Animal traction in eastern Upper Volta

Material Information

Animal traction in eastern Upper Volta a technical, economic and institutional analysis
Series Title:
MSU international development papers
Barrett, Vincent
Michigan State University -- Dept. of Agricultural Economics
United States -- Agency for International Development
Place of Publication:
East lansing MI
Dept. of Agricultural Economics, Michigan State University
Publication Date:
Physical Description:
xiii, 118 p. : ill., map ; 27 cm.


Subjects / Keywords:
Draft animals -- Burkina Faso ( lcsh )
Agricultural implements -- Burkina Faso ( lcsh )
Agricultural innovations -- Burkina Faso ( lcsh )
Agriculture -- Economic aspects -- Burkina Faso ( lcsh )
Rural development -- Burkina Faso ( lcsh )
Technical assistance, American -- Burkina Faso ( lcsh )
bibliography ( marcgt )
non-fiction ( marcgt )
Spatial Coverage:
Burkina Faso


Includes bibliographical references (p. 116-118).
General Note:
Cover title.
General Note:
Contract AID/afr-C-1314 between the U.S. Agency for International Development and the Dept. of Agricultural Economics, Michigan State University.
Statement of Responsibility:
by Vincent Barrett ... et al..

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact Digital Services ( with any additional information they can provide.
Resource Identifier:
08948683 ( OCLC )
0731-3438 ; ( ISSN )

Full Text
A Tehial Economican
VicetBaret Grgr Lasie, ilok
198 Eas Lasn, Mihia 48824

MSU INTERNATIONAL DEVELOPMENT PAPERS Eric W. Crawford, Carl K. Eicher and Carl Liedhoim, Co-Editors
The MSU International Development Paper series is designed to further the comparative analysis of international development activities in Africa, Latin America, Asia and the Near East. The papers report research findings on historical, as well as contemporary, international development programs. The series includes papers on a wide range of topics, such as alternative rural development strategies; non-farm employment and smallscale industry; housing and construction; farming and marketing systems; food and nutrition policy analysis; economics of rice production in West Africa; technological change, employment, and income distribution; computer techniques for farm and marketing surveys; and farming systems research. While the papers mainly convey the research findings of MSU faculty and visiting scholars, a few papers will be published by researchers and policymakers working together with MSU scholars on research and action programs in the field.
The papers are aimed at teachers, researchers, policy-makers, donor agencies, and international development practitioners. Selected papers will be translated into French, Spanish, or Arabic. A list of available papers and their prices may be obtained from:
MSU International Development Papers Department of Agricultural Economics
Agriculture Hall
Michigan State University
East Lansing, Michigan 48824-1039
Individuals and institutions in Third World countries may receive single copies free of charge.

Vincent Barrett,' Gregory Lassiter,- David Wilcock,-/ Doyle Baker,/ and Eric Crawford6/
1/This paper was prepared under Contract AID/afr-C-1314 and published
under Grant AID/afr-0929-G-SS-2011-00 between the U.S. Agency for International Development and the Department of Agricultural Economics, Michigan State University, East Lansing, Michigan.
/Consultant and former member of the MSU Research Team in the Eastern Region of Upper Volta.
/Associate Professor, Department of Animal Science, Agricultural Business, Education, and Mechanics, California State University, Chico, California.
/Extension Economist, Department of Food and Resource Economics, University of Massachusetts, Amherst, Massachusetts.
Z/Graduate Research Assistant, Department of Agricultural Economics, Michigan State University, East Lansing, Michigan.
/Assistant Professor, Department of Agricultural Economics, Michigan State University, East Lansing, Michigan.
MSU is an Affirmative Action/Equal Opportunity Institution

SSN 0731-3438
0 All rights reserved by Michigan State University, 1982.
Michigan State University agrees to and does hereby grant to the United States Government a royalty-free, nonexclusive and irrevocable license throughout the World to use, duplicate, disclose, or dispose of this publication in any manner nd for any purpose and to permit others to do so.
ublished by the Department of Agricultural Economics, Michigan State niversity, East Lansing, Michigan 48824, U.S.A.

List of Table~s
List of Figures
Section 1 -- INTRODUCTION 1
1.1 Background 1
1.2 Objectives of the Report 2
1.3 Supporting Data Used in This Report 2
2.1 Overview of the Hoe Production System 6
2.2 The Farming System with Animal Traction 6
2.2.1 Managerial Requirements of Large Animals 7
2.2.2 Use of New Implements and Agronomic Techniques 7
2.2.3 Intensification of Land Use and
Maintenance of Soil Fertility 9
2.2.4 Changes in Crop Mixtures 9
2.2.5 Supporting Services Required for a
Successful ANTRAC Program 9
2.3 Summary Characteristics of Hoe and ANTRAC Households 9
3.1 Previous Programs 13
3.2 The 1974-80 Eastern ORD Program 16
3.2.1 The Eastern ORD's Approach to the
Extension of Animal Traction 18
3.2.2 The Credit System 19
3.2.3 Marketing 24
3.2.4 Animal Insurance 24
3.2.5 Types of Draft Animals 25
3.2.6 Animal Trai ni ng 25
3.2.7 Livestock Extension 26 Dry Season Feeding 26 Fattening of Old Oxen for
Resale as Meat Animals 28
3.2.8 Equipment 28
3.2.9 The Supply of Spare Parts* 31
3.2.10 Conclusion 31

4.1 Overview of the Traction Sample 33
4.2 Purchase, Sale, and Training of Draft Animals 33
4.2.1 Animal Purchases 33
4.2.2 Castration and Placement of Nose Rings 35
4.2.3 Animal Training 35
4.2.4 Sale of Draft Animals 36
4.3 Animal Maintenance 37
4.3.1 Stabling and Feeding 37
4.3.2 Cost of Feeding Animals 37
4.3.3 Animal Health 39
4.3.4 Treatment of Diseases and Mortality Rates 40
4.3.5 Animal Insurance 43
4.4 Use of Animal Traction for Tillage 43
4.4.1 Ownership and Use of Tillage Equipment 43
4.4.2 Performance of Tillage Operations 45
4.4.3 Equipment Repairs 51
4.5 Use of Animal Traction for Transport 51
4.6 Estimated Production Effects of Animal Traction 54
4.6.1 Area Effects of Animal Traction 54
4.6.2 Changes in Cropping Emphasis 57
4.6.3 Yield Effects 59
4.7 Impact of Animal Traction on Household Labor Allocation 63
4.7.1 Allocation of Household Labor to Cropping Activities 64
4.7.2 Proportional Allocation of Household Labor
to Farm, Nonfarm, and Leisure Activities 68
5.1 Introduction 72
5.2 Farm Household Income 74
5.3 Cash Flow 80
5.3.1 Annual Cash Flow 80
5.3.2 Monthly Cash Flow 83
5.4 Medium-Term Income Analysis 86
5.4.1 Net Benefits Assuming No Production Increase 88
5.4.2 Area and Yield Effect from Plowing 93

5.4.3 Plowing and Weeding 95
5.4.4 Plowing, Weeding, and Phosphates 98
5.4.5 Summary 99
6.1 Overview of Findings 103
6.2 A General Strategy for Improving the Adoption and
Effectiveness of Animal Traction Technology 109
6.3 Recommendations for Improving Animal Traction Programs 109
6,3.1 Developing a Viable Locally Adapted Technical Package 109
6.3.2 Extension and Farmer Training 110
6.3.3 Financing ill
6.3.4 Marketing .113
6.3.5 ANTRAC Equipment Services 114
6.3.6 Livestock and Veterinary Services 114
6.4 Conclusions 115
Bibliography 116

Table Page
1.1 Distribution of the 480 Farm Households Surveyed
in 1978-79 by Zone, Village, and Sub-Sample 4
2.1 Summary Characteristics of Farm Households
in the Oxen and Donkey Zones, 1978-79 11
3.1 Number of Animal/Equivalent Units Distributed
by the EORD Up to March 31, 1980 17
3.2 Eastern ORD Medium-Term Credit Repayment Performance:
Collection Ratios and Percentage of Portfolio
in Arrears for 1976-1980 22
3.3 Forage Species Showing Potential in Eastern Upper Volta 27
4.1 Number of Traction Units Owned by Animal Traction
Farmers, 1978-79 Season 34
4.2 Percentage of Animals Trained by Different People 36
4.3 Average Cash Expenses for Maintaining Traction Animals,
by Quarter, 1978-79 Crop Season 38
4.4 Distribution of Annual Cash Expenses for Animal Feeding 38
4.5 Incidence of Health Problems and Accidents and
Percentage of Potential Draft Animal Work Time Lost 41
4.6 Animal Sickness: Amount of Lost Work Time and
Percent of Draft Animals Affected 42
4.7 Type of Tillage Equipment Owned by Oxen and Donkey
Farmers by Years of Experience with ANTRAC 44
4.8 Number and Percentage of Farmers
Who Owned and Used Equipment 46
4.9 Relationship Between.Experience with Animal Traction
and the Percentage of Farmers Who Performed Different
Tillage Practices 47
4.10 Distribution of Annual Fieldwork Time for Donkeys
and Pairs of Oxen, 1978-79 49
4.11 Average Area and Proportion of Area Where
Tillage Operations Were Performed 50
4.12 Number of Carts Owned and Average Time Worked Per Cart 51

Tible Pg
4.13 Total Time Carts Were Used and
Percentage of Time Rented Out 52
4. 4 Effect of Carting on Utilization Rates of
Animal Units in 1978-79 53
4.45 Household Size and Area Cultivated
by Zone and Use of ANTRAC 55
4.16 Percentage of Area Planted to Different Crops
by Zone, 1978-79 58
4.17 Yields for Major Crops Under Hoe, Oxen, and
Donkey Cultivation in ANTRAC Study Zones, 1978-79 60
4.18 Results of Natural Phosphate Fertilizer and
Plowing Trials Conducted in the Eastern Region in 1979 62
4. 9 Average Worker Equivalent Hours of Family Labor Used
Per Hectare for Major Field Activities on All Crops
40 in ANTRAC Zones 65
4.0 Allocation of Potential Labor Time to Leisure,
S Farm, and Nonfarm Activities 69
5. Total Value of Crop Production, 1978-79 73
5. Farm Household Annual Income Statement, 1978-79 75
53 Summary Farm Household Income Characteristics 79
5 Annual Cash Flow Statement, 1978-79 81
5. 5 Monthly Cash Flow Statement for Hoe Households 84
5.6 Monthly Cash Flow Statement for ANTRAC Households 85
5.7 Assumed Increases in Area, Yield, and Total Value of
Production for Oxen and Donkey Cultivation Under
Three Levels of Technology Adoption 89
5.8 Donkey Plowing, No Area or Yield Effect 90
5.9 Oxen Plowing, No Area or Yield Effect 92
5 10 Donkey Plowing with Area and Yield Increase 94
5 11 Oxen Plowing with Area and Yield Increase 94
5 12 Donkey Plowing and Weeding 96

Table Page
5.13 Oxen Plowing and Weeding 97
5.14 Donkey Traction: Plowing, Weeding, and
Phosphate Fertilizer 100
5.15 Oxen Traction: Plowing, Weeding, and
Phosphate Fertilizer 100
5.16 Internal Rate of Return to Selected Oxen
and Donkey Traction Investment Cases 101

Figure Page
1.1 Map of Eastern ORD Survey Area 5
3.1 ARCOMA/COREMMA Til lage Equipment 30
4.1 Allocation of Man-Hours Worked on Cropping Activities
by Period and Major Activity 67
4.2 Seasonal Indices of Man-Hours Worked on
Household Fields, Non-Cropping Farm Activities,
and Non-Farm Activities 71

The authors would like to thank the following persons for providing editorial comments during various stages in the writing of this report: Chris Delgado, Carl Eicher, Gordon Gemmill, Bill Morris, Gordon Potts, and Tom Whitney. The authors take responsibility for errors or omissions herein.

1.1. Background
In the 1970s, there was a dramatic increase in development assistance to the West African Sahel, a contiguous belt of countries including Mauritania, Senegal, Mali, Upper Volta, Niger, Chad, Cape Verde Islands, and The Gambia. This increased assistance came largely in response to the severe 1969-1973 drought. Following the drought, Sahelian states, with the assistance of donors, launched a number of projects to increase food production and achieve regional self-sufficiency.
In 1975, the U.S. Agency for International Development (USAID) and several other donors helped the Government of Upper Volta launch a medium-term recovery project in the Eastern Region of Upper Volta, an area covering about 50,000 km2 with a 1979 population of about 440,000 (Mehretu and Wilcock, 1979). The Integrated Rural Development Project had two major objectives.- One was to help develop the institutional capacity of the regional rural development authority, the "Organisme Regional de Developpement de l'Est" (Eastern ORD or EORD). The second was to increase agricultural production and rural incomes in the region through the introduction of animal traction (ANTRAC) cultivation techniques, supported by EORD technical, extension, credit, and marketing services. The use of animal draft power was expected to improve farm productivity by alleviating labor constraints and by integrating crop and animal production. The project's
strategy of developing the operational capacity of the Eastern ORD (EORD) and of introducing donkey and oxen cultivation technology followed a pattern used in other ORDs in Upper Volta.-/
The technical assistance component of the USAID project was provided through a contract with the Department of Agricultural Economics at Michigan State University (MSU) beginning in May, 1977. The MSU team provided technical assistance to the Eastern ORD, and also carried out farm production, marketing, and regional planning surveys over a three-year period, 1978-1980. During the
project, the number of trained personnel employed by the EORD increased six-fold to over 400 persons throughout the region. The number of farmers using animal
traction (ANTRAC) increased dramatically over the project period, from 180 in 1974 to approximately 1,740 in 1979/80.
lSee USAID (1974) for a description of the IRD Project.
YSee Eicher et al. (1976) for an early review of the IRD Project implementation strategy, and Mehretu and Wilcock (1979) for an introduction to the Eastern Region.

1.2. Objectives of the Report
This report analyzes the technical, economic, and institutional impact of the animal traction program, the major component of USAID assistance to the EORD Integrated Rural Development Project. The use of animal traction has been popular in West Africa since the 1930s, owing partly to apparently favorable experiment station research results, and partly to cases of successful adoption, e.g., in Senegal and Mali. However, as Sargent et al. (1981) found in a review of 125 projects involving ANTRAC in francophone West Africa, there is very little e idence on two questions: (1) the performance of ANTRAC under actual farmer c nditions; and (2) the effect of ANTRAC independent of other elements of the t chnical package such as improved seeds, fertilizer, etc.
The specific objectives of this paper are to: (a) describe the 1975-81 EORD animal traction program with emphasis on its institutional features and h storical context; (b) evaluate the farm level technical and economic impact of t hIe ANTRAC program on the basis of empirical studies carried out in 1978 and 1779; and (c) recommend changes in content and approach which may improve future ANTRAC programs in the Eastern Region of Upper Volta and elsewhere in West A rica.
1.3. Supporting Data Used in This Report Most empirical findings presented in this report are based on analysis of data from the 1978-79 farm survey conducted by the Bureau of Economic Analysis and Planning of the Eastern ORD in cooperation with the MSU contract team.1-/During the period of May 1, 1978 to April 30, 1979, the economic activities of 480 farm households selected from 27 villages were monitored.- The sample is stratified across 12 zones that were purposively selected in order to represent the broad agro-climatic variation found in the EORD. Within each zone, a sample of "traditional" agricultural households (those using currently available agriculural methods based almost entirely on hand hoe cultivation techniques) was randomly selected. In addition, a purposive sample of the relatively most successful animal traction farm households, as identified by local extension personnel, was selected in five zones in order to represent the "performance
1/This survey was carried out with support from contract AID/afc-C-1314 etween USAID and Michigan State University.
.-2For additional details on the objectives, structure, and methodology of
he survey, see MSU Contract Team, "Six-Month Report: December 1977-May 1978," p. 31-54.

frontier" or potential of this new technology.- After attrition, the sample
consisted of 355 hoe farming households and 125 ANTRAC households. The distribution of these 480 households across the 27 villages and 12 agro-climatic zones of the EORD is presented in Table 1.1. The analysis in this report focuses on the five ANTRAC zones--Piela, Diabo, Logobou, Diapangou, and Ougarou--whose location
is indicated in Figure 1.1 by the circled capital letters "TA." Sample villages with no animal traction are indicated by triangles.
Farm families were interviewed on a wide range of farm, off-farm, and household activities. The survey employed the "cost route" method of data collection, based on recurrent weekly or monthly interviews to obtain information on household resource allocation. Labor use in all farm field activities was obtained through weekly interviews with one-third of both hoe and ANTRAC households.
Other information used in the report comes from forage, plowing, and fertilizer trials conducted by the livestock specialist. Observations and informal interviews conducted by all team members provide additional background for the analysis.
IBecause of the recency and geographical dispersion of the program, the majority of EORD ANTRAC users in 1978 were recent adopters who had hardly begun to use their ANTRAC equipment or experience any benefits from it. For this
reason, a non-random sample was used to carry out a "most favorable case" evaluation of ANTRAC in order to provide an indication of ANTRAC potential under EORD conditions.

Number of Sampled Households by Sub-Sample
Traditional Traction
Agr climatic Zone Village (TRAD) (ANTRAC)
1. B gande 1. Balemba 18
2. Komboassi 18
2. Mani 3. Lanyabidi 18*
4. Bombonyenga 18
3. Piela 5. Dabesma 18
6. Piela (ANTRAC) 18
4. Diabo 7. Mocontore 18
8. Lantaogo (ANTRAC) 18
26. Diabo I (ANTRAC) 17
27. Diabo II (ANTRAC) 18
5. ogobou 9. Namponkore 18*
10. Kindi Kombou 18*
11. Logobou (ANTRAC) 18
6. artiaga 12. Bomondi 18*
13. Dupcaali 18
7. onde 14. Ouobgo 17
15. Kondogo 18*
8. iapangou 16. Tilonti 18
17. Diapangou (ANTRAC) 18
9. Sotou (N. de Fada) 18. Botou (N. de Fada) 18*
19. Ougarou (N. de Fada) 19*
10. ,antchari 20. Mantchangou 17
21. Mohadagou 18
11. ugarou 22. Poniokondi 18
23. Ougarou (ANTRAC) 18
12. ama 24. Tindangou 16
25. Kpcaali 16
TOTAL 355 125
*Indicates villages where the chief was purposively included in the sample in order to assure village support for the survey. Due to the non-random nature of the selection
process for these seven village chiefs, they are excluded from the analyses in this

30 Mies
0 X.10 eleq
1104^0109 -n
(T A)
m"'to-m rn
TA ko
TA Snoy AMA 00
TA m
.b main Road
Secondary Flood Trail
0AWA 00 Sector HoodquvWt
Village A SAMO#d Villaoil,
So.pled Anmal Trutoft A"" % Enumeslot Ftesidofto

2.1. Overview of the Hoe Production System
Farming in the Eastern Region consists almost entirely of hoe agriculture and livestock production by small farmers. Virtually everyone farms and most people raise goats, sheep, or less frequently, cattle. The principal crops are sorghum and millet. Family labor is the key agricultural input. Even including the government and service sectors, there are few opportunities for full-time wa e employment.-/ Because both input and product markets are poorly developed, few economic transactions are monetized and most crop production is consumed by rual households. Despite the low productivity of existing agricultural techno ogy, moderately fertile soils enable small agricultural surpluses to be produed in average years. Nonetheless, infrastructural and institutional const aints make it difficult for farmers to participate in the market economy.
The low productivity of hand hoe production systems is largely a function of th limited area that hoe farmers can cultivate within the relatively short span of the rainy season (550 mm. of rainfall spread over 3.5 months in the extreme no then part of the Eastern Region to 1,100 mm. of rainfall over 5.5 months in the south). An adult can plant and weed only about 1.25 hectares (see Table 2.1 at the end of this section). Because of high variability of rainfall and high ev poration rates at the beginning and end of the rainy season, effective rainfa 1 is low and the timing of planting can be extremely critical. High midseason rainfall stimulates the growth of weeds which can greatly reduce yields.
2.2. The Farming System with Animal Traction
In the literature on technical change, mechanization is considered to be
labor-saving with little, if any, impact on yields (Bieri, de Janvry, and Schmitz, 1972; Binswanger and Ryan, 1977). Proponents of ANTRAC in West Africa hIve attributed much broader benefits to ANTRAC. By replacing hoe cultivation,
A TRAC potentially allows farmers to expand acreage and improve yields. Acreage e mansion is possible through a reduction in labor time required per hectare. Fcr example, animal weeding is three to four times faster than hand weeding for a given area. Higher yields result in the short run from better and more timely p rformance of tillage, and in the long run from improved soil fertility due to
! Wilcock (1981) for a detailed description of rural small-scale enterprises in the Eastern Region.

incorporation of manure and crop residues.1 Savings in labor time due to ANTRAC may be devoted to other activities of value to the household. Use of animaldrawn carts can facilitate crop removal and marketing and provide a source of income from custom transport where the demand for that service exists.
Full adoption of ANTRAC entails several major changes in the traditional farming system: (1) learning to manage large animals; (2) using new implements and agronomic techniques; (3) intensifying land use and maintaining soil fertility; (4) changing the crop mix; and often (5) substantial borrowing to finance purchase of the ANTRAC package. ANTRAC adopters also become more dependent on
outside institutions for input supply, repair and maintenance, animal health services, credit, and extension advice.
2.2.1. Managerial Requirements of Large Animals
Hoe farmers in the Eastern Region of Upper Volta and in most parts of the Sahelian region of West Africa have had limited experience with livestock other than goats, sheep, and poultry. A farmer adopting animal traction must be able
to select the appropriate species, breed, and age of animal. He must learn to train and maintain these large animals on his farm. The necessary managerial skills are complex and thus take time to acquire. Training animals, especially oxen, is a new farm task which can be daunting for an inexperienced farmer. Animal feeding requires knowing how to conserve forage and how to formulate rations. The farmer must reallocate his labor force to provide forage and pasture for his animals. Lastly, he must learn how to maintain the health of his animals through good stabling techniques and the use of prophylactic and curative veterinary practices.
2.2.2. Use of New Implements
and Agronomic Techniques
ANTRAC technology entails new tillage techniques, and hence a series of unfamiliar economic and agronomic decisions. The farmer must decide whether the returns from certain practices will be sufficient to justify investment in the necessary equipment and labor. For example, seedbed preparation is critically dependent upon the timing and the quantity of rain. If rains come late, the hoe farmer does not have the time to prepare all his fields even though he knows this is beneficial. A delay in planting will reduce yields. Plowing with a moldboard
"/Maintenance of soil fertility on continuously cultivated land permits a transition from extensive bush fallow farming to intensive "sedentarized" farming, which is implicitly regarded as desirable by some ANTRAC advocates.

plow is quicker and easier but can still delay planting because a good, soaking rain is necessary before the sunbaked, hardened soil can be worked. Whereas a traditional farmer can plant immediately following the first rains with the daba (hand hoe), the traction farmer plows first before planting. There is thus a trace-off between the benefits of plowing and early planting. Irregularity of
early rains and the probability of drought periods in May and June exacerbate thi dilemma.
The farmer must also decide whether to carry out scarification, which does not provide the same agronomic benefits as plowing, but is quicker.- In areas with low rainfall (600 mm.) and sandy soils, scarification may be the best method of eedbed preparation for large fields because the rigid scarifier tines can enable the farmer to work the soil even before the first heavy rains. Other
dec sions involve whether to plow under green manure at the end of the rainy sea on, or to prepare the seedbed using a ridger to avoid water-logging.
Sowing techniques.may also need changing. For example, weeding is a major constraint which can be overcome by animal-drawn weeders, but this requires the farmer to plant in lines by hand or with an animal-drawn seeder. If a farmer plants by hand, he must use a line tracer or a rope. Animal-drawn seeders can increase the speed at which fields are planted, providing the land has been des umped and derooted and the soil is moist.
Ridging, or basin-listing, increases infiltration and helps prevent lodging during the latter part of the season. Ridging can be performed with either a rider or a moldboard plow, although the latter is relatively inefficient since it requires at least two passes down a row.
Other implements can be used such as a peanut lifter for harvesting groundnuts and weeding, a harrow for breaking up clods prior to use of a seeder, and a
-lage size ridger for making drainage ditches and erosion control bunds. Carts, al hough very costly, have a multitude of uses and enable farmers to work their an mals throughout the year, maintaining their state of training and amortizing their cost. An animal-drawn water-lift system (the dalou) permits irrigation of smll garden parcels and increases animal use.
The maintenance of all this equipment poses a new set of problems for the famer who often has little experience with machines. In the Eastern Region, for ex ple, there are presently only a few blacksmiths who know how to manufacture
I/Scarification refers to light tillage (1 to 5 cm. deep) using a springto thed cultivator, illustrated in Figure 3.1.

spare ANTRAC parts and repair equipment. By contrast, tools used in the manual farming system are manufactured locally and are readily available.
2.2.3. Intensification of Land Use and
Maintenance of Soil Fertility
A move from shifting cultivation to animal traction and the permanent utilization of fields accelerates the use of soil nutrients. Farmers must therefore learn to maintain the fertility of their land through the use of green manure, animal manure, compost, and chemical fertilizers. Farmers must also learn to prevent erosion. If improperly done, plowing increases the amount of soil washed away. This can be counteracted by contour plowing and the construction of bunds.
2.2.41. Changes in Crop Mixtures
Under the traditional system, farmers produce primarily to meet their own
consumption needs. The adoption of ANTRAC greatly increases the farmer's capital needs and requires him to increase the area under cultivation or to alter the cropping pattern in order to produce a larger marketable surplus.
2.2.5. Supporting Services Required for
a Successful ANTRAC Program
In order to assist farmers to make a successful transition from hoe cultivation to animal traction, a wide range of supporting services are required. These include:
1. a credit system to help farmers finance the purchase of equipment and
2. an extension service, particularly for farmer training in the use of
ANTRAC technology;
3. a livestock service to vaccinate draft animals and to give farmers
advice on selection and feeding of animals;
4. a system to supply equipment, spare parts, and repair facilities;
5. a system of on-farm, adaptive (farming systems) research to identify
problems and develop technological packages which are appropriate to
local market, agronomic, and family economic conditions; and
6. a marketing system which enables farmers to sell their surplus production.
2.3. Summary Characteristics of Hoe and ANTRAC Households
Due to large agro-climatic differences across the 12 zones sampled in the 1978-79 farm survey, it is necessary to assess the impact of ANTRAC by comparing

ANTR C households with a control group of hoe farmers within the same agroclimatic zone. Thus, the hoe farmer control group analyzed in this report consists of 106 households sampled within the five ANTRAC zones (see Table 1.1). Further, the performance of oxen or donkey traction is compared to that of hoe agriculture only within the relevant oxen zones (Diabo or Ougarou) or donkey zon s (Piela, Diapangou, or Logobou). Because the ANTRAC and hoe farming
samples vary in size within the individual zones, a weighting procedure was used
in alculating mean values for the two oxen and three donkey zones.2/
Table 2.1 presents summary characteristics of the oxen, donkey, and hoe farers surveyed. Both oxen and donkey farmers have larger families than their hoe farmer counterparts, a scale factor which should be kept in mind whenever com aring household means of ANTRAC and hoe farmers. ANTRAC farmers, particularly hose with oxen, also have a larger work force, as indicated by the number of "ac ifs" (or active workers, persons of age 15 to 54). The ratio of dependents to total persons in oxen households (.53) is similar to that of their control far ers (.54), while it is much higher for donkey farmers (.63) than control farmers (.55).
The total area cultivated is higher for ANTRAC farmers than for hoe farmers, but on a per person basis the difference is not statistically significant. The mos relevant measure of land intensity, the area cultivated per active worker,
1/In oxen zones, 90 percent of ANTRAC farmers use oxen traction. Of the ANTRAC farmers in donkey zones, 85 percent use donkeys. Fifteen (12 percent) of ANTRAC households are excluded from the calculation of statistics for oxen zones an donkey zones because they represent exceptions to the classification system. Of these 15 households, 7 are donkey farmers residing in oxen zones (3 at Ougarou an 4 at Diabo), and 8 are oxen farmers in donkey zones (3 at Piela, 1 at Logobou, and 4 at Diapangou).
!A simple average calculated for all oxen (or donkey) households would not be comparable to an average of all hoe households from the same zones because the ho and ANTRAC sample sizes can differ for each zone. Among oxen zones, for
example, the Diabo zone accounts for 54 ANTRAC users and 18 hoe households, while th Ougarou zone provides 18 ANTRAC and 18 hoe households. If based on unweighted household averages, a comparison between hoe and ANTRAC farming in the oxen
zo es would be biased because the hoe average would overrepresent Ougarou. To re olve this problem, when the ANTRAC and hoe sample size differs within a zone, th smallest of the two is weighted more heavily when calculating intra-zone va ues. Computationally, this means that the hoe subsample in Diabo is given a we ght of 3 when statistics are calculated for oxen zones. For donkey zone
statistics, the ANTRAC subsample in Logobou is given a weight of 2. This
weighting system is not used when variances are calculated, however, because of the bias it would create.

All ANTRAC Zones Oxen Zones Donkey Zones
Number of Households Evaluateda 106 110 36 65 70 45
Persons per Household 7.75 11.21 6.67 11.14 8.83 11.27
Adult Workers per Household 3.50 4.71 3.04 5.27 3.96 4.14
Total Area Cultivated (ha)b 4.30 6.59 3.96 7.13 4.64 6.04
Proportion of Area in:
Millet and Sorghum (%) 80.1 74.7 79.1 77.5 81.0 71.8
Groundnuts 9.6 9.6 10.3 6.8 8.8 12.4
Maize 3.0 3.4 3.3 3.8 2.7 2.9
Cotton 0.2 1.9 0.1 2.1 0.3 1.7
Rice 2.2 2.8 1.9 3.5 2.4 2.0
Soybeans 0.5 3.8 0.5 3.9 0.4 3.6
Other Crops 4.6 4.0 4.8 2.4 4.4 5.6
Total Area Cultivated per Person (ha) 0.560 0.588 0.593 0.640 0.526 0.536
Total Area Cultivated per Actif (ha) 1.26 1.39 1.29 1.33 1.22 1.45
Value of Livestock Owned (FCFA) 122,491 316,545 139,185 372,767 105,320 258,510
Percent of Non-Farm Income 18.7 12.5 32.0 7.5 0.6 21.5
Age of Household Head (years) 53 43 60 44 46 42
Education of Household Head (years) .42 1.06 .44 .64 .40 1.65
aBEcause of time and resource constraints, complete area data was collected for only two-thirds of these households. In a random one-third sub-sample only sorghum and millet fields were measured. While harvest data was collected on all crops for all households, area data presented in this table are based only on measurements from two-thirds sub-sample.
bEstimated for the entire sample by using household size in the one-third subsample to project the non-sorghum/millet area.

is h igher among the animal traction farmers. This will be explored in greater det il in section 4.6 of this report.
Cropping emphasis does not differ greatly between ANTRAC and hoe farmers. Note that the proportion of area devoted to sorghum/millet is very high, ranging from 72 to 81 percent across the subsamples. There is a slight reduction in the proportion of total area in sorghum and millet, along with a minor increase in rel tive areas of maize, cotton, rice, and soybeans in ANTRAC households. Oxen and donkey farmers have similar cropping patterns, although the latter de-emphasize sorghum and millet to a greater degree.
ANTRAC farmers tend to be more wealthy than hoe farmers. The value of
liv stock owned by ANTRAC farmers is more than twice that of hoe farmers, a dif erence only partly attributable to the ANTRAC animals alone. The proportion of household income originating from nonfarm sources is similar for ANTRAC and hoe farmers, although it is highly variable by zone. The heads of ANTRAC
households tend to be younger and better educated than their hoe farmer counterparks.
We will return to these comparisons in Chapter 4 and to their economic implications in Chapter 5. First, however, in Chapter 3 we will examine past attempts to introduce ANTRAC technology in the Eastern Region, and the institutio al structure and performance of the EORD ANTRAC program.

3.1. Previous Programs-1
There have been at least six different efforts to introduce animal traction in the Eastern Region since the early 1940s when the colonial government decided
that animal traction should be used to increase the production of industrial crops, particularly cotton and groundnuts. This first scheme failed because:
(a) the equipment, designed to be pulled by European oxen, was too heavy for the
local animals; and (b) farmers were forced to use the equipment and were punished if they did not. A second scheme in the early 1950s used the same equipment but tried renting out the package instead of forcing farmers to use it. Unfortunately, fear of the intentions of the colonial administration and poorly adapted equipment limited acceptance of this program as well.
The third effort, based on "pilot farms", was begun in 1954. The idea was to bring together in one package all of the ingredients necessary for a successful animal traction farming system. The package consisted of a trained pair of oxen, plus a plow (28 or 36 kg Kirpies), a weeder (Ebra), a harrow (Puzenat), a
cart, and gardening tools. The farmer was also given a stable for the oxen with a cement floor for collecting urine, a compost pit, a pit silo, a rack for drying hay, and a shed for equipment and grain storage. The package, financed by FIDES and the local Provident Societies (Societes de Prevoyance), cost an average of 128,000 FCFA per farm in 1956 and was provided free to the farmer.-/ In return,
the farmer contracted to adopt the practices recommended by the extension service for a duration of three years. These included composting litter and manure, hay
and silage conservation, and a crop rotation system to be implemented on a 2-hectare demonstration plot measured off for him. One-half hectare of this was to be used for a garden and orchard. The rest was to be put into either rice (on bottom lands) or a rotation of cereal crops, groundnuts or cotton, and a green manure crop.
l/This chapter is based on interviews with extension agents who have worked
in the region for 20 years, various reports by the Agricultural Production Bureau of the Eastern ORD, an analysis of some portions of the farm survey data, and other footnoted sources of information.
-/One U.S. dollar was approximately equal to 220 FCFA during the 1978/79 survey period.

By 1958, this scheme had failed for the following reasons:
1. Farmers felt they were working for the extension service. There was no
real understanding of the benefits of the farming practices being extended. When they followed the agents' advice, it was because they felt an obligation to do so. Farmers were flattered by the attention they received, but if there was any relaxation in supervision, they generally abandoned the new technology.
2. The extension service provided insufficient technical support to farmers.
3. Although farmers were allowed to choose between cotton or groundnuts,
they had to plant one of them. These were the only crops with a
guaranteed market. There was no effort to improve the marketing of cereals which farmers preferred to grow in order to feed their
4. A lack of spare parts and skilled blacksmiths to repair the equipment
caused worn equipment to lie idle.
From 1958 to 1961, the "Rural Communities" (Collectivites Rurales) provided injerest-free loans for farmers to purchase animal traction packages, with the proviso that the recipients would serve as model farmers in their areas. The package consisted of oxen bought and trained by extension agents, and then resold to the farmers with the necessary equipment. This appeared to work fairly well fo a time. One extension agent remembers that in Yobri (Diapaga Sector), women were a major force in introducing the package. They pushed their husbands to accept animal traction in order to avoid the arduous labor of seedbed preparati n, particularly for rice. Although this system was a partial success, it was pl gued by the same problems which the earlier "pilot farm" scheme faced--lack of a profitable market for cereals, absence of spare parts and repair facilities, and weak technical support for the farmers. The program ended in 1961 when a
shift in policy orientation at the national government level eroded support for animal traction in favor of the introduction of tractor plowing.
The fifth program in the Eastern Region took place in the late 1960s and early 1970s, when the extension service arranged for about 20 volunteer farmers to be sent to the Center for Agricultural Training (Centre Polyvalent de Formation Agricole) at Matourkou, near Bobo-Dioulasso. The goal of this scheme was to provide farmers with in-depth training in the use of animal traction technology. E~ch trainee was-provided with a pair of oxen, a cow, a plow, and a weeder. In addition, every two families were given a harrow, a seeder, and a cart. If he d sired, the farmer could pay'for his equipment over the three-year duration of

the training course. Each trainee was allotted a 7 hectare (ha) farm on which he was required to plant 2 ha of cotton and groundnuts, 1.5 ha of cereal crops, and 0.5 ha of forage crops. Another 2.5 ha was to be kept in fallow pasture land. The Center provided all chemical inputs (fertilizer and insecticides) and marketed all surplus production. The cost of inputs was deducted from the farmer's income.
Within the artificial environment of the center at Matourkou, this scheme worked fairly well. Farmers could expect to make a net income of 30,000 FCFA in 1965. But on their return home, they encountered some of the same problems which defeated the four previous efforts: lack of reliable markets and inadequate supply of spare parts and other supporting services. In spite of this, many of these farmers have continued to use animal traction and their farms are quite impressive. They do have a good technical background and it is from this group that the EORD has recently recruited three of its oxen trainers.
A sixth effort to implement animal traction was the voluntary work of a private French farmer, Maurice Colas, who spent almost 20 years in the Eastern Region in the 1960s and 1970s. While he trained farmers and supplied them with credit for oxen,- his main innovation was provision of repair facilities and local manufacture of spare parts for existing plows. He also designed a locally manufactured donkey cultivator called the Houe Fada. Many of the people that he trained are now employed by the Eastern ORD as skilled workmen.
These projects shared several common features which led to partial or complete failure:
1. Most of the schemes had a predetermined cropping plan and animal/equipment package. Faced with an inflexible farm plan, the farmer tended to feel that he was working for the extension service. Since the emphasis
on cotton and groundnuts was not backed up with an applied research program at the farm level, these crops were not well adapted to farmer
objectives and resources.
2. The absence of stable, profitable markets for cereal crops made it
difficult for farmers to use cereal surpluses in good years to meet loan
3. With the exception of Maurice Colas' efforts, the absence of spare part
supplies and repair facilities rendered worn-out plows and weeders useless.
1In-1974 when he left Diabo, Colas had loans outstanding to 174 farmers for oxen. The recovery of these loans was taken over by the EORD.

4. Most efforts were based on the "model farmer" approach and thus reached
only a small number of people. These farmers were often isolated
(either as individuals or as groups) from the supply of spare parts, veterinary services, the technical advice of extension agents, and other vital supporting services. It was easy for them to become discouraged and give up the package.
5. The projects all relied on pairs of oxen as draft animals, at least for
plowing. While these animals are stronger and appreciate in value over time, they are much more expensive and difficult to feed than are donkeys or even single oxen. Single oxen, which have proven very effective for weeding and ridging, were never tried and it was only in
the late 1960s that donkey plowing was introduced on a wide scale.
6. All schemes concentrated on the use of the moldboard plow for seedbed
preparation. While this has undeniable benefits, it is slow and can delay planting significantly. Under certain circumstances, scarification can be more beneficial.
7. Each effort to introduce animal traction in the Eastern ORD attempted to
impose the technology and make it work within a short 3- to 5-year
period. Because the system involves the adoption of a complex technology and revolutionary change in farming practices, there was no hope of
success in such a short time.
3.2. The 1974-80 Eastern ORD Program During 1974-80, the EORD relied on extension of animal traction as the main st ategy to increase dryland agricultural production. Animal traction was also th key to seedbed preparation in bottomland rice production systems, after an initial plowing of virgin land by tractors. Although EORD personnel have been involved in the extension of animal traction since the organization's inception in 1968, large-scale effort began in 1974 with the introduction of 26 traction un, ts-11 in the Diabo Sector. The program expanded rapidly, with over 1,700 units in use by April, 1980. Forty-two percent of these units were powered by oxen.
Ta le 3.1 shows the evolution of the program up to April, 1980 by EORD sector and by type of traction unit. In spite of this considerable effort, less than 5
1/A "traction unit" typically refers to either a pair of oxen or a single do key plus the corresponding ANTRAC equipment. Single ox or horse traction is uncommon in the Eastern Region.

EORD Sectors
Type of
Animal Agricultural Comin- ORD
Unit Campaign Bogande Yanga Diabo Diapaga Fada Kantchari Matiakoali Pama Total
Oxen 74-75 + 75-76 11 24 4 39
Units 1976 1977 71 1 27 1 100
1977 1978 64 10 122 26 29 47 35 30 363
1978 1979 7 19 8 5 3 14 5 29 90
1979 1980 40 10 21 16 28 10 10 135
TOTAL 111 39 233 48 111 76 40 69 727
Donkey 74-75 + 75-76 9 72 12 59 152
Units 1976 1977 42 9 35 86
1977 1978 47 6 52 68 70 12 8 3 266
1978 1979 13 4 84 52 49 41 19 7 269
1979 1980 34 3 19 61 58 38 23 3 239
TOTAL 103 13 269 202 271 91 50 13 1,012
Total 74-75 + 75-76 9 83 12 83 4 191
1976 1977 113 10 62 1 186
1977 1978 111 16 174 94 99 59 43 33 629
1978 1979 20 23 92 57 52 54 24 36 359
1979 1980 74 13 40 77 86 48 23 13 374
TOTAL 214 52 502 250 382 167 90 82 1,739

per ent of farms in the Eastern Region were equipped with a usable traction pac age as of 1980, even when equipment placed before 1974 is included.13.2 1. The Eastern ORD's Approach to the
Extension of Animal Traction
The EORD extension program for animal traction differs from most past progr s in that it focuses not on a few model farmers, but on reaching a large number of farmers throughout the Department. It is difficult to provide isolated model farmers with essential support services. The EORD's approach assumes that if enough farmers adopt the technology, the demand for these services and the economic benefits derived from increased production will be sufficient to warrant the investment necessary to improve the services.
The EORD has begun to create the extension services necessary to support mass adoption of ANTRAC technology. There were 155 ANTRAC extension agents in 19 0, or about 1 agent per 360 farm families, compared to less than 40 agents in thd early 1960s. Eight specialists in agricultural credit publicize loan preconditions and ensure the smooth running of the credit program. Six livestock
ag nts have been trained to improve health care and feeding of traction animals. Te oxen trainers have also received short training courses on advising farmers how to choose and train their animals as well as helping extension agents give demonstrations of plowing and weeding techniques.
To reach the maximum number of farmers, extension agents work mostly through village groups (Groupements Villageois). This facilitates the dissemination of technical information. Demonstration of new techniques on collective fields rIduces risk to individual farmers. A farmer must be a member of a village group a d have the approval of the group president before he can receive credit. In a dition, the group guarantees his loan.
Too much is currently expected of the extension agents. The new ones are
yo ung men, fresh from school. They often lack the experience and confidence necessary to give farmers proper support. A partial listing of their responsib lities includes:
1. The extension of ANTRAC technology and improved agronomic practices
(such as the use of new varieties, fertilizers, and insecticides) and
the conservation of forages;
-Based on: 1975 population figures of 402,720 and a 1.93 percent annual growth rate; 7.27 members per household; 95 percent of households involved in farming; and an estimate that 500 units distributed before 1974 are still func:ioning.

2. Determination of credit worthiness and repayment capacity;
3. The distribution and collection of short- and medium-term credit;
4. The organization of village groups and support of group activities such
as cereals banks, pharmacies, and shops;
5. Statistical reporting;
6. Sales of inputs;
7. Conducting adult literacy classes; and
8. Management of supplies.
It is doubtful that anyone could do all of these activities effectively even in the small zone under each agent's control. Furthermore, agents are often transferred in and out of a village so quickly that they do not have time to establish working relationships with the population.
3.2.-2. The Credit System
The EORD provides low-interest, medium-term credit for both animals and equipment. Short-term credit is also available for seasonal inputs such as fertilizer and insecticides. An indication of the importance of the credit system is that 89.5 percent of oxen packages and 50.4 percent of donkey packages were bought on credit. For donkey traction, there is a four-year repayment .schedule. For the higher investment in the oxen package, there is a five-year timetable. Both schedules include a one-year grace period to allow farmers to obtain a complete package and to begin to implement the new farming system. This is an important element as farmers rarely receive all of the components of the package during the first year.
A precondition of medium-term loans is that one-third of the area cultivated must be planted in cash crops to enable the farmer to pay back his loans. There is an important difference between the EORD system and previous ones, however, in that the farmer can decide what constitutes a "cash" crop. Thus, farmers can choose between industrial crops, whose markets are relatively well organized in
some areas, and sorghum or millet, which contribute to food needs but whose markets are slightly less certain due to transport and organizational difficulties.
It is well beyond the scope of this paper to provide a detailed evaluation of the performance of the EORD medium-term credit program. However, due to the
vital importance of this supporting service, a few summary observations have been drawn from Tapsoba (1981a), the key study of the EORD credit program.

To begin, it is useful to underline a few of the distinctive characteristics of the EORD credit system. First, the EORD does not play a significant role in mobilizing local financial resources for investment in agriculture. It primarily employs funds from outside sources in its lending program. Of a total of app oximately 85 million CFA ($425,000) distributed in short- and medium-term credit between 1975 and 1980,.only 9.5 percent came from the EORD's own funds. Jusj over 40 percent of the total came from the USAID Integrated Rural Development (IRD) project alone. Second, loans are provided to farmers in kind (equipmen and animals) rather than in cash. Third, the EORD engages in a wide range of other agricultural and rural development activities in addition to the provision of redit. Most credit transactions with farmers have been carried out by extension agents who have, as we have seen, a wide range of other functions to perform.
In evaluating the institutional performance of the EORD credit system, we can examine two key financial indicators: the real cost of lending and the rate of loan repayment. First, the real cost of lending is defined here as the EORD's cos of administering each 100 CFA of loans outstanding. This can be calculated
for a given year by dividing the value of the total loan portfolio outstanding by an estimate of operational costs incurred in that year. Costs can be estimated conservatively as 10 percent of the salary costs of EORD "credit staff" (sector and subsector chiefs, extension agents, and those headquarters staff directly involved in credit administration) and 10 percent of EORD operating costs including vehicle operation. On this basis, annual EORD costs in the three most recent cre it years are:
Cost of Administration
of Each 100 CFA of Total
Credit Year Loan Portfolio Outstanding
1977-78 27.6 CFA
1978-79 30.0 CFA
1979-80 19.2 CFA
These cost figures vary principally 'according to the total value of loans outstanding at the end of the credit year. Other methods used to calculate the real cost of EORD lending also show that the recurrent costs of the EORD acting as n agricultural lending institution are extremely high (Tapsoba, 1981a). The e high costs led the GOUV to create a new nationwide agricultural lending ins itution, the CNCA,1/ to administer agricultural credit in cooperation with the regional ORD's.
1-CNCA, "Caisse Nationale de Credit Agricole."

The second performance indicator is the rate of loan repayment. As the
literature on agricultural credit stresses, a wide range of methods of calculating this measure may be used. The two figures examined here are based on a total portfolio of 71,528,000 CFA of animal traction credit extended between 1976 and 1980: the collection ratio (including payments on arrears) and the percentage of the medium-term portfolio in arrears on an annual basis. These figures are shown in Table 3.2. The annual collection ratio has varied between 38 and 57 percent. The percentage of portfolio in arrears has risen steadily from 2 percent in 1977 to 28 percent in 1980. These figures are quite disturbing since these indicators would normally reflect EORD performance favorably." However,
they indicate that the E0RD system has not been performing adequately in terms of loan collection. If the collection ratio is not substantially improved, the amount of original working capital available for new loans will be severely limited in future years, and there may be negative psychological effects. on both E0RD personnel and farmer borrowers.
Two questions must be raised at this point. First, can this poor performance be justified? Second, have measures been taken which can potentially improve the performance of the system? The answer, in both cases, is a qualified yes.
The following factors largely explain poor system performance to date:2/.
1. The extremely rapid growth of EORD personnel and number of loans over
the 1975-1978 period clearly outstripped EORD administrative capacity.
2. All field actions in the vast and sparsely populated Eastern Region are
complicated by the almost total lack of reliable road and communications infrastructure.
3. The EORD is not primarily a credit institution. EORD field staff did
not begin to receive detailed field training in credit admnistration until 1978.q' Specially trained credit workers (one per EORD sector)
were not available until 1979.
"/This is because the collection ratio includes payments on installments in arrears in addition to current installments and because the percentage of portfolio in arrears will tend to remain low in early years since the value of the portfolio has been increasing rapidly.
VOther reasons for loan repayment delinquency are contained in Stickley and Tapsoba (1979).
!/This resulted in a detailed credit training manual, ORD de l' Est, BDC (1979).

(in thousands of CFA)
1976/77 1977/78 1978/79 1979/80
1. Total Portfolio at End
of Credit Year 10,729 52,774 59,177 71,528
2. Installments Due 206 2,336 14,027 24,584
3. Total Repayments Made-/ 78 1,332 6,092 11,300
4. Arrears at End of
Credit Years 170 1,331 9,639 19,736
5. Collection Ratio 38% 57% 43% 46%
(Row 3 Row 2)
6. % of Portfolio in Arrears 2% 3% 16% 28%
(Row 4 Row 1)
a/Includes payments on arrears. Source: Tapsoba (1981a), based on EORD Credit Account computer printouts.

4. Temporary or permanent attrition in EORD field personnel, particularly
sector and subsector chiefs (due to reassignment, prolonged absences for training programs, firing, and resignation), has meant that in very
few localities has there been stable, continuous contact between the
EORD and farmers taking loans.
5. Personnel movements and rapid growth in the volume of loans had a
disastrous impact on credit record keeping and accounting. This situation became so acute that it was necessary in 1978 to conduct a detailed sector by sector inventory simply to establish reliable records on the
numbers and status of outstanding loans.!! This was the first step taken in a complete reorganization and computerization of credit record keeping, accounting, and administration which was completed in 1979.-U
6. Poor record keeping had a strong negative impact on loan collections in
two ways. First, in many instances, farmers were simply not contacted by EORD personnel when credit installments became due. All evi dence
suggests that when farmers in the region fully understand their credit
obligations, they take them seriously. The vast majority make every effort to make required payments even if they must liquidate household
assets to do so. Second, poor record keeping and extremely frequent personnel movements led to loose program administration which permitted substantial embezzlement of farmer credit payments by some extension agents and subsector chiefs. In some individual cases, this embezzlement amounted to over 200,000 CFA.
7. Finally, improvements in the EORD rural credit system instituted in
1978 and 1979 did not show a positive impact on system performance in 1979-80 since there was an agent boycott of credit collections through
most of the collection period.
The following factors should contribute to substantial improvement in credit administration and repayment:
1. The rate of growth of E0RD personnel has stabilized since 1978 and the
increase in the number of new loans has been fairly moderate.
2. Current field personnel have received substantial practical training in
credit administration.
I/The results of this inventory are reported in ORD de l'Est, BDC (1979).
21This program is discussed in Stickley (1980a, 1980b, and 1980c).

3. The new computerized system of credit administration has been in effect
for about one year. This system has streamlined required paperwork and
has made it more possible to assure program continuity even if there are continued personnel movements. It has also substantially decreased the potential for the embezzlement of credit payments. In fact, of the payments received during 1979-80, a substantial proportion involved voluntary reimbursement to the EORD of funds previously misused by
field personnel.
4. Computer reports can now indicate those sectors and subsectors where
repayment rates are low, calling for greater collection effort.
3.2.8. Marketing
ORD's have traditionally provided marketing services for cash and food crops. Marketing was a key component in the intensive zone strategy of the USAID IRD roject. However, the decision by the GOUV to concentrate cereals marketing acti ities in the hands of OFNACER (Office National des Cereales) has severely handicapped the EORD. While OFNACER provides a base price for cereals, its acti ns are fairly limited and unpredictable. Thus, farmers still complain that marketing is their number one constraint. In addition, lack of readily available capital and transportation facilities and poor national price policies have discouraged EORD cash crop marketing.
The EORD has, however, intervened in the grain marketing system by supplying short-term credit to village groups so they can buy grain at harvest time for local "cereal banks", when farmers need money and when the price of grain is low. The group then resells the grain at a later date when prices have risen. In some areas, the profits, which in the past have gone to large grain merchants, now sta) within the region. Unfortunately, this is being done only on a limited scale.
3.2.4. Animal Insurance
For oxen traction users, the animals represent the most expensive part of the package; losing one can spell disaster for a farmer, and may lead him to abar don animal traction. A significant new addition to the ANTRAC package in the Eas ern Region has been compulsory animal insurance for both donkeys and oxen. For a total of 15,000 FCFA for a pair of oxen and 3,750 FCFA for a donkey, the anim als are insured for a period of five years against death by sickness or unavoi dabl e acci dents.

3.2.5. Types of Draft Animals
The EORD offers credit and assistance to farmers to purchase three different types of animal power systems: two oxen, one ox, or one donkey.- Each system has benefits and drawbacks:
1. A pair of oxen provides the most power. Their value appreciates during
the working years, but they are difficult to feed and initially very expensive. The 1980 price of a good three year old pair was about 70,000 FCFA but the EORD offered a maximum loan of only 55,000 FCFA.
As a result, farmers often bought young or sickly animals when they
could not afford to pay extra for a good pair.
2. A single ox is cheaper, easier to feed, and easier to handle than a pair
of oxen. Crop maintenance operations can be performed more easily using one strong ox than using a pair of weak oxen. By comparison with a
donkey, the single ox is stronger and it appreciates in value, although
it is more expensive and more difficult to maintain.
3. The donkey is the cheapest animal (about 15,000 FCFA in 1980), and the
easiest to train, handle, and maintain. However, the value of a donkey
does not appreciate and it is not as strong as oxen.
This choice of animal power system provides the flexibility that is needed to adapt ANTRAC technology to a wide range of family sizes, family wealth positions, and local agro-climatic conditions. For example, a dryland farmer with sandy soils in the northern part of the Eastern Region might choose donkey traction because of the difficulty of feeding oxen and because he does not need as much power to cultivate his sandy soils. Another farmer, next to an irrigated perimeter, might prefer oxen because rice straw is available to feed them and because oxen are needed to prepare heavy bottomland soils.
3.2.6. Animal Training
In past projects, pre-trained animals have been sold to farmers. The
teaching of training techniques was left to extension agents. This failed since the agents often did not have the necessary knowledge, time, or inclination. The EORD currently teaches farmers how to choose and train animals themselves. This
"/There are only two farmers in the Eastern Region who currently use horse traction. Horses are as expensive as an ox, difficult to maintain, and, since they are considered prestige symbols, farmers are reluctant to use them for work. They walk faster than oxen, but the sustained effort of the small horses found locally is not as great.

is n important ingredient in a self-sustaining system. The EORD has now hired 10 Oxen trainers (bouviers) to help farmers. This program seems to be working very well. The results of a survey show that 92.3 percent of the farmers who had received their oxen and equipment between January and June of 1979 had had assistance in training their animals. The amount of time worked by the animals during the first year was also significantly increased by the efforts of the bou~i ers.1
3.2 7. Livestock Extension
The EORD livestock bureau has a veterinary service as well as an animal pro auction extension service. The veterinary service is generally understaffed and poorly equipped. As a result, only 26 percent of oxen and 3 percent of don eys purchased for the 1979-80 campaign were vaccinated.-/ The EORD hired six new livestock extension agents in 1980 to help vaccinate animals. However, the supply of equipment and delivery of veterinary medicines is still inadequate and
must be improved.
The animal production extension service is a new effort for the EORD, which has traditionally concentrated on vaccination programs. Livestock extension agents will be responsible for teaching farmers how to formulate rations for dry season maintenance of draft animals and for fattening older oxen for resale as meat animals. Dry Season Feeding
Forage production for dry season feeding is important in maintaining animals in good shape for the plowing season, the time when maximum effort is deim anded. Table 3.3 shows results from observation trials on numerous forage species carried out by the EORD to determine those best suited to local farming systems.- Species with promising production potential are available for most ecological zones in the Eastern Region. The best alternatives seem to be:
(a) planting pure stands of short-cycle mung beans in late July; or (b) intercropping high yielding cowpea varieties with cereal crops. The disadvantage of
(b) is that serious insect problems occur if more than 25 percent of the field
1/For more details, see ORD de l'Est, BPA (1980), pp. 12-13.
-/Based on a survey of 117 farmers made in December, 1979. See ORD de l' st, BPA (1980), p. 21.
3/See Barrett (1979b).

Minimum Dry Matter
Rainfall Production
Species Common Requirement Preferred at Fada
and Varieties Names (mm) Soil Type Kg./Ha.a Observations
Phaseolus Siratro 600 Well drained, 1,005 Produced well in pure stands as
Macroptilium sandy well as in grass-legume mixtures. Remains green into
Stylosanthis Brazil 900 Poorer soils 916 Takes two years to establish.
Gracilis lucerne clay or
Phaseolus Mung 600 1,004 Short-cycle (45 days to floweraureus bean ing produces best in a pure
"MG55" stand.
Vigna Cowpegs 600 Clay, well 2,750 There are many problems with
unguiculata (Niebe) drained insects in pure stands.
"Vita I"
Brachiaria Congo 1,000 Rich, well 3,718
ruziziensis grass drained
Sorghum Forage 500 Clay, well 4,050 Difficult to convince farmers
bicolor sorghum drained to grow sorghum for forage.
Sorghum Columbus 400 Clay, bottom 2,570 Although less productive,
almum grass lands easier to vulgarize.
yields of pure stands during the 1978-79 cropping season.

is pl nted in cowpeas. Also, most high yielding varieties of cowpeas are spreading t pes which interfere with weeding.
3.2.7. 2. Fattening of Old Oxen for
Resale as Meat Animals
armers already know the value of raising oxen for meat. In three cases during the early part of 1980, farmers in the Fada Sector were able to sell oxen for 70,000 to 90,000 FCFA a piece. In each case, the farmers had worked their animals four years, and in each case they bought a new pair for about 70,000 FCFA.
The marketing outlet provided by ONERA (Office National de l'Exploitation des ressources Animales) should encourage fattening schemes, since they buy by weig t and by animal quality. ONERA is willing to buy animals where lots of at
leas 10 animals can be assembled in each village of a zone supplying at least 30 anim ls. However, the EORD program to encourage farmers to fatten old oxenl/ has not orked very well owing to lack of available supplemental feeds during the dry seas n.
3.2. Equipment
The EORD began its animal traction program by selling imported, Bourguignon plow A six-inch plow (Tom 14) was sold for donkey traction and a nine-inch plow (BM2M) for oxen. Three-tined weeders and ridgers that can be mounted on the BM2M have been manufactured locally by COREMMA. The equipment is favored by farm rs because of its light weight, but it has become difficult to find spare party.
Plows produced in Upper Volta were introduced in the Eastern ORD in 1976. The irst to be distributed was the HV1B, a nine-inch ox plow made in Ouagadougou. This is a solid plow, although farmers complain that it is too heavy and the hand es are too high. Three- and five-tined weeders were also available. Unfortuna Iely, many of the parts (i.e., plowshares and wheels) are not interchangeable with later models and again farmers have had difficulty finding spares.
In early 1977, an equipment assembly plant was opened by the EORD in Fada N'G urma as part of the COREMMA (Cooperative Regional de Montage de Materiel Agricole) system. Since then, it has produced 2,150 plows, 2,400 weeders, 1,250
riders, and 400 carts. The EORD manages this installation. Duty-free raw materials are sold to the Fada plant by ARCOMA (Atelier Regional pour la Constr ction de Materiel Agricole) which is the parent organization. ARCOMA does
1/See ORD de 1'Est (1977).

most of the heavy work requiring expensive machines such as bending the plow beams. They are also responsible for making design changes in existing equipment and for developing new tools for specialized tasks such as bund formation.
COREMMA produces two lines of equipment, illustrated in Figure 3.1. The first is the HV2B multi-cultivator for oxen. It is based on a nine-inch plow with a steel beam. Besides the normal trapezoidal plow shares, a new model with a chisel point is now available to help break hardened soils before the rains start. The cultivator, which accepts the plow handle and wheel, has five flexible tines. It comes with scarifying shares and weeding shares. Rigid tines can also be mounted on the cultivator frame for scarification of hard soils. As of 1980, these were only available for demonstration purposes. An adjustable-wing ridger can be fitted to the plow beam. A prototype peanut lifter has also been produced.
The second line is the HV2A multi-cultivator designed for donkey traction.
It is based on a six-inch moldboard plow, and it has a three-tined cultivator as an accessory. There is no ridger available so ridging must be done with the plow.
An advantage of the COREMMA equipment is that certain parts are interchangeable between the donkey and oxen multi-cultivators. This reduces the cost of changing from one type of traction animal to the other. Furthermore, it reduces the number of parts which must be stocked by the EORD. ARCOMA has already made
some improvements in the equipment. For 1981, the plows were given adjustable handles, the cultivator frame and HV2A plow beam were made lighter, and the cultivator tines were reinforced.
COREMMA manufactures three types of carts. These include two donkey carts which are designed for a maximum load of 500 kgs. In the first, the chassis is set up above the wheels, whereas in the second, the chassis sits directly on the
axle and it has a sheet metal bed and sides. Farmers generally prefer the latter because it is more maneuverable and has metal sides which facilitate the transport of sand and gravel. The ox cart has a one-ton capacity. Again, the chassis is set above the wheels and the cart is delivered without sides. Farmers have often stated that this cart is too bulky for use on bush trails. It has not had a very wide acceptance, except in Bogande where there are many roads and little competition with other types of transport.
The EORD has developed a prototype animal-drawn, water-lift system (the dalou) which is now installed in Tiparka, near Fada. This will be of special

Flexible Tined Weeder
5 Tines
9 Inch Moldboard Plow
6 Inch loldboard Interchangeable fletree
Plow Scarifying Shares ier

interest for irrigation of gardens in such areas as Logobou and Piela where the water table is fairly high during the harmattan season.
COREMMA has been able to produce a large number of multi-cultivators but the equipment often arrives on the farm too late to be used during the first year. While the actual assembly operation is efficient and well-run, the supply of raw materials is erratic and slow. This is partially due to ARCOMA not delivering inputs on time. However, there is much room for improvement of the financial and logistical management of COREMMA in order to avoid such pitfalls as ordering the wrong size or type of part and setting inappropriate prices. If the EORD
continues to manage the operation, they should seek personnel with experience in machinery manufacture and marketing.
3.2.9. The Supply of Spare Parts
The CNPAR (Centre National de Perfectionnement des Artisans Ruraux) in Ouagadougou has trained about 10 blacksmiths from the region as a first step in creating a network of skilled workers to repair and provide spare parts for the ANTRAC equipment package. CNPAR planned to open a regional training center in Fada in late 1981.
Since there have not been enough of these specialized blacksmiths in all parts of the region, the EORD has taken an active role in assuring the supply of spare parts. It has contracted with the existing blacksmiths to manufacture plow shares, plow soles, and yokes. The contracts enable the EORD to control the quality of the material and provide incentives to the blacksmiths for timely delivery. The blacksmiths, armed with the EORD's order, are able to get quality raw materials on credit from SACS (Service d'Assistance, de Conseil et de Soutien), a branch of the CNPAR. The EORD gets certain stock parts such as nuts and bolts from ARCOMA.
Although the EORD is now trying to standardize the equipment delivered to farmers, there are 9 different types of plows, 5 weeders, 4 ridgers, and 3 carts which are currently being used in the region. The supply of parts for all this equipment is a logistical nightmare. There are 26 different parts on the HV2B
multi-cultivator alone which will eventually wear out. While many of these parts are interchangeable with parts on other plows, there are probably over 100 parts which must be kept in stock for all of the material currently being used.
3.2.10. Conclusion
The Eastern ORD is a relatively young organization which has enormous responsibilities. Because of rapid expansion in recent years, both in terms of

personnel and projects undertaken, there has been a lack of coordination and plaF ning of ANTRAC activities. The recent arrival of new upper-level management per sonnel who have good technical backgrounds should help to overcome this problem.
The EORD's current program differs from earlier attempts to introduce animal traction in that it is based on the establishment of a "critical mass" of part icipating farmers and on the widespread provision of institutionalized suppor services for all phases of farming with draft animals. It remains to be seen whether the EORD can maintain and improve these support services over the nex five to ten years, as will be necessary if animal traction is to become a via le, self-sustaining technology in Eastern Upper Volta.

This chapter uses farm survey and other data to provide an in-depth analysis of the technical impact of the Eastern ORD animal traction program at the farm level. The topics covered in this chapter include the husbandry of draft animals, farmer use of recommended tillage practices, the use of work animals for
transportation, and an assessment of the impact of ANTRAC technology on total acreage, yields, income, cropping patterns, and allocation of household labor. The technical analysis of animal traction at the farm level will set the stage for an economic analysis of the program in Chapter 5.
4.1. Overview of the Traction Sample
As we have pointed out earlier, there are important demographic and socioeconomic characteristics which distinguish our purposive sample of 125 animal
traction farmers from the randomly selected sample of 355 traditional households using hoe cultivation. The traction farmers in our sample were identified by extension agents as the most successful ANTRAC users in the region. The objective was to select a sample of the "best" animal traction farmers in order to assess the potential of animal traction in the region. As a result, the surveyed ANTRAC farmers were slightly atypical in that they were wealthier- and more educated than average hoe farmers. Due to the recovery of the EORD ANTRAC
program, the level of experience was not high among surveyed ANTRAC farmers, despite the selection process. Fifty-eight percent of oxen farmers had less than three years of experience with ANTRAC, compared to 23 percent of donkey farmers.
The breakdown of animal traction farmers surveyed in 1978-79 by type of draft animal used for the majority of fieldwork operations is presented in Table 4.1. The relative proportion of oxen and donkey traction in the sample reflects the incidence of these two draft animal types in early 1978, when the sample was selected. At that time, only 42 percent of the draft animal units distributed by the EORD were donkeys. By 1980, however, donkey traction had become more widespread than ox cultivation by a margin of about 1.4 to 1.
4.2. Purchase, Sale, and Training of Draft Animals
4.2.1. Animal Purchases
Extension agents are supposed to participate in all ANTRAC animal purchases for which EORD credit is provided. For farmers in our sample with less than
-When evaluated on the basis of key consumer durables that are not strictly essential for survival, such as bicycles, mopeds, radios, and kerosene Iamnps, see Lassiter (1980), Table 15, page 23.

Typ of Animal Used Number Number of Traction Units
fo the MajorbtY of
o Field Work- Farmers Pairs Single
Oxe 72c/ 83 7
Don eys 52d/ 3.5 75
124e/ 86.5 82
a/T action Unit = 2 oxen or 1 donkey. b/H horse traction is not analyzed in this report because only one farmer in the
original sample of 125 traction farmers owned a horse.
-/For the 72 farmers who primarily used oxen, 63 owned one pair, 8 owned two
Fairs, and 1 owned four pairs. No farmer used a single ox.
d/f the 52 farmers who used donkeys: 1 used a borrowed animal; 31 owned 1; 17
owned 2; 2 owned 3; and 1 owned 4 donkeys.
e/-leven of the 124 farmers owned both oxen and donkeys. Six of them used their
oxen and five did not.

three years of experience, extension agents assisted in the purchase of 83 percent of their oxen but only 8 percent of their donkeys.
Crossbreeds (Zebu x Taurin) are generally recommended by extension agents for draft purposes because of their tolerance to trypanosomiasis and their more
compact conformation. However, 66 percent of the oxen bought by our sample farmers were Zebus.- This could be due to their larger size and their greater avail abi I ity.
4.2.2. Castration and Placement of Nose Rings
Both castration and the placement of nose rings are designed to facilitate the handling of oxen. In our sample, 88 percent of the oxen were castrated and 92
percent had nose rings. Most of the animals which had not received these operations came from Ougarou area where ANTRAC has only recently been introduced.-3 In Diabo, these services are performed by farmers themselves in many cases. This suggests a need for EORD assistance in performing these services in areas such as Ougarou where ANTRAC has been recently introduced.
Farmers in our sample reported that 70 percent of their oxen were castrated when they were 3 years or older. This is not a recommended practice because the trauma associated with castration is greater for older animals. Since farmers must buy what is available (generally bulls), they should either buy them at a young age and castrate them or leave them intact. The latter is probably
preferable, since spirited animals are desirable for draft purposes.
4.2.3. Animal Training
Table 4.2 shows the percentage of animals trained by different categories of people.- Eighty-seven percent of the donkeys and 80 percent of the oxen were
trained by the farmer or a friend. Only 15 farmers had their oxen trained by an extension agent. Extension agents generally do not get involved in animal training because they often lack training and confidence.5/
l/This figure is approximate because it was difficult for interviewers to recognize the breed of oxen in all cases.
2/Approximately 60 to 70 percent of all cattle in the EORD are Zebu.
3/Twelve out of 18 uncastrated oxen and 7 out of 15 without nose rings were located in Ougarou.
4/There were no "bouviers" (farmers hired by the EORD as animal trainers) at the time of our survey.
5/Thirteen out of the fifteen farmers whose animals were trained by extension agents were from Ougarou.

Trainer Donkeys Oxen
(Percent) (Percent)
Owner 73 77
Ext nsion Agent 13.5 20
Fri nd 13.5 3
TOTAL 100.0 100.0
The average age of oxen surveyed was 3.4 years at purchase. Fifty-two
percent were bought at age 4 or 5 when it is more difficult to train the animals and when they are more expensive. All working oxen surveyed were between the
ages of 3 and 8 years old. There was no evidence of farmers working their animals bey nd their maximum age of approximately eight years.
4.2 4. Sale of Draft Animals
Farmers sold 19 oxen and 10 donkeys during the 1978/79 survey year.- Five of ,he oxen were sick and farmers managed to salvage 52 percent of the purchase pri e by selling them. Another seven oxen were healthy but they were sold after onl two years' work because they were either too lazy or too aggressive to be trainedd./ Farmers had worked the remaining seven oxen for four or five years so the animals were in prime condition when they were sold. Farmers, on the
ave age, sold these animals for over three times the original purchase price.
s Eight of the ten farmers who sold donkeys owned two or more and can be con idered to be donkey traders. The average donkey sale price was 73 percent ab ve the original purchase price. All of the donkeys sold were in good health
at the time of sale.
l/The number of draft animal sales is probably underestimated because we fo nd it difficult to distinguish between draft oxen sales and the sale of ordinary cattle. In addition, our sample represents an incomplete cross-section oflthe different stages of draft animal growth. Because of the large number of
relatively new ANTRAC farmers sampled, animal sales would be low.
2/The average sale price was 55 percent above the original purchase price.
Fo r of the five farmers involved in these sales came from Ougarou, another in ication of the poor performance of farmers where ANTRAC is being introduced fo the first time.

4.3. Animal Maintenance
4.3.1. Stabling and Feeding
Extension agents recommend that animals be kept on the farm in order to avoid contact with disease-bearing herds from outside the region and to avoid the greater danger of accidents when they are kept in the bush with Fulani herders. The EORD credit system requires that animals be kept at home throughout the year but farmers are sometimes forced to entrust their animals to the nomadic herders when they do not have adequate feed resources or when they do not have enough labor to feed them during the dry season. During the 1978/79 survey year, twothirds of the oxen in our sample were kept in the "concession"'/i at night and put out to pasture during the day. Most farmers who kept their oxen in the bush had less than two years of experience with ANTRAC.
Ninety percent of donkey farmers kept their animals at the "concession" at night during three-fourths of the year because of the danger of leaving animals unprotected in the bush. When donkeys are left in the bush, they must fend for themselves because the Fulani are not willing to herd donkeys. Also, many more
donkeys than oxen were used for carting during the dry season so farmers had more reason to keep them at homne.?.
Seventeen farmers (24 percent) experienced difficulties in feeding their oxen. A lack of feed and watering facilities was cited as a major problem. Most farmers (53 percent) were inexperienced. As was the case with oxen, 25 percent
of donkey farmers complained of having problems with feeding their animals.
4.3.2. Cost of Feeding Animals
In general, the cash expense of maintaining a pair of oxen is about four times that for a donkey. This is an important consideration for a farmer faced with the choice between the two types of animals. The expenses for feeding
donkeys and oxen are presented in Tables 4.3 and 4.4. Approximately 50 percent of the cash costs of feeding oxen are incurred in the period of May through July when animals are required to work the hardestY~ Salt and grain accounted for the largest amounts during this period for both types of animals. Farmers with
"Homestead or compound.
.Y.Thirteen farmers used their donkeys and only one farmer used his oxen for carting in the dry season.
!/This is also the period when farmers purchased the largest amount of food
for their families.

FCFA By Quarter
T$pe of May-July August- Nov.-Jan. Feb.-April
nimal 1978 Oct. 1978 1979 1979 Total
Oxen (2) 1,993 355 660 796 3,804
Donkey (1) 466 130 244 98 938
Type of Expense Oxen Donkeys
(Percent) (Percent)
Mi let and Sorghum 13.0 22.4
Fo age 11.2 15.1
Mi let and Sorghum Bran 6.9 8.9
Sa t 36.1 19.9
Me dici ne 18.2 10.5
Ro pe 4.8 18.9
Labor 7.0 4.3
Other 2.8 0.0
TOTAL 100.0 100.0

oxen, however, spent an average of 337 FCFA on medicine during the survey year. Expenses were lower during the August through October period when forage is plentiful and animals are generally in their best shape. In November through
January, cash expenses increased and a major proportion was spent on purchasing crop residues to store for the dry season.- The major expenses for oxen in the fourth quarter were salt, medicine, and grain.
4.3.3. Animal Health
Eighty-eight percent of oxen and 37 percent of donkeys in our sample had been "vaccinated" at least once in 1978.2Y In order for a prophylactic treatment to be effective against trypanosomiasis, the animals should be treated at least
three times and preferably four times in heavily infested areas. An ideal
program would also vaccinate at least against rinderpest and contagious bovine pleuropneumonia. According to the Chief of the Livestock Service, he has neither the staff nor the equipment to carry out a vaccination program of this magnitude.
Morbidity rates are quite high among surveyed animals in 1978-79--24.7 percent for oxen and 20.2 percent for donkeys. "Trypanosomiasis" and "mange or streptotrichosis" were cited by farmers as the most serious health problems for oxen.4/ The most serious problem for donkeys in terms of the number of work days lost was described by farmers as "Polio." This may be osteomalacia caused by a phosphorus deficiency. Most of the cases occurred during the dry season when P
levels in crop residues and range grasses are extremely low. However, brucellosis is very common in the area and symptoms associated with that disease could be ascribed to "polio."
l/Oxen farmers spent 44 percent of their total cash expenses on forages in this third period and donkey farmers 52 percent. Most farmers do not produce
enough crop residues on their own farms to feed their animals for the entire year. The average amounts spent annually per household on supplemental forages was relatively low (490 FCFA for oxen and 205 FCFA for donkey farmers).
2/The figure for oxen is perhaps somewhat unrepresentative because a study carried out by BDC in 1978 showed that only 54 percent of all oxen in the Eastern ORD had been vaccinated.
!/Personal communication, Dr. Derra, March, 1980.
4/This may be an overestimate. Trypanosomiasis is difficult to diagnose even for experienced veterinary assistants. Moreover, the Gourmantche and More words for trypanosomiasis are often used as catchalls for chronic diseases.

Table 4.5 shows that animal health problems were common and sometimes due to farmer negligence (lacerations). The table also shows that the incidence of illness falls more heavily on oxen than on donkeys.
During the survey year, the oxen of 31 farmers (43 percent) fell sick at least once. Eleven farmers, or 21 percent of the households which owned donkeys, had animals which fell sick during the year. Because of the importance of timely seec bed preparation and weeding it is essential that animals are in good health
during this period. However, 22 percent of the ox teams had at least one animal too sick to work during this crucial period whereas only 6 percent of the donkeys wer sick.
4.3 4. Treatment of Diseases and Mortality Rates
During the survey year, there were 60 cases of illness of oxen and donkeys
ser ous enough to prevent them from working but farmers contacted a veterinary assistant or an extension agent in only 43 percent of these cases. Farmers may hay avoided the high cost of treatment by EORD personnel- or doubted the
uti ity of seeking help from the ill-equipped and understaffed Livestock Servic .
Farmers get most of their veterinary medicines from sources outside the EOR Only a third of the purchases of commonly used medicines such as penicillin, injections, and pills were purchased from the EORD. The balance was either bou ht from a pharmacy or the black market. Home remedies, such as motor oil, salt, sugar, and various herbal concoctions, were used in 46 percent of the cases of disease.
Although the number of cases of sickness decreased during the harvest and dr season periods, the severity of cases generally increased, as can be seen by the average number of days lost per week in Table 4.6. This can be partially explained by poorer nutrition and decreased resistance to disease; especially during the dry season. Since many animals would not have worked even if they were healthy during these two periods, farmers probably applied a different standard as to what constituted an "animal too sick to work," so even this high morbidity rate may be underestimated.
l/The costs may be high relative to treatment by indigenous methods. The ac ual costs of the Livestock Service are heavily subsidized but farmers often must pay for the gasoline required to bring the veterinary assistant to his farm.

Oxen Donkeys
Number Number
Number of Days Percentage Number of Days Percentage
of Lost Due to of Total of Lost Due to of Total
Type of Problem Cases Sickness Time Lost Cases Sickness Time Lost
1. "Trypanosomiasis"
(Trypanosomiasis, intestinal worms,
liverfluke disease, any other
chronic wasting diseases) 14 420 34.7
2. Mange or Streptotrichosis 7 357 29.5
3. "Boils on Mouth" (Foot and
Mouth Disease, BVD) 3 140 11.6
4. Broken Horn or Laceration on Head 8 136 11.2 5 28 5.3
5. "Polio" (Osteomalacia, due to a
P. Deficiency, or Arthritis,
Bursitis, due to Brucellosis) 3 79 6.5 2 168 31.5
6. Sunstroke or Heat Exhaustion 2 42 3.5
7. "Boil on Eyes" (Pink Eye, either
viral or bacterial, or due to
Thelasia Worms) 2 23 1.9 5 74 13.9
8. Digestive Disorders (Constipation
or Diarrhea) 2 10 0.8 1 56 10.5
9. Snake Bites 1 3 0.6
10. Unspecified 2 5 0.3 3 204 38.2
TOTAL 43 1,212 100.0 17 533 100.0

Plowing Time Weeding & Ridging Harvest Period-/ Dry Season-/
May 1-July 16 July 17-Oct. 1 Oct. 2-Jan. 28 Jan. 29-April 30
(11 weeks) (11 weeks) (17 weeks) (13 weeks)
Oxen Donkeys Oxen Donkeys Oxen Donkeys Oxen Donkeys
Number of Animals Sick 19 5 19 3 8 6 8 5
Percentage of Animal Units
Which Lost Work Time
Due to Sickness 21.8 6.0 21.8 3.6 9.2 7.2 9.2 6.0
Average Number of Days Lost
by Sick Animal Unit. 9.9 4.8 29.4 6.7 26.4 64.2 31.9 20.8
Average Number of Days
Lost Per Week 0.9 0.4 2.7 0.6 1.6 3.8 2.5 1.6
-/The harvest period is the time when crop residues and standing hay are plentiful and of relatively good
quality so nutrition should be fairly good.
b/The dry season is the period when range hay is of lowest value and crop residues are often missing.

The mortality rate among oxen was much higher than that for donkeys. Eight oxen died during the 1978-79 survey year for a mortality rate of 4.5 percent.The single donkey death during the year represents a mortality rate of 1.2 percent. This rate is significantly lower than for oxen and is a reflection of
the endurance of donkeys.
4.3.5. Animal Insurance
During the 1978/79 survey year, only 20 percent of the oxen owned by the surveyed farmers and 13 percent of the donkeys were insured. These figures are
probably much higher now because of the EORD's subsequent requirement that all animals bought on credit must be insured. Since insurance is a new concept for
both farmers and extension agents, it will take some time for the concept to become an accepted management strategy.
The insurance rates of 3,000 FCFA per year for oxen appear to be reasonable, given the 4.5 percent mortality rates observed.2/ For donkeys, however, the rate of 750 FCFA per donkey per year seems high and we recommend lowering it to 600 FCFA per year. Even if the observed mortality rate of 1.2 percent increased to 3 percent per year and a 10 percent administration fee is included, the rate
of 600 FCFA per donkey per year would cover the 18,000 CFA replacement cost of a donkey.
4.4. Use of Animal Traction for Tillage
4.4.1. Ownership and Use of Tillage Equipment
Table 4.7 shows the relationship between the length of experience with ANTRAC and the amount of equipment owned. All of the sample farmers owned a plow. Farmers with greater ANTRAC experience owned more equipment. Sixty
percent of the farmers with more than three years of experience owned weeders or
ridgers, while only 19 percent of farmers using oxen for less than three years owned a weeder or ridger.
1Three died from unknown causes, two from snake bites, one from a broken leg, one from a pulmonary disease, and one from an accident.
-It is not known how the original rate of 3,000 FCFA/year/pair was calculated. It does seem reasonable if one assumes the following: (a) a 5 percent
annual mortality rate; (b) 10 percent of the farmer's premium is used to pay administrative costs; (c) average salvage value of 8,000 to 10,000 FCFA; and
(d) only partial replacement cost reimbursement in order to discourage false claims. This means that if an animal died and the farmer collected insurance he would only have to pay between 8 and 12 percent of the cost of purchasing a new pair of oxen.

Ox Traction Farmers Donkey Traction Farmers
Equipment 0-2 Years 3-6 Years 7-28 Years 0-2 Years 3-6 Years 7-28 Years
Plow Only 34 6 6 9 20 7
Plow and Weeder 0 0 1 3 5 2
Plow, Weeder, Ridger 7 8 9 0 4 2
Plow and Ridger 1 0 0 0 0 0
Total Number 42 14 16 12 29 11
Percent Who Owned
a Weeder and/or
a Ridger 19 57 63 25 31 36
!/This is based on equipment owned by May 1978 and does not include weeders and ridgers which farmers
received during June and July of 1978.

About two-thirds of the ANTRAC farmers in the sample did not own a weeder because it was either too expensive (26 percent) or the EORD had not made them available for purchase (34 percent). Only 7 percent of the farmers said they did not want to purchase a weeder. Three-quarters of the farmers did not own a ridger. Thirty percent said they wanted one but the EORD had not yet supplied one, and 44 percent said they thought the ridger was too expensive. Only 4
percent said they did not want one.
In general, Table 4.8 shows that most farmers who owned the tillage equipment used it. All of the sampled farmers owned a plow. The eight oxen farmers who did not plow had only one or two years of experience with ANTRAC. Donkey farmers tend to use this equipment less than oxen farmers. When asked why they
did not use the weeder or ridger, 42 percent of the farmers said their donkeys were not strong enough to pull the equipment. Fifty percent gave the following reasons for not using the equipment: they had not received the equipment until
the year of the survey, animals were not trained, or they had not planted in rows. Of the oxen owners not using their weeding equipment, 60 percent said they had not planted in lines while 40 percent reported that their animals were not well enough trained to do the work.
4.4.2. Performance of Tillage Operations
Table 4.9 shows that the length of experience with ANTRAC is a major determinant of whether farmers used oxen or donkeys in performing different tillage operations. Since plowing is the easiest operation to master and the benefits of plowing are the most readily apparent, it is understandable that 82 percent of the farmers with less than two years experience plowed and all farmers with more than two years of experience plowed.
Although fewer farmers weed and ridge,1 a greater proportion of experienced farmers performed the operations. Twenty-seven percent of experienced donkey farmers and 56 percent of experienced oxen farmers weeded with ANTRAC. Oxen farmers who weeded had an average of 11 years of experience, whereas donkey
farmers had only 6 years of experience.-/ Farmers must have confidence to be able
-The number of farmers who weeded and ridged is different than the number
who used weeding and ridging equipment because some used their plows for these operations.
-/Eleven out of 31 farmers who weeded had three years or less of experience
with ANTRAC. Ten of the 11 lived in Diabo where traction has been used for 30 years.

Plowing Weeding Ridging
Number of Number of Percent Number of Percent Number of Percent
Type of Farmers Farmers Who Farmers Who Farmers Who
Draft in Who Owned Used Who Owned Used Who Owned Used
Animal Sample Plows Plows Weeders Weeders Ridgers Ridgers
Donkey 52 52 96.2 16 37.5 6 33.3
Oxen 72 72 88.9 25 68.0 25 68.0

No. of Years of Type of No. of Percentage of Farmers Who:
Experience With Draft Farmers
Animal Traction Animal in Sample Plowed Weeded Ridged
0-2 Years Donkey 12 83.3 16.7 16.7
3-6 Years Donkey 29 100.0 27.6 20.7
7-28 Years Donkey 11 100.0 27.3 9.1
Sub-Total Donkey 52 96.2 25.06/ 17.0-2 Years Oxen 42 81.0 9.5 9.5
3-6 Years Oxen 14 100.0 35.7 42.9
7-28 Years Oxen 16 100.0 56.3 50.0
Sub-Total Oxen 72 88.9 25.0 25.0d/
a/Including seven farmers who used their donkey plows to weed.
b/Five farmers used their donkey plows to ridge. C/One farmer used his ox plow to weed. d/one farmer used his ox plow to ridge.

to w ed. Furthermore, in order to weed and ridge with ANTRAC, farmers must learn to p ant in rows. But farmers are reluctant to plant in rows because it involves line,* tracing, an additional time-consuming operation. More importantly, weeding requ res well-trained animals to walk between the rows without damaging the plants. Also, the weeding equipment is expensive. For oxen traction, for
examp le, the 24,460 FCFA price of a five-tined weeder and a ridger is 3,170 FCFA
greater than the price of a nine-inch plow..!/ The three-tined donkey weeder cos s 10,510 FCFA but many farmers prefer using their donkey plow to avoid the expense 'and because they feel their donkeys are too weak to pull the weeder. The use of a plow is not as efficient as the use of a 3- or 5-tined weeder.2'
Table 4.10 presents the percentage distribution of fieldwork time for donkey~ and pairs of oxen. Assuming that animals are used an average of 4 hours per working day, about half of the farmers used their donkeys and oxen less than 50
houns or 12 days during the entire 1978/79 survey year. This is an extremely low rat( of utilization and it has important cost recovery implications because the anin als must be fed throughout the year whether they work or not.
Table 4.11 reveals that donkey farmers plowed a slightly larger area than oxei farmers. Also, donkey farmers plowed 85 percent of their total area under cul ivation while oxen farmers only plowed 59 percent. Not only were surveyed
oxen! farmers less experienced with ANTRAC, but it also takes longer to learn how to .se the ox traction package.
Oxen farmers weeded and ridged a greater proportion of total area than donkey farmers. Farmers in the two zones with the most experience with ANTRAC, Diarbo and Diapangou, weeded and ridged the largest proportion of their land.3 For farmers who performed the operations in these zones, oxen weeded 46 percent of cultivated area and donkeys 39 percent; oxen ridged 52 percent of total area while donkeys ridged 46 percent.
Farmers did very little custom field work. Donkey farmers spent less than 5 percent of their time on custom work whereas oxen farmers spent less than 14
1/1981 EORD prices for COREMMA equipment: 5-tined weeder, 18,410 FCFA; and a ridger, 6,050 FCFA.
2Farmers using a six-inch plow should theoretically make at least five pauses in a row to weed the 80 cm. width, but since the earth thrown up by the pl covers weeds, normally only three passes are used.
_3/Farmers from Diabo and Diapangou had an average of 5.5 years and 6 years of experience with ANTRAC, respectively.

Number of Percentage of Draft Units Which Worked:
Draft Animal
Type of Units in 0.5 to 25 to 50 to 100 to Over 150
Draft Animal Sample 0 Hours 25 Hours 50 Hours 100 Hours 150 Hours Hours
Donkey 61 1.6 18.0 26.2 29.5 21.3 3.3
Oxen 80 11.3 8.8 25.0 33.8 15.0 6.3
-/The Piela zone is excluded because of unreliable data during the rainy season.

Average Percentage Average Percentage Average Percentage
Type of Area of Total Area of Total Area of Total
Draft Plowed Area Weeded Area Ridged Area
Zone Animal (ha) Plowed (ha) Weeded (ha) Ridged
Diabo Oxen 5.01 67.4 1.20 16.2 1.06 14.2
Ougarou Oxen 0.69 13.4 0 0 0 0
Sub-Total Oxen 4.06 59.2 0.94 13.7 0.82 12.0
Piela Donkey 4.88 94.2 0 0 0.30 5.7
Logobou Donkey 3.70 75.4 0.52 10.6 0.31 6.3
Diapangou Donkey 5.11 80.5 0.99 15.6 1.35 21.3
Sub-Total Donkey 4.53 84.8 0.52 9.8 0.65 12.2

percent. Plowing was the only type of custom work performed. Since seedbed
preparation must be completed in a relatively short time, most farmers do not have time available for custom work. Income from custom work was likewise
minimal, averaging only 63 FCFA for donkey farmers and 530 FCFA for oxen farmers.
4.4.3. Equipment Repairs
During the survey year, 15 farmers had difficulty using their equipment because of breakdowns. Oxen traction farmers spent an average of 1,380 FCFA f or repairs and donkey traction farmers spent 1,264 FCFA on repairs.
4.5. Use of Animal Traction for Transport
Sample farmers showed a strong preference for donkey carts. Thirty-five
farmers in the sample owned donkey carts and only seven owned oxen carts." The
preference for donkey carting is also reflected in Table 4.12 which shows that donkey carts are used over four times as much as oxen carts, Most striking,
however, is the fact that of total time worked by the different animals, 49 percent of donkey time was spent carting whereas oxen spent only 4 percent of their time carting.
Donkey Carts Ox Carts
Number Owned 357
Average Hours Worked Per Cart
During 1978 136 29
Some farmers specialize in carting wood for fuel and construction. Wood
transport accounted for 52 percent of total time spent carting. The transport of agricultural produce accounts for another 21 percent.?. Construction material, water, and forages each accounted for between 5 and 7 percent of total time
"/This includes five farmers who owned both a donkey cart and an oxen cart. Two preferred to use their donkey cart and three preferred an oxen cart.
.?.This includes transport of the farmer's own agricultural produce as well as custom transport of the produce of other farmers.

carting. Most carting (45 percent) is done during the dry season when farmers build houses, water is scarce, and farmers have slack time. Seventy-two percent of the time spent carting in the dry season was spent carrying wood.
iDonkey transport was concentrated in Logobou and Diapangou. Logobou is isolated and farmers rely on carts for much of their transport work. Diapangou is i n a relatively more monetized region on the main road and has been served by extension activities for a long time. The 20 carts in these zones accounted for 80 p .rcent of the carting time and 84 percent of the revenue earned from carting in or total sample.
Table 4.13 reveals that the rental of carts accounts for only 18 percent of the total carting time..!/ The average cash income from rental of donkey carts was 1,981 FCFA; and 743 FCFA for ox carts. This seriously underestimates the inclie generated by carting because it does not include revenue in-kind from carting or income from the sale of the products transported. For example, the majol- revenue from carting is from firewood which was either sold directly or used by the farmer as fuel to produce other products, such as dolo or bean cakes. MostI of the recorded carting rental income (88 percent) came from the transport of drops, baggage, and construction material.
Donkey Carts Ox Carts
Average Time Worked Per Cart Per Year (Hours) 136 29
Per engage of Time Used on Personal Work 72.4 80.8
Percentage of Time Loaned Out 9.4 8.4
Per entage of Time Rented Out 18.3 10.8
I/Only 16 of the 35 donkey carts were rented; these 16 carts earned an aver age of 4,334 FCFA during the year. Only two of the seven oxen carts were rented and they earned an average of 2,600 FCFA. Rental rates for donkey carts were lower (87 FCFA/hour) than those for oxen carts (236 FCFA/hour), although thelatter is based on only two observations. Oxen carts have twice the capacity of the 500 kg donkey carts.

Aside from hauling the farmer's own produce and firewood, the rationale for carting is that it generates income, increases the utilization of draft animals during the non-agricultural season, and helps keep the animals trained. Table 4.14 shows that the average total time worked by animal units during the year is more than doubled by using a cart. Moreover, farmers with carts use their
animals throughout the year whereas farmers without carts only use animals during the rainy season. This finding would seem to recommend a further promotion of the use of carts. However, the market for carting services in an area can easily become saturated. Farmers should carefully study the market for transportation services before they invest in acart.Table 4.14 EFFECT OF CARTING ON UTILIZATION RATES OF ANIMAL UNITS IN 1978-79
Average Total Hours Worked Per Animal Unit May-July Aug.-Oct. Nov. 1978- Feb.-April 1978 1978 Jan. 1979 1979 Total
Farmers With Carts 55.9 26.4 25.1 47.2 154.6
Farmers Without Carts 53.4 10.6 -- 64.0
"/There are several important factors to take into consideration, including the number of carts in the area, the demand for carting services, and the state
of the roads in the area. In Soudougui, farmers reported that they were no
longer using their oxen carts because of poor roads. In Bogande, there are more good roads and farmers seem satisfied with their oxen carts.

4.6. Estimated Production Effects of Animal Traction
The purpose of this section is to estimate the effect of ANTRAC use on cult vated area, crop mixture, and yields based primarily on the 1978-79 farm survey data. In addition, data from the 1979 plowing and fertilizer field trials are sed to provide better controlled estimates of the yield effect of ANTRAC.
Caution is necessary in interpreting the survey results. An "ideal evaluation' of animal traction would follow particular farmers over a 5-10 year time span In this study, however, we have had to rely on comparisons between hoe and ANTR C farmers based on data from a single farm year, 1978-79. This has complicate the analysis in two ways. First, the impact of ANTRAC can only be inferred from cross-section comparisons between ANTRAC farmers and hoe farmers. Causality is difficult to establish because performance differences between the two subsamples may not be attributable to ANTRAC alone. Due to the sampling procedure, surveyed ANTRAC farmers tend to be more educated, innovative, and successful than typical hoe farmers. Second, the large size of the Eastern Region and
its idely diverse agro-climatic, demographic, and socioeconomic characteristics all contribute to substantial variation in key performance indicators. Thus, som performance indicators vary more between zones than they do between hoe and
traction farmers within each zone. Despite a large total sample size, the
relatively small zonal cell sizes (particularly for area and labor data) make it difficult to separate the performance effects of ANTRAC from these exogenous factors. For this reason, multiple regression and analysis of variance were used whe e necessary to control for the effect of non-ANTRAC variables.
4.6 1. Area Effects of Animal Traction
As shown in Table 4.15, modest increases in cultivated area were found among ANTRAC households. Total area cultivated (column 3) is substantially higher for ANTAC farms than for hoe farms, but this is primarily due to the larger overall household size. Both total family size (column 1) and the number of active wor ers1 per household (column 2) are appreciably higher for ANTRAC. Traction households cultivate 5 percent more land per capita (column 4) than their hoe counterparts, but this difference is not statistically significant.-/
I/Defined as persons from 15 to 54 years old.
2/Unless otherwise noted, all tests of statistical significance in this re ort are based on a 95 percent confidence level.

1 2 3 4b 5b 6
Number of Number of Total Area Area Area Percentage
Household Active Workers Cultivated Cultivated Cultivated Change in Area
Members (Persons per per House- per per Active
Type During 15-54 Years Household hold Member Active Worker Worker Due
of Rainy Season Old) (Hectare) (Hectare) (Hectare) to ANTRAC
Diabo Oxen 6.83 11.34 3.08 5.16 3.98 7.44 0.63 0.70 1.40 1.49 + 6.4
Ougarou Oxen 7.92 12.10 3.92 6.20 3.61 5.15 0.48 0.46 0.95 0.85 +10.5
Sub-totalc Oxen 7.09 11.52 3.28 5.41 3.89 6.89 0.59 0.64 1.29 1.34 + 3.9
Pieia Donkey 6.33 10.67 2.83 4.44 3.56 5.18 0.59 0.50 1.29 1.32 + 2.3
Logobou Donkey 9.48 11.73 4.54 3.91 4.68 4.91 0.47 0.54 l.ll 1.51 +36.0
Diapangou Donkey 7.75 12.00 3.33 4.70 4.12 6.35 0.57 0.56 1.39 1.43 + 2.9
Sub-totalc Donkey 8.26 11.53 3.81 4.24 4.26 5.34 0.53 0.54 1.22 1.44 18.0
TOTAL ALL FARMERSc 7.68 11.53 3.55 4.83 4.08 6.10 0.56 0.59 1.26 1.39 10.3
aThese figures are calculated only for the households for which there was area data, which includes only two thirds of the sample in each zone. Thus, they differ slightly from those in Table 2.1 which provides estimates for the entire sample. Donkey or horse farmers in areas %here oxen traction predominates were excluded. Thus, 1 farmer from Ougarou and 4 from Diabo are eliminated. Likewise, oxen farmers in areas where donkey traction predominates were excluded so that 3 from Piela, I from Logobou, and 2 from Diapangou are eliminated. Another farmer is missing from Ougarou because he had no area data.
bThe zonal averages for Column 4 and 5 are the averages of each household's value. In earlier reports we presented a figure which represents the zone treated as one farm, i.e. we have divided the average farm size by the average number of household members and active workers to arrive at the figure.
CFor the sub-totals, and totals, village averages were weighted differently because sub-sample sizes differed. See footnote 1, page 13.

A better indication of the area effects of ANTRAC is area per active worker (col .mn 5), since adult labor is the major input in the region's agriculture and is the primary constraint on farm production. As can be seen in column 6, ANTRAC
is a sociated with a 4 percent increase in area per active worker for the oxen zone. Due to poor performance in the Ougarou oxen zone where there was actually a decrease in area per active worker of about 10 percent, the overall increase for pxen traction is not statistically significant.-' The increase in donkey zone of 18 percent is statistically significant as is the increase of 10 percent for raction farmers overall.
In fact, differences in acreage per active worker vary more between geograp ical zones than they do within each zone when comparing ANTRAC and hoe farmeIrs. Farmers in the Diabo zone, for example, cultivate 60 percent more area
-per orker than farmers in Ougarou. The possible effect of agro-climatic or locational variables in determining area cultivated per worker explains our use of within-zone paired comparisons between hoe and traction farmers.
It was expected that the impact of ANTRAC would be greater for more experienced users. Surprisingly, analysis of the survey data did not demonstrate a statistically significant relationship between cultivated area per active worker and Lhe number of years of ANTRAC experience. For practical reasons, our sample was not specifically designed to facilitate analysis of the effect of ANTRAC exp rience. Given the short history of ANTRAC and the geographical pattern of
its implementation in the EORD, most farmers tend to have the same level of exp rience in a given zone.- Thus, it was not possible to select a subset of suc essful ANTRAC farmers that was stratified by both zone and experience. In add tion, variation in performance was high even within a given experience category and zone.-/ Other indicators of experience were analyzed, but none indicated a correlation between experience and superior performance.
Due to the limited use of weeders among sampled farmers, only a small amount of data is available to estimate the area effect of the complete ANTRAC package.
1-The decrease in area per worker in Ougarou occurs for reasons mentioned ear ier--lack of experience with the technology and lack of extension support.
2/For example, in Ougarou, 90 percent of ANTRAC farmers had only 2 years of experience or less.
3/Within the one- to two-year experience category in the Lantaogo village of the Diabo zone, the area per active worker ranges from 0.74 to 3.1 hectares.

Nonetheless, for the 12 oxen farmers with complete weeding data, the area cultivated per worker was 22 percent above that for oxen farmers who did not weed with draft power. Such an area effect was not found for donkey traction. Based on only 9 farms with donkey weeding data, the area per worker was 4 percent below
that of non-weeders. This result is probably not representative of donkey weeding under optimal conditions, however, since most of the donkey weeding was
performed with a plow.-' By contrast, most oxen weeding was conducted with a weeder. In addition, farmers weeding with oxen had more experience than those using donkeys (11 versus 6 years of experience, respectively).
In summary, the evidence of an acreage effect due to the use of ANTRAC technology is mixed. While we have shown a 10 percent average increase in area
cultivated per active worker for animal traction users, this varies substantially among zones and between donkey and oxen farmers. The effect of ANTRAC on area cultivated was negative or minimal in those zones where ANTRAC has been recently introduced. For oxen farmers, the use of weeders was strongly related to increased area per worker while this was not demonstrated for donkey farmers practicing weeding.
4.6.2. Changes in Cropping Emphasis-;-'
Since sorghum and millet are the major staple foods in the region, the area planted to these crops is determined primarily by the food consumption requirements of the household. The large area in these crops is the most significant aspect of cropping mixtures in the Eastern Region. Area planted to these crops
accounts for almost 80 percent of area cultivated for hoe farmers and 75 percent of area cultivated for traction farmers (see Table 4.16). In all zones except Logobou, traction farmers have larger absolute areas of these two cereals in order to feed their larger families. However, the proportion of total area planted to cereals is slightly smaller in all zones except Diabo.-/
lAs discussed in Section 4.4, weeding with a plow is much slower than when a weeder is used.
2/For further discussion of cropping patterns and the characteristics of individual crop enterprises, see Lassiter (1981).
3/The higher proportion of area planted to sorghum and millet in Diabo may be due to the fact that fields in the Diabo traction sample are generally much
older and less productive than those for any other zone (see Table 9, p. 16 of Baker and Lassiter, 1980).

Oxen Zones Donkey Zones Averages For
Oxen Donkey
Diabo Ougarou Piela Loobou Diapangou Zones Zones All Zones
Crop Hoe Antrac Hoe Antrac Hoe Antrac Hoe Antrac V.6 Antrac Hoe Antrac Hoe Antrac Hoe Antrac
Sorghum/MilJet 77.8 79.4 81.8 71.3 78.9 70.9 77.2 67.3 90.8 78.4 79.1 77.5 81.0 71.8 80.1 74.7
Maize 1.0 2.5 11.2 9.6 3.5 3.8 2.2 2.3 3.2 3.6 3.3 3.8 2.7 2.9 3.0 3.4
Rice 2.1 2.2 1.5 7.4 2.2 0.8 3.5 2.9 0.3 1.4 1.9 3.5 2.4 2.0 2.2 2.8
Groundnuts 13.0 8.1 1.4 2.0 14.1 21.8 9.8 11.6 1.8 6.1 10.3 6.8 8.8 12.4 9.6 9.6
Bambara Nuts 1.5 0.5 0.9 0.8 1.0 1.0 0.8 0.8 0.6 0.4 1.3 0.5 0.8 0.6 1.1 0.5
Soybeans 0.5 3.2 0.5 3.2 0.2 0.4 0.3 4.7 0.8 4.5 0.5 3.9 0.4 3.6 0.5 3.8 co
Cotton -- 2.6 0.4 0.8 -- -- 0.6 2.7 -- 0.1 0.1 2.1 0.3 1.7 0.2 1.9
Tubers 0.6 0.6 0.7 1.7 -- -- 2.3 2.0 0.4 0.5 0.6 0.9 1.4 1.0 1.0 0.9.
Gumbo -- -- 1,0 1.6 0.1 0.1 2.2 0.6 0.1 0.5 0.2 0.2 1.3 0.5 0.8 0.3
Other Crops 3.3 1.0 0.6 1.6 -- 1.1 1.1 5.1 1.8 4.5 2.7 1.0 1.0 3.4 1.8 2.0
Cash Cropsb 15.6 16.1 3.4 13.3 16.5 23.0 14.2 21.9 3.0 12.0 13.0 16.7 12.0 19.7 12.5 18.0
Legumesc 15.0 11.8 2.7 6.0 15.3 23.2 10.9 17.1 3.2 11.1 12.3 11.6 10.0 16.7 11.1 13.8
aThese figures were weighted in the same manner as Table 2.1.
b;Cash Crops" are rice, peanuts, soybeans and cotton.
CLegumes are peanuts, Bambara nuts and soybeans.

Overall, there is little major difference between the cropping mixtures of
ANTRAC and hoe households. The largest increases in cultivated area are in soybeans, cotton, and rice. However, the shift into cash crops-/ is very slight, representing an increase in the proportion of area from 13 percent for hoe
farmers to 18 percent for traction farmers. Further, the total area of cash crops per household is small--1.2 ha for ANTRAC and 0.5 ha for hoe farms.
4.6.3. Yield Effects
The estimates of the yield effect of ANTRAC are based on data from both the 1978-79 farm survey and from 1979 plowing and fertilizer field trials. The yield estimates presented here are weighted average yields.-/ These yield estimates are lower than yield plot estimates presented in an earlier report (Baker and Lassiter).3/
The crop yields presented in Table 4.17 are generally higher for ANTRAC subsamples for most crops except soybeans. However, because of the small sample size for minor crops, only maize yields in all zones and groundnuts in oxen zones represent statistically significant yield increases. The most striking feature concerning yields is that they are quite low in general and dramatically so in the donkey zones, where half the sample (Piela and Diapangou) suffered severe drought. This drought effect should be considered when evaluating the relative performance of the donkey traction technology itself.
Average yield differences between hoe and traction farmers do not provide conclusive evidence of the impact of ANTRAC because they do not control for many
1Defined as rice, groundnuts, soybeans, and cotton.
2-Computationally, the weighted average yield represents the total harvest of a given crop for all households in a specific zone divided by the total area in that crop. An alternative estimate of yields is the average of the yields of each household without regard to the area cultivated by each household. The
latter method gives high estimates because the high yields of small farms with only small compound fields are weighted the same as the lower yields of large farms using more extensive cultivation practices. The former method has some intuitive appeal because it provides a yield estimate for an "average hectare." Only unweighted yields are used to compute variances for tests of significance.
3/Subsequent analysis has led us to believe that yield plots overestimate yields for a variety of reasons--the "border effect," placement bias favoring better sections of a field or better fields, harvest error by the farmer, or "lost" plots in abandoned fields which are excluded from the analysis.

Zones Oxen Zones Donkey Zones
Millet and Sorghum 466 468 555 554 377 381
Groundnuts 213 238 59 179 366 296
Maize 425 686 500 746 349 585
otton 108b 171 118b 253 97b 88
ice 442 465 329 630 554 300
oybeans 283b 197 241b 294 324b 99
aYields presented here are weighted averages per hectare based on stimates of total household production in 1978-79. They are alculated only for the two-thirds random subsample of farmers for which complete cultivated acreage was measured.
bThese estimates are based on a small number of observations epresenting less than one hectare of cropland per zone.

factors that can obscure the results.-/ In addition, the yields for ANTRAC farms presented in Table 4.17 are averaged across all cultivated fields, whether or not they were plowed or weeded with animal power. Regression analysis of sorghum/millet yields showed that yields were positively related to number of workers and labor inputs per hectare, and negatively related to farm size; however, the use of animal traction was not significantly related to yields.
The potential yield effects of animal traction were also evaluated based on
data from field trials conducted in 1979. The effects of plowing and the
application of natural phosphate fertilizer were tested in controlled experiments under farm-level conditions on 19 peanut fields and 24 sorghum/millet fields.-2 The farmers chosen were a subsample of the 1978-79 farm survey's ANTRAC sample. Farmers from each of the five traction zones were represented.
Trial plots of 360 square meters were set out in each farmer's peanut and sorghum fields, and then divided into four subplots measuring 6 meters by 15 meters. Before the harvest, a yield plot of 3 meters by 3 meters was delimited in the center of each subplot. The four subplots consisted of: (a) a control plot which was scarified manually with a hoe; (b) a second plot which received a broadcast application of 150 kilograms per hectare of natural phosphate3/ which was incorporated in the soil by manual scarification with a hoe; (c) a third plot which received the same amount of phosphate incorporated by animal plowing; and
(d) a fourth plot which was plowed using animal traction, but received no phosphate.
As shown in Table 4.18, Treatment 4 (animal plowing only) produced an average increase in peanut yields of 18.2 percent. This amounts to an average 105.9
-In particular, the large agro-climatic variability makes yield comparisons difficult, despite the paired zonal comparison method used. Fields within even one kilometer of each other can receive substantially different amounts and
patterns of rainfall.
2-/Originally, 42 farmers participated in the trials. However, reliable harvest data were obtained only for the 24 sorghum fields and 19 peanut fields.
-/The natural rock phosphate used comes from Kodjari in the Diapaga sector of the EORD. It is composed of 64 percent Tricalcium phosphate and has 30 percent content in phosphorus pentoxide (P 0 ) and a 40 percent content in quicklime or calcium oxide (CaO). It is cvughed to a fineness such that 90 percent of the particles are smaller than 0.09 millimeters in diameter. The
solubility of the phosphate varies between 25 and 33 percent per year depending
on rainfall and other factors.

Sorghum/Millet Groundnuts
Confidence Confidence
Interval Interval
at 95% at 95%
Treatment Certainty Treatment Certainty
reatment Results Level Results Level
yield of control plot (land prep. by hand no fertilizer) Kg/Ha 531.0 +117.5 582.1 +128.8
land prep by hand; 150 Kg. phosphate. Percent increase over control plot yield 30.2% + 20.5% 11.4% + 11.4%
ANTRAC plowing and 150 Kg. phosphate. Percent increase over control plot 65.0% + 26.2% 26.8% + 21.1%
ANTRAC plowing, no fertilizer. Percent increase over control plot 16.7% + 14.6% 18.2% + 12.4%
Source: EORD BPA Plowing/Phosphate Trials, 1979.

kg/ha increment in yields which would give the farmer an increase in gross income of 5,749 FCFA per hectare.- For sorghum and millet, there was an average increase of 16.7 percent in yields due to plowing. The yield increment was 88.7 kg/ha which would give a gross return to the farmer of 3,548 FCFA per hectare.2/
There is also a significant yield increase due to plowing under natural phosphate fertilizer (Table 4.18, Treatment 3). This fertilizer is readily available and inexpensive in the Eastern Region.- The recommended method of
application is incorporation by plowing. Plowing under 150 kg of the phosphate
resulted in an average 65 percent increase in yields of sorghum and millet, and a 26.8 percent increase for groundnuts. This represents a yield increment of 345.2 kg of sorghum and millet which would give the farmer a net benefit4/ of 10,808 FCFA per hectare. For groundnuts, the yield increment was 156 kgs/ha which represents a potential 5,469 FCFA net profit per hectare. These results are
similar to those found elsewhere in Upper Volta (Bikienga et al., 1980; Stoop and Pattanyak, 1980).
4.7. Impact of Animal Traction on Household Labor Allocation
In this section we evaluate differences between hoe farmers and traction farmers in their use of labor in farming activities and in the relative amount of time spent on farm and nonfarm enterprises.- First, we examine the impact of animal traction on the hours spent on various field activities. Key issues here include: (1) is there a savings in field labor time, as measured in worker
-/This calculation is based on the official price of 54.29 FCFA/kilogram of groundnuts.
2/Based on official (OFNACER) base price of 40 FCFA/kg.
3/The official EORD price for natural phosphate is 20 FCFA/kg.
4/These calculations of "net" benefit assume no cost for plowing with animal traction. An economic analysis of fertilizer and plowing, including the costs of ANTRAC use, is provided in Section 5.4.4.
5/This section is based on analysis of detailed labor data collected for one-third of the farm households surveyed. Three different questionnaires were
administered weekly to collect data on labor spent on field activities, livestock herding and care, and the overall allocation of family time. Section 4.7.1. is based on the first two questionnaires, and Section 4.7.2. on the third, which is less accurate since it recorded data only for the day previous to the interview (i.e., a one day per week labor enumeration).

equivalent hours- per hectare; (2) does the seasonal pattern of labor use for
cropping activities shift as a result of using animal traction; and (3) in which activities are labor requirements increased or decreased. Second, we examine the
distribution of total household labor. The question is whether traction farmers allocate relatively more or less labor to non-cropping activities than do hoe farm rs.
4.7.). Allocation of Household Labor to Cropping Activities
Table 4.19 presents a summary of the weighted average worker equivalent
(WE) hours per hectare allocated to the three principal labor activities--seeding, soil tillage,-/ and harvest--for hoe, oxen, and donkey households. On aver Lge, ANTRAC households devoted 174 WE hours less labor per hectare than hoe hous holds. This represents a reduction of 25 percent in average labor time per hectare. Sixty-eight percent (or 119 WE hours per hectare) of this reduction occurred in the category of soil tillage. As one might expect, the average labor redu tion per household is substantially greater in oxen zones (31 percent) than in donkey zones (20 percent). As shown by multiple regression analysis, the labo- reduction among ANTRAC households was statistically significant, cont-olling for other variables.
The savings in field labor associated with ANTRAC are somewhat offset by the additional labor required to feed and maintain draft animals. Delgado (1979) conc uded that the absolute level, timing, and quality of the added labor for oxen traction are prohibitively high. Analysis of the 1978-79 Eastern Region survey y data (Lassiter, 1982)-/ indicates, however, that the labor requirements
-iWorker equivalent man-hours are calculated by weighting hours worked by different age and sex categories by a coefficient of work productivity. See
footnote b, Table 4.19, for the weighting coefficients used.
2/soil tillage includes soil preparation (including plowing, weeding, and ridging), whether done by hand or with ANTRAC.
!/Based on livestock maintenance data from 12 oxen traction households and 12 onkey traction households. Weekly labor was collected from only 41 of the 125 surveyed ANTRAC households. Of these 41, 21 owned only draft oxen, 14 owned only donkeys, and 6 owned both. Unfortunately, it was not possible to analyze the labor data from 15 oxen households and 8 donkey households. This is mainly bec use the questionnaire used to record animal care labor data did not usually distinguish labor for draft animal care from labor for herding other animals, particularly range cattle. In addition, many of these unanalyzed households kept their draft animals on the range and made little or no use of them as work animals. The donkey data are estimated from 16 donkeys, averaged on a per donkey basis over 12 households. Half the donkeys included were used for carting as well as for field operations.

Page 65
Missing From
0 0

Page 66
Missing From
0 0

20 19
18 II ///
it ILeft Column: Hoe Farmers Planting
17 II Right Column: Traction Farmers
Soil Tillages
16ifLL J
15 j i Harvesting
14 a l I I
~It Il i ether
13 I i I I.
liltl 1,1 Ii ..'"i
4J 6l
11 1 ll .1111
aI 1.1 1 11 g ,, ll ""
. 3 I I I:I I Il I I I
1 1Ill III I "
5/1- 5/29- 6/26- 7/24- 8/21- 9/18- 10/16- 11/13- 12/11- 1/8- 2/5- 3/5- 4/25/28/78 6/25 7/23 8/20 9/17 10/15 11/12 12/10 1/7/79 2/4 3/4 4/1 5/30/79
aOther includes land preparation, building gralneries, etc. Also, if man-hours worked on a major activity such as planting
represented less than 10 percent of the hours worked in a month, that activity was included in other.

sub amples are remarkably similar throughout the year. Over 50 percent of the ann al hours per hectare are worked during the three peak months (May 29 through Aug st 20) in both subsamples (54 and 53 percent for hoe and traction households, respectively). The proportional savings in labor time per hectare for traction households is slightly greater in the peak season than for all months during the yea-. There is a slight shift toward periods 3 and 4 (late June to late August) amoig traction households as compared to hoe households whose labor use peak occurs during periods 2 and 3 (late May to late June). This suggests that
greater labor savings may be possible as more traction farmers adopt ANTRAC weeding, which should save labor in periods 3 and 4.
The histograms in Figure 4.1 further show that the cropping season lasts through early December. Planting is done primarily in periods 1 to 3 (through lath July). Soil tillage is the dominant use of labor through mid-September. Har vesting of early crops begins to be a major demand on labor time in mid-September, continuing through mid-December. During the remainder of the year nearly all time is spent on other cropping-related activities, including land clearing, drying and storing grain, constructing grain storage huts, etc.
4.7.2. Proportional Allocation of Household Labor
to Farm, Non-Farm, and Leisure Activities
The allocation of potential labor time to leisure, farm and non-farm activities is shown in Table 4.20. Part A presents a simple breakdown into total hours worked and hours spent resting, walking, ill, or visiting a local market. Figur s are presented for both oxen and donkey zones since there are important zonal variations in the way work time is allocated. Because of the small sample sizes and the one day per week enumeration of data, the aggregated figures in the first tw columns are the most reliable indicators of differences among the subsamples. In terms of the proportional allocation of work time, there is very little difference between the subsamples. As would be expected, members of traction households spend a slightly greater proportion of their time on livestock raising and agricultural trading. However, contrary to expectations, traction households allocated a greater proportion of their work time to household fields, but le~s to household chores, agricultural transformation and other activities.
As can be seen in Table 4.20, women in animal traction households spend a greater proportion of their working time on household fields as compared to women in households using hoe tillage. They also spend proportionally less time on ho sehold chores than do women in hoe households. In large part, this apparent effect of ANTRAC can be attributed to the differing demographic structures of the

Zones Oxen Zones Donkey Zones
Number of Households 36 41 12 23 24 18
A. Hours (Hours per Day)
Hours Workedc
Men 5.5 5.4 5.5 5.8 5.4 5.0
Women 7.4 5.8 7.3 5.8 7.5 5.7
Household 6.5 5.6 6.5 5.8 6.4 5.4
Hours Resting, Walking, Ill
Men 6.5 6.4 6.5 6.2 6.6 7.0
Women 4.6 6.2 4.7 6.2 4.5 6.3
Household 5.5 6.4 5.5 6.2 5.6 6.6
B. Proportional Allocation of Hours Worked (Percent)
Household Fields
Men 57.4 56.2 59.1 62.0 55.6 50.4
Women 26.7 35.6 27.6 44.5 25.8 26.6
Household 40.5 44.1 41.0 52.3 40.0 35.8
Livestock Raising
Men 12.5 14.4 11.1 19.7 14.0 9.2
Women 2.7 4.2 4.4 6.3 1.0 2.2
Household 6.9 8.6 7.2 12.3 6.7 4.9
Household Chores
Men 3.7 3.5 4.3 2.8 3.1 4.2
Women 46.0 35.2 46.8 31.2 45.1 39.2
Household 27.7 22.0 28.7 18.4 26.6 25.6
Agricultural Transformation
Men 0.5 0.8 0.2 1.0 0.8 0.6
Women 18.5 14.8 18.2 14.4 17.8 15.1
Household 10.4 8.9 10.5 8.4 10.3 9.5
Agricultural Trading
Men 2.9 5.4 0.9 1.3 4.8 9.1
Women 0.9 3.9 0.2 0.4 1.6 7.4
Household 1.8 4.4 0.5 0.8 3.0 8.0
Men 23.0 19.8 24.4 13.2 21.7 26.5
Women 5.7 6.3 2.8 3.2 8.7 9.5
Household 12.7 12.0 12.1 7.8 13.4 16.2
potential labor time is defined as a 12 hour day; the daylight hours in the EasternRegion. Leisure is derived as a residual of time not accounted for on work activities.
bFigures for sub-samples within zones are weighted averages for all people in the sub-sample
households. Figure for combined zones is a simple average since zonal effects are important but subsamples are unequally represented in each zone. Note that this only partially corrects for agroclimatic effects within sub-samples since it does not eliminate variance due to different village characteristics within zones.
CActual hours are used in this analysis rather than man-hours since work productivity indices were not obtained for non-cropping activities.
dlncludes work on fields of others, livestock raising for others, wage labor, construction, artisanal activities, small industry, and schooling.

sample households. Traction households, as noted above, are larger and have more women to share household chores. To the extent that the time, needed for household chores increases less than proportionally to increases in household size, each woman in the larger ANTRAC households can spend less working time on household chores. Despite the increased proportion of women's work time going to fiel work, the total hours of fieldwork by women, as well as by men, is lower in ANTR C households. The reduction in fieldwork labor is slightly greater for men, 30 percent, as compared to a reduction of 21 percent for women.
A final labor allocation question often raised in relation to animal traction programs is whether animal traction enables a more even distribution of laboY over the year. In Figure 4.2, seasonal indices for labor inputs in household field activities, non-cropping farm activities, non-farm activities and tot 1 labor time are shown for hoe versus traction households. It can be seen tha the labor profiles for the two subsamples are very similar. Labor used on
houIehold fields and non-farm activities is more evenly distributed over the year in raction households. The coefficient of variation for monthly labor inputs to hou ehold fields is .68 for traction households and .79 for hoe farming householis. The coefficient of variation for non-farm activities is .59 and .69, respectively. These figures tend to support the view that animal traction reduces labor requirements for household fields during the peak season, enabling a rore even distribution of labor for non-farm activities. On the other hand, labor for non-cropping farm activities is less evenly distributed among traction ho seholds.
Turning to the distribution of all household labor inputs over the year, it ca be seen that total labor is more evenly distributed than any of the component
caIegories. This results from the fact that non-farm labor is primarily counter-seasonal to cropping labor. Thus, the coefficients of variation for total la1Ibor inputs are .19 for hoe farmers and .21 for traction farmers. This suggests th t animal traction only slightly reduces the seasonal variation of labor inputs in o cropping and non-farm activities.
In summary, the main impact of animal traction on household labor allocation is a significant reduction in total WE hours worked per hectare in crop production, even when the labor for draft animal care is taken into account. While tiere is a slight shift in the peak cropping season for traction households, the pofile of the agricultural calendar remains quite similar for both subsamples. Regarding total household labor allocation, there again appears to be little d fference among subsamples. It appears that traction households spend slightly m re of their work time on livestock and agricultural trading.

------------------------------ Traction
A. Total
B.0on-ar 150 100
50N 0
C. Non-piFarm
C. Householdin Fld 25
200 52-62-72-82- 91- 1/6 11-1/1 /~25 42
5/5/28/7 6/2 7/2 8/2 1 9/1 8 10/1 6- 2 11 12/11 1/7/ 2/ 3/4 4/ 2-/3/7

5.1. Introduction
n this section, financial and cash flow analyses are used to evaluate the econo ics of oxen traction, donkey traction, and traditional hoe technology at the f rm level. These analyses are based on EORD survey data from 110 ANTRAC and 106 hbe farming households from the 1978-79 crop season.- In addition,,the
mediu -term financial impact of animal traction is assessed on the basis of 10year income projections under various assumptions of current and potential ANTRA performance. The components of total farm and total household incomeY2
analy ed in this chapter include: (a) the actual value of cash transactions,
(b) tie imputed value of unsold farm and nonfarm production,-/ and (c) the imputed values of in-kind costs.
creage and yield data, first discussed in the previous chapter, have been used to derive the total value of crop production per household in 1978-79, as pres nted in Table 5.1. The value of each of the major crops is presented sepa ately in order to show its relative economic importance. The prices used are eighted average sales prices for the 1978-79 season. The relative share of the different crops in the value of production is similar to that demonstrated by the allocation of cultivated acreage to each. Sorghum and millet are by far the most important, contributing 63 to 85 percent of the total value of crop prod ction for each of the four subsamples. Their reduced importance in the donkey zones tends to reflect the localized effect of drought on sorghum and millet yields. Among hoe farmers, cowpeas, groundnuts, maize, and rice are the next most important crops economically. These crops are of similar importance to ANTR C farmers, except that the order is slightly different: rice, cowpeas,
group nuts, and maize.
1For more details on these technologies, see Chapter 2 for an overview of regional farming systems and sections 4.6 and 4.7 for detailed descriptions of field crop statistics and household labor allocation. Also see the bibliography for a full listing of previous farm survey reports.
2Total farm income refers to all income attributable to the following farmrel ited activities: crop production, livestock raising, crop and livestock track ing, and crop processing. Total household income refers to income generated by all activities, farm and nonfarm.
!/The largest component of unsold production consists of crops directly consumed by the household which represent 80 to 90 percent of the value of crop pro .uction.

Oxen Zones Donkey Zones
Crop Price/Kga FCFA % FCFA % FCFA % FCFA %
Sorghum 45.5 59,821 69.9 81,167 54.3 36,449 43.6 36,871 39.2
Millet 45.5 12,792 14.9 29,206 19.6 18,775 27.5 13,217 14.1
Niadib 45.5 0 0.0 14 0.0 7,319 8.8 10,476 11.1
Maize 39.6 2,663 3.1 6,767 4.5 2,827 3.4 4,129 4.4
Groundnuts 68.9' 1,971 2.3 4,857 3.3 1,114 1.4 11,447 12.2
Bambara Nuts 59.0 1,103 1.3 1,242 0.8 1,109 1.3 1,105 1.2
Cowpeas 73.2 5,192 6.1 10,035 6.7 8,105 9.7 6,551 7.0
Soybeans 72.4 239 0.3 4,302 2.9 282 0.3 1,013 1.1
Sesame 57.6 6 0.0 39 0.0 421 0.5 561 0.6
Cotton 67.4 58 0.1 1,818 1.2 73 0.1 916 1.0
Rice 90.2 1,746 2.0 9,909 6.6 7,127 8.5 7,726 8.2
TOTAL 85,591 149,356 8$,601 94,012
Sorghum & Millet & NiadI 72,613 84.8 110,387 73.9 62,543 74.8 60,564 64.4
aThis represents the weighted average selling price realized by sample households during the 1978-79 survey period.
bA 60-day, short season millet Yariety grown in the Logobou area.

5.2. Farm Household Income
statement of annual income for sampled farm households is presented in Tabl 5.2. Net household income is defined as the sum of five income components: crop production, livestock raising, trade in agricultural products, agricultural proc ssing and the sale of gathered crops, and "other" sources of income.
The overall picture which emerges is one of a subsistence economy with a surp isingly low degree of monetization. Looking at the major source of total farm household income, crop production, only 8 to 15 percent of the value of crop
prod ction was sold by the four subsamples in 1978-79. Both ANTRAC and hoe
farm rs use only minimal amounts of purchased seed, fertilizer, or hired labor. The eotal value of crop production is higher among ANTRAC farmers than among hoe farm.Irs. However, since oxen households are 67 percent larger and donkey households are 28 percent larger than traditional households, this difference mainly reflects the larger scale of ANTRAC farms. On a per capita basis, the value of cro production, relative to hoe farmers, is only slightly higher (+4.5 percent) for oxen traction and moderately lower (-11.9 percent) for donkey traction.
Table 5.2 shows that the introduction of animal traction imposes a variety of roduction costs which are substantially higher than those encountered by hoe f arIers. For example, ANTRAC-related variable costs in 1978-79 (feed grain pur hased or supplied by the household and ANTRAC animal maintenance costs) amo nted to 5,544 FCFA for oxen farmers and 4,134 FCFA for donkey farmers, inceases of 59 percent and 44 percent, respectively, above hoe 'levels. Annual fixed costs (excluding depreciation on animals) were 8,224 FCFA for oxen and 6,243 FCFA for donkey farmers, fully 154 and 127 percent higher, respectively, than for hoe farming.
Due to its potential for increasing returns from crop production, ANTRAC can gen erate direct revenues which may help cover some of these added costs through either contract services (plowing or carting) or the sale of mature oxen which appreciate greatly in value during the time they work as draft animals. Among ox n farmers, such contract revenues were not substantial, amounting to only 660 FC A, because of inexperience and a lack of popularity of oxen carts. However, th value of appreciation on oxen, estimated at 10,000 FCFA per ox per year, more th n covered all ANTRAC-related costs in 1978. Among donkey farmers, contract re venues were more substantial than for oxen farmers, due almost exclusively to a

Oxen Zones Donkey Zones
Number of Households 36 64 72 46
1. Crop Production Enterprise
Revenue d
Value of Crop Production 85,591 149,356 83,601 94,012
of Which, Value Sold 6,661 9,680 9,569 13,798
Contract Plowing Revenues 0 + 524 0 + 70
Contract Transport Revenues 0 + 136 0 + 1,635
Variable Costs
Purchased Seed 484 583 784 1,273
Value of Household Seed 4,175 7,930 4,490 6.981
Fertilizer and Insecticides 28 402 153 788
Wage Labor 250 490 217 315
Grain Purchased for ."Invitatlon" Field Laborb 0 31 0 48
ANTRAC Feed Grain (Purchased) 0 640 0 328
ANTRAC Feed Grain (Value of Household Grain)a 0 1,672 0 2,826
Other ANTRAC Maintenance Costsc 0 3.232 0 980
Fixed Costs
Repairs to ANTRAC Equipment .0 68 0 26
Replacement Parts for ANTRAC Equipment 0 1,012 0 1,075
Interest Payments for ANTRAC Cregit 0 1,915 0 47
Depreciation on ANTRAC Equipment 0 5,229 0 5,095
Depreciation on ANTRAC Animalse 0 +22,645 0 2,081
Repairs of Other Tools and Equipment 36 67 61 77
Depreciation on Other Tools and Equipment 1,996 3,170 2,324 2,678
Net Revenue
Net Revenue from Crop Production 78,622 146,220 75,572 71,099
II. Livestock Enterprise
Sales of Animals 3,652 27,693 17,337 33,281
Sales of Animal Products 345 5,434 1,684 680
Animal Purchases 5,556 -26,961 -10,127 -30,924
Feed and Maintenance Expenses 411 1,031 3,084 1,641
Net Revenue
Sub-total 1,970 5,135 5,810 1,396
Ill. Agricultural Trading
Value of Sales (Net of Transport Costs) 1,594 2,877 7,867 17,913
-aTue of Purchases (Net of Transport Costs) 1,358 3,406 6,682 -19,402
Depreciation 61 599 234 215
Change in Value of Inventoriesa 0 + 2,058 9 + 3,626
Net Revenue
Sub-total 175 930 942 1,922
IV. Agricultural Transformation & Gathered Crops Revenue
STe of Transformed Crops 1,052 2,744 1,994 8,185
Sales of Gathered Crops 513 3.406 994 631
Purchases of Variable Inputs 797 2,718 1,489 9,897
Depreciation on Equipment 240 254 797 339
Net Revenue
Sub-total 528 3,178 702 1,420
NET FARM INCOME 77,355 155,463 83,026 72,997
V. Other Sources of Income
Gross Returns to Non-Ag. Trading & Artisanal
Activities 38,422 14,822 8,858 34,385
Salaries 11 484 0 4,817
Pensions 0 5,807 9 860
Inheritance & Net Cash Gifts 382 14 2,673 5,972
Variable Costs of Non-Ag. Trading & Artisanal
Activities 1,120 7,943 5,156 -24,457
Depreciation 572 641 527 1,535
Net Revenue
Sub-total 36,359 12,543 511 20,042
TOTAL NET HOUSEHOLD INCCME 113,714 168,006 83.537 93,039

Notes }n Table 5.2
a. Crop values based on average sales prices listed in Table 5.2.
b. lInvitation" labor refers to festive work parties of a reciprocal nature in
wich food and sorghum beer are the primary in-kind payment. These cash
purchases of grain substantially understate the real costs of invitation lbor which primarily utilize household food stocks, rather than purchased
g ain.
c. Chiefly non-grain feeding expenses, salt, and medicines for animal mainten nce.
d. Refers to cart rental services. This does not include the sale of carted
products, such as firewood, but only the rental of the cart for transport
u e.
e. T e following straight-line depreciation schedule was used for ANTRAC equipment and animals. (Note that the values in parentheses represent appreciation.)
Estimated Annual
1978 Price Working Life Salvage Value Depreciation A TRACK Item (FCFA) (Years) (FCFA) Rate (FCFA)
Oxen Traction:
Pl( 18,250 10 2,000 1,625
We der (5 Teeth) 19,635 7 1,500 2,591
Ri (Iger 6,470 5 500 1,194
Ac essories 7,225 5 250 1,395
Ca t 44,735 10 3,000 4,174
1 x 35,000 4 75,000 (10,000)
Donk y Traction:
Plow 11,320 10 1,000 1,032
We der 17,200 8 1,500 1,963
Ri iger 4,850 6 500 725
Accessories 5,185 5 200 997
Ca 44,735 10 2,000 4,274
1 Donkey 18,000 7 3,000 2,143

small number of high-income generating donkey carts.- However, these donkey contract revenues were offset by the fact that donkeys depreciate in value.
The average net revenue from cropping activities was 86 percent greater for oxen farmers than hoe farmers in the same study areas with the estimated appreciation of oxen alone contributing a 16 percent increase. For donkey farmers, average gross revenue from cropping is 12 percent higher but, when higher ANTRAC costs are included, donkey farmers have net cropping revenues 6 percent lower
than hoe farmers in the same zones. The poor income performance of donkey
traction is influenced by the fact that a drought seriously reduced overall yields in donkey zones. The timing of the drought (early June through July in
Piela and an early end of rains in Diapangou) may have disproportionately penalized an ANTRAC cropping strategy by prohibiting plowing in Piela and severely penalizing late planting in both villages. In addition, this drought was more
severe in some ANTRAC villages than the neighboring control villages due to rainfall differences in Diapangou and soil differences in Piela. In general, however, donkey traction has lower costs than oxen traction, which can make donkeys appealing to small, poor, or inexperienced farmers.
While crop production is by far the largest source of total farm household income, Table 5.2 presents four other important components. Although livestock raising and crop trading do not contribute substantial net revenues on the average, they have important cash flow benefits for many ANTRAC households. Proportionately twice as many ANTRAC farmers engaged in crop trading and two to
three times as many in cattle trading (Ouedraogo and Wilcock, 1980, p. 36). Although the net revenue from crop trading is higher for ANTRAC households, their net revenue from livestock raising is higher than for hoe households only in oxen zones. Unfortunately, a major component of the net revenue of the livestock
enterprise, the growth in value of herds, is not included in Table 5.2, which may explain these ambiguous results.2/
1'It should be noted that donkey cart revenue in Table 5.2 may substantially underestimate the potential revenue from carting in certain areas because this
figure reflects only direct cart rental fees and not the sale of hauled goods such as firewood, water, or gravel. As shown in Chapter 4, over half of donkey
cart time was used to haul firewood. In heavily populated areas such as the donkey zones--Diapangou, Piela, and Logobou--there is a flourishing market for firewood. Firewood carting can be a substantial source of revenue; in some areas, donkey wood haulers can easily net 10,000 FCFA per month.
--/The analysis of the change in the value of farm livestock inventories over the 1978-79 season was considered beyond the scope of this report because of the
large amount of time and computer resources it would have entailed.

agricultural transformation and gathered crops contribute little to total
incom Certain transformation activities, for which much of the production may be co sumed but little sold (such as weaving, food preparation, and dolo making), can r sult in negative net revenues. A surprising proportion of total household income is generated from other sources of income. The mean values of such income for dome zones is somewhat exaggerated by a few individuals who are either skilled laborers (such as masons or carpenters) or retail traders.
able 5.3 indicates the absolute and relative importance of the five major inco e components for sampled farmers. Four components are summed to make up net farm income; all five components are included in total household income. The table also includes several efficiency measures: net crop production revenue, net farm income, and net household income on a per capita, per active worker, and per ectare basis. On all efficiency measures (except net household income per acti e worker), donkey farmers had lower incomes than the hoe farmer control grou As stressed earlier, this is largely due to drought in donkey areas which affected traction farmers more than hoe farmers.
For oxen farmers, the results are more encouraging. All standardized income mea ures are higher for oxen households for both net crop production revenue and tot 1l farm income. Total household income measures for oxen farming are lower tha for hoe farming, however, due to unusually high levels of other income attr ibutable to 2 of the 36 hoe farmers in the oxen zones. For oxen farmers, net farm income is 20 percent higher per capita, 16 percent higher per active worker, and 12 percent higher per hectare, than for hoe farmers in the same zones.
By any measure, income is clearly very low for both hoe and animal traction far ers. Hoe farmers had a net household income per capita of 13,255 FCFA ($60.25) a year. The ANTRAC households had a per capita average income of 1,586
FCFA lower than that for hoe farmers. It is clear that such low per capita
incomes represent a level of material poverty which severely limits the potential fo savings and investment among most farmers in the Eastern Region.
In summary, the survey data indicate that the income effects of ANTRAC were no substantial during the 1978-79 season. Oxen farmers had a modestly higher pe" capita farm income than hoe farmers, while that of donkey farmers fell below ho farming levels. Unfortunately, these survey data are limited in their ability to provide a fair test of ANTRAC profitability. Although the ANTRAC s ple represents farmers considered by extension agents to be relatively more s ccessful in using traction, the program is young and most of these farmers are

Oxen Zones Donkey Zones
Value of Major Sources of Income FCFA FCFA FCFA FCFA
I. Crop Production 78,622 146,220 75,572 71,099
II. Livestock Raising 1,970 5,135 5,810 1,396
Ill. Crop Trading 175 930 942 1,922
IV. Agricultural Processing 528 3,178 702 1,420
V. Other Sources 36,359 12,543 511 20,042
NET FARM INCOMEa 77,355 155,463 83,026 72,997
NET HOUSEHOLD INCOME 113,714 168,006 83,537 93,039
Relative Importance of Sources of Income Percent Percent Percent. Percent
I. Crop Production (% of total) 69.1 87.0 90.5 76.4
II. Livestock Raising 1.7 3.1, 7.0 1.5
III. Agricultural Trading 0.2 0.6 1.1 2.1
IV. Agricultural Processing 0.5 1.9 0.8 1.5
V. Other Sources 32.0 7.5 0.6 21.5
Efficiency Measures FCFA FCFA FCFA FCFA
Net Crop Production Revenue per Person 11,787 13,126 8,559 6,309
Net Crop Production Revenue per Active Worker 25,863 27,745 19,084 17,174
Net Crop Production Revenue per Hectare 19,854 20,508 16,287 11,771
Net Farm Income per Person 11,597 13,955 9,403 6,477
Net Farm Income per Active Worker 25,446 29,450 20,968 17,632
Net Farm Income Hectare 19,534 21,804 17,894 12,085
Net Household Income per Person 17,049 15,081 9,461 8,256
Net Household Income per Active Worker 37,406 31,879 21,095 22,473
Net Household Income per Hectare 28,716 23,563 18,003 15,404
aNet Farm Income in the sum of major income components I through IV.

still in the process of learning the different elements of the package. Traction teams are typically underutilized, particularly for weeding and ridging. Many farmers are inexperienced and many EORD ANTRAC services are not fully operational. Traction animals tend to be fed and cared for at less than recommended level As a result, most ANTRAC farmers surveyed were not yet achieving substant al productivity increases. Thus, the income effect of ANTRAC should rise in future years as performance improves both at the farm and program level.
5.3. Cash Flow
5.3.1 Annual Cash Flow
lthough net income is a very useful measure of economic welfare, the farm cash low position has important practical consequences in an area such as the Eastern Region. Due to poorly articulated markets plus a low level of monetization n the rural economy, where less than 10 percent of agricultural production is sold, farmers are limited in their ability to generate cash. Cash flow
probl ms can inhibit the adoption of interventions such as ANTRAC which have
highly variable cash expenditures and revenues. The EORD credit program alleviates this problem partly, but not entirely, as seen below.
he annual cash flow statement presented in Table 5.4 shows the extremely low 1 vel of annual cash flow for both ANTRAC and hoe farmers. Of the already low net cropping revenues and household net incomes, only a very minor proportion is realized as cash revenue. While the absolute value of cash inputs into crop production is small, the cash inputs in hoe farming are quite large relative to the amount of cash revenue generated, varying from 26 to 30 percent of the value of sales in the two subsamples. Oxen traction, moreover, causes a dramatic increase in the current cash costs of crop production. Even when contract
revenues are included, net current cash costs amount to 6,524 FCFA or 67 percent of cr p sales, leaving a net cash revenue of only 3,156 FCFA (item 5 in Table 5.4). Current cash expenses for donkey traction are less than those for oxen; most re recovered through cash revenues generated from contract services. Thus, the et current cash costs of crop production amount to only 3,325 FCFA for donkey farmers, or 24 percent of crop sales. In terms of net cash flows generated from crop production, donkey traction outperformed hoe agriculture, both on an absol te and a per capita basis, while oxen traction did not.-/While 66 percent of the ANTRAC sample farmers had an outstanding ANTRAC loan from the EORD, many did not make repayments in 1978-79 due to poor credit (footnote is continued on page 82)

Oxen Zones Donkey Zones
Crop Production, FCFA
1. Value of Sales 6,661 9,680 9,569 13,798
2. Non-ANTRAC Inputs 1,752 2,682 2,879 2,621
3. ANTRAC Related Current Cash Expenses 0 4,502 9 2,409
4. ANTRAC Related Revenues 0 + 660 0 + 1,705
5. Net Cropping Cash Revenue 4,909 3,156 6,681 10,473
6. Major Food Purchases 4,966 -11,617 9,505 -20,782
7. Net Cropping Cash Surplus 57 8,461 2,824 -10,309
Livestock Production
8. Revenues 3,997 33,127 19,021 33,961
9. Expenditures 5,967 -27,992 -13,203 -32,565
Agricultural Trading
10. Revenues 1,594 2,877 7,867 17,913
llb Expenditures 1,358 --3,406 6,682 -19,402
Agricultural Processing and Gathering
12. Revenues 1,565 6,150 2,988 8,816
13. Expenditures -1,037 2,972 2,286 -10,236
Other Sources of Income
14. Revenues 38,051 21,127 6,194 46,C34
15. Expenditures 1,120 7,943 5,156 -24,457
Capital Expenditures
16. Non-ANTRAC Equipment Purchased 183 126 504 276
17. ANTRAC Equipment Purchased 333 640 0 2,399
18. Borrowing and Reimbursements Received 1,155 6,853 2,854 9,169
19. Loans and Repayments 1,870 -19,237 3,138 7,724
20. Net Cash Flow 34,437 .643 5,131 8,525

The problem of current cash expenditures for ANTRAC is even more acute when one considers cash cropping revenue net of cash food expenditures. In Table 5.4, net c sh surplus from cropping (defined as net cash revenue from cropping less major cash purchases of food) can be considered as a rough measure of a household' ability to produce enough food to feed itself and generate a cash surplus from he crop production enterprise alone. Neither traditional nor ANTRAC farmers generated such a surplus in 1978, a result most marked in donkey zones due to poor rainfall. Although 1978 was a year for rebuilding on-farm grain stocks following several years of sporadic drought which may have caused low cash sales, only the hoe farmers sold more crops than they purchased. This implies that ANTRAC farmers, at least during the 1978-79 season, were less food selfsufficient than hoe farmers and that they had to generate current cash revenues
from sectors other than crop production in order to meet the annual cash flow requirements of both ANTRAC adoption and food needs. Thus, in Table 5.4, one sees that the net cash deficit of the ANTRAC cropping enterprise (represented under net cropping surplus) is offset by positive cash flows from the other econ mic sectors, principally "other sources of income."
In summary, the current cash costs of ANTRAC create a serious annual cash flow problem, caused primarily by capital expenditures on equipment and animals plus the scheduling of financing.- The modest output increases attributable to
1(continued from page 80) collection efforts by the EORD. Repayment
behavior therefore has a substantial effect on the annual cash flow position shown in Table 5.4, at least for oxen farmers. Of the 19,237 FCFA in cash loans and repayments of oxen traction farmers, 12,697 FCFA represents repayments on EORD medium-term ANTRAC loans. Since ANTRAC credit collection efforts were
delayed until after the survey period, donkey farmers reimbursed only 284 FCFA to the EORD during the 1978-79 season. A more reasonable assessment of the cash flow effects of ANTRAC financing would be to use the value of ANTRAC credit rep yment that the typical oxen or donkey farmers from our survey should have paid in 1978-79. For the median ANTRAC farmer in each subsample, this implies a cash repayment of 22,600 FCFA for oxen farmers and 14,175 FCFA for donkey farmers. If these repayment values are taken into account (net of ANTRAC credit repayments already included in Table 5.4), the annual net cash flow becomes
-10 546 FCFA for oxen farmers and -5,366 FCFA for donkey farmers.
-2/This cannot be fully depicted in an annual cash flow statement because the pro lem involves year-to-year changes in cash flow position. Further, our sample of ANTRAC farmers does not accurately represent a cross-section of farmers at the various stages of ANTRAC financing that are characteristic of the EORD mediumterm credit program.

oxen traction are not accompanied by increases in crop sales because some of the additional output is consumed. This, plus high current cash costs, makes oxen a less attractive investment than donkeys on a current cash basis. While oxen
traction is financially more rewarding than donkey traction (although this is somewhat exaggerated by the effects of drought on our donkey sample), an oxen traction farmer must have other sources of liquidity to carry him through cash deficit years until capital gains are realized from the sale of his oxen team. Donkey traction also requires alternative sources of liquidity but to a lesser extent because of its lower cash costs. The lower and less variable cash costs of donkey traction may account for its popularity in spite of its poor financial performance in 1978-79. These factors are examined further in the next section.
5.3.2. Monthly Cash Flow
The annual cash flow statement obscures important changes in the cash flow position during the cropping season. Tables 5.5 and 5.6 present the monthly cash flow for hoe and ANTRAC households, respectively.-Y In the hoe sample, the monthly cash balance in Table 5.5 is quite low, as one would expect, but also surprisingly uniform throughout the year. Crop sales are heavier in the December through March post-harvest period but relatively high levels of sales are maintained throughout the year, except for late August through mid-October. Cash
expenditures for crop production inputs are concentrated during the May through August cultivation season; however, these are so low that net cropping cash revenue remains positive on average. Major food purchases are also concentrated in this period, the latter part of which represents the "hungry season," creating an important cash deficit. This cash deficit appears to be covered by higher levels of either livestock sales, crop sales, or other sources of income.
In contrast, the monthly cash balance of ANTRAC farmers, presented in Table 5.6, is far more variable throughout the year. The timing of crop sales is
slightly more seasonal for ANTRAC farmers than for hoe farmers. In addition,
cash expenditures for crop production inputs are greater, making net cash revenue from cropping negative throughout most of the cultivation season. Although
ANTRAC-related cash expenses are less seasonal than other agricultural input costs, they still represent the most important cause for this negative net
11Since monthly statements present cash flow values that 'are averaged over a large sample, the average monthly stream of cash revenues and expenditures is smoother than it would be for an individual household. The monthly cash flow
statement for a typical hoe farming household would show infrequent transactions (particularly for non-crop related activities) and occasional large transactions (particularly for livestock or agricultural trading).

May 1- May 29- June 26- July 24- Aug. 21- Sept. 18- Oct. 16- Nov. 13- Dec. 11- Jan. 8- Feb. 5- Mar. 5- Apr. 2Cash Flow Item May 28, 1978 June 25 July 23 Aug. 20 Sept. 17 Oct. 15 Nov. 12 Dec. 10 Jan. 7, 1979 Feb. 4 Mar. 4 Apr. 1 May 30, 1979
Crop Production
1. Value of Sales 1,277 411 632 470 205 285 472 952 1,224 1,104 547 1,840 893
2. Non-ANTRAC Inputs 826 630 479 -217 -U107 47 817 51 9 4 38 63 -102
3. ANTRAC Related Current Cash expenses 0 0 0 9 0 0 0 0 0 0 0 0 0
4. ANTRAC Related Revenues 0 0 0 0 0 0 0 0 0 0 0 0 0
5. Net Cropping C(ash Revenue 451 219 153 244 98 238 385 901 1,215 1,100 509 1,777 791
6. Major Food Purchases -1,375 -1,164 -1,559 -652 -897 -753 322 -213 471 -330 -341 -384 -638
7. Net Cropping_ Cash Surplus 924 -1,383 -1,406 -408 -799 -515 63 688 744 770 168 1,393 153
Livestock Raising
8. Revenues 1,380 1,714 1,285 1,141 581 647 2,279 204 961 1,375 711 1,279 1,112
9. Expenditures -1,252 800 -1,379 -762 -697 -880 -1,683 -321 535 -443 -870 -426 -770
Agricultural_ Trading
10. Revenues 1,613 843 251 224 839 66 124 51 200 566 106 1,055 302
11. Expenditures 66 311 16 -280 86 -249 776 -829 -1,185 -462 -356 19 -467
Agricultural Transformation and_ Gathering
12. Revenues 145 306 28 192 108 302 187 119 230 54 349 289 245
13. Expenditures 81 141 113 55 -58 -138 103 -205 84 -174 88 54 8
Other Sources of Income
14. Revenues 421 50 1,486 1,039 554 575 630 948 336 596 803 4,669 720
15. Expenditures 258 85 425 -377 -380 77 316 -164 904 41 -420 -181 -137
Capital Expenditures
16. Non-ANTRAC Equipment Purchased,! 19 0 29 6 42 38 33 29 125 32 12 3 24
17. ANTRAC Equipment Purchased 1 91 12 0 0 0 0 0 0 0 0 0 0
18. Borrowing and Reimbursements Received 108 133 137 418 136 325 346 62 382 54 206 20 77
19. Loans and Repaymnents 356 187 221 -263 -169 -214 244 -164 355 -148 -277 14 -104
20. Net Cash Flow 710 48 414 863 13 -196 474 360 332 2,115 320 8,008 1,099

--- ---- - --. .... -...-.-___-_. --_--------- _--- --- .May I- May 29- June 26- July 24- Aug. 21- Sept. 18- Oct. 16- Nov. 13- Dec. 11- Jan. 8- Feb. 5- Mar. 5- Apr. 2Cash Flow Item May 28, 1978 June 25 July 23 Aug. 20 Sept. 17 Oct. 15 Nov. 12 Dec. 10 Jan. 7, 1979 Feb. 4 Mar. 4 Apr. 1 May 30, 1979
Cr_o-p_Prod -!!- --t-ion
1. Value of Sales 645 1,849 552 454 757 409 690 487 2,469 776 704 3,476 751
2. Non-ANTRAC Inputs -1,173 620 829 676 166 168 178 65 37 53 25 54 190
3. ANTRAC Related Current Cash Expenses -1,147 394 378 119 121 58 109 310 170 105 239 223 215
4. ANTRAC Related Revenues 59 176 221 39 26 35 24 89 142 18 21 60 60
5. Net Cropping Cash Revenue -1,616 1,011 434 302 496 218 427 201 2,404 636 461 3,259 406
6. Major Food Purchases -4,274 -2,958 -1,909 743 452 517 707 -1,874 -1,187 807 -1,046 -1,503 -1,476
7. NetCrpoping Cash Surplus -5,890 -1,947 -2,343 -1,045 44 299 280 -1,673 1,217 171 585 1,756 -1,070
Livestock Raising
8. Revenues 1,559 5,048 6,907 681 7,799 11,122 2,236 19,134 9,760 2,011 1,159 2,643 8,214
9. Expenditures -3,021 -4,433 -4,300 -3,763 -6,112 5,676 -7,107 -7,081 -3,662 -3,162 -5,576 -4,866 -3,570
Agricultural Trading
10. Revenues 307 2,023 928 303 156 707 467 1,526 951 36,046 '1,173 14,807 3,496
11. Expenditures 667 178 471 489 -1,180 -18,707 -2,652 -3,203 528 -8,234 741 -4,691 -1,388
Agricultural Transformation and Gathering
12. Revenues 145 697 286 132 613 270 466 468 505 436 621 1,190 1,307
13. Expenditures 240 452 397 177 665 449 300 399 377 468 489 630 324
Other Sources of Income
14. Revenues 2,515 3,551 2,634 4,019 3,667 4,733 3,062 2,903 1,730 2,215 3,913 3,050 2,263
15. Expenditures 896 742 -1,835 -2,955 -1,355 1,061 -3,628 -1,374 129 693 -1,682 754 987
Cap i ta Epeditu r es
16. Non-ANTRAC Equipment Purchased 14 4 0 9 18 57 33 29 7 20 11 15 0
17. ANTRAC Equipment Purchased 172 596 475 98 464 59 8 0 41 I 4 65 303
18. Borrowing and Reimbursements Received 149 346 312 272 448 2,704 1,601 715 196 286 606 246 533
19. Loans and Repayments -1,324 826 -1,662 573 432 654 -1,610 -1,523 -1,505 -2,597 -1,751 -1,633 -1,466
20. Met Cash Flow -7,549 2,487 416 -3,702 2,501 7,426 -7,786 9,464 8,110 25,648 -3,367 11,038 6,705