Small Ruminant
Production Techniques
International Livestock Research Institute
ILRI
Training
Manual
3
ILRI Training Manual 3
Small Ruminant
Production Techniques
H. Ibrahim
International Livestock Research Institute
P.O. Box 30709, Nairobi, Kenya
ISBN 92–9146–044–3
Correct citation: Ibrahim H. 1998. Small Ruminant Production Techniques. ILRI Manual 3. ILRI (International
Livestock Research Institute), Nairobi, Kenya. 207 pp.
Table of Contents
List of Tables .......................................................ix
List of Figures ......................................................xi
Foreword .........................................................xiii
About the Manual ...................................................xiv
Module 1: An overview of small ruminant production in Africa ......................1
1.1 Performance objectives.................................................1
1.2 Agro-ecological zones in sub-Saharan Africa ...............................1
1.2.1 Agro-ecological zones ...........................................1
1.2.2 Population pressure and natural resources ............................4
1.3 Major production systems in sub-Saharan Africa ............................5
1.3.1 Traditional types of livestock production systems ......................5
1.3.2 Modern types of livestock production systems ........................6
1.3.3 Management within livestock production systems......................7
1.3.4 Distribution of small ruminants in sub-Saharan Africa ..................7
1.4 Function, role and productivity of small ruminants ..........................11
1.4.1 Special features of small ruminants ................................11
1.4.2 Small ruminants as sources of meat and milk ........................11
1.4.3 Other contributions of small ruminants in sub-Saharan Africa ...........12
1.5 Constraints to and opportunities for increased productivity....................12
1.6 Exercises ...........................................................13
1.7 References and reading materials ........................................14
Module 2: Livestock farming systems research .................................15
2.1 Performance objectives................................................15
2.2 Introduction.........................................................15
2.3 Livestock farming systems: Key concepts .................................17
2.4 Livestock farming systems conceptual model ..............................18
2.5 Livestock farming systems process ......................................20
2.5.1 The diagnostic phase ...........................................21
2.5.2 The design phase ..............................................24
2.5.3 The testing phase ..............................................24
2.5.4 The extension phase ............................................25
2.6 Participatory rural appraisal (PRA) ......................................25
2.6.1 Why PRA? ...................................................25
2.6.2 PRA techniques ...............................................26
iii
2.6.3 Virtues of PRA ...............................................27
2.7 Categories of livestock farming systems research ...........................27
2.8 Exercises ...........................................................28
2.9 References and reading materials ........................................28
Module 3: Reproduction in African small ruminants .............................29
3.1 Performance objectives................................................29
3.2 Introduction.........................................................29
3.3 Physiology of reproduction.............................................29
3.3.1 Basic concepts of reproduction ...................................30
3.4 Reproductive performance .............................................35
3.4.1 Annual reproductive rate ........................................36
3.4.2 Litter size ....................................................38
3.4.3 Lamb mortality ................................................40
3.4.4 Lambing interval ..............................................41
3.4.5 Proportion of reproductive ewes ..................................42
3.5 Management of reproduction ...........................................42
3.5.1 Techniques for inducing oestrus and ovulation .......................42
3.6 Exercises ...........................................................44
3.7 References and reading materials ........................................44
Module 4: Health and diseases............................................47
4.1 Performance objectives................................................47
4.2 Introduction.........................................................47
4.3 Background information...............................................47
4.3.1 Group I: Epidemic diseases ......................................49
4.3.2 Group II: Parasitic and viral diseases ...............................49
4.3.3 Group III: Diseases associated with intensification ...................50
4.4 Major diseases of small ruminants in sub-Saharan Africa .....................50
4.5 Helminthosis ..........................................................
4.5.1 Trematodes ...................................................56
4.5.2 Nematodes ...................................................57
4.5.3 Cestodes (tapeworms) ..........................................57
4.5.4 Control strategies for helminthosis.................................58
iv
4.6 Exercises ...........................................................59
4.7 References and reading materials ........................................60
Module 5: Genetic improvement of African small ruminants ........................61
5.1 Performance objectives................................................61
5.2 Introduction to genetic concepts.........................................61
5.3 Characterisation and conservation .......................................62
5.3.1 Rationale for characterisation and conservation.......................67
5.3.2 Characterisation and conservation methodology ......................68
5.4 Genotype × environment interaction .....................................71
5.5 Utilisation of biodiversity ..............................................71
5.5.1 Selection .....................................................71
5.5.2 Cross-breeding ................................................72
5.5.3 Inbreeding....................................................74
5.6 Introduction to improvement ...........................................74
5.6.1 Nucleus breeding schemes .......................................75
5.6.2 Genetics of diseases resistance....................................76
5.7 Exercises ...........................................................77
5.8 References and reading materials ........................................77
Module 6: Nutrient requirements and feed resources .............................79
6.1 Performance objectives................................................79
6.2 Introduction.........................................................79
6.3 Feed intake .........................................................81
6.3.1 Using digestibility data..........................................81
6.3.2 Using markers.................................................81
6.4 Nutrient value of feed .................................................82
6.4.1 Energy.......................................................82
6.4.2 Protein.......................................................82
6.4.3 Minerals .....................................................82
6.4.4 Vitamins .....................................................82
6.4.5 Water .......................................................82
6.5 Evaluation of feed....................................................83
6.5.1 In vivo measurement of digestibility ...............................83
6.5.2 In vitro methods for measuring digestibility .........................83
6.6 Feed resources ......................................................83
6.6.1 Natural pastures ...............................................83
6.6.2 Improved pastures .............................................84
v
6.6.3 Cultivated pastures or fodder crops ................................84
6.6.4 Conserved forage ..............................................85
6.6.5 Crop residues .................................................86
6.6.6 Industrial by-products...........................................87
6.6.7 Fodder trees ..................................................89
6.7 Feeding techniques ...................................................89
6.8 Exercises ...........................................................90
6.9 References and reading materials ........................................90
Module 7: Economics of small ruminant production .............................93
7.1 Performance objectives................................................93
7.2 Introduction to production relationships...................................93
7.2.1 Production relationships ..........................................94
7.3 Partial budget analysis ...............................................100
7.3.1 Logic behind partial budget analysis ..............................100
7.3.2 Data required for partial budget analysis ...........................102
7.3.3 An example of partial budget analysis for small ruminants .............105
7.4 Exercises ..........................................................108
7.5 References and reading materials .......................................109
Module 8: Biometrics techniques .........................................111
8.1 Performance objectives...............................................111
8.2 Introduction........................................................111
8.3 Background to basic biometry .........................................113
8.3.1 Methods of describing data .....................................113
8.3.2 Descriptive statistics...........................................116
8.3.3 Precision of estimated parameters ................................117
8.3.4 Analysis of variance ...........................................119
8.3.5 t-test .......................................................123
8.3.6 Blocking ...................................................124
8.3.7 Heterogeneity of variance.......................................124
8.3.8 Regression analysis and analysis of covariance ......................125
8.3.9 Analysis of discrete data........................................127
8.4 Planning experiments ................................................128
8.4.1 Stages in planning experiments ..................................129
8.4.2 Experimental units ............................................131
8.4.3 Randomisation ...............................................131
8.4.4 Designing surveys ............................................132
8.4.5 Sizes of experiments and sample surveys ..........................134
vi
8.5 Presenting results ...................................................136
8.6 Exercises ..........................................................137
8.7 References and reading materials .......................................140
Module 9: Research project protocols ......................................141
9.1 Performance objectives...............................................141
9.2 Introduction........................................................141
9.3 Research project protocols ............................................142
9.4 Planning ..........................................................142
9.5 Components of a protocol.............................................142
9.5.1 Background and justification ....................................143
9.5.2 Objectives...................................................143
9.5.3 Expected outputs .............................................143
9.5.4 Work plan ...................................................144
9.5.5 Budget......................................................145
9.5.6 Logical framework ............................................146
9.6 Writing strategies ...................................................147
9.6.1 Some writing guidelines........................................148
9.7 The review process ..................................................150
9.8 Exercises ..........................................................148
9.9 References and reading materials .......................................149
Module 10: Guidelines to communication ...................................151
10.1 Performance objectives..............................................151
10.2 Introduction.......................................................151
10.3 Planning communication in a research project............................151
10.4 Communication guidelines ...........................................152
10.4.1 Guidelines to oral presentation..................................152
10.4.2 Guidelines to interpersonal communication........................154
10.4.3 Guidelines to visual skills......................................155
10.4.4 Guidelines to writing/editing skills ..............................156
10.5 Writing scientific papers.............................................157
10.5.1 Planning the paper ...........................................157
10.5.2 Writing the first draft .........................................161
10.5.3 Style ......................................................165
10.5.4 Review by colleagues.........................................165
10.5.5 Submission to a journal .......................................165
vii
10.6 Writing review articles ..............................................166
10.6.1 Types of review articles........................................166
10.6.2 Literature search .............................................167
10.6.3 Writing the review article ......................................167
10.7 Posters and overhead transparencies....................................167
10.7.1 Posters for scientific meetings...................................167
10.7.2 Preparing overhead projection transparencies.......................171
10.8 Exercises .........................................................172
10.9 References and reading materials ......................................172
Module 11: Transfer of technology ........................................175
11.1 Performance objectives..............................................175
11.2 Introduction.......................................................175
11.2.1 Key concepts ................................................176
11.3 Technology transfer systems .........................................178
11.4 Transfer of technology linkages with research............................179
11.4.1 Why do we need links? .......................................179
11.4.2 Organisational issues.........................................179
11.4.3 Types of linkage processes ....................................180
11.4.4 Management of linkages ......................................182
11.5 Extension for integrated crop–livestock systems .........................182
11.6 Extension for pastoralist systems ......................................183
11.7 Case studies ......................................................183
11.7.1 Case study 1: Fodder banks in Nigeria............................183
11.7.2 Case study 2: Alley farming in West Africa .......................186
11.7.3 Case study 3: Crossbred goats for milk production in Burundi .........189
11.7.4 Case study 4: Dual-purpose goats in Kenya........................190
11.8 Exercises .........................................................191
11.9 References and reading materials ......................................191
viii
List of Tables
Table 1.1 Agro-ecological zones of sub-Saharan Africa. ..................................3
Table 1.2 Human population projections for sub-Saharan Africa, 1990 2025..................4
Table 1.3 A classification of small ruminant production types and systems
in sub-Saharan Africa. ....................................................5
Table 1.4 Livestock owning in agropastoral and pastoral societies in Kenya and Chad. ..........7
Table 1.7 Distribution (%) of domestic ruminant livestock by agro-ecological zone and
geographic region, sub-Saharan Africa.......................................10
Table 1.8 Ruminant density and carrying capacity in agro-ecological zones
in sub-Saharan Africa.....................................................10
Table 1.9 Small ruminant meat and milk production by region in sub-Saharan Africa. .........11
Table 1.10 Major constraints and opportunities affecting the potential of small ruminant
production in sub-Saharan Africa. ..........................................13
Table 2.1 Approaches to research. ..................................................21
Table 3.1 Productivity Index II for sheep in sub-Saharan Africa. ..........................37
Table 3.2 Estimates of litter size in African sheep.......................................38
Table 3.3 Estimates of embryonic losses in sheep. ......................................38
Table 3.4 Estimates of abortion rate in sheep and goats in Africa...........................39
Table 3.5 Estimates of losses due to the slaughter of pregnant females in sub-Saharan Africa. ...39
Table 3.6 Estimates of preweaning mortality rate among sheep in Africa. ...................40
Table 3.7 Post-mortem diagnosis of causes of mortality among Horro and Menz sheep
in Ethiopia. ............................................................41
Table 5.1 Selected examples of breeds of sheep and goats in sub-Saharan Africa. .............64
Table 5.2 Information on African small ruminant genetic resources in the FAO data bank. ......68
Table 5.3 Status of African small ruminant genetic resources..............................68
Table 5.4 Traits used as a basis for selection in small ruminants. ..........................71
Table 5.5 Examples of goat and sheep breeds developed through cross-breeding in sub-Saharan
Africa. ................................................................73
Table 6.1 Types of animal data used to diagnose animal nutrition problems. .................81
Table 7.2 Checklist of costs involved in producing ruminants. ...........................104
Table 7.3 Checklist of benefits of producing ruminants. ................................104
Table 7.4 Small ruminant production parameters under four treatments in the
Ethiopian highlands, 1989. ...............................................106
Table 7.5 Gross return (EB/ewe) of feeding and health management experiments for
sheep, Debre Birhan, Ethiopia, 1989. .......................................106
Table 7.6 Results of feeding and health management experiments for sheep, Debre Birhan,
Ethiopia, 1989. ........................................................107
Table 8.1 Data set for the effect of supplementation on health and growth rate of weaned
Red Maasai and Dorper lambs. ............................................112
Table 8.2 Number of animals grouped by packed cell volume (PCV) category. ..............113
ix
Table 8.3 Number of animals grouped by faecal egg count (FEC) category..................114
Table 8.4 Descriptive statistics of variables measured in the example data set................117
Table 8.5 Example of a recording sheet..............................................130
Table 8.6 Weight gains (kg) of lambs. ..............................................136
Table 9.1 Schedule of experimental matings to generate genetic parameters in the
endoparasite project, Debre Birhan, Ethiopia. ................................144
Table 9.2 Budget format for a protocol. .............................................145
Table 11.1 Areas of interdependence between research and technology transfer. ..............177
x
List of Figures
Figure 1.1 Sub-Saharan Africa: Main agro-ecological zones. ...............................2
Figure 2.1 Models for information flow among researchers (R), farmers (F) and
development agents (D). .................................................18
Figure 2.2 General conceptual model for livestock systems. ...............................19
Figure 2.3 Systems hierarchy. ......................................................19
Figure 2.4 Farming system research process............................................20
Figure 3.1 The ovarian cycle of ewes. ................................................31
Figure 3.2 Hormonal control of oestrus in ewes ........................................31
Figure 3.3 Hormone patterns during the oestrus cycle of ewes..............................32
Figure 3.4 Monthly variation in number of oestrus cycles of Menz ewes in Ethiopia............33
Figure 3.5 Frequency distribution of the duration of oestrus cycle length in Menz ewes. ........33
Figure 3.6 Monthly distribution of lambing activity in Menz sheep under natural mating. .......34
Figure 3.7 Causes of poor reproductive performance in sheep. .............................36
Figure 3.8 Effect of birth weight on lamb mortality in Menz sheep. .........................41
Figure 4.1 Relationships among factors affecting health. .................................48
Figure 4.2 Relationships among host, environment and disease causal agent. .................48
Figure 4.3. The life cycle of the liver fluke (Fasciola spp). ................................56
Figure 4.4 The life cycle of gastro-intestinal nematodes (Haemonchus, Ostertagia,
Trichostrongylus and Cooperia). ...........................................57
Figure 4.5 The life cycle of a tapeworm (Taenia multiceps)................................58
Figure 5.1 A thin-tailed sheep. ......................................................62
Figure 5.2 A West African Dwarf ram. ...............................................63
Figure 5.3 A fat-rumped Blackhead Persian sheep. ......................................63
Figure 5.4 West African Dwarf goats. ................................................65
Figure 5.5 Crossing types. .........................................................74
Figure 5.6 Outline of the structure of an open nucleus breeding scheme in C te d Ivoire.........75
Figure 5.7 Example of an optimum structure of a nucleus breeding programme. ...............76
Figure 7.1 Constant marginal productivity. ............................................95
Figure 7.2 Increasing marginal productivity. ...........................................95
Figure 7.3 Decreasing marginal productivity ...........................................96
Figure 7.4 Stages of production and the relationships between marginal product (MP) and
average product (AP). ...................................................97
Figure 7.5 Least cost combination of two inputs.........................................99
Figure 7.6 Hypothetical product transformation curve. ...................................99
Figure 7.7 Net benefit curve, feeding and health management trials, Ethiopian
highlands, 1989. .......................................................108
Figure 8.1 Histogram of frequency distribution of faecal egg counts (FEC). .................114
xi
Figure 8.2 Scatter plot of packed cell volume (PCV) versus faecal egg count (FEC) with
mistake in PCV value (31 rather than 13%) for lamb 94.........................115
Figure 8.3 Box plot of faecal egg count (FEC) for each breed supplementation group..........115
Figure 8.4 Regression line of packed cell volume (PCV) on log (FEC+50). .................126
Figure 8.5 Weight gains (kg) of sheep in the form of a bar chart...........................137
Figure 10.1 Format determines how you design your poster. ..............................168
Figure 10.2 Layout of posters. ......................................................169
Figure 10.3 Balance among elements in a poster. .......................................170
Figure 11.1 Linking research and technology transfer agencies to farmers....................176
xii
Foreword
The International Livestock Research Institute (ILRI) trains research scientists from national research
systems (NARS) in sub-Saharan Africa. As part of this training, ILRI has conducted several training
courses on small ruminant production techniques. The knowledge and skills learned in the courses
enabled the participants to conduct research using an integrated approach to resolving constraints to
increasing small ruminant productivity in smallholder farming systems.
This manual has been developed from the training courses and is intended for research scientists
who are actively involved in research on small ruminants, or soon will be. The manual is designed for
multiple users, including participants in courses on small ruminants or individuals who desire to learn on
their own.
Many individuals have contributed to the preparation of this manual. As a training materials
specialist, Dr H. Ibrahim wrote the first draft of the modules from the latest course materials as well as
material available with ILRI's Information Services. The modules were then reviewed by ILRI scientists
involved in teaching the course. A sample of former ILRI trainees were surveyed to obtain feedback. The
manual was internally reviewed by Dr S. Lebbie, Co-ordinator of the Small Ruminant Network associated
with ILRI, and externally by Dr L. Reynolds, Animal Scientist, Consultant, UK.
This manual emphasises an integrated approach to research on small ruminants. The manual starts
with describing small ruminants in various agro-ecological zones in sub-Saharan Africa with emphasis on
the prevailing production systems. The user is then introduced to the concepts and methodologies of
farming systems research to better understand the needs of the smallholder farmer. The core of the manual
presents knowledge and skills in the disciplines of reproduction, diseases, genetics, biometry, economics
and communication. The manual ends by stressing the importance of transferring research results to
farmers and the role of researchers in this vital process.
Even though this manual was written to stand on its own, it is part of ILRI’s series of training
material on small ruminants. This material includes audio-tutorial modules, slide series and posters. We
view them as complementary to each other.
ILRI is grateful to the European Union for funding the production of this manual through its
financial support to the Small Ruminant Network (SRNET).
Dr M.E. Smalley
Director, Strengthening Partnerships with National
Agricultural Research Systems
ILRI
xiii
About the Manual
This manual is part of ILRI's effort to provide training materials for use in courses or self-teaching. The
manual aims at providing an overview of small ruminant research knowledge and methods to help young
scientists in Africa improve the situation in this vital area. Eleven modules cover the topics in an integrated
approach to improving small ruminant productivity in sub-Saharan Africa.
The first module gives an overview of the role of small ruminants in production systems in relation
to agro-ecological zones in sub-Saharan Africa. This is followed by the description of farming systems
approaches. The core of the manual covers reproduction, diseases, breeding and genetics, nutrition,
economics, biometry and communication. The reader is then exposed to approaches for writing protocols
for funding research. The last module covers the role of researchers in transferring research results to the
farmers. Relevant reading materials support each module.
Each module is designed such that the performance objectives are placed at the beginning to
facilitate learning. The body of the module includes text, illustrations, reference and reading materials.
Exercises are placed at the end of each module to stimulate the user for further learning. Topics that further
illustrate the main concepts in the body of the module are placed in boxes.
Many individuals have contributed to the preparation of this manual. Dr H. Ibrahim prepared the
modules and the following ILRI staff contributed to them:
• Dr S. Lebbie reviewed modules 1, 2 and 9
• Dr E. Mukasa-Mugerwa wrote and reviewed module 3
• Dr S. Tembely wrote and reviewed module 4
• Dr C. Valentine Yapi wrote and reviewed module 5
• Dr N. Ummuna reviewed module 6
• Dr B. Shapiro reviewed module 7
• Dr J. Rowlands wrote and reviewed module 8
• Mr P. Neate and Ms A. Nyamu reviewed module 10
• Drs M. Mohamed-Saleem and M. Jabbar reviewed module 11
Scientists in NARS are encouraged to use the manual in their training courses. It is hoped that users
find the manual useful.
Habib Ibrahim
Training Materials Specialist
ILRI
xiv
Module 1: An overview of small ruminant
production in Africa
1.1 Performance objectives
Module 1 is intended to enable you to:
1. Describe the effect of agro-ecological zones in sub-Saharan Africa on small ruminant production.
2. Describe the major small ruminant production systems in sub-Saharan Africa.
3. Describe the distribution of small ruminants in sub-Saharan Africa by zone and region.
4. List the special attributes of small ruminants.
5. List constraints to productivity of small ruminants in sub-Saharan Africa.
6. List the opportunities for improving small ruminant productivity.
1.2 Agro-ecological zones in sub-Saharan Africa
1.2.1 Agro-ecological zones
Sub-Saharan Africa can be divided into five agro-ecological zones: desert, arid, semi-arid, subhumid,
humid and the cool highlands (Figure 1). These zones vary in climate, natural resources and human
population density. This classification is based on amount and distribution of rainfall, effect of altitude on
temperature and length of annual plant growing period. The parameters defining the zones (rainfall and
plant growth days) and percentage area across the different regions of sub-Saharan Africa are shown in
Table 1.1. In this module we will focus on small ruminant production in the arid, semi-arid, subhumid and
humid zones where most of these animals are raised. However, since natural resources and human and
livestock populations vary from one zone to another, special sections will be devoted to the effect of
population pressure on crop–livestock integration.
1
1.1 Performance objectives
1.2 Agro-ecological zones in sub-Saharan Africa
1.3 Major production systems in sub-Saharan Africa
1.4 Function, role and productivity of small ruminants
1.5 Constraints to and opportunities for increased productivity
1.6 Exercises
1.7 References and reading materials
Figure 1.1. Sub-Saharan Africa: Main agro-ecological zones.
2
Arid zone
The arid zone receives 0 to 500 mm of rainfall annually which sustains plant life for less than 90 plant
growth days (Table 1.1). The small amount of rainfall and its erratic distribution prevent cropping in most
years; some cropping takes place in oases or irrigated areas. Plant cover consists of short annual grasses,
legumes, scattered shrubs and trees. Farmers use most of the trees and shrubs for fuel. The arid zone
occupies 36% of the area of sub-Saharan Africa, most of which is in West and East Africa. Opportunities
for livestock development are limited but existing techniques could be improved upon if not to increase
productivity, to at least sustain it.
Semi-arid zone
The semi-arid zone receives 500 to 1000 mm of rainfall annually which sustains 90 to 180 plant growing
days. The lower rainfall areas of this zone are used for grazing. Cropping and crop–livestock systems
dominate the areas with higher rainfall. Farmers grow millet, sorghum, groundnut, maize and cowpeas.
This zone occupies 18% of the area of sub-Saharan Africa and is found in all regions except central Africa.
Subhumid zone
The subhumid zone receives 1000 to 1500 mm of rain annually which sustains plants for 180 to 270 plant
growing days. Food and cash crops are grown, including cassava, yams, maize, fruits, vegetables, rice,
millet, groundnut, cowpeas and cotton. Some of these crops are used to make feeds such as cotton seed
cakes and the residues of other crops are available for livestock feeding. In some areas of this zone farmers
grow soyabean and leguminous forage crops. This zone occupies 22% of sub-Saharan Africa, mainly in
southern and central Africa.
Humid zone
The humid zone receives more than 1500 mm of rainfall annually which sustains plants for 270 to 365
plant growing days. This zone consists of rain forest and derived savannahs. The soils have high levels
of iron and aluminium and low levels of phosphorus. Therefore the organic matter content is low, and the
soils are fragile and easily degraded when the vegetative cover is lost. The zone occupies 19% of
sub-Saharan Africa, mostly in central and West Africa. This zone contains relatively few livestock and
3
Table 1.1. Agro-ecological zones of sub-Saharan Africa.
Zone Definition
Area (%)
Rainfall range
(mm)
West
Africa
Central
Africa
Eastern
Africa
Southern
Africa
Area of
zone
(%)
Total
area
(×10
6
km
2
)
Arid <90 pgd 0–500 54 1 52 20 36 7.7
Semi-arid 90–180 pgd 500–1000 20 7 18 34 18 4.0
Subhumid 180–270 pgd 1000–1500 16 29 16 38 22 4.8
Humid >270 pgd 1500+ 10 59 2 7 19 4.1
Highlands
a
20°C n.a. 0 4 12 1 5 1.0
Total 100 100 100 100 100
Total area
(×10
6
km
2
7.3 5.3 5.8 3.2 21.6
n.a. = data not available; pgd = plant growth days.
a. Defined as areas at 1500 m altitude within the semi-arid, subhumid, and humid zones where the mean daily temperature
during the growing period is less than 20°C.
Source: ILCA (1987) after Jahnke (1982).
has limited potential for livestock development, particularly because of the threat of the trypanosomosis-
transmitting tsetse fly.
Highlands zone
The cool highlands zone includes areas above 1500 m altitude that have a mean daily temperature of less
than 20°C. This zone represents 5% of the total area of sub-Saharan Africa, most of which lies in eastern
Africa; half of this zone is in Ethiopia. The soils are deep fertile Vertisols and Nitosols.The zone receives
bimodal rainfall (1000 mm annually) and there are two growing seasons. Farmers grow forages and
animal traction assumes more importance as the population pressure encourages crop–livestock
integration. The cool highlands are a high potential area for crop–livestock integration.
1.2.2 Population pressure and natural resources
The human population in sub-Saharan Africa currently is increasing by 3.1% a year (Table 1.2). The
projections (Table 1.2) indicate a decrease in growth rate after the year 2010 but it will still be relatively
high by the year 2025. Based on projections of population increases and limited land availability, McIntire
et al (1992) hypothesised that population pressures on agricultural land in sub-Saharan Africa will drive
agriculture toward intensification.
One form of intensification due to population pressure is the evolution of crop–livestock systems.
This system will be more economically efficient and sustainable than other systems when the following
conditions prevail:
• traditional production methods, with low input technology, are used
• inputs are scarce
• markets are poorly developed.
The following scenarios may develop under different population pressures and land resources:
Scenario 1. Low population pressure + low animal disease stress symbol Herding
Scenario 2. High population pressure Crop–livestock systems
Scenario 3. Very high population pressure + urbanisation Specialised systems
4
Table 1.2. Human population projections for sub-Saharan Africa, 1990–2025.
Year Human population
(millions)
Annual growth rate
(% 5-year period)
1990 498 –
1995 580 3.1
2000 676 3.1
2005 784 3.0
2010 902 2.8
2015 1,028 2.6
2020 1,159 2.4
2025 1,294 2.2
The first scenario results from using fallow periods rather than manure to restore the fertility of the
cropping land. Labour requirements for the fallow system are low. The second scenario develops as
population pressure makes increased food availability essential and therefore cropping is intensified.
Continuous cropping leads to a decline in fertility and hence the need to introduce animals into the system
to provide manure. Since pasture and fallow are no longer options for feeding livestock, crop residues start
to play a more important role. The third scenario evolves when urban centres become the high population
centres and markets and technologies are developed (fertilisers replacing manure, tractors replacing
animal traction, concentrates replacing crop residues). As the population continues to increase, the
efficiency of crop–livestock systems plateaus and attaining a higher level of productivity requires the use
of higher input technologies. The demand for animal products becomes higher as the incomes of city
dwellers increase. This leads to specialisation such as market-oriented peri-urban dairy production and
livestock fattening, especially of ruminants. The demands of urban dwellers may also lead to specialised
cropping such as vegetables, forages and cash crops. The evolution of crop–livestock integration is more
complex than presented here. Each of the above scenarios can be further elaborated once you impose agro-
ecological zones as factors.
1.3 Major production systems in sub-Saharan Africa
Livestock production systems in sub-Saharan Africa can be divided into two major types, namely
traditional and modern (Table 1.3). They can be distinguished mainly through the three production
factors, land, labour and capital.
1.3.1 Traditional types of livestock production systems
The traditional livestock production systems can be sub-divided into four: pastoral, agropastoral,
agricultural and urban/peri-urban (Table 1.3). The production systems are segregated according to the
following factors:
5
Table 1.3. A classification of small ruminant production types and systems in sub-Saharan Africa.
Type System Management
Main production
factors Nutrient source
Traditional Pastoral Nomadic/
semi-sedentary
Land Range
Agropastoral Transhumant/
sedentary
Land/labour Range/crop
by-products
Agricultural Sedentary Land/labour Crop by-products/
household waste/
forage
Urban/peri-urban Sedentary Labour Household waste/
feed
Modern Ranching Sedentary Land/capital Range/forage
Feedlot Sedentary Capital/labour Feed/storage
Station Sedentary Land/labour/
capital
Range/forage/feed
• degree of dependence on livestock products (of household or production unit) for income or food
• type of agriculture practised in association with livestock
• mobility and duration of movement.
Based on the above criteria, the characteristics of these production systems are discussed below.
Pastoral system
• More than 50% of household income (production unit) comes from livestock/livestock products.
• In pure arid areas there is little or no cropping activity; further south of the Sahara rainfed
agriculture is practised.
• Livestock owners/helpers travel long distances year-round looking for forage and water along
specific orbits (in oases frequency of movement is reduced and a little agriculture is practised, i.e.
semi-sedentary).
Agropastoral system
• Between 10 and 50% of household income comes from livestock/livestock products.
• In semi-arid regions agriculture is practised together with management of livestock.
• The system is either transhumant or sedentary (limited movement). Transhumance refers to the
whole system moving periodically, e.g. migration to the wetter fringes of the arid and semi-arid
zones in the south during the dry season and returning in the wet season.
Agricultural systems
• Less than 10% of household income is derived from livestock/livestock products.
• Livestock production is secondary to crop production.
• The system is sedentary.
Urban/peri-urban
• Involves fattening of small ruminants to sell during festivities or to city dwellers.
• The operators are usually wage earners in the city who invest cash in sheep and goats for short-term
profit.
The system is sedentary with cut-and-carry feeding for fattening (it can be natural grass cut and sold
to urban dwellers or specialisation in forage marketing may emerge).
Land assumes more importance in the pastoral system whereas labour is more important in
agropastoral, agricultural and urban systems. Depending on the extent of cropping, hired labour is needed
to supplement family labour. As intensification increases, labour becomes more important than land. The
urban system is capital intensive since farmers invest cash to buy sheep and goats to fatten for sale during
peak demand. In this system, the high wages for labour are the main constraint to cutting and carrying
forage to zero grazing livestock. In the pastoral system the small ruminants obtain their feed from
rangeland. In agropastoral systems rangeland is used but crop residues assume importance. Crop residues,
household waste and forages are used for feeding small ruminants in the agricultural systems. Urban
systems provide sheep and goats with household waste, grain and protein concentrates if the number of
animals is large or use a lot of labour to cut and carry grass to the animals.
1.3.2 Modern types of livestock production systems
The modern livestock production systems can be divided into the ranching, feedlot and station systems
(Table 1.3). These production systems are characterised by high inputs and are all sedentary. Land is
6
important and expensive in the ranching system while the feedlot system is capital intensive and is
characterised by the use of large amounts of feed. Sometimes in the feedlot system, cut-and-carry forages
are used to provide feed and this activity requires labour.
1.3.3 Management within livestock production systems
Management within livestock production systems is governed by the various objectives and strategies of
the farmers. Table 1.4 shows details of ownership pattern for a few ethnic groups in Kenya and Chad. The
objectives of the use of goats and sheep within ethnic groups in a few countries in sub-Saharan Africa are
shown in Table 1.5. It is clear that farmers favour multiple uses of small ruminants over a single use,
especially for goats.
Management is affected by the amount of rainfall (Table 1.6). The common features for the arid,
semi-arid and highland areas are:
• as moisture increases, there is movement from nomadism to fattening strategies
• as we move from dry to wet zones, we move from free grazing to more confinement of animals
• the flock sizes decrease as we move to less harsh environments
• sheep are more predominant in the arid zone whereas goats and sheep predominate in the semi-arid
zone; sheep ranching is predominant in the Kenya highlands and is highly commercial.
Farmers follow a range of strategies to ensure the best use of existing resources within an ecological
zone (Table 1.6). The farmers in the pastoral system exercise some control over breeding by keeping their
better stock for this purpose. In rare cases good genes are taken out of the stock when farmers sell their best
animals.
1.3.4 Distribution of small ruminants in sub-Saharan Africa
Livestock populations vary across the agro-ecological zones of sub-Saharan Africa (Table 1.7).
According to FAO (1989) statistics sub-Saharan Africa has 162 million cattle, 127 million sheep, 147
million goats, 13 million camels, 11 million pigs, 8 million donkeys, 3 million horses, 1.5 million mules
and 631 million chickens.
7
Table 1.4. Livestock owning in agropastoral and pastoral societies in Kenya and Chad.
Country Kenya Chad
Tribe Maasai Karapokot Zioud Salamat
Gondeye-
Tchein
Production
system Pastoral Agropastoral Pastoral Agropastoral Agropastoral
Average holdings
(head/household)
Cattle 157.3 11.8 36.4 133.3 2.1
Sheep 44.0 5.4 43.5 2.0 1.3
Goats 83.1 13.6 45.0 46.3 4.7
Table 1.5. Some examples of management objectives related to flock structures.
Sheep
1
Goats
2
Country/
ethnic group
Males Females Males Females
Use Total Castrates Total Breeding Use Total Castrates Total Breeding
Mauritania/Moor Meat/hair 22.9 6.2 78.1 58.6 Milk/meat 20.2 1.2 79.8 55.1
Mali/Fulani Meat/wool 25.5 11.3 74.5 55.9 * * * * *
Chad/Arab Meat/milk 26.5 Few 73.3 53.7 Milk/meat 28.3 Few 71.1 48.1
Sudan/Baqqara Meat 22.2 0.0 77.8 57.7 Milk/meat 23.6 0.0 76.4 51.2
Kenya/Maasai Meat (fat) 31.4 15.4 68.6 54.2 Meat (fat)/
milk
33.8 10.3 66.2 48.3
1. Percentages of total number of sheep.
2. Percentages of total number of goats.
* = data not available.
Table 1.6. Management of goats and sheep in the arid, semi-arid and highland zones of sub-Saharan Africa.
Micro-management
Sheep      Goats
Size of flock/
herding group
Climatic regime
(rainfall, mm)
Macro-
management Examples Day Night Day Night Sheep Goats
Arid  (200) Nomadic Mauritania/Moor
Ethiopia/Afar
Sudan/Kababish
Mali/Twareg Loose flock
Open camp
Penned
Penned
Open camp
Penned
100--500 30--80
    (300) Transhumant Niger/Twareg
Chad/Zhagawa
Kenya/Turkana
Open camp
Penned
Loose flock Open camp
Semi-arid   (400)
      (500)
      (600)
Semi-
sedentary
Sedentary
Ethiopia/Afar
Sudan/Baqqara
Mali/Fulani
Kenya/Maasai
Sudan/Daju 
etc.
Tight flock
Loose flock
Tight flock
Tight flock
Penned
Penned
Open camp
Penned
Penned
Loose flock
Tight flock
[Dry season not
herded, Crop
season
tight flock]
Penned
Penned
Penned/tied
Penned
Tied
200--250
50--150
20--60
200--500
20--80
5--10
40
30--100
20--80
40--120
5--40
Mali/Bambara Tight flock
Tied
Penned/tied Tied Tied 0--10 2--20
Stall feeding West Africa/
Mouton de case
Kenya/Thenges
Tied Tied 1--5 1--5
Highlands
(>1000)
Extensive
paddocks
Kenya/large
ranches
Paddocks 500--1000
In the drier areas, which are not suited for crop production, rangeland is used for livestock.
Livestock populations are therefore higher in the arid and semi-arid zones than in other areas (Table 1.8).
The humid zone and the wetter parts of the subhumid zone do not sustain large populations of livestock
due to the presence of the tsetse fly, which transmits the disease trypanosomosis. The cool highlands are
less threatened by diseases and thus the zone has a high potential for livestock production. In this zone and
in the marginal area between the wet and dry zones mixed crop–livestock systems have evolved in
response to high population pressure.
The western and eastern parts of sub-Saharan Africa have the largest population of ruminant
livestock in the region (Table 1.8). Livestock species differ in their biomass and in order to express
livestock numbers in terms that reflect the large differences in body size of the species, livestock biomass
is aggregated into tropical livestock units (TLU). A TLU is the equivalent of an animal of 250 kg live
weight. The values for ruminant density and carrying capacity are shown in Table 1.8. Carrying capacity is
high in the arid zone while the humid zone has the lowest figures because of the diseases afflicting
ruminant animals. Ruminant density in sub-Saharan Africa is 0.37 TLU/person, it is 0.20 TLU/person in
Asia, 0.33 TLU/person in West Asia and North Africa and 0.67 TLU/person in Latin America. The value
for the world is 0.30 TLU/person.
10
Table 1.7. Distribution (%) of domestic ruminant livestock by agro-ecological zone and geographic
region, sub-Saharan Africa.
Agro-ecological zone/
geographical regions Cattle Sheep Goats Camels
All domestic
ruminants
1
Arid 20.7 33.7 38.2 100.0 29.8
Semi-arid 30.6 22.9 26.3 0.0 27.1
Subhumid 22.7 14.4 16.5 0.0 19.6
Humid 6.1 8.3 9.4 0.0 6.1
Highland 19.9 20.8 9.6 0.0 17.4
Total 100.0 100.0 100.0 100.0 100.0
West 24.8 34.2 42.3 15.2 26.3
Central 6.6 4.1 6.4 0.0 5.8
Eastern 54.1 59.5 46.2 84.8 56.3
Southern 14.5 7.2 5.2 0.0 11.6
Total 100.0 100.0 100.0 100.0 100.0
1. Calculated from TLU.
Source: ILCA (1987), after Jahnke (1982).
Table 1.8. Ruminant density and carrying capacity in agro-ecological zones in sub-Saharan Africa.
Agro-ecological zone
Ruminant density
(TLU/person)
Carrying capacity
(ha/TLU)
Arid 1.00 10–30*
Semi-arid 0.35 4–8
Subhumid 0.25 3–6
Humid 0.10 –
Highlands 0.40 2.5–3.5
*Carrying capacity is 30 for areas with <250 mm annual rainfall.
1.4 Function, role and productivity of small ruminants
Livestock play an important role in the lives of people in Africa. Animals provide income and serve as
protection against drought years.
1.4.1 Special features of small ruminants
Small ruminants in Africa represent 21% of the world's small ruminant population. The population of
sheep in Africa represents 17% of the total world population while goats represent 30%.
Farmers keep small ruminants for many reasons, the major ones being that they are a source of food
(milk and meat), fibre (wool and skins), cash and a form of savings.
Other special attributes include:
• sheep and goats are highly adaptable to a broad range of environments
• goats are more effective at grazing selectively
• small ruminants have short generation cycles and high reproductive rates which lead to high
production efficiency
• the efficiency of converting feed into milk is higher in goats than in other dairy animals
• certain breeds of sheep and goats (e.g. Red Maasai sheep and West African Dwarf goats) are
tolerant of diseases such as helminthosis
• small ruminants are small enough to be consumed by an average rural family in a day or two hence
no refrigeration facilities are needed.
1.4.2 Small ruminants as sources of meat and milk
Animal products are of great importance to the diets of pastoral people. Small amounts of animal protein
correct amino acid deficiency in diets, particularly of small children, in sub-Saharan Africa.
Meat from small ruminants represent 18% of the total meat eaten by people in sub-Saharan Africa.
Small ruminant meat consumption is 2.1 kg/year per person (11.7 kg/year for all meat). Across the
regions in sub-Saharan Africa small ruminant meat production is distributed as shown in Table 1.9.
Table 1.9. Small ruminant meat and milk production by region in sub-Saharan Africa.
Production
Region Goat Sheep (mutton) % of SSA total
Meat output (’000 t)
West Africa 240 135 20.0
East Africa 199 265 25.3
Southern Africa* 26 10 5.5
Central Africa 30 17 8.5
Milk volume (’000 t)
East Africa 1414 1052 84.2
West Africa 288 139 14.6
Southern Africa* 12 – 0.8
Central Africa 21 3 0.4
*Data from the Republic of South Africa are not included.
11
Negligible quantities of small ruminant meat from sub-Saharan Africa are exported, but no accurate
statistics are available. South Africa exports large quantities of meat and with the sanctions against the
country lifted, the statistics for exports of meat from sub-Saharan Africa will change. Special markets for
small ruminants are expanding, e.g. goats exported from Uganda and Somalia to the Middle East.
Milk from small ruminants is an important source of nutrition for the population in East Africa.
This is not the case in southern and western Africa. The contribution of milk from small ruminants across
the regions is shown in Table 1.9.
1.4.3 Other contributions of small ruminants in sub-Saharan Africa
Employment and income
Farmers use small ruminants as savings that generate cash when the environment is harsh, e.g. during
drought years. Small ruminants are sold to raise money to replace large ruminants lost during droughts.
Small cottage industries based on small ruminant products have started to flourish in African cities, e.g.
coarse wool for carpets, hides/skin products.
Sustainability of agriculture
Small ruminants are integrated in most production systems and this enhances the sustainability of the system:
• the animals produce manure which is returned to the system (nutrient cycling) to benefit vegetable
gardens, food crops and cash crops
• sheep eat weeds reducing weed control costs, e.g. in coffee and coconut plantations
• crop residues are utilised by small ruminants, e.g. straw from rice and wheat and maize stovers are
converted into animal products
• small ruminants make use of plants in marginal lands and convert them into animal products
• small ruminants meet social and cultural needs, e.g. payment of dowry, celebrations, gifts to family members
• the animals are a source of savings to be sold when funds are needed (spreading risk).
1.5 Constraints to and opportunities for increased productivity
The human population in sub-Saharan Africa will more than double by the year 2025 (Table 1.2). Increase
in population, coupled with urbanisation, will result in higher demand for meat per capita. The resources
will be overutilised and more land will be needed to produce food crops. Competition for land between
livestock and crops will therefore be more acute. If we continue to produce livestock and their products at
the current rates, the increase in livestock production will lag behind human population increases. The
potential of small ruminants is not yet fully tapped. FAO (1989) estimated that animal production should
be increased by 4% annually to meet the demand of the human population by the year 2010. The factors
that affect the productivity of the small ruminants in sub-Saharan Africa are:
• natural resources
• feed supply
• health
• genotype
• farming systems and animal management
• land resources
• policy constraints
• institutional constraints.
12
Table 1.10 lists these factors against the constraints that limit the productivity of small ruminants,
reducing their potential output, and the opportunities that can be tapped to achieve the potential.
1.6 Exercises
1. Between 1990 and 2025 there will be enormous demographic and social changes such as
population increase, urbanisation and income change across sub-Saharan Africa. Describe the
effect these changes will have on small ruminant production using examples from your region or
country.
13
Table 1.10. Major constraints and opportunities affecting the potential of small ruminant production in
sub-Saharan Africa.
Constraints Opportunities
Natural resource management
Fragile ecosystem
Soil infertility
Water scarcity
Sustainable crop–livestock system
Sustainable pastoral system
Application of fertilisers
Recycling of nutrients with manure
Mixed crop–livestock systems under trees
Feed and feeding
Quality and quantity of pastures, forages
and crop residues
Crop residue utilisation
Fodder banks
Stall feeding
Cultivated forages
Leguminous trees and alley cropping
Improved pastures
Mineral and protein supplementation
Animal health
Peste des petits ruminants
Contagious caprine pleuropneumonia
Helminthosis
Trypanosomosis
Tick-borne diseases
Sheep pox
Vaccines against diseases
Health management services
Genotype
Low productivity
African genotypes poorly characterised
Adaptive, reproductive traits
Breeding for disease resistance
Socio-economic factors
Poor extension services
Poor infrastructure
Deficient inputs
Price policies and marketing restrictions
Weak public institutions
Inefficiency in drug distribution
Strong market potential
Flexible livestock resource
Low fixed production costs for milk and meat
2. Small ruminant production is strongly influenced by agro-ecological zones in sub-Saharan
Africa. Elaborate on this statement.
3. There are several factors which impose constraints on small ruminant production. Indicate the
significance of these factors by agro-ecological zone by completing the table using the following
scale:
xxx = very important; xx = moderately important; x = some importance; blank = no importance.
Agro-ecological zone
Factors Arid Semi-arid Subhumid Humid Highlands
Crop residue
Helminthosis
Trypanosomosis
Low productivity
1.7 References and reading materials
FAO (Food and Agriculture Organization of the United Nations). 1990. Production Yearbook 1989.
Volume 43. FAO, Rome, Italy. 346 pp.
ILCA (International Livestock Centre for Africa). 1987. ILCA’s Strategy and Long-term Plan. ILCA,
Addis Ababa, Ethiopia. 99 pp.
ILCA (International Livestock Centre for Africa). 1991. A Handbook of African Livestock Statistics.
Working document 15. ILCA, Addis Ababa, Ethiopia. 66 pp.
Jahnke H.E. 1982. Livestock Production Systems and Livestock Development in Tropical Africa.
Kieler Wissenschaftsverlag Vauk, Kiel, Germany. 253 pp.
Lebbie S.H.B., Rey B. and Irungu E.K. 1994. Small Ruminant Research and Development in Africa.
Proceedings of the Second Biennial Conference of the African Small Ruminant Research
Network AICC, Arusha, Tanzania, 7–11 December 1992. ILCA (International Livestock Centre
for Africa)/CTA (Technical Centre for Agricultural and Rural Co-operation). ILCA, Addis
Ababa, Ethiopia. 268 pp.
McIntire J., Bouzart D. and Pingali P. 1992. Crop–Livestock Interaction in Sub-Saharan Africa. The
World Bank, Washington, DC, USA. 246 pp.
Winrock International. 1992. Assessment of Animal Agriculture in Sub-Saharan Africa. Winrock
International Institute for Agricultural Development, Morrilton, Arkansas, USA. 125 pp.
14
Module 2: Livestock farming systems research
2.1 Performance objectives
Module 2 is intended to enable you to:
1. List the key concepts of livestock farming systems research.
2. List the steps in the farming systems research process.
3. Describe the components of the livestock systems research process.
4. Describe the four main categories of livestock farming systems.
2.2 Introduction
The farm is a complex system. Often, researchers design experiments without considering the complexity
of the system. This chapter is designed to give you an overview of this complex system and ways to
approach your research with the needs and goals of resource-poor farmers in mind.
Livestock research in sub-Saharan Africa needs to address the factors that have led to the poor
performance of the subsector during the last two decades: inappropriate policies, institutional
deficiencies, resource constraints and a failure to design experiments and technologies applicable to the
farmer's situation; we will address the last factor in this chapter. The systems approach to research is
recognised as the most appropriate means of gaining knowledge of the factors which influence production
decisions at the farm level.
This chapter will not cover methods of data collection. Such information is available in two
working papers published by ILCA during 1990 (see Box 2.1). This chapter will provide you with a basic
understanding of the farming systems research approach and a few references to guide you if you wish to
go into more detail.
15
2.1 Performance objectives
2.2 Introduction
2.3 Livestock farming systems: Key concepts
2.4 Livestock farming systems conceptual model
2.5 Livestock farming systems process
2.6 Participatory rural appraisal (PRA)
2.7 Categories of livestock farming systems research
2.8 Exercises
2.9 References and reading materials
16
Box 2.1: Livestock farming systems research: Methods and type of data to be
collected for various themes
To conduct research on livestock systems, you need to acquire knowledge and skills. A holistic
approach is taken, that requires an understanding of the  relationship between the  livestock
subsystem and the rest of the farming system. Hence, a study of the livestock subsystem cannot be
undertaken in isolation, and the researcher must also investigate the relevant parts of the cropping
system.*
A. Methods
• Informal surveys
• Developing hypotheses about systems relationships/linkages
• Questionnaire design
• Sampling methods
• Methods of data collection for household budgets
• Methods for animal production: aging by dentition, recall surveys, progeny history method,
direct observation and measurement
• Methods for assessing the composition of feed, sampling crop residues
• Sampling to detect disease presence and prevalence rates
• Methods for estimating livestock numbers, offtake and acquisition and to understand
marketing functions
• Breeding practices
• Organisation, presentation and analysis of results
B. Type of data to be collected
• Baseline data
? general characteristics of the study area
? specific characteristics of farmer/pastoralist within each target group
? available technologies which might be applicable to the problems identified
• Data for labour input
? labour available to the household
? labour use over a period of time by age and sex categories
• Household budgets and assets
? assets inventory data
? income data
? expenditure data
• Animal production
? inter-species composition of the production system (households) livestock holding
? herd/flock structure
? reproductive performance
? mortality Box 2.1 cont...
2.3 Livestock farming systems: Key concepts
A farming system is a set of interacting activities which can be synergistic or competitive.
The activities are:
• managed towards satisfying the farmer’s objectives (short- or long-term objectives)
• decided by one level of management (farmer or household)
• compatible with the environment.
The farmer in this complex system chooses among various resources (e.g. labour, land, capital etc)
to produce outputs. Resources and outputs could be:
• synergistic, e.g. nitrogen-fixing legumes contribute nitrogen to other non-fixing crops
• competitive, e.g. crop residues for feeding sheep vs crop residues for mulching the soil.
Farmers manage activities in the farming system with a set of goals in mind. A decision can be made
by the farmer or collectively by the family. This decision is determined by the objectives and by the risk
the farmer is willing to take. When the farmer has complex objectives, he/she makes complex decisions.
The farming system can best operate when it is geared towards stability. Therefore we can infer that:
• a farming system is more than the sum of its components as there is interaction among the
components
• introducing an innovation modifies the whole system; it shifts the balance through synergy or
competition.
In order for research results to suit complex farming systems and to be accepted by farmers, it is
essential that researchers understand the system. For the research to be effective (i.e. to meet the needs of
the farmers) it should take into account the following issues:
17
Box 2.1 cont...
? growth and weight gain
? outputs (meat, milk, hides, skins and traction)
• Animal nutrition
? Production effects: liveweight gain, condition scores, traction power and milk and wool
production
? Amount of feed consumed: feed intake
? Composition of feed consumed: oesophageal or rumen fistula samples, faecal samples
and grazing behaviour studies
? Feed data: digestibility, energy value of feed and crude protein content
• Animal health
? Passive data: veterinary records, records of diagnostic or research laboratories and data
from slaughter houses, quarantine stations and checking points
? Active data: data supplied by farmers, data from laboratory analysis and productivity
data
* For further information, see ILCA (1990).
• Research should be geared towards solving the clients’ problems, therefore, a bottom-top approach
is best.
• Research should be viewed as an investment for development.
• Information flows between researchers, development agents and farmers should be redefined
(Figure 2.1).
• Researchers should recognise the rationale behind farmers’ decisions.
• Researchers should consider small farms as complex systems managed by farmers.
• The research should be multidisciplinary.
In the past farmers were used to test information without any thought being given to their needs.
Recent research has shown that the information flow in the interactive model (Figure 2.1) is essential for
the scientist to exchange information with the farmer. Starting and ending with the farmer (farmer-to-
farmer model) allows for direct exchange of information between the scientist and the farmer. The ideal
way to exchange information is to involve the three parties: farmer, researcher and the development agent.
It is important for the scientist to understand why the farmer does what he/she does because this allows
research to provide the solutions needed by the farmer.
2.4 Livestock farming systems conceptual model
As researchers our objectives for following a farming systems research approach are to:
• gain knowledge about cropping and livestock systems
• conduct research, based on the knowledge of the system, that will provide the information
necessary to design interventions for improving the performance of the livestock system.
Implementing these objectives compels us to look at problems using a more holistic approach. It
also leads us to use multidisciplinary teams to conduct the research. A conceptual model was developed by
Lhoste (1986) to describe these relationships (Figure 2.2).
18
Figure 2.1. Models for information flow among researchers (R), farmers (F) and development agents (D).
B31B29
B32B29
R D F
RDF
R
F
D
B43B6cB61B73B73B69B63B61B6c B6dB6fB64B65B6c
B49B6dB70B72B6fB76B65B64 B63B6cB61B73B73B69B63B61B6c
B6dB6fB64B65B6c
B49B6eB74B65B72B61B63B74B69B76B65 B6dB6fB64B65B6cB33B29
Livestock systems are part of the agricultural systems hierarchy (Figure 2.3). Studying livestock
systems in isolation will make it difficult to see or deal with the total picture. This approach does not
necessarily conflict with scientists’ specialisation in specific areas.
19
B4cB69B76B65B73B74B6fB63B6b B66B61B72B6dB65B72
B54B72B69B62B65B2c B66B61B6dB69B6cB79
B41B73B73B6fB63B69B61B74B69B6fB6eB2c B67B72B6fB75B70B73
B41B67B65B6eB74B73 B64B65B63B69B73B69B6fB6e B6cB65B76B65B6c
B4eB65B65B64B73B2fB70B72B6fB6aB65B63B74
B53B74B61B74B75B74B65
B54B65B72B72B69B74B6fB72B79
B53B74B72B75B63B74B75B72B65
B50B72B69B6dB61B72B79 B70B72B6fB64B75B63B74B69B6fB6e
B55B73B65B2fB61B6eB69B6dB61B6c
B45B76B6fB6cB75B74B69B6fB6eB2fB71B75B61B6eB74B69B74B79B2fB73B65B61B73B6fB6eB2f
B71B75B61B6cB69B74B79B2fB79B65B61B72B73
B48B65B72B64B2fB66B6cB6fB63B6b
B41B6eB69B6dB61B6cB73B2fB73B70B65B63B69B65B73B2fB62B72B65B65B64B73
B4eB75B6dB62B65B72B2c B63B6fB6dB70B6fB73B69B74B69B6fB6e
B44B79B6eB61B6dB69B63B73
B45B76B6fB6cB75B74B69B6fB6e B6fB76B65B72 B74B69B6dB65
B54B69B6dB65B54B69B6dB65
B50B72B6fB64B75B63B74B69B76B69B74B79
B48B65B72B64 B6dB6fB76B65B6dB65B6eB74 B73B74B72B61B74B65B67B69B65B73
B53B70B61B63B65 B6dB61B6eB61B67B65B6dB65B6eB74
B4cB61B6eB64 B74B65B6eB75B72B65 B73B79B73B74B65B6d
B46B6fB64B64B65B72 B73B79B73B74B65B6d
B46B6cB6fB77 B6fB66 B6fB72B67B61B6eB69B63 B6dB61B74B74B65B72
B46B65B65B64B69B6eB67B2fB62B65B68B61B76B69B6fB75B72
B53B70B61B74B69B61B6c
B48B65B72B64 B70B72B61B63B74B69B63B65B73
B2fB6cB69B76B65B73B74B6fB63B6bB2f
B65B6eB68B61B6eB63B65B64 B76B61B6cB75B65
Agrarian system
Cropping
system
Fodder
system
Livestock
system
Farming system
Figure 2.3. Systems hierarchy.
Source: Lhoste (1986)
.
Figure 2.2. General conceptual model for livestock systems.
2.5 Livestock farming systems process
Livestock farming systems research methodology may be influenced by the nature of the particular
system under study but the process involved for all types of farming systems is similar. This includes the
terminology used and the sequencing of the process. The following features are prerequisites for the
farming systems approach:
• no preconceptions about the system
• it should be improvement oriented
• all factors affecting the household should be examined
• it should involve multidisciplinary teams
• farmer participation is essential
• solutions should be evaluated in biological, environmental, economic and political terms.
The livestock farming systems process involves four main interlinked phases, namely the
diagnostic, design, testing and extension phases.
Figure 2.4 illustrates the sequencing and relationships among these phases. We will describe each
phase in more detail. Comparing the bottom-up and top-down approaches to research further illustrates
the relationship between the phases (Table 2.1). Farmer involvement is central to the bottom-up approach.
It allows us to know the farmer’s capabilities when improving existing technologies or proposing new
ones. The ability of the farmer to economically afford or socially accept the new technology is easier to
identify in this approach. The top-down approach, used by most researchers in the past, contributed to the
high rate of failure in adoption of technologies by farmers. Solutions generated through the top-down
approach may not address the needs of the farmer.
20
B44B69B61B67B6eB6fB73B69B73 B6fB66 B63B6fB6eB73B74B72B61B69B6eB74B73 B4fB6e B66B61B72B6d
B44B65B73B69B67B6e B6fB66 B73B6fB6cB75B74B69B6fB6eB73
B54B65B73B74B69B6eB67 B6fB66 B73B6fB6cB75B74B69B6fB6eB73
B44B69B61B67B6eB6fB73B69B73 B6fB66 B6eB65B77
B63B6fB6eB73B74B72B61B69B6eB74B73
B4fB6e B66B61B72B6d
B4fB6e B73B74B61B74B69B6fB6e
B61B6eB64
Figure 2.4. Farming system research process.
Table 2.1. Approaches to research.
Bottom-up approach Top-down approach
Location Activity Location Activity
Farm Diagnosis Research station New technology
Research station Alternative technology
Farm Technology testing Farm Farm testing
2.5.1 The diagnostic phase
Diagnosis offers answers to the following questions:
1. What is going on? What is the farmer doing? What goes into the system?
2. What is the problem?
3. What is the cause of the problem?
4. What could be done to solve the problem?
Knowing the answers to the first three questions will allow you to (i) address question 4 using a multi-
disciplinary approach, (ii) focus your research, (iii) save time and (iv) use limited resources fruitfully.
The main objectives of the diagnostic phase are:
• To describe the production system
Researchers use this phase to highlight the circumstances and practices of the representative
farmers. Understanding the household activities and the wider features of the system is necessary
for technology design.
• To identify constraints to resource use
Farmers’ main objectives are to produce for household consumption and for sale and to avoid risk
and save for hard times. It is important to identify the factors which prevent farmers from meeting
these objectives.
• To understand what causes the constraints
Unless the underlying forces pertaining to the constraints are understood, solutions will always be
elusive.
• To determine opportunities for further improvement
This involves screening available technologies for their suitability. This information will allow the
researcher to decide on an appropriate course of action. Sometimes improvement on farmers’ practices
can be a solution, e.g. prevalence of parasites in spite of dipping may be due to resistance of ticks to the
dipping solutions, improper dipping or reduced animal resistance due to underfeeding. Surveys of
indigenous knowledge are recommended as part of meeting this objective.
Diagnostic tools
A few diagnostic tools will be briefly described. For these tools to be effective researchers need to use their
feet, ears and eyes as well as brains to diagnose farmers’ problems.
1. Surveys
Surveys can be formal or informal and comprise one or multiple visits. To meet the objectives of surveys a
researcher should:
• have adequate knowledge about the target population of farmers
21
• use good questionnaire techniques
• use good sampling techniques
• use skilled surveyors with adequate logistic support (transportation, incentives, knowledge of the
area, knowledge of the local language etc)
• involve or consult with a biometrician.
The number of observations and amount of data per observation are important factors in determining data
quality. An equilibrium should be reached so that the number of observations match the available resources.
The information collected during a survey should be:
• qualitative and quantitative knowledge on the functioning of the livestock system
• knowledge on variability of livestock systems
• quantitative knowledge on systems performance (function of the farmers’ memory and confidence).
Lengthy interviews lead to loss of concentration on the part of the farmer which affects the accuracy
of the answers given. Respondent fatigue can also occur when surveys are conducted over a long time.
Formal survey. In a formal survey, a well-structured or semi-structured questionnaire is developed. The
questionnaire contains a number of questions to be answered by the farmer. The farmers are randomly
selected and the survey is designed to collect data on particular aspects of the production system. Formal
surveys can also be used to redefine target groups and identify priorities for research.
Informal survey. In this type of survey no structured questionnaire is needed; information is gathered in an
informal way. One needs only to remember the key issues in asking the questions. The reliability of the
information depends on the questions and the way in which they are asked. Farmers are busy and in both
formal and informal surveys we must not waste their time unnecessarily. The best approach is a
combination of both structured and unstructured questionnaires. Both formal and informal surveys should
be complemented by observations made during farm visits.
Single-visit survey. In some cases one visit may be enough to gather enough information from a farmer.
This is especially so when the farmer is busy during the peak of the seasonal calendar (when the maximum
number of activities are being conducted at the same time). This approach is useful for farm inventories
and qualitative information.
Multiple-visit survey. Multiple visits are used when we want to verify information or we require more in-
depth information. This approach establishes better rapport between researchers and farmers.
Researchers have a chance to show interest in the farmer’s personal life but they should be careful not to
probe matters of privacy or taboo (e.g. income, number of children) directly. This type of survey is
particularly useful for semi-structured interviews, and for when farmers are less busy.
2. Farm monitoring
In this method the team of monitors visits farmers over an extended period and records detailed data, e.g.
reproduction performance, changes in environment and livestock dynamics. This approach involves
collecting continuous data. Depending on the nature of investigation, data collection can take one month
to several years. The main disadvantage of this method compared with surveys is its high cost. Its
advantages over the survey method are that it can be used to double-check information from rapid surveys
and it relies less on the farmer’s memory (particularly related to the performance of animals).
To monitor flocks of small ruminants on a farm:
• choose flocks to be monitored
• collect information on resources (territory, flock and family)
• conduct retrospective studies (female productive history)
• pay regular visits to collect data.
22
2.5.2 The design phase
As mentioned above, analysis of information during the diagnostic phase allows us to identify the
problems and their causes (if lucky). Diagnosis is followed by designing a set of research activities to
solve the problems (see Biometrics Module). The purpose of designing solutions is to develop
technologies that will address the constraints and to adapt the technologies to suit the situation.
Considerations during designing
• Nature of the farms which vary in location and management
• Farmer’s capabilities: education and ability to handle technology
• Degree of management control: researcher-managed trials (researcher plans and executes),
researcher–farmer-managed trials (researcher plans and farmer executes) and farmer-managed
trials (farmer plans and executes)
• Type of trial
• Number of test animals
• Type of statistical analysis
• Available resources (time, money and expertise)
Research should focus on technologies which are compatible with the resources and objectives of
the farmer and consistent with the features of the system.
Major rules for designing solutions
• Adapt technologies already developed
• Evaluate all feasible options
• Exploit possibilities in the system
In the short term it is useful to exploit existing options. In any production system major long-term
increases can only be obtained by overcoming constraints to production. Whether exploiting existing
technologies or developing new ones to improve the system, the economic evaluation is important, as
farmers will not use solutions which are not profitable.
2.5.3 The testing phase
The research agenda formulated in the design phase will comprise trials to be carried out at research
stations (on-station trials) or in farmers’ fields (on-farm trials). The trials may be short or long term.
Farmers and policy makers are eager for results and research impact. To sustain their participation, we can
demonstrate some spin-offs so that they can look forward to the final result. To maintain the interest of
farmers, especially in long-term projects, consistent monitoring and spot analysis of data will help identify
opportunities for mid-way interventions. Many researchers delay analysis until the end of the project thus
missing many opportunities for spin-offs. Looking at the trend of results can help identify some valuable
solutions which will satisfy the farming community.
Considerations
• Farmers should be involved in the trial process and evaluation of technologies on farm
• Experiments should be simple
• The involvement of farmers outside the target area should be considered
• Farmers’ reactions to the alternative technologies should be monitored
23
On-station experimentation
On-station research is needed to arrive at the best alternative among the many tested options. The best
option is then tested on farm. However, on station we can only investigate biological factors. Social
factors that are of major importance are best investigated on farm.
On-farm experimentation
Simple designs should be used. For each design there are two treatments: the new technology and the
farmer’s usual practice (control). Testing can be done on one or many farms. We should test alternative
technologies in which we have confidence based on data from on-station experimentation. We should be
careful that our technologies are not developed for high resource input as this will lead to their rejection by
farmers, e.g. crossbred cows tested under good management on station will not perform as well where the
farmer has no access to resources similar to those available at the research station.
We need to collect data that will lead to technologies that are technically and socially acceptable
and allow us to evaluate performance biologically and economically.
2.5.4 The extension phase
The extension phase occurs after the most appropriate technology option has been identified and tested
and found to be acceptable to farmers. The objective of this phase is to extend the option to as many
farmers as possible. Monitoring and evaluation are needed to assess impact, but this will be largely
outside the scope of a research station and is the responsibility of the extension service. It is essential to
monitor:
• production levels and the use of inputs
• resource allocation (e.g. land for crops vs animals)
• institutions (e.g. markets)
• farmers’ options—acceptance or rejection of technologies.
Even though this phase is led by the extension service, it requires input from farmers and
researchers; such contact should be maintained to ensure feedback.
2.6 Participatory rural appraisal (PRA)
Participatory rural appraisal is:
• An opportunity to learn from the target people about their situation. Target groups can be involved
at several points in the process:
? investigation, identification and prioritisation of needs
? planning and implementation
? monitoring and evaluation.
• A method that gives researchers and farmers an opportunity to exchange knowledge. This will
enable researchers to design programmes that will meet the needs of the farmers.
2.6.1 Why PRA?
1. The success rate of agricultural projects has been unsatisfactory (a ratio of 3 successful to 55
failed projects in the 1960s) because:
• projects failed to adequately acknowledge the interactions among the component parts of farming
systems including ecological, biological and socio-economic elements
24
• there was limited or no participation by the target groups in conceiving, planning, implementing,
monitoring and evaluating the projects.
2. Evolution through the decades
• It started with rapid rural appraisal (RRA), a technique which ignored participation of the target
group.
• Rural project appraisal was developed as a new form of RRA; it involved rapid assessment
procedures and has the same limitations as RRA.
• PRA overcame the limitations of the two techniques by involving the target group at every stage of
the research process.
2.6.2 PRA techniques
Problem identification
PRA techniques are preliminary exercises that provide information about constraints facing the farming
system. They complement more formal methods, e.g. surveys.
• Open discussion (non-structured)
Problems are identified through open non-structured discussion. The main features of the
discussion are:
? the clients lead the discussion (‘we can do it’ attitude) and the researcher acts as a facilitator
only
? the researcher steps in only to seek clarification; otherwise he/she watches, listens and learns.
• Semi-structured (topic investigation)
? specific answers are obtained for specific questions
? the researcher’s role is limited to introduction and organisation
? target group leaders take lead to ensure vigour in the discussion
? the researcher’s role in the discussion is catalytic.
Gathering data
There is always a wealth of information embedded in a variety of resources. PRA is a quick way to make
this information available to the researcher before project planning. To simplify the collection of data, it is
classified into many types depending on the source, location, time and social considerations.
• Secondary data: publications, maps, census reports, documents, especially grey literature, and
aerial photographs.
• Field-generated data (primary data): This is spatial data collected jointly with the farmer.
? Sketch maps of the farm are drawn to show resources, activities, opportunities and problems.
? A transect of the farm is drawn to show items the sketch map may not have included.
? Farmers and their families draw farm sketches to show land resource and farm size, and to give
an idea about operational strategies.
• Time-related data (temporal)
? A time line is developed showing important local, national and international events observed
by the community.
25
? A trend line is developed showing important changes in the community, why there are changes
and the community’s attitude to change (since research aims at change, such information is essential).
? Seasonal calendars are developed to show seasonal problems or opportunities in average years.
• Social data (people-related information)
? Farm household (HH) interviews are conducted to gather socio-economic information and
characteristics of a particular community or farm.
? Information is collected about institutions to provide knowledge about various groups and
organisations within the community (including churches, schools, women groups etc), to
show insight into the relationship among the institutions and to determine how the community
views its institutions (e.g. ranking the contribution of institutions to development).
Ranking and scoring
The impact of problems at the farming systems level can be of varying magnitude. Ranking and scoring
are ways of prioritising different interventions and technical solutions that are relevant to and adoptable by
farmers. This involves placing things in order of importance at a particular time. Ranking could be done
for problems and opportunities.
2.6.3 Virtues of PRA
• Participatory and hence client oriented
• Enabling, as farmers feel the researcher is interested in them
• Empowering because the clients say what needs to be done and the researcher listens
• Semi-structured, leading to local participation in a systematised fashion
• Flexible, gender sensitive and applies to all classes
• Iterative and thus exploits all alternatives and possibilities
• Exploratory and hence looks at all possible situations
• Involves training and orientation which enable the farmer to rationalise and evaluate the products
(e.g. weighing small ruminants, weighing milk)
• Focused
• Ensures sustainability
2.7 Categories of livestock farming systems research
Category 1: Farming systems research sensu stricto
Farming systems research in this sense studies the farming system as is. The studies are academic
and scholarly rather than practical.
Category 2: On-farm research
This is research on farmers’ fields with farmers participating and with a farming system
perspective. In this category the researcher:
• assumes that only the farmers’ experiences can reveal their needs
• isolates a subsystem and studies it in enough depth to obtain new perspectives
• proceeds to on-farm experimentation with the farmers’ contribution.
26
Category 3: Development of new farming systems
The objective is to invent, test and exploit new systems. This involves radical restructuring of
farming systems instead of the step-wise change followed in Category 2, e.g. developing an intensified
dairy system in place of an extensive pastoral system.
2.8 Exercises
1. Give examples from livestock–crop interactions that show (i) synergistic effects of resources
and outputs (ii) competitive effects of resources and outputs.
2. Lhoste (1986) developed a global conceptual model for livestock systems research. Does this
model fit the themes from your region? Describe the similarities and the differences.
3. The livestock farming systems research process has four phases. List these phases in sequence
and explain what will happen to the value of your research results if any of the phases is
omitted.
4. List the categories of livestock farming systems research. Which of these categories suits the
available human, financial and physical resources in your situation and why?
2.9 References and reading materials
ILCA (International Livestock Centre for Africa). 1990. Livestock Systems Research Manual.
Volume 1. ILCA Working Paper 1. ILCA, Addis Ababa, Ethiopia. 287 pp.
ILCA (International Livestock Centre for Africa). 1990. Livestock Systems Research Manual.
Volume 2. ILCA Working Paper 1. ILCA, Addis Ababa, Ethiopia. 55 pp.
Lhoste P. 1986. Le diagnostic sur les systèmes d’élevage. In: Landais E. (ed), Actes de l’atelier sur les
méthodes pour la recherche sur les systèmes d’élevage en Afrique intertropicale, Mbour
(Sénégal), 2–8 Février 1986. Etude et Synthèses de l'IEMVT (France) 20. IEMVT (Institut
d'élevage et de médecine vétérinarie des pays tropicaux), Maison-Alfort and ISRA (Institut
sénégalais de recherches agricoles), Dakar, Senegal. pp. 39–59.
27
Module 3: Reproduction in African small ruminants
3.1 Performance objectives
Module 3 is intended to enable you to:
1. Define and describe the basic concepts of small ruminant reproductive physiology.
2. Compute annual reproductive rates for sheep and goats using field production data.
3. List and compute flock production indices.
4. Describe factors that lead to reproductive wastage in small ruminants.
5. Understand and use techniques for controlling reproduction in ewes.
6. List the methods available to improve the reproductive performance of sheep and goats.
3.2 Introduction
In sub-Saharan Africa sheep and goats provide food and fibre, and produce manure which contributes to
agricultural output. The majority of sheep and goats are raised by smallholder farmers on a small scale;
profitability is low due to low market weight, overall low reproductive efficiency and high mortality.
Small ruminant productivity results from the interaction of reproductive efficiency, growth rate, yield and
quality of desired products. Therefore good reproductive efficiency is a pre-requisite for efficient
production systems.
3.3 Physiology of reproduction
The basic concepts of small ruminant reproductive physiology are fairly similar across breeds but
differences may occur due to genetic and/or environmental factors. In general, small ruminants in Africa
are year-round breeders. As a result these animals lamb or kid all year round except where breeding is
regulated through specific management systems such as controlled mating. This section will consider the
physiological events as they occur during a full reproductive cycle since it is these events that affect the
overall reproductive output which in turn affects flock productivity.
29
3.1 Performance objectives
3.2 Introduction
3.3 Physiology of reproduction
3.4 Reproductive performance
3.5 Management of reproduction
3.6 Exercises
3.7 References and reading materials
3.3.1 Basic concepts of reproduction
Puberty
Puberty is the time when reproductive organs become functional. In the female, it is defined as the time
when the first functional oestrus takes place and the earliest age at which reproduction can occur. In some
cases, however, ovulation without oestrus occurs before the first observed oestrus. At other times, animals
may show oestrus which is not associated with ovulation. The age at which puberty is attained is
determined largely by genotype or breed, nutrition, season and other environmental factors (e.g. climate).
As a consequence, large variations occur between and within breeds. The age at first lambing may also
vary as a result. Full development of the reproductive organs is not reached until puberty.
Males manifest a marked increase in testis weight and volume with age which coincides with the
appearance of the primary spermatocyte and separation of the preputial sheaths. Among females too, the
reproductive organs increase in size in line with general body growth. Thus faster growing animals will
reach puberty earlier than slow growing animals. In both males and females, puberty generally occurs
when animals reach 45–60% of the mature weight for the breed.
Pubertal events are under the influence of pituitary gonadotropic hormones—follicle stimulating
hormone (FSH) and luteinising hormone (LH). The anterior pituitary gland of the prepubertal animal
contains high amounts of gonadotropins but the level of these hormones is low in the plasma during this
phase of growth. The responsiveness of the gonads (testes and ovary) to gonadotropin only increases
during growth and as puberty approaches. The hypothalamus also plays part in modulating activities
leading to puberty.
Oestrus and ovulation
Oestrus, or heat, is defined as the time when the female is receptive to the male. The critical marker is the
male partner (ram or buck). The male is used to detect females in heat by applying dye markers to its belly.
This dye rubs off on the backs of females. The immobility of the receptive female is most often used as the
critical criterion for identifying oestrus. Oestrus is difficult to detect among ewes alone as they do not
exhibit homosexual behaviour. While the short duration of oestrus may make it difficult for farmers to
detect heat, the overall duration is adequate to ensure fertility if the ram:ewe ratio is favourable and rams of
adequate fertility and libido are running with the flock. Using a teaser ram is the most practical way to
detect oestrus in sheep since other observable signs of oestrus (swollen vulva, mucous secretions etc) are
most often absent or anatomically difficult to detect. A teaser ram is a castrate that mounts the ewes but
does not mate with them.
The ovarian cycle of the ewe is shown in Figure 3.1. The oestrus cycle of sheep is divided into the
follicular and luteal phases. In general the oestrus cycle is said to start on the second day of the follicular
phase. This phase lasts about three days during which small follicles develop on the ovaries. The luteal
phase starts at ovulation and lasts 14 days. An extended ovarian cycle may result from prolonged follicular
or luteal phases. The hypothalamic–pituitary–gonadal axis controls the occurrence of oestrus and the
duration of the cycle (Figure 3.2). The hypothalamus releases the luteinising hormone releasing hormone
(LHRH) which stimulates the release of FSH and LH. The gonadal steroids oestrogen and progesterone
regulate the production of LHRH. FSH and pulses of LH stimulate the development of follicles on the
ovary. Once the follicles mature they produce and secrete oestrogen, especially oestradiol which is
essential for starting sexual receptivity.
30
Example: Puberty in Menz sheep in the Ethiopian highlands
A study (Mukasa-Mugerwa et al 1993) that looked into the reproductive performance of Menz sheep in
the Ethiopian highlands reported that ewe lambs attain puberty (first oestrus) at 10 months of age and
16.9 kg mean weight or 56% of mature body weight. The onset of puberty was earlier with higher
weaning weight and it is probable that poor nutrition can delay puberty by one season.
Figure 3.1. The ovarian cycle of ewes.
Figure 3.2. Hormonal control of oestrus in ewes.
31
B50B72B6fB65B73B74B72B75B73
B4fB65B73B74B72B75B73
B41B6eB6fB65B73B74B72B75B73
B44B69B6fB65B73B74B72B75B73 B4dB65B74B6fB65B73B74B72B75B73
B50B72B65B67B6eB61B6eB63B79
B32B44B31B31 B2d B31B32 B44
B32B44
B4eB53
B53
B53
B56B61B72B69B61B62B6cB65
B31B34B35B44
B43B6fB6eB63B65B70B74B69B6fB6e
B53 B3d B53B65B61B73B6fB6eB61B6c B4eB53 B3d B4eB6fB6eB2dB73B65B61B73B6fB6eB61B6c B44B3dB44B61B79B73
Hypothalamus
Anterior pituitary
Ovary
Follicle
Corpus
luteum
Uterus
LH releasing hormone
Follicle stimulating
hormone
Luteinising
hormone (LH )
Prolactin
Oestrogen
Progesterone
Prostaglandins
LH is initially secreted in small amounts but large outbursts of the hormone follow and this starts the
ovulation process. The follicles collapse and develop into a corpus luteum which produces progesterone
that inhibits other follicles from developing. In the normal oestrus cycle, the regression of the corpus
luteum is controlled by prostaglandins produced by the uterus. The levels of LH, oestradiol and
progesterone in the blood of a normal ewe are shown in Figure 3.3. The ewe in oestrus becomes restless
and shows increased activity. The sexual behaviour of the ewe can be divided into two phases: searching
for the sexual partner and the sequence of interactions leading to copulation. Among the activities
exhibited are sniffing, nudging, mounting and copulation.
.
The duration of oestrus in the ewes is 32 to 50 hours for sheep and goats but is highly variable among
breeds. During or following oestrus, ovulation of the mature follicle occurs under the influence of LH. Time of
ovulation after the onset of oestrus has been estimated as 30 hours for sheep and 30–36 hours for goats.
Oestrus cyclicity and synchronisation
Sheep in the tropics tend to experience oestrus cycles throughout the year, every 14 to 19 days with a mean
of 16 days (Figures 3.4 and 3.5). The further one gets away from the equator, the longer the oestrus cycle, a
sign of the effect of photoperiod on the ewe reproductive cycle. In a study in the Ethiopian highlands 48
Menz ewes were observed over one year (Mukasa-Mugerwa et al 1993). The ewes were maintained with a
vasectomised ram. Results showed that 22% of the cycles were short (<13 d), 56% were normal (14–19 d),
11% were long (20–26 d), 8% were missed (27–40 d) and only 3% indicated anoestrus (>40 d).
Exceptionally long cycles are often associated with particular seasons of the year.
Herdsmen need to observe oestrus to effectively manage mating and lambing events, but the
observation activity is time consuming. To reduce the time spent in observing oestrus, it is essential to
synchronise it. This is often critical when mating and lambing need to be matched with availability of
labour or nutritional sources.
Oestrus synchronisation is best achieved by using progestogen or prostaglandin F2?. Progesterone
can be administered by daily injections, orally, as subcutaneous implants or in intravaginal sponges. When
intravaginal sponges are placed in the vagina for 12 days, most of the ewes come into heat within 2–3 days
of sponge removal. Using intravaginal sponges has the advantage of ease of application, no detrimental
effect on pregnancy and comparable fertility rates as for untreated controls. In addition to these treatments
to induce oestrus, prostaglandin F2? (uterine luteolycin) can be used to very effectively synchronise
oestrus in sheep with fairly good fertility rates.
32
B33B35
B33B30
B32B35
B32B30
B31B35
B31B30
B35
B30
B33
B32B2eB35
B32
B31B2eB35
B31
B30B2eB35
B30
B35B30B31B36B31B35B31B30B35B30B31B36B31B35
B4fB76B75B6cB61B74B69B6fB6e B4fB76B75B6cB61B74B69B6fB6e
B45B45
B44B61B79 B6fB66 B6fB65B73B74B72B75B73 B63B79B63B6cB65
B4cB48 B28B6eB67B2fB6dB6cB29
B4cB48 B50B72B6fB67B65B73B74B65B72B6fB6eB65 B4fB65B73B74B72B61B64B69B6fB6cB45 B3d B6fB65B73B74B72B75B73
B33B2eB35
B50B72B6fB67B65B73B74B65B72B6fB6eB65 B6fB72
B4fB65B73B74B72B61B64B69B6fB6c B28B6eB67B2fB6dB6cB29
Source: Gatenby (1986).
Figure 3.3. Hormone patterns during the oestrus cycle of ewes
Follicular populations, ovulation rate and litter size
Small ruminants have the genetic ability to carry more than one litter in a pregnancy. The ovary of the adult
ewe has a large pool of very small or primordial follicles from which it has been estimated that 3–4
follicles develop every day. These follicular populations appear to influence the ovulation rate which is a
primary factor limiting litter size. Since ovulation rates are genetically controlled, improvement in this
trait can be and has been achieved by selection. In breeds with low litter size, increased twinning can be
achieved by using pregnant mare serum gonadotrophin (PMSG), usually administered after sponge
withdrawal.
33
60
50
40
30
20
10
0
JFMAMJJASOND
Number of observations
B4dB6fB6eB74B68
N = 502 oestruses
40
30
20
10
0
6055504540353025201510
5
<
Duration of c ycleindays
Short (<14 d) = 23.7%
Normal (14 - 19 d) = 57.1%
Long (20 - 26 d) = 10.2%
Missed (27 - 40 d) = 6.6%
Anoestrus (>40 d) = 2.4%
B50B65B72B63B65B6eB74B61B67B65 B6fB66 B63B79B63B6cB65B73
N = 24 ewes
N = 245 oestrus cycles
Source: Mukasa-Mugerwa and Lahlou-Kassi (1995).
Figure 3.5. Frequency distribution of the duration of oestrus cycle length in Menz ewes.
Figure 3.4. Monthly variation in number of oestrus cycles of Menz ewes in Ethiopia.
Figure 3.6. Monthly distribution of lambing activity in Menz sheep under natural mating.
Breeding and artificial insemination
Females can be bred naturally or artificially. Under natural mating, a ratio of 50 to 1 ewes/ram can be used.
The ram must be mature and have good libido, particularly when the ewes are synchronised into oestrus.
Under artificial mating, a larger number of ewes may be bred by the same ram. Ram fertility is generally
related to the production of large numbers of high quality spermatozoa.
For artificial insemination (AI) semen is collected, processed and stored before insemination.
Collected semen is used immediately or stored in liquid nitrogen as frozen pellets.
Insemination of sheep is preferably done by the intra-uterine method; the intra-cervical method can
also be used. If AI is done intra-cervically, a large number of spermatozoa should be used. To ensure
success, insemination should be done at the onset of oestrus in goats (1–12 hours) and at the end of oestrus
in sheep. The time of ovulation, duration of sperm transport in the female genital tract and survival time of
the male and female gametes are factors that should be considered to achieve success in AI. Using
hormones to synchronise oestrus also plays an important part in determining the timing of AI. Double
insemination (12 hours apart) may be attempted to improve success. Usually sperm deposition is made
48–58 hours after sponge removal when PMSG is used.
34
Example: Effect of environment and nutrition on manifestation of oestrus
Even though ewes in the Ethiopian highlands manifest oestrus throughout the year there is a
depression coinciding with the peak of the rainy season (July to September). Progesterone profiles
of the ewes reveal that failure of certain ewes to show oestrus in the wet period is, however, due to
increased frequency of ovulations that are not accompanied by oestrus (silent heats). Rams also
experience a depression in their fertility during the rainy season (Mukasa-Mugerwa and Lahlou-
Kassi 1995). This reproductive pattern is reflected in the pattern of lambing activity in Menz sheep
which also peaks during the rainy season (Figure 3.6).
B31B35
B31B30
B35
B30
B4aB61B6e B46B65B62 B4dB61B72 B41B70B72 B4dB61B79 B4aB75B6e B4aB75B6c B41B75B67 B53B65B70B74 B4fB63B74 B4eB6fB76 B44B65B63
B33B35B30
B33B30B30
B32B35B30
B32B30B30
B31B35B30
B31B30B30
B35B30
B30
B25 B4cB61B6dB62B69B6eB67B73
B28B6eB3dB31B37B30B29
B52B61B69B6eB66B61B6cB6c B28B6dB6dB29
B52B61B69B6eB66B61B6cB6c B28B31B39B37B39B2dB38B39B29 B4cB61B6dB62B69B6eB67 B61B63B74B69B76B69B74B79
Fertilisation
Fertilisation is the introduction of hereditary material from the sire through sperm activation of the ovum;
the sperm and the ovum join to form a new cell. It is the most important step in reproduction.
In small ruminants the sperm has to undergo physiological changes in the female genital tract
before it can fertilise the ovum. Insemination before ovulation (6–24 hours) is therefore recommended.
The fertile life of a sperm is 30 to 48 hours and is even shorter for the ovum (less than 24 hours).
Gestation and parturition
Gestation is the period from fertilisation to delivery of the foetus. Its duration is 148–150 days for sheep
and goats and is influenced by age of the dam, litter size and environmental factors (nutrition).
During pregnancy the blood vessels in the vulva and vagina increase. The cells of the cervix secrete
mucus to form the ‘cervical mucus seal’ and the uterus enlarges. The Graafian follicle transforms into a
corpus luteum which persists throughout pregnancy, secreting progesterone which is essential for
maintaining the pregnancy. However, after the first trimester the sheep’s placenta can produce enough
progesterone to maintain the pregnancy even if the corpus luteum is removed. The placenta is formed to
maintain the exchange of material between the foetus and the mother.
The process of parturition in small ruminants is more complex. Foetal and maternal roles have been
studied and reviewed by several scientists (Mukasa-Mugerwa et al 1994; Mukasa-Mugerwa and Lahlou-
Kassi 1995). There are subtle differences in the prepartum course of events culminating in parturition in
sheep and goats. In sheep progesterone levels in maternal plasma decline 7 to 15 days before term.
However, reducing the influence of progesterone on the sheep uterus is not critical for labour initiation. In
goats, an abrupt decline in progesterone which occurs 24 hours before delivery is the critical event for
activating the uterus. Oestrogens increase during the last day of pregnancy in ewes and rise gradually in
goats. These events culminate in stimulating the muscles of the uterus to contract. The foetus and placenta
are expelled and this is followed by the regeneration of the uterine wall. About 24 days are needed for the
uterus to return to its normal size (involution). Oestrus cycles may start again as early as 14 to 21 days after
parturition. In some cases silent ovulation may occur before behavioural oestrus. Suckling delays the
onset of oestrus but this observation is variable in the tropics.
Pregnancy diagnosis in sheep and goats
To diagnose pregnancy in small ruminants you can use:
• rectal palpation after two months of pregnancy
• abdominal palpation in advanced pregnancy
• hormonal tests
• Doppler or Ultrasonic techniques
• direct palpation after laparotomy
• vaginal biopsy for the presence of cornified epithelial cells
• laparoscopy
• foetal electrocardiography
• immunologic methods: antigens specific to pregnancy.
3.4 Reproductive performance
For small ruminants to be highly productive, good levels of reproductive performance need to be
achieved. Reproductive efficiency can be assessed by measuring fertility, prolificacy (litter size),
fecundity (fertility × prolificacy) and survival. The following sections will deal with these factors briefly.
For more details see the section on reading materials (Section 3.7).
35
Major factors that affect productivity in small ruminants include reproductive rate, survival rate,
growth rate and policies that affect prices of livestock and their products. Figure 3.7 shows how these
factors can lead to poor reproductive efficiency and how they relate to each other. This chapter deals with
manipulation of the animal environment and of genetic factors to correct for impairment in reproductive
performance. The physiological control of puberty will be described as will manipulation of physiological
processes in relation to the environment and the hormone flow in sheep and goats.
3.4.1 Annual reproductive rate
The impact of reproduction on sheep and goat productivity is best estimated by the annual reproductive
rate (ARR) which is defined as the number of lambs or kids weaned per ewe or doe of reproductive age per
year. It is generated as:
ARR = S(1? M)/I (Equation 1)
where S is litter size, M is the rate of lamb or kid preweaning mortality and I is lambing or kidding interval
in ewe or doe in years. Equation 1 can also be used to estimate the flock productivity by including the
number of lambs/kids weaned per ewe/doe in the flock per year as follows:
ARR’ = S(1? M)P/I (Equation 2)
where P is the proportion of ewes/does in the flock of reproductive age. However, there are several ways of
assessing the reproductive rate in small ruminants, e.g. for sheep:
• Fertility = number of ewes lambing/number of ewes available for mating.
• Prolificacy = litter size = number of lambs born alive/number of ewes lambing.
• Fecundity = fertility × prolificacy = number of lambs born alive/ewes available for mating.
• Lambing rate = number of lambs born/ewes available for mating (often expressed as a percentage).
• Weaning rate = number of lambs weaned/ewes available for mating (often expressed as a
percentage).
36
Figure 3.7. Causes of poor reproductive performance in sheep.
B44B65B6cB61B79B65B64
B70B75B62B65B72B74B79
B49B6eB63B72B65B61B73B65B64 B70B72B65B6eB61B74B61B6c B6cB6fB73B73B65B73 B49B6eB63B72B65B61B73B65B64 B70B65B72B69B6eB61B74B61B6c
B6dB6fB72B74B61B6cB69B74B79
B48B69B67B68 B6cB61B6dB62
B6dB6fB72B74B61B6cB69B74B79
B41B62B6fB72B74B69B6fB6e
B31B2e B47B65B6eB65B74B69B63 B66B61B63B74B6fB72B73
B32B2e B45B6eB76B69B72B6fB6eB6dB65B6eB74B61B6c B66B61B63B74B6fB72B73 B28B64B69B73B65B61B73B65B2c B6eB75B74B72B69B74B69B6fB6eB2c B63B6cB69B6dB61B74B65B2c B6dB61B6eB61B67B65B6dB65B6eB74B29
B46B65B77 B6cB61B6dB62B73 B77B65B61B6eB65B64B2fB65B77B65B2fB79B65B61B72
B4cB6fB77
B6fB76B75B6cB61B74B69B6fB6e
B72B61B74B65B73
B4cB6fB77
B66B65B72B74B69B6cB69B73B61B74B69B6fB6e
B72B61B74B65B73
B50B6fB6fB72
B69B6dB70B6cB61B6eB74B61B74B69B6fB6e
B72B61B74B65B73
B45B61B72B6cB79 B65B6dB62B72B79B6fB6eB69B63
B6dB6fB72B74B61B6cB69B74B79
B53B74B69B6cB6c B62B69B72B74B68B73
The above relationships show that reproduction rate is influenced by litter size, young mortality and
interval between parturitions.
Productivity of tropical sheep is said to be low because of scarce documentation or poor
measurement of outputs such as meat, milk and manure. Since milk, blood and manure are rarely
measured, productivity is expressed in terms of ‘per unit weight of breeding ewe’ rather than ‘per
unit input’. Justification for this is based on the proportionality of the number of breeding ewes to
the quantity of real inputs needed by the flock. Initial lamb weaning weight and subsequent
parturition interval were used by Wilson et al (1985) to construct three flock production indices
which can be expressed as follows:
1. Index I (kg/ewe) = (total live weight of litter at 90 days × 365)/subsequent parturition interval
(days).
2. Index II (g/ewe) = Index I/post-partum weight of dam (kg).
3. Index III (g) = Index I/post-partum weight of dam0.73 (kg).
Index I is derived in kg/ewe and represents lambing interval as a percentage of a year and considers
lamb performance until weaning. Since the dam is the source of input until weaning, it was necessary to
assess productivity in g/kg of dam weight (Index II) or in kg/kg metabolic weight (Index III). When lambs
die before weaning, the index is zero.
Table 3.1. Productivity Index II for sheep in sub-Saharan Africa.
Breed Location Index II* (g/ewe)
West African Dwarf Senegal 0.47
Afar Ethiopia 0.51
Baggara Sudan 0.60
Bambara Mali 0.56
Maasai Kenya 0.47
Menz Ethiopia 0.52
Barbados Blackbelly Latin America 0.53
Blackhead Brazil 0.38
Pelibuey Latin America 0.43
West African Dwarf West Africa 0.42
Mean 0.48
*Index II (g/ewe) = [(total live weight of litter at 90 days × 365) subsequent parturition interval (days)]/post-partum weight of
dam (kg).
37
Examples: Annual reproductive rate and productivity indices
The mean annual reproductive rate (ARR) in Africa is 1.2 (Mukasa-Mugerwa and Lahlou-Kassi 1995)
while that reported for Menz sheep in the Ethiopian highlands was 1.4 (Gatenby 1986; Wilson 1989).
Estimates of productivity indices (Index II) from sheep flocks are shown in Table 3.1. The productivity
of the West African Dwarf sheep was studied in Senegal. Index I had an average value of 11.5 kg/ewe
and Index II was 0.47 g/kg dam weight. Flocks were studied in semi-arid Africa and Index II ranged
from 0.47 to 0.60 g/ewe. It was found that Index II is the most useful of the three indices. Fitzhugh and
Bradford (1983) calculated Index II for sheep in Latin America and West Africa. Values for Index II
range from 0.42 g/kg dam weight for West African Dwarf sheep to 0.53 g/year for Barbados
Blackbelly. The values are similar to the ones calculated by Fall et al (1983).
3.4.2 Litter size
Most estimates of litter size in tropical sheep range from 1.0 to 1.5 indicating that twinning rate is
generally between 0 and 50%. Mean litter size may be less than one, indicating that the rate of pre- and
perinatal losses in the form of stillbirths is greater than the twinning rate in the flock. In Africa there are
some breeds which are recognised to be prolific, e.g. West African Dwarf and the Moroccan D’Man, as
they usually deliver two or more young. Table 3.2 gives some estimates of litter size in African sheep.
Table 3.2. Estimates of litter size in African sheep.
Breed Location Litter size (number)
Afar Ethiopia 1.05
Awassi Egypt 1.12
Bornu Mali 1.05
Desert Sudan 1.30
Dorper Botswana 1.10
Maasai Kenya 1.02
Maure Chad 1.07
Nungua Ghana 1.13
Uda Niger 1.07
West African Dwarf Côte d’Ivoire 1–1.31
West African Dwarf Senegal 1.12
Yankasa Nigeria 1.23
Average 1.14
The major determinant of litter size is the number of eggs shed, i.e. the ovulation rate. This is
modified by such factors as fertilisation rate and embryonic and foetal losses before parturition (Table
3.3). Although the ovulation rate of D’Man ewes averages 2.5, only about 42% of the eggs develop to full-
term foetuses. Ovulation rate varies among breeds. It increases with ewe age up to 6–7 years and is greatest
in the first half of the breeding season among seasonal breeders. The trait is highly repeatable within
individuals. Since ovulation rate is genetically controlled, it is possible to select for it and achieve
improvement. Litter size is influenced by environment, age and ewe body weight at mating. The body
weight is affected by nutrition. Litter size can be increased 10 to 40% by improving the nutrition
management of the premating ewe or by treatment with gonadotropins. Litter size in sheep and goats can
also decrease through abortion (Table 3.4). Abortion can be caused by diseases, poor nutrition, litter size
and dam age; genetic factors also play a role in inducing abortion.
Table 3.3. Estimates of embryonic losses in sheep.
Breed Location Embryonic losses (%)
Mixed USA 32.7
Romney New Zealand 22.1
Hampshire USA 28.6
Merino Australia 24.0
Merino Australia 22.8
Menz Ethiopia 22.0
38
Table 3.4. Estimates of abortion rate in sheep and goats in Africa.
Breed Location Abortion (%)
Sheep
Balami Nigeria 3.0
Fat-tailed Rwanda 2.9
West African Dwarf Côte d’Ivoire 3.7
Mossi Burkina Faso 10.7
Goats
Sahel Chad 4.6
Sahel Mali 12.6
Nubian Sudan 8.0
Cross Kenya 10.0
Sokoto Niger 6.5
Malawi Malawi 11.7
Embryonic and foetal losses can occur due to slaughter of pregnant females. Failure to diagnose
early pregnancy can partly explain this behaviour; however, in times of crisis farmers may deliberately
sell pregnant animals. It is estimated that 20 to 70% of the does and ewes slaughtered in sub-Saharan
Africa are pregnant (Table 3.5).
Table 3.5. Estimates of losses due to the slaughter of pregnant females in sub-Saharan Africa.
Location
Estimate (%)
Sheep Goats
Ethiopia 70 –
Mali 20 32
Chad 44 72
Togo
Average
40
40
39
42
Perinatal lamb losses greatly minimise reproductive rate. They occur in the first few days of life
and can be substantially reduced through management. Small lambs are always weaker than those with
larger birth weight; very large lambs may be lost due to dystocia. Lambs born singly often have a higher
survival rate than twin or triplet births; twins and triplets have a lower survival rate due to delivery
39
Example: Losses due to slaughtering of pregnant ewes in the Ethiopian Highlands
Most farmers in the Ethiopian highlands sell their sheep to generate income. These sheep are
slaughtered for human consumption. Farmers are usually reluctant to sell pregnant ewes. Ewes that
have conceived but without the farmer’s knowledge end up being sold. A study found that 70.1% of
the slaughtered females (for market consumption) were pregnant, 24% with twins (Mukasa-
Mugerwa and Tekelye 1988). A high proportion of these females were pregnant in their third to fifth
months of gestation. In this example pregnant sheep were slaughtered all year round. The authors
extrapolated reproductive efficiency from these data if the pregnant ewes were not slaughtered and
were allowed to lamb. The fertility rate would have been 73.2%, fecundity 125.1% and twinning
rate 25.1%.
malposition coupled with low birth weights. The weight change of the dam during pregnancy significantly
influences lamb birth weight. Improved nutrition during the last trimester of pregnancy, when 60 to 70%
of foetal growth takes place, increases dam weight. As ewes become older and more experienced, their
ability to milk, deliver and care for their young improves, leading to a decline in perinatal losses.
3.4.3 Lamb mortality
In traditional flocks an estimated 10 to 50% of the lambs die annually before weaning. The mortality rate
varies between years and is affected by lamb age, litter size, birth weight, ewe nutrition and dam parity.
Mortality can result from the combined effect of birth weight and mothering ability especially in terms of
milk production. Ewes that are undernourished often give birth to low-weight young that are likely to die
of starvation. Supplementing ewes with extra feed during the second and third trimesters produces heavier
lambs with higher chances of survival. While milk production and mothering ability improve with age,
there is a peak age beyond which dams lose condition and the ability to rear the young. Lambs born in the
wet season have low survival rates because of cold stress and disease infection especially pneumopathies.
Preweaning mortality rates among sheep in Africa are shown in Table 3.6. These losses affect weaning
rate and hence productivity as the resources invested in the mother are wasted.
Table 3.6. Estimates of preweaning mortality rate among sheep in Africa.
Breed Country Mortality
East African Fat-tailed Rwanda 18
West African Dwarf Côte d’Ivoire 10
West African Dwarf Ghana 32
West African Dwarf Senegal 33
Sahel Mali 30
Maasai Kenya 32
Menz Ethiopia 29
East African Blackheaded Uganda 21
Mean 25
40
Example: Lamb mortality
Mukasa-Mugerwa et al (1994) found that Ethiopian Menz breed lambs which weigh <2.0 kg at birth
are more likely to die due to starvation, mismothering and exposure (the SME syndrome). The rate
of survival increased from 37% (1 kg birth weight) to 69% (1 to 2 kg birth weight) and 98% (2–3 kg
birth weight); beyond 3 kg birth weight no losses were experienced (Figure 3.8). As lambs grow
older, their chances of survival are higher: mortality is 25–45% in the first year, 10% during the
second year and <2% in age groups over two years. Birth weight (<3 kg) and breed (pure bred)
greatly accentuated the cause of death. Studies on Horro and Menz sheep in Ethiopia revealed
similar conclusions: starvation, mismothering and pneumonia were major factors for mortality
among the lambs (Table 3.7). Trauma and accidents were not important.
Table 3.7. Post-mortem diagnosis of causes of mortality among Horro and Menz sheep in Ethiopia.
Mortality (%)
Cause of mortality Horro Menz
SME* complex 20 15
Pneumonia 20 29
Gastro-intestinal problems 9 13
Septicaemia 14 17
Still-birth 6 7
Joint and navel infection 4 1
Endoparasites 4 1
Trauma and accidents – 1
Undetermined 20 16
*SME = starvation, mismothering, exposure.
Source: Mukassa-Mugerwa and Lahlou-Kassi (1995).
Figure 3.8. Effect of birth weight on lamb mortality in Menz sheep.
3.4.4 Lambing interval
The interval between two parturitions is the lambing or kidding interval. Lambing interval has three
phases: the gestation period, the post-partum anoestrus period and the service interval. Extended lambing
intervals commonly arise from long post-partum anoestrus intervals, repeated cycles of service intervals
without conception, embryo death or abortion. Lambing interval generally declines with age suggesting
that young ewes take longer to regain condition after lambing. Good management strategies remove
breeding ewes with problems from the flock. Farmers usually keep such ewes in the hope that the situation
may improve; this results in wasted resources in terms of feed, and a lower flock production index.
41
B31B32B30
B31B30B30
B38B30
B36B30
B34B30
B32B30
B30
B30 B30B2eB35 B31 B31B2eB35 B32 B32B2eB35 B33 B3eB33
B42B69B72B74B68 B77B65B69B67B68B74B2c B6bB67
B4dB6fB72B74B61B6cB69B74B79 B30B96B39B30 B64B61B79B73
B4dB6fB72B74B61B6cB69B74B79 B30B96B37 B64B61B79B73
Mortality (%)
Lambing intervals in Africa ranged from 230 to 437 days (Wilson 1989). About 65% of Menz sheep
in the Ethiopian highlands have a lambing interval of eight months and are capable of three lambings in
two years except when mating is managed.
3.4.5 Proportion of reproductive ewes
The fraction of a flock that is of reproductive age is dependent on age at puberty, the length of
reproductive life of ewes, lamb mortality and the ratio of rams to ewes. Traditional animal owners
generally maintain flock structures for maximum output while minimising risks. The average value for
ewes and does of reproductive age (P) in tropical flocks commonly ranges from 0.45 to 0.75 with an
average of 0.55. Using equations 1 and 2 for the values S=1, M=0.15, I=0.75, P=0.55 we can arrive at the
following situation:
ARR = S(1?M)/I = 1.0(1?0.15)/0.75 = 0.85/0.75 = 1.1 lambs per ewe per year
ARR’ = 0.6
Values established for African sheep ranged from 0.8 to 1.2. All the components of ARR are inter-
related and must be considered collectively to effect meaningful improvement. For example, a 10%
increase in litter size does not by itself increase ARR by 10% because a rise in litter size will be
accompanied by increased lamb mortality and extended lambing intervals unless this is accompanied by
improvement in overall management. Breeding ewe lambs at an early age may raise the value of P but the
impact of this may not be achieved due to small litter size, high lamb mortality and longer lambing
intervals.
3.5 Management of reproduction
One of the major advantages of tropical sheep is that they can cycle throughout the year. Most of these
sheep graze natural pastures or utilise crop residues. The quantity and quality of these feed sources
fluctuate seasonally leading to periods of serious nutrition deficits which compromise the ewes' natural
ability for continuous breeding. This results in fluctuating lamb production which does not necessarily
peak at the time when market prices are highest. Understanding and manipulating the reproductive cycle
of sheep could help to alleviate this problem. Better control of sheep reproduction would also allow
producers to meet the needs of export as well as expanding urban markets. These factors require better
breed improvement programmes that will use technologies such as AI and multiple ovulation and embryo
transfer (MOET) to maximise rates of gain for important carcass traits.
3.5.1 Techniques for inducing oestrus and ovulation
The male or ram effect
As mentioned earlier, farmers introduce a ram into a previously isolated flock as the breeding season
approaches to induce ewes to breed earlier in the mating season. The effect appears to result from an
increase in LH which induces the maturation of the pre-ovulatory follicle. The ewes’ ability to respond to
this effect depends on the depth of anoestrus.
Use of exogenous hormones
Reproduction is the sum total of oestrus, ovulation, fertilisation and parental foetal growth and
development. These processes can be induced with hormones but to achieve the best results this must be
supported by genetic selection and proper management. The main aims of hormonal control of
reproduction are to:
• synchronise breeding during the mating season
• increase ovulation rate and subsequent multiple births
42
• induce fertile mating during the anoestrus period
• induce early puberty.
Three types of hormones are often used to control sheep reproduction: progestogen, oestrogen and
gonadotropin.
Progestogens. These are female sex hormones that are produced naturally in the animal or made
synthetically. Progesterone is produced by the corpus luteum (CL) during the luteal phase before the
animal shows oestrus. It is used to synchronise oestrus and ovulation during the breeding season. When
ewes are in the anoestrus phase, progestogens are used to prepare the uterus for pregnancy and sensitise the
animals to respond better to hormones that can stimulate oestrus and ovulation. Using progestogens to
synchronise oestrus is discussed in Section 3.3.
Oestrogens. Oestrogens are also female sex hormones produced naturally by the ovary or synthesised
chemically. In the animal they are produced in large amounts by the follicle just before or during oestrus.
When used in combination with progesterone, oestrogens are valuable in sensitising the ewe to better
respond to the effects of ovulatory hormones and help to prepare the uterus for pregnancy.
Gonadotropins. These are hormones that cause ovulation (e.g. LH and FSH) and are produced by the
pituitary gland. The most commonly used is pregnant mare serum gonadotropin (PMSG). It is produced
by the uterine lining of the mare in the first phase of gestation. The human chorionic gonadotropin (HCG)
hormone is also used. The hormone PMSG is best given on the day progesterone treatment stops. PMSG at
the rate of 200 to 300 iu increased twinning in Menz sheep (Mukasa-Mugerwa and Lahlou-Kassi 1995). It
is noteworthy that an immune response which may last for up to one year may develop from repeated use
of PMSG. The normal recommendation is to apply PMSG and HCG in a ratio of 8:1 on mating ewes
synchronised with sponges. The combined use of sponges plus PMSG appears to have potential for
improving lamb production for specialised situations, e.g. a condensed lambing period for research
purposes where small sire groups are desired. Particular attention to details of management of mating are
important, e.g. 2% of sponges get lost and need immediate replacement, a high incidence of multiple births
is expected and better management at and soon after lambing is required to minimise perinatal losses.
Use of melatonin
Day length is perceived by the eye and translated into a circadian pattern of melatonin secretion which
increases during the time of longer darkness. Several slow-release formulations of melatonin are
available:
• slow biodegradable subcutaneous implant, e.g. REGULIN
• an intra-ruminal glass bolus
• a hydrogel polymer-based implant. This approach is, however, less relevant in the tropics.
Use of laparoscopy for AI in sheep
A major obstacle to maximising the benefits of genetic selection and improvement programmes in sheep,
in contrast to cattle and pigs, is the unsuitability of existing AI procedures. The factors leading to this
situation are short shelf-life of semen (<10 h) and the low conception rate (20 to 30%) following the
insemination of frozen semen via the cervical route. Apparently this arises from poor semen motility after
thawing. The laparascopic technique has helped to overcome some of these limitations and make the use
of frozen semen more practical. This technique results in conception rates ranging from 47 to 79%
(compared to >90% when fresh semen is used).
Potential role of laparoscopy in embryo transfer (ET)
The rate of genetic progress for important traits can be doubled if each ewe contributes 10 offspring to the
breeding programme. This is possible using the multiple ovulation and embryo transfer (MOET) technique.
The major obstacle to MOET is variation in superovulatory response using current gonadotropins.
These hormones are given during the last two to three days before sponge removal or injection of
43
prostaglandin to synchronise oestrus. Although the laparascopic embryo recovery technique gives 15%
less embryo recovery in contrast to surgical procedures, it allows for repeated flushing even at monthly
intervals and does not interfere with subsequent ewe breeding potential.
3.6 Exercises
1.
Define oestrus and describe practical ways to detect it.
2. Improved nutrition and health would lead to improved reproductive output. Do you agree with
this hypothesis? Design experiments to test this hypothesis in the following situations: a. pastoral
b. agropastoral c. urban.
3. The following table lists the reproductive rates in sheep in three African countries.
Country Litter size
(number)
Parturition interval
(days)
Annual reproductive rate
(young/year)
Ethiopia 1.08 262 1.50
Kenya 1.05 312 1.23
Sudan 1.17 437 0.98
What conclusions can you draw from the difference in parturition interval among countries?
4. Even though sheep in the tropics can cycle throughout the year, in many places in sub-Saharan
Africa the quantity and quality of feed fluctuate seasonally leading to periods of serious
nutritional deficits which affect the ewe’s natural ability for continuous breeding. Manipulating
the reproductive cycle of sheep, therefore, could help alleviate this problem. Describe how you
can achieve this objective using hormones.
3.7 References and reading materials
Abassa K.P. 1995. Reproductive Losses in Small Ruminants in Sub-Saharan Africa: A review. ILCA
Working Document. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia
and IDRC (International Development Research Centre), Ottawa, Canada. 169 pp.
Fall A., Diop M., Sandford J., Wissocq Y.J., Durkin J. and Trail J.C.M. 1982. Evaluation of the
Productivities of Djallonké Sheep and N'Dama Cattle at the Centre de recherches
zootechniques, Kolda, Sénégal. ILCA Research Report 3. ILCA (International Livestock Centre
for Africa), Addis Ababa, Ethiopia. 70 pp.
Fitzhugh H.A. and Bradford G.E. 1983. Productivity of hair sheep and opportunities for
improvement. In: Fitzugh H.A. and Bradford G.E. (eds), Hair Sheep of Western Africa and
Americas: A Genetic Resource for the Tropics. Westview Press, Boulder, Colorado, USA. pp.
23–52.
Gatenby R.M. 1986. Sheep Production in the Tropics and Sub-tropics. Tropical Agriculture Series.
Longman, London, UK. 351 pp.
Mukasa-Mugerwa E. and Lahlou-Kassi A. 1995. Reproductive performance and productivity of
Menz sheep in the Ethiopian highlands. Small Ruminant Research 17:167–177.
Mukasa-Mugerwa E. and Tekelye B. 1988. The reproductive performance of Ethiopian highland
sheep. Animal Reproduction Science 17:95–102.
Mukasa-Mugerwa E., Zere Ezaz and Viviani P. 1990. Plasma concentrations of progesterone during
oestrus cycles of Ethiopian Menz sheep using enzyme immunoassay. Small Ruminant Research
3:57–62.
44
Mukasa-Mugerwa E., Negussie A. and Said A.N. 1994a. Effect of postweaning level of nutrition on
the early reproductive performance and productivity indices of Menz sheep. Journal of Applied
Animal Research 5(1):53–61.
Mukasa-Mugerwa E., Sovani S., Toe F. and Lahlou-Kassi A. 1994. Reproductive responses
associated with ram introduction in mature tropical Menz ewes after a period of isolation.
Animal Reproduction Science 36:243–251.
Mukasa-Mugerwa E., Anindo D., Lahlou-Kassi A., Mutiga E.R. and Sovani S. 1993. Seasonal
variation in ovarian and oestrus activity of tropical Menz sheep as affected by plane nutrition.
Reproduction, Nutrition, Development 33(6):585–595.
Mukasa-Mugerwa E., Said A.N., Lahlou-Kassi A., Sherington J. and Mutiga E.R. 1994b.
Birthweight as a risk factor for perinatal lamb mortality, and the effects of stage of pregnant ewe
supplementation and gestation weight gain in Ethiopian Menz sheep. Preventive Veterinary
Medicine 19:45–56.
Ngere L.O. and Aboagye G. 1981. Reproductive performance of the West African Dwarf and the
Nungua Blackhead sheep of Ghana. Animal Production 33:249–252.
Reynolds L. 1979. Breeding performance and growth rate of the indigenous Malawi goat. Bunda
College of Agriculture Research Bulletin 10:90–100. Bunda College, Bunda, Malawi.
Wilson R.T. 1989. Reproductive performance of African indigenous small ruminants under various
management systems: A review. Animal Reproduction Science 20:265–286.
Wilson R.T. and Traore A. 1988. Livestock production in central Mali: Reproductive performance
and reproductive wastage in ruminants in the agro-pastoral system. Theriogenology
29:931–944.
Wilson R.T., Peacock C.P. and Sayers A.R. 1985. Preweaning mortality and productivity indices for
goats and sheep on a Maasai group ranch in south-central Kenya. Animal Production
41:201–206.
Yapi C.V., Boylan W.J. and Robinson R.A. 1990. Factors associated with causes of preweaning lamb
mortality. Preventive Veterinary Medicine 10:145–152.
45
Module 4: Health and diseases
4.1 Performance objectives
Module 4 is intended to enable you to:
1. Describe briefly the symptoms of diseases of economic importance in sheep and goats.
2. Describe in detail helminthosis as it affects sheep and goats.
3. List control strategies for major diseases of small ruminants.
4.2 Introduction
In spite of the numerical and economic importance of small ruminants in developing countries of the
tropics, productivity is generally low owing to inadequate feed resources, poor management and diseases.
Losses from diseases alone reduce productivity by 50–60% (Pritchard 1988; Winrock 1992). The
recurrent loss in productivity and profit is often due to parasitic infections, particularly helminth infections
which are a common and major problem for small ruminant production in most parts of the tropics.
One of the main problems faced by African smallholder farmers in controlling diseases of small
ruminants is lack of adequate veterinary services. These services are often provided by government
agencies which lack facilities for surveillance and diagnosis of diseases and are unable to maintain
adequate vaccine supply/production (Winrock 1992). To alleviate these constraints these services need to
be strengthened in both governmental and private organisations.
4.3 Background information
Diseases can be defined as any deviation from health. Healthy animals are alert, active and show interest in
their environment. Disease can also be defined as a change in the normal condition of the animal caused by
any invading living organism including parasites, bacteria, viruses, protozoa and fungi. Infectious
diseases which are spread from one animal to another directly or indirectly are called contagious diseases
while those that are spread by a vector are referred to as vector-borne diseases. The definition of a disease
implies knowledge of the normal condition of the animal, but this information is beyond the capacity of
this manual. Readers are urged to familiarise themselves with this knowledge. By carefully watching the
animals and learning how they behave it is possible to observe deviations which may be caused by
diseases. Many factors are important for optimum health and productivity. These include good livestock
47
4.1 Performance objectives
4.2 Introduction
4.3 Background information
4.4 Major diseases of small ruminants in sub-Saharan Africa
4.5 Helminthosis
4.6 Exercises
4.7 References and reading materials
management and animal husbandry. Paramount among these is nutrition. Management and nutrition
interact to maintain optimum health (Figure 4.1). The relationship among the host (ruminant animal), the
agent (the causal living organism) and the environment demarcates the line between a healthy or diseased
animal (Figure 4.2). To illustrate this concept: sub-Saharan Africa harbours the vector (tsetse fly) which
transmits trypanosomosis to ruminant animals, a case where the environment determines the occurrence
of trypanosomosis. The type of environment may have an adverse effect on the condition of the animal,
predisposing it to disease. The climate may favour vegetation that also harbours vectors such as snails
which serve as intermediate hosts for Fasciola gigantica and F. hepatica, the two common liver flukes of
sheep and goats.
Figure 4.1. Relationships among factors affecting health.
Figure 4.2. Relationships among host, environment and disease causal agent.
48
There are a few terms that need to be defined as they are used frequently both in this manual and in
the literature:
• Incubation period: The period from when the invading organism enters the animal’s body to when
signs of disease are shown.
• Symptoms: Signs or changes in the health of the animal.
• Morbidity: The number of animals in a herd or flock which are affected by disease. The morbidity
rate is usually expressed as the percentage of animals that are clinically affected compared with the
total number of animals exposed to the same risks.
• Mortality: The number of animals dying from disease. The population mortality rate is the
percentage of all exposed animals which die.
• Immunity: The ability of the animal to resist or tolerate the disease. This could be innate immunity
(natural resistance) or acquired immunity (exposed or vaccinated).
• Immunisation: Administration of vaccines to protect animals from diseases by imparting artificial
immunity.
• Tolerance: A form of resistance acquired through contact with the disease through many generations.
• Susceptibility: The animal is affected by the disease on exposure as it has no tolerance of, immunity
or resistance to the disease.
• Diseases are clustered into three groups based on the epidemiological features that take into account
the relative importance of the environment, production system and the pathogens (Winrock
International 1992) .
4.3.1 Group I: Epidemic diseases
Group I includes highly contagious epidemic diseases that constitute a widespread risk and cause high
mortality rates and economic losses (Provost 1991). These diseases:
• are independent of the production system and the environment
• are widely distributed (regional and international)
• cause moderate to severe disruption of production or depressed productivity.
The most important diseases in Group I are: peste des petits ruminants (PPR), foot-and-mouth
disease (FMD), contagious caprine pleuropneumonia (CCPP), Rift Valley fever and blue tongue. The
main control strategy is immunisation of the animals. If not controlled, Group I diseases lead to high
mortality rates. Due to effective immunisation and preventive efforts by national and international
agencies, the diseases are being gradually brought under control in East Africa and are almost eradicated
in West Africa.
4.3.2 Group II: Parasitic and viral diseases
Group II constitutes the main disease constraints to ruminant productivity in sub-Saharan Africa. They are:
• infectious and largely parasitic diseases
• widely distributed
• strongly influenced by the environment and the production system
• known to depress productivity and cause mortality.
Group II diseases are the parasitic and viral diseases and they are mainly transmitted by vectors.
The most important diseases in this group are insect borne (including trypanosomosis), tick-borne
diseases, helminthosis, heartwater, brucellosis, contagious agalactia and Q fever. Vaccines to control
49
most of Group II diseases are not readily available. The most common control methods include the use of
chemotherapeutic drugs but resistance to these drugs is easily developed. The cost of the drugs and/or
irregular availability limit their use by smallholder farmers. Effective and economically sound
technologies for controlling these diseases are therefore needed.
4.3.3 Group III: Diseases associated with intensification
Group III includes both infectious and non-infectious diseases that are associated with significant
mortalities. These diseases cause serious economic losses through reduced productivity (Provost 1991).
They:
• are endemic in nature (that might be viral, bacterial or protozoal), significantly depress animal
productivity, but are very often insidious
• become increasingly important as production systems are intensified
• are often associated with poor nutritional status of the animals, genotype, management practices
and environmental characteristics
• cause reproductive wastage.
Since sub-Saharan Africa is moving towards intensification of animal agriculture, these diseases
will cause great losses and need to be controlled if farmers are expected to adopt intensification-related
technologies. The most important diseases in this group include: soil-borne diseases (anthrax), infectious
reproductive tract diseases (brucellosis), diarrhea and pneumonia of the newborn, tuberculosis (tbc) and
sheep and goat pox. Control measures for these diseases are achieved through management, effective
immunisation and proper disposal of animals which die from infections.
4.4 Major diseases of small ruminants in sub-Saharan Africa
This section gives a summary of the major diseases of sheep and goats in sub-Saharan Africa (Tables 4.1,
4.2 and 4.3). For convenience we will group the diseases into three types:
• Respiratory diseases are characterised by symptoms of coughing, running nostrils and poor condition in
general. This group includes enzootic pneumonia, pasteurellosis, lungworm infections and CCPP.
• Diseases of the digestive system may be associated with parasites, bacteria and viruses and are very
important in Africa.
• Other diseases include those that affect the reproductive system leading to infertility, abortion, still
birth or early death of lambs; these diseases are very important in economic terms. Diseases of the
skin include external parasite infections and foot-rot. The systemic diseases include PPR, sheep
pox, cowdriosis, anthrax, enterotoxaemia and trypanosomosis. There are other diseases of sheep
and goats which are related to the nervous system or are caused by lack of essential nutrients. Since
these are minor we will not deal with them in this manual.
To achieve maximum productivity in sheep and goats, it is important to prevent and control the
diseases mentioned above. We also believe that research resources should be focused on the diseases that
cause the most economic losses at the farm level. This module will therefore describe helminthosis in
detail because:
• infection is widespread in sub-Saharan Africa
• most of the other diseases have received attention, and preventive measures such as vaccination are
available to control them
• helminthosis severely reduces productivity
• little research has been done on this disease with the aim of providing cheap and effective control
measures
50
Table 4.1. Diseases of sheep and goats of economic importance in sub-Saharan Africa: viral diseases.
Diseases and causal organisms Occurrence Clinical signs Economic impact Prevention and treatment
1. Rift valley fever: infectious—non-contagious (for sheep) —Zoonosis
• characterised by fever, abortion
and hepatitis
• caused by a virus; cattle and goats
and humans are susceptible and
frequently attacked
• vector-borne disease transmitted
by mosquitoes
• all breeds and sexes
• seasonal occurrence: moist warm
months
• declined feed consumption
• fever
• death of lambs and ewes
• weakness
• death within 24 hours
• high mortality (95%)
• abortions
• protein deficiencies
• cost of preventive programmes
• vaccinations
2. Nairobi sheep disease (NSD)
• acute infectious (non contagious)
• characterised by fever, diarrhoea
and hoemorrhagic gastro-enteritis
• caused by a virus
• vector-borne disease transmitted
by ticks
• all breeds and sexes
• more commonly attacks adult
animals than lambs
• goats and possibly some wild
ruminants are susceptible
• high temperature
• diarrhoea
• respiration difficult
• mucopurulent exudate
• deaths
• cost of preventive programmes
• threat of spread to new
geographic area
• protein deficiency
• control of brown ticks
• vaccination
3. Peste des petits ruminants (PPR)
• rinderpest-like disease of small
ruminants
• caused by a virus
• goats and sheep • fever, ocular and nasal discharge
• coughing and sneezing
• oral necrosis
• diarrhoea, enteritis and pneumonia
• morbidity and mortality rates
higher in goats than in sheep
• this results in reduced
productivity
• vaccination
• quarantine measures
• slaughter of infected animals
4. Sheep pox (Variola ovina)
• acute, contagious and often fatal
disease
• characterised by generalised
papules on the skin and plagues
on mucous membranes
• caused by a virus
• all breeds, sexes and ages of
sheep; young more susceptible
than adult animals
• fever
• dermal eruption
• physical depression
• loss of appetite
• muculae form on eyelids, nostrils,
lips, cheeks, perineum, prepuce and
udder
• economic losses result from high
mortality, reduced meat, milk and
wool production
• animal vaccination in endemic
areas
• proper disposal of dead animals
• antibiotics to control bacterial
infection of papules
5.Contagious caprine pleuropneumonia (CCPP)
• caused by Mycoplasma mycoides
ssp caprae/M. m. capr
• occurs in many African and Near-
East countries
• only affects goats
• dyspnoea
• fever, pneumonia
• diarrhoea
• high mortality rate
• losses due to high morbidity and
mortality
• antibiotics
Table 4.2. Diseases of sheep and goats of economic importance in sub-Saharan Africa: bacterial diseases.
Diseases and causal organisms Occurrence Clinical signs Economic impact Prevention and treatment
1. Enzootic pneumonia (pasteurellosis)
• acute infectious disease of sheep charac-
terised by pneumonitis and pleuritis
• caused by interaction of Chlamydial,
mycoplasma, bacteria, viruses and
environmental stresses
• all breeds and sexes • fever
• nasal discharge
• respiration accelerated, difficult and
often accompanied by coughing
• morbidity up to 50%
• deaths
• reduced live weight
• cost of treatment
• avoid excessive crowding
• Sulfonamides
• antibiotics
2. Anthrax (splenic fever; charbon)
• acute contagious septicaemia
• characterised by splenic swelling and
blood, exudations from body openings
• caused by a sporing bacterium; soil-
borne disease; zoonosis
• all breeds and sexes
• more common in adult animals
• fever
• loss of appetite
• congestion of mucous membranes
• sudden death
• deaths
• quarantine enforcement
• cost of immunisation
• treatment practices
• dead animal disposal
• incapacitation of people affected
• dead animals, bones, skin,
bedding—properly disposed of
• individual treatment at early stages
• Antibiotics
• immunisation
3. Black disease (infectious necrotic hepatitis)
• acute infectious toxemia of sheep
• characterised by liver necrosis and
sudden death
• caused by interaction of bacteria and
liver flukes
• black disease develops in sheep one
year of age and older
• animals in excellent physical
condition are most likely to develop
the disease
• acute in nature and short course
• usually escape detection in the
clinical phases
• depressions in co-ordination
• Fever
• morbidity 15%
• mortality 100%
• controlling flukes (use of
anthelmintics),
• proper disposal of dead animals
• vaccine against infection from
Clostridium novyi
4. Brucellosis
• contagious bacterial disease of cattle,
swine, goats, sheep, man
• characterised by abortion, genital
infections and formation of localised
lesions in various body tissues
• all breeds
• frequent occurrence associated with
close housing and intensive
management methods
• primary signs: abortion, although
all infected pregnant animals do not
necessarily abort
• retained placenta
• vaginal discharge in females, loss
of sexual desire and infertility in
male
• economic losses result from abortions,
stillborn lambs, early deaths of lambs,
cost of prevention programmes
• vaccinations of young stock
• sanitary measures
• testing of adult animals
Table 4.3. Diseases of sheep and goats of economic importance in sub-Saharan Africa: parasitic diseases.
Diseases and causal organisms Occurrence Clinical signs Economic impact Prevention and treatment
1. Verminous pneumonia (lung worm infections)
• chronic and prolonged infection of
sheep and goats
• characterised by bronchitis and
broncho pneumonia
• all breeds, sexes and ages
• seasonal occurrence (mostly during
cold seasons)
• coughing, nasal discharge, rapid
breathing
• death may result from
accompanying bacterial infections
• unthriftiness
• retarded growth
• reduced production of meat, milk
and wool
• deaths
• cost of prevention and treatment
• pasture management
• vaccination
• anthelmintics
2. Parasitic gastro-enteritis (PGE)
• acute or chronic disease of sheep
• caused by a wide range of species
of gastro-intestinal nematodes
• most breeds, sexes, ages of sheep
(evidence for breed resistance)
• seasonal occurrence
• anaemia
• reduced live weight
• Inappetence
• sometimes diarrhoea
• oedema
• unthriftiness
• depressed growth
• cost of preventives and control
programmes
• deaths
• reduced milk production
• adequate nutrition
• avoidance of overstocking
• pasture management
• use of anthelmintics
• resistant breeds
3. Liver fluke infections (fasciolosis; distomatosis)
• acute parenchymal hepatitis and
chronic cholangitis
• characterised by unthriftines, loss
of weight, anemia, oedema and
eosinophilia
• caused by two species of
trematodes Fasciola gigantica and
F. hepatica
• most breeds, sexes, ages (some
evidence for breed resistance)
• occurrence of the disease in goats,
cattle camels, wild ruminants
• anaemia
• bottle jaw
• Inappetence
• loss of live weight
• inefficient feed conversion
• retarded growth
• deaths
• condemnation of infected livers
• cost of preventive and treatment
programmes
• pasture management (avoid
swampy areas)
• anthelmintics
• breeding for resistance
4. Trypanosomosis
• acute or chronic disease
• characterised by remittent fever,
anaemia, reduced weight
• caused by protozoan parasites
• domestic and game mammals • peracute, acute and chronic forms
• fever, anaemia oedema of limbs,
scrotum and eyelids
• neurological disturbances
• enlargement of peripheral lymph
nodes
• reduced productivity • control of tsetse flies
• reduce reservoir hosts in the wild
fauna
• treatment of infected animals with
trypanocidal drugs
• resistant breeds
• even though drugs are available on the market to control helminthosis, they are expensive and
beyond the reach of the smallholder farmer; in many places the parasites are developing resistance
to these drugs.
4.5 Helminthosis
Helminth infections are of economic importance in the tropics. Helminthosis is one of the diseases causing
great economic losses worldwide to sheep and goat production (Gatenby 1986). This disease constrains
animal production by causing poor growth and impaired fertility.
Helminthoses are infections caused by internal parasites called helminths. These parasites can be
divided into two major groups: the Platyhelminthes or flat worms (trematodes and cestodes) and the
Nemathelminthes or round worms (nematodes). There is great variation in the prevalence and the
geographic distribution of helminth infections in small ruminants in sub-Saharan Africa. Helminths
generally spend part of their life cycles outside their definitive hosts (i.e. sheep or goats), either in the
ground, on grass or within invertebrates such as snails, insects or earthworms. Temperature, rainfall and
the type of soil determine the occurrence of a given parasite species. Therefore, the knowledge of their
epidemiology in different agro-climatic zones is essential for successful control strategies. One approach
to achieve this aim is to conduct a cross-sectional (vertical) study in which all categories of animals are
examined according to age, sex and breed at a particular time to provide the point prevalence. A more
appropriate approach is the longitudinal (prospective) study in which groups of animals are monitored
over an extended period of time, preferably three years, to take climatic variations into account.
Pathological conditions associated with helminth infections are characterised by loss of appetite,
anaemia, infertility or mortality. The identification of parasites present can be carried out either:
• during post-mortem examination of animals for adult and immature morphological characteristics
or
• through identification of parasite eggs in faecal samples from live animals (for detailed procedures
see Box 4.1).
Assessing the level of infection is done by faecal egg count (FEC) which is an indirect method of
quantifying worm burdens in live animals, differential worm counts at necropsy, faecal culture, packed
cell volume (PCV) and pasture larval counts.
Helminthosis affects small ruminant productivity and may lead to death of infected animals,
especially young ones. The major effects of the disease are slow growth (of infected young animals),
weakness and lowered resistance to other diseases.
The largest number of parasites are found in the intestinal tract from where they can absorb food and
suck blood from the wall of the intestine. A heavy worm load may block the intestinal tract. Parasites may
lead to direct production loss, i.e. death of the animal, or indirect production losses such as:
• impaired growth
• reduced milk, meat and fibre production
• reduced reproduction performance
• reduced work performance
• condemned organs
• high cost of control measures.
The pathogenicity of the diseases depends on the parasite species involved, the number of parasites
and their stage of development, and age, sex and the nutritional status of the host animal. General
symptoms arising from helminth infections include loss of appetite, digestive disturbances, loss of
condition, anaemia, pale mucus membranes, constipation or diarrhoea, coughing and symptoms of
bronchitis.
54
55
Box 4.1. Helminth parasites of small ruminants: diagnostic procedures
Internal parasites of small ruminants cause severe reductions in the productivity of sheep and goats in
sub-Saharan Africa. Economic losses occur due to mortality, low productivity and high costs of
anthelmintics. Effective control measures rely on appropriate diagnostic methods and identification of
resistant breeds or individual animals. This box gives a brief description of the most commonly used
procedures for diagnosing helminth parasites.
A. Parasite identification
At the start of any parasitological investigation it is essential to identify which parasite species are
present in an area by conducting initial surveys. If this information is already available, this step can be
skipped. Identification to the species level is done through examination (i) post-mortem or (ii)
identification of parasite eggs/larvae in faecal samples from live animals.
B. Major measurement techniques
1. Faecal egg count (FEC). To diagnose parasites in sheep and goats, it is essential to recover the
parasites, eggs or larvae from the digestive tract or faecal material. Resistance has normally been
assessed by faecal egg count which was shown to be a good indicator of worm burdens in the animals
(especially lambs and kids); care should be taken as this may not be the case in all species. To identify
and quantify the eggs:
• collect faecal samples
• separate eggs from faecal material and concentrate them
• using a microscope, examine and count.
2. Faecal culture. Sometimes it is necessary to differentiate parasite species that have similar egg
shape. This differentiation is achieved by using faecal cultures which provide a suitable environment
for the helminth eggs to hatch into larvae (the infective stage). The same steps used for FEC are used
followed by:
• preparation of faecal culture
• isolation and identification of larvae from cultures.
3. Isolation of larvae from herbage. This is a procedure for isolating, identifying and counting larvae
from samples of grass taken from a grazing area. Many factors, such as warm temperature and
adequate moisture, affect the number of larvae in herbage and therefore care should be taken in
interpreting data from this procedure.
4. Packed cell volume (PCV) determination. An effective way to detect infection by parasites is
measuring the faecal egg count. Sometimes disease infection can proceed and show its effect before
the eggs appear. One of the early effects of parasite infection is anaemia. The packed cell volume
(PCV) technique allows the assessment of anaemia to detect such cases even before eggs appear. In this
technique we measure red blood cells in a sample of circulating blood. Since anaemia can be caused by
many other agents, it is advisable to do this test as well as the FEC when the eggs appear. The procedure
requires sampling blood into a test tube containing a coagulant. The blood sample is centrifuged at
12,000 rpm for 4 minutes. The tube is placed in the reader and the reading is expressed as a percentage
of packed red cells in the total volume of whole blood.
For detailed information on procedures read MAFF (1977), Hansen et al (1990) and Bekele et al
(1992).
The following section describes in detail the two groups of parasites of sheep and goats mentioned
above: Platyhelminthes (trematodes and cestodes) and Nemathelminthes (nematodes).
4.5.1 Trematodes
Fasciola gigantica and F. hepatica, the two major liver flukes of sheep and goats, cause a disease referred
to as fasciolosis. Both species are found in the cool tropics of the African highlands, particularly in the
Ethiopian highlands. Several species of snails serve as their intermediate hosts. The life cycle of the liver
fluke is shown in Figure 4.3. The eggs, which are laid by the adult worm in the bile ducts of the livers of
infected animals, are passed into the environment through the faeces. When deposited in a wet
environment, a miracidium (larva) develops, escapes from the egg and actively swims to seek the right
snail species. Upon penetration into the snail intermediate host, the miracidium passes through various
stages. The developed larvae (cercariae) leave the snail, attach to the grass and encyst to become
metacercariae, the infective stage. Grazing animals eat the encysted cercariae along with the grass for
initial infection. Hay or fresh grass cut and carried can transfer metacercariae to confined animals. The
duration of the life cycle is 2.5 to 3 months depending on the parasite species.
Figure 4.3. The life cycle of the liver fluke (Fasciola spp).
The symptoms of the disease vary depending on the stage of development of the parasite within the
definitive host and the duration of the infection. In sheep, liver flukes cause an acute disease which results
in sudden death or a chronic disease (in more resistant breeds) whose symptoms are diarrhoea, anaemia,
jaundice and oedema (bottle jaw). In chronic cases, animals lose weight progressively and may ultimately
die from the infections. The disease is diagnosed by examining faecal samples, by serological techniques
and liver function tests. We look for the parasites in the liver and examine the faeces for the presence of
eggs (for further details see Box 4.1). Liver damage, arising from liver fluke infection, causes high
mortality.
56
4.5.2 Nematodes
A number of nematode species can parasitise sheep and goats. The nematodes that infect small ruminants
are small thread-like worms (usually 10–30 mm long) localised in the intestine. During development, a
nematode moults at intervals shedding its cuticules. In the complete life cycle, there are four moults, the
successive larval stages being designated L
1
,L
2
,L
3
,L
4
and finally L
5
which is the immature adult. The
life cycle of nematodes with a few exceptions is usually direct with no intermediate host involved (Figure
4.4). The survival, development and transmission of the nematode infective larvae are mostly influenced
by climatic conditions (moisture, temperature and rainfall). Susceptibility of the host is determined by
many factors: age, breed, physiological status (parturition, depression of immunity), level of immunity
(previous exposure) and plane of nutrition. Infection of sheep and goats by round worms is manifested by
two major types of symptoms, namely, lack of appetite and anaemia. Young animals are affected more
than adults.
Mortality in young animals can reach 25%. The diseases resulting from nematodes can best be
diagnosed by recovery of worms at necropsy and by examination of the faeces (Box 4.1).
Figure 4.4. The life cycle of gastro-intestinal nematodes (Haemonchus, Ostertagia, Trichostrongylus
and Cooperia).
4.5.3 Cestodes (tapeworms)
Tapeworms (cestodes) are long, segmented, flat white worms, a few species of which may infest sheep.
They are located in the small intestine of the host. The tapeworms range from 0.5 to5minlength and from
2 to 20 mm in width. They shed white rings that can be observed easily in the faeces of infested animals.
The rings contain large numbers of eggs.
Sheep may also harbour the larval stage of a tape worm of dogs and jackals (Taenia multiceps)
(Figure 4.5). Dogs become infected by ingesting raw offal (brain) of dead animals containing the cystic
(larval) form of the worm. The cysts develop into adult stages within the definitive host (dogs and jackals).
When the segments containing the eggs become ripe, they break off and pass into the environment with
57
faeces. Sheep ingest eggs while grazing, and embryos from the fertile eggs enter the blood stream and are
carried to the spinal cord or the brain. Infected sheep develop symptoms of unsteady gait and walking in
circles, a condition called circling disease.
Figure 4.5. The life cycle of a tapeworm (Taenia multiceps).
4.5.4 Control strategies for helminthosis
Control strategies are usually aimed at keeping infection at a level compatible with production at an
acceptable cost (for details of treatment and control strategies see Hansen and Perry 1990). Control is best
effected through livestock and pasture management, chemotherapy, immunisation and through a good
knowledge of local epidemiology.
• Chemotherapy involves use of modern anthelmintics. Effective drugs are not readily available in
many developing countries and their high cost limits their use by smallholder farmers. In addition,
most farmers lack information on how to use the medication. The frequent use of anthelmintics
leads to parasites developing resistance to the drugs. Smallholder farmers may also use indigenous
medicinal plants to treat their sick animals. However, no formal studies have been carried out to
assess the effectiveness of these local treatments.
• Pasture management is done through the cut-and-carry technique, proper rotational grazing and
avoiding marshlands. This strategy is ideal but is difficult to implement in the traditional production
system. However, it is important to develop awareness of grazing management techniques among
the farmers.
58
• Vaccines for controlling internal parasites are still not available for sheep and goats; there are major
problems with derived antigen variations. However, there is a commercial vaccine for lung worms
in cattle.
• Vector control through environmental management is a feasible strategy, e.g. using molluscicides
against snails. The limitations of this strategy are the toxicity of the chemicals against non-target
animals and organisms and the high cost to the smallholder farmers.
• Breeding sheep and goats that are resistant to helminths is a long-term strategy. This minimises the
risk of parasites becoming resistant to anthelmintics and reduces the cost of intervention. Research
at ILRI and its collaborative networks with national agricultural research systems (NARS) is
proceeding to achieve this goal. Success has been achieved in both Australia and New Zealand for
cattle resistance to ticks and sheep resistance to nematodes.
• Improved nutrition is an important strategy for helminth control. Immunity to gastro-intestinal
nematode infections can be enhanced in grazing animals through economical supplementation with
protein-rich concentrates.
Improvement in the control of helminth infections in sheep and goats offers opportunity to increase
small ruminant productivity in sub-Saharan Africa. Livestock owners will continue to rely on the use of
anthelmintics for quite sometime before effective vaccines against helminth infections become available.
Therefore, sustainable control measures should combine strategic anthelmintic treatments, based on
sound epidemiological data, and other means of control such as pasture management, use of resistant
breeds and biological control methods.
4.6 Exercises
1. Cost-effective control programmes for helminth infection in small ruminants should be based on
sound epidemiological knowledge of the time relationship between contamination of pastures
and seasonal availability of infective larvae in a given geographic area.
a. List experiments you would conduct and data you would use to verify this hypothesis.
b. Since the use of anthelmintic drugs is costly, what long-term alternative control measures will
you target?
2. Management and the environment interact to maintain optimum health. Using examples from
your region, describe the health status of small ruminants in your area in relation to this
statement.
59
Example: Improved nutrition as a strategy for helminth control in small ruminants
Research has demonstrated that improved nutrition can enhance immunity to gastro-intestinal
nematode infections in grazing small ruminants. Nutrition can be improved by supplementing the
animals with protein-rich concentrates. For example urea and cottonseed cake can be used as
concentrates but this supplementation is expensive and beyond the means of many smallholder
farmers. To solve this problem, researchers identified low-cost energy and protein supplements that
can be added to low quality roughage diets (Shaw et al 1995; Knox and Steel 1996). The use of low-
cost supplements such as urea–molasses blocks (UMB) has been proven effective to enhance an
animal’s ability to utilise the available diet and assist the animal to withstand infection (Anindo et al
1998). At times the parasite infection can be very high. During such emergency situations,
medicated UMB can be used for short periods.
4.7 References and reading materials
Angus M.D. 1978. Veterinary Helminthology. 2nd edition. William Heinemann, Medical Books Ltd,
London, UK. 323 pp.
Anindo D., Toe F., Tembely S., Mukasa-Mugerwa E., Lahlou-Kassi A. and Sovani S. 1998. Effect of
molasses–urea-block (MUB) on dry matter intake, growth, reproductive performance and
control of gastrointestinal nematode infection of grazing Menz ram lambs. Small Ruminant
Research 27:63–71.
Bekele T., Kasali O.B. and Rege J.E.O. 1992. Repeatability of measurements of packed cell volume
and egg count as indicators of endoparasite load and their relationship with sheep productivity.
Acta Tropica 50:151–160.
Charray J., Humbert J.M. and Levif J. 1992. Manual of Sheep Production in the Humid Tropics of
Africa. CABI (Centre for Agriculture and Bioscience International), London, UK. pp. 36–59.
FAO (Food and Agriculture Organization of the United Nations). 1991. Livestock Production and
Health for Sustainable Agriculture and Rural Development. Background document 3.
FAO/Netherlands Conference on Agriculture and the Environment, S-Hertogenbosch, The
Netherlands, 15–19 April 1991. FAO, Rome, Italy.
Hansen J. and Perry B. 1990. The Epidemiology, Diagnosis and Control of Gastro-intestinal
Parasites of Ruminants in Africa. A Handbook. ILRAD (International Laboratory for Research
on Animal Diseases), Nairobi, Kenya. 121 pp.
Jensen R. and Swift B.L. 1982. Diseases of Sheep. 2nd edition. Lea & Febiger, Philadelphia, USA.
330 pp.
Knox M. and Steel J. 1996. Nutritional enhancement of parasite control in small ruminant production
systems in developing countries of Southeast Asia and the Pacific. International Journal of
Parasitology 26:963–970.
MAFF (Ministry of Agriculture, Fisheries and Food). 1977. Manual of Veterinary Parasitological
Laboratory Techniques. Technical Bulletin 18. Her Majesty’s Stationery Office, London, UK.
129 pp.
Mboera L.E.G. and Kitalyi J.I. 1994. Diseases of small ruminants in central Tanzania. In: Lebbie
S.H.B., Rey B. and Irungu E.K. (eds), Small Ruminant Research and Development in Africa.
Proceedings of the Second Biennial Conference of the African Small Ruminant Research
Network, AICC, Arusha, Tanzania, 7–11 December 1992. International Livestock Centre for
Africa (ILCA)/Technical Centre for Agricultural and Rural Co-operation (CTA). ILCA, Addis
Ababa, Ethiopia. pp. 117–120.
Pritchard W.R. 1988. Ways that veterinary medicine can alleviate hunger in Africa. Journal of
American Veterinary Medical Association 192:1701–1705.
Provost A. 1991. Report on Animal Health in sub-Saharan Africa. Executive Summary. Winrock
Institute, UK.
Shaw K.L., Nolan J.V., Lynch J.J., Coverdale O.R. and Gill H.S. 1995. Effects of weaning,
supplementation and gender on acquired immunity to Haemonchus contortus in lambs.
International Journal for Parasitology 25:381–387.
Winkler J.K. 1982. Farm Animal Health and Disease Control. 2nd edition. Lea & Febiger,
Philadelphia, USA. 230 pp.
Winrock International. 1992. Assessment of Animal Agriculture in Sub-Saharan Africa. Winrock
International Institute for Agricultural Development, Morrilton, Arkansas, USA. 125 pp.
60
Module 5: Genetic improvement of African
small ruminants
5.1 Performance objectives
Module 5 is intended to enable you to:
1. Define concepts of animal genetic resource characterisation and conservation.
2. Identify indigenous breeds of sheep and goats in sub-Saharan Africa.
3. Describe the main features of the important indigenous breeds of sheep and goats in sub-Saharan
Africa.
4. Describe methods for characterisation and conservation of small ruminant genetic resources in
sub-Saharan Africa.
5. Describe selection methods for small ruminant improvement.
6. Describe cross-breeding methods for improvement of small ruminants in sub-Saharan Africa.
5.2 Introduction to genetic concepts
Small ruminants in sub-Saharan Africa, like all livestock in the continent, evolved to survive in harsh
environments at the expense of all other factors (production included). Hence most African breeds under
ideal conditions are not as productive as temperate breeds, although under harsh conditions, temperate
breeds may be less productive than tropical breeds. The challenge facing scientists in Africa is how to
improve production traits of the indigenous breeds without losing their capacity to survive in harsh
environments.
Farmers try to improve the productivity of their small ruminants by using the best producers in
their flocks for breeding. Characteristics that can be passed on are due to genetic factors, which constitute
the genotype; other characteristics cannot be passed on and are due to the environment. The two
61
5.1 Performance objectives
5.2 Introduction to genetic concepts
5.3 Characterisation and conservation
5.4 Genotype × environment interaction
5.5 Utilisation of biodiversity
5.6 Introduction to improvement
5.7 Exercises
5.8 References and reading materials
components (genotype and environment) taken together are termed the phenotype. The main difficulty is
to separate these two components once the phenotype value of the character (performance) is measured.
5.3 Characterisation and conservation
Most sheep and goats in Africa are of the indigenous types. There are many types and numerous breeds in
sub-Saharan Africa (Table 5.1). Detailed information about other breeds can be found in: Devendra and
Burns (1983), Gatenby (1986) and Peacock (1996). Sheep breeds indigenous to sub-Saharan Africa can be
categorised into (i) thin-tailed (ii) fat-tailed and (iii) fat-rumped. Thin-tailed breeds are most common in
the dry tropics, e.g. the African long-legged type which is predominant in the Sahara. This type is usually
large (Figure 5.1) and well adapted to a migratory existence. Examples include: West African Long-
legged, the Dongola, Northern Sudanese and Zaghawa. In West Africa the thin-tailed sheep, found in the
humid areas, are smaller than their dry tropic counterparts and often referred to as the dwarf or forest sheep
(Figure 5.2). Fat-tailed sheep are predominant in eastern Africa. Breeds of this type are: Abyssinian, East
African Blackheaded and Maasai. Fat-rumped types predominate in north-east Africa but are also found in
South Africa and other countries of the southern region (Figure 5.3). In South Africa wool breeds were
imported from Europe to the temperate environments in the country, e.g. South African Merino. Breeds
for harsher environments were developed from African hair sheep, e.g. the Blackhead Persian. Cross-
breeding between wool and hair sheep was done in South Africa, e.g. Dorper.
Figure 5.1. A thin-tailed sheep.
62
Figure 5.3. A fat-rumped Blackhead Persian sheep.
63
Figure 5.2. A West African Dwarf ram.
Table 5.1. Selected examples of breeds of sheep and goats in sub-Saharan Africa.
Name of breed Location Characteristics Use
Sheep
African Long-fat-tailed
Tanzania, South Africa,
Malawi, Zimbabwe,
Madagascar
Tail usually long, fat and broad
with hanging tip, twisted,
cylindrical, strap-shaped, funnel-
shaped or carrot-shaped
Meat and hair
Blackhead Persian South Africa White with black head and neck,
polled, fat-rumped
Hair and meat
Fulani Senegal to Cameroon West African long-legged type,
white, pied or coloured
Meat
Menz Ethiopia Black or dark brown with white on
head, neck and legs, weight 20–30
kg female, male usually horned,
female usually polled, fat-tailed
with woolly undercoat
Meat and
coarse wool
West African Dwarf Southern West Africa White, usually with red or black
spots, occasionally black hair,
male with mane, weight 30 kg
male, females 20 kg, male horned,
female polled
Meat
Goats
Afar Eritrea, Ethiopia Leggy, curved horns Milk and meat
Sudanese Desert Sudan Both sexes are horned but those of
males are longer twisted horns,
long-legged, coat usually grey
Meat and
skins
West African Dwarf West and central Africa Short legs, small, weighing about
20 kg and are about 50 cm in
height
Meat
Boer South Africa Colour is white, medium size, ears
long, broad and droopy
Meat, milk
and skins
Small East African Kenya, Uganda, Tanzania Small light horns and ears, colour
variable, small, adult weight
around 30 kg
Meat and
skins
Indigenous goats fall into two main groups: (i) long-eared and (ii) short-eared. Goats are also
classified as (i) large (ii) intermediate and (iii) dwarf types (Figure 5.4). Large types, which usually have
long legs, are found at the edge of the Sahara and in southern Africa. The intermediate types are mainly
found in east and north-central Africa. The dwarf types are mainly found in humid West Africa.
.
64
Exotic breeds of small ruminants were introduced mainly by South Africa for commercial
purposes. There are large populations of Merino sheep (Kenya, Lesotho, Zimbabwe), Karakul sheep
(Angola, Botswana, Namibia, Rwanda) and Angora goats (Lesotho, Madagascar). Other countries have
exotic breeds but of smaller populations.
There is pressure from exotic breeds on indigenous sheep and goats through indiscriminate cross-
breeding and upgrading. Since these indigenous breeds have evolved under a stressful environment (poor
nutrition, high temperatures and high incidence of disease), they are well adapted and it is important to
characterise and conserve their genetic resources. Under experiment station conditions, the indigenous
African goats and sheep breeds showed good growth and reproduction performance, indicating their
potential. Thus improvements in management can give responses of better productivity. However,
indigenous breeds are not distinct from one another and few characterisation studies have been carried out.
Some 40 years ago, governments and non-government organisations discussed and published
documents on conservation of animal genetic resources. Many conferences, seminars, symposiums and
other meetings were held to raise the awareness of the public about the conservation issue. The Food and
Agriculture Organization of the United Nations (FAO) played an important role through cataloguing,
maintaining and using of animal genetic resources (AGR) at the very beginning of its activities. This effort
was targeted towards policy makers. The United Nations Conference on Environment and Development
(UNCED), which was held during 1992 in Rio de Janeiro, Brazil, gave due attention to the issue of
biodiversity. ILRI, formerly ILCA (International Livestock Centre for Africa), started activities on
conservation and characterisation of livestock during 1977. These activities led to the creation of a
network that dealt with trypanotolerant animals in Africa. Other organisations dealing with biodiversity in
animals in Africa included the Organization of African Unity (OAU) and CIRAD–EMVT (Centre de
coopération internationale en recherche agronomique pour le développement–Département d'élevage et
de médecine vétérinaire des pays tropicaux, France); CIRAD–EMVT focused its activities in
francophone Africa.
The first step to conservation is to learn which breeds or populations to conserve. It is therefore
necessary to define the current situation through characterisation and evaluation of breeds or populations
before embarking on any programme of conservation. The concepts of conservation and characterisation
are defined in Box 5.1.
65
Figure 5.4. West African Dwarf goats.
66
Box 5.1. Conservation and characterisation: Definition of major concepts
Animal genetic resources
Genetically unique breed populations, formed throughout the domestication process, within each
animal species used for the production of food and agriculture together with their immediate wild
relatives.
Domestic animal diversity
Genetic variation or genetic diversity existing among the species, breeds and individuals, for all
animal species that have been domesticated, and their immediate relatives, to meet human needs for
food and agriculture.
Conservation
Conservation embraces preservation, maintenance, sustainable utilisation, restoration and
enhancement of the natural environment to meet immediate and short-term requirements for food
and agriculture and this diversity remains available for future use. It can be in the:
• Narrow sense: preservation—storing of genetic resources that are not currently popular
among farmers for possible use in the future.
• Broad sense: all operations involved in the management of genetic resources, such as
inventory, monitoring and characterisation of genetic resources, the development and better
use of the adapted indigenous genetic material, the maintenance of unique resources to help
meet future requirements and the establishment of better access to a wider variety of these
resources particularly for research and development.
In situ conservation. Active breeding of animal populations for food and agriculture, such that the
diversity is used both in the short term and maintained for the longer term. Operations pertaining to
in situ conservation include performance recording schemes, development (breeding) programmes
and ecosystem management and use for sustainable production for food and agriculture. In situ
conservation of wild relatives is generally called in situ preservation. Live populations are
maintained in their adaptive environment or as close to it as practically possible.
Ex situ conservation. Ex situ conservation involves collection and cryopreservation of samples of
breeds in the form of live semen, ova or embryos, which can be used to generate animals. DNA
techniques do not constitute ex situ conservation.
Characterisation. Characterisation is the enumeration and description of:
• genetic identity of an animal species, breed, strain or population
• the environment to which such species, breed, strain or population is adapted, partially
adapted or not adapted.
This involves the distillation of all available knowledge (published or unpublished) which
contributes to the reliable prediction of genetic performance in defined environments.
Monitoring of genetic diversity. Monitoring of genetic diversity involves surveys and registration
of all available breeds, populations or groups of interbreeding animals. Information collected
includes whether a breed is going to be endangered, be a candidate for conservation and whether
conservation should be performed as a concerted effort of several countries with the same breeds.
Breed. Population with similar genetic traits common to the group.
Landrace. Population that has lived sufficiently long locally to adapt to the various conditions of
the locality.
5.3.1 Rationale for characterisation and conservation
Why characterise? The main reasons behind characterisation of animal genetic resources are:
• threat to the indigenous livestock
• a number of different breeds are named after their location, ethnic group or by physical
characteristics and may not necessarily be genetically different
• most censuses in developing countries are done by species which does not give an accurate picture
of the population trend of individual breeds over time in order to determine populations at risk of
extinction
• little knowledge is available about existing breeds to understand:
? unique qualities of the breed, e.g. particular adaptation to the local environment
? the potential contribution to productivity if treated as improved breeds
? the potential to make the greatest variety for future contribution and determine special genes
that could be useful in the future
? the genetic variability (within and between breeds) for future needs.
There is a threat to livestock populations and some breeds may be approaching extinction. Factors
that lead to extinction are:
• demand for higher animal production due to growing human population, improved standards of
living and urbanisation
• natural disasters (drought, famine, disease outbreaks, flooding, earthquakes etc) which decimate
flocks
• socio-political instability
• negligence (low commercial value, ignorance)
• loss of genetic variability through:
? use of a limited number of species and a few specialised breeds within species (since the
production of improved breeds is high, less productive animals are displaced)
? intensive selection for important traits facilitated by artificial insemination (AI), i.e. small
number of males are being used for reproduction
? upgrading of native breeds with superior breeds, including exotics.
The actions listed above lead to genetic erosion which is accentuated by international trade and
ease of transportation throughout the world. As part of the global biological diversity, there is an
urgent need to conserve and preserve African animal genetic resources in order to meet the needs of
future generations.
Most developing countries, particularly in Africa, lack activities relating to animal genetic
resources. Most conservation activities in Africa have been directed towards plant genetic resources and
wild life. Many meetings have been held to discuss biodiversity, but small ruminants have never been
considered. Little information exists on African Small Ruminant Genetic Resources (SRGR) in the FAO
data bank (Table 5.2). This is due to lack of suitable data recording facilities in most African countries.
Table 5.3 shows the status of small ruminant genetic resources in Africa. At present five breeds of
sheepare already extinct, one breed of goat and three breeds of sheep are endangered and 10 breeds of
sheep are decreasing in number (Box 5.2).
67
Table 5.2. Information on African small ruminant genetic resources in the FAO data bank.
Number of references
Species Number of breeds Population
information
Production
information
Sheep 133 47 46
Goats 59 19 21
Cattle 175 84 51
All* 434 179 120
* Buffalo, cattle, goats, sheep, pigs and horses.
Table 5.3. Status of African small ruminant genetic resources.
Breed status Species Country
Extinct Hottentot sheep
Steekhaar (variety of Ronderib Africander)
Walrich Mutton Merino
Wooled Persian sheep
Permer sheep
South Africa
South Africa
South Africa
South Africa
Nigeria
Critical or
endangered
Pafuri goat
Bezuidenhout Africander
Namaqua Africander sheep
Nungua Blackhead sheep
Mozambique
South Africa
Namibia, South Africa
Ghana
Decreasing in number Black Maure sheep
Macina sheep
Variety Koundoum
Dongolla sheep, Murle, Dinka, Nuba Mountain
Touareg sheep
Ronderib Africander sheep
Mauritania, Mali,
Senegal, Mali,
Niger
Sudan
Mali, Niger, Chad
South Africa
5.3.2 Characterisation and conservation methodology
The first step to conservation is to know which breed to conserve. Characterisation consists of collecting
information on available stocks and the environment in which the stocks are performing. Basic
information includes:
• preliminary characteristics such as type/breed/variety; predominant location and climatic
conditions; utility, management and production systems; physical and special genetic
characteristics; production traits; and population status
• DNA information about genetic distinctiveness, and genes responsible for valuable traits.
Ex situ conservation is done through cryopreservation of animal germplasm by storing sperm,
oocytes and embryos. For this technique to be effective, progeny of at least 25 sires should be stored. The
facilities in Africa for this technique are not well developed. The germplasm can be collected in Africa and
stored in countries where facilities are available.
In situ conservation is the preservation, multiplication and utilisation of indigenous breeds in their
native habitats and maintenance of pure breeds or strains. The minimum number recommended for sheep
is 60 ewes and the maximum is 1500 ewes. Special conservation flocks can be established as part of the
cultural heritage for the endangered breeds or breeds in critical status. Nucleus breeding flocks can be
established in co-operative breeding programmes to supply breeding stocks to farmers. This conservation
method requires long-term funding and therefore it is practical in Africa if donor funding is secured. In
situ conservation requires good management of the conserved flock. Appropriate breeding programmes
should be designed and properly implemented. These programmes should:
68
• balance breeding objectives for all important traits and allow economic weights on each of them to
reflect the local or national situation
• prevent deterioration of fitness or adaptive and functional traits
• optimise selection strategies with respect to the long-term effects of inbreeding and genetic
variance
69
Box 5.2. Criteria to determine breeds at risk
Breeds are classified into one of the seven categories defined by the Food and Agriculture
Organization of the United Nations (FAO) United Nations Environmental Programme (UNEP) as
follows:
1. Extinct
It is no longer possible to recreate the breed population. This situation becomes absolute when there are
no breeding males (or semen), breeding females (oocyte) or embryos remaining.
2. Critical
The total number of breeding females is less than 100 or the total number of breeding males is less or
equal to 5.
or
The overall population size is slightly above 100 but decreasing and the percentage of females
being bred pure is below 80%.
3. Endangered
The total number of breeding females is between 100 and 1000 or the total number of breeding
males is greater than 5 but less than or equal to 20.
or
The overall population size is slightly below 100 and increasing and the percentage of females
being bred pure is above 80%.
or
The overall population size is slightly above 1000 and decreasing and the percentage of females
being bred pure is below 80%.
4/5. Critical-maintained/Endangered-maintained
This is the situation when active conservation programmes are in place or populations are maintained
by commercial companies or research institutes.
6. Not at risk
The total number of breeding females and males are greater than 1000 and 20, respectively.
or
The overall population size approaches 1000 and is increasing and the percentage of females
being bred pure is close to 100%.
7. Unknown
Information on population figure is unknown.
• harmonise local or national breeding goals with sustainable agriculture, i.e. integration of
crop–livestock production
• utilise genes from the world gene pool that are not in conflict with the breeding goals.
Genetic markers and breed identification as tools in conservation
A genetic marker for a trait is a DNA segment which always exhibits the same characteristics when
present in a genotype. Genetic markers serve several uses in conservation of genetic resources but are not a
method. Three major uses include:
• Validation of pedigree through the verification of parentage: A correct pedigree is important for
developing breeding strategies that minimise inbreeding.
• Estimation of the level of genetic diversity: Analyses of genetic marker variation allows you to
objectively measure the genetic variation within the breed.
• Identification of the breed: Gene markers are used to determine whether an individual is a member
of a particular breed or to compare breeds by calculating genetic similarity or genetic distance. This
exercise allows the estimation of the distinctiveness of a particular breed in relation to the other
breeds of the same species.
The most common types of DNA markers are: RFLP (Restriction Fragment Length
Polymorphism), VNTR (Variable Number of Tandem Repeats), microsatellites, RAPDs (Randomly
Amplified Polymorphic DNAs) and PCR (Polymerase Chain Reaction).
The choice of conservation method depends on the safety offered and the cost involved. There are
advantages and disadvantages to using in situ over ex situ conservation methods. The advantages are:
• In situ conservation allows the evaluation, identification of animals with unique characters and
their utilisation in genetic improvement programmes.
• In situ conservation allows animals to adapt to changes over time to physical environment and
diseases.
• Methods for freezing semen or oocyte and techniques for embryo preservation are not available for
all species.
• Live animal conservation can serve an aesthetic and cultural need.
• In situ conservation can generate funds through sales of live animals. Revenue for local
communities can be raised by charging a fee to tourists who want to see the conserved animals.
The disadvantages are:
• Live animal conservation is costly in terms of facilities and supervision.
• In situ conservation is subject to inbreeding and random genetic drift due to the small size of
conserved animals.
• In situ conservation is subject to natural and artificial selection.
• Live animal conservation is subject to natural calamities such as drought, floods and disease
epidemics.
Breeds to be given priority in conservation programmes should be selected based on their
comparative advantage obtained through objective evaluation. Priority should be given to:
• breeds that have reached critical or endangered status
• genetically diverse stocks
• stocks with unique characters
• stocks with high overall economic merit.
70
In the majority of African countries, opportunities to raise productivity through the use of temperate
breeds are minimal because these exotic breeds need an improved production environment, a costly
exercise. Therefore there is need to improve the indigenous types to raise their production potential.
5.4 Genotype × environment interaction
The genotype × environment interaction is defined as the change in the relative production performance of
two or more genotypes measured in two or more environments (Gatenby 1986). It may involve changes in
rank order of the genotypes in different environments and/or changes in the magnitude of the phenotypic,
genetic or environmental variances. It may not be true to assume that parents selected based on their
superiority in one environment will still be superior in any other environment. It is therefore advantageous
to design selection programmes with local populations because:
• there will be ample time for concurrent improvement in management, disease control and feeding
• losses to adaptation in local environments are minimal
• management is much easier.
5.5 Utilisation of biodiversity
Genetic variation is the raw material for animal improvement and a means to overcome selection limits. Animal
breeding aims to create genetic changes in the population. The classical techniques for genetic improvement are:
• selection (which animals to keep as parents)
• cross-breeding (which breed combination).
The heritability value is the main criterion for choosing between selection and cross-breeding but in
many cases the two choices are combined.
5.5.1 Selection
Selection is the process in which the best animals are chosen to be parents of the next generation. The
frequency of genes resulting in animals with favoured phenotypes is increased. The aim of selection is to
achieve improvement in quantitative and measurable performance traits of economic value. Examples of
traits that are most often used as a basis for selection are given in Table 5.4.
Selection can be either direct or indirect. The higher the heritability, the greater is the rate of improvement
by selection. The heritability can be classified as high (h
2
>0.50), medium (0.25<h
2
<0.50) or low (h
2
<0.25).
Table 5.4. Traits used as a basis for selection in small ruminants.
Productivity factors Traits
Reproduction Age at first lambing
Fertility
Lambing
Lambs raised to weaning
Growth Birth weight
Weaning weight
Adult weight
Feed conversion efficiency
Stress Resistance to diseases
Milk yield Milk yield
Fat content
Protein content
Lactation length
71
The response to selection (R) is the difference between the mean phenotypic value of the offspring
of the selected parents and the phenotypic value (before selection) of the population from which the
parents were selected. R depends on the superiority of the selected animals in relation to the mean of the
population from which they come, the extent to which the superiority is inherited (h
2
) and the interval
between generations (L). It is estimated as:
R=h
2
.S
The selection differential (S) is the mean phenotypic value of the individuals selected as parents
expressed as deviation from the population mean.
The annual genetic progress (?G) obtained through selection is expressed as:
?G=(h
2
.S)/L
where h
2
is heritability, S is the selection differential and L is the generation interval.
The selection differential, S, is expressed in units of the phenotypic standard deviation (?
P
)to
standardise its value:
?G/?
P
=(h
2
.S)/(L.?
P
)
Selection methods
Different methods are used to combine traits in a selection programme.
Tandem selection. Priority is given to one trait at a time. When the selection goal for that trait is achieved,
another trait is considered. Progress with this method depends on the genetic correlation between traits. It
is the least efficient method.
Independent culling. A minimum level is set for each trait. Potential animals must meet all the levels.
There is no compensation for poor performance.
5.5.2 Cross-breeding
Cross-breeding aims at (i) combining all desirable characteristics of two or more breeds in one progeny
type and (ii) exploiting the hybrid vigour or heterosis.
Heterosis is attained when the average performance of the crossbred offspring is better than the average
performance of the parents. Heterosis is due to dominance or overdominance. Cross-breeding increases the
number of heterozygous gene pairs. The greater the genetic difference between the parental populations, the
greater are the heterozygosity and the heterosis expected. Heterosis depends on the genetic diversion. There is a
reciprocal relationship between heterosis and heritability. Traits that demonstrate the most hybrid vigour are
those that have the lowest heritability, e.g. fitness and reproductive traits. Expression of heterosis is
environment-dependent and greater heterosis is shown in stressful environments.
Cross-breeding is useful in the creation of new breeds. The desirable qualities in crossbred
foundation stocks are combined to form new synthetic breeds or strains. Cross-breeding is useful in
upgrading stocks. With continuous back-crossing with males (or females) of the desired breed, the
proportion of genes from the original population decreases with each generation to half the one present in
the previous generation. After five generations, there is little difference between the genotypes of the
upgraded animals and animals of the desired breed. The original population can be changed completely.
Cross-breeding also results in the production of crossbred animals for slaughter. Sires of one breed are
mated to females of another breed to produce superior commercial progeny, not used for breeding. For
example, a prolific sire breed is mated with hardy local ewes to produce fat lambs. Examples of breeds
developed from cross-breeding are shown in Table 5.5.
Cross-breeding has its drawbacks:
• it requires more management than pure breeding
• in order for it to continue, pure breeds should be maintained.
72
Table 5.5. Examples of goat and sheep breeds developed through cross-breeding in sub-Saharan Africa.
Breed Location
Goats
Vogan (Djallonké × Sahelian)
Boer {(Africander, South African common) ×
(indigenous Bantus, European, Angora,
Indian Blood)}
Pafuri (Boer × Landim)
Togo
South Africa, Rwanda, Kenya,
Lesotho, Swaziland,
Zimbabwe, Burundi
Mozambique, Namibia
Long-legged Hair Sheep
Toronké (Sambourou × Maure)
Vogan (Oudah × Djallonké)
Varalé (Toubire × Peul-Peul)
Ingessana (Sudan Desert × Nilotic)
Meidob (Zaghawa × Sudan Desert)
Mali
Togo
Senegal
Sudan
Sudan
Fat-rump Hair Sheep
Nungua Blackhead (Blackhead Persian × Djallonké)
Wiltiper (Wiltshire Persian) × (Wiltshire Horn × Blackhead
Persian)
Van Rooy {Van Rooy White Persian} × {Blackhead) Persian ×
[Rambouilet × Ronderib Africander]}
Ghana
Zimbabwe
South Africa
Coarse Wool Sheep
Okuma (Dorper × Kirdi)
Permer (Blackhead Persian × German mutton Merino)
Bezuidenhout Africander {Wooled Persian} {Africander ×
[Wooled Persian × Blackhead Persian]}
Dorper {Dorset Horn × Blackhead Persian} White Dorper
[Dorsian, Dorsie]}
Vandor (Dorset Horn × Van Rooy)
White Wooled Mountain {German mutton Merinox (Dorset
Horn × Blackhead Persian)}
Gabon
Nigeria
South Africa
South Africa
South Africa
Fine Wool Sheep
Afrino {South Africa Mutton Merino ×
[Ronderib Africander x South Africa Merino]}
Döhne Merino (German Mutton Merino ×
South Africa Merino)
Dormer {Dorman [Dorset Horn × German Mutton Merino
Letelle Merino (Derived from Ramboillet)
Walrich Mutton Merino (Précoce × South African Merino)
South Africa
South Africa
South Africa
South Africa
South Africa
Crossing can be done in any one of the following ways (see Figure 5.5):
Single two-way cross. Pure bred females (or males) of one breed are mated with pure-bred males (or
females) of a second breed to produce the first filial generation F
1
.
Three-way cross. Males (or females) of a third breed are mated with females (or males) of the single
two-way cross. Three-way crosses can lead to superior traits because they utilise hybrid vigour.
73
Four-way cross. Populations A and B are first crossed with each other and so are D and C. The resultant F
1
generations of the first two crosses are then crossed to produce the final four-way cross.
Two-way rotational cross. Males (or females) of each of the two breeds are crossed alternately.
Three-way rotational cross. Males (or females) of each of the three breeds are crossed alternately.
Figure 5.5. Crossing types.
5.5.3 Inbreeding
Inbreeding results from mating related individuals, i.e. individuals with common ancestors. This usually
happens when the size of the population decreases. Selection tends to increase inbreeding because it
reduces population by restricting reproduction to a few animals.
The consequences of inbreeding are:
• decrease in growth rate, disease resistance, reproductive performance and viability
• increase in the proportion of homozygotes, therefore decrease in the proportion of heterozygotes
and increased frequency of undesirable traits due to recessive genes.
Increase in population size and careful monitoring of mating programmes help to avoid inbreeding
or reduce its level to a minimum.
5.6 Introduction to improvement
Genetic improvement requires an accurate and continuous recording system. Selection is associated with
recording the performance of individual animals. Accurate records are needed to identify the best and
poorest performing animals on-farm; at the national or regional levels records of the population are
needed. Performance recording is also an aid for making management decisions. It is difficult to keep
records for small ruminants in most African countries, especially for those raised by smallholders under
range management conditions. This is due to the vast distances that need to be covered by the record
keepers, the constant movement of the animals and the reluctance of the smallholders to provide accurate
information so as to avoid taxation.
74
B41 B42
B46
B41 B42
B46
B46
B46
B44B43
B46
B46
B43
B42B41
B31
B32
B31
B31 B31
B32
B54B68B72B65B65B2dB77B61B79 B63B72B6fB73B73
B46B6fB75B72B2dB77B61B79 B63B72B6fB73B73
B54B77B6fB2dB77B61B79 B63B72B6fB73B73
B54B77B6fB2dB77B61B79 B72B6fB74B61B74B69B6fB6eB61B6c B63B72B6fB73B73
B54B68B72B65B65B2dB77B61B79 B72B6fB74B61B74B69B6fB6eB61B6c B63B72B6fB73B73
B41B42
B46B41
B46 B42
B46B41
B46B42
B41
B42
B46B43
B46 B41
B46B42
B46B43
B46
B46
B31
B32
B33
B34
B35B35
B34
B33
B32
B31
5.6.1 Nucleus breeding schemes
Small ruminants in Africa are fed extensively on rangeland or on agricultural by-products in dry areas. In
the humid and subhumid zones of Africa, small ruminants are kept in small numbers by smallholders. It is
almost impossible to conduct within-flock selection under such conditions. An alternative is to establish a
nucleus flock in a group breeding programme.
A nucleus breeding programme is a centralised improvement programme in which very superior
animals are brought together from supply farms to form an elite nucleus flock. Farmers have to agree to put
together their superior animals. The nucleus or base may remain open to the best animals from the supply
flocks. This system is therefore called the open nucleus breeding scheme.
Once the nucleus is established, an efficient recording and selection programme can be
implemented. With the actual production environment in Africa, this system seems much easier to realise
than conventional breed improvement programmes. Institutional or government flocks can serve as the
nucleus and then be linked with private and small farmers.
The nucleus breeding scheme is much used in Australia and New Zealand where it is referred to as a
co-operative breeding scheme. An open nucleus breeding scheme has been used in Côte d’Ivoire since
1984 to improve growth of the indigenous Djallonké sheep (Figure 5.6). An optimum nucleus breeding
programme for sheep must have a minimum of 300 breeding females and be able to distribute 70 to 110
rams annually to participating farmers (Figure 5.7). By rotating rams among farmers, inbreeding can be
avoided. The modelling of an open nucleus breeding programme has shown that the rate of genetic gain
may be increased by 10–15%.
Figure 5.6. Outline of the structure of an open nucleus breeding scheme in Côte d’Ivoire.
The advantages of open nucleus schemes can be summarised as follows:
• animals entering the nucleus are tested under farm conditions
• selection is based on records for traits of economic value
• the improvements are quickly spread as farmers participating in the schemes receive their
replacement sires from the nucleus
75
33% (10 43%
)
B50B65B72B66B6fB72B6dB61B6eB63B65 B74B65B73B74B69B6eB67
B52B61B6d B66B6cB6fB63B6b B28B6fB6e B73B74B61B74B69B6fB6eB29
B53B65B6cB65B63B74B69B6fB6e B62B61B73B65 B66B6cB6fB63B6bB73
B28B41B76B65B72B61B67B65 B6eB75B6dB62B65B72 B6fB66 B65B77B65B73B2fB79B65B61B72 B3d B38B39B34B35B29
B4fB74B68B65B72 B66B6cB6fB63B6bB73
B28B4eB6fB6eB2dB70B61B72B74B69B63B69B70B61B74B69B6eB67 B66B61B72B6dB73B29
B43B75B6cB6cB73
B41B76B65B72B61B67B65 B34B33B33
B72B61B6dB73B2fB79B65B61B72
B43B75B6cB6cB73 B28B33B31B25B29
B52B61B6d B6cB61B6dB62B73 B31B39B25 B28B31B35B96B32B37B29
B46B6cB6fB63B6bB73 B28B30B96B32B31B29
1st category
rams
36% (16 50%
)
2nd category r
ams
• a rapid generation turnover can be maintained
• inbreeding is avoided
• objectives are maintained for many years
• small farmers benefit from co-ordinated effort, policy, pooled experience and shared facilities.
Figure 5.7. Example of an optimum structure of a nucleus breeding programme.
5.6.2 Genetics of diseases resistance
The exploitation of genetic resistance to disease is becoming important in livestock development
programmes. This is further encouraged by the high cost of control measures, complexity of control
recommendations and ecological damage from control treatments to the environment. Candidate diseases
for genetic resistance are trypanosomosis, helminthosis and tick-borne diseases. Packed cell volume
(PCV) and enzyme-linked immunosorbent assay (ELISA) have been used as selection criteria during
trypanosome infection in cattle (see Box 5.3).
5.7 Exercises
76
B4eB75B63B6cB65B75B73
B6e B3d B33B30B30
B4fB74B68B65B72 B66B6cB6fB63B6bB73
B43B6fB2dB6fB70B65B72B61B74B69B6eB67 B66B6cB6fB63B6bB73 B6e B3d B33B30B30B30
B6dB61B6cB65B73
B34B30B96B36B30
B6dB61B6cB65B73 B66B6fB72
B73B6cB61B75B67B68B74B65B72
B36B30B96B31B32B30
B66B65B6dB61B6cB65B73
B42B65B73B74 B79B6fB75B6eB67
B38B6dB61B6cB65B73
B61B6eB64 B36B30 B66B65B6dB61B6cB65B73
B36B30
B66B65B6dB61B6cB65B73
B37B30B96B31B31B30
B6dB61B6cB65B73
Box 5.3. Research on genetic resistance to helminths in tropical small
ruminants: Red Maasai sheep
Helminthosis is of considerable significance in a wide range of agro-climatic zones in the tropics
and represents one of the most important constraints to small ruminant production. Losses in
production, costs of antihelmintics and death of infected animals are major problems. Current
control methods focus on anthelmintic treatment or controlled grazing to reduce pasture
contamination. In the tropics, these control methods are limited by the high cost of antihelmintics,
their uncertain availability, increasing frequency of drug resistance and limited scope for controlled
grazing. A sustainable and low-cost solution is breeding for genetic resistance.
Resistance to endoparasites has been assessed by faecal egg counts (FEC) or packed cell
volume (PCV). FEC has been shown to be a good indicator of worm burdens especially in young
animals. This correlation can be affected by the parasite species. PCV is a good indicator of worm
burdens such as Haemonchus contortus (Baker et al 1994).
Box 5.3. cont...
1.
There is pressure from exotic breeds on indigenous sheep and goats in sub-Saharan Africa. Do
you agree with this statement? Describe the situation in your locality or region. If you agree with
this statement, what steps do you suggest should be taken to solve this problem?
2.
Most of conservation activities in Africa have been directed towards crops and wildlife. What
reasons can you suggest for giving attention to genetic conservation of small ruminants?
5.8 References and reading materials
Baker L.R., Reynolds L., Lahlou Kassi A., Rege J.E.O., Bekelye T., Mukasa-Mugerwa E. and Rey B.
1994. Prospects for breeding for resistance to endoparasites in small ruminants in Africa—a new
ILCA research programme. In: Lebbie S.H.B., Rey B. and Irungu E.K. (eds), Small Ruminant
Research and Development in Africa. Proceedings of the Second Biennial Conference of the
African Small Ruminant Research Network, AICC, Arusha, Tanzania, 7–11 December 1992.
ILCA (International Livestock Centre for Africa)/CTA (Technical Centre for Agricultural and
Rural Co-operation). ILCA, Addis Ababa, Ethiopia. pp. 223–227.
Becker W.A. 1984. Manual of Quantitative Genetics. 4th Edition. Academic Enterprise, Pullman,
Washington, USA. 188 pp.
77
Box 5.3. cont...
There is clear evidence that in the subhumid zone of coastal Kenya, Red Maasai sheep, which
are indigenous to Kenya, are more resistant to endoparasites than the introduced Dorper sheep
breed. Red Maasai lambs, recorded between birth and one year of age, have significantly lower FEC
and higher PCV and lower lamb mortality than Dorper lambs. In addition, Red Maasai ewes have
lower FEC and higher PCV than Dorper ewes during the same lambing–lactation period. The
relative resistance to endoparasites of the Red Maasai sheep at the Kenyan coast confirms earlier
reports from research conducted in the Kenyan highlands. The results demonstrate that this breed
expresses its endoparasite resistance when climatic conditions are conducive to high and persistent
levels of challenge with the blood-sucking parasite Haemonchus contortus.
In addition to confirming genetic variation in resistance to endoparasites, this shows
evidence of variation among African breeds. Estimates of heritability of both PCV and FEC
increase with age, with significant heritabilities among lambs at about 10 months. At this age the
heritability of FEC is 0.220.07 for all lambs, but higher in Dorper-sired lambs (0.320.13) than Red
Maasai-sired lambs (0.110.07). This suggests that after many centuries of natural selection under
endoparasite challenge, the Red Maasai sheep have become fixed for some of the important genes
for resistance, yet still retain important within-breed genetic variation for resistance.
In contrast to the clear breed difference in Kenya, the breed difference for resistance in
Ethiopia is less marked with the Menz sheep being a little more resistant to endoparasites than the
Horro sheep. While genetic variation for resistance to endoparasites within breeds was virtually
non-existent in young lambs in Kenya, this was not the case in Ethiopia. The heritability of FEC in
lambs at weaning (3 months of age) in Ethiopia is 0.240.08, which is very similar to the heritability
of 10-month-old lambs in Kenya.
Helminthosis is a global animal health constraint, affecting the tropical and temperate regions.
Breeds of livestock that have originated and evolved under the severe disease challenge conditions in the
tropics are more likely to possess disease resistance traits. Widening the use of tropical livestock genetic
resources to enhance sheep and goat production is the challenge of the future.
Bodo I. 1995. Minimum number of individuals in preserved domestic animal populations. In:
Proceedings Third Global Conference on Conservation of Domestic Animal Genetic Resources,
Kingston, Ontario, Canada, 1–5 August 1994. Rare Breeds International, Kenilworth, UK.
Falconer D.S. 1981. Introduction to Quantitative Genetics. 2nd edition. Longman, London, UK. 340
pp.
FAO (Food and Agriculture Organization of the United Nations). 1968. Report of the 2nd ad hoc
Study Group on Animal Genetic Resources. FAO, Rome, Italy. 26 pp.
FAO (Food and Agriculture Organization of the United Nations). 1971. Report of the 3rd ad hoc
Study Group on Animal Genetic Resources (Pig Breeding). FAO, Rome, Italy. 21 pp.
FAO (Food and Agriculture Organization of the United Nations). 1977. Mediterranean Cattle and
Sheep in Crossbreeding. FAO Animal Production and Health Paper 6. FAO, Rome, Italy. 37 pp.
Gatenby R.M. 1986. Sheep Production in the Tropics and Sub-tropics. Longman, London, UK. 351
pp.
Henderson C.R. 1973. Sire evaluation and genetic trends. In: Proceedings of the Animal Breeding
and Genetics Symposium, Village Polytechnic Institute and State University, Blacksburg,
Virginia, 29 July 1972. American Society of Animal Science, Champaign, Illinois, USA.
Henderson C.R. 1984. Applications of Linear Models in Animals. University of Guelph, Guelph,
Ontario, Canada. 462 pp.
Hopkins I.R. 1978. Some optimum age structures and selection methods in open nucleus breeding
schemes with overlapping generations. Animal Production 26:267–276.
Jackson N. and Turner H.N. 1972. Optimal structure for a co-operative nucleus breeding system.
Proceedings of Australian Society of Animal Production 9:55–64.
James J.W. 1977. Open nucleus breeding systems. Animal Production 24:287–305.
Lauvergne J.J. and Sovenir P.Z. 1993. Actors in keeping and managing genetic resources of farm
animals in 1992. Animal Genetic Resources Information 12:5–25. FAO (Food and Agriculture
Organization of the United Nations), Rome, Italy.
Rege J.E.O. 1994. Indigenous African small ruminants: a case for characterization and improvement.
In: Lebbie S.H.B., Rey B. and Irungu E.K. (eds), Small Ruminant Research and Development in
Africa. Proceedings of the Second Biennial Conference of the African Small Ruminant Research
Network, AICC, Arusha, Tanzania, 7–11 December 1992. ILCA (International Livestock Centre
for Africa)/CTA (Technical Centre for Agricultural and Rural Co-operation). ILCA, Addis
Ababa, Ethiopia. pp. 205–211.
Ruane J. 1992. The global data bank for domestic livestock. In: Rege J.E.O. and Lipner M.E. (eds),
African Animal Genetic Resources: Their Characterisation, Conservation and Utilisation.
Proceedings of the Research Planning Workshop held at ILCA, Addis Ababa, Ethiopia, 19–21
February 1992. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp.
49–54.
Turner H.N. 1982. Basic considerations of breeding plans. In: Gatenby R.M. and Trail J.C.M. (eds),
Small Ruminant Breed Productivity in Africa. ILCA (International Livestock Centre for Africa),
Addis Ababa, Ethiopia. pp.1–16.
78
Module 6: Nutrient requirements and feed resources
6.1 Performance objectives
Module 6 is intended to enable you to:
1. List strategies to improve ruminant livestock nutrition.
2. Describe feed intake for ruminant livestock.
3. Describe nutrient requirements: protein, energy, minerals, vitamins and water for ruminant
livestock.
4. Calculate crude protein from nitrogen analysis data.
5. List in vivo and in vitro methods for measurement of digestibility.
6. Describe feed resources: natural pastures and ways to improve them, artificial pastures,
fodder trees, crop residues and industrial by-products.
7. Describe methods for conserving forages.
8. List feeding techniques.
6.2 Introduction
Preceding modules described the production systems in sub-Saharan Africa and they stressed the
following facts:
• one of the major factors limiting productivity of small ruminants in sub-Saharan Africa is poor
nutrition.
• variations in birth weight may be related to the plane of nutrition at the time of conception and
during late pregnancy in ruminants.
• neonatal losses can be closely correlated with birth weight, which is a reflection of maternal
nutrition during the last two months of gestation.
79
6.1 Performance objectives
6.2 Introduction
6.3 Feed intake
6.4 Nutrient value of feed
6.5 Evaluation of feed
6.6 Feed resources
6.7 Feeding techniques
6.8 Exercises
• undernourished animals are highly likely to contract diseases.
• feed supply is limited during the dry season.
The above factors necessitate that nutrition must be considered in any strategy to increase
productivity. In this module we will highlight (i) nutrient requirements and (ii) availability and use of feed
resources. More details about ruminant nutrition (methods of feed evaluation, nutrition strategies) can be
accessed in manuals produced by ILRI (ILCA 1990; Osuji et al 1993).
Insufficient forage during the dry season is a major constraint to ruminant productivity. Forage
availability determines the overall carrying capacity of the land. In sub-Saharan Africa human population
is increasing, forcing farmers to use grazing areas for arable farming. Even though this may have a
negative effect on pastoralists, it also results in a symbiotic relationship. Pastoralists can graze their
animals on crop residues in the dry season and the settled farmers benefit from the animals' manure.
To assist you in learning the following sections, Box 6.1 contains the definitions of some basic
concepts to nutrition and feed. Solutions to nutritional constraints depend on the nature of the system. In
pastoral systems the scope is limited as the situation is complicated by land tenure and stocking rates. In
mixed crop–livestock systems there is potential for improvement. The following are a few strategies to
improve ruminant livestock nutrition through technological solutions:
• Crop improvement: This pathway includes (i) selection of crop varieties which yield residues of
high nutritive value and/or in large quantity (ii) changing crop combinations, e.g. intercropping
with legumes and (iii) changing time of planting.
• Livestock management strategies: This involves changing livestock management to match feed
availability with livestock feed requirements.
• Pasture improvement strategies: This involves improving the management of the range principally
through increasing the amount and quality of forage.
• Feed supplementation strategies: These involve using fodder banks, fodder trees, by-products such
as oilseed cakes and meals, and urea–molasses blocks to supplement crude protein shortages.
Scientists conduct research to develop feeding systems that provide sets of tables that give nutrient
requirements of the animal and the nutritive value of feed, both expressed according to the feed evaluation
system. The feeding systems are prepared to help farmers calculate rations for their animals. Scientists
80
Box 6.1. Definition of concepts related to nutrition and feeds
Crude protein. The amount of protein in the feed available to the ruminant. It is calculated using
nitrogen content of the feed corrected by a factor of 6.25 based on the assumption that feed protein
contains, on average, 16% nitrogen. Assumes that all faecal nitrogen is in the form of proteins.
Crude fibre. A fraction of feed that contains the cell wall and consists of carbohydrates (hemicellulose
and cellulose) and lignin.
Digestibility. The digestibility of a feed determines the amount that is actually absorbed by an animal
and therefore the availability of nutrients for growth, reproduction etc.
Energy. Derived from organic constituents of food and used as fuel for body functions and production.
Maintenance. This is the condition when an animal's energy requirements are in balance (equilibrium)
and it is not reproducing or producing output.
Metabolism. Metabolism is the sum of all the physical and chemical processes taking place in living
organisms. For example the excretion of waste products (manure and urine) from the body is part of the
metabolic process.
conduct research to find solutions to nutritional problems. This involves studying the relationship between
nutrition and performance, management and grazing conditions. To achieve this goal, data are collected to:
• examine the effect of nutrition on production
• determine the amount of feed consumed
• determine the composition of the feed consumed.
This module presents some concepts and methods used to collect these data. Table 6.1 summarises
the types of data. For further details, you are referred to the reading materials.
Table 6.1. Types of animal data used to diagnose animal nutrition problems.
Objective Type of data
Production effects Liveweight gain, condition scores, traction power,
milk and wool production
Amount of feed consumed Feed intake
Composition of feed consumed Rumen fistula samples, faecal samples and grazing
behaviour
Detailed methods for data collection are described in ILCA (1990).
6.3 Feed intake
Feed intake is the amount of feed an animal consumes. It can be estimated by using either digestibility data
or markers.
6.3.1 Using digestibility data
When digestibility data are available, intake can be estimated by multiplying the dry-matter (DM) weight
of faeces by a digestibility factor. The factor is known as the feed:faeces ratio and is expressed as:
DM intake = 100/(100 ? digestibility) × DM weight of faeces where digestibility is in per cent
For example if the dry-matter weight of faeces of an animal is 870 g/day and the percentage digestibility of
the feed consumed is 60%, then the amount of dry feed consumed would be:
DM intake (g/day) = (100/40) × 870 = 2175
6.3.2 Using markers
Digestibility and intake data can be derived from the indigestible components of a diet, known as
‘markers’. Markers are classified as internal, if they are ordinarily present in the diet (e.g. lignin), or as
external if they are added to the diet (e.g. chromic oxide, iron oxide and barium sulphate). They are used
when the measurement of feed intake and faecal output is difficult.
The formula used to estimate faecal output is:
Faecal output (g) = Grams of marker in feed per day/Concentration of marker in faeces
For example an animal is dosed with 50 g of chromic oxide per day to determine its daily faecal output.
The concentration of marker in the dry-faeces sample is 5.75% and the dry-matter weight of the faecal
output is:
Faecal output = 50/0.0575 = 870 g
If the proportion of marker in the diet is 3.4%, then we calculate dry-matter intake as follows:
DM intake (g/day) = 870 × (5.75/3.4)
81
6.4 Nutrient value of feed
It is important to know the supply of different nutrients to ruminant livestock in systems in relation to their
need for these nutrients. This module will cover energy, protein, water, minerals and vitamins. Methods to
estimate amounts of these nutrients can be found in ILCA (1990).
6.4.1 Energy
The energy yield of a source of feed (such as natural pasture) can be estimated from its dry-matter weight
per unit area. Feeds with a high biomass per unit area are often low in energy since they also contain a high
proportion of indigestible fibrous matter.
6.4.2 Protein
Protein is the basic structural material from which all body tissues (e.g. muscles, nerves, blood cells) are
formed. It is therefore essential for production and maintenance and cannot be replaced by other nutrients
in the feed. The requirements of the animals are for essential amino acids, which are the building blocks of
the protein. This is expressed as protein requirements, where crude protein (CP) is obtained by using this
equation:
Crude protein = % nitrogen × 6.25
where the value 6.25 is based on the assumption that feed protein contains, on average, 16% nitrogen.
Crude protein is highly related to grain yield at the time of harvest (Church and Pond 1982). On a per
hectare basis, grain yield was related to CP in the leaf. Such relationships can be used to indicate the
availability of CP in the different sources and/or at different stages of plant growth.
When estimating the CP content of browse plants and crop residues, it should be kept in mind that
the presence of certain chemical compounds (e.g. tannins) in these feeds can affect the availability of
nitrogen to the ruminant.
It is important to know that ruminants are able to synthesise protein from non-protein nitrogen
sources (e.g. urea) by microbial action in the rumen.
6.4.3 Minerals
Ruminants require essential minerals as part of their feed for tissue growth and regulation of body
functions. If one of these minerals is missing in the diet, symptoms of deficiency occur; if any mineral is in
excess, the animal experiences toxicity. Usually under extensive livestock systems in the tropics, mineral
imbalances are rarely seen. Analysis for minerals should only be attempted if mineral deficiencies are
clearly evident. Even then, if other nutrients such as energy or crude protein are more limiting (as is likely
to be the case in African rangelands), the mineral constraint should be dealt with only after the primary
deficiencies have been rectified. A more detailed account of symptoms of mineral deficiency and the role
of minerals in ruminant nutrition can be found in basic nutrition texts (e.g. Church and Pond 1982).
6.4.4 Vitamins
Ruminant livestock require vitamins, in very small amounts, to regulate body functions. Although vitamin
deficiencies do occur in ruminant animals in the tropics, in general they are of minor importance. Under
more intensive systems (e.g. peri-urban dairy), providing vitamin supplements is advised. Many of the
vitamins are synthesised in the rumen, especially many of the members of the vitamin B complex and
vitamin C. If animals are exposed to the sun, they are able to manufacture vitamin D. Vitamin A is
essential during pregnancy. Small ruminants can obtain provitamin A from green forages and convert it
into vitamin A. Young herbage and cereal grains contain adequate amounts of vitamin E.
6.4.5 Water
Ruminants require water because:
• it affects voluntary intake; less water leads to inadequate intake of dry matter
• it is needed to maintain the water content of the body.
82
For animals kept under the pastoralist system, the frequency of watering is very important. During
the dry season water is available only from wells or some lakes/streams. This leads to overgrazing around
the watering points forcing pastoralists to move their animals far away from water sources to obtain
pasture. The situation presents a dilemma to nomads because water intake increases as watering frequency
decreases and food conversion efficiency becomes lower as watering intervals increase.
Heat stress increases the water requirement of sheep. It is recommended that sheep and goats be
given water ad libitum. The water requirements increase when the ewe is pregnant or lactating. Lactating
ewes require 50% more water during early lactation and 25% more during late lactation. The extra water
required is equivalent to the amount of milk produced.
6.5 Evaluation of feed
The chemical composition of a feed determines its nutritive value. The difference in energy value of feeds
is due to their differing digestibility. The digestibility of a feed determines the amount that is actually
absorbed by an animal and therefore the availability of nutrients for growth, reproduction etc. A few
methods to evaluate feed for digestibility are listed. [For more details about the methods consult Osuji et al
(1993a)].
6.5.1 In vivo measurement of digestibility
Feeding trials
A feeding trial is used to obtain data to calculate digestibility values. Faeces from several animals is
collected for at least one week following an adaptation period of two weeks. The following equation is
used:
Digestibility = 100 × (DM
food
? DM
faeces
)/DM
food
where DM
food
is the dry matter consumed and DM
faeces
is the dry matter excreted in the faeces.
Nylon bag technique
You place samples of feed in small nylon bags, which are then placed in the rumen through a fistula. The
degradation of the feed is recorded after two days.
Indigestible marker technique
You can mix chromic oxide with the feed. The concentration of chromic oxide in the faeces is used to
calculate digestibility.
6.5.2 In vitro methods for measuring digestibility
• You can use equations to calculate digestibility from chemical analysis.
• Incubate a feed sample in rumen fluid and then transfer it to a pepsin solution. Include one or two
standards which were measured in vivo to monitor accuracy.
• You can use the Menke in vitro gas production technique. In this method gas is produced when feed
is incubated in rumen fluid. The amount of gas is correlated with digestibility. This method
determines the gas produced over a 24-hour incubation period.
6.6 Feed resources
6.6.1 Natural pastures
Natural pastures are the most important feed resource for sheep and goats in sub-Saharan Africa.
Pastures differ in their species and canopy coverage. The canopy coverage depends on the amount of
moisture in the soil and sunlight. UNESCO (1979) adopted the following classification for natural
pastures:
83
• Savannah: It is composed of grasses and shrubs or trees. The acacia grasslands of East Africa fall
into this category.
• Grasslands are composed of pure grasses. There is little true grassland in the tropics. Aquatic
grasslands are found in southern Sudan, around Lake Chad and in the seasonal flooding areas of
East Africa.
The potential carrying capacity of the pasture is determined by the quantity of vegetation. The
productivity of sheep and goats on pasture is affected by the quality of the vegetation. The quantity and
quality of pasture is affected by:
• climate and season
• burning: accidental fires or burning for pasture management
• legume content: the mixture of legumes and grass has a better quality than grass alone.
At the start of its growth, natural pasture has a high water and protein content which is not
advantageous for animals because of limited energy intake; cellulose increases thereafter. Natural pasture
with 30–40% cellulose content is adequate for the maintenance requirements of sheep. Higher cellulose
content becomes detrimental because rumen micro-organisms are unable to cope with its degradation.
This makes pastures with high cellulose content low in nutritive value and hence unable to support the
nutrient requirements of small ruminants. The highest quality of grasses are reached during stem
elongation and the quality decreases after heading.
6.6.2 Improved pastures
The yield and quality of natural pastures can be improved by:
• changing the species composition, e.g. introducing a legume such as Stylosanthes guianensis in a
natural pasture to increase the protein yield per hectare
• improving management, e.g. grazing control or applying fertilisers.
Improving pasture in the dry zones of the tropics is limited by cost such that only simple measures can
be implemented. There is more scope for improvement in the subhumid and humid zones of sub-Saharan
Africa provided a solution is found for control of the tsetse fly and associated small ruminant diseases.
Introducing new species must be done carefully. An aggressive species will dominate the less
aggressive ones. Introduced species should be able to:
• grow in soils of low fertility
• withstand stress, i.e. heat, moisture and waterlogging
• withstand severe grazing pressure
• spread on hill slopes to help control erosion.
The list of species and the methods used to establish and maintain pasture are beyond the scope
of this manual. Specialised texts can be consulted for more details (see Crowder and Chheda 1982).
The successful use of improved pastures depends on grazing control. The use of barbed wire or
electric fences is beyond the means of resource-poor farmers. Recently the use of fodder trees or
nonpalatable shrubs as fences was demonstrated as a possible means of controlling grazing (Charray
et al 1992). However, the use of this technology is limited by labour and inadequate knowledge of the
management of trees/shrubs.
6.6.3 Cultivated pastures or fodder crops
Even though natural pastures are used extensively in sub-Saharan Africa, they can only support the
desired productivity of sheep and goats to a certain level and for a short period. Cultivated pasture or
fodder crops are grown with the aim of:
• increasing forage production per unit area
• improving the feed value of the grasses.
84
When choosing a fodder crop we should ask the following questions:
• Is it easy to sow? And establish?
• How is it to be used?
• What is the length of the vegetative cycle?
Fodder crops are not yet widely used in sub-Saharan Africa because of the high cost of inputs such
as land, labour and fertiliser. These crops are attractive when grown as:
• Catch crops growing in a period when the land is otherwise unused.
• Break crops, which improve the yield of subsequent crops by improving soil fertility or reducing
pests and diseases.
• Intercrops, grown between the rows of another crop to improve soil fertility, provide shade or
control erosion.
Fodder crops must give higher yields or have better quality than pastures. Legume fodder is known
to have these attributes. However, feeding excessive legumes to sheep and goats may cause bloat (the
accumulation of gases in the rumen) in the animals.
Fodder banks were found to improve the nutritional status of livestock during the dry season
(Mohammed-Saleem 1985). These are special plots of legumes grown by pastoralists adjacent to their
homes to serve as supplements to dry-season grazing. For Nigerian pastoralists, ILRI developed fodder
bank technology packages that use low inputs. The farmers received the technology well and researchers
and extension teams worked with them to overcome difficulties related to establishment, utilisation and
regeneration.
Supplementation of roughage
Dry natural pastures can only provide low-quality roughage and therefore small ruminants will benefit
from any form of supplementation of the roughage. Supplements usually consist of at least one of the
following:
• concentrates
• molasses
• non-protein nitrogen (e.g. urea)
• minerals.
They can be fed to animals separately, incorporated into a complete diet or as feedblocks. Important points
to consider when supplementing roughage:
• Supplementation with molasses must be accompanied with urea.
• Urea is not palatable to sheep; mixing it with molasses overcomes this problem.
Supplements to roughage are beneficial to small ruminants because:
• they provide essential nutrients to the rumen micro-organisms.
• the enhanced activity of micro-organisms in the rumen results in better digestibility
• they provide nutrients for the sheep.
6.6.4 Conserved forage
During the dry season, the quality of pasture is low and forages are scarce. One way to solve this problem is
to conserve forages. Conservation aims at retaining the feed value of the forage. Fodder can be conserved
as hay or silage.
Hay
Hay is dried fodder. To produce hay of high quality, it is essential to harvest the fodder at the right time.
When fodder is harvested too early, its moisture content is too high resulting in hay with reduced dry-
85
matter content. If the fodder harvest is delayed, the plants develop a high lignin content. Hay made from
fodder with high amounts of lignin is of poor quality because its digestibility is low. In the humid zone
heavy rainfall forces farmers to delay fodder harvest until the dry season resulting in fodder with high
lignin content. In addition, at the end of the rains most crops are ready for harvest so hay-making competes
with crop harvests for labour.
Hay can be sun dried to 10–15% moisture content. Low air humidity is needed to achieve this level
of moisture in the hay. This is easy to achieve in the arid and semi-arid zones but difficult in the humid zone
where the rainfall and relative humidity are higher.
There are many methods for making hay. The simplest and cheapest method is drying the forage on
the ground. The grass is cut (using a machete) early in the morning and is spread on the ground in the field.
The heaps of forage are turned over many times to avoid picking up moisture from the soil. Sophisticated
methods for drying forages are available but they are costly and beyond the reach of smallholder farmers.
Hay should be protected from rain and the sun as exposure to these factors reduces its feed value. Sheep
and goats may refuse hay of low quality unless they are starving. Hay can never be equal in feed value to
the fresh forage from which it is made.
How do farmers use hay?
• Feed it to animals on maintenance diet.
• Feed it to lambs and kids to accelerate rumen development and the rumination process.
• Use it as roughage for lambs destined for fattening on rations with large amounts of high-energy,
high-protein concentrates.
Silage
Silage is produced by fermenting the sugars in green herbage under anaerobic conditions. The desired
product is reached when enough acids are produced (low pH which prevents bacterial decomposition) and
fermentation stops. Silage can be stored for at least a year.
Silage can be made in silos of various shapes. The easiest method is to use pit and trench silos which
can be built using simple materials. The silo and its wall must be completely airtight to prevent the fodder
putrefying. The bottom of the silo must be covered with stones to allow the liquids resulting from
fermentation to drain off. To ensure successful silage making:
• After harvesting the forage, keep it at low moisture by exposing it to the sun. The harvested forage
should not be exposed to rain.
• Chop the forage into fine pieces (with a machete) of not more than 10-cm length.
• Compress the forage in the silo to remove as much air as possible (treading with feet can be used to
compress forage in small silos).
• Cover the silo as soon as possible with compacted straw or a sheet of plastic (if available) and place
soil on top (a layer of about 50 cm). Compact the upper layer of the soil to maintain good coverage.
6.6.5 Crop residues
There are many crop residues which can be used for feeding sheep and goats. Sheep and goats make better
use of these agricultural by-products in crop–livestock systems. Farmers with no animals may use crop
residues to improve fertility of the soil or sell the residues.
Straw
Straw is mostly a by-product of cereal crops that is used as roughage for ruminants. Straw of the
following crops is used in sub-Saharan Africa: maize, millet, sorghum, wheat, barley, rice and teff.
Residues from other crops are also used on small scale: sugar-cane, cocoa, banana, cotton, cassava
and legume crops.
Straw has low nutritive value. The energy content of straw ranges from 5.5 to 9.6 MJ ME/kg DM.
Energy values vary with the cereal variety and the management of the residue after grain harvest. Straw is
86
high in lignin, which lowers its digestibility. Straws of most crops are low in crude protein and minerals
(especially phosphorus). The nutritive value of straw and its coarse physical form limit the activity of
micro-organisms in the rumen and contribute to the low rate of passage through the digestive system. All
these factors result in low voluntary intake. You can improve the feed value of straw by:
• Adding supplements of energy, protein and minerals, which will give an adequate supply of
nutrients to both the rumen micro-organisms and the animal. Research at ILRI (Getachew et al
1994; Umunna et al 1995) has shown that supplementation with forage legumes is a sustainable
way of improving the feed value of poor-quality crop residues especially for resource-poor African
smallholders. The results indicate that Stylosanthes guianensis, Lablab purpureus and Sesbania
sesban were useful as supplements for sheep in Ethiopia.
• Grinding or chopping the straw or heating it with steam to increase voluntary intake. The cost of
treatment is high because it requires labour.
• Treatment with chemicals, particularly alkalis (e.g. sodium hydroxide, potassium hydroxide,
calcium hydroxide and sodium carbonate) and nitrogenous compounds (urea and ammonium
hydroxide). These treatments will increase digestibility and nitrogen content but they are expensive
and availability of chemicals may pose problems for small farmers. This practice is therefore
suitable only for research stations and a limited number of farmers.
Bran
Bran is produced from cereals and it consists of the outer parts of the hulled grains, some broken grain and
germ. Bran is palatable and is a valuable feed for small ruminants. Its crude protein content is 12%. The
main problem with bran from rice and maize is its relatively high oil content (around 10%). Fatty acids
turn rancid with storage and this lowers the nutritive value and palatability of the bran.
6.6.6 Industrial by-products
Cottonseed cake
Cottonseed cake, when available, is an excellent feed for sheep and goats. It has a high content of protein
(25–40%), fat (10–23%) and cellulose (25–30%). Once the animal gets used to cottonseed cake, intake is
increased. The best results are obtained by mixing the cake with salt and molasses. This was tried with
West African Dwarf ewes and a mixture of 50:50 molasses and cottonseed cake gave good results
(Charray et al 1992). The high protein content allows the amount of cottonseed cake to be reduced without
detrimental effects. Unfortunately cottonseeds contain the toxic substance gossypol which reduces the
87
Example: Feeding sorghum stover
Sorghum (bird-resistant Seredo) was grown at Melkassa in the Rift Valley in Ethiopia. Fourteen
days after hand-harvesting grain in November 1988, stover was hand-cut and transported to the
research station at Debre Zeit. The stover was chopped (tractor-mounted Alvan Blanch Chaff
Cutter) and stored for 1 month before feeding to 24 bucks (12 months old) and 24 rams (15 months
old). Animals were individually penned and given 155 g/day of cottonseed cake and offered one of
the three amounts of stover ad libitum (25, 50 or 75 g/kg animal mass per day). Salt lick and water
were provided. This continued for 75 days. The results showed that both goats and sheep are
capable of selective feeding leading to increased intake and growth when they are offered
increasing amounts of chopped sorghum stover ad libitum. Growth rate of rams was double that of
bucks. The generous feed strategy used offers a method of alleviating the low nutritive value of
sorghum stover as a dry-season feed for small ruminants.
Source: Aboud et al (1993).
growth of young animals but has little effect on adults. There are cotton varieties that are glandless and
their seeds are free of gossypol.
Oilseed cakes
Oilseed cakes (oilmeals) are by-products of processing a variety of oil crops: groundnut, sunflower and
soybean. These cakes are like cottonseed cakes: rich in protein and fatty acids. They can be used alone or
mixed with molasses for effective results. Research at ILRI (Osuji et al 1993b) showed that sunflower
cake was utilised effectively by Menz sheep in the Ethiopian highlands in terms of rumen microbial
nitrogen synthesis, nitrogen retention and growth. The addition of small amounts of energy such as
crushed maize grain increased microbial nitrogen synthesis, nitrogen retention and liveweight gain. The
cost and availability of the cakes will determine the likelihood of their adoption by farmers.
Brewery by-products
When beer is made, the residues are the spent grains and yeast. Sheep and goats readily accept these as
feed. Sources for these by-products are beer factories, which are increasing in sub-Saharan Africa,
(including home-made installations). The by-products from home installations are richer in energy and
protein than the residues from the factories.
Sugar industry by-products
The by-products from sugar-cane factories are dried sludge, molasses and bagasse.
Sheep can be fed all three by-products. Farmers mostly use molasses, a thick dark brown liquid which
contains 50–65% sugar with little protein or water. It is thus a high energy feed. When added to cottonseed
cakes, it increases the intake of coarse and less readily accepted cakes.
It can be added to the herbage during silage production. Molasses therefore is an excellent feed
provided it is supplemented with protein and minerals.
6.6.7 Fodder trees
The integration of sheep and goats with tree crops is common practice in many places in sub-Saharan
Africa. Sheep and goats are allowed to graze under trees, e.g. goats and sheep feed on the leaves and pods
of Faidherbia albida in the parklands of West Africa. The animals are fed the branches of trees and
subsequently manure is deposited under the trees. Farmers in the Kenya highlands feed leaves of
Leucaena to small ruminants and cattle. The dry-matter yield of Leucaena is between 2 to 20 tons per
88
Example: Supplementation with molasses
At ILRI's research station in Debre Zeit, Ethiopia, four ruminant fistulated crossbred (Bos taurus ×
Bos indicus) steers (mean live weight 468 kg) were fed twice daily a basal diet (Diet MO)
comprising native grass hay ad libitum (free eating) and 4 kg dry matter (DM) of wheat bran. Three
other diets were tested: 1. Native grass hay plus 750 g wheat bran per kg DM and 250 g molasses per
kg DM; 2. Native grass hay plus 500 g wheat bran per kg DM and 500 g molasses per kg DM; and 3.
Native grass hay plus 250 g wheat bran per kg DM and 750 g molasses per kg DM. The native grass
hay was predominantly Pennisetum species. Wheat bran and molasses were mixed and fed together
from a bucket. Increasing the level of molasses did not affect the intake of hay. There was no effect
on digestibility of wheat bran. It was concluded that replacing wheat bran with different levels of
molasses did not affect the function of microbes in the rumen. Rumen microbes utilised energy
equally from molasses fermentation compared with wheat bran.
Source: Osuji and Khalili (1994).
hectare per year. The tree leaves contain mimosine which is toxic to sheep. Many other tree species are
used for fodder, e.g. Sesbania sesban, Calliandra spp and Gliricidia sepium. Fodder from trees is
especially useful during the dry season when it is used to supplement roughage or hay. Research at ILRI
concluded that Erythrina abyssinica has high forage potential and can effectively serve as a cheap source
of protein supplement for low quality diets during the dry season for resource-poor farmers with stall-fed
sheep and goats (Larbi et al 1993). Supplementation of pasture in Zimbabwe with 110 g Acacia
angustissima per goat per day was found to result in an increase of 12 g per goat per day.
6.7 Feeding techniques
This module will only provide a list of feeding techniques. For more information consult Charray et al
(1992).
The following are common feeding techniques:
• Grazing: Pastoralists feed their sheep and goats on rangeland. Sheep and goats are free to graze
for at least 8 hours per day. Pastoralists have developed complex management systems for
grazing.
• Continuous grazing: This method is used when an area of forage is allocated for sheep and
goats to graze freely. The enclosed area is calculated based on the carrying capacity of the
pasture.
• Rotational grazing: This method leaves time for the grass to recover between successive periods of
use. The stock grazes a plot for a specified time then moves to another plot, returning to previously
grazed plots when they recover.
• Zero-grazing (also known as cut-and-carry): In this method feed is brought to the sheep and goats
which remain confined.
6.8 Exercises
1.
A lamb is given a diet of 11.0 MJ ME per kg DM. Its DM intake is 511 g per day. Assuming it eats
ad libitum, and that energy is the limiting nutrient, how fast will it grow?
2. Do farmers in your region conserve forages to feed to small ruminants? If so, describe
indigenous methods used by farmers in your area to conserve feed for their small ruminants.
89
Example: Feeding multi-purpose trees as supplements
Many multi-purpose trees have been tested as supplements to feed of livestock ruminants. In this
example we will focus on results of an experiment that used Erythrina abyssinica. The literature has
ample examples of other fodder trees. The feeding value of E. abyssinica leaves was evaluated with
native sheep and goats at Soddo, southern Ethiopia, in a 21-day digestibility trial and an 80-day
growth trial. In the digestibility trial, wilted leaves were fed ad libitum (freely), while in the growth
trial, animals were offered a basal diet of Pennisitum purpureum with 3 levels (0, 500 and 1000
g/head per day) of E. abyssinica leaf. When levels of E. abyssinica leaf supplementation were
increased, liveweight gains of sheep and goats increased. Sheep gained weight faster than goats. It
was concluded that E. abyssinica has high forage potential and can effectively serve as a cheap
source of protein supplement for low quality diets during the dry season for resource-poor farmers
with stall-fed sheep and goats.
Source: Larbi et al (1993).
3. Mark the correct answer:
A. Multi-purpose trees are useful as supplements because they can provide nitrogen and they
all lack toxic substances.
B.
Multi-purpose trees are useful supplements because they can provide nitrogen but a few
contain toxic substances.
C.
Multi-purpose trees are useful supplements because they can provide nitrogen in the diet
of the ruminant and the rumen can detoxify any toxic substances that a few MPT species
may contain.
4. Smallholder farmers use more crop residues than industrial by-products. List two reasons for
this behaviour.
5. List the techniques used by farmers to feed small ruminants.
6.9 References and reading materials
Aboud A.A.O., Owen E., Reed J.D., Said A.N. and McAllan A.B. 1993. Feeding sorghum stover to
Ethiopian goats and sheep: effect of the amount offered on intake, selection and performance. In:
Gill M., Owen E., Pollott G.E. and Lawrence T.L.J. (eds), Animal Production in Developing
Countries. Occasional Publication 16. BSAP (British Society of Animal Production), UK.
Charray J., Humbert J.M. and Levif J. 1992. Manual of Sheep Production in the Humid Tropics of
Africa. CAB (Commonwealth Agricultural Bureaux) International, London, UK and CTA
(Technical Centre for Agricultural and Rural Co-operation), Wageningen, The Netherlands.
Church D.C. and Pond W.G. 1982. Basic Animal Nutrition and Feeding. 2nd edition. John Wiley and
Sons Inc., New York, USA. 403 pp.
Crowder L.V. and Chheda H.R. 1982. Tropical Grassland Husbandry. Longman, London, UK.
Devendra C. and Burns M. 1983. Goat Production in the Tropics. CAB International, Slough, UK. 183 pp.
Gatenby R.M. 1986. Sheep Production in the Tropics and Sub-tropics. Longman, London, UK. 351
pp.
Getachew G., Said A.N. and Sundstal F. 1994. The effect of forage legume supplementation on
digestibility and body weight gain by sheep fed a basal diet of maize stover. Animal Feed Science
and Technology 46:97–108.
ILCA (International Livestock Centre for Africa). 1990. Livestock Systems Research Manual.
Volume 1. ILCA Working Paper 1. ILCA, Addis Ababa, Ethiopia. 287 pp.
Jarrige R. 1989. Ruminant Nutrition: Recommended Allowances and Feed Tables. Institut national de
la recherche agronomique (INRA), John Libbey Eurotext, London, UK. 389 pp.
Larbi A., Thomas D. and Hanson J. 1993. Forage potential of Erythrina abyssinica: Intake,
digestibility and growth rates for stall-fed sheep and goats in southern Ethiopia. Agroforestry
Systems 21:263–270.
MAFF (Ministry of Agriculture, Fisheries and Food). 1975. Energy Allowances and Feeding Systems
for Ruminants. Technical Bulletin 33. HMSO (Her Majesty’s Stationery Office), London, UK.
Mohamed-Saleem M.A. 1985. Forage legumes in agropastoral production systems within the
subhumid zone of Nigeria. In: Kategile J.A. (ed), Pasture Improvement Research in Eastern and
Southern Africa. Proceedings of a Workshop held in Harare, Zimbabwe, 17–21 September 1984.
IDRC (International Development Research Centre), Ottawa, Canada. pp. 222–243.
Osuji P.O. and Khalili H. 1994. The effect of replacement of wheat bran by graded levels of molasses on
feed intake, organic matter digestion, rumen fermentation and nitrogen utilisation in crossbred (Bos
taurus × Bos indicus) steers fed native grass hay. Animal Feed Science and Technology 48:153–163.
90
Osuji P.O., Nsahalai I.V. and Khalili H. 1993a. Feed Evaluation. ILCA Manual 5. ILCA
(International Livestock Centre for Africa), Addis Ababa, Ethiopia. 36 pp.
Osuji P.O., Sibanda S. and Nsahlai I.V. 1993b. Supplementation of maize stover for Ethiopian Menz
sheep: Effects of cottonseed, noug (Guizotia abyssinica) or sunflower cake with or without
maize on intake, growth, apparent digestibility, nitrogen balance and excretion of purine
derivatives. Animal Production 57:429–436.
Umunna N.N., Osuji P.O., Nsahlai I.V., Khalili H. and Saleem M.A. 1995. The effect of
supplementing oats hay with either lablab, sesbania or wheat middlings and oats straw with
lablab on the voluntary intake, nitrogen utilization and live weight gain of Ethiopian Menz sheep.
Small Ruminant Research 18:113–120.
UNESCO (United Nations Educational, Scientific and Cultural Organization). 1979. Tropical
Grazing Land Ecosystems. UNESCO, Paris, France.
Van Soest P.J. 1982. Nutritional Ecology of the Ruminant. O and B Books Inc., Oregon, USA. 374 pp.
91
Module 7: Economics of small ruminant production
7.1 Performance objectives
Module 7 is intended to enable you to:
1. List basic production function relationships.
2. Write the functions for factor–product, factor–factor and product–product relationships.
3. Describe the production relationships as they relate to small ruminant production.
4. Define partial budget analysis.
5. Describe the logic behind partial budgets.
6. Describe the framework for partial budget analysis.
7. List the essential data required for computing partial budget analysis.
8. Given production parameters data, corresponding cost and prices data, compute partial budgets.
9. Given computed partial budgets, interpret the results.
7.2 Introduction to production relationships
Agricultural economists are usually members of multidisciplinary teams solving problems of livestock
agriculture. The economist applies economic concepts and tools to estimate economic profitability of a
technology. This input can be done during any phase of the research process: technology development,
testing and evaluation on-farm. Details of economic theories and tools are beyond the scope of this
module. In sub-Saharan Africa most research teams lack an economist because only a few people have
specialised in this field. Since the majority of the audience for this manual are non-economists, this
module will briefly touch on relevant economic concepts and focus on partial budgeting, a tool that we
think can be used by non-economists to quickly gain insight into the profitability of the technologies
without having to wait until an economist conducts detailed analysis of the data. Farmers need to make
decisions about allocation of their resources on a day-to-day basis as well as on a long-term basis. This
includes decisions related to the whole farm such as what crops to grow, what animals to raise, what type
of system, what inputs to use and how to market. A farmer has to answer four basic questions:
• What should the farm produce?
• How should it be produced?
• How much should be produced?
• How should it be marketed?
93
7.1 Performance objectives
7.2 Introduction to production relationships
7.3 Partial budget analysis
7.4 Exercises
7.5 References and reading materials
To effectively evaluate a technology, we need to know the resources available to the farmer and his
risk considerations. Farmers take into account: the family subsistence requirements, marketing
opportunities and availability of resources (land and labour). It is important also to consider factors related
to a farmer's choice of production technologies. New technologies are not automatically superior to
traditional practices. The farmer wants to produce at the least cost and with little deviation from traditional
practices. Marketing factors, such as prices, transportation and storage facilities need to be considered into
our analysis. We will concentrate on how economics helps to answer the first three questions since we are
interested in raising the productivity of small ruminants and will concentrate therefore on technological
change.
7.2.1 Production relationships
Farmers obtain outputs when they use inputs (factors of production). The farmer decides the amount of
inputs he will use to meet his output goals. The relationship between factors of production and output are
best described by the following equation:
Y = f(X
1
,X
2
,X
3
...X
n
),
where Y is the output (e.g. meat yield, milk yield) that is obtained as a result of using inputs X
1
,X
2
,X
3
etc
(e.g. tons of hay, vaccination, fertiliser) and f is the symbol for function. The output and production factors
are called variables. Variation in one of these quantities will lead to variation in the other. The equation
Y=f(X) is termed the production function which means Y is a function (f) of X. We will describe the three
basic relationships that involve these variables:
• Factor–product relationship: how output is related to a single variable input holding other inputs fixed.
• Factor–factor relationship: how output (held constant) is related to two or more variable inputs.
• Product–product relationship: how the relative quantity of two or more outputs is related to a fixed
quantity of inputs.
Factor–product relationship
In this relationship one input is varied and the level of other inputs is kept constant. The production
function to express this is:
Y = f(X
1
? X
2
,X
3
...X
n
),
where the symbol ? means that only X
1
is varied while X
2
to X
n
are fixed. This relationship could lead to
constant output, increasing output or decreasing output. The concepts marginal product, constant
marginal product, increasing marginal product, decreasing marginal product and average product will be
used to describe the factor–product relationship.
Marginal product. Marginal product is the change in output due to a change in variable input. The
following function describes it:
Marginal product = ?Y÷?X
The symbol ? means change and is known as ‘delta’. For example:
Milk = f(feed? land)
where the change in milk output is ?Y and the change in feed ?X. ?Y÷?X means the change in amount of
milk due to a specified change in amount of feed; the factor land is held constant. This relationship is used
in agricultural decision making. This relationship compares the value of an additional quantity of output to
the cost of the additional input required to produce the marginal product.
MP
milk
= ?milk produced ÷?feed quantity
Thus the farmer will choose an input and decide what quantity of this input to use and how much
output to produce.
94
The marginal product can be constant and this happens when, within a range of input levels, the
application of each additional unit of input, yields equal increments in output (Figure 7.1). This
relationship is rare in agriculture. Sometimes, for a range of input levels, each additional unit of input
results in increasing increments of output (Figure 7.2). This is possible when the fixed input is in excess in
relation to the variable resource. This situation is also rare in agriculture. The most typical relationship in
agricultural production is decreasing marginal product (Figure 7.3). For a range of input levels, each
additional unit of input yields an incremental output which is less than that from the previous increment in
input.
Figure 7.1. Constant marginal productivity.
Figure 7.2. Increasing marginal productivity.
Figure 7.3.
95
B31B39B30B30
B31B37B30B30
B31B35B30B30
B31B33B30B30
B30
B31B30 B32B30 B33B30
B34B30
B41B6dB6fB75B6eB74 B6fB66 B66B65B72B74B69B6cB69B73B65B72 B28B6bB67B29
B41B6dB6fB75B6eB74 B6fB66 B70B72B6fB64B75B63B74 B28B6bB67B29
Decreasing marginal productivity.
Average product. This is the ratio of output to input, i.e. Y÷X, where Y is the total output and X is the
total input. The following relationships exist between average product (AP) and marginal product
(MP):
• when MP>AP, AP is increasing
• when MP=AP, AP is at maximum
• when MP<AP, AP is decreasing.
AP measures the technical efficiency of a process of production.
Relationships between marginal product and total product. The following relationships hold between
marginal product and total product as the level of input increases:
• when total product is increasing at an increasing rate, marginal product is positive and increasing
• when total product is increasing at a decreasing rate, marginal product is positive and decreasing
• when total product is maximum, marginal product is zero
• when total product is decreasing, marginal product is negative.
The production relationship (Figure 7.4) is divided into stages I, II and III. The use of resources in
stage I only is uneconomical. In such a case the farmer is using too little of the variable input given the level
of fixed inputs that are available. One can increase the level of variable input up to the border between
stage I and stage II to reach higher AP. If this is not possible, then some of the fixed inputs can be shifted to
other uses to produce more products. It is also irrational to add inputs beyond the point of maximum TP
(8X in Figure 7.4) since the marginal product is negative, i.e. too much input is used relative to the fixed
inputs. In this case it is advisable to use less of the variable resources. Stage II represents the area of
rational production, i.e. economic relevance. We need to know the production function and input/output
price ratio to be able to locate the precise economically efficient level of variable input.
96
246810
0
6
12
18
24
Amount of concentrate (kg)
Amount of
milk (kg)
Do farmers maximise production? No, farmers maximise profit but they do not necessarily
maximise production. When TP is maximum (Figure 7.4) MP is zero, i.e. Y=0 and MP=Y/X=0. Do
farmers stop at the border between stage I and stage II, i.e. at X=5? No. Depending on the price, farmers
keep using additional input even though MP is decreasing.
Figure 7.4. Stages of production and the relationships between marginal product (MP) and average
product (AP).
Hypothesis: Farmers maximise profit
Prices of inputs ? costs
Prices of outputs ? revenue
Profit is the difference between revenue and cost:
Profit = Revenue ? Cost
Revenue = Output (Y) × Pricey (Py)
Cost = Input (X) × Pricex (Px)
Profit = (Y × Py) ? (X × Px)
Maximum profit = Revenue ? Cost
(Production within stage II)
Maximum profit = Marginal revenue (MR)
= Marginal cost (MC)
Maximum profit occurs when added return (marginal revenue) equals added cost (marginal cost)
and production is occurring within stage II:
MR=MC
MR = ?Y × Py
MC = ?X × P
x
97
TP= Total product I ? Total product increasing at increasing rate
AP= Average product II ? Total product increasing at decreasing rate
MP= Marginal product III ? Total product decreasing
8
MP
AP
Total input, X
Total output, Y
54
42
30
18
6
42 6
TP
I II III
0
10
where P
y
is the price of output and Px is the price of input,
?Y × P
y
= ?X × P
x
?Y÷?X=P
x/
P
y
Px ÷ P
y
= input/output price ratio
MAX profit = MR = MC
In real farm situations farmers do not use levels of inputs that maximise profits because:
• the farmer’s knowledge of the value of resources is imperfect and he/she is unsure of input/output
relations
• the farmer is faced with risk such as uncertainty in future prices or future yields; to reduce the risk,
the farmer discounts future returns
• the farmer is short on capital to buy the necessary inputs that will maximise profit.
Factor–factor relationship
In this production relationship we have a number of variable inputs (factors). The following function is
used to express this relationship:
Y = f(X
1
,X
2
/X
3
,X
4
...Xn),
where X1 and X2 are variable. The aim in this relationship is to get the least-cost combination of inputs
that give the desired level of output. A farmer can use different combinations that produce the same level
of output. The substitution process has limits as a minimum of certain inputs are needed for a given output.
Factor–factor relationships can be increasing, constant or decreasing. The only rational economic
relationship is when the marginal rate of substitution between inputs is decreasing. The farmer will need to
know the relative price differences of the inputs and the output prices. This will help to decide which
combination of inputs to use to produce a certain level of output, i.e. the farmer desires the least-cost
combination of inputs that maximise profit. This can be obtained when the marginal rate of factor
substitution equals the inverse factor price ratio:
MR = MC, i.e. MR ? MC=0
(MC is at the minimum cost)
MC = ?X
1
.P
x1
+ ?X
2
.Px
2
=0
?X
1
.P
x1
= ?X
2.
P
x2
Slope = ?X
1
÷ ?X
2
=PX
2
÷PX
1
That is when the value of the marginal input added (?X
2.
P
x2
) equals the value of the marginal input
reduced (?X
1
.P
x1
), the resource combination is optimal. Figure 7.5 shows this relationship: when the
slope of isocost line (PX
2
/PX
1
) touches the isoproduct curve (?X
1
/?X
2
).
Product–product relationship
In this relationship two outputs are produced when the level of inputs is fixed. The farmer desires to
produce optimal combination of outputs for a number of fixed inputs. The farmer aims at maximising
revenue since the cost is fixed. The two outputs can be complementary, supplementary or competing for
the fixed resources. Therefore, marginal analysis is used to decide on the best combination of outputs.
Figure 7.6 shows a hypothetical transformation curve describing the three possible combinations. The
slope of the product transformation curve is ?Y
1
/?Y
2
. The two outputs generate different revenues. The
equation for total revenue from the two outputs:
R=Y
1
P
Y1
+Y
2
P
Y2
98
where R is total revenue, P
Y1
is the price of output Y
1
and P
Y2
is the price of output Y
2
. The optimum
output combination is reached when ?Y
1
/?Y
2
=P
Y2
/P
Y1
. This occurs when the slope of the isorevenue
line PY
2
/PY
1
is equal to the slope of the product transformation curve, ?Y
1
/?Y
2
(see Figure 7.6). The
combination is complementary in region A–B, i.e. increasing the production of Y
1
will increase the
production of Y
2
. For example legume crops fix nitrogen which is used by the other intercrop (a non-
legume). Region B–C is for outputs that are supplementary, i.e. increasing the production of Y
1
does not
affect the production of Y
2
. For example keeping a few milking goats on the farm does not affect the
productivity of producing corn as the sources of labour are not competitive. As long as both products have
positive prices, it is not economical to produce in region A–B or B–C. It is advisable to keep producing
until we reach the competitive point, i.e. the point where the isorevenue line touches the product
transformation curve.
99
B31B32
B31B30
B38
B36
B34
B32B2eB35
B35B31B30B31B35 B32B30B32B35
B49B73B6fB63B6fB73B74 B6cB69B6eB65
B49B73B6fB70B72B6fB64B75B63B74 B63B75B72B76B65
B41B6dB6fB75B6eB74 B6fB66 B62B65B72B73B65B65B6d B28B6bB67B29
Amount of con
centrate (kg)
Y2
A
B
D
Y1
Product translation curve
Isorevenue line
A B Complementary region
B C Supplementary region
C D Competitive re gion
C
Figure 7.5. Least cost combination of two inputs.
Figure 7.6. Hypothetical product transformation curve.
7.3 Partial budget analysis
Partial budget analysis is the tabulation of expected gains and losses due to a relatively minor change
(marginal) in farming method or technology, e.g. feeding concentrates to sheep. The new technology
could be technically feasible but this is not a necessary condition for adoption by farmers; the new
technology must be profitable. Therefore it is important for the team of scientists developing the new
technology or improving the existing one to determine the profitability of the technology. However, most
national research systems do not have enough economists, and biological scientists often lack the ability to
conduct economic analysis. This module therefore aims at describing a simple technique to determine the
profitability of technologies at farmer level—the partial budget analysis. Livestock scientists need to be
equipped with basic economic tools to evaluate small ruminant technologies. Partial budget analysis can
be used to estimate the likely economic impact of a technology, before it has been tested (ex ante analysis),
and after the trial has been completed, when actual data is available.
Partial budgets list only those items of income and expenses that change. They (i) measure change
in income and returns to limited resources (ii) provide a limited assessment of risk and (iii) suggest a range
of prices or costs at which a technology is profitable. After the changes are determined, the relationship is
shown by the following equation:
[Added returns + Reduced costs] ? [Added costs + Reduced returns] = [Profit or
loss]
As stated earlier, the farmer's goal is to maximise profit and not necessarily only productivity.
Determining the farmer’s goal is a good start. For example the household may set as a goal increasing
income during drought periods and achieve this goal by introducing sheep in addition to cropping. The
advantages and disadvantages of partial budgeting are listed in Box 7.1.
7.3.1 Logic behind partial budget analysis
The farmer's goal is to maximise returns. The following equation describes this relationship:
NI=TR?TC
where NI is the net income (e.g. from sale of sheep), i.e. the amount of money left when total costs (TC) are
subtracted from total returns (TR). Total cost (TC) is composed of the cost of all inputs: variable costs
(VC) and fixed costs (FC). Since FC are the same for comparing the new technology and the technology
practised by the farmer, then:
TC=VC
The farmer wants to know if the new technology will increase his income while the researcher
wants to know which technology is potentially more attractive economically. The change in income can
be expressed as follows:
??? = TR ? ?TC
??? = TR ? ?VC
This relationship can be used as a rule of thumb for partial budgeting, since ?FC = 0 because by
definition fixed costs do not vary.
If capital is not a constraint, choose the highest ?NI. Since higher benefits require higher costs, it is
necessary to compare the extra (or marginal) costs with the extra (or marginal) benefits.
Marginal rate of return (MRR) is another way of taking the cost factor into account. It measures the
?NI which is generated by each additional unit of expenditure (?VC):
MRR = ?NI/?VC
MRR is also a measure of additional capital invested in a new technology and its effect on net return.
100
Researchers make recommendations to farmers about what to adopt. To make these
recommendations, the following economic criteria should be observed (provided the technology is in line
with the farmer’s goal, sustainable etc):
• If net income remains the same or decreases, the new technology should not be recommended as it
is not more profitable than the farmer’s current technology.
• If net income increases and variable costs remain the same or decrease, the new technology should
be recommended as it is more profitable than the farmer’s current technology.
• If both net income and variable cost increase (the most common case), the MRR should be
considered. The greater the increase in ?NI and the higher the MRR, the new technology is more
economically attractive.
101
Box 7.1. Advantages, disadvantages and appropriate situations for partial
budget analysis
Advantages
• The technique is simple (it can be performed with a hand calculator) and easy to learn for non-
economists.
• It examines only net changes in costs and benefits, therefore it is effective for assessing
economic viability of single intervention technologies.
• It requires less data than whole-farm budgeting since fixed costs are not examined.
• The data required for partial budget analysis are collected for almost all other economic analyses.
• It allows early conclusions about the adaptability of a new technology.
Disadvantages
• There is a danger of neglecting the fact that farmers’ resources are limited and sometimes
knowledge about the resource base may be lacking.
• Often scientists do not understand the farmer’s objectives; the partial view of a farming system
might obscure the secondary character of a given farm component.
• The time for various farming activities which may overlap with or contradict a new schedule is
often overlooked.
• Linearity of factors is assumed by using small-scale input/output data for a large-scale
operation. The assumption that an increase in a unit of resource will increase the profit
proportionally is questionable. Similarly, assuming that most new technologies do not require
additional management skills is also questionable.
Appropriate situations for partial budget analysis
Partial budget analysis is appropriate where:
• a single component must be analysed
• inputs and outputs are measurable and easy to price
• animal yields vary little between farms
• profitability is the major concern rather than issues such as equity and income distribution
• fixed costs do not change
• economic evaluation is required for new technologies which are not yet well developed.
Partial budgets for animal enterprises are essential because:
• animals have a longer life cycle than food crops
• animals use non-market inputs (crop residues, native pastures, labour and land) and market prices
often do not exist
• animals have multiple outputs (e.g. meat, milk and manure)
• the life cycle of animals is not synchronised which makes it difficult to measure.
7.3.2 Data required for partial budget analysis
The most important step in performing partial budget analysis is the proper identification of data on the
costs and benefits associated with the alternative technologies. Generally the following are essential data
that must be collected:
• quantities of inputs which vary between alternative technologies
• prices of these variable inputs
• yields or productivity levels resulting from the alternative technologies
• prices of the outputs valuing non-market inputs or products opportunity cost (the value of the
resource or product in its next best alternative use, e.g. family labour vs market labour wages).
Important products of small ruminants include reproductive capacity (offspring), milk yield, meat
yield (weight gain), manure, skins and wool.
Inputs depend on the technology being used. Input costs should include cash costs (e.g. feed) and
non-cash costs (family labour, capital costs, depreciation costs, crop residues and household wastes).
A logical framework for the steps of the partial budget analysis is given in Box 7.2. All benefits and
costs should be calculated using farm-gate prices (the actual price which the farmer pays for the inputs or
receives for his products). Input prices should account for all costs, a farmer selling animals will encounter
transportation costs, storage charges and marketing costs. If a technology affects the quality of the sheep
and goats, market prices should reflect the quality factor.
102
Box 7.2. Partial budgeting format
1.
Additional income: list the items of income from the alternative plan that will increase or not
be received from the base plan.
2.
Reduced expenses: list the items of expense for the base plan that will be avoided or reduced
with the alternative plan.
3.
SUBTOTAL (?TR):1+2
4. Reduced income: list the items of income from the base plan that will decline or not be
received from the alternative.
5. Additional expenses: list the items of expense from the alternative plan that are not present in
the base plan.
6. SUBTOTAL (?VC):4+5
7. NI difference (?TR–?VC): 3–6: A positive (negative) difference indicates that the net
income of the alternative exceeds (is less than) the net income of the base plan by the amount
shown.
8. If there is capital constraint: MRR = ?NI ÷ ?VC
Table 7.1 shows the breakdown of the benefits and costs for economic analysis of small ruminants.
Care should be taken in determining the value to assign to inputs and output when calculating partial
budgets. In some cases opportunity cost needs to be used. For example, manure provides a low cost
fertiliser for crops and in this case it can be given the cost of reduction in fertiliser use. Family labour is
another example. Rural wages for hired labour can be used to evaluate family labour. When labour is not
an option for household members, it is difficult to evaluate. Animals or animal products that are consumed
by the household can be valued against the purchase value in the market. Tables 7.2 and 7.3 present
checklists of costs of and benefits from producing ruminants and pigs which must be considered when
computing partial budgets.
Table 7.1. Breakdown of benefits and costs for economic analysis of small ruminants.
Benefits Costs
Primary
Increase milk yield
Increase litter size
Culls (ewes, rams, does and bucks)
Goat meat and lamb or mutton
Mohair and wool
Hides
Manure
Horn and hooves for processing into feed
supplements and other products
Meat and milk by-products
Primary
Cash costs
Feed costs: hay, salt and minerals,
concentrates, grains, feeds for the young
Non-feed costs: veterinary medicine and
drugs, vaccine
Wages for hired labour
Pasture rent
Non-cash costs
Family labour
Depreciation of facilities, equipment etc
Other non-market feed costs
Secondary
Urine
Weight increase (realised when theanimal
is sold)
Weeding–grazing
Farmers’ preference
Attractiveness of colour and appearance
Seed distribution
Research
Secondary
Carry disease
Destruction of crops, trees, trampling of
land (which causes ecological damage)
103
Table 7.2. Checklist of costs involved in producing ruminants.
Type of cost Form of cost
Primary costs
Variable costs
Feed costs
Concentrates Cash
Grass and hay Cash/non-cash
Mineral/supplements Cash
Grain Cash/non-cash
Water Cash
Other costs
Medicines/vaccines Cash
Veterinary services Cash
Breeding fees Cash/non-cash
Supplies Cash
Milk hauling/marketing Cash/non-cash
Transportation Cash/non-cash
Utilities Cash
Hired labour Cash
Other labour(family/exchange) Non-cash
Stock replacement Cash/non-cash
Fixed costs
Housing/bedding Cash/non-cash
Beginning stock Cash
Land rent Cash/non-cash
Depreciation
Taxes, interest
Secondary costs
Destruction of crops Cash/non-cash
Trampling of land Non-cash
Noise, disease Non-cash
Foul odour (animal and/or manure) Non-cash
Threat to safety of small children Non-cash
Table 7.3. Checklist of benefits of producing ruminants.
Type of benefit Form of benefit
Primary benefit
Milk Cash/non-cash
Meat Cash/non-cash
Work (large ruminants) Cash/non-cash
Hides Cash/non-cash
Manure Cash/non-cash
Horn and/or hooves for feeds Cash/non-cash
Horn and/or hooves for art Cash/non-cash
Table 7.3. cont...
104
Table 7.3 cont...
Meat and milk by-products Cash/non-cash
Transport (large ruminants) Cash/non-cash
Capital asset Cash
Reproduction Cash/non-cash
Secondary benefit
Urine Non-cash
Weeding/grazing Non-cash
Aesthetic value Non-cash
Religious value Non-cash
Pet value Non-cash
Entertainment and sports Non-cash
Research Non-cash
7.3.3 An example of partial budget analysis for small ruminants
We will use an example to illustrate the steps used to calculate and interpret the results of partial budget
analysis. The example will use data from experiments conducted in the Ethiopian highlands (Ehui and Rey
1992). The goal of the experiment was to solve nutritional and health problems faced by sheep in the
Ethiopian highlands. Four treatments were used with 50 ewes:
• Treatment I (control): animals were under normal grazing, no drenching or feed supplementation
• Treatment II: animals were drenched with anthelmintics (Ranide and Panacur) three times per year
• Treatment III: animals were supplemented with 300 g wheat bran/ewe per day and 150 g noug
cake/ewe per day in addition to normal grazing.
• Treatment IV: animals were both drenched and given feed supplementation
The results indicated that:
• nutritional supplementation leads to reduced mortality, heavier animals and higher lambing rates
• control of endoparasites leads to reduced mortality but had no significant effect on lambing rates or
weight.
Production parameters for each treatment are listed in Table 7.4. The variable costs were found to
be feeds, anthelmintics, labour for administering drenching, breeding stock depreciation charges and
capital costs. The cost items and prices were as follows:
Cost item Price (EB)
Drenching 3.00/50 ewes
Radine 0.41/treatment
Panacur 1.23/treatment
Wheat bran 6.38/100 kg
Noug cake 16.33/100 kg.
1. US$1=EB2.07
105
Table 7.4. Small ruminant production parameters under four treatments in the Ethiopian highlands,
1989.
Production parameters Treatment I Treatment II Treatment III Treatment IV
Annual reproduction rate (%) 1.02 1.02 1.35 1.35
Survival (%)
0–4 months 0.75 0.85 0.85 0.85
4–8 months 0.85 0.90 0.90 0.90
8–12 months 0.95 0.95 0.95 0.95
Effective lambing rate (%) 0.62 0.74 0.98 0.98
Live weight at 12 months (kg) 17.48 17.88 18.78 18.78
Liveweight productivity per ewe (kg) 10.80 13.25 18.43 18.43
Number of ewes/ram 16.0 16.0 16.0 16.0
Average weight/head 24.77 24.77 28.67 30.37
Breeding stock mortality rate (%) 0.18 0.16 0.16 0.16
Treatment I = normal grazing; Treatment II = drenching with anthelmintics; Treatment III = feed supplementation; Treatment IV
= both drenching and feed supplementation.
Using the data in Table 7.4, the cost and prices listed above, the gross and net return and marginal
rate of return (MRR) were calculated for each treatment. The summary of results are shown in Tables 7.5
and 7.6. As an example, the marginal rate of return for Treatment II over Treatment I was computed as
follows:
?VC = VC (TII) ? VC (TI) = 13.81 ? 12.35= 1.46
??? = ?? (???) ? ?? (??) = 83.92 ? 79.06 = 4.86
??? = ?? ÷ VC = 4.86 ÷ 1.46 = 3.34 = 334%
Table 7.5. Gross return (EB/ewe) of feeding and health management experiments for sheep, Debre
Birhan, Ethiopia, 1989.
Treatment I Treatment II Treatment III Treatment IV
Mean price/kg LW (EB) 2.57 2.57 2.57 2.57
Mean liveweight productivity
at 12 months (kg/ewe)
10.80 13.25 18.42 18.42
Average weight/head (kg) 24.77 24.77 28.67 30.37
Gross return (EB)/ewe
1
91.4 97.72 121.03 125.40
Treatment I = normal grazing; Treatment II = drenching with anthelmintics; Treatment III = feed supplementation; Treatment
IV = both drenching and feed supplementation; EB = Ethiopian Birr.
1. [(Mean liveweight productivity at 12 months (kg/ewe) × 2.57)+(Average weight/head (kg) × 2.57)].
106
Table 7.6. Results of feeding and health management experiments for sheep, Debre Birhan, Ethiopia,
1989.
Treatment I Treatment II Treatment III Treatment IV
Gross returns (EB) 91.41 97.72 121.03 125.40
Cash costs
Feed
Noug cake – – 8.94 8.94
Wheat bran – – 6.98 6.98
Veterinary – 0.41 – 0.41
Ranide – 1.23 – 1.23
Panacur – 0.54 – 0.54
Labour
Non-cash costs
Capital cost (at 30% per annum) – 0.65 4.78 5.43
Breeding stock depreciation charges 12.35 10.98 12.7 13.46
Total costs that vary (VC) 12.35 13.81 33.41 36.99
Net return (NR) 79.06 83.92 87.62 88.41
Net return over control – 4.86 8.56 9.35
Marginal rate of return (MRR) (%) – 334.00 19.00 22.00
Treatment I = normal grazing; Treatment II = drenching with anthelmintics; Treatment III = feed supplementation; Treatment
IV = both drenching and feed supplementation; EB = Ethiopian Birr.
Figure 7.7 illustrates the net benefit curve for the data. The figure also shows the values for MRR for
the added benefits of TII over I, TIII over II, and for TIV over III. The following are interpretations of the
partial budget analysis results:
• Treatment IV yields the highest return over the control followed by Treatment III.
• If cash is not a constraint, feed supplementation would be recommended to farmers.
• The increase in benefit is accompanied by an increase in cost; farmers would not adopt this
technology if it were recommended to them.
• The added benefits from Treatment II are EB 4.86 and a 334% marginal rate of return.
• The added benefits from Treatment III are EB 3.71 and a 19% marginal rate of return.
• The added benefits from Treatment IV are EB 0.78 and a 22% marginal rate of return.
We can conclude that:
• treatments with supplemental feeding (Treatment III) and supplemental feeding with drenching
(Treatment IV) cannot be recommended to the farmers
• drenching (Treatment II) can be suggested for on-farm testing as it yielded a very high marginal rate
of return (334%) on station.
107
Figure 7.7. Net benefit curve, feeding and health management trials, Ethiopian highlands, 1989.
7.4 Exercises
1. The following table shows data for variable amounts of feed used by a farmer and the
corresponding milk output.
Feed Milk Marginal product
X
?X
Y
?Y
()*
(kg) (kg) (kg) (kg)
2 ( )* 10 ( )* ( )*
4 ( )* 30 ( )* ( )*
6 ( )* 60 ( )* ( )*
8 ( )* 100 ( )* ( )*
10 ( )* 150 ( )* ( )*
( )* : Fill in the parenthesis with the appropriate figures.
Make the necessary computations and use your knowledge on production relationships to fill in the
( )* with the missing words, letters and numbers.
2. Complete the partial budget for on-farm fattening sheep experiment at Hidi/Debre Zeit,
Ethiopia, in the wet season during 1991.
Treatments
• Treatment 1: Grazing (plus 400 g teff straw per head/day)
• Treatment 2: Grazing plus supplementation (100 g Sesbania, 50 g noug cake, 8 g salt, 5 g bone meal
and 50 g wheat bran per head/day)
• Treatment 3: Pen-feeding (700 g teff straw, 100 g Sesbania leaves, 50 g noug cake, 8 g salt, 5 g bone
meal and 50 g wheat bran per head/day)
Data (Ethiopian Birr/sheep) for variable costs, revenue from sales are given below:
108
B28B31B30B29
B28B31B31B29
B28B31B31B31B29
B28B31B29
B32B32B25
B31B39B25
B33B33B34B25
B39B30
B38B35
B38B30
B31B30 B31B35 B32B30 B32B35 B33B30 B33B35 B34B30
B54B6fB74B61B6c B76B61B72B69B61B62B6cB65 B63B6fB73B74B73 B28B45B42B2fB65B77B65B29
B4eB65B74 B62B65B6eB65B66B69B74 B28B45B42B2fB65B77B65B29
Treatment group*
12 3
NVT VT NVT VT NVT VT
Revenue from sales 73.50 97.34 154.11 131.26 120.48 114.48
Total variable costs 44.36 71.65 101.11 90.50 90.00 87.00
Sheep purchase 28.50 29.10 29.40 28.65 29.65 28.85
Feed
Teff straw 5.76 5.76 – – 10.08 10.08
Sesbania – – 6.19 6.19 6.19 6.19
Noug cake – – 1.54 1.54 1.54 1.54
Wheat bran – – 0.76 0.76 0.76 0.76
Salt – – 0.76 0.76 0.76 0.76
Bone meal – – 0.24 0.24 0.24 0.24
Labour
Per diem 1.0 1.0 1.0 1.0 1.0 1.0
Labour 8.0 8.0 8.0 8.0 8.0 8.0
Veterinary drugs (Ranide) – 1.10 – 1.10 – 1.10
Transportation (sheep) 1.10 1.10 1.10 1.10 1.10 1.10
Net return
Net return over Control
Marginal rate of return (%)
*NVT = no veterinary intervention; VT = veterinary intervention.
(A)
Calculate: (i) net return (ii) net return over control and (iii) marginal rate of return (%).
(B) Draw conclusions in relation to (i) what the most profitable treatment is (ii) your
recommendation to the farmers concerning feeds (iii) your recommendation to the farmers
concerning veterinary services.
7.5 References and reading materials
Amir P. and Knipscheer H.C. 1987. A conceptual framework for the economic analysis of on-farm
trials with small ruminants. In: Devendra C. (ed), Small Ruminant Production Systems in South
and Southeast Asia. Proceedings of a Workshop held in Bogor, Indonesia, 6–10 October 1986.
IDRC (International Development Research Centre), Ottawa, Canada. pp. 308–391.
Amir P. and Knipscheer H.C. 1989. Conducting On-farm Animal Research: Procedures and
Economic Analysis. Winrock International, USA, and IDRC (International Development
Research Centre), Ottawa, Canada. 244 pp.
109
Ehui S. and Rey B. 1992. Partial budget analysis for on-station and on-farm small ruminant
production systems research: Method and data requirements. In: Rey B., Lebbie S.H.B. and
Reynolds L. (eds), Small Ruminant Research and Development in Africa. Proceedings of the
First Biennial Conference of the African Small Ruminant Research Network, ILRAD, Nairobi,
Kenya 10–14 December 1990. ILCA (International Livestock Centre for Africa), Addis Ababa,
Ethiopia. pp. 91–104.
ILCA (International Livestock Centre for Africa). 1990. Livestock Systems Research Manual.
Working Paper 1. Volume 2. ILCA, Addis Ababa, Ethiopia. 125 pp.
110
111
Module 8: Biometrics techniques
8.1 Performance objectives
Module 8 is intended to enable you to:
1. Define the basic concepts of biometry.
2. Define descriptive statistical parameters.
3. Compute variance, standard deviation, coefficient of variation, standard error, confidence interval
of normal, binomial and Poisson variables.
4. Write statistical models to fit a set of data.
5. Compute analysis of variance.
6. Use the t-test.
7. Compute regression analysis.
8. Compute analysis of covariance.
9. Use the Chi square (?
2
) test.
10. Plan experiments.
11. Use randomisation techniques.
12. Know how to design sample surveys.
13. Compute sample sizes for experiments or sureys.
14. Present results.
8.2 Introduction
Animal research comprises both experiments and surveys conducted either on a station, on a farm or in the
field. Such research has to be properly designed and conducted, and the results must be statistically
analysed and interpreted. Typically, comparable groups of animals are subjected to specified treatments
for some predetermined period and the results of these imposed treatments, as measured by one or more
variables, are used to estimate different types of responses.
8.1. Performance objectives
8.2 Introduction
8.3 Background to basic biometry
8.4 Planning experiments
8.5 Presenting results
8.6 Exercises
8.7 References and reading materials
112
The scope of this course is to introduce some of the basic principles of experimental design and
statistical analysis. The purpose of good experimental design is to ensure that the results obtained are free
of bias and that interpretations can be made which are uncomplicated by the existence of uncontrolled
factors. Statistical analysis, from calculation of treatment means and standard errors to the more elaborate
techniques of analysis of variance, is a useful and necessary tool for researchers. However, only
experiments of sound design permit sound statistical analysis and interpretation.
The data set that will be used throughout much of this module is from an experiment carried out to
study the effect of supplementation of weaned lambs on their health and growth rate (Table 8.1). Sixteen
Table 8.1. Data set for the effect of supplementation on health and growth rate of weaned Red Maasai
and Dorper lambs.
Record
ID
No. Breed Sex Supplement Block
Weaning
weight
Weigtht at
6 mo. (kg)
PCV
(%)
FEC
(epg)
Weight
gain (kg)
1 349 1 2 1 1 8.0 8.4 10 6500 0.4
2 326 1 2 1 1 9.0 9.6 11 2650 0.6
3 393 1 1 1 2 12.0 12.1 22 750 0.1
4 71 1 1 1 2 12.3 14.1 15 5200 1.8
5 271 1 1 1 3 13.0 13.2 19 4800 0.2
6 382 1 2 1 3 15.5 16.3 24 2450 0.8
7 85 1 2 1 4 16.3 17.7 27 200 1.4
8 176 1 2 2 4 15.9 17.2 21 3000 1.3
9 286 1 2 2 1 11.0 13.1 21 1600 2.1
10 183 1 1 2 1 9.9 11.2 21 450 1.3
11 21 1 2 2 2 11.6 12.6 25 2900 1.0
12 122 1 1 2 2 12.5 14.3 25 300 1.8
13 374 1 1 2 3 14.6 17.4 19 2250 2.8
14 32 1 2 2 3 14.2 16.4 22 2800 2.2
15 282 1 2 2 4 16.3 19.7 20 750 3.4
16 94 1 1 2 4 16.7 17.2 13 5600 0.5
17 127 2 2 1 1 8.5 9.1 26 1350 0.6
18 216 2 2 1 1 9.5 10.3 19 1150 1.1
19 133 2 1 1 2 11.1 12.7 30 200 1.6
20 249 2 1 1 2 9.8 11.4 28 0 1.6
21 123 2 2 1 3 12.6 13.6 23 600 1.0
22 222 2 2 1 3 12.3 14.5 24 1500 2.2
23 290 2 2 1 4 13.3 15.3 22 1950 2.0
24 148 2 1 1 4 14.1 15.9 26 500 1.8
25 142 2 2 2 1 9.2 12.5 25 850 3.3
26 154 2 2 2 1 9.5 13.2 35 700 3.7
27 166 2 1 2 2 10.7 13.8 29 400 3.1
28 322 2 1 2 2 9.6 13.0 26 800 3.4
29 156 2 1 2 3 11.2 14.0 28 1550 2.8
30 161 2 2 2 3 12.2 15.6 22 550 3.4
31 321 2 1 2 4 13.1 16.9 25 1250 3.8
32 324 2 1 2 4 14.8 19.1 24 1100 4.3
113
Red Maasai and 16 Dorper lambs were treated with an anthelmintic at weaning (approximately 3 months
of age) and assigned at random to supplemented and non-supplemented groups (eight lambs in each
group). All lambs grazed on pasture for a further 3 months. At night they were housed and lambs in the
supplemented group were fed cottonseed cake and bran meal. Data recorded included body weight at weaning
and at six months of age, packed red cell volume (PCV) and faecal egg count (FEC) at 6 months of age.
Some of the questions we will attempt to answer will be:
• Did supplementation improve weight gain?
• Did supplementation affect PCV and FEC?
• Were there any differences in weight gain, PCV or FEC between breeds?
• Is the experiment well designed?
• Is the size of experiment (number of animals per group) sufficient to detect differences between
breeds and diets?
8.3 Background to basic biometry
8.3.1 Methods of describing data
Research data consist of recorded observations or measurements made on experimental units, e.g.
animals. We shall discuss later how to decide what should be the basic experimental unit in a statistical
analysis. Measured characteristics such as body weight, growth rate, PCVand FEC are called variables.
One of the first steps in analysing research data should be to examine the frequencies of different
values recorded for the variables measured. This will help organise the data in some way and summarise
their main features and may also help to spot data errors or extreme values.
Frequency table
A useful starting point is to examine the range spanned by the data. To do this, one finds the lowest and
highest value of a variable (e.g. PCV) and divides the range into a reasonable number of intervals. The
numbers of data values that occur within each interval are counted and made into a table (Table 8.2). After
calculating the number of animals in each class interval, the relative frequency, which is the percentage of
values contained in each interval, can then be calculated. This shows how the data are distributed. Table
8.2 shows that a majority of animals have PCVs between 19 and 30. Four animals have PCVs below and
one animal has a PCV above this range. The frequency distribution for FEC is very different with the
majority of values at the lower end of the range (Table 8.3).
Table 8.2. Number of animals grouped by packed cell volume (PCV) category.
PCV class interval
(%)
Frequency of animals
Relative frequency
(%)
10–12 2 6
13–15 2 6
16–18 0 0
19–21 7 22
22–24 8 25
25–27 8 25
28–30 4 12
31–33 0 0
34–36 1 3
Total 32 100
114
Table 8.3. Number of animals grouped by faecal egg count (FEC) category.
FEC class interval
(epg) Frequency of animals
Relative frequency
(%)
0–700 12 38
800–1500 8 25
1600–2300 3 9
2400–3100 5 16
3200–3900 0 0
4000–4700 0 0
4800–5500 2 6
5600–6300 1 3
6400–7100 1 3
Total 32 100
Histogram
Tables can be presented pictorially by means of a diagram, known as a histogram. A histogram helps
identify the shape of the distribution. Sometimes the histogram is more or less symmetrical, with the bulk
of the data gathered near the centre and the proportions of data on each side roughly balancing each other.
Sometimes the histogram is skewed. This means that the data are rather bunched up to one side. Figure 8.1
gives an example of the skewed distribution of FEC. There are more data values to the left than the right.
Biological data often belong to a normal distribution; the frequency distribution is ‘bell-shaped’; the
distribution of PCV approximates to this shape.
B30B96B37B30B30 B38B30B30B96B31B35B30B30 B31B36B30B30B96B32B33B30B30 B32B34B30B30B96B33B31B30B30
B33B32B30B30B96B33B39B30B30 B34B30B30B30B96B34B37B30B30 B34B38B30B30B96B35B35B30B30 B35B36B30B30B96B36B33B30B30 B36B34B30B30B96B37B31B30B30
B43B6cB61B73B73 B69B6eB74B65B72B76B61B6c B28B46B45B43B29
B30
B32
B34
B36
B38
B31B30
B31B32
B31B34
B46B72B65B71B75B65B6eB63B79 B6fB66 B61B6eB69B6dB61B6cB73
Figure 8.1. Histogram of frequency distribution of faecal egg counts (FEC).
Scatter plots
Graphs derived by plotting one variable against another are commonly used to examine relationships
between two variables. Figure 8.2 shows a scatter plot of PCV and FEC. Such a scatter plot can also
indicate outliers (i.e. data values which are rather extreme and appear different from the others). Such
outliers may be due to data errors. Figure 8.2 shows one possible outlier. Let us assume that during data
entry lamb 94 has had its PCV digits transposed by mistake from 13 (see Table 8.1) to 31. With such a high
FEC its PCV value is somewhat away from the others towards the top right corner of the diagram.
Examination of the input data shows that a mistake has been made in data entry.
Figure 8.2. Scatter plot of packed cell volume (PCV) versus faecal egg count (FEC) with mistake in
PCV value (31 rather than 13%) for lamb 94.
Of course, when all the data are described together like this, variations due to diet and breed are
hidden. If possible, it is best to summarise the data in each group separately. Quite a nice way is to use what
is known as Tukey’s box and whisker plot (Figure 8.3), which displays the range, median and quartiles
(with the middle half of the data contained within the box and the remainder described by the vertical
lines) for each group alongside each other.
Figure 8.3. Box plot of faecal egg count (FEC) for each breed–supplementation group.
115
B30 B31B30B30B30 B32B30B30B30 B33B30B30B30 B34B30B30B30 B35B30B30B30 B36B30B30B30 B37B30B30B30
B46B45B43 B28B65B70B67B29
B35
B31B30
B31B35
B32B30
B32B35
B33B30
B33B35
B34B30
B50B43B56 B28B25B29
B30
B31B30B30B30
B32B30B30B30
B33B30B30B30
B34B30B30B30
B35B30B30B30
B36B30B30B30
B37B30B30B30
B31B30 B31B31 B32B30B32B31
B42B72B65B65B64 B73B75B70B70B6cB65B6dB65B6eB74B61B74B69B6fB6e
B46B45B43 B28B65B70B67B29
8.3.2 Descriptive statistics
Mean
Tables and graphs, such as those already described, are often useful and revealing. They are always the
first step in attempting to summarise a set of data and to look at its structure. One of the first items of
interest is a measure of a central or representative overall value.
The most commonly used measure is the arithmetic mean (usually referred to simply as the
mean). The mean is obtained by dividing the sum of all the observations by the number of observations.
The calculation of the arithmetic mean may be demonstrated for the Dorper lambs (breed 1) that were non-
supplemented (diet 1) by using their weights at 6 months.
mean = (8.4 + 9.6. + ...+ 17.2)/8 = 108.6/8 = 13.575 kg
The mean is a kind of central value. However, the mean can be somewhat misleading if the
distribution is skewed, i.e. if there is a high proportion of large values or a high proportion of small values.
The problem is that a few extreme values unduly influence the value of the mean. In these situations the
median may be a very useful alternative. The median is the value lying in the middle, such that one half of
the values are above and one half below. Putting the 6-month body weights in order, we get 8.4, 9.6, 12.1,
13.2, 14.1, 16.3, 17.2, 17.7; the median value is therefore 13.65 kg. If there is an odd number of values, the
median is the middle value. In both cases, the numbers must be arranged in ascending order (from smallest
to largest) before calculating the median.
These two values (for mean and median) are in this case very similar. For FEC, however, we might
expect the mean and median to be more different. Thus, if we put FECs for the same animals in order, we
get 200, 750, 2450, 2650, 3000, 4800, 5200, 6500. Thus, the median is 2775 while the mean is 3194, which
is somewhat higher. For a variable such as FEC, for which the frequency is skewed, the mean will often be
greater than the median.
Variance and standard deviation
Biological processes yield variable data. The most common measures of variability are the variance and
standard deviation. The variance of a group of observations is calculated as the sum of squares of
deviations from the mean divided by one fewer than the number of observations. Thus the variance of the
6-month body weights for breed 1, diet 1 is calculated as:
[(8.4 ? 13.575)
2
+ (9.6 ? 13.575)
2
+....+ (17.2 ? 13.575)
2
]/7
=[(?5.175)
2
+ (? 3.975)
2
+....+(3.635)
2
]/7 = 82.755/7 = 11.8221 kg
2
An alternative formula, which is easier to use with a large number of observations, is to
calculate the sum of each value squared, subtract the total squared divided by the number of
observations, and divide the answer by the number of observations less one. Using this method the
variance becomes:
[8.4
2
+ 9.6
2
+ .... + 17.2
2
? (108.62/8)]/7
= (1557.0 - 1474.245)/7 = 82.755/7 = 11.8221 kg
2
and the answer is the same.
While it is often convenient to use the variance based on squared deviations as a measure of
dispersion or variation of the data, it is often usual also to think of this variation in terms of the original
units. By taking the square root of the variance one returns to the original scale of measurement. The
square root of the variance is known as the standard deviation (SD).
Thus the standard deviation = ?11.8221 = 3.438 kg.
Coefficient of variation
We have shown that the variation among a set of observations can be measured using the variance or the
standard deviation. There are occasions when we wish to compare the relative variation between two
116
variables having different means, e.g. birth weights of calves and adult weights of cows. Since the adult
weights would be expected to have a larger standard deviation because of the larger values involved, a
comparison of the two standard deviations may not be very helpful. We might also be interested in
comparing the relative variation between different characteristics or traits of the same animals, e.g.
weaning weights and PCV of calves. For such comparisons the coefficient of variation (CV) is calculated
as the standard deviation expressed as a percentage of the mean. Thus, for 6-month body weight for breed
1, diet 1, the coefficient of variation:
CV = (SD ÷ mean) × 100 = (3.438/13.575) × 100 = 25.3%
The CV is independent of the unit of measurement. For body weight or PCV, for example,
coefficients of variation of the order of 10% might normally be expected. Growth rates, however, often
have higher coefficients of variation. This example shows that the spread of 6-month weights in this group
of lambs is wider than one would normally expect.
The descriptive statistics for the 32 animals in the example data set are summarised in Table 8.4.
Note the high CVs for all variables, particularly for weight gain and FEC. The comparatively large CVs
for body weight and PCV may have been due to the exposure of these lambs to high levels of helminth
infestation.
Table 8.4. Descriptive statistics of variables measured in the example data set.
Variable Mean Median
Standard
deviation
Coefficient of
variation
Weaning weight (kg) 12.19 12.25 2.49 20.4
6-month weight (kg) 14.11 13.90 2.82 20.0
Weight gain (kg) 1.92 1.80 1.18 61.4
Packed cell volume (%) 22.72 23.50 5.35 23.6
Faecal egg count (epg) 1770 1200 1686 95.3
Number of animals = 32.
8.3.3 Precision of estimated parameters
Standard error
Suppose the experiment were repeated several times. Each time we would calculate a mean, say for breed
1, diet 1, each mean would be different. Thus there would be variability among means just as there is
variability among individual observations in each sample. However, since each mean is averaged over
eight observations, the variation among means would be smaller than the variation among individual
observations. The variance of a mean can be calculated as the average of the variances calculated each
time the experiment is conducted divided by the number of observations used for calculating each mean.
The standard deviation of the mean is known as the standard error.
The standard deviation is a useful measure of the variation of an individual observation. The
standard error is a useful measure of the variation of a mean. Normally we would not repeat an experiment
to calculate the average variation of the mean. Instead, we pretend that the variance of individual
observations will remain constant from experiment to experiment (in practice it will not vary much) and
use the value already calculated as an estimate of this average. Thus, we can then calculate, for the 6-
month body weights of sheep in breed 1 fed diet 1:
variance of mean = 11.8221/8 = 1.4778 kg
2
standard error = 1.4778 = 1.22 kg
117
118
The standard error or variance of a mean decreases as the number of observations increases.
Thus, for example, if 24 sheep were used for each breed, then, for breed 1 diet 1:
variance of mean = 11.8212/12 = 0.9926 kg
2
standard error = 1.00 kg
Note that this standard error is lower than the value calculated above with 8 sheep.
Confidence interval
The standard error can be used to develop what is known as a confidence interval. A confidence interval
is a range between upper and lower limits, which is expected to include at a given level of probability the
true (or population) mean value. This is the value for which the sample in the experiment provides an
unbiased estimate.
Usually we talk about the 95% confidence interval. This is the interval in which the true mean
should lie with a 95%, or 19 times in 20, chance of being correct. Similarly, the 99% confidence interval
gives the range within which we expect the true mean to lie with a 99%, or 99 times in 100, chance of being
correct. The approximate 95% confidence interval can be calculated as the sample mean plus or minus
twice the standard error. Similarly, the 99% confidence interval is the sample mean plus or minus
approximately 2.6 times the standard error. Thus the approximate 95% confidence interval for the mean 6-
month body weight of lambs of breed 1 fed diet 1 is:
13.58 ± 2 × 1.22 = (11.14 to 16.02) kg
Similarly, the approximate 99% confidence interval is:
13.58 ± 2.6 × 1.22 = (10.41 to 16.75) kg
As one might expect, the 99% confidence interval is wider than the 95% one. Another way of
thinking about this is to say that if the experiment were repeated 100 times then we would expect that on 99
occasions the sample mean would fall within the range 10.41 to 16.75 kg, and on the other occasion it
would fall outside.
It should be emphasised that the above formula is only approximate; instead of 2 or 2.6 we should
strictly use a t-value with the appropriate degrees of freedom; t-values will be discussed later.
Binomial and Poisson variable
Sometimes response variables take on the values 0 or 1. These could be, for example, mortality or disease.
The variables are sometimes described as discrete variables, i.e. they can only take on discrete values.
Variables such as body weight are known as continuous variables, i.e. they can take on any value within a
reasonable range to a given degree of accuracy. Discrete variables such as mortality and disease often
belong to distributions that are not normal, but binomial or Poisson distributions. Discrete variables
taking on values 0 or 1 are also sometimes called binary variables. For example, suppose that 4 lambs in a
sample of 20 die before weaning. We can record a death as 1, a lamb that survived as 0. The proportion of
lambs that died is p = 4/20 = 0.2. This is simply the mean of the 0s and 1s, i.e. (1+1+1+1+0....+0)/20 = 4/20
= 0.2. Mortality is typical of a variable that is associated with a probability of occurrence, p. Such variables
typically belong to a binomial distribution.
We have seen that the mean = p, and that this can be calculated in an analogous way to that of a
normally distributed variable. The binomial distribution, however, has a special expression for the
variance, namely p(1 ? p). Thus if p = 0.2, then variance = 0.2 × 0.8 = 0.16.
The standard deviation and standard error are similarly calculated, i.e.
standard deviation = ?0.16 = 0.4
standard error = ?(0.16/20) = 0.089
119
The distribution of certain binary data is often closer to that of a Poisson distribution, a distribution
associated with rare events. Incidence of trypansomosis, for example, can often be considered as being
associated with a Poisson distribution, particularly if the prevalence is very low (e.g.<10%). The mean is
the same as that for the binomial distribution, namely p. The variance, however, is different from that of
the binomial distribution and is also p. Thus, if the prevalence of trypanosome infections is 0.06 in a
sample of 100 cattle, (i.e. 6/100 are detected parasitaemic), then the variance is also 0.06 and
standard deviation = ?0.16 = 0.24
standard error = (?0.06/100) = 0.024
In practice when p is small (<0.1), it does not matter whether one assumes a binomial or Poisson
distribution.
8.3.4 Analysis of variance
Statistical models
So far these course notes have been prepared without using mathematical, algebraic equations. This has
been deliberate. However, statistical concepts are often presented mathematically and it is important for
biologists not to be frightened by the formulae that sometimes appear in scientific papers. Formulae are
often the easiest way to describe the statistical methods that have been applied.
If we look at the structure of our example data set we can see that each variable, weaning weight, 6-
month body weight, PCV, FEC and weight gain, can be characterised by other attributes. For example,
records 1–8 all refer to lambs from breed 1. They are also characterised by the fact that these lambs were
not supplemented. Thus, we can say that each record is made up of a component due to breed and a
component due to diet. We shall use a letter to refer to each of these components, namely b (breed), d
(diet). These components are often referred to as factors or parameters. It is common to give a subscript to
each parameter level. For example, we have 2 breeds. Let us refer to them as b
1
,b
2
. Similarly, we have 2
diets, d
1
,d
2
. Thus, for example, records 1–8 are each made up of b
1
+d
1
.
Similarly, records 9–16 are made up of b1 + d2 and records 25–32 of b
2
+d
2
. It is usual to use the
letter y to refer to the variable to be analysed and to use a combination of subscripts to refer to the particular
data value. Thus:
y
111
=b
1
+d
1
for record 1
y
124
=b
1
+d
2
for record 12
y
225
=b
2
+d
2
for record 29
Note that the first subscript refers to breed, the second to diet and third to lamb number (1–8) within each
breed–diet category.
Normally b and d are defined as deviations from the overall mean for which the symbol µ is
normally used. Also, since there are fewer factor levels (2 + 2) than there are lambs (32), each observation
will deviate from the sum of the 2 parameter levels by an amount usually described by the letter e or ?,
known as the error or residual term.
Thus, the above equations are written in full as:
y
111
= µ +b
1
+d
1
+e
111
y
124
= µ +b
1
+d
2
+e
124
y
225
= µ +b
2
+d
2
+e
225
120
Thus, each observation is made up of an overall mean, a term for breed, a term for diet and a
residual. These equations can be written more generally as:
y
ijk
= µ +b
i
+d
j
+e
ijk
(i=1,2;j=1,2;k=1....8)
The purpose of analysis of variance is to estimate the overall significance of differences among
each set of parameter means. It separates variations due to breed and diet from each other, and compares
the magnitudes of these different sources of variation with the variation which is left over among residual
terms. In using analysis of variance, two main assumptions are used: firstly, that the eijks have the same
distribution for each parameter or factor (i.e. are not more variable for one than another) and, secondly,
they are distributed normally. Thus, it is not strictly valid to use analysis of variance for discrete variables
that are associated with binomial or Poisson distributions.
In developing the idea of analysis of variance it will be simplest to start with a model based on just
the four groups of lambs categorised by breed and diet in the data set, but ignoring the breed–diet structure.
The model we shall first use, with g signifying group, is:
y
ij
= µ +g
i
+e
ij
(i = 1....4; j = 1....8)
It is important to emphasise the need for writing down the statistical model to be fitted to a set of
data before embarking on a statistical analysis. It helps the researcher decide the form the analysis of
variance should take.
One-way analysis of variance
The model:
y
ij
= µ +g
1
+e
ij
is an example of a one-way analysis of variance. Such an experiment is often referred to as a completely
randomised design.
The degrees of freedom (df) for the variation among groups is calculated as one fewer than the
number of groups (i.e. 4?1). The total degrees of freedom = 32 (number of lambs)?1. The residual degrees
of freedom is calculated by subtracting the among groups df from the total df.
The following calculations calculate the sums of squares (SS) and mean squares (MS).
Total. This is the same calculation as used previously for the variance, i.e. the sum of squares of individual
observations minus the square of the total divided by the number of observations. Mathematically we can
write this as:
?y
ij
2
? (?y
ij
)
2
/32
Thus, if we apply this formula to the weight gains of the 32 lambs, we obtain:
total SS = 160.78 ? 61.4
2
/32 = 160.78 ? 117.8112 = 42.9688
Groups. The formula for the among group sums of squares uses the totals for each group divided by the
number of lambs in each group. It is:
?Y
i
.
2
/8 ? (?Y
ij
)
2
/32
The capital letter Y, together with the dot, means the sum of the individual yijs summed over the subscript
that the dot is replacing, in this case, j. Thus:
Y
i
.=y
11
+y
12
+....+y
18
The among group SS for weight gain is therefore:
(6.6
2
+ 15.1
2
+ 11.9
2
+ 27.8
2
)/8 ? 61.4
2
/32 = 148.2525 ? 117.8112 = 30.4413
121
The among group mean square equals the among group sum of squares divided by the among group
degrees of freedom. Thus:
Among group MS = 30.4413/3 = 10.1471
Residual. We calculate the residual sum of squares by subtracting the among group sum of squares from
the total sum of squares. Thus:
Residual SS = 42.9688 ? 30.4413 = 12.5275
Residual MS = residual SS/residual df = 12.5275/28 = 0.4474
Putting the results of these calculations into the analysis of variance we get:
Source of variation df SS MS F
Among groups 3 30.4413 10.1471 22.7
Residual 28 12.5275 0.4474
Total 31 42.9688
The residual MS estimates the average variation among individuals within groups and is an
estimate of the average within group variance of weight gains. The among group MS represents the
additional variation brought about by the differences in mean weight gains among the four groups. The
magnitude of this additional variation can be assessed by dividing the among group MS by the residual MS
to give what is known as the F-value. Thus:
F = 10.1471/0.4474 = 22.7
We use this value to compare with values in F-tables with 3 and 28 degrees of freedom. If we look at
the column headed 3 df at the top of each of the F-tables and then come down this column until we reach 28
df on the left hand side, we find the values 2.95, 4.57, and 7.19 for 5%, 1% and 0.1% levels of probability,
respectively. If our observed F-value exceeds any of these values, then we say it is statistically significant.
The value of 22.7 exceeds the value of 7.19. Thus, the analysis of variance shows that there is a highly
significant variation in liveweight gain among the four groups of lambs at the 0.1% level of statistical
significance. This is normally described using the notation (P<0.001). In other words, the probability of
there being no differences among the 4 groups of lambs is less than 1 in 1000.
Factorial analysis of variance
The groups of lambs in the experiment have been assigned a particular structure. In other words the 4
groups are composed of combinations of 2 diets and 2 breeds, namely:
Supplementation
Breed Yes No
112
234
This design, sometimes described as a 2×2 factorial design, is a very useful way of getting as much
information from an experiment as possible. Thus we can look at the effects of dietary supplementation on
weight gain of two breeds at the same time. What is more, we can look at the way that dietary
supplementation and breed may interact. If there is no interaction then we mean that any effect of dietary
supplementation is the same for each breed. If there is an interaction then the effect of supplementation is
different for different breeds.
We can rewrite our statistical model as follows:
y
ijk
= µ +b
i
+d
j
+ (bd)
ij
+e
ijk
where (bd)
ij
is a notation used to signify the interaction of breed and diet.
122
If the analysis shows there to be no significant interaction then we could express the model as:
y
ijk
= µ +b
i
+d
j
+e
ijk
In other words, the model simply adds to breed the same estimate for diet irrespective of breed.
Sometimes we refer to such effects as being additive.
To estimate sums of squares for breed and diet we split the sums of squares we have already
obtained for groups, namely:
?Y
ij
.
2
/8 ? (?yijk)
2
/32
into a component for breed:
?Y
i
..
2
/16 ? (?
yijk
)
2
/32
and a component for diet:
?Y
.j.
2
/16 ?(?
ijk
)
2
/32
The expression we had earlier for the among group SS was:
(6.6
2
+ 15.1
2
+ 11.9
2
+ 27.8
2
)/8 ? 61.4
2
/32 = 30.4413
The between breed SS becomes:
(21.7
2
+ 39.7
2
)/16 ? 61.4
2
/32 = 10.1250
(where 21.7 and 39.7 are total weight gains for breed 1 and breed 2, respectively).
The between diet SS becomes:
(18.5
2
+ 42.9
2
)/16 ? 61.4
2
/32 = 18.6050
(where 18.5 and 42.9 are total weight gains for diet 1 and diet 2, respectively).
This results in sums of squares of 30.4413, 10.1250 and 18.6050 for group, breed and diet, respectively.
The sum of squares for the interaction can be obtained by subtraction of the last two numbers from the first, i.e.
30.4413 ? 10.1250 ? 18.6050 = 1.7113. We can then complete the analysis of variance table as follows:
Source of variation df SS MS F Probability
Breed (B) 1 10.1250 10.1250 22.63 <0.001
Diet (D) 1 18.6050 18.6050 41.58 <0.001
Interaction (B×D)
1 1.7113 1.7113 3.82 0.06
Residual 28 12.5275 0.4474
Total 31 42.9688
This analysis demonstrates a highly significant difference between breeds in average weight gain, a
highly significant effect of supplementation with cottonseed cake and bran meal and an interaction
approaching statistical significance. But how precisely were these effects manifested? We need to look at
the mean values and calculate their standard errors to determine the precision with which they are
estimated. We can then apply earlier knowledge to calculate confidence limits for each mean.
Mean values for weight gain (kg) are as follows:
Supplementation
Breed No Yes Mean
1 0.82 1.89 1.36
2 1.49 3.48 2.48
Mean 1.16 2.68 1.92
123
Standard errors (SE) can be calculated as the square root of the residual variance divided by the
number of observations. From the analysis of variance the residual variance = 0.4474 kg
2
. Therefore, SE
of any mean in the body of Table 8.1 = ?(0.4474/8) = 0.236. Similarly, SE of overall mean of
supplementation (yes or no) = ?(0.4474/16) = 0.167; SE of overall breed mean = ?(0.4474/16) =
0.167 kg.
8.3.5 t-test
We now have a good idea from the last section that there are highly significant differences between breeds
and diets. We shall now formalise the method for analysing differences between residual means when
there are more than two breeds or diets. Instead of the F-test, which considers the variation among any
number of levels of a factor, we use the t-test which looks at the variation between just two levels. The
formula for the t-test is:
t = (difference between two means)/(SE of difference between two means)
The SE of the difference between two means (SED) is calculated as the square root of the variance
multiplied by (1/n
1
+ 1/n
2
) where n
1
and n
2
are the numbers of observations respectively making up the
two means to be compared. Thus:
SED = [variance × (1/n
1
+ 1/n
2
)]
When n
1
=n
2
the formula reduces to:
SED = ?(2 × variance of mean) or ?2 × SE.
To better illustrate the use of the t-test let us return to the one-way analysis of variance and pretend
that the four group means now represent sheep fed four different legume species. Thus to compare the
difference between weight gains for legume species 1 and 2 we get:
t = (1.89 ? 0.82)/(?2 × 0.236) = 1.07/0.334 = 3.20
We compare this value with t-values in the t-table using the number of degrees of freedom, namely
28, for the residual variance. With 28 df we get t-values of 2.05, 2.76 and 3.67 at 5%, 1% and 0.1% levels
of significance, respectively. The value of 3.20 is less than 3.67 but greater than 2.76, and so the difference
in weight gain for these two species is significant (P<0.01). However, if we compare species 2 and 3 we
get:
t = (1.89 ? 1.49)/(?2 × 0.236) = 1.20
which is not significant.
Least significant difference
It is sometimes more convenient to calculate the least significant difference (LSD) needed for the
difference between two means to be significant. This is calculated as:
LSD=t × SED
Thus, for the 4 legume species means
LSD = 2.05 × 0.334 = 0.68 at the 5% level of significance
= 2.76 × 0.334 = 0.92 at the 1% level of significance
= 3.67 × 0.334 = 1.23 at the 0.1% level of significance
We can then compare differences between the respective means with these values to determine their
significance. Thus, for example,
1.89 ? 0.82 = 1.07 : significance level P<0.01
124
3.48 ? 1.89 = 1.59 : significance level P<0.01
1.49 ? 0.82 = 0.67 : not significant
8.3.6 Blocking
One aim of good experimental design is to ensure that the residual mean square is as small as possible. One
way to do this is to ensure that animals are as uniform as possible before the experiment starts. Another is
to ensure that animals with similar attributes are assigned equally to the different factors being compared.
The most common type of design is a randomised block for which animals might be blocked or grouped,
for example, on the basis of body weight, age, lactation, pregnancy status or herd. An experiment might
also be replicated in time. Each replicate then constitutes a block.
Because of the wide range of weaning weights among lambs in this experiment it might be sensible
to group the lambs on the basis of weaning weight. This in fact is what was done (see Table 8.1). Lambs in
each breed were sorted into weaning weight order and formed into four blocks of four with the lightest four
in the first and heaviest four in the fourth block. Lambs were then randomly assigned to the two levels of
diet within each block. This assignment is shown in the data set in Table 8.1. When this is done, the model
formula becomes (with r for block):
Y
ijk
= µ +b
i
+r
ik
+d
j
+ (bd)
ij
+e
ijkm
(i=1,2;j=1,2;k=1,2,3,4;m=1,2)
Note that there are eight possible values of rik: r
11
,r
12
,r
13
,r
14
for breed 1 and r
21
,r
22
,r
23
,r
24
for breed 2.
The subscript m has two values, one for each of the two lambs assigned to each of the two diets in each
block.
The formula for calculating the among block within breed sum of squares is the sum of:
?Y
1jk1
2
/4 ? (?y
ljkm
)
2
/16 calculated for lambs of breed 1
?Y
2jk1
2
/4 ?(?y
2jkm
)
2
/16 calculated for lambs of breed 2
The new residual sum of squares is then calculated by subtracting this sum of squares from the residual
sum of squares previously calculated. The completed analysis of variance for weight gain then becomes:
Source of variation df SS MS F Probability
Breeds (B) 1 10.1250 21.68 <0.001
Blocks/breeds 6 2.2538 10.1250 0.80 NS
Diet (D) 1 18.6050 0.3756 39.84 <0.001
B × D
1 1.7113 18.6050 3.66 0.07
Residual 22 10.2737 1.71113
Total 31 42.9688
This table shows us something about the design of the experiment. If we compare the F-value of
0.80 for among blocks with the 5% value in the F-table with 6 and 22 degrees of freedom, namely 2.55, we
see it is much smaller. In other words, differences among blocks were not significant. In this experiment,
therefore, assignment of lambs to blocks was of no benefit in reducing, in terms of weight gain, the
residual variance. In other words, weight gain post-weaning was unrelated to weaning weight. This does
not mean that it was a bad thing to assign sheep to diets on the basis of their initial weight, for in other
instances we may find that blocking by initial weight is beneficial.
8.3.7 Heterogeneity of variance
The variances for weight gains among individuals can be calculated separately for each breed–diet group and
should be 0.3793, 0.9041, 0.2955, 0.2107 kg
2
, respectively. The residual variance used in the analysis
125
of variance is the average of these. Examination of these different estimates shows that the variance was
smaller for groups 1, 3 and 4 than for group 2. The analysis of variance we have undertaken assumes that
each variance calculated for each group estimates the same ‘population’ variance. This may not be true. In
this example the differences in variances are not much to worry about, and one can assume that the weight
gains of lambs in group 2 were by chance more variable than in the other groups. The conclusions are
clear. The analysis of variance technique is robust. In other words it handles data for which assumptions,
such as homogeneity of variance of symmetric, normal distributions, do not strictly apply. But what
alternative methods are at our disposal? Faecal egg count (FEC), for instance, has a very skewed, non-
normal distribution and an analysis of variance undertaken on these data may not be justified.
Transformations
Firstly we could transform FEC to a different scale by using a calculation such as log (y), log (k+y) or ?y.
Each of these calculations reduces the skewness of the data and makes the residual variance less
dependent on the mean. The logarithm transformation is stronger than the square root one. The log (k+y)
transformation, where k is a constant >0, is used instead of log (y) when individual values of y occur which
equal 0. The disadvantage of this method is that, although it may produce a more valid statistical analysis,
it is often not so easy to present the results. This is because the means and standard errors calculated in the
statistical analysis are now on a different scale.
Sometimes a transformation is not appropriate and an alternative method would be to split the data
into separate parts and to do a separate analysis of variance on each part. It is conventional, however, to
analyse FEC on the log scale. An arcsin (sin
-1
?y) or square root transformation is more appropriate for
disease prevalence data, the square root for when the prevalence is low. However, there are now new
methods using what is known as logistic regression or log linear models which are more suitable for such
data. These methods are too advanced for this module.
Results can be expressed on the original scale by calculating the geometric mean (antilog of the
transformed mean) and the antilog of the 95% confidence limits for the transformed mean. Take, for
example, the Dorper lambs (breed 1) supplemented with cottonseed cake and bran meal. The mean and
standard error for FEC calculated on the log scale (loge y) are 7.26 and 0.36, respectively. The 95%
confidence range for this mean is 7.26 ± t × 0.36 where t, with 7 degrees of freedom = 2.36. Thus, 95%
confidence interval for mean = (6.41, 8.11). If we transform these values back to the original scale using
the antilog (or e
x
) transformation, then mean = 1422, range = (608 to 3327).
Thus, we can say that the geometric mean is 1422 eggs per gram (epg) and that the 95% confidence
interval within which the true ‘population’ mean lies is between 608 and 3327 epg. Three important points
need to be highlighted:
• it is not permitted to calculate an antilog for the standard error to produce a standard error on the
original scale
• the upper and lower confidence interval values are not equidistant from the mean
• if log
e
(50 + y) is used for the transformation, then 50 must be taken off the antilogs after
transformation back to the original scale in order to calculate the correct geometric mean and range.
Another variable that is often transformed to logarithms is antibody titre. Serum is usually diluted in a
geometric series, i.e. with a constant ratio between successive dilutions, e.g. 1/2, 1/4, 1/8, 1/16 etc. Thus,
frequency distributions of titres tend to be lognormal and so the logarithmic transformation is to be desired.
8.3.8 Regression analysis and analysis of covariance
Regression analysis
Regression analysis is a form of analysis of variance. The only difference is that the statistical model is
written in a form that is based on the relationship between a y variable and one or more other continuous
variables, say x, z...., often referred to as independent variables, rather than with discrete sets of
parameter or factor levels.
Thus y
i
=a+bx
i
+e
i
fits a straight line between the y
i
and x
i
values with slope b, known as the
regression coefficient, and crossing the y-axis at the point a, known as the intercept. We shall not
consider methods of calculation here, but they result in a similar analysis of variance table to those derived
previously. For example, consider the relationship between PCV and loge (50 + FEC) and ignore the fact
that lambs were from different breeds and fed different diets. Statistical analysis of these data (using
calculations not described) lead to the following analysis of variance:
Source of variation df SS MS F Probability
Regression 1 374.5087 374.5087 21.9 <0.001
Residual 30 513.9600 17.1320
Total 31 888.4688
Regression equation is y = 46.0 ? 3.297x
Regression coefficient = ?3.297
SE of regression coefficient = 0.705
Correlation coefficient (r or p) = 0.65
We can test for the statistical significance of the regression coefficient by using the t-test:
t = b/SE = ? 3.297/0.705 = 4.68
Comparing with t-values in the t-table with 30 degrees of freedom, we see that the regression
coefficient is significant (P<0.001). We come to the same conclusion using the F-value in the analysis of
variance table. Thus, there is a highly significant association between PCV and FEC. The degree of this
association is described by the correlation coefficient. A correlation coefficient can vary between 0 (no
correlation at all) and 1 (perfect association, i.e. the points lie on a straight line). In this case r = 0.65. The
scatter diagram of PCV and loge (FEC + 50) is illustrated in Figure 8.4 with the regression line imposed.
The spread of points illustrates a reasonable association between PCV and logarithm of FEC, although the
association between PCV and FEC untransformed (Figure 8.2) may be slightly better.
Figure 8.4. Regression line of packed cell volume (PCV) on log (FEC+50).
126
B33B34B35B36B37B38B39B31B30
B35
B31B30
B31B35
B32B30
B32B35
B33B30
B33B35
B34B30
B6cB6fB67 B28B46B45B43 B2b B35B30B29
B50B56B43 B28B25B29
Analysis of covariance
Analysis of covariance is a simple extension of analysis of variance. Independent variables not
incorporated in the design of the experiment are included as covariates. Thus, in the study of the effect of
supplementation on the weight gains of the two breeds of sheep, weaning weight was taken into account in
the definition of blocks.
Sex, however, was not. We can, however, include sex in the model as a covariate with values 1 or 2
for males and females, respectively (Table 8.1). Covariates may also be continuous variables. The full
model appears as follows:
y
ijkm
= µ +b
i
+r
ik
+d
j
+ (bd)
ij
+ßx
ijkm
+e
ijkm
where ß is the regression coefficient of weight gain (yijkm) on sex (xijkm). The analysis of covariance
appears as follows:
Source of variation df SS MS F Possibility
Breeds (B) 1 10.5302 10.5302 23.31 <0.001
Blocks within breed 6 2.9192 0.4865 1.08 NS
Diet (D) 1 19.3548 19.3548 42.82 <0.001
B × D
1 1.8982 1.8982 4.20 0.05
Sex 1 0.7870 0.7870 1.74 NS
Residual 21 0.4518 0.4518
Total 31 42.9688
The analysis adjusts, or corrects, individual parameter level means for the differences in numbers of
males and females making up the means. The regression on sex, whilst not significant, has nevertheless
reduced the residual mean square. Interestingly, when adjusted for sex, the F-value for blocks is now also
greater than 1. Note that the total sum of squares is no longer the sum of the individual squares in the
analysis of variance table. This is because each sum of squares is now corrected for differences between
sexes.
When there is a choice to use an independent variable as a covariate or to define it as a blocking
factor in a designed experiment, the second option is usually preferred, provided adequate numbers of
degrees of freedom are retained for the residual term. If there had been equal numbers of females and
males for each breed in Table 8.1, then it would have been possible to block lambs on the basis of body
weight for each sex.
8.3.9 Analysis of discrete data
Sometimes response variables measured in an experiment may be discrete, not continuous. Such
variables, often referred to as binary variables, can take two values, 0 or 1. One example is mortality;
others are proportions of animals conceiving, proportions lambing, or proportions with a given disease
etc. This module briefly mentions the simplest method of handling such data.
Chi-square (?
2
) test
For simple experiments one can group the data into a two-way table and perform a ?
2
test. Thus, for
example, if the lambs used in this experiment were drawn from a group of 55 Dorper and 42 Red Maasai
lambs and 14 and 4, respectively, died before weaning, then we can compare differences in mortality by
forming a table, often known as a contingency table, as follows:
127
Alive at weaning
Breed No Yes Total
Dorper 14 41 55
Red Maasai 4 38 42
Total 18 79 97
The above values are the ones observed. The next step is to calculate the values that would have
occurred had the mortality rates in the two breeds been the same. These are known as expected values. To
do this calculation, expected values for each of the four cells in the body of table are obtained by
multiplying together the two totals in the corresponding row and column and dividing by the grand total.
Thus, expected values are:
55 × 18/97 55 × 79/97
42 × 18/97 42 × 79/97
10.2 44.8
7.8 34.2
The formula for the ?
2
test is:
?(o
i
? e
i
)2/e
i
where o
i
= observed value and e
i
= expected value (i = 1, 2, 3, 4).
Thus ?
2
= (14 ? 10.2)
2
/10.2 (41 ? 44.8)
2
/44.8 + (4 ? 7.8)
2
/7.8 + (38 ? 34.2)
2
/34.2
= 1.416 + 0.322 + 1.851 + 0.422
= 4.01
The number of degrees of freedom = (2 ? 1) × (2 ? 1)=1
The 5% and 1% values for ?
2
with 1 degree of freedom are 3.84 and 6.63, respectively. Thus we can
say that the mortality rate before weaning in the Dorpers of 14/55 (25%) was significantly higher than that
of 4/42 (10%) in the Red Maasai (P<0.05).
When any value of ei is less than 5, the above formula should be replaced by:
?(| o
i
? e
i
| ? 0.5)
2
/e
i
where |o
i
? e
i
| is the positive value of the difference between o
i
and e
i
. This is because the ?
2
test is only an
approximation to the distribution of numbers in a contingency table when the numbers are small.
8.4 Planning experiments
So far we have considered in some detail the statistical analysis of experimental results. We have
encountered the concept of blocking to help improve the precision of an experiment. Now we shall
discuss in more detail the general aspects of good experimental design. Careful planning of any
experiment is obviously important, but it is even more essential for certain types of livestock trials
and for on-farm experiments. Experiments involving calving, lambing or kidding may be particularly
awkward since it may be difficult to control birth dates. This means that different animals may enter
the experiment at different times. In addition, livestock trials are often fairly long term, when
compared to crop trials, and need intensive data recording. Animals are also valuable. This makes
detailed planning essential.
128
8.4.1 Stages in planning experiments
Objectives
All experiments should have clear, well-defined and realistic objectives. This may seem obvious, but is
frequently overlooked. An example of a vague objective is: ‘To examine the effect of nutrition on
reproductive performance’. More precise objectives might be: ‘To quantify the response to treatment with the
trypanocidal drug Samorin in early pregnancy on conception rates of goats exposed to a high trypanosomosis
risk’, or ‘To determine the effect of anthelmintic treatment on weight gains of lambs to weaning in different
breeds of sheep’. The scope of the experiment should also be clear: which breed(s) will be used, what weight
and age/parity ranges will be studied, what management system will be followed etc.
Data collection
At the planning stage, one needs to decide what data to collect and how they should be measured. For
example, sera samples may need to be stored and frozen. Is this practically feasible in an on-farm trial?
With large on-farm trials in particular, it might be useful to run a small pilot study before the main trial to
reveal and resolve any unexpected problems in data collection. Data recording sheets should be designed
before data are collected. They should ideally allow for both easy field recording and easy statistical
analysis without the need to copy data from one sheet to another. The main methods of statistical analysis
should be decided before the trial is started. There is little point in collecting a mass of data if you do not
know how you are going to analyse them.
Some guidelines (not necessarily rules) for good practice in data recording are as follows:
1. As far as possible data that are collected manually should be entered on the same recording sheet, i.e.
do not put body weight on one sheet, PCV on another etc. Results of subsequent automated analysis
(e.g. ELISA antigens) should not, however, be copied to this sheet. It is better to devise a system for
entering these data separately and merging with the manually recorded data in the computer.
2. Do not transcribe data from one sheet to another. This is a time when many errors occur.
3. Try to involve two people in measurement and recording so that they can cross-check each other.
4. Decide beforehand on accuracy required for data recording (e.g. to the nearest kg, 0.5 of a kg etc),
and take care when reading values from an instrument recorder. Values of 25 or 30% for PCV,
for example, are, in practice more often recorded than values of 26 or 29% when using a gauge to
measure the height of the packed red cell volume read from a straw after centrifugation.
5. Enter animal identification on the recording sheet in the order in which animals are weighed and
the samples collected. This may be preferable, particularly in large field trials, to using recording
sheets on which animal numbers have already been written.
6. When recording information on breed, age etc, think how you might want to use it. Information is
usually better recorded numerically when the numerical code is easy to remember (e.g. 1 – male;
2 – female). Remember, however, that the computer can easily recalculate values, so do not do
mental arithmetic in converting names to numbers if you are likely to make mistakes.
7. Do not do calculations by hand that the computer can do for you, e.g. calculation of weight gain.
8. If entering data into the computer, preferably get someone else to verify the data you have
entered.
The way a recording sheet might be produced is illustrated in Table 8.5. Note that numbers have been
used where possible. Age is a difficult variable to record in a mixed age population. One way around the
problem is illustrated in the table. It might be preferable to separate trypanosome species in the recording sheet
as shown, rather than list species under one column headed ‘trypanosomes’. This is easy to do and means that
you do not have to get the computer to separate species later. The value 2 represents a score for intensity of
infection (1–5), in this case for Trypanosoma vivax. Note that all variables and attributes are entered together
in columns on the same sheet with animals listed down the left hand side of the page.
129
Table 8.5. Example of a recording sheet.
Sample
no.
Animal
ID Breed
Calf/
adult Age PCV T. congolense T. vivax T. brucei
1 301 1 1 12 mo. 26 0 0 0
2 126 1 2 4 y 24 0 2 0
3 313 2 2 6 y 29 0 0 0
4 219 2 1 9 mo. 27 0 0 0
Experimental design
You will need to decide how many animals or farms to use based on previous knowledge of the likely variation
you might expect. If you do not have sufficient resources you will have to consider either a modification of your
objectives, carrying out the trial in stages over a number of years, or abandoning the trial altogether. There is
little point in proceeding with a trial if you are fairly sure that it cannot meet your objectives!
Experiments should be kept as simple as possible in terms of design, execution and analysis. The
most efficient designs are often the most simple to analyse. For simplicity of analysis there should be equal
numbers of experimental units assigned to each treatment. Such designs are known as balanced designs.
While this is usually relatively easy to organise for crops, it is not always so easy with livestock and on-
farm trials. Unbalanced experiments need sophisticated statistical analysis which cannot readily be done
without statistical computer software. Sometimes the analysis has to be done that way, either because it
was not possible to make the experiment completely balanced in the first place, or because animals may
have died during the course of the experiment.
Assignment of animals. Practical arrangements should be simple to avoid problems, confusion and
errors when implementing an experiment. However, this should not be at the expense of other
considerations, such as avoiding systematic errors. For example, with an on-station trial in individual
pens, it might be convenient to put all animals on the same diet in adjacent pens. Animals on different
diets will then be in different parts of the barn. However, the barn may not be a uniform environment;
some regions may be more exposed to, say, wind and cold air. The differences between diets may be
confounded with differences in environment, making the results impossible to interpret.
Randomisation should be used to overcome these sorts of problems, even though it may introduce
some practical complications.
Precision. An experiment will give an estimate of the difference between two treatments, e.g. the
difference in their effects on mean growth rate. The precision of an estimate reflects how close the
estimate is likely to be to the true or population value and, as we have already seen, this is measured in
statistical terms by its standard error. The smaller the standard error the greater the precision.
The precision of an experiment will depend on the size of the experiment (e.g. the number of animals
involved). Precision increases as the number of experimental units per treatment increases. It also depends on
the inherent variability of the experimental material. An experiment with large variation among experimental
units will have less precise estimates than one for which there is very little variation among units.
Replication or blocking. Estimating the precision of an experiment depends on the amount of random
variation. Therefore, it is essential to be able to calculate the standard deviation of the data. In order to
do this there must be data from more than one experimental unit in each treatment. Thus the
treatments must be replicated.
Very low levels of replication such as two experimental units per treatment may not be sufficient to
give an adequate estimate of the standard deviation nor to give reliable estimates of treatment means. As a
general guide one should always try to ensure that at least six degrees of freedom are available for
estimating the residual variance in the analysis of variance. The precision of an experiment is also
increased by increasing the number of times each treatment is replicated.
130
It has previously been explained that a good choice of blocks will make the variation among blocks
greater than the residual variation within blocks. This is also a good way of improving the precision. In the
data example (Table 8.1), blocking based on initial weight did not reduce the residual variance. In other
situations, however, blocking based on body weight will be beneficial and, therefore, is a good principle to
adopt. Sometimes there may be several choices for suitable replicates or blocks, e.g. body weight, age,
sex. Research workers are often tempted to make treatment groups similar in all respects by balancing
such attributes across treatments. This is wrong and can introduce systematic errors. This method gives no
scope for randomisation of animals to treatments.
It is essential to allocate treatments at random to experimental units. This is to eliminate any
subjective bias which may occur (no matter how honest one tries to be). Randomisation minimises the risk of
systematic errors and helps to ensure that each treatment is represented fairly in the trial. The aim of
randomisation is to give each experimental unit an equal chance of being allocated to any of the treatments.
8.4.2 Experimental units
The best formal definition of an experimental unit is that it corresponds to the smallest division of the
experimental material such that any two units may receive different treatments in the actual experiment.
For example, in an on-station feeding trial, if animals can be fed individually then any one animal can
receive any of the feeds. An animal is the experimental unit in this case, and animals are randomised to
treatments. However, if individual pens are not available, then animals must be group fed. Any animal
must then be given the same diet as its pen-mates, will compete with its pen-mates for food and therefore
cannot be considered independent of them. In this case, a pen is the experimental unit, the pen is
randomised to treatments and all statistical calculations to compare treatments should be carried out on a
pen basis. Group feeding may have the advantage, though, of being closer to farm practice.
Experimental units need to be carefully thought about and defined in on-farm trials. With many types of
treatments it will not be possible for a farmer to give one treatment to some animals and a different treatment to
others. All animals in one herd or flock will be given the same treatment. In many such cases a herd or flock will
be the experimental unit. Data will have to be summarised on a herd or flock basis before statistical analysis is
undertaken. The correct definition of the experimental unit is crucial for the validity of most statistical analyses.
Split-plot designs
These are more common in crop than in animal experiments. They involve two levels of experimental
units, a ‘main plot’and a ‘subplot’. Thus, for example, an experiment might use a number of flocks, half of
which receive regular anthelmintic treatments. Within each flock, sheep might be randomised into two
groups, one receiving trypanocidal prophylaxis, the other not. For anthelmintic comparisons, flock is the
experimental unit, for trypanocidal comparisons, the individual animal is the experimental unit. When
using split-plot designs one has to be careful to ensure that adequate numbers of replicates are available at
the main-plot level since there is more variation among flocks than among sheep within flocks.
8.4.3 Randomisation
Once a suitable method of blocking has been decided, one needs to design an adequate number of blocks
which can allow equal numbers of experimental units to be assigned to each treatment. Thus, in the
example, lambs were ranked in order of body weight and eight sets or blocks of four lambs formed. The
four lambs in each block were then assigned at random to each of the two diets. There are different
mechanisms for assigning experimental units at random.
Drawing lots
A card or piece of paper can be prepared for each treatment, or cards can be taken from a pack of playing
cards. The cards can then be shuffled and drawn like raffle tickets. The treatment marked on the first card
drawn is allocated to the first animal and so on.
131
Using tables of random digits
Many statistics books give tables of random numbers. Alternatively, the digits from telephone
numbers in a telephone directory, for example, could be used. As an example, suppose we wish to
allocate six treatments to six animals in a block, and suppose the sequence of random digits listed in a
table is:
673822908569831470580818615432....
The first animal in the block would have treatment 6, the next number, 7, would be ignored since
there is no treatment 7. The second animal would get treatment 3, 8 is ignored and the next animal would
receive treatment 2. The next 2 in the sequence would be ignored because treatment 2 has already been
allocated, and so on. The full randomisation for the first block would then be:
animal 123456
treatment 632514
Continuing, we get for the second block:
animal 123456
treatment 516432
The lines below show again the random sequence; the digits that are ignored either because a digit is not in
the range 1 to 6 or it has been used before in the block are underlined.
6738229085698314 70580818615432
Block 1 Block 2
With five or fewer treatments, one can use two or more digits to represent a treatment. For example,
with three treatments, digits 1, 4, and 7 could represent treatment 1, digits 2, 5 and 8 treatment 2 and digits
3, 6 and 9 treatment 3. Then only 0 would be ignored. With more than ten treatments, random digits can be
taken in pairs giving random numbers from 0 to 99. Again different pairs of digits can be used to represent
the same treatment.
Other methods
A dice used in children’s games can be used instead of random digits for six or fewer treatments. Note
though that one should not use two dice for up to twelve treatments, since numbers such as 7 (3+4, 2+5 or
1+6) are far more likely to occur than 12 (6+6), and the number 1 cannot occur at all.
8.4.4 Designing surveys
An essential component of a survey is a sound sample design. In most cases it is impractical to collect data
on all individuals of a population, but a representative subset or sample of the population can allow one to
draw inferences about the characteristics of the population. There are several methods to select animals for
sampling. The methods can be classified in two types, one type that does not and one type that does rely on
random sampling techniques.
Objectives
As in designing experiments it is important to clearly state the objectives and to list the variables to
be estimated. It is also important to define the target population and to ensure that the sampled population
is representative of the target population. It would be inappropriate to use data from diagnostic
laboratories or abattoirs, for example, and then to make inferences about the occurrence of disease in an
animal population of a country.
132
Sampling frame
The population to be sampled must be divided into sampling units before selecting the sample. A unit can
vary from an individual animal to an aggregate of individuals, such as pens, farms, dip tanks, villages etc.
The list of all sampling units in the sampled population is called the sampling frame.
Non-random sampling
Two methods can be used: convenience sampling and purposive sampling. In convenience sampling, a
sample is selected because it is easy to obtain, e.g. a herd near the investigator. In purposive sampling, the
selection of units is based on known exposure or disease status. It is thus an inadequate technique for
obtaining data to estimate the occurrence of the disease in a population. Both methods may create serious
biases.
These methods are not based on probability sampling techniques and therefore, it is not possible to
give a level of confidence to the mean. These methods are not usually encouraged since they may produce
inaccurate population estimates and are often as expensive to perform as random sampling methods.
There may be good reasons on occasions for using such methods, but the experimenter needs to be aware
of their limitations.
Random sampling
A number of methods are available, e.g. simple random sampling, systematic random sampling, stratified
sampling and cluster sampling. Confidence limits of the population values can be calculated. These designs
are more complex to analyse but the validity of the results is worth the investment in time and money.
The objective of random sampling is to select a percentage of the population using a formal
randomisation process as has already been described. Such a selection procedure is required for the
investigator to calculate the precision of the sample estimate as measured by the standard error of the
mean. In practice, random sampling provides the investigator with the assurance that his sample will be
representative of the population being investigated.
To obtain a simple random sample of n from a population of N animals to determine the prevalence
of trypanosomosis by antigen-ELISA analysis, for example, one would randomly select n numbers from
between 1 and N, each number identifying an animal in the sampling frame. Thus, the animals selected
would be distributed randomly throughout the sample population.
The estimated standard error of the sample mean:
=s/?n× ?(1 ? n/N)
where s is the calculated standard deviation, which equals ? p(1 ? p) for a binary variable.
Note that this is different from the standard error used previously in the analysis of designed
experiments. It has the additional value ?(1 ? n/N) which is included to take into account the fact that the
sample is from a population of finite size N. If the population is infinite (or very large), this term can be
ignored.
Systematic random sampling
In systematic sampling the n sampling units are selected from the sampling frame at regular intervals (e.g.
every 10 animals). The interval k between animals (in this case 10) is the integer closest to N/n. When
systematic methods are used, the starting point in the first interval is selected at random. Systematic
sampling is a practical way to obtain a representative sample, and it ensures that the sampling units are
distributed evenly over the entire population. It is particularly useful for sampling animals arriving at a dip
tank. Then, we can assume that the sample is random and a standard error can be calculated. When
sampling animals at a crush on a farm, however, the systematic sample may give an accurate estimate of
the mean, but it may not be strictly random because the weaker animals may arrive last at the crush. If there
is any doubt about whether one can pretend that a sample is random or not, and the sampling frame is
known in detail in advance, then it will usually be advisable to select the whole sample at random.
133
Stratified random sampling
With this technique, the sampling frame is divided into blocks known as strata based on factors (e.g. herd,
area, age of animal) likely to influence the level of the primary variable (e.g. prevalence of helminthosis)
being estimated. A simple random or systematic random sample is then selected within each stratum.
Stratified random sampling improves the precision of the sample estimate, because only the within
stratum variation contributes to the variation (standard error) of the mean. Different sampling percentages
can be used in the various strata. For example, if 60% of animals are from the Red Maasai breed and 40%
from the Dorper breed, then the sample of size n can be selected from the breeds in the same ratio (e.g.
60:40).
When stratified sampling is used, information on all sampling units, with respect to the factors
forming the strata, must be known before drawing the sample. In general, the number of factors used for
stratification should be limited to those likely to have a major impact on the mean value of the primary
variable to be estimated (e.g. prevalence of helminthosis).
An estimate of the population mean is given by the formula:
y = ??
i
y
i
/N = ?N
i
/N
where y
i
is the sample mean in stratum i of size N
i
.
The standard error of y = (1/N) × ??[?
i
2
× s
i
2
/n
i
× (1 ? n
i
/N
i
)].
Cluster sampling
In cluster sampling, the first sampling unit is larger than the particular unit of interest, and sampling takes
place at several stages. For example, cluster sampling in two stages could include a sample of primary
units (e.g. farms) and then, within the farms, a sample of secondary units (e.g. animals).
Cluster sampling is often used because of its practical advantages and flexibility. The number of
primary (n
1
) and secondary (n
2
) units may be varied to account for different costs of sampling primary
versus secondary units, as well as the variability among primary units and among secondary units within
primary units.
Whenever possible one should select sampling units at each stage with probability proportional to
the number of population individuals in the stratum or cluster. This minimises the standard error and
stabilises the sample size. Usually cluster sampling generally requires more animals to achieve the same
precision as simple random sampling, but the costs will be lower. The formulae for estimates of means and
standard errors can be found in statistical text books.
8.4.5 Sizes of experiments and sample surveys
Experiment
Before conducting an experiment it is important to decide on how many animals, flocks, herds etc one may
need in order to estimate treatment means to a desired precision. In other words, we need to know how
many replicates or blocks we need in order to have a reasonable chance of detecting a real difference at a
given level of statistical significance.
We have already seen that the standard error of the difference between two means, each of n
observations, is given by the formula:
SED = ?(2 × variance/n) = ?2SE
One might have a preconceived idea of the approximate size of the standard deviation or variance to
be expected. Thus, if the experiment described in our example were to be repeated using two different
breeds, we could assume that we might expect a residual variance of about 0.5 (it was 0.4518 in the
experiment). Suppose we decided that we needed to determine a difference of 1.0 kg in weight gain
between two groups to be statistically significant (P<0.05). We can then apply the formula.
134
LSD=t × SED = t ×?(2 × variance/n)
Squaring both sides and manipulating terms we get:
n=2t
2
×variance/LSD
2
Substitutingt=2(approximate value with large number of degrees of freedom) we get:
n=(2×4×0.5)/1.0
2
= 4.0
Therefore we can expect to require four animals per group to show a difference in weight gain of 1.0 kg as
significant between diets (P<0.05).
With a completely randomised design with 4 groups (breed × diet) we would have 12 (i.e. 4 × 3)
degrees of freedom for the residual mean square. t
12
= 2.18 where t
12
is the t-value with 12 degrees of
freedom. Replacing t
2
with 2.18
2
in the above formula we can revise our sample size estimate to:
n=(2×2.18
2
× 0.5)/1.0
2
= 4.75.
Therefore we would in fact need five animals per diet.
Surveys
Suppose we wish to carry out a survey to estimate the trypanosome prevalence in a given population of
cattle. To estimate the size of a random sample we follow the same steps as in the previous section. First we
decide how accurately we need to determine the population mean µ, i.e. how close to µ we want our
sample to be. Approximating to a normal distribution, we can write:
t=(y? µ )/se(y ? µ)
Write y ? µ =L
Then:
SE(L) = SE (y) since µ is fixed, i.e. it is the mean of the population) = ?[p(1– p)/n]
Therefore:
t=L/?[p(1 ? p)/n]
Squaring both sides and manipulating the equation we get:
n=[t
2
p(1 ? p)]/L
2
As before we first substitutet=2forlevel of significance P<0.05. It is clear that in order to estimate n we
need to have some idea of p beforehand. If we are sampling from a finite population of size N then a further
step is needed. In this step we adjust the required sample size to n1 using the formula:
1/n
1
= 1/n+1/N
To demonstrate how to calculate L, suppose p = 0.4. If we decide that we want our sample to provide
an estimate within 10% of this value then:
L = 0.4 × 0.1 = 0.04
Then:
n=2
2
× 0.4 × 0.6/0.04
2
= 600
If the population size is 1200, then:
1/n
1
= 1/600 + 1/1200
and so n
1
= 400.
135
8.5 Presenting results
A set of numerical results should usually be presented as a table or graph rather than be included in text.
Text becomes difficult to read if too many numbers are included. Well presented tables and graphs can
concisely summarise information which would be difficult to describe in words alone. Poorly presented
tables and graphs, however, can be confusing or irrelevant. In general, tables are better than graphs for
giving detailed numeric information, whereas graphs are better for indicating trends and making broad
comparisons.
Tables and graphs should, ideally, be self-explanatory. In other words, the reader should be able to
understand them without detailed reference to the text. The title should be informative and rows and
columns of tables and axes of graphs should be clearly labelled. Graphs and tables should be as simple as
possible, while having sufficient detail to be useful and informative.
In scientific papers statistical information, such as standard errors and significance levels, is usually
required. This may not be necessary for articles for a more general readership or for slides or
transparencies for use in a lecture. Such statistical information should always be presented in such a way
that it will not obscure the main message of the table or graph.
One often finds tables in scientific papers clustered with little superscripts a, b, c etc, or symbols *,
*** to signify significant differences. Authors need to ask themselves whether these are really necessary,
as they tend to obscure the main ingredients of the table. In some cases they will feel they are necessary and
help with the understanding of the table. Think twice, however, before doing this. Normally tables are
much better if they are simple and present the data at their face value. Often only means and standard errors
are required. The summary values given in Table 8.6 are all that are needed to present the mean results for
the breed/dietary supplementation example, and, provided the reader understands standard errors, he/she
will be able to draw his/her own conclusions on how biologically significant the results are. Significance
levels can be quoted in the text as a guide to the reader, but the reader should be able to draw his own
conclusion on how ‘significant’ he/she feels the results are.
Table 8.6. Weight gains (kg) of lambs.
Supplementation
Breed No Yes Mean
1 0.82 1.89 1.36
2 1.49 3.48 2.48
Mean 1.16 2.8 1.92
SE of differences between means in body of table = 0.24.
SE of differences between overall breed and dietary means = 0.17.
The number of digits and decimal places presented in a table should be the minimum number that is
compatible with the purpose of the table. Thus, in this example three decimal places for weight gain, e.g.
0.825 may be too many, since the third decimal place may have little practical significance; conversely,
one decimal place, e.g. 0.8, 1.9, could lose too much information. As a general rule, standard errors should
be written with the same or one more decimal place than the mean.
The data in Table 8.6 are presented in the form of a bar chart in Figure 8.5. Graphical presentation
often helps to highlight certain aspects of the results. Individual standard errors for each diet can be
presented as a vertical line above a bar. If it is necessary to include standard errors of differences between
means, these can be included as a footnote or in the text. A graph should never duplicate the results
displayed in a table in the same report.
136
Figure 8.5. Weight gains (kg) of sheep in the form of a bar chart.
8.6 Exercises
1. Calculate the mean, median, standard deviation and coefficient of variation of the PCVs of the
eight lambs in breed 1 in Table 8.1 fed diet 1.
2. Using the results of Exercise 1, calculate the standard error and 95% confidence interval for
mean PCV of breed 1 lambs fed diet 1.
3. In a sample of 80 sheep 5 are detected with trypanosomes. Calculate the standard error of the
prevalence of trypanosomes assuming first a Poisson distribution and second a binomial distribution.
Calculate the 95% confidence interval using the formula given for the normal distribution.
4. Write down a statistical model to describe the factors which you might expect to affect incidence
of helminthosis during a 3-year study of 10 flocks of sheep.
5. Calculate the variance for weight gain of the eight lambs within each breed–supplementation
group in Table 8.1 and show that the average of these variances is the same as the residual mean
square (0.4474) calculated in the one-way analysis of variance by group.
6. Using the following means and standard errors calculated for weight gain:
Supplementation
Breed No Yes Mean
1 0.82 1.89 1.36
2 1.49 3.48 2.48
Mean 1.16 2.68 1.92
SE of mean in body of table = 0.236
SE of overall mean by supplementation or by breed = 0.167
Calculate approximate 95% confidence intervals for the various means. For example, for breed 1,
non-supplemented, you should get a 95% confidence interval of 0.82 2 × 0.236 = (0.35 ? 1.29). Briefly
explain what these results show you.
137
SE
Breed
0
1
2
1
Not supplemented
Supplemented
3
4
2
12
SE
SE
Weight gains (kg)
SE
7. Complete the following analysis of variance table for PCV using the data in Table 8.1.
Source of variation df SS MS F
Breed (B) 1
Diet (D) 1
B × D1
Residual 28 560.3750
Total 31 888.4688
Interpret the findings.
8. Using the results of Exercise 7, calculate SED for comparing mean PCVs of breeds 1 and 2,
calculate the t-value and show that t
2
= F value for breed in the analysis of variance table.
9. Using only data for non-supplemented lambs (in Table 8.1) carry out a t-test to compare FEC
between breeds, firstly using values on their original scale, secondly transforming FEC (y) to
loge (50+y). The value of 50 is used because it is the least detectable FEC.
10. Two flocks were sampled for trypanosomosis. In one flock 12 of 32 animals were found to be
parasitaemic (38%), and in the other 4 of 19 were parasitaemic (21%). Are these trypanosome
prevalences significantly different?
11. Two hundred and forty breeding does (120 Mashona type and 120 Matebele type) purchased
from communal areas are being used in a 6-year continuous mating study. Does were randomly
allocated to eight treatment groups within breed. A 2 × 2 × 2 factorial design was used. Two
levels of feed supplementation (no supplementation and supplementation), dosing (no dosing
and dosing with anthelmintics) and dipping (no dipping and dipping at regular intervals) were
evaluated. After three years of study deaths within one of the treatment groups for the Mashona
breed have resulted in no does or kids remaining in that group. The experiment has another three
years to run. Should the researcher fill this group with new does or continue to analyse the
experiment with one treatment group missing? Write down any questions you would like to
address to the researcher.
12. You plan to carry out a farm trial to assess the effect of anthelmintic treatment on growth rate of
lambs to weaning. You have 10 flocks, each on a different farm, at your disposal. It will be
possible to divide the flocks on four of the farms so that each half grazes in a different area. The
flocks on the other farms cannot be subdivided in this way. It has been decided that it would not
be desirable to mix treated and untreated sheep in the same grazing pasture. Design two
experiments, one based on just the 4 flocks which can be divided, the other based on all 10 flocks
to carry out the objective. Write down the statistical models in each case and sketch the form of
the analysis of variance table. Which are the experimental units in each case?
13. Suppose that the lambs in the example data set (Table 8.1) do not belong to two breeds but that
they are all Red Maasai and the code used for breed is instead anthelmintic treatment (yes or no).
Redesign the experiment to investigate the effect of supplementation in the presence or absence
of treatment and reassign the lambs to treatment and supplementation using the table of random
numbers at the end of these notes.
14. Suppose a stratified random sample of 40 cattle is chosen: 20 from 180 zebu and 20 from 120
N’Dama cattle. Suppose 10 of the zebu and 2 of the N'Dama are found with trypanosomes in their
blood. Calculate the estimated mean trypanosome prevalence in the population of 300 cattle and
its standard error.
15. The mean trypanosome prevalence is to be estimated from 200 cattle in 20 herds, which range in
size from 36 to 180 animals. It has been decided to select 10 herds at random with probability
proportional to their size and then to select 20 animals at random from each herd. Describe how
you would select the 10 herds.
138
16. Calculate how many replicates you would need to determine a difference of 0.5 kg as significant
(P<0.05) between any group in the breed x diet experiment described in Table 8.1 if your
estimated standard deviation was 0.8 kg.
17. Suppose you wish to estimate the size of the sample needed to estimate antigen prevalence rates
of Theileria parva in zebu calves. We estimate that the likely prevalence is 50%. Calculate how
many calves you will need to sample if the population mean prevalence in a group of 400 calves
is to be estimated to within a 10% prevalence range in accuracy (namely 45–55%). What will be
the predicted accuracy of this sample if the actual prevalence turns out to be 80%?
18. There are at least three ways that Table 8.6 could be improved for publication in a scientific
paper. List these. Are there any changes that you would make to this table if you were to present it
as a transparency in a talk?
19. Using the following analysis of variance table:
Source of variation df SS MS F Probability
Breeds (B) 1 10.1250 10.1250 21.68 <0.001
Blocks/breeds 6 2.2538 0.3756 0.80 NS
Diet (D) 1 18.6050 18.6050 39.84 <0.001
B × D
1 1.7112 1.7112 3.l66 0.07
Residual 22 10.2738 0.4670
Total 31 42.9288
and applying the t-test to the means and standard errors summarised in exercise 6, write, in no more than
three sentences, a report of the results.
8.7. References and reading materials
Cochran W.G. and Cox G.M. 1992. Experimental Designs. John Wiley and Sons, Inc., New York,
USA. 634 pp.
Cox D.R. 1987. Planning of Experiments. John Wiley and Sons, Inc., New York, USA. 296 pp.
Snedecor G.W. and Cochran W.G. 1982. Statistical Methods. Iowa State University Press, Iowa,
USA. 507 pp.
Steel R.G.D. and Torrie J.H. 1960. Principles and Procedures of Statistics. McGraw-Hill, London,
UK.
Thrusfield M. 1986. Veterinary Epidemiology. Butterworth & Co. (Publishers) Ltd. 280 pp.
139
Module 9: Research project protocols
9.1 Performance objectives
Module 9 is intended to enable you to:
1.
List components of a protocol.
2.
Plan for writing a protocol.
3.
Write a protocol.
4.
Co-ordinate the review process for your protocol.
9.2 Introduction
As a scientist working on small ruminant production, your goal is to solve the problems of resource-poor
farmers through research. This research should focus on the farmers’ problems and needs to be an integral
part of the research of your institution and maybe the region you are in. You may also be a member of a
network, e.g. the Small Ruminant Network (SRNET). In all these situations, you need to plan, develop
and submit your research ideas in a format that can be reviewed and approved by the appropriate body. To
achieve this objective, you need to plan your research and develop a project protocol. Funds for research
are becoming scarce and scientists have to compete for the few resources available. This competition can
be for resources provided by your national research institution or a donor agency; in both cases you need to
develop and write a protocol that will convince the donors to fund your research. This module will enable
you to develop the skills to do so. This module will also cover a few topics related to implementation when
your protocol is approved: administration, financial information and reporting. These topics are usually
not part of a protocol unless the supporting institution wishes them to be.
141
9.1 Performance objectives
9.2 Introduction
9.3 Research project protocols
9.4 Planning
9.5 Components of a protocol
9.6 Writing strategies
9.7 The review process
9.8 Exercises
9.9 References and reading materials
9.3 Research project protocols
What is a research protocol? It is an instrument that describes your intentions to conduct research to solve a
problem. It is a document that details what needs to be done:
• It can have one or several proposals.
• It is a project portfolio (including one proposal or several proposals).
In countries where funding is mostly obtained from donor institutions, it can be called a grant
proposal. A protocol is more inclusive than a proposal, a term that is used by many in place of protocol. A
protocol can contain one or more proposals, e.g. ILRI submitted to the EU a protocol with three grant
proposals, one for small ruminants, one for cattle and one for nutrition.
9.4 Planning
Several prerequisites will help you in planning, developing and finally writing a good protocol:
• You have good research skills, are experienced and have published articles on previous research.
• You are committed to implementing the research.
• You are linked to a stable multidisciplinary team and/or network.
• Your institution is committed to research.
When planning your research, you should:
• be innovative
• provide a clear rationale
• be focused
• be timely
• address specific problems that farmers have.
If the proposal has many experiments, then the experiments should be feasible and in a logical
sequence.  You should know the resources available to you before writing a protocol: personnel,
equipment, supplies, land, animals etc. You should also know the study area and the needs of the farmers.
Knowledge about other institutions is important as joint protocols may be called for by the complex nature
of the problems.
The planning process states:
• the problem: the why
• the importance of the problem
• what needs to be done
• how the problem will be addressed: the work plan
• resources (human and financial) needed to accomplish the plan.
9.5 Components of a protocol
The attributes of a good project protocol:
• Well defined problem statement.
• Clear, well focused and realistic objectives.
• The work plan must not address too many questions.
• The budget must be realistic and adequate for the intended purpose.
142
This section will give guidelines on what goes into each component of a protocol. Only
experience will teach you to perfect your protocol. Different institutions (e.g. your research
organisation, a network or a donor) may require more components or break down the components
into more categories. The attributes of a protocol listed above are the main components that are
usually required for most protocols.
9.5.1 Background and justification
This section of the protocol should:
• state the problem and rationale for solving it
• clearly describe the problem
• state what we already know
• state what needs to be done.
This section is used to demonstrate your understanding of the subject and to justify the need to
conduct the proposed research. You should show why the information that will be obtained through
research will be useful; this means explaining how the information can be used to solve the problems of the
farmer. Work done by others that is relevant to your work should be clarified; avoid controversial issues in
this section.
9.5.2 Objectives
This section lists and the overall objective and the specific objectives. The specific objectives are items in
a list that will contribute towards attaining the overall objective. You should describe:
• the research that you intend to accomplish
• the hypotheses that you will test.
Remember that your success will be evaluated against the objectives you have set out to
achieve.
9.5.3 Expected outputs
Expected outputs are indicators used to verify whether the objectives have been achieved. To illustrate
how outputs should be written here are two examples from a protocol submitted by ILRI to the European
Union to fund the Small Ruminant Network (SRNET):
• Output i : Three issues of newsletter published
• Output ii: Training manual for the small ruminant production technique course will be published
and distributed
143
Example:
Let us use the project ‘Gastro-intestinal parasitism in small ruminants’ to illustrate the two types of
objectives:
• An example of an overall objective of the project: To characterise breeds of sheep and goats in
sub-Saharan Africa for genetic resistance to internal parasites
• An example of a specific objective that contributes towards the overall goal: Determine the
specificity of genetic variation for resistance in terms of different parasite genera
Outputs constitute the terms of reference of the project. They should be written to reflect the results
for which the project is held directly accountable and for which it is given resources. When outputs are
achieved, the project is complete.
9.5.4 Work plan
The work plan describes the materials needed and the methods to be used. It sets out what will be done and
how to achieve it. The description should be clear and detailed for every experiment. There are two
different types of work plans:
• A work plan that describes a single experiment.
• A work plan that describes more than one experiment.
The second type of work plan could be constituted of sequential trials. It could also describe an
umbrella project that will cover a series of trials.
Each trial included in the work plan should have the following attributes:
• Objectives: The objective(s) of each trial should relate to the overall and specific objectives of the
protocol.
• Expectations: Expectations relate to information on how the objectives will achieve the overall
output.
• Time frame: The time frame shows the timetable for the activities in each trial (see Table 9.1
for an example). The targets should be treated as flexible dates but it is important to set such a
timetable to guide the team implementing the experiment(s). In such a framework, activities
should be listed in chronological order for every year of the lifetime of the protocol. You
should be as specific as possible, e.g. feeding schedule, blood sampling dates etc should be
listed.
• Materials and methods: The work plan will include the materials needed in every experiment. What
goes into this part? In a nutshell it should describe what is to be done and how it will be done. You
may find the following terms used but they all mean the same thing:
Materials and methods ? Methodology ? Procedures ? Activities
The materials and methods can be described together or separately. The main guideline is that no
detail should be left out. The agency that you wish to sponsor your protocol will need to know what needs
to be done, why and how.
• Statistical analyses: Experimental design refers to various biometrical designs used in
experimentation (see Module 8). In the protocol you must:
? outline the experimental design and procedures
? for new methods, explain why they are better than the old ones
? explain control treatments clearly
? explain how data will be collected and analysed
? state and describe/define the materials that will be used: crops, animals, feeds
etc
? list collaborators (scientists, institutions)
144
Table 9.1. Schedule of experimental matings to generate genetic parameters in the endoparasite
project, Debre Birhan, Ethiopia.
Mating
Lambing in
dry season
Lambing in
wet season
1
st
mating May 1992 January 1993
1
st
lambing October 1992 June 1993
2
nd
mating May 1993 January 1994
2
nd
lambing October 1993 June 1994
3
rd
mating May 1994 January 1995
3
rd
lambing October 1994 June 1995
Last date of recording October 1995 June 1996
9.5.5 Budget
• The budget section shows the resources that are needed to implement the protocol. Major features
of the budget are:
• It has to be based on the real situation (i.e. be realistic). This can be done by surveying the market
costs of items in the budget. Past experience can be of great help.
• Inflation rates should be included.
• A percentage for contingency should be included.
Items that are usually included in protocol budgets (Table 9.2) are:
• Personnel costs: This includes the list of principal and support staff and labour (permanent and
casual). Consultants are listed in this section.
• Equipment costs: This can include laboratory, field and office (computers, audio-visuals)
equipment and vehicles.
• Operation costs include experimental material (animals, feed, drugs etc), travel, training,
communication, information, workshops and office supplies.
• Contingencies
• Administrative costs (overheads)
• Budget explanatory notes: This part is placed below the budget table. It explains items in the budget
that may need further details or qualification.
145
Table 9.2. Budget format for a protocol.
Budget category Year 1 Year 2 Year 3 Total
Personnel costs
1. Professional staff
2. Support staff
3. Consultants
Subtotal
Equipment costs
1. Lab equipment
2. Animal handling facility
3. Audio-visual equipment
4. Vehicles
5. Computers
Subtotal
Operational costs
1. Animals
2. Feeds
3. Drugs
4. Travel
5. Training
6. Communication
7. Information
8. Workshops
9. Office supplies
Subtotal
Administration cost
Subtotal
Contingencies
Subtotal
Total
Budget explanatory notes.
9.5.6 Logical framework
Many donors require a logical framework to accompany the protocol. The logical framework serves to
improve the planning, implementation and performance appraisal processes by clarifying project design
and making it transparent to the borrower (the project team) and lender (the source of funding). The logical
framework is a table of information about the protocol that:
• stimulates you to think of what you want to do and how
• can be used as reference to monitor the progress of the project as it provides the indicators to use and
the means of verification
• shows the basic assumptions you made before the project started.
146
The logical framework consists of a 4 × 4 matrix, with a vertical hierarchy of objectives:
• goal
• purpose
• output
• activity.
The horizontal components are:
• summaries of the objectives at each level
• performance indicators for achievement of those objectives
• the sources needed to verify the indicators
• the important assumptions for moving from one level of objectives to the next higher level.
The ‘pre-conditions’ is an extra box at bottom right describing important factors which must be in
place for the project to be initiated. Examples are: a. DNA sequencer available (for a genomics project) b.
liquid nitrogen available (for an animal breeding project) c. country statistics available (for a policy
analysis or economics project). An example of a logical framework is given in Box 9.1.
147
Box. 9.1. An example of a logical framework: Logical framework for the Small
Ruminant Network Project
Narrative summary Measurable indicators Means of verification Important assumptions
Goal
Improve and sustain
small ruminant
research and enhance
human resource
capacities in the region
1. Increase in new
small ruminant
production
technologies and
research
methodologies
2. Increase in national
support for small
ruminant research
1. Final project report
2. Level of small
ruminant activities in
the participating NARS
3. Level of national
support for small
ruminant research
1. Commitment to
regional collaboration
continues
2. Small ruminant
research and
development remains a
priority in
the region
3. Networking remains
an acceptable
mechanism for regional
collaboration
Purpose
1. Build small
ruminant research
capacity and
competence in the
NARS
Strengthen
collaboration
between and within
NARS
3. Promote national
small ruminant
network formation
4. Optimise use of
available human and
other resources.
1. Better small
ruminant research
programmes
2. Better management
of research projects
3. Strong
NARS NARS and
NARS ILRI linkages
4. Active small
ruminant network
activities
5. Better utilised
resources (human and
infrastructure)
1. Project annual
reports
2. Project evaluation
reports
3. Increased network
activities
4. Level of
institutional
collaboration and
interaction
1. Existence of
enabling policy
environment
2. National
government s
investment in small
ruminant research
continues
3. Support for
networking is sustained
at regional and national
levels
Box 9.1. cont...
Suggestions for developing a logical framework
1. Involve as many stakeholders as possible in preliminary discussions.
2. Devote adequate time to develop a first draft of the log frame. It is important and deserves the
team’s best efforts.
3. Regard it as a very useful exercise which is essential for the success of the project and an absolute
requirement for obtaining funding.
148
Box 9. 1. cont...
Output
1. Well-established
regional collaborative
research activities
2. Strong research
capacity in the NARS
3. Relevant and
adoptable research
methodologies and
technologies
developed
4. Large volume of
georeferenced data
base on small ruminant
genetic
resources in the region
5. Strong institutional
linkages
6. Enhanced research
leadership capacity
1. Eighteen on-going
research projects in
participating countries
2. Number of
technologies and
research tools
developed and tested
3. Level of interaction
between collaborating
institutions and
scientists
4. Level and quality of
information
accumulated on small
ruminant genetic
resources
5. Level of
management of
research and
accountability
6. Established national
small ruminant
networks
1. Number of research
grants awarded
2. Research project
reports
3. Research portfolios
4. Publications at local
and international levels
5. Level of NARS
participation in
network activities
6. Level of interaction
between and within
institutions and
scientists
7. Number of active
national Small
Ruminant networks in
the region
1. Research
programmes are
scientifically sound and
well managed
2. All collaborators are
committed
3. There is an enabling
environment
4. Funds allocated for
project implementation
are available on time
Activities/inputs
1. Steering Committee
meetings to plan and
sanction network
activities
2. Research planning
and development
meetings
3. Implementation of
approved research
protocols
4. Documentation and
publication
5. Counselling and
training
6. Co-ordination
1. Steering Committee
to hold two meetings
2. Number of research
planning and
development meetings
3. Number of on-going
research projects
4. Research reports
and published papers
5. Training
programmes completed
or on-going
6. Planned network
activities implemented
1. Records of
committee meetings
2. Reports on project
activities
3. Newsletters,
proceedings and
published papers
4. Number of training
courses and scientists
trained
5. Completed network
activities
1. Effect of natural
disasters is minimal
2. Infrastructure
(laboratories, farms etc)
is available
3. Inputs/resources are
available
Project Budget (US$)
1. Personnel 168,090
2. Services 4,250
3. Supplies 12,500
4. General
expenses 12,100
5. Travel 14,500
6. Capital 3,700
Total 215,140
4.
Begin with purpose as this is the objective of the project/programme. In developing purpose,
ensure it will contribute to goal (which should be pre-defined). Remember, success is purpose
achieved.
5.
Identify important assumptions, i.e. those which must be fulfilled in order for the project to
achieve the goal.This provides an indication of the feasibility of the project.
6. Define outputs required to achieve purpose. Remember you are contractually obliged to produce
these. When outputs are achieved, the project is completed.
7. Define activities required to achieve outputs.
8. Identify important assumptions at the activities, outputs and goal levels. Do not underestimate
the value of this step as it can lead to a revision of outputs and/or activities.
9. Detail objectively verifiable indicators (OVIs) for goal, purpose and output and show a budget
summary at the activity level
10. Identify means of verifying at all levels.
11. Use a first draft as a basis for further discussion with all stakeholders.
12. Revise draft.
13. Send to selected independent reviewers for comment before finalising.
9.6 Writing strategies
Protocols differ depending on the nature of the problem that needs to be addressed and the format
requested by the institution to which you will submit the proposal. This section will provide you with
general rules to help you write a good, well presented protocol.
• Step 1: Begin to write your protocol early as it takes time to write a good protocol
• Step 2: Set a time outline for yourself. This will leave you ample time to think, plan, prepare,
outline, write, revise, have colleagues review, revise, polish, get administrative approval, and mail
the protocol.
• Step 3: Use the checklist provided by the donor or network.
9.6.1 Some writing guidelines
1. Make an outline for each component of the protocol to safeguard against missing important points.
2. Include well designed and labelled tables and figures.
3. Be accurate, clear and consistent:
• Use a logical sequence of presentation of the ideas.
• It is advisable to avoid big words and jargon. Your donor reviewer may not have time to go to the
dictionary.
• Avoid ambiguity.
• Terminology and abbreviation should be the same throughout the text.
• Appropriate sections of the protocol should agree with each other.
4. Do not include irrelevant information.
5. Use simple words, short, direct sentences and short paragraphs.
6. Try to be positive but do not overstate your case.
If editing facilities are available in your institution, it is a good idea to use them.
149
9.7 The review process
Discussing your ideas and sharing the first or second drafts of your protocol with colleagues is a good step
to help you avoid pitfalls. It will give you an idea of whether others understand what you want to convey. It
is helpful if your institution has a review system for all protocols before they reach their destination. It is a
good idea to set some guidelines for your reviewers on the feedback you need. We suggest the following
points:
• scientific and technical significance of the research
• originality
• adequacy of methodology
• suitability of the facilities
• appropriateness of the budget
• appropriateness of time to carry out the research.
Request the reviewers to summarise the strengths and weaknesses of the protocol and provide
recommendations.
9.8 Exercises
1. To write your protocol, the following are the main components:
? Background and justification
? Objectives
? Expected outputs
? Work plan
? Budget
? Logical framework
To practice, start with your ideas for a research project and answer the following questions for all of
the above sections:
• What is the purpose of this section of the protocol?
• What should the scope of this section of the protocol be?
• How should the subject be introduced?
• What are the main ideas to be included in this section?
• Which of the items are essential? not essential?
9.9 References and reading materials
Reif-Lehrer Liane. 1989. Writing a Successful Grant Application. Jones and Bartlett Publishers,
London, UK. 283 pp.
150
Module 10: Guidelines to communication
10.1 Performance objectives
Module 10 is intended to enable you to:
1.
List the steps for planning the communication requirements of your research project.
2. Use guidelines for (i) oral communication, (ii) interpersonal communication, (iii) visuals and
(iv) writing and editing.
3. Write papers for publication in refereed journals.
4. Write review articles.
5. Prepare posters.
6. Prepare overhead projection transparencies.
10.2 Introduction
Scientific research to solve the problems of livestock systems is a prerequisite to generating technologies.
It is then the duty of the researcher, extension agent and others in government and non-governmental
organisations to communicate this research to farmers, development agents, other researchers, donors and
the general public.
This module is written for the livestock researcher. It broadly describes ways to communicate and
provides reading materials that can be consulted for more details. Since the field of communication is
wide, researchers are encouraged to seek information on skills not covered in this module from
communication text books.
10.3 Planning communication in a research project
We have illustrated how to write protocols in Module 9. Communication of results should be part of the
research plan. Communication activities must be planned at the onset of your research project. Most
151
10.1 Performance objectives
10.2 Introduction
10.3 Planning communication in a research project
10.4 Communication guidelines
10.5 Writing scientific papers
10.6 Writing review articles
10.7 Posters and overhead transparencies
10.8 Exercises
10.9 References and reading materials
researchers think of communicating their results on an ad hoc basis and this usually happens near the end
of the project. This behaviour results in poor communication of the results.
• As part of the plan for your project, develop the plan of communication activities. For example:
seminars, annual reports, newsletter articles, refereed journal articles, workshop proceedings and
project reports.
• Write the objectives for each communication activity.
• Define the expectations and agree on the role of each member of the research team in the project.
• Discuss issues with professional communicators (if you have them in your institution).
• List the agenda for action in order of priority.
• Agree on the deadlines for outputs.
• Implement the plan.
• Monitor the implementation of the plan.
• At the end of project evaluate the outputs of the communication plan as part of the overall
evaluation exercise.
10.4 Communication guidelines
In order to communicate properly you need to acquire the necessary skills. In this module we will focus on
guidelines for oral presentation, interpersonal communication, visual communication and writing and
editing. We hope you learn and use these guidelines to communicate your research results to the various
audiences (farmers, general public, researchers, extension agents, other development agents and donors).
Depending on the situation, you may find yourself in need of other skills. Since it is beyond the scope of
this manual to cover all communication skills a list of reading materials is included at the end of this
module.
10.4.1 Guidelines to oral presentation
To communicate the results of your research, you often need to present your results in scientific meetings
or to farmers. In both cases you need to acquire oral presentation skills. You need not fear speaking in
scientific meetings or in meetings with farmers provided you learn the skills needed for this purpose (see
Tripathi 1991).
What do we mean by presentation? Presentation means communicating your message by sight and
sound via the co-ordinated act of mind, body, language and voice. Presentation skills can be divided into
stage craft and delivery.
Stage craft
Stage craft is the effective management of stage devices. The stage is the place from where you make your
presentation.
• The stage can be set into different configurations. For effective presentations, the front centre and
the left side of the stage are the strongest. Movement on the stage should be co-ordinated for the
most effective relay of the message. You should move forward if you want to stress a point,
backward to show rejection and laterally to signal change in topics.
• The speaking platform in the stage is a table or podium. It is placed on the stage to facilitate
placement of notes, transparencies and other material that you may need for the presentation.
• A public address system, composed of speakers and microphones, is sometimes available. If the
system is not functioning well, raise the tone of your voice to keep the audience alert to your
messages. Tips for using a microphone include:
152
? talk over the microphone, not to it.
? do not handle sheets of paper near the microphone.
• You may like to use a board (e.g. whiteboard) or flipcharts.
• You may need to use projection equipment. It is advisable to learn how to use the projection
equipment (overhead or slide projector) before the meeting. You need to check that the equipment
is functioning well before your presentation.
Delivery
Delivery is the communication of the message by verbal and non-verbal means. To master some measure
of control over your delivery, you need to follow five simple rules:
• Preview the materials
• Practise your presentation
• Prepare the environment
• Prepare your audience
• Present
When you begin your presentation, approach the audience with a confident air. If the presentation is
the first for you, you may feel nervous. Tips to feel at ease:
• Prepare or select your illustrations (slides or overhead transparencies) very well. Number your
illustrations and arrange them in order before the meeting. Before the meeting, load the slides into
the projector and do a trial run.
• Feel that the audience is there to learn from you and exchange information with you rather than
evaluate you.
• Have no fear that you will lose your status.
It is advisable to show enthusiasm as this will pass to the audience. You need to observe the
following tips in your delivery:
• Adopt a simple conversational style in a relaxed and confident composure. Look at the audience no
matter how shy you may feel.
• Use an effective voice that is audible, intelligible, expressive and pleasant.
• Use simple and direct language; avoid jargon.
• Speak at a reasonable rate (100 to 125 words per minute) so that your audience can follow you.
• Present in a lively form; avoid dull, monotone presentation.
• Avoid distracting mannerisms (e.g. smoothing hair, playing with pens, too much body movement etc).
• At the beginning of your presentation thank the chair and the audience for giving you the chance to
present and at the end thank them for their attention.
If unexpected events happen during your presentation (e.g. bulb of overhead projector burns), do
not panic. Always have a contingency plan for such events. The audience will sympathise with you and
probably admire you if you can continue the presentation in spite of the problem at hand.
After you deliver your presentation, allow time for discussion. A presentation without discussion
will lack interaction with the audience. Tips for your conduct during this important part of the
presentation:
• Keep calm while answering the questions.
• Do not evade questions; if you cannot answer the question or do not have the data to support your
answer, say ‘I do not know’ which will gain you the respect of the audience.
153
• Treat every question or comment seriously even if they appear elementary to you; the listener has
the right to learn.
10.4.2 Guidelines to interpersonal communication
Like all communication processes, interpersonal communication is a two-way process. To be successful
in your career you need to understand and apply positive interpersonal methods. In this section we will
deal with the following tools in interpersonal communication: verbal communication, non-verbal
communication and active listening.
Verbal communication
You need verbal communication in every aspect of your career. You need to communicate verbally with
your colleague scientists, farmers and development agents. A basic rule in verbal communication is
simplicity. Simple language is more effective with most audiences. You can use jargon with listeners who
understand it, otherwise it will distort your message. Your intention is to communicate and not intimidate.
Tips for good verbal communication habits:
• Empathy: Try to put yourself in the other person's place. Do it with an open mind.
• Respect: Show respect to your listener and this will improve exchange of information.
• Eye contact: Establish eye contact with your listeners; this shows confidence and transparency.
• Grammar: Use correct grammar and avoid slang.
• Pace or speed: Speak at a reasonable rate as you may loose the listener if you speak too fast or too
slow.
Your listeners often will hear what they want to hear. Factors that lead to this state are different
attitudes, cultural and educational backgrounds and difficulties with language.
The best way is to keep trying until you get the proper feedback. Try to understand the factors that
may hinder communication first and try to find ways to avoid the problems.
Non-verbal communication
In all cultures, people communicate with gestures, facial expression and body language. You can use non-
verbal communication effectively if you understand the cultural background of the listener.
• Positive body language: relaxed appearance, body leaning forward, arms outspread and hands at
side.
• Negative body language: tense appearance, chin in hand and hands hidden.
• Personal distance: If you happen to know the listeners you can speak to them at a distance of one
metre, e.g. person-to-person communication with a colleague or a visiting farmer.
• Social distance: If you are involved in business or formal communication, maintain a distance of
about two metres, e.g. conversation during a meeting or in group discussion.
• Public distance: During exhibitions or field days, you can maintain a longer distance while
speaking without offending your listener. Use a microphone if available.
Active listening
To understand what the speaker is saying one must actively listen. Passive listening is good in certain
circumstances, e.g. psychiatric treatments. Active listening is the active involvement of the listener
(receiving the message) with the speaker (sending the message). Active listening can:
• show the speaker that you want to hear what he wants to say
• allow you to show the speaker that your are interested in open, honest communication
• lead to effective discussion
154
• assist in recognising problems
• aid in solving problems.
To be an active listener you should avoid the following behaviour:
• When giving feedback do not threaten, preach, blame, ridicule or interrogate.
• When listening, keep attentive and do not become distracted.
10.4.3 Guidelines to visual skills
This section of the module is targeted at scientists or extension/development agents who are not
necessarily specialists in visuals. It is intended to help you grasp the concepts of visuals so that you may be
able to:
• interact with artists while they are preparing visuals for you
• prepare simple visuals for written papers, posters and slides in the absence of visual specialists.
For more details about visual skills please refer to the reading materials especially Tripathi (1991).
Role of visuals
Visuals in your scientific writing or presentation can do one or all of the following:
• demonstrate a real situation
• highlight key points
• explain trends
• summarise information
• gain and hold the attention of an audience
• save time by avoiding excessive verbal explanation (a picture is worth a thousand words).
Visual format
There are two visual formats:
• Projected format which includes overhead transparencies and slides.
• Non-projected format which includes posters, flip charts and specimens.
In the following sections we will deal with preparation of overhead transparencies and posters. For
more details on these or other formats please refer to the reading material.
Considerations in preparing visuals
The two main considerations are (i) clarity of the message and (ii) clarity of the image.
The following are tips for clear message and image visuals:
• Your message must be relevant and simple. The projected visual must show what is described. A
good, effective visual has one fact or idea and is uncluttered. Divide complex information into
chunks that are easily understood. It is better to show more visuals than to try and cram every
concept into one or a few visuals.
• The recommended maximum number of text lines is nine including title and subtitle.
• The recommended number of words and/or data is 25 to 40.
• Try to use tables of maximum 6 rows × 6 columns.
• If you are to project tables or graphs from journal articles or books, try to produce your own version
with summarised information; tables or graphs from books usually contain much information.
155
• Each visual must have a simple, short title that conveys the essence of the message and stimulates
interest.
• Your image will be more visible if there is clear contrast between content (text, illustration) and
background. One way you can achieve contrast is by using colour. You can use light colours for
images (content) and dark colours for background. For text, white lettering on a blue background or
yellow on a black background are recommended. You can use other combinations to avoid too
many colours in one visual.
• Use lower case letters for the text as they are easier to read than capital letters. You can use capital
letters at the beginning of sentences or in short titles.
• Select a readable typeface such as Times.
• The size of the letters depends on the viewing distance. You can check the legibility of a 35 mm
slide by holding it at about 35 cm from your eyes. If you can read it, then the audience should be able
to read it on the screen. To quickly check the legibility of an overhead transparency, lay it on the
floor on a white paper and if you can read standing over it, then the audience should be able to read it
from a screen.
• Sequence your visuals to flow from introduction to the body of your presentation to conclusion.
Graphic design for visuals
It is essential to understand the principles of graphic design for effective communication through visuals.
Even though you may not be a graphics specialist, it is useful to learn and use principles of graphic design.
For detailed information about graphic design consult Arntson (1988).
Nowadays it is easy to acquire, learn and use computer graphics software. The first step is to learn
from good examples which are available in publications in the library.
• Start with a plan and a format.
• Consider the equipment you will use. We recommend computer programs such as Harvard
Graphics and Freelance.
• Use simple typefaces.
• Learn to use vacant space to add to your message.
• Change the default space specifications between lines if you have different requirements.
• Keep your graphic design simple.
• Try to use illustrations if possible. You can use photos or select from the clip-art in the computer.
Beware of copyrighted art material as it may get you into legal entanglement. If you need it badly,
obtain copyright from the publisher.
10.4.4 Guidelines to writing/editing skills
This section will provide tips on writing and editing. For details refer to the reading material.
Writing tips
• Write short sentences.
• Keep the paragraphs short; editors do not like very long paragraphs.
• Use short, simple words in place of long words with the same meaning.
• Use personal words (like you and we, a person's name, direct quote etc) which will give your
writing a human touch.
156
• Action verbs grab the interest of the reader better than to be verbs. Use verbs in the active rather than
the passive voice.
When you have no access to an editor
• Check your paper or report before you submit it.
• Avoid mixed tenses. For example:
Wrong sentence: He planted the seeds and milks the goat.
Correct sentence: He planted the seeds and milked the goat.
• Avoid redundancy. For example:
Wrong sentence: The farmer was growing the same identical cultivar as his neighbour.
Correct sentence: The farmer’s cultivar was identical to his neighbour’s.
• Be careful of plural and singular verbs. For example:
Wrong sentence: A mixture of fertiliser, herbicide and inert material are not available to
the poor farmer.
Correct sentence: A mixture of fertiliser, herbicide and inert material is not available to
the poor farmer.
• Avoid careless repetition. For example:
Wrong sentence: You want to emphasise a point, so you repeat it, using different words
for emphasis.
Correct sentence: You want to emphasise a point, so you repeat it using different words.
• Be careful about mixed construction. For example:
Wrong sentence: We informed the farmer to prepare his field early, that he should plant
at the onset of the rains, and he should weed in time.
Correct sentence: We informed the farmer to prepare his field early, to plant at the onset
of the rains and to weed in time.
10.5 Writing scientific papers
We do research to solve the problems of the resource-poor farmer. However, we need to communicate the
research results to our colleagues. Communication channels to do so include newsletters, workshops and
workshop proceedings, reviews and refereed journals. This section will deal briefly with how to write
papers for refereed journals. For more details consult the reading materials section. We recommend you
consult O’Conner (1991) and Stapleton et al (1995).
10.5.1 Planning the paper
The logical time to plan your paper is at the time you planned your research. When you complete the
experiments, decide whether your results are suitable for publication. A good rule to follow: Aim to
publish a few first-class articles than many mediocre ones. Quality will do you more good than quantity.
How to know that a paper is worth publishing? A paper worth publishing meets one or all of the criteria
listed below. It:
• contributes significantly to experimental or theoretical knowledge
• contributes significantly to observational knowledge
• translates known principles into applications.
157
If your work measures up to these criteria, then proceed with writing the paper. One way to test
whether your results fulfill one of these requirements is to write down the conclusions you have reached.
Measure these conclusions against what is known. Discuss them with a colleague. This preliminary
exercise will show you whether there are gaps in your arguments and observations. When you reach this
stage and are satisfied that your work meets the criteria, find time to write your paper. Simple rules to
observe at this stage:
• Do not slice a piece of work into several short papers. Publishing several papers from the same work
is suitable only for very large investigations.
• Use your own original work to write the paper. Using other people’s work is plagiarism; this action
is fraudulent and unacceptable.
• The paper must be original and the work should not have been published in another journal.
This is another form of plagiarism. This rule does not prevent you from publishing a brief note
or preliminary communication; if you have done so, you can still submit the work as a full
paper.
Choice of journal
The following are important factors in choosing the journal:
• The scope of the journal: Does the journal publish research papers or review articles?
• What type of research papers does the journal publish? You are a specialist on small ruminants and
you want to send your manuscripts to journals that deal with your research area.
• The audience: Is your work of interest to a wide general audience in your field of research or to a few
specialists?
• Which journal is the audience you hope to reach likely to read?
• Which are the well established journals in your field?
• Which journals require authors to be members of the society that owns or sponsors the
journal?
• Which journals provide reprints?
• How often is the journal published?
• Does the journal have page charges?
Once you answer the questions, make a list of possible journals. Choose the journal that will publish
quickly and that has a style you can cope with. Once the journal is decided upon, obtain and read the
‘instructions to the author’.
Outline
An outline will help you to write the paper systematically building it step by step. It is an essential part of
the plan. The outline can be a topic outline or a sentence outline.
• Topic outline
? List points under the main headings.
? Sort the points you listed under the main headings into a logical order.
? Give the points levels using numbers. Example:
1.0 Introduction
1.1 Agro-ecological regions
1.2 Small ruminant distribution in agro-ecological regions
• Sentence outline
158
This style uses small, short sentences for the outline. Each sentence is then expanded. The example
above in the topic outline can be transformed into a sentence outline as follows:
Agro-ecological regions in Africa are five. Distribution of small ruminants in agro-ecological
regions depends on diseases, feed availability, policy issues and population pressure.
Authorship
Research on small ruminant productivity is best done by multidisciplinary teams. The contribution of the
different members of the team differs. When it comes to publishing, the question of authorship is
discussed. Sometimes it can lead to serious disagreements. The following criteria can be used to settle this
issue:
• Each author should have participated sufficiently in the work to take public responsibility for the
content.
• Credit for authorship should be based on substantial contributions to conception and design,
analysis and interpretation of data, drafting the article, reviewing the manuscript critically for
important intellectual content and giving final approval of the version to be published.
• Those who should not be included as authors are persons who:
? participated only in the acquisition of funds
? participated only in collecting data
? only supplied you with research material
? only performed general supervision of the research group.
To give due credit to such persons, you can mention them in the acknowledgment section.
All those listed as authors should give their consent. There is need to agree on the order of authors’
names. The following guidelines may be used:
• The first author is the person who did most of the work and all or most of the writing.
• Co-authors’ names are listed in an order that reflects the amount of work done by each.
• If the contribution of authors is equal, then list names alphabetically. Another solution to this
situation is to assign the place of first author to the members of the team in turn for a series of papers
that will be published from the work.
• If you are the first author, find out exactly how each co-author wants their name on the
publication.
How to organise writing among the team of authors
The team should agree on the purpose, scope and outline of the paper before writing starts. The following
are suggestions on how to handle the process:
• One person in the team should do the writing to be edited and reviewed by others.
• Another approach: Each member of the team should write a section. One person should edit the
final draft so that the paper is consistent in style.
• The team can meet to discuss the draft; these meetings should be short.
• All authors should approve the final draft of the paper before it is submitted to the journal.
There is no one style of doing this and teams are free to develop the process that will suit their
situation.
Preparing tables and figures
Readers often scan a paper quickly to assess whether they want to continue reading it or not. They most
often refer to the abstract, the tables and figures. Therefore attention should be given to these important
elements in your paper. Each table or figure must stand on its own and must contribute to the text. Decide
159
whether the data will be presented as a table or a graph. Check the instructions which the journal provides
for preparing tables and graphs before you do them. You should prepare the tables and figures before you
start writing.
Tables are used to:
• report precise numerical information
• show data for component groups
• present numbers when they are not enough to make a graph.
A table consists of title, column headings, row headings, the rows and columns containing data and
explanatory notes when needed.
Tips for designing tables:
• Keep the structure simple.
• Focus on what message you want the reader to get out of the table.
• Arrange the data from left to right and from top to bottom
• Make the width of the table suitable to fit the columns of the journal.
• Write concise titles for tables.
• Arrange the columns in each table in a way that will show the readers the conclusions you want
them to draw.
• Make the column headings brief.
• The rows and columns containing the data should show numbers to accuracy. Numbers should be to
the nearest significant figure. When the numbers are large and loaded with zeros, use appropriate
factors or units for column headings, e.g. 1,000,000 can be written as 1000 in the field cell and the
column heading read ×1000.
• Give the statistics such as probability values, standard error of mean etc. Follow the journal style.
• If there is a need, use footnotes to explain data.
Figures that are often used in papers include graphs, maps, photographs and algorithms and flow
charts. For complex illustrations and photographs you need to use the services of a professional. If your
institution lacks such services, hire freelance artists or photographers. Most computers have the software
required to prepare graphs. Therefore we will provide you with tips on how to prepare graphs. If you want
to attempt other types of figures on your own, we suggest you consult communication text books that deal
with these techniques (see the reading material).
Caution: Producing professional quality maps and photographs requires expensive equipment and time to
master the required skills.
Graphs are used to:
• illustrate trends and the relations between variables in data
• present numbers when there are too many of them.
There are different types of graphs:
• Line graphs are curves demonstrating the relationship between two variables. Line graphs are often
used for comparison over time. They show trends more easily than other types of graphs.
• Bar graphs or bar charts show data for one variable. Use bar graphs when there is no continuum (e.g.
time) between the experimental points. You can also use them when the results can be subdivided
and compared.
• Histograms show the frequency of distribution of the data points for each class of a variable (e.g.
weight).
• Pie charts show the size of component parts.
160
Tips for drawing a graph
• Try to use the services of a graphic specialist. If this service is not available to you, use computer
software such as Harvard Graphics. You need to learn the skills necessary to perform this job.
• Do not try to cram your graph with too much data but do not waste space either.
• Three to four curves in one line graph figure should be the maximum.
• When the curves cross, use different line thickness to separate them.
• Draw your curves as straight lines between data points or curves (smooth) fitted by an equation
(smooth curves do not have to pass through every point). The computer will fit the curve if
instructed to do so.
• You should place the dependent variable on the y-axis (e.g. yield) and the independent variable on
the x-axis (e.g. time or fertiliser input).
• Draft short but informative descriptions for the axes; use lower case lettering.
• Use the same symbols when the same entities occur in different graphs.
• Use the same x-axis co-ordinates for different figures if values in them are to be compared.
• Do not draw rectangles around graphs unless the journal instructs you to do so.
• If you extrapolate a curve beyond the observed point, draw the extrapolated part as a dotted
line.
• To differentiate between curves of similar types of line (e.g. solid) and thickness, use different
symbols for points joined by lines ( e.g. m, •) or similar symbols for points joined by lines of
different type or thickness (-, -). Use the symbols preferred by the journal.
• Draw a vertical line to show the standard deviation or the standard error of the mean. Make these
lines thinner than the other lines. In the legend tell the reader what these lines represent.
• Whenever possible, label curves directly rather than using a key. Position the labels horizontally.
• Label the x and y axes briefly and simply. Use the labelling to show (i) the variable and (ii) the unit
of measurement.
• Choose the scales of the axes carefully. If the axis does not start at zero, mark a break; mark
calibrations clearly.
10.5.2 Writing the first draft
After you have prepared the outline, your notes and draft tables and figures, you are ready to write the body
text. Leave wide margins and print in double spacing to allow corrections or comments to be written in the
empty spaces.
You can start writing the sections in a systematic way or start with easy sections. Most people find
the materials and methods to be the easiest place to start. Follow your outline, but it should only be used as
an aid.
A typical scientific paper consists of:
• Title
• Authors
• Postal addresses
• Abstract
• Introduction
• Materials and methods
• Results
161
• Discussion
• Acknowledgment
• References
Most journals follow this order.
The main headings are the reference points that make the structure of paper clear to the reader. The
conventional main headings (abbreviated as an acronym IMRAD) are Introduction, Methods or Materials
and Methods, Results and Discussion. Results are often combined with discussion but it is better to keep
these sections separate if you can.
Title
The title should answer the question: What is the paper about? Most people find it easier to write the title
after they write the first draft. The final version of the title must help the reader and not you. The journal
may have specific requirements for the maximum length of the title. Do not use trade names or jargon in
your title. If you write a series of papers on the same subject, you can either give each paper a separate title
or use a general title with numbered subtitles. Journals differ in their policy on this issue. Some journals
may ask you for a shorter version of your title to use as a running head (header) or as a footer.
The title of your paper should describe the main subject. It is the part of the paper that is most often
read by researchers when scanning the journals or when conducting a computerised literature search. It is
used in abstract journals, citation index journals and in data bases. Therefore it is important to give it
attention. A good title should:
• be brief (around 25 words)
• describe the contents of the papers accurately
• avoid jargon and formulae
• contain key words for indexing and retrieval.
Abstract
An abstract answers the following questions:
• Why did you start?
• What did you do?
• What did you find?
• What do your findings mean?
The abstract usually comes at the beginning of the paper unless the journal states otherwise. Even
though the content of the abstract reflects the paper, it is written to stand on its own.
• It is an abbreviated, accurate representation of the contents of your paper, without added
interpretation or criticism. It is usually not more than 250 words.
• It should be informative and should present as much of the quantitative and/or qualitative
information contained in your paper as possible. It reports the objective of the research, its scope,
the methods used, the main results and the principal conclusions.
• It uses short and simple sentences. Try to write the abstract as a single paragraph. If the abstract is
long, you can split it into two or more paragraphs. Write it after you have completed the other
sections of the paper.
Things not to do in the abstract:
• Do not include information that is not in the paper.
• Do not include tables or diagrams.
• Do not include citations.
Key words
162
Some journals may ask you to provide key words for indexing and cataloguing entries. Choose the most
important and most specific terms you can find in the paper. If you are keen that your work reaches as
many scientists as possible, write specific key words; remember that readers search in the libraries using
these key words.
Introduction
Readers will first scan the figures and tables and read the abstract. They then move to the introduction.
Therefore the introduction must convince them to continue reading.
The introduction answers the following question:
• Why did you do the work?
An introduction should be relatively short and should contain:
• Information on what is already known or in use supported by references to the most important
publications on the subject.
• The hypothesis.
• Support for the arguments to do the research.
• Experiments you intend to do to test the hypothesis.
Materials and methods
This section answers the following questions:
• What materials did you use?
• How did you use them?
• Why did you choose these methods?
What do you include in the materials and methods section?
• Describe the experimental design.
• Describe what you did and how you did it. A basic rule to follow: Give enough information for
your colleague scientists to repeat your experiment if they wish, or assess how reliable your
methods are.
• If the methods are already published in widely circulated journals, give the reference instead of
repeating all the details. If you use a new method, describe it in full. If you made changes to
published methods, describe the changes you made.
• Describe the materials used.
• Describe briefly the geographical area.
• Use generic names for chemicals. Do not use trade or local names unless there is no generic name
for them.
• Use the full taxonomic names for animals, plants and micro-organisms.
• For small ruminants, give the age, sex and genetic and physiological status.
• Describe the method in a logical order including the procedures.
• State which statistical methods you used. If you used complex statistical methods, explain
them.
• Indicate which computer program you used, if any.
Results
The results section answers the question:
• What did you find or see?
163
What do you include in the results section?
• Write the section so that it can stand on its own.
• Present the results in a logical order. The best order is the one used for the protocol.
• Reduce the massive amount of data you have collected to means along with standard errors or
standard deviations.
• Include results from controls as well as treatments.
• Include only results relevant to your argument; these can agree or disagree with your hypothesis.
Suppressing negative results is unethical. A negative result is as important as a positive one as it
may direct the reader to a new hypothesis.
• You may wish to combine the results and discussion sections. In this case results are to be presented
briefly while stating what the findings mean.
• Repeat in text only the most important findings shown in tables and figures.
• Refer to every table and figure by number in the text.
Statistics, to be used in the results section, give estimates of variability or precision of the findings.
Follow the journal’s policy when presenting statistical parameters in the results section, in tables and
figures. (See Module 8 for more details). The following are general tips:
• Use well known statistical tests.
• Report the data such that readers can assess the degree of experimental variation.
• Use standard deviation to show the variability among individuals. Use standard error of mean to
show the precision of the sample mean. State the number of measurements used to arrive at the
mean.
• If data are from a skewed distribution, give the median and the range.
Discussion
The discussion section answers the following questions:
• What do your findings mean?
• What are the implications of these findings for further study?
What do you write in the discussion section?
• Answer the specific questions you stated in the introduction.
• Show how your findings relate to existing knowledge.
• Describe what is new in your work.
• Discuss the results and hypotheses.
• Present facts and refrain from speculation.
• Focus on the main facts and avoid temptation to refer to every detail of your work.
• Use the end of the introduction as your starting point in the discussion section.
• Use past tense for results and present tense for the conclusions you draw from them.
• Conclude the section by stating the main results, conclusions and recommendations for future
work.
Reference list
All references in the text should be included in the reference list and vice versa. Follow the reference style
of the journal. Avoid secondary citation, i.e. Jones (1953), cited in Black (1994); cite only those
publications you have read yourself. Do not rely on any other author’s interpretation.
164
10.5.3 Style
The rules for grammar and style are well covered in CBE (1994) and Benjamin (1989).
Tips for style
• As often as possible use the first person (I or we).
• Try to use the active voice most of the time. You can still use the passive voice on rare occasions.
• Use the past tense for observations and results of completed actions. For example:
Ewe nutrition influenced lamb birth weight.
• Use the present tense for generalisations and conclusions. For example:
Losses are particularly high in lambs weighing 2.0 kg at birth.
• Avoid jargon.
• Be simple and concise.
• Be sure of the meaning of every word.
• Use verbs instead of abstract nouns. Abstract nouns end in -ion. For example:
Wrong: Investigation of risk factors.
Correct: The risk factors associated with these problems need to be investigated.
Wrong: No exploration of this possibility has yet been carried out.
Correct: Nobody has explored this possibility.
• Break up noun clusters and stacked modifiers. These are strings of nouns and adjectives which give
no idea of which modifies what. For example:
Wrong: Guidelines enforcement for laboratory worker safety is a management responsibility.
Correct: Management is responsible for enforcing guidelines for the safety of laboratory
workers.
10.5.4 Review by colleagues
Request your colleagues to review your paper before you submit it to the journal. The following are steps
to do so:
• Present your work as a seminar. Feedback will pin-point the strengths and weaknesses in your
arguments. Note that tables and figures for a seminar are different from those for a journal.
• Select two to three colleagues and ask them to review your paper. Give them a deadline as
sometimes it takes a long time for busy colleagues to do the job. If your institute practises the review
process routinely, follow the institute rules.
• You may send your paper to a reviewer from outside your institute.
• Incorporate the comments from the reviewers and then submit the final draft to the journal.
10.5.5 Submission to a journal
The following are steps that may help you submit your manuscript to the journal:
• Write a covering letter to the editor.
• Mail the required number of copies.
• Enquire if you hear nothing in about eight weeks.
165
• Reply to the editor appropriately when you receive one of the following answers: (i) paper accepted
(ii) paper will be considered if you make the requested change or (iii) paper rejected.
If revision is requested
If you agree that the changes requested by the editor will improve the paper, then work on the manuscript
to make the changes. Always work with your co-authors and get their agreement.
• Return the manuscript with changes incorporated.
• Send a covering letter thanking the editors and attach the list of the changes you have made. If you
disagree with one of their suggestions, explain why you disagree. Your reasons must be compelling
and clearly argued.
• If you feel the paper is good as it stands and you disagree with all their suggestions, there is no need
to waste your time. Submit the paper to another journal.
If the paper is rejected
• If the editor rejects your paper because it is outside the scope of the journal, submit the manuscript
to another journal.
• The editor may reject your paper because it is too long and needs changes. If you agree with this and
the rejection letter includes an offer to reconsider the article, make the changes and resubmit the
paper. Otherwise change the style and submit your manuscript to another journal.
• The editor may inform you that the paper is rejected because of serious flaws or lack of information.
If you agree with this assessment, put the manuscript away and obtain more information to fill in the
gaps. If you feel the editor is wrong, send the paper to another journal.
• If you are not convinced by the assessment of the editors, write a short, polite letter explaining why
you think the paper should be reconsidered.
Copyright
You can use published or unpublished information. If the information is published by others you need to
obtain written permission from whoever has the copyright to this material. If you fail to do so, your paper
may not be published by the journal or you will be legally pursued. It is your responsibility to obtain
permission.
10.6 Writing review articles
When you are advanced in your career, you may be requested or you may want to write a review article to
synthesise information in your field and share your views with your colleagues in your field of
specialisation. Review articles have a wider perspective than research papers and writing such articles will
help your own development.
10.6.1 Types of review articles
Review articles are detailed summaries of the work published on a specific topic, e.g. feed utilisation by
small ruminants. There are two types of review articles:
• Type I: Review articles that evaluate published work on the topic and provide an up-to-date
synthesis of it.
• Type II: Review articles that summarise the work without evaluating it.
When you accept to write a review article, look at the journal’s instructions to authors and at
reviews in the current issue. The journal may define the scope of the review. You must be clear on whether
you are asked to write type I or II.
166
10.6.2 Literature search
Once you have agreed with the journal on the topic, you need to conduct a literature search before you
begin to write. The following are steps to do so:
• Decide on the main headings of the review.
• Make an outline of the article and plan your reading strategies.
• Make notes on your procedures, on the criteria you will use and facts of special interest.
• Use filing cards to keep your notes or you can use a computer. Guard these cards carefully.
• Start with background publications (text books and earlier general reviews).
• Do a computer search if the facility is available. In the case of small ruminants or livestock contact
ILRI for assistance.
10.6.3 Writing the review article
• A basic rule is to keep the audience in mind.
• Write a clear introduction which can accommodate both the specialist and the non-specialist. In this
section describe the background and state the question your review will examine.
• Write the body of the review which usually focuses on a description of the literature search and ends
with conclusions and recommendations for future work. Unlike research papers, review articles do
not have standard headings such as materials and methods, results and discussion. The headings in
the body of the review depend on which aspects of the topic are being discussed. In this section
inform the reader how you searched the literature: by conventional methods, a computer search or a
combination. If the review is a synthesis and analysis of studies, explain how you assessed the
reliability of the studies covered. Explain the methods you used to analyse the literature, e.g.
statistical methods. Evaluate the selected articles objectively:
? Analyse variations in the findings critically.
? Present both sides of a controversy dispassionately and present your position.
• Write your conclusions clearly. In the conclusion section write the summary of the main findings
and make recommendations for new research.
• Write the reference list. Observe accuracy by taking details from the original publications. This
section is very valuable to researchers and mistakes may waste their time. Write the references in
the style of the journal in which you will publish the review and ensure that all references which you
cite in the text are included in the reference list and vice versa.
• If the review journal asks you to write an abstract, try to make it informative by describing briefly (i)
the purpose of the review (ii) the search method (iii) the main findings (iv) the main conclusions and
(v) the recommendations.
10.7 Posters and overhead transparencies
You may present your visual material using an overhead or a slide projector, e.g. flip charts, films, video
clips or computer multimedia or a combination of these. In this section we will limit ourselves to preparing
posters for scientific meetings and transparencies for overhead projectors.
10.7.1 Posters for scientific meetings
It is often necessary to communicate information in a display format. Posters will do this effectively.
Posters are communication tools used for publicity and advertising. They are linked to meetings, shows
and public sales. In meetings they are used to display brief information about a theme. The poster must
attract onlookers to view it and stimulate them into further action. This section will deal with posters for
scientific meetings.
167
The content of a poster
After you have decided on the topic for your poster, outline the content of your poster keeping the
audience in mind. In general, the following are the main sections of an outline of a poster for
livestock research:
• Title
• Introduction (it should be brief)
• Materials and methods used
• Results
• Conclusions
Posters for farmers or the general public will not necessarily have the same sections. They usually
present a visual message accompanied by a few words. Whether the poster is for farmers or scientists, the
title should be informative, short and meaningful.
Format and design of posters
The format and design of posters depend on:
• How much space is available? What are the vertical and horizontal lengths?
• What shape is the space?
• Are there restrictions on the size of lettering?
You must bear in mind that your audience will not have time to read too many words. So you should
convey the meaning through visual means if possible, e.g. striking pictures, use of colour (strongly
coloured cards make good background for text in white). There is no single design formula. A basic poster
design should organise the shapes following the principles of design: balance, use of space, use of text
type, dynamics of colour and layout (see Arntson 1988). If you need to break the rules for a good reason,
go ahead. In all cases, the arrangement of the design elements (title, subheadings, body text, visuals) must
be simple and pleasing to the eye. Steps for designing a poster:
• The first step is to find the information about the format before you embark on designing. Posters
can have a horizontal (landscape) or vertical (portrait) format (Figure 10.1).
168
Figure 10.1. Format determines how you design your poster.
• Draw to scale a sketch plan of the presentation.
• The plan will guide you on how many illustrations and how much text you can fit into the allotted
space.
• Make your display easy to carry especially if you have to travel to the meeting place. If you are
allotted a space of 185 cm-wide by 120-cm long, you can have pages of text and figures fixed to 8
pieces of thin card (38 × 54 cm). The cards will occupy two rows of four pieces each and leave room
for a title card. The title card should be large enough to be placed at the top left followed by your
name and address (Figure 10.2).
• Number the text and illustration sheets on the upper right side (Figure 10.2); numbered sheets help
readers to find their way quickly through information. It is a good idea also to number the sheets at
the back to help you carry them in the right sequence to the conference site. The thin cards (38 × 54
cm) will each hold an A4 sheet containing text or illustrations. Use your own variations but for best
results keep the number of different sizes of texts and illustrations to a minimum.
• Design an uncluttered display; this will help a reader who approaches the poster to stop, read,
understand and move on in a very short period.
• When you arrive at the meeting place, fix the text and illustrations to the cards. Use glue or paste that
does not dry too fast; this will allow you to adjust positions easily. You can also use double-sided
adhesive tape to fix your sheets.
• Prepare a handout for interested viewers who may want to consult the information later. Design the
handout to include the title, your name and address and a summary of the poster display. The
handout should be short and as beautiful as the poster.
Design considerations for posters:
• The major design elements are the space occupied by the title, the body of text and the visual. One
of the elements should be larger than the others to give contrast.
• Leave space around the design elements to avoid crowding. The open space area should be 30 to
40% of the poster area.
169
Figure 10.2. Layout of posters.
• Balance the elements, e.g. a large element on one side can be balanced by two smaller elements on
the other. You can deviate by creating imbalance or stress which may be desirable to draw attention
or direct the eye to certain elements of the design (Figure 10.3).
• Use horizontal lettering as vertical lettering is difficult to read. To create contrast, a large vertical
line or illustration can be effective to attract the viewers' attention to the poster.
Materials for preparing posters:
• Select the best material you can afford to produce your poster.
• Smooth hard finish paper: You use this paper if you use ink. This combination will allow you to
produce sharp finished lines.
• Heavy printing paper.
• Coverstock paper.
• Cardboard.
• Felt tipped markers for lettering.
• Commercially available letters: These include rub-down transfer letters and pressure-sensitive
paper or vinyl letters. Transfer letters are expensive and rubbing them time consuming.
Tables and graphs for posters:
• Use figures rather than tables if you can.
• Write short and informative titles for figures and tables; these titles should be the same size as the
subheadings.
• Keep legends short and place them below the figures.
• Make labels on graphs run horizontally.
• Lettering for posters:
• Produce lettering for the title and subheadings using a computer if you have the facility. If you do
not have a computer, use dry-transfer or stenciled lettering. We suggest you use bold sans serif type
170
Figure 10.3. Balance among elements in a poster.
for these parts of the poster. Make the capital letters of the title at least 4 cm high and lower case
letters at about 2.5 cm high.
• Use lower case (small) letters for the body of the text, except when it is necessary to use capital
letters (start of sentence or genus). Make capitals 8 mm high and the small letters 4 mm to facilitate
reading from 1 m away. Do not use capital letters for the body of the text because they cannot be
easily read from a distance.
• Write your name and address with letters a little larger than the body of the text.
• For subheading titles, use capital letters of 1.0 to 1.6 cm high and lower case letters of 0.6 to 1.0 cm
high.
• Line thickness for letters should be reasonably bold for good legibility. Rule of thumb: ideally line
thickness should be one-sixth the height of the letter.
Visuals for posters:
The visual (illustration, photograph) is most often an important part of the poster. Posters for extension
purposes usually contain visuals. The visual should:
• present an idea that will appeal to the audience
• be simple, compelling and to the point
• carry or reinforce the message.
Put yourself in place of the viewer and choose the most appealing and informative visual that will
communicate the message.
Displaying posters
To display your poster:
• place it where it is highly visible, i.e. at eye level or above
• make sure there is adequate lighting.
10.7.2 Preparing overhead projection transparencies
Transparencies are large, transparent acetate sheets on which information can be written and projected.
They are handy to use in presenting results during scientific meetings. The overhead projector is widely
used and is available in most meeting rooms. It can be used with normal room lights. Tips for using the
overhead projector:
• be sure that an electric outlet is available and working
• face the audience while using the projector
• switch the projector light off between transparencies to protect the audience from the glare.
Transparencies contain words or graphics (including illustrations). They are easy to prepare.
They range from inexpensive, simply made transparencies to multicoloured transparencies with several
parts (or overlays). Acetates of various quality are available to make transparencies. The equipment
used to prepare transparencies ranges from specialised machines to inexpensive thermal and
electrostatic copiers.
Preparing transparencies:
• Prepare the main points for each transparency.
• Use 10 lines or less with 7 words per line in each transparency.
• Use a legible typeface.
• Type text/data first on paper.
171
• Prepare illustrations (or obtain from clipart or other sources). Illustrations must be few.
• Design the layout of the transparency.
• Insert text/data and illustrations.
• Photocopy material onto the transparency using the equipment available to you.
10.8 Exercises
1. During your career you will be submitting papers to journals using different reference
systems. Write the steps you will take to arrange the huge number of references that will
accumulate. Indicae how you will do this using the computer and how you will do it manually.
2. When you start writing your paper, cut yourself off from the outside world. Do you agree
with this statement? Elaborate on your reaction and write your own view of how you would
go about it.
3. A newspaper editor approached you and asked you to write an article about your research for
the newspaper. How would it differ from the style and format you use to write your research
paper? Try to change this title of a research paper to a title for the newspaper article:
Effects of restricting the water intake of dairy goats upon their milk yield
10.9 References and reading materials
Arntson A.E. 1988. Graphic Design Basics. Holt, Rinehart and Winston, Inc., London, UK. 214 pp.
Benjamin B. 1989. Elementary Primer of English Grammar. Futura, London, UK.
Calvert P. 1990. The Communicator’s Handbook: Techniques and Technology. Maupin House,
Gainsville, Florida, USA. 301 pp.
CBE (Council of Biology Editors). 1994. Scientific Style and Format. The CBE Manual for Authors,
Editors, and Publishers. 6
th
edition. Cambridge University Press, Cambridge, UK.
Lindsay D. 1984. A Guide to Scientific Writing. Longman, Melbourne, Australia.
Montagnes I. 1991. Editing and Publication: A Training Manual. IRRI (International Rice Research
Institute), Manila, Philippines and IDRC (International Development Research Centre), Ottawa,
Canada.
O’Connor M. 1991. Writing Successfully in Science. Chapman and Hall, London, UK. 229 pp.
Reynolds L. and Simmonds D. 1982. Presentation of Data in Science. Martinus Nijhoff, The Hague,
The Netherlands.
Tripathi B.R.1991. Roles of Visuals in Scientific Presentations. Communication Instruction Series.
Audiotutorial Module 1. ILCA (International Livestock Centre for Africa), Addis Ababa,
Ethiopia.
Tripathi B.R.1991. Speaking at Scientific Meetings: Organizing the Message. Communication
Instruction Series. Audiotutorial Module 2. ILCA (International Livestock Centre for Africa),
Addis Ababa, Ethiopia.
Tripathi B.R. 1991. Speaking at Scientific Meetings: Oral Presentation Techniques. Communication
Instruction Series. Audiotutorial Module 3. ILCA (International Livestock Centre for Africa),
Addis Ababa, Ethiopia.
Tripathi B.R. 1991. Writing a Scientific Paper. Communication Instruction Series. Audiotutorial
Module 4. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia.
172
Reynolds L. and Simmonds D. 1981. Presentation of Data in Science: Publications, Slides, Posters,
Overhead Projections, Tape-Slides, Television—Principles and Practices for Authors and
Teachers. Martinus Nijhoff, The Hague, The Netherlands.
Stapleton P., Youdeowei A., Mukanyange J. and van Houten H. 1995. Scientific Writing for
Agricultural Research Scientists: A Training Reference Manual. WARDA (West African Rice
Development Association), Bouaké, Côte d’Ivoire and CTA (Technical Centre for Agricultural
and Rural Cooperation), Wageningen, The Netherlands. 28 pp.
Youdeowei A. and Kwarteng J. 1995. Development of Training Materials in Agriculture: A Course
Manual. WARDA (West Africa Rice Development Association), Bouaké, Côte d’Ivoire. 104
pp.
173
Module 11: Transfer of technology
11.1 Performance objectives
Module 11 is intended to enable you to:
1. Define transfer of technology.
2. List the technology transfer systems.
3. Describe the role of research in the transfer of technology.
4. Explain extension needs of pastoralists and farmers in crop–livestock systems.
5. Describe at least one case study from your region.
11.2 Introduction
As a researcher your job is to develop technologies that will increase livestock productivity. But this effort
will not have any impact unless the technologies are transferred to the farmers. Although your main
function is research, you cannot afford to ignore linking your research to technology transfer efforts. This
module outlines the mechanisms involved in linking your research to transfer of technologies to resource-
poor farmers. In most countries of sub-Saharan Africa there has been little government support for the
agricultural sector and, within this sector, small budgets have been allocated to research and extension
services. This resulted in failure to transfer new technologies. A few success stories can be cited where
extension programmes received good finance, were well structured and were targeted on high-value
commodities. An example of a successful extension effort is dairy development in Kenya.
Many factors led to the poor transfer of technologies. Among these is the separation of extension
from research programmes. As a consequence many extension agents have little information to share with
the farmers. This module will focus on research–extension linkages. Your role in the research–transfer of
technology continuum will be clarified.
175
11.1 Performance objectives
11.2 Introduction
11.3 Technology transfer systems
11.4 Transfer of technology linkages with research
11.5 Extension for integrated crop–livestock systems
11.6 Extension for pastoralist systems
11.7 Case studies
11.8 Exercises
11.9 References and reading materials
11.2.1 Key concepts
Technology transfer is a two-way flow of technical information and materials among farmers, researchers
and those who disseminate technologies. This definition extends the meaning of extension beyond
association with the traditional public sector to involve services provided by other institutions such as
non-governmental organisations (NGOs), private firms, educational institutions and producers'
associations. Our definition also implies that technologies are generated by all parties and diffused by all.
There is a continuum with no hard boundaries (Figure 11.1).
Figure 11.1. Linking research and technology transfer agencies to farmers.
The main processes of research are:
• Discovery: The process of collecting information and/or searching for relationships among
variables. This process includes basic and/or strategic research.
• Exploratory development: The process of identifying, understanding and controlling the
interaction between a proposed technology and the socio-economic and physical environments.
This process includes applied research.
• Technology consolidation: Involves transforming the results of research into a new technology
package. It includes adaptive research.
The main processes of technology transfer:
• Technology production: The process of providing materials (inputs and/or information) in
sufficient quantities to those who are responsible for delivery.
• Delivery of technologies to farmers: The technology to the farmers; this can be done through many
channels.
• Monitoring and evaluating the use of technologies: Studying the adoption rate by farmers and the
impact of the technology.
176
B52B45B53B45B41B52B43B48 B49B4eB53B54B49B54B55B54B49B4fB4eB53
B54B65B63B68B6eB6fB6cB6fB67B79 B67B65B6eB65B72B61B74B69B6fB6e
B46B41B52B4dB45B52B53 B41B53 B52B45B53B45B41B52B43B48B45B52B53
B4fB6eB2dB66B61B72B6d B74B65B73B74B69B6eB67B2c B76B65B72B69B66B69B63B61B74B69B6fB6e
B61B6eB64 B64B65B6dB6fB6eB73B74B72B61B74B69B6fB6e B6fB66 B74B65B63B68B6eB6fB6cB6fB67B69B65B73
B54B45B43B48B4eB4fB4cB4fB47B59B54B52B41B4eB53B46B45B52
B49B4eB53B54B49B54B55B54B49B4fB4eB53
B52B65B73B65B61B72B63B68 B69B6eB73B74B69B74B75B74B69B6fB6eB73
B45B78B74B65B6eB73B69B6fB6e B75B6eB69B74B73
B4cB6fB63B61B6c B6cB65B61B64B65B72B73B2fB69B6eB73B74B69B74B75B74B69B6fB6eB73
B49B6eB70B75B74 B73B75B70B70B6cB79 B61B67B65B6eB63B69B65B73
B4eB47B4fB73
B46B41B52B4dB45B52B53
B41B64B6fB70B74B69B6fB6eB2fB6eB6fB6eB2dB61B64B6fB70B74B69B6fB6e
B50B61B72B74B69B63B69B70B61B74B69B6fB6e B69B6e B72B65B73B65B61B72B63B68
B50B61B72B74B69B63B69B70B61B74B69B6fB6e
B69B6e B72B65B73B65B61B72B63B68
B44B69B66B66B75B73B69B6fB6e B6fB66 B74B65B63B68B6eB6fB6cB6fB67B79
B46B61B72B6dB65B72B2dB74B6fB2dB66B61B72B6dB65B72
B65B78B74B65B6eB73B69B6fB6e
B4dB61B69B6eB6cB79 B69B6e B6fB6eB2dB66B61B72B6d
B72B65B73B65B61B72B63B68 B61B72B65B61B73
B44B69B61B67B6eB6fB73B74B69B63
B73B75B72B76B65B79
B45B78 B70B6fB73B74
B65B76B61B6cB75B61B74B69B6fB6e
Institutions involved in research and transfer of technology are many. Many researchers may be
involved in transferring technology. In this module we will use information on research methodologies
from Module 2. On-farm research:
• applies a problem-solving approach
• identifies and describes major farming systems in which the technology is needed
• adapts technology at farm level for farmers sharing the same problem
• involves farmers participating as partners
• obtains feedback to improve the relevance of research
• involves the collaboration of research, extension and development agencies for effective transfer of
technologies.
We need to stress the point that technology transfer is used in this module in a broad sense to
encompass extension activities. Transfer of technology differs from extension in the following ways:
• It provides inputs and services such as training on use of technologies or inputs, indigenous
knowledge, credit, health services, infrastructure, entrepreneurship etc. Extension, the way it is
applied in many countries, is mostly limited to information delivery.
• Government and non-governmental institutions can transfer technologies.
• Transfer of technology is a two-way flow of information. Farmers, development agents and
researchers exchange information during the process.
Researchers should be involved in the process of technology transfer because:
• research is conducted to solve farming problems
• it is essential to translate research results into action at the farm level
• the researcher needs input from extension and development agents during research planning and
implementation and for transfer of technology activities.
Therefore there is a continuum from research to transfer of technology that necessitates the
participation of all parties: farmer, researcher and technology transfer agents (Table 11.1). Studies by
ISNAR (Kaimowitz et al 1989) confirmed the hypothesis that the traditional weakness of the links
between research and technology transfer was in part due to lack of awareness of their
interdependence.
Table 11.1. Areas of interdependence between research and technology transfer.
Areas for interdependence The need for interdependence*
Problem definition Transfer agents > researchers
– help them understand farmers’ problems
– facilitate contact with farmers
Researchers > transfer agents
– identify the scientific approaches to solve a specific problem
Experimentation Transfer agents > researchers
– identify representative location for trials
– develop realistic experimental designs
– explain objectives to farmers
Researchers > transfer agents
Table 11.1 cont...
177
Table 11.1. cont...
– establish proper procedures
– provide advice during implementation
Technology adaptation Transfer agents > researchers
– help farmers choose the right option
Researchers > transfer agents
– propose the options for modifying technology to suit local
conditions
Technology verification
– Transfer agents > researchers
– help manage on-farm trials
Researchers > transfer agents
– ensure scientific rigour
– process data from trials
Technology packaging
– Transfer agents + researchers > farmers
– inputs required by new technology are available to farmers
Researchers > transfer agents
– explain the context in which a new technology should be
applied
Provision of information Transfer agents > researchers
– produce and deliver effective messages to farmers
Researchers > transfer agents
– ensure that the messages are accurate
Provision of services Transfer agents > researchers
– provide specialised communication services to disseminate
technologies
Researchers > transfer agents
– provide services such as soil testing, pest identification and
seed certification
Provision of feedback Transfer agents >researchers
– provide the feedback of farmers about the new technology
Researchers > transfer agents
use feedback to research-identified problems
* > means contribution to.
11.3 Technology transfer systems
There is no agreement on the best way to transfer technologies in sub-Saharan Africa. The extension
systems in the region differ widely because of differences in countries, prior colonial structures and the
current interaction with international research and development agencies.
We can group the technology transfer systems practised in sub-Saharan Africa as follows:
• Conventional hierarchical approach. This system employs a large number of field staff who work
directly with the farmers and are supported by subject-matter specialists.
178
• Co-operative organisations approach. In this system the organisations provide advice, services and
inputs to the farmers.
• Profit-oriented agencies approach. In this system, the agencies provide specialist services of high
value crops or livestock to farmers. The agencies make profit selling services or inputs.
• Farming systems research and extension approach. In this system a multidisciplinary approach is
followed. The team looks at the total farm unit and technology development is a collaborative effort.
Most of the countries in sub-Saharan Africa use the conventional hierarchical approach. Some
countries, through funding from the World Bank, modified this system by introducing the training-and-
visit system. A common feature of the conventional hierarchical approach is the separation of crops and
livestock technology transfer systems. If combined, there is bias towards crops. There is now pressure
from farming communities and donor agencies to combine the different crop and livestock extension
services into a single service. To achieve any impact, this service must link farmers with researchers.
11.4 Transfer of technology linkages with research
If you conduct research to develop technologies, you need to link your research to farmers and
technology transfer groups. Researchers link to farmers through on-farm research (Figure 11.1) to
ensure that the results of research solve farmers’ priority problems. Researchers link to transfer of
technology groups through validation of technologies and pilot production projects (Figure 11.1).
The link ensures impact through a wider dissemination of technologies. The two links complement
each other and one cannot replace the other.
11.4.1 Why do we need links?
Links between researchers, farmers and technology transfer groups ensure that:
• research is focused on priority needs and problems of farmers, therefore results from experiment
stations are used to solve farmers’ problems
• farmers and technology transfer workers keep up with research developments
• technologies are tailored to suit local conditions
• relevant technologies are widely promoted and adopted
• farmers have access to the information, inputs and services required to adopt technology
• researchers have access to farmers’ knowledge.
11.4.2 Organisational issues
To link research to transfer of technology we need to develop an organisational structure. An
organisational structure can be achieved through:
1. Grouping research and technology transfer tasks
• The most important interactions should reflect the grouping. Units should be set up according to the
degree of interaction needed to meet a specific objective.
• Grouping people whose work might overlap.
• The difficulty with this approach is that no group takes responsibility for translating research results
into specifications for a new technology, packaging that technology and providing the materials
necessary for its use.
2. Merging research and technology transfer units
In this approach two or more units in the same institution have a common supervisor who is responsible
for integrating their functions and work. Merging is assumed to increase efficiency, facilitate
communication and collaboration and promote a shared goal between researchers and technology
179
transfer agents. This approach has its problems and is not as easy to implement as it may sound. It requires
the following conditions to succeed:
• Researchers and transfer of technology agents share a common goal and area of concern, e.g. to
improve productivity of livestock.
• Financial and human resources are available for linkage activities.
• The size of the new institution is manageable.
• The management of the new institution must aim at making the groups work together.
3. Co-ordination units or positions
This approach can use subject-matter specialists, research–extension liaison officers or pre-extension
services. Unlike direct supervision, these positions have no authority over the units they will co-ordinate.
Co-ordination units or positions can handle technical matters better than direct supervision as they have
the time and resources to use for integration, to develop expertise and to recognise and fill gaps. However,
this approach has many disadvantages:
• institution managers may overlook their responsibilities in the belief that the task is already under
control
• the co-ordinated groups or units may ignore liaison staff because the co-ordination positions have
neither formal authority nor support from management
• liaison staff may resent the job of only linking others who are doing the interesting
work.
This approach may therefore lead to additional barriers (new procedures, objectives, language) rather than
removing existing ones.
4. Permanent committees
These committees help promote horizontal and vertical integration. The members may be drawn from
various units or institutions. For such a committee to succeed:
• all parties consider it necessary
• members of the research–transfer of technology groups have a stake in its work and implement its
decisions or recommendations
• it is action oriented and not only a forum for discussion
• its mandate is specific
• senior management in all involved institutions support it
• meetings are well planned and chaired and follow agreed-upon procedures.
11.4.3 Types of linkage processes
In the previous section we dealt with the structural issues of linkage. This section will deal with the process
which will apply links between on-station and on-farm research as well as between research and transfer of
technology.
Planning and review processes
Planning and review meetings have been widely used to forge linkages between research and technology
transfer operations. To be successful tools, they require resources (staff time and operating funds). These
processes function through:
1. Joint problem diagnosis
Joint problem diagnosis is best done in the field. Informal diagnostic field survey proved effective in
bringing together on-station researchers, on-farm researchers, technology transfer agents and farmers.
The survey helps on-station researchers learn about farmers’ problems and appreciate the potential
benefits of the technology.
180
2. Joint priority setting
This exercise is very useful but is difficult to implement. It can be a potential source of conflict as its
expected outcome is the work agenda. One group may perceive another as trying to dictate the agenda. The
problems can be overcome by including senior managers in the exercise.
3. Joint programming and review meetings
These meetings should be kept small and should concentrate on a particular commodity (e.g. livestock)
and region. A meeting is successful when it is attended by people who can make decisions and implement
actions.
Collaborative activities
1. Collaboration in trials, surveys and dissemination
This is a very effective mechanism for linkage. Formal arrangements are needed to facilitate
collaboration between on-farm researchers and transfer of technology agents at the field level. Such
agents should be trained in skills needed for trials or surveys. Researchers need to be trained in
dissemination skills.
2. Joint decision making on recommendations
This can be a problem for linkages if not treated properly by senior management. Management needs to
develop a system that agencies can use to reach consensus on flexible mechanisms for assigning the
relevant agencies in joint decisions and recommendations. Management should also assign agencies to
implement these decisions and recommendations.
3. Field visits
Joint field visits are very effective in strengthening linkages. For this activity to be sustained managers
should allocate adequate operational funds.
4. Informal consultations and sharing of tasks
This system is a low-cost way of collaborating. It is very effective for links between on-station and on-
farm research and should supplement, but not replace, formal mechanisms. The system is successful when
staff are willing to share additional responsibilities and when institutions are flexible. It depends on
individual relationships.
Resource allocation
The resources allocated to research–transfer of technology reflect its importance to managers. The major
resources are staff time and operational funds.
1. Guidelines for allocating time
Research staff rarely allocate time to transfer of technology activities; they prefer to use their time solely
for research. Managers need to provide researchers with clear guidelines on the time to be spent on
activities for technology transfer: joint meetings, field visits etc. Managers should, however, be careful
not to affect the core professional activities of the researcher.
2. Allocation of funds
Senior management should provide operational funds to meet transport, fuel etc. Lack of funds will slow
or stop collaboration. Collaborative activities should be line items in the programme budget and the funds
should be the responsibility of those who will use them. Collaborating institutions should share in
providing the funds.
3. Staffing
The rotation or secondment of staff is an effective way of getting the necessary human resources for
linkage activities. It is essential to set a clear time frame for rotation or secondment.
181
Communication
Communication for research–transfer of technology links must be planned at the onset of activities and
funds must be provided. Commitment, skills and talents are usually needed.
1.
Reports, audio-visual material and publications
Extension workers are interested in short, easy-to-read and understand (preferably in local languages and
without scientific jargon) publications, e.g. brochures, bulletins, leaflets. Research results should be made
available to all involved in the research–technology transfer continuum. Researchers should not wait until
the results are published as it takes a long time for papers to appear in journals. In addition, extension
workers may not have access to the journals. Institutions should invest in audio-visual resources, human
resources and infrastructure. Visuals and radio messages are often more effective than printed material.
2.
Seminars and training courses
Seminars and training courses are effective means to foster links. These activities are specially effective
among groups that are new to each other or do not understand each other’s work.
11.4.4 Management of linkages
Senior management of the institutions or units involved in research–transfer of technology links need to:
• participate in research–transfer of technology activities to demonstrate their importance
• build linkage responsibilities into the job description of all involved staff
• provide guidelines to staff on how time and funds are allocated to linkage activities
• provide funds for linkage activities which are shown as budget line items
• reward participation and minimise hardships.
11.5 Extension for integrated crop–livestock systems
The population pressure in sub-Saharan Africa is driving the cropping and livestock systems towards
integration. Farmers and governments are therefore seeking to integrate the currently separate extension
services for crops and livestock. This should lead to efficient delivery of viable technologies and efficient
use of financial, physical and human resources. The integrated service should aim at improving
farmer–researcher–extension linkages. Unfortunately, many livestock departments in Africa are resisting
the process for fear of being submerged into crop departments. Nigeria took this bold step in 1991 and their
experience shows that the integration process will not be easy.
The majority of extension agents in sub-Saharan Africa have crop production backgrounds and
their educational level is low. They are therefore unable to provide advice on animal production to
farmers. The agents should be able to provide effective advice on common livestock needs such as feeding
regimes, calf feeding, clean milk production, fodder production, vaccination, deworming, dehorning,
castration, breeding management systems, marketing, resource allocation etc.
Animal production specialists working with or in the extension service are essential to provide
more advanced advice. They need to work with farmers’ training centres. As farmers become
sophisticated, the need for more specialised information becomes acute. Training can best be done by the
farmers’ organisations and newly sprouting businesses. Government services should co-ordinate and
encourage these efforts. An example of services that can be provided by private organisations is dipping.
There is a need to find new and innovative ways to provide information to farmers. Sub-Saharan Africa
can benefit from the example of information dissemination in Asia where progressive farmers have been
the most important source of information for other farmers.
182
11.6 Extension for pastoralist systems
Pastoralists differ from sedentary farmers in many respects and therefore need different transfer of
technology approaches. The management requirements for livestock differ between pastoralists and
sedentary farmers. The acceptance of new technologies also differs between the two groups. There is no
specific technique for transferring technologies to pastoralists.
Experts on rangelands should work closely with pastoral groups to improve the systems. The
experts should have the knowledge and skills to do the job; they should also have high quality
interpersonal communication skills. In addition, they need to be familiar with the communities they serve.
Knowledge of the local language is a great advantage.
Pastoralists may need information on the following aspects of livestock improvement:
• animal health
• better feeding technologies
• improved breeding programmes
• seeds for food crops and animal traction technology
• processing of animal products (milk, dried meat, hides and skins)
• improving water supplies.
The type of extension work and who should do it will vary depending on the needs and movement of
the pastoralist. Usually government services to pastoralists are limited to health services. To remedy this
situation governments should provide well trained multi-purpose subject matter specialists. To solve the
problem of movement, these specialists can meet with the pastoralists at watering points or certain
camping sites. The pastoralists should be encouraged to select certain extension agents to perform health
services and other simple extension requirements. This will relieve the burden of the subject-matter
specialists who will limit their intervention to scheduled visits.
11.7 Case studies
In this section of the module we will use case studies to illustrate the concepts of transfer of technology and
research–transfer of technology links. As the stories occurred in diverse conditions, they might not cover
all the aspects of transfer of technology or the linkage mechanisms. Still we consider them useful and hope
you use them as guidelines while pursuing your own activities.
11.7.1 Case study 1: Fodder banks in Nigeria
Background
Nigeria has 56 million small ruminants. The majority of livestock owners are semi-settlers and live in the
harsh environment found in the north of Nigeria. During the last two decades pastoralists have settled in
the subhumid zone.
The problem
• Natural pastures in the subhumid zone of Nigeria are the main source of feed for ruminants.
However, the pastures do not meet the nutrient requirements of the animals throughout the year.
• Crude protein is sufficient for only three months.
• By the end of the dry season, it is common for animals to have lost 10–20% of their body weight.
Losses associated with nutritional stress are reproductive wastage, deaths and prolonged calving
intervals.
183
• Even though feed supplementation can be used to alleviate nutritional stress, supplementation with
agro-industrial by-products is beyond the reach of smallholder farmers because of their high cost
and shortage.
• Attempts to increase productivity of the rangeland using conventional methods (bush control,
selective burning and rotational grazing) fail because the rangelands are grazed communally and
are burned accidentally.
Research
To deal with the problem of poor nutrition, researchers introduced the fodder bank concept (de Leeuw et al
1994; Saleem and von Kaufmann 1995). The researchers observed that farmers only grew forages when
there was no competition with crops for labour and when the forages were protected, even though the
forages were needed as animal feed. The researchers also observed that animals were allowed to graze on
regerminated legumes. It was therefore hypothesised that a technology that could provide fodder and
impose no competition for labour with crops could be appealing to farmers. This was an example of
researcher–farmer interaction that led to refinement of the farming system. In 1980 ILRI started
experiments on farmers’ fields undersowing the sorghum crops with stylos (Stylosanthes guinensis cv
Schofield and Cook and S. hamata cv Verano). The results of this on-farm research showed that:
• land preparation for sowing stylos was minimal
• intercropping with forage conflicted with weeding of the crop
• itinerant herds attacked the legume pastures and it was necessary to protect the legumes with
fencing.
ILRI concluded that the techniques available to integrate forage production with food crops were
inadequate, agropastoralists were ready to grow forage crops provided there was adequate protection and
there was a need to increase feed resources in arable farming systems and fallow lands.
ILRI conducted different levels of researcher-managed/executed and farmer-managed/executed
trials. Various techniques for incorporating forage legumes (mainly Verano and Cook stylo) into
compatible cereal–forage crop mixtures were assessed. The sowing date of forage legumes in cereal crops
was found to differ for each cultivar. The best crop geometry that produced the highest yield of grain and
fodder per land unit was found to be inter-row sowing of stylo with two rows of sorghum and one row of
soybean on alternate ridges.
ILRI conducted a series of trials to refine the technology:
• Fodder improvement in the fallow lands
• Nutrient and water-management trials
• Screening of legumes
• Nodulation trials
• Wet-season grazing trials.
The technology
Out of ILRI’s research the fodder bank technology emerged. A fodder bank is an ‘enclosed area of
concentrated forage legumes reserved for dry or wet season supplementary grazing’. A fodder bank can be
established by farmers as follows:
1. Fencing fodder banks
• A good site is selected (stony or waterlogged areas avoided).
• A small unit of land is fenced and sown with stylo. The size of the fodder bank (1–4 ha) depends on
the number of animals in the herd.
• Materials for fencing include wooden posts and barbed wire, metal fences and live fences using
suitable trees or shrubs.
• Fire breaks are established to prevent accidental fires during the dry season.
184
2.
Preparing land for fodder banks
Land is best prepared using tractors but this is expensive and so the technology uses alternative methods:
• kraal animals on the fenced site and burn before sowing
• undersow legumes in crops at the appropriate time, in the last year of the cropping cycle.
3. Sowing seed to establish a stylo fodder bank
• Use a seeding rate of 15–20 kg per hectare.
• Sowing is done after it has rained sufficiently. This ensures that seed-harvesting ants will not gather
seeds during dry spells. The ants can also be prevented by sowing a slurry of mixed stylo seed with
wood ash or manure.
• Stylo seeds have hard seed coats and therefore exhibit hardseededness which can be reduced by
scarification and hot water treatment to encourage better germination. This is, however, a
disadvantage if all the sown seeds germinate at the same time followed by a dry spell. It is better to
sow a mixture of softened and hard seeds.
• Mixing stylo seeds with 150 kg superphosphate per hectare before broadcasting generally gives
satisfactory establishment of a good fodder bank.
4. Weed control
Animals can graze fodder banks early in the season to selectively control the fast-growing grasses which
otherwise smother the slow-growing legumes.
5. Dry season management of the fodder bank
• Grazing begins as required during the dry season.
• The animals can graze the fodder bank either early in the morning or in the evening after returning
from free grazing.
Linkage mechanisms
On-station research by ILRI was followed by on-farm research (results highlighted above) which led to
the fodder bank technology. The technology was packaged by the researchers and National Livestock
Projects Division (NLPD) for transfer to livestock owners. Features of the linkage mechanism used to
transfer the fodder bank technology in the subhumid zone of Nigeria:
• extension staff worked closely with researchers
• collaboration was strengthened during a survey and frequent visits to the research sites and farmers’
fields
• research staff trained extension staff.
Transfer of technology
• NLPD extension staff in collaboration with researchers from the National Livestock Project and
ILRI conducted a rapid appraisal for several states. Results showed that few farmers were aware of
the fodder bank technology. To correct for this, NLPD publicised the technology through many
radio programmes.
• The World Bank provided credit to farmers between 1987 and 1991 to establish fodder
banks.
Impact and lessons
The following are highlights of adoption rates and benefits of the technology. Detailed studies and follow-
up are needed to document the large-scale impact across Nigeria if any.
185
• By 1986 about 25 fodder banks were established.
• By 1991 farmers had established 637 fodder banks under the supervised loan scheme of the World
Bank Second Livestock Development Project.
• Adoption rates were increased when credit was provided to establish the fodder banks and when
pastoralists were allocated land.
• Adoption rates were high among pastoralists who owned land or settled inside grazing reserves.
The benefits of using fodder banks are:
• they provide supplement feed during the dry season
• animals grazing fodder banks have increased growth rates, milk production and calf survival
• pastoralists are forced to sell fewer underfed stock
• fencing increases security by preventing trespassing
• confining animals inside the fenced fodder bank saves labour costs
• crops following legumes benefit from nitrogen build-up.
• the fenced fodder bank can be used for rotational cropping.
The extension package now includes rotational cropping showing beneficial interaction between
cropping and the fodder bank technology. The problems facing adoption are:
• The high inflation rate that followed the devaluation of the Nigerian currency increased the cost of
establishing fodder banks. However, many farmers still believe the benefits of fodder banks
outweigh the cost.
• Attempts to introduce fodder banks in the semi-arid zones of Nigeria were less successful than in
the subhumid zone because of lower rainfall and a shorter growing season.
11.7.2 Case study 2: Alley farming in West Africa
Background
The humid zone of Africa receives over 1500 mm of rainfall annually and extends over the coastal area of
West Africa, eastern Madagascar and part of Mozambique. In this case study we will focus on the humid
zone of Nigeria which contains 9 million small ruminants. Farmers keep small ruminants to generate cash
for special needs. Animals are kept in varying systems ranging from free range to confinement. Among the
important breeds of small ruminants in West Africa are the West African Dwarf goats and sheep which
possess a degree of tolerance of trypanosomosis.
The problem
Due to population pressure and the need for land for crops, farmers increasingly confine goats and sheep
all year round and cut-and-carry pasture to feed the animals. As tropical grasses mature, they are less
palatable as lignin content is high and crude protein is low. The productivity of goats and sheep is therefore
low due to poor nutrition. Small ruminants are infected with many diseases in the humid zone. Thus better
nutrition is hypothesised to increase tolerance of the effects of trypanosomiasis and other diseases and
thus enhances productivity.
Research
To solve the above problem, ILRI started its Humid Zone Programme (HZP) in 1978 in Ibadan, Nigeria.
This programme collaborated with the International Institute of Tropical Agriculture (IITA) which
developed the alley cropping technology. In alley cropping, the prunings from leguminous trees are
applied as mulch during the cropping season. To increase the productivity of small ruminants, ILRI
considered the possibility of using the prunings during the non-crop dry season as protein-rich supplement
to traditional feeds. ILRI called the new technology (a modification of alley cropping) alley farming
186
(Jabbar et al 1996). A research programme was pursued on-station and on-farm with the following main
components:
1. Agronomic studies to modify alley cropping for using tree leaves as mulch and fodder
• Suitable tree–grass combinations and their methods of establishment and management were
developed.
2. Animal nutrition studies to determine animal response to browse supplementation and
develop feeding strategies for utilising limited feed supply
• Leucaena leucocephala and Gliricidia sepium were found to be the most promising trees for alley
farming. L. Leucocephala contains mimosine which is toxic to animals if fed in high amounts. A
technique was developed for growing Leucaena sp and Gliricidia sp in alternate rows and offering
them to the animals in a 1:1 ratio to avoid possible toxicity.
• The optimum level of supplementation was established for young and mature animals.
Supplementation generally increased the survival of lambs and kids. Digestibility and nitrogen
balance studies showed that utilisation of fresh or dry Leucaena was similar. Any surplus fresh
material from wet seasons can be dried and stored for use in dry seasons. Intake of G. sepium was
higher in the dry form than in the fresh form.
3. On-farm studies were conducted
• On-station results on biological performance were validated.
4. Socio-economic studies to assess the benefits of feed supplementation and identify factors
related to adoption potential
• Including livestock in the farming systems increased return by 13 to 16%.
• Including small ruminants in the alley farming system increased its sustainability.
The technology
1. Technology 1: Fodder trees in alley farming systems
• Grow L. leucocephala and G. sepium in alternate rows. Both trees can be established by direct
seeding in association with crops. Establishment can be improved by applying phosphorous and
nitrogen and inoculating with effective strains of Rhizobium.
• Grow crops in the alleys formed by the two leguminous trees. If the farmer is growing maize, four
rows of the crop are recommended.
• During the cropping season, the trees are pruned and the prunings are applied as mulch to improve
fertility and control erosion. Mulching done during the later part of the cropping season is not
beneficial to the crop.
Pruning is done during the later part of the cropping season and during the remaining part of the dry
season to provide the animals with a protein rich feed supplement to the traditional diet. Supplementation
is given at about 600 g dry matter per day for young animals and 800–900 g dry matter per day for adults.
2. Technology 2: Intensive feed gardens
An alternative technology to resolve the competition between trees and crops is planting small feed
gardens with trees-only or tree–grass combinations.
The trees are sown at 50 cm spacing in trees-only small feed gardens. The trees are pruned at 12-
week intervals.
The trees are established 2.5 m apart with 2 rows of grasses. Recommended grasses are Brachiaria
mutica and Panicum maximum.
187
Linkage mechanisms
In this case study the research–transfer of technology involved researchers from ILRI, IITA (through
AFNETA) and a number of education and research institutes in Nigeria. Extension and development
organisations were partners in on-farm research. Farmers participated in on-farm research and other
activities needed to transfer technology. An extension worker from the Ministry of Agriculture was
trained by ILRI and was placed with ILRI technical staff in the villages where the research and transfer of
technology were done. Other extension staff participated on an irregular basis.
Extension agents and researchers collaborated in the selection of sites for on-farm experiments.
The combination of transfer of technology (semi-extension developmental approach) and research
(on-farm experimental approach) resulted in two advantages:
• Farmers learned the technology package through hands-on experience and so they owned it.
• Time from on-station experiments to adoption was reduced.
Transfer of technology
Many farmers neighbouring the on-farm experiments decided to adopt the technology and planted alley
farms. Attracted by the interest of farmers in and around the research villages, the National Livestock
Projects Division (NLPD), a World Bank-funded federal government agency in Nigeria, showed interest
in the alley farming technology package.
In 1989 NLPD established pilot projects in several states in southern Nigeria. The projects included
establishing plant nurseries in selected villages and distributing tree seedlings to farmers. NLPD provided
additional services: a package of animal health care, credit and advice. Staff from NLPD received training
from ILRI scientists on how to establish and maintain the alley farms. The NLPD staff then passed this
information on to the farmers.
In 1991 FAO launched a two-year pilot project in Ghana, Togo and Nigeria to disseminate alley
farming through national extension and development agencies. The mechanism:
• FAO assessed the alley farming situation in those countries in consultation with extension services,
development agencies and universities.
• Training was provided for both extension agents and farmers.
Impact
1. The benefits of alley farming were demonstrated.
2. Initial adoption by farmers was low and remained around the on-farm research areas. This
was because:
• Alley farming is a complex technology. Farmers need to invest time to learn the technology,
especially establishment and pruning.
• Even though the technology was profitable, it needed labour input for pruning at a time when it
competed with weeding the crop. Even though the total labour requirement was the same between
alley farming and non-alley farming systems, this labour competition partly contributed to the slow
rate of adoption.
3. The transfer of technology mechanism of NLPD and the FAO project did not function well
in spite of special effort. However, their input is yet to be evaluated.
188
11.7.3 Case study 3: Crossbred goats for milk production in Burundi
Background
The Burundi highland plateau is part of the Great Lakes highlands. The high population pressure has
resulted in shorter fallow periods and cropping on steep land and in the swamps. This evolution pushed
cattle out of the system and farmers replaced them with goats.
The problem
There is a need to improve animal productivity to provide for household nutrition and cash. Farmers
needed dual-purpose goats to produce milk and meat.
Research and technology development
A national network on small ruminants was created in 1987 to link research and development projects in
Burundi. The national university, Facagro, co-ordinated the network. At the university research was
conducted on:
• crossbred goats
• characterisation of goats
• utilisation of Leucaena, Panicum maximum and Pennisetum purpureum and rice straw by goats.
A package of improved does, improved feed resources and disease control was tested on 21 farms in
Mutaho, Burundi. The goat development project in Ngozi started in 1980 with funding from German
Agency for Technical Co-operation (GTZ). The project targeted 3500 smallholders. The project staff
crossbred the local goats with Alpine bucks. Research found that this crossbred increased milk and body
size. The project established buck centres where farmers can bring their does for mating. During 1988,
1989 and 1990, more than 8000 matings were performed. ILRI scientists in Burundi joined the
research–transfer of technology effort to determine how the livestock system fitted in the farming system
and to identify opportunities for increasing the productivity of livestock in the system (Rey 1993). A total
of 240 farmers were selected, 120 of whom brought their does for mating. The researcher observed that:
• farmers who adopted crossbred dairy goats had farms
• farmers established fodder crops before adopting crossbred goats
• there was a decrease in off-farm employment
• many farmers back-crossed the Alpine–Small East African does with Small East African bucks to
limit the negative effect of a high level of Alpine genes.
Linkage mechanisms
Collaborative research was developed between the Ngozi goat project and ILRI. The thrust of the research
was technology transfer. The Ngozi goat project staff collected day-to-day data while ILRI staff visited
the project to follow-up on data collection. The university facilitated the participation of graduate
students.
Transfer of technology
The project established an extension unit to encourage farmers to participate. In addition to cross-
breeding, the unit provided farmers with information on improved housing, management, fodder crops
and disease control. To commercialise milk production, the project built a factory to process cheese and
organised marketing of the milk.
Impact
The Ngozi project operated only for a limited period. When the project terminated, the breeding of
crossbred goats stopped in 1991. Therefore it was not possible to assess adoption rates or impact. This is
189
an example of a project which is not sustainable as it depends exclusively on external funding. When the
donor stopped funding, the projected activities ceased.
11.7.4 Case study 4: Dual-purpose goats in Kenya
Background
Due to the rapid increase in human population in Kenya, the demand for livestock products is increasing.
The Kenyan government set a goal for improving the diet and standard of living of rural people through
enhancing the productivity of goats. The Kenya Agricultural Research Institute (KARI) established the
Small Ruminant Collaborative Research Support Programme (SR-CRSP) in 1979. The programme is
part of a collaborative effort involving research institutions in Bolivia, Brazil, Indonesia, Kenya,
Morocco, Peru and the United States of America. It is focused on production of milk, meat and manure
from goats on small farms.
The problem
The population pressure in western Kenya has increased rapidly causing a decrease in farm size. The
prevailing agricultural system could no longer support the needs of smallholder farms for food and cash.
SR-CRSP diagnosed the farming situation in western Kenya and formulated the following hypotheses:
• Dual-purpose goats can convert unused crop residues and browse to milk and meat.
• Keeping three to five does instead of a cow can add a small but year-round supplement of milk
protein to family diets.
• A typical doe can produce three kids in two years, thus increasing marketing prospects and meat for
family consumption.
• The technology to promote dual-purpose goats must be low-cost, low-risk and complementary to
cropping activities.
Research
The Small Ruminant Collaborative Research Support Programme (SR-CRSP) was implemented in three
phases:
• Phase 1 1980–82: The programme characterised social, economic and biological activities of
production systems.
• Phase 2 1983–85: The programme conducted on-station research on goat breeding, health,
nutrition and agronomy. It also conducted on-farm research on agronomy, goat nutrition and health
management, and research on preliminary cost–benefit and social feasibility studies.
• Phase 3 1986–91: The programme performed large-scale economic and social evaluation of dual-
purpose goats’ performance in production systems.
SR-CRSP emphasised the farming systems approach to ensure that the research was relevant to
farmers’ needs.
The technology
Technology packages were ready for expanded testing during 1988 (Semenye and Hutchcroft 1992). The
technology package was:
• Two-way and four-way cross goats composed of equal parts of two exotic breeds, Toggenburg and
Anglo-Nubian, and two indigenous breeds, East African and Galla.
• The goats were dual-purpose to produce milk and meat.
• The goats were fed crop residues and agricultural by-products. This base ration was supplemented
with foliage from trees (Leucaena or Sesbania), Napier grass or sweet potatoes vines.
• The goats were confined and the forages were cut and carried to the unit.
190
Linkage mechanisms and transfer of technology
The principal institutions involved were SR-CRSP of KARI and various units in the Ministry of
Agriculture. SR-CRSP also worked with scientists from Egerton University and the University of Eastern
Africa at Baraton.
• Linkage with the universities was done through graduate studies. Thirty-seven students earned
their masters or PhD degrees. They conducted research relevant to the programme.
• Extension field assistants were trained by SR-CRSP staff on the new technology packages.
• On-farm demonstrations and group meetings were held. They led to interactions among extension
field assistants, farmers and scientists.
• SR-CRSP used a memorandum of understanding between the programme and collaborating farmers.
This detailed the rights and duties of the farmers; it placed ownership of the goat with the farmer.
• SR-CRSP exhibited its technology packages in local agricultural shows. These provided feedback
on how the public perceived the technology packages.
• SR-CRSP conducted on-farm research in Emuhaya, Hamisi and Vihiga in Kakamega District,
Maseno in Kisumu District and Boro in Siaya District.
• During on-farm testing participating farmers were given one doe to start a dual-purpose goat flock.
During the large-scale testing, new farmers were given two does in 1988.
Impact
Socio-economic monitoring showed that the technology packages were adopted enthusiastically by
farmers in the relatively higher elevation, higher rainfall and more densely populated villages. Farmers in
less densely populated villages at low elevation and with less rainfall were not highly receptive of the
technology. This was due to the bias in research to high elevation areas and the fact that farmers in low
elevations had larger farms and could afford cows.
11.8 Exercises
1. Transfer of technology can link with research through formal or informal mechanisms. List
mechanisms linking research to transfer of technology in your region. Are the linkages effective?
If not, explain why.
2. Agricultural technology system (ATS) consists of all the individuals, groups, organisations and
institutions engaged in developing and delivering new or existing technology. The involvement
of the various parties listed in this definition requires various types of skills. List these skills and
the basic activities you think the groups should be engaged in.
11.9 References and reading materials
Claar J.B. and Watts L.H. (eds). 1983. Knowledge Transfer in Developing Countries: Status,
Constraints, Outlook. Proceedings of a Conference on International Extension at Steamboat
Springs, Colorado, July 1983. Colorado State University, Fort Collins, USA. 59 pp.
Jabbar M.A., Larbi A. and Reynolds L. 1996. Alley Farming for Improving Small Ruminant
Productivity in West Africa: ILRI’s Experiences. Socio-economic and Policy Research Working
Paper 20. ILRI (International Livestock Research Institute), Addis Ababa, Ethiopia. 96 pp.
Kaimowitz D., Snyder M. and Engel P. 1989. A Conceptual Framework for Studying Links between
Agricultural Research and Technology Transfer in Developing Countries. Linkages Theme
Paper 1. ISNAR (International Service for National Agricultural Research), the Hague, The
Netherlands. 30 pp.
191
de Leeuw P.N., Mohamed-Saleem M.A. and Nyamu A.M. 1994. Stylosanthes as a Forage and
Fallow Crop. Proceedings of the Regional Workshop on the Use of Stylosanthes in West Africa,
Kaduna, Nigeria, 26–31 October 1992. ILCA (International Livestock Centre for Africa), Addis
Ababa, Ethiopia. 340 pp.
Merill-Sands D., Kaimowitz D., Sayce K. and Chater S. 1991. The Technology Triangle: Linking
Farmers, Technology Transfer Agents and Agricultural Researchers. Summary Report of an
International Workshop held at ISNAR, The Hague, 20–25 November 1989. ISNAR
(International Service for National Agricultural Research), The Hague, The Netherlands. 118
pp.
Rey B. 1993. Contribution of crossbred goats to milk production and social welfare in Burundi. In:
Kategile J.A. and Mubi S. (eds), Future Livestock Industries in East and Southern Africa.
Proceedings of the Workshop held at Kadoma Ranch Hotel, Zimbabwe, 20–30 July 1992. ILRI
(International Livestock Research Institute), Addis Ababa, Ethiopia. pp. 69–72.
Saleem M.A. and von Kaufmann R. 1995. Fodder bank: Improving the nutrition of cattle in the
subhumid zone of West Africa. In: Ruiz M.E., Sere C. and Li Pun H. (eds), Animal Production
Systems Global Workshop. Proceedings of the Global Workshop on Animal Production Systems
held in San José, Costa Rica, September 15–20 1991. IICA (Inter-American Institute for
Cooperation on Agriculture), San José, Costa Rica.
Semenye P.P. and Hutchcroft T. 1992. On-farm Research and Technology for Dual-purpose Goats.
National Printing Press Ltd., Kisumu, Kenya. 144 pp.
192