THE SOUTHERN FULANI FARMING SYSTEM IN UPPER VOLTA:
A MODEL FOR THE INTEGRATION OF CROP
AND LIVESTOCK PRODUCTION IN THE
WEST AFRICAN SAVANNAH*
Christopher L. Delgado
*A major part of the fieldwork for this report was financed under
contract REDSO/WA 77-107 with the United States Agency for International
Development. The 211-D grant from the U.S. Agency for International Develop-
ment to the Center for Research on Economic Development of the University of
Michigan provided supplementary support for write-up and analysis.
This paper is published under grant (AID/afr-G-1261) from the Agency
for International Development to Michigan State University.
**The author is an assistant research scientist at the Center for Research
on Economic Development (CRED) of the University of Michigan.
TABLE OF CONTENTS
1 THE FULANI FARMING SYSTEM IN
TENKODOGO: RESEARCH OBJECTIVES,
ISSUES, AND METHODOLOGY........................... 1
Rationale of the Study........................... 1
Specific Hypotheses to be Investigated............ 6
The Field Research................... ............ 7
Outline of the Chapters........................... 10
2 THE RESEARCH SITE AND ITS FULANI
Location and Characteristics of the Research
The Fulani of Greater Oueguedo.................... 22
The Economic Roles of the Fulani in
Tenkodogo and the Cattle-Entrusting
System............................. .... .*.....* 25
3 HOUSEHOLD LABOR USE ............................... 33
Household Labor Availability...................... 33
The Agricultural Calendar........................ 35
The Division of Labor by Age and Sex............... 38
Household Labor Allocation by Major
Sector of Activity........................... 41
Household Labor Allocation to Crop Work........... 43
Household Labor Allocation to Livestock............ 48
4 LAND, CAPITAL, LIVESTOCK, AND MANURE.............. 56
Fulani Land Use.................... ............. 56
The Fulani Inventory of Capital Goods.............. 65
Fulani Livestock Holdings......................... 65
The Use of Manure by Fulani
Herdsmen in the Tenkodogo Area............... 71
5 OUTPUT, INCOME, AND EXPENDITURE .................. 76
Crop Output....................................... 76
The Change in the Value of Household
Stock Attributable to Net
Herd Growth.................................... 82
The Timing of Herd Growth and Net Household
Revenue from Cattle
Sales and Purchases.......................... 85
Other Revenue from Livestock Operations........... 89
Annual Cash Income and Subsistence
Production for an Average Sample
Fulani Household.............................. 94
Agricultural Production and Cash
Income Compared for Average
Fulani and Mossi Households
in Oueguedo................................... 98
Fulani Household Expenditures
Each Month by Sector.........................101
6 A MODEL OF HOUSEHOLD RESOURCE ALLOCATION
AND PRODUCTION ............. .................... 105
The Basic Model...................................106
Comparison of Optimal Production
Strategies Under Different
The Opportunity Cost of Scarce
Resources and Constraints to
Conclusions from the Modeling Exercise ............117
7 SUMMARY AND POLICY RECOMMENDATIONS.................118
A Summary Profile of the Fulani Farming
System in Tenkodogo..........................118
A New Model for the Integration of
Stockraising with Crop Production
and the Need for Policy Intervention.........123
Recommended Policy Actions........................126
Appendix A ........................................ 133
LIST OF TABLES
2.1 Percentage Distribution of Population in
Upper Volta by Ethnic Group 1960-61........... 17
3.1 Average Number of Fulani Workers in Each
Household by Age and Sex Category
Compared with Peasant Farms................... 34
3.2 Tenkodogo Fulani Agricultural Calendar
(1977-78) ..................................... 36
3.3 Division of Labor in Fulani Households............ 39
4.1 Mean Fulani Household Field Area Cultivated
Compared to the Mossi and Bisa................. 59
4.2 Mean Number of Agricultural Implements
and Storage Facilities Available
on Each Farm................................... 66
4.3 Summary of Fulani Herd Sizes in
Oueguedo for 1977........... ................ 68
5.1 Yield Plot Results for Fulani Sample Crop
5.2 Computation of the 1977 Returns to
Major Fulani Crop Enterprises.................. 81
5.3 Herd Growth from 1976 to 1977..................... 84
5.4 Computation of the Value of Average
Household Cash and Subsistence
Production in the Livestock Sector............ 96
5.5 A Comparison of Annual Agricultural
Production and Cash Income
Between Fulani and Mossi
Households in Oueguedo........................ 99
5.6 Average Fulani Sample Household Expenditures
in Each Sector by Month........................102
5.7 The Percentage of Household Expenditures
Allocated to Each Sector by Month.............103
6.1 The Fulani Household Agricultural
Production Model............................... 111
6.2 Summary of the Production Strategies
Which Maximize Farm Income Under
Different Assumptions ......................... 113
6.3 Dual Variable Values in the Optimal
Solutions to Various Models................... 116
LIST OF ILLUSTRATIONS
2.1 Location of Tenkodogo in Africa................... 14
2.2 Ethnic Boundaries and Research Area................ 15
2.3 Relation of the Research Area to the
Onchocerciasis Control Program ................ 20
3.1 Mean Total Household Labor Hours Each
Fortnight by Category of Work................. 42
3.2 Ratio of Mean Crop to Livestock Work Hours
on Sample Fulani Farms, 1977................... 44
3.3 Mean Household Hours Allocated to Each
Major Agricultural Operation, by Fortnight.... 45
3.4 Comparison between Peasant-Farmer and Fulani
Households of Labor Hours Allocated
to One Hectare of Food Grains................. 47
3.5 Mean Household Hours Allocated to Cattle
Work per Fortnight.................. .......... 49
3.6 Mean Household Labor Hours. Allocated to
Grazing Cattle................................ 50
3.7 Mean Household Hours per Animal Kept............... 52
3.8 Economies of Scale in Cattle-Raising.............. 53
3.9 Comparative Labor Requirements over
Time for Crops and Livestock.................. 55
4.1 Bush Fields in Oueguedo.......................... 61
4.2 Transhumance Routes of the Oueguedo
Herds during the Dry Season................... 64
4.3 Age Structure of Fourteen Fulani Herds............. 70
5.1 Average Monthly Entrances and Exits
from Household Herds.......................... 86
5.2 Average Value of Sales and Purchases of
Fulani-owned Cattle for Sample
Households (over 19 Households) ................ 87
5.3 Average Household Production and Sales
of Milk Each Month by Volume.................. 91
The idea for this study grew out of the presence of the author in
Upper Volta from October 1975 until June 1977, in connection with the
Entente Livestock Project of the University of Michigan (U.S.A.I.D. con-
tract AFR C 1169). During this period the author was primarily
involved in research concerning the Mossi and Bisa farming systems of
the southeastern part of the country. As will be made explicit in
chapter one, the experience permitted him to contact and become friendly
with a group of semi-sedentary Fulani herdsmen living in the same area.
Given the desire of this group to be given the same hearing as their
Mossi and Bisa neighbors, and the existence of a trained research team --
including an excellent Fulani enumerator -- qualified to work with them,
the decision was made to seek additional funding to gather the data
which form the basis of this report. This was provided by U.S.A.I.D.
contract REDSO/WA 77-107 and the 211-D grant of the Center for Research
on Economic Development of the University of Michigan.
The author extends his sincere thanks to John Eriksen, agricultural
economist at REDSO/WA, Professor Elliot Berg, then director of C.R.E.D.,
and Professor Kenneth Shapiro, director of C.R.E.D.'s Entente Livestock
Study. Each of them helped pave the way for this research. Acknowledge-
ments are also due to friends at the U.S.A.I.D. mission in Ouagadougou,
who provided helpful advice on many occasions. Special thanks are due
to Bob Dakan and Gr6goire Ouedraogo in this respect.
The greatest debt incurred in the process of the fieldwork concerns
Laurent Ouedraogo, head cartographer at O.R.S.T.O.M., Ouagadougou.
Laurent provided sage advice throughout the running of the project, based
on the experience of over a dozen years' research in Upper Volta. As a
great friend and esteemed colleague, he made frequent trips to the research
site. He kept the field accounts, payed and supervised the enumerators
during the author's lengthy absence from Upper Volta. A major part of
the success of the project is due to his tenaciousness and willingness to
help. Yelle Diessongo and Sidiki Diallo pedalled through rainy storms
on starless nights and forded flooded washes to get the basic data from
the semi-weekly interviews. Their loyalty in doing the rounds of outlying
Fulani farms was the subject of much local comment in the research area.
The author extends his greetings to the canton chief of Oueguedo,
the Fulani chief of Oueguedo, and the Fulani chief of Pouswaka. Without
their cooperation, this study would have come to naught. Finally the
long-suffering Fulani sample members deserve special mention. They showed
exemplary patience and concern in tolerating a stranger prying into their
affairs and disturbing their rest twice a week for seven months. It is
the author's most fervent wish that this study may be of some use to their
cause. It is to them that this work is dedicated.
THE FULANI FARMING SYSTEM IN TENKODOGO:
RESEARCH OBJECTIVES, ISSUES, AND METHODOLOGY
This paper attempts to provide a benchmark study of the farming
system encountered among a group of semi-sedentary Fulani herdsmen in south-
central Upper Volta. The analysis supports the view that, from an
economic viewpoint, the social product is maximized by increased integra-
tion between the Fulani and their sedentary farming neighbors at the
level of production. The essential point is that this integration can
best occur through a symbiotic relationship between Fulani herding specia-
lists and sedentary crop producers. This contrasts with the view that
the policy objective of the integration of livestock and crop production
is served by encouraging the Fulani to cultivate a substantial amount of
crops. In effect, this paper proposes supporting the division of labor
between stockraising and cropping on ethnic lines. The view here is that
the loss of production externalities brought on by specialization can be
more than compensated for by the increased efficiency of the division of
labor. At the same time, closer links between the Fulani and neighbor-
ing farming groups permit both herdsmen and farmers to enjoy the major
production benefits of mixed farming.
Rationale for the Study
Donor agency and governmental policy in Upper Volta since the great
Sahelian drought of 1974 has stressed the importance of moving the focus
of national livestock production from the northern to the more humid
southern and central regions of the country [R.H.V., M.D.R. (1976); Tyc,
(1975); U.S.A.I.D. (1975)]. The primary rationale for this strategy is to
lessen the dependence of the crucial livestock sector on the erratic
rainfall patterns of the Sahel ibidd). A secondary objective of this
strategy is to help sedentary farmers, who typically do not keep cattle,
to reap the production benefits of mixed farming practices involving
cattle (op. cit., p. 3). These consist principally of the fertilization
of crops with cattle manure and the feeding of animals with crop residues.
The benefits of mixed farming seem doubly attractive in the context of
the typical southern Voltaic smallholding characterized, in part, by de-
clining soil fertility, low cash income, hand tools with low productivity
and a shortage of protein in the diet.
Despite the assured desirability of a southern livestock industry
from the national perspective, and its apparent desirability from the point
of view of the individual rural household, very few of the sedentary farm-
ers who occupy most of the inhabitable southern area do, in fact, main-
tain large stock on the farm. In Tenkodogo, many farmers own cattle, but
it is invariably entrusted to Fulani herdsmen who live on the periphery
of the village and perform herding services for the predominant peasant
farming group. This aspect of the relationship of the Fulani to crop-
The term "peasant" is used throughout this paper to refer to
sedentary farmers in Upper Volta, whose primary economic activity is
the cultivation of crops for subsistence consumption. The term is con-
venient for distinguishing members of these groups from the Fulani, who
are less under the authority of their chiefs than, say, the Mossi or the
Bisa of Tenkodogo, and who are also considerably more integrated into the
cultivating peoples in the agricultural zones south of the Sahel is
pervasive throughout West Africa [Hurault, 1964; Miller, 1967; Queant
and Rouville, 1969; Barry, 1975; Horowitz, 1975; Rupp, 1975; Delgado,
1977; R.H.V., C.V.R.S., 1977]. Given the very strong policy interest
in increasing livestock production in the Sahelian countries, and given
an equally strong orientation in favor of the integration of crop and
livestock production, it is essential for planners to be aware of the
nature of the Fulani production processes.
Unfortunately, little is known about the agricultural production
relationships of Fulani pastoralists south of the Sahel. There is a fairly
voluminous anthropological literature dealing with Fulfulde-speaking
peoples, primarily in the Sahelian regions [Hopen, 1958; Stenning, 1959;
Dupire, 1962, 1970; Riesman, 1974]. Anthropologists have long been con-
cerned with the key economic and social issues of the relationship of
Fulani herdsmen to surrounding peasant-farming groups farther south
[Hurault, 1964; Miller, 1967; Queant and Rouville, 1969; Horowitz, 1972,
1975; Rupp, 1973, 1975; Diarra, 1975]. More recently, agricultural
economists have become interested in quantifying the exchange relationships
between the Fulani and their farming neighbors in these zones, primarily
for planning purposes [Barry, 1975; Delgado, 1977; R.H.V., C.V.R.S., 1977].
Finally, there is a small literature that attempts to give a quantitative
description of the basic production processes of the Fulani in the more
southern zones of the Savannah [Queant and Rouville, 1969; Van Raay, 1975;
Rochette, 1976; Delgado, 1977].
While the last four studies provide a point of departure, a great
deal of work remains to be done in order to understand the production
trade-offs within the Fulani farm and to identify resource or institu-
tional constraints which effectively prevent increases in output. These
objectives require detailed information from a random sample of Fulani
households concerning the flow of inputs to the production process
over the relevant season, and the resulting flow of outputs. Specifi-
cally, it is necessary to estimate the resource requirements for the
production of a given quantity of crops and livestock. Labor is the
most important and the most difficult to specify among these. Human
work is the primary factor of production in Tenkodogo, and it is the
availability of this resource during certain peak periods which effective-
ly determines the quantity of crop output among peasant-farmers [Delgado,
1978]. However, the labor requirements for producing, for example, one
head of cattle depend greatly upon the size of the herd involved. Further-
more, the timing of labor input is crucial to crop production.
This study applies the methodology of the classic "farm management
survey" of agricultural economics to provide a conceptual framework for
the simultaneous considerations of these issues. This technique allows
the analyst to distinguish the similarities in the agricultural produc-
tion process among farms within the group studied. These similarities
The precedents for this in West Africa are reviewed in Delgado
, chapter one, which is, in fact, a case in point.
can be viewed as components of a farming system "which transforms in-
puts into agricultural outputs and which undergoes changes over time"
[Ruthenberg, 1976, p.13]. Therefore, the Fulani farming system in a
given area consists of the method by which a typical production unit in
that area combines resources to get outputs, and its relationship to other
production and consumption units in the same zone.
Specifically, the study of the farming system indicates production
strategies that are in the overall economic interest of the people con-
cerned. Thus, an activity which appears to be profitable when viewed in
isolation (for example, crop cultivation), may be less attractive when
the ramifications of the enterprise on other pursuits arealso considered
(less attention to livestock). Furthermore, the study of the farming
system points out those policy actions which would allow farmers to in-
crease their incomes, if they so desire.
No claim is made at any point in this paper that the Fulani, or any
other group of people, will slavishly adapt their behavior to follow
production strategies which maximize farm revenue. However, knowledge
of what these strategies are under different assumptions provides an
important element in appraising actual behavior. In particular, this
information can show that certain well-known policy prescriptions
designed to change farmer behavior with the stated goal of increasing
household income are, at best, tenuous and, at worst, erroneous. Specifi-
cally, this study will investigate seven interrelated hypotheses.
Specific Hypotheses to be Investigated
(1) Despite appearances to the contrary, the Fulani of Tenkodogo
are highly integrated into the cash economy. Policy actions designed
to "bring the Fulani into the marketplace" are therefore futile.
(2) Even though the Fulani have a much higher cash income than
their farming neighbors, the overall value of production (including
subsistence) per worker is the same for the two ethnic groups. Apart
from the individual's job preferences, there is currently no economic
incentive for the Fulani to become Mossi-like crop farmers, or vice-
(3) The current Fulani practices in Tenkodogo of cultivating relatively
large areas of very low-yielding millet and of keeping low-return en-
trusted cattle are part of a de facto profit-maximizing strategy, given
the relevant resource constraints.
(4) The simultaneous production of millet and cattle leads to a
severe labor bottleneck in late July. This seasonal labor shortage
effectively constrains the expansion of output.
(5) Under current structures, the sale of manure is an economically
feasible enterprise that can boost income for Fulani sellers and Mossi
(6) A rise in the returns to entrusted cattle has the effect --
under current prices and assuming that grain is available for purchase --
of making specialization in livestock the economically optimal production
strategy, to the detriment of the extensive cultivation of millet. This
result is strengthened if manure can also be sold for its opportunity
cost to market-gardeners.
(7) Fulani specialization in livestock under these conditions
maximizes overall economic welfare in Tenkodogo.
These hypotheses will be tested by constructing a profile of the
typical Fulani production unit. The first two questions are purely
factual, and can be established by the compilation of the relevant data.
The next five hypotheses require a means for ascertaining optimal pro-
duction strategies and resource constraints when dealing with a multi-
tude of inputs and outputs. As a practical matter, the data from the
Fulani farm profile will be used for this purpose in a maximizing linear
The Field Research
The field research for this paper grew out of a larger study by the
author among the Mossi and Bisa farming populations of the Tenkodogo
area in southeastern Upper Volta. The original fieldwork involved
data collection from May 1976 through May 1977, from a sample of sixty
Mossi and Bisa farms distributed between two villages along ethnic
lines. The report of this work [Delgado, 19781 demonstrates that Mossi
and Bisa farmers in the research area do better economically by en-
trusting cattle to nearby Fulani herdsmen, rather than by attempting to
engage in mixed farming using large ruminants.I
During the process of this research into the potentials for mixed
farming by the peasant populations, the author came to know, and be
known by, the Fulani herdsmen living on the periphery of the Mossi
sample village, Oueguedo. Good social contact gradually grew into a
working relationship. Using an interpreter, the author was able to per-
sonally interview approximately twenty distinct Fulani households in
the greater Oueguedo area, from January through May 1977.3 The data
gathered were primarily of a stock nature (i.e., number of people in
the household), since insufficient time and resources were available at
that late date for a proper farm management survey using frequent visits
to the same compound. The sample households were chosen randomly by
travelling the length and breadth of the research area on a moped.4
The original study is part of the Entente Livestock Project of the
Center for Research on Economic Development of the University of Michigan,
U.S.A.I.D. Contract AID/afr-c-1169.
2The 57 questions that guided the series of interviews are repro-
duced as questionnaire L of Delgado  Appendix C. They primarily
concern the relationship of the Fulani to the Mossi and necessary animal
husbandry practices in the area. Also see Appendix B of this report.
3The notion of household in dealing with the Fulani poses little
difficulty for production research, since the people who work together
with the same animals and fields tend to live together in a geographically
distinct farm. By contrast, one Bisa compound, for example, may contain
several households, each one with a distinct set of fields that serve
only to nourish members of the household in question.
4The sample can be said to be "stratified" in the sense that an effort
was made to include households from all over the area. Once the tradi-
tional authorities had given their assent, all the sample members agreed
in principal to cooperate.
The first impressions about the Fulani in the Oueguedo area,as perceived
through this sample, are summarized in Delgado .
It soon became apparent that the really interesting economic issues
involving the southern Fulani, such as those sketched out in the previous
section, would require detailed data about the flow of labor and other
resources to the various products of the Fulani farms. Given a willing-
ness on the part of the sample to continue with repeated interviews after
May 1977, an effort was made to set up a farm management survey to cover
the 1977 cropping season.
Semi-weekly farm management survey interviews, exactly similar to
those conducted the previous year among the Mossi and Bisa, were con-
ducted by a specially-trained Fulani enumerator recruited locally,
assisted by the most experienced of the interviewers from the original
study,2 The data collected by this method concerned the allocation of
labor to all activities and the use of manure from June 1 until December
31, 1977. The semi-weekly data were subsequently aggregated into fort-
nights. Questionnaires from the original study concerning the house-
hold census, a field census, herd inventories, and farm tool and building
1Research funds for field expenses from June 1977 to February 1978
were provided by U.S.A.I.D. contract REDSO/WA 77-107. Unfortunately,
prior commitments of the author's time precluded a longer study that
would have covered the entire calendar year.
2The details of this data collection procedure, which with different
coding manuals is equally applicable to herdsmen and crop farmers, are
contained in Delgado, , chapter three.. Also see Appendix B to this
inventories were administered to the Fulani sample during this period.1
Rather complicated eight-page questionnaire, administered monthly, pro-
vided additional data on the flow of sales, purchases, barter goods,
loans, and gifts of all commodities.2
The scarcity of trained Fulfulde-speaking manpower on the research
team and the large distance between the sample households forced the
limitation of the labor-flow interviews to only twelve households out
of the twenty. The research area was divided into three parts, and four
compounds were randomly selected for interviewing in each.3 The inventory
questionnaires and the monthly sales-and-purchases interviews were
administered to nineteen households. While data on the Fulani popu-
lation are tenuous, information from the tax rolls indicates that the
nineteen farms represent roughly one third of the herdsmen who reside
permanently in the greater Oueguedo area.
Outline of the Chapters
The location and characteristics of the research site are discussed
in chapter two. The chapter also gives a brief introduction to the Fulani
Questionnaires A, B, H, and K respectively, contained in Appendix
C to Delgado ; reproduced in Appendix B to this report.
2The experience synthesizing the results obtained from this docu-
ment indicates that it needs to be simplified and reduced in size.
This allowed the research team to visit four compounds a day, which
is an upper limit, given the remoteness of the area and the complete
lack of roads. Even the footpaths were difficult to navigate after a
One of the twenty having withdrawn subsequent to a disagreement
with the enumerators.
population living in this region, and their economic relationship with the
predominant Mossi farming group.
Chapter three examines the availability and allocation of labor with-
in an "average" household. This data is used to calculate labor require-
ments per hectare for each crop, and per animal for each species kept,
each fortnight of the cropping season. Chapter four unites the data on
the use of resources other than labor to produce crops and livestock.
Average household livestock holdings are interpreted as a manifestation of
capital availability. Manure is dealt with both as a resource in crop
production and as a possible output from the Fulani production system,
with a high opportunity cost in Mossi market gardening.
Chapter five describes the monthly flow of crop and livestock out-
puts from the average household. A comparison of annual cash incomes
with the Mossi shows that the herdsmen are more integrated into the mar-
ket economy than their peasant farming counterparts. A comparison, never-
theless, of the overall value of production per active worker shows that
the Fulani and Mossi are equally well-off from an economic standpoint.
An analysis of household expenditures demonstrates that foodstuffs,
particularly staple grains, account for the majority of purchases.
Chapter six serves to construct a model of the typical sample farm,
based on the average values for household inputs and outputs reviewed in
earlier chapters. The analysis attempts to clarify the trade-offs be-
tween products in production and the constraints to increased output.
Chapter seven includes a summary profile of the typical Fulani
production unit in the research area and argues for a new model for
the integration of crop and livestock activity in the Savannah, based
on the traditional relationships between Fulani herdsmen and peasant
crop.farmers. The chapter argues that there is an urgent need for policy
action in this domain, since bad feelings over crop damage and cattle
theft are weakening traditional ties. The study concludes with the
argument that there is considerable scope for policy-induced improve-
ments in the pattern of relations between the Fulani and their farming
neighbors. The most striking examples of desirable policy intervention
concerns improved communications permitting the exchange, via the mar-
ket place, of herding services and manure for food grains.
THE RESEARCH SITE AND ITS FULANI INHABITANTS
The environmental characteristics of the research site and intro-
duction to the principal ethnic groups living therein are contained in
Delgado f1978], chapter 2. The present chapter gives an updated descrip-
tion of the 1977 Fulani herder study research site, which represents a
portion of the research area covered by the 1976 farm management survey
among the Mossi and Bisa. The criteria for the choice of location and
brief introductions to the Mossi or Bisa ethnic groups are to be found
in the original study. However, an expanded introduction to the Fulani
is provided below for the convenience of the reader. The characteristics
of the research site are fairly typical of the sudannic Savannah in West
Africa, with the exception of a relatively high population density. The
Fulani inhabitants of the area, on the other hand, closely resemble other
pastoralists in southern Upper Volta.
Location and Characteristics of the Research Site
The research site chosen is in the sub-prefecture of Tenkodogo, in
the center-east district of the country (see Figure 2.1). The greater
research area for the 1976 farm management survey, from which villages
were selected and within which all subsequent interviewing took place,
comprises 450 square kilometers in a rectangle approximately thirty kilo-
meters long and fifteen kilometers wide (see Figure 2.2). The town of
1The center-east Administrative district (Prefecture) corresponds
exactly to the borders of the "Koupela" O.R.D., so-called because O.R.D.
headquarters are situated in the sub-prefecture of Koupela.
LOCATION OF TENKODOGO IN AFRICA
scol: 1/66 000 000
SOURCE: Delgado 1978, p.19.
ETHNIC BOUNDARIES AND RESEARCH AREAS
Tenkodogo is close to the midpoint of the eastern boundary of the rectangle.
The exact research area for the Fulani study, for which no precise
measurements are available, is defined by the contours of Oueguedo canton
and environs, settled primarily by the Mossi. Scattered pockets of
Fulani herdsmen maintain more or less permanent quarters on the periphery
of the Mossi living areas.
The 1976 population residing permanently in the Fulani study re-
search area included approximately 7,000 persons. Fulani pastoralists
accounted for nearly 5 percent of this figure. Table 2.1 shows that this
concentration of Fulani is below the average for the Mossi Plateau as a
whole. It should be remembered, however, that the figures for the re-
search area only include those herdsmen who are permanent residents of the
Oueguedo area. If semi-transient herds are also included, the proportion
of Fulani herdsmen present at any given time in the research zone is likely
to be much higher.
The overall population density in Oueguedo canton is relatively high,
at ninety-five inhabitants/km2. This may be compared to our average of
forty inhabitants/km2 in the greater region and a national average of twenty
Besides the canton of Oueguedo, as defined by the Voltaic admini-
stration, the greater research area includes the independent villages
of Pouswaka and Gando which, for political reasons, report directly to
the Chief of Tenkodogo. These villages form a part of the greater
Oueguedo region geographically and economically, and are in fact separated
from Tenkodogo by Oueguedo village. Since many of the Fulani herders
who keep cattle for Mossi farmers living in Oueguedo reside in this region,
it was included in the greater research area. This makes use of second-
ary data more difficult, since data for Pouswaka and Gando are often
aggregated into totals for Tenkodogo canton.
PERCENTAGE DISTRIBUTION OF POPULATION IN
UPPER VOLTA BY ETHNIC GROUP 1960-61
Sahelian North of Savannah Area) Total
Mossi 8.2 82.5 48.0
Bisa 0.5 8.5 4.7
Fulani 81.2 7.0 10.4
Other 10.1 2.0 36.9a
Total 100.0 100.0 100.0
SOURCE: RHV, Service de Statistique et de la Mecanographie, Enquete
D6mographique Par Sondage en Republique de Haute-Volta 1960-61, Volume I,
aLobi 7.0%, Mandingo 6.9%, Bobo 6.7%, Senufo 5.5%, Gourounsi 5.3%,
Gourmantche 4.5%, Miscellaneous 1.1%.
The town of Tenkodogo is roughly two-thirds Bisa and one-third Mossi.
The traditional ruler of central eastern Upper Volta, the chief of Tenkodogo,
is one of the three Mossi kings. The town is a district headquarters for
the Voltaic administration (chef-lieu de prefecture), with a population
of 8,000 people. A consumer-oriented market is held every three days;
large numbers of people attend from Oueguedo and surrounding areas. Oueguedo
proper is eight kilometers to the north-west of Tenkodogo market.
An all-weather road through the town was completed in 1974, linking
the Togo frontier (100 kilometers to the south) with the capital, Ouagadougou
(185 kilometers to the northwest). Tenkodogo is at an ethnic crossroads
in Upper Volta, since the region is astride the southernmost limit
of the Mossi "Plateau," a name applied to the northwest-southeast band
running across Upper Volta that contains most of the Mossi population.
Gourmantche territory begins some fifty kilometers to the east of the
The land around Tenkodogo is representative of the Savannah, approxi-
mately 700 kilometers inland from the Ghanaian coast at the longitude of
Accra. The trees are primarily deciduous fire-resistant varieties. After
the rainy season, abundant grass cover is available, including high quality
fodders such as Andropogon Gayanus in wetter areas [Benoit, 1974, pp. 20-23].
Grasses rarely grow over one meter and large areas are consumed by busy
fires every dry season. Mango and shea nut trees are found in abundance.
The research area is primarily crop-oriented. Where there is a con-
flict between crop-growing and stockraising, it is usually the latter
which suffers [Delgado, 1977, 1978]. The Mossi farmers of the area de-
rive their cash incomes from crop sales rather than livestock, even if
the latter is defined to include poultry [O.R.S.T.O.M., 1975, II (3),
Figures 17 and 18].
The density of the cattle population in the research area is approxi-
mately nine head per square kilometer. Cattle density is higher in
Loanga canton to the immediate south of the research area, which includes
large, nearly unpopulated, stretches of the White Volta Valley. The
density figures may be compared to national estimates for 1974 of 6.3
animals per square kilometer in the southern Savannah and twelve in the
central Savannah. Using the same classification of areas, Tenkodogo
would fall near the line of demarcation between central and southern zones.
Thus, it can be said that the research area carries approximately an
average load of cattle.
The Tenkodogo region is subject to trypanosomiasis-bearing insect
infestation. For this reason, most of the cattle kept are of mixed breeds,
with smaller trypano-resistant varieties well represented. The presence
of tsetse flies only becomes apparent when moving into the southern portion
of the greater research area, near the White Volta. It is also in this
region that onchocerciasis has had a visible effect on the human popula-
tion. Figure 2.3 shows the approximate relation of the greater research
site to the Onchocerciasis Control Program.
In sum, the research area is, on most counts, typical of the sudannic
Savannah belt stretching across West Africa, immediately to the south of
the Sahel. It is a little too dry to be of much value as a major cash
crop area, yet its agriculture makes it self-sufficient in food, with a
small surplus during years of average rainfall.
The rainfall for 1976 measured at the O.R.D. field station in Tenko-
dogo was considerably below average, particularly during the crucial
period from the end of July through the beginning of August. On the other
hand, rainfall in September and October was very high, which interfered
with the pollination of the millet ears and led to very low grain forma-
tion on each plant (ibid.). There was a widely-held belief within the re-
search area during the 1976 harvest that 1977 would be a deficit year
when most families would have to either drastically reduce consumption
or purchase grain from outside sources. This was later confirmed by events.
The 1977 rains, on the other hand, started quite late in June, yielding
FIGURE 2.3 : Relation of the Research Area
to the Onchocerciasis Control Program
Source : World Health Organization (1973) fig.
64, updated by the author.
Planned Project Zones
U Suggested Project Zones
Ghana Upper Volta Frontier
a correspondingly poor harvest of millet to feed people through 1978.
Generally speaking, rainfall in the area is sufficiently high to
permit the growing of cash crops. A twenty-five year average of annual
precipitation is 950 mm in eighty days of rain. The principal limitation
on cash crops appears to be the low fertility of the soils. Soils in the
northernmost parts of the greater research area are typically composed of a
thin layer of tropical ferruginous soil over clay and sand. To the south
soils are dark, with a content of over 30 percent clay on top of granite-
derived clay and sand mixtures. Formerly fertile soils are poor over
the entire area due to intensive cropping with little fertilization and
insufficient fallow. The crops most often sold are peanuts, rice from
lowland rainfed plots, vegetables, and red sorghum.
The principal food crops are pearl millet (the staple), red sorghum,
and cowpeas. Small amounts of maize are cultivated on organically ferti-
lized garden plots. A small amount of cassava is grown as a hedge against
crop shortfalls. The research region is more than self-sufficient in
overall cereals production in years of average rainfall. Taste and price
factors, however, may encourage the exchange of red sorghum for pearl
millet produced in other regions. The conventional wisdom indicates that
the Tenkodogo area is a potential "breadbasket" for the Koupela O.R.D.
[Garey and Storm, 1972].
1Jeune Afrique  pp. 14-15. In 1976, rainfall was under 800
mm in 56 days.
2Data taken from the O.R.S.T.O.M. 1/500000 soils map of the area.
See O.R.S.T.O.M., .
The one characteristic of the research area which may be unrepre-
sentative of the Savannah area as a whole, as opposed to Upper Volta in
particular, is the relatively high population density of 41 inhabitants
per square kilometer. However, the problems associated with population
density are precisely those which exacerbate the difficulties of intensi-
fying livestock activities in Savannah areas. Since increased levels of
population density are likely to occur in the near future, it is in this
kind of environment that the constraints on increasing livestock pro-
duction in southern areas should be defined.
The Fulani of Greater Oueguedo
The Fulani are a pastoral group spread over the length and breadth
of West Africa. They are also the predominant population group in the truly
Sahelian part of Upper Volta in the Djibo-Dori-Sebha axis and the second
most numerous ethnic group in Upper Volta after the Mossi [Jeune Afrique,
1975, p. 26]. Scattered pockets of Fulani live in the savannah areas of
Upper Volta, the most important concentrations being in the Nouna-Bobo-
Dioulasso axis in the west, the Ouagadougou-Kaya axis in the center, and
the two isolated pockets of Kantchari in the far east and Tenkodogo in
the center east ibidd., pp. 28-29).
The Fulani in the research area live in groups of one or two isolated
compounds, usually in bush areas away from Mossi or Bisa settlement. Among
the Fulani, a compound refers to a geographically distinct cluster of
woven straw huts. Unlike the Mossi and Bisa, there is no wall around
the "compound." Very often a Fulani compound consists of only members
of a nuclear family, or several nuclear families with the same husband in
the case of polygamy.
Unpublished research by Tahirou Diao a Fulani born in the Oueguedo
area, indicates that the herdsmen of that region arrived from the Macina
area of Mali in the 18th century. This was before the arrival of the
current ruling Mossi family. The Fulani of Oueguedo should be considered
permanent residents of the area, rather than recent arrivals from the north.
This last finding may be contrasted with research on the Fulani living on
the east bank of the White Volta Valley, whose case may be analogous to
that of the Fulani in Loanga living in the remoter parts of the canton.
During a census conducted in July 1976, Rochette [1976, p.10] found that
40 percent of the Fulani in the area had arrived within the last three years,
20 percent within the last two years.
The Fulani of Oueguedo were politically absorbed by the Mossi con-
quest of Tenkodogo. Thus, the Mossi and Bisa canton chiefs of the Ten-
kodogo area -- vassals of the Mossi king of Tenkodogo -- each have "their"
Fulani, a qualifier still in use today.2 Each peasant canton chief has
a Fulani advisor, called the Fulani chief. The Fulani chief reports to
the canton chief.
In Oueguedo, the principal role of the Fulani chief is to act as a
go-between for the herdsmen and the canton chief. In particular, he
1Cited with permission. For details, Mr. Diao can be contacted via
the Catholic Relief Service Office in Ouagadougou. This finding was con-
firmed by both the canton chief of Oueguedo and Fulani community members.
2Being called this, incidentally, galls the Fulani more than any other
qualifier, in my experience.
collects the taxes and remits them to the chief. If the canton chief
has received a complaint concerning a herdsman, it is the responsibility
of the Fulani chief to bring the accused to him for judgement.
In some respects, the Fulani chief appears much like a Mossi village
chief. However, his Fulani "subjects" are spread over several villages.
While the peasant village chief has considerable real power in the
allocation of land, the Fulani chief must operate primarily through moral
suasion. Furthermore, the Mossi peasants readily accept the authority of
the village and canton chiefs. The Fulani herdsmen, on the other hand,
appear to have a more independent position vis-a-vis both.
Nevertheless, real power resides with the canton chief, and it behooves
all who wish to continue living in Oueguedo to avoid his wrath. By Mossi
custom, the descendants of Fulani families who occupied land in Oueguedo
before the arrival of the ruling Mossi clan have a right to some land.
In practice, the precise location of that land depends on the good will of
the canton chief.
The Fulani in the research area are all Muslims, in contrast with the
Mossi and Bisa, who are primarily Animist but also include a moderate num-
ber of Christians and a small number of Muslims. There is some fraterni-
zation between Muslim peasants and Fulani herders, but little inter-
marriage. There is a moderate degree of intermarriage between Mossi and
Bisa peasants within each religious category.
The Fulani family structure is renowned principally for the in-
dependence of the women and the individualism shown by sons, who often
move far away from parental authority after marriage. A woman's male
relatives have particular responsibilities towards her children in
matters of cattle ownership and inheritance.
The Fulani of Oueguedo may be classified as "semi-sedentary" in
their herding practices, since part of the family unit occupies the family
compound year-round.1 On the other hand, some of the young men take
the herds to better pasture and water points in early May, returning two
months later to help with cultivation.
The Economic Roles of the Fulani in Tenkodogo
and the Cattle-Entrusting System
Like their Mossi and Bisa neighbors, the Fulani in Tenkodogo culti-
vate crops primarily for their own use. However, they have two additional
economic roles. They function as herd managers for Mossi peasants and
civil servants, who "entrust" cattle on a permanent basis into their care.
They also raise their own cattle,which is their traditional and preferred
calling. Each of these functions will be examined in turn.
During the rainy season, the Fulani of Oueguedo usually cultivate
fields in a circle around the compound. Cattle are kept at night in a
thornbush corral, just beyond the fields. The Fulani claim to particu-
larly dislike agricultural activity, but admit to its necessity.
1This classification follows that of Rochette .
2This section draws heavily upon Delgado , pp. 37-42. Addi-
tional information on the economic roles of the Tenkodogo Fulani is
contained in Delgado , pp. 23-43.
3In this, the Fulani of Oueguedo resemble their northern cousins
in the Djibo area. See Riesman , p.76.
The crop mixtures planted resemble those of the neighboring Mossi.
Red sorghum is planted in the immediate vicinity of the compound. The
location of the corral from the previous dry season is planted with
maize and sometimes cotton. The rest of the cropped area around the com-
pound is intercropped with red sorghum, millet, and cowpeas.
These results can be compared with the findings of Rochette [1976,
p. 3] for the White Volta Valley area fifty miles northwest of the re-
search site. He found that the Fulani there very rarely grow red sorghum,
but do grow white sorghum. In Oueguedo, white sorghum is almost never
grown. Both the peasants and the Fulani claim that this is because yields
are low. Unlike the peasants, the Fulani will eat red sorghum as a food
grain, since the gritty dumpling it produces is palatable with milk, but
reputedly unappetizing without. All the Fulani interviewed, however,
claimed that millet was the staple.
Red sorghum is grown intensively in the whole Tenkodogo area, since
yields are reportedly very good.3 The principal use for red sorghum
is for making beer which the Fulani, as Muslims, are not supposed to drink.
However, red sorghum may be considered a "cash crop," even if not sold.
The Fulani can receive help in their agricultural tasks from their peasant
neighbors by having peasant women brew beer with their sorghum. They
then distribute it to the work force in the fashion of Mossi labor invi-
tations. Most families affirm that they need to buy 2 to 3 sacks (about
250 kg total) of millet per year.
iProbably due to a mineral deficiency in the soil.
The peasants eat red sorghum only as a last resort, as in the case
of the U.S. food aid in 1974.
3This assertion will be supported further on in this study.
The primary occupation of the Fulani is stockraising. They function
in this capacity both as herd managers for Mossi cattle owners and as
livestock owners themselves. The cattle under the care of one household
are usually kept in one corral, adjacent to the compound. The Fulani
prefer to keep the number of herd in one corral under forty in order to
provide some insurance against the spread of disease. Easier control
and, perhaps, greater discretion also provide motivations for individual
herds not exceeding this size. A relatively rich household may have
several corrals. A compound may also have several smaller corrals if
married sons elect to stay in the compound of their father.
A typical household herd contains about forty animals, two thirds
of which may belong to peasants. In return for herding services, the
Fulani occasionally receive small gifts of cash (500 CFA) and cola nuts;
the usual renumeration is possession of the milk from the cows. Women sell
all the surplus milk on their own account, while the head of household has
rights over the crops, with the exception of small amounts of cotton grown
by the women. The milk yield is often so low from December until May that
the herders leave all the milk to the calves at this time.
Besides milk from cows belonging to peasants, a traditional element
of the remuneration for herding services has been the possession of the
1This will be elaborated upon below.
manure in the corrals. Every Fulani with whom the author spoke confirmed
that dung is collected and spread on the fields. Recently, however,
peasant proprietors have begun to request use of manure, sending their
children with baskets to collect it. This appears to be a manifestation
of the awareness of the declining fertility of in-village soils, in addition
to the expansion of market gardening and the traditional use of manure
as a building material. Cattle droppings form an essential ingredient of
indigenous cement. The Fulani, for their part, resent these requests and
regard usufruct of the cattle manure as their right -- a proprietor who
wants some should pay for it.1
Management decisions concerning the portion of the herd belonging
to peasants are most often left to the herder, with final approval com-
ing from the owner. Purchases and sales of healthy animals are made at
the request of the proprietor. A peasant expresses to his herdsman the
desire to buy an animal; the herdsman then transmits the interest to other
herdsmen by word of mouth. The radius of search may include herds in tran-
sit south and sellers up to fifty kilometers away. When the herdsman
finds the type of animal sought by the peasant, he negotiates over the price
with the seller. The latter is another herdsman who may be representing
either himself or a peasant. When a tentative price is agreed upon, the
herdsman brings the prospective proprietor to see the animal, or simply
gets his agreement based on description and price. The peasant gives the
money to the herdsman who concludes the transaction and collects the animal.
I am indebted to Tahirou Diao for this point.
When a peasant proprietor wishes to sell a given animal, he informs
his herdsman. The Fulani locates a buyer. In the case of very old
cattle, this may be an agent for a butcher from Tenkodogo. Animals in
their prime are often sold to collectors from the cattle market in
Pouytenga, fifty kilometers north of the research area. The herdsman re-
ports a prospective price to the owner who either accepts or declines.
There is no easy way to know whether the actual sale price corresponds
to the amount of money the peasant approves. If the proprietor is pleased
with the sale price, he may give 1000 CFA to the Fulani.
Urgent sale cases occur when an animal is sick. Judgement as to
the animal's chances are left to the herdsman. Most of the Fulani inter-
viewed said that they would sell an animal without warning the proprietor,
if necessary. The remainder said that all the proprietors of their herds
were sufficiently close that they would make every effort to get approval
before sale. When a sick animal dies in the bush, the Fulani brings the
head of the animal to the proprietor as proof of the death. When a sick
animal is sold, the proprietor must accept the herdsman's word on the price
received, which may be one fifth of that of a healthy animal. If an animal
belonging to a peasant is sick, but not in danger of death, the herdsman
treats the animal and eventually informs the owner. If an animal is
taken to the veterinary service in Tenkodogo, the proprietor is warned in
advance since he must pay for vaccinations and medicine.
The herdsman can decide when to begin milking a cow and when to wean
a calf without informing the proprietor. The herdsman must inform the
owner of all births. The offspring always accrue to the proprietor and
are almost never given to the herdsman. It is difficult for the herds-
man to be dishonest in this matter, since proprietors visit the herds
often enough to be aware of calving.
The annual migration in search of better pasture is the herding opera-
tion that requires the greatest bond of trust between peasant proprietors
and herdsmen. The majority of Oueguedo herds leave the area of peasant
settlement for over two months, beginning in early May and ending in July.
At this time the herds may be sixty kilometers away from the peasant owners.
All management decisions must be made by the herder alone. The herdsman
cannot abandon the rest of the cattle to bring back the head of a dead
animal for the proprietor to see, yet many animals die at this time when
cattle are weakest, just after the dry season.
One third of twenty Fulani families interviewed denied informing
owners when they leave on transhumance. The other two thirds replied
that they informed either the owners or the Mossi chiefs of Oueguedo or
Pouswaka before leaving. If an owner objected to his cattle being taken,
his animals are left with Fulani household members who stay in Oueguedo,
to fend as best they can. An owner very rarely interferes, however, with
the management of the cattle by the herdsmen.
Besides herding cattle for the peasants, the Fulani breed and keep
their own cattle. Property rights over cattle among the Oueguedo Fulani
are basically the same as for Fulani elsewhere in Upper Volta. Formally,
all cattle in the household corral that do not belong to peasants, or
For the western region, an excellent account is contained in T.
Queant and C. de Rouville [19691, I, pp. 181-89. For the north, see
Riesman , p.95.
other owners outside the household, "belong" to the head of household.
However, each family member can "own" cattle within the household herd,
in the same sense that an automobile may belong to an American teenager, even
if it is registered in his parents' name. The head of household can
sell his own cattle and that of his unmarried sons without the consent
of the latter, although this will provoke dissension within the family.
In theory, the husband cannot sell cattle belonging to a wife without her
consent, although he can exert considerable pressure on her to agree,
should the need arise.
Unmarried children and wives cannot sell their own cattle without
the consent of the head of household. A married son living in the com-
pound will not sell any of his cattle kept in his father's herd without
informing him in advance. When children marry, they acquire full property
rights over the cattle they own. Cattle belonging to a daughter are moved
to the corral of her husband at the time of marriage. A son can choose to
build his own corral and compound, at which point his cattle transactions
are completely independent of his father. If a woman seeks a divorce or
if she abandons her husband, her cattle revert to him. If the husband
seeks to repudiate the wife, without due cause in the view of her family
(such as flagrant adultery), he must return her cattle to her male relatives.
The Fulani of Oueguedo have five legal ways of acquiring ownership
of cattle. These are inheritance from a parent, gifts from a father or
a maternal uncle to a child, acquisition from a son-in-law as bride-
wealth, gifts from a proprietor of cattle for herding services (rare),
and purchase of young cattle with proceeds from the sale of old stock.
Often, cattle acquired as bridewealth will be returned to the son-in-
law in the form of a gift to the children of the match, at their birth.1
Although not necessarily the same animals.
HOUSEHOLD LABOR USE
The timing and quality of labor input to different enterprises com-
prises a key element in traditional African agricultural systems. This is
particularly true to the extent that crops, weeds, and pasture are dependent
upon the seasonality of rainfall. An examination of the Fulani agricultural
calendar for the rainy season in Tenkodogo shows a careful adaptation to
the constraints imposed by the simultaneous cultivation of crops and
maintenance of cattle. Herding is always the major Fulani pastime, and
their crop enterprises are roughly half as labor intensive per unit of land
as those of their Mossi neighbors. Somewhat inconclusive evidence suggests
that traditional stockraising is characterized by increasing returns to
scale up to roughly forty head per household. Both crop and livestock
enterprises require peak labor inputs in July and November, raising the
possibility of severe labor shortages at these times.
Household Labor Availability
The average Fulani household in the sample is composed of a male head
of household under sixty years of age, his wife, four children (possibly
including a younger sibling), and perhaps an aged parent. The mean house-
hold composition is given in Table 3.1, next to comparable figures for
neighboring Bisa and Mossi households. Statistical tests suggest that the
average herder household is larger than the average farming household in
AVERAGE NUMBER OF FULANI WORKERS IN EACH
HOUSEHOLD BY AGE AND SEX CATEGORY COMPARED
WITH PEASANT FARMS
Stratum: Bisa + Mossi Fulani
(No. of Households) (N = 41) (N = 19)
Category Mean S.D. Max Min Mean S.D. Max Min
M 8-14 .63 .77 3 0 .89 .94 3 0
L 15-60 1.66 .85 4 0 1.95 1.08 5 1
S 60+ .17 .38 1 0 .37 .50 1 0
E 8-14 .48 .64 2 0 .63 .76 2 0
A 15-60 2.01 1.09 5 1 2.26 1.69 7 0
E 60+ .24 .49 2 0 .32 .58 2 0
Total Family 5.22 2.36 11 2 6.42 2.95 13 2
Tenkodogo. However, a Mossi or Bisa compound is often larger than its
Fulani counterpart, since the former may contain several households, while
the latter contains only one. A household is defined as those people who
work on a regular basis on the fields, or with the herds, of the head of house-
The Agricultural Calendar
The agricultural calendar in Tenkodogo during the 1977-78 season is
given in Table 3.2. The rains began in earnest relatively late in the
season, in mid-June. The operations listed in Table 3.2 occur in sequence
depending upon the date of the beginning of the rainy season. Generally,
Fulani pastoralists in the Savannah recognize five major seasons.
Dungu is the period from late June until mid-September, corresponding
to the rainy season. The household herd is grazed near the compound, but
away from cultivated fields. Forage and water are abundant, however sub-
stantial time must be spent weeding crops and moving the corral to control
parasites. Dairy production is at its peak at this time and maize is ready
for harvest in August.
A test for the equality of the means in the last line of Table 3.1
was able to reject the null hypothesis of equality at the 5 percent
2This is the logical unit of analysis for the study of the labor
3The distinction of five seasons and the Fufulde terms are taken
from Van Raay , pp. 108-109.
TENKODOGO FULANI AGRICULTURAL CALENDAR (1977-78)
(Divided into Fortnights)
9 May 22
23 May 5
6 June 19
12 Sept.- 25
26 Sept.- 9
10 Oct. 23
- 6 Nov.
- 20 Nov.
- 4 Dec.
- 18 Dec.
- 1 Jan.
- 15 Jan.
Principal Activity of Sample
Herds away on transhumance
First rains; first planting of
sorghum; herds begin returning
Maize planted; second planting
of sorghum; millet planting begins
Millet planting; all herds have
returned to Oueguedo; first
weeding of cereals
Second weeding of cereals; ridging;
herds are grazed in bush areas
Maize harvest;period of maximum
milk yield and sale of dairy pro-
Relative slack; greatest period of
crop damage incidents in peasant
Sorghum harvest; end of rains.
Beginning of cold season:
gathering, threshing, and
storage of cereals and legunes;
cotton harvest; livestock grazed
on field stubble close to home
Surface water dries up; wells dug;
livestock grazed on field stubble near
Mossi villages; no more surplus milk
Hot season: cattle herds leave
for river valleys on
Yamde is the period from after the harvest of red sorghum in
September until the cold period in December. It is a time of plenty
since the new harvest is gathered and cattle are at their best after
feeding on the abundant rainy season grasses. Nyaile corresponds to
the end part of this period when livestock are grazed on crop stubble in
Dabunde is the cold dry season from December through February.
Cereals are threshed and winnowed in the harmattan winds from the north-
east. Crop residues in the fields become scarcer, and the herd is often
grazed on the fields inside the Mossi village. This is the period when
manure is moved from corrals onto the sorghum fields. Surplus milk for
human consumption is no longer available.
Cheedu corresponds to the hot dry season from February through April.
Surface water disappears, necessitating a substantial amount of labor to
dig wells in lowland areas. By mid-March most of the Oueguedo herds have
left the research area for the annual transhumance. The Fulani take the
animals up to sixty kilometers away into the river valleys at this time
for water and pasture.
Seeto is the transition period after the very first "mango" rains
in April, but before the beginning of the rainy season in late May or
early June. Among the Mossi and Bisa this is the prime period of field
preparation. Most of the young Fulani men are away with the herds in
the river valleys at this time. Cattle mortality reaches its peak now,
due to heat stress and the lack of forage. The herds return to Oueguedo
after the start of the rainy season when the new grasses make their
first appearance. For this reason, the Fulani often plant after their
The Division of Labor by Age and Sex
The purchase, maintenance, and sale of livestock is primarily..a
male concern. Small boys help with herding chores during the entire year.
From December until March, the herd may be left solely in the care of
children twelve years of age or younger. During the rainy season, however,
the animals are supervised by older boys and men. This is because the
animals are harder to control at this time due to the proximity of crops.
Great care must be exercised to ensure that the animals do not damage the
plants in peasant bush fields. A greater than normal effort is required
in August to prevent cattle from eating grasses that are too wet [Van Raay,
1975, p. 110]. The motivation for this reflects a concern about diseases
connected with wet forage and the fear that the animals will eat less dry
matter, leading to a smaller weight gain (ibid.).
During the hot season from March to May, adult male labor is also
required to water stock from wells. The annual three month transhumance
which occurs at this time requires approximately one adult male herdsman,
full-time, per fifteen to twenty head of cattle. The division of labor
observed in the sample is portrayed in Table 3.3.
The planting, weeding, harvesting, and threshing of millet and sor-
ghum, the food staples, are nearly the exclusive domain of men. Women
help, however, with cutting the ears of grain off harvested stalks (like
DIVISION OF LABOR OF FULANI HOUSEHOLDS
( x= Primarily Performed by This Group)
Visit Cattle Proprietors
Guard at Night
Fetch Feed for Calves
Feed Salt and Millet
Bran to Dairy Cows
Plant and Weed fields
Harvest: Cut Stalks
Cut Ears of Grain
Years and up
Years and up
(rest of year)
(rest of year
if not enough
TABLE 3.3 (cont.)
LABOR 20 FULANI HOUSEHOLDS
Years and Up
Years and Up
Taking Care of Children
Milk and Butter
Vegetables and oil
Mossi women) and the spreading of manure on the fields.
Maize and cotton, nevertheless, are primarily cultivated by women.
Maize provides a staple during the rainy season, just prior to the
harvest. The women spin the cotton from their own fields.
Milking and the preparation and sale of dairy products are exclusively
female concerns. In practice, surplus milk is available for human con-
sumption from June until December.
Household Labor Allocation by Major Sector of Activity
The average number of hours allocated by a sample household each fort-
night to different categories of work is depicted in Figure 3.1. The
graph of labor allocation covers the interviewing period from June through
December. The upward slope of all the lines in Figure 3.1 during the first
five fortnights is attributable to two related phenomena. First, the
young men are returning home from the seasonal transhumance, up until the
beginning of July. Their labor hours are not included in the graph un-
less they are physically present in the household. Second, late July
(fortnight 6) represents the peak period of work during the rainy season.
The total number of hours worked each fortnight by the average household
attains a maximum during the second weeding of millet and sorghum, which
occurs at this time. Domestic and nonagricultural work are relatively
stable during the agricultural season. Crop work gradually tapers off
after July, except for brief flurries of activity in late September and
mid-November. These periods correspond, respectively, to the harvest of
red sorghum and millet. Most interestingly, livestock work also requires
a maximum amount of labor during the weeding of cereals in late July.
MEAN TOTAL HOUSEHOLD LABOR HOURS
EACH FORTNIGHT BY CATEGORY OF WORK
-e e-0 -
Nov. 21 Dec. 4
Bo- r --
-..- .------ .- .
/ /x. Sept. 12-25
-'r Dec. 19 Jan. 1
T i' I I I I I I I i J I I I i I
22 4 6 8 10 12 14 16 18
This is because the herds are relatively close to the fields at this
time and require a great deal of supervision to prevent crop damage.
The wetness of pastures and the corral also increases the incidence of
parasitic diseases and care must be exercised to keep the animals healthy.
While the principle occupation of the Fulani sample members is always
in the livestock sector, crop work requires a significant commitment of
their time during the height of the rainy season. Figure 3.2 shows that
the ratio of crop to livestock work hours for the average household is
always inferior to unity. However, total crop labor hours are equivalent
to more than half the work devoted to livestock in June. It seems very
likely that this labor commitment to crops is not costless in terms of
foregone livestock output. This may be the case in terms of the number
of animals kept and in terms of the care of the herd during a period of high
incidence of cattle disease. The linear programming model below will help to
clarify some of the costs of Fulani diversification into crop raising.
Household Labor Allocation to Crop Work
Figure 3.3 depicts the total number of labor hours allocated by the
average sample household each fortnight to the three major agricultural
operations of seedbed preparation, weeding, and harvesting. It is note-
worthy that seedbed preparation, which is the operation that would most
benefit from the use of animal traction, requires relatively little labor
input. Furthermore, this input occurs well before the seasonal peaks in
late July. Weeding, on the other hand, constitutes the major demand
Sept. 26 Oct. 9
Dec. 19 Jan. 1
S I I I I I
2 4 6
I I I I I I I I I
10 12 14 16 18
RATIO OF MEAN CROP TO LIVESTOCK WORK HOURS
ON SAMPLE FULANI FARMS, 1977
-MEAN HOUSEHOLD HOURS ALLOCATED TO EACH MAJOR
AGRICULTURAL OPERATION, BY FORTNIGHT
Mean Household Labor Hours
SWeeding and Ridging
S Seedbed Preparation
200 .. .. ... Harvesting and Processing
Sept. 26 Oct. 9
0 2 4 6 8
on household labor resources. However, weeding labor input occurs
during the peak dairy season. Fulani women in the research area typically
do not participate in the weeding of cereals and are in fact fully
occupied in milking, preparing cultured milk and butter, and selling pro-
duce in the local market.
Given the preponderance of weeding work, the reluctance of Fulani
women to engage in it, and the high labor requirements for livestock at
this time, it is not surprising that the Fulani devote far less time
per hectare of food grains than their Mossi and Bisa counterparts.
Figure 3.4 compares the average number of hours allocated each fortnight
by Mossi and Bisa farmers on the one hand, and Fulani sample members on
the other, to one hectare of red sorghum and millet intercropped with cow-
peas. This mixture represents the most common food-growing arrangement
in Tenkodogo. Clearly the Fulani mode of cultivation in the research
area uses far less labor input per unit of land than the Mossi and Bisa
farming system. Other things being equal, lower yields are therefore to
be expected from the Fulani efforts, because of greater weed growth and
less timely harvesting. It will be shown below that the judicious use
of cattle manure as fertilizer in fact allows the herdsmen to surpass the
crop yields of peasant farmers on small portions of the Fulani fields.
1The data for forty-one Mossi and Bisa farms are taken from Delgado
 chapter 4. They relate to the 1976 growing season. Some allowance
should be made for the fact that the rains began earlier in 1976 than in
1977, implying that some operations were in fact performed a little earlier
Mean Household Hours
COMPARISON BETWEEN PEASANT-FARMER AND FULANI HOUSEHOLDS
OF LABOR HOURS ALLOCATED TO ONE HECTARE OF FOOD GRAINS
Mossi and Bisa House-
20 July 3
Aug. 29 Sept. 11
2 4 6. 8
Household Labor Allocation to Livestock
Given the exclusion from the interviewing period of five dry
months from January to May, the labor allocation to cattle reported
below does not reflect the increase in cattle labor brought on by the
need to water animals from wells and to graze them farther away from the
compound during the dry season.- The estimates here, however, do serve
to show the relationship between the cropping and pastoral calendars
during the agricultural season.
An examination of the total number of hours allocated to cattle work
by members of the average household each fortnight, depicted in Figure
3.5, shows two peak periods in addition to that of late July discussed in
the previous section. Prior to the sorghum harvest in late September
and early October and after the millet harvest in mid-November, the
average fortnightly labor allocation to cattle increases notably. These
peaks correspond to periods of relative slack in crop cultivation. The
care of animals is restricted to essentials during the busy season in
raising crops; however, they receive more attention to compensate for
this when the herdsmen have the time.
The extra hours worked with the cattle herds during the September
and late November peak periods mostly involve longer grazing of the animals.
This extra work is performed primarily by adult male herdsmen, as shown
in Figure 3.6. The peak labor allocation to cattle in late July also
involves increased grazing work, however the input of adult male workers
is relatively constant at this time. The extra herding labor required
to maintain the animals during this, the most humid part of the rainy
Mean Household Hours
MEAN HOUSEHOLD LABOR HOURS ALLOCATED
EACH FORTNIGHT TO CATTLE WORK
Nov. 21 Dec. 4
1 1 1 1
Mean Household Labor Hours
MEAN HOUSEHOLD HOURS ALLOCATED
TO GRAZING CATTLE
Sept. 12-25 Nov. 21 Dec. 4
June 20 July 3 -- -
-- (Male 15-60)
I I I I I I I I I I I I I I I
2 4 6 8 10 12 14 16
season, is provided by other workers (boys, girls, and older people).
The men in their prime weed the food grain crops at this time. Although
not shown in Figure 3.6 which covers only the cropping season, the labor
input to cattle tasks of male workers in their prime decreases after the
end of December. It will rise again in early March when surface water
runs out. Men then dig wells and perform the very arduous labor of
drawing water for the animals.
The labor required to look after one head of cattle and one small
ruminant is ascertained by dividing the number of hours allocated to each
of these pursuits in a given fortnight, in each household, by the number
of animals kept by the family in question. The mean of these figures over
households yields the average labor allocation each fortnight per animal
kept. The results, which are given in Figure 3.7, are useful in a linear
programming analysis of farm production strategies. Because of the possi-
bility of economies of scale in stockraising, however, these figures are
valid only for herds of approximately the same size as the sample herds.
The average in this case is 43 head of cattle and 16 sheep and goats.
Results based on admittedly scanty data suggest that per animal labor
requirements for cattle decline with herd growth when less than forty
animals are kept, but remain relatively constant thereafter. Figure 3.8
depicts the means of fortnightly labor requirements for, respectively,
five herds with from fifteen to thirty head, four herds with forty to
fifty head, and three herds with up to ninety head each. While the curve
for the smallest herd size lies appreciably above the other two, the latter
are quite close to each other. The tentative interpretation is that the
Mean Household Hours per Animal Kept
1 \ *-
/ \ -.
V Cattle (for an average
herd size of 43 animals)
(for an average herd size
of 16 animals)
6 8 10 12 14
Mean Hours per
Animal Each Fortnight
ECONOMIES OF SCALE IN CATTLE-RAISING
(Herds of 15-30
(X = 26, N = 5)
. Herds of 40-50 Head
j(x = 47, N = 4)
/ Herds of 60-90
(X = 75, N = 2)
I I I I I I 1
2 4 6 8
I I I 1 1
12 14 16
overhead of corral maintenance and supervisory labor continues to be
spread by herds in excess of forty animals. Further labor savings per
animal are likely to decrease as herd size increases, however. Further
research involving a much larger number of herds would be necessary to
satisfactorily resolve this issue.
The conflicting labor requirements for typical livestock and crop
enterprises of sample farms are portrayed in Figure 3.9. The primary
competition for resources between the two sectors occurs in late July.
The replacement of millet and sorghum cultivation with maize would have
the effect of decreasing labor bottlenecks in late September and mid-
November, but also of increasing the pressure on resources in August. The
linear programming analysis below serves to define more precisely the nature
and magnitude of resource constraints in production strategies which in-
volve the production of both crops and livestock.
Mean Household Hours
per typical Herd Size
COMPARATIVE LABOR REQUIREMENTS OVER
TIME FOR CROPS AND LIVESTOCK
size = 43
--- -- One hectare maize
-. *. +* One hectare sorghum
- Small ruminants,avera-
ge herd size = 16
0 2 4 6 8 10 12 14 16
LAND, CAPITAL, LIVESTOCK, AND MANURE
Even though the Fulani of Oueguedo are first and foremost herds-
men, sample households manage to cultivate an area equivalent in size, on
the average, to two-thirds of the crop area typically planted by neighbor-
ing Mossi and Bisa farmers. The Fulani household plots benefit greatly
from the manure of the cattle corrals. The neighboring Mossi bush fields,
however, often suffer because of the proximity of the herds during the rainy
season. A greater incidence of crop damage cases in recent years under-
scores the increasing pressure of man and beast upon the land. The adop-
tion of Mossi farming methods by the Fulani of Oueguedo -- when it suits
them -- is never more evident than during an inventory of capital goods
present in herder compounds. The Fulani household cultivates using the
same facilities and equipment as their Mossi neighbors when allowance is
made for the smaller area planted. On the other hand, substantial house-
hold herds are also kept, with surprising proportions of Mossi-owned and
older animals. The dung from the cattle represents a valuable resource
for the Fulani, being used either on Fulani grain crops during the rainy
season or Mossi vegetable gardens in the dry season.
Fulani Land Use
The Fulani sample members grow crops on permanent fields around each
compound. Livestock are grazed on unoccupied land during the crop season,
on harvested fields immediately thereafter, and in the relatively lush
but disease-ridden river valleys during the hot dry period from March
to June. Questions of land allocation at the household level, therefore,
revolve around cropping patterns. The major issue at the community level,
on the other hand, concerns the allocation of land resources between
individual cropping enterprises and collective grazing. The carrying
capacity of the pastures with respect to cattle is a related, but subsidiary,
Fulani Crop Cultivation Compared to the Mossi and Bisa Table 4.1
shows that the average household grew approximately two-thirds of the area
of crops produced by sample Mossi and Bisa farms. Unlike peasant hold-
ings, which are dispersed in up to twenty fields spread over a large area,
Fulani crops are grown in a single circle around the compound. There is
less variety of crops on herder farms, with 93 percent of holdings devoted
to a mixture of millet, red sorghum and cowpeas. The Fulani attempt to
take advantage of the high organic content of the soils stemming from the
presence of cattle. Red sorghum is planted on the choice strips of manured
farmland and relatively large stands of maize and cotton are cultivated
on the site of the previous dry season's corral.
The area cropped per active household member is roughly half that of
the average Mossi or Bisa farm. However, this figure is misleading. Both
men and women work on all crops on the peasant smallholdings. Fulani women,
on the other hand, confine their crop work to maize and cotton, which repre-
sent only a small fraction of the cultivated area. The relevant figure
for comparison is the acreage of millet and sorghum grown per agricultural
worker, which should include both men and women in the calculations for
the Mossi and Bisa, and women only in those for the Fulani. The bottom
line of Table 4.1 demonstrates that by this yardstick the Fulani cultivate
more extensively per worker than their neighbors. This serves to confirm
the finding in the previous chapter to the effect that the Fulani use fewer
labor hours to cultivate a hectare of food grains than the specialized
Fulani Livestock Grazing and Competition for Resources with Peasant
Bush Fields Cattle graze most of the year on natural pastures of the
sort described by Benoit . These consist primarily of seasonal grasses
that grow up to a yard high. The edges of watercourses and other lowland
areas often contain andropogon guyanus, highly prized by the Fulani for
forage. During the early dry season, Mossi crop enterprises are on the
whole complimentary to stockraising, since cattle can browse on crop stubble.
The exceptions to the rule are dry season irrigated gardens, which are pro-
gressively crowding livestock away from the much-prized lowland areas.
The Mossi fences made of woven millet stalks are not always sufficient
to keep large stock, such as cattle, away from the plants.
During the wet season, crop and livestock activities are in direct
opposition to each other with respect to land use. An unsupervised herd
of forty hungry animals can wipe out half a hectare of grain in a matter
of minutes. Naturally, feelings run high in such cases, after the peasant
has worked hard to coax the crop out of the infertile ground. In 1976,
one peasant in the central Oueguedo shot six sheep belonging to the Mossi
This section draws heavily upon Delgado , Chapter 5.
MEAN FULANI HOUSEHOLD FIELD AREA CULTIVATED
COMPARED TO THE MOSSI AND BISA
(Standard error of the means in brackets; all figures in hectares)
Ethnic group and
number of households Fulani Mossi & Bisa
N = 12 N = 58
area of fields per 2.46 3.85
household (1.13) (2.75)
Average household area of 2.30 3.27
millet and sorghum fields (1.08) (2.58)
Average household area of .16 .02
maize fields (.07) (.027)
Average field area per household .37 .79
member over 8 years of age (.14) (.38)
Average area of millet and sorghum .83 .67
fields per agricultural worker (.36) ( )
aThe means are calculated over households. The Fulani observations relate to the 1977-78 agricultural
season, whereas the figures for Mossi and Bisa farmers are drawn from 1976-77 [Delgado 1978, p.1441.
bThe agricultural work force relevant to grain production is the whole family in the Mossi and Bisa
case and males only in the Fulani case (see text)
Estimated; certainly less than .79 hectaresper worker.
Estimated; certainly less than .79 hectaresper worker.
canton chief that had slipped their tethers to invade his house garden.
In hierarchical Mossi society, this might be thought a tremendous affront
to the chief, yet public opinion in Oueguedo was on the farmer's side.
The fields most often involved are the Mossi bush fields cleared
and planted, in many cases, after the Fulani have installed themselves
nearby. The density of habitation and population growth in the Tenkodogo
area are continually pushing the frontier of village fields outward, as
in many parts of the arable savannah. In many cases, the Fulani are
forced to move further away from the center. In addition, land that has
been occupied by the Fulani for any period of time is coveted for its
agricultural potential due to manure fertilization.1 This may encourage
settlement close to Fulani living areas. Figure 4.1 shows bush fields
belonging to Mossi sample members sandwiched between Fulani sample house-
holds. The instances of crop damage from livestock inevitably increase
as peasant fields expand into Fulani areas. An increase in the density
of the cattle population beyond a certain point will tend to drive herds
closer to cropped areas, thereby aggravating current frictions. The in-
crease may, however, reflect the expansion of the cropped area as much as
the growth of herd numbers.
Five out of a total of sixty-two Mossi and Bisa families interviewed by
the author in the Tenkodogo area in 1976 reported at least one instance of
substantial damage to bush fields from cattle in the growing season.2
I am not aware of overt cases of Fulani being thrown off their land
in Oueguedo in recent times in order to plant. However, I have personally
seen several instances of Fulani families being driven off the land thirty
kilometers to the south, the household burned over and cleared for plant-
ing, all in a space of two weeks.
It was not possible to ascertain how many incidents the Fulani sample
had been involved in during 1976.
MOSSI BUSH FIELDS IN OUEGUEDO
SOURCE: Delgado 1978 p.155.
Often the aggrieved peasant cannot find the herdsman who is responsible
for the ravages, although he will ask to find out who was nearby at the
time of the occurrence. If he is unsuccessful in his quest, the damaged
portion of the field is a net loss.
When the offending herd is identified, the peasant makes an appeal
for redress of grievances to his village chief in the first instance, and
then to the canton chief. The canton chief sends for the Fulani chief in
charge of herders in the area concerned and briefs him. The latter is
responsible for producing the offender. In the case of substantial damage,
the case is submitted to the tribunal of the first degree in Tenkodogo.
This court is composed of a middle-level civil servant and two elders of
the ethnic groups concerned. In cattle questions, the advisors would
usually be a Fulani and a representative of the peasant group concerned.
The court requests an assessment of damages from the local representative
of the O.R.D., who sends an agent to note the area involved and the degree
of damage. The tribunal disposes of the case based on the O.R.D. findings
and the principles of customary law of the ethnic groups concerned. The
defendant has thirty days to appeal the decision.
If the defendant chooses to appeal, the verdict is handed down by
the tribunal of the second degree. This consists of a court held by the
Sub-Prefect or Deputy Prefect of Tenkodogo. In theory, a decision may be
Organism Regionale de Developpement or regional development authority.
2Sous-Pref&t and Secretaire-Gen6rale du Department, respectively.
appealed to the regional magistrate in Fada N'Gourma (136 kilometers
by road) and thence to the Supreme Court in Ouagadougou.
During the 1976 growing season, numerous cases were handled by the
canton chiefs. The author was able to locate the minutes of seven cattle
damage trials in the tribunal of the first degree.1 "Several" cases were
heard before the tribunal of the second degree, although the officials
concerned could not remember the exact numbers nearly a year later. No
one at the Tenkodogo administrative headquarters could remember a cattle
damage case ever being appealed to the magistrate in Fada N'Gourma.2
Where clear evidence of damage by the accused is established, which
was usually the rule in cases going to the prefecture, fines and compensa-
tory damages are levied. These might run from forty to several hundred
dollars, a considerable expense for a herdsman. If a convicted herdsman
pleads inability to pay, he is kept in jail until his relatives pay for
The penalties for being held responsible in a crop damage case are
sufficiently severe to ensure great care by the herdsmen during the sea-
sonal migration, or "transhumance." When animals leave during the dry
season, they can take the most direct routes through the villages to get
to the river valleys, as shown in Figure 4.2 The return trip takes place
when the grain sprouts are knee-high. At this time, herdsmen are very
careful to remain on the government-sanctioned cattle trails, often not
sleeping for the two to three days it takes to move the herds back to
1There may have been more.
2This procedure is usually adopted only if long prison sentences are
possible, such as in the case of cattle theft on a grand scale.
the rainy season pastures, according to sample members.
The Fulani Inventory of Capital Goods
The farm buildings and agricultural implements present on Mossi and
Bisa sample smallholdings were noted during the 1976 survey conducted
by the author. Comparable interviews of fourteen Fulani sample compounds
used the same questionnaire in 1977. The results of both sets of inter-
views are given in Table 4.2 Casual observation and the informal compari-
son of the means for the two sets of sample members strongly suggest that
the Fulani are equipped for crop cultivation in much the same fashion as
their peasant neighbors.
The noticeably smaller average number of hatchets and machetes on
Fulani farms is consistent with the lower number of active agricultural
workers per household. The proportion of Fulani to peasant grain storage
facilities is approximately equal to the ratio of the average total field
area cultivated by households in each sample.
The Fulani inventory of capital goods includes corrals, which are
typically made of thorn bushes arranged in a circle of six meter radius,
around a shade tree. Households usually have one such corral for every
thirty to forty herd of cattle, for ease of management and health reasons.
Calves are tethered to stakes, as are small ruminants during the cropping
season. Livestock represent the primary use and manifestation of Fulani
Fulani Livestock Holdings
A preliminary cattle census of Fulani herds in the Oueguedo area was
MEAN NUMBER OF AGRICULTURAL IMPLEMENTS AND
STORAGE FACILITIES AVAILABLE ON EACH FARM
Bisa & Mossi
N = 57
N = 14
Machetes Sickles Silos
SOURCE: Delgado 1978, pp. 371-72. The means are taken over households.
conducted by the author in 1976 [Delgado, 1977]. A subsequent detailed
study of fourteen of these herds occurred in late 1977. The results
in this paper are based on the latter census. The earlier figures in
fact seriously understate the extent of livestock holdings. The 1977
figures reflect a greater herder confidence in the research team, members
of which were allowed to count the animals. The figures on animal numbers
are essentially correct. The figures on age structure of the herds are
tentative, given some confusion in establishing the age of each animal
with the herdsmen. Often the figures represent the best guess of the enumer-
ator. Results concerning the number of animals actually owned by the Fulani
as opposed to entrusted cattle must be interpreted with care and with a
measure of skepticism. The research team had no means of verifying the
ownership, as opposed to the physical possession of animals. In the same
way that it is not difficult to establish the level of overall deposits in
a bank, but hard to know the extent of an individual's savings, the actual
ownership of cattle by household members is a tricky subject. The figures
here are given as estimates of the correct order of magnitude.
Fulani Herd Inventories Approximately 60 percent of the cattle in
the fourteen Fulani herds surveyed in 1977 belonged to sources outside the
household. These consist primarily of peasant farmers in the immediate
vicinity of Tenkodogo. The average size of the cattle herds in 1977
1Also reported in Delgado , chapter 6, from which this section
2The 1976 survey of nineteen Fulani herds reported in Delgado 
counted 120 non-Fulani proprietors of which approximately one-sixth lived
outside the rural area of Tenkodogo.
was forty-three animals. Each household also kept an average sixteen sheep
and goats, with the emphasis on the former. These were all owned by the
household in question. Being Muslims, the Fulani do not keep swine. Table
4.3 presents a summary of the data.
Between December 1976 and December 1977, there was a net increase of
4 percent in the cattle herds of Fulani sample members.2 At the same time,
there was a net decline of over 36 percent in the flocks of sheep and goats.
The primary explanation for this surprisingly large decrease in stocks is
the sale of sheep and goats to buy millet, since the 1976 harvest was
particularly poor. Another possible explanation is that herdsmen are re-
constituting their personal cattle herds using small ruminants
SUMMARY OF FULANI HERD SIZES IN OUEGUEDO FOR 1977
Type of stock Mean Standard Deviation Maximum Minimum
Total cattle 43.14 19.16 90 17
Fulani-owned Cattle 17.29 14.63 50 2
Sheep 9.71 8.33 30 0
Goats 6.43 5.28 22 1
Horses .07 .27 1 0
Donkeys .36 .82 3 0
1These results may be contrasted with the 1976 survey, which found an
average herd size of thirty-eight head, of which 30 percent were owned by the
Fulani.[Delgado 1977, p.34.]
Based upon herder recall in December 1977, rather than a comparison
of two surveys which are not fully comparable, owing to a greater degree
of access to herds in 1977.
raised since the drought in 1974. Both these assertions require more
data and further research for adequate analysis.
The Age Structure of Fulani Herds Data on the age structure of
the Fulani herds show that over half the animals kept are females five
years and older. A surprisingly high figure of one-fifth of the herd is
represented by males five years and older. A top-heavy age pyramid for
Fulani herds reflects in part the high mortality among younger stock.I
Another explanation may be a herder preference for older cattle who have
proven their endurance and fertility, the younger stock being sold to
collectors for the Pouytenga cattle market.2
Figure 4.3 presents the data on Fulani herd structure, in age pyra-
mid form. The most salient aspect of herd structure among the Tenkodogo
Fulani is the large number of older female animals. This is consistent
with the view that pastoralists are anxious to maximize the milk output
from the herds. More surprising is the relatively high proportion of older
male animals. Nearly one animal in five among the 605 cattle surveyed
was a mature male (including steers).
The explanation that the older male animals belong to peasant farmers
who are primarily concerned with a stable store of value is not supported
by the evidence. A survey of Bisa livestock holdings entrusted to Fulani
herdsmen (not necessarily those in the herder sample) showed that not
1Said to be from 35 to 45 percent in Upper Volta. [Peretti, 1976,
I, p. 63.]
2According to sample members, Pouytenga is the destination for young
animals after they are sold. This is a major collecting market and the
regional starting place for cattle drives to the capital or to the coast.
Age Structure of Fourteen Fulani Herds
9% 6 37 %
9% years 14 %
90 years 90
30 20 10 i0 0 1 20 30 40
50 20 10 0 0 1.0 20 30 40
% ( Based on 605 animals.)
one of the thirty households interviewed owned a mature male animal,
although one farmer in three owned cattle [Delgado,1978, p. 178]. Further-
more, the Bisa ownership strategy appears to be consistent with a strategy
of growing out immature animals and selling them before age six.1
A plausible explanation of the age and sex structure of the average
Fulani herd in Tenkodogo is that herdsmen seek to constitute diversified
herds with a high proportion of older cattle who have proved their re-
sistance to disease and drought. A higher than average mortality rate
among younger stock from 1972 until 1974 served to thin out the correspond-
ing entries of the age pyramid (3-4 year olds). This is consistent with
[Peretti, 1976, I, p.661, who notes that drought took a high toll of immature
animals among northern herds at that time.
The Fulani concern with milk production is well documented in the
literature. Use and sale of cattle manure in crop producing areas, how-
ever, has received virtually no attention. [Van Raay, 1975, pp. 57-60] notes
the importance of cattle wastes for crop growing in the West African Savan-
nah. The Fulani practice of grazing peasant village fields against payment
in the form of night-paddocking residues is also cited. However, the
economics of the Fulani use of manure remain unclear.
The Use of Manure by Fulani Herdsmen in the Tenkodogo Area
The Fulani of Oueguedo use cattle manure in three distinct ways.
1Delgado [1978, chapter 6] contains a discussion of the returns to
cattle ownership under different assumptions about growing-out strategies.
First, it is spread on their own fields of sorghum and maize. Second,
it is in effect exchanged for millet and the good will of the Mossi
village during "night-paddocking invitations." Third, it is sold to market
gardeners for cash.
Most of the sample heads of household claimed during interviews that
they collected cattle manure from the corral and spread it on their fields
during the dry season. This could not be verified by the labor use sur-
vey because of its limitation to the agricultural season. The outstand-
ing yields achieved for red sorghum and maize by sample members tend to
support the view, however, that the Fulani are very conscious of the value
of manure as fertilizer. The physical labor of collecting and spreading
dung is alleviated by moving the corral several times into positions close
to the living area during the dry season. These patches are subsequently
planted with maize and cotton.
The "night-paddocking invitation" is an institution whereby a herds-
man will take the animals under his care into the Mossi village; this occurs
mostly during February and March. Field stubble is sufficiently scarce at
that time that the animals are grazed throughout the village, but especially
alongside watercourses. The herd is gathered onto the fields of the pea-
sant host at night, without the benefit of a corral. In exchange for the
droppings thus acquired, the farmer feeds the herdsman who must remain
with the animals at all times to avoid damage to dry season gardens in
the village. The latter are fenced with woven millet stalks, which are
not sufficient to deter large stock. All the herdsmen interviewed stated
that they disliked participating in these arrangements because they are
very tiring. The cattle, it seems, are prone to get up to graze at
all hours of the night.
The relative economic advantages of night-paddocking arrangements are
unclear. Van Raay [1975, p.58] claims that the institution was declin-
ing in importance in the Zaria region of Northern Nigeria because of a
feeling on the part of the farmers that the sacrifice of food is not
worth the extra fertility of the soil. However, a study of the four such
arrangements entered into by Tenkodogo Fulani sample members in 1976
shows that the value of the dung left on farm fields exceeds the value of the
payment received in millet [Delgado, 1977, pp. 46-50]. The exact figures
are tenuous, since it is difficult to accurately assess the quality and
market value of the manure involved. The sample members who did partici-
pate in invitations, nonetheless, were adamant that they did so only be-
cause they owed the farmer involved a favor. They claim that in earlier
times such practices were more prevalent, but that the advent of vulner-
able dry season gardens on in-village lowland areas has made the institution
too much bother to be worthwhile.
If Mossi market gardening has contributed to the decline of night-
paddocking, it has also considerably increased the demand for manure as a
saleable commodity. Some preliminary evidence from the 1976-77 cropping
season showed a market value of 4,000 CFA per metric ton of manure de-
livered to the customer [Delgado, 1977, p.46]. This figure is based on
only three observations, and may somewhat overstate the price that is
received by the average Fulani sample member. The essential point is that
the concept of the sale of cattle dung for cash has been established in
the research area.
The expansion of market gardening in the Tenkodogo area subsequent to
the opening of a paved road to the capital in 1974 is evident to even
the casual observer familiar with the area. This enterprise takes place
primarily in the four months between December and March, when other crop
activity is at a minimum. The timing of vegetable production effectively
precludes direct Fulani participation, since it competes heavily for the
labor and lowland resources at the disposal of the herdsmen for watering
A very conservative estimate of the returns per hectare of vegetables
is 135,000 CFA. If crops are irrigated by hand on land that is fertilized
with cattle manure, the additional resource requirements for cultivation of
one hectare of vegetables is roughly estimated as the labor of two people
for four months. At a wage of 250 CFA per day, this is equivalent to
60,000 CFA in labor costs. If the land is otherwise unused in the dry
season, the returns of the manure that makes cultivation possible are
75,000 CFA per hectare. Estimating the manure requirements for one hectare
of vegetables at twenty-five metric tons, the average return to one metric
ton of dung is 3,000 CFA. If the cost of carting one ton of manure from
the Fulani corrals to the vegetable producer does not exceed 1,000 CFA, then
1This figure was used in Delgado [1978, chapter 7]. It represents
the return under adverse soil and water conditions with labor-intensive
(traditional) technology. By comparison, the Memento del'Agronome,
Republique Frangaise [1974, p.131] cites tomato yields in Africa of
thirty-two tons per hectare, with a 1976 retail value of more than 1.5
The manure requirements is based on a fertilizer requirement for
one hectare of tomatoes of 98 kg of N, 27 kg of P 05, and 143 kg of K 0.
Republique Frangaise [1974, P. 3131. One ton of dry cattle dung is assumed
to contribute five kg of N, three kg of P205, and six kg of K20. Ibid,
the "shadow price" of cattle dung is 2,000 CFA. This figure, which is
very likely on the low side, represents the opportunity cost of one
metric ton of dry cattle manure in Fulani corrals (i.e., F.O.B.). It is
appropriate for the modeling below that this estimate is on the conserva-
tive side. Results which show the treatment and transportation of dung
for fertilizer to be a profitable activity are only strengthened by the
use of a low valuation for the end product.
In any event, the value of 2,000 CFA will be used below as an alterna-
tive household production option with respect to the usual Fulani practice
of using the manure from their cattle on sorghum and maize. The next chap-
ter attempts to establish the details concerning the output, income, and
expenditures of an average household in the sample.
OUTPUT, INCOME, AND EXPENDITURE
The Fulani output of goods and services in the research area com-
prises the cultivation of crops for subsistence and the maintenance of
livestock for subsistence and cash. Cropping patterns reveal the
judicious use of cattle manure to improve yields. The results below
separate production between the crop and livestock sectors, ignoring
for the present that the cereal output observed is not independent of the
number of cattle kept. The components of household livestock production
dealt with below concern herd inventory changes, the sale and purchase of
Fulani-owned cattle, other revenue activities concerning livestock (such
as dairy, small ruminant sales, egg and poultry production), and miscel-
laneous income from herding and the sale of manure. A comparative analysis
of Fulani production patterns with those of their Mossi neighbors shows that
the pastoral people are in fact considerably more integrated into the market
economy than the peasant farmers. An analysis of the monthly flow of expen-
ditures by an average Fulani household shows that foodstuffs, particularly
grains, account for the majority of purchases.
The most striking aspect of Fulani crop raising in the research area
1The data gathered do not permit an accurate estimate of nonagricultural
income, however this appears to be a relatively minor source in the case
of the Fulani of Oueguedo.
is the careful use of cattle manure on a small part of the area cultivated,
in order to produce very healthy yields of maize and red sorghum. Small
plots of maize grown on the site of old corrals and a narrow band of red
sorghum planted in a circle around the living area often produce yields
double those obtained by Mossi farmers on their best house fields, even
though the latter have been fertilized with night soil and goat dung.
Table 5.1 contains the results of forty-three yield plot measure-
ments made during the 1977 harvest.I The mean yield of red sorghum of
2,343 kg of dry grain per hectare represents something of a triumph for
traditional agriculture, especially since the land in question has been
continuously cropped for a dozen years, on the average. Millet grown in
single stands on the same heavily-manured plots yields a respectable average
of 945 kg of dry grain per hectare, which is a fair return in a year noted
for adverse rainfall conditions for millet.
These results must be tempered with four major qualifications. First,
yield plot data often overestimate the true yields involved, especially on
small plots and with only a small number of observations. Second, the soil
characteristics of Tenkodogo are particularly suited to red sorghum, which
makes the prediction of similar results for manure fertilization in other
areas somewhat tenuous. Third, the yields of maize were measured in terms
of humid ears, with only a "guesstimate" as to the appropriate conversion
1Details of the yield plot measurement methodology may be found in
Delgado [1978, pp. 57-58].
YIELD PLOT RESULTS FOR FULANI
SAMPLE CROP ENTERPRISES
(Dry grain in kilograms per hectare)
Traces of cotton
Type of Land
Light or no
Light or no
aThis is based on the broad assumption that the ratio between dry
storable grain and humid husked ears of maize is one to five. The yields
were measured in the latter state.
factor for dry storable grain. Fourth, and most importantly, these re-
sults apply only to the most heavily manured areas, which represent approxi-
mately one half of a hectare on the average farm, or roughly one-fifth
of the total area cultivated.
The millet and cowpea mixture grown on the remaining four-fifths of
the land holdings of the typical farm is a different story. Table 5.1
shows that an extremely low average yield of 81 kg per hectare of millet
and 198 kg, per hectare of cowpeas was obtained in 1977. This may be
compared with mean yields of 415 kg./ha for millet and 126 kg/ha for
cowpeas obtained in 1976 on similar land by Mossi farmers in Oueguedo.
Delgado [1978, p. 189]. Both 1976 and 1977 were poor years for millet, al-
though by common accord the latter year was worse, explaining part of the
difference in results.
The single most important explanation of the relatively low yields
of millet obtained by Fulani sample members on unfertilized land appears
to be the relatively low level of labor input demonstrated in Table 3.4.
1This is because sample members use the maize ears very soon after the
harvest. Since this crop ripens in August when food is short, it was not
reasonable to ask participants to set part of the harvest aside for thresh-
ing and weighing. The rough figure of one part dry grain to five parts
total weight of the freshly harvested ear is sufficiently conservative to
avoid overestimation of yields.
2The difference between the mean yields for Mossi and Fulani farms
is statistically significant at the 1 percent level. Unfortunately, the
research team did not sample Mossi yields in 1977, due to lack of manpower.
A fully satisfactory comparison, therefore, is not possible.
This leads to the conjecture that the Fulani farming system in Oueguedo
substitutes manure for labor in grain production. Where manure applica-
tions are considerable, the results most likely surpass what would have
been obtained with very labor intensive cultivation in the absence of
manure fertilization. However, the small amount of time devoted to
millet cultivation on unfertilized land is associated with low yields and
makes this activity a seemingly dubious proposition.1
Table 5.2 contains the computation of the imputed value of crop
production on an average sample household in 1977, using the average yields
from Table 5.1. Virtually none of this output is sold, although red sorghum
is occasionally bartered on a one for one basis for millet with neighbor-
ing Mossi farmers. The exercise in Table 5.2 is valid primarily for mak-
ing broad comparisons between the value of Fulani crop output and that of
the Mossi farmers' output observed in Oueguedo during 1976. It is also
useful for gauging the proportion of Fulani output attributable to crops
as opposed to livestock, ignoring the returns to manure in crop production.
The linear programming exercise in the next chapter will deal with
this issue further.
2The evaluating prices are the 1976 harvest prices increased 10 percent
for inflation. This provides an easy yardstick for comparison of 1977 crop
to livestock output, even if the Fulani crop figure does not necessarily
represent the exact sum a herdsman might have obtained if he had sold an
entire crop output in 1977. The 1977 harvest prices were not available at
the time of writing. The 10 percent increase in prices is designed to
give a better comparison with 1977 livestock revenue. The value of Fulani
crop output is simply downgraded by 10 percent for comparison with 1976
Mossi crop output.
COMPUTATION OF THE 1977 RETURNS
TO MAJOR FULANI CROP ENTERPRISES
(a) (b) (c) (d) (e)
1977 Approximate Number of Value of pro-
Average 1977 value hectares culti- duction by the
Individual yield Average price of one vated by the aver- average house-
Enterprise crops in.kg./ha. in CFA/kg. hectare in CFA age household hold in CFA
manured land Red sorghum 2,343 22 51,546 .34 17,526
Maize and Maize 1,546 35 61,610 .16 9,858
heavily Cotton 250 30
Average value of household crop production = 40,579 CFA
1976 harvest prices from Delgado , increased by 10%
= the sum of (a x b) within crop category
Table 4.1 and notes
= (c x d)
The value of 1977 crop output on the average Fulani sample household
is one third that of the 1976 Mossi sample farm, when both outputs are
evaluated at 1976 prices using the same methodology.1 This gives a more
accurate appraisal of the relative importance to the two groups of crop-
raising than simply comparing the average household areas cultivated by
each one. Mossi households grow a substantial proportion of high-value
and high-labor input vegetables and root crops, unlike the Fulani. The
next two sections attempt to quantify the value of 1977 Fulani household
livestock production. This is measured in the form of herd growth, cash
income from livestock, and the imputed value of by-products (such as
milk) consumed on the farm.
The Change in the Value of Household Stock
Attributable to Net Herd Growth
The net change in average household herd sizes between 1977 and 1978
pertains to two main issues. First, data on the net growth of farm live-
stock holdings shed light on the relative importance herdsmen attach
to small ruminants and cattle. Specifically, this information indicates
if herdsmen are using sheep and goats to build up cattle herds ravaged
by recent droughts. Second, the data on household income from livestock
cannot be adequately interpreted without knowing how much off-take strat-
egies increase or deplete capital resources, specifically herd sizes.
Fourteen Fulani sample households were surveyed in December, 1977,
as to the net changes since the previous cold season in the number of
A discussion of the methodology and estimates for Mossi farms in
1976 is given in Delgado [1978, p. 218].
cattle, sheep, and goats kept by the household. Table 5.3 presents the
data averaged over the fourteen households. Household cattle herds grew
at an average of 4 percent per annum, but holdings of small ruminants de-
creased by 56 percent. With a roughly seven to one value advantage of
cattle over sheep and goats, this represents a net increase in the value
of household stock of 14,500 CFA, or 1.3 percent of the value of average
Given the small number of observations and the high degree of variation
represented by the ratio of the means to their standard deviations in
Table 5.3, caution should be used in the interpretation of these results.
This is particularly true in that this result pertains to the overall herd,
and not just the Fulani-owned section. When relating herd growth to the
revenue from animal off-take, the question arises as to what portion of the
growth of the cattle herds is in fact attributable to Fulani-owned animals.
In the absence of better data, the assumption here is that the Fulani-owned
section grew at the same rate as the overall household herd, which also
includes entrusted animals. Since small ruminants are all owned by the
Fulani, the effect of this assumption is to diminish the appreciation of
In the interests of preserving a working relationship with the sample,
no effort was made to establish changes in the Fulani-owned, as opposed
to the (overall) Fulani-managed household herds. Counting noses presents
much less problem than identifying owners. The small ruminants were all
owned by the household who kept them.
2Based on an average flock of sixteen animals and a net decrease in
numbers of nine small ruminants.
Based on the figure for average holdings of forty-three cattle and
sixteen small ruminants and valuation prices of 25,000 CFA per head of
cattle and 3,500 CFA per small ruminants.
HERD GROWTH FROM 1976 to 1977
Cattle Sheep Goats
Net herd change averaged
over 14 households
(animals): 1.85 -6.71 -2.36
Standard deviation of
the mean over 14
households: 10.13 9.31 5.36
Max of the households: 23 2 10
Min of the households: -10 -30 -8
Net annual change in
the average household
herd size: +4% -69% -37%
Value of addition
household herd : 46,250 -23,485 -8,260
Portion of value added
attributable to Fulani-
owned animals 18,500 -23,485 -8,260
aNew additions valued at 25,000 CFA per head of cattle and 3,500
CFA per small ruminant. These figures correspond to the average 1976 sale
prices of three-year-old animals. There is no way of assessing the age of
the new acquisition for the purposes here. The prices used, however,
have the advantage of preserving a value ratio of approximately one to
seven between cattle and small ruminants. The figures are assumed to
include the value added by the general aging of the herd as well as the
increase in numbers due to new entrants.
Based on the finding that 40% of the household cattle are actually
owned by the Fulani. The small ruminants are all owned by the household
that keeps them. In practice, there was no easy way to discover how much
of the increases were in Fulani-owned stock.
Fulani-owned cattle herds relative to the disinvestment represented by
the decreases in the flocks of sheep and goats. In Table 5.3, this
translates into the finding in the last line of the table that the net
average change in the value of the animals owned by a sample household
actually decreased by 13,250 CFA, or a decline of 2.8 percent.
The Timing of Herd Growth and Net Household
Revenue from Cattle Sales and Purchases
Household cattle herds grow larger through animal entrances which
are attributable to births, gifts, entrustments, or purchases. Stock
holdings diminish with deaths, losses, or sales. Figure 5.1 shows the net
difference between entrances and exits for the average household herd,
including entrusted cattle. Figure 5.2, on the other hand, portrays the
value of sales and purchases of Fulani-owned cattle specifically.
Figure 5.1 shows that household herds grew significantly in June.
The large number of entrances at that time are primarily due to births,
although these figures include a number of net purchases of cattle by Mossi
farmers who have entrusted the animals to herdsmen in the sample. Figure
5.2 shows that the Fulani participants purchased virtually no animals of
their own in June. The excess of entrances over exits decreases monthly
until October. During October and November, the number of animals kept by
the household actually diminishes. This is primarily attributable to
profit-taking by peasant proprietors, who take advantage of the yearly
peak in per animal prices that prevail at this time. A glance at Figure
5.2, on the other hand, shows that this is precisely the period when the
Fulani are net buyers of cattle.
SFIGURE 5.1 Average Monthly Entrances and
\ Exits from Household Herds
Jul Aug Sept
\ \ _------)--------
AVERAGE VALUE: OF SALES AND PURCHASES OF FULANI OWNED CATTLE
FOR SAMPLE HOUSEHOLDS (OVER 19 HOUSEHOLDS)
(Sales of Cattle)
/ (Purchases of Cattle)
/ (Sales Minus
July August September October
Subsidiary information provided by sample members indicates that the
Fulani prefer to sell off mature male and cull female stock in August.
The funds thus generated serve to finance purchases of food grains and
very young stock, at the end of the rainy season.
In sum, there seem to be significantly different commercial approaches
to livestock between Fulani herdsmen when dealing with their own animals
and Mossi cattle owners. The peasant proprietors appear to be slightly
more willing to take risks by purchasing cattle cheaply at the end of the
dry season, when the risk of mortality is high. They also take advantage
of high prices during the fall. The Fulani, on the other hand, prefer not to
risk purchasing animals at the end of the dry season. The available data
are inconclusive with regard to the season preferred by the Fulani for
selling, although some increase in activity occurs in August. Further-
more, Fulani sample members clearly prefer to restock their herds in the
September to November period. The young stock can then be grazed by the
owner for at least six months before facing the hot dry season.
The excess of the value of sales of Fulani-owned cattle over purchases
between June and December 1977, averaged over sample households, is 30,030
CFA. The period in question includes both the restocking of herds with
immatures and sales of older animals in August and December. When extra-
polated on a twelve-month basis, this yields an estimated annual net revenue
from sales and purchases of cattle amounting to 51,480 CFA per household.
The suitability of this figure may be ascertained by comparing it to the
sum of 55,250 CFA obtained by multiplying an average herd size of 17
animals by a 13 percent off-take rate and a valuation of 25,000 CFA per head.
Other Revenue from Livestock Operations
Fulani households in Oueguedo also obtain revenue from four other
main areas of livestock activity, both in the form of cash and subsistence
production. These are: the consumption and sale of milk, the sale of small
ruminants, the consumption and sale of eggs and poultry, and miscellaneous
payments for herding services or the provision of cattle manure. Sub-
sistence consumption of beef and mutton is rare, but does occur during high
feast days. For all practical purposes, this consumption is arbitrarily
estimated at two small ruminants per annum, corresponding to the Muslim
feasts of Ramadan and Tabaski.
Dairy Products Milk sales depend upon the level of production and
the number of subsistence consumers within the household. The relationship
observed indicates that for an average household with six to seven active
members, the first seventy to eighty liters are consumed in the form of
fresh milk. An average of just over one half of additional production be-
yond eighty liters is sold, usually in the form of a yogurt-like curdled
milk which keeps well in the heat.
Van Raay [1975, p.49] indicates that the average daily surplus milk
per nursing cow in Zaria, Nigeria,is .77 liters. The Fulani herds sur-
veyed by Van Raay closely resemble sample cattle in Tenkodogo, which is
at approximately the same latitude as Zaria. Van Raay's data (Ibid.) show
that the peak level of milk production is just over one liter per day in
August, per lactating cow. On a relative scale, this is disappointing
level of performance. However, somewhat inconclusive evidence for the
Tenkodogo herds show that Van Raay's estimates, if anything, overstate
the situation in Upper Volta. No surplus milk is available in Tenkodogo
from January until May. Furthermore, the peak level of production per
lactating cow in August appears to be closer to .6 liters.
The average household production of surplus milk and sale of milk are
charted in Figure 5.3. Butter sales are relatively minor, and are there-
fore not shown. Based on these figures and a rather conservative price
estimate of 25 CFA/liter, the imputed annual value of subsistence milk
production for the average household is 21,500 CFA. Milk sales for 1977
brought an estimated 6,665 CFA into the household.1
Sales of Small Ruminants Sales of sheep and goats undoubtedly
constitute a vital part of the sedentary Fulani production system. They
represent a faster maturing and more "divisible" store of wealth than
cattle. Van Raay [1975, p.97] evaluates one small ruminant as being equiva-
lent to one seventh of a steer when computing forage requirements. It
is interesting that this is approximately the ratio between the average
sale prices of three-year-old animals of each species. Sheep and goats
are used as gifts to prospective fathers-in-law, as ceremonial table fare,
and for commercial sale. Goat milk is not used for human consumption.
It is unfortunate that adequate data were not gathered to document
the value of sales and purchase of small stock by Fulani sample members.
1The actual price of milk was 50 CFA per calabash of "approximately
one liter" in June, falling to 25 CFA in rural Oueguedo in August. Milk
prices in Ouagadougou are substantially higher.
2This is partly attributable to sample member reluctance to discuss
the amount of "small change" they might have on their person, and partly
due to inadequate encouragement of the enumerators to pursue the point.