviewed works are shown in Tables 1, 2 and 3. Kitalyi (1982) is a typical example of an experiment using 389 Table0.Efrectorsupplem ntreedingongr,thteh'-'dlam sana'dsc nfin dimetabo icc ge . Animal(sex species,br ed7 age,LW) Supplementaryre d Amount Totalintakeperd y Protein-energy ratio Composition (g/d)DM0kgME'MJCP)DCP/MJME Growth rate Sym bol 7ig.0Rer. Malegoatkids LocalTanz., 0'month Malel mb, BMandHP 0-07month 0-0kgLW Castratemale lamb,BHP 0-7month 0 0 0 0 0 0 0 0 0 MB-CSC-Moll-Min (70:0:0:0) MB-CSC-Moll-Min (00:0:0:0) MB-CSCPns -Moll-Min (00:0:N:0:0) MB-CSC-Min (70:07:0) MB-Min,(05:5) MB-LB-Min (00:00:0) MB-LB-Min 00:0:0) 0 200 030 200 0- 00 )0 )0 )0 0.0 0.03 0.7 0.0 0.■ 0.77 0.00 0.7 0.0 0.0 0.0 0.0 0.0 0.0 7.0 0.0 0.9 0.0 0 0 34 41 07 30 00 7 77 0.0 0.0 0.0 0.0 0.0 0.0 70 0- 0.0 0) 0) 0) (X) 0) X 30 0) 30 N+00 00±0 00±7 0±07 0±0 00±0 0±7 0+7 70±7 Kitalyi (0.) Massae (000) 00-0 kgLW 0Metabolisableener'(ME)calcu atedrr m.t kordig sti lerganicm ter(DOM),r omt adig s iblenutrients(TDN asexplainedinthetext. Abbrensatnson*a dnsymbols: RM-edMaasaiMoll-Molas esDM-rymatter BHP-lackHeadPersianMin-ineralm xtureE-Metabo s bleen gy MB-aizebranS7C-unrlowerc kDCP-igestiblecrudeprotei CSC-ottonseedcakLBIablabbeanm lN-um eroranim linexperime t CO8 o - .o E.g> >>Eco -1— 0} •* N ^r CO 5 2 +1 +1 +1 +1 o O ro o> o <3 eg CM eg CO >. nr O) ■5 s c CO i -2 .c o CD 2 8 Co c o O) c "5 cd 3 6 l CM • 55 co o 2 ° CO 0. o Q -5 UJ 2 CD 2 Q c _ eS r E 8 S IIcoc :|8 io r^. C\ CO cd + 1 «- +1 CM co CO °> 3 $ in d 8 d 8 IO d 8 eg CVI in ■o CO eg d CO c a £1 CO o O in d 8 o co O CO CD n, co J? co co -n O co ^ Z jc CO _ '. co CD C0 N CO CD — tL lO CO CO £ c CO cu £ s a x cu CD E CO co CD £ J o T3 C CO CO CD O) «- TJ CD C 13 " CO TJ CD C o 1 E CO a> a CD a> JS O "D "Sc CA COo CO™ a> TJ CO -Q ■£ CD CD o C CO CD•a E -2 9) CD "5 E CL•c 3 a. 8 CO (0 CO o o o CO c CO CO CO -2 CD • c CO 5c C 0 CD CO U o CO a 3 CO | CD CD o o -J £ II c TJ JQ _l c < _l «- * 391 Table3.Efrectorsuppl mentaryreedingongrow hatazl mbsandatkids. Rer. Nyaki (000) Hai (000) Susuma (000) Rg.0 Symbol 0) 0) 0) 30 0) 30 (X) Growth rate (g/d) ) 0 00 07±0 00±30 0±ns 00±7 Protein-energy ratio (g/DCP/MJME) 0.0 0.7 0.0 7.0 7.0 7.0 0- Totalintakeperday1 DCP(g) ' 00 00 00 7 00 00 ME(MJ) 0.0 0.0 0.0 7.0 7.0 7.0 7.0 DG(kg) 0.00 0.0 0.0 0.70 0.70 0.70 0.70 Amount ■reed* (g/d) 0 00 000 0ro 0ro 0ro 0ro Supplementary Composition MB-CSC(07:0) MBICSC(00:7) MB-CSC-Min (00:00:0) Cass-CSC-Min (■:00:0) MB-CSC-(70:00) Cass-CSC(7:05) N 0 0 0 0 0 0 0 Animal(species, sex,breedage, LW) 0month,00kg Malegoatkids 7emalegoatkids Malel mb RMandBHP Tanz.-Norw. crosses, ro0month, Tanz.-Norw. crosses 0-0months, 00-07kg 0-7kg Abbrensatnson: Cass=assava,driedtubeel onot sr rl0. 0.Intakerromgrassesassumedtbe.00kgDMp rd ywithcon entorMJEdCP SM(4ef.hoo,1986andMu ikambele,90insable2) *Indoor7yakialsosupplementedRhod sgrass,w ilHaindS sumaentedG atemalar slibitum. 7orotherabbreviations,s eTable0rootnote. supplements with increasing contents of protein. The results clearly demonstrate that supplementation of 200 g concentrate with only 10% oil cake (11% CP) improved growth rate by two times. No further improvement was obtained by increasing the protein content of the mixture. Nyaki (1981) and Massae (1984) observed about three-fold increase in growth rates by using 400 or 500 g of supplement with moderate protein content. The effect of high-protein allowance to low-energy diets was especially demonstrated by the works of Shoo (1986) and Muhikambele (1990) using supplements of Leucaena leucocephala or cottonseed cake, respectively (Table 2). In both cases, the supplement had no effect on the growth rate. The relationship between growth and intake of energy or protein for all eight studies is illustrated in Figure 1 a and 1 b, respectively. The data show that growth rates are highly correlated with energy intake (r =0.85), while the correlation to protein intake (all points) is poor (r = 0.67). Figure 2 illustrates growth rates in relation to the amount of supplemented concentrate. The correlation is fairly high (r = 0.77). Results and Discussion The relationship between growth rates and intakes of protein and energy was already illustrated by ARC (Agricultural Research Council) in 1980 (Figure 3). The data presented in Figure 1 b are in good agreement with the ARC illustration. Increase of protein intake on a low level of energy intake gives no improved growth. With increasing intakes of energy, growth rate increased and the requirement for protein increased. The encircled points in Figure 1 b show protein intake vs the growth when energy is non-limiting. Under these conditions growth rate is highly correlated to protein intake (r = 0.94). Thestippled lines in Figure 1 a and 1 b show feeding norms for a 15-kg goat kid according to Devendra and McLeory (1982). These feeding norms assume that protein and energy are fed in a ratio of 6 g DCP per MJ ME. Other feeding norms also give similar ratio for protein/energy, eg NRC (1976) stated 6.5 g DCP/MJ ME. The protein/energy ratios of the supplemented diets referred to in the present review (Tables 1 , 2 and 3) in most cases exceed that of feeding norms. Table 4 summarises the data in Tables 1 , 2 and 3, showing average values of intake, protein/energy ratio and growth rates for unsupplemented groups and for groups given concentrates with "moderate" or "high" protein content. The data show that unsupplemented diets have low intakes, low protein/energy ratio and low growth rate. Supplementation increased energy intakes by 50% in average (from 4 to 6 MJ) and increased intakes of digestible protein by 100 or 200 /i. The average effect on growth rates was about 100%, independent of the 393 Figure 1 . Growth rate of lambs and goat kids versus intakes ot energy and protein. Growth (g/day) / 100 o°i/ Y = 16x-34 r- = 0.85 (oand x) 4 5 6 7 Energy intake (MJ ME/day) Growth rate (g/day) 100 10 20 30 40 50 60 70 Protein intake (g DCP/day) 80 Encircled points selected from graph b, as least-protein intake for same growth, i.e. protein limiting for growth. Stippled lines show feeding norms (Devendra and McLeroy, 1982). 394 Figure 2. Growth rate of lambs and goat kids versus daily allowance of concentrate feed. Growth rate (g/day) 100 Y= 0.13 x + 29 r= 0.77 100 200 300 400 Amount of concentrate (g/day) 500 Data from Tables 1, 2 and 3. protein level. The data strongly point to the energy intake as the major limiting factor for growth under the conditions of the studies reviewed. The importance of energy supplementation is emphasised by the findings of Hai (1988) and Susuma (1989) both using cassava chips to increase the energy content of the supplement without changing the protein content (Table 3). The minor increase in energy intake from the supplement can not explain the observed increase of growth rate (about 50%). However, it is most likely that the groups on the cassava based concentrates had a higher uptake of grasses, resulting from a positive effect of cassava starch on the digestibility of the grass. The effect of energy on utilisation of protein is mainly due to the fact that protein synthesis is a very energy demanding process. The synthesis of microbial protein in the rumen requires the presents of digestible carbohydrate to provide 395 Figure 3. The response of growing ruminants to the combined effects of energy and protein (after ARC, 1980). A&e Increasing protein intake energy. The synthesis of proteins in the animal body, for example muscle protein, require the supply of energy from combustion of nutrients like acetate and glucose. If the supply of energy rich nutrients is deficient, then an increased intake of protein will result in an increased transamination and deamination of amino acid. Increased amounts of ammonia are released, converted into urea in the liver and further excrete in the urine (Riis, 1983). This explains that increased protein intake on low energy allowance results in increased excretion of urinary nitrogen (Shoo, 1986; Muhikambele, 1990). Such diets result in increased nitrogen intake and increased nitrogen excretion, but no significant improvement in nitrogen balance or growth. Furthermore, the addition of high protein to low energy diets may have a negative effect as illustrated in Figure 3, lower curve. This is likely due to the energy expense of urea synthesis. 396 CO ! CM !■■ 8 ! eg in r^ +1 to § b ci +1 01 c o Z c E CD — ® a a c "O o> 5 o S o> 2 c Jc * O) o i o. T3 c " CM a> - • c « 01 »= 5S • § I » ^ O w c * *t o> a> I 8. t: O ■D ~ J3 O ,« .O H E £>* 9 w « 0.05) (c) *, **, *** = significant at P < 0.05, P < 0.01 and P < 0.001 respectively. Compared to lambs on treatment L, lambs on treatment H ate significantly more food in total (765 vs 558 g/day; P<0.001) and per unit metabolised weight (87 vs 66 g/kg W0'75; P<0.01) during the first eight weeks of the study (Table 2). This 406 Figure 1 . Liveweight-time relationship as affected by plane of nutrition. Liveweight (kg) 6 7 8 9 Time (weeks) 10 11 12 13 14 15 16 ' Treatment HH ■ Treatment LH I Treatment HL . Treatment LL was expected as lambs on H were given more concentrates. During the last eight weeks of the experiment, there was also a highly significant difference (P <0.001 ) between treatments in hay and total dry-matter consumption. Hay consumption was highest in animals on LL, followed by animals on LH, HL and HH in a decreasing order. DM intake expressed as total g/day and kg metabolic body weight was highest for LH animals, followed by HH, LL and HL in decreasing order (Table 2) and differences between treatments were significant (P<0.001). Feed conversion ratio (DM intake/g liveweight gain) was lowest in animals on treatment LH and highest in animals on treatment HL. 407 When the whole experimental period is considered, animals on treatment HH consumed more dry matter (767 g/day and 83 g/kg W° 75) followed by LH (705 g/day and 77 g/kg W075), HL (600 g/day and 76 g/kg W075) and LL (564 g/day and 74 g/kg W075). However, differences in feed intake per metabolic body size were small and insignificant (P>0.05). On average, animals on LH and HH were superior in feed utilisation (9.7 vs 1 4.6 FRC) compared to animals on LL and HL. Feed utilisation efficiency (reciprocal of FRC) was highest in animals on HH and lowest on animals on LL but the difference in FRC between LH and HH animals was insignificant (P>0.05). Slaughter characteristics Table 3 shows the effect of dietary treatment on slaughter characteristics. Animals on treatment H slaughtered at the end of the first eight weeks weighed and dressed higher (P<0.05) than the animals on treatment L Percentage gut fill, kidney and kidney fat were about equal for animals on both dietary treatments. Gut fat was, however, significantly (P<0.01) higher in lambs on the high plane of nutrition. Table 3. Effect of feeding regime on killing-out characteristics. Per cent of ismpty body weight Liveweight at slaughter (kg) I Kidney and kidney fat End of 8th week Carcass wt Gut fat Gut fill L 15.55" 45.21" 0.55" 35.19 0.36 H 20.68b 51.49b 4.34b 28.38 0.38 F-value 6.18* 7.49' 54.10" 2.62 NS 0.65 NS End of 1 6th week LL 17.33" 44.98" 1.74" 36.24b 0.68" HL 21.18"" 50.03b 4.28c 32.94" 0.60" LH 23.40b 51.60b 13.68b 29.15" 0.76" HH 24.25 53.16b N/A N/A N/A F-value 5.03* 7.85" 8.03" 9.80" 4.00* At the last slaughter weight (16 weeks; Table 3), animals in treatment HH weighed more (P<0.01) at slaughter than the others with the differences between these lambs and those on treatments LH, HL and LL being 0.85, 3.07 and 6.92 kg 408 respectively. The differences in dressing percentage between lambs on treatment LH and those on HL and HH of 1.57 and 1.56 were not significant whereas the difference between lambs on LH and LL of 6.62% was significant (P<0.01). Animals on LL treatment had the lowest proportions of gut fat (P<0.01) and kidney plus kidney fat (P<0.05) and the highest (P<0.01) proportion of gut fill. Carcass composition (physical and chemical) There were significant (P<0.001) differences in carcass weight between treatments (Table 4) and these mainly reflected differences in slaughter weight of these animals. These differences were also reflected in absolute weights of carcass components. Weights of carcass tissues in Table 4 are therefore expressed as percentage of carcass weights to partly correct for carcass weight effects. Table 4. Effect of feeding regime on carcass composition. Per cent carcass wt Ratios Treatment Carcass wt (kg) Lean Fat Bone Lean to fat Lean to bone End of 8th week L 5.14" 67.8 5.0" 25.6" 15.4" 2.7* H 8.12b 64.0 18.1b 17.9b 3.6b 3.6" F-value 18.65*" 3.2 NS 79.4*" 51.8"* 28.2 " 18.1" End of 1 6th week LL 5.49" 62.2" 14.0" 23.8" 5.3 2.9 HL 7.78b 65.5b 13.6b 20.9b 4.9 3.1 LH 9.00b 60.7° 19.2b 20.1b 3.4 3.1 HH 9.65b 57.0d 26.0C 17.0C 2.0 3.4 F-value 14.59*" 13.5* 15.5" 56.6*" 2.7 NS 0.7 NS By the eighth week, animals on treatment H had significantly (P<0.001) higher proportions of fat and lower proportions of bone but the superiority of 3.8% of lean in animals on L was not significant (P>0.05). Lambs on treatment L had 1 1 .8% more lean to fat ratio but 0.9% less lean-to-bone ratio than lambs on treatment H. Chemical composition data in Table 5 show that carcasses of animals on H were superior to those of animals on L in dry matter and ether extract but inferior in crude protein and ash. 409 Table 5. Effect of feeding regime on chemical composition of the carcass of lambs. Chemical composition (%, fresh weight) DM CP EE ASH End of 8th week L 31. 1a 18.6a 4.2a 13.4 H 40.4b 16.8b 9.3b 12.1 F-value 140.1*" 27.8*" 48.1"' 3.8 NS End of 1 6th week LL 36.4a 16.3 15.8a 10.4 HL 39.8a 17.8 18.1b 11.0 LH 41.5" 16.1 19.5b 13.4 HH 45.8b 14.4 27. 1c 11.2 F-value 9.4" 1.86 NS 17.3" 1.02 NS At the end of the 16th week, lambs on treatment HL had the highest percentage of lean meat followed by lambs on LL, LH, and HH (P<0.05) in descending order. Percentage carcass fat in lambs on HH treatment was higher (P<0.01) than that in lambs on LH, LL and HL by 6.8, 12.0 and 12.4 units, respectively, while proportions of bone was highest (P<0.001) in lambs on treatment LL followed by those on treatments HL, LH and HH in decreasing order (Table 4). Treatment effects on lean:fat and lean:bone ratio were small and insignificant (P>0.05). Table 5 shows that the plane of nutrition significantly (P <0.01 ) influenced the dry matter and ether extract composition of the carcasses, with the highest value being observed in animals on HH treatment. Changes in carcass lean and fat Table 6 shows that both the increases in carcass lean and fat were highest in lambs on treatment H during the first 8 weeks. Increase in lean from the 9th to the 16th weeks in lambs on LH was 3.8 times that of animals on HH (31.1 vs8.3 g/day), but lambs on HH were slightly superior in the rate of fat deposition. When the whole experimental period is considered, increase in lean in LH and HH lambs were very similar but fat increase rate was 1.5 times more in lambs on HH. Fat and lean increase rates were also higher in HL compared to LL lambs during this period. 410 Table 6. Effect of feeding regime on change in carcass lean and fat (glday). L H LL HL LH HH Lean Start-8th week 4.1 32.6 .... 9th week-1 6th week - - 0.2 7.5 31.1 8.3 Start-1 6th week - - 2.1 17.6 20.0 20.4 Fat Start-8th week 0.8 9.8 9th week-1 6th week - - 9.8 5.3 14.9 18.3 Start-1 6th week - - 5.3 9.6 13.6 20.3 Efficiency and economic returns High efficiency of lean meat production is always desirable. Efficiency is sometimes defined as the ratio of total lean tissue to total food consumed during the period under evaluation but under practical situations carcass rather than lean weight is used. In this study, efficiency of carcass deposition was 0.02, 0.4, 0.6 and 0.06 kg of carcass per kg of food (hay and concentrate combined) for the LL, HL. Table 7 shows the economics of treatments imposed in this study. The highest profit margin per carcass sold was obtained from animals on LH, followed by those on HH, HL and LL Table 7. Effect of feeding regimes on profitability of lamb production. LL HL LH HH Concentrate consumed (kg/animal) Cost of concentrate consumed (T.shs./animal) Gain in carcass weight (kg/animal) Revenue from carcass gain/animal (T.shs. 150/kg) Marginal profit (T.shs. /carcass) 8.4 26.6 26.6 44.8 58.8 186.2 186.2 313.6 0.94 3.23 4.45 5.10 141.0 484.5 667.5 765.0 (823.5) (1167.0) (1350.0) (1147.5) 82.2 298.3 483.3 451.4 411 Discussion Performance response of animals to low and high plane of nutrition during the first eight weeks of the experiment was as expected and was mainly due to higher concentrate intake, which resulted in increased growth rate, higher slaughter weight, dressing percentage and proportion of carcass fat, and lower proportion of lean, (Owen, 1976). During the growth period between the 9th and 1 6th week, lambs on treatment LH were superior in growth rate (104 g/day) to those on treatment HH (63 g/day) despite the fact that they were on the same plane of nutrition. This is likely due to the phenomenon of compensatory growth (Osbourn and Wilson, 1960; Keenan andMacManus, 1969; Keenan, etal, 1970; Goodchild and Mtenga, 1982). In this period, animals on treatment LH consumed more feed per unit metabolic body weight and utilised it more efficiently than their counterparts on treatment HH. This partly accounts for their superiority in growth rate. In a concurrent study, digestibility coefficients for dry matter, crude protein and crude fibre were 59.1, 59.3, 67.0, 65.0, 71.2, 70.2, 77.6, 76.5%, and 63.5, 60.3, 65.3 and 85.3%, respectively, for animals on treatments LL, HL, LH and HH. Similar observations have been reported by other workers with sheep (Allden and Young, 1 964; Allden, 1970; Graham and Searle, 1979). The reason for lower daily gains for lambs on HH at 9th-16th week period compared to H lambs at 0-8-week period was not clear but possibly the lambs had passed the phase of accelerated growth rate, Orskov et al (1976) observed that when lambs change from high to low dietary regime, they tend to eat less and grow slower than lambs on low plane of nutrition. The results of this study confirm their observation. Winchester and Ellis (1957) commented that the refed animal tends to grow at a rate appropriate to its physiological age rather than its chronological age. The present study confirms this statement in that lambs on low plane of nutrition were physiologically younger as measured by live-weight, and lean and fat contents. On changing from a low to a high plane of nutrition, these animals grew faster and almost caught up with animals maintained continuously on a high plane of nutrition. Dressing percentage is both a yield- and value-determining factor and is therefore an important parameter in assessing performance of meat-producing animals. Lambs on H treatment dressed higher than lambs on L treatment in accordance with findings reviewed by Owen (1976). Mukhtar and El-Hag (1978) found that tropical sheep fed roughage diet only dressed 10% units lower than those fed 3:1 concentrate to roughage ratios. In this study, differences in liveweight, age and level of fatness in the carcass could have accounted for variation in dressing percentage (Berg and Butterfield, 1976; Owen, 1976) which was 35% for L lambs and 28% for H lambs. The percentage gut fill was 36, 32, 29 and 20 for LL, HL, LH and HH lambs, respectively. 412 Berg and Butterfield (1 976) reported that fat is the most labile tissue in the animal's body and can easily be manipulated by nutrition and management. The treatment effects on gut, kidney and carcass fat in the present study can be attributed to energy and protein in the diets of these animals. In accordance with the views of Berg and Butterfield (1976), lambs on a low plane of nutrition were leaner and had less carcass dry matter, fat and energy contents but higher crude protein, bone and ash. It was shown that compensatory growth was partly due to the deposition of more lean and hence more protein and less fat and therefore less energy (Sheehy and Senior, 1942). Recent studies have also supported this hypothesis (Butterfield, 1966; Keenan and McManus, 1969; Drew and Reid, 1975). This hypothesis is also supported in this study (Tables 4, 5 and 6). Some workers have reported a similar body composition for both realimented and continuously grown lambs (Graham and Searle, 1975; 1979; O'Donovan, 1974). Others have reported increased fat and reduced protein percentage in realimented sheep. The phenomenon of compensatory growth has been reviewed in detail by Goodchild and Mtenga (1982). Factors such as age at the onset of restriction, severity of undernutrition, body weight at slaughter, sex and breed may singly or in combination cause disparity in results between workers. Conclusion This study shows clearly that Tanzanian lambs have the ability to compensate in terms of growth if switched from a low to a high plane of nutrition and this seems to be brought about by a number of factors such as increased food intake, greater digestibility of the diets and higher protein deposition. It appears that the merits of supplementation are the prevention of mortality, improved growth rates and the production of better carcasses. The phenomenon of compensatory growth can be exploited through strategic supplementation to produce enough and good quality but relatively cheap meat at a time of relative scarcity. However, more research work is needed on the applicability of this phenomenon in field situation taking into account all factors associated with it, including the availability of supplements. More data are also required on the economics of feeding manipulation in different parts of Tanzania. References Allden W G. 1 968. Undernutrition of the merino sheep and its sequelae. 1 . The growth and development of lambs following periods of nutritional stress. Australian Journal of Agricultural Research 19:621-638. Allden W G. 1970. The effects of nutritional deprivation on subsequent productivity of sheep and cattle. Nutrition Abstracts and Reviews 40: 1 1 67-1 1 84. 413 AlldenWG and Young R S. 1964. The summer nutrition of weaner sheep: Herbage intake following periods of differential nutrition. Australian Journal of Agricultural Research 15:989-1000. AOAC (Association of Official Analytical Chemists). 1975. Official methods of analysis. 12th edition. AOAC, Washington, DC, USA. Berg R T and Butterfield R M. 1976. New concepts of cattle growth. Sydney University Press, Sidney, Australia. 240 pp. Bohman V R. 1955. Compensatory growth of beef cattle. The effect of hay maturity. Journal of Animal Science 14:249-255. Butterfield R M. 1966. The effect of nutritional stress and recovery on the body composition of cattle. Research in Veterinary Science 7:168-179. Drew K and Reid J T. 1975. Compensatory growth in immature sheep. I. The effects of weight loss and realimentation on the whole body composition; II. Some changes in the physical and chemical composition of sheep half-carcass following feed restriction and realimentation; III. Feed utilization by sheep subjected to feed deprivation following by realimentation. Journal of Agricultural Science (Cambridge) 85:193-220. El Hag G A and Mukhtar A M S. 1 978. Varying levels of concentrates in the rations of Sudan Desert sheep. World Review of Animal Production 14(4):73-79. French M H. 1944. Development of livestock in Tanganyika Territory. IV. The development of mutton qualities in Masai and Grade Block Head. Persian sheep. Empire Journal of Experimental Agriculture 12:207-218. Goodchild A V and Mtenga L A. 1982. Compensatory growth in ruminants. Proceedings of the Tanzanian Society of Animal Production 9:176-195. Graham N McC and Searle T W. 1975. Studies of weaner sheep during and after a period of weight stasis. I. Energy and nitrogen utilization. Australian Journal of Agricultural Research 26:343-353. Graham N McC and Searle T W. 1979. Studies of weaned lamb before, during and after a period of weight loss. Australian Journal of Agricultural Research 30:513-531. Keenan D M and McManus W R. 1969. Changes in the body composition and efficiency of mature sheep during loss and regain of live weight. Journal of Agricultural Science (Cambridge) 72:139-147. Keenan D M, McManus W R and Freer M. 1 970. Voluntary intake of food by mature sheep following restricted feeding. Journal ofAgricultural Science (Cambridge) 74:477-485. Mtenga L A and Nyaky F P. 1985. Effect of supplementation on performance of Black Head Persian and Red Masai lambs. Bulletin ofAnimal Health and Production in Africa 33:43-49. O'Donovan W M. 1974. Developmental changes in the bodies of Dorper sheep. III. Effects of different patterns of growth on live body-mass and body composition of weaned Dorper lambs. Rhodesian Journal of Agricultural Research 12:99-111. Orskov E R, McDonald I, Grubb D A and Pennie K. 1976. The nutrition of early weaned lamb. IV. Effects on growth rate, food utilization and body composition of changing from a low to a high protein diet. Journal of Agricultural Science (Cambridge) 86:411-423. Osborne J Band Mendel LB. 1915. The resumption of growth after long continued failure to grow. J. Biol. Chem. 23:439-446. Osbourn O F and Wilson P N. 1960. Effects of restricted food upon growth and development of cockerels. Journal of Agricultural Science (Cambridge) 54:278-289. Owen J B. 1976. Sheep production. Bailliere Trindall, London, UK. 436 pp. 414 Searle T W and Graham N McC. 1975. Studies of wearier sheep during and after a period of weight stasis. II. Body composition. Australian Journal of Agricultural Research 26:355-361. Sheehy E J and Senior E J. 1942. Storing cattle at different levels of nutrition. J. Dept. Agric. Eri. 39:245-254. Speedy A W. 1980. Sheep production: Science into practice. Longman Group Ltd, London, UK. 195 pp. Steel R G D and Torrie J H. 1980. Principles and procedures of a biometrical approach. McGraw-Hill Book Company, New York, USA. 633 pp. Winchester C Fand Ellis N Ft. 1957. Delayed growth ofbeef cattle. Technical Bulletin 1 159. USDA (United States Department of Agriculture), Washington, DC, USA. 415 Lamb production from Stargrass- Silverleaf Desmodium pastures Anna Warambwa Henderson Research Station Private Bag 2004, Mazowe Zimbabwe Summary The possibility of finishing lambs on Stargrass-Silverleaf pasture was investigated. Twenty-five male and female lambs (30 Dorpers and 20 Merinos) were used in the trial. The lambs were weaned at 12 weeks of age and stratified according to breed, sex and mass and randomly assigned to five groups. The first group was slaughtered to give an estimate of the initial carcass mass of the lambs on experiment, the second group consisted of lambs receiving the conventional high energy (10-12 MJ/kg DM): protein finishing diet. The other three treatments were made up of lambs grazing on low (800 kg livemass/ha), medium (1000 kg livemass/ha) medium (100 kg livemass/ha) and high (1200 kg livemass/ha) stocking rates. The lambs fed on the concentrate diet grew significantly (P<0.001) faster (251 gld) than those on low (103 gld), medium (101 gld) and high (87 gld) stocking rates. There were no significant differences in growth rate between lambs on the different stocking rates. Production per hectare in terms of livemass was significantly higher (P < 0. 05) on the high stocking rate (617 kg/ha) than on the low (483 kg/ha) and the medium (595 kglha) stocking rates. Stargrass dry-matter yield was significantly higher (P<0.001) than legume yield. Crude protein content increased whilst crude protein per cent decreased after a month in both grass and legume herbage. The growth rates obtained on the pasture were low and lambs failed to reach slaughter mass. A period of pen-fattening before slaughter would benefit the lambs. 417 Finition d'agneaux sur pâturages de chiendent et de Desmodium uncinatum Anna Warambwa Resumé Cette étude visant à évaluer la possibilité d'entreprendre la finition des agneaux surpâturage de chiendent et de Desmodium uncinatum a été effectuée sur 25 mâles et un nombre égal de femelles (30 Dorper et 20 Mérinos) sevrés à l'âge de 12 semaines. L'échantillon, stratifié fut réparti au hasard au sein de cinq groupes de traitement. Les animaux du premier groupe ont été abattus au départ en vue d'une estimation du poids de carcasse au début de l'essai. Le deuxième groupe fut composé d'animaux recevant la ration classique de finition à forte teneur en énergie (10-12 MJ/kg MS) et en protéine. Enfin, les agneaux des trois autres groupes furent élevés sur pâturage, avec une charge faible (800 kg de poids vif /ha), moyenne (1000 kg de poids vif/ha) ou élevée (1200 kg de poids vif /ha). La croissance des animaux sur concentré fut significativement (P<0,001) plus rapide (251 gljour) que celle de leurs homologues élevés sur pâturage, que la charge soit faible (103 gljour), moyenne (101 gljour) ou élevée (87 gljour). Aucune différence significative n'a été enregistrée entre les taux de croissance de ces trois derniers groupes. La production de viande fut significativement plus élevée avec la forte charge (617 kglha) qu'avec la charge moyenne (595 kglha) ou faible (483 kglha). La production de matière sèche de chiendent fut significativement (P<0,001) plus élevée que celle de la légumineuse. La teneur en protéines brutes de l'une et l'autre espèce végétale s'est accru alors que son taux a commencé à baisser à partir du premier mois. Les taux de croissance sur pâturages furent faibles et le poids atteint par les agneaux à la fin de l'expérience inférieur au poids classique à l'abattage: Une période d'embouche en claustration avant l'abattage s'avère nécessaire. 418 The role of caged layer waste as a nitrogen supplement to fibrous crop residues commonly fed to sheep and goats S.B. Kayongo, MM. Wanyoike, P.N. Mbugua, P.N. Nyagah and T.E. Maitho Faculty of Veterinary Medicine University of Nairobi P O Box 29053, Nairobi Abstract A study was carried out to assess the value of caged layer waste (CLW) as a nitrogen (N) supplement to fibrous crop residues commonly fed to sheep and goats in Kenya. Three crop residues namely maize stover (MS), maize cobs (MC) and wheat straw (WS) were mixed with graded levels of sun-dried CLW from 0, 10, 20, 30, 40, or 50% (as-is basis). The mixtures were incubated for 12, 24, 36, 48, and 72 hours in the rumen of five fistulated Galla goats to study dry-matter degradability (DMD) using the nylon-bag technique. In another study, the effect of method of processing poultry waste on DMD in the reticulo-rumen was investigated. CLW was processed according to the following methods: sun-drying for 4 or 8 days; ensiling for 21 or 42 days; deep-stacking for 21 or 42 days; or fumigation with formaldehyde, four or eight times. The results showed that the DMD of crop residues significantly (P<0.01) increased with increasing level of CLWinclusion up to 40%, beyond which there was no further increase. DMD of all crop residues and at all levels of CLW inclusion increased with increasing duration of incubation in the rumen. Maize stover and maize cobs had a similar (P>0.05) DMD which was higher (P<0.01) than that of wheat straw. The results of the study suggested that poultry waste could be used as a supplemental source of N to fibrous crop residues to increase their degradability in the reticulo-rumen. Among the methods of processing studied, deep-stacking resulted in a higher (P<0.01) DMD than the other methods. Duration of sun-drying or deep-stacking or number of times the CLW was fumigated did not (P>0.05) affectDMD in the rumen. However, ensiling for 42 days gave 419 a higher (P<0.01) DMD of CLW than 21 days. DMD was higher (P<0.01) for longer incubation time irrespective of the processing method. It was concluded that the processing methods used did not adversely affect degradability of CLW in the reticulo-rumen. Utilisation des fientes de volaille comme complément azoté de résidus de récolte fibreux dans l'alimentation d'ovins et de caprins S.B. Kayongo et M. M. Wanyoike Résumé Cette étude a été effectuée pour déterminer la valeur des fientes de pondeuses élevées en batterie comme complément azoté pour des résidus de récolte fibreux. Des fientes séchées au soleil ont été mélangées à trois types de résidus de récolte, à savoir les tiges de maïs, les rafles de maïs et la paille de blé chaque fois à raison de 0, 10, 20, 30, 40 et 50%. Ces mélanges ont été incubés pendant 12, 24, 36, 48 et 72 heures dans le rumen de chèvres fistulées, afin d'en étudier la dégradabilité de la matière sèche par la technique des sachets de nylon. Une seconde étude a porté sur l'effet de divers traitements des fientes sur la dégradabilité de la matière sèche de l'aliment dans le réseau. Les traitements utilisés étaient le séchage au soleil pendant 4 ou 8 jours, l'ensilage pendant 21 ou 42 jours, l'empilage pendant 21 ou 42 jours, et la fumigation au formaldéhyde. Il ressort des résultats obtenus que l'addition de taux de fiente croissants augmente significativement la dégradabilité des résidus de récolte (P<0,01), jusqu'à 40%, seuil au-delà duquel elle s'est stabilisée. Cette dernière a toujours augmenté avec la durée d'incubation, quel que soit le taux de fiente considéré. Les taux de dégradabilité de la paille et des rafles étaient comparables (P > 0, 05), mais supérieurs à celui de la paille de blé (P<0,01). Il ressort de ces résultats que les fientes de volaille pourraient être utilisées comme source d'azote complémentaire pour augmenter la fermentescibilité des résidus de récolte fibreux dans le réseau. En ce qui concerne les traitements, la dégradabilité des aliments traités par empilage était supérieure (P<0,01)à celles obtenues avec tous les autres procédés. La durée du séchage au soleil et de l'empilage ainsi que le nombre de fumigations n'avaient pas d'effet 420 (P>0,05). Par contre, la dégradabilité étaitplus élevée (P<0,01)pour l'ensilage de 42 jours que pour celui de 21 jours, et quel que soit le traitement considéré, augmentait (P<0,01) en outre avec l'allongement de la période d'incubation. Les traitements des fientes de volaille, particulièrement l'empilage, peuvent donc améliorer l'utilisation des résidus fibreux de culture en accroissant leur dégradabilité. 421 Session 5 Small ruminant genetic resources and breeding Ressources et amelioration genetique des petits ruminants Breeding strategies for small ruminant productivity in Africa G.H. Kiwuwa Department of Animal Science, Makerere University, Kampala, Uganda Summary The combined population of small ruminants in sub-Saharan Africa numbering 350 million is approximately 31% of the world total and substantially contributes to human nutrition and the economic requirements of the African people. However, the majority (62%) of the small ruminants (sheep and goats) are raised in arid andsemi-arid zones where the nomadic production systems do not render systematic genetic improvement feasible. However, animals found in the aridIsemi-arid areas have, for a long time, been naturally selected for the harsh conditions and should be the best genotypes for those areas. A relatively smaller (38%) proportion of small ruminant populations are traditionally kept in humid, semi-humid and highland ecozones. Human populations in these areas practice sedentary agriculture and agropastoral production systems. It is in these zones that strategic small ruminant breeding schemes could be applied. However, under smallholder production systems in the agricultural and agropastoral areas, conventional breeding methods are constrained by communal grazing, small flock sizes, single-sire flocks, low levels of literacy and lack of proper recording practices. Under these circumstances breeding strategies will have to be modelled along stratified schemes based on central nucleus elite breeding and multiplier flocks to generate sires for distribution to smallholder farmers. As recording improves, cooperative breeding schemes based on central nucleus herds without concurrent multiplier herds could also be adopted following on-farm performance evaluation for selection of elite females for comprehensive testing in the on-station nucleus herd. The implications of instituting a successful small ruminant breed improvement complemented by on-farm performance testing are also discussed. 423 Stratégies de sélection visant à accroître la productivité des petits ruminants en Afrique G.H. Kiwuwa Résumé L'ensemble des effectifs ovins et caprins de l'Afrique subsaharienne s'élève à 350 millions de têtes et représente environ 31% du cheptel mondial. Ces animaux, qui permettent aux populations africaines de subvenir à une bonne partie de leurs besoins alimentaires et économiques, sont dans leur grande majorité (62%) élevés dans les zones arides et semi-arides du continent, où les systèmes de production pastorale à caractère nomade font obstacle à la mise en oeuvre de programmes systématiques d'amélioration génétique. En revanche, l'hostilité du milieu a opéré une sélection naturelle qui a donné lieu à la constitution de génotypes particulièrement adaptés à la région. Une part relativement moins importante (38%) des populations ovines et caprines est élevée traditionnellement dans les zones humide et semi-humide et dans les hauts plateaux de l'Afrique. De type agricole sédentaire ou agropastoral, les systèmes de production en vigueur dans ces régions autorisent la mise en oeuvre de programmes stratégiques de sélection, encore que l'exploitation en commun des pâturages, la taille réduite des troupeaux, l'insuffisance du nombre de géniteurs au sein des troupeaux, le faible taux d'alphabétisation et l'absence de méthodes adaptées de relevés de données qui caractérisent les systèmes de la petite exploitation agricole rencontrés dans ces zones viennent s'opposer à l'utilisation des méthodes classiques de sélection. Dans ces conditions, les stratégies adoptées en matière de sélection devront se modeler sur des schémas stratifiés s'appuyant non seulement sur des troupeaux de sélection composés d'animaux d'élite mais aussi sur des troupeaux de multiplication destinés à produire des mâles reproducteurs aux fins de diffusion auprès des petits exploitants. Avec l'amélioration des contrôles, des schémas de sélection à l'échelle collective, fondés exclusivement sur des noyaux de sélection, pourront être mis en place. L'évaluation des performances au niveau des exploitations aura pour objectifde sélectionner des mères d'élite. Celles-ci seront intégrées au noyau de sélection de la station en vue d'une évaluation exhaustive de leurs performances. Les conditions indispensables à la réussite d'un programme d'amélioration génétique des petits ruminants et de l'évaluation en exploitation sont également analysées. 424 Introduction The goat, a key actor in the small ruminant production theatre of animal agriculture in Africa was once referred to by the colonialists as "a problem child of African agriculture", causing soil erosion and deforestation and being an absolute nuisance in crop-based farming systems. Not until work carried out in the late 1950s at Serere Research Station in Uganda proved these allegations to the contrary, did the scientific world realise that small ruminants are not villains but have a role to play in the overall livestock production systems on the African continent. This meeting, today, bears witness to one of the major achievements of the last two decades, regarding animal agriculture in Africa. That achievement has been the fact that greater awareness of the role that small ruminants play in subscribing to the animal protein pool for human consumption has been created. Yet there are still numerous constraints against increased small ruminant productivity. Some of these have been spelt out by Adeniji (1989) and include low genetic potential for milk and meat production, inadequate nutrition and poor management practices, prevalence of diseases and reproductive wastage and limited market outlets for small ruminant products. Secondary limitations cover a range of logistic inadequacies including inappropriate technologies, poor extension services and linkages between research and extension, lack of economic incentives and unavailability of inputs for the utilisation of research findings. In this address, privilege has been accorded to the speaker to switch on the state of the art and focus on breeding strategies for small ruminants in Africa. In the process, attempt will be made to highlight some of the critical bottlenecks against genetic improvement through known conventional methods and suggestions will be made in relationship to unconventional approaches that could be pursued for various farming systems in the different ecological zones that characterise the environments of sub-Saharan Africa. State and Focus Sheep and goat populations in Africa have recently been put at about 191 and 159 million heads, respectively. These estimates represent about 30% and 32% of the world's total population of sheep and goats, respectively. Distribution trends by ecozones indicate that the arid areas take a substantial share of the African small ruminants, accounting for 36% of sheep and 39% of goat populations. The semi-arid, the subhumid and the humid zones account for 22%, 14% and 8% of sheep and 27%, 16% and 9% of the goat populations, respectively. 425 It would appear therefore that the majority of small ruminants (58% sheep and 66% goats) are kept in arid and semi-arid zones which predominantly practise pastoral-nomadic system of livestock production. Implications from distribution trends are clear. Breeding strategies for nomadic flocks face contradictions between adaptation and absolute productivity. The type of small ruminant breeds for the harsh environment are determined by the ability to survive prolonged droughts, walk and cover long distances in search of water and forages, and utilise coarse forages for growth, reproduction and lactation (Osinowo and Abubakar, 1989). Breeding strategies for intensive small ruminant production presuppose a sedentary rather than a nomadic system of production. It is under the subhumid and humid ecozones, where communities practice agropastoral and sedentary agricultural farming system, that breed improvement technologies are likely to be successfully adopted. These areas, however, constitute roughly 35% to 40% of the small ruminant population out of which about 10% are on the extreme humid side with the attendant problems arising out of the stressfully humid climate, low nutritive value of fodder and major disease vectors such as ticks, tsetse flies and helminthic parasites. Breeding for heat and disease tolerance would be desirable but could in the process contradict the genetic improvement of the economically productive traits. However, irrespective of the strategic options, it would be highly desirable that breeding plans match small ruminant breed resources to the available feed, management and socio-economic resources in a given ecozone. The bottom line to future approaches in genetic improvement should be to promote the utilisation of the more promising and adapted indigenous breed resources rather than opting for introduction of temperate goat and sheep germplasm that are often ill-adapted to the tropical conditions. The main goal would be to develop simple breeding procedures involving farmers at the grassroot level and be able to register progress towards predictable and sustainable levels of small ruminant meat and milk productivity in agricultural and agropastoral farming systems around Africa. Constraints In designing breeding strategies for small ruminant improvement in Africa one should not overlook a number of constraints that must be overcome before strategies are put in place. A few that come to my mind include: • Lack of adequate data on breed/type characterisation and standardisation of breed names within and between the five regions of Africa (North, West and Central, East and South). It is possible that there could be a lot of duplication of breed names by country or region. Determination of genetic distances among some of the populations may therefore by useful. The developments 426 in genome mapping and biochemical genetics may, in future, be able to provide information on DNA sequences that can be used to measure genetic distances and thereafter narrow the present breeds/strains to a smaller number that can be evolved into a dominant type for a given ecozone. • Much as on-farm breeding data recording schemes would promote participatory research, extension and feedback attributes of a strategic breeding plan, flock mobility under the extensive nomadic systems would pose a major problem. In smallholder systems, small flock sizes coupled with communal herding and single sire flocks also creates problems for isolating sire, farm or flock effects from the true genetic effects. These problems immediately suggest that breeding strategies will have to be modelled along group breeding schemes with on-station nucleus breeding units to provide tested sires to participating farmers. • Small ruminants, particularly goats, tend to perform less satisfactorily under on-station than on-farm environment. Peters (1989) attributes this phenomenon to differences in management practices. On-station production systems follow modern husbandry methods (fencing, routine disease control, restricted breeding, housing etc.). These lead to reduced feed selection possibilities, higher disease risks and restricted mating freedom. Hence, transposing on-station results on breed comparison to field conditions may not necessarily reflect the true life situation on the range or free choice herding systems in villages. On the other hand, on-farm studies deal mainly with non-genetic factors and may not lead to proper assessment of the desired genotypes. • Much as some sheep breeds (Dorper Blackhead Persian, Yankasa, Balami, Uda:Ouda etc.) and goats (West African Dwarf, Sudan Desert, Nubian, Galla, Mubende, Boer, Red Sokoto and Somali) have been reasonably documented, there are other types that have not received sufficient attention to explore their full potential. Low genetic potential among the majority of small ruminants is often assumed and breeding plans to cross the indigenous stocks with exotic germplasm are implemented regardless. This is a constraint in a sense that breeders are pushed by economic forces to adopt germplasm for immediate (short-term) benefits without projecting sustainability consequences. The results have been improvement in productive traits at the expense of fitness traits, which in all, could lead to disaster in the long run. Breeding Strategies Strategic planning for small ruminant breeding must, from the start, therefore, put into consideration a number of factors. These include the production system (i.e. nomadic, sedentary, agropastoral, mixed farming etc.), the priority order of 427 economic traits (i.e. meat, milk, fibre skins etc.), the national flock structure and the socio-economic mix, including level of literacy, product consumption habits, market outlets and type of ownership (i.e. smallholder, pure or mixed ranching). With regard to the production system, the majority of small ruminant populations are located in arid and semi-arid zones of Africa. The owners pursue a nomadic production system which is ecologically imposed but makes full exploitation of arid and semi-arid ranges possible. Nomadism, however, does not render itself to systematic genetic improvement of productive traits other than ensuring and promoting traits for survival. Strategies for nomadic systems will, for some time, be restricted to disease control and regulation of carrying capacities by creating socially acceptable market outlets. When we turn to the prioritisation of traits, and assuming that breeding strategies will be best suited to the agropastoral and smallholder crop production systems, two issues come to the fore: First, with the exception of the milking Nubian-Riverine and the Abyssinian goats, most of the goat and sheep breeds are kept for meat. Those that produce limited amounts of milk (40-80 kg/year) are mainly located in arid and semi-arid environments (i.e. Sudan Desert/Sahel, Maradi/Red Sokoto, Small East African upland semi-arid, Sudan Desert sheep and Ethiopian fat-tailed sheep). Having noted already that breeding efforts should focus on sedentary farming and agropastoral production systems, one would be inclined to exclude milk production as a key trait for genetic improvement. The effort to breed for milk production would involve importation of exotic dairy goat breeds for cross-breeding. The associated costs of importation followed by ill-adaptation of the crossbreds to the ecosystems in the sub-Saharan region cannot be justified. Increases in milk yield from 0.5 kg to 2 kg per day are possible, but would still be less than what could be obtained from similar efforts with cattle. Second, accepting that milk can more effectively be obtained from cattle than from goats or sheep under intensive agro-farming practices, the next consideration is to decide between sheep and goats as meat sources. Sheep can be cheap sources of mutton as they tend to weigh more than goats. However, by the nature of their coat, sheep may not adapt well to the humid climatic zone. Their natural ecosystem is the highlands whose total geographical area is relatively small with extremely high human population densities as compared to the semi-humid and semi-arid areas. Sheep breed, improvement strategies for semi-arid zones may be hampered by the nomadic nature of the production systems in those areas. Furthermore, due to their close-to-the-ground grazing behaviour, sheep kept together with cattle can easily cause soil erosion, particularly in the hilly and 428 mountainous areas of eastern Africa. In the West African region, breeds of sheep such as the Yankasa, Uda and Balami may have a role to play in the agropastoral production systems of that region. Overall, it would appear that for the humid and semi-humid zones breeders should concentrate on goat improvement for meat and skins which offer decisive advantage as goats fit in well with peasant smallholder production systems in those areas. The third issue to consider is that modelling breeding strategies along the conventional methods of progeny testing over a wide range of farm conditions, often supported by recording and artificial insemination services will not work for small ruminants. In most of sub-Saharan Africa, record keeping for meat animals has been difficult to institute due to range production systems where animals stay away from homesteads for some period of time and kidding or lambing are asynchronous with seasons. Furthermore, ownership of small ruminants ownership is in small numbers and is scattered far and wide in the area. Concentrating large numbers of small ruminants for conventional breeding techniques may not be feasible in the near future, as most progressive farmers and agropastoralists would rather invest the high capital outlay into cattle than goats or sheep. Under such circumstances it would be strategic for breeders to pursue procedures that would enhance superior gene flow from nucleus elite herds to the farmers, supported by performance testing under smallholder production systems. In short, and according to Olayiwole and Adu (1989), since conventional selection methods are not feasible for small flocks where recording is inadequate or hard to institute, breeding programmes for such a situation should be carefully planned, easy to operate, sufficiently flexible and should incorporate the utilisation of available genetic resources. Within a given breeding scheme emphasis on meat should supersede consideration for other secondary traits such as milk and fibre (skins being a by-product of meat production improvement efforts). Strategic small ruminant breed improvement could then be approached from three angles: a) screening and selection within the locally adapted breeds for prolificity, faster growth rate, larger body size and conformation. The programme should be carried out in elite nucleus flocks to generate superior sires for distribution to smallholder farmers. b) crossbreeding or inter-breeding between two most promising local breeds that developed in nearly similar ecozones but that can complement each other's attributes. c) blending several local breeds irrespective of their original environment with the hope of evolving a synthetic breed for a given production system 429 Strategic Breeding Schemes Three possible approaches are hereby proposed with the following assumptions: a) Meat is the primary target for small ruminant improvement. b) Body size at maturity is generally categorised into three groups: small (20-30 kg), medium (30-45 kg) and large (45-60 kg). c) Heritability estimates for growth characters are reasonably medium to high (0.30-0.45) and litter size, though it may register low heritability estimate, responds somehow to selection. d) The majority (over 90%) of small ruminant production is still under traditional systems of production but can change gradually to modern systems with improved extension services. Open nucleus elite breeder and multiplier flocks The scheme involves assembling animals through purchases to a central station. The number of animals (does or ewes) should be sufficiently large (300-500) with about one-third to act as control group. High selection pressure is thereafter imposed on the offspring and the dams according to productivity indices for male and female offspring on the one hand and selection of dams to stay in the herd, on the other. The progeny selection index takes into account the reproductive rate of the dam and the offspring's growth rate and weight at weaning, or better still up to one year to avoid maternal effects at weaning stage. Dam selection index is on the basis of fertility and is calculated at the end of each breeding season. It considers the dam's present and previous records in relationship to those of its contemporaries. These indices have been described by Osinowo and Abubakar(1989). The selection pressure would allow the top 10% of the males to be retained in the elite herd for further breeding while the next top 30-40% would be transferred to form the flock sire multiplication scheme after performance testing and the 50-60% would be culled. On the side of the dams, about one-third would be culled remaining during each cycle and replaced by the progenies of the top (10%) performing offspring. The next one-third of the females would be moved to the flock sire multiplier station. The flock sire multiplier station would ensure an increase of the number of males available for distribution to farmers after further screening. The system would ensure improved males from the elite breeder flock entering multiplier flocks, from which greater numbers of improved males are generated for distribution to smallholder village flocks. 430 Open nucleus elite breeder and participatory farmer flocks The procedure under this scheme would be essentially similar to the first except for the starting and ending stages of animal selection. It also does away with the intermediate sire-multiplier flock stage and requires full cooperation between participating farmers and the Elite Nucleus Breeders. The scheme requires some minimum level of literacy on the part of the beneficiaries (farmers) and outright goodwill to release and receive animals between the station and the farm flocks. In essence the proposed scheme is a private cooperative breeders' organisation pivoted around a public government institution that funds the central nucleus breeding flock. The procedure assumes that there has been preliminary on-farm performance appraisals of animals that would enter the nucleus flock. The scheme involves organising groups of interested participants on a provincial or state basis to discuss the concepts, aims and benefits, including the legal and genetic aspects of the scheme. Each member of the cooperative group, assisted by the field staff, contributes the top-performing females to the elite nucleus central station. This is the concept of preliminary screening of the population for elite females based on weight for age and histories on prolificity. The next step would be to decide on the exchange rate for the top females contributed to the central nucleus flock, in return for the highly selected sires to pass back to the cooperative farmers. Perhaps a ratio of five females contributed to one tested sire and two super dams received back from the station could be a useful starting point. Within the central nucleus flock, the top performing females will acquire an elite status as more performance data accumulate. These females must then be bred to the top (1-5%) sires for use within the nucleus flock. Under this scheme faster genetic progress comes from the relatively high selection pressure made possible in the initial on-farm screening operation followed by equally intensive selection and the technology transfer and extension service between the central nucleus station and the farmer-contributor of superior foundation animals. On-farm performance testing Much as intensive selection programmes under centralised elite nucleus flocks generate superior germplasm for distribution to farmers, the final benefits can only be gauged through on-farm performance evaluation. Performance testing identifies and quantifies the nongenetic constraints in order to improve management, disease control and feeding practices. Since we cannot breed beyond the environment the improved animal's genetic abilities should 431 periodically be tested against the prevailing environmental conditions on the farmers' farms. Performance evaluation is a logistical complement to genetic improvement schemes. It diagnoses constraints pertaining to a given production system, provides on-spot checks and analysis of husbandry problems, and generates feedback information to breeders and farmers alike. Very often, the results obtained from central station tests are of little relevance to traditional production systems and may not contribute much towards understanding of the specific adaptation ability of animals to farmers' conditions. One further attraction of on-farm performance testing, is that it provides information on location-specific production conditions that could lead to breed improvement options that are appropriate to the system. Individual animal appraisals and breed comparisons, however, can best be carried out under controlled on-station conditions. Successful implementation of on-farm performance evaluation is, however, likely to be constrained by a number of problems. Peters (1 989) gave a comprehensive outline on the subject and highlighted the following aspects: a) on-farm recording must be based on a large sample of flocks carefully stratified following a baseline survey. b) comprehensive recording of farm covariates is required to explain systematic differences and major reasons for variation in performance. c) where small ruminants are reared on an extensive scale, flock mobility is often a major problem to performance testing. d) in smallholder systems, small flock sizes and communal herding also create problems for isolating farm or flock effects. e) seasonality of production would facilitate accurate recording of reproductive data, but asynchronous production is most common in African systems and demands regular visits and data recording in herds participating in the on-farm performance schemes. f) in order to delineate covariates such as season, disease, age of dam, parity, litter size and sex effects and adjust the data accordingly, one would require a data base spanning over a period not less than two years. g) individual animal identification may face difficulties when herds have mixed ownership and continuity may be disrupted by decisions of the owner to sell or slaughter the animal. However, there are a number of complementarities between on-farm performance evaluation and central nucleus herd stations. According to Peters (1 989), genetic improvement schemes on the latter require a field base if they are 432 to be successful. Furthermore, where nucleus breeding units are integrated with on-farm performance evaluation, immediate improvement through selection of superior foundation animals (see strategy 2) to attain faster and more effective progress can be achieved. A successful on-farm performance evaluation scheme should be concerned with the whole farm environment; cover all components of the farm environment; identify factors limiting production within the system; initiate integrated research to find solutions based on on-station controlled procedures; test breeding or production interventions (on-farm); evaluate the consequences of the interventions; and be instrumental in the design of small ruminant programmes (Peters, 1989). Conclusion Breeding strategies for small ruminant improvement in Africa should focus on meat production characteristics, with skins as a by-product. Multiple births phenomenon should be effectively exploited to provide the maximum meat output per unit of input. Under smallholder production systems of humid and subhumid zones of Africa, conventional breeding methods are constrained by communal grazing, small flock sizes, single-sire flocks, low levels of literacy and poor recording schemes. A partial solution has been suggested to institute elite nucleus breeding stations, screen and test animals (sires) for distribution to farmers. These schemes, however, in the long run should be complemented by simple but systematic on-farm performance evaluation programmes to reinforce and monitor continual genetic improvement. References Adeniji K 0. 1989. Statement on the improvement of small ruminants in West and central Africa. In: Adeniji K O (ed), Improvement of small ruminants. Proceedings of the Workshop on the Improvement of small ruminants in West and central Africa held in Ibadan, Nigeria, 21-25 November 1988. OAU (Organization of African Unity), Nairobi, Kenya. pp. 43-59. Olayiwole M B and Adu I F. 1989. Past and present research on sheep and goat breeding in Nigeria. In: Adeniji K O (ed), Improvement of small ruminants. Proceedings of the Workshop on the Improvement of Small Ruminants in West and Central Africa held in Ibadan, Nigeria, 21-25 November 1988. OAU (Organization of African Unity), Nairobi, Kenya. pp. 61-69. Osinowo O A and Abubakar B Y. 1989. Appropriate breeding strategies for Small Ruminant production in West and Central Africa. In: Adeniji K O (ed), Improvement of small ruminants. Proceedings of the Workshop on the Improvement of Small Ruminants in West and Central Africa held in Ibadan, Nigeria, 21-25 November 1988. OAU (Organization of African Unity), Nairobi, Kenya. pp. 71-64. Peters K J. 1989. Trends in on-farm performance testing of small ruminants in sub-Saharan Africa. In: Wilson R T and Azeb Melaku (eds), African small ruminant 433 research and development. Proceedings of a conference held at Bamenda, Cameroon, 18-25 January 1989. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia, pp. 439-469. 434 The performance, potentials and limitations of the West African Dwarf goat for meat production in the forest belt of Ghana A.K. Tuah\ M.K. Buadu\ F.Y. Obese^ and K. Brew2 'Department of Animal Science, Faculty of Agriculture University of Science and Technology Kumasi, Ghana 2Cocoa Research Institute of Ghana P 0 Box 8, Tafo E/R, Ghana Abstract Production records (1970-1989) for a flock of West African Dwarf (WAD) goats kept on the farm of the University of Science and Technology, Kumasi, in the forest belt of Ghana were analysed. The mean age at first kidding and the mean kidding interval were 543. 18 and 284.26 days, respectively. The month of kidding did not (P<0.05) affect kidding interval. The prolificacy of the does was 185.5%. Type of birth affected (P<0.01) birthweight (1.207, 1.145, 1 .008 and 0.847 kg for singles, twins, triples and quadruplets, respectively) and weaning weight (5.49, 5.30, 4.86 and 4.34 kg for singles, twins, triplets and quadruplets, respectively). Pre-weaning growth rate (39.6,33.1,31.6 and 29.0 glday for singles, twins, triplets and quadruplets, respectively) was not (P<0.05) affected by type of birth. Pre-weaning mortality rate was affected (P<0.01) by type of birth (34.92, 34.78, 36.25 and 55.56% for singles, twins, triplets and quadruplets, respectively). The breed has good qualities such as adaptation to the environment, trypanotolerance, high prolificacy and good kidding interval but growth rate is low. 435 Performances, potentialités et limites de la chèvre naine d'Afrique de l'Ouest pour la production de viande dans la zone forestière au Ghana A.K. Tuah, M.K. Buadu, F.Y. Obese et K. Brew Résumé Des données de production rassemblées de 1970 à 1989 sur un troupeau de chèvres naines d'Afrique de l'Ouest élevé à l'University of Science and Technology de Kumassi, dans la zone forestière du Ghana, ont été analysées. L'âge moyen à la première mise-bas et l'intervalle moyen entre parturitions ont été respectivement de 543, 18 et 284,26 jours. Le mois de la mise-bas n'a pas eu d'effet significatif (P<0,05) sur l'intervalle entre parturitions. Le taux de prolificité des femelles était de 185,5%. Le type de naissance, qui avait une influence significative (P<0,01) sur le poids des chevreaux à la naissance (1,207, 1,145, 1,008 et 0,847 kg respectivement pour les naissances simples, les jumeaux, les triplets et les quadruplets), n'avait en revanche aucun effet significatif (P<0,05) sur le rythme de croissance avant le sevrage (39,6, 33,1, 31,6 et 29,0 gljour respectivementpour les naissances simples, lesjumeaux, les triplets et les quadruplets). Le type de naissance avait un effet significatif (P<0,01) sur la mortalité avant sevrage (34,92, 34,78, 36,25 et 55,56 respectivementpour les naissances simples, lesjumeaux, les triplets et les quadruplets). Dotée de qualités telles que l'adaptation au milieu, la trypanotolérance, une prolificité élevée, un faible intervalle entre parturitions, cette race de chèvres est en revanche limitée par un faible taux de croissance. Introduction Chevon is the meat used most for preparing food in public restaurants (popularly known as "chop bars") in Ghana. The West African Dwarf goat (WAD), the commonest breed in Ghana, is known for its prolificacy and trypanotolerance. Its small size (15-21 kg; Devendra and McLeroy, 1982) makes it suitable for butchering in small communities where there are no cold-storage facilities. It is therefore not surprising that the government intends to encourage farmers to increase the production of goats during the implementation of its medium-term agricultural development programme (Ministry of Agriculture, 1988). The 436 expected increase in the population of goats (1 901 000 in 1987 to 9 172 000 by the year 2000) is greater than the expected increase in the populations of sheep (1 989 000 in 1987 to 6 319 000 by the year 2000) and cattle (1 170 000 in 1987 to 1 853 000 by the year 2000). Since the population of goats is expected to increase about ninefold, it is necessary to assess the performance of the dominant breed in all the ecological zones of the country so as to know its potentialities and limitations. This will help the policy makers to decide whether new genes should be introduced or not. The aim of this paper therefore is to assess the performance of the WAD goat on the farm of the University of Science and Technology, Kumasi, which is located in the humid forest zone of Ghana. Materials and Methods Location, climate and management of the flock The data used in this study were obtained from the production records of the flock of WAD goats belonging to the University of Science and Technology, Kumasi. The study covered the period January 1969 to December 1970. The location, climate and vegetation of the University farm have been described by TuahandBaah (1985). Kiddings were recorded and kids were weighed within 24 h after birth. The kids were ear-tagged for identification and were weaned when they were four months old. The males not needed for breeding were castrated using the burdizzo at the time of weaning. The breeding males ran with the females throughout the year, thus there was no controlled seasonal breeding. Routine drenching against endoparasites was carried out fortnightly. Dipping was also carried out fortnightly to control ectoparasites. The goats were kept under a semi-intensive system throughout the year as described by Tuah and Baah (1985). They were housed in pens at night and allowed to graze during the day. Straw or wood shaving bedding was provided in the pens. The animals grazed mainly Cynodon spp interspersed with Centrosema pubescens from morning up to 1 700 h GMT. They were supplemented with wet brewer's spent grains. Methods of parameter computation and statistical analysis Age at first kidding was computed as the difference (in days) between birth and first kidding dates of does. Kidding interval was calculated as the difference (in days) between two successive kiddings for all does with more than than one kidding record. 437 Prolificacy was calculated as the percentage of kids dropped of does kidding. Percentage incidence of twins, triplets and quadruplets was also calculated as Number of multiple births x100 Total number of births Secondary sex ratio was calculated as the ratio of the total number of male kids dropped to the total number of female kids dropped. The chi-square test was used to test for the significance of the observed ratio from the expected (Snedecor and Cochran, 1967). The effects of type of birth and sex on birthweight, weaning weight and pre-weaning growth rates were analysed separately using the completely randomised design (CRD) and Tukey's Honestly Significant Test (Steel and Torrie, 1980). The effect of month of kidding on kidding interval was analysed using completely randomised design (Steel and Torrie, 1980). Mortality rate was calculated as follows No. dead x100 No. born The effects of type of birth on mortality rate was determined using chi-square with the mortality rate of singles used as expected (Steel and Torrie, 1980). Results and Discussion Age at first kidding and kidding interval is shown in Table 1 . The mean age at first kidding was less than two years (543. 18 ±192.10 days) . Goats on the University farm are more precocious than the sheep kept on the same farm. Tuah and Baah (1985) reported the mean age at first lambing on the same farm to be 638 days with a range of 338 to 1542 days. The mean kidding interval (284.26+100 days) was, however, greater than the mean lambing interval (264 days) reported by Tuah and Baah (1 985) and the mean kidding interval reported for the same breed (258 days) by Vohradsky and Sada (1973) on the Accra Savanna plains. The minimum kidding interval is similar to that reported by Ali et al (1973; 151 days) for Black Bengal goats. The month of kidding did not significantly affect the length of the kidding interval (Table 4; SE 7.33). This is an indirect evidence showing Table 1 . Reproductive performance traits of the West African Dwarf goat. No. of Range (days) Mean (days) Standard deviationrecords Age at first kidding 56 Kidding interval 207 323-1061 152-731 543.18 284.26 192.10 100 438 that on the farm, kidding interval is not affected by changes in quantity and quality of forages which occur during the various seasons of the year as these forages form the staple diet of these animals. It is also possible that the practice of feeding wet brewers' spent grains ameliorated the effects of nutritional stresses, hence the no significant effect of the month of kidding on kidding interval. It must also be noted that in the forest belt the dry seasons are not as severe as in the savanna areas of the country and feed shortage is not very great. The prolificacy of the goats was 1 85.5%. On the same farm, goats were more prolific than the Djallonke sheep (Tuah and Baah, 1985; 130.3%). The incidences of single births and births of twins, triplets and quadruplets were 35.69, 45.61, 16.15 and 2.55%, respectively. It is not desirable to have a prolificacy value of more than 200% as kid mortality is increased with an increase in prolificacy (Table 2). In this country, artificial milk replacers are not fed to the extra kids which cannot get access to teats. In an earlier study covering a ten-year period (1969-1978), Tuah (1989) reported that prolificacy in goats on the same farm was affected (P <0.01) by the age of dam. The values for the various age groups involving 234 kids were 1 24. 1 , 196.8, 182.8, 206.7, 277.8 and 233.3% for does aged one, two, three, four, five and six years, respectively. Table 2. The effect of month of kidding on kidding interval. Month of Kidding interval Standard kidding No. of records (days) Range (days) deviation January 27 257.37±5.47 166-431 67.82 February 9 347.38±9.47 186-510 106.37 March 27 276.19±5.47 163-521 90.31 April 19 252.42±6.52 167-404 73.28 May 14 323.93±7.59 169-731 151.51 June 16 258.25+7.10 167-470 90.02 July 9 291.67±9.47 192-391 75.24 August 14 280.07+7.59 202-435 73.59 September 13 255.85±7.88 152-477 110.91 October 26 245.04±5.57 154-486 70.42 November 12 315.50±8.20 205-715 149.91 December 21 249.76±6.20 196-358 43.70 439 The effects of type of birth on birthweight, weaning weight and pre-weaning growth rate are shown in Table 3. Birthweight was affected (P<0.01) by type of birth. Generally birthweight decreased with increase in litter size. Robinson et al (1977) reported that, for lambs in utero, as the number of foetuses increases, the number of caruncles attached to each foetus decreases thus reducing the feed supply to the foetus and hence the birthweight of the lambs. Pre-weaning growth was not affected by the type of birth perhaps because most of the kids were raised as singles and twins (Table 4). Hunter (1957) observed that each twin lamb suckling its mother obtained 68% of the milk obtained by a single lamb, hence the differences in the pre-weaning growth rates of single-born lambs and twin-born lambs. The overall mean pre-weaning growth rate (32.88/day) falls just outside range reported by Reynolds (1989) for kids kept in Nigeria (17.4 g/day to31.9g/day). Weaning weight was affected (P<0.01 ) by type of birth, decreasing with increase in litter size. Sidwell et al (1962) reported that type of birth had the greatest influence on weaning weight. In an earlier study Tuah (1989) reported that birthweight and weaning weights of kid were not affected by sex, season of birth, age and parity of dam. The mean overall post-weaning (4-12 months) growth rate (24.04/day) was less than the mean overall pre-weaning growth rate (32.8 g/day) but well within the range reported by Reynolds (1989; 14-28.3 g/day) for the same breed in Nigeria. Table 4 contains the pre-weaning mortality rate of kids. Mortality rate increased as litter size increased but only the mortality rate of quadruplets was significantly (P<0.01) different from that of the other birth types. The overall mortality rate of 36.3% was higher than the rate (20.95%) reported by Tuah and Baah (1985) for lambs on the same farm. The rate is, however, similar to the rates reported by Wilson et al (1985; 30%) and Vallerand and Branckaert (1 975; 35%) for unweaned lambs. The major causes of pre-weaning deaths in kids on the same farm according to Tuah (1989) were starvation (24.27%), pneumonia (16.50%), helminthiasis (20.38%), bacterial infection (6.80%), diarrhoea (5.8%) and headwater (4.85%). About 21% of the mortalities occurred during the neonatal period (0-7 days). These causes of mortality can be controlled if proper management practices are instituted. Potentialities and limitations of the breed for meat production The breed is prolific (185.6%) and this trait needs no improvement since mortality rates of kids can be high with prolificacy values exceeding 200%, especially when kids which have no access to teats are not artificially fed in this country. Mortality rates of kids are high (36.34%) but this source of wastage can be reduced with proper management practices. Kidding interval is greater than the ideal of 240 days. It must be observed here that there is no culling of poor producers on the university farm. Perhaps if culling of poor producers is carried out, the kidding 440 CD >. C T3 c2 c'j 2 i 2 5.Q) 2 CD O c to I co a 1 co £ co CO CO 5 o> c .5 O) coi co ». ifco e I* 5 &€ CO oi CO co r- d +i CO co 8 8 ° O) o 8 +i CO £ (O a g> eg C\J 1 C\J COT- Cg O8 £1 — d d c +l +1 (0 s in a> o T 5 cM r~ s eg o d +l co 8 CM + 1 O a> eg d in c\i eg CO o 5. CD CO 8 <3> CO CO I8 d d d d +l +1 +1 +1 8 s CO in in ■* ^r £ CO o d +1 CO d cS a 8 i-. ~ r~ c ~ o o CO c=> t v CO D.CO ■—■ ° £ o £• i to c a> a> a i: it C ■D a; CO CO co CD I a,c 5 CO I .o c CO I o E O) c 5 CO c o -5■*. o CO I eg eg co eg to eg CO 8 to CO CO r.- CO LO CO CO * COCO eg in eg O) CO eg CO o T CO CO CO co 5! 8 eg a, . 3 U) sfi CO c: c J) .c CO O) 0) c c a> S c CO it JO £ CO CD CD T3 c o i o O o CD c 2 o T3 £ o T3 >. .D CDco z CO to CO oco .ns .5 CO 3 T3 c co c o o o o T3 03 h- ;z 5 2 i^ 5 o d v 2^ 441 interval could be shortened. The mean age at first kidding (543.18 days; 18.1 months) is higher than the value reported for Katjang goats in Malaysia (15-16 months) by Devendra (1966). The greatest limitation of the breed is its slow growth rate. There is the need to carry out selection within the breed or to introduce new genes to improve the trait. Introduction of new genes should be done with caution so as not to adversely affect the good characteristics of the breed. The feed conversion efficiency of the breed should also be studied. Acknowledgements The authors are grateful to all the workers on the livestock farm of the University of Science and Technology for keeping the records on the animals. The Head of the Department of Animal Science is thanked for allowing the publication of the data. Ms. Gladys A. Ndziba is thanked for typing the script. References Ali S Z, Hoque M M and Hasnath M A. 1973. A study on the growth and reproductive performance of Black Bengal goats under farm conditions. Indian Veterinary Journal 50:438-440. Devendra C. 1966. Goat breeds of Malaysia. The Malaysian Agricultural Journal 45:268-274. Devendra C and McLeroy G B. 1982. Goaf and sheep production in the tropics. Intermediate Tropical Agricultural Series. Longman Group Ltd, Harlow, Essex, UK. 271 pp. Hunter G L 1957. The maternal influence on size in sheep. Journal ofAgricultural Science 48:36-59. Ministry of Agriculture. 1988. Medium-term agricultural development programme. Working paper 4. Livestock Subsector. Ministry of Agriculture, Accra, Ghana. Reynolds L. 1989. The integration of livestock in alley farming. Paper presented at the inaugural meeting of the Alley Farming Network for Tropical Africa (AFNETA), 1-3 August 1989, IITA, Ibadan, Nigeria. Robinson J J, McDonald I, Fraser C and Crafts R M J. 1977. Studies on reproduction in prolific ewes. I. Growth of the products of conception. Journal ofAgricultural Science (Cambridge) 88:539-552. Sidwell GM, Everson DO and Terrill C E. 1962. Fertility, prolificacy and lamb livability of some pure breeds and their crosses. Journal of Animal Science 21:875-879. Snedecor G W and Cochran W G. 1967. Statistical methods. 6th edition. Iowa State University Press, Ames, Iowa, USA. 593 pp. Steel R G D and Torrie J H. 1980. Principles and procedures of statistics — a biometrical approach. Second edition. McGraw-Hill Book Company, New York, USA. 633 pp. Tuah A K. 1989. The causes of high mortality rates in immature small ruminants management in West and central Africa. In: Adeniji K O (ed), Improvement of small ruminants. Proceedings of the Workshop on the Improvement of Small Ruminants 442 in West and Central Africa held in Ibadan, Nigeria, 21-25 November 1988. OAU (Organization of African Unity), Nairobi, Kenya. pp. 137-148. Tuah A K. 1989. Ruminant livestock production in Ghana. Invited paper presented at the Second Annual General Meeting of the Ghana Society of Animal Production held at Legon, Ghana, 5-6 July, 1989. Tuah A K and Baah J. 1985. Reproductive performance, pre-weaning growth rate and pre-weaning lamb mortality of Djallonke sheep in Ghana Tropical Animal Health and Production 17:107-113. Vallerand Fand Branckaert R. 1975. La race ovine Djallonke au Cameroun. Potentialites, zootechniques, conditions d'elevage, avenir. (West African Dwarf sheep breed in Cameroon. Breeding potentialities, husbandry conditions, future). Revue d'Elevage et de Medecine Veterinaire des Pays Tropicaux 28 (4): 523-545. Vohradsky F and Sada I. 1973. West African goat in Ghana. I. Reproduction and death rate of kids. Sbornik Vyosoke Skolyo Zemedelske V Praze (Czechoslovakia) 6:161-172. Institut Tropickeho a Subtropickeho Zemedelstvi. Wilson R T, Traore A, Peacock C P, Mack S and Agyemang K. 1985. Early mortality of lambs in African traditional livestock production systems. Veterinary Research Communications 9 (4): 295-301 . 443 Comparative performance of improved meat goats in Malya, Tanzania S.M. Das and D.S. Sendalo Zonal Research and Training Centre, Central Zone, Livestock Production Research Institute, P O Box 202, Mpwapwa, Tanzania Abstract Records on liveweight, growth and reproductive performance of Blended (BLD), Blendedx Galla (BG) and Kamoraix indigenous goats for a period of over 10 years (1970-1980) were analysed. For BLD, BG and Kl goat kids, birthweight averaged 2.9, 2.9 and 2. 7 kg (P<0.01), respectively, averaged 13.4, 12.1 and 12.7 kg (P<0.05), and liveweight at 72 weeks of age averaged at 29.8, 27.1 and 29.3 kg (P<0.05), respectively. Sex and birth type had significant effects on all liveweight measures while year of birth did not. Mean litter size for six parities for BLD, BG and Kl does were 1.35, 1.21 and 1.43, respectively. It was concluded that BLD goats with higher liveweights and efficient reproductive performance compared to BG and Kl goats should be used for crossbreeding purposes to improve the meat production from goats in the rural areas. Comparaison des performances de caprins a viande a Malya (Tanzanie) S.M. Das etD.S. Sendalo Resume Des donnees rassemblees pendant 10 ans (1970-1980) sur le poids vif, la croissance et les performances de reproduction de caprins Blended (BLD), Blended x Galla (BG) et kamorai x race locale (KL) ont ete analysées. Le poids a la naissance moyen des BLD, BGetde KL fut respectivement de 2,9, 2,9 et 2,7 kg (P<0,01) pour un poids moyen au sevrage de 13,4, 12,1 et 12,7 kg (P<0,05) et un poids vif de 29,8, 27,1 et 29,3 kg (P<0,05) a I'age de 72 semaines. 445 Contrairement a I'annee de la naissance, le sexe du chevreau et le type de naissance eurent des effets significatifs sur tous ces paramètres. L'age a la première parturition tut respectivement pour desBLD, desBGetdesKL.de 780,3 ±8,5, 766,8±17,3et757,0±24,7 jours . Evalue pour six parturitions, I'intervalle moyen entre mises-bas fut respectivement de 379,6±12,5, 388,2+13,2 et 373,5±11,6 jours, pourune taille de la portee moyenne de 1,35, 1,21 et 1,43 chevreaux. Compte tenu de leur poids vif plus eleve et de leurs meilleures performances de reproduction, les BLD devraient etre preferes aux 8G et aux KL dans les programmes de croisements destines a améliorer la production de viande des caprins en milieu rural. Introduction Small ruminants have been frequently used as a source of meat in both rural and urban areas of Tanzania, although meat from goats is both scarce and expensive. The high demand for goat meat led to the improvement of indigenous goats through crossbreeding with bigger breeds so as to increase the per capita animal protein consumption levels of Tanzanians. A breeding programme to develop a meat goat breed based on indigenous goat was started in the early 1960s. The indigenous goats termed as "Small East African goat" by Devendra and Burns (1983) was crossed with Kamorai and Boer breeds to develop the Blended goat. The present study reports on growth and reproductive performance of Blended goats at the Livestock Research Centre, Malya. Progeny crosses between Blended x Galla and Kamorai x Small East African goat (SEA) genotypes were used for comparative purposes. The results in this study would indicate the benefits of using the meat goat genotypes developed to be used in improving the indigenous goat population. Materials and Methods The study was undertaken at the Livestock Research Centre (LRC), Malya, situated near Lake Victoria at an altitude of 1250 m above sea level. The mean precipitation at the Centre ranges from 640 mm to 1020 mm extending from late October to early May. The natural pasture is dominated by Hyparrhenia spp with isolated patches of Setaria sphacelata, Cynodon dactylon and Chloris gayana. The main browse plant was Acacia tortilis. Records for this study were obtained from LRC, Malya and the National Livestock Registry at Mpwapwa. A total of 4824 records on liveweights and 1258 kidding records covering over a 10-year period (1970-1980) of Blended (BLD), Blended x Galla (BG) and Kamorai x SEA (Kl) were used in the present study. Computation of the genetic constitution of BLD goat showed the following proportions: 446 Kamorai 55%, Boer 30% and indigenous 1 5%. Blended x Galla goats considered in this study were 50% Blended and 50% Galla, while Kl goats were 50% Kamorai and 50% indigenous. The schematic evolution of blended goats, BG and Kl is shown in Figure 1. Figure 1 . Evolution of Blended, Blended x Galla and Kamorai x indigenous goats. t*r& Blended goat 1. Kamorai55%, Boer30%, Indigenous 15% mrh *—h XBlended x Galla 1/2 Blended 1/2 Galla Kamorai x Indigenous Iff? 1/2 Kamorai 1/2 Indigenous 447 Managerial practices of goats at LRC, Malya did not vary during the period of study, whereby kids weaned at 16 weeks of age were subjected to grazing on natural pasture. At the age of 72 weeks, maiden does were mated for the first time, and mating period extending for 50 days (mid-July to early September) was conducted once per annum. The ratio of buck to does during mating was about 1:40. Kids at birth were ear-tagged for identification and weights were recorded at birth, and at ages of 1 6 weeks (weaning) , 24 weeks (post-weaning) , 48 weeks (puberty) and at 72 weeks (maturity). Weights were also recorded at mating and kidding for females and at yearly interval for males. Main diseases of goats encountered at the centre included tapeworm infestation, pneumonia and foot-rot. Routine treatment and disease control measures (deworming and dipping) was regularly done for all goats at the Centre. Traits investigated in this study included birthweight, weaning weight, weights at 24, 48 and 72 weeks of age and growth rate. Reproductive traits investigated included size of litter, percentage of birth with twins and sex ratio of kids born. The liveweight data were analysed by a least-squares method. Results Mean liveweights of kids from birth to 72 weeks of age by genotype, sex and birth-type are shown in Table 1 . Blended kids having 2.9 kg birthweight, 13.4 kg weaning and 29.8 kg at 72 weeks of age were observed to have significantly higher liveweights than BG and Kl kids. At 48 weeks of age Kl kids were significantly heavier than BG kids. Male kids exhibited higher liveweights (P<0.05) than females, while single-born kids were heavier than twins at all ages of weight recording. Growth rates from birth to weaning, 24, 48 and 72 weeks of age are shown in Table 2. Blended kids showed higher growth rates from birth to weaning and from birth to 72 weeks of age. Single-born kids exhibited higher growth than the twins from birth to weaning and to 24 weeks of age, while at higher ages their growth rates were not significantly different. Blended does showed higher incidence multiple births with twinning rate of 40.9%, overall for six parities compared to BG does, whereas Kl does had higher twinning rates compared to both BLD and B does (Table 3). Litter size computed by dividing the total number of kids born by the total number does kidded showed a tendency to increase from first parity to fifth parity and a reduction in sixth parity. Sex ratios for kids bom were observed to be slightly less males than females but were not significantly different across parities and genotypes in the study. 448 Table 1 . Mean liveweights of kids at the Livestock Research Centre, Malya, by genotype, sex and birth-type. Age in weeks Birth 16 24 48 72 Mean liveweight (kg) Overall 2.6 12.7 14.4 19.8 28.7 Breed1: BLD 2.9a 13.4a 14.8a 19.9a 29.8a BG 2.6b 12.1b 14.3b 19.6b 27.2b Kl 2.7b 12.7c 14.5b 20.4a 29.3a Average SE 0.01 0.13 0.17 0.21 0.33 Sex: male 2.7a 13.5a 15.5a 20.9a 30.2a female 2.5b 11.3b 13.3b 19.1a 20.4b Average SE 0.01 0.15 0.19 0.25 0.38 Birth -type Single 2.9a 13.3a 15.6a 20.6a 29.3a Twin 2.4b 11.6b 13.4b 19.4 27.2b Average SE 0.03 0.14 0.19 0.23 0.36 Means within a column of class with different letters differ significantly (P<0.05). 1. BLD = Blended; BG = Blended x Galla; Kl = Kamorai x Small East African. Discussion Liveweights at birth and weaning of BLD, BG and Kl kids in the present study are lower from those reported for BLD kids at Malya and Kongwa by Das (1989). It was also observed that weights at birth, weaning and at 72 weeks age of goat kids in this study were lower than those reported for Boer goats by De Haas (1978), Malawi goats by Ayoade and Butterworth (1982) and Tsana goats (APRU, 1984). Blended goat kids showed higher liveweight and average daily gain compared to BG and Kl kids. These parameters were also higher than that of Black Bengal goats reported by Singh et al (1983) and Black Bengal x Beetal crossbred goats reported by Kanaujia et al (1986). At birth, males were only 0.2 kg heavier than females, which was smaller than the 0.3 kg reported by Kyomo (1978) for Small East African goats. Singles were heavier than twins and grew faster from birth to 24 weeks of age. Similar observations were reported by Nath and Chawla (1978) for Beetal, Alpine and their crossbred kids. The pre-weaning daily gain of BLD kids of 94.3 g/day exceeded that of BG and Kl kids by 6.4 and 12.8, respectively. Similar 449 Table 2. Mean daily gain of goat kids at the Livestock Research Centre, Malya, by genotypes, sex and birth-type. Age in weeks 0-16 0-24 0-48 0-72 Mean daily gain (g) Overall 88.5 71.3 52.8 52.6 Genotype1:BLD 94.3a 72.1 52.3 54.6a BG 88.7b 70.4 52.3 50.2b Kl 82.7b 71.3 53.5 53.4a Average SE 0.83 0.94 0.44 0.37 Sex: male 94.1a 76.0a 53.1a 54.7a female 81.4b 67.1b 51.0b 50.3b Average SE 0.85 1.02 0.47 0.45 Birth-type single 95.4a 76.3a 52.2 53.8 twin 80.5b 65.9b 51.5 51.4 Average SE 0.91 1.06 0.35 0.31 ab = Means within a column of a class with different letters differ significantly (P<0.05). 1. BLD = Blended; BG = Blended x Galla; Kl = Kamorai x Small East African. observations were made by Kassuk (1974) who attributed lower pre-weaning average daily gain of BG and Kl kids to lower milk yields of their dams. Weaning weight would reflect mothering ability of dam as well as the inherent growth potential. Thereafter growth potential would predominate. In this study, males were significantly heavier and grew faster from weaning onwards (P<0.05). Their differences increased from 16 to 72 weeks of age implying that sex effects are more pronounced with age after puberty. Blended and Kl does showed higher multiple births than BG does. The incidence of multiple births tended to increase from first parity to fifth parity. Twinning per cent of does in the present study were lower than that reported for Black Bengal and Beetal x Black Bengal does in India by Kanaujia et al (1986). The sex ratio of kids born are similar to that reported by Sinha and Sahni (1 981) for Indian goats. 450 Table 3. Kidding performance of goats at the Livestock Research Centre, Malya. Parameter Sex ratio Parity n Twinning % Litter size (% males ) Overall BLD1 799 40.9 1.44 49.8 BG 248 34.3 1.39 47.5 Kl 193 43.5 1.46 47.8 I BLD 279 19.4 1.19 52.9 BG 92 7.6 1.08 52.5 Kl 52 36.5 1.41 49.3 Il BLD 204 35.3 1.38 55.3 BG 53 32.1 1.32 48.6 Kl 34 36.5 1.41 49.3 III BLD 155 47.1 1.48 50.7 BG 38 44.7 1.45 43.6 Kl 34 50.0 1.52 48.1 IV BLD 60 41.6 1.45 46.8 BG 27 40.7 1.41 47.4 Kl 19 57.9 1.58 46.7 VBLD 60 56.7 1.57 46.7 BG 27 44.4 1.44 45.5 Kl 18 50.0 1.50 48.5 VI BLD 41 46.3 1.46 46.7 BG 11 36.5 1.35 47.1 Kl 10 46.0 1.40 47.0 n denotes number of does kidded 1. BLD = Blended; BG = Blended x Galla; Kamorai I Small East African The results of this study indicate that BLD goats, given proper management and nutrition, have a potential as a meat-producing goat. The mean age at kidding could be lowered by mating goats immediately after puberty. The superiority of BLD goats over BG for both growth rate and reproductive performance suggests that there is little point in crossing Blended goats with Galla goats. 451 Acknowledgements We wish to thank the International Foundation for Science (IFS) for sponsoring this study. We are also grateful to the Commissioner for Research and Training (MALD) for granting permission to present this paper at the First Small Ruminant Network Conference in Nairobi. References APRU (Animal Production Research Unit). 1984. Goafs, livestock and range research in Botswana 1981-1982, Report. APRU, Botswana, pp. 100-204. Ayoade J A and Butterworth M H. 1982. The relationship between birth and weaning weights in kids of Malawian local goats, Boer and their crosses. Tropical Animal Production 7(2):113-115. Das S M. 1989. Preliminary results on evaluation and breeding of Blended dairy goats in Tanzania. In: Wilson RT and Azeb Melaku (eds), African small ruminant research and development. Proceedings of a conference held at Bamenda, Cameroon, 18-25 January 1989. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 536-545. Devendra C and Burns M. 1983. Goaf production in the tropics. 2nd edition. Commonwealth Agricultural Bureau, Slough, UK. 183 pp. De Haas H J. 1978. Growth of Boer goat crosses in comparison indigenous Small East African goat in Kenya. Der Tropenlandwint 79 (April):7-12. Kanaujia A S, Pander B L, Vinayak A K and Kalra S. 1 986. Seasonal variation in reproductive parameters of does: a note. Indian Journal of Animal Production and Management 2(4): 168-1 70. Kassuk A A. 1974. Development of improved meat goats. Annual Research Report, Livestock Research Centre, Malya, Tanzania, pp. 5-6. Kyomo M L. 1978. Meaf from goats in Tanzania. PhD thesis, University of Dar-es-Salaam, Dar-es-Salaam, Tanzania. 272 pp. Nath I and Chawla D S. 1978. A study of birth weights of Beetal, Alpine and Beetal x exotic crossbred kids. Indian Veterinary Journal 55(4):306-309. Singh C S P, Singh D K, Singh S, Nath S and Mishra R H. 1983. Growth rate in crossbred and purebred Black Bengal goats. Tropical Veterinary and Animal Science Research 1(1):63-68. Sinha N Kand Sahni KL 1981. Secondary sex ratio in goats. Veterinary Research Journal 4(2): 157-1 59. 452 Dairy goat breeding in Malawi: Gestation length, birthweights and growth of the indigenous Malawi goats and their Saanen crosses S.K. Kama andJ.W. Banda Bunda College of Agriculture Box 219, Lilongwe, Malawi Abstract Some Malawi local goats were mated to Saanen bucks and others to local bucks under the same management at the Bunda College farm. Gestation length, birthweights, weaning weight, weight gain per day and weights at one year were observed and compared. The average gestation length of local dams sired by Saanen bucks (146.90+2.68 days) was not significantly different from that of dams sired by local bucks (147.76±3.44 days). Sex of kids and litter sizes did not affect gestation length and although younger dams and dams kidding in the hot wet season had shorter gestation lengths than older dams and those kidding in the cool dry and hot dry seasons, respectively, the differences were not statistically significant. With the exception of twin-born males, birthweights ofSaanen crosses were significantly higher than local kids. Likewise, weaning weights and weights at one year were higher in the Saanen crosses than in the local kids. Weight gains per day were higher before weaning than after and were higher in Saanen crosses than in the local kids. Mortality was 22% in Saanen crosses and 21% in the local kids. 453 Elevage de caprins laitiers au Malawi: durée de la gestation, poids à la naissance et croissance de la race locale et de ses croisements avec des Saanen S.K. Karua etJ.W. Banda Résumé Des chèvres de race locale du Malawi ont été croisées avec des boucs Saanen ou de race locale à la ferme du Bunda College of Agriculture. La durée de la gestation, le poids des chevreaux à la naissance, au sevrage et à l'âge de 1 an et le gain journalier de poids ont été enregistrés et comparés. La durée moyenne de la gestation des femelles saillies par des mâles Saanen (146,90 ±2,68jours) n'étaitpas significativement différente de celle des femelles saillies par des boucs de race locale (147, 76±3,44 jours). Ni le sexe des chevreaux ni la taille de la portée n 'avaient d'effet sur la durée de la gestation. Sans présenter des différences significatives, celle-ci était plus courte pour les chèvres plus jeunes et pour les naissances intervenant au cours de la saison chaude et humide par rapport à la saison fraîche et sèche ou à la saison chaude et sèche, Exception faite des jumeaux mâles, le poids à la naissance des croisements de Saanen était significativement supérieur à celui des chevreaux de race locale. De même, les poids au sevrage et à l'âge de 1 an de ces croisements étaient significativementplus élevés que ceux des chevreaux de race locale. Le gain de poids journalier était plus élevé avant qu'après le sevrage et plus élevé chez les croisements de Saanen que chez les chevreaux de race locale. Le taux de mortalité des premiers était de 22% contre 21% pour les seconds. Introduction The zebu cow has remained the major source of milk in rural areas of Malawi where 89% of the people live. Improved dairy cattle breeds are concentrated in areas surrounding urban centres designated as Milk Shed Areas. Figures on milk consumption in rural areas are not available. However, FAO (1 986) reported that total milk production in Malawi was 40 million kg which gave an estimated per capita consumption of 5 kg. With a population which is 454 increasing at 3.2% per annum but with about 0.95 ha of land per capita (Munthali, 1987), the likelihood of improving this per capita milk consumption is remote. The goat, due to its small size and hence lower feed and capital requirements, can supplement cattle as a source of milk specially for rural areas. Keeping goats is already popular among the smallholder farmers in Malawi. A survey conducted by Banda and Phiri (1990) indicated that goat milk is not discriminated against and is actually preferred in some cases. However, milk production from local goats is low (Mwenefumbo and Phoya, 1982; Banda, 1989). As an effort to improve milk production from goats, crossbreeding Saanen bucks to Malawi local goats commenced on an experimental basis at Bunda College of Agriculture in 1988. So far, observations have been on gestation length, birthweights, weaning weight and growth to one year. Materials and Methods Indigenous (local) Malawi does were mated randomly either to Saanen bucks or to local bucks at the Bunda College farm (latitude 14°35'S, longitude 33°50'E, 1200 m above sea level). The doe were herded on natural pastures together with a vasectomised buck. Any doe noticed on heat was separated and mated to a fertile buck. This provided an opportunity to record accurately the mating day, kidding day and calculate gestation length. Also recorded at kidding was kid weight, litter sizes and sex of kid. All male kids in both groups were castrated at two weeks of age. Age of the dams was determined using dentition (Wilson and Durkin, 1985). All kids together with their dams were fed indoors on Star grass (Cynodon nlemfuensis) and Napier grass (Pennisetum purpureum) for the first one month. Thereafter they were herded on natural pastures together with the rest of the flock. No supplementary feeding was done and the dams were not milked. Mortality was also recorded. The data were analysed using Statistix 2.0 Program on the Unisys Computer available at Bunda College. Results and Discussion Gestation length Effect of breed of sire on gestation length The average gestation length of dams sired by Saanen bucks was not different from that of dams sired by local bucks (Table 1). Although previous data on 455 gestation length of Malawi local goats are not available, that of Saanen has been reported to be 150 days (Peaker, 1978). This seems to indicate that the sire of the litter has no effect on gestation length. Overall gestation length was 147.17±2.58 days and agrees with reports by Wilson (1957) on the Mubende goat in Uganda (146 1/2 days). Table 1 . Gestation length (days) of local dams sired by Saanen and pure local bucks. No. of Gestation According to kiddings length SE Overall 74 147.17 2.58 Breed of sire: Saanen 56 146.90 2.68 Local 18 147.76 3.44 Age of dam: 1 pair (14-18 months) 9 144.89 3.10 2 pairs (19-23 months) 7 145.60 5.17 3 pairs (24-31 months) 35 147.80 2.85 4 pairs (over 32 months) 23 146.95 2.69 Litter size: Single birth 62 147.14 2.86 Multiple births 12 147.09 3.05 Sex of kid: Male 38 146.94 2.79 Female 36 147.34 2.79 Season of kidding: Cool dry season (May to Aug) Hot dry season (Sep to Nov) 26 147.23 3.20 Hot wet season (Dec to Apr) 29 146.46 2.92 Effect of age of dam and season of kidding on gestation length Although dams that kidded in the wet season and were younger had shorter gestation lengths than the others, the observed differences were not statistically significant. However, results reported by Asdell (1929) seem to have shown significant age and seasonal differences. 456 Effect of litter size and sex of kid on gestation length Gestation length in dams kidding single kids (147.14+2.86 days) was about the same as in dams kidding multiple kids (147.09±3.05 days). Similarly, the gestation length in those kidding male kids (146.94±2.7 days) were the same as those kidding female kids (147.34+2.98 days). The results therefore agree with Asdell (1929) that gestation length is not affected by litter size and sex of kid. Birth and mortality of kids With the exception of males born as twins, birthweights of Saanen crosses were significantly higher than those of pure local (Table 2) . Birthweights of males were also higher than that of females in both groups. Mortality rates of kids were 22% in local kids and 21% in Saanen crosses. It is difficult to suggest reasons for these mortality rates since post-mortem examination of dead kids was not conducted. Growth performance As in birthweight, weaning weights and weights at 1 year (364 days) were significantly higher in the Saanen crosses than in pure locals in both male and female kids (Table 3). Weight gains per day were also higher in the Saanen crosses than in the pure locals both before and after weaning. Unlike the Saanen crosses where males had higher weaning weights (1 2.80±5. 1 4 kg) and weight at one year (25.00±3.00 kg) than females (11.96±2.77 kg at weaning and 22.50± 1.78 kg at one year), in the pure local kids, the females had higher weaning weight (8.33±0.7 kg) and weight at one year (16.16+1.55 kg) than males (7.60+1.53 kg at weaning and 14.99± 1.66 kg at one year). Similarly, weight gains per day were higher in males (92.32 gm before and 48.41 gm after weaning) than in females (82.67 gm before 41 .82 gm after weaning) in the Saanen crosses, but higher in females (56 gm before and 31 gm after weaning) than males (50 gm before and 29 gm after weaning) in pure local kids. This could be attributed to the effect of castration. However, similar results have been reported by Kasowanjete et al (1987) in the Malawi goats. Conclusions So far, these observations indicate that Saanen crosses are performing better than pure local kids (higher birthweights, weaning weights, weight gains per day and weight at one year) under the same management. However, this was at a research station and further investigations are planned for village conditions. Similar studies are planned for 3/4 and 1 /4 Saanen under both station and village management before a firm recommendation is made. 457 CO in co t I CD O) -2 "5 o Ic OQ Q. .O c CO 52. ~3 u o ~i x c CD c to CO 00 I I OQ oi e J3 in tD* C\J d O +l ;= +1 in o c CO .O CD CO CO in . c c;f '5 $T3 O) c t fog f E~ CO +1 « — ra «"° O)C0 I >. CO 13 — w O) CD — Q.CD asD 2 5 O) o c — |+l eg r■. co co eg eg O) CD c CO CO go «- eg eg N>»■ co co o co »- O) ^ t i- eg co 8 CO CD inin CO s +1 8 +1 +1 o in O) CO in eg «t CD C\J 8 2 Scb in V £ CO N,— in o in eg ^j c\i +1 +1 +l +1 8 8 o COCOCO eg y r-^ CO 5 Li- 2 458 References Asdell S A. 1929. Variations in the duration of gestation in the goat. Journal of Agriculture Science (Cambridge) 19:382-396. Banda J W. 1989. Milk production potential of goats and sheep in Malawi. Phase I. Genotypic and seasonal influences on milk yield and milk composition. Final Technical Research Report. Contract Research Committee, Ministry of Agriculture, Lilongwe, Malawi. Banda J W and Phiri C D. 1990. Investigations into factors influencing the choice and composition of milk and milk products in Malawi. Journal of Consumer Studies and Home Economics 14. FAO (Food and Agriculture Organization of the United Nations). 1986. 1985 FAO production year-book. Volume 39. FAO Statistics Series 70. FAO, Rome, Italy. 330 pp. Kasowanjete M, Stotz D and Zerfas H S. 1 987. Goaf development programme in Malawi. ILCA Small Ruminant and Camel Group Newsletter 8:15-21. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. Munthali JT. 1987. Cattle fattening on basal diets of maize stover and groundnut tops in Malawi. In: Little D A and Said A N (eds), Utilization of agricultural by-products as livestock feeds in Africa. Proceedings of a workshop held in Blantyre, Malawi, September 1986. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia, pp. 45-52. Mwenefumbo A L and Phoya R K D. 1982. Composition and milk yield of Malawian local goat. Tropical Animal Production 7(1):71 . Peaker M. 1978. Gestation period and litter sizes in the goat. British Veterinary Journal 134:379-383. Wilson P N. 1957. Studies of the browsing and reproductive behaviour of the East African Dwarf goat. East African Agricultural Journal 23(2):138-147. Wilson R T and Durkin J W. 1985. Livestock production in central Mali: Reproductive components in traditionally managed sheep and goats. Livestock Production Science 19(3-^):523-529. 459 Composition and yield of milk from non-dairy goats and sheep in Malawi J W Banda, 1 ,2 J Steinbach2 and H-P Zerfas3 'University of Malawi, Bunda College of Agriculture, Box 219, Lilongwe, Malawi University of Giessen, Tropical Sciences Centre, Division of Animal Production, Ludwigstrasse 21, 6300 Giessen, Germany 3Malawi-Germany Livestock Development Programme (MGLDP), Box 30549, Lilongwe 3, Malawi Abstract Milk yield, milk composition and growth rates of offspring from goats (Iocal-LL, Boer-BB and their crosses-BL) and of sheep (Iocal-LL, Dorper-DD and their crosses-DL) were estimated over two kiddingllambing seasons for 12 weeks of lactation using three techniques of milk-yield estimation, namely kid/lamb suckling, hand-milking and oxytocin + hand-milking. Mean total milk production in the dry and wet seasons was 84.6± 3.1, and 66.8 ± 2.6kg, respectively, for goats; 50.2 ± 1.7 and 49.3 ± 7.9 kg, respectively, for sheep. Only the seasonal difference in goats was significant (P < 0. 00 1) . Estimates of milkyield by suckling and oxytocin techniques were similar. Hand-milking produced 26.8% and 40.5% less milk than the other two techniques in goats and sheep, respectively. The yields of LL (70.6 ±2.5 kg) and BB (75.0 ±3.8 kg) were similar, but significantly lower (P<0.001) than those of BL does (81.6 ±2.3 kg). Local, DL and DD ewes produced, respectively, 37.8±1.8, 48.9 ±1.6 and 62. 7 ±2. 1 kg milk and the differences among these genotypes were significant (P<0.001). Dorper ewes did not perform as well in the rainy season as they did in the dry season. The overall mean levels of total solids, fat, SNF, ash, protein, lactose and energy for goats were 17.4%, 6.8%, 10.6%, 0.88%, 4.5%, 4.7%, and 4.44 MJ/kg, respectively. The respective values for sheep were 18.3%, 6.0%, 12.3%, 0.94%, 5.2%, 4.9% and 4.39 MJ/kg. Within each species, significant differences in compositional values between 461 seasons, among techniques of milk-yield estimation and among genotypes were observed and are discussed. Overall mean birth, weaning weight at 17 weeks and pre-weaning daily liveweight gains for kids were 2.77±0.08kg, 13.0+ 0.41 kg and 87.0+3.5 g, respectively. The values of sheep were, respectively, 3.19± 0.05 kg, 18.5+0.32 kg and 128.9 ±2.5 g. In goats, season of kidding had significant effect on all these traits, but genotype had effect only on kid birthweight. In sheep, season and genotype exerted significant influences in all the lamb traits. Correlations between milk yield and growth rate ofthe young were significantly high andpositive. Regressions of growth rate on milk production for kids were poorer than those for lambs. Composition, production et classement du lait de caprins et d'ovins non laitiers au Malawi J.W. Banda, J. Steinbach et H.-P. Zerfas Résumé Cette étude a été effectuée sur des caprins de race locale (LL), Boer (BB) et leurs croisements (BL) ainsi que sur des ovins de race locale (Lo), Dorper (Dû) et leurs croisements (DLo). La production et la composition du lait ainsi que le rythme de croissance des petits ont été estimés pour deux parturitions et une période totale de 12 semaines. Trois méthodes ont servi à évaluer la production de lait, à savoir la tétée, la traite manuelle et l'association de l'injection d'oxytocine à la traite manuelle. En outre, 206 personnes ont répondu à un questionnaire destiné à déterminer l'accueil fait au lait de chèvres et de brebis et aux produits dérivés par rapport à celui de la vache et aux produits identiques dérivés. La production de lait des chèvres s'élevait à 84,6±3,1 et 66,8±2,6 kg respectivement pour la saison sèche et la saison humide contre 50,2±1,7 et 49,3±1,9 kg respectivement pour les brebis. La saison n'avait d'effet significatif (P<0,001) que sur la production des chèvres. Les valeurs obtenues par la tétée et l'association de l'injection d'oxytocine à la traite manuelle étaient analogues. La traite manuelle permettait d'obtenir plus de lait que les deux autres méthodes, les chiffres étant supérieurs de 26,8 et 40,5% respectivement pour les caprins et les ovins. La production des chèvres de race locale (70,6+2,5 kg) et celle des Boers (75, 0+ 3,8 462 kg) étaient analogues mais significativement (P<0,001) supérieures à celle deleurs croisements (81,6±2,3kg). Quant aux brebis Lo, DLo et DD, leurs productions étaient significativement différentes, s'établissant respectivement à 37,8±1, 8, 48,9±1,6 et 62,7±2,1 kg. La production de saison sèche des brebis Dorper était supérieure à leur production de la saison des pluies. La teneur moyenne en matière sèche, en matière grasse, en matière sèche non graisseuse, en cendres, en protéine, en lactose et en énergie du lait de chèvre était respectivement de 17,4%, 6,8%, 10,6%, 0,88%, 4,5%, 4, 7% et 4,44 MJ/kg contre respectivement 18,3%, 6,0%, 1 2, 3%, 0, 94%, 5,2%, 4,9% et 4,39 MJ/kg pour le lait de brebis. Au sein de chaque espèce, des différences significatives avaient été mises en évidence et décrites, notamment en ce qui concerne la composition suivant la saison, les valeurs relatives aux différentes techniques d'évaluation et les génotypes. Comparés à leurs homologues nés et élevés pendant la saison des pluies, les petits nés et élevés pendant la saison sèche avaient un poids à la naissance plus faible, mais un poids vif à 12 semaines, au sevrage et après le sevrage et un taux de croissance plus élevés (généralement au niveau de 0,001). Les paramètres de la croissance étaient systématiquement meilleurs chez les agneaux que chez les chevreaux. En ce qui concerne la croissance, les différences génotypiques étaient plus prononcées chez les ovins que chez les caprins. Par ailleurs, il existait une corrélation élevée et positive entre la production de lait et la croissance des petits (r = 0,54 et 0,81 respectivement pour les caprins et les ovins). Il ressort des résultats de cette enquête que les principaux obstacles à la consommation du lait de chèvre et de brebis sont le faible niveau de l'offre (P<0,001) suivi du fait que traditionnellement, les populations concernées n'ont pas pour habitude de traire ces espèces animales et de consommer leur lait. En ce qui concerne le test réalisé sur le goût, le lait de chèvre venait en tête, suivi du lait de brebis, le lait de vache venant en dernière position. Introduction Research with non-dairy goats and sheep both in temperate and tropical areas has assessed the variation in the milking ability of dams, and demonstrated that the amount of milk produced by various breeds at various stages of lactation has a strong influence on kid and lamb growth during the pre-weaning period, with 20 to over 60% of the variation in weaning weight accounting for the volume of milk produced (Peart, 1982). 463 Due to influence of milk yield on weaning weights and due to the large variation in milking ability among the wide ranging genotypes available in the tropics, several methods for estimating milk production, namely kid/lamb suckling, hand- milking and hand-milking after oxytocin injection, must be tested and the best selected (Steinbach, 1988). Results reported in literature on the use of suckling and oxytocin methods are very variable. Moreover, in extensive systems of management, season exerts significant indirect effects on the quantity and quality of feedstuffs as well as on the levels of diseases and parasites which, in turn, affect the milk production ability of the animals. For example, Sacker and Trail (1966) reported higher milk yields in goats and sheep during the rainy season while Mittal et al (1977) reported significant seasonal effects in Barbari goats, but not in Jamnapari goats. Malawi has about 1.6 and 0.8 million goats and sheep, respectively, of various genotypes (local, imported and their crosses). The introduction of dairy goats between 1986 and 1987 raised further questions as to the acceptability of their milk. Such information is needed to develop optimal breeding, development and marketing strategies for both meat and milk production programmes. The objectives of this study were: • to describe and evaluate the present and potential milk production capacity of the existing local and imported small ruminant stock (goats and sheep) • to compare different goat and sheep genotypes with respect to the composition of their milk • to assess the potential influence of season on milk production and milk composition of goats and sheep in Malawi • evaluate and compare the acceptability of goat and sheep milk relative to cow milk and • to make recommendations for on-farm improvements, policy changes and draw suggestions for further research. In addition, the growth rate and milk conversion efficiency of kids and lambs were studied. Materials and Methods The study was carried out at Lifidzi Goat Breeding Centre near Salima in Central Malawi. Lifidzi Goat Breeding Centre lies at 13°50' south and 34°30' east with an altitude of 600 meters. The Centre is one of the four farms run by the Malawi-German Livestock Development Programme (MGLDP) established in 1983. 464 In the study on milk yield and milk composition, data were collected from the same 20 goats and 38 sheep in the dry season and in the rainy season. In addition to the animals used in both seasons, 5 goats and 7 sheep were used during the dry season only and 18 goats and 5 sheep during the rainy season only. Local (LL), Boer (BB) and crossbred (BL) goats and local (LL), Dorper (DD) and crossbred (DL) sheep were used. Each genotype was composed of first and second parity animals in the dry season and, second and third parity animals in the rainy season. Milk production of these animals was estimated by kid/lamb suckling, hand milking and oxytocin + hand-milking techniques on 3 different days in each of the 1 2 weeks following parturition. The techniques were alternated every three weeks. Animals were incorporated each week as they kidded or lambed. All these factors were arranged in a 2 x 2 x 3 factorial design. The breeding and management of all the animals are presented in Figure 1. In the dry season, all dams reared kids/lambs of their own genotypes. In the rainy season, LL does reared 50% BL kids, BL does 75% BL kids and BB does BB kids; LL ewes reared 50% DL lambs, DL ewes 75% DL lambs and DD ewes DD lambs. All animals were kept under the same feeding and general management practices. Throughout the trials, the animals grazed natural pastures from 0700 to 1700 hours except for one hour at mid-day. At night they were kept in raised houses fitted with slats. During milking only, animals on the trial received a concentrate mixture composed of 91% maize bran, 8% dried leucaena leaves and 1% common salt (NaCI) at a rate of 300 g per animal/day. Refusals were weighed. The concentrate mixture contained 12.7% crude protein and 19.7 MJ GE (about 13 MJ ME)/kg DM. The initial milk yield for each animal was determined about one week post-partum. Thereafter, milk yields were determined every week until the 12th week of lactation. On the days of yield determination, kids and lambs were separated from their dams for two 4-hour periods: 0700 to 1 1 00 hours and 1 1 1 0 to 1 51 0 hours. For the suckling method, the young were allowed to suckle at 1 1 00 hours and at 1510 hours. They were weighed before and at completion of suckling. The difference in weight was taken as the milk production during the separation period. For the hand-milking technique, animals were milked out before going out for grazing. The amount of milk wao not measured at this time. At 1 100 and 1500 hours, they were milked out by hand for 10 minutes each time and the amount of milk measured. The daily milk production for both methods was calculated by summing up the two 4-hourly yields and multiplying the results by 3. The oxytocin technique (oxytocin + hand-milking) was done by milking the animals after injection of oxytocin at the beginning of each milking. Ten international units or 1 ml of oxytocin was injected intramuscularly into the rear flank. After 3 to 5 minutes animals were milked out rapidly until no more milk could be withdrawn. Only one milk yield determination was carried out for health reasons. The amount of milk obtained during this single separation period was multiplied by 6 to obtain daily milk production. 465 I I E a o £ I ! I < ft El to to i i.3 I 2 E^ J? < 5 o c E co 0) 5 & 05 1 5 CO 0) c O) o CD a CO CD E CO £ 5 Q r= (D E Q- 5 f ™ E i i CO to y g E3 .9 E 2 _co 1 i 2 < 2 co co suiey 466 During the days and times that hand-milking and oxytocin techniques were employed, 50 to 100 ml of milk samples were obtained. The samples were immediately frozen at about -20°C and stored for analysis in order to determine the chemical composition of the milk. Liveweights of dams and the young were taken at parturition and at weekly intervals until the young were weaned at 1 7 weeks of age. At the time of weighing, condition scores of the dams were done. The scale used for the score was that described by Russel et al (1969) except that for goats the subcutaneous fat deposition component was ignored. All the milk samples collected were analysed for fat by the Babcock method, for protein (N x 6.38) by the Kjeldahl method and for total solids and ash by methods described by AOAC (1980). Lactose was determined by Nelson's colorimetric determination of sugars. Solids-not-fat (SNF) concentration was derived from total solids and fat concentrations by difference. Energy values were calculated using equations developed by Economides (1986) for sheep and goats as follows: a. Goats: Y= 1.64 + 0.42X, b. Sheep: Y= 1.94 + 0.43X, where Y is the calorific value of milk in MJ/kg and X, is the per cent fat. Dependent variables were milk yields, milk composition, liveweights and growth rate. Fixed least squares models described by Harvey (1977) for use on data with unequal subclass numbers were used. They included fixed effects of season, technique of milk-yield estimation, genotypes and their first order interactions. The effects of parity and the number of kids/lambs suckled were included as regressions. The model for the analysis of milk composition included in addition, the effect of lactation stage. Results Milk production and lactation pattern Analysis of variance and tests of significance for total lactation yields in goats and sheep are given in Table 1. In goats, season of kidding, technique of milk-yield estimation and genotype exerted significant effects on the total 12-week lactation yield. None of the interaction effects was significant. Parity of the dam had a quadratic relationship to the total lactation yield. The linear effect of the number of kids suckled was significant. In sheep, technique of milk yield 467 § m If) o a>CO CM dN d CO CM 2 ,ns o CD CM N CO O) s T- CD CO co o in in o o 0) o LO CM ^J en k d CO CO t— O) s CM CO o r^ K * CM o CO CO CM en 1O CO N a) '■" T a> T~ CD CO CO 00 O in o CM d CM 00 t- «" ^~ CO CM ■* CM CO CO CD CO CM CO CM l-r CO CM N 00* CO CM CM CD in 8CM ■or in d !nCO in CO CO CO N CO 1^ in CM ■>t d cb CD CD i- * s. CD 00 CM in CO o c CD o o ~ & 5 CO t O if o CO c o 5 T3 CO ■D co 0) c 3 JC in CX cj >. >. 3 ." CO O) 'c 0CI ^ « o s o o c 0ra E : = cr 468 Figure 2. The relationship between milk production (MP) and litter size suckled (L) and parity (P) of the dam in goats. Milk yield (kg) 100 I Singles L. Twins MP = 16.78 + 13.86L + 32.67P - 6.07P2 Figure 3. The influence ofgenotype, technique of estimation and season on total 1 2 week milk yield in sheep. Milk yield (kg) f J LL sheep 80 DL sheep DD sheep OHMTHMT Genotype ST = suckling; HMT - hand-milking OHMT = oxytooin - hand-milking Dry season Rainy season Season 469 estimation, genotype, genotype by technique and genotype by season interaction effects significantly influenced total lactation yield. Parity had only a quadratic effect on milk yield. The linear effect of litter size suckled was highly significant. The least squares means and standard errors for total milk yields in both goats and sheep are presented in Table 2. In goats, milk production of does kidding in the dry season was significantly higher (P<0.001) than that of those kidding in the rainy season, irrespective of genotype or technique of milk-yield estimation. Estimates of milk yield obtained by the suckling technique and the oxytocin technique were similar. The quantity of milk harvested by hand-milking was 36.5% less (P<0.001) than that harvested by the other two techniques. Boer x local crossbred goats showed higher (P<0.001) 12-week milk production over the local and Boer goats which gave similar milk-yield estimates. The regression of the total 1 2-week milk production on parity was 1 2.83 and -5. 1 5 for the linear and quadratic terms, respectively. Total milk production increased by 13.64 kg with the increase in number of kids suckled. Coefficients of variation obtained from within-genotype analysis of variance were 24.5%, 118.9% and 20.9% for local, Boer x local crossbred and for boer goats, respectively. In sheep, mean total milk yields were similar. The overall difference in 12-week milk production estimated by the lamb-suckling or oxytocin methods (58.6 or 56.6 kg, respectively) was not significant. Over the whole lactation period, hand-milking estimates were 40.5% less (P<0.001) than those of either the lamb-suckling or the oxytocin methods. Dorper ewes produced the largest quantities of milk and local ewes produced the least. Dorper x local crossbred ewes produced total milk yields which were intermediate between the two other genotypes. Differences among these genotypes were significant (P<0.001). Milk production performance among genotypes differed according to the technique milk yield estimation employed (P<0.05) and according to season of lambing (P<0.001). The local ewes produced 47.6% and 30.9% higher milk yield by the suckling method than by the hand-milking and oxytocin methods, respectively. Dorper x local crossbred and Dorper ewes on the other hand gave higher milk yields by the oxytocin method than by the hand milking and suckling methods. Dorper ewes produced 1 5% less milk in the rainy season than in the dry season, but the other two genotypes produced slightly more milk in the rainy season tan in the dry season. The regression equation of total milk production (MP) on litter size (L) and parity (P) was MP = 21.31 + 16.33L + 16.33 L + 12.84P - 3.38P2 The effect of litter size suckled and parity on total milk yield are presented in Figure 4. The average daily milk yields and lactation pattern for goats and sheep are presented in Figure 5. The average daily milk yield in goats increased to a maximum of 1076 g at two weeks post-partum and then decreased steadily till the end of lactation at 12 weeks when production averaged 962 g/day. In sheep, 470 Table0.Leastsquaresmeansndstand derro s(kg)rotot lmi ky elg tsh p. 7 0.30 0.0 0.0 0.0 0.0 0.0 0.00 0.00 0.0 Goats Mean 0.7 0.. 0.0a ..0b 0.0a 0.0b 70.0a 00.. 0.0a 7o. 300 70 0 0- 0- ■ 00 0 22 Maineifect Subclass Overall Season Dry Rainy Technique Suckling Hand-milking Oxytocin Genotype Local Crosses Exotic 7o. Sheep Mean 7 000 70 00 ■.0 00.0 ■.00 0.00 0.70 0.00 0.00 0.0 0.0 0.0 0.0 2.0 00.0b 0.00a ■.0b 07.0a 0.0b ns.70c . 0 0 0 00 ns Meanswithinvariablegroupsb ringd ife entl tersf rsigni icantly(P<0. o).Thw thoule t rs0 0n twyer nce theanalysisorvariance. -J Figure 4. The relationship between milk production (MP) and litter size suckled (L) and parity (P) ot the dam in sheep. Milk yield (kg) 70-i 60 50 40 30 20 10 0 Singles i Twins MP = 21.31 + 16.33L + 12.84P - 3.38P2 Figure 5. Overall lactation curves of sheep and goats. Av. milk yield (g/day) 1100t v 900 800 ■ Goats ' Sheep ,-A. \ \ 6 7 8 Lactation week 10 11 12 472 on the other hand, the average daily milk yield was practically linear in decline, maximum production of 745 g occurring in the first week, decreasing throughout lactation until 12 weeks of lactation when production was 508 g/day. Milk composition The composition of goats and sheep is shown in Table 3 and Figure 6. Local and BB goats had essentially the same milk composition in terms of total solids, butterfat, lactose and energy and these values except those for lactose, were significantly higher (P<0.001) than those for the Boer x local crossbred goats. There were no significant differences between BL and BB goats for SNF but SNF was significantly higher (P<0.001) in LL goats. Ash content was highest in BL goats and lowest in BB goats. Protein content was highest in LL goats and lowest in BB goats. Local sheep produced milk with the lowest total solids (P<0.05), butterfat (P<0.001) and energy (P<0.001) contents, but had the highest levels of SNF (P<0.001), ash (P<0.01), protein (P<0.001) and lactose (P<0.001). The differences between DL and DD ewes in total solids, butterfat, SNF, ash and energy were not significant. However, the contents of protein and lactose in DD ewes were least and lower (P<0.001) than those in DL ewes whose values were intermediate. Live weights and growth rates of kids and lambs Mean liveweight and liveweight gains of kids from birth to weaning at 17 weeks are given in Table 4. Although kids born during the dry season were lighter (P) at birth than those born during the rainy season, they weighed more at weaning (P) and gained more (P<0.001) weight until statistically significant differences among dam genotypes were detected only for kid birth-weights. Kids born to LL does were lightest and those born to BB does were heaviest (P). Those born to BL does were intermediate. However, between birth and weaning, no statistically significant differences among dam genotypes for kid growth traits were observed. Least squares means and standard errors of lamb live weights and liveweight gains from birth to weaning are shown in Table 5. Although the mean birthweight of lambs born in the dry season was lower (P) than that of those born in the rainy season, the performance of the former group was significantly superior to that of the latter group until weaning. Performance of lambs born to and reared by LL ewes was the least (P<0.001) in all growth traits. Weaning weights and liveweight gains were highest (P<0.001) in lambs born to and reared by DD ewes. The performance of those born to and reared by DL ewes was mostly intermediate However, only the difference in birthweight between DL and DD ewes was 473 Table0.contents(%)ot alsolids,butterratapro einigmi k. 7 0.7 0.7 0.00 0.17 0.17 0.17 0.17 0.ro 0.ro 0.7 0.7 Protein Mean 0.0 0.0b 0.00a 0.77b 0.0a 0.07a ..0a 0.00a 0.0a 0.0. 0.0a 7 0.00 0.00 0.00 0.0 0.0 0.0 0.0 0.0 0.0 0.07 0.07 7at Mean 0.70 0.70 0.00 0.00a 0.00ab 0.07b 0.00b 7.00c 7.70c 0.0a 7.07b 7 0.07 0.0 0.00 0.00 0.' 0.' 0.' 0.00 0.00 0.00 0.00 Totals lids Mean 07.0 '.0b 30.7a 07.. 07.0a 07.0a 30.0a 07.0bc 00.0c 07.0a 07.7b 7o. 000 00 00 07 07 00 00 00 00 00 00 Maineifect Lactationweek Technique Han0-milking Subclass Overall Season Rainy 1-0 00 00 7ns 0-0 000 Oxytocin Dry Meanswithinvariablegroupsb ring0iife entletterd if rsigni ica tly(P<0.17).Thosew thoutsn td . 2 Figure 6. Changes in the composition of milk from sheep and goats at two weekly intervals. Total solids (% 20 . % Ash 0.96- 0.94- 0.92 0.90- 0.88 0.86-1 0.84- 0.82-1 0.80 --. y / S % Protein 5.6-i 5.4- 5.2 5.0 4.8 4.6 4.4 4.2 4.0 % Lactose 5.4- Energy (MJ/kg) 5.5- 475 Figure 7. Changes in liveweight and change condition scoresol sheep and goats during the experimental period. Liveweight (kg) 36- 35 34 33- 32 31 30 • • Goats • • Sheep -\ r ~ i i i 5 6 7 8 Lactation week 10 12 significant (P<0.05). The differences between other performance traits between the two genotypes were not significant. Changes in liveweight and body condition scores of goats and sheep during lactation are shown in Figure 7. In goats, the overall weight (±SE) was 34.7+0.25 kg and the mean body condition score was 2.87+0.05. In sheep, overall mean weight was 31.1 ±0.16 and mean body condition score was 2 70±0.04. 476 Table0.hecontent(%)ort alsolids,r tapro inihe pmi k. 7 0.7 0.7 0.7 0.0. 0.00 0.7 0.00 0.00 0.00 0.7 0.7 Protein Mean 0.ro 0.00b 0.09a 0.0a 0.00a 0.0a 0.0a 0.70b 0.0b 0.07b 0.0a 7 0.17 0.07 0.07 0.7 0.7 0.7 0.7 0.7 0.7 0.00 0.00 7at Mean 0.7 0.7b 0.0a 0.00a 0.30a 0.0b 0.00c 7.00d 7.0e 0.0a 0.70b 7 0.00 0.00 0.00 0.0 0.0 0.0 0.30 0.30 0.30 0.07 0.07 Totalsolids Mean '.00 0.30b 07.70a 07.0a 07.00a 07.0b 00.00c 0.0d 0.30e 07.0a '.0b No 070 000 300 070 070 070 307 300 000 00 07 Maineifect Lactationweek Technique Hand-milking Subclass Overall Season Rainy 0ro 0-4 00 7ns 0-0 000 Oxytocin Dry Meanswithinvariablegroupsb r ngdi fe entletterf rsign rica tly(P<0.17).Thosew thoutle rn tdi . LU CO o o l2 t)v c CO o CD s 2 01 to to CD c a iff c 111 *w co a CD 10 0) cd -c CO CO CO■Q o c O CO c i2 CO to CD o 5 O) .c e o 9o a LU5 o co o :* 1 «oo CD _l x: o c CO 2 o CD 2 c w CD *»- CO■o cD « LU■o CO c to to .cOc aCO 6CDE CO c CD CO t_ CO 3 CD Cr 5 CO CO CO CD -J «' CD -O ■ K t in o Q Q O. £ P QO CO CO O CM CO O o o »- o o o o o d d o d d d d d d ■O .a CO O LO CO LO CO CO CO -Q LOo mCO CD d d t1 LO f~ CO «* w O o o o o o o o o o CO o o o CO 8 CMO CO CO CO 8 8 C\J CO COo o o o o O o o o b o d o d d d d d o d o d o 8 o d .o. n -Q -Q o «? O CO r^ a> N CD o o d O O T O O N. CO Q CM CO o o o o o d d d d d CO CO CO O O CO o o SCO Q CM CM © O O O d d d d d o CO CO CO d d to «- to o c CDCO CO to CD O o CO Li. a o z s. 3U- CO >. en 2>* n o CO CO CO CO CM CO LO CO -O O LO o o CO CO CM o LO T ^r n n CO CO .O n -Q 8 8 o •f LO CO LO CO o> CM a> in CO to CM d LO «* 'T LO o d V ex >. o o o t o o 8 cO) r- r^ 8 CM CO COo o o O o o o d o o d O d o .E '5 CD 3= CO CO n -Q O o CO T3 o t o LO ■c»■ CM CM toLO CO O) CO i— CO CO LO CM d LO ■6 t CD o o o 8 LO LOo o d d d CM O to a. o Q) CO CO o eg 8 CM 8 o d o o CM CM o o d d m z co c CD CD O O O ct 3 5= 8 O) CO c * CO CD n c/> Q. =1 o CO a> LO CD -j f J3 M CO ^ CO O > c c CO !c T3 O) I OCO CDS to il C ^> CO U- LU CD z 2 CO 478 Discussion In the present study, goats produced significantly less milk in the rainy season than in the dry season which disagrees with the observations of Mukundan and Bhat (1983). During the year when rainy season trials were set up, rainfall had increased from 976 mm to over 1 300 mm which was obtained from October 1 988 to early April 1989. These much longer and intensive rains disturbed the normal grazing period and pattern. In addition, the level of worm infestation observed in faeces and the higher level of disease incidences during this season could have accounted for the productive performance. The non-significant difference in estimates of total and daily milk yields between suckling and oxytocin techniques throughout lactation does not agree with the results of Mill and Steinbach (1984) who obtained higher estimates when oxytocin technique rather than when suckling technique was used. Hand-milking produced 36.5% less milk than either of the other two methods. Similar findings have been reported by Ueckermann et al (1974). These findings and the present results might show that all the breeds used in the present study have never been selected for milk production and indicate very strong mothering instinct. Differences in dam genotypes in total and daily milk production were evident. Previous estimates of milk production of the three goat genotypes are limited. Levels attained by the LL does are much higher than those reported for other tropical/subtropical breeds (Devendra and Burns, 1983). The breed is, however, surpassed by far by the Beetal and Jamnapari of India and the Damascus of Cyprus. The estimates of Boer goats are very much lower than those observed on the same breed elsewhere in Africa (Ueckermann et al, 1974). The present milk-yield estimates of BL does are 1 12% of the mid-parent average, 116% of the LL goats and 109% of BB goats. This might suggest that crossbreeding the BB and the LL goats could be beneficial for increased milk production. The total milk yields of sheep in the dry rainy seasons were similar. Similar findings were reported by Aboul-Naga et al (1981a). Much earlier, Sacker and Trail (1966) observed that the growth rate of lambs of the East African Blackheaded sheep lambing in the dry season was restricted. Total milk intake by lambs (suckling technique) was similar to total milk production of the ewes (oxytocin + hand-milking). The oxytocin + hand-milking technique usually gave no second milk let-down or residual milk in the udder as confirmed by palpation. This might suggest that either the ewes had adjusted their daily milk secretion to equal that consumed by their lambs or that all the lambs were consuming all the milk the ewes were capable of producing as suggested by Wohlt et al (1984). Similar findings were reported by Doney et al (1979) and Aboul-Naga et al (1981b). 479 Although previous estimates on the sheep breeds used in the present study are not available, the levels attained by LL ewes are similar to those of Malpura and Chokla in India, higher than those of West African Dwarf sheep in Nigeria (Gatenby, 1986), but surpassed by Rahmani, Ossimi and Barki (Aboul-Naga et al, 1981a). Ueckermann et al (1974) reported that of the South African non- woolled sheep, Dorper breed produced an average daily milk yield of 1262 g in 100 days. The present yield of about 750 g/day recorded in 84 days is too low and might be indicative of the existing insufficient nutritional environment. The pattern of daily milk yields in goats significantly rising to maximum within two weeks and then decreasing afterwards to end of lactation was also observed by Ehoche and Buvanendran (1983) in the Red Sokoto goat. The pattern of milk yields as observed in sheep was also reported by Gatenby (1986). The overall contents of total solids, fat, SNF, ash, protein, lactose and energy observed in the present study are much higher than those of all other breeds, except those of the West African Dwarf and the pygmy goats, as summarised by Parkash and Jenness (1968) and Devendra and Burns (1983). The contents in sheep are lower than those reported for Middle Atlas of Morocco (Gatenby, 1986). The changes in the composition of sheep and goat milk in this study are different from the results reported for dairy goats in the temperate areas (Parkash and Jenness, 1968). Kids and lambs born in the dry season weighed less at birth, but weighed and grew more up to weaning. The lighter birthweight in the dry season might probably be due to nutrition of the dam in the last third of pregnancy. Most goats and sheep gave birth between May and June when rains had stopped and plan of nutrition was getting low. Peart (1982) concluded that when nutrition of the dams during the last stages of gestation is low or restricted, birthweight of kids and dam milk yields in the first lactation weeks are reduced. However, the higher performance of these kids and lambs until weaning is probably due to the higher yields of milk and milk components produced by the dams during this reason. Observations made in the present study are contrary to expectations and to most published results (Sousa et al, 1987; Hendy, 1987). The controversy might have arisen due to differences in location, management, environment and genotype. Genotypic effect of dams was evident for differences in kid birthweight, but not for weights and gains to weaning. The least and heaviest birthweights in kids born to local and Boer does, respectively, are expected. Similar results were observed in sheep. The intermediate birthweights of kids born to crossbred goats may imply some additive gene action among the dams for kid birthweight and an advantage over the local goats as regards this trait. These results are contrary to the report published by De Haas (1977). Later, similar findings to the present results were reported by Ayoade and Butterworth (1982). However, similarity among genotypes in most of the pre-weaning traits might indicate that the milk yield differences may have been inadequate for the kids to show differences. In 480 sheep, lambs born to and reared by local lambs performed the poorest. Weaning weights and pre- weaning gain differences in lambs born to and reared by Dorper on the one hand and crossbred ewes on the other hand were not significant, implying that the differences in milk yield only were not adequate to show differences in growth traits of the lambs. Conclusion Goats produced more milk than sheep irrespective of genotype used. The milk yields are much lower than those for temperate and other Indian-recognised breeds. Due to reproductive problems in exotic genotypes, especially in Boer goats, no clear conclusions can be drawn. The content of all milk constituents in sheep, except fat, was higher than that in goats. The average composition is much higher than that of most tropical genotypes recorded. Lambs grew faster than kids. The low milk yields of the goats and sheep used could probably be increased by both selection and crossbreeding with well known high milk producing breeds. Although no valid recommendations could be drawn from this short-term study, the study has brought out some specific areas and problems that need attention. Some suggestions to be made need to be backed up by long-term studies through multidisciplinary approach. Suggestions for further research • If at all feasible, more research is required to include a larger number of Boer goats in order to come up with valid conclusions. • There should be a continuation of the evaluation and description of the performance potential (milk and meat) of all the existing stock to cover all the genotypes in all the zones of the country. • Evaluation of the milk production potential should be extended to include the newly introduced Saanen breed and its crosses with local goats. Assessment should include even post-weaning period to effectively evaluate milk production persistency, a character suitable for dairy production. The three milk measurement techniques should be tried to find the differences in milk production response. The suckling method needs refinement. • Possible causes of mortalities and disease incidences should be found. There is need to search several new treatment strategies to improve the health status of the small ruminants. Causes of reproductive problems in Boer goats and Dorper sheep should be promptly investigated. • Testing of several lamb and kid rearing regimes. This should include different suckling regimes and time of weaning, different feeding systems and housing types to help reduce mortality. Effect of milk removal on performance of the young should be examined. 481 • Test several feeding, health, breeding and management technologies on the existing stock and evaluate their economic viability. • Selection of the existing small ruminant stock for growth rate of the offspring and milk production should be expanded or established. • Comparative technical and economic evaluations of goat and cattle milk and meat production enterprises at farm and station level need to be conducted. Acknowledgements The authors wish to thank the Malawi-German Livestock Development Programme for the provision of animals, facilities and other materials used in the study. The funds used in this work were obtained from the German Academic Exchange Service (DAAD), Contract Research Committee of the Ministry of Agriculture (Research), Malawi and the University of Malawi Research and Publications Committee. References Aboul-Naga A M, El-Shobokshy A S, Marie I F and Moustafa M A. 1981a. Milk production from subtropical non-dairy sheep. 1. Ewe performance. Journal of Agricultural Science (Cambridge) 97: 279-301 . Aboul-Naga A M, El-Shobokshy A S and Moustafa M A. 1981b. Milk production from subtropical non-dairy sheep. 2. Method of measuring. Journal of Agricultural Science (Cambridge) 79:303-308 AOAC (Association of Official Analytical Chemists). 1980. Official methods of analysis. 13th edition. AOAC, Washington, DC, USA. 1018 pp. Ayoade J A and Butterworth M H. 1982. The relationships between birth and weaning weights in kids of Malawian local goats, Boer and their crosses. Tropical Animal Production 7(2): 1 13-115. De Haas H J. 1977. Boer crossbred goats in Kenya: A comparison of growth rates. Animal Research and Development 6:72- 76. Devendra C and Burns M 1983. Goaf production in the tropics. Commonwealth Agricultural Bureau, Slough, UK. 183 pp. Doney J M, Peart J N and Smith W F. 1 979. A consideration of the technique for estimation of milk yield by suckled sheep and a comparison of estimates obtained by two methods in relation to the effect of breed, level of production and stage of lactation. Journal of Agricultural Science (Cambridge) 92:123-132. Economides S. 1986. Comparative studies of sheep and goats: Milk yield and composition and growth rate of lambs and kids. Journal of Agricultural Science (Cambridge) 106:477-484. Ehoche O W and Buvanendran V. 1983. The yield and composition of milk and pre-weaning growth rate of Red Sokoto goats in Nigeria. World Review of Animal Production 19:19-24. Gatenby R M. 1 986 Sheep production in the tropics andsubtropics. 1st edition. Longman Group Ltd, Essex, UK. pp. 202-225. 482 Harvey W R. 1977. User's guide for LSML76. Mixed model least-squares maximum likelihood computer program. Ohio State University, Columbus, Ohio, USA. (Mimeograph) Hendy CRC. 1987. Production of the indigenous goats under traditional management in southern Tanzania. In: SantantaOP, Aliomar Gabriel daSilvaand Foote WC (eds), Proceedings of the IV International Conference on Goats. Volume II. International conference held in Brasilia, Brazil, 8-13 March 1987. Departamento de Difusao de Tecnologia-DDT, Brasilia, Brazil, p. 1544 (Abstract 342). Mill H and Steinbach J. 1984. Genotypische unterschiede in der Milchleistung bis zum Zitpunk des Absetzens in Abhangigkeit von der Melktechnik Seminar iiber Ziegenforschung in Cap Serrat/Sria, Tunis, Oktober, 1984. Germany, pp. 150-161. Mittal J P, Agarwal M P and Bist K S. 1977. Effect of breed, age, season and para on milk secreting capacity of goats. Indian Veterinary Journal 54:449-453. Mukundan G and Bhat P N. 1983. Lactation curve in Malabari goats and their Saanen half-bred. I. Milk production. Indian Journal of Animal Science 53 (6): 666-669. Parkash S and Jenness R. 1968. The composition and characteristics of goat's milk: A review. Dairy Science Abstracts 30:67-87. Peart P N. 1982. Lactation of suckling ewes and does. In: Coop I E (ed), World animal science. C. Production- systems approach. 1. Sheep and goat production. Elsevier Publishing Co., Amsterdam, The Netherlands. Russel A J F, Doney J M and Gunn R G. 1969. Subjective assessment of body fat in live sheep. Journal of Agricultural Science (Cambridge) 72:451-454. Sacker G D and Trail J C M. 1966. The effect of year, suckling, dry season, and type of dam (ewe or gimmer) on milk production in East African Blackheaded sheep as measured by lamb growth. Journal of Agricultural Science (Cambridge) 66: 93-95. Sousa W H, Corriea W S, Nery J K, Lima F A M and Pant K P. 1987. Influence of the breeding season on birth and survival of Caninde kids. In: Santanta O P, Aliomar Gabriel da Silva and Foote W C (eds), Proceedings of the IV International Conference on Goats. Volume II. International conference held in Brasilia, Brazil, 8-13 March 1987. Departamento de Difusao de Tecnologia-DDT, Brasilia, Brazil, p. 1487 (Abstract 263). Steinbach J. 1988. Experiences with the evaluation of the production potential of local and improved goat breeds in Northern Tunisia. Animal Research and Development 28:100-114. Ueckermann L, Joubert D M and van der Stein G J. 1974. The milking capacity of Boer goat does. World Review of Animal Production 10 (4): 73-83. Wohlt J E, Foy W L, Jr, Kniffen D M and Trout J R. 1984. Milk yield by Dorset ewes as affected by sibling status, sex and age of lamb, and measurement. Journal of Dairy Science 67(4): 802-807. 483 Les performances de différents niveaux de croisements: chèvre alpine et petite chèvre de l'Afrique de l'Est à la Station d'élevage caprin de Vyerwa (Burundi) C. Ntahimpereye Projet Caprin Ngozi, B.P. 45 Ngozi, Burundi Résumé A partir de 1980, le projet de développement caprin dans la Buyenzi a réalisé un programme de croisement d'absorption de la petite chèvre d'Afrique de l'Est par des chèvres alpines importées, à la Station d'élevage de Vyerwa. Cette communication résume les performances des différents génotypes produits (50%, 75%, 87% de sang alpine) de 1981 à 1990. Le croisement entre ces deux races caprines améliore sensiblement les performances de croissance et laitière de la petite chèvre d'Afrique de l'Est, et ce, dès la première génération: • poids à 3 mois: 10,0 kg, 14,3 kg, 14,1 kg, 15,2 kg pour respectivement F, (sang alpin), fl, (3/4 alpin), R2 (7/8 alpin) et la race pure alpine, alors que la R.L. (race locale) pèse 8,5 kg au même âge. • La production laitière allantjusqu 'à 236 litres pour les F,, 355 litres pour les Ri par lactation tandis que la race locale n'a que 33-36 litres en 90-100 jours. Ce croisement paraît toutefois affecter négativement les performances de reproduction (fertilité, fécondité, intervalle entre mise-bas) 485 Performance of Alpine x small East African Dwarf goats with varying proportions of Alpine blood at Vyerwa Livestock Breeding Station, Burundi C. Ntahimpereye Abstract In 1980, the Goat Development Project in Buyenzi region introduced a crossbreeding programme at Vyerwa Breeding Station aimed at upgrading the local breed by crossing it with imported Alpine goats. This paperpresents a summary ofdata on the performance ofanimals with different levels of Alpine blood (112, 3/4, 7/8 Alpine) recorded between 1981 and 1990. Crossbreeding markedly increased milk yield and adult liveweight relative to the local race, even in the first cross, but resulted in poorer reproductive performance (lower prolificity and fecundity and longer parturition interval). Introduction Les caprins, ovins, comme d'ailleurs les bovins de races locales sont peu productrices et leurs performances sont insuffisantes pour rentabiliser les investissements et les frais d'entretien occasionnés par l'intensification de l'élevage. Dans certaines régions où la densité démographique est élevée comme la région du Buyenzi (nord du Burundi: 300 à 400 hab./Km2), l'élevage bovin en système traditionnel semble en nette régression au profit des petits ruminants. Cette diminution entraîne une baisse sensible des quantités de viande et surtout de lait. Compte tenu du faible niveau de production laitière de la race bovine locale, cette même production pouvant être obtenue à partir des chèvres laitières importées, un croisement d'absorption entre la chèvre locale et la chèvre alpine a été effectué dans cette région par le placement des boucs de race alpine dans des centres de monte. Les performances de reproduction, de croissance, et de production laitière des différents génotypes obtenus en station par croisement de la petite chèvre d'Afrique de l'Est avec les boucs de la race alpine font l'objet de la présente communication. 486 Matériel et méthodes Les performances ont été relevées de 1981 jusqu'en 1990 à partir d'un troupeau caprin de 250-300 têtes élevé à la Station d'élevage de Vyerwa au Burundi. Le troupeau utilisé était composé de génotypes suivants: • Race pure alpine (R.A.) • Les croisés Fi (R.L; R.A.) • Les croisés R, (1/4 R.L; 3/4 R.A.) • Les croisés R2 (1/8 R.L; 7/8 R.A.) Ces croisés ont été obtenus par le croisement d'absorption des mâles importés (R.A.) étant mis sur les femelles petite chèvre de l'Afrique de l'Est. Le troupeau R.A. sert de noyau de production de géniteurs destinés à la diffusion en milieu rural dans les centres tandis que les croisés étaient élevés dans le but d'avoir les données sur les paramètres zootechniques. Le relevé des paramètres de croissance s'est effectué par pesée sur balance des chevreaux, à la naissance et une fois tous les mois. Le calcul de poids type à 30 jours, à 60 jours, à 90 jours et à un an est fait par moyenne simple. Relevés des performances laitières: Alors qu'avant 1985 tous les chevreaux (R.A. et croisés) étaient allaités naturellement, après la détection de l'arthrite encéphalite des caprins en 1985-1986, tous les chevreaux de race alpine pure sont allaités artificiellement tandis que les croisés étaient toujours soumis à l'allaitement maternel. Ainsi pour les R.A., toute la quantité de lait produit est pesée directement, et pour les croisés la quantité de lait consommé par les chevreaux a été calculée par le pesage des chevreaux avant et après l'allaitement. Le sevrage a lieu à 15 kg à peu près à l'âge de 3 mois. Les données ont été stockées pendant 9 ans et l'enregistrement sur ordinateur (Logiciel IDEAS) a eu lieu en septembre 1990 par Hahn (1990). Résultats et discussion Les paramètres de reproduction Les taux suivants ont été déterminés pour étudier les performances de reproduction (tableau 1): taux de fertilité apparente (nombre de mises-bas par rapport aux femelles en âge de reproduire); taux de fécondité (nombre de 487 chevreaux nés par rapport aux femelles en âge de reproduire); taux de prolificité (nombre de chevreaux moyen par mise-bas) et intervalle entre mise-bas. En conclusion, le croisement d'absorption entre la petite chèvre d'Afrique de l'Est et le bouc de race alpine a une influence négative sur les paramètres de reproduction. Tableau 1 . Paramètres de reproduction à la Station de Vyerwa (Ngozi) Type génétique Paramètres F1 R1 R2 alpine Taux de fertilité apparente (%) 114,3% 109,9 % 103% 124,7% Taux de fécondité (%) 181,1 % 172,9 % 161,2% 191,4% Taux de prolificité (%) 161,4% 157,8% 155,3% 148,9% Intervalle entre mise-bas (jours) 313 335 357 329 Les performances de croissance Le tableau 2 présente les valeurs pondérales moyennes des chevreaux à la naissance, à 30 et à 90 jours. A la naissance les mâles de tous les niveaux de croisement sont plus lourds que les femelles, et les chevreaux nés simples sont plus lourds que ceux nés jumeaux et triples. Ces poids sont toujours supérieurs à ceux des chevreaux de race locale (tableau 3). A 1 mois les mâles sont encore plus lourds que les femelles chez tous les génotypes sauf les F1 où le poids est le même pour les deux sexes. Les R, et R2 (mâles et femelles) sont plus lourds que les chevreaux de race alpine. Cela peut être dû au système d'allaitement, puisque les R, et R2 restaient avec leurs mères tout le long de la journée, alors que les R.A. avaient un allaitement artificiel, à partir de 1985. Cette différence n'est cependant plus notable au sevrage (90 jours). A cet âge, on note surtout que les chevreaux demi-sang sont plus légers que les trois autres types. Les performances laitières Les résultats des performances laitières des différents niveaux de croisement sont exprimés dans le tableau 4 (Schmidt, 1990). Ces résultats ont été relevés à la même station de Vyerwa sur une période n'arrivant pas à 9-1 0 ans comme les autres paramètres déjà vus. 488 o o c CO CO 3 c eg■o § eg 3 c Co I T3 CD■o CO co -c co■O o o> o co -CO CD O c CO o a. <\i 3 CO o Si a 8 a> co 3^ CD CO" CO CO. CO IO o CO CO CO CM o o < rr CO. f-" in in iri CO cm" ro Il co" c § =- * 1 1 ST CO S CO" CO * i S. CD CO r^" c CM 5 o CM s 5" CO o CO ? CM CO 5" CO cr CO eg in" ro" CM CO cm. , . O cm" CD CM § CO CO E CO o o" lZ - CM en 1 I d o" - CM 'at < 5>_ CO CO CO CO LD CO CO CM CM. CO in CM 5" CO CO CO CO" CO cf co" CO CM co" 2i CO CM co" s in CM s COCO CM J CO P CO 1^ CMCM 2 co" CM. co" N co" co" — co" N-" , cm o co' CM CO CO ID CM CO in CO 5" CO CO 2 CO o , . 2, c-" 3_ r-" CO co" N in h-" O) CO o? 2. s o «0 o in" CM CM. CO 1 CO s CO 5" O "5 < in" tn" in" m" CO o cm" CO 55" o 5" 8 m 5T CM CMCO O) COco" CO" o co co" co" cm" COCM co" 3 CO 3" CM CO IO s CO I in o" 35" ^* a>cm" 1 CM 2 CO co" cm" CO" co" co" co" "3 rr 1 enco 1 CO § cm"2- I in ^r^ CO & CO CMco" cm" co" co" cm" cm" co" "3 cm CO CD CO CM 2 CO CM cm" o CO 9 in cm" CO CD s CO CO CM CO cm" cm" cm" CM cm" II cm' a> CD O 0) O a> o eg c o o CO o» co o c c c c a> c 1 a> c 1 C CO 0> CO 1 =It s* 0) if 34 c 0) »■§ S 9- >. o E 1.9- Z i: (D>. CO in T— CT> SI cn CO C\J cr> CO LO 8ii o> CM r- a> CO a> 3 2 8 LO CO ID 8 en CO a> C\J CO CO CO in o■ a u ■ n d V K cn in CO 8 co Q. to a 5 s o> C0 o CO 01 O)r- >• 8 CO 8 in cn CO o c O) w « ■Q u B 4 Q. CO CO CO r^ CO f CO cn CO Q. 03 n (S n 3 o CO «t 8 CM toCO co CCO LO m CO CD CO co o CM LO 8 5= CO x: CO a. g> CD 3 CD S COCM cncn o cn in «- m cm mcn cn cn cn* * » ^- CM O «» CM CD «- «- CM (O V * «- CM K "' O)° ^ "> s LO CD CO CO CO O) O) O)■" *- *~ x: r 5 CO o CO o CD O > cn in 3cn CM 8 cm ? o 8 m §5 »- « .- cn r 5 CO X) in co 8? | 2 s 3 E 5 J3 CD O. a. o CD X CD cn o a T3 C CO CO XI« 502 variables affecting age at eruption were year of birth and to a lesser extent type of birth. Daily weight gain from birth and age at first eruption was correlated (P<0.01). Gain between successive eruptions were also correlated (P<0.01) with their respective ages at eruption. This, as in goats, may have been a reflection of nutrition from birth to first eruption. Since only breeding females were available from second to fourth eruption, the effects of year of birth on weight and weight change were confounded by physiological status of animals which was not taken into account in the present analysis. Sheep and goats differed (P<0.01) in age of eruption of their second to fourth pairs of permanent incisors, the magnitude of the difference increasing between each pair (Table 2). The mean interval between first and second, second and third and third and fourth was 151,114 and 227 days for goats and 1 84, 1 92 and 183 days, respectively, for sheep. Overall, goats attained full mouth at a significantly (P<0.01) younger age than sheep (Table 2). Goats were lighter (P<0.01) than sheep by 6, 8, 3 and 4 kg for first to fourth eruptions, respectively. Daily weight gain from birth to first eruption differed (P<0.01) between species with sheep gaining 14 g per day more than goats. Table 2. Mean age and 95% range (days) at eruption of permanent incisors for goats and sheep. Dentition Species Least squares contrast Age (goat minus sheep) Range First pair Goats 287 423 377-487 Second pair Sheep 247 418 Goats 145 581 Sheep 252 605 Third pair Goats 120 691 Sheep 240 798 Fourth pair Goats 114 918 Sheep 213 -24' -96' -55 982 367-469 509-653 525-685 597-786 700-896 773-1064 823-1141 Significant at P < 0.01 . 503 Commercial Implications Though other studies have found no direct correlation between normal skeletal development and incisor development (Tagle and Helman, 1943), age determination facilitates market transactions and identification of carcasses of possible superior eating quality. A goat grading system, taking age classification of carcasses (which are different from sheep) into account is proposed. The large variation in ages at eruption of incisor teeth recorded here resulted in some overlapping of the distributions for successive pairs (Table 2). This limitation should be considered when using dentition to determine age. Acknowledgement We would like to thank all members of staff of the Small Stock Section for care of the animals and data collection. We also thank Dr. T. Smith for helpful discussions. References Otesile E B and Obasa1u M F. 1982. Relationship between age and rostral teeth development in Nigerian goats. In: Proceedings of the Third International Conference on Goat Production and Disease, University of Arizona, Tucson, Arizona, USA, 10-15 January 1982. Dairy Goat Publishing Company, Scottsdale, Arizona, USA. p. 349. SAS (Statistical Analysis System) Institute. 1987 . SAS/STAT guide for personal computers. 6th edition. SAS Institute Inc., Cary, North Carolina, USA. Sibanda R. 1990. Productivity of indigenous goats under an accelerated kidding management system. MPhil thesis, University of Zimbabwe, Zimbabwe. Starke J S and Pretorius A G. 1955. Dentition of sheep as an indication of age. Farming, South Africa 30:53-56. Tagle E L and Helman M B. 1943. Condideraciones sobre la edad del ovino. Animal Breeding Abstracts 15:34. Ward H K, Richardson F D, Denny R P and Dye P J. 1979. Matopos Research Station: A perspective. Rhodesia Agricultural Journal 76:1-18. Wilson R T and Durkin J W. 1984. Age at permanent incisor eruption in indigenous goats and sheep in semi-arid Africa. Livestock Production Science 1 1 :451-455. 504 Growth traits of the Dorper sheep I. Factors influencing growth traits B.A.O. Inyangala\ J.E.O flege1 and S. Itulya2 'Animal Production Department, University of Nairobi P O Box 29053, Nairobi, Kenya 2Animal Science Department, Egerton University, P 0 Box 536, Njoro, Kenya Abstract Data on 969 Dorper lambs collected over a 10-year period (1978 to 1987) at Ol'Magogo were used in this study. Lamb traits studied were weights from birth to yearling and absolute growth rates between adjacent stages of growth. All the fixed effects studied influenced growth in one way or another. Sex was highly significant for all traits (P<0.01) except growth rates between weaning and six months (GR2), and six to nine months (GR3), and season of birth influenced all traits significantly (P<0.05) except birthweight (BIRTHW), growth rates between weaning and six months (GR2), and between nine and twelve months (GR4). The effect ofparity was confined to pre-weaning traits (P<0.05), while period of birth was significant (P<0.05) on all traits except yearling weight (ADJTMW) and growth rate between weaning and six months (GR2). Parametres de croissance du moutort Dorper I. Facteurs agissant sur les parametres de croissance B.A.O. Inyangala, J.E.O. Rege etS. Itulya Resume Ceffe efude repose sur les donnees recueillies pendant 10 ans (1978 a 1987) sur 969 agneaux de race Dorper a OL'Magogo Naivasha. Les paramètres etudies sont le poids des agneaux de la naissance, a l'age de 1 an, et leurs taux de croissance absolus a intervalles de 505 temps consécutifs. Tous les effets fixes examines influengaient la croissance d'une manière ou d'une autre. L'effet du sexe etait tres significatif pour tous les paramètres etudies (P<0,01), a l'exception des taux de croissance entre le sevrage et l'age de 6 mois (GR2) et entre 6 et 9 mois (GR3). La saison de naissance a montre un effet significatif (P<0,05) sur tous les paramètres excepte sur le poids a la naissance (BIRTHW), et les taux de croissance entre le sevrage et l'age de 6 mois (GR2) et entre 9et12 mois (GR4). Alors que l'effet du rang de la naissance se limitait aux seuls paramètres examines avant le sevrage, la saison de la naissance avait un effet significatif (P<0,05) sur tous les paramètres etudies excepte sur le poids a 1 an (ADJTMW) et le taux de croissance entre le sevrage et l'age de 6 mois (GR2) Introduction The Dorper breed was developed in South Africa around 1942, mainly at Grootfontein College of Agriculture from initial crosses between Blackhead Persian ewes and Dorset Horn rams. The Dorper has a reputation for its adaptability to rather harsh environment. Characteristics of the breed include the ability to walk long distances and forage well in permanently dry areas and in times of drought, good mothering ability in the ewes, high ram fertility and vigour, excellent carcass conformation for good mutton production in comparison with the indigenous breeds and an unrestricted breeding season. In Kenya, Dorper sheep accrue income mainly from meat and sale of surplus stock. Bodyweight and rate of gain are among the most economically important and easily measured traits of meat animals. Although weight is an important objective in selection, knowledge of the particular phase of the animal's growth upon which to base selection is of utmost importance. Studies on the environmental factors influencing growth traits are few for sheep reared under Kenyan conditions. Gumedze (1979) studied fertility in Dorper sheep at Ol'Magogo. Ouko Odenya (1982) and Kiriro (1986) studied lamb traits only up to weaning. A majority of studies have reported the performance of temperate sheep. There is an extreme paucity of information regarding the performance of sheep raised in Kenya. The purpose of this study was to investigate environmental sources of variation influencing growth traits in a flock of Dorper sheep at various stages of growth, from birth to yearling. Materials and Methods The data used in this study consisted of growth records of 969 Dorper lambs. These data were collected from records of the Sheep and Goat Development 506 Project (SGDP) based at OI'Magogo, a substation of the National Animal Husbandry Research Centre (NAHRC), Naivasha, between 1978 through 1987. Bodyweights analysed included weight at birth, at three months (weaning), at six months, at nine months and at months (yearling). Lambs were nursed by their dams up to weaning. Each lamb record included sire, dam and lamb identifications, type of birth and sex. Three seasons of birth were defined on the basis of the monthly rainfall distribution. Two rainy seasons were identified with April and May (season 1) forming the peak of the long rains while October and November were classified as a season of short rains (season 2). The remaining months were classified as a dry season (season 3). Parities were defined based on the number of times the ewes had lambed giving rise to parities 1 , 2, 3 and 4, the latter comprising ewes with 4 or more lambings. Due to disproportionate distribution of data across years which resulted into discontinuity in the data, it was not possible to include actual years in the analysis. Therefore, to adjust for differences in weather conditions across years, periods of birth were defined by grouping adjacent years on the basis of annual rainfall pattern. This otherwise uncommon grouping method was implemented after examination of meteorological data, and was felt to be the best method under the circumstances. Period 1 consisted of the years 1978-1980 which received the highest rainfall, period 2 (1981-1985) received intermediate amounts while period 3 (1986-1987) received the lowest. The absolute growth rates were derived by taking the difference in weight within the period and dividing it by the time interval in days. The absolute rate of gain for each lamb was calculated over five growth periods namely: Birth to weaning (GR 1 ) , weaning to six months (GR2) , six to nine months (GR3) , nine to 1 2 months (GR4) and birth to 12 months (OVRGRT). These, together with the bodyweights, constitute the 10 traits analysed in this study. In view of the differences in actual age at which weights were taken, the latter were pre-adjusted as follows: ADJWWT (adjusted 90-day weaning weight) = GR1 X 90 days + birthweight (BIRTHW) ADJSMW (adjusted six-months weight) = GR2 X 90 days + ADJWWT ADJNMW (adjusted nine-months weight) = GR3 X 90 days + ADJSMW ADJTMW (adjusted 12 months weight) = GR4 X 90 days + ADJNMW. 507 Adjustment for fixed effects (sex, season of birth, period of birth, parity and dam breed) was achieved by including them in the model. Statistical analysis: The statistical model used to relate observations with independent variables was as follows - Y.ikimn = U + a, + bi + Ck+d| + fm + Oiiidmn where Yiiwmn = the ijklmnp"1 observation. u = an underlying constant for the trait ai = effect of the i'h sire (i= 1 63) assumed random, N(0 ) b| = effect of the jlh season of birth (j= 1,2, 3) ck = effect of the k,h parity (k= 1 , 2, 3, 4) d, = effect of the Ith period of birth (I= 1, 2, 3) fm = effect of the m'h sex (1 =male, 2=female) Siiikm = random error associated with the ijklmnp"1 observation; N(O, ) The Least Squares Method (Harvey, 1982) in which sires were cross-classified with fixed effects was used in the analysis of the data. Results and Discussion Analysis of fixed effects Analysis of variance for bodyweights and growth rates are presented in Tables 1 and 2, respectively. Corresponding least squares means are presented in Tables 3 and 4 by subclasses. Sex Least squares means (Tables 3 and 4) indicate that male lambs performed, in all cases, considerably better than their female counterparts. These differences were significant (at least at P <0.05 level) for all the traits except for pre-weaning growth rate (GR2) and (GR3) for which it was not significant. Consistent superiority of male lambs has been widely reported (Dass and Acharya, 1970; Vesely and Robison, 1970;Magidetal, 1981; Fitzhugh and Bradford, 1983; and Kiriro, 1986). This has been attributed to hormonal differences between sexes and their resultant effects on growth (Velardo, 1958; Bell et al, 1970). 508 ao Q. o Q ■c O .52 eg CO o CM •2- O ,- 3 -c a) X ro o .2 «co > c/5 a> CD re m CO CO CL a. LU 1 «> a>« ,>* If 2 ? c 2 CD o W > 0-0- EC CC C3 u. S 8 t- cm co eg r r CO c £ O c E Is I I c o o *« « c t II II o DC> o cJ! 9> 03 -C "C O c o E 509 £?3 Q< 5 O O CO 00 m q en o o tv t^ 1^ ego o ■>r eg 5~ +1 +1 +1 +1 +1 in +1 +1 + 1 +l in in +1 +1 +1 T Z °» +1 CM CO 8 m Q< CO CO CO IT) CO CO o co 00 O) CM o a> eg O) O) CM o w IO eg mCM CM CM CO CM eg eg Q < c a> a* CO a a in CO 8 8 CO seg eg eg eg eg CO +1 +i +i +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 eg o> f- 8 § egeg O) CO § COCO t io o m 8 s 8 IOCO IOCO in 8 inCO CO CO CO CO CO CO io N CO m fc. ,- in C- in in o a> in CO 00 CO 00 CM CO 00 CO Ol O) eg O) p d 6 6 d y^ d d T- d d r^ d t— +1 +i +1 +1 + 1 +l +1 +1 +1 +1 + 1 +1 +! eg O) in in O) eg 8 *— eg CM t^ «— <^CD CO CO CO CO CD r-t N CO t O v0 in CO in eg in ■>T t TT ■>r * eg K CM eg eg eg eg eg CVJ eg CM CM eg eg eg o ti CO Tl •§ 1 a to c■o u C8 C 0 £CO CO 2 ?r J3m Oen E c- m » eg 0) -j £■ n o *> O) a • a l- O o +1 CM in d +1 a CO CO s d +1 CM CM d CM CD d +1 00 CO ■? in d d +1 +1 * «- in co d ob eg «- cm «- o +1 CO o> o +1 eg o 8 UJ CO +1 c ns CD E CD a CO 9 CM eg f- CM in CO 8 eg eg co o> eg o +1 in 0)0)0)0K o +1 tv d CM n o o o 00 en o o CM eg K CO ■>r«— T— t— t— 1- o «— *- f t■ t- «-" *— o o o o o o o o o o o o o +1 +1 +! +l +1 +l +1 +1 +1 +1 +1 +1 +1 CM eg eg o CO § § 8 t ■>»■ tv 5> 8O CO T *r *T 1- t ■«S- CO ■«■ ^- i- ■t CO ■* «- eg co CO o t- r- N eg in «- cm »- eg co o 1 510 l- tr ^ o5 cc S> >=*. a t ? §1 CO ^ -s si s c s- o Q £5 ? 0 »■■ a "5 0) o> COc «- TJ "5 SI0) O 0- CD■o s. .0) 0) o CO O) 0) CD —1 co «f o 05 XI a t- in 8 § g g 8 8 o o o O o o +1 +1 +1 + 1 +1 +1 CD 05 s IT! § CO a> CM t- oCJ o o *- O o o o o co «- «- 8000000 o +1 so 8 8 8 o +1 CO 10 o 0 8 0 +1 id in o o CJ5 o o 0 +i CO 5o d O o d Tl in o o 5 8 8o 0 +1 05 8 o o +1 I^ in o o o +1 CM in o d o +1 05 in o d o d +1 o 05 oo o +1 CO in o d o o +1 CO 8 o d +1 CO co o o o ti CO in o d o +1 0) in o d o +1 CO m o o o Tl in in o 0 o +1 CD in o 0 in 8 8 d +1 CD d in O T in m O 8 8 8 0 O 0 0 0 +1 +1 +1 +1 t! 05 s CM00 CMCO O 8 fc CM CO CM C5 CD CM UJ 1) tl 0) €C CO xi 05 E CO 0— 0) c 0J E 05 5 CO 5 05 li- 0> CO r o o o d o +1 +1 I^ I^ (Tl CNJ o o d d 00000o o tl o tl m in o d CO o I^ I^ in «- o o tl 8 8 8 8 5 8 8 o tl i^ CM■u S 5> X) XI CO a. 511 Season of birth The effect of season of birth was significant for all traits except birthweight, pre-weaning growth rate and growth in the interval six to nine months. The effect of season of birth arises from seasonal variation in the physical environment resulting from changes in weather conditions (including rainfall amounts, temperature, and humidity) which directly affect feed availability, especially in a situation (such as is the case in this study) with no supplementary feed. Seasonal influence on a trait such as birthweight operates through its effect on the dam's uterine environment mostly in late gestation (Eltawil et al, 1970). Such factors operating in seasons prior to lambing will be manifested in birthweight. This may explain the higher (albeit non-significant) birthweight of lambs born in the dry season, lambs born in the long dry season during the critical stages of gestation. Such lambs would be expected to weigh more at birth compared to those whose dams underwent a nutritionally stressful period during gestation. It is, therefore, expected that the season when the ewe is in gestation is likely to play a more important role in birthweight than the actual season of birth. On the other hand, season of birth plays an important role in growth performance indirectly through its influence on the dam's nutrition (and hence amount of milk available to the unweaned lamb) and later, directly, through its effect on the pasture availability and quality on which the lamb is subsequently weaned. Growth traits are known to have positive correlations, both genetic and phenotypic (Osman and Bradford, 1965; Dzakumaetal, 1978; Mavrogenis et al, 1980; Stobartetal, 1986). The significance of season of birth for early growth performance may thus be responsible, as a carryover effect, for its significant influence on growth traits up to yearling. That growth rate from weaning to six months was not significantly influenced by season of birth — although season of birth was a significant source for growth rate before and after this period — may be due to the post-weaning stress which may have obscured the effect of season of birth. In general, seasons 1 and 3 were associated with better performance than season 2 (the short rains). The impact of this advantage early in life appears to be perpetuated up to yearling. As far as the bodyweights are concerned, season 3 was consistently associated with superior performance. From the results of this study, it is evident that lambs born in the dry season performed better than those born in the rainy season. It is therefore recommended that breeding at Ol'Magogo should be planned so that lambing would occur in the dry season. This way, would lambs benefit from favourable prenatal nutrient availability via their dams. However, lambing in the dry season has serious nutritional implications to the ewes. This is a time when ewes require good pastures so as to support their young and improve their own body condition. The best compromise would be to time lambing to occur towards the end of the dry season just preceding the rainy season. 512 Parity of dam Parity was a significant source of variation for birthweight, pre-weaning growth rate and adjusted weaning weight (P<0.01). Least squares means (Tables 3 and 4) indicate that as far as pre-weaning growth rate and weaning weight are concerned, the performance of lambs improved with increase in parity. That young ewes tend to produce smaller lambs at birth has been documented (Dass and Acharya, 1970; Wilson, 1987). First parity ewes are still growing and thus must provide for their own growth in addition to the foetal demand. It is generally known that mothering ability, especially milk production, increases with parity; older ewes are larger in body and are better milkers (Dass and Acharya, 1970; Eltawil et al, 1970; Wright et al, 1975; Stobart et al, 1986). Hence, influence of the superior maternal environment of such ewes is expected to be translated into better lamb performance up to weaning. It was, therefore, not surprising that post-weaning growth performance was not significantly influenced by parity. The effect of parity of dam on lambs is thus imparted as maternal influence whose direct influence is limited to the nursing period. Period of birth Period of birth was a significant source of variation for all traits except birthweight and yearling weight. As has been alluded to in Materials and Methods, period of birth was defined by grouping adjacent years which, from meteorological data, generally had a similar rainfall pattern. To this end, the significance of this factor was only important because it facilitated the adjustments of records for the effect of 'periods'. Any particular period, on its own, has no important bearing on the interpretation of the results and therefore did not have any management implications. Conclusion In order to improve breeding value, selection must be based on genotypic rather than environmental superiority. Thus, variation due to definable environmental effects must be removed by use of suitable adjustment factors. Ideally, these adjustments would be developed individually for each management unit (i.e. flock), but only rarely is sufficient data available to allow this. The next possible level for development of adjustment factors is within either definable groups of management units, or definable genetic groups. It is necessary that all known sources of variation influencing the traits of importance be included in the model of analysis, otherwise the results of the study may not be reliable. If less factors are included there is often a tendency to obtain overestimates of the genetic parameters of interest. In this study, for example, it would have been quite useful to include such factors as age of dam, and type of birth and rearing as some of the adjustment factors. However, type of birth and rearing was not included in 513 the model of analysis because genetic parameters were also being estimated based on paternal half-sibs only, while age of dam was not available in the original data set. The trait of interest ought to be investigated in the environment under which it is to be typically expressed, since this environment may be the one which is necessary for revealing certain desirable or undesirable genes or, in contrast, the sought-after genetic differences may be of little importance or indistinguishable in this environment. For instance, traits evaluated at OI'Magogo should not necessarily be expected to behave in the same way in a quite different environment, such as the Kenyan highlands. Genetic improvement of many traits might be more effectively accomplished by evaluating them, and others correlated with them, under particular environmental circumstances or during a specific period in the life cycle, depending upon the particular genes involved and the correlations among them. Acknowledgement This work received financial support from the Norwegian Agency for International Development (NORAD). Permission to use project data was granted by the Ministry of Livestock Development, Government of Kenya. We are indebted to the support and cooperation we received from staff of the Sheep and Goat Development Project, Naivasha. References Bell G H, Davidson J N and Scarborough H. 1970. Textbook of physiology and biochemistry. Longman Group Limited, Edinburgh, UK. Dass G S and Acharya R M. 1970. Growth of Bikaneri sheep. Journal of Animal Science 31(1):1-4. Dzakuma J M, Nielsen M K and Doane T H. 1978. Genetic and phenotypic parameter estimates for growth and wool traits in Hampshire sheep. Journal of Animal Science 47(5): 1014-1021. Eltawil L A Hazel L N, Sidwell G M and Terrill C E. 1970. Evaluation of environmental factors affecting birth, weaning and yearling traits in Navajo sheep. Journal ofAnimal Science 31 (5): 823-827. Fitzhugh H A and Bradford G E (eds). 1983. Hair sheep of western Africa and the Americas: A genetic resource for the tropics. Winrock International Livestock Research and Training Center, Morrilton, Arkansas, USA. 317 pp. Gumedze J S E. 1979. Factors affecting fertility in a flock of Dorper sheep. MSc thesis, University of Nairobi, Nairobi, Kenya. Harvey W R. 1982. User's guide for LSML76. Mixed model least-squares maximum likelihood computer program. Ohio State University, Columbus, Ohio, USA. (Mimeograph). 514 Kiriro PMM. 1986. Estimate of genetic and phenotypic parameters tor the Dorper, Red Masai and their crosses in Naivasha (OI'Magogo), Kenya. MSc thesis, Texas University, Texas, USA. Magid A F, Swanson V B, Brinks J S, Dickerson GE and Smith G M. 1981. Border Leicester and Finnsheep crosses. I. Survival, growth, and carcass traits of Fi lambs. Journal of Animal Science 24:766. Mavrogenis A P, Louca A and Robison O W. 1980 Estimates of genetic parameters for preweaning and postweaning growth traits of Chios lambs. Animal Production 30:271. Osman A H and Bradford G E. 1965. Effects of environment on phenotypic and genetic variation in sheep. Journal of Animal Science 24:766. Ouko Odenya W. (n.d.). Relationship of coat cover and production traits in the Dorper breed of sheep. MSc thesis, University of California-Davis, California, USA. 74 pp. Stobart R H, Bassett J W, Cartwright T C and Blackwell R L. 1986. An analysis of body weights and maturing patterns in Western Range ewes. Journal of Animal Science 63 (3): 729-740. Velardo J T. 1958. The endocrinology of reproduction. Oxford University Press, New York, USA. Vesely J A and Robison O W. 1970. Genotype-sex interaction in sheep. Journal ofAnimal Science 31 (2):273-277. Wilson R T. 1987. Livestock production in central Mali: Environmental factors affecting weight in traditionally managed goats and sheep. Animal Production 45:223-232. Wright L A, Thrift F A and Dutt R H. 1975. Influence of ewe age on productive characters of Southdown sheep. Journal of Animal Science 4(2):517-521. 515 Growth traits of the Dorper sheep. II. Genetic and phenotypic parameters RAO. Inyangala\ J.E.O. flege1 and S. Itulya2 'Animal Production Department, University of Nairobi P O Box 29053, Nairobi, Kenya 2Animal Science Department, Egerton University P O Box 536, Njoro, Kenya Abstract Data on 969 lambs collected over a 10-year period (1978 to 1987) on Dorper sheep at Ol'Magogo were used in this study. Lamb traits studied were weights from birth to yearling and absolute growth rates between adjacent stages of growth. Heritability estimates from paternal half-sib analysis were: 0.15±0.07, 0.18+0.08, 0.39±0.11, 0.55±0.13, 0.53±0.13, 0.14+0.07, 0.28±0.09, 0.59±0.14, 0.49±0.12 and 0.49±0.12; for birthweight (BIRTHW), adjusted weaning weight (ADJWWT), adjusted six-months weight (ADJSMW), adjusted nine-months weight (ADJNMW), adjusted 12-months weight (ADJTMW), pre-weaning growth rate (GR1), weaning to six months (GR2), six to nine months (GR3), nine to 12 months (GR4), and birth to yearling (OVRGRT), respectively. Post-weaning heritability estimates were generally higher than those for pre-weaning suggesting a lower direct genetic influence early in life. Genetic and phenotypic correlations estimated between weights were mainly positive (0.15 to 0.99 and 0.02 to 0.98, respectively). There were generally high genetic andphenotypic correlations between adjacent weights, between weights and rates of growth, and among growth rates themselves. 517 Paramètres de croissance du mouton Dorper II. Paramètres génétiques et phénotypiques B.A.O. Inyangala, J.E.O. Rege et S. Itulya Résumé Cette étude repose sur les données recueillies pendant 10 ans (1978 à 1987) sur 969 agneaux de race Dorper à OL'Magogo Naivasha. Les paramètres étudiés sont le poids des agneaux de la naissance à l'âge de 1 an, et leurs taux de croissance absolus à intervalles de temps consécutifs. Les coefficients d'héritabilité obtenus à partir de demi-frères paternels étaient de 0,15+0,07 pour le poids à la naissance, de 0,18+0,08 pour le poids au sevrage corrigé, de 0,39+0,11 pour le poids à 6 mois corrigé, de 0,55 ±0,1 3 pour le poids à 9 mois corrigé, de 0,53+0,13 pour le poids à 12 mois corrigé, et en ce qui concerne la croissance, de 0,14+0,07 avant le sevrage, de 0,28+0,09 du sevrage à l'âge de 6 mois, de 0,59±0, 14de6à9 mois, de 0,49±0, 12 de 9 à 12 mois, et de 0,49 ±0, 12 de la naissance à l'âge de 1 an. Les coefficients d'héritabilité des caractères étudiés après le sevrage étaientpresque toujours supérieurs à ceux des caractères étudiés de la naissance au sevrage, ce qui signifie que l'influence directe du patrimoine génétique de l'animal est moins importante au cours des premiers mois de la vie qu 'au cours des mois suivants. Les coefficients de corrélation génétique et phénotypique estimés entre les poids étaient essentiellement positifs, allant respectivement de 0,15 à 0,99 et de 0,02 à 0,98. Par ailleurs, il existait de fortes corrélations génotypiques et phénotypiques entre les poids moyens aux intervalles consécutifs, entre les poids et les taux de croissance et entre les taux de croissance eux-mêmes. Introduction The potential for genetic improvement is largely dependent on the heritability of the trait and its genetic relationship with other traits of economic importance upon which some selection pressure may be applied. Information on heritabilities is essential for planning efficient breeding programmes and for predicting response to selection. Genetic correlations, on the other hand, are essential in predicting indirect responses to selection and are needed in order to determine the optimum weighting and expected response to multiple trait selection. Statistical techniques presently utilised for the estimation of breeding values in selection programmes rely on estimates of genetic variation within and between traits of economic importance. The overall impact of any selection programme 518 will depend on the direct and correlated responses that result from selection on the selection criterion. These responses can be predicted a priori by using estimates of genetic and phenotypic relationships between all traits of economic importance. In Kenya, Dorper sheep are kept mainly for meat production. Thus, traits affecting economic viability include those associated with growth. Body weight and rate of gain are among the most economically important and easily measured traits of meat animals Although weight is an important objective in selection, knowledge of the particular phase of the animal's growth upon which to base selection is of utmost importance. Genetic and phenotypic parameter estimates are scarce in sheep reared under Kenyan conditions and where such information is available, analytical methods used tend to be inadequate. The purpose of this study was to estimate genetic and phenotypic parameters of growth traits in a flock of Dorper sheep at various stages of growth from birth to yearling. Materials and Methods The data used in this study consisted of growth records of lambs in a flock of Dorper sheep. These data were collected from records of the Sheep and Goat Development Project (SGDP) based at GTMagogo, a substation of the National Animal Husbandry Research Centre (NAHRC), Naivasha, between 1978 through 1987. A total of 969 records were available and were thus analysed for genetic and phenotypic parameters. Bodyweights analysed included weight at birth, at three months (weaning), at six months, at nine months and at 12 months (yearling). Lambs were nursed by their dams up to weaning. Each lamb record included sire, dam and lamb identifications, type of birth, sex and dam breed. Three seasons of birth were defined on the basis of the monthly rainfall distribution. Two rainy seasons were identified with April and May (season 1) forming the peak of the long rains while October and November were classified as season of short rains (season 2). The remaining months were classified as dry season (season 3). Parities were defined based on the number of times the ewes had lambed giving rise to parities 1,2,3 and 4, the latter comprising ewes with four or more lambings. Due to disproportionate distribution of data across years which resulted into discontinuity in the data, it was not possible to include actual years in the analysis. Therefore, to adjust for differences in weather conditions across years, periods of birth were defined by grouping adjacent years on the basis of annual rainfall pattern. This otherwise uncommon grouping method was implemented after examination of meteorological data, and was felt to be the best method under the circumstances. Period 1 consisted of the years 1978-1980 which received 519 the highest rainfall, period 2 (1981-1985) received intermediate amounts while period 3 (1986-1987) received the lowest. The absolute growth rates were derived by taking the difference in weight within the period and dividing it by the time interval in days. The absolute rate of gain for each lamb was calculated over five growth periods namely: Birth to weaning (GR1), weaning to six months (GR2), six to nine months (GR3), nine to 12 months (GR4) and birth to 12 months (OVRGRT). These, together with the bodyweights, constitute the 10 traits analysed in this study. In view of the differences in actual age at which weights were taken, the latter were pre-adjusted as follows: ADJWWT (adjusted 90-day weaning weight) = GR1 X 90 days + birthweight (BIRTHW) ADJSMW (adjusted six-months weight) = GR2 X 90 days + ADJWWT ADJNMW (adjusted nine-months weight) = GR3 X 90 days + ADJSMW ADJTMW (Adjusted 12-months weight) = GR4 X 90 days + ADJNMW Adjustment for fixed effects (sex, season of birth, period of birth and parity) was achieved by including them in the model. Genetic and phenotypic parameters estimated were heritability of each trait, and genetic and phenotypic correlations among these traits as traits of the lamb. Statistical analyses The statistical model used to relate observations with independent variables was as follows: Yiikimn = u + a, + bi + c„ + di + fm + eiiklmn where Yiiwmn = the ijklmnpth observation. u = an underlying constant for the trait a* = effect of the ilh sire (i = 1,...,63) assumed random, N(0, o2) bi = effect of the jlh season of birth (j=l, 2, 3) C = effect of the k,h parity (k=1, 2, 3, 4) di = effect of the Ith period of birth (1=1 , 2, 3) fm = effect of the mlh sex (1 =male, 2=female) e,ikimn = random error associated with the ijklmnplh observation; N(0, aZ) e 520 Parameters were estimated from covariances of relatives using paternal half-sib (PHS) analysis based on the Least Squares Method (Harvey, 1982) in which sires were cross-classified with fixed effects. The analytical complications introduced in half-sib analysis by use of litters was avoided by including only lambs born as singles. Heritability was estimated from variance components as h2 = 4 o2 l(o2 + o2) s s w where h2 = heritability estimate ct2 = sire variance component s o2 = variance of records within sires w The approximate method of Swiger et al (1 964) was used to estimate the standard error of the heritability estimate. Genetic correlations between two traits 1 and 2, r12 were estimated as r12 , os1s2/ (as1os2) where os1 s2 is the sire covariance component for the two traits and os1 and os2 are square roots of respective sire variance components. Results and Discussion Genetic and phenotypic parameters Heritability Heritability estimates are summarised in Table 1. Heritability of birthweight was found to be 0. 1 5±0.07. This estimate compares favourably with estimates of 0. 1 9 (Osman and Bradford, 1965) and 0.21 (Dzakuma et al, 1978). The estimate for weaning weight was 0.18±0.08 which was lower than 0.28+0.1 1 (Stobart et al, 1986) and 0.36±0.12 (Mavrogenis et al, 1980). Thus, in general, both estimates for birth and weaning weight were lower than those in the literature. Heritability for six-months weight was estimated at 0.39±0.1 1. This was slightly above the estimates of 0.28±0.10 (Dzakuma et al, 1978) and 0.21 (Ercanbrack and Price, 1972). The estimate for weight at nine months (0.55±0.13) compares favourably with the range of 0.30 to 0.50 (Rae, 1982). However, few studies have reported heritability estimates of nine-months weight. That for 12-months weight was 521 1able.Herisibi ities3gene i0a dph otypi0o resitionsr mater alh lr-s bnalys(Dorp rs ). OVRGR1 0.57 ±0.22 0.23 ±0.1 0.32 ±0.17 0.32 ±0.07 0.■ ±0.05 0.31 ±0.23 0.22 ±0.1 000 ±0.3 0.27 ±0.17 0.-0 ±0.1 GR2 007 ±0.D 0.21 ±0.23 -0.3 ±0.1 -0.15 ±0.- 0.53 ±0.3 0.3 ±0.20 ±0.1 ±0.13 0.-0 ±0.1 0.37 GR3 0.15 ±0.1 -0.02 ±0.23 -0.15 ±0.1 0.07 ±0.3 0.53 ±0.3 -0.05 ±0.25 -0.22 ±0.3 0.51 ±0.2 -0.3 0.22 GR2 -0.3 ±0.21 0.1 ±0.- 0.3 ±0.03 0.50 ±0.3 0.15 ±0.1 0.D ±0.31 0.- ±0.01 -0.31 -0.03 0.2 GR1 0.1 ±0.31 0.19 ±0.3 0.2 ±0.3 0.11 ±0.- 0.3 ±0.- 0.2 ±0.07 -0.D -0.00 0.02 0.32 1raits ADJ13D 0.3 0.- 0.70 ±0.15 0.D ±0.15 0.70 ±0.00 0.53 ±0.3 0.25 0.1 0.32 0.50 0.31 ADJDD 0.3 ±0.D 0.51 ±0.D 0.- ±0.11 0.55 ±0.3 0.72 0.D 0.1 0.51 -0.15 0.3 ADJ33D 0.D ±0.22 0.3 ±0.1 0.31 ±0.11 0.01 0.55 0.3 0.05 -0.27 -03 0.2 ADJDWT 0.15b ±0.3 0.- ±0.03 0.- 0.52 0.3 0.3 -0.17 -005 0.3 0.1 Birth 0.15a ±0.07 0.32c 0.1 0.3 0.25 0.15 -0.3 0.11 0.11 0.3 1raits BIR1HD ADJWWT ADJ33D ADJDD ADJ13D GR1 GR2 GR3 GR2 OVRGR1 aHeritabilitiesw thstandarderro sb l wthemonp im rydi gonal. b.Geneticcorrela ionswithstandarderrorsb l wth maboveedi gonal. c.Phenotypiccorrelationsb owthediag nal. BIR1HD=birthweightGpre-weaningg o that ADJDDT=adjustedweaningightGR2weaningtosixmonths ADJ3MD=adjustedsix-monthsweig tGRixtn nemonths ADJDD=adjustednine-monthsweig tGR2n net1mo ths ADJ1MD=adjusted1-monthsweightOVRGR1birthtyea ling. estimated at 0.53+0.13 and was higher than values reported in other studies: 38±0.23 (Dass and Acharya, 1970); 0.26±0.11 (Stobart et al, 1986); and 0.11+0.11 (Dzakuma et al, 1978). Heritability estimates of growth rates progressively increased from 0.14±.07 for pre-weaning growth rate to 0.59±0.1 4 for growth rate between six to nine months, then declined to 0.49+0.12 for rate of gain between nine to 12 months. The range of heritability for pre-weaning growth rate in the literature (Hundley and Carter, 1956; Givens et al, 1960; Vogt et al, 1967) was 0.18-0.37. Overall growth rate had a heritability estimate of 0.49±0.12. However, this was higher than 0.29+0.11 (Stobart et al, 1986). Rae (1982) reported a range of 0.20-0.40 for post-weaning growth rate while a high figure of 0.58±0. 1 4 has been reported in Columbia sheep (Ercanbrack and Price, 1972). It is clear from these and other results that post-weaning growth generally has higher heritability estimates than pre-weaning growth. This would indicate that environmental factors, in relation to additive genetic factors, had more influence on early lamb gain than on gains later in the lamb's life. This may be attributed to the high maternal influence associated with lamb growth performance early in life. High maternal influence has a tendency to increase the component of variance environmental to the lamb thereby lowering heritability estimates (Harrington et al, 1962; Thrift et al, 1973). The best time to evaluate an animal's additive genetic value for a desired trait is under circumstances which assure maximum expression of the genes, that is under conditions when heritability is highest, provided that the genetic expression at this time is highly correlated with the genetic expression during the time period in life when the trait is most valuable or important (Ercanbrack and Price, 1972). However, one should not ignore the effect of time of evaluation on generation interval. Moreover, the desired genotype might be more accurately evaluated through another highly heritable trait, or an index, which is highly correlated genetically with it. These results, therefore, indicate that to select lambs for their own genetic merit for weights and gains, it would be best to use bodyweight at nine months as the selection criterion rather than weaning weight as is often practised. This trait should be superior to weaning weight or pre-weaning growth rate since it is much less influenced by maternal effects which tend to obscure the direct additive genetic effect for growth. Although the generation interval may not be considerably shortened, overhead costs should certainly be curtailed when the rest of the lamb crop is disposed of at this stage. On the other hand, selection of ewes as dams must be based on lamb performance pre-weaning and at weaning. In any case, the objective should be to choose a practical selection criterion which will maximise the annual rate of progress for the trait to be improved without seriously impairing merit in important correlated traits (Ercanbrack and Price, 1972). 523 Genetic correlations Genetic correlations between growth traits at various ages was generally high and positive ranging from 0.15±0.24 to .99+01 (Table 1). Genetic correlations between adjacent traits were generally higher than the ranges reported in the literature: 0.21-0.77 for correlation between birth weight and weaning weight (Osman and Bradford, 1965; Mavrogenis et al, 1980), 0.13-0.22 for birthweight and 1 2-months weight, and 0.21-0.72 for weaning weight and 1 2-months weight (Dzakuma et al, 1978; Stobart et al, 1986). The high genetic correlations among these traits suggests that selection for any one of these traits would result in considerable positive change in other traits. The complications brought about by maternal environment early in the lamb's life (hence the lowered estimate of direct genetic effects) may be overcome by utilising correlated response. One could concentrate on traits with high heritability as long as there exists a high positive correlation with other traits of economic value. Moreover, selection directed towards weights at later ages would minimise response in birthweight and possible increased frequency of dystocia (Shelton, 1964; Thrift et al, 1967; Olson et al, 1976; Martin et al, 1980; Mavrogenis et al, 1980). However, selection for weights at later ages would be expected to lead to increased yearling weights which is desirable for meat animals, but may be associated with increased maintenance costs for breeding animals and those kept for wool production. Genetic correlations between growth rates and various weights and between growth rates themselves ranged from medium and negative to high and positive (-0.42 to 0.99). Statistically, most of the estimates were not significantly different from zero. Such (negative) estimates could be explained by the effect of compensatory growth obscuring underlying genetic relationships. Thus, in general, genetic correlations were low to high and positive. There is, therefore, potential for exploiting correlated response for most of these traits. Traits such as birthweight, weaning weight and pre-weaning growth rate could be incorporated in a selection index aimed at selecting for yearling weight since these traits are highly correlated with 1 2-months weight. However, care must be taken to avoid problems related to dystocia which may result due to selection for increased birthweights. Phenotypic correlations Phenotypic correlations between weights were all positive and generally high (0.19-0.74) (Table 1). As was the case for genetic correlations among weights, phenotypic correlations between adjacent weights were higher. The correlations tended to decrease as the time interval separating the observed weights increased. In the case of genetic correlations between growth rates and weights, and among growth rates themselves, the phenotypic correlations ranged from low and negative to high and positive values (-0.17 to 0.98). The fact that there exists a 524 negative (although low) correlation between pre-weaning growth rate and growth rate in the intervals weaning to six months and six to nine months (-0.16 and -0.06, respectively), may indicate that a faster pre-weaning growth rate because of substantial milk available tended to be followed by slower post-weaning growth rate. It seems logical to suggest, in general, that the low negative correlations which appear in some growth periods such as between weaning to six months and nine to twelve months (-0.08), and six to nine months and nine to 12 months (-0.16), are consequences of compensatory growth and not antagonistic as such. There was, as would be expected, a consistent positive phenotypic correlation between overall growth rate and all other growth rates of the lamb before yearling indicating common environmental effects. Conclusion From the results of this study, it is concluded that selection for weight and weight gain would best be based on post-weaning traits though the generation interval is likely to be slightly longer. Positive correlated response should be expected in other correlated traits due to the generally large and positive genetic correlations. Preliminary selection could also be conducted during the pre-weaning period since Inyangala (1989) reported favourable repeatability estimates for birth weight, pre-weaning growth rate and weaning weight. However, the high maternal influence on pre-weaning must not be ignored as it tends to mask the true genetic merit in the lambs. Acknowledgement This work received financial support from the Norwegian Agency for International Development (NORAD). Permission to use project data was granted by the Ministry of Livestock Development, Government of Kenya. The authors are indebted to the staff of the Sheep and Goat Development Project, Naivasha, for their support and cooperation. References Dass G S and Acharya R M. 1970. Growth of Bikaneri sheep. Journal of Animal Science 31(1):1-4. Dzakuma J M, Nielsen M K and Doane T H. 1978. Genetic and phenotypic parameter estimates for growth and wool traits in Hampshire sheep. Journal of Animal Science 47:(5):1014-1021. Ercanbrack S K and Price D A. 1972. Selecting for weight and rate of gain in noninbred lambs. Journal of Animal Science 34:713. Givens C S Jr, Carter R C and Gaines J A. 1960. Selection indexes for weaning traits in spring lambs. Journal of Animal Science 19:134. 525 Harrington R B, Brothers D G and Whiteman J V. 1962. Heritability of gain of lambs measured at different times and by different methods. Journal of Animal Science 21:78. Harvey W R. 1982. User's guide for LSML76. Mixed model least-squares maximum likelihood computer program. Ohio State University, Columbus, Ohio, USA. (Mimeograph) Hundley W S Jr and Carter R C. 1956. Heritability of growth rate and market grade in mutton lambs. Journal ot Animal Science 15:1226 (Abstract). Inyangala B A O. 1 989. Genetic and phenotypic parameters for growth traits of the Dorper and Dorper x Red Maasai sheep. MSc thesis, University of Nairobi, Nairobi, Kenya. Martin T G, Sales D I, Smith C and Nicholson D. 1980. Phenotypic and genetic parameters for lamb weights in a synthetic line of sheep. Animal Production 30:261 . Mavrogenis A P, Louca A and Robison O W. 1980. Estimates of genetic parameters for preweaning and postweaning growth traits of Chios lambs. Animal Production 30:271. Olson L W, Dickerson G E and Glimp H A. 1976. Selection criteria for intensive market lamb production: Growth traits. Journal of Animal Science 43:78. Osman A H and Bradford G E. 1965. Effects of environment on phenotypic and genetic variation in sheep. Journal of Animal Science 24:766. Rae A L. 1982. Breeding. In: Coop I E (ed), World animal science. C. Production-system approach. 1. Sheep and goat production. Elsevier Scientific Publishing Company, Amsterdam, The Netherlands, pp. 15-55. Shelton M. 1964. Relation of birth weight to death losses and to certain productive characters of fall-born lambs. Journal of Animal Science 23:355. Stobart R H, Bassett J W, Cartwright T C and Blackwell R L. 1986. An analysis of body weights and maturing patterns in Western Range ewes. Journal of Animal Science 63:729. Swiger L A, Harvey W R, Everson D O and Gregory K E. 1964. The variance of intraclass correlation involving groups with one observation. Biometrics 20:818. Thrift F A, Whiteman J V and Ovajera A A. 1967. Evaluation of udder scoring in a ewe flock. Journal of Animal Science 26:206. Thrift F A, Whiteman J V and Kratzer D D. 1973. Genetic analysis of preweaning and postweaning lamb growth traits. Journal of Animal Science 36:640. Vogt D W, Carter R C and McClure W H. 1967. Genetic and phenotypic parameter estimates involving economically important traits in sheep. Journal ofAnimal Science 26:1232. 526 Milk yield of Cameroon Dwarf Blackbelly sheep R.M. Njwe and Y. Manjeli Department of Animal Production, Institut National du Développement Rural (INADER), University Centre of Dschang, B.P. 222, Dschang, Cameroon Abstract Milk yield of Cameroon Dwarf Blackbelly sheep was estimated by the indirect method of weighing lambs before and after sucking during a period of 12 weeks. Twenty-two lactations were evaluated between February and December 1985. Average daily milk yield ranged from 0.31 to 0.50 kg for ewes suckling single lambs. Maximum weekly milk yield (3.52 kg) was attained during the 4th week of lactation after which production declined. Differences in milk yield between ewes and weeks of lactation were highly significant (P<0.01). Milk yield per lactation was positively correlated with the age of the ewe at the beginning of lactation. Production laitière de moutons Djallonké au Cameroun R.M. Njwe et Y. Manjeli Résumé La méthode de la double pesée d'agneaux (avant et après la tétée) a été utilisée pour déterminer la production laitière de brebis Djallonké à ventre noir camerounaises. Vingt deux lactations de 12 semaines de durée ont été estimées entre février et décembre 1985. La production journalière moyenne fut comprise entre 0,31 et 0, 50 kg pour des brebis allaitant un seul agneau. La semaine de production maximale fut la 4e après la mise-bas (3, 52 kg); la production a ensuite décliné. La production laitière a varié de manière significative (P<0,01) d'une brebis à l'autre et d'une semaine de lactation à l'autre. La production totale par lactation estpositivement corrélée avec l'âge des brebis au début de la lactation. 527 Introduction Sheep are generally not milked in Cameroon. However, it has been established that milk production has direct effect on weight gain by lambs. Barnicoat et al (1956) reported a positive correlation between growth rate of New Zealand Romney lambs (between 0 and 12 weeks) and fat content, and total milk consumption with correlation coefficients of 0.72 and 0.61, respectively. Milk production during lactation is an important factor affecting maternal ability in sheep. Estimates of milk production by Cameroon Dwarf Blackbelly sheep do not exist. The present study focused on estimating daily and total milk production during the lactation period of Cameroon Dwarf Blackbelly sheep; establishing the lactation curve; and investigating the relationship between age of animal at the commencement of lactation and total milk production during the lactation period of 12 weeks. Materials and Methods Animals The sheep used for the study are popularly called Blackbelly sheep as a result of the distinct black abdomen and dominant brown colour on all other parts of the body. They are a strain of the West African Dwarf sheep found along the Atlantic coast from Senegal to Angola. This sheep is a hair type with a thin tail. It measures 55 to 65 cm at the withers. The head is small with a straight profile. The ears are small and droop a little. Males have short spiral horns. In general, females are hornless but sometimes they have spurs. The hair is short and quite smooth but is sometimes rough. Adult males weigh 25-35 kg while adult females weigh 20-25 kg. Birthweight ranges from 1.5 to 2.0 kg. Average age at first lambing is 534 days while the lambing interval is 228 days. The sex ratio is about 49 per cent male and 51 per cent female. Prolificity has been estimated at 107 per cent while fertility and fecundity rates are 1 13 and 121 per cent, respectively. This sheep is thought to be the ancestor of the Barbados Blackbelly sheep (Epstein, 1971). It is found in the forest region of Cameroon, Gabon, Congo and Zaire (FAO, 1 979) . Mason (unpublished data) reported that 1 0.5 per cent of sheep in the Central and South Provinces of Cameroon were the Blackbelly sheep. The sheep had been collected from the East and South Provinces of Cameroon in 1980 and maintained at the University Farm at Nkolbisson near Yaounde until 1 984 after which in December 1 984 they were transferred to Dschang on the high plateau (with an altitude of 1450 metres, annual rainfall of 2200 mm and average temperature of 20-22°C). 528 Feeding of lactating sheep The experimental period was between February and August 1985. Lactating sheep were fed cut fresh Pennisetum purpureum daily with concentrate supplement. The composition of the supplement is presented in Table 1 . Enough forage was provided to each lactating ewe such that a residue of about 20 per cent was always left. Each animal received 200 grams of concentrate supplement which was served at 08.00 hours everyday. Table 1. Composition of concentrate and Pennisetum purpureum fed to lactating Cameroon Blackbelly sheep. Concentrate* Pennisetum purpureum" Dry matter 88.05 25.60 Crude protein 19.36 10.59 Crude fibre 10.38 28.70 Ether extracts 4.80 2.40 Nitrogen-free extracts 54.87 49.96 Ash 10.59 8.35 * 100 kg concentrate composed of 50 kg maize, 30 kg wheat bran, 18.5 kg soybean and 1 kg bicalcium phosphate and 0.5 kg salt. ** Mean composition of P. purpureum samples for entire experimental period. Assessment of milk production The assessment of milk production started after the first three days of lactation to enable lambs suckled colostrum. Lambs were kept separated from ewes and only brought together at periods of suckling. Milk production of the ewes was estimated by the indirect method of weighing lambs before and after suckling using a sensitive balance. Estimates were carried out on alternate days up to weaning at 12 weeks. During the first four weeks lambs were allowed to suckle at 2-hour intervals between 8 a.m. and 6 p.m. Between the 5th and 12th week of lactation, four suckings at 3-hour intervals between 8 a.m. and 5 p.m. were allowed. Milk production data collected for four days of the week were extrapolated for the week of seven days. Weekly milk production was assessed for 22 ewes during lactation periods of 12 weeks. The data were subjected to analysis of variance using the randomised complete block design where week of lactation represents treatments and each ewe was a replicate. Total milk production of ewes was correlated with age to determine the relationship between them. 529 Results Weekly milk production by Cameroon Dwarf Blackbelly sheep are presented in Table 2. From parturition there was an increase of milk production from 2.95 kg per week to a peak of 3.52 kg at four weeks and thereafter a gradual decline to 2.10 kg at 12 weeks of lactation. Table 2. Mean weekly milk production by Cameroon Blackbelly sheep. Total milk Number of Average milk Lactation week production (kg) observations production (kg) 1 64.91 22 2.95 2 73.67 22 3.35 3 76.95 22 3.50 4 77.46 22 3.52 5 70.68 22 3.21 6 65.07 22 2.96 7 64.44 22 2.93 8 62.97 22 2.86 9 58.31 22 2.65 10 55.42 22 2.52 11 48.89 22 2.22 12 46.25 22 2.10 Average milk production per lactating ewe is indicated in Table 3. Total milk production per lactation ranged from 28 to 42 kg, corresponding to 0.33 to 0.50 kg per day. Age at the time of parturition (days) was positively correlated with total milk production (kg) and the correlation coefficient of 0.71 was significant (P<0.05). The regression equation describing the relationship was as follows: Y = 21.497 + 0.021x(r = 0.71) where Y = milk production per lactation (kg) X = age of ewe at parturition (days). 530 Table 3. Milk production by Cameroon Dwarf Blackbelly sheep during the first three months of lactation. Age of animals Total milk Average daily Animal number (days) production (kg) production (kg) 1 na 27.79 0.33 2 na 32.40 0.39 3 528 26.00 0.31 4 707 36.81 0.44 5 844 42.15 0.50 6 na 34.03 0.41 7 na 35.00 0.42 8 597 37.59 0.45 9 na 37.06 0.44 10 na 39.59 0.48 11 628 37.06 0.44 12 na 39 59 0.48 13 na 37.18 0.44 14 847 31.27 0.43 15 435 34.21 0.41 16 635 33.33 0.40 17 na 34.21 0.41 18 na 30.63 0.36 19 na 31.47 0.37 20 na 37.12 0.44 21 758 36.33 0.43 22 877 40.18 0.48 na = not available. Discussion Maximum daily milk production was attained by the Cameroon Dwarf Blackbelly sheep at the 4th week for lactation. This implies a more gradual increase of milk production to attain a peak when compared to some breeds for sheep that do so more rapidly and earlier. Coombe et al (1960) reported that milk output reached a peak at 10-20 days; while Ricordeau and Denamur (1962) reported that Prealpes sheep attained milk production peak at 25 days. The daily peak milk production of 0.50 kg obtained with Cameroon Dwarf Blackbelly sheep is low when compared to 1 .25 litres per day with Prealpes sheep (Ricordeau and Denamur, 1962). Spedding (1963) has indicated that peak milk yield in sheep varies with breeds. Bonsma (1929) reported peak daily milk 531 production in Merino sheep of 0.77 kg while Owen (1957) indicated a value of 2.23 kg. The yield of 28-42 kg of milk by Cameroon Dwarf Blackbelly sheep during a lactation period of 96 days is low when compared to 48 and 53 kg by Ossimi and Rahmani sheep, respectively (Sharafeldin and Mostageer, 1961), for a lactation period of 84 days and 100-150 kg by Awassi sheep for a period of 160 days (Koseoglu and Aytug, 1961). The positive relationship between age and milk production obtained in the present study with Blackbelly sheep agrees with reports by Gruer (1 959) that total milk production per lactation by Precoce, Stara, Zagora and Plovdiv sheep increased upto six to seven years before declining. However, the oldest ewe used in the investigation of the age/lactation total milk relationship had not attained three years. The evaluation of milk production needs to be extended to other local breeds for which we do not have any information at the moment. References Barnicoat C R, Murray P E, Robert E M and Wilson G S. 1956. Milk secretion studies with New Zealand Romneyewes. Part VIII. Experiments on early weaning of lambs. Journal of Agricultural Science 48 (Part 1):22-24. Bonsma F N. 1 929. Factors influencing the growth and development of lambs, with special reference to cross-breeding of Merino sheep for fat lamb production. University of Pretoria Agriculture 48. University of Pretoria, South Africa. Coombe J B, Wardrop I D and Tribe D E. 1960. A study of milk production of the grazing ewe, with emphasis on the experimental technique employed. Journal of Agricultural Science 54:353. Epstein H. 1971. The origin of the domestic animals of Africa. Volume 2. Africana Publishing Corporation, New York, USA. FAO (Food and Agriculture Organization of the United Nations). 1979. FAO production year-book 1978. Volume 32. FAO, Rome, Italy 287 pp. Gruer V. 1959. Correlations between milk production, wool yield and body weight in sheep. Mez.Sel.hoz. z. 3(2):109- 118. Koseoglu H and Aytug C N. 1961. Studies on the milk production of Awassi sheep which are breeding at Cukurova Stock Farm. Lalahan Zootek. Arast. Enst. Derg. 1(10): 100-1 10. Owen J B. 1957. A study of the lactation and growth for hill sheep in their native environment and under lowland conditions. Journal of Agricultural Science 48:387-412. Ricordeau G and Denamur R. 1962. Production laitier des brebis Prealpes du sud pendant les phases d'allaitement, de sevrage et de traite. Annates de Zootechnie 1 1 (1):5-38. Sharafeldin MA and Mostageer A. 1961. Suckling in Ossimi and Rahmani lambs. Journal of Animal Production (U.A.R.) 1:53-59. Spedding CRW. 1963. Sheep production and grazing management. Bailliere,Tindalland Cox, London, UK. 532 Session 6 Small ruminant development in Africa: implication for research Développement de l'élevage ovin-caprin en Afrique: consequences pour la recherche Research highlights from the Small Ruminant Collaborative Research Support Program: The Dual-Purpose Goat (DPG) P.P. Semenye, A.N. Mbabu, B. Mwandotto, F.B. Nyaribo, J.M. Onim and F. Rurangirwa Small Ruminant Collaborative Research Support Program P 0 Box 252, Maseno, Kenya Summary The Small Ruminant Collaborative Research Support Program (SR-CRSP) in Kenya is engaged in a multidisciplinary research programme with a farming-system perspective. Its overall objective is to carry out research on dual-purpose goat production systems, to enhance production of food protein through meat and milk, and to improve cash flow through sale of goats and products in small-scale intensive farming systems. Research highlights are presented on feed production, preservation and utilisation, goat production and management, development and performance testing of a composite dual-purpose goat breed, immunoprophylaxis and diagnosis of contagious caprine pleuropneumonia (CCPP), heartwater and helminthiasis. The economy of raising dual-purpose goats, their acceptability and the utilisation of theirproducts by farmers in western Kenya are also highlighted. La recherche au Small Ruminant Collaborative Research Support Program du Kenya: la chevre a aptitudes mixtes lait/viande P.P. Semenye, A.N. Mbabu, B. Mwandotto, F.B. Nyaribo, J.M. Onim et F. Rurangirwa Resume Le Small Ruminant Collaborative Research Support Program du Kenya (SR-CRSP) est un programme de recherches 533 pluridisciplinaires dont les travaux sont essentiellement axés sur l'étude des systèmes de production. Son mandat consiste à effectuer des recherches sur les systèmes d'élevage caprin en vue de promouvoir la production de lait et de viande et partant d'augmenter le niveau d'alimentation protéique des petits éleveurs tout en accroissant le revenu monétaire de ces systèmes de production intensifs sur de petites exploitations grâce à la vente d'animaux sur pied, de viande et de lait. Les principaux résultats obtenus sont présentés, qui concernent notamment la production, la conservation et l'utilisation des aliments du bétail, la conduite des troupeaux caprins, la mise au point d'une race composite de caprins destinée à la production de lait et de viande et l'évaluation de ses performances ainsi que l'immunoprophylaxie et le diagnostic de la péripneumonie contagieuse des caprins, de l'hydropéricardite et des helminthiases. Il aborde également les aspects économiques de ce type d'élevage, l'accueil qui lui est fait par les paysans de I ouest du Kenya ainsi que l'utilisation qu'ils font des produits ainsi obtenus. Introduction The twofold objectives of the Small Ruminant Collaborative Research Support Program (SR-CRSP) in Kenya are: (1) to carry out research on dual-purpose goat (DPGs) to develop production systems which would contribute to increased production of food protein through milk and meat and improved cash flow through marketable goats and goat products in small-scale intensive farming systems of Kenya, and (2) to provide training which would enhance the long-term capability for research and development with goats in Kenya and in other tropical countries. With this approach, it has been necessary to characterise existing farming systems to determine how DPG component can be introduced and sustained in western Kenya production systems. The research problem, therefore, has been to develop an animal of appropriate genotype, a heafth-nutrition-management package appropriate to the small farm resource base, and to ensure the component is economically and sociologically acceptable. Towards these goals a holistic and multidisciplinary approach has been taken by SR-CRSP, Kenya. Research, therefore, has been conducted on-station and on-farm in testing and evaluating the technical feasibility of DPGs including their management and nutrition, as well as their economic viability and social acceptability. On-farm research is conducted on 150 small-scale farms in Siaya, Kisumu and Kakamega districts. What follows are research highlights from on-farm and on-station trials conducted by SR-CRSP, Kenya. 534 Results and Discussion Constraints Following farming system characterisation, constraints were identified that would affect successful introduction and management of DPGs. They are appropriate genotype, shortage of pastures due to land pressure for production of staple and cash crops, diseases and lack of husbandry and production performance measures of DPGs. These constraints are not unique to western Kenya, but are found in many other areas with mixed farming practices. The approach undertaken to resolve them can be extrapolated in other areas. The tollowing research efforts have been undertaken to resolve the above constraints. Appropriate genotype As no indigenous nor exotic breed fits the dual-purpose goat production characteristics ideal for Kenya, a new breed is necessary. Towards this goal indigenous goat breeds were evaluated for dairy potential and then strains and individuals from within a breed were selected for breeding with proven exotic dairy breeds. A composite dual-purpose goat was planned from a synthesis of two local goat breeds, the Galla (G) and the East African Goat (EA) and two exotic breeds, the Toggenburg (T) and Anglo Nubian (N). The breeding plan for evolving a composite dual-purpose goat breed is justified in performance as shown in Table 1 , whereby F1 's and four-way crosses outperform the Galla and East African Goat in milk production. Table 1 . Daily milk yield least squares means ± SE (g) *. Breed type ** Mean SE East African Goat 88.9 207.9 Galla 542.4 143.2 F1's 709.0 98.4 4-way crosses 1077.8 153.0 * After Mwandotto et al (1990). ** F1's and 4-ways-pooled. Feed shortage Following on-farm determinations of dry-matter intake, it is clear that the most limiting nutritional factor is quantity rather than quality of available forages. This claim is substantiated in Table 2 which shows crude-protein (CP) content and dry-matter digestibility (DMD) of dietary component are above average, while dry-matter intake (DMI) is below average for a pregnant or lactating doe. 535 Table 2. Mean nutritive values of the diet of dual-purpose goats in western Kenya. DM% CP% DMD% DMIkg %BWT ME N 36 36 36 36 36 36 Mean 28 16 55 0.7 2.4 1.8 SD 12 3 6.5 0.3 0.8 0.6 Due to high population densities, up to 900 people per sq km in some areas in western Kenya, sub-division of farms has substantially reduced pasture and fallow lands rendering reduction in cattle population (Russo et al, 1983). Lack of adequate pasture has led to intensive management methods such as tethering and stall-feeding. In response to this challenge SR-CRSP scientists are solving the problem through integration of DPG with crop agriculture. This is achieved through dual-purpose crops, crop residues and by growing fodders and legume trees in alley cropping or soil conservation practices. Table 3 gives an indication of the quality of some of the forages derived from dual-purpose crops and crop residues. In testing the hypothesis that integration of DPG with crop production would result in adequate production of forage for goats, a model farm was established at Maseno. The farm is 0.5 ha in size and is designed to produce enough major food crops for the farmer's needs as well as feeds for four goats or one adult cow. Food crops (maize, beans and sweet potato) one adult cow. Food crops (maize, beans and sweet potato) are planted on 60% of the land, while forages, Napier grass, two species of Leucaena, Sesbania and Calliandra, that are planted both as borders and alleys took 40% of the land. In one year this farm produced food crops and forage as shown in Table 4. Four DPGs were exclusively fed from the model farm forages for the whole year with large surpluses realised every month. Diseases The main goat diseases in western Kenya and in Kenya at large are contagious caprine pleuropneumonia (CCPP), heartwater and haemonchosis. SR-CRSP has had success in these diseases in vaccine and diagnostic kit development. Contagious caprine pleuropneumonia In an endeavour to evaluate the inactivated and lyophilised CCPP vaccine developed by SR-CRSP and Kenya Agricultural Research Institute (KARI) scientists, 50 goats were immunised with a single dose of vaccine and 50 others 536 Table 3. Chemical composition and digestibility values (mean + SE) of forages derived from dual-purpose crops and crop residues in western Kenya. Forage CP % NDF% ADF % ADL % IDMD % Sugar-cane tops 7.4+3.1 73.1±8.7 37.5+9.6 3.7+1.9 42.6±5.2 (12) (10) (10) (10) (10) Maize stover 5.4±2.7 69.9±10.1 37.6±±9.1 12.1±2.0 55.3±7.4 (401) (52) (49) (43) (31) Maize thinning? 20.7±4.5 63.5±6.9 (151) " (15) Maize fresh leaves 13.5+5.1 72.8±13.1 35.4±16.6 14.4±12.6 59.0±8.0 (92) (21) (18) (16) (10) Finger millet 7.4±2.4 71.7±4.7 47.3±4.5 10.0±3.0 50.8+6.4 (stover) (21) (9) (9) (9) (13) Sorghum stover 12.6±1.8 69.2±4.2 43.7±5.4 22.8±2.3 50.3 (serena) (9) (6) (6) (6) (1) Cassava fresh 17.6±2.7 39.2±3.9 29.6±6.6 leaves (10) (29) (29) Rice stover 4.1 ±1.1 40.7±8.2 (5) (3) Sweet potato 17.9±3.2 51.5+9.6 34.3+4.0 18.5±6.5 62.6±9.8 vines (100) (22) (21) (19) (15) Bean haulms 7.2±2.5 42.3±6.5 25.8±8.0 12.9±5.2 54.7±7.7 (100) (19) (13) (20) Pigeon pea fresh 25.1 ±1.9 56.5±7.3 - leaves [30] (23) (3) Note: ( ) Number of samples analysed, — analysis not done. Table 4. Food and forage output from a model farm on- station*. Size 1/2 ha. Food crops Maize beans and sweet potato. Area taken 60% Forage Napier grass Food production 1908 kg or 21 (90 kg) bags of maize 50 kg dry beans 273 kg sweet potato tubers Forage production 26 kg DM - fresh maize thinnings 505 kg DM - fresh maize leaves 5286 kg DM - sun-dried maize stover 53 kg DM - bean haulms 2526 kg DM - fresh Napier grass 1303 kg DM - fresh forage - Leucaena 80 kg DM - fresh sweet potato vines * Two cropping seasons per year. 537 were left as unimmunised controls. All the goats were left on the same farm for two months and then moved to Kabete Veterinary Laboratory farm for challenge. The two groups were mixed and allowed to graze during the day and at night they were housed with five goats with experimental challenge of CCPP. During the first two weeks after movement of the goats to Kabete, 12 vaccinates and 14 unimmunised goats died of diarrhoea. None of these goats had evidence of CCPP infection. Of the remaining 38 vaccinated goats, two became infected with CCPP and had a fever for two days. None of the 38 vaccinated died. At the end of the experiment (four months later), all 38 goats were euthanised and examined for lesions and organisms. Only the two with fever had mild adhesion of the cardiac lung lobes to the thoracic wall and no F38 organisms were isolated from any of the 38 vaccinated goats. In contrast, 30 of the 36 unimmunised goats developed fever and 27 died of CCPP. The three unimmunised goats that developed fever and survival had adhesions of the cardiac and apical lobes of the lungs to the thoracic wall. The six control goats without clinical signs did not have lesions or detectable organisms at the end of the experiment. In summary, 75% of the unimmunised goats died when challenged by contact with CCPP-infected goats and none of the vaccinated goats died. Heartwater Heartwater is a severe infection of goats, sheep and cattle caused by rickattsia (Cowdria ruminantium). In goats, the disease causes a very high mortality and can be devastating in susceptible populations. Transmission is by ticks, primarily Ambloyomma variegatum, which are widely distributed in Kenya. Currently, diagnosis is made by post-mortem examination of brain tissue. A more convenient method of diagnosis is needed as well as a method of prevention. A published method was used to grow C. ruminantium in vitro in bovine endothelial cells. Establishment of the tissue culture system at Kabete made available a renewable source of organisms. Larger quantities of organisms provided Deoxyribonucleic Acid (DNA), protein and infectious organisms for a variety of studies and represents an important advance in our research program on heartwater. Endothelial cells were infected with Kiswani isolate of heartwater, made in Kenya. DNA purified from C. ruminantium organisms isolated from infected endothelial cells was used to make genomic library. Size-fractionated fragments were ligated into puc19 and used to transform Escherichia coli. Two cloned genes were identified that were from the heartwater organism and one of the genes pcr9, a 744 base pair (bp) fragment has been characterised as DNA probe. Cloned C. ruminantium gene was isolated from plasmids, labelled with biotin and the specificity evaluated. The 744 bp insert from the recombinant plasmid pc* hybridised to C. rum/nanf/urn-infected endothelial cell DNA. The 744 bp insert 538 DNA probe did not hybridise to DNAs from uninfected endothelial cells, leukocytes, Babesia bigemina, Mycoplasma sp F38 Anaplasma marginale, A. ovis, A. centrale, Chlamydia psittacci, Chlamydia trachomatis, Ehrlichia risticii and 1 1 other bacteria. The 744 bp pc's DNA probe detected C. ruminantium in Amblyomma variegatum ticks fed on C. ruminantium -infected goats but not in ticks fed on uninfected goats. Tick infection with C. ruminantium was confirmed by transmission of headwater to uninfected goats either by feeding of nymphs fed as larva on infected goats or by subinoculation of tick gut homogenates. The pcr9 insert DNA probe detected C. ruminantium in midguts of A. variegatum nymphs infected as larvae and in midguts of A. variegatum adult ticks infected as nymphs but not in midguts from control nymphs and adults. Nucleic acid sequence of this pct9 insert DNA was determined for eventual development of oligonucleotide primers for use in polymerase chain reactions in order to increase sensitivity of assays using the DNA probe. Haemonchosis Research on haemonchosis focused on the identification of genetically resistant or tolerant goats as one long-term solution to control this important internal parasite. This research emphasis is based on the findings that several expensive treatments per year are required to prevent clinical disease and associated production losses. The assumption that genetically resistant or tolerant goats can be identified is based on the identification of a resistance trait in sheep and on our work that indicated susceptibility differences occurred among breeds of goats. Natural Haemonchus contortus infection of 300 kids was evaluated from two months to one year of age. Faecal samples were collected from the rectum and parasite eggs per gram of faeces determined. During the observation period, some kids had very low or no eggs while others had high egg counts. To extend the observation of differences among individuals to natural infection, two groups of goats were selected for further study. Group A contained 1 1 goats which had a mean egg count of 201 ±200. The two groups were confined in pens, cleared of worm infection using Ivermectin and challenged with 500 H. contortus larvae per kg bodyweight. Faecal samples were collected once a week for eight weeks and eggs per gram determined. Group A goats had significantly (P<0.05) lower eggs per gram (mean 726±213) than group B (mean 1643+463). Also, individual goats in both groups had low numbers of eggs indicating resistance to the challenge infection. Some animals in group B had very high infection rates indicating that individual differences may be greater than group differences. This study emphasises the importance of experimental challenge to assess genetic resistance of goats to H. contortus and that individual differences in susceptibility occur among goats. Further studies are in progress to define more clearly the occurrence and basis for differences in susceptibility. 539 On-farm evaluation As the name implies, on-farm evaluation calls for the researcher to take measurements on-farm and for interaction with the farmer, on the subject of the technology or intervention with regard to its biological and socio-economic aspects. In looking at the biological aspect, the on-farm performance of DPGs was compared with on-station over a period of two years. Results of the comparison are given in Table 5. Judging by these results, it is evident that the on-station DPGs are out-performing their counterparts on-farm. Causes leading to these differences give guidance for further investigations and trials. For example, the kidding rate on-station is 98% as opposed to 65% due to managerial rather than biological problems. Farmers are not fully conversant with heat detection and buck availability is limited to a few farms. This is not a researchable problem but one that requires education and more bucks. Table 5. Comparison of on-station with on-farm performance of dual-purpose goats in western Kenya. Variable On-station On-farm Number of does Doe survival Kidding (%) Kid survival (%) Kid weight at the age of one year Milk offtake (kg) Average doe weight (kg) Productivity/doe/year (kg) Productivity/metabolic body weight/year (kg) What is the verdict of farmers collaborating in SR-CRSP on-farm research? So far, research findings indicate that SR-CRSP DPG technology packages (techpacks) have been most enthusiastically taken up in the relatively high altitude (1630 m), high rainfall (1800 mm) (Onimetal, 1986), and high population density (700-1000 per km2; Kenya, Republic, 1979) clusters of Muhanda and Hamisi. Conversely, the techpacks have been least enthusiastically taken up in the relatively low altitude (1300 m), low rainfall (1100 mm), and low population density (200-400 per km^ clusters of Masumbi and Lela. For example, both Muhanda and Hamisi have consistently maintained a high proportion of their participating farmers growing all the recommended forages as well as, on average, the highest number of forage trees and a substantial area devoted to production of Napier grass and sweet potato vines (Table 6). Masumbi cluster, and to a certain extent Lela cluster lag behind on both counts. 200 150 95 80 98 65 90 85 20 18 120 60 38 35 27 11 1.8 0.8 540 o -c to 3 o -c CD a a o a 03 CD W c to CO 0) a c CO T3 C ca .CO 2 CO -Q CO CO CO■Q C CO CO c CD CO U CO 8,2 g£ cd a »t CO a CO Z 10 en CM ^3- CO eg in in CO CO > a. w CO T m o> CO CO CO O) ci eg CO n .CO 2 CO rQ co CD to 2 o Tf CO in C\J 1O o to a> o CO s> 3 a o 8 CO A .co E E CO CO CO X 2 3 o XI 541 N Cut-and-carry Open-grazing 21 70 10 14 54 16 34 100 0 27 69 24 10 46 73 Similarly, both Muhanda and Hamisi clusters are noted to be most consistent in feeding the cultivated forages to the DPGs. By contrast, Masumbi farmers merely feed their DPGs with cuttings from growing on shrubs fallow land while Lela farmers prefer to "open graze" their DPGs rather than feed them with large amounts of cultivated forages in the cluster (Table 7). Table 7. Proportion of participating farmers doing 'cut-and-carry" or "open-grazing" Cluster Hamisi Masumbi Muhanda Rabuor Lela Linear programming (LP), a procedure for determining efficient resource allocation which uses mathematical procedures consistent with economic concepts to determine optimal farm plans, showed the following based on the results from Table 7, with regards to DPG enterprise (Nyaribo, 1989): • Dairy production generated higher net returns than food crop production. This makes sense since milk and meat are both high (cash) value commodities. • The DPG's market potential is very good and can easily out-compete the local zebu cattle. • As we move from the medium to the large farm, there is a modest 7 per cent increase in livestock sales, but a significant 31 per cent increase in net farm income. This means that with additional factors of production (capital, land and management) farmers can increase their incomes significantly through livestock sales. It is interesting to note that there were no livestock sales in the small Hamisi farms (0.69 ha) of which only 0.59 ha was cultivable land while the LP model production is at variance with actual observed practices since it does imply that the smallest farms in the area keep livestock under non-optimal conditions for risk aversion reasons and other reasons not related to profitability. In fact, the small farms have been observed to be more attracted to the DPGs and its techpacks than larger farms that are more in commodity agriculture. It is anticipated, with increased production level of the DPGs and particularly, with milk production of over one kg per doe/day, the larger farms are likely to capitalise on it as they are currently doing with cash crops. 542 Despite these trends, on-farm research results indicate that DPGs contribute positively towards nutritional and economic welfare of participating farmers in western Kenya. From an average of two does a small supply of milk has been produced and used as a nutritious or medicinal drink, in making tea, and vegetable and fish sauces. Culled does and some of their followers have been slaughtered for home consumption, while others have been sold for cash to meet essential needs, for example, paying school or medical fees, buying clothes and farm inputs. In addition, DPG manure has increased productivity of crop land resulting in higher yields of maize and other crops. Acknowledgements We gratefully acknowledge material and financial support from Kenya Agricultural Research Institute (KARI) and the United States Agency for International Development (USAID) Grant No. DAN 1 328-G-SS-4093-00 and technical assistance from collaborating scientists and farmers from western Kenya. We also express our sincere thanks to Ms. Philomena Pacha for word-processing the final draft. References Kenya, Republic. 1979. Population census. Volume II: Analytical report. Central Bureau of Statistics, Ministry of Finance and Planning, Nairobi, Kenya. MwandottoBAJ, AhuyaCO, Ruvuna F and Taylor. 1990. Milk yield potential of the Kenya dual-purpose goat. In: Proceedings of the Eighth Small Ruminant-CRSP Workshop, Nairobi, Kenya. Small Ruminant Collaborative Research Support Program, Nairobi, Kenya. Nyaribo F B. 1989. Integrating dual-purpose goats on small scale farms in PTO western Kenya: A linear programming analysis. PhD dissertation, Washington State University, Washington, USA. Onim JFM, Mathuva M N, Hart R, Fitzhugh H A and Otieno K. 1986. Recommendation domains for dual purpose goat research in Nyanza and western provinces of Kenya. In: Proceedings of the Fifth Small Ruminant-CRSP Workshop, Nairobi, 4-6 November 1986. Small Ruminant Collaborative Research Support Program, Nairobi, Kenya. Russo S, Hart R, Otieno K, Owino J and Onim J. 1983. Feed production research for smallholder agriculturalists in western Kenya. Paper presented at CIMMYT Networkshop on Draft Power and Livestock Feeding in Eastern and Southern Africa, 4-6 October 1983, Mbabane, Swaziland. 10 pp. 543 Programme de développement de l'élevage des petits ruminants au Togo: essai de mise au point des possibilités d'association de la production vivrière et de l'élevage ovin I.Y. Pessinaba Centre d'appui technique de Kolokopé B.P. 65, Atakpame, (Togo) Résumé Afin de contribuera la satisfaction de la demande en viande, le Togo a instauré un Projet national petits ruminants dont les trois volets sont présentés: l'encadrement du secteur traditionnel, la création d'élevages semi-intensifs, et le développement du centre d'appui de Kolokopé. Les recherches menées par le projet visent à l'intégration de l'agriculture vivrière et de l'élevage ovin- caprin : valeur alimentaire des sous-produits de cultures, qualité des savanes, utilisation du fumier. Small ruminant development programme in Togo: Effort to integrate crop and sheep production I.Y. Pessinaba Abstract A national small ruminant development project has been set up in Togo with the aim of increasing meat production. It has three components: extension activities within the traditional farming systems; creation of semi-intensive production units; and development of the supporting centre of Kolokope. Research carried out focuses on the integration of small ruminants production with arable, in particular on the feeding value of agricultural by-products, savannah quality and manure utilisation. 545 Introduction Bien que l'élevage occupe au Togo par rapport à l'agriculture une place modeste dans l'économie nationale, son développement constitue une des conditions nécessaires pour atteindre l'objectif que s'est fixé le pays depuis une vingtaine d'années: l'autosuffisance alimentaire. Dans ce cadre, un Projet national petits ruminants a été mis en place au Togo, dont nous présenterons ici les principales activités. Malgré les tentatives et efforts de développement de l'élevage des ovins et caprins, force est de constater que les paysans ont du mal à associer cette activité aux cultures vivrières. Nos recherches sont donc axées sur l'étude des éventuels avantages mutuels entre la culture vivrière et l'élevage des petits ruminants au niveau du paysan-éleveur et de définir dans quelles mesures nous pourrons les valoriser davantage. Une synthèse de ces résultats est présentée en deuxième partie. Programme de développement de l'élevage des petits ruminants au Togo Pour atteindre les objectifs lui étant assignés, réduire le déficit en viande du Togo et augmenter les revenus paysans, le Projet national petits ruminants (PNPR) comprend trois volets: celui de l'encadrement du secteur traditionnel, celui de la création et du suivi des élevages semi-intensifs, et le Centre d'appui technique de Kolokopé. Encadrement L'encadrement du secteur traditionnel vise à l'amélioration des pratiques de l'élevage traditionnel dans les domaines de l'alimentation, de l'habitat et de l'hygiène, ainsi qu'à la protection sanitaire contre la peste des petits ruminants et le charbon bactéridien (vaccination), et d'autre part contre les parasitoses gastro-intestinales et les ectoparasites (déparasitages). Les actions entreprises au niveau du secteur traditionnel sont reprises auprès des élevages semi-intensifs et intensifs avec un suivi plus marqué auxquelles s'ajoutent celles de l'amélioration génétique. Les actions d'encadrement et de suivi (sanitaire et zootechnique) touchent 1220 élevages villageois répartis dans les régions maritimes, des plateaux, centrales et des savanes, qui comportent des troupeaux de 1 0 à 20 têtes; et 250 élevages semi-intensifs répartis dans les mêmes régions économiques du pays et comportant des troupeaux de 25 à 200 têtes. 546 Les campagnes de vaccination et de déparasitage ont permis une baisse moyenne du taux de mortalité de 25% à 8%, alors que les actions d'amélioration de l'alimentation, de l'habitat, de l'hygiène et de la reproduction ont permis d'enregistrer une amélioration du taux de productivité, qui est passé de 81% à 134%. Dans le nouveau programme, ces actions seront poursuivies en considérant trois niveaux d'encadrement: Niveau 1 : encadrement sanitaire du secteur traditionnel. Il correspond à la protection sanitaire du cheptel promue par des campagnes de sensibilisation (radio, affiches), à la vaccination et au déparasitage, ainsi qu'à un conseil zootechnique élémentaire (castration, habitat, complémentation minérale). Niveau 2: encadrement zootechnique du secteur traditionnel. Il vise à accroître la rentabilité des élevages pratiquant déjà les actions du niveau 1 . Il les renforce et ajoute les thèmes de l'amélioration, de l'habitat (parcs et abris traditionnels) et de l'alimentation (pâturages, complémentation et supplémentation, mangeoire, etc.). Niveau 3: établissement de fermes ovines améliorées (FOA). Il vise la poursuite de l'implantation et du développement des élevages semi-intensifs qui sont sélectionnés après une période de test dans le niveau 2 d'encadrement. Les cultures fourragères sont introduites et un système de prêts mis en place: prêt d'installation (construction et équipement de bergeries) remboursable en espèces en quatre ans, crédit de fonctionnement (aliments, médicaments sur douze mois) remboursable en espèce en dix-huit mois (8% d'intérêt annuel), et prêt d'animaux reproducteurs retournables en animaux après un différé de deux ans. Le Centre d'appui technique du Kolokopé L'effectif actuel du Centre est d'environ 2 000 têtes réparties comme suit: 5 troupeaux totalisant 760 brebis reproductrices; 1 troupeau de 40 géniteurs mâles confirmés; 4 troupeaux constitués de jeunes mâles et femelles en croissance. Le Centre d'appui technique de Kolokopé s'est vu assigné comme objectifs la production et la sélection des géniteurs améliorés et leur diffusion dans les élevages encadrés, ainsi que l'exécution de programmes de recherche appliquée pour préparer les thèmes à vulgariser. La sélection des animaux améliorateurs est basée essentiellement sur la vitesse de croissance des jeunes mâles et sur la conformation des jeunes femelles. Les mâles sont livrés pour la reproduction à l'âge de 12 mois avec un poids de 26 à 27 kg, tandis que les femelles sont mises à la reproduction à l'âge de 8 mois, avec un poids de 14 à 15 kg. Le programme de sélection permet de livrer annuellement: 547 • 300 à 400 agnelles, 30 à 50 béliers reproducteurs dans les élevages; et • 400 à 500 têtes à la boucherie (animaux non retenus pour l'élevage ou à réformer). Les thèmes de recherche appliquée, au niveau du Centre d'appui de Kolokopé sont les suivants: • expérimentation de plusieurs types de gestion de troupeaux: lutte continue, lutte organisée avec "rattrapage", lutte organisée synchronisation de l'oestrus, etc.; • insémination artificielle, avec l'assistance de la FAO: 264 brebis du Centre ont été inséminées à l'aide de semences récoltées sur des béliers de grande valeur du Centre de sélection ovine de Bouaké. 70% des inséminations furent fécondantes, et 22 agneaux mâles ont été sélectionnés pour la reproduction; • expérimentation des programmes de lutte contre les tiques en faisant varier le rythme de balnéations et en utilisant des produits nouveaux (étude faite en collaboration avec l'école des sciences de l'Université du Bénin, Lomé; • mise en place de programme d'essais en collaboration avec le CIPEA sur la complémentation alimentaire d'agneaux avec Leucaena et pois d'Angole (Cajanus cajan), et sur la capacité de production fourragère de Leucaena et pois d'Angole en fonction de la densité du semis et de l'âge des plants. Etude des possibilités d'association entre la production vivrière et l'élevage ovin Les investigations ont concerné l'évaluation de la production et la détermination de l'appétibilité et de l'ingestibilité des résidus et les sous-produits agricoles (paille, sons, fanes et épluchures de tubercules) disponibles au niveau du paysan; les fourrages des pâturages naturels et artificiels (parcelles de Panicum maximum C1 et T58) ainsi que la constitution et l'analyse du fumier ramassé dans les parcs (bergeries). Ces travaux se sont déroulés au Centre d'appui technique de Kolokopé du Projet (Centre de sélection et de recherche d'accompagnement) et chez les paysans-éleveurs au niveau du secteur amélioré encadré par le Projet. Matériel et méthodes Le disponible en sous-produits agricoles est évalué en pesant la paille de riz récoltée sur des carrés choisis au hasard dans un champ et les fanes d'arachide sèches stockées, les sons de céréales et les épluchures de manioc disponibles (après séchage). La matière verte et la matière sèche des prélèvements de 548 fourrages opérés dans les pâturages naturels du Centre de Kolokopé ont été déterminées. Leur appétibilité est évaluée par la rapidité avec laquelle les animaux entament le sous-produit servi; le taux d'ingestion est déterminé par le rapport de la quantité de sous-produit ingérée et du poids vif total des animaux. Les crottes de mouton sont récupérées par balayage des parcs (deux fois dans la semaine) et pesées après passage au tamis. La production est évaluée en fonction des moues de conduite des lots d'animaux constitués: • pâture (8 h par jour) et stabuiation (1 6 h environ) d'un troupeau de brebis et d'agneaux, avec complémentation (graines de coton) et supplémentation minérale (pierre à lécher); • stabuiation permanente d'un troupeau mixte et d'un troupeau de jeunes mâles en croissance (fourrage vert, foin, graines de coton et pierres à lécher). Après précompostage des crottes mises en tas pendant 14 jours, quatre traitements ont été réalisés pour le fumier: en tas sous abri, en tas couvert, en fosse sous abri, et en fosse non couverte. Des échantillons ont été prélevés un et deux mois après la date de la confection du fumier, à des profondeurs différentes (20 cm, 50 cm, 80 cm) en vue de doser les différents éléments fertilisants après détermination de la matière sèche: N (méthode Kjeldahl), C (méthode Anne), P205 (parcalorimétrie), K2O (par photométrie de flamme), CaO et MgO (par spectrométrie d'absorption atomique). Présentation et discussion des résultats Sous-produits de cultures Le disponible en paille de riz (tableau 1) est le double de la production de grain (respectivement 961 g/m2 et 514 g/m2, moyenne de 9 échantillons). La paille de riz bien séchée, très riche en matières minérales, peut être conservée et servie comme aliment de lest chez les moutons. Actuellement exploitée dans ce sens par les paysans du nord du Togo, elle est rarement, sinon jamais, utilisée par ceux du sud du pays. Tableau 1. Production et ingestion de panicules vides de sorgho Superficie récoltée Sorgho paniculé Panicules vides Consom mation Taux d'ingestionRefus 20 m2 1 000g 250 g 95 g 155 g 38% 549 Le tableau 1 montre que le taux de production de panicules vides de sorgho est de 25%. En considérant un rendement moyen de 500 kg à l'hectare de sorgho paniculé en culture associée (contre 850 kg à 1 tonne en culture pure de sorgho) , le disponible est de 125 kg/ha de panicules vides. Le taux d'ingestion est de 38% pour des moutons habitués; ce taux, faible, peut être augmenté en abandonnant délibérément quelques grains de sorgho sur les panicules vides. En dépit de son faible taux de production au niveau artisanal (19%), le son sec de maïs, de par sa richesse en MPD (80 g/kg de MS) et sa valeur énergétique (0,85 UF/kg de MS), constitue un bon complément dans l'alimentation des ovins. Le tableau 2 permet d'évaluer la production d'épluchures sèches de manioc à 2 400 kg par hectare. Malgré leur taux de production faible (8,70%), les épluchures sèches de manioc sont très utilisées par les paysans-éleveurs dans la complémentation alimentaire de leurs animaux. Les tests sur l'ingestibilité de ces épluchures sèches montrent un taux de consommation de 80% prouvant ainsi la bonne appétibilité de ce sous-produit. Tableau 2. Production artisanale d'epluchures de manioc Taux Epluchures Epluchures d'epluchures fraîches sèches fraîches Production Racines fraîches d'6pluchures Superficie sèches 20 m2 55 200 kg 14 500 kg 4 800 kg 26% 8„70% Un hectare d'arachide produit 675 kg de fanes, dont le taux de consommation est d'environ 90%. La conservation et l'utilisation de ces fanes d'arachide séchées sont des pratiques courantes dans la complémentation alimentaire des ovins. Disponibilité fourragère Sur les différents types de pâturages au Centre de Kolokopé (naturels, naturels entretenus, et naturels améliorés), deux séries de coupes espacées de 60 jours ont été réalisées et ont donné les résultats illustrés dans le tableau 3. Le pâturage naturel entretenu correspond à l'apport de fumier de ferme après réduction de l'effectif des espèces semi-ligneuses et ligneuses. Il s'agit de la pratique actuellement vulgarisée auprès des paysans-éleveurs. Le pâturage naturel amélioré correspond à un labour suivi d'un enrichissement de la flore avec d'autres espèces (Panicum maximum). Il n'est pas encore de pratique 550 courante au niveau du paysan-éleveur, car il necessite du materiel qui n'est pas encore a sa portée. Tableau 3. Production de fourrages pour ovins %deMS 1™ coupe i Production (kg de MS/ha)Paturage Espece e coupe Nature1 Graminees locales (1) 44,62 32,40 6 120,40 Naturel entretenu Graminees locales 42,66 31,17 7 198,82 Naturel Graminees ameliore locales + Panicum maximum 33,73 28,01 5377,70 (1) Especes recensees: Hyparrhenia diplandra Hyparrhenia sinuthiana Hyparrhenia chrysargyrea Cymbopogon gigantes Sorghum arundinaceum Panicum maximum Sehizqclyrium rupestre Hyperthelia dissoluta Imperata cylindrica Production de crottes et de fumier de mouton En ce qui concerne la disponibilite en crottes de mouton, on constate (tableaux 4 et 5) que la quantité de crottes produites augmente avec celle de la matiere seche ingérée et que le temps de parcage conditionne positivement la disponibilite des crottes dans le pare. 551 Tableau 4. Aliments ingéres par les troupeaux stabules et taux d'ingestion Type de troupeau Aliment distribue Troupeau mixte stabule Quantite Taux distribute (kg) d'ingestion Troupeau de jeunes males stabules Quantite Taux distribute (kg) d'ingestion Graines de coton 42 65% 58,650 65% Foin 30,800 31% 46,450 44% Fourrage vert 162,600 44% 147,050 42% Complement mineral 1,100 11% 0,550 5,5% Tableau 5. Production de crottes en fonction du mode de conduite du troupeau Mode de conduite Production de crottes Total 32 jours 1 kg* vif/j 1 kg* vif/an 25 kg* vif/an Pature et stabulation d'un troupeau mixte 48,750 12 4,380 105,500 Stabulation permanente d'un troupeau mixte 59,550 15 5,415 136,875 Stabulation permanente de males en croissance 66,700 18 6,570 164,250 ' Production de crottes ramenee a 1 kg et a 25 kg poids vif. 552 Tableau6.Compositiondescrotteseurumiermou onaprese xm po age( v cnng a ndeav ri )p v aSOcmdeprorondeur MS% MO Compositionminerale( n%dMS) P2O5K 0 CaOMg Typede crottes 7umier Crottes rraTches03000 dujour 00070, 330ro0 0030 0000 0ns30 0,.30 0,00 0,00 000 000 0nsns 000- 070 0,00 0,00 0nsro 0,300 0,00 0000 000 00 000 0307 0300 0ns0 00. 000 00,0 0300 30,0 0000 000 030ns 000 Crottespr - compostees 7ossenon couverteTasnon couvert Tassous abri 7ossesous abri 0700 0ns0■ s00,nns '0,■0 La stabulation permanente au niveau du paysan influe cependant négativement les productions de l'élevage (perte de poids, avortements éventuels, poids faibles des agneaux à la naissance, etc.) car le paysan-éleveur ne maîtrise pas encore les contraintes de cette technique de stabulation permanente. En conséquence, les conseils prodigués au paysan-éleveur portent sur le mode de conduite sur les parcours avec un temps de pâture assez long (8 h environ par jour) suivi d'un repos au parc. La composition minérale des crottes fraîches et précompostées (tableau 6) montre que les taux des différents éléments fertilisants se révèlent identiques à l'exception de ceux du carbone et de l'azote. Le tableau 6, qui résume aussi les analyses des différents types de fumier, permet de recommander la préparation du fumier sous abri (en tas ou en fosse) de préférence au fumier non couvert, car il fournit ainsi une meilleure concentration en éléments. La constitution d'un fumier excrément permet de profiter de la proportion, même faible, des éléments fertilisants des crottes de mouton, et fait actuellement l'objet de vulgarisation. Conclusion Ce programme de développement de l'élevage des ovins et caprins suivi de recherche d'accompagnement vise l'amélioration du secteur traditionnel de cet élevage. L'étude a permis d'une part de dégager la disponibilité des résidus et sous-produits agricoles et leur utilisation, d'autre part d'apporter des améliorations dans l'association agriculture-élevage ovin en vue de mieux implanter l'agropastoralisme au stade paysannat. La mise à la disposition des paysans-éleveurs des animaux reproducteurs sélectionnés, la maîtrise des thèmes de vulgarisation par ceux-ci et enfin l'association de l'élevage à l'agriculture (exploitation des jachères pour pâturages, conservation et utilisation des résidus et sous-produits agricoles pour l'alimentation des animaux, bonne utilisation du fumier de mouton pour la fertilisation des champs, etc.) constituent une des solutions pour la réalisation de l'équilibre entre productions végétale et animale. 554 Recherche et développement des caprins et ovins au Burundi /. Nsabiyumva Département de la faculté des sciences agronomiques (FACAGRO), Bujumbura, Burundi Résumé La pression démographique sur le foncier dans les hautes terres de l'Afrique de l'Est donne au petit élevage un potentiel croissant dans des systèmes intégrant agriculture et production animale. Le Département de zootechnie de la faculté d'agronomie de l'Université du Burundi est chargé de la coordination du réseau national petits ruminants et a constitué le Centre de recherche universitaire sur le petit élevage. Ce Centre gère la station zootechnique de Maramvya et deux programmes de recherche en exploitation: Le PIPRA (Programme intégration petits ruminants agriculture) dans la commune de Mutaho, et des essais d'association de l'élevage ovin au reboisement et aux palmeraies. Small ruminant research and development in Burundi /. Nsabiyumva Abstract The high population density in the East African highlands encourages small ruminant production in integrated crop-livestock production systems. This paper reports on research activities in Burundi on integrating sheep production with reafforestation and palm-tree plantation systems. These studies focus on improving forage resources, animal health and flock management. 555 Introduction Dans les régions densément peuplées des hautes terres, l'élevage bovin perd régulièrement du terrain. Une politique de reboisement ambitieuse essaie de limiter les effets de l'érosion mais empiète systématiquement sur les parcours communautaires de faible valeur bromatologique. L'effectif de chèvres et moutons semble augmenter parallèlement du déclin du gros bétail. Cependant, les systèmes de production traditionnels restent peu performants et le taux d'exploitation ne s'élève que très faiblement. Deux types de solutions sont à l'étude pour accélérer l'intensification de la production du petit bétail: l'intégration de l'élevage des petits ruminants à l'agriculture et l'association de l'élevage ovin aux reboisements et à l'élaiculture. Projet intégration petits ruminants agriculture (PIPRA) Les recherches poursuivies depuis 1986 ont porté essentiellement sur les possibilités d'intégrer l'élevage des petits ruminants à l'agriculture dans une des régions naturelles les plus densément peuplées du pays (Nord-Kirimiro-commune de Mutaho). L'enquête zoo-économique préalablement menée sur 60 exploitations agricoles a fait ressortir certaines caractéristiques: • réduction extrême de la superficie cultivable des exploitations; • agriculture intensive avec jachères de durée très courte, limitées le plus souvent à quelques ares en 1" saison; • effectifs animaux faibles, l'espèce bovine est en voie de régression; • intérêt identique pour la chèvre et le mouton, sans préférence caractérisée. En 1987, six exploitations-pilotes, associant l'élevage caprin à l'agriculture suivant un contrat-type, ont été installées et encadrées. Un certain nombre de contraintes ont pu être identifiées; difficultés d'installation de certaines cultures fourragères, soit par manque de semences, soit par utilisation d'une technique non appropriée; réceptivité laborieuse des exploitants à certaines techniques semi-intensives; effectifs restreints de reproducteurs du centre de Murongwe ne permet pas de satisfaire les souhaits de tous les demandeurs. L'expérience acquise au cours des quatre dernières années dans l'installation et l'encadrement de 21 exploitations conduit à formuler un certain nombre d'observations: 1.1. Le projet a un succès certain; le nombre de volontaires dépasse largement les possibilités en intrants animaux et fourragers. Et les préférences sont aujourd'hui orientées vers le mouton. 556 1.2. Les conditions requises des éleveurs pour leur accorder un prêt de 5 reproductrices sont au départ correctement remplies. Par la suite, diverses difficultés surgissent: 1.2.1. Les cultures fourragères ne sont pas exploitées rationnellement. La supplémentation est épisodique. La récolte du fourrage n'est pas pratiquée en temps voulu. D'autre part, plusieurs de ces cultures fourragères peuvent être utilisées à d'autres fins qu'à l'alimentation des animaux. Il en résulte évidemment que les performances des animaux s'en ressentent. La diffusion, chaque année, de nouvelles boutures/graines de cultures fourragères et l'introduction du creep-feeding chez les éleveurs devraient pouvoir améliorer la productivité de ce système d'élevage. 1.2.2. Les aliments non conventionnels disponibles dans le rugo ne sont pas utilisés rationnellement. Dans cette optique, il a été expérimenté avec succès une supplémentation avec une adventice de bonne valeur alimentaire: Acanthus sp. qui envahit les jachères, les fossés et les talus. Les éleveurs hésitent cependant à l'utiliser, parce qu'elle est difficile à récolter et que les épines dont elle est recouverte, favoriseraient l'apparition de l'echtyma contagieux. Les noyaux d'avocats ont donné d'excellents résultats mais leur récolte est saisonnière. 1 .3. Les prévisions de projection de troupeaux se confirment. Au bout de deux années, l'objectif des 10 UZPR est atteint et les remboursements sont généralement terminés. Malheureusement, les éleveurs ont alors tendance à vendre trop rapidement leurs animaux, y compris les bonnes reproductrices. Il craignent en effet la malveillance de leurs voisins et pensent se mettre à l'abri de cette dernière par une commercialisation hâtive. 1 4. Intégration petits ruminants/agriculture: elle n'est pa réelle à l'heure actuelle. On parlera plutôt d'une association entre une composante agricole et une composante élevage. Cependant, en 1987-1988, une approche systématique des contraintes a été réalisée et son analyse devrait permettre de proposer des solutions rationnelles et satisfaisantes. 1.5. Méthodologie 1.5.1. Afin d'augmenter les possibilités d'approvisionnement en intrants animaux et fourragers: 557 1.5.1.1. Développement des troupeaux caprin et ovin de multiplication à Murongwe afin de pouvoir satisfaire les besoins des éleveurs locaux. L'effectif moyen de 50 UZPR devrait être atteint dans les deux troupeaux. Ceci devrait permettre d'approvisionner 15 à 20 éleveurs nouveaux par an. Tous les paramètres classiques continuent à être collectés et interprétés, afin de les comparer utilement à ceux obtenus dans les rugos. Les faibles superficies de parcours à Murongwe sont encore réduites par le débordement régulier de la Mubarazi en saison des pluies. L'effectif actuel de 83 reproductrices semble difficile à dépasser et même à maintenir. Toutefois, diverses améliorations pourraient être trouvées en intégrant davantage les petits ruminants aux parcours et ressources fourragères des taurillons élevés à Murongwe pour la production de fumier. Au cas où ces améliorations ne pourraient être obtenues, le développement de la Station de Maramvya permettra à l'avenir de produire les géniteurs nécessaires. La Station de Murongwe servira alors de Centre de transit et de quarantaine, avant la diffusion aux éleveurs. 1.5.1.2. Installation de centres de Saillie. En effet, afin de contrôler efficacement l'amélioration génétique du troupeau de la commune, les saillies se font exclusivement avec les mâles sélectionnées de la Station. Les jeunes mâles, produits en Station ou dans les rugos encadrés, sont: - soit castrés et mis à l'embouche, - soit retenus pour la reproduction aux centres de saillie, selon les performances enregistrées sous la mère. Ils sont ensuite confirmés par test de descendance. Deux centres de saillie sont installés à Murongwe et Gitongo. 1.5.1.3. Les éclats de souche de Setaria, Tripsacum, Pennisetum, plants de Leucaena et semences d'avoine sont nécessaires. Actuellement, 50 000 éclats de souche de Setaria, et 50 kgs de semences d'avoine sont distribués annuellement aux exploitants. Les graines de Leucaena divesifolia sont fournis de telles sorte que chaque exploitant puisse établir en saison sèche une pépinière de 2 000 plants. Les ressources en 558 éclats de souche de Tripsacum et de Pennisetum sont suffisantes chez les exploitants eux-mêmes pour leur permettre d'étendre leurs cultures. 1 .5.2. La multiplication des exploitations petits ruminants devrait aboutir à moyen terme, à leur regroupement au sein d'une coopérative, dont la Station de Murongwe représentera le support logistique, destiné à fournir les intrants animaux et végétaux, et assurer l'encadrement vétérinaire et zootechnique. 1.5.3. Intégration petits ruminants/agriculture. Un accent tout particulier est porté: 1.5.3.1. aux problèmes de fertilisation. La diffusion de fumières améliorées devrait permettre de produire un abondant fumier de qualité si, toutefois, une litière suffisante est apportée. Dans un premier temps ce fumier de petits ruminants est expérimenté sur le haricot cultivé en marais et sur colline. Dans un second temps les expérimentations s'étendraient sur d'autres cultures. 1.5.3.2. La faible croissance des plants de Leucaena chez les exploitants devrait être corrigée par l'inoculation des graines. 1 .5.3.3. L'utilisation des aliments non conventionnels sera de plus en plus généralisée après qu'un bilan fourrager aussi précis que possible soit affecté au sein de chaque exploitation. Sylvo-pastoralisme Les opérations portent essentiellement sur l'élevage ovin et concernent deux secteurs: les reboisements et les palmeraies. Reboisement L'association entre élevage ovin et reboisement, entreprise à RUGAZI en 1981-1982 avec le concours de la CCE a été interrompue faute de financement. Elle a été reconduite à Vyanda, sur des reboisements à Pinus Eliotii grâce à la collaboration du Département des forêts et à un financement conjoint FAC/Banque mondiale. Elle est poursuivie, depuis 1987 à Ryarusera sur des boisements à eucalyptus. Les programmes mis en place qui tendent essentiellement à mettre au point les systèmes de gestion les plus appropriés et les plus complémentaires des deux composantes, sont à long terme et devront se poursuivre. 559 Les troupeaux reproducteurs installés sont à l'heure actuelle en pleine évolution et aucun problème ne se pose quant à l'écoulement des agneaux produits. Il est à noter que c'est le troupeau de Ryarusera qui fournit le plus de géniteurs. Palmeraies Un programme identique est prévu pour les palmeraies de l'IMBO-SUD, qui présentent l'avantage supplémentaire d'un excellent fourrage représenté par la plante de couverture, à savoir Pueraria phaseloldes, abondante pendant les premières années de la plantation. La station zootechnique de Maramvya Depuis sa création la Station Zootechnique de la FACAGRO possède un projet petits ruminants. La capacité d'accueil est actuellement de 150 reproductrices pour chaque espèce et l'on compte plus de 600 ovins et caprins à la Station de Maramvya. Les recherches actuelles ou à venir portent sur: • la caractérisation et la connaissance des races locales: toutes les données concernant la croissance et la reproduction, sont collectées et sont exploitées afin de sélectionner les meilleurs reproducteurs. Ceux-ci sont entre autre diffusés au niveau des projets comportant un volet développement des petits ruminants. • l'alimentation: des études dans ce domaine ont porté sur le comportement des animaux au pâturage, la digestibilité du Leucaena leucocephala, du Panicum maximum et du Pennisetum purpureum et la paille de riz. Divers niveaux de complémentation à base de son riz sont testés sur les mères allaitantes et les jeunes non sevrés. • la reproduction: des recherches plus fondamentales sont menées afin d'établir des profils hormonaux et de les mettre en relation avec les caractéristiques de reproduction de nos races locales. • pathologie: un bilan des maladies des petits ruminants permet d'établir les taux de mortalité et de morbidité et d'instituer des programmes prophylactiques adéquats. • un encadrement des éleveurs de la commune vient d'être lancé afin de collecter des données d'exploitation et d'assurer un suivi sanitaire rigoureux. 560 Le programme national de sélection ovine (PNSO) de Côte d'Ivoire: mise en place du contrôle de performances en ferme A. Oya Société de développement des productions animales (SODEPRA) Centre de Bouaké B.P. 1366, Bouaké 01, Côte d'Ivoire Résumé En 1983-1984, a été créé le Programme national de sélection ovine (PNSO) dont l'objectif principal est d'améliorer, par la voie de sélection en race pure, le format du mouton d'origine locale: le Djallonké. Pour ce faire il a été mis en place une base de sélection où des contrôles de performances sont effectués, qui conduisent à présélectionner dans les élevages, les agneaux sur le critère de leur poids à 80jours. Ces agneaux deviennent, après une autre sélection (en station), les béliers sélectionnés que le projet diffuse ensuite dans et hors de la base de sélection. Cette base de sélection mise en place progressivement compte aujourd'hui 10 500 brebis réparties dans 100 élevages disséminés sur tout le territoire ivoirien. Par an, ce sont 400 à 500 agneaux présélectionnés que le PNSO achète aux éleveurs adhérants à ce système de contrôle des performances zootechniques des animaux. Une meilleure intégration financière et technique des actions de sélection dans les activités des projets d'encadrement de la SODEPRA (Société de développement des productions animales) devraitpermettre de résoudre les problèmes que rencontre le PNSO. 561 The national sheep selection programme (PNSO) in Cote d'lvoire: Establishment of on-farm performance evaluation A Oya Abstract The objective of the National Sheep Selection Programme (PNSO) introduced in 1983-1984 was to improve the body conformation of the local West African Dwarf sheep through selection within the breed. To achieve this, it established a nuclear flock on which initial selection is performed on-farm, based on performance evaluation, with lambs being selected on their weight at 80 days old. Following further selection on-station, rams are distributed within the nuclear flock and to other flocks. The nuclear flock now consists of some 10 500 ewes in 100 flocks throughout Côte d'lvoire. Each year, PNSO purchases between 400 and 500 preselected lambs from breeders participating in this livestock performance evaluation system. Further financial and technical integration of the selection activities into the development projects of SODEPRA (Livestock Production Development Society) should solve the problems currently encountered by PNSO. 562 LIST OF PARTICIPANTS BOTSWANA E. SENYATSO Agricultural Research Institute Private Bag 0033 Gaborone BURUNDI A. de G. HABONIMANA Institut des Sciences Agronomique du Burundi B.P. 795 Bujumbura Cyrille NTAHIMPEREYE Projet Caprin de Ngozi B.P. 45 Ngozi lldephonse NSABIYUMVA FACAGRO B.P 2940 Bujumbura BENIN Mawule M. Syivie HOUNZANGBE ep ADOTE FSA Universite Nationale du Benin B.P. 526 Cotonou BURKINA FASO Aime NIANOGO SN-lDRI Universite de Ouagadougou; B.P. 7021 Ouagadougou -03 CAMEROON M. AYISSI Department of Agricultural Economics, Dschang University Centre P.O. Box 110 Dschang M. Ruben NJWE Department of Animal Science, Dschang University Centre; P.O.Box 110 Dschang REPUBLIQUE CENTRE AFRICAINE J. Yaguelke ANDE B.P. 1509 Bangui CONGO Alphonse M. BATALOU Project National du Mayombe B.P. 2499 Brazzaville Vincent DOULOU Unite de Recherches sur les Systemes de Production Animates URSPA/CRVZ; B.P. 8041 OMS Brazzaville COTE D'lVOIRE Ahouakan ASSOHOU SODEPRA CENTRE BP1336 Bouake 563 Alain OYA FRANCE SODEPRA CENTRE Jean-Pierre BOUTONNET BP1366 Dept. de Sociologie et d'Economie Bouake, Rurale; INRA, Montpellier 9 place viala ETHIOPIA Montpellier Bernard REY ILCA GABON P.O.Box 5689 I. KOUDIARATU Addis Abeba Ministere de I' Agriculture TLX: 5587 MINAGRI GIRMA ADUGNA Libreville Faculty of Veterinary Medicine Addis Ababa University GAMBIA P.O.Box 34 Debre-Zeit Derek CLIFFORD ITC Simeon EHUI MB 14 ILCA Banjul P.O. Box 5689 Addis Ababa GHANA A.K. TUAH TEKELYE BEKELE University of Science and ILCA Technology P.O. Box 5689 Department of Animal Science Addis Ababa Faculty of Agriculture BEYENE CHICHAIBELU Kumasi University of Agriculture KENYA P.O.Box 138 Camillus O. AHUYA KARI Alemaya P.O.Box 25 SahrLB. LEBBIE Naivasha ILCA P.O.Box 5689 Addis Abeba BERHANE KIFLEWAHID IDRC Francois THIAUCOURT Box 62084 National Veterinary Institute Nairobi B.P. 379 Debre Zeit A.B. CARLES University of Nairobi Jean Jacques TULASNE Mission Veterinaire Francaise en P.O. Box 29053 Kabete - Nairobi Ethiopia B.P. 1053 GITHIGIA MAINA Addis Ababa University of Nairobi 564 NYANGALA Dept. of Animal Production University of Nairobi P.O. Box 29053 Nairobi S.B. KAYONGO Dept. of Animal Production University of Nairobi P.O. Box 29053 Nairobi Len REYNOLDS ILCA P.O.Box 80147 Mombasa B.A.J. MWANDOTTO NAHRC SR/CRSP Box 25 Naivasha OKEYO A. MWAI Dept. of Animal Production University of Nairobi P.O. Box 29053 Nairobi Robert SHAVULIMO Dept. of Animal Health, Egerton University P.O.Box 536 Njoro RUTAGWENDA Dept. of Animal Production University of Nairobi P.O. Box 29053 Nairobi Paterson P. SEMENYE SR-CRSP P.O. BOX 252 Maseno P.O. Box 5781 1 Nairobi R.G. WAHOME University of Nairobi P.O. Box 29053 Nairobi S.K. KUNYU MOLD-SEP P.O. BOX 254 Naivasha LESOTHO J.P. HUNTER SWEDFOREST CONSULTING AB, National Forest Enterprise of Sweden C/O LAPIS Project; P.O.Box 333 Maseru 100 M. MOLAPO Agricultural Research Institute Maseru MALAWI James W. BANDA Department of Animal Science Bunda College P.O. Box 219 Lilongwe Soka K. KARUA Department of Animal Science Bunda College of Agriculture P.O.Box 219 Lilongwe MALI Mamadou DOUMBIA CRZ/ SOTUBA B.P. 262 Bamako D. WACHIRA KARIHQ Saidou TEMBELY Laboratoire Véterinaire Central 565 P.O. Box 2295 Bamako TAHIROU TANGARA Centre de Recherche IK. ODUBOTE Dept. of animal Sciences; Obafemi Awolowo University lle-ife Zootechnique Sotuba LO. WOSU BP262 Dept. of Veterinary Medicine, Bamako University of Nigeria MAROC Nsukka A. LAHLOU KASSI W.Akin HASSAN Institut Agronomique et Velerinaire Dept. of Animal Science; Usmanu Hassan II Danfodiyo University B.P. 6202 P.M.B. 2346 Rabat Sokoto Lahsen DERQAOUI NIGER Dept. de Reproduction Animate et A. YENIKOYE Universite de Niamey B.P. 237 Niamey Insemination Artif icielle BP6202 Rabat MAZOUZ RWANDA Dept. de Reproduction Animale et d'lnsemination Artificielle G. SIBOMANA ISAR B.P. 138 Butare BP6202 Rabat MOZAMBIQUE SENEGAL C. CONCEICAO Instituto Produccion Animal Magatte, NDIAYE Laboratoire d'Elevage et de Recherches Veterinaires BP2057 Dakar-Hann P.O. Box 1410 Maputo NIGERIA GO. OYEDIJI Olivier FAUGERE Programme PPR B.P. 2057 Dakar-Hann University of Ibadan, Nigeria P.O. Box 5320 Ibadan IF. ADU SIERRA LEONE Dept. of Animal Production & Health I. C. HASSAN P.M.B. 1096 9C THOMPSON BAY Abeokuta Free Town 566 SOMALIA HASSAN M. HASSAN Faculty of Veterinary Medicine and Animal Husbandary P.O. Box 1738 Mogadishu SWAZILAND D.T. MAVIMBELA University of Swaziland TANZANIA E.E. MASSAE Regional Livestock Development Office P.O. Box 3084 Arusha LA. MTENGA Department of Animal Science and Production; Sokoine University of Agriculture; P.O.Box 3004, Chuo Kikuu Morogoro S.M. DAS Zonal Research and Training Centre, Central Zone; Livestock Production Research Institute P.O.Box 202 Mpwapwa D.S.S. SENDALO Livestock Research Centre West Kilimanjaro P O.Box 147 Sanyajou W.D. MAFWERE Department of Animal Science Faculty of Agricultur, Sokoine University of Agriculture P.O Box 3004, Chuo Kikuu Morogoro TOGO Y.I. PESSINABA Projet National Petits Ruminants B.P. 65 Atakpame Y.N. HADZI Projet National Petits Ruminants Centre d'Appui Technique de Kolokope B.P. 65 Atakpame TUNISIA G. KHALDI Dept. Zootechnie, INRAT Avenue de I'independance Ariana 2080 NEFZAOUI Department Zootechnie, INRAT Avenue de I'independance Ariana 2080 TCHAD Daniel BOURZAT Projet Regional Petits Ruminants, Laboratoire de Farcha B.P. 433 N'djamena UGANDA G.H. KIWUWA Department of Animal Production, Makerere University Entebbe UNITED KINGDOM M. GILL Natural Resources Institute Central Avenue Chatham Maritime Chatham Kent ME4 4TB 567 ZAMBIA J. KATONGOLE University of Zambia Lusaka ZIMBABWE AnnaWARAMBWA Henderson Research Station Private Bag 2004 Magowe LR. NDLOVU Department of Animal Science University of Zimbabwe P.O.Box MP 167; Mount Pleasant Harare Lindiwe SIBANDA Department of Animal Science University of Zimbabwe c/o Dr LR. Ndlovu; Post Box MP 167 Mount Pleasant, Harare O. MATIKA Matopos Research Station P.B. K5137 Bulawayo Peter HATENDI Grasslands Research Station Private Bag 3701 Marondera 568 ISBN 92-9053-258-0 Printed at ILCA, Addis Ababa, Ethiopia