ILCA Research Report No. 14
Livestock production in central Mali
Long-term studies on cattle
and small ruminants
in the agropastoral system
R.T. Wilson
December 1986
INTERNATIONAL LIVESTOCK CENTRE FOR AFRICA
ADDIS ABABA, ETHIOPIA
ORIGINAL: ENGLISH
ABSTRACT
ILCA has been conducting a long-term study on livestock production in central Mali since the beginning
of 1976. This report presents results based on data collected over a 6-year period from 1978 to mid-1984.
In Part I the livestock production systems in the zone, management practices and herd and flock demog
raphy are described. Cattle and small ruminant productivity is discussed in detail in Parts II and III and
recommendations based on the results of the studies are given in Part IV.
KEYWORDS
/Sahel//Mali//cattle//sheep//goats//livestock production systems//productivity//livestock management/
/herds//reproduction//growth//mortality/
RESUME
En 1976, le CIPEA a entrepris une étude à long terme sur la production animale dans le centre du Mali.
L 'analyse des données recueillies de 1978 à 1984 estprésentée dans leprésent rapport, qui comprend quatre
parties. La première est consacrée à une description des systèmes d'élevage, de la conduite des troupeaux
et de la démographie animale. Les deuxième et troisième chapitres traitent des divers aspects de la produc
tivité des bovins et des petits ruminants, et le quatrième présente des recommandations fondées sur les
résultats de l'étude.
MOTS CLES
/Sahell/Malil/bovinl/moutonl/chèvrel/système d'élevage//productivité//exploitation du bétail//troupeau//
reproductivité//croissance//mortalité/
ISBN 92-9053-081 -2
ii
CONTENTS
LISTOFTABLES v
LIST OF FIGURES vii
PREFACE ix
ACKNOWLEDGEMENTS ix
SUMMARY x
SYNTHESE xi
GENERAL BACKGROUND xiii
1. INTRODUCTION 1
General 1
Cattle types and distribution 1
Small ruminant types and distribution 1
2. THE STUDY AREA 7
Location 7
Climate 7
Vegetation 7
3. LIVESTOCK PRODUCTION SYSTEMS 9
Livestock in the Sahel zone 9
Livestock production in Mali 9
Ecological zones 9
Distribution of people and animals 12
Production systems 14
Subsystems in the agropastoral system 19
4. MATERIALS AND METHODS 20
The study herds and flocks 20
Field methods 20
Data analysis 20
Additional sources of information 21
5. MANAGEMENT 22
Ownership patterns 22
Movement 22
General management 26
Management case studies 28
in
Mouton de case in the Niono area 28
Management of Marina sheep in the inundation zone 32
6. HERD AND FLOCK DEMOGRAPHY 35
Cattle population structure 35
Small ruminant population structure 36
CATTLE PRODUCTIVITY 41
7. REPRODUCTIVE PERFORMANCE 43
Age at first calving 43
Calving interval 43
Season of calving 44
Number of calves born per breeding female in the herd 44
Discussion 45
8. GROWTH AND WEIGHT 47
Birthweight 47
Growth to maturity 47
Breeding female postpartum weight 47
Effects of climate on growth and mature weight 47
Discussion 50
9. MORTALITY AND OFFTAKE 57
Abortions 57
Mortality to 4 years of age 57
Mortality in adult cattle 57
Offtake 57
Discussion 57
10. PRODUCTIVITY 60
Meat production: A case study at Niono slaughterhouse 60
Numbers and seasonality of animals slaughtered 60
Sex and age structure 60
Carcass weights 60
Contribution of cattle to meat supply 60
Milk production 60
Productivity indices 60
SMALL RUMINANT PRODUCTIVITY 65
11. REPRODUCTIVE PERFORMANCE 67
Age and weight at early parturitions 67
Litter size, parturition interval and annual reproductive rate 70
Effects of climate on the period of birth and on litter size 71
Numbers of births by age class 74
Discussion 74
Early reproductive performance 74
Main reproductive traits 77
Climatic effects on reproduction 80
12. GROWTH AND WEIGHT 82
Birthweight 82
Growth to maturity 82
Breeding female postpartum weights 87
Effects of climate on specific age and sex groups 88
Discussion 90
13. MORTALITY AND OFFTAKE %
Abortions 96
Preweaning mortality %
Mortality after weaning 97
Offtake 97
14. PRODUCTIVITY 99
Milk production 99
Wool production 99
Productivity indices 99
Appropriate improvement paths for goats and sheep in Mali 101
CONCLUSIONS AND RECOMMENDATIONS 105
15. CONCLUSIONS AND RECOMMENDATIONS 107
The nature of traditional livestock production systems 107
The current data set 107
Major problems to improved productivity 108
Future research 108
REFERENCES 109
APPENDIX Ill
LIST OF TABLES
Table 1. Rainfall for Niono, 1977-1983 8
Table 2. Distribution of ruminant livestock by species in the Sahel countries 10
Table 3. Distribution of people and animals by ecological zone in Mali 16
Table 4. Dependence on livestock in different Sahelian production systems 17
Table 5. Main characteristics of livestock production systems in West Africa 18
Table 6. Management and environmental factors in the millet and rice subsystems
in central Mali 19
Table 7. Holdings of livestock and farm implements according to official statistics for a sample
of villages in the millet and rice subsystems of central Mali 23
Table 8. Ownership pattern and flock sizes for small ruminants
in the agropastoral system of central Mali 24
Table 9. Ownership of cattle and other stock by different
ethnic groups in central Mali 25
Table 10. Structure of two cattle management units in the Niger inundation zone,
by main profession of owner 25
Table 1 1 . Seasonal distribution of livestock according to ownership groups 28
Table 12. Regression analyses and correlations between the ownership of moutons de case
and other classes of livestock 31
Table 13. Percentage distribution of all confined sheep by sex 31
Table 14. Percentage distribution of confined male sheep by age group 31
Table 15. Percentage of owners offering different types of feed to moutons de case in the rice
and millet subsystems and percentage of owners offering who buy the products 32
Table 16. Stratification of Macina flocks with demographic characteristics
and management objectives of each group 34
Table 17. Cattle herd structures (%) in central Mali 36
Table 18. Age and sex of cattle populations in three urban areas in central Mali 37
Table 19. Flock structures of goats owned by some ethnic groups in central Mali 39
Table 20. Flock structures of sheep owned by some ethnic groups in central Mali 39
Table 21 . Analysis of variance of calving interval for agropastoral cows in central Mali 44
Table 22. Least-squares means for calving interval of agropastoral cows in central Mali 44
Table 23. Analysis of variance of cattle weights at specified ages 48
Table 24. Least-squares mean weights (kg) for agropastoral cattle at different ages 49
Table 25. Phenotypic correlations between calf weights 51
Table 26. Analysis of variance of postpartum weights of dams 52
Table 27. Least-squares means for postpartum weights (kg) of dams 52"
Table 28. Regression analyses of weight (kg) on time for different classes of cattle
and of rainfall (mm) on time, 1977-1983 56
Table 29. Observed mortality rates for agropastoral cattle to 4 years of age,
presented by different variables 58
Table 30. Analysis of variance of cattle productivity indices 63
Table 31. Least-squares means of cattle productivity indices 63
Table 32. Analysis of variance of weight at first conception and ages at first and
second parturitions for goats and sheep in central Mali 68
Table 33. Least-squares means for weight at first conception and ages at first and second parturitions
for goats and sheep in central Mali 69
Table 34. Numbers of parturitions and percentages of young by parturition type for sheep
and goats in central Mali 70
Table 35. Reproductive performance of agropastoral goats and sheep, based
on owners' recall data 72
Table 36. Analysis of variance of litter size, parturition interval and annual reproductive rate
for sheep and goats in central Mali 73
Table 37. Least-squares means for litter size, parturition interval and annual reproductive rate
of sheep in central Mali 75
Table 38. Least-squares means for litter size, parturition interval and annual reproductive rate
of goats in central Mali 76
Table 39. Data from the polynomial analyses of reproductive characteristics
of goats and sheep in central Mali 78
Table 40. Number of parturitions per breeding female by age class 81
Table 41 . Analysis of variance of weights-at-age for goats in central Mali 83
Table 42 . Least-squares means of weights (kg) at specified ages for goats in central Mali 84
Table 43 . Analysis of variance of weights-at-age for sheep in central Mali 85
Table 44. Least-squares means of weights (kg) at specified ages for sheep in central Mali 86
Table 45. Phenotypic correlations between weights at various ages for goats
and sheep in central Mali 88
Table 46. Analysis of variance of postpartum weights of goats and sheep in central Mali 88
Table 47. Least-squares means for postpartum weights (kg) of goats and sheep in central Mali 89
VI
Table 48. Weight changes by sex and age in central Malian goats and sheep, expressed as
percentages of mean annual weight 92
Table 49. Regressions of weight (kg) on time for mature female sheep and goats
and of rainfall (mm) on time, central Mali 95
Table 50. Analysis of variance of preweaning mortality
in goats and sheep in central Mali 96
Table 51. Least-squares means for preweaning mortality (%)
in goats and sheep in central Mali 97
Table 52. Total annual offtake by system, sex of animals and species 97
Table 53. Mean wool production by Marina sheep 100
Table 54. Analysis of variance of productivity indices for goats and sheep in central Mali 101
Table 55. Least-squares means of productivity indices for goats and sheep in central Mali 102
Table 56. Ratios of comparative advantages for variables on Index II, to be used
in designing improvement pathways 103
APPENDIX
Correlation coefficients and significance levels for calves born in 1978-83 and rainfall
at specific periods before calving Ill
LIST OF FIGURES
Figure 1 . A Sudanese Fulani stud bull at the Station du Sahel, Niono 2
Figure 2. Distribution of cattle types in central Mali 2
Figure 3. A goat of the West African Dwarf type 3
Figure 4. A goat of the West African Long-legged type 3
Figure 5. Toronke variety of the Fulani sheep at Niono market 4
Figure 6. Black Maure ram in a harvested field 4
Figure 7. A ram of the Tuareg type in central Mali 5
Figure 8. Marina wool sheep 5
Figure 9. Distribution of the main sheep types in central Mali 6
Figure 10. Climatic normals for the Niono area, 1930-1976 8
Figure 1 1 . Livestock densities in the Sahel countries 11
Figure 12. Human population densities in the Sahel countries 11
Figure 13. Numbers of TLU per rural inhabitant in the Sahel countries 12
Figure 14. Annual rainfall and ecological zones in Mali 13
Figure 15. Number of cattle per rural inhabitant in Mali 14
Figure 16. Number of goats and sheep per rural inhabitant in Mali 15
Figure 17. Ownership of ploughs and oxen pairs in a central Malian village 24
Figure 18. Pastoral cattle feeding on stubble in an agro-sylvo-pastoral system in central Mali 26
Figure 19. Landscape units of the Niger inundation zone and seasonal distribution
of Macina sheep 27
Figure 20. A net pouch of baobab string used to restrain calf suckling 28
Figure 21 . Annual pattern of cattle management in a sedentary village in central Mali 29
Figure 22. A mouton de case ram in a stall in central Mali 30
Figure 23. Ownership of moutons de case in the rice and millet subsystems 30
vii
Figure 24. Length of confinement for moutons de case in the rice and millet subsystems 31
Figure 25. Work oxen ploughing a field in central Mali after the early rains 35
Figure 26. Population structure of a Fulani herd with milk as a primary function 37
Figure 27. Distribution of oxen pairs by current working life in a central Malian village 38
Figure 28. Frequency distribution of calving intervals in cattle 43
Figure 29. Seasonal distribution of conceptions and calvings for cattle in central Mali 45
Figure 30. Number of calves born per cow in central Malian herds 46
Figure 3 1 . Generalised growth curve to 4 years of age for cattle of known birth date 50
Figure 32. Relationship between season of birth and growth of cattle to 3 years of age 51
Figure 33. Relationships between postpartum weights of cattle and parity,
season and year of calving 53
Figure 34. Long-term growth patterns for calves born in May each year from 1978 to 1982 54
Figure 35. Seasonal weight changes in mature oxen and female cattle with 1,2,3 and 4 pairs
of permanent incisors 55
Figure 36. Long-term weight changes in oxen and breeding cows related to changes in rainfall 56
Figure 37. Age-specific hazard and cumulative survival rates for agropastoral cattle
in central Mali 59
Figure 38 . Sex and age composition of animals slaughtered at Niono slaughterhouse 61
Figure 39 . Species contribution to meat supplied from Niono slaughterhouse 62
Figure 40. Distribution of weights at first conception for goats and sheep in central Mali 67
Figure 4 1 . Ages at first and second parturitions for goats and sheep in central Mali 67
Figure 42 . Frequency distribution of parturition intervals for sheep and goats in central Mali 71
Figure 43. Best-fit estimates from the polynomial analyses of reproductive characteristics
of sheep and goats in central Mali 79
Figure 44. Total number of parturitions per month and average number of young per parturition
per month for goats and sheep in central Mali, 1978-1983 80
Figure 45 . Generalised growth curves from birth to maturity for goats and sheep in central Mali ... 87
Figure 46. Relationship between year of birth and subsequent growth to 3 years
of age in central Malian goats 87
Figure 47. Relationships between postpartum weights and parity, season and year of birth
of small ruminants in central Mali 90
Figure 48. Seasonal variations in weight of corresponding age and sex groups of small ruminants
in central Mali 91
Figure 49. Long-term growth patterns for female sheep born in the millet subsystem
in May to June and in October to November each year from 1978 to 1982 93
Figure 50. Long-term variations in weight of sheep and goats in the rice and millet subsystems
of central Mali 94
Figure 5 1 . Mean lactation curve and representative yields for Marina sheep 100
PREFACE
In the past, little attention was given to traditional
systems of livestock production. Indigenous
breeds of livestock were considered to have low
genetic potential and to contribute only marginally
to agricultural production. Moreover, with very
few exceptions, the term 'livestock' was almost in
variably used as a synonym for cattle: goats and
sheep, if considered at all, were regarded as mere
harbingers of degradation and drought.
In recent years, some of the old attitudes to
traditional livestock production have changed but
there is still very little information available on
the productivity of local animal breeds under
traditional management. This report presents the
findings of research work carried out by the Inter
national Livestock Centre for Africa (ILCA) on
cattle and small ruminant production in central
Mali.
The author of this report, who was the Senior
Animal Scientist and later Team Leader of
ILCA's research programme in Mali, is now
Head of the Small Ruminant and Camel Group at
ILCA's headquarters in Addis Ababa, Ethiopia.
ACKNOWLEDGEMENTS
Data were collected for this report over a period of more than 6 years. I could not possibly thank indi
vidually all the staff of ILCA's programme in Mali who were involved in the study but I am, needless to
say, grateful to each and everyone for their contributions.
Issa Haidara and Aba Mariko were largely responsible for collecting the field data, and latterly they
were supervised in this by Adama Traoré. Analyses were carried out in conjunction with ILCA's Com
puter Unit in Addis Ababa and Robin Sayers, Jeff Durkin and Darrell Light provided invaluable assist
ance and advice. Engda Girma cleaned and corrected the data.
Mary Wilson suffered much throughout in addition to typing many drafts over the years. Azeb
Melaku, assisted by Abeba Zenebe, struggled with appalling calligraphy, oversized tables and yet
further drafts while remaining both undismayed and cheerful.
Use Alipui is thanked for editorial assistance and the staff of the ILCA Publications Division for
designing, typesetting and printing this report.
 SUMMARY
This study provides data on livestock production systems in the West African Sahel with special refer
ence to Mali. Management practices are described and details of demography in relation to herd and
flock function are provided.
Data on cattle productivity were collected during a 6-year study of nine herds. Average age at first
calving was 49.5 months and mean parturition interval was 22 months. Calving was highly seasonal and
correlated with rainfall at 9 and 10 months previously. Total lifetime production of young was about
three calves per breeding cow. Growth was slow and mature weights were not reached until 5 years of
age. Weight losses occurred each year during the dry season even in young, growing stock. Over the 6-
year study period, mature weights of oxen declined by about 100 kg and those of cows by about 30 kg,
both being correlated with a reduction in annual rainfall. Abortions totalled 3.3% of all parturitions and
total deaths to 4 years of age were 31.6% . Adult mortality was about 5% and offtake about 8.4% . The
three productivity indices calculated (weight of calf produced per breeding cow per year, per kg breeding
cow per year and per kg metabolic weight of breeding cow per year) were 34.4 kg, 0. 16 kg and 0.70 kg.
Data on small ruminants were collected from 40 flocks. Average age at first parturition was 15.9
months, the mean kidding interval for goats was 9.6 months and for sheep 8.6 months, while litter sizes
were 1.19 and 1.04 for the two species respectively. The average numbers of parturitions for breeding
females present in the flocks were 2.2 for goats and 1 .8 for ewes. Growth was rapid and continued unin
terruptedly to 3 years of age with much less marked seasonal variations than in cattle. Abortions in goats
and sheep were 12.6% and5.1% respectively while total preweaning mortalities for the two species were
34.4% and 23.4% . At 4 years of age, only 38% of all goats and 30% of all sheep born were still in the
flocks. The productivity indices for goats were 14.6 kg, 0.494 kg and 1.23 kg and those calculated for
sheep were 28.4 kg, 0.867 kg and 2.22 kg.
SYNTHESE
Ce rapport est consacré aux résultats d'une étude sur les systèmes d'élevage sahéliens menée en Afrique
occidentale, au Mali. Les modes de conduite du troupeau, les variables démographiques et leurs liens avec
la finalité de l'élevage y sont présentés.
La productivité bovine a été calculée sur un échantillon de neuf troupeaux suivis pendant six ans.
L'âge moyen au premier vêlage est de 49,5 mois et l'intervalle moyen entre mises bas de 22 mois. Les
vêlages sont groupés et analysés en rapport avec la pluviométrie observée neufà dix mois auparavant. Les
observations portent en moyenne sur trois veaux par vache reproductrice présente dans le troupeau. Les
gains de poids sontfaibles et le poids adulte n 'est pas atteint avant l'âge de 5 ans. Des pertes depoids inter
viennent chaque année pendant la saison sèche, même chez les jeunes en croissance. Le poids adulte des
mâles a diminué d'environ 100 kg pendant l'étude et celui des femelles de 30 kg. Ce phénomène est lié à la
diminution de la pluviométrie. Le taux d'avortement est de 3,3% et le taux de mortalité cumulé jusqu'à
4 ans est de 31,6%. La mortalité des adultes s'élève à 5,0% et le taux d'exploitation est évalué à 8,4%. Les
indices de productivité (poids de veau produitpar reproductrice et par an, par kg de vache adultepar an et
par kg de poids métabolique de reproductrice par an) sont respectivement de 34,4 kg, 0,16 kg et 0, 70 kg.
Chez les petits ruminants, la productivité a été calculée sur un échantillon de près de 40 troupeaux.
L'âge moyen à la première mise bas est de 15, 9 mois et l'intervalle moyen entre mises bas de 9,6 mois chez
la chèvre et de 8,6 mois chez la brebis. La taille moyenne de la portée est de 1,19 chez les caprins et de 1,04
chez les ovins. Le nombre de naissances par reproductrice présente dans le troupeau est de 2,2 pour les
chèvres et de 1,8 pour les brebis. La croissance est rapide et continuejusqu'à 3 ans et les pertespondérales
de saison sèche sont moins importantes que chez les bovins. Le taux d'avortement atteint 12,6% chez les
caprins et 5,1% chez les ovins et la mortalité au sevrage 34,4% et 23, 4%. Le taux de survie des animaux de
4 ans est de 38% chez les caprins et 30% chez les ovins. Les indices de productivité atteignent 14,6 kg,
0,494 kg et 1,23 kg chez les caprins et 28,4 kg, 0,867 kg et 2,22 kg chez les ovins.

PART ONE
GENERAL BACKGROUND

1. INTRODUCTION
GENERAL
Almost all livestock output in developing coun
tries results from animals kept in the traditional
sector. However, attempts to improve output are
based almost exclusively on transferring tech
nologies developed under experimental station
conditions. While local types of animals are often
used on these stations, usually little is known of
the actual performance 'across the fence' under
traditional systems of management.
In an attempt to overcome this deficit of in
formation, the International Livestock Centre for
Africa (ILCA) has embarked on a series of sys
tems studies in various ecological zones. Central
Mali is part of the Sahel ecoclimatic zone where
ILCA has been conducting a long-term study
since the beginning of 1976. The results presented
in this report are from data collected over a period
of more than 6 years from 1978 to mid-1984.
In 1981 the livestock population of Mali was
estimated to include 5.1 million cattle, 6.3 million
sheep, 7 million goats and 173 000 camels (FAO,
1981). These figures were similar to those re
ported before the 1968-1973 Sahelian drought. In
terms of numbers, cattle are thus very important
in the Malian livestock economy. They represent
more than 70% of the total liveweight of all
domestic ruminants, and therefore are the major
consumers of the fodder resources available.
CATTLE TYPES AND DISTRIBUTION
Most cattle in Mali are in the Sahelian and Sahelo-
Sudanian zones in the central belt of the country
between about 12° and 17° latitude. Annual rain
fall in this area varies from 100 mm in the north to
1200 mm in the south. North of 17° there is virtu
ally no permanent livestock population. To the
south, livestock management is complicated by
the presence of the tsetse fly, although there are
some Bos taurus cattle of the N'Dama and West
African Shorthorn types in this zone, and it is pos
sible to maintain reasonably high levels of pro
duction under chemical prophylaxis (Logan et al,
1984).
The predominant breed of cattle in central
Mali is the Sudanese Fulani (Mason, 1969). A
male animal of this breed is shown in Figure 1. In
the extreme west near the Senegal border, some
Gobra are present; Maure cattle occur seasonally
along the Mauritanian frontier and the Azaouak
breed belonging to the Tuareg is found in the east.
Figure 2 shows the principal features of central
Mali and the distribution of the main cattle types.
SMALL RUMINANTTYPES
AND DISTRIBUTION
Based on the demographic structure of the live
stock population and the mean population
weight, goats and sheep account for 22.2% of the
total domestic ruminant biomass in central Mali.
They are thus an important resource not only in
the livestock economy but also in the total
economy of the country. Their ability to produce
meat, milk and other products even under the
harsh environmental conditions of the semi-arid
zone and at periods of the year when cattle are not
producing makes them very important in the
livelihood of traditional pastoralists and agropas-
toralists.
In the study area, direct subsistence is a
major objective of small ruminant husbandry.
Meat, milk and fibre all contribute to this sub
sistence while surplus males are sold off for cash.
In central Mali the majority of goats are of
the West African Sahel type (Mason, 1969). In
some urban areas and in the south small numbers
of the dwarf type of goat, common in the more
heavily wooded and forested humid zones, are
found (Figure 3). Although attempts are some- ligible and there seems to be no justification for
times made to classify the Sahel goats (Figure 4) this practice,
into breeds, the difference between them is neg-
Figure 1 . A Sudanese Fulani stud bull at the Station du Sahel, Niono.
 
Figure 2. Distribution ofcattle types in central Mali.
 
Figure 3. A goat ofthe West African Dwarftype.
 
Figure 4. A goat of the West African Long-legged type.
 
The variation in sheep is greater, and there
are three distinct types of the West African Long-
legged sheep (Mason, 1969). These are:
• The Toronké variety of the Fulani breed
(Figure 5), which is owned by Fulani trans-
humants who migrate seasonally over
most of central Mali. Animals of this vari
ety (and its crosses with the West African
Dwarf) are also owned by most of the
agropastoral sedentary farmers of the re
gion. This variety is the one used in the
long-term studies described in this report.
• The long-haired Black Maure (Figure 6),
which is found in the dry season along the
Mauritania/Mali border and southwards
for distances of up to 100 km. The flocks
are kept as pure as possible so that the hair
used in tent-making will be of good qual
ity. As most Moors keep flocks of the
short-haired White Maure or Tuabir sheep
in addition to the black breed, some cross
ing and dilution does occur.
• The white, pied or fawn Tuareg breed
(Figure 7), which is kept mainly by nomadic
or transhumant Tuaregs. In central Mali,
the breed is found mainly in the Gourma
region but it also extends north of the Niger
river to the Adrar des Moras and eastwards
into the Niger Republic.
In addition to these hair sheep, one of the
few true wooled breeds of Africa also occurs in
central Mali. This is the Macina breed (Figure 8),
Figure 5 . Toronké variety ofthe Fulani sheep at Niono market.
 
Figure 6. Black Maure ram in a harvestedfield.
 
-&*S£f?.y*£
Figure 7. A ram ofthe Tuareg type in central Mali.
 
Figure 8. Macina wool sheep.
 
which is kept mainly for its coarse wool which is
used in local blanket and garment manufacture.
The owners are almost all agropastoral Fulani
who also cultivate rice. The sheep is confined
largely to the inundation zone of the River Niger,
but short-distance migration takes place each
year to avoid the seasonal floods.
The distribution of these types of sheep in
central Mali and adjoining areas is shown in
Figure 9.
Figure 9. Distribution ofthe main sheep types in central Mali.
 
2. THE STUDY AREA
LOCATION
Central Mali, defined as the area of the country
between latitudes 12° and 16° N, was the area of
study covered by this report. Most of the detailed
work was carried out using Niono (14°15' N,
6°0' W) as the centre of operations. Niono is the
location of the Station d'Elevage et de Re-
cherches Zootechniques du Sahel, one of the
research stations operated by the Livestock
Division of the Malian National Institute for
Research in Livestock Production, Forestry and
Hydrobiology. This station was used as a base by
the ILCA research team, and ILCA scientists col
laborated with the government staff throughout
the period of the study. Most of ILCA's field
work was carried out off-station among tradi
tional livestock producers. Most herds studied
were within a 40-km radius from Niono except for
three situated at Segou, some 120 km to the
south. Additional specific studies were underta
ken for varying periods of time throughout cen
tral Mali.
CLIMATE
Central Mali has a Sahelian-type climate charac
terised by highly seasonal, unimodal rainfall fal
ling mainly in the period from June to September.
Precipitation is highest in late July or early Au
gust. The long-term (1930-1976) average annual
rainfall at Niono is about 550 mm, but since the
start of the notorious Sahel drought of 1968 to
1973, the rainfall has been considerably less.
During the 1978-1983 study period the annual
average rainfall was less than 400 mm; details
(averaged for five stations within 40 km of Niono)
are given in Table 1 . Occasional minor falls of rain
in months other than those shown in the table are
ignored, as this rain does not contribute to primary
production.
Mean temperatures in the area vary from a
high of 31°C in May and June to a low of 23°C in
January. Humidity is highest in June and through
out the rainy period of July to September; it is
lowest in February and early March. A climatic
diagram for Niono, which can be considered typi
cal of much of the central Mali area, is given in
Figure 10. Fuller details of the rainfall regime of
the area can be found in Wilson et al (1983).
VEGETATION
In the south of the study area, where annual rain
fall is generally in excess of 800 mm, the vegetation
is a typical north-Sudanian deciduous woodland
with many Combretaceae (Terminalia acroptera,
T. avicennioides , Combretum ghazalense) and
Leguminosae in the tree layer. Tree density is of
the order of 150-3000 per hectare with a canopy
cover of 20 to 50% . A feature of this zone is the
large number of economically useful species re
tained within the agro-sylvo-pastoral system,
such as Parkia biglobosa (the locust bean tree),
Butyrospermum parkii (the shea-butter nut or
kariti tree) and Adansonia digitata (the baobab
tree). The last is an introduction in this zone and
is valued for its bark (for string-making), for its
leaves (dried and used as a vegetable) and for a
variety of other uses. The herbaceous layer is
composed of medium to tall perennial and long-
cycle annual grasses, including Sporobolus
pyramidalis, Andropogon gayanus, A. pseuda-
pricus, Diheteropogon hagerupii and Loudetia
togoensis. Primary production of the herbaceous
layer is of the order of 3.5 1 DM/ha.
In the central zone of 300 to 600 mm annual
rainfall, some Combretaceae remain in the tree
layer but there is, as one moves north, a greater
dominance of thorny Leguminosae, including
Pterocarpus lucens, Acacia seyal, A. laeta and
 Table 1. Rainfallfor Niono, 1977-1983.
Month
Rainfall (mm)
Mean
1977 1978 1979 1980 1981 1982 1983
May 21.8 3.6 28.7 7.3 31.0 9.8 0.0 14.6
June 40.3 32.4 68.7 40.4 40.8 41.4 2.3 38.0
July 77.5 162.6 78.9 96.1 107.6 60.6 45.7 89.9
August 177.3 149.2 141.4 122.2 171.9 179.7 102.2 149.1
September 67.2 90.1 85.3 66.5 13.0 46.0 15.1 54.7
October 0.0 10.8 7.9 9.7 0.0 0.0 0.0 4.1
Total 384.1 448.7 410.9 342.2 364.3 337.5 165.3 351.9
Dichrostachys cinerea. Sclerocarya birraca and
Anogeissus leiocarpus are important in the south
ern part of this central zone. On sandy terraces,
Acacia albida is a conspicuous component of the
woody flora and is much valued as a provider of
dry-season forage of high protein content. The
number of trees per hectare varies from 500 to
2000, with a canopy cover of 5 to 15% in the
north. Grasses are mainly annuals, including
Elionurus elegans, Schoenefeldia gracilis and
Chloris prieurii, with a maximum standing crop
biomass (SCB) of 2500 kg DM/ha in the south and
1200 kg DM/ha in the north.
In the north-Sahelian zone where annual
rainfall is below 300 mm, taller trees are replaced
by shrubby growth, including many thorny
species and some Euphorbiaceae. Typical species
are Grewia bicolor, Boscia senegalensis , Maerua
crassifolia and Acacia tortilis. The maximum
number of trees per hectare is 500, with a canopy
cover not exceeding 10%. The field layer is domi
nated by short-cycle annual grasses including
Eragrostis tremula, Cenchrus biflorus, Aristida
mutabilis and Tragus bertheronianus . One of the
few perennials on dune sands is Panicum tur-
gidum . The SCB of the field layer does not exceed
800 kg DM/ha.
In central Mali, the vegetation pattern is
complicated by the complex geomorphology of
the so-called 'dead' delta and the present inunda
tion zone of the Niger river, also known as the
'live' delta. Fuller details of vegetative composi
tion and primary productivity can be found in
Wilson etal (1983).
Figure 10. Climatic normalsfor the Niono area, 1930-1976.
Rainfall
(mm)
160 I
oi A
J M M J S N
Month
Maximum and
minimum temperature
(°c)
4CH
30+*"
20
10
^«-«-»~».
^•"•••1^
J M M J S N
Month
Maximum and
minimum RH
Day length
(hr s)n I3-,
90-
70;
50
30:
12-
IIv \10
J M M J S N
Month
J M M J S N
Month
3. LIVESTOCK PRODUCTION SYSTEMS
LIVESTOCK IN THE SAHEL ZONE
As shown in Table 2, the Sahel proper (including
the same ecological zone in the Sudan republic)
has a total estimated livestock population of over
33 million tropical livestock units (TLU, these
being the 'unite de betail tropical' of francophone
authors), each equivalent to 250 kg liveweight.
This is over a quarter of the livestock population
of tropical Africa (CIPEA, 1981), which suggests
that the Sahel is an extremely important region
in terms of animal production. Over the past
15 years, the Sahel, previously neglected, has
been a focus of world attention, as the 1968-1973
drought clearly showed the vulnerability of this
area and the pressing need to solve its develop
ment problems. In many parts of the Sahel, live
stock are the only means of subsistence, since the
extreme variability of the rainfall makes cropping
very risky. Combining crop with animal produc
tion improves food security. In some areas of the
Sahel, cropping would not be possible were it not
for the presence of livestock which provide an al
ternative source of food for the human population
when the periodic grain deficits caused by poor
rainfall occur. Livestock in the Sahel therefore
play a part which is not simply related to mere
numbers.
Livestock distribution in the West African
Sahel shows a high degree of diversity. In some
parts, a clear congruence can be seen between
areas of medium-to-high densities of livestock
and areas of dense human populations (Figures 1 1
and 12). This is the case in western and central
Senegal and, even more markedly, in the inten
sive agricultural zone of southern Niger, where
intense population pressure and shortage of
fallow land forces animals to move north during
the cultivation season. However, this correlation
does not occur in Burkina Faso or Mali, where
there is a relatively high density of animals in the
western part of the Niger inundation zone and low
densities on the Dogon and Mossi plateaux,
where farmers are too poor to invest in animals
and do not use cattle for traction. A map of live
stock numbers per rural inhabitant (Figure 13)
indicates the extent to which the rural population
is dependent on pastoral activities, clearly show
ing the importance of livestock in the north of the
region as compared to the south (except for south
ern Niger and southern Senegal).
The important pastoral areas in Mali are
found in the central and northern parts of the
country, especially in the two administrative divi
sions of Nioro du Sahel and Nara (where pastures
along the Mauritanian border are used by
Moors), the northwest part of the Niger delta
(where Fulani animals are grazed) and the
Gourma-Rharous, Gao and Menaka divisions
(which are part of the Tuareg pastoral zone).
Other important pastoral areas are in Niger (not
ably the Wodabe and Tuareg areas in the west and
centre of the country) and in the Dori region of
northern Burkina Faso (where both Fulani and
Tuareg herds are found).
LIVESTOCK PRODUCTION IN MALI
Ecological zones
Crop and animal production systems in Mali are
influenced primarily by rainfall. The mean annual
rainfall isohyets for Mali and the main ecological
zones delineated by them are shown in Figure 14.
However, the area flooded annually by the Niger
river has its own characteristic vegetation and
land use. On the basis of rainfall and annual flood
ing it is possible to define four broad ecological
zones.
The arid zone. Dominated by pastoral pro
duction, this zone includes all land receiving less
Table0.Distributionofrum nantl vestockbysp ci stheSa elco ntri s.
TotalLU
00 00 04 1000 3g0
070 010
(TLU)" 00
570 40 000 00 000
Goats
('000ead)
000 010 00 0600
040 00 00
(TLU)"
000 000 010 00 00 000
S0eep
('000ead)
0000 00 0100 0000 040
i00
100
(TLU)''
00 0i 010 200
2000 00
Cattle
('000ead)
0000 00
00
000 000 00 00
(TLU)«
0 00
0
00 00
00
Camels
('000ead)
0 00 0 00 00 00 00
Country Senegal
Mali
BurkinaFaso
Niger C0ad Sudan
Total0ead
*TLUswerecalculat dontherollowingb sis:"00l=0.T ;=adofc=t e07;nsh prg tLU.
Sources:FAO(060);CIPEA0).
*;it--'>---* ***"'-^
than 600 mm annual rainfall (400 mm with 80%
probability), but excluding the inundation zone of
the Niger. The arid zone includes two major belts.
The first consists of those areas receiving less than
200 mm where cropping is not possible and
where, north of the 100 mm rainfall isohyet, the
vegetation becomes typically Saharan. In the sec
ond belt, often termed the northern Sahel and
where rainfall is usually between 200 and 400 mm ,
some rainfed agriculture is practised but is ex
tremely risky, since the coefficient of variation of
annual rainfall is 25 to 35%. Thus defined, the
arid zone accounts for well over half the land area
of Mali.
Figure 11. Livestock densities in the Sahel countries.
MAURITANIA
Livestock density (TLU/km2) 
1 = 1 camel, 1.37 cattle c
8.33 sheep/goats
Figure 12. Human population densities in the Sahel countries.
Human population density (people/km2)
 
11
 The semi-arid zone. This area is associated
with pastoral and low-potential rainfed crop pro
duction and is located approximately between the
600 and 1000 mm isohyets. It includes the south
ern Sahel and north-Sudanian vegetation zones,
extending in a relatively narrow belt across the
southern half of the country. Rainfed millet is the
main crop, being replaced by sorghum, ground
nuts and cotton towards the south. The coeffi
cient of rainfall variation is 20 to 25% .
The subhumid zone. Associated with high-
potential rainfed crop production, this zone lies
south of the 1000 mm isohyet, covering only the
southern fringes of Mali. It includes the southern
Sudanian vegetation belt. The higher and more
reliable rainfall permits rainfed cultivation of cot
ton, while sorghum and to a lesser extent maize
are the dominant food crops.
The inundation zone of the Niger river. As
sociated with floodplain grasslands and some
farming, this natural region forms a fourth zone
where vegetation is strongly influenced by the
annual flood. Its extent is coincident with the dis
tribution of Macina sheep (Figure 9). The zone
includes the rich grasslands grazed by Fulani pas-
toralists, areas of traditional cropping, and areas
of rice cultivation under semi-controlled irrigation.
Distribution of people and animals
Rural population densities in Mali are given in
Figure 12. The pattern shows three main areas.
The first is a large area of very sparse settlement
(<6 people/km2), roughly equivalent to the pas
toral zone with its predominantly livestock-based
economy (the 'pure' pastoral production system),
but also including the hilly area in the southwest
of the country, which is heavily infested with the
tsetse fly. The second area has medium to dense
settlement (>14 people/km2) and includes: the
Dogon Plateau and adjacent areas; Dire, a small
district with an extensive irrigation potential re
sulting from the nearby course of the Niger; all
the remaining irrigable areas of the floodplain;
the Office du Niger, a controlled irrigation
scheme owned and normally managed by the
state; and the intensive rainfed agricultural areas
of southeast Mali together with the area around
Bamako. The remainder of the country consti
tutes the third area, which has moderate settle
ments (6-14 people/km ).
Figure 1 1 shows the density of livestock per
km2 in TLU. For Mali, the livestock pattern indi
cates two main areas. The first has low animal den
sity and lies in two different parts of the country:
in the pastoral zone where large tracts of land are
scarcely used because of lack of water, and in
southwest Mali where tsetse flies restrict livestock
production. Both correspond with areas of low
human settlement. The second area is one of high
animal density in the Niger floodplain, corres
ponding closely with a production system in which
pastoralism is associated with irrigated cultiva
tion. This pattern occurs to a lesser extent in the
Figure 13. Numbers of TLUper rural inhabitant in the Sahel countries.
MAURITANIA NIGER
2°
VOLTA Y/'//////''///////'//,^:^mm
TLU/rural inhabitant
□ 0 -0.39
(53 0.4-0.79
HI 0.8- 1.19
HH 1.2- 1.59
ES >l.6
 
wsmmmwmimm JNft, mft''',,//''".
mm
TLU=i camel, 1.37 cattle or
8.33 sheep/goats
500 km
12
Figure 14. Annual rainfall and ecological zones in Mali.
Rainfall station
Rainfall station with data checked
for reliability
 
North -Sahelian
South- Sahelian
North -Sudanian
South- Sudanian
intensive rainfed agricultural areas of extreme
southern Mali.
The number of TLU per rural inhabitant is
shown in Figure 13, giving an impression of the
dependence on livestock or the livestock-based
wealth in different parts of the West African
Sahel. In the case of Mali, the two highest density
categories (i.e. >1.2 TLU/rural inhabitant) cor
respond closely with the northern pastoral zone
and the Niger inundation zone, i.e. with the pure
pastoral and floodplain-associated pastoral pro
duction systems. The area with the highest ratio
of animals to people (4.4 TLU/rural inhabitant) is
the Gourma-Rharous district in central Mali.
Outside the northern pastoral and inundation
zones, livestock numbers per rural inhabitant are
much lower, especially in the sparsely inhabited
southwestern parts of the country.
If human and livestock densities are com
pared, it appears that the areas with high numbers
of livestock per rural inhabitant are principally in
the north and east of the country, whereas those
with a high human population are mainly in the
south. The two zones scarcely overlap at all, indi
cating the extent to which agriculture and pas-
toralism are still separate activities. The exceptions
to this situation are the central Malian districts of
Mopti, Macina, Dire" and Niafunké, all of which
are in or near the Niger floodplain and contain
both agricultural and pastoral systems.
The areas of high cattle density (Figure 15),
containing an average of more than one head of
cattle per rural inhabitant, occur in a belt along
the southern edge of the pastoral zone: in the
Nioro du Sahel and Nara districts along the
Mauritanian border; in the Niger floodplain; in
the Gourma-Rharous district; and in the Menaka
district in the east. This pattern corresponds with
the production systems in which pastoralism is as
sociated with either dryland or floodplain cultiva
tion, as well as with the southern fringe of the
'pure' pastoral system.
13
Figure 15. Number ofcattle per rural inhabitant in Mali.
 
The distribution of sheep and goats per rural
inhabitant (Figure 16) is somewhat different. The
three highest density areas, containing more than
2.2 sheep or goats per rural inhabitant, form a belt
corresponding roughly with the pastoral zone and
the 'pure' pastoral production system. Sheep and
goats have a more northerly and easterly distribu
tion than cattle, with the highest densities lying in
the Gourma-Rharous, Gao, Bourem and Menaka
districts. They are much less important in the rest
of Mali, especially in the south. With the exception
of Macina wool sheep, small ruminants are less
numerous in most of the Niger floodplain which is
first and foremost a cattle zone.
The estimated distribution of people and
livestock by ecological zone in Mali (Table 3)
shows that the rural population is concentrated
mainly in the semi-arid zone which, although it
covers only 16% of the country, contains 45% of
its inhabitants. The subhumid zone to the south
contains just over one quarter of Mali's rural
population, while the arid zone and the Niger
river inundation area contain another 15% each.
The arid or pastoral zone contains over 90%
of Mali's camels and 43% of its sheep and goats,
but only 21% of its cattle. However, it should be
remembered that the cattle, sheep and goats from
the Niger floodplain also use the rangelands of the
pastoral zone for 6 months of the year, as do many
cattle from the extensive rainfed agricultural
areas farther south. Conversely, animals from the
pastoral zone use pastures and crop residues in
agricultural areas after the harvest and until the
following rainy season.
Table 3 highlights the importance of the
semi-arid zone, which contains one third of Mali's
cattle and nearly one third of its total TLU, al
though it is far smaller in area than the arid zone.
Production systems
Two main criteria have been used to define the
animal production systems of Mali: the first is the
degree of dependence on pastoral products for
the gross revenue or food supply of the household
or production unit; the second is the particular
type of agriculture associated with the livestock
system. Other criteria, such as the duration and
distance of livestock movement, were considered
as less important. While livestock movement may
be an important aspect of an animal production
system, it is contingent upon the system, often
having the effect of diverting attention away from
the main criterion, which is the degree of depen
dence on the animals raised.
14
Figure 16. Number ofgoats and sheep per rural inhabitant in Mali.
 
No. of small ruminants/rural inhabitant
£23 1. 1 1-2.21
V/A 2.22-3.32
3.33-6.65
>6.66
Somewhat arbitrary limits were set when the
degree of dependence on livestock products was
classified. A system in which more than 50% of
gross revenue (the value of subsistence plus mar
keted production) or more than 20% of house
hold food energy was directly derived from live
stock or livestock-related activities was classified
as a pastoral system. One which derived between
10 and 50% of gross revenue from livestock, in
other words 50% or more from agriculture, was
classed as an agropastoral system. A third system,
in which less than 10% of revenue was derived
from livestock, might have been classified as
'agricultural' system but lay outside the scope of
this study as was a possible fourth 'urban' system.
It should also be pointed out that the concept of
gross revenue includes the theoretical value of
camels or other species as transport animals, and
the value of cattle for traction and manure pro
duction. An indication of the degree of depen
dence on livestock in different production sys
tems in Mali and the West African Sahel is given
in Table 4.
The two main criteria thus give rise to two
major production systems, pastoral and agropas
toral, each with a number of subsystems. These
subsystems are shown in Table 5, in which the
main characteristics of each are listed. Three pas
toral subsystems can be identified. The first is a
'pure', mainly camel-based, system in the north
ern arid zone, characterised by high mobility, and
having almost no direct links with agriculture.
The second, found in the northern central and
northwestern semi-arid areas, is one in which ani
mal production is associated with dryland crop
ping, with some cultivation and the exchange of
manure for stubble grazing. Cattle, goats and
sheep are the main species raised. In the third
subsystem, specific to the inundation area of the
Niger river and its hinterland, animal production
is linked with floodplain grazing and farming;
cropping is more important here and cattle are the
main species raised.
Under agropastoral ism. three subsystems
can again be identified in which mobility is very
low, cropping is the major component, and the
main species raised is cattle, often used for
draught. In the first of these subsystems, found in
the central semi-arid regions, animal production
is associated with the rainfed cropping of millet,
mostly for subsistence. In the second, found in the
'dead' delta of the Niger, livestock are raised by
producers who are under land tenure contract
with the Office du Niger irrigation scheme, and
the main crop grown is rice. In the third subsys
tem, located in the southern subhumid zone, ani
mal production is a minor component associated
with both cash and subsistence cropping, with
15
Table0.Distributionofpeopleandnima sbyecol 0icalz neiMa .
a
r0 r08) r00) r00) r00) r000)
T0'
r'000)
0000 050 000 000
05
0oats
r0 r00) r00) r00) r0) r000)
S0eepand
r'0000ead)
05 00 00 50
00
Cattle
r0
ron) r00) r5) r05) r000)
r'0000ead)
00 550 0000 05 050
Camch
r0 r5)
r0) - -
r000)
r'0000ead)
050 0 - - 00
Roralpopolation
r0 r0) r00) r00) r0) r000)
r'000)
00 055 055 000 55
Area
r0 r00) r05) r00)
r0)
r000)
rkm2)
0on on0 00
50 000
Zone
Aridzone
Semioaridz ne Sob0omidzone
InnerN0 rdelta
Total
"Onecamel=0.0TLU;onheadofc=t,7og atro0e p0.0TLU.
millet, sorghum, groundnuts and cotton being the
main crops. The technical part of this report deals
mainly with the first and second of these agro-
pastoral subsystems.
An approximation of the distribution of
livestock in the different production systems is
given in Table 5. This suggests that 30% of live
stock in the semi-arid zone belong to the agro-
pastoral system, while the remaining 70% belong
to the pastoral system associated with rainfed
cropping. In the arid zone, 50% of the cattle and
70% of the small ruminants are raised under the
pure pastoral subsystem, the remainder being as
sociated with the rainfed millet subsystem.
The relative importance of the domestic
species found within the systems and subsystems
varies considerably. The only exception relates to
camels, which are confined almost exclusively to
the pure pastoral subsystem in the arid zones al
though a few are found in the pastoral/dryland
cropping subsystem. However, even in the pas
toral system, camels are numerically unimportant
compared to cattle or sheep and goats. Because of
their size and cash value, cattle certainly consti
tute the most important species in all the systems,
economically if not culturally, and irrespective of
whether their role is primarily milk production (as
in the pastoral systems), draught power (as in the
agropastoral systems), or a combination of these
two roles plus meat production where system
boundaries are not clear-cut. In the urban system,
donkeys would replace cattle, to be followed
closely by goats, as the most important species. In
terms of TLU, cattle dominate the livestock sec
tor in Mali (Table 3), although the numbers of
both sheep and goats are probably considerably
and consistently underestimated. Sheep are ex
tremely important in the pastoral system, espe-
Table 4. Dependence on livestock in different Sahelian production systems.
System,
country,
ethnic group
Percentage
of gross
revenue
from
livestock
Percentage of total kcal
consumed
Pure pastoralism
Milk Meat Cereals Year Source
Mali
Northeastern
Tuareg 99 68 8 24 1971 (1)
Niger
Tuareg 80 51 3 47 1963 (1)
Fulani 96 39 2 58 1963 (1)
Chad
Annakaza - 48 - 24 1950 (1)
Pastoralism/
rainfed cropping
Burkina Faso
Fulani 78 12 3 85 1977 (2)
Niger
Fulani - 24 2 74 1963 (4)
Tuareg - 33 2 65 1963 (4)
Tuareg - 17 3 80 1976/7 (3)
Pastoralism/
floodplain farming
Mali
Fulani 57 25 - 75 1958 (5)
Agropastoralism
Mali
Bambara 10 0.5" 0.8 95 1974/5 (6)
Burkina Faso
Mossi 10 - - - (2)
* Probably underestimated.
Sources: (1) Swift (1979a, b); (2) Delgado (1978); (3) Eddy (1979), but kcal values recalculated using 850 kcal/kg for milk and 1450
kcal/kg for meat; (4) Swift (1979a); (5) Swift (1979a) quoting Gallais (1977); (6) IER (1975).
17
Table0.M inchar cteristicsoflivestockproduct onsy temsnWeAf i a.
Associatedwit0
rainfedagriculture
medium
medium/00 00 medium/00
cas0crop, 00-400 0.040.0 0.0-0.0 medium
Agropastoralsystems
00
fieldscultivated,anim ltraction
cropresiduesaimpo tant
lowandf rs0 rtdistancesuringt emain
cultivationseason
060 low
Associatedwit0
irr0ation
00
variable
i0.040.0
00 medium/00
0.04.0
00 low low
low/medium
low
Associatedwit0
rainfedagriculture
o0subsistence
0
600-100 0.0-0.0 medium
low/medium
0.04.0
00
medium medium
00 0g0
Associatedwit0
irrigation
00
variable
o0fields cultivated
i0.0-4.0
0i0/very0
medium/00
i.04.0
00inwet
season
medium/00
low/medium low/medium low/medium
low
Pastoralsystems
Associatedwit0
rainfedagriculture
10 00L00
somecultivat"n,
exc0angeofmanure
0.040.0 low/medium low/medium 0.04.0
medium,fixed
base medium low 00 low
medium
Rice 00
<600 weak
0.0-0.0 verylow
low
0.04.0
0g0,nofixed
base
0g0 0g0 00 low low
Contributionflivestock
torevenue(%)
Rainfall(mm/year)
Relationswi 0agriculture
NumberofTLU/000a
Carryingc pac ty
People
Animals
TLUperperson*
Senegal
BurkinaFaso *Rangeorratios.
C0aracteristic Importancein
Mali
Mauritania
NigerMobility
daily in the northern and floodplain areas. In the
agropastoral system, goats are much more im
portant than sheep, at least at the level of the indi
vidual producers.
In the agropastoral system and especially in
the irrigation and cash crop subsystems, work
oxen are an extremely important element in the
total livestock holding. Donkeys in these subsys
tems also play a large, if usually unpublicised, role
in providing transport for agricultural products
and fuelwood. Horses are of minor importance in
all systems, perhaps vested with the only remain
ing traces of the prestige syndrome once consid
ered to be the principal preoccupation of live
stock owners.
Herds are built up by a variety of processes
including inheritance, gifts, loan arrangements
and natural increase. In the pure pastoral subsys
tem, some animals may be purchased with the
proceeds of caravan trading or with cash from
wage labour, often earned outside Mali. In the
pastoral/dryland cropping subsystem, some agri
cultural profits may be reinvested in livestock, as
they also are in the floodplain subsystem. In both
of the two last-named subsystems women can own
considerable numbers of animals, acquired by
dowry and from the proceeds of milk sales. In the
agropastoral system, work oxen are often ac
quired through credit facilities arranged with a
relevant organisation such as the Office du Niger.
Livestock also play a role in savings, the re
turns from agriculture being diverted to this end
in the absence of alternative forms of investment.
In fact, there is considerable evidence that rapid
changes in ownership are taking place and that
these are not only among farmers. The new own
ers include civil servants, service personnel and
merchants as well as agriculturalists who invest
large amounts of cash in livestock. This tendency
brings with it a rapid increase in the use of salaried
herdsmen and the growing acquisition of grazing
rights by non-pastoralists, a trend which may well
be undesirable from the point of view of an equit
able distribution of the country's resources. Inte
gration within the market economy is also in
creasing, in some cases rapidly, although there is
still a wide range of market involvement, from al
most wholly subsistence-based production to
cropping for cash alone.
Subsystems in the agropastoral system
Table 6 lists some of the main differences between
the millet and rice subsystems which might ac
count in large part for performance differences
between the two. The rice subsystem appears to
have certain advantages over the millet subsys
tem, which may lead to increased productivity. In
the rice subsystem, for instance, the stall-feeding
of castrate sheep, known throughout fran
cophone West Africa as the mouton de case sys
tem, is much easier owing to the greater availabil
ity of crop residues (Kolff and Wilson, 1985).
Water is also abundant all the year round,
whereas in the millet subsystem it is restricted
during the dry season.
Table 6. Management and environmentalfactors in the millet and rice subsystems in central Mali.
Factor Millet subsystem Rice subsystem
Water availability
Crop residues
Dry-season fodder
Supplementary feeding
Energy expenditure
Restricted in dry season
Limited in time, quantity
and quality
Very limited, some early
browse in hot dry season
Generally not practised
(little surplus cash) ;
salt occasionally provided
High in dry season due
to long distance to water
and sparse food availability
Abundant all year round
Longer season of availability (later
in dry season) better quality
Weed regrowth on irrigated fields
Fairly common, especially for sheep:
rice bran, legume haulms, leaves of Khaya
senegalensis. Work oxen sometimes fed,
but supplements are of insufficient quantity
and poor quality
Lower for longer period of time on account
of proximity of water and longer
growing period
19
 4. MATERIALS AND METHODS
THE STUDY HERDS AND FLOCKS
Following a preliminary survey in early 1978,
herds and flocks in six locations in the millet sub
system and in four locations in the rice subsystem
were selected for study. These units were then fol
lowed on a regular basis until June 1984. At the
start of the survey there were approximately
600 head of cattle in the herds and some 700 head
of goats and sheep.
At the end of June 1984, when the analysis
was undertaken, there were records for over 1150
cattle and 5000 small ruminants, both these totals
including those animals that were present at the
beginning of the study. Most of the animals enter
ing the herds did so by birth, although there was a
small number of purchases and some animals
were brought in as gifts, loans or exchanges.
HELD METHODS
At the start of the survey, all cattle were individu
ally identified by means of numbered plastic or
metal ear tags. Details of the animal's age (deter
mined by dentition and/or by owner questioning) ,
its function in the herd and its reproductive his
tory (if a female) were obtained. From these data,
an individual record card was established. All ani
mals subsequently entering the herds were
treated in the same way, and individual record
cards were also created for these animals.
Visits were made to cattle herds on a regular
basis , usually at intervals of about 4 to 6 weeks but
more often during the main calving period of late
April to early August. Owners cooperated in de
fining the events - births, deaths, purchases and
sales - which had taken place during the time pe
riod which had elapsed since the previous visit.
Details of all these events were entered on the in
dividual record cards or new record cards were
created, as appropriate.
Goat and sheep flocks were visited at inter
vals of about 15 days. Young animals were
weighed on most of these occasions, while older
animals were weighed on every second visit.
Lighter animals (estimated as less than 100 kg)
were weighed on most of the regular visits by sus
pending them in a sling attached to a dial-type
spring balance. Heavier animals were weighed in
a proprietary brand, transportable weigh crush.
Weighings of heavier cattle were made at less fre
quent intervals, usually four times each year:
these were timed to coincide with the end of the
rains (September/October), the beginning of the
dry period (late November to early January), the
early part of the hot dry period (March/April) and
the end of the dry season and beginning of the
rains (July). On occasions, logistical and adminis
trative problems made it impossible to adhere to
this timetable. Some weighings of adult cattle
therefore appear in the records for other months
of the year.
DATA ANALYSIS
The individual records for each animal were en
tered on a computer and the following subsets
were established from the main data base, relat
ing to the animal's identity and aspects of its pro
ductivity:
• Identity: individual number, system of
management, herd identity, reason for
entry, sex, parity, type of birth, date of
entry, reason for exit, dam number.
• Reproductive performance: age at first
parturition, date of parturitions, interval
between successive parturitions, identity
of young, sex of young, the parity (order in
its dam's reproductive record) of young.
Where previous reproductive history of
cows of mature age was unknown, the first
20
recorded parturition was set arbitrarily at
three, and where this parameter was un
known for goats or sheep, the first
recorded parturition was set at four.
• Weight: two subsets were established; one
for animals of known birth date from
which it was possible to calculate growth
rates and the effect of some variables on it,
and one for animals of unknown birth date
which was used primarily for establishing
weights for age in the older age groups and
for determining the magnitude of seasonal
weight changes in these animals.
• Dentition pattern: for a small number of
animals this was used to determine ages at
eruption of permanent incisors, but the
main use of this subset was in the weight-
for-age data at older ages.
• Milk production: small sets of records
were established on milk offtake for
human consumption for a few animals in
one herd of cattle and for the total milk
production of some Macina sheep.
Usual compilation and statistical techniques
using a pocket calculator were used for many pre
liminary analyses and as an aid to verification and
preparation of data. Most final analyses were car
ried out using recognised software packages, par
ticularly SPSS, BMDP, GLIM and the Harvey
(1977) model for the generalised least-squares
procedure suitable for use on data with unequal
subclass numbers. The estimated least-squares
means generated by this analysis, being adjusted
for the unequal subclass numbers, may differ
from the observed means.
Effects in the Harvey model usually included
the fixed effects of:
- subsystem: millet, rice
- month of parturition or birth
- parity : 1 , 2 . . . .n and '9' when unknown but
considered to be >3 for cattle and >4 for
small ruminants:
- type of birth: single, twin, triplet or multi
ple where twin and triplet births were com
bined
- sex: female, male
- herd within system
- year, and
- interactions of some of the above main ef
fects.
When adequate data were available, the ran
dom effects of dam within flock or within system
were used in a mixed model.
The residual mean square was used as the
error term in the Harvey model to test the signifi
cance of all differences. Linear contrasts of least-
squares means were computed to determine the
significance of differences between groups.
For cattle, a total of 519 births from 274 cows
or heifers and 244 calving intervals were recorded
during the study. The weights of animals of
known birth date totalled almost 6000 records: all
weights, including those from mature animals, to
talled over 10 000 records.
In small ruminants, 3605 parturitions gave
rise to 4049 young and more than 2000 parturition
intervals. Almost 40 000 weights were recorded.
ADDITIONAL SOURCES
OF INFORMATION
Considerable additional information was ob
tained from primary or secondary sources and has
been incorporated in this report. In particular,
this information related to the production systems
(Section 3), to the general field of management
(Section 5), to herd demography (Section 6) and
to some aspects of reproductive performance
(Sections 7 and 11) and productivity (Sections 10
and 14).
This information was obtained by means of:
• specific short-term studies,
• direct observations in the field,
• structured and unstructured interviews
with owners and herders,
• data collected by students of the Rural
Polytechnic Institute, supervised by ILCA
staff and published as dissertations,
• information extracted from internal re
ports (Programme Documents) of the
Arid and Semi-arid Zones Programme,
and
• other relevant literature.
These sources have been acknowledged in
the text and listed in Part 5 of this report.
21
 5. MANAGEMENT
OWNERSHIP PATTERNS
The numbers of households and people who own
cattle, as well as the actual numbers owned, vary
between the rice and the millet subsystems. The
farming system practised is also reflected in the
ownership of farm equipment. In Table 7, some
of these differences are shown for the two subsys
tems in central Mali. The higher percentages evi
dent for rice farmers result partly from a readier
access to credit through the parastatal irrigation
scheme, the Office du Niger.
Ownership patterns established from official
statistics give only a general idea of the variations
in numbers of people owning livestock and of the
importance of the animals in the whole system. In
the Niono area the Office du Niger carries out its
own 'census' of animals within the irrigated
perimeter. Its figures are additional to the official
administrative census and there is seldom any re
lationship between the two. Our own observa
tions have shown that the Office du Niger has no
apparent interest in small runminants, and these
classes of stock are largely ignored in their census.
The figures for small ruminants are not only mis
leading in terms of ownership and total numbers
but also in respect of the ratio of goats to sheep. In
the official census, where goats and sheep are
counted separately (it being evident from Table 7
that this is not always the case), individual families
are usually shown to have equal numbers of each.
Our own observations confirm the numerical im
portance of goats over sheep at a ratio of 5.4:1
in the millet subsystem and 1.4:1 in the rice sub
system. Table 8 details the ownership patterns
of 43 small ruminant flocks in the Niono area
censused by us.
The ownership of principal means of produc
tion, work oxen and ploughs, is also rather un
even, especially within the millet subsystem. As
can be seen from Figure 17, most families own
only one plough while a smaller number own pairs
of oxen. It might be inferred from this that there
is a tendency for a plough to be acquired before
oxen, and that the hiring and renting of oxen are
not uncommon practices in the area.
Some additional data relating to ownership
patterns are presented in Table 9 for various
ethnic groups and subsystems. It needs to be
stressed that all the animals in individual owner
ship are rarely herded as a single unit; a family
may split its herd and put a number of animals
into several other groups for the purposes of day-
to-day management. In the case of cattle such
groups vary in size from about 50 to 200-300
head, depending on the system, the nature of the
terrain and the time of year. Occassionally, bigger
herds of over 500 head may be constituted. The
ownership structure of such herds is usually very
complex (Table 10).
MOVEMENT
The extent of movement varies with a number of
factors. These include the need to keep animals
away from crops during the growing season, the
principal occupation of the owner, the necessity
of searching out the best fodder and water re
sources in the dry season, and the need to escape
from the Niger floodplain during the period when
this is inundated.
The longest annual movements are under
taken by the Fulani who normally reside in the
inundation zone. Beginning in July each year,
these people make treks of 300 km to the Mema
and to southeast Mauritania. The reverse journey
is made in October after the short Sahel rains
have stopped and the annual flood in the inunda
tion zone begins to recede. Similar but shorter
seasonal treks are made into the Seno-Mango,
southeast of the floodplain, with some herds con-
22
Table0.Holdingsflivestockandfarmimplem ntccor ingtoffic alst tisti srmplev lagenthem tricubsystemcentM i.
Carts
%
owning
4.1 0.0 0.0 0.0 01.0 00.0 0.0 0.0 0.0 00.0 0.i 0.0 00.0 1.1 00.1 4.0
Plou0s owning 00.0 00.0 0.0 00.i 00.i 00.0 00.0 00.0 4.0 0.0
0i.0
00.0 00.0 00.0 00.0 00.0
/o
Donkeys
rangein numbers 0- 0-
i40
0-
i-0 i-00
i
0- 0- 0- 0- 0- 0- 0-
i-0
0
o0ing
/o 00.0 10.0 00.0 00.0
g.0
0.0 00.0 00.0 0.0 0.0 00.0 4.i 00.0 4.0 0.0 0.0
rangein numbers 400 -00 -0 -0 -4 -0 -0 -0
-00 -00 -i0 -00
- 0 -
-0
Goats
o0ing
0 0 0 0 0 0 0 0 0 i 00 0 0 0
/o 10.0 10.0 0.0 1.0 00.0 1.0 00.0 4.0 00.0 0.0 0.0 0.0 0.0 0.0 00.0 0.1
rangein numbers -00 -00 -0 -0 -4 -0
- - 0
-0
0
-00
-
-1 -0
S0eep
o0ing
0 0 0 0 0 0 0 2 0 0
4L07
0.0 10.0 0.0 00.0 00.0 1.0 0.0 0.0 1.0 0.0 0.0 0.0
0i.0
0.0 0.0 0.1
rangein numbers*
-60
-00
-60 -0 -0
-00 -0 -0 -0 -4 -40 -60 -00
-10
-4
-00
Cattle 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
%
owning
g.0
00.0 1.0 00.0 10.0 1.0 0.0 01.0 10.0 01.i 00.0 01.0 00.0 01.0 0.0 00.0
Subsystem/village
(No.of0ouse0olds)
Pogo(00)
Kamono(lO)
Sissako(0)
Teninzana(0)
S0uine(00)
T0ng(60)
Siraouma(0)
Ndebougou(i4)
Ntila(00)
Bamada(4)
N0(4) N0(00) N0(4)
Ni0(01) Bl(0)
B0(4)
Millet" Riceb
*Nilholdingsexcludedroraspec es.
bDatarort0emill tsubsys emwe eakenromhadmini trativecensus;hcOrficuN g rnsus.
to
 Table 8. Ownership pattern andflock sizes for small ruminants in the agropastoral system ofcentral Mali.
Millet villages Rice villages
Goats Sheep Goats Sheep
Total number of flocks 16 27
Number owning species 16 7 26 15
Number owning goats only 9 12
Number owning sheep only 0 1
Mean flock size " 38.19 7.46 8.96 6.41
Mean flock sizeb 38.19 12.56 9.31 11.53
Range in flock size 2-91 0-58 0-23 0-58
* Irrespective of whether the holding of one or both species is nil.
b Nil holdings excluded.
Figure 17. Ownership of ploughs and oxen pairs in a
central Malian village.
No. of fomilies
12-
10-
8-
6-
4-
2-
0 2 4 C
No. of ploughs
0 2 4 6
No. of oxen pairs
tinuing south into Burkina Faso. The Moors also
spend the rainy season in Mauritania and move
south to Mali in the dry season. Here, they pas
ture their animals on the crop residues of the
Bambara sedentary cultivators, providing ma
nure in exchange for water from wells which the
farmers have dug in their fields (Figure 18).
The animals owned by the cultivators of the
Office du Niger irrigation scheme migrate over
shorter distances. Except for a few work oxen,
cattle in this subsystem move out of the irrigated
area at cropping time. On returning to the irri
gated area after the harvest, they benefit from the
abundant crop residues, and certain groups, nota
bly work oxen, may receive some special sup
plementary feeding. Cattle which move least are
those of the rainfed millet cultivators and those
owned by the urban populations in such centres as
Ségou, Mopti and Niono. The latter are kept by
town dwellers, including civil servants and mili
tary personnel, mainly as an investment or as a
source of milk. A supply of milk for sedentary
family members in other subsystems is assured by
retaining a few milking cows at the homestead
when other cattle go on their annual transhu-
mance.
In the rainfed millet cropping zone sheep are
managed under sedentary conditions, and sea
sonal movements, other than short-distance ones
away from the farmed areas during the growing
season, are not part of the management strategy.
Some supplementary feeding is practised either
from crop residues or from cut and carried browse
(especially Pterocarpus lucens) during periods of
food scarcity. In the irrigated rice areas of the Of
fice du Niger the patterns are similar but more
supplementary feeding is possible on account of
the greater availability of crop residues and the
fodder provided by weedy regrowth in fields and
on the borders of irrigation canals. Goats are
more important than sheep in these primarily
agricultural areas, outnumbering sheep in a ratio
varying from 1.7 to 2.7:1.
A transhumant system of management - in
which animals migrate seasonally but return to a
fixed base - is practised by the agropastoral Fu-
lani of the inundation zone, the Fulani of the
Gourma, the Moors and some Tuareg who live
close to the Niger river.
The Fulani of the inundation zone cultivate
rice under the annual flood regime of the Niger.
Because of the flood they are forced to move most
of their sheep on to the adjacent dry areas each
year. The extent of this movement can be judged
from Figure 19. The Fulani also have a sophisti
cated system of flock stratification (Table 16)
which involves leaving some sheep behind on the
24
Table 9. Ownership ofcattle and other stock by different ethnic groups in central Mali.
No. (average or range) of animals owned per household
Group/subsystem No. of Average
households household Cattle Goats and sheep Camels Donkeys
size
Bambara/sedentary , millet 30 16.7 25.0 31.7 0.0 2.1
Rimaibe/sedentary, millet 29 6.4 5.2 20.3 0.1 1.7
Fulani/transhumant, millet 27 7.5 24.5 9.5 0.3 0.1
Fulani/professional herders 10 5.0 1.0 35.0 0.0 0.0
Moora/transhumant, millet 9 6 3-5 30-50 10-20 3-4
Tuareg'/transhumant, millet - 17 30-50 50-60 2-3 2-3
* From Fofana (1974).
Table 10. Structure of two cattle management units in
the Niger inundation zone, by main profes
sion ofowner.
Profession
of owner
Structure (%)
Herdl Herd 2
(11 = 220) (n = 263)
Professional herder 23 18
Other livestock owner 37 32
Cultivator 2 3
'Investor' 13 30
Unknown 25 17
Total number of owners 48 29
village mounds in the flood area to provide milk
for those family members remaining to cultivate
the rice. Flock sizes are large with 69% of all
flocks having more than 100 head. Goats are of
little importance in this system, being outnum
bered by sheep in the ratio of 7: 1 .
The Moors spend the rainy season in
Mauritania where they cultivate small patches of
bulrush millet (Pennisetum typhoides) on an op
portunistic basis. They are drawn south to Mali
during the dry season where fodder conditions on
the open range are marginally better. The Moors
have a close relationship with the settled cul
tivators in the areas in which they spend vthe dry
season. This involves the cultivators' digging
wells on their farmed land and offering the water
to the Moors. The latter then camp on the farmed
areas and their animals enrich the soil with their
droppings and urine. Sheep are of relatively less
importance to this group than to the Fulani of
the inundation zone but still outnumber goats in
the ratio of 1.6:1. Flock sizes generally exceed
80 head.
Transhumant Fulani roam over most of cen
tral Mali, following more or less fixed orbits, each
one particular to its own area. Cattle are para
mount to their way of life but small ruminants
occupy an important subsidiary position, with
sheep being outnumbered by goats in a ratio of
about 1.0:1:6. The average flock size is about
25 head. Conflicts between the Fulani and settled
cultivators are not infrequent towards the end of
the cropping season, as Fulani cattle encroach on
cultivated areas in their constant search for
nourishment. On the other hand, the farmers are
to some extent dependent on the Fulani for ma
nure and they also make use of traditional Fulani
husbandry skills by having them herd their own
animals. Fulani in the drier areas of the Gourma
have less contact with cultivators and have wider
seasonal orbits than their contemporaries in the
areas where rainfall is more regular.
Some Tuareg use the Niger river as a perma
nent focal point and cultivate either rainfed or
'falling-flood' crops to supplement income from
livestock. Most Tuareg, however, are more typi
cally nomadic with no permanent base and with
almost all their income deriving from livestock.
This would include cash income from transport
of, for example, salt by camels, or of grain from
cultivators to the market by camels or by don
keys. Sheep are milked by this group (as they are
to some extent by the Moors), the milk being con
served in times of plenty as a form of hard curd.
Sheep are more important than goats to the
Tuareg, the ratio being 1.4:1 and flock size about
50 animals.
Some idea of the complexity of the move
ment patterns as practised by different ethnic
groups in central Mali can be obtained from
Table 11.
25
Figure 18. Pastoral cattlefeeding on stubble in an agro-sylvo-pastoral system in central Mali.
 
GENERAL MANAGEMENT
Most livestock owners in developed countries
would not consider the practices pursued by cen
tral Malian traditional owners to be 'manage
ment'. Nonetheless, the sum of such practices
must be considered to constitute management.
Breeding in cattle or goats is not controlled,
and the seasonality that occurs (Sections 7 and 11)
is a result of the natural environmental conditions
acting through the feed supply. Breeding control
is occasionally practised for sheep, especially by
the more mobile ethnic groups: the method used
is the kunan, a cord stretched from the scrotum to
the prepuce, which deflects the penis at erection
thus preventing intromission.
Management of young stock is practised:
young animals are not allowed out to graze until
they are 3 to 4 months of age, and they are pre
vented from suckling when milk from their dams
is required for humans (Figure 20). During this
period, calves are restrained in groups by a simple
individual rope halter attached to a long general
line, also of rope. Lambs and kids are treated
similarly or kept in the owners' houses. There is
evidence (Diallo and Wagenaar, 1983) that some
owners are aware of the milking capabilities of
their cows and milk most heavily those cows with
higher yields. Calves from such cows still grow at
a faster rate than the offspring of lower-yielding
cows.
The control and manipulation of the demog
raphy of the herds (Section 6) also indicates a
level of management, the population structure
being adjusted so that the herds or flocks are most
productive under the prevailing circumstances.
The seasonal movements already described in this
section also constitute a carefully planned man
agement strategy.
An example of the annual management of
cattle in a sedentary village which takes in trans-
humant cattle during the dry season, is given in
Figure 21 . In such a village , it is very probable that
the village-owned cattle will be herded by a pro
fessional herdsman of a different ethnic group
who receives milk and possibly part of the calf
crop in lieu of all or a portion of cash wages. He
may also be allowed to keep one or two of his own
animals in the herd.
In the agropastoral system, management at
the micro-level is almost invariably comprised of
herding by day (either in individually owned
herds and flocks or in herding groups which are
26
composites of several owners' animals herded on a
rotational basis or by a professional herder) and
close penning or individual attachment at night.
In the dry season, and particularly in the upland
millet areas, it is not unusual for goats (and occa
sionally sheep and cattle) to be unguarded during
the day. The practice of not herding animals often
results in considerable numbers of 'lost' animals.
The numbers of lost animals vary between flocks.
This is just one of the factors contributing to the
great variations in overall flock productivity
which will be discussed further in the relevant sec
tions on cattle and small ruminants.
Figure 19. Landscape units ofthe Niger inundation zone and seasonal distribution ofMacina sheep,
a) Landscape units b) Seasonal distribution
I5°N —
I4»N —
 
i ) October
—I5°N
I4°N
 
5°W 4°W
LEGEND
a)
SR Sahel rangeland
TZ Transition zone
I P Inundated plains
EP Elevated plains
ii ) February
 
b) Density (head/km?)
V7A <5.0
tyMfy 5.0-14.9
j 15.0-30.0
■ ■ >30.0
ili) March
 
27
Table 11. Seasonal distribution oflivestock according to ownership groups.
Wet-season
Dry-season grazing areas
grazing Deep wells in
'Sahel' and millet
Irrigated rice
areas
Delta area
areas 'old' burgu 'new' burgu
Northern pastoral areas
('sahel')
Maure Fulani Tuareg dependants Fulani
with traditional
rights in the delta
without traditional
rights in the delta
Rice subsystem Maure dependents Rice farmers Merchants using
Fulani herders
Fulani dependants
Millet subsystem Millet farmers Merchants Civil servants Fulani livestock traders
In an agricultural village, the management of
work oxen differs from that of the general herd.
In the dry season, they are herded with the bulk of
the village cattle on fields close to the homestead.
Late in the dry season or very early in the wet sea
son, they are withdrawn from the main herd
which goes on its annual short transhumance. The
oxen are then kept in the house compound at
night while being allowed to continue grazing
freely throughout the day, when not required for
ploughing or transport purposes. Supplementary
feeding is not very common (although its inci
dence is perhaps increasing) but some household
salt, or salt of other type, may be provided during
this season.
MANAGEMENT CASE STUDIES
Mouton de case in the Niono area
Almost all West African animal husbandry is of
an extensive nature. Occasionally a few animals,
particularly sheep, are kept under more intensive
conditions. This smallholder fattening is gener
ally known as the mouton de case system and can
be defined as one in which sheep are tethered or
confined near the house of the owner and receive
a supplement of good-quality roughages and con
centrates (Figure 22). The system uses male sheep
in the main, these being selected out of the flock
or specially bought for fattening over a relatively
short period. It is most common where a high
Figure 20. A netpouch ofbaobab string used to restrain calfsuckling.
 
28
Figure 21. Annual pattern ofcattle management in a sedentary village in central Mali.
ERD MANAC .
 
proportion of the population is devout Moslem,
as one of the principal reasons for this fattening
system is to have a sheep for slaughter (or for sale)
at the annual religious festival known as Tabaski.
Although this system is supposedly common,
it is often difficult to distinguish between it and
more extensive ones (Coulomb et al, 1980). In
Chad, Dumas (1978) felt that several problems
needed to be overcome before a definition of
mouton de case could be finalised. In an attempt
to define more precisely this special case, a spe
cific study was undertaken. Management data
were collected from five rice and six millet villages
with totals respectively of 123 and 71 confined
sheep.
Of the 159 and 148 households visited in the
rice and millet subsystems, 39% and 24.3% re
spectively kept moutons de case. The differences
between the subsystems just failed to be signifi
cant at the 5% level (X^ = 3.43, d.f. = 1). The
mean number kept per owning household was al
most the same being 1.98 in the rice subsystem
and 2. 19 in the millet subsystem , with a considera
bly greater percentage of millet households own
ing two sheep (Figure 23).
Not all animals being fattened are confined
for whole periods of 24 hours, this factor in part
leading to the difficulty of defining a mouton de
case. Of the 194 sheep in the 11 villages visited a
total of 54 (27.8%) were not continuously tied up.
There were no significant differences (P>0.05)
between the rice (28.5%) and millet (26.8%) sub
systems in this practice.
The relationships between the keeping of
moutons de case and three other classes of stock
(work oxen, all other cattle and total sheep) were
also examined. Table 12 shows the regression
analyses calculated on the data obtained on own
29
Figure 22. A mouton de case ram in a stall in central Mali.
 
ership of the other three classes of stock. Signifi
cant positive correlations were found between the
keeping of moutons de case and all cattle in the
rice subsystem. There were no significant correla
tions between the keeping of moutons de case and
all cattle in the millet subsystem and total sheep
kept in either subsystem.
Not all confined sheep are kept with slaugh
ter or sale in view. A proportion is kept for breed
ing, this being equivalent to 15.4% in the rice vil
lages and 25.7% in the millet villages. The differ
ence just failed to be significant between systems
Figure 23. Ownership of moutons de case in the rice
and millet subsystems.
a) Rice subsystem
% of owners
50"
40-
30
20
10
b) Millet subsystem
23 4>4 I2 34>4
No. of moutons de case
at the 5% level (X2 = 3.62, d.f. = 1). Only two
of the 39 sheep said to be kept for breeding, one
in each subsystem, were males. No females were
kept for slaughter or sale purposes. The break
down by sex and by age of sheep kept confined for
whatever period is shown in Tables 13 and 14.
There were no significant differences (P>0.05)
between subsystems for either sex or age distribu
tion.
There were no significant differences
(X2 = 3.48, d.f. = 1, P>0.05) between the sub
systems in sheep being fattened for sale - 56.7%
in the rice villages and 78.4% in the millet villages
- but there was a significant difference
(X2 = 7.14, d.f. = 1, P<0.05) in those being
kept for home slaughter in the rice subsystem
(40.3%) and in the millet subsystem (19.6%).
The fate of the remaining 2.9% and 2.0% in the
respective subsystems had not been decided at the
time of the enquiry. A slight majority of all ani
mals (59.6% and 41.2% respectively in the two
subsystems) was destined for slaughter at the cur
rent year's Tabaski ceremony.
In the rice subsystem, 76.4% of confined
sheep were bought compared to 38% in the millet
subsystem, this difference being significant at the
5% level (X2 = 6.44, d.f. = 1). Birth in the
owner's flock was the reason for entry for 23.6%
of sheep in the rice subsystem and 50.7% in the
30
Table 12. Regression analyses and correlations between the ownership ofmoutons de case and other classes oflivestock.
System
Class of Calculated Statistical value
stock regression
a b d.f. r P
Work oxen 3.6 25.6 4 0.943 * *
All cattle 1.8 31.5 : 0.979 *
All sheep 0.3 43.4 3 0.155 n.s.
Work oxen 5.7 20.4 4 0.926 *»
All cattle 2.5 35.0 2 0.890 n.s.
All sheep 1.1 39.0 3 0.850 n.s.
Rice
Millet
Notes: a = slope, b = intercept.
**P<0.01; 'P<0.05;n.s. = not significant.
Table 13. Percentage distribution of all confined sheep
by sex.
Sex
System
Rice Millet
Male, entire
Male, castrate
Female
79.9
4.9
15.4
72.9
1.4
25.7
Table 14. Percentage distribution of confined male
sheep by age group.
Physiological
age
System
Rice Millet
Milk teeth
Permanent incisors
1 pair
2 pairs
3 pairs
4 pairs
Unknown
50.0
21.2
4.8
0.1
0.0
24.0
49.0
23.5
7.8
3.9
15.7
15.7
millet subsystem, the difference again being sig
nificant at the 5% level (X2 = 9.88, d.f. = 1). In
the millet subsystem, owners were unable to give
the precise entry reason for 11.3% of animals.
The number of months during which male
sheep (excluding animals kept for under 1 month)
had been confined is summarised in Figure 24.
The most common period was between 4 and 6
months in both subsystems (34.6 and 35.3% of all
sheep). However, in the rice subsystem, signifi
cantly (X2 = 5.97, d.f. = 1,P<0.05) more sheep
(24% of all animals) were confined for a period of
7 to 9 months than in the millet subsystem (9.8%).
The reverse was the case for animals confined for
more than 12 months, with significantly less sheep
(X2 = 12.85, d.f. = 1, P<0.001) being confined
in the rice subsystem (8.7%) than in the millet
subsystem (39.4%).
A wide variety of green fodder roughages
and crop byproducts was fed to the sheep. Sea
sonal availability was obviously a major factor in
determining what feedstuffs were fed at any given
period. It was, however, the system which had the
greatest effect on the type of feed fed.
Table 15 summarises for each subsystem
questionnaire responses on the agricultural
byproducts and salt offered by 62 owners with
123 sheep in the rice subsystem and 36 owners
with 79 sheep in the millet subsystem. Rice prod
ucts were fed by a significantly greater number of
owners in the rice subsystem than in the millet
subsystem. The number of rice growers feeding
cotton seed was also significantly greater than
that of millet growers feeding it. More rice owners
fed sweet potato haulm than did millet owners, al
though this difference was not significant. Millet,
Figure 24. Length of confinement for moutons de case
in the rice and millet subsystems.
a) Rice subsystem
% of moutons
de case
b) Millet subsystem
50-
40-
30-
20-
10-
- 3 4-6 7-9 9-1 2 > 12 1-3 4-6 7-9 9-12
Confinement period (months)
»2
31
 Table 15. Percentage ofowners offering different types offeed to moutons de case in the rice and millet subsystems and
percentage ofowners offering who buy the products.
Type of feed
Percentage of
owners offering
Rice Millet
subsystem subsystem
(n = 62) (n = 36)
Percentage of
those offering who buy
Rice Millet
subsystem subsystem
(n = 62) (n = 36)
Rice (Oryzasativa)
Straw
Grain
Bran
Millet (Pennisetum typhoides)
Grain
Bran
Cotton (Gossypium hirsutum) seeds
Cowpea (Vigna sinensis) haulm
Groundnut (Arachis hypogaea) haulm
Sweet potato (Ipomoea batatas) haulm
Salt
27.4 0.0 0.0 0.0
27.4 0.0 5.9 0.0
88.7 17.7 40.0 100.0
24.2 77.8 53.3 10.7
54.8 99.4 44.1 5.9
16.1 5.6 100.0 100.0
45.2 75.0 3.6 0.0
17.7 75.0 0.0 0.0
52.6 36.1 0.0 0.0
67.7 83.3 100.0 100.0
cowpea and groundnut products were fed by a sig
nificantly greater number of millet- than rice-
growing owners. The percentage of owners offer
ing salt did not differ significantly between
subsystems.
With the general exception of salt and cotton
seed, most other products were home produced
(Table 15). Exceptions to this general statement
are that all the millet farmers who fed rice bran
had to buy it and a proportion of rice farmers who
fed millet products also needed to buy those.
These figures need to be viewed with some cir
cumspection as feedstuffs were not bought en
tirely for feeding to moutons de case, and their use
for fattening animals was often to the detriment of
the work oxen for which much of them were
primarily destined.
Feedstuffs of non-agricultural origin were
also provided in both subsystems although in a
much wider variety in the rice than in the millet
subsystem. In the rice subsystem, in addition to
grass, the leaves of a locally grown Ipomoea sp.
were most commonly fed, followed by those of
Khaya senegalensis (an introduced shade and
timber tree whose leaves are not selected on the
tree by free-ranging animals) and Pterocarpus lu-
cens. Leaves from a further nine plant species
were identified in the rice subsystem and other
species were undoubtedly fed. In the millet sub
system, P. lucens was the unique non-agricultural
feed offered. All feedstuffs in this category were
cut and carted by family labour although there is
an active commercial trade in most of these items
in the Niono urban centre.
Management of Macina sheep
in the inundation zone
Macina sheep are managed under a transhumant
system, its base being in the 'inundated' (i.e. the
lower-level) plains of the inundation zone. The
hydraulic regime of this zone is such that the area
is under the flood waters of the Niger from late
July or August through to February or March, al
though the southern parts of the zone begin to dry
out in November or December following the re
treat of flood water northwards. The majority of
animals must thus leave the zone on annual mi
gration in early July. The migration route for
most of the sheep is to the north and west, where
there is typical Sane li an pasture. The emigration
from the inundation zone coincides with the short
Sahelian rainy season so the sheep are, at least to
some extent, still in conditions (green pastures
and a humid climate) which are favourable to
them. The extent of the seasonal movement and
its intensity can be gauged from Figure 19 which
shows the densities and distribution of small ru
minants at the end of the rainy season, when the
central zone is still inundated, and in February
and March, when the floods have receded. While
distances are relatively short, there is an almost
total evacuation of animals from the 'inundated'
plains to the 'transition' zone in the northwest
and, to a lesser extent, to the Sahel' pastures
32
farther northwest and to the east. This movement
is paralleled by that of cattle although these ani
mals move over much longer distances.
During the flood season some 80% or sheep
are found outside both the 'inundated' and 'ele
vated' plains of the central zone, leaving some
50 000 head only in this area at a density of about
three animals per km2. These sheep are mostly on
the small eminences within the floodplains on
which permanent village bases are established.
After the flood has fallen the situation is almost
exactly reversed, with 70% of all small ruminants
found on the plains at a density of 8 per km on the
elevated and 20 per km2 on the inundated plains.
The Fulani shepherds recognise four sea
sons: indiewde, which is roughly the post-rains pe
riod of October and November; dabunde, the
cold dry period from December to February;
cedu, the hot dry season of March to June; and
ndungu, the rainy season of July, August and Sep
tember. Major movements are governed by the
seasonal calendar which in itself is subordinate to
the flood regime. The wet-season transhumance
cycle is dictated to some extent by the needs of the
sheep for a mineral supplement, which is usually
provided by salt earths although salt may occa
sionally be bought.
The main body of sheep move back into the
inundated plains at the start of the cedu, but the
factors which govern the actual timing include the
extent to which the pastures have dried out, the
occupation of the area by cattle-owning groups
who have precedence, and the whim of the 'chief
shepherd' who is the elected representative of a
village or area and who is usually the largest
owner. For obvious reasons it was not possible to
obtain much information on actual ownership
patterns. Management units (flocks) in the trans-
humant system are usually large but some stratifi
cation is practised. The inundation zone is a series
of water meadows topographically flat, with good
visibility and a dense herbage cover. There are
thus advantages in maintaining large grazing
groups under the responsibility of one or two her
ders. During the period spent in the inundation
zone itself these groups are seldom composed of
less than 300 head and more than 500 head is not
uncommon. In an intensive ground survey, 69%
of all groups in the flood zone were greater than
100 head in size and a further 23% were between
50 and 100 head. The analysis of aerial survey re
sults showed a similar situation with some 75% of
groups having more than 100 head of sheep in the
'Sahel' and 'transition' land units during the rains.
That these were mainly Macina groups was easily
confirmed by the homogeneity of colouring as op
posed to the varied colours of mixed flocks of hair
sheep and goats.
Because of the need to ensure a constant sup
ply of milk and meat for the family members re
maining in the inundation zone during the flood
season, a sophisticated system of stratification has
developed. This is undoubtedly based in part on a
similar system used by cattle-owning Fulani, al
though the terms used and the type of groupings
are not generally the same. The principal charac
teristics of this stratification are shown in Table 16.
In 85% of cases weaning takes place at about
4 months, usually by removing the lambs to
another management group. Where this is not
possible the teats are blocked with dung or tied
with a cord. Castration is carried out either
shortly after weaning (at 6 or 7 months in 30% of
cases) or at about 1 year (48%). The common
practice is attrition of the cord by beating with two
sticks although the open method is occasionally
used. There is little or no attempt to control
breeding season, but most owners (74%) thought
that the main tupping period was during the rains,
giving rise to births in the latter part of the cold
season or in the early hot season.
Shearing, if such it can be called, is done with
a double-bladed knife. In general this activity can
be considered the poorest of the management
practices of the traditional husbandman. Indi
vidual sheep are shorn as many as four times in a
year. It is perhaps worth noting that the French
colonial power enacted legislation regulating the
number (two) and the timing (before and after the
rains) of shearings: at independence this law
passed into the Malian code but, while still on the
Statute Book, its existence is ignored by the own
ers and the responsible livestock authorities have
neither the personnel nor the logistic support to
enforce it. There is generally no specific period
assigned to shearing nor are all sheep in one flock
shorn together. There is thus a constant source of
new wool for spinning and weaving by the local
village industry, with supply being perhaps
geared to periods when labour would otherwise
not be utilised.
33
 Table 16. Stratification ofMarinaflocks with demographic characteristics and management objectives ofeach group.
Flock
name
Group
size
Use
Composition
General Males
(%)
Females
on X
(%)
Beydi Generally
small
Nurse
flock 1
Newly lambed
females, ewes
in advanced
pregnancy, weak
and aged animals
26 74 Kept in the village; herded by infants.
Regular movements of animals
into and out of this group.
Tarancaradji Medium Sale/
slaughter
Largely male,
generally young,
with some older
females
GO 40 Kept in the village, generally not
herded.
Njamiri Small Slaughter Overwhelmingly
male
95 5 Individually tied and zero-grazed ;
women's responsibility.
Bucal Medium Milk Predominantly
female
25 75 Individual ownership but commonly
grazed on reserved pastures by paid
Bendi
Horey Large Wool/ Predominantly
meat female
24
herder or by family labour in rotation .
Household milk supply. Most of
village goats are in this group.
Similar to bucal. Term used
mainly by owners of hair sheep.
76 Main flocks which transhume .
Reserve for constituting
other groups as required.
Milked by herders as required.
34
6. HERD AND FLOCK DEMOGRAPHY
With the exception of a relatively small number of
animals owned purely as an investment, cattle are
kept principally to provide milk for subsistence or
draught power for agriculture (Figure 25). Minor
functions of cattle are as drawers of carts, for lift
ing water from deep wells (especially in the north
of the area), and to provide manure for fertilizing
fields. Some transhumant groups use oxen as bur
den animals for the transport of people, personal
effects and camp gear.
Small ruminants are kept primarily as a
source of meat and as a ready source of cash.
Marina sheep, however, also provide wool and
milk (Section 5, Table 16), Maure sheep provide
hair, and a large proportion of goats is milked at
least some time in the course of a life cycle.
CATTLE POPULATION STRUCTURE
The age and particularly the sex composition of
herds is regulated to a great extent by the main
herd functions. Table 17 shows some typical herd
structures in the different production subsystems
of central Mali. Although the functions of these
particular herds are known, in this case it would
be possible to deduce the functions from the de
mographic structure alone. The high percentage
of mature castrates in the two main cultivating
groups is a result of the requirement for draught
Figure 25 . Work oxen ploughing a field in central Mali after the early rains.
 
35
 Table 17. Cattle herd structures (%) in central Mali.
Class of
stock
Ethnic group, type of management and use
Bambara, Mixed, Fulani', Fulani", Rimaibe, Tuareg",
nomadic,
milk and
transport
sedentary, sedentary, transhumant, transhumant, sedentary,
milk and draught and milk and milk milk and
draught milk transport draught
Males
<1 year 5.6 7.2 9.9 9 15 16.3
1-3 years 8.5 9.3 12.5 6 13 9.3
>3 years 32.7 46.3 15.6 6 8 12.2
Total 46.8 62.8 38.0 21 36 37.8
Females
<1 year 6.2 5.7 10.2 10 15 9.8
1-3 years 11.3 11.1 15.2 24 12 14.7
>3 years 35.7 20.2 36.1 45 38 37.6
Total 53.2 37.0 62.0 79 65 62.1
Castrates as %
oftotal herd 32.3 43.4 9.3 4 7 12.9
' From Sangare (1983).
oxen. These herds are not, however, able to
maintain themselves by recruitment from breed
ing, and draught animals have to be purchased
from other ethnic groups.
In herds where milk is the major product (the
detailed structure of such a herd is given in Figure
26), the proportion of breeding females is in the
region of 40% of all animals: about two-thirds of
these females, or a total of 25 to 27% of the herd,
can be expected to be lactating at any one time.
Even in the urban herds, which are generally con
sidered to have milk as the primary output, these
figures hold true (Table 18). It needs to be re
membered that, where milk is a product, the milk
taken for human consumption is at the expense of
the calf, as no concentrates or substitutes are fed
other than some coarse roughages. This practice
has repercussions on calf growth and survival.
Oxen are usually trained to the plough and
start work at about 4 years of age. It is possible
that in both the millet and rice subsystems oxen
are being kept to an age at which they are rela
tively inefficient producers of power. If this is the
case (and Figure 27 provides some evidence that it
is), it results from the extremely slow growth rate
and late maturity of this class of stock, the difficul
ties in replacing them, particularly in recent
years, and (even when supplied on credit) the
amortisation cost. There will be a continuing de
mand for work oxen. The best means of satisfying
this demand is probably to improve early nutri
tion to ensure entry into the work force at a
younger age. This will result in either prolonging
their working life or permitting their retirement at
a younger age, giving better carcass quality.
Whether limited feed resources should be de
voted to this target or whether they should be
used for breeding cows and currently working ani
mals needs further study.
Fulani transhumant herds provide a reserve
of mature male animals for transfer as draught
animals to the neighbouring agropastoral sys
tems. The herds of the latter, in particular those
of the rice subsystem, have too few breeding cows
whose fertility levels are too low to ensure herd
stability. As a result mature draught animals must
be purchased to maintain herd numbers and
structure. Our assessment of the numbers of work
oxen in the irrigated areas is in general accord
with the returns of the Office du Niger although in
some villages up to 70% of the cattle are recorded
as work oxen and figures of 55% are not uncom
mon. Such an unbalanced herd structure has obvi
ous repercussions on the stocking rate, and it is
necessary for each head of cattle in the rice sub
system to be counted as 0.94 TLU and for each in
the millet subsystem as 0.90 TLU. This compares
with a 'normal' value for cattle herds of0.73 TLU.
SMALL RUMINANT POPULATION
STRUCTURE
As for cattle, the age and sex composition of small
ruminant flocks is regulated by the main functions
of the flocks, but to a considerably lesser degree.
36
Figure 26. Population structure ofa Fulani herd with milk as a primaryfunction.
Males
Age
(years)
12 +
I I
10
9-10
8-9
7-8
6-7
5-6
4-5
3-4
2-3
1-2
Females
i
12 8
Ti
4
0-1
0 0
% of total herd
4
Table 18. Age and sex structure ofcattle populations in three urban areas in central Mali.
Nionoa
(n = 588)
S6goub
n = 901)
Mopti
(n
and Sevarec
Age
(years)
( = 743)
M(%) F(%) M(%) F(%) M(%) F(%)
<1 12.76 14.46 12.65 14.54 18.71 23.15
1-2 4.59 3.06 6.10 7.55 2.29 7.81
2-3 4.42 4.42 3.00 5.11 0.94 1.75
3-4 1.02 5.27 1.44 3.33 0.27 2.83
4-5 4.42 21.77 1.66 2.66 0.40 12.65
>5 4.59 19.22 7.77 34.18 2.02 27.19
Total 31.80 68.20 32.62 67.37 24.63 75.38
Sources: * Fayinke(1980);b Maiga (1980); c Ballo(1980).
Examples of flock structures in a number of
central Malian societies are given in Table 19 for
goats and in Table 20 for sheep. Male sheep and
goats (whether entire or castrated) in general con
tribute about 25% of total numbers. Offtake of
males (in addition , perhaps, to a higher death rate
at very young ages) commences early, before
6 months of age in the better-managed flocks. In
strict husbandry terms, males capable of breeding
are in excess of the 2.5% actually required, but
the possibilities of loss and of temporary sterility
due to nutritional or disease stresses have to be
considered. Moor sheep flocks have a ratio of
breeding males to females of 1:13 while that of
Fulani flocks is 1:15. For both these ethnic groups
sheep provide wool as well as meat, and mature
37
 Figure 27. Distribution of oxen pairs by current
working life in a central Malian village.
No. of oxen pairs
28 1
24
20-
16-
12-
2 4 6 8
Period of use (years)
10
entire males therefore contribute to flock econ
omics in addition to fulfilling a breeding role.
With the exception of the Tuareg nomadic flocks
only in Fulani Macina sheep flocks do castrates
contribute significantly to flock numbers - a
further indication of the importance attached to
wool production by the Fulani.
No comparative data on pre- and post-
drought flock structures are available. This is un
fortunate but it is probable that with the rapid rate
of flock turnover and high levels of natality any ef
fects would be of an ephemeral nature. With the
exception of the structure of Macina flocks (Table
16) there was little evidence in central Mali of any
system of flock stratification other than the short-
term separation of kids and lambs from their
dams at very young ages.
38
Table0.Flockstructuresofg atswnedbysomethnicgroupsice tralMa .
Breeding8
0.1 00.0 0.0 00.0 00.i 00.0
<4
mont0s
Females(%)
03.0 4.0 00.0
0i.0
08.0 00.0
mont0s
>4 0.0 0.0 00.0 00.0 0i.0
00.0
Total
60.0 01.0 60.0
g.0
60.0 1.0
Castrates
0.0 0.0 0.0 0.0 0.0 1.0
<4
mont0s
0.0 0.0 4.0 0.0 00.0 0.0
i(%)
Male;
mont0s
>4 0.0 0.0 0.0 0.0 0.0 0.0
Total
00.0 0.0 00.0 0.0 0.0 0.0
Totalno.
ofanimals
000 00 00 00 000 00
Management
system
Pastoral/nomadic
Pastoral/trans0umant Pastoral/trans0umant
Agropastoral/trans0umant
Pastoral/trans0umant
Agropastoral/sedentary
Et0nic group
Tuaregb Tuareg Fulani" Fulanic Moor
Bambara
*Femalesoldert0 n00months.
bPeacock(0080).
cFulaniagropastoralistsit0eNigernundationz "owna sMarinashe p.
Table0.Flockstructuresofshe pwnedbysomethnicgrou sice tralMa .
Et0nic group
Management
system
Totalno.
ofanimals
Males(%)
Females(%)
Total
>00<
mont0s
Castrates
Total
>i0<
mont0s
Breeding*
Tuareg Tuareg Fulanib Fulanic
Moor
Bambara
Pastoral/nomadic
Pastoral/trans0umant Pastoral/trans0umant
Agropastoral/trans0umant
Pastoral/trans0umant
Agropastoral/sedentary
00
660 000 041 0000 00
0.0 4.0 0.0 0.0 00.0 01.0
0.1 0.0 0.0 00.0 0.0 0.0
4.0 0.0 00.1 00.1
i0.0
0i.0
4.0 0.0 0.0 0.0 0.0 0.0
0i.0
0.0 03.0 0.0 00.0 10.0
00.0 0.0 0.0 0.0
0i.0
0.0
00.0 00.0 0.1 00.0 1.0 1.0
01.0 10.0 00.0 01.1 0.0 00.0
"Femalesolderthan00months.
bPeacock(0000).
'FulaniagropastoralistsitheNigernund tionzo"ownMaris0e p.

PART TWO
CATTLE PRODUCTIVITY

7. REPRODUCTIVE PERFORMANCE
AGE AT FIRST CALVING
A total of 20 heifers whose date of birth was
known had themselves given birth by the end of
the study. The mean age at first calving of these
animals was 1505 ± 103.3 days (49.5 ± 3.34
months).
CALVING INTERVAL
The observed calving interval (x ± s.d.) was
665 ± 202.2 days (n = 244) or just under
22 months, the coefficient of variation being
30.3% . As can be seen from Figure 28, there is a
distinct bimodal pattern to the distribution of
these intervals with peaks at 12 to 16 months and
at 23 to 26 months, and some indications of a third
minor peak at around 36 months.
The analysis of variance of this trait (Table 21)
demonstrates that parity (i.e. dam's age) signific
antly affected the length of the next parturition in
terval. The least-squares means for the main vari
ables affecting calving interval are given in Table
22. The observed sample mean (665 days) differs
somewhat from the computed overall least-
squares mean (644 days) as the latter is adjusted
for the unequal subclass numbers. The general
Figure 28. Frequency distribution ofcalving intervals in cattle.
Percentage of
all intervals
10-
9
8-
7-
6-
5-
4-
3-
2-
n = 244
12 21 24 27 30
Calving interval (months)
33 36 39
43
 trend appeared to be one ofreducing interval with
increasing parity until older ages were reached
when the interval between carvings increased
sharply. There was a clearly longer interval after
the birth of a male calf than after that of a female.
Month of birth had a considerable but non-signifi
cant effect on the subsequent parturition interval,
which ranged from a minimum of 565 days after a
January calving to a maximum of 723 days after a
post-rains calving in October and November.
There were also considerable though non-signifi
cant differences between herds within each of the
two subsystems.
Table 21. Analysis of variance of calving interval for
agropastoral cows in central Mali.
Source of
variation
d.f. MSxKT
System
Month
Parity
Sex
Herd (millet)
Herd (rice)
Remainder
1 109
11 21
4 142"
1 46
3 40
4 21
220 40
••P<0.01.
Table 22. Least-squares means for calving interval of
agropastoral cows in central Mali.
Variable n x (days) ± SE
Overall 244 644 18
System
Millet 104 617 26
Rice 140 671 21
Parity
1 772 646a 29
2 47 602a 33
3 32 654ab 39
4 39 585a 34
,9- 54 734b 32
Sex
Female 136 630 21
Male 108 659 24
Note: '9' are births of unknown parity assumed to be £3.
Within variable groups, means followed by different letters dif
fer significantly (P<0.05). Variable groups without any letters
did not show a significant difference in the analysis of variance.
When the age at death of a calf was intro
duced as a covariant, the subsequent parturition
interval was lengthened significantly (t = 2.68,
P = 0.011) for each day of survival of the calf, the
relationship being:
Parturition interval = 499.5 + (age at calf
death x 0.318) days
The repeatability (± SE) of the 244 intervals
for a total of 166 cows was 0.182 ± 0.0138.
SEASON OF CALVING
The monthly distribution of 452 calvings (relating
to all animals born in completed years of the
study) is shown in Figure 29 which also shows the
season of conception, assuming this occurred ap
proximately 9 months previously. Some 56% of
all calves were conceived during the 3-month
period of the rains.
When correlation coefficients were calcu
lated for the number of births per month and rain
fall 9 and 10 months previously, there were highly
significant correlations (respectively r = 0.562
and 0.563, d.f. = 70, P<0.001) for the pooled
data over 6 years. Correlations were significant
for rainfall both 9 and 10 months earlier for indi
vidual years. Correlations between rainfall 11 and
12 months previously and the number of calves
born were not significant. Details of these coeffi
cients and significance levels are given in the
Appendix.
NUMBER OF CALVES BORN
PER BREEDING FEMALE IN THE HERD
Reconstruction of cow histories (by observation
and by questioning of owners for cow calving data
before the study began) indicated that the 274 cows
in the study which had calved at least once had
given birth to a total of 797 calves. Cows having
given birth to one to eight calves respectively ac
counted for 23.0, 21.2, 23.7, 16.1, 10.9, 1.8, 1.8
and 1.5% of all cows, the average number of
calves born per cow being 2.91. Numbers of
calves born to those cows still in the herds (i.e.
whose reproductive careers were still in progress)
at 30 June 1984 are shown in Figure 30: the aver
age number born to each of these animals was
3.02.
There is a slight anomaly in these calcula
tions due to assuming that the first calf of mature
cows of unknown history is the third for that ani
mal. However, changing this to four would hardly
affect the general conclusion that very few cows
bear more than five calves. The average lifetime
production is of the order of three calves per cow.
Most cows therefore either died or were culled
from the herd at about 9 years of age (this being
age at first parturition + interval to second calf +
interval to third calf + time of weaning third calf) .
44
Figure 29. Seasonal distribution ofconceptions and calvingsfor cattle in central Mali.
Percentage of all births
20-
15-
10-
n=452
Month of birth
Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov
Cold dry Hot dry Rains Post- rains
Month of conception
Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb
Hot dry Rains Post- rains Cold dry
DISCUSSION
It has been suggested that indigenous cattle in the
tropics normally drop their first calf at 3 to 4 years
of age (Mahadevan, 1966). Under improved man
agement where seasonal nutritional stress can be
reduced, it is indeed possible to achieve first calv
ing at 43 months. This has been the case at Niono
on the government research station (ILCA/IER,
1978) under exactly the same climatic conditions
as pertained in this study. Nonetheless, even
under these improved conditions, in series of bad
rainfall years or when supplementary feed
supplies could not be provided, age at first calving
was delayed. Traditional owners in the dry areas
of Africa have , for many years , complained of de
layed first calvings due to "lack of fodder re
sources" - although they do not always associate
this lack with overstocking in general. There has
been to date little empirical evidence for this as
sertion, but the firm data on age at first calving
from 20 heifers in this study support the view that
first parturitions do not generally occur until the
beginning of the fifth year of a cow's life.
The calving interval of 665 days is equivalent
to an approximate calving rate of 55% (365 x 100/
665). In Sudan, under similar climatic conditions,
a rate of 40% was calculated for sedentary cattle,
this improving to 65% for migratory cattle where
nutritional stresses could be expected to be miti
gated to some extent (Wilson and Clarke, 1976a).
Based on questionnaire surveys rather than on
observation, estimated rates for other sedentary
cattle in Mali are 55% (Coulomb, 1970). Rates of
60% (Coulomb, 1971) and 63% (Wilson and
Wagenaar, 1983) for transhumant cattle have
been calculated from questionnaire data collected
in Niger. Where nutritional stresses are reduced
by research station management, calving inter
vals can be progressively reduced - at Niono, for
example, from 561 to 423 days from 1966 to 1973
(ILCA/IER, 1978).
The effects of year were not obtained during
this study and month of calving had no significant
effect, possibly because of the generally long calv
ing interval. However, the general effect of
month of calving was that which would be ex
pected if nutrition were the major factor in repro
ductive rate. Thus a calving in the cold dry season
allowed conception during the next rainy season,
leading to shorter intervals than the mean . The ef
fect of parity was also as expected since older,
weaker cows took longer to recover from preg
nancy and lactation stress, this resulting in longer
intervals. The effect of the sex of the calf is inter-
45
Figure 30. Number of calves born per cow in central
Malian herds.
Percentage of cows
25o
20-
n=274
 
"i 1 r
5 7
No. of births
esting, male calves apparently delaying concep
tion by putting more strain on cows. A similar calf
sex effect has been observed in red deer in Scot
land (Clutton-Brock et al, 1982) where calving in
tervals 11 days longer after males were attributed
to longer gestation intervals, heavier calf weights
and more frequent and more intensive suckling by
male calves. The effect of the age at death of the
calf on calving interval can be attributed to the
suppression of lactation anoestrus.
The repeatability of 0.182 is better than the
average of all published data for this trait in tropi
cal cattle, calculated as 0.131 by Payne (1970). It
is as good as, or better than, other African re
cords: 0.05 for Northern Sudan Zebu (Alim,
1964) and 0.09 to 0.18 for the Small East African
Zebu (Mahadevan et al, 1962). However, it is still
a relatively low value and it is more than likely
that the observed variation is due to environmen
tal rather than hereditary additive factors. Seek
ing improvement of this trait through breeding
would therefore appear to have little chance of
success, at least until better control of the external
environment is achieved.
Nutritional effects on the reproductive pat
tern are clearly demonstrated in the seasonality of
calving (Figure 29). The results obtained in this
study confirm other observations made in dry
northern Africa. In Sudan, 62.5% of all calves
were born in the period April to June (Wilson and
Clarke, 1976a). In Niger, owners reported that
40% of calves dropped in May and August
(Coulomb, 1971) and in a recent survey, also in
Niger, 57% of calves were born in July to Sep
tember (Wilson and Wagenaar, 1983). This
phenomenon is not confined to dry northern Af
rica but occurs also in southern Africa, where
breeding restrictions are not practised and where
rainfall is seasonal. In Swaziland, both Brown
(1959) and Butterworth (1983) have reported
marked seasonality in calving. The latter author
found a normal unimodal curve of calf births, with
57% and 55% of the annual total occurring in Oc
tober to December in the High Veld and Middle
Veld respectively. The values of r in Butter-
worth's study, which included 7 years' data, were
0.89 and 0.83 for rainfall 10 and 11 months previ
ously, and 0.74 for 9 months previously. Repro
ductive response in the current study, which prob
ably occurred even before the onset of weight
gain, was thus more rapid than in Swaziland.
Additional evidence that nutrition is the
major constraint to reproduction in the tradi
tional environment comes from the bimodal dis
tribution of calving intervals shown in Figure 28,
with those cows failing to conceive in the first
rainy season not conceiving until the next. This
pattern has also been observed in Niger where
38% of calving intervals were around 11 to 13
months and 24% around 24 to 25 months: there
were no intervals at 19 and 20 months and only
one at 18 months (Wilson and Wagenaar, 1983).
Poor nutrition is in large part responsible for
the low average number of calves born to cows in
the herd. The average of 2.9 calves per cow results
from late first parturitions, long calving intervals
and early exhaustion of the body reserves leading
to an early culling age. Cows are unproductive for
long periods of their life but continue to consume
natural fodder resources, thus further aggravat
ing an already critical feed supply situation. Im
proved fodder production or better utilisation of
existing production would do much to reduce nu
tritional stress, particularly in the dry season, and
thus improve the reproductive rate.
46
8. GROWTH AND WEIGHT
BIRTHWEIGHT
The observed birthweight (x ± s.d.) of 107
calves born during the 1978-1984 period was
16.6 ± 2.67 kg. The analysis of variance of this
trait (Table 23) demonstrates that only parity (i.e.
dam's age) has any significant effect. The least-
squares means for all the variables considered are
shown in Table 24.
GROWTH TO MATURITY
For cattle of known birth date, a generalised
growth curve to 4 years of age is shown in Figure 31 .
Analyses of variance and least-squares means for
weights at various ages are shown in Tables 23 and
24 respectively. The observed sample means used
in Figure 31 differ somewhat from the computed
least-squares means shown in Table 24, as these
last are adjusted for the unequal subclass num
bers. The season of birth had a highly significant
effect on growth throughout the life of the animal
to its maturity (Figure 32). The year of birth also
had consistent significant effects on weight at all
ages (except birth) up to 9 months of age. Among
herds, differences in growth rate were also signifi
cant. Coefficients of variation for weights at vari
ous ages were: 30 days, 20.00% ; 90 days, 20.63% ;
180 days, 19.98%; 210 days, 19.16%; 270 days,
18.49%; 365 days, 18.43%; 550 days, 17.06%;
730 days, 16.43%; and 1095 days, 14.55%.
Phenotypic correlations between calf weights at
different ages are shown in Table 25.
Cows did not reach final mature weights until
after 5 years of age. Mature weights were 230.4 kg
in the rainfed millet subsystem and 224.5 kg in the
rice subsystem. Males - exemplified in this study
by work oxen in the irrigated subsystem - had
average mature weight of 297.2 kg at ages above
6 years, although the largest individuals weighed
up to 430 kg.
BREEDING FEMALE
POSTPARTUM WEIGHT
The observed postpartum weight (x ± s.d.) was
214.7 ± 33.63 kg with a coefficient of variation of
13.04%.
The mean squares from the analysis of var
iance are set out in Table 26 which shows that par
ity, season and year exerted highly significant
effects on postpartum weight while differences
among herds were also significant in both rainfed
and irrigated subsystems. The least-squares means
for postpartum weights are given in Table 27 while
graphic representations of the main variables af
fecting this parameter are shown in Figure 33.
The postpartum weight was reduced by
0.016 kg for each day of reduction to the previous
parturition, this effect being non-significant.
EFFECTS OF CLIMATE ON GROWTH
AND MATURE WEIGHT
Growth curves for calves born during the main
calving season in each year from 1978 to 1982 are
shown in Figure 34. The figure clearly shows the
effects of the fluctuating feed supply on the over
all development of weight over medium- to long-
term periods.
The weight of mature oxen fluctuated from
88.9 to 107.2% of their mean weight when
weights were pooled over two-monthly periods
(Figure 35). When weights for single months were
used, the minimum early July weight (240.5 kg)
was found to be only 72.5% of the maximum
weight of 331 .9 kg in November. Similar ranges of
variation were found for all ages of breeding
females, reinforcing the picture shown in Figure
34. Adult females with four pairs of incisors var
ied from 87.7 to 110.2% of their mean weight.
The most perturbing long-term effect is
shown in Figure 36, supported by the data in
47
I
1
,aI
.•
V3 O o *t s© ,ir i^. tN o * O
s S -* 5
os'
!
1 a c d
en
00
T3
" m r-i rN * w rn a
os
•
w-i mUi MD q r- w. r-
2 a S
rN
^ t 1 a s g
-o
;
^ rM Wi *» » oo r^ -* rr,
£ © i 3
u-i
1 s i
*i ~ - CI 1*1 t t*-i T m w -
,n ri
c/n m u-, 00 V in 00 "* sC
2 ci § 1 § s § ? S
r4
oo
HI C""l u-j i^"j
T3
-- -- n m T m * r^ m
1
^ ,* v, o b V So ^r O -t r-
5 a a gg- §
s & 3^u-1
,d
m tn rsi wH T m rr,
S
BB io o-i r^ u-i o 00 m Os
s s ^ Os 3 i oi 2
%i M - e-i m T in -q- m m g,d
«
• •
So
£
V3
Pi
r- t*i C>i rN W1
s 38 8
os" s i g dr- § pic
f,
,o M
r-i
•
m T m (-" S
v-i rr, sO
• t- G^ r^ tn « o
s r-i R d3 *
u-i s S w-1 3
^ - ~ m C*i « tn «t r-j n o
T3
#
r-i
CO 00 *M o — » rst 0] O r^ n-i
r^ rs s 5 P
00 •i Pi si S
M «-j s© c-i ,t ci 0 r-
,b s£i
a CM s s k s 1 i S
r-j ~ tn tn r-^ a sb u-i
,d
■* - m - ,** m T I I 8
1
c
-
0
E
1
c
1 1
1 1
8 6
1 1
CJ
1| K £ I
s
d
V
48
>s
IM
■o
Vi
s
B
s. IM
•3
8
c
e, IM
i
S
vi
B
>, IM■o
m
so
m
a
s. 1*
CO■o
O
rj
°
>% IM«•o
o
iM
B
s. IM
i
e
C
s,
"
■s
8
B
R
IM
•3
S
B
f
IM
a
C
a■:
>
8 P
1 I *
O« -it
s a
& SB
S S
co « Xt
SSSS SR
fi ©
co « n Xi
00 lfi CN H
a Ji h- os
* yo si R
- x. x. -
vi S vi ■*
— pi ro ^
s? s
00 •* — vi
os o © ©
pi ITi Os ©
so ■* pi as
co Xi co eg
a $ s s
8 = 38
£ 3
* 2 C
1 uS 1 I
co Xi co
pi rn m
s V, VI VI a s r * 3i S m ^V V,
,7 CM
CN
S m
pi cm s 5 I"2pi
?, *
s
S S sBf! X © c*J uo. ?
pi
VI Vi W1 ?; # — .C msD "A ^v. Vi
2 2 O
go so
o> u-l 0>
— (n g 5 3 3 =pi
■fi
A u co
— S * s m s St 6 3C So iN O;
5 R S S Pi OO OS
ro «n op" rn *rn •* mm pi
5
os ro s * m ^ pi
PI
3- 00 Tt o- a ? -
a P 3 8 3 s a s
re
2
a a a a a a a a a W » r^pi in <M
B
CN 3 pi S3 St 5
JC Hi g R wPI
X. X.
r- El In - « 3 3C S ?i ?
ro r- r- r-^ r^ r-' ob oo -C r» r-^ oo
s ?. in *f a a 2 * w, S 2 2
j^n
« X X. X X
* « «^ M »
* C*1 00 X ■-«
»' d ■»' h d
m pJ — in r5
Sop oo m' © rj
\o r■ vt * o
Os (N srt ^ Os sO
^ te m * n
5 « Os in so
8 s s s s
- x, -. X. -r X X --»■ 00 P s i 3 oo sQ m Os rni r^ cn r^ r- N
s s p p — r^ — sor- K r- ro OS (*1 OS VI r^ cm so m —S r- r^ r- r^P» N « r^
gt ■ w w m sp (- © i * S c- P nsO Vi ^ — * * w -
1 1 a 5* K so m 5M« -«r-m , (N 2 pi m osi m **• so i> ©- 2 «9 -■ m^ so ^ s© s©3 3 s s sO sO so so 00 OS OS 00so so m woi
o m m so
* - t r- ?2«$5
OQ N — ^ SJSSJN - iS N
« r- rj p s a a = s 8
B
« ra a co Xl
'5
3 a a a m 6 a m m— Z — V, VI
i8 8 £ S 8 8*82
73
s
2 w 8 a Sf SS5 2 1
.5
8
x — r- a-
— oo m —
co Xi X io co
K
r> r^ in oo
ro r- i- oo P 6r- -c po r-
« X A «
s s a 5
eg xi co ce
» * » JVI i> in Os
vi W m q
in * m i?i
» 00 r- so r-
« Xt « ca
5 s 2 *
3 R P S
£2SS S35?as
# i Is assss
s * 8 a a
— pi _- so" pi
in in m in in
sQ O sO 00 —
I rN os i> o so 1
, 3 I S " "S ,
I Os r- sO fi i-" I
rsi sp" o m ej
in sn « m «
i> c> » *-* *
I rs) 3b r^ 00 Fo ,
u Xi
, i I !§ si S ,
00 (N vi 00 vi
* m m * kh
X) X)
CO CO co « X
Os sO (N Os Os ,
1 n 5 ^ * m I
os r^ os vi o
m rn cn (n ^r
I Jn m 00 * 00 1
M JD W « co ■
i> m « os sp in— X v. C cri ot I
Tf « vi sp vj sQ
N m (N N N N
iiS2SS3* 1 Si $ I 5 I I H I
mplr-ipi ^-op«inPi
r-^o— pSpioso
viviood r^oosi>^
« O t oo— — po c*-. 8 5 s a a -a
i
25.86 24.05 26.31 26.69 sESSP
e
1
O
m N h «
r* « i> cn
in •* » in g•* *n m ?s| in i
00 r^ 00 i>
in so m ^ i>
I
M12I alilllii 1 «^-2 I « 2 « £
v
5;
49
Figure 3 1 . Generalised growth curve to 4 years ofagefor cattle ofknown birth date.
Liveweight (kg)
200
 
T
24
Age (months)
T
36 48
Table 28. The poor rainfall (in relation to the
long-term mean) encountered at the beginning of
the study period has continued and even been ag
gravated. As a result, the mean mature weight of
work oxen fell by about 1 .4 kg per month over the
whole period. Similar, although less spectacular,
losses were recorded in mature females.
DISCUSSION
Growth rates were very slow, averaging only 185 g/
day from birth to presumed weaning at 7 months
of age. From this age to 1 year growth slowed
even further to an average of only 121 g/day. This
period corresponds to postweaning and, for the
majority of calves born during the late hot dry
season, would coincide with the worst period of
the year from February or March through to May
or June. In fact, calves born during the hot dry
season had a postweaning gain of only 59 g/day
compared with 174 g/day gained by those few
calves that were born in the post-rains season and
thus benefitted from the following year's rainy
season during their postweaning period.
As can be seen from Figure 32, the significant
effects of season of birth on growth were not only
marked hut followed a logical pattern related to
suckling and then to the availability of fodder
right up to the age of 2 years.
Although between- and among-year effects
were significant up to 9 months of age, there was
no clear relationship with total annual rainfall for
the year in question. It is probable that the pat
tern of rainfall within the season and variation in
management practices between years masked any
direct effects. System did not exert a significant
influence on growth but except at birth, cattle in
the irrigated rice subsystem were consistently
heavier than their contemporaries in the rainfed
subsystem up to adult weights. Males were lighter
than females after weaning, this probably indicat
ing that slightly more care was given to female
calves or that less milk was removed for human
consumption from dams of female calves. From
1 year of age onwards males continued to be
heavier than females, with an advantage of 3.0%
at 1 year, 6.0% at 2 years, 10% at 3 years, 17% at
4 years and 32% at mature weights. Differences in
weight between the sexes were not significant up
to 3 years of age.
The significant differences in weight which
occurred among herds have not yet been
50
Figure 32. Relationship between season ofbirth and growth ofcattle to 3 years ofage.
Live weight (kg)
200-1
150-
100'
50-
20
 
Season of birth
• Cold dry
x Hot dry
* » Rains
* * Post-rains
—i 1 r-
0 2 4 6
~i—
12
1
18
Age (months)
24
—l
36
Table 25. Phenotypic correlations between calf weights.
Age Age (days)
\uays,
30 90 180 210 270 365 550 730 1095
30 1.00 0.47 0.01 0.14 0.20 0.23 0.16 0.38 0.37
90 1.00 0.31 0.25 0.04 0.04 0.31 0.26 0.15
180 1.00 0.91 0.74 0.46 0.62 0.38 0.37
210 1.00 0.86 0.67 0.61 0.52 0.51
365 1.00 0.44 0.66 0.72
550 1.00 0.62 0.51
730 1.00 0.75
1095 1.00
51
 Table 26. Analysis of variance ofpostpartum weights of
dams.
Table 27. Least-squares means for postpartum weights
(kg) ofdams.
Source of
variation
d.f. MS
System
Sex
Parity
Season
Year
Herd/millet
Herd/rice
Error
1
1
3
3
6
3
4
414
797.73
857.92
19 354.09***
10 286.64***
6 433.46***
3 282.96**
2 812.88**
783.95
***P<0.001;**P<0.01.
explained in terms of the differing management
practices or individual abilities of the owners or
herders.
There is an absence of a clear inflection point
in the growth curve in early life. Indeed, growth
rates from 1 to 2 years at 134 g/day and from 2 to
3 years at 143 g/day exceeded the postweaning
gain to 1 year of 121 g/day. Calves and growing
stock obviously suffered severely from poor nutri
tion. In the rice subsystem, the probably slightly
better feed situation was offset by heavier para
site burdens, especially by liver flukes (Traore,
1984).
Postpartum weights in general followed the
pattern that could be expected for parity and sea
son. Primiparous cows weighed approximately
202 kg, compared with a mean weight of 192 kg
for all females at 1460 days. This weight, obtained
from a much larger sample (n = 107) than the
small number (n = 20) of animals whose age at
first calving was known for certain (Section 7),
adds firmly to the hypothesis that first calving
generally occurs at just over 4 years of age. The
present main calving season is in the late hot dry
season, this having repercussions on cow weight
(and therefore ability to survive and supply milk)
as well as the postweaning growth of calves, as
already noted.
The low and diminishing rainfall over the 7-
year study period (1984 weights were included in
this analysis) resulted in a long-term and appar
ently sustained reduction in postpartum weights.
Weights for the years 1982 to 1984 inclusive were
all below the 7-year mean, the lowest being 4.7%
below the overall average and 12.5% below the
maximum weight observed in 1980.
Seasonal effects were evident not only on
growth rates but also on the intra-annual change
Variable
Overall
System
Millet
Rice
Sex of calf
Female
Male
Parity
1
2
3
.9
Season
Cold dry
Hot dry
Rains
Post-rains
Year
1978
1979
1980
1981
1982
1983
1984
Herd/millet
5
7
8
15
Herd/rice
53
GO
64
67
69
433 222
±SE
2.73
192 220 3.28
241 223 3.06
231 223 3.07
202 220 3.05
107 202a 3.67
86 227b 3.85
59 227b 4.42
181 231b 3.00
69 220a 3.83
283 209b 2.49
64 232c 4.08
17 227ac 7.15
15 227abc 7.78
31 227a 5.52
97 238b 3.50
85 222a 3.63
106 216c 3.53
85 212c 3.68
14 215c 7.97
23 217ab 6.48
25 218a 6.31
88 215a 3.65
56 232a 4.30
78 213a 3.82
52 219ab 4.58
57 225ab 4.47
31 232b 5.61
23 228b 6.37
Note: '9' are dams with known parities >4 and unknown
parities >3.
Within variable groups, means followed by different letters dif
fer significantly (P<0.05). Variable groups without any letters
did not show a significant difference in the analysis of variance.
in weight of cattle within specified age groups
(Figure 35), including those under 4 years of age
which should still be actively growing. In addition
to the effects on reproductive performance and
overall output of breeding females, the conse
quences are serious for other aspects of agropas-
toral operations. Work oxen enter the cultivation
season in very low condition. Compensatory re-
52
covery in weight is further delayed in this class of
stock as they have little time for grazing during
the period of agricultural operations and are not,
in general, given adequate high-energy sup
plementation to satisfy their needs.
The long-term decline in mature weight (Fig
ure 36) is seen to be correlated, at least in part,
with a similar decline in the amount of rainfall.
Mature weights ofoxen were lower by about 80 kg
at the beginning of 1984 compared with the initial
weights in 1978, this decline being equivalent to
about 4% per year. In cows, while absolute weigh
changes were less than for oxen - 40 kg lighter
(2.7% per year) in the rice subsystem and 27 kg
lighter (1.8% per year) in the millet subsystem -
the consequences on productivity are likely to be
more serious. These long-term changes in indi
vidual body weight are a result not only of the re
duced rainfall leading to lower fodder availabil
ity, but also of a probable increase in cattle num
bers over the study period. These weight changes
do not appear to have resulted, so far, in de
creased calving rates or increased mortality, but
the time cannot be far distant when these will
occur.
Figure 33. Relationships between postpartum weighs of
cattle and parity, season and year ofcalving.
Liveweight
(kg)
230H
220-
210-
200
Parity
230i
220-
210-
200
 
D J F M A M J J a| s 0 N
Cold dry Hot dry Rains Post-rains
Season
240n
230
220- '
210
 
78 79 80 81 82 83 84
Year
53
Figure 34. Long-term growth patternsfor calves bom in May each yearfrom 1978 to 1982.
Liveweight
(kg)
250-,
200-
150-
100-
50-
 
1978 ' 1979 ' 1980 r
1981 ~~' i982 ' 1983 ' 1984
Yeor
54
Figure 35. Seasonal weight changes in mature oxen andfemale cattle with 1,2,3 and 4 pairs ofpermanent incisors.
Liveweight
(kg)
350-
310-
270-
230-
190-
150
110
oxen 
3 pairs
2 pairs.
I pair
~D ' J ' F '~M ' A ' M ' J ' J ' A ' S-" 0 """n '
Month
Post-
rainsCold dry Hot dry Rains
Season
55
Figure 36. Long-term weight changes in oxen and breeding cows related to changes in rainfall.
Liveweight
(kg)
350-i
310-
270-
230-
190-
150-
110
Rainfall (mm)
400
200-j
Calculated
regression
 
Irrigated rice
subsystem
Rainfed millet
subsystem
Dec
1978
Dec
1979
—I
Dec
1980
—I—
Dec
1981
Dec
1982
I
Dec
1983
 
Table 28. Regression analyses" ofweight (kg) on timefor different classes ofcattle and ofrainfall (mm) on time, 1977-
1983.
Subsystem
and class
Number of
observations a b r P
Rice
Mature work oxen 835 -1.37 378.6 -0.3856 0.0000
Females 4 pairs incisors 1364 -0.56 259.3 -0.2279 0.0000
3 pairs incisors 341 -0.57 237.8 -0.2322 0.0000
2 pairs incisors 272 -0.36 200.7 -0.1821 0.0013
Millet
1 pair incisors 246 -0.51 183.3 -0.2265 0.0002
Females 4 pairs incisors 1070 -0.38 252.2 -0.1972 0.0000
3 pairs incisors 229 0.48 175.5 0.1924 0.0017
2 pairs incisors 193 -0.03 181.4 -0.0188 0.3978
1 pair incisors 155 -0.40 175.8 -0.2194 0.0030
Rainfall 7 -33.40 452.1 -0.7871 0.0450
* y = ax + b, intercept at January 1977.
56
9. MORTALITY AND OFFTAKE
ABORTIONS
A total of 15 (3.32%) of all the 452 births re
corded were abortions.
MORTALITY TO 4 YEARS OFAGE
Details of deaths to 4 years of age are given in
Table 29. The calculated overall figure was 31.6%
of all animals born (excluding abortions - if these
are included, the figure would be 34.9%), al
though a true figure would be slightly higher than
this as a number of animals had not reached
4 years. Deaths to weaning at about 7 months of
age were about 9% (Figure 37), with the major
risk of dying during the first month of life. There
was a low death rate between weaning and 1 year
of age. A major crisis period for young animals
was during their second year of life , this being fol
lowed again by a low death rate between 2 and
3 years of age and an almost negligible risk of
dying in the fourth year.
As can be seen from Table 29, the system,
the season of birth and the year of birth had signif
icant effects on the levels of mortality.
MORTALITY IN ADULT CATTLE
Cattle over 4 years old (those with four pairs of
permanent incisors if actual age was not known)
were considered to be adult. The number of adult
cattle years was calculated for the whole study
period as 1877.5. A total of 94 adult cattle died
(including eight lost and not recovered). The
death rate was calculated as 94/1877.5 x 100, this
being equal to 5.01% .
OFFTAKE
Offtake was calculated as sales plus slaughter (for
social purposes or in extremis) plus animals gifted-
out permanently. A few slaughters and sales
(especially the latter) took place before animals
were mature. Calculated on the same basis as
adult mortality, total offtake was 8.36% com
posed of 0.53% slaughtered, 6.98% sold and
0.85% gifted animals.
DISCUSSION
Abortion - at least in late pregnancy when it can
be more easily detected - has been a minor prob
lem since 1978. The main periods at which death
occurred were the first months of life and the sec
ond year of age. It is probable that some young
animals did not receive sufficient attention in
early life to enable them to survive. The second-
year death rate most probably resulted from an
accumulation of general stress and disease fac
tors.
No specific diseases were identified as major
causes of death during this phase of the study, al
though liver fluke has since been identified as a
contributing factor in the general debility
(Traore\ 1984). A constantly low -and decreasing
in the long term - level of nutrition plays a major
overall part in animal losses.
Overall, the abortion rate recorded in this
study was considerably lower than that found on
the Station du Sahel from 1966 to 1976, and the
total mortality rate was not much higher than the
26% to 3 years of age recorded there (ILCA/IER,
1978).
57
Table0.Observedmortalityrates'foragrop toralc to4yea sfag ,pres ntedydiff r ntv riab es.
System
%
Sex
%
Parity
Seasonofbirt0
%
Yearofbirt0
Millet
Rice
0.000
00.0 4.0 4.0
1i.0
4.0 1.0
i0.01
000
060
000 00 00 001
0
0.000
0.0 4.0 0.0 00.0 0.00
Colddry Hotdry
Rains
Post-rains
i
0.100
4.0 10.0
1i.0
0.0 00.0
i.1
0 0 0 0 •9
0
0.01i
00.0 0.0
i.0
Female
Male
0
0.000
0.0 0.0 0.0
i
Mantel-CoxTest
Statistic
d.f.
Significance
x=00.0%.
&
F0ore0 .A0e-specifichazardandcumul tives rvivalatesf0 opastoralt leintrM i.
Ageospecific hazardrater%)
5i
a.Overall
15 ICH 5 0
 
5
I5H
5 5. 0 20 15
b.System
 
15
c.Sex
H^v
Cumulativesurvival
rater%)
Iyo,
 
340
Ageryears)
 
Age-specifichazard
rater%)
5o1'
d.Parity
 
Cumulativesurvival
rater%) lOOi 340
Ageryears)
 
10. PRODUCTIVITY
MEAT PRODUCTION: A CASE STUDY
AT NIONO SLAUGHTERHOUSE
Numbers and seasonality of animals slaughtered
In the year from March 1979 to February 1980, a
total of 1075 cattle, 643 sheep and 5794 goats were
slaughtered at Niono slaughterhouse. Chi-square
tests showed highly significant differences
(P<0.001) in the monthly numbers of animals
slaughtered. There were significant increases
(P<0.01) in the numbers of cattle (1210) and
sheep (756) slaughtered in the following year
(March 1980-February 1981) but not (P>0.05) in
the numbers of goats (5858) . A very small number
of camels - 20 to 30 head per year - were also
slaughtered.
Contribution of cattle to meat supply
As can be seen from Figure 39, cattle provided
most of the marketed meat in Niono town, this
being equivalent to about 60% ofthe total supply.
At certain times of the year, however, they pro
vided less meat than goats, indicating some com
plementarity of these two species. Beef available
to the urban population of Niono through com
mercial channels was 7.5 kg per person in 1979/
1980 and 7.8 kg in 1980/1981, excluding edible
offal. However, cattle do not provide proportion
ately as much offal as do small ruminants due to
the significantly smaller proportion of lungs (X2
= 9.13, d.f. = 1, P<0.01) and liver (X2 = 10.40,
d.f. = 1, P<0.01) condemned for diseased states
among small ruminants.
Sex and age structure
For all three species, females were the dominant
sex slaughtered although to a lesser extent in the
case of goats. This is well illustrated by Figure 38
which shows that 68% of cattle, 76% of sheep and
54% of goats slaughtered at Niono were females.
The greater numbers of females slaughtered are
due to a considerable extent to the much greater
demand for male animals for export (in the case of
cattle and sheep), and also to the demand for
work oxen.
For all species, females were generally in the
older age ranges having outlived their useful re
productive life. Broken-mouthed animals total
led some 25% of all females in the oldest age
group (both full and broken-mouthed females are
shown with four pairs of permanent incisors in
Figure 38).
Carcass weights
The overall weighted mean carcass weight for
cattle was 109.4 kg. This is 32.6% lower than the
official figure of 145 kg used in national statistics.
MILK PRODUCTION
Few data were obtained on this characteristic.
Average milk offtake for human consumption
was estimated to be 1.088 litres per day with a
range of 0.502 to 2.490 litres for individual ani
mals at different stages of lactation. Lactation
length averaged 297 ±81 days in the range 200 to
506 days. Total lactation offtake for human
consumption was estimated at 323 litres.
PRODUCTIVITY INDICES
The characters of calf weight at 12 months of age,
reproductive performance (expressed as the par
turition interval) and calf survival (the death of a
calf resulting in a zero index for that cow) have
been used to construct productivity indices for
individual cows.
The three indices used were calculated as:
Calf weight
Index I Weight of calf at 12 months x 365
produced per =
Subsequentcow per year
parturition interval
60
2o
OJ2
II
 
o ro
o
T 1
oo O
COro
 
T3 .c
w
5 o>53 3 o
2 o o3 • - d
00 o o
W c E
£ 2 c
a £ o IO
i
o
1 1 1
E (? o
o o o O
r^ 10 lO 1- ro oj
E
iu
— ff o
E
(- Q.
O
<t -C
0) o
ro a> Q.
w >. II
CO
0)
D
o
5 o c
o
ro
a.
E
0)
i-
1-
CO
61
Cow postpartum
weight
Index I
Index II Weight of calf Index I
produced per =
kg liveweight
of cow per year
Index HI Weight of calf
produced per =
kg metabolic
weight of cow
per year
In total , 247 indices were calculated and were
used in least-squares analysis to test for the effects
of different variables on the indices. Year effects
were not tested because of the long parturition in-
Cow postpartum
weight '.73
tervals in relation to the total period of the study.
The results of the analysis of variance for the
three productivity indices are shown in Table 30.
The least-squares means of the three produc
tivity indices are given in Table 31. Only parity
had any significant effect on productivity, first
calvers being less productive than all other classes
of cows except third calvers. No other variable
exerted a significant effect, this probably being
due to the overall poor level of performance.
Herd productivity would be less by some 5%
than that given in Table 31 due to deaths of
breeding females.
Figure 39. Species contribution to meat suppliedfrom Niono slaughterhouse.
Meat supply
('OOOkg)
16-1
14-
12-
10-
6-
4-
2-
I
 
i—i—i—i—i—i—i—i—i—TTn—i—i—i—i—i—i—i—i—i—i—r
FMAMJJASON JFMAMJJA SON
1979 1980
62
Table 30. Analysis of variance ofcattle productivity indices.
Source of
variation
d.f. Index I
Index II
(xlO4)
Index III
(xlO4)
System
Season
Parity
Sex
Herd/millet
Herd/rice
Error
1 17.0 9.2 12.6
3 391.0 120.5 1 819.5
3 2 751.9*" 246.8 5 530.4*
1 99.6 103.5 1677.8
3 559.1 41.1 827.8
4 457.9 108.2 1 935.1
132 457.8 108.2 1 935.1
*P<0.001;*P<0.05.
Table 3 1 . Least-squares means ofcattle productivity indices.
Variable
Index I (kg) Index 11(g) Index III (kg)
n X X X
247 34.40 164 0.698
247 36.20 163 0.704
106 36.61 161 0.702
141 35.96 166 0.707
39 37.57 168 0.722
167 36.44 176 0.747
30 30.09 134 0.587
11 41.04 176 0.762
74 28.16a 140 0.586a
50 52.09b 181 0.782b
32 33.88ab 155 0.677ab
91 41.02b 178 0.722b
136 35.62 157 0.677
111 36.95 170 0.732
11 40.45 168 0.739
12 43.04 181 0.787
50 32.80 148 0.636
33 30.17 146 0.644
39 34.45 171 0.730
33 33.20 147 0.619
39 33.67 149 0.643
16 40.71 187 0.798
14 37.77 176 0.745
Observed mean
Overall LS mean
System
Millet
Rice
Season
Cold dry
Hot dry
Rains
Post-rains
Parity
1
2
3
4+
Sex
Female
Male
Herd/millet
5
7
8
15
Herd/rice
53
60
64
67
69
Within variable groups, means followed by different letters differ significantly (P<0.05). Variable groups without any letters did not
show a significant difference in the analysis of variance.
63

PART THREE
SMALL RUMINANT PRODUCTIVITY

11. REPRODUCTIVE PERFORMANCE
Reproductive performance has been analysed
under three main headings: early reproductive
performance; overall reproductive performance
including litter size, parturition intervals and an
nual reproductive rate ; and effects of a number of
climatic variables on reproduction. Most data re
sult from the long-term study although some data
from specific short-term studies have been
included where considered appropriate.
AGE AND WEIGHT AT EARLY
PARTURITIONS
Histograms of the observed distributions of
weight at first conception are given in Figure 40
for both goats and sheep. The mean weights
(x ± s.d.) for goats were 16.6 ± 3.46 kg in the
range 6.8 - 29.5 kg and for sheep 22.8 ± 4.29 kg
in the range 12.8 — 45.6 kg.
At first parturition the age of goats averaged
485 ± 128.87 days in the range 275 - 1104 days
Figure 40. Distribution ofweights atfirst conceptionfor
goats and sheep in central Mali.
No. of animals
30
20
10-
a. Goats
n = 22l
t} jfl &*,n-p
20-
10-
b. Sheep
n = 207
-r-i >
9
r-R-
12
flfl
5 18 21 24
Liveweight ( kg)
HW
27 30 33
while for sheep age at first parturition was 480.2 ±
115.27 days. At second parturition the figures for
goats and sheep respectively were 760.7 ±
144.46 days in the range 471 — 1300 days and
756.3 ± 128.45 days in the range 546 - 1221
days. The frequencies of the observed distribu
tions for these data are shown in Figure 41.
Figure 41. Ages atfirst andsecondparturitionsforgoats
and sheep in central Mali.
No. of parturitions
30- -, a. Goats
-. n = 225
20- J-
10-•f nTiTh- " n n
10-
mrilUlT -Ti 111 rmrn.n
40H
30
20
10-
10-
&
b. Sheep
n = 2IO
m Th-n
,JlMnh=;l20r
9 12 15 18 21 24 27 30 33 36 39
Age (months)
The mean-squares values for the three traits
are laid out in Table 32 while the least-squares
means are given in Table 33. There was a signifi
cant effect of system on weight at first conception,
with animals from the rice area being significantly
heavier (P<0.05 for goats and P<0.001 for sheep)
than those from the millet subsystem. Month of
birth and maternal parity also had a highly signifi
cant effect on sheep weights. Sheep weighed sig
nificantly (P<0.001) more than goats at first con
ception. System was the only variable to have a
significant effect on age at first and second partu-
67
O ^ (N Q ^ CO **
"-i & m * c*I i0 Q
t- F- r- i-" os o «
c
o,a,c
3
c«
a
T3
a
u
so
<
o
0
on
5
00 a
1
1
I
,5
§,
a
a
c/3
s
H « H J. H
S 5 S. 8
a S s 2
^H r-l t-i f^ CN iri cs
8
|
1
2
Oi X N
5s ^ io
CO 5s VI
o
O
tN 3
cs -* o
"5.
c
o
a.
a.B
u
j=
on
on
S; 3
h « .-i r- cs <rI os
2
1
o
O
t/3
s *© m r- o ohi «n v> en th
— — th r-* 3
V
a.
o
Va.
t
Si
C «
I!
■o .5
o £
1 52 1
C/5 2
8
E
va.
68
1
a
b
c6^
■5
ft.
ST
| I^ •c■a 3
c c
a «
1 i0 ISac 9k
■6.
-a
a
I
T3 'S
§ 8
a
0 8« Hft.
5! c
c CO
0 acio
rn oo
in (N
°1 *?
cn Os o>r^t6osNmpi■*tin©>-i'-it-^'-h
r-r-r-r-r-r-«r-
r- oo
r■ r-
00 Os
os vj
PS
'Q^'irJciQaofN^OQsN(^
oor-«osojror-mt^osorj op r-so in
t- p-
so rn rn ih h H h "8322
Wgcs*-5psosr^oiQ,03^-
NoBSopF.5Mrto*tNt^
SP-spsososoOsW-lW-iOQcm©
ui os <s "o aO'-icsao*-ir^m^'m©«oo
io« 9Pi0 P^rno^oornOsOsiris-oso
op os 5 aj io*t>wNa^tNr-oc>'-,J
r; ©
o p-
("-r^r■r^r-roinoo
(NPo;in*sornOs©
I
s
8
■o
1
o
in rn t>r-;0'0'*eninoo
co © rJr^Om^ps^wS
Os--h»(N©0p^HiO
^'ro^,^'"ors©oom©rn^' oo os OssQrnrnmso©r^
HHrfiNHHH <t ^ «o rl Os© rN ^ ^ rn N "-" «-i
2i.
00 00
Pi Pi
•O w
(N O
it,
we te■§■§:=• 81 „fi.
A rt cfl X A C -C fl
(N<nO'—ON'COssSi os i>
Pis
«-i©oo.-i^ri>-os~-
oo os (^^HCM'-h*-'«-h*-ii-hrnMrn(N
rntwi*-'©rn"or*rnrn«r^
r^wSr^oor^r^r^«r^(^««
s £ §3
i> ooao(N«'«frnoooo
~- vi in * ^ N i-i
aoaq«Osoqr;(Nr;
s*?saas'-
III
-Is5-8
111 1111 fill II
i& 5* —
a.•c
"8 u
in S-
(N pn ^ «o * r-
69
 ritions, with the effect being confined to sheep.
Although in goats the system effect was not signif
icant, those reared in the rice areas also had their
first and second parturitions earlier. Flock within
system had significant effects on all the traits, with
the effects being much more marked in the natur
ally less well endowed millet area.
There was no significant correlation between
age at first parturition and the interval to the sec
ond parturition either for goats (r = 0.20,
P>0.05) or for sheep (r = 0.03, P>0.05).
LITTER SIZE, PARTURITION INTERVAL
AND ANNUAL REPRODUCTIVE RATE
Annual reproductive rate (ARR) was calculated
as a function of litter size and parturition interval
according to the formula: litter size x 365/sub-
sequent parturition interval. Calculations for par
turition interval and ARR were limited to animals
having given a previous birth in 1983 in order that
short intervals would not bias the results.
During the study period there was a total of
3605 parturitions giving rise to 4049 young. The
distribution of births and young is shown in
Table 34. Mean litter size (x ± s.d.) for goats
was 1.19 ± 0.41 and for sheep 1.04 ± 0.21. Litter
size increased from 1.04 for primiparous goats to
a maximum of 1.39. In sheep, litter size varied
from 1 .01 for primiparous females to 1 . 13 in older
animals.
The observed distribution of parturition in
tervals is shown in Figure 42. The mean for goats
was 291 ± 105.2 days and for sheep 261 1 76.3
days. The longest interval for goats (298 ± 99.2
days) occurred between the first and second
parities and the shortest (208 ± 45.4 days) after
the ninth parity. For sheep, the longest interval
(208 ± 63.6 days) was again after the first parity,
with a decrease in the length of the interval in sub
sequent parities. Overall annual reproductive
rates, calculated from the two previous param
eters, were 1.49 for goats and 1.45 for sheep.
These data from the long-term study can be
compared with those from a rapid survey carried
out in early 1978 which attempted to establish the
distribution of types of birth and litter size by age
class. The main results from this small study are
shown in Table 35.
Mature animals with an unknown breeding
history were not included in the least-squares
analyses. Parity in this set of analyses refers to the
order of births in the reproductive career of the
animal whose records are being analysed and not,
as in the section on early reproductive perform
ance, to the parity of its dam. Parities greater than
4 were grouped and treated as fourth parity and
triplet and twin parturitions were grouped as mul
tiple births. A separate analysis was carried out
for each parameter for each species. The models
considered the random effect of dam within sys
tem and within flock and the fixed effects of sys
tem, flock, season and year of parturition, type of
birth, sex of offspring and parity. During prelimi-
nary analyses, system (rainfed millet and irrigated
rice) was seen to have significant effects on some
of the traits being studied. Because of the limita
tions of the models, the main analyses were there
fore carried out by systems with dams within
flocks as the random variable. Orthogonal
polynomials were fitted to some of the main vari
ables in an attempt to further understand some of
the influences acting on them.
Levels of significance for the different
sources of variation acting on the three measured
traits are shown in Table 36. System influenced
significantly all three traits in sheep but none at all
in goats. All further analyses resulting from this
analysis of variance were however carried out by
system with flock introduced as a fixed effect. The
practical results of the effects of system on sheep
Table 34. Numbers ofparturitions and percentages ofyoung by parturition type for sheep and goats in central Mali.
Type of birth
Sheep Goats
No. % No. %
Total births 1650 100.0 1955 100.0
Total young 1722 100.0 2 327 100.0
Single births 1 579 95.7 1593 81.5
Young born as singles 1579 91.7 1593 68.5
Twin births 70 4.2 350 17.9
Young born as twins 140 8.1 700 30.1
Triplet births 1 0.1 12 0.6
Young born asi triplets 3 0.2 36 1.5
70
Figure 42. Frequency distribution ofparturition intervals for sheep and goats in central Mali.
Percentage of all
parturitions
3CH
20o
10o
a. Sheep
n=984
b. Goats
n = 1 1 1 1
8 10 12 14 16 18 20 22 24 4 6 8 10
Parturition interval (months)
=F I'M
12 14 16 18 20 22 24
were that those in the rice subsystem had a better
total reproductive performance than those in the
millet subsystem. This was probably due to the
better year-round feed conditions in the former
subsystem, a hypothesis supported by the lack of
significance in the seasonal effect on all three
traits for sheep.
The least-squares means for the three traits
are shown in Table 37 for sheep and in Table 38
for goats. Table 39 provides data on the best-fit
polynomials for variables where significance oc
curred in the main analyses, and Figure 43 shows
in graphic form the calculated orthogonal fits.
Other than the effects of the flock factor, and with
the exception of the effects of year and parity on
litter size in goats in the rice subsystem, all signifi
cant effects are confined to one or the other
species in the millet subsystem.
EFFECTS OF CLIMATE ON THE PERIOD
OF BIRTH AND ON LITTER SIZE
For these analyses the numbers of births and the
mean number of young per birth were calculated
for each 10-day period from 1978 to 1983. Simple
correlations and multiple linear regressions were
then calculated between these data and six climatic
variables for each of the 15th to 20th 10-day periods
prior to those being analysed for numbers of par
turitions and litter size. BMDP programmes
(Dixon and Brown, 1983) were used for the statis
tical analyses.
The distribution of 1702 parturitions for
goats and 1500 for sheep by month is shown in
Figure 44 which also shows the mean litter size by
month. There were highly significant differences
among months for the numbers of parturitions in
goats (X2 = 553.8, d.f. = 11, P<0.001) and for
those in sheep (X2 = 108.0, d.f. = 11, P<0.001).
The minimum number of parturitions (54) in
goats occurred in August and the maximum (323)
in November. In sheep, the minimum number
(88) was in May, the maximum (174) being in
September.
Litter size in goats varied from a maximum of
1.31 kids in March to a minimum of 1.14 in both
August and October. In sheep, maximum litter
size of 1.14 lambs was achieved in April with a
minimum of 1.01 in September. An analysis of
variance showed significant differences among
months for both goats (F = 5.55, d.f. = 11,1657,
P<0.001) and sheep (F = 3.90, d.f. = 11,1444,
P<0.001).
The maximum number of conceptions in
both goats and sheep took place in May to July
while the maximum number of ova appeared to
be shed in October and November in goats and
perhaps slightly later (in November and De
cember) in sheep. The smallest number of con
ceptions in goats took place in February and
March, while in sheep there was a fairly clearly
defined period of poor conception rates extend
ing from October through March. Monitoring ac
tual oestrus activity in sheep in the Egyptian sub-
tropics, Aboul-Naga et al (1985) found it to be ir
regular throughout the year but with a minimum
in April. Monitored by hormone levels in Niger,
in climatic conditions similar to those at Niono,
minimum oestrus activity and longest cycles were
found from January to April (Yenikoye et al,
1982). In the latter case, abnormally long cycles
contributed to the low numbers of conceptions
observed: there was, however, a rapid upsurge in
oestrus activity from May onwards.
Correlations between goat parturitions and
climatic variables were positive and highly signifi
cant for maximum temperature (r = 0.30 to
71
Table4.Reproductiveperformancofag opastoralg a sndsheep,b s dnow s'rec lldat .
Overall
000 10 00 0 000 000i i.0
Temporary
00 0 0 0
0.0
lpair
00 1
0 00
60
i.60
Goats
0pairs
10 0 0
00 00
0.0
1pairs
1 04 0 000 010
i.00
0pairs
010 00 00 0 00 000
0.0
Overall
010 00 4 0 00 00
0.0
Temporary
07 0 0 0
0.0
0pair
i.0
S0eep 0 0
0
0 4
0pairs
22
4
0 0 00
i.00
1pairs
00 00 0
4 0
0.00
0pairs
0 00 0
0
00 40
0.00
Numberinsample Totalparturitions Totalyoungb rn Averagelitters z
Typeofbirt0
Single Twin
Triplet
Table10.Analysisofvariancelittersiz ,p turitionintervalandarep odu veatfhe pg sc ntM li.
Trait/species
Sourceofvariation
System
Flock
Dam/Flock
Season
Year
Parity
Sexof young
Typeof
birt0
Litters z
Alls0eep MS"s0eep RS"s0eep Allgoats MSgoats RSgoats
Parturitionin erval
Alls0eep MSs0eep RSs0eep Allgoats MSgoats RSgoats
Annualreproductiverate
Alls0eep MSs0eep RSs0eep Allgoats MSgoats RSgoats
n.s. n.s. n.s. n.s. n.s. n.s.
n.s. n.s. n.s. n.s. n.s.
n.s. n.s. n.s. 060 **
*
n.s. n.s. n.s. n.s. n.s. n.s. _ - - - - -
* n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.
n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. * * n.s.
n.s. ** il.S. ♦* t n.s. n.s. n.s. n.s. * * n.s.
**
* n.s. * n.s. n.s. •* ** n.s. n.s. n.s. n.s.
« * 60 • *« • n.s. * n.s. n.s. n.s.
•♦*P<0.000;**P<0 ;'P<n.s.=notsignificant.
*MS=milletsubsystem;Rrice .
S3
 0.47), minimum temperature (r = 0.25 to 0.42)
and day length (r = 0.27 to 0.40) for all of the
10-day periods from the 15th to the 20th previ
ously. A similar pattern was observed for sheep
with correlations ranging from 0.29 to 0.39 for
maximum temperature, from 0. 17 to 0.37 for min
imum temperature, and from 0.25 to 0.42 for day
length. There were significant positive correla
tions also between the parturition interval and
both maximum and minimum humidity at the
15th period for both goats and sheep and between
parturition interval and rainfall at the 15th period
for sheep, but no other periodic conditions were
significant for these parameters and some were
negative.
The litter size of goats showed significant
correlations with maximum temperature
(r = 0.14 to 0.19) at the 17th to 19th periods and
with both maximum humidity (r = 0.16 to 0.22)
and minimum humidity (r = 0.14 to 0.22) at the
17th to 20th periods previously. In sheep, litter
size was significantly correlated only with
maximum humidity (r = 0.16 to 0.26) and mini
mum humidity (r = 0. 15 to 0.26) for all of the
15th to 20th periods in the case of the former and
from the 17th period in the case of the latter.
When stepwise multiple regression was
done, the final equations containing terms signifi
cant at the 5% level accounted for a low propor
tion of the variance. The maximum variance ac
counted for by parturitions in goats was 25.8% at
the 17th period previously, with maximum tem
perature accounting for 19.9% of this and day
length for the rest. For sheep, the maximum var
iance accounted for, at the 18th period, was
17.7% with maximum temperature contributing
14.6% and day length the rest.
The total amount of variance accounted for
by litter size was much lower, being 6.7% at the
19th period with minimum humidity accounting
for 3.7% and minimum temperature for 2.9%. In
sheep, most variance (6.7% and 6.8% respectively)
was accounted for by relative humidity at the 19th
and 20th periods previously.
NUMBERS OF BIRTHS BY AGE CLASS
During the rapid study already referred to it was
possible to obtain an indication of the numbers of
births by each class of age of female. The main re
sults are provided in Table 40. For sheep, ewes
with milk teeth and one to four pairs of perma
nent incisors had given birth 0.43, 1.00, 2.05, 2.57
and 3.62 times respectively, the overall number of
births per breeding female in the flock being 1 .84.
For goats, the numbers of births were 0.33, 0.70,
1.39, 1.95 and 3.71 for.the individual age groups
and 2.22 births for all does in the flock. For both
species there was a rapid reduction in the number
of females having given birth more than five
times.
DISCUSSION
Early reproductive performance
The mean age at first kidding of 463 days is in
good agreement with the few other existing data
for goats of a similar type managed under tradi
tional systems in the West African semi-arid
zone. For example in Chad, it has been found that
age at first parturition varies from 418 to 502 days
depending on the region (Bertaudiere, 1979;
Dumas, 1980). In Niger, Gerbaldi (1978) found
that Sahel-type goats first kid at 401 ± 24 days. In
southern Niger, the Red Maradi goat had an age
of 427 ±13 days at first kidding (Haumesser,
1975).
The age at first lambing of 474 days also
agrees well with other published reports. In
Chad, mean age at first lambing varied from 395
to 502 days depending on the region (Gerbaldi,
1978; Haumesser and Gerbaldi, 1980).
Management practices to control breeding
are generally designed to delay the age at first par
turition. This applies whether the management is
traditional, as in the case of the Maasai pastoral
system (where an apron is used to cover the penis
of the male and some flock stratification is evi
dent) where the age at first kidding is 556 ±
119 days and at first lambing 549 ±112 days
(Wilson et al, 1984), or under station conditions,
as in Senegal (where males are separated from the
females when breeding is not desired and females
are mated only on attaining a specified minimum
weight) where age at first lambing is 572 ±
24 days (Fall etal, 1982).
The greater weight at first conception in the
irrigated rice subsystem reflected the differences
in food availability between the rice and millet
subsystems, and the generally faster growth rates
(from birth weights of 1.8 kg for goats and 2.6 kg
for sheep in both subsystems) and heavier mature
weights achieved in the rice subsystem (31.8 kg
for goats and 35.9 kg for sheep), compared to the
millet subsystem (28.1 and 32.1 kg respectively
for goats and sheep). Earlier physical maturity in
the rice subsystem at first conception also appar
ently results in slightly earlier physiological
maturity, reflected in the 35-day younger age at
first parturition.
The only other variable to affect significantly
the age at first parturition is maternal parity, with
74
Table 37. Least-squares meansfor litter size, parturition interval and annual reproductive rate ofsheep in central Mali.
Variable Litter size
Millet subsystem
Parturition
interval
reproductive
Annual
Litter size
Rice subsystem
Parturition
interval
reproductive
Annual
rate rate
n
No. of
young
n Days
Young/
year
n
No. of
young
n Days
Young/
year
Overall 909 1.036 502 290 1.53 445 1.031 257 259 1.63
Flock"
1 85 1.012 40 304 1.37 86 1.032 31 257 1.59
: 35 1.021 19 251 1.69 160 1.089 102 237 1.85
3 39 1.152 26 262 1.99 74 1.045 49 228 1.73
4 50 1.022 27 317 1.40 58 0.984 36 241 1.65
5 207 1.011 115 265 1.61 24 1.006 13 310 1.37
6 101 1.016 65 271 1.56 43 1.033 26 279 1.57
7 54 1.016 28 308 1.40 - - - - -
8 78 1.010 45 302 1.42 - - - - -
9 31 1.117 20 264 1.80 - - - - -
10 122 1.033 59 288 1.54 - - - - -
11 32 1.002 15 365 1.18 - - - - -
12 37 0.999 19 303 1.36 - - - - -
13 38 1.053 24 275 1.57 - - - - -
Average SE 0.0236 18.1 0.096 0.0316 19.6 0.115
Season
Cold dry 271 1.053 147 297 1.46 101 1.039 45 262 1.62
Hot dry 206 1.074 117 285 1.63 143 1.054 90 256 1.74
Rains 217 1.004 117 267 1.62 103 0.991 60 257 1.62
Post-rains 215 1.018 121 312 1.40 98 1.042 62 260 1.52
Average SE 0.0144 13.0 0.054 0.0327 16.0 0.086
Year
1978 90 1.019 - - - 11 0.932 - - -
1979 101 1.023 81 306 1.48 26 1.054 16 222 2.01
1980 159 1.020 110 298 1.44 70 1.062 44 275 1.55
1981 159 1.023 122 296 1.47 79 1.070 60 258 1.66
1982 144 1.042 105 300 1.50 105 1.022 82 264 1.56
1983 166 1.065 84 252 1.75 107 1.048 55 274 1.36
1984 90 1.057 - - - 47 1.032 - - -
Average SE 0.0263 15.3 0.078 0.0751 24.8 0.165
Parity
1 271 1.034 141 292 1.51 145 1.012 84 271 1.47
2 212 1.055 115 287 1.59 105 0.992 60 257 1.57
3 155 1.026 89 277 1.59 80 1.038 46 248 1.69
>4 271 1.028 157 306 1.42 115 1.078 67 257 1.78
Average SE 0.0177 14.2 0.067 0.0396 17.8 0.104
Sex
Female - - 272 281 1.56 - - 116 256 1.55
Male - - 230 300 1.49 - - 141 261 1.70
Average SE 11.7 0.039 14.2 0.067
Type ofbirth
Single - - 487 269 - - - 242 253 -
Multiple - - 15 312 - - - 15 264 -
Average SE 15.4 - - 16.8
Rocks in the millet and rice subsystems carrying the same number are not necessarily under the same ownership.
75
 Table 38. Least-squares meansfor litter size, parturition interval and annual reproductive rate ofgoats in central Mali.
Millet su bsystem Ricesubsystem
Variable Litter size
Parturition
interval
reproductive
Annual
Litter size
Parturition
interval
reproductive
Annual
rate rate
n
No. of
young
n Days
Young/
year
n No. of
young
n Days Young/
year
Overall 1104 1.154 597 298 1.53 310 1.177 166 297 1.63
Hock"
1 71 1.182 29 272 1.58 38 1.323 21 286 1.75
2 174 1.236 102 267 1.79 39 1.275 21 279 1.82
3 173 1.160 106 255 1.84 16 1.253 7 251 1.93
4 42 1.310 20 270 1.83 50 1.062 28 295 1.49
5 39 1.106 21 328 1.39 147 1.084 82 295 1.52
6 23 1.245 13 297 1.68 20 1.066 7 377 1.29
7 22 1.043 11 325 1.28 - - - - -
8 41 1.160 15 293 1.73 - - - - -
9 311 1.182 171 302 1.44 - - - - -
10 32 1.049 17 280 1.52 - - - - -
11 30 1.095 16 343 1.24 - - - - -
12 60 1.050 35 334 1.31 - - - - -
13 39 1.152 19 313 1.27 - - - - -
14 47 1.178 22 293 1.52 - - - - -
Average SE 0.0534 21.7 0.132 0.0599 27.9 0.161
Season
Cold dry 260 1.245 140 316 1.50 69 1.161 27 287 1.67
Hot dry 252 1.186 146 291 1.62 75 1.190 40 296 1.52
Rains 184 1.076 114 281 1.56 39 1.144 31 283 1.79
Post-rains 408 1.107 197 305 1.44 127 1.214 68 322 1.55
Average SE 0.0301 12.0 0.072 0.0543 21.9 0.134
Year
1978 61 1.164 - - - - - - - -
1979 99 1.184 70 364 1.40 - - - - -
1980 148 1.208 101 304 1.68 46 1.298 33 303 1.72
1981 211 1.184 153 283 1.69 66 1.126 48 301 1.49
1982 223 1.110 153 278 1.51 67 1.269 49 313 1.54
1983 196 1.081 120 261 1.38 68 1.188 36 272 1.77
1984 166 1.143 - - - 63 1.004 - - -
Average SE 0.0579 17.7 0.112 0.0888 24.5 0.158
Parity
1 370 1.004 197 274 1.37 102 1.028 54 299 1.41
2 256 1.093 157 294 1.36 67 1.060 41 295 1.54
3 186 1.205 106 299 1.65 58 1.277 30 271 1.87
>4 292 1.311 137 326 1.74 86 1.343 41 324 1.71
Average SE 0.0368 14.7 0.095 0.0664 24.8 0.161
Sex
Female - - 304 295 1.55 - - 81 303 1.60
Male - - 293 301 1.51 - - 85 291 1.67
Average SE 9.9 0.052 18.6 0.095
Type of birth
Single - - 545 304 - - - 152 283 -
Multiple - - 52 293 - - - 14 312 -
Average SE 11.5 21.5
Note: * Flocks in the millet and rice subsystems carrying the same number are not necessarily under the same ownership.
76
offspring of young and old mothers maturing later
than those from dams in the intermediate (3-
5 years) age groups. The reasons for this are not
clear as there is no obvious relationship with
weight at first conception.
On the whole the effects of environmental
factors on age at first parturition are small. This
was also found to be the case in Senegal where
neither year, nor month nor type of birth had any
significant effect on age at first lambing (Fall et al,
1982). It would appear, therefore, that little is to
be gained by attempting to control breeding sea
sons to optimise age at first parturition. Mortality
of young born to primiparous females is almost al
ways higher than that of those born to older
females. Although delaying first parturition to an
age when the female is nearer to its full physical
development reduces the death rate to some ex
tent, the slower growth rate of the young is re
flected in the lower productivity of this class of
dam when compared with that of older dams.
There is no evidence from traditional systems that
early first parturitions result in subsequent di
minished reproductive capacity and it is therefore
probably worthwhile selecting for an early first
parturition, accepting the higher death rate and
lower productivity resulting from this and an
ticipating an extra parturition and a longer total
reproductive life. Such a strategy might also en
able a greater proportion of breeding females in
the middle age groups to be maintained in rela
tion to primiparous and older females, thus
leading to further improvements in total flock
productivity.
Main reproductive traits
The effects of parity are those which might logi
cally be expected. Litter size increased linearly
with age in goats in both subsystems. The lack of
significance in sheep is probably due to the overall
low level of twinning. The quadratic form of the
parturition interval of sheep in the millet subsys
tem can be explained by the adaptation of the
ewes to the different partitioning of resources as
they grow, reach maturity and then senesce.
Overall, however, the effects of parity on parturi
tion interval are slight in sheep in the millet
subsystem and do not affect sheep in the rice sub
system. The parturition intervals for goats are not
affected by parity in either subsystem. The signif
icant effect of parity on annual reproductive rate
is confined to goats in the millet subsystem, re
sulting from increased litter size with advancing
age of the dam.
The effects of year are difficult to interpret.
It is possible, however, that the maximum annual
reproductive rate in goats in the millet subsystem
during the middle period of the current study is
related to the recovery of animals and food re
sources from the effects of the 1968-1973 Sahel
drought, followed by a decline due to the effects
of lowered rainfall during the early 1980s and
increased pressure on feed resources as a result
of a build-up of stock numbers. Why the same
phenomenon is not significant in sheep in the
same subsystem is not clear.
Although the effects of flock are significant
for most variables, it has not yet been possible to
isolate the specific mechanisms which cause the
differences. In traditional systems in Africa, ac
cess to the basic resources of grazing and water is
theoretically equal for all flocks. It is possible that
individual management abilities play a role in ac
quiring better use of these resources by ensuring
that animals have adequate time at grazing and
are given sufficient water on a regular basis. Other
practices such as keeping night-holding pens or
picket areas free of manure, and simple veteri
nary interventions, may also influence flock
reproductive performance.
A comparison of these results with those
available from similar studies indicates that tradi
tionally managed goats and sheep in Mali have a
reproductive performance close to the African
average. Observed litter sizes for goats range
from 1.23 in Kenya (Wilson et al, 1984) to 1.51 in
Sudan (Wilson, 1976c) and for sheep from 1.05 in
Kenya to 1.14 in Sudan (Wilson and Clarke,
1976b). In Sudan, goats have been recorded as
having parturition intervals of 238 days and in
Kenya of 306 days. Intervals for sheep range from
275 days in Sudan through 301 days in Niger
(Haumesser and Gerbaldi, 1980) to 320 days in
Kenya. In other studies where goats and sheep
have been recorded together under the same con
ditions, goats have always shown clearly superior
annual reproductive rates. In central Mali, more
detailed research, particularly on management
skills, is needed to explain the superiority of
sheep over goats.
In the systems studied there appears to be
little possibility of improving total reproductive
performance by attempting to shorten the parturi
tion interval. Evidence from other tropical areas
indicates that the figures recorded in this study
are optimal. Creole goats in Guadeloupe average
237 days and the Indian Malabar 300 days (Garcia
and Gall, 1981). Blackbelly sheep in the West
Indies average 248 days and Pelibuey-West
Africa sheep in Central America have an average
interval of 245 days (Fitzhugh and Bradford,
1983).
77
Table4.D tafromthepolyn mialanalysesfrep oductivechar cteristicsg andsh pin alM l .
Speciesandtrait
System
Variable
Ort0ogonal
regression
Regressionconstants -
4.00 -0.00 0.10
-0.0000
-
-0.6000
0.100 -0.000
-
-0.00i
0.00 0.00 0.01
-0.600 -060
0.6000 0.000
0.010
-0.01
-0.010
0.i40
0.00
00.0 40.0 00.0
i.00 i000 i.000 i.000
i.i4 40.0
i.00
0.00
Linear
Quadratic
Cubic
Quadratic Quadratic
Linear Linear Cubic Linear
Quadratic Quadratic
Linear
Season Season
Year Parity
Season Season Parity Year Parity Year Year Parity
Millet Millet Millet Millet
Rice
Millet Millet
Littersize
Parturitionin erval
Annualreproductiverate
Litters z
Parturitionin erval
Annualreproductiverate
S0eep Goats
-0.00
-0.000
00
Figure 43. Best-fit estimates from the polynomial analyses ofreproductive characteristics ofsheep and goats
in central Mali.
No. of young/birth
1.3-
1.2
I.Io
1.0
VMS') goats
-« MS sheep
Hot Rains Post-
dry dry rains
a) Litter size
RS'J goats
 
Cold ' 'R ' 1980 1981 1982 1983 1984
RS goats/
'•"/MS goats
-1 1 1 1
2 3 >4
Days
340-
310
280-
250-
220
190
Sheep
s« «'
b) Parturition interval (millet subsystem)
Goats
X X
-1 1 1 1
VSheepX
-1 1 1 1 1
»^^ Sheep _^-«
^^« »^^
Cold Hot Rains Post- 1979 1980 1981 1982 1983
dry dry rains
c) Annual reproductive rate (millet subsystem )
-i 1 1
2 3 >4
No. of young/year
1.8-
1.7-
1.6- r(
1.5-
/Sheep
1.4-
1.3 1 1—
Goats
Cold Hot Rains' Post- 1979' 1980' 1981 ' 1982' 1983'
dry dry rains
Goats
I ' 2 ' 3 ">4 '
MS = millet subsystem; RS= rice subsystem.
79
 Figure 44. Total number of parturitions per month
and average number ofyoungperparturition
per month for goats and sheep in central
Mali, 1978-1983.
a) Goats
No. of parturitions
300-
b) Sheep
200-1
100
50-1
n=!702 r
n=!500
No.of young/parturition
1.3o
1.2
J M M J S
m
N J
Month
n-rh-rffl
M J S N
The potential for improving the annual re
productive rate by increasing litter size seems
more feasible. Where selection has been prac
tised, the number of kids per birth for Creole
goats in Guadeloupe was 2.33, for the Damascus
in Israel 1.76 and for the Beetal and Jamnapari in
India 1.70 and 1.45 respectively (Garcia and Gall,
1981). Sheep litter sizes of 1.84 in the Barbados
Blackbelly and 1.24 in the Pelibuey in the Carib
bean have been recorded (Fitzhugh and Brad
ford, 1983). The possibilities of increasing litter
size by crossing with the more prolific African
types from similar environmental areas might also
be considered.
Manipulating flock structure so that as many
of the more productive females of older parity as
possible are maintained, could also improve re
productive performance. Encouraging first par
turitions at as early an age as feasible and select
ing breeding stock on the basis of performance at
first parturition (Ozakuma et al, 1982) would also
boost the overall annual reproductive rate.
Climatic effects on reproduction
The positive correlation of the number of parturi
tions with day length found in this study is con
trary to that found in seasonally breeding sheep in
temperate latitudes and is also contrary to the
postulate of Hafez (1951 ; 1952) that in the tropics,
the breeding season of sheep is not affected by
light i.e. day length. The pattern of births in this
study is, however, similar to that recorded in
Chad where 60 to 80% of births occurred from
November to February (Dumas, 1980). The prin
cipal climatic source of variance is ambient tem
perature, this phenomenon also having been
noted previously (Lees, 1971). The combined, in
teracting effects of temperature and day length on
oestrus activity of sheep have also been noted in
Niger (Yenikoye et al, 1982). No previous studies
appear to have attempted to separate the effects
of a simple conception from those of litter size.
The lack of significant correlations with rain
fall is rather surprising. In cattle in this area, par
turition is very highly correlated (r = 0.56,
d.f. = 70, P<0.001) with rainfall 9 months pre
viously (Section 7). As rainfall is almost uniquely
responsible for primary production - except for
the appearance of some browse - and as almost no
supplementary feeding is practised in the systems
studied, it does not appear that nutritional state is
the proximal factor in the conception rate al
though it probably has more influence on the ovu
lation rate. The similar patterns of conception
and of distribution of litter size in both goats and
sheep are also puzzling in view of their different
dietary habits.
More detailed experimental studies under
controlled environmental conditions are required
to explain fully the factors affecting fecundity and
prolificacy in tropical goats and sheep.
80
o
O
c
o•a■C
3
2
B
9
*-i ^ Vi
O i0 sO 00 (N
N •* sO 0 i-3
Os OO *h OO
<■»
3 E — <n o tt
81
 12. GROWTH AND WEIGHT
BIRTHWEIGHT
The observed birthweight (x ± s.d.) of 581 goats
born in the years 1978 to 1984 was 2.2 ± 0.64 kg.
The mean birthweight for 613 sheep during the
same period was 2.9 ± 0.88 kg.
The analysis of variance of this trait for goats
is shown in Table 41 : sex, parity, type of birth and
year of birth of the kid exerted significant effects
on its birthweight. The least-squares means for
kid birthweights are given in Table 42. This shows
that, as might be expected, males were heavier
than females at birth, young of older females
weighed more than those from younger dams,
and twins (and triplets) were lighter than kids
born as singles. Although the year of birth had a
significant effect on birthweight, the observed
variation can not be clearly related to the rainfall
pattern over the study period.
The analysis of variance for birthweight in
sheep (Table 43) shows that basically the same
sources of variation acted on this trait as in goats
although in addition there were highly significant
differences due to system. The least-squares
means for lamb birthweights are given in Table 44.
Lambs born in the irrigated rice subsystem were
heavier than those from the rainfed millet subsys
tem, their dams presumably benefitting from the
better nutritional conditions in that zone. The
other sources of variation exerted effects similar
to those for goats.
GROWTH TO MATURITY
Generalised growth curves for goats and sheep
from birth to maturity are given in Figure 45.
Analyses of variance and estimated least-squares
means for weights at various ages are shown in
Tables 41 and 42 for goats and in Tables 43 and 44
for sheep. The observed sample means in Figure
45 differ from the least-squares means in Tables
42 and 44 as the data in the tables are adjusted for
the unequal subclass numbers.
In goats, the system under which they were
reared exerted a relatively minor effect (except at
150 and 240 days of age) until 1 year of age. The
effects of parity were significant only up to 240 days
and the effects of birth type, with the exception of
some anomalous results around 2 years, were also
quickly lost. Season of birth exerted a significant
effect on weight only at 10 and 30 days. Sex had a
consistently significant effect on weight as did
year of birth. The effects of year of birth on
growth to 3 years of age are shown in Figure 46.
Kids born in 1981 exhibited a similar growth rate
to those born in 1980, and kids born in all sub
sequent years again showed slower growth rates
resulting in lighter weights-at-age when com
pared to kids born during 1978 and 1979.
In sheep, system had a highly significant ef
fect on weight until mature weights were reached
at about 3 years of age. The effects of the other
sources of variation were similar to those ob
served for goats, with the exception of the effects
of year of birth which were lost at about 1 year of
age.
Average daily gains (calculated on the least-
squares means) for goats from birth to 150, 365
and 1095 days were 58.0, 49.3 and 28.7 g/day
respectively. In sheep, the average daily gains
were 88.7, 66.9 and 33.3 g/day to the same ages.
Asymptotic weights were achieved in sheep at
about 3 years while goats were almost 4 years old
before their growth curve leveled off.
Phenotypic correlations between weights at
various ages up to 2 years are shown in Table 45.
In general the correlations were better for sheep
than for goats but all were highly significant for
both species at all ages.
82
I a S R s 2
— m — m m , , , , , g
a § 2 SS ?! §
11(11
5«O^ClClos(*lO2 2
i/1 VI & (s 0s r*
» in (s ©
B,■8
I
S°5S rn oO s0 ^
n^nmnnftnpi
(Noe*r-r-fiooo©^o
^H f-* c^i •"-• c^ v1 fi rn ^ fi fi
» r- in ** so in ■* os p (N Tt
i* « ri ^ ~- "o
- So
1
c
a
a o»»w^pp,oo^r^sopp
I
a■S.
I
1
cn^-msoinoo,«rnso —
O
V
0.
V
PL.
[2
I
5 I
£ • s
en tfi Ph CQ >-
M M M M M
c c - c c
£ 8 1 fc/3 &i 0* H
o
V
OH
83
IM
Vii
a
IK
O
«
a
|M
?!
r*
O
|M
a
in
a
IM
3
a
IM
i
Q
■3
IM
5; 8
"a
1
<i>
a
Q
.5
3
IMa
©
8
,2.
8 a
00
a■a
Ju |M
»1
S
a
a
oo
^
a<
5> IM
B
i o
"&
a
i
8 IM
§
o- o
*n
i e5
csi
-,-
u
3
>
to
H
& c* » * m » m oo oo^r-r^ ^^^o
ei oo v-i
,c O *
rn N rs( in in 3 rN ci I I I
N * N 09 iQ ofl c-
ro W ci pi Pi cr M W
2 S 2 ft S i i i
s8 os sq N os *; ^^
Rod ffj * o-j c-i m ,»
Pi ci ci m ci m ci
O- U-, — »C pi « o R
SIS "82S SK S 8 S 8 H 2 5! R i , i
t^ os x P r* ci ^r in "Co
8 8 s r sSn s s
4 i5 * « sso -r xs ir, •*»":*"iff; I
R S S K K 8 8 8 R
§ § a s a s ss a s s s 2 r r a s ,
,O 00 ■so
a
4 s
s as a R a a a
.-« m <o K 9 in Pi S S h
J! (S OS CN rM
2 8 a a a 8 8
r* rN Os w-i © 3s ci ©
ss g I SRpg
a a a a nun
? ? ;; ; c ? a 8 8 00 ,
& $ a A 3 S 2 I (5 2 ,
assa a a s5 a 8s a
ssa? ?3Rsasi
m »
o o —
©s t>
X) x>
i 1 3
m
3 «
2 «
M
so § 3
vl-i
Xi
q O oo ,s* s
2 2
1 1,s» sO VO s 2
i
2 5 3 §
sD 00 5s g i a 2 § SO g 5 S3 1
*
m 2 * 1 3SO « s*
i8 w ,s Xi 3 X> fi. fi " fi. - CsJ ,O
O
oo *! * SO 00
o o -
I1
ci m
3 S g B
o s § § B s SO R g S iO g p '
!
« X>
in Os e 5 X)
oo
so
00 00 00
a
Xi
Jj *
OO 00 r-' r-' ad 00 *I >
§ S 2 § 8 I a o § 3 3 H 00 8oo M s s w a rs| — i
a§ 1
Xi
jl t g i 1
MJ
so
t t t t t t' 0
§ 1 g 3 § 1 s 5 IS oo5 §13 s 8 S & a s s 5 Iso
8 8B 8 ? | i ^ jj SP
"n mo! csjci n pi m n m
I IS H S S 3 2 g
3g Mia I
rsi rsi ri psi
H § 8
curnm n n Orsinrj mmrMrjcimc-i
CsJ ■-
rsi csirsi fsirsi trs rsicsi H ^
in m
jlijlf— 1||}
5 g 3 I 3 i g S I r a
rt N pi n ci i»i fi N i»i ci (si
X> XI XI Xi
a gj a a aalSSIa
psirsifsipj m n - ri rsi (si pj
»P2K SS38P32
alll j 1 1 1 II I i I
84
I1
1
?
a
t
& I
N Oj N N * N O}
x £ a s a s a
~ -* m — mm, , | | Iji
Pjgsaaassssss
•^ ^ m ^ (n fi fi r^ ^ *4 (^ ^f
pi -i m ■* m m os
O 00
_i _-, i-, .- n VI rr. r-, O c", (N IN
M5sss;s?i'a!j
m <n m o> cn
s § p s a s p $ a
m«fimmmosmm«o
(j", oo "«* — m v, t C v, X po, ■C
pi^pi«mmo>pisri
| | g 3 R S 2
oor^rNoo^m<»^Ospi*-•C>
s n s a - -
rn^-rn«m«n^rn«
— m « m m »
- S
f s. x I ^ I■3 V) iS, a. CG >"
£ M M M H *
o B e a b a cO fi O ■ H
£ 8 1 £ 1 J«i * a. h « >■
s
5
85
1
1
8
r> «- m ei FN so o o
X K 5 9! m s s
m ■* 9 r^ 00 r- a■* CN
CN 5 6 t^ m Tr P■
SO a (N s 8 m s Spi
in r*i
a s
a a
a
iN ■■*
g 3
ca X■* ""}
cn (^
1 8
a
s a
Pi iN
3 S
in ifi r- «
2 3 8 5
IS X X Xfi fi fi fi
m ro 00 "•*
(s m cN *
* **1 *
p\ n pi
3 7. 5
m oo m cn
r- oo "n so
in oo o m
s K Pi s
§ g at §
s r £ £
m oo •-■ so
S cn r4 in
CO JO Xi Xi
os pi F in
T "0 * -C
vj i»- so a m « C) «
& 2- £i 59 ?B SC-»d'
pi n n A n ^ pi pi
"»■ a «n r-
R 3
m pi
§ 2
o ^t
cn
oo r^ cm B &> so
r^ o m
"O sO ^; l> ^H Os
so vi » j io in
m fi C*l r*i fi d
OO ^ « N
# W v n
pi pi i*i fi
so N so ■«■
2 5
cn m «-
N « N «
N N N S
a S3§
•n oo pi
s a
m pi Noxnso i> r> re m
? 8 S a S ?! R5SS
s " 2 a s a csias
pi © i-^ (s ©
« r- vi 4 m
pi pi fI (-n fi
SU"
N ^ 00
i> in •*
i> in
pi pi
rpr? sn
N m *>
*-i © **i
m S5 m
& sg £
Xi XI
ca ca ca
i> in m
3 N (S R a s
ssass?
Xi Xi
m o
a a s
v * ^
s s s
SSSf 8
CO CS Xi
e> cn so
r^ r^ vi
« oc P ft
« iri 2 w
s § s a ssss HSs 3 § i 3 g g s
8
(SO
a £ £ £r» oo ■* rJ
6 h N N
u ■g ■a
a & 4 s
N « f. 6
'8
j?
3"«
(N N
co X
so in
in in
r- oe
pi pi
5 s a a
S c> A pi
in m so so
I S i 2
m pi < *
a a s 5
m xi Xo ,p
in in 5 t
rsi cN m' rsi
S 3 § 8 I a s
5s O
S §
£ (S
CO XI
m mr- •*
c*i rsi
= s s s?
sd m sbm"
Po m cn c>
m m m cn
» * n r^
m' ri ^ «
a a
Xi xl
g cs a
so m Q
56 o So
so so in
in so in
22a
oo f O
-H OS oo
^ cri c*i
sr5 3 § a
S R 8 S
N N n N pi cn rsi
1 1 M
vi * m vi
a 3 S s
" s a a
fi fi i*i pi
3 S I p
3 S P S
N N N N
§ i
f) 0O (*i Os2 s p a m o $ S S ? 8 S S S S
5
xi
H
11=8 1 -3 I n w n *• |tl Jlill |iSig|ii
86
Figure 45 . Generalised growth curves from birth to maturity for goats and sheep in central Mali.
Liveweight (kg)
30-
 
24 36
Age (months)
60
Figure 46. Relationship between year of birth and sub
sequent growth to 3 years of age in central
Malian goats.
Liveweight
(kg)
Year of birth
» « 1978
30-
20-
10-
 
Age (years)
BREEDING FEMALE
POSTPARTUM WEIGHTS
The observed means (x ± s.d.) for this character
were 26.0 ± 5.16 (n = 1729) for goats and 30.0 ±
5.33 (n = 1536) for sheep.
The mean squares from the analysis of var
iance are set out in Table 46. The weight of goats
was significantly affected by the system in which
they were reared, the parity (age) of the animal,
the type of parturition and the season and year
during which the doe gave birth. The effects of
flocks within systems were also highly significant.
There were also significant effects due to the in
teractions of season with the system, with parity,
parturition type and year. In sheep, postpartum
weight was affected by the same sources of varia
tion except season and year of parturition, and, of
the interactions, only that of season x system was
significant.
The estimated least-squares means for post
partum weights are given in Table 47 and graphic
representations of the main variables affecting
them are shown in Figure 47. Weights were higher
for both goats and sheep in the irrigated rice sub
system than in the rainfed millet subsystem, and
in both species females that had given birth to
twins (or triplets) were heavier than those that
87
 Table 45. Phenotypic correlations between weights at various agesfor goats and sheep in central Mali.
Age
(days)
Age (days)
30 90 150 240 365 550 730
Goats
10
30
90
150
240
365
550
Sheep
10
30
90
150
240
365
550
0.89
0.84
0.67 0.56 0.45 0.33 0.32 0.36
0.73 0.62 0.49 0.38 0.42 0.40
0.84 0.70 0.52 0.62 0.55
0.79 0.52 0.62 0.55
0.67 0.56
0.62
0.58
0.73
0.72
0.65 0.61 0.68 0.64 0.52 0.55
0.82 0.78 0.74 0.70 0.64 0.66
0.88 0.77 0.69 0.70 0.61
0.79 0.63 0.62 0.53
0.83 0.75
0.80
0.68
0.74
0.77
had given birth to singles. Primiparous females in
both species were lighter than multiparous ones,
the progression being linear from first to fourth
and greater parities. Goats at first kidding (82.2%
of the overall mean) were relatively lighter than
Table 46. Analysis ofvariance ofpostpartum weights of
goats and sheep in central Mali.
Source of
Goats Sheep
variation d.f. MS d.f. MS
System 1 755.9*" 1 1 350.3***
Sex 1 0.5 1 55.6
Parity 3 5 257.3*** 3 2 852.8***
Type of
parturition 1 154.3*** 1 179.6***
Season 3 58.8** 3 22.0
Year 6 43.6*** 6 5.7
Flock/millet 13 312.6*** 14 411.2***
Flock/rice 5 177.4*** 16 170.5**
Season x System 3 87.8*** 3 60.4**
Season x Sex 3 5.0 3 20.9
Season x Parity 9 22.0* 9 18.2
Season x Type of
parturition 3 40.2* 3 20.9
Season x Year 16 23.0* 16 15.8
Error 1 661 11.5 1 456 15.9
***P<0.001; **P<0.01; *P<0.05.
sheep at first lambing (87.9% of the overall mean)
but both species achieved the overall mean weight
by the time of second parturition.
The effects of season were rather different on
the different species, goats being lightest in the
cold dry season and then gaining weight in the hot
dry season while sheep were lightest after a par
turition during the hot dry season. The effects of
year, although not significant in sheep, had simi
lar trends in both species but the overall changes
over the period of study were not marked.
EFFECTS OF CLIMATE ON SPECIFIC
AGE AND SEX GROUPS
Seasonal weight changes of selected age and sex
groups of both species in both subsystems ob
served in relation to the overall mean weight-
for-age are shown in Figure 48. Maximum and
minimum weights in relation to the overall mean
for all age and sex groups are given in Table 48.
The magnitude of the seasonal effects was much
less than that on cattle (as might also be inferred
from a comparison of Tables 23 and 24 with
Tables 41 to 44). There were no major response
differences due to seasonal effects between sub
systems or between species.
In general, minimum weights were observed
for each sex, age class and species of stock in the
millet subsystem in June and maximum weights in
the post-rains period in October and November,
although there was more variation in the timing of
maximum weight than there was in the timing of
least weight.
Table 47. Least-squares means forpostpartum weights (kg) ofgoats and sheep in central Mali.
Variable
Goats
SE
Sheep
SE
Overall
System
Millet
Rice
Sex
Female
Male
Parity
1
2
3
>4
Type ofparturition
Single
Multiple
Season
Cold dry
Hot dry
Rains
Post-rains
Year
1978
1979
1980
1982
1983
1984
Flock/millet
Best
Worst
Flock/rice
Best
Worst
1729 25.8 0.18 1536 31.5 0.35
1348 24.7a 0.17 1 059 29.8a 0.37
381 26.9b 0.27 477 33.1b 0.42
809 25.8 0.20 770 31.3 0.38
920 25.8 0.19 766 31.7 0.36
422 21.2a 0.24 381 27.7a 0.40
291 25.3b 0.27 293 31.5b 0.42
207 27.2c 0.28 218 32.8c 0.43
809 29.4d 0.18 644 33.9d 0.36
1406 25.3a 0.17 1468 30.4a 0.19
323 26.2b 0.25 68 32.5b 0.65
435 24.9a 0.29 436 31.0 0.51
448 25.7a 0.36 415 30.9 0.43
311 26.8b 0.29 357 32.9 0.94
535 25.6a 0.36 328 31.1 0.69
108 24.8d 0.49 153 31.4 0.54
150 26.4bcd 0.35 161 31.8 0.47
389 26.5b 0.22 293 31.5 0.41
308 25.9ac 0.24 262 31.5 0.44
307 25.6ad 0.26 274 31.3 0.43
126 25.3ad 0.36 115 31.4 0.54
100 27.8 0.37 97 33.6 0.53
37 20.7 0.58 67 23.8 0.61
36 29.7 0.58 4 42.8 2.06
72 24.6 0.43 14 27.8 1.13
Within variable groups, means followed on different letters differ significantly (P<0.05). Variable groups without any letters did not
show a significant difference in the analysis of variance.
In the rice subsystem there was slightly more
variability in the timing of both the minimum and
maximum weights. As in the millet subsystem,
female goats attained maximum weights in the
post-rains period, but sheep tended to be heaviest
in the late cold dry and early hot dry periods, this
probably being related to the availability of
weedy regrowth in the harvested rice fields at that
time.
Resulting in large part from the low effects of
season on weight, the growth curves of goats and
sheep do not exhibit the same marked type of
saw-tooth pattern as that observed in cattle. Ex
amples of actual growth curves for female sheep
in the millet subsystem are given in Figure 49.
Long-term weight changes over the study
period are shown in Figure 50 for mature female
sheep and goats in both subsystems. Table 49 pro
vides information on the regression coefficients of
these long-term weight changes. Although, as
with cattle, the general trend in mature weight has
been downwards, the monthly and total losses
have been much less spectacular. Animals in the
rice subsystem suffered less than those in the mil-
89
Figure 47. Relationships between postpartum weights
and parity, season and year of birth ofsmall
ruminants in central Mali.
Liveweight
(Kg)
35
 
30
25-
D J F M A M J J A S 0 N
Cold dry Hot dry Rains Post-rains
Season
30
25
,« — — « <
78 ' 79 ' 80 ' 81 '82 '83 '84 '85 '
Year
let subsystem, sheep weighing 5.1% less at the
end of 1983 than they did at the beginning of 1978
while in the millet subsystem sheep weighed 5.8%
less. Goats in the rice subsystem actually weighed
2.5% more at the end of the study than at the be
ginning, but goats in the millet subsystem suffered
most of all and were 9.9% lighter at the end of
1983 compared with the beginning of 1978.
DISCUSSION
The growth rates observed during this study were
within the range reported for other African goats
and sheep of similar mature size and are not indi
cative of any serious nutritional constraints when
compared to the growth rates of cattle in the same
environment. The gain in goats to 1 year after
weaning was at 85% of the preweaning gain and in
sheep at 75%: these figures compare with a post-
weaning gain in cattle of only 65% of the pre-
weaning gain. The effects of season on postwean-
ing gain were not significant in sheep and, although
significant in goats between 8 months and 1 year
of age, these were not marked.
The effects of year were generally significant
on weights-at-age and these can in general terms
be related to the decline in rainfall and the re
duced availability of total feed resources. The
lighter weights at relatively mature ages (2 and
3 years) of animals born later in the study are what
would be expected from the decline in weight noted
for specific ages and sexes of animals in Figure 50
and Table 49. The inflection point in the growth
curve occurred generally at about 12 months of
age but was less marked for goats than for sheep.
Postpartum weights followed the pattern
that could be expected in relation to the age of the
dam. Postpartum weights in relation to the season
of birth differed in pattern between goats and
sheep. While maximum postpartum weights were
observed in both species after a birth in the rainy
season, lowest weights were noted for goats dur
ing the cold dry season and for sheep during the
hot dry season. Goats presumably benefitted nu
tritionally from leaf development by some browse
species as humidity increased during March and
April. Postpartum weights as affected by season
did not show the same trend as for cattle. For both
goats and sheep they remained fairly constant
about the mean over the 6-year study period.
The relative lack of seasonal effects on
growth was also reflected in the intra annual
weight changes in specific age and sex groups for
both goats and sheep. Although some weight
losses occurred in all age groups due to seasonal
effects (Table 48 and Figure 48), these were much
less than in cattle. In general terms weights did
not deviate from the mean by more than 6% in
either direction for each species, sex and age
group and in both subsystems. This should be
compared with deviations of about 15% for
female cattle and up to or more than 20% for
work oxen. In practice, the effects of being reared
in the rice subsystem appeared to be to delay the
loss of weight in the cold dry season but there was
a much less marked increase in growth rate in
goats and sheep than in cattle in this subsystem
due to the availability of weed growth after the
rice harvest (March-April). Resulting from the
less severe seasonal effects, growth to maturity
proceeded with less fluctuations for goats and
sheep than for cattle (Figures 48 and 35 respec
tively), and this applied to whatever season of
year that young were born.
The long-term decline in mature weight was
also much less serious in goats and sheep (Figure
50) than in cattle (Figure 36). Losses on a com-
90
Figure 48. Seasonal variations in weight ofcorresponding age and sex groups ofsmall ruminants in central Mali.
Liveweight (kg)
43i
42
41-
40-
39
38-|
3 7
36
35-
34
33-
32
31-
30
29-
28-
27o
26
25
24
23
 
b.Goots
Irrigated rice subsystem
x x Females (4 pairs)
x x Males (I pair)
Rainfed millet subsystem
• • Females (4 pairs)
• • Males (I pair)
 
DJFMAMJ JA SO N D J F M A M J A S 0 N
Month
Cold dry Hot dry Rains Post-rains Cold dry Hot dry Rains
Post-
rains
22
Season
parative basis were in general less than half those supply are clearly evident in these observations
suffered by cattle. Both species lost less weight in on weight. The results also show that in the
the rice than in the millet subsystem. medium term, the forage supply has deteriorated
The better adaptation of both goats and in comparative terms more drastically for cattle
sheep to the harsh conditions of the semi-arid en- than for small ruminants,
vironment and to the generally fluctuating feed
91
Table 48. Weight changes by sex and age in central Malian goats and sheep, expressed as percentages ofmean annual
weight.
Species, sex
and age*
Millet subsystem Rice subsystem
Goats
nb xc
Weight change (%)
min max
nb xc
Weight change (%)
min max
Males: 1 642 25.1 93.0 106.5 167 24.7 91.1 107.9
2 380 30.3 95.6 104.6 51 30.4 93.5 108.3
3 309 34.7 87.6 110.5 49 29.8 94.1 104.9
4 385 40.9 96.8 104.9 14 30.8 90.9 107.1
Females 1 1342 20.4 91.8 105.0 218 20.9 94.1 109.3
2 1026 22.5 91.9 106.3 145 23.3 95.4 109.2
3 1300 24.9 96.2 103.1 179 26.0 96.7 105.4
Sheep
4 4 257 28.0 94.0 104.3 789 28.8 97.5 107.8
Males: 1 574 34.8 92.5 105.7 199 40.7 89.6 104.7
2 221 37.7 96.6 107.3 62 45.1 91.2 113.2
3 147 43.8 92.4 110.5 10 51.2 95.7 111.1
4 442 45.9 96.4 104.4 - - - -
Females: 1 1 159 25.7 91.7 105.3 566 29.6 94.6 102.6
2 908 27.8 96.2 107.1 428 31.4 96.1 104.1
3 1090 29.7 94.0 103.8 461 33.1 94.5 104.0
4 3088 31.2 94.2 104.4 835 33.1 93.0 104.9
* Age given as pairs of permanent incisors.
b Number of observations in each class of stock .
' Mean annual weight in kg.
92
rD
I
-a
c
5
I
 
Oct
s3s r
B • r
-co
.-» to
ao
0)
-5
I
If)
1 I 1
O
1
o in IT)
CM CM
-3 CD
OJ
GO
-a
-j CO "£
0) o
- 2
Q
CO
0)
CO a,
S
CD
93
o
o
 
1—
c
I
i 1 1 1 1 1 1 1 1—i r
,5
1
|I
5
3 5
 
3.
E
I r-
> CO
-1 fO
O
CO
-2
00
1-2
o
CO
Q
ao
"5 —
I
-1-
If)
(O
to
IO
-I 1 1-
CO — o
10 B W (M
CO N
CM CO
10
tNJ
O
5
CO TJ-
00
CD
coro
00
CD
5
Q
co co
00
_ 0->
CO _
00
CD
o>
-2
~I—if) ■*
CO CO
94
Table 49 . Regressions of weight (kg) on time"for maturefemale sheep and goats and ofrainfall (mm) on time, central
Mali.
Subsystem/
species
Number of
observations a b r P
Rice
Sheep 657 -0.024 33.6 -0.08% 0.0216
Goats 789 0.010 28.4 0.0331 0.3533
Millet
Sheep 1727" -0.026 32.1 -0.1117 0.0000
Goats 1727b -0.041 29.8 -0.1714 0.0000
Rainfall" 7 -33.4 452.1 -0.7871 0.0450
* Monthly change; intercept at December 1979.
b Random sample selected from 3 088 observations for sheep and 4 257 observations for goats.
c Intercept at January 1977 and values for years not months.
95
 13. MORTALITY AND OFFTAKE
ABORTIONS
Abortions in goats totalled 12.6% of all births
while in sheep the corresponding figure was
5.1%.
PREWEANING MORTALITY
The unadjusted mean preweaning mortality (to
150 days of age) was 34.4% in goats and 23.4% in
sheep. In goats about 30% of all preweaning
deaths (other than abortions) were stillbirths or
occurred on the first day of life and a further 15%
of kids died during the first week. In sheep some
17% of preweaning deaths were either stillbirths
or first-day mortalities and a further 10% of
deaths took place during the first 7 days. In goats,
66% of all deaths recorded during the study oc
curred before weaning while in sheep only 46% of
deaths took place during that period.
The mean squares from the analysis of
variance for preweaning mortality are given in
Table 50. The main sources of variation influenc
ing this parameter were the system, the season of
birth, parity and flock in the millet subsystem. In
goats, the year of birth had a highly significant ef
fect and the sex of the kid also influenced the level
of mortality. In sheep, the type of birth exerted a
significant effect on the death rate.
Least-squares estimates of mortality are laid
out in Table 51. For both goats and sheep there
were more deaths in the millet than in the rice
subsystem. In both species, young born in the
cold dry season had a greater chance of dying than
those born at other times of the year, the next
highest mortality rate being suffered by young
born during the hot dry season. Lowest death risk
was met by young born during the rainy and post-
rains seasons. In goats there was an almost linear
reduction in the mortality rate from 1978 to 1983.
The trend for sheep was similar although the ef
fect of year of birth on mortality in this species
was not significant. More young of primiparous
females died than of multiparous dams although
Table 50. Analysis of variance ofpreweaning mortality in goats and sheep in central Mali.
Source of
variation
Goats Sheep
d.f. MS d.f. MS
System 1 1.04* 1 0.95*
Season of birth 3 0.72* 3 0.52*
Year of birth 5 2.46*** 5 0.27
Parity 4 3.15*** 4 1.35***
Sex 1 0.85* 1 0.00
Type of birth 1 0.41 1 1.06*
Flock/millet 13 0.91*** 14 0.58***
Flock/rice 5 0.30 16 0.27
Error 2 001 0.21 1517 0.17
' P<0.001; *P<0.05.
%
Table 51 . Least-squares means for preweaning mortal
ity (%) in goats and sheep in central Mali.
Goats
Variable
Sheep
Overall
System
Millet
Rice
Season
Cold dry
Hot dry
Rains
Post-rains
Year
1978
1979
1980
1981
1982
1983
Parity
1
2
3
4
>5
Sex
Female
Male
Type ofbirth
Single
Multiple
2 035 35.0
1 603 38.6a
432 31.3b
378 39.5a
651 36.9ab
369 32.5bc
637 30.9c
1 563 28.0
1 067 32.1a
496 23.9b
322 32.9a
518 28.1ab
376 23.0b
347 27.9ab
158
233
422
456
411
355
51.5a
35.6bc
33.6bc
38.7b
31.2c
19.1d
178
189
321
291
285
299
34.4
27.8
27.5
28.0
27.6
22.6
429 47.9a
307 38.8b
258 31.4bc
168 30.7bc
873 26.1c
387 38.6a
299 22.7b
222 24.2b
150 28.3b
512 26.2b
998 32.9a 800 27.9
1037 37.0b 763 28.1
1 375 33.3
660 36.6
1 432 23.0a
131 33.0b
Within variable groups, means followed by different letters dif
fer significantly (P<0.05). Variable groups without any letters
did not show a significant difference in the analysis of variance.
only in goats was this reduction significant be
tween the second and older parities. Sex and type
of birth resulted in the expected trend in mortal
ity, with more males than females and more
offspring of multiple than of single births dying.
MORTALITY AFTER WEANING
Mortality in the period between weaning and
1 year of age amounted to only about 5% (i.e. al
most 10% on a yearly basis) in both goats and
sheep. The death rate after 1 year was higher in
sheep than in goats, such that at 4 years of age
only 30% of sheep born were still alive while 38%
of goats survived to 4 years. For the calculation of
these last two figures animals removed for slaughter
or sale were obviously excluded from the analysis.
The mean annual mortality rate over 1 year
of age was about 12.7% in goats, being somewhat
higher in sheep. The trend observed in prewean
ing mortality rate was maintained in adult life: a
greater percentage of animals died in the millet
than in the rice subsystem and the death rate was
higher in males than in females.
Slightly more than 17% of goat deaths were
due to animals being 'lost' or taken by predators -
mainly dogs - and almost 10% of deaths in sheep
were from these causes.
OFFTAKE
As for cattle, offtake was calculated on an annual
basis as sales plus slaughter (for social purposes or
in extremis) plus animals gifted out permanently.
The data by species and system are provided in
Table 52.
Table 52 . Total annual offtake by system, sex ofanimals
and species.
System and sex
Offtake (%)
Millet subsystem
Females
Males
Rice subsystem
Females
Males
Overall
Goats Sheep
13.8 23.5
9.0 14.7
38.7 45.5
26.6 34.6
14.8 15.3
53.1 78.7
19.3 26.8
For female goats approximately 30% of off
take was in the form of slaughter, about 8% in the
form of gifts and 62% was through sales. For male
goats about 39% of the animals were slaughtered,
6% were gifted and 55% were sold. Male goats
were generally slaughtered before 1 year of age
(63.5% of all slaughterings) while a slightly lesser
percentage (52.0) was sold before 1 year of age.
Females were slaughtered (39.8% before 1 year
old) and sold (32.5% before 1 year old) later than
males. The overall average age at offtake was
about 14 months, being about 15 days earlier in
the rice than in the millet subsystem and about
100 days earlier for males than for females al
though castrated animals were kept on to older
ages than entire males. Average weight at offtake
was 20.7 kg, entire males weighing about 1 kg
more than females with castrates being 6 and 8 kg
heavier than males in the millet and rice subsystems
respectively.
In sheep about 25% of female offtake was in
the form of slaughter, only 2% as gifts and 73%
97
as sales. For male sheep about 29% were
slaughtered, 2% were gifted and 69% were sold.
A similar percentage of male sheep (58.4) to that
of male goats were slaughtered before 1 year old
but a higher percentage (68.2) was sold before
this age. As for goats, females were slaughtered
(40.0% before 1 year old) and sold (44.5% before
1 year old) later than males. The overall average
age at offtake was similar to that of goats at about
14 months but was very much later in the millet
subsystem (16 months) than in the rice subsystem
(12 months) and offtake of females in the former
subsystem averaged at about 22 months com
pared with only 13 months in the latter. Average
weight at offtake was 26.8 kg, being 24.8 kg in the
millet zone (females 23.0 kg, males 24.7 kg, cas
trates 43.1 kg) and 29.8 kg in the rice zone
(females 24.9 kg, males 31 .5 kg, castrates 38.4 kg).
98
14. PRODUCTIVITY
MILK PRODUCTION
Few data were obtained on this characteristic. For
Marina sheep under delta-type conditions at
Niono, a lactation curve was established by an ap
plication of the rectangle formula for evaluating
definite integrals. Data on milk yields were col
lected on average once per month. Lambs were
penned away from their dams for a 24-hour period
and allowed to suckle after 12 and 24 hours. They
were weighed immediately before and im
mediately after suckling, and any remaining milk
was hand pulled. While this method will not allow
total milk production to be ascertained, it is prob
able that most is accounted for. All records relat
ing to 10-day periods from parturition (1 to 10, 11
to 20 etc.) were pooled and the mean figure was
assumed to be the yield at 5, 15 .... n days. Each
rectangle in the lactation curve thus covered a 10-
day period. The mean lactation curve and some
individual yields are shown in Figure 51. Total
lactation yield is of the order of 50 kg with lacta
tion length varying from 85 to 165 days. The lacta
tion was assumed to be completed on the day of
the visit when a ewe was dry although she may
well have stopped suckling up to 1 month pre
viously: lactation lengths are therefore probably
about 15 days less than shown in Figure 51 with a
true mean length in the region of 130 to 140 days.
WOOL PRODUCTION
Wool was shorn twice a year to give two periods
corresponding to 'wet' and 'dry' seasons. Wool
production over a 3-year period is shown in Table
53. Annual yields of greasy wool at 684 g are
equivalent to growth of 1.87 g/day. As might be
expected, daily wool production in the dry season
(1 .70 g) was very much less than in the wet season
(2.22 g). Males and castrates yielded more wool
than females although female yields were closer
to those of males in the wet season when energy
and protein requirements are obviously in excess
of those needed for maintenance, pregnancy and
lactation.
PRODUCTIVITY INDICES
Productivity indices were constructed in the same
manner as for cattle, except that weight of the lit
ter at presumed weaning at 150 days was used as
the measure of production.
The mean squares from the analysis of var
iance are shown in Table 54 for both goats and
sheep. In general the sources of variation which
exerted significant effects on the indices were the
same for both species and had similar levels of sig
nificance. The sex of the young did not affect the
productivity of female goats. Year effects on
sheep were not significant and were relatively
weak on goats.
The least-squares means for the productivity
indices are laid out in Table 55. The rice subsys
tem had superior indices for both species. The ef
fects of season were such that goats giving birth in
the hot dry or rainy seasons were more productive
than those kidding in the post-rains season, while
births in the cold dry season resulted in lowest
productivity. In sheep, births in the rainy season
were superior in terms of productivity to those in
the hot dry season which in turn were superior to
both post-rains and cold-dry season births.
Productivity indices of first-parity ewes were
clearly inferior to the indices of all other parities
but in goats, this distinction was only evident on
Index I, and when the resource-based indices II
and HI were calculated, the productivity of
primiparous does did not differ significantly from
that of second-parity females and most of those
from the sixth parity upwards.
Ewes giving birth to males and twins had bet
ter indices than those giving birth to females and
singles. There were no differences in goat produc-
99
Figure 5 1 . Mean lactation curve and representative yields for Macina sheep.
Milk yield
(gxl02/day)
12
Mean lactation curve
Representative individual yields
10-
 
130 150
Days since lambing
Table 53. Mean woolproduction" by Macina sheep.
Sex and variables
Dry season Wet season Year
x ± s.d. x ± s.d. x ± s.d.
Males + castrates
No. in sample
Total yield (g)
Growth per day (g)
Females
No. in sample
Total yield (g)
Growth per day (g)
Overall mean
Total yield (g)
Growth per day (g)
Interval between
shearings (days)b
11.3 ± 4.51
468.3 ± 72.86
2.20 ± 0.412
27.3 ± 7.02
263.2 ±73.54
1.21 ± 0.107
365.7 ±67.86
1.70 ± 0.200
16.3 ± 6.11
377.4 ± 33.68
2.53+ 0.300
28.0 ± 5.29
281.1 ± 25.72
1.92 ± 0.932
392.2 ± 5.19
2.22 ± 0.393
216 43 151 27
13.8 ± 5.30
835.7 ± 52.54
2.29 ± 0.245
27.7 ± 5.69
534.1 ± 64.98
1.46 ± 0.181
684.9 ± 42.83
1.87 ± 0.119
365
* Based on 3 years' data.
b When shearing intervals totalled more or less than 1 year, an equal number of days was added to or subtracted from each season.
100
Table 54 . Analysis of variance ofproductivity indicesfor goats and sheep in central Mali.
Source of
variation
Goats Sheep
d.f. Index I Index II Index III d.f. Index I Index II Index III
System 1 2 619*** 2.37*** 15.82*** 1 6 050*** 4.98*** 32.66***
Season 3 759*** 1.33*** 7.21*** 3 1660*** 2.35*** 13.84***
Year 5 211* 0.28* 1.63* 5 229 0.18 1.30
Parity 8 1 083*** 0.55*** 4.51*** 7 1 848*** 0.90*** 7.15*"
Sex 1 4 0.00 0.00 1 2 304*** 1.89** 12.86***
Type of birth 2 4 308*** 4.22*** 27.82*** 1 5 208*** 3.42*** 24.74***
Flock/millet 13 351*** 0.36*** 2.21*** 13 362** 0.44** 2.54**
Flock/rice 5 1 156*** 1.16*** 7.42*** 7 849*** 0.59** 4.05***
Error 1 152 93 0.13 0.73 934 163 0.17 1.08
*P<0.001; "P<0.01; *P<0.05.
tivity arising from sex of young, this being most
probably due to the relative lack of precocious
sexual dimorphism in goats at weaning. Although
twin-bearing does had indices superior to those
giving birth to singles, the productivity of dams
giving birth to triplets was inferior to those having
twins and equal only to those giving birth to
singles: this was due largely to the high death rate
in triplets.
The variables acting on the productivity indi
ces (litter weight at 150 days, mortality, parturi
tion interval and dam postpartum weight) were
correlated to the productivity indices themselves
in order that comparative advantages could be
calculated and improvement pathways designed.
Correlations between weaning weight and
the three indices were the highest of all the char
acters considered (about 0.90 for both goats and
sheep). Higher weaning weights would obviously
lead to increased output as reflected by the indi
ces. The figures presented are weights weaned
per female, and for individual surviving young
these are biased downwards by the inclusion of
the zero weights of their dead counterparts. Ac
tual weaning weights of surviving animals were in
the region of 18.2 kg for kids and 24.7 for lambs.
Weights of young weaned per female would
therefore be increased not only by acting directly
on this character but also by attempting improve
ment through a reduction in mortality.
Correlations between viability and the indi
ces were high (0.72) and significant (P<0.001).
Reducing mortality would greatly improve the in
dices not only directly but indirectly through
higher weaning weights. Although the viability
rates in this study appear to be low, they are com
parable to those encountered in many other tradi
tionally managed flocks throughout Africa (Wilson
et al, 1984; 1985). Clear identification of the
causes of mortality in small ruminants and reduc
tion or elimination of these would enable in
creased output to be achieved over much of
Africa.
There were highly significant (P<0.001)
negative correlations between parturition interval
and all three indices for both goats (r = —0.26)
and sheep (r = —0.36). A shortened parturition
would at first sight, therefore, appear to offer pos
sibilities for improving total productivity. It has
however been shown (Wilson et al, 1983) that in
this environment intervals of less than 240 days
result in a much higher mortality than intervals
longer than this. Attempting to reduce the par
turition interval in sheep would, then, increase
productivity very little but there appears to be
scope for reducing that of goats.
Correlations between postpartum weight
and all three indices were positive for both goats
(weight/index I = 0.36, II = 0.18, III = 0.24)
and sheep (0.37, 0.14 and 0.20 respectively) and
significant (P<0.01). Increasing breeding female
weight either permanently through genetic im
provement leading to greater body size or tempo
rarily by means of strategic feeding would lead to
higher indices. The correlations are the lowest of
all the characters considered and more rapid im
provement may arise from concentrating on
improving other characters first.
Appropriate improvement paths for goats
and sheep in Mali
The effects due to flock - considered here to be
the basic management unit - have not yet been
discussed. There were usually highly significant
101
 Table 55 . Least-squares means ofproductivity indices for goats and sheep in central Mali.
Variable
Goats Sheep
Index I Index II Index III Index I Index II Index III
n
(kg) (g) (kg)
n
(kg) (kg) (kg)
Overall 1 191 14.6 494 1.23 973 28.4 867 2.22
System
Millet 934 12.1a 419a 1.04a 672 24.7a 761a 1.95a
Rice 257 17.1b 569b 1.42b 301 32.1b 973b 2.01a
Season
Cold dry 270 12.3a 406a 1.02a 245 26.2a 784a 2.01a
Hot dry 309 15.8b 560b 1.37b 267 28.9b 904b 2.30b
Rains 246 15.9b 540b 1.34b 239 32.1c 997c 2.54c
Post-rains 366 14.2c 470c 1.18c 222 26.4a 785a 2.02a
Year
1978 77 11.5a 392a 0.98a 95 27.9 849 2.17
1979 141 15.8b 536b 1.33b 121 29.7 900 2.30
1980 252 15.0b 500b 1.25b 204 27.4 845 2.15
1981 283 14.6b 484b 1.21b 205 29.9 911 2.34
1982 252 14.7b 506b 1.25b 203 27.2 830 2.12
1983 186 15.8b 546b 1.35b 145 28.1 870 2.23
Parity
'0' 357 15.5ad 522ad 1.30ac 149 30.0ab 900a 2.32a
1 274 9.3b 398b 0.93b 253 21.0c 713b 1.78b
2 215 11.3c 430bc 1.04b 194 27.7a 878a 2.23a
3 151 14.0ae 495acd 1.22a 146 30.6b 939a 2.40a
4 % 15.8ad 552d 1.36c 92 29.6ab 891a 2.30a
5 54 17.4d 581d 1.46c 57 29.7ab 857a 2.21a
6 21 17.8de S48abd 1.40abc 34 29.8ab 889a 2.27a
r 14 20.4d 633d 1.61c 48 29. lab 872a 2.23a
>8 9 9.6abc 289ab 0.74ab - - - -
Sex
Female 601 14.5 495 1.23 514 26.8a 823a 2.10a
Male 590 14.6 494 1.23 459 29.9b 912b 2.33b
Type ofbirth
Single 970 12.9a 451a 1.11a 930 22.2a 709a 1.79a
Twin 214 20.4b 685b 1.71b 43 34.6b 1 026b 2.64b
Triplet 7 10.4a 345a 0.86a - - - -
* '0' are unknown parities considered £4; for sheep parity 7 is all parities 27.
Within variable groups, means followed by different letters differ significantly (P<0.05). Variable groups without any letters did not
show a significant difference in the analysis of variance.
differences among flocks for all the characters
and for the three indices. The causes of these dif
ferences have not been determined although it is
probable that they are related to the owner's or
herder's management abilities and strategies or to
his preference for one or the other species. The
greatest differences in productivity are, in fact,
encountered in this source of variation. If the dif
ferences are indeed related to management and
the practices relating to improved productivity
can be indentified and extended to other manage
ment units, overall productivity might be in
creased considerably. Such an improvement
strategy would be appropriate to a majority of
owners as it is already being practised by some
with similar means of access to virtually the same
resources.
The incorporation of other measures, de
duced from the analyses discussed here, into an
overall plan would provide an effective, inte
grated and cost-efficient first step in improving
total productivity from small ruminants. Some
102
comparative advantages accruing to Index II for
within-variable differences are shown in Table 56.
By inserting these, along with some other mea
sures, into a sequential improvement scheme in
roughly ranked order, a practical, technologically
adapted programme for development can be de
signed.
As can be seen from Table 56, apart from
overall management practices, the best results
will most probably come from selecting for twin
ning and by attempting to manipulate demog
raphic structure so that as many females aged 4 to
5 years as possible are in the flocks. Creating
some of the conditions of the rice subsystem in the
rainfed millet subsystem would also improve pro
ductivity. Controlling the breeding season so that
young are dropped during the hot and dry or rainy
seasons also confers considerable advantages but
these would have to be carefully assessed in rela
tion to the potential loss of productivity if births
were to be limited to only one per year. The two
factors over which least control can be exerted,
sex of young and year, fortunately do not greatly
influence productivity.
Table 56. Ratios of comparative advantages for vari
ables on Index II, to be used in designing im
provementpathways.
Variable Goats Sheep
Flock
Best to worst millet 1.83 1.55
Best to worst rice 2.97 1.73
System
Rice to millet 1.35 1.29
Season
Hot dry/rains to cold dry 1.45 1.28
Year
Best to worst 1.20 1.09
Parity
Sixth to first 1.45 1.29
Type of birth
Twin to single 1.54 1.44
Sex
Male to female 1.00 1.10
103

PART FOUR
CONCLUSIONS AND RECOMMENDATIONS

15. CONCLUSIONS AND RECOMMENDATIONS
THE NATURE OF TRADITIONAL
LIVESTOCK PRODUCTION SYSTEMS
Until recently traditional systems of livestock
production have not been subjected to detailed
study and analysis. Little, therefore, has been
known of their productivity and problems. But
this lack of knowledge has not prevented a strong
corpus of opinion being propounded.
Such opinion considers, regardless of the
domestic species in question, that:
• Indigenous African livestock types are of
inherently poor genetic make-up;
• Nutrition is inadequate;
• Disease is a major problem;
• There is a lack of a suitable marketing in
frastructure.
The proposed solutions to these constraints
are, equally, of a depressing similarity:
• Importation of supposedly genetically
superior stock to up-grade or replace the
native one;
• Provision of concentrate feedstuffs;
• Mass vaccination; and
• Creation of a modern marketing system
that is usually by inference - if not stated
explicitly - state-controlled and state-run.
These hypothetical problems and solutions
fail to take account of the nature and role of tra
ditional livestock husbandry in a mixed fanning
system. Domestic animals in these agropastoral
systems are at best an adjunct to subsistence or
cash crop production. Owners are able to devote
only a part of their labour to them and, across
households, the amount of time given and the
skill provided to stock management vary widely.
Sweeping solutions applied to all species are
not likely to succeed and will not make the best
use of resources, whether these be financial or
natural and whether they be internal or external.
THE CURRENT DATA SET
The studies analysed and reported here are
unique in semi-arid Africa. They provide data
over a period of 6 years from two agropastoral
subsystems and allow direct comparison of the
three major species of domestic ruminants under
the same environmental and management condi
tions. The results clearly demonstrate the
superiority of both species of small ruminants
over cattle in terms of meat production (Tables 31
and 55). The complementarity of the species is
evident in terms of their contribution to human
nutrition (Figure 39). The markedly seasonal na
ture of the reproductive process in cattle (Figure
29) and its inefficiency (Table 22) are highlighted
and contrasted with the much less seasonal (Fig
ure 44) and far more rapid (Figure 42, Table 38)
process in both goats and sheep. The dependence
of cattle on the limited grazing resources results,
in the short term, in massive fluctuations in
weight within a year (Figures 34 and 35) and over
the long term has resulted in an alarming reduc
tion in mature body size (Figure 36, Table 28).
Goats, and even sheep, with more eclectic dietary
habits, are less subject to both seasonal (Figure
48, Table 48) and long-term (Figure 50, Table 49)
weight changes. In contrast to the superior effi
ciency of small ruminants in most of the observed
parameters, their early death rate (Table 51) is
much higher than that of cattle (Figure 37).
These studies did not, unfortunately, con
tinue into the dry season of 1984/1985 which fol
lowed the disastrous rainfall years of 1983 and
1984. Little that is objective can therefore be said
about response to really severe drought condi
tions but there is considerable circumstantial evi
dence, obtained from visits to the herds and flocks
and through discussions with the owners, that
goats and sheep suffered less severely than cattle .
107
 MAJOR PROBLEMS TO IMPROVED
PRODUCTIVITY
The results reported here have enabled the iden
tification of the principal factors mitigating
against improved productivity. In cattle these
problems are:
• Poor reproductive performance as re
flected in late age at first calving and long
intervals between calves; and
• Nutritional stress, both seasonally and in
the long term, leading to massive fluctua
tions in weight in both growing and mature
cattle and overall slow growth rates which
contribute to delayed sexual maturity in
females and to advanced ages at which
males are capable of providing draught
power and being fit for slaughter.
In small ruminants the major problems are:
• High levels of preweaning mortality due to
a variety of interacting causes; and
• Continued relatively high levels of mortal
ity in sub-adult and mature stock due
mainly to a seasonally recurring complex
of respiratory diseases.
In both cattle and small ruminants, the dif
ferences in productivity between herds and flocks
with access to the theoretically same resources in
dicate that the individual management ability of
owners or herders could be a major constraint.
FUTURE RESEARCH
With the exception of veterinary inputs, solutions
to the identified constraints should be provided,
as far as is feasible, from within the existing sys
tem. With this proviso in mind, future research
should concentrate on:
• Identifying management and/or socio
economic factors leading to the observed
differences in flock productivity;
• Isolating the causes of the poor reproduc
tive performance by cattle, which are
probably nutritionally rather than physi
ologically or disease related;
• Overcoming the severe nutritional crisis in
cattle by encouraging the production of
fodder and forage crops, including browse
species, from the agricultural component
of the system, although in view of the
shortfall in total feed availablity it is prob
able that specific target groups will need to
be identified and accorded priority;
• Determining the specific causes (manage
ment, nutrition, health) of preweaning
mortality in lambs and kids; and
• Developing a package of prophylactic and
curative veterinary measures based on
local antigens to the respiratory disease
complex.
108
REFERENCES
Aboul-Naga A M, Aboul-Ela M B and Hassan F. 1985.
Oestrus activity of Suffolk, Mutton Merino and
their crosses with subtropical Ossimi sheep. /.
Agric. Sci. (Camb.) 104: 27-34.
Alim K A. 1964. Proc. 5th Int. Congr. Anim. Reprod.
and A.I. (Triento). Section 2(3): 128.
Ballo A. 1980. Etude des troupeaux bovins de la com
mune de Mopti. Mémoire de fin d'études. Institut
polytechnique rural, Katibougou, Mali.
Bertaudière S. 1979. Région de recherche d'Afrique
centrale. Rapport d'activité 1978. Institut d'Ele
vage et de Médecine Vétérinaire des Pays
Tropicaux, Maisons-Alfort, France.
Brown DL. 1959. The Nguni breed of cattle. A descrip
tive review. Emp. J. Exp. Agric. 27: 176-190.
Butterworth M H. 1983. Seasonality of calving amongst
unimproved cattle in Swaziland. Trop. Agric.
(Trin.) 60: 218-220.
CIPEA. 1981. Tendances et perspectives de l'agricul
ture et de l'élevage en Afrique tropicale. Document
de travail N° 5. CIPEA, Addis-Abeba, Ethiopie.
Clutton-Brock T H, Guiness F E and Albon S O. 1982.
Red deer: Behaviour and ecology oftwo sexes. Uni
versity of Chicago Press, Chicago.
Coulomb J. 1970. Projet de développement de l'élevage
dans la région de Mopti (République du Mali): étude
du troupeau. Institut d'Elevage et de Médecine Vé
térinaire des Pays Tropicaux, Maisons-Alfort,
France.
Coulomb J. 1971. Zone de modernisation pastorale du
Niger. Economie du troupeau. Institut d'Elevage
et de Médecine Vétérinaire des Pays Tropicaux,
Maisons-Alfort, France.
Coulomb J, Serres H and Tacher G. 1980. L'élevage en
pays sahélien. Agence de coopération culturelle et
technique et Conseil international de la langue fran
çaise. Paris, France, p. 27.
Delgado D. 1978. Livestock versus foodgrain produc
tion in southeastern Upper Volta: A resource allo
cation analysis. Centre for Research on Economie
Development, University of Michigan, Ann Arbor,
USA.
Diallo A and Wagenaar K T. 1983. La production de
lait. In: Wilson R T, de Leeuw P N and de Haan C,
Recherches sur les systèmes des zones arides du Mali-
résultats préliminaires. Research Report No. 5.
ILCA, Addis Ababa, Ethiopia.
Dixon W J and Brown M B. 1983. BMDPstatistical soft
ware. University of California Press, Berkeley,
USA.
Dumas R. 1978. Etude sur l'élevage des petits ruminants
du Tchad. Institut d'Elevage et de Médecine Vé
térinaire des Pays Tropicaux, Maisons-Alfort,
France.
Dumas R. 1980. Contribution à l'étude des petits rumi
nants du Tchad. Rev. Elev. Méd. Vét. Pays Trop.
33:215-233.
Eddy E. 1979. Landand labour use ofmixedfarms in the
pastoral zone of Niger. Centre for Research on
Economic Development, University of Michigan,
Ann Arbor, USA.
Fall A, Diop M, Sandford J, Wissocq Y J, Durkin J and
Trail J C M. 1982. Evaluation oftheproductivities of
Djallonke sheep and N'Dama cattle at the Centre de
Recherches Zootechniques, Kolda, Senegal. Re
search Report No. 3. ILCA, Addis Ababa, Ethiopia.
FAO. 1979. Production yearbook, Volume 33. FAO,
Rome, Italy.
FAO. 1981. Production yearbook, Volume 35. FAO,
Rome, Italy.
Fayinke B. 1980. Etude des troupeaux de la ville de
Niono: composition, nutrition, production.
Mémoire de fin d'études. Institut polytechnique
rural, Katibougou, Mali.
Fitzhugh H A and Bradford G E. 1983. Hair sheep of
West Africa and the Americas: A genetic resourcefor
the tropics. Westview Press, Boulder, Colorado,
USA.
Fofana B S. 1974. L'élevage bovin dans la boucle du
Niger. Thèse de 3e cycle. Université de Haute Nor
mandie, Rouen, France.
G al lais J. 1977. Le delta intérieur du Niger. Etude de
géographie régionale. IFAN, Dakar, Senegal.
Garcia D and Gall C. 1981. Goats in the dry tropics. In:
Gall C. (ed.) Goat production. Academic Press,
London, UK.
Gerbaldi P. 1978. Divers rapports sur les petits rumi
nants. Institut d'Elevage et de Médecine Vété
rinaire des Pays Tropicaux, Maisons-Alfort, France.
Hafez E S E. 1951. Sexual season of the ewe and day
light environment. Nature (London) 166: 822.
109
 Hafez E S E. 1952. Studies on the breeding season and
reproduction of the ewe. /. Agric. Sci. (Camb). 41:
189-265.
Harvey W R. 1977. User's guide for least-squares and
maximum likelihood computerprogram. Ohio State
University, Columbus, Ohio, USA.
Haumesser J B. 1975. Quelques aspects de la reproduc
tion chez la chèvre rousse de Maradi: comparaison
avec d'autres races tropicales et sub-tropicales.
Rev. Elev. Méd. Vét. Pays Trop. 28: 225-234.
Haumesser J B and Gerbaldi P. 1980. Observations sur
la reproduction de l'élevage du mouton Oudah
nigérien. Rev. Elev. Méd. Vét. Pays Trop. 33: 305-
313.
1ER (Institut d'économie rurale). 1975. Enquête agro
socio-économique dans la zone d'extension de
l'opération de riz Ségou. 1ER, Bamako, Mali.
ILCA/IER. 1978. Evaluation of the productivity of the
Maure and Peul breeds of cattle at the Station du
Sahel, Niono. ILCA Monograph No. 1. ILCA,
Addis Ababa, Ethiopia.
Kolff H E and Wilson R T. 1985. Livestock production
in central Mali: The "mouton de case" system of
smallholder sheep fattening. Agric. Syst. 16: 217-
230.
Lees J L. 1971. Some aspects of reproductive efficiency
in sheep. Vet. Rec. 88: 86-96.
Logan L L, Godwin J T, Tembely S and Graig J M.
1984. Maintaining zebu Maure cattle in a tsetse-
infested area of Mali. Trop. Anim. Health Prod. 16:
1-12.
Mahadevan P. 1966. Breeding for milk production in
tropical cattle. Tech. Bull. No. 17. Commonwealth
Bureau of Animal Breeding and Genetics, Farn-
ham Royal, U.K.
Mahadevan P, Galukande E B and Black J C. 1962. A
genetic study of the Sahiwal grading-up schemes in
Kenya. Anim. Prod. 4: 337-342.
Maiga A W. 1980. Etude des troupeaux bovins de la ville
de Ségou: composition, nutrition, production.
Mémoire de fin d'études. Institut polytechnique
rural, Katibougou, Mali.
Mason I L. 1969. A world dictionary oflivestock breeds,
types and varieties (2nd Edition). Tech. Comm.
No. 8 (Revised) of the Commonwealth Bureau of
Animal Breeding and Genetics, Farnham Royal,
UK.
Ozakuma J M, Whiteman J W and McNew R W. 1982.
Repeatability of lambing rate. J. Anim. Sci. 54:
540-543.
Payne W J A. 1970. Cattle production in the tropics.
Volume 1, Breeds and breeding. Longmans, Lon
don, UK.
Peacock C P. 1983. Phase exploratoire d'une étude des
systèmes de production animale dans le Gourma
malien: données de base sur le cheptel caprin et
ovin. Document de programme N° 92c. CIPEA,
Bamako, Mali.
Sangare M. 1983. Phase exploratoire d'une étude des
systèmes de production animale dans le Gourma
malien: données de base sur le cheptel bovin. Docu
ment de programme N° AZ92d. CIPEA, Bamako,
Mali.
Swift J J. 1979a. West African pastoral production sys
tems. Centre for Research on Economie Develop
ment, University of Michigan, Ann Arbor, USA.
Swift J J. 1979b. The economics of traditional nomadic
pastoralism: The Tuareg of the Adrar n'Ifora (Mali).
D.Phil.Thesis. University of Sussex, UK.
Traoré A. 1984. Incidence de la fasciolose chez les ru
minants du système agro-pastoral associé à la
riziculture irriguée dans la zone de Niono. Docu
ment de programme N° AZ 121. CIPEA, Bamako,
Mali.
Wilson R T and Clarke S E. 1976a. Studies on the live
stock of Southern Darfur, Sudan. II. Production
traits in cattle. Trop. Anim. Health Prod.
8: 47-57.
Wilson R T and Clarke S E. 1976b. Studies on the live
stock of Southern Darfur, Sudan. III. Production
traits in sheep. Trop. Anim. Health Prod. 8: 103-
114.
Wilson R T. 1976c. Studies on the livestock of Southern
Darfur, Sudan. IV. Production traits in goats. Trop.
Anim. Health Prod. 8: 221-232.
Wilson R T, de Leeuw P N and de Haan C. 1983. Re
cherches sur les systèmes des zones arides du Mali:
résultats préliminaires. Research Report No. 5.
ILCA, Addis Ababa, Ethiopia.
Wilson R T and Wagenaar K T. 1983. Enquête sur la
démographie des troupeaux et sur la production
chez les animaux domestiques dans la zone du pro
jet Gestion des Pâturages et Elevage de la Répu
blique du Niger. Programme Document No. AZ
80. ILCA, Bamako, Mali.
Wilson R T, Peacock C P and Sayers A R. 1984. Aspects
of reproduction in goats and sheep in south-central
Kenya. Anim. Prod. 38: 463-467.
Wilson R T, Peacock C P and Sayers A R. 1985. Pre-
weaning mortality and productivity indices for goats
and sheep on a Maasai group ranch in south-central
Kenya. Anim. Prod. 41: 201-206.
Yenikoye A, Pelletier J, Andre D and Mariana J C.
1982. Anomalies in ovarian function of Peul ewes.
Theoriogenology 17: 355-364.
110
APPENDIX
Correlationc efficientoando0n0f cancel velforalvbord r 0t 0p dr in allp ificodoo e0.
0.000 0.000 0.050 0.000
Mean
P
0.0500
0.on00
0.00 -0.0on0
r
0.50 0.000 0.00 0.005
r030)
P
on80
0.055 0.005 0.055
-0.0000
r
0.005 0.05 0.005
0.050
on5r050)
P
0.050 0.0505 0.000 0.0005
rmm)inprevioooyear
r
0.003 0.55 0.00 0.550
on5r000)
P
0."00
0.5"
0.0000 -0.555
Yearofbirt0randtotalain ll
r
0.000
0.05
0.00
0.50
r0on)
P
on00
0.000
0.00r
0.5%
0.055
r
0.00
0.50
0.05
0.00
r000)
P
050 0.5000
0.on5
0.000 -0.0000
r
0.05 0.000 0.000 0.50
r50)
P
on0 0.on0
0.0r0
-0.0500 -0.0050
Rainfall atmont0o previoooly
r
0 00 5 5

THE CONSULTATIVE GROUP ON INTERNATIONAL
AGRICULTURAL RESEARCH
The International Livestock Centre for Africa (ILCA) is one of the 13 international agricultural research
centres funded by the Consultative Group on International Agricultural Research (CGIAR). The 13
centres, located mostly within the tropics, have been set up by the CGIAR over the last decade to pro
vide long-term support for agricultural development in the Third World. Their names, locations and re
search responsibilities are as follows :
 
Centro International de
Agricultura Tropical (CIAT),
Colombia: cassava, field beans,
rice and tropical pastures.
Centro International de
Mejoramiento de Maiz y Trigo
(CIMMYT), Mexico: maize and
wheat.
Centro International de la Papa
(CIP), Peru: potato.
International Centre for
Agricultural Research in the Dry
Areas (ICARDA), Syria:
farming systems, cereals, food
legumes (broad bean, lentil,
chickpea), and forage crops.
International Board for Plant
Genetic Resources (IBPGR),
Italy.
International Crops Research
Institute for the SemioArid
Tropics (ICRISAT), India:
chickpea, pigeon pea, pearl
millet, sorghum, groundnut,
and farming systems.
International Livestock Centre
for Africa (ILCA), Ethiopia:
African livestock production.
International Rice Research
Institute (IRRI), the Philippines:
rice.
International Institute of
Tropical Agriculture (IITA),
Nigeria: farming systems, maize,
rice, roots and tubers (sweet
potatoes, cassava, yams), and
food legumes (cowpea, lima
bean, soybean).
International Laboratory for
Research on Animal Disease
(ILRAD), Kenya: trypano
somiasis and theileriosis of cattle.
West Africa Rice Development
Association (WARDA),
Liberia: rice.
International Service for
National Agricultural Research
(ISNAR), the Netherlands.
International Food Policy
Research Institute (IFPRI),
USA: analysis of world food
problems.
 ILCA RESEARCH PUBLICATIONS
Research Reports
1. Tendances et perspectives de l'agriculture et de I'élevage en Afrique sub-saharienne, par C. de
Montgolfier-Kouevi et A. Vlavonou. 1983.
2. Cattle herd dynamics: An integer and stochastic modelfor evaluating production alternatives, by
P. KonandreasandF.M. Anderson. 1982.
3 . Evaluation ofthe productivities ofDjallonke sheep andN'Dama cattle at the Centre de Recherches
Zootechniques, Kolda, Senegal, by A. Fall, M. Diop, J. Sandford, Y.J. Wissocq, J. Durkin and
J.C.M. Trail. 1982.
4. Research on farm and livestock productivity in the central Ethiopian highlands: Initial results, by
G. Gryseels and F.M. Anderson. 1983.
5. Recherches sur les systemes des zones arides du Mali: resultats préliminaires. eds. R.T. Wilson,
P.N. de Leeuw et C. de Haan. 1983.
6. The water resource in tropical Africa and its exploitation, by G.A. Classen, K. A. Edwards and
E.H.J. Schroten. 1983.
7. Livestock water needs in pastoral Africa in relation to climate andforage, by J.M. King. 1983.
8 . Organisation and management of water supplies in tropical Africa, by S .G . Sandford . 1 983 .
9. Productivity of Boran cattle maintained by chemoprophylaxis under trypanosomiasis risk, by
J.C.M. Trail, K. Sones, J.M.C. Jibbo, J. Durkin, D.E. Light and Max Murray. 1985.
10. Economic trade-offs between milk and meat production under various supplementation levels in
Botswana, by P. A. Konandreas, F.M. Anderson and J.C.M. Trail. 1983.
11. Crossbred dairy cattle productivity in Arsi region, Ethiopia, by G.H. Kiwuwa, J.C.M. Trail,
M.Y. Kurtu, Getachew Worku, F.M. Anderson and J. Durkin. 1983.
12. Evaluation of the productivity ofcrossbred dairy cattle on smallholder and Governmentfarms in
the Republic ofMalawi, by Kwaku Agyemang and Lidie P. Nkhonjera. 1986.
13. Productivity of transhumant Fulani cattle in the inner Niger delta of Mali, by K.T. Wagenaar,
A. Diallo and A.R. Sayers. 1986
International Livestock Centre for Africa
P.O. Box 5689
Addis Ababa, Ethiopia