and Training Support Project
Institute d'Etudes et de Recherches Agricoles (IN.E.R.A)
in collaboration with
A.I.D. Contract No. 624-0270-C-00-0012-00
The editors gratefully acknowledge the organizational skills of Mrs. Katy G. Ibrahim and Ms.
Wendy Shorter who were responsible for the preparation of this document.
Appreciation is also extended to the skilled French and English translators, Jean-Marc Boffa,
Sylvie Kauffmann, Jennifer Jakovljevic, and Christian Gautier for their contributions.
This is a technical report of research in progress at theInstitut d'Etudes et de Recherches
Agricoles (IN.E.R.A). Although the ARTS project has ended, agricultural research in Burkina
Faso is just beginning. The work done during the ARTS project will be a starting point, not
an end in itself. Results will be further tested and conclusions will continue to be refined.
In this context, the usual summary of project results written by the campus staff and technical
assistance team would be inappropriate. Instead this technical report is composed of "working
papers" co-authored by the Burkinabd and American scientists. These are working papers in
the sense that they are not final products, but rather documents written to help researchers
organize their materials and communicate preliminary results to those colleagues whose
comments and criticism are essential elements in refining and polishing scientific
This document is by nature a partial record of the ARTS project accomplishments. Some
activities remain to be completed. Others have yet to be analyzed and written up.
French and English versions of the technical report are being published to facilitate
communication. The original language of a working paper is noted in the table of contents.
Because of timing and logistical problems authors have not had the chance to proofread the
translations of their papers. The version in the original language should be considered the
definitive version, if the word "definitive" can be applied to a working paper.
As with all working papers, comments and criticism from readers is welcome. Comments
should be addressed to the authors of individual working papers.
A complete summary of project objectives and activities can be found in the administrative
report entitled "Integrated Research in Agricultural Production and Natural Resource
Management: Agricultural Research and Training Support (ARTS) Project, Burkina Faso,
1990-1994, Administrative Report."
Table of Contents
Section I Soil and Water Conservation
Principal Production Constraints and Findings of a Selection of
RSP Production Research, 1990-1994 ......................... .. ............. 1
J. Dickey, E. Sankara, A. Sohoro, and S.J.-B. Taonda***
Survey: Farmer Evaluation of Zai in Donsin, 1993 .. .. .. ............ ... ....... .... 17
E. Robins and M.-C. Sorgho*
Economic Analysis of Compost Production in Southwestern Burkina Faso. .............. 29
S. Amadou, M. Bertelsen, and S. Ouedraogo**
Use of Unconventional Products in Animal Health Case: The case of
Draft Animals in Burkina Faso.......... ....... ......... .. .. ... ... ..... 37
Evolution of Sedimentation, Surface Micro-Morphology, and Millet Production in
Response to Soil Conservation Practices on an Eroded Site at Yilou, Burkina Faso ........ 43
N.F. Kambou, S.J.-B. Taonda, R. Zougmor6, D. Kabor6, and J. Dickey*
Characterization of the Soil-Plant System in Bush Fields in the Tropical North
SudanianRegion ofBurkina Faso ......................... ................ 55
S.J.-B. Taonda, J. Dickey, P. S6dogo, and K. Sanon**
Economics of Rock Bunds, Mulching and Zai in the Northern Central Plateau of
Burkina Faso: A Preliminary Perspective ........... ........ .......... ...... .. 67
D. Kabor6, F. Kambou, J. Dickey, and J. Lowenberg-DeBoer***
The Value of Research on Indigenous Knowledge: Preliminary Evidence from
the Case of Zai in Burkina Faso.............. ...... .......................83
M. Bertelsen and S. Ou6draogo**
The Economics of Rock Bund Construction on Sorghum and Millet Fields in
Burkina Faso ................ ........ .......................... .. ... ...... 91
D. Kabor6, M. Bertelsen, and J. Lowenberg-DeBoer***
Section II Cropping Systems
Farmers' Evaluations of On-Farm Tests and the State of Village Natural Resources
(Results of the Farmer Opinion Surveys 1990-1994) ............... ........... .. 05
Women's Agricultural Strategies in the Central Plateau, Burkina Faso .................. 121
M.-C. Sorgho and E. Robins***
Performance of Several Sorghum Lines Under Striga hermonthica Infestation:
Preliminary Results from Kawara, Burkina Faso ..................... ....... 131
E. Sankara, J. Dickey, L. Butler, and G. Ejeta**
Analysis of the Production System and Farming of Rice By Women in the
South West of Burkina Faso: The Case of Kawara Women in the Como6 ........... 135
On-Farm Performance of Forage/Grain Varieties of Sorghum and Cowpea:
Agronomic Evaluation by Researchers and Farmers. .. ........... ......... .. 145
A. Sohoro, S.J.-B. Taonda, J. Dickey, and E. Robins*
Sorghum Variety-Fertilizer Interactions in the Village of Kamsi ...................... 155
A. Sohoro, S. Ou6draogo, and J. Dickey***
The Riskiness of Alternative Phosphate Sources in Burkina Faso. .................. 165
V. Hien, D. Kabor6, S. Youl, and J. Lowenberg-DeBoer***
Upland Rice: An Alternative For Cash Crop Diversification to Stabilize Farm
Incomes in Western Burkina Faso................. ......................175
E. Sankara, A. Sidib6, and J. Dickey***
Dynamic Recommendation Domains in Burkina Faso: A Geographic Information
System as a Tool for Facilitating the FSR-Extension Connection.................. 183
M. Bertelsen, S. Ouedraogo, J. Dickey, S.J.-B. Taonda, E. Robins, and D. Kabor6***
Section III Integration of Livestock and Crop Production
Supplementary Feeding of Milking Cows with Cottonseed Meal for Improved
Dry Season Productivity in Western Burkina Faso ................... ..... ..... 191
A. Lalba and J. Dickey*
Use of Unconventional Products in Animal Health Care: The Case of Draft
Animals in Burkina Faso ................. ..................... ....... ...... 201
Dynamics and Management of Animal Traction in Western Burkina Faso:
A 1990 Diagnostic Study...................... .......... .... ....... .207
Section IV Agroforestry
Researching Tree Management Strategies in Thiougou Village, Central Plateau,
Burkina Faso.... ......... ........ ............................. 225
The Economics of the West African Parklands Agroforestry System:
Preliminary Evidence from Two Central Plateau Villages in Burkina Faso ........... 235
M. Bertelsen and D. Kabor6**
Economic Analysis of Some Activities of Women Linked to the Use of
Non-Woody Products of Local Forest in the South West of Burkina.. ..... ...... ... 251
Establishment and Management of karite (Vitellaria paradoxa) Parklands in
Sudanian Burkina Faso .. ......... ....... .... ...... .................. 259
J-M. Boffa, L. Lompo and D.M. Knudson***
Section V Research Extension Issues
Farmer Participation in a New FSR Program in Burkina Faso, West Africa ............ 281
E. Robins, W. Fiebig, and S.J.B. Taonda**
Extending Agricultural Innovations in the Multi-Ethnic and Multi-Cultural
Societies of Western Burkina Faso ........................................... 297
D. Ilboudo and E. Robins***
Collaborative Agricultural Systems Research in Burkina Faso 1992.............. ...313
E.Robins, P. Sanou, S. Ou6draogo, and D. Ilboudo***
Definition of New Intervention Zones of the R.S.P. Program with
Geographical Information Systems ....... ...................... .... .... .. ... .317
B. Djaby, M. Bertelsen, and S. Ou6draogo
Section VI Agricultural Policy and Land Tenure
Land Tenure and Farm Productivity in Western Burkina Faso............... ......... 325
The Challenge to Create a Durable Agriculture: The Experience of Burkina Faso ........ 333
Section VII Baseline Socio-Economic Studies
Farmer Participation in Generating Village Socio-Economic Profiles ................. 341
E. Robins, M. Bertelsen, and D. Kabor6**
The Settling of Western Burkina: Future Trends for Village Societies ................. 353
Women's Activities in the Village Research Sites of the RSP Western Zone ............ 371
D. Ilboudo and P. Lingani*
List of Abbreviations and More Words ............................... 393
Original Document in French
** Original Document in English
*** Supplied by authors) in English and French
SOIL AND WATER CONSERVATION' ,:
Principal Production Constraints and Findings of a Selection
of RSP Production Research, 1990-1994
J. Dickey, E. Sankara, A. Sohoro, and S.J.B. Taonda
The approach of RSP during the period of the ARTS Project was similar to RSP's continuing
methodology which consists of the following general steps:
* Diagnostic study (problem definition, definition of important demographic economic,
natural, social parameters)
* Identification of potential solutions, including preliminary discussions with farmers, as well
as appropriate development and research personnel
* On-farm technology testing to provide a common base of experience with technology among
farmers and researchers of various disciplines
* Field observation (observations of production, natural resource evolution, farmer opinion,
farm expenditures and income), laboratory analyses (statistical, chemical, physical, etc.), and
evaluation of results. Farmers, and ideally collaborators in development and research, take
part in field evaluation
* Restitution of results to collaborators, especially to farmers, and refinement of results by
collaborators, especially farmers.
.This process is necessarily iterative, but to the extent possible, each cycle should produce
something usable by farmers, development, and/or research. The formats for communication
with these groups are informal and formal. Formal means include field days, meetings, reports,
technical bulletins, and scholarly presentations, articles, theses, and dissertations. Informal
means are visits among small groups of individuals.
The activities of the ARTS Project were limited mostly to the Central and Western Zones of
RSP, which contain 8 village sites (Figure 1). These sites encompassed a wide variety of
socioeconomic and biophysical production conditions, some of which are summarized in
Sites and ___
..........., ... .. ... .
:; : .. .. ..... O -(=-. .I-i =,--.. ., .
,as o : ::' ". .: .
....I. .. . .n .... n
: :: : ::: i::: : .. .. .
: .. .. ..:.... .. .E, .:l ::.r e :: ::: .
.. ... .. .. :. .. **.. ..** ;..o *
.cmft+ .i- **
:, ,: :., ,.L,, ., ,- .*
: .. :, ,
Table 1. RSP Village Sites, Central and Western Zones
Zone Village/ Animal Principal Crops Level ofNR Dominant
Annual Traction Degradation Migratory
Central Donsin Absent Sorghum, millet Extreme Old and recent
Central Kamsi Little Sorghum, millet Advanced Recent
Central Thiougou Dominant Sorghum, millet Moderate to Recent
723 advanced emigration
Western Yasso Dominant Sorghum, maize, Moderate to Old and recent
802 cotton advanced immigration
Western Kayao Dominant Sorghum, maize, Moderate to Recent
819 cotton advanced immigration
Central Tiano Dominant Sorghum, maize Moderate Recent
Western Kawara Little, but Maize, sorghum, Slight Old immigration
1010 increasing rice
Western Dimolo Little, but Maize, sorghum, Slight Slight, recent
1015 increasing yam immigration
aPost-1969. For West African sites, rainfall means including post-1969 data are considered a better indicator of
current rainfall expectations than means including pre-1970 data. This is because a relatively rainy period of what
appears to be a long-term climate cycle came to an end in about 1969. Post-1969 means for RSP villages are from
150 to 200 mm less than means for pre-1970 data.
A number of the papers contained in this collection are the results of work on individual
research themes that made up a coordinated, evolving research program. Each work had a
place in the above research process, and many were interrelated. This paper is not an
exhaustive listing of all research activities. Rather, it ties a selection of activities to the
principal constraints they were meant to address. The objective is to share some of the overall
logic of the production research activities, and thereby to provide a general context for the
various individual studies reported in other papers.
Major constraints to sustainable, secure, and productive management of natural, other capital,
and labor resources in Burkina Faso can be divided into primary (fundamental), secondary
(derived from fundamental constraints), and tertiary (derived from secondary constraints) groups.
1. Shortage of uncultivated land for clearing and cultivation, pasturage, or gathering (of wood,
pharmacopoeia, mulching materials, or other forest products). This is generally due to
demographic pressure on the land resource.
2. Erratic rainfall, especially early- and late-season drought periods, as well as occasional and
very damaging mid-season droughts.
3. Pest and disease pressure on production (e.g. Striga spp.,maize streak virus, downy mildew,
numerous cowpea insect pests).
4. Insufficient access to markets (e.g. maize sold in Yasso at 2500 CFA/100 kg, January, 1994)
and erratic, often low pricing (e.g. the 1993 cotton crisis).
5. Storage pests and rots, especially for their effect on seed and setts.
Traditional, low-input, rotational bush fallow farming systems require large land areas to
function productively, since a large percentage of the land must be in fallow at any one time.
Demographic pressure is such that this is no longer possible in many areas. Rainfall quantity
and distribution are rarely ideal and often result in significant stress on plants and animals in
the production system. Pests and diseases, especially when crop vigor is reduced by other
stresses, can limit production and destroy planting material. These are cited as primary
constraints because they exist in the presence or absence of other constraints, although they do
give rise to other (secondary) constraints.
6. Insufficient village political organization to solve resource-use problems (e.g. insecurity of
usufruct landholders, chronic cultivator-pastoralist conflicts).
7. The degraded state of land (including soil and vegetation) resources. For the soil this
includes erosion, crusting, reduced water holding capacity and organic matter contents, as
well as chemical constraints to fertility (largely shortages of N and P) and localized acidity.
Wild plant and weed communities have become less diverse, more dominated by plants
selectively preserved by farmers (e.g. Hibiscus sabdarifa or da, Butyrospermum paradoxum
or karit6), relatively unpalatable forage and browse, and weed populations capable of
surviving continuous cropping and frequent cultivation (e.g. Digitaria horizontalis, Striga
Primary constraints on land require that this resource be managed skillfully by village
communities. Traditional land tenure systems provide the community with a rich assortment of
tools. However, these systems are often understandably in the process of adapting to changing
demographic conditions (population density, ethnic makeup, and families' economic
objectives), and a deteriorating resource base. During this period of adaptation, when
production systems must be transformed in profitable steps, substantial physical and biological
degradation of the land base can occur. Organization for land management (constraint 6) and
land degradation (constraint 7), therefore, are secondary constraints by virtue of their
dependence on constraints (1) and (2).
8. Shortage of biomass (animal wastes and vegetative matter) with which to nourish people,
crops, livestock and wildlife.
9. Scarcity of labor and animal traction in certain regions, making cultivation and materials
transport (of organic fertilizer, mulch, and stones, for example) for land management very
10. Ignorance of resource management possibilities and probable results.
11. Access to capital (partially resolved for cotton producers, who use cotton credit to purchase
fertilizer used on a number of other, notably grain crops).
12. Poor animal health due to insufficient nutrition and veterinary care.
Degradation of the land resource and failure to address this degradation by changing village
land management (secondary constraints 6 and 7) requires that farmers seeking sustained
production invest labor and capital in land maintenance (fertility, soil and water conservation),
pest control, and animal feeding (on distant pastures or cut forage), however, these efforts are
often insufficient. These unmet needs give rise to the tertiary constraints listed above.
II. RESEARCH ACTIVITIES
Thematically, the activities can be broken down into the following categories:
Crops and cropping systems
1. Varietal evaluation, including analysis of yield stability
2. Disease and pest management
3. Evaluation of agricultural machinery and its use
4. Training in methods of seed/sett production and storage
5. Evaluation of crops or cropping systems
6. Improved feed production, livestock feeding, and animal health
Environmental monitoring, conservation, and regeneration
7. Soil fertility, including use of mineral fertilizers, fabrication and use of organic fertilizers,
and definition of fertility requirements of varieties
8. Land conservation and reclamation
9. Characterization of land degradation and variability
10. Environmental monitoring.
Findings: Crops and Cropping Systems. Farmers' needs and measures they are willing and
able to take to resolve them vary greatly across Burkina Faso. While much of this variability is
due to the nature of the individual farmer, some community characteristics help one to
understand and organize thinking about farmers and their behavior. The environmental
conditions and the importance of cotton in the production unit strongly influence the way
people farm in Burkina., This, in turn, has a large impact on the relevance of a given
technology to farmers. The influence of these factors on farmers' use of inputs and animal
traction for land preparation and weeding are described below. Clearly, cropping environments
that are untilled, unfertilized and weedy will welcome different kinds of innovations than fields
that are tilled, fertilized, and weeded on time. This makes it critical to understand these
patterns of land management before discussing research results for specific technologies.
Prolonged and substantial public investment in cotton production (extension of information on
improved practices, credit for equipment purchase and production costs, input availability, and
organized marketing) has borne fruit. Farming practices in cotton production areas are in
general strikingly different from practices outside these areas.
There are areas with favorable environmental conditions for crop production, which for various
reasons do not currently produce much cotton. Naturally, these areas benefit much less from
investments made to encourage cotton production.
Farmers in the Central Zone, generally not cotton producers, lack access to credit and run a
greater risk of crop failure. This is especially so in the dryer northern areas, where rainfall is
scarce and relatively erratic in distribution. Farmers therefore favor a strategy that minimizes
capital inputs, thus limiting the capital put at risk. Five examples along this transect, taken in
RSP villages, follow. Refer to Table 1 for additional information about these villages.
* At Kayao and Yasso a high proportion of farmers have and use animal traction, (for land
preparation, weeding, and transportation), cotton is widely produced, and chemical fertilizers
are habitually applied to field crops (not only to cotton, but also to cereals, especially maize).
The combination these practices with higher, less variable rainfall than in much of the rest of
Burkina Faso provides field conditions with relatively high yield potential and low risk of
At Tiano there is relatively little cotton, yet farmers have a rather entrepreneurial approach
to farming, (many are recent migrants). Many farmers have animal traction, and are prepared
to plow and apply chemical fertilizer to achieve good maize yields on large areas.
At Kawara and Dimolo, little cotton is produced, and the use of animal traction,
mechanization, and inputs is much less prominent than at Kayao and Yasso. Dimolo youths
are frequently absent, gone to C6te d'Ivoire to find wage labor jobs. Farming of rice on
hydromorphic soils is a major activity at Kawara.
At Thiougou, many farmers have animal traction and plow, and organic fertilizer is produced
(some with Burkina phosphate rock phosphate added) and used intensively in village
fields. Practically no fertilizer of any kind is used in bush fields. Maize in large fields is
considered too risky. Farmers are sometimes willing to purchase pesticide for cowpea
At Donsin, animal traction is very rare, farmers recently began very labor-intensive use of
manure and compost in zai in 1993, and practically no mineral fertilizer or pesticide is
purchased. Cash is scarce and needed to pay for food during the hungry season (July,
August). Farmers are clear in their priorities; many will not spend money on pesticide or
SKamsi is somewhat particular due to the large amount of dependency on income from
immigrants, which changes farmer motivation.
The use of inputs by farmers in the cotton-production region means that new varieties find
themselves in much the "improved" (relatively low-stress) conditions that the breeders who
developed them anticipated. As a result, varietal introduction is relatively easy and effective.
That is, a new variety has a reasonable chance to out-yield the local variety in a farmer's field,
so if its culinary quality is satisfactory, it runs a good chance of some adoption and of making a
positive impact on production. Testing is straightforward: put the variety in the field under
conditions stipulated in the breeder's technical bulletin, because the breeder knows what the
variety needs to perform, and the farmer frequently can approximate these conditions.
On the other end of the spectrum, the strategy of farmers at Donsin (on all fields) and at
Thiougou (in the bush), however, is in direct conflict with the dominant milieu of varietal
selection and with management stipulations in technical bulletins associated with improved
varieties. These bulletins are conceived to allow these varieties, given adequate rainfall, to
yield well, (or to realize their potential). Low to moderate doses of chemical or organic
fertilizer are generally recommended, and in the case of cowpea, so is treatment with
insecticide. When planning varietal tests in cultural environments that do not match the
bulletins' recommendations, the question becomes who to please, the breeder or the farmer?
The research approach in the Central Zone was therefore two-tiered, with new varieties
introduced under at least two fertility regimes, to get an idea of how varieties would do under
farmers' and "improved farmer-managed" conditions, and therefore of their stability across the
real spectrum of fertility conditions. The results for sorghum, millet, and cowpea were
surprising: local varieties frequently out yielded improved varieties at both ends of the fertility
spectrum, exceptions being the millet IRAT P8 (virtually identical to the local), millet IKMP-3
(which yielded slightly better than the local without fertilizer), and the cowpea local Donsin
(which did not respond to fertility as well as improved varieties).
Improved peanuts yielded well and were similar to locals across the fertility spectrum, except at
Thiougou, where the local was slightly superior.
Several good, disease resistant maize varieties had been produced by INERA and were in
theory already passed to extension, however these varieties were relatively unknown in all RSP
villages. Introduction of these maize varieties was of interest to RSP for the following reasons:
* It benefited cooperating farmers.
* It responded to a need for new alternatives in the face of difficulties in the cotton sector.
* It provided a common base of experience in food crop production for farmers and the RSP
teams. (All aspects of maize production and the adoption of new technology were studied by
the various disciplines, not just yield performance.)
* Productive working relationships and dialogue between cooperating farmers were developed
around this and other production activities.
At Tiano, maize varieties SR22 and DMR.W each produced about 3 Mg/ha (at recommended
fertilization rates), out yielding local maize by about 20%. The variety SR22 was so productive
in Kayao, Yasso, and Kawara that it was rapidly adopted. The early maize varieties KPB and
KPJ were tested and adopted by some, especially in Dimolo, as crops to provide some early
relief from the hungry season. Tested hybrid maize did not perform well, and in retrospect
probably had little place in on-farm tests.
Another promising innovation is the introduction of upland rice cultivation, based on the
availability of early varieties adapted to upland conditions, (FKR-5 and FKR-33). Yields
averaged over 2 Mg paddy rice/ha, and maximum yields were upwards of 5.5 Mg/ha. Yield
stability of FKR-33 was superior, while given adequate rainfall, FKR-5 had higher yield
potential. In general, the crop is sensitive to dry spells of over 10 days. The preceding crop is
important, with preference in the order rice < upland cereals < cotton < garden crops. Hand
planting and weeding consume an enormous amount of labor and are the major constraints to
increased area in upland rice. Mechanized and chemical weeding were examined in 1993, each
showing some promise. The INERA Water, Soil, Irrigation, and Agricultural Mechanization
program (ESFIMA) worked on tools for mechanized planting at close spacing during 1993, and
RSP will test these tools in 1994. RSP has done some work on hand-planting rice at wider
spacing, also with some success. Burkina imports a great deal of rice. If prices stay
sufficiently high, there is tremendous productive potential for rice in Burkina Faso.
Untimely weeding is a major upland cereal production constraint in the Western Zone. As for
rice, herbicides are on the market and used by some farmers. Without prejudging the
environmental or production advantages of chemical weed control, tests in maize (on-station
and on-farm) began in 1993. Controlwas frequently achieved, but the economic benefit is less
clear than for rice.
Sorghum, on the other hand, suffers greatly from Striga infestation, especially on continuously
cultivated land. Striga is cited by farmers in Yasso and Kawara among their principal
production constraints, and as a principal indicator of land degradation. Where land cannot be
left fallow, rotation, resistant varieties, and chemical control are options. Six resistant lines
were tested in an infested sorghum field at Yasso in 1993. All lines showed significant
resistance relative to the local check. Taken together, the six lines had 13% of the Striga
infestation level of the local check, and produced 242% of its grain yield. Spraying with 2,4-D
was also effective, and relatively inexpensive due to the low cost of this herbicide.
Mildew attacks millet in much of Burkina Faso, with the yield reduction depending on severity.
The seed treatment Apron Plus, (100 g metalaxyl, 60 g carboxine, and 340 g furathiocarbe, all
per kg of powder), provides protection from mildew at about 2000 CFA/ha, equivalent to about
25 kg millet/ha. RSP has participated in widespread testing of the product in farmers' fields
and developed a technical bulletin for the use of Apron Plus aa well as another bulletin to help
estimate the potential economic benefit of using Apron Plus in a given field. Impact appears to
depend on the region, particularly on the native level of mildew infestation.
As for use of insecticide on cowpeas, the approach up to 1994 was to treat as needed, and to
experiment with cereal intercropping as a means of reducing the absolute need for insecticides.
Given the response of many farmers (indicating unwillingness to purchase insecticide), a test
oriented toward reducing the number of insecticide treatments is planned for 1994. The test
will include some of the more resistant cowpea materials that are currently available. Among
monitored farmers, few at Donsin treat their cowpeas, yet they manage to get acceptable
harvests, even with improved materials, which they generally consider more susceptible to
insect attack. This has not been the case at Thiougou, where insect attack is much more severe.
Farmer interest in cowpea as a source of food, cash, and fodder (roughly in that order) is strong.
Production quality and quantity are very important issues for grain/forage varieties of cowpea
(7/180-4-5) and sorghum (ICSV-1049). In Donsin, where grain production for human
consumption is the overwhelming preoccupation, the lower yields (relative to KVX395-4-4 and
local), the recommended forage harvest date (after one pod picking), and propensity for
shattering limited interest in 7/180-4-5. On the other hand, Thiougou farmers, while noting that
the grain production is somewhat low, appreciated the forage quality, and seemed willing to
sacrifice second and third pod pickings to get quality forage. The need to nourish draft animals
motivated farmers most strongly to produce forage.
The caudatum heritage of the sorghum ICSV-1049 renders it particularly sensitive to head rots,
and makes the grain difficult to process into flour. It also appears to be sensitive to late-season
drought. At least in the south (Thiougou), delayed planting (as recommended) can largely
resolve the problem of head rot, if the end of the rains is timely. As with cowpea, the feed
value of the stalk, (which is considerable animals prefer it and will destroy fields to get it),
is more appreciated in the south, where animal traction is more common and where food
production is more assured.
This leaves a large hole in what RSP has to offer to farmers in the way of cereal grains.
Through 1993, no sorghum or millet has competed favorably with local varieties for a
substantial spot in the farming system. In 1994, the sorghum varieties Sariaso-9 and Sariaso-10
will be tested in all sites. Sariaso-9 is an improved local selected at Saria (about 710 mm/year
average annual rainfall since 1970), and Sariaso-10 is another improved caudatum. Only
Sariaso-9 will be put under farmers' management. In the south, planting date trials will be run
jointly with the Sorghum, Millet and Maize Program (SOMIMA) to refine this aspect of
To further respond to severe limitations to pasturage and corresponding farmer interest in forage
production (at Yasso and Kayao), several alternatives for feeding are being explored:
* Production ofAndropogon guayanus in vegetated bands in fields
* Improvement of feed value of crop residues by grinding and mixing with urea
* Supplement with cottonseed cake, a byproduct of cotton processing
* Production of forage crops, like dolic and sudan grass (Sorgho sudanense)
Benefits from these feeding systems can include:
* Improvements in animal health and reductions in mortality
* Increased weight gains (recorded for milking cows and calves) and milk production
* Strengthening of draft animals before the cropping season
* Concentration of animal wastes and association with compost pits or bins
So far, farmer interest in these alternatives is running high. The economic benefits of
supplementary feeding seem to be high when meat prices are high (as they have been since the
CFA devaluation in January 1994), and milk prices are good (in zones like Kayao and Yasso,
where milk consumers outnumber milk producers).
A general problem for all crops is production and storage of high-quality planting materials.
There are two cases in which high levels of farmer and RSP interest have combined to help
solve this problem. When SR22 trials were at first successful in 1991, large, seed-production
fields were established during 1992, and this constituted much of the seed supply for the
widespread planting of this variety in 1993. Likewise, several kg of sudan grass was
successfully multiplied at Yasso in 1993 for more widespread production of seed and forage in
Seed production and post-harvest storage is also the focus of several other 1994 activities. The
problem of storage, especially seed storage, is most pronounced for cowpea. It is sensitive to a
number of storage pests (notably bruchids). Seed production and storage is also a hurdle when
farmers wish to retain a new variety in their cropping system: a sustainable local seed stock
must be established. For this reason, production and storage of quality seed is the major theme
of the monitoring of farmers' use of improved varieties in 1994 and in other, specific, seed-
production activities for sudan grass, cowpea, and dolic. Siting and maintenance of fields will
be monitored jointly, and training in effective storage techniques will take place at harvest time.
Likewise, training in and farmer-testing of sett multiplication for yam began this year at Tiano.
This activity will be followed up by activities in yam production and storage.
Yams for market also suffer from storage problems. Large yams are preferred for the premium
they bring in the market, but they store poorly. One of the varieties involved in the yam
activities is Florido, which is smaller in size and lends itself to planting in ridges instead of the
traditional mounds. It could provide a product to store and sell later in the year, when large
yams are off of the market and small yams bring a good price. Also, when produced in ridges,
it can produce roughly double the yield of the traditional variety in mounds. Setts multiplied
for both local and Florido during 1994 will be planted in these two systems for comparison in
1995. Marketing will also be monitored.
The above work was oriented toward understanding and improvement of crops and cropping
systems. RSP also devoted much of its research effort to the broader, natural-resources context
of crop production. Among other things, this includes land conservation and reclamation.
III. FINDINGS: ENVIRONMENTAL MONITORING, CONSERVATION, AND
The RSP Geographer produced maps of natural resource distribution and use for each village,
based on aerial photographs and field investigations. These maps and the associated knowledge
base have been invaluable tools for orienting subsequent work in sustainable resource
management and production. Many of the impressions of land use in this report are derived, at
least in part, from those mapping efforts.
One of the greatest challenges facing Burkina Faso is the identification and implementation of
sustainable land use in the most productive regions. These same regions also receive many
immigrants. As areas of bush and old fallow are grazed, logged, or cleared and farmed, their
capacity to produce fuel, feed, food, fiber, and pharmacopoeia is often degraded, sometimes
irreversibly. If this land resource is lost, this largely agricultural country risks becoming one of
the worlds' "basket cases". For this reason, several research activities have focused on various
aspects of land degradation, conservation, and reclamation.
As with other cultural practices, soil management problems and the measures farmers are
willing to take to solve them vary greatly by region. Understanding of the pattern of this
variation is critical if one is to interpret farmers' reactions. Major driving forces of this
variation are population density, rate and history of migration, and climate. A sample transect
of these factors and of land management across Burkina, stopping in RSP villages (Table 1)
* Dimolo (1015 mm rainfall/year mean from 1970 to 1993), where bush land is available
for pasture or for clearing of new fields, and the rate of immigration is relatively low. Soil-
related problems are predominantly fertility, and the main action to resolve them is to fallow
land. Animal traction exists but does not predominate, and is used mainly for plowing.
* Kayao and Tiano (819 and 848 mm rainfall/year, respectively ), where the rate of
immigration is very high, and pasture is becoming scarce with the clearing of more land for
cultivation. Immigrants and natives alike are motivated to cultivate extensively because
cultivation secures land for future use. Animal traction is predominant and is used for
plowing and weeding. For some farmers, soil fertility problems are increasingly addressed
with chemical fertilizers purchased with credit for cotton production, and compost production
* Thiougou (723 mm rainfall/year) has some neighborhoods that border the forest. Immigrants
have recently settled this land. Land management in these areas is similar to that of Tiano
and Kayao, except for the absence of cotton. Otherwise, Thiougou and Yasso (802 mm/year
rainfall for Yasso) received large, early (30-80 years ago) waves of migration. Migration has
now slowed due to crowding. In most areas of these villages, very little bush remains to be
cleared, and fallow periods on cleared land have become increasingly rare. Animal traction
is widely used for plowing and weeding, and low fertility and Striga hermonthica (a parasitic
weed symptomatic ofcontinuous cereal cultivation) have become major production
constraints. Credit for cotton production has provided a means for chemical fertilizer
purchase in Yasso, but there is increasing interest in compost production as an alternative
and/or complement. Compost production is widespread in Thiougou, but does not suffice for
fertilization of bush fields. Rock bunds were constructed on Thiougou village fields about 10
years ago. With the land shortage and a stabilizing population, farmers in both villages are
increasingly prepared to consider land conservation measures, with strong preference for
measures requiring a low capital and labor demand.
SDonsin (647 mm rainfall/year) has a relatively stable farming population and practically no
land available for clearing for new settlement. Pasturage is extremely limiting, Striga
infestation is widespread, large areas of village land have become so degraded as to be
completely denuded and abandoned, and crop production and the food supply are insecure.
Farmers are highly motivated to stem what they consider an environmental/economic crisis.
On remaining arable land, farmers increasingly use traditional soil and water conservation
measures to improve current and future production (grass mulching, leaving strips of native
vegetation on clearing, laying down of branches roughly on the contour or in gullies to
reduce water erosion). Two years ago, the non-governmental organization Foster Parents'
Plan began to support rock bund construction, (training in contour identification, free
transport of rocks from quarries to fields) and assisted natural regeneration of Piliostygma
spp. This precipitated the sudden and widespread installation of rock bunds on much of the
cultivated land. RSP initiatives in soil and water conservation have been rapidly adopted by
most of the village, as will be discussed later. Animal traction is practically absent, and
although fertility constrains production, cash scarcity and high risk of crop failure result in
very little mineral fertilizer purchase. Composting was introduced by RSP, and some farmers
have adopted it.
The character of land and weather dictates management possibilities and, to some extent, their
relative costs and benefits. African landscapes encompass extreme variability in productivity
and vegetation. High levels of short-range (over distances of several meters) soil variability, for
example, are well-documented in West Africa, and pose special management problems. Several
of RSP's research projects were aimed at land characterization. These projects included
soil/landscape/vegetation mapping, soil sampling and description on short transects to answer
specific questions, woody plant community inventory, and detailed evaluation of the process of
land degradation under cultivation. The longer-range objective of each study was to provide
basic knowledge for the development of sustainable systems of land management.
The large-scale mapping efforts will not be discussed at length. Soil or soil/landscape maps of
Kamsi and Thiougou villages were prepared, largely to meet the requirements of two
dissertation studies. A vegetation mapping effort was begun at Donsin by the agroforesters. At
each site, the soil mapping information contributed heavily to the planning of research. In
Thiougou, in particular, it permitted the researcher to control for soil type in a very exacting
study of land degradation.
In the context of another dissertation on the socioeconomic role of trees in parklands, woody
plant communities maintained by farmers in their croplands were characterized. This
demonstrated the variability of farmer woody-plant management strategies, the importance of
variables such as the presence of animal traction in farmer decision making, the predominance
of shea nut trees in cleared and uncleared land, and that the pattern of woody plant evolution
with time under cultivation is not a simple rapid reduction in woody plant numbers and species
Two small studies of soil variability were undertaken. Sampling across a transect heading up-
slope from a rock bund resulted in a detailed description of erosion, sedimentation, and
revegetation across this zone. In a second study, soil samples were taken late in the cropping
season across transects of evident soil variability. Corresponding harvest observation plots
permitted the correlation of soil and crop results in a village field, in a bush field under
prolonged, continuous cultivation, and in a newly-cleared field. Fieldwork for these two studies
was informally organized as on-the-job training in micro-profile description and in diagnostic
soil sampling in farmers' fields, respectively. Each study required about 1 day of field time.
This sort of activity could be used far more frequently to fill data gaps with at least some
provisional information, and to create educational fieldwork opportunities with various resource
people. The economists made effective use of rapid techniques to respond to specific, well-
defined questions (e.g. estimation of opportunity costs of capital, estimation of partial budgets
for soil conservation practices).
In the regions of immigration and little chemical fertilizer use, intensively managed village
fields (plowed and receiving substantial organic fertilization) produce well year after year, even
after many years of continuous cultivation. Bush fields, on the other hand, are mined of their
fertility and abandoned. Reduced land availability has resulted in reduction or elimination of
fallow periods in the rotation, so that fallow no longer achieves the needed regeneration of
fertility in bush fields. The nature of land degradation during the years after clearing was
studied in great detail. Changes in crop production as well as soil chemical and physical
characteristics were monitored and described. Nitrogen immobilization and/or alelopathy
limited sorghum yield in the first season after clearing. Thereafter, production decayed
exponentially from a maximum during the second year after clearing. The same practices that
allow sustainable production in the village (protection with rock bunds, application of 2.5 Mg
manure/ha) succeeded in bringing the oldest bush field back from 13 to 74% of the productivity
level of newly cleared land.
Several other activities were focused on joint evaluation of soil and water conservation (SWC)
practices with farmers. In Tiano and Kayao, where farmers have little time for these practices,
but nevertheless sense a need for a fertility maintenance program, composting activities have
begun. In Thiougou and Yasso, where farmers will consider some intensified efforts to
conserve land and water, soil conservation, activities are beginning. In spite of the history of
rock bunds at Thiougou in village fields, the labor and transportation constraints to their
installation makes less demanding technologies (such as vegetated strips) more attractive for the
more extensive bush. Land tenure can figure even more importantly in the choice ofan SWC
technology. In regions such as these, most farmers are on loaned land, and therefore lack the
necessary land tenure security, and often the right to install in durable SWC technologies (i.e.
rocks, trees). Furthermore, cultivation of land by an immigrant (usufructuary) can augment his
chance of borrowing that land the next year. For the autochthonous farmer, it helps to morally
justify a future refusal to loan the land. This motivates both, therefore, to cultivate a lot of land
to retain the option of future cultivation. This works against intensive investment in
maintenance of a given field. This pattern is most noticeable where the rate of immigration is
currently rapid and farmers are therefore scrambling for land.
Donsin has had little recent immigration and therefore land tenure is rather secure for most
residents, which helps explain their rapid and enthusiastic adoption of rock bunds when Foster
Parents' Plan began to support their installation. Along with the intensive traditional methods
of SWC (e.g. grass mulching) that are widely used at Donsin, this suggests a very different
situation from Yasso and Thiougou. Farmer opinion surveys and field observations indicated
that the problems of food insecurity are substantial and also motivated investment in SWC.
Difficulties during stand establishment and with mid- and late-season droughts have all plagued
Donsin during RSP's time there and before. Farmers have begun to act.
Indeed, farmer response to the introduction of zai (manure-filled planting holes), another labor-
intensive SWC technology, proves this. After a field trip of RSP researchers and Donsin farmer
representatives to a zone where zai are used, zai was tested in several farmers' fields alone and
in combination with grass mulching, and compared to grass mulching and planting bare ground.
In the year of introduction of zai at Donsin, 70% of the production units experimented with
them, experiencing variable, but generally very positive results. Fully 80% of production units
plan to plant into zai in 1994. The technology is adapted to farmer labor constraints, since most
or all of the additional demand is during the off-season, when labor is available. The ability of
zai to reduce risk of drought stress by capturing and holding water around a healthy plant
reduces farmers' risk, especially during stand establishment. Time for re-sowing is reduced,
which increases labor time available for weeding. Due to more timely establishment and
perhaps to reduced stress from drought, nutrient deficiency, and weed competition, crop yields
are doubled, or tripled when grass mulching is included.
With the rapid adoption of this technology, the major limiting factor for many farmers,
especially those with few livestock, will become organic fertilizer. While composting is being
practiced by a number of households, its potential to increase organic fertilizer volume in the
village has yet to be realized. The scarcity of feed and pasture complicate manure recovery
schemes. To increase manure recovery, some producers are beginning to stable animals during
the night and constrain daytime grazing to defined areas, from which the manure is collected.
Manure is being traded and purchased at Donsin. This ensemble of considerations could
motivate forage production by reclamation of rangeland or by forage field-crop production.
A significant proportion of Donsin's land area is highly degraded, producing practically no
vegetation (zi-pel6). Although some of these lands were once among the most fertile at Donsin,
they are generally considered by farmers to be lost for productive uses. A thematic trial by
ESFIMA and field experience of farmers and NGO's in Yatenga and Bam provinces
demonstrated that the combination of rock bunds with zai with or without mulch would reclaim
such land for cereal production. After reclamation, clearly other land uses (pasture, forestry)
could be imagined, however cereal production remains a strong motivator of hungry, rural
families to reclaim land. These results were applied to a zi-pel6 at Donsin, where mulch, zai,
and zai with mulch treatments produced 430, 1060, and 2290 kg of sorghum grain/ha,
respectively. Average grain yields in farmers' fields this year was in the neighborhood of 400
kg/ha. As with zai in farmers' fields, these methods of land reclamation are experiencing some
adoption on other zi-pel6. The progress of several farmers who have reclaimed substantial
areas for cereal production will be monitored in 1994.
Cereal production, or something with a similar payoff, may be needed to pay for the labor-
intensive practices of zai, rock bunds, and grass mulching. However, more productive pasture
may be.sufficient to motivate investment in less expensive reclamation technologies, like small
earth bunds in the form of half-moons. One farmer has planned to collaborate with RSP in the
evaluation of half-moons for pasture reclamation. The idea is that the bunds slow runoff and
favor greater infiltration on eroded, slowly-permeable, shrink-swell clay soils. The wetted soil
swells, then dries and shrinks, leaving a somewhat roughened, cracked land surface. Infiltration
from the next storm is greatly increased. The greater volume of water stored in the soil
supports morevegetation, and the pasture begins to improve.
By any measure, Donsin has demonstrated extraordinary dynamism in the rapid evolution of
natural resources management in the village. Likewise, the rapid and widespread adoption of
SR22 in maize-production zones signals that an important production constraint was identified
and addressed by RSP in the south and west. The enthusiasm for forage (dolic and sudan grass)
production and upland rice are producing very successful results. If RSP were a development
agency working in these few village sites, this would be enough to claim success.
However, RSP's job is to support development institutions so that they may increase their
impact in a much greater number of villages. Therefore, RSP must now describe what has
happened in each of these instances in such a way that development institutions can identify
similar situations and replicate these successes on a larger scale.
Survey: Farmer Evaluation of Zai'in
E. Robins and M.C. Sorgho
"Zai" is a technology indigenous to Burkina and used in Yatenga on the Central Plateau to
reclaim exhausted soils for cultivation. It involves digging a hole 20 cm wide and 15 cm deep.
This is then filled with manure and seeds are planted in it. This "zai" creates a moist
environment for the plants, which favors germination, growth and crop yield.
In the Namentenga (also in the Central Plateau), where soil degradation resembles what used to
be the case in Yatenga, "zai" has been recently introduced.
The Farming Systems Research Program (RSP) values the experience acquired by the Yatenga
farmers, and sees its pertinence for Namentenga. Therefore, to facilitate teaching Namentenga
farmers how to fight soil degradation, a "farmer to farmer" visit was organized by the RSP in
May, 1993. A group of farmers from Donsin, which is one of the RSP research sites in
Namentenga, traveled to Ouahigouya, in Yatenga, to observe water and soil conservation
Zai was one of the technologies observed during this visit. Donsin farmers were already aware
of the existence of this technology. One farmer from the neighboring village of Bonam had
tried it the previous year. The latter's highly positive experience encouraged the Donsin
farmers to follow his example and try it for themselves. After their visit to Yatenga, these
farmers therefore undertook the extension of zai technology to their own village. The number
of farmers volunteering to try zal was phenomenal more than 70% of villagers tried it in 1993.
The RSP, together with its collaborating program, ESFIMA, were part of this effort.. They
collected data from 6 farmers testing zai. Researchers also demonstrated zai on "zi-pC16" ( i.e.
exhausted; infertile land which no longer produces anything). The results of this are reported
elsewhere (the agronomic report by the RSP center team, 1993).
In order to assess the results of the zai trial in the village, the RSP chose a random sample
among the farmers surveyed. This report presents the results of this study, which aimed to
improve our knowledge of the agricultural impact of zai methods, based on field trials and
evaluations made by the farmers themselves.
A stratified and random sub-sample was taken from a sample group of farms, itself chosen
randomly for another research purpose (i.e. the socio-economic profiles of village farms). This
enabled us to see whether the variation in the farmers' experiences could be attributed, among
other things, to their socio-economic status. The sub-sample comprised 37 farmers, who were
divided into 4 categories. Another 5 farmers, none of whom had been in the original sample,
and who were therefore unclassified, also tested zai methods, following protocol previously
established by RSP. These 5 are also included in the sample for this study. The distribution of
the sample group is described in Table 1. The 42 (out of 186) farmers studied represented 23%
of the village farms.
Table 1. Farmers surveyed on the performance of Zai
Producers 1st 2nd 3rd 4th not Total
Category Category Category Category classified
group 7 7 7 16 5 42
sample 17 17 17 38 11 100
% of the
village 12 12 24 52 -- 100
A questionnaire was compiled and tested. The questions, often non directive (or "open")
focused on two main themes:
1. The evaluation of zai used in the previous trials; and
2. the intentions of the farmers concerning the reuse of zai the following year.
The replies were coded, recorded and analyzed with the help of Quattro-Pro software.
II. RESULTS: EVALUATION OF ZAI
The Trial and the Yield. 71% of the villagers tried zai. The majority of them reported a
positive impact on crop yield. This is consistent with results from agronomic tests. Refer to
Figures 1 and 2.
The yield of zai parcels (zai alone) was two thirds better than that of the control area (i.e. the
plot with neither zai' nor mulch) according to farmers' estimates. (For example: 10 baskets, as
opposed to 6 baskets for the same surface area). Some farmers reported an increase of up to
three times the yield on the zai plot.
The majority of farmers say that the greatest differences in yield (when comparing plots with
and without zai planting) occurs on the most degraded land. One farmer, on the contrary, held
the opinion that zai had more impact on moist, fertile land. Neither the position of the zai plot
within the toposequence, nor the number of years that the parcel had been established, were
considered factors determining yield.
Where the yield of a zai parcel was not so good as, or even worse than the control parcel, the
following explanations are proposed: Striga may have been present on the plot; sowing may
have been late; the site was chosen badly (on a farmer house site, or in a water-logged
depression). Another reason could have been failure to close over the zai holes adequately.
The observations of farmers using zai and mulch together are not conclusive. Some said that
the mulch helped protect the young plants during a period of drought; others saw this as
doubling the labor involved when doing both, and proposed instead to choose one or the other
method, rather than combining the two. The choice of technique also depended on the
availability of help, on manure supplies, and on the characteristics of the site. This will be
discussed again later,
Twelve farmers did not try using zai in 1993. They said that they did not have manure, or that
the family's labor was taken up mulching, building rock bunds, or clearing land, or that old age
prevented strenuous work. However, 9 of the 12 declared their intention to use za' the
following year ( see section 4).
Germination and reseeding. Germination was more successful in the zai parcels than in others:
seedlings came up at a rate of 91%, as opposed to 77% in other parcels. All parcels were
Yield of the Zai Parcel
in comparison to the control
threatened by a period of drought in the last two weeks of June. During this time, producers
observed that the need to resow was reduced on the zai parcels, where plants showed greater
resistance to drought.
For the whole sample, the zai parcels were resown 1.5 times; other parcels were reseeded twice.
Figure 3 shows the number of producers who had to reseed zai' or non-zai parcels.
(number of producers)
Zai parcels were resown less frequently, and by fewer producers. As has already been noted,
germination was better in the zai parcels. It could therefore be concluded that zai parcels favor
better germination. Zai's greater resistance to drought was also brought out in comments made
In assessing the need to reseed zai parcels, a number of factors were cited. These are
summarized in Figure 4. Observations made by the farmers themselves complete this
Reasons Given For ReSowing Zai Parcels
za'f made incorrectly
'0 5 10 15 20 25 30 35 40
/% of responses
The drought affected all parcels, both zai and non-zai. Its impact was felt more on the latter,
but the difference between the two in terms of the area that had to be re-seeded was not
calculated. The lack of practical knowledge concerning how to most effectively implement zai
methods presents significant problems to those considering adopting such methods. Errors
were made, for example:
Digging holes incorrectly: either not deep enough (so that seeds were washed away by rain); or
dug too early (so that wind formed a hard crust on top, making it necessary to dig again, more
deeply, before sowing). Also, sometimes, not enough manure was used.
Sowing was done incorrectly in some cases. Instead of waiting for the first rains before sowing
producers say that it is better to sow on dry soil.
Sowing directly onto manure is not advised. Instead, soil and manure must be mixed well in
order to promote deep root growth (otherwise, the plant will not withstand drought).
Pests were often responsible for the failure of the first sowing. In particular, there were
infestations of Striga in the parcel and worms in the manure. Shea nut leaves and debris
blown by the wind blocked the holes.
Choice of site may also be a source of problems. If a parcel is located in the way of natural
drainage, water washes the manure away. In low-lying areas, maintenance is too difficult. One
site which was situated where there used to be a house suffered from soil that was both too dry
and too acid.
Expectations from Zai parcels. Producers employing zai methods for the first time enjoyed an
increase in productivity both in parcels that were already fairly productive, as well as in those
parcels that had been degraded. Three quarters of those surveyed were expecting to improve
production with zai. For the others, this represented a trial: new technology that was the talk of
the village, and they took part in the trial for that reason.
For whatever reason given for trying zai at the outset, producers were clearly interested in
repeating the trial next year, as can be seen in Figure 5. The comments of both those who were
and were not interested in the trial are recorded in the next section.
Trying Zai Next Year
(percent of responses)
IV. PERSPECTIVES: GOALS FOR NEXT YEAR
What to do next year and where. Eighty one percent of producers are planning to use zai next
year. Of the five that are not interested, three decided not to try in 1993. Why not? Reasons
for not trying include a lack of manure, or of manpower, or a preference for mulching. Two
others did try in 1993, but were not satisfied with the results.
For the remainder, zai represented a breakthrough, to quote one participant: "The RSP showed
us how to reclaim exhausted soil ". The great majority of producers plan on using zai' next
year, and use it on larger areas (see Figure 6). Some producers plan on reusing last year's
holes, taking advantage of the remaining manure left there.
No Response (7.1%)
Factors that influence the choice of site are many (see Figure 7). In general producers start by,
tackling poor, dry soils, which yield almost nothing. Some producers have targeted the zi-p,,
or koun koubris.
Expected Placement of Zai in 1994
(percent of responses)
with rock bunds
parcel near the house
low fertility soil -",
0 5 10 15 20 25 30 35 40 45 50
Expected Size of the Zai Parcel in 1994
(in comparison to 1993)
Smaller (5.9%) i-Same (11.8%)
Gravel-based soil (well drained) is also considered appropriate. Where transport (of manure) is
a challenge, compound land is used for the zai site.
The predominance of mulching in Donsin indicates its importance. In spite of their satisfaction
with zai, farmers do not intend to abandon mulching.
Certain recommend using both zai and mulching together. For others, zai is more appropriate
for use on the most unproductive parcels, such as zi-p616 land, and mulching is used for the
Inputs: Fertilizer. Equipment, and Labor. It is clear that the availability of manure and
manpower in each family are prerequisites for the use of zai. Almost every producer expressed
his intentions for next year in terms of a precise and pragmatic scenario comprising various
activities which make competing demands on his time. Work (involving either mulch-
collection the construction of zai or of bunds) starts in January and continues through until the
first rains. The area involved is limited by the amount of labor available. Manure, compost
and household waste are more or less available, but if it is necessary to go and collect manure
from the fields that also takes time. The farmers therefore manage their limited resources
according to a given strategy. Manpower becomes an issue especially when the number of
able-bodied workers in the family is less than four. One result that is not without interest is
noted in Figure 8, namely that farmers who come from the highest socio-economic class in the
village do not experience the same constraints in this respect.
by socio-economic group
3O |- .....................................
first group third group
second group fourth group
Figure 8. % of households
The availability of manure to make zai does not pose a problem for the majority of producers
(see Figure 9). Animals, and especially small ruminants are numerous in Donsin. Even if one
farmer does not have enough of his own animals, he can get manure from a neighbor, parent, or
from pasture land. Some farmers have their own compost pits; others collect household waste.
In any case, producers worry more about the problem of distribution than the availability of
manure. This question is tackled in the discussion of the impact of zai in the future.
Expected Source of Fertilizer
for zai next year
No Response (19.0%)-
; -Livestock (69.0%)
Producers are interested in using the appropriate equipment for preparing zal. Traditional hoes
are not strong enough to dig the zai pits. Instead, it is suggested that ice-picks or Yatenga axes
be used to make the task less costly.
Techniques of choice for successful zal. Farmers recognize the importance of mulch, rock
bunds, and other measures taken to fight resource degradation such as earth bunds and
scattering tree branches. Zai is perceived as another such technique, but one which does not
necessarily go with those mentioned above. Forty two percent (42%) of farmers have already
used one CES technique on their land.
Farmers make their own suggestions for successful zai cultivation. For example, they
recommend leaving ample space between rows to encourage plant growth. In addition, they
recommend that zai pits should not be dug too deeply, since this tends to increase the worm
population in the manure. Zai that are dug too closely together inhibit the growth of shoots.
Lastly, manure is preferred to household waste.
The Impact of Zai on the Village. The farmers agreed almost unanimously that farming with
zai would result in a reduction in the amount of land cultivated. The sheer demands zai makes
on labor make it impossible to carry on cultivating the same surface area as at present.
Moreover, the presence of rock bunds, or other CES measures will establish defined field
limits. The farmers therefore foresee an end to itinerant farming. At the same time, the new
methods would bring improvements to the soil.
A definite need for fertilizer is foreseen. Farmers talk of stalling their animals as one solution.
However, one must admit that the local population did have a certain reticence to this method.
This is most probably due to their concerns about how to feed their livestock. RSP experience
has shown that the Donsin villagers are not interested in producing fodder. Sheepfolds, cited by
quite a lot of villagers, are a realistic alternative for the production of organic material.
Donsin villagers consider making sturdier hoes. The RSP conducted tests on several models
which had been made in Yatenga. At present, any model can be chosen and a copy made up by
a blacksmith in Boulsa. The SANREM project also plans to help the Donsin villagers to
develop appropriate tools.
Two further observations complete this discussion. Mutual aid group work in Donsin could
resolve the labor shortage but this practice is not well developed in the village and its future
development is uncertain.
Wandering animals prevent sowing at the beginning of the season. In principle, the village
chief puts a stop to this when he considers it time to start sowing. If sowing is to be done
before the first rain, wandering animals should be stopped earlier. The likelihood of this
happening is equally uncertain.
The role of RSP. The villagers were consulted as to how the RSP and its collaborators
(researchers, extensionists etc.) could best help them prepare the zai next year.
Typically this kind of question provokes a passive response, of the "give us this, give us that"
kind. This kind of reply indicates a strongly felt need for the right equipment to prepare the zai
and was quite a common response in the village. More important still are those replies which
manifest the villagers' constructive interest in making the best tools themselves or in improving
the making of zai should the material and knowledge be supplied. (Diagram 10). Education
figures prominently in the current role of the RSP in the village.
The importance of strong hoes has already been mentioned. Wheelbarrows and carts are also
important. A better functional knowledge was requested, i.e. how tomake good compost, and
knowledge of how to make the zai themselves (specifications for digging the pits, how much
manure to use, spacing, seeding, etc.). Questionnaires recorded equally dynamic and concrete
replies from farmers who sought further knowledge of agroforestry. There were also requests
like "show our blacksmiths how to make the three-pronged hoe".
Conclusions and Recommendations. The interest of Donsin farmers in zai has been firmly
established. The villagers themselves decided to adopt zai technology on a trial basis, even
without the intervention of the RSP, whose role was limited to such things as facilitating the
"farmer to farmer" visit, demonstrating how to recuperate a zi-p616, and proposing that the 6
partners test zai-mulching. The intention to use zai next year has been clearly expressed. A
growth of 15% by comparison with this year can be anticipated. This would mean that 80% of
villagers will use zai in 1994.
The advantages of zai for the population have been clearly expressed: the superior retention of
moisture; better resistance to drought; good germination record as well as good yield; and the
restoration of zi-pls (exhausted land). However, the ultimate adoption of this technology
depends on its resource requirements, in particular the labor and organic matter required.
Hence the necessity for each farmer to make his own evaluation according to his particular
priorities and the availability of the required resources.
Action to be taken. For the RSP program, there are still details about zai technology that need
to be communicated to the farmers, because this will be indispensable to ensure future success.
Providing answers, for example, to questions such as:
* How to obtain the best level of profitability on parcels of quality varying from fertile arable
land to eroded or "zi-pel6" parcels.
* Under what conditions is it better to mulch rather than use zai, or when should both methods
* In what form should the manure be and how should it be applied?
* How much reseeding is required with different za itechniques?..."
Preparatory activities must also be considered, such as the availability of manure which implies
animal management (stalling and feeding). Associated activities, such as the creation of rock
bunds or agroforestry are also elements in this equation. Equally, there are actions to be taken
to address equipment problems. The making of sturdier hoes, discussed above, is underway, as
is the test of hand-pulled carts for transporting the organic matter.
Finally, the development of the villagers' awareness of the issues and methods is crucial. To
date, the RSP plans a feedback session in Donsin, to be followed by a demonstration of the
correct method for preparing zai. For this, organizers plan to use the "farmer to farmer"
approach again. Group testing is also appropriate, promoting collaboration in the form of
association or grouping of farmers.
One farmer suggested that the RSP could help villagers if it provided "the goats and chickens
that must be sacrificed In order to satisfy customary obligations" (and thereby ensure rain).
The RSP does not have the ability to provide rain, but it can conduct research. RSP's goats are
their on-farm trials, and their chickens are the education programs. The customary obligations
that the RSP will help the villagers satisfy are those which restore the land for future
generations, and which leave a legacy which ensures agriculture of a truly lasting nature.
Economic Analysis of Compost Production in Southwestern Burkina Faso
S. Amadou, M. Bertelsen, S. Ouedraogo
Soils in Burkina Faso, and more specifically, in the West are being continually degraded. This is
due to a number of factors, including the growth in surface area under cultivation as a result of
emigration and the consequent demographic pressure on the land. At the same time, the
withdrawal of input subsidies has meant that farmers have had to shoulder increasing expenses,
while seeing their incomes drop.
There is ample research evidence that the use of organic matter, including the use of compost,
contributes effectively to restoring soil fertility, whereas the continual use of mineral fertilizer
ultimately leads to a loss of fertility.
However, very little is known about the economic aspects of composting. The latter should go
hand in hand with the research undertaken on technical aspects, with a view to indicating the
different users and eventually to directing researchers towards inexpensive innovations.
The object of this study is to make an economic analysis of the preparation of compost, as
compared with the cost of buying mineral fertilizer, notably NPK, which is widely used. The
projected results will:
* Provide information for researchers and decision-makers on the chances of adopting
composting, considering the socio-economic difficulties which are associated with compost-
* Help interested farmers make the decision to adopt composting by making available to them
the socio-economic data collected.
* Contribute to the identification of new directions for research, including research on the use of
certain materials, with a view to reducing the cost of compost production.
* Help lift any constraints which discourage the adoption of composting.
The site chosen for this study was the village of Kayao, situated in the Province of Mouhoun, in
Burkina Faso. Kayao is one of the four sites chosen by the Western Zone Farming Systems
Research Program (RSP), in the Western Zone of the Institute for Agricultural Study & Research
(INERA- Institut d'Etudes et de Recherches Agricoles) of Burkina Faso's Western Zone.
As in most regions of Southwestern Burkina Faso, Kayao is a zone which receives significant
immigration. According to INERA typology, it is representative of the Western zone's poorly
equipped, cotton-producing areas. Cultivation methods are extensive ones, which under the
pressure of emigration considerably reduces the availability of arable land; thereby provoking land
conflicts between immigrants and indigenous peoples. (Ou6draogo, 1991)
Moreover, because of the extension of cultivated areas and the reduction of fallow land, Kayao
lands have mostly been degraded, or are on the way to becoming so. (Sanou, 1991)
In order to address the challenges in this zone, the RSP has put in place a system which integrates
crop and the livestock, production systems. This was done by testing feeding regimes on draft
cattle and compost making.
In order to evaluate the profitability of compost by comparison with mineral fertilizers, a
questionnaire was prepared which was designed to collect data on compost production.
Researchers conducting the survey held weekly interviews with participating farmers to collect
information on the following variables:
* labor involved for all operations
* inventory of materials used
* number of animals
* types of inputs used
* volume of compost pits.
A total of ten participant-farmers were chosen, selected according to their willingness to invest in
this technology and acquire or possess the required cattle and plant material. These variables
were supplemented by secondary data (bibliography and price surveys).
The infrastructure and the research technicians already in place enabled us to conduct these
surveys with ease.
The research hypothesis is that the cost of soil improvement with compost is less than that with
mineral fertilizer, and that the latter could be used more efficiently as a complement to, rather than
a replacement for compost..
Equipment. The acquisition of appropriate materials constitutes the principal difficulty involved
here. In fact, the necessary investment in materials for the first year appeared to be very costly for
farmers whose income is limited. According to our data, acquisition of materials amounted to
about 90,600 Fcfa. However, most farmers already possessed at least some of the necessary
tools, including carts, dabast, mattocks and buckets. By deducting the cost of materials already in
the possession of the farmers, the cost of materials that still had to be bought amounted to 13,000
Fcfa: The latter represents the price of an pick, a spade, two basins and a barrel, none of which
were generally owned by the farmers.
However, the annual depreciation cost is low given the long life of materials and the compost pit.
In addition, some tools, such as the pick, are mostly used only when digging the pits.
Water. The survey makes clear that the availability of water constitutes a barrier to composting.
Most of our participant-farmers were unable to follow the recommended composting techniques
because of lack of water. According to the survey, the nearest water source was approximately 6
kilometers from the village where the compost pits were located.
Labor. The survey reveals that two kinds of labor are used including family members and paid
The participating farmers depended on hired labor more for digging the pits (46% of total digging
labor time) than for filling the pits. Only 4% of labor for filling pits was hired.The table below
indicates the average time spent, in man-hours per cubic meter, on digging and filling operations,
Table 1. Average labor time for production of one cubic meter of compost in'Kayao.
Task Labor time (Hours) Standard deviation,
Digging out pit j8.88 2.48
Filling in 7.18 1.5
Number of observations = 10
It is clear from Table 1 that digging timevaries greatly. The cost of digging is estimated at 165
Fcfa/m3. The cost of filling is estimated at around 19 Fcfa per hour of digging, averaging out
costs for the duration of the operation. Even if this figure appears to be very low, in comparison
to the opportunity cost (100 Fcfa) during the field work, this was expected, because the digging
operation took place during the "hungry season. Sidib6 (1993) found an even lower opportunity
cost, for female labor (10 Fcfa per hour) in the village Yasso, in Kossi province situated in
Southwestern Burkina Faso.
Analysis of the compost production costs: Hypotheses of the analysis.
* The pit is assumed to last for five years.
* Equipment depreciation.
1 dabas are short handled hoes used in Africa
Table 2 indicates the projected equipment lifetime and the rate of use of the equipment for
compost production. These elements served as a basis for calculating straight line depreciation.
Equipment Price (Fcfa) Number Replacement Rate of use Depreciation
rate/year (%) (Fcfa)
Pick 3,900 1 2 years 10% 975
Mattock 400 1 1 year 100% 400
Spade 1,000 1 5 years 100% 200
Basin 5,000 2 2 years 100% 2500
Daba hoe 800 2 1 year 100% 800
Bucket 1,500 1 2 years 50% 375
Cart 75,000 1 5 years 10% 1500
Barrel 3,000 1 5 years 50% 300
Opportunity cost of labor. For the digging, the labor cost used reflects the actual amount paid by
producers, that is 165 Fcfa/m3. To estimate the cost of filling in the pit, the opportunity cost of
labor in Kayao was used, i.e. around 100 Fcfa throughout the cropping season. It is
recommended that the Compost pit be filled during the cropping season so that it can benefit from
Estimates of the amount of time taken to empty the pit were not available when the analysis was
made. It was therefore assumed that the length of time taken to empty a pit was equivalent to half
the time taken to fill it, i.e. 3.59 hours/m3 The same applies to time taken to transport and
spread the compost, for which secondary information was used.2
Comparative analysis of soil amendments. Table 3 shows that the production of one cubic meter
of compost requires an investment of 2,476 Fcfa. To obtain the recommended amount of
compost per hectare3, which is 2.5 tons (CFA, 1991) requires an investment of 26,914 Fcfa (see
Table 3). According to Segda (1991), crop yields comparable to those obtained using the above
amount of compost were obtained in Sourou by using 300 kg/ha of NPK, at a cost of 28,500
Fcfa/ha, with subsidized NPK priced at 95 Fcfa per kg.
Table 3. Production costs of compost per cubic meter, in Kayao.
Category Costs (FCFA)
Labor (digging)5 33
2 Based on INERA/RSP Western Zone data on agricultural labor times.
3 hectare = 2.471 acres
4 Depreciation costs of equipment which lasts for the same length of time as the pit (5 years).
5 Time taken to dig ( 8.8 hours) x primary tillage costs (19 Fcfa per hour)/5 yrs.
Burkina Phosphate 552
Given the transport and spreading costs for both fertilizers, Table 4 shows that there is very little
difference between the investment required when using NPK as opposed to compost, at least for
the first year of use, i.e.: 28,054 Fcfa for compost, and 28,614 Fcfa for NPK.
However, Segda (1991) has pointed out that compost remains effective for a two-year period.
On the other hand, if mineral fertilizer is to be the sole soil amendment, it has to be applied
To establish a proper comparison between investment costs, they must be actualized. Table 4
therefore shows the actualized costs of using NPK or compost, respectively, as soil amendments.
This is based on the assumption that the capital opportunity cost is at 1000%9.
Table 4. Comparative amendment costs: NPK versus Compost.
Category Compost NPK
Production or Purchase 26,914, 28,500
Transport to field (5 kin.)12 912 57
Spreading over 1 ha. 228 57
Total 28,054 28,614
Present value (c.op.1,000%)13 28,054 2,60114
10% of production cost
7 Time taken for filling (7.18 hours) x opportunity cost of labor (100 Fcfa per hour).
8 Depreciation costs of equipment whose lifetime is shorter than that of the pit (5 years).
9 This hypothesis was based on the results of Lowenberg et al. (1993) which indicated that most of the small
lucrative activities had an annual rate of profit of about 1000%.
10 Calculation based on the assumption that one pit, with a volume of 10.87 m3, to produce the recommended
amount of 2.5 tons/ha of compost, which remains effective for 2 years.
' Recommended amount in order to obtain the same yield as produced by 2.5 tons of compost.
12 Calculated on the basis that 16 carts-full of compost, each of which takes an average of 1.30 hr. to transport to
13Opportunity cost of capital.
14 Present value of the cost of NPK, for the second year.
Table 4 also shows that the comparative investment cost between these two amendments is only
very slight. One notes that the cost of compost in this case is only 10% lower than that of NPK.
The difference is approximately 3000 Fcfa.
In some cases, a discount rate of 1000% seems very high. This is why Table 5 shows the costs of
NPK and compost, at several levels of opportunity costs (40%, 100%, 500%, and 1000%).
Table 5 shows that there are very significant differences, related to the various opportunity costs.
For an opportunity cost of 40%, for example, the amendment cost of NPK appears to be almost
twice as high as that of compost (28,054 Fcfa for compost, as opposed to 49,053 Fcfa for NPK).
Even so, Table 3 indicated that there was very little difference between these costs when the
opportunity cost is at 1000%.
Table 5. Comparative costs of amendment: NPK versus Compost, according to opportunity cost
Capital opportunity 40% 100% 500% 1000%
Compost 28,054 28,054 28,054 28,054
NPK 49,053 42,921 33,383 31,215
The economic analysis of compost production shows that, regardless of the bio-physical soil
improvements, the use of compost for soil improvement is also more profitable than the use of
mineral fertilizer (NPK).
However, this profitability is strongly influenced by the opportunity cost of capital. It is therefore
very important to factor in the latter as decisive when calculating the present value of the cost of
NPK. (The present value of the investment in NPK was used because two investments in a two-
year period produce the same effect as only one investment in compost over the same period).
Thus, the relative profitability of composting diminishes when costs of capital run very high (at
1000%, for example).
In spite of the fact that the financial profitability of composting has been demonstrated, there still
remain other obstacles which may hinder its adoption.
* There are obstacles linked to the difficulties of acquiring the necessary equipment to make the
compost pits. These obstacles are mostly due to the high investment required relative to the
low incomes of producers, and to the fact that some equipment is used only the first year.
However, this second problem could be solved by buying and managing equipment
The scarcity of water during the dry season constitutes another obstacle to adopting
composting for soil amendment. In fact, the distance from water sources is discouraging for
all but a few producers, who must first of all supply their own needs. It is indispensable that
the decision-makers consider the problem of water availability as one which not only involves
ensuring supplies for the population, but which is also vitally linked to improving soil
* The many interviews with producers revealed that a large proportion of them were unaware of
the method and effects of composting. They preferred to wait and see for themselves the
effects on soil fertility of this first trial of composting, that they would then compare with their
own practices using manure from livestock corrals. It is therefore necessary to emphasize the
importance of training and information using demonstrations.
G.F.A. 1991. Les perspectives de la culture attel6e au Burkina Faso. Annexes.
Lowenberg-DeBoer, J., Tahirou Abdoulaye, and Daniel Kabore. 1993. The Opportunity Cost of
Capital for Agriculture in Sahel: Case Study Evidence from Niger and Burkina Faso.
Ou6draogo, S. 1991. Influence des modes d'acces A la terre sur la productivity des exploitations
agricoles: le cas de la Zone Ouest de Burkina Faso. Thesis. University Nationale de la
C6te d'Ivoire/Centre de Recherches Economiques et Sociales
Sanou, Patrice. 1991. L'insertion spatiale des producteurs agro-pastoraux dans la Zone Ouest du
Burkina: cas deDimolo, Kawara, Kayao, Yasso. Rapport d'activit6. INERA/RSP zone
Ouest. [Activity report. INERA/RSP Western Zone].
Segda, Z. 1991. Contribution a la valorisation agricole des r6sidus de culture dans le Plateau
Central du Burkina Faso. Inventaire des disponibilites en matibre organique et etude des
effects de l'inoculum Micro 110 IBF dans la pratique du compostage.
Sidib6, A. 1991. Analyse economique de quelques activities des femmes lies a l'utilisation des
products forestiers non ligneux dans la zone Ouest du Burkina Faso. Communication
presented au symposium sur les parcs agro-forestiers des zones semi-arides d'Afrique de
l'Ouest. [Paper presented to the Symposium on the Agroforestry parklands of the Semi-
Arid Zones of West Africa]. Ouagadougou, Burkina Faso: 25-28 October 1993.
Use of Unconventional Products in Animal Health Care:
The Case of Draft Animals in Burkina Faso
A survey conducted in 1990 in the four Western RSP village sites revealed a large cattle herd size
in these villages. Herds of the two cotton-producing sites, where animal traction is widely used,
include a large number of draft oxen (Badini 1990). The survey shows that no veterinary services
are available in any of these villages. The national vaccination campaign is conducted every year
in the villages between February and June.
During the agricultural season farmers are nevertheless faced with animal health problems during
the agricultural campaign, which require assistance from a livestock or veterinary professional. In
principle, farmers have had no training in primary care of animals and should consult a
,veterinarian. However, the closest animal health station is about thirty kilometers away from the
village. What do farmers resort to during the agricultural season when their draft oxen are ill?
In order to address this question, a preliminary survey was conducted in Kayao1, one of the RSP
villages, which represented a sub-zone characterized by subsistence cropping with some cotton
production and a high rate of animal traction in 1990. There are 223 draft oxen, 88 donkeys, and
1189 cattle, in the village and 34 % of the villagers use animal traction (Badini 1990).
The objectives of this survey were to:
* Gather knowledge on farmer practices,
* Inventory the unconventional products used for draft animal health care,
* List the supply sources for these products, application methods and treatment costs.
Data was collected through informal interviews with farm family members possessing draft oxen.
The interviews especially focused on young individuals who work with these animals and maintain
them. These individuals monitor the behavior of draft animals and know them best. The survey
was initiated during the 1991-1992 season and was completed in January 1994.
A questionnaire was used throughout the survey. However, it should be noted that this
questionnaire was not uniformly administered because some questions dealt with sensitive issues.
Most people interviewed found it difficult to answer the questions, sensing that they had to
1Survey participants: Sanou Sogo and Dabilou Salam, technician-enumerators Fandi6 Kank6ki, village enumerator
compromise their own interests. Clandestine vaccination is normally prohibited in Burkina Faso.
Therefore, a combination of interviews and direct observations of farm practices was used.
Every year, the selected sample included 25 production units using animal traction and
participating in RSP tests. However, observations were made on all farms having draft oxen.
Disease diagnosis by farmers. Farmers use their experience as a basis for diagnosing animal
diseases. They do not have the necessary knowledge of specific pathogens, but often recognize
aggravating factors such as change in climate, diet, work schedule, etc. Once an animal displays
an abnormal behavior such as sadness, fatigue, lethargy, refusal of food, etc., farmers examine the
animal (body temperature by touch, feces, eyes, snout, hair, nasal secretions, etc.). More
knowledgeable farmers are called upon to determine the cause of the disease when they cannot do
Reasons for the use of unconventional products. The survey results indicate that:
* To call the livestock agent of the village of Oronkua to Kayao is rather difficult, because
appointments are often missed and the agent is often in field missions and so cannot be easily
found at his office2;
* Rarely can the prescribed remedy be obtained at the livestock extension station. Therefore,
farmers have to go to a larger center such as the CRPA at 90 or 190 km distance. This causes
a delay between the time of the visit and the beginning of the treatment. Additional problems
arise when farmers cannot procure the product (because products are frequently out of stock)
or when subsequent services of the agent are required, as in the case of injectable products.
Problems which discourage farmers from using conventional products are in order of importance:
* Difficulties of access to conventional products (70%)
* Unavailability of the livestock extension agent (20%)
* Confidence in unconventional products (10%)
Products used. Products used to treat draft animals are of various origins. They include drugs,
pesticides, antibiotics and other illicit pharmaceutical products.
Table 1 indicates that none of the products used is specifically intended for animal care, Most
pharmaceutical products are fraudulently imported from Ghana. They are confiscated by the
police during checks of the local markets.
One of the advantages of unconventional treatments for farmers appears in Table 2. Treatment
cost varies between 30 and 250 Fcfa, which is inexpensive relative to conventional treatments and
affordable to those raising cattle.
Antibiotics are all called "toupai" (literal translation); This is because they are used to care for all
diseases. They are of two kinds: Totapen and tetracycline, both in capsules.
Paracetamols identified are of several sorts:
* Daga: package of three 500 mg pills
* Alagme: package of twelve 500 mg pills
* Phenic: Paracetamol in 500 mg and 250 mg dosages
The range of paracetamols is quite broad.
Table 1. Percentage of farmers interviewed having used one of 6 unconventional
products at least once to treat 8 frequent diseases (N=25)
Fever Diarrhea Wounds Skin Fatigue Skin General ill-
parasites disease being
Antibiotics 75- 77 -- 25 95 80
Tetracycline ____ '
Paracetamols 70 25 10 -- 61 28 96
Daga, alagme__ _____
Sumicidine -- -- -- 98 -- -- --
"Dissolution" -- -- 84 -- -- -- --
Instant coffee -- -- -- -- 11 25 --
Undetermined 5 -- -- -- 6
aNineteen farmers, for example, out of 25 (75%) said they used antibiotics against fever.
blnsecticide used in cotton production
CManufactured to repair bicycle inner tubes
Table 2. Description of unconventional products and their costs
Disease Product Form/Unit Unit cost Dose Cost
(F CFA) (F CFA)
Fever Totapen Capsule 25 3 75
Fever Tetracycline Capsule 15 4 60
Fever Daga Pill 50 2 100
Fever Daga Pill 50 2 100
Fever Alagme Pill 25 4 100
Diarrhea Totapen Capsule 25 10 250
Diarrhea Tetracycline Capsule 15 14 210
Wounds "Dissolution" Bottle 150, 1 150
Skin parasites Sumicidine Water jug <5000 <<1 200
Strepto-trichosis Totapen Capsule 25 6, 150
Skin disease Coffee Packet 30 1 30
Gen. ill-being Tetracycline Capsule 15 8 120
Gen. ill-being Totapen Capsule 25 4 100
Directions for use. The directions for use of products depends not only on the manner they are
administered but also on the disease to be treated, as one product can be used for the treatment of
Products can be orally administered in three ways:
* Diluting it in drinking water. The animal is left for some time without drinking water, then,
once thirsty, given a reasonable quantity of product-water mixture to be entirely drunk.
* Forcing the animal to swallow the entire product by pulling on its tongue
* Mixing it in food. This method is little used because of animal refusal.
The most widespread practice is intra-muscular injection in the shoulder or flank is also used.
Some "specialized" farmers being especially skillful are called upon for injections. Two solvents
are used to dilute products. Water is most widely used. Drinking water is used without any
special treatment. It is not boiled nor filtered. "Koutoukou", an alcoholic drink derived from
sugar distillation, is also used by some people. It is a popular drink and is sometimes called "who
pushed me", or "zonzon", or "pat&c6". A liter of the sort with high alcohol content (about 90
degrees) costs 500 Fcfa, while the sort of low alcoholic content (9 to 20 degrees) is sold for 300
Fcfa a liter. The latter type is used as a solvent.
Syringes are generally designed for a single use, but they are used repeatedly until useless.
Syringes sold in pharmacies have a capacity of 5 to 10 ml. When animals are familiar with the
person doing the treatment, it is petted on the head for an easy intra-muscular injection.
Otherwise, the animal has to be immobilized by the head or taken down.
Example of preparation of injectable products against fever. The origin of the fever is generally
not identified. It is noted that the animal is sad with bristled hair and high body temperature. The
treatment includes tetracyclines and "daga" paracetamol diluted in a bottle of "koutoukou" or
water, which has to be strongly shaken for proper mixing. The solution is pumped into the
syringe without the needle. Air is drawn out of the syringe before the intra-muscular injection.
In general, the dosage used is not well defined. It varies according to the person administering
the treatment, the stage of development of the disease and the age of the animal.
Examples of products for external treatment.
Wounds. Mostly used for this purpose is the "dissolution", a glue product used to repair cycle
inner tubes. It is an aseptic wound dressing. This product replaces Synexa and K'otrine more and
more frequently. The glue is applied on the wound in a thin layer. A few moments later, the
product is dry and protects the wound from flies and other insects. The dressing is repeated until
the wound heals.
Tick control. There are two products used against ticks; both are pest control products used in
cotton production: Sumicidine and decis. The first method consists in mixing eight to ten
volumes measured with a peanut shell (7 to 10 ml) in a bucketful of water and spraying the
mixture on the body of the animal with a backpack sprayer. In the second method, fresh cattle
feces are kneaded in the chemical liquid and the animal is coated with this paste.
General fatigue. The animal is tense. It cannot stand sustained effort. In order to give him a
boost, three "daga" pills are administered orally or by injection.
Therefore, the main diseases treated which are commonly encountered are affections that do not
require the intervention of a livestock specialist, with the exception of skin streptotrichosis.
Vaccination accidents. These medication practices are not without consequences. The use of
products or treatment equipment can be traumatic to animals.
Solvents. Water used as solvent is not treated in any special way and may come from
questionable sources. It may contaminate animals with other diseases.
Products. Antibiotics of unknown origin without written directions and known expiration dates
may be dangerous. The use of bacteriostatic or bactericidal chemicals without the advice of a
veterinarian can be overdone. Potential results include accidents and resistance to antibiotics.
Syringes. These can be sources of trauma and infection to animals, as they are not disinfected
beforehand and are used repeatedly. Interviews revealed that there are accidents due to
* Bleeding. A poorly practiced injection may result in bleeding which can then turn into
* Oedema. There is swelling where the injection was done. It can worsen and cause an abscess.
Animals may even have reduced mobility on the side of the leg affected by the injection.
Interviews indicate that in order to care for their draft animals, farmers develop their own
strategies and use fraudulent pharmaceuticals, pesticides, and other products (coffee, glue, etc.).
The difficulty of access to veterinarian services and the high cost of conventional remedies
(including transportation costs to Didbougou or Bobo-Dioulasso) on one hand, and, the
availability of products and solvents (water, local alcohol) and of soiled equipment (used syringes)
on the other hand make these practices easier and help to perpetuate them. Even though these
practices are applied to animals, they give rise to serious problems for farmers and risk the health
of their livestock.
Badini, 0. 1990. Census of livestock management systems of Western Burkina Faso. RSP
Western Zone Report.
Evolution of Sedimentation, Surface Micro-Morphology,
and Millet Production in Response to Soil Conservation
Practices on an Eroded Site at Yilou, Burkina Faso
N. F. Kambou, S. J.-B. Taonda, R. Zougmore,
B. Kabor6, and J. Dickey
Land degradation is widespread in Burkina Faso. Approximately 24% of the country's land
area is highly degraded, and most of this land is concentrated in central Burkina on the Mossi
Plateau (Figure 1). Current land management practices and population pressure contribute to
the high rate of arable land degradation, and therefore threaten environmental quality and food
security in the short and long run. Applied research in the area of on-farm soil and water
conservation is a priority activity of the Water, Soils, Fertility, Irrigation, and Agricultural
Machinery program of the National Institute for Agricultural Studies and Research.
Degradation zones in
Years of farming without adequately protecting land from intense rainfall and runoff have
resulted in formation of extremely dense, bare surfaces capped by a one-to-two-cm-thick crust.
These areas are known locally as zi-pel6, or literally "white places". A number of soil
High 66914 24
Total 274286 100
reclamation practices have been successfully applied by farmers to these sites. Three of the
most successful practices are:
* Rock bunds placed on the contour to slow runoff and encourage sedimentation
* Mulching with grass or crop residues to protect the soil surface from raindrop impact,
encourages termite activity and porosity, reduces evaporation, and adds some organic matter
* Zai which are small holes that are filled with manure or compost, and then seeded with a
Usually, a single practice is applied to a given land area, or rock bunds are combined with one
of the other practices. Our objectives were:
* To begin to describe the mechanisms, rate, and extent of soil reclamation effects above and
just below the first rock bund in a watershed
* To evaluate the impact of mulching and zai, alone and in combination, on soil surface
condition and millet production
The sites chosen were zi-pele (2% slope) at Yilou and Nioniogo (Figure 2), where highly
variable annual rainfall averages about 600 mm (Figure 3). In general, the soil was
approximately 30 cm of very compact sandy loam over a laterite pan. A (laterite) rock bund of
about 30 cm height and 100 m length (Figure 4) was constructed on the contour, and an area
(downhill) was fenced off for cultivation, since zi-pele receive a great deal of domestic animal
traffic during the growing season. The following soil management treatments were applied
during the 1991 to 1993 seasons at Yilou in three randomized complete blocks (RBCD's), and
in 1993 in 4 RBCD's at Nioniogo inside the fenced areas (Yilou site shown in Figure 4):
* Direct sowing into the bare soil
* Grass mulching before sowing
* Zai' dug on an 80 cm x 80 cm square grid (zai in lines), then grass mulching
* Zai dug on an 80 cm x 80 cm square grid
* Zai at a similar density, but scattered irregularly
On each plot at Yilou, the same treatment was applied in each of the three years. Millet stover
from each plot was harvested, stored, and reapplied to the same plot just before the next season.
The millet variety IKMV-8201 was planted each year. All plots received 100 kg/ha of 14-23-
14-6-2 (N-P-K-S-B) at planting and 100 kg/ha of urea (46-0-0), for a total of 60-23-14-6-2.
Weeds were hand-pulled before they competed with the crop during 1991 and 1992 at Yilou
and in 1993 at Nioniogo. Weed pressure at Yilou in 1993 required a weeding with a hand hoe.
Soil surface condition and sedimentation were observed in the plots after the 1992 season at
Yilou, as was millet growth and yield in all years at Yilou and in 1993 at Nioniogo.
SMain road, year-round practicability.
Main road, intermittent practicability.
_______ Test Location
= >. -)
SYilou 1992, Total 510 mm
Median per decade
20th and 80th percentiles
at Mand 1970-1990
Average Total Annual Rainfall
1970-1992 = 616 mm, Mane
1962-1969 = 762 mm, Mane
SOURCE: IGN, 1959, TOPOGRAPHIC MAP OF WEST AFRICA, 1/500,000
After the 1992 season, we described the sedimentation and micro-profiles in 5 shallow soil pits
on a 30-m transect above the two-year-old rock bund at Yilou (Figure 4).
I I I
1 **I V i\: \ I
SClay and silt deposits
Clay and gravel deposits
-- I Rock Bund
S FR, Frst replication
S RI Second replication
RI Third replication
T Control plot
T4 Zal + mulching
Ts Scatted za
4 71 Plant cover
Z Bare, eroded surface
III. RESULTS AND DISCUSSION
The pattern of sedimentation uphill from the rock bund at Yilou was approximately as follows
SEDIMENTATION FRONT UPHILL FROM A ROCK BUND, YILOU, 1992:
Profile 5 Profile 4 Profile 3 Profile 2 Profile 1
0 ... Rock Bund
i.; mm. -
1 --0; m----. -- ---------
-l l1 i i l l ll"-- -
250 mm -
Organic matter made up of algae under
Clayey plasmic horizon creating an
Coarse elements (gravel).
Clay, silt and fine sand mixture with
Clay, silt and sand mixture with 10%
Hyper proportion of coarse elements
SProfile 1 (0.81 m from bund): Several mm of clay underlain by layers of silt and fine sand to
5 mm, fine and medium sand layers to 10 mm. Underlain by gravelly sandy loam. (Figure 6).
* **** *, p.
* ... .
a. o. .
* e *
'- : i ".
I I I I I I I
Profile #1 (0.81 m from rock bund)
1 Sedimentation layer made up of sorted sand: from top to
bottom clay, silt and fine sand
Color 10YR 6/3 Thickness
2 Sandy layer: Thickness: 5 mm
3 Mixture of clay, silt, sand and 30 to 35% coarse
10 YR 6/3 Thickness: 85mm
4 Mixture of clay, silt and 25% coarse material.
10 YR 6/6 Thickness: 235 mm
5 Rock pan
10 YR 6/6
* Profile 2 (6.2 m from bund): Laterite pea gravel to 8 mm, underlain by a 0.5-mm clay
deposit, another mm of fine sand, and 10 mm of sandy loam with significant organic
matter, for a total depth of 22 mm. This underlain by gravelly sandy loam (Figure 7).
..... ..... ....
.. ...... .. .
* Profile 3 (27.7 m from bund): Pea gravel scattered thinly on top of 15 mm of massive, sandy
loam, the surface sealed in a crust of very low porosity. This over gravelly loam (Figure 8).
Profile #3 (27.67 m from rock bund)
1 Highly compacted clayey layer (Thickness = 15mm).
Erosion crust preventing all infiltration. Plasmic horizon.
Presence of roots of 10 YR 5/8 color
Horizon color = 10 YR 6/2
2 Thickness = 70 mm. Over 60% coarse material. Clay
silty cement with organic matter. Grey color = 10Yr 6/3
3 Thickness = 70 mm. Over 40% coarse material. Clay-
sandy cement with organic matter. 10 YR 7/6
4 Rock pan with ferruginous modules. Thickness = 75 mm
10 YR 7/8
IILII 230 mm
Profile #2 (6.2 m from rock bund)
1 Coarse gravelly material
2 Plasmic film of clayey texture.
Color: 10 YR 7/3 size: 0.5 mm
3 Fine sand on sorted material.
10 YR 5/3 size: 3.5 mm
4 Mixture of silt and fine sand with high organic matter
10 YR 6/2 Size: 10mm
5 Mixture of fine (clay, silt and fine sand) and less than
10% of coarse elements
10 YR 6/3 Size 168 mm
6 Coarser elements than in 5. Transition zone with change
10 YR 6/6
7 Higher (35 to 40%) proportion of coarser elements.
10 YR 6/4 Size: 60 mm
8 Rock pan
10 YR 6/6
* ; .
. .. C 9 *
.'r '. '. & :'-
.* '< v .
SProfile 4 (40.8 m from bund): Massive sandy loam to 25 cm, the eroded surface sealed.
This overlaying gravelly sandy loam (Figure 9).
..bL b~ *~Q
Profile #4 (40.81m from rock bund)
25 mm 1 Highly compact clayey layer forming an erosion crust
preventing all infiltration. Plasmic horizon. Thickness -
Color 10YR 5/2
2 Thickness = 60 mm. Over 60% coarse material. Clay-
silty cement with organic matter. Grey color 10 YR 6/5
3 Over 40% coarse material. Clay-silty cement.
Thickness = 115 mm
10 YR 6/4
4 Rock pan
Thickness = 60 mm 10 YR 7/6
S. : .
SProfile 5 (50.8 m from bund, beginning to leave the zi-peld): Surface layer (4 mm) of
organic material, including algae under a layer of low, dead vegetation. This underlain by 12
4mm Profile #5 (50.81 m from rock bound)
16 mm 1 4 mm thick micro-horizon. Organic matter made up of
Color 10YR 4/2
2 Clayey plasmic horizon Thickness = 12 mm 10Yr 5/3
3 Mixture of clay, silt, sand and 30 to 40% gravel.
Thickness = 80 mm Roots 10YR 6/3
4 Thickness 104 mm. Horizon layer is darker than
horizon 3 and lighter than horizon 4. 5% small gravel.
Kaolimitc clay of 1/1 type
10 YR 4/3
5 Rock pan. 40% or more small gravel
10 YR 7/6
6 Horizon harder than horizon 5. Laterite.
mm of clay loam. Gravelly sandy loam begins again at 16 mm. First presence of plant tops
or roots in transect (Figure 10).
Most of the deposits that can benefit plant life are within a few meters of the bund. This zone
gradually gives way to the sealed, massive, eroded surface that characterizes the zi-pel6, which
in turn gives way to surfaces covered with some topsoil at the top of the zi-pele.
Below the rock bund, the erosive force of the water is reduced, but the surface conditions
created by the soil and water conservation treatments in the trial influence surface
characteristics and how much of this water actually infiltrates. The surface soil conditions in
the 5 treatments can be summarized as follows:
* Direct sowing: Massive, sealed, eroded surface remains, some gravel deposit on surface.
* Grass mulch: Surface protection by the mulch and termite activity under the mulch allowed
rehabilitation of porosity in the surface layer, and there were several mm of deposition of
clay to fine sand.
* Zai with grass mulch: Conditions between the zai' much as above, but surface local to zai
had higher porosity and reduced crust formation, and of course higher organic matter.
* Zai" in lines: Surface local to zai" had higher porosity and reduced crust formation, but this
effect was less than where zaif were combined with mulch. Areas between lines of zai
eroded, massive, sealed.
* Scattered zai: Much as above, but scattered zai' seem to slow surface flow enough to
increase deposition and reduce crusting of surfaces between zai.
Only rainfall data for 1992 at Yilou is available. Total annual rainfall in 1992 (510 mm) was
well below the post-1969 average of 616 mm (Figure 3). Dry conditions were primarily at the
beginning of the season, when rains began some weeks late, and during two dry mid-season
decades, one in mid-July and one in mid-August. Such conditions provide an opportunity to
evaluate the ability of soil and water conservation measures to reduce crop drought stress.
Yield and harvest index means and mean comparisons for individual site/years and for all 4
site/years grouped together are presented in Table 1. In general, all treatments but direct
sowing allowed for locally acceptable levels of cereal production on this otherwise baren land.
The various levels of grain and stover productivity indicate the rate of site reclamation for crop
production under the various practices. Table 1 also presents F values and probability levels for
comparisons among groups of treatments. The nature of these comparisons is illustrated in
Table 2 for the example of grain yield in the combined analysis. In the example, an F value of
80.8 is associated with a probability level of less than 0.001 for the comparison of direct
sowing with all other treatments taken together, suggesting a clear difference between grain
yield under direct sowing and under the other treatments. The only other comparison that is
significant for that analysis was between aligned and nonaligned zai (F=6.8, P=0.03).
Table 1. Yield and harvest index means and mean comparisons for individual and grouped site/years.
Grain yield (kg/ha) Stover yield (kg/ha) Plant height
Yilou Nion Tous Yilou Nion Tous Yilou Nion Tous
__iogo logo _ogo
Treatment 1991 1992 1993 1993 Avg. 1991 1992 1993 1993 avg. 1991 1992 1993 1993 Avg.
Dirct vs. others 0 161 253 49 111 0 539 777 83 329 0 123 36 51
Mulching 0 1334 829 37 614 73 2038 1553 247 921 0 200 124 110
Zai & mulching 118 1553 680 232 449 1142 2084 914 1345 1369 137 202 165 168
Zai lines 86 722 876 195 510 987 1170 1453 1492 1292 130 189 174 165
Zai not scattered 121 1084 1109 146 579 1854 2358 2328 2029 2133 139 192 153 160
Average 65 971 749 132 453 811 1638 1638 1039 1209 81 181 130 131
Comparison F value
Direct vs. others 14.5 65.9 27.0 12.1 80.8 71.3 41.5 21.6 14.4 70.9 306.3 31.9 56.7 59.1 113.7
Mulching vs. 5.0 15.5 0.2 25.3 0.1 1.9 1.1 0.0 17.8 8.0 62.0 0.4 0.0 6.4 58.9
Zai+ mulching 19.6 10.0 5.7 3.5 2.9 105.0 1.4 27.9 1.4 5.1 448.5 0.5 38.9 0.0 0.6
Zailines vs.. 1.7 5.3 2.4 1.7 6.8 32.6 19.4 16.8 1.8 37.8 1.6 0.0 24.1 1.2 0.3
scattered za'i ___
SProbability of an orthogonal comparison
Direct vs. others 0.005 <.001 <.001 0.005 <.001 <.001 <.001 0.002 0.003 <.001 <.001 <.001 <.001 <.001 <.001
Mulching vs. 0.06 0.004 0.6 <.001 0.8 0.2 0.3 0.95 0.001 0.02 <.001 0.6 0,98 0.03 <.001
Zai+ mulching 0.002 0.001 0.04 0.009 0.13 <.001 0.3 <.001 0.3 0.05 <.001 0.5 <.001 0.9 0.5
Zai lines vs. 0.2 0.05 0.2 0.2 0.03 <001 0.002 0.003 0.2 <.001 0.2 0.9 0.001 0.3 0.6
scattered zai __
Table 2. F Probability of an orthogonal comparison
Treatments Direct Mulching Zai+ Zai' in Scattered
sowing _mulching lines Zai
All years and sites
Direct Sowing <.001
Zai in lines 0.03
The following observations can be made:
During 1991 at Yilou and 1993 at Nioniogo, none of the treatments produced much grain,
however stover production was mostly between 1000 and 2000 kg/ha for all treatments that
included zai; and less than 300 kg/ha for other treatments.
* In 1992 and 1993 at Yilou, all treatments except direct sowing produced average grain yields
greater than 600 kg/ha. All of the treatments except direct sowing, therefore, can provide
acceptable levels of crop production under the rainfall and fertility conditions of this
experiment, but mulch alone cannot do so during the first year of reclamation.
Direct sowing means were the lowest in every analysis (except for grain yield of the mulch-
only treatment at Nioniogo), and is not suitable for soil reclamation, even below a rock bund
with adequate chemical fertilizer.
Considering the grain and stover yields (stover is also of value for fuel, construction, and
basketry in Burkina), and the possibility to have at least some grain the first year, (100-kg/ha
grain yields are not uncommon in the region), the treatments with zai' offer some advantage
during early reclamation. If no fertilizer had been added (often the case in farmers' fields),
the relative fertility contribution of the zai' would probably have been more pronounced.
The combination of grass mulch with zai gave the highest average grain yield (1550 kg/ha in
1992), a truly astonishing level of productivity on this type of land by local standards. In a
drier year, the potential for superior water conservation by this practice (relative to zai
without mulch) might produce greater differences, making it a possible risk-reduction
farming strategy. This advantage is not sustained in all conditions as reclamation progresses,
and did not hold at Yilou in 1993.
There appears to be some merit in scattering zai" as opposed to putting them in lines (all three
years at Yilou, not at Nioniogo), probably due to the elimination of wide, unobstructed inter-
row spaces for surface water flow.
The evolution of the soil surface above the bund included (1) deposition of several alternating
layers of fine and coarse materials near the bund, (2) deposition of coarse materials, including
gravels further away from the bund, and (3) formation of a surface seal on compact, eroded
surfaces farther from the bund. In general, a significant increase in porosity of the surface layer
was observed where significant sedimentation occurred, and the soil was beginning to support
plant life again. Below the bund, zai augmented soil porosity and reduced crusting locally.
Mulching encouraged termite activity, having the same type of effect as zail but to a lesser
degree and over the whole soil surface. Scattered zai increased sedimentation between zai.
Millet production was between 0 and 250 kg/ha the first year in all treatments. In the second
and third years of reclamation, yield relationships among reclamation practices varied, but all
reclamation practices retained a clear advantage over direct sowing. Maximum average annual
treatment grain yield was 1550 kg/ha.
Characterization of the Soil-Plant System in Bush Fields
in the Tropical North Sudanian Region of Burkina Faso
S. J.-B. Taonda, J. Dickey, P. Sedogo, and K. Sanon
The economy of Sahelian countries is mostly based on traditional subsistence agriculture.
There are two kinds of fields in the land management system of Burkina Faso, namely village
and bush fields. Local organic inputs such as crop residues, animal droppings, household
refuse or compost are applied on village fields which are located next to the residential areas.
In rare cases, exogenous inputs such as mineral fertilizer are used there. More and more
frequently, fields are protected from erosion caused by very intense rainfall events by erosion
control practices. Production in village fields represents a minor part (10%) of total production.
Remaining crop production comes from bush fields. However, extensive agriculture based on
nutrient mining is practiced. Fallowing used to be the only form of soil fertility restoration. In
most Sahelian countries, demographic pressure has resulted in the reduction or elimination of
fallow periods. In the absence of viable systems of fertility restoration, this land management
trend has been accompanied by soil degradation.
Burkina Faso's lands are a part of this crisis, which is particularly acute on the Central Plateau,
where 2 to 20% of the population (depending on the region) has permanently emigrated during
the last decade (INSD, 1985). Emigration is directed toward neighboring coastal countries and
especially toward southern and southwestern Burkina Faso. Migrants preferentially settle in,
less populated or newly settled zones where land is available and soils are fertile (Milleville,
In frontier zones, the intensive and extensive land management systems (of village and bush
fields from the Central Plateau respectively) are replicated. Lack of means precludes viable
application of intensive management of bush fields. The fundamental question to be addressed
is the following: How can production be sustained while maintaining or improving land and
soil resources? This question can only be answered by taking into account the socio-economic,
biological, physical, and chemical dimensions of the dynamics of the soil degradation process.
Authors such as Nye and Greenland (1960), Charreau (1972), Siband (1972) and Pichot (1974)
have studied changes in soil physical and chemical properties of cultivated and fallowed fields
in research station and farm environments. This study focuses on on-farm soil degradation in a
frontier zone of Burkina Faso's southern Central Plateau.
Hypotheses which underlie the study are the following:
SA frontier zone presents the opportunity to analyze changes in soil physical and chemical
properties at various stages of field evolution, whereas this process has run its course on
degraded lands, or has not yet begun on untouched natural lands.
*Diverse agricultural practices, which will induce different types of field evolution, may
coexist in a frontier zone.
In order to control variability in the study, only bush fields cultivated with animal traction were
selected. Most farmers in frontier areas use animal traction. Furthermore, intensification of
production, which is becoming more pronounced, must rely on animal traction because labor
availability is limited. Some farmers in the studied village cultivate by hand.
The objective of the first part of this study, which is discussed in this paper, is to characterize
soil physical and chemical property changes during the years after clearing of cultivated soils
on farms in semi-arid zone (700 to 900 mm of rainfall).
II. MATERIALS AND METHODS
Chemical and physical properties of soil were studied on a soil chronosequence.
The Soil "Chronosequence". Thiougou, a frontier village in the Southern part of the Central
Plateau was selected. It is located between 11 29' and 11 24' north latitude and 0O49' and 0*54'
west longitude. A study of village soils was carried out and a physiographic map of the village
Among the village's soils, modal leached tropical ferruginous soils (sols ferrugineux tropicaux
lessives modaux, probably Paleustalfs after U.S.D.A., 1975) were selected for this study due
to their dominance of Thiougou's land area. Moreover, they are the most widely cultivated
soils in the village and the surrounding region. A map of all village fields and fallows was
developed, from which fields belonging to this soil mapping unit could be listed. These fields
were grouped according to duration of cultivation, including fields of ages 0 (fallow), 1 (new
field), 2, 4, 5, 7, 9 and 17years. Several studies were undertaken on this chronosequence,
including: x A study of soil physical and chemical properties on eleven farmer fields described
in this paper x A study of soil productivity on 30 plots which will be reported elsewhere
(Taonda et al., 1995).
Measurement Methods for Physical Properties. Bulk density was measured on undisturbed soil
cores. Textural separates were measured by the Robinson pipet method and texture was
characterized according to the textural triangle. Moisture content at various pF levels was
measured by weight after equilibration on a membrane press. These samples had been sieved to
less than 2 mm diameter and repacked into cylinders. Field moisture content was based on
auger sampling and oven-drying at 105'C.
Roughness of the soil surface was measured using the Guillobez method (1991). In this
method, a one-meter-long frame containing a series of moveable, straight vertical rods which
can be freed to match the irregularities of the terrain is levelled above the highest ground
surface along its length. The standard deviation of prodruding rod lengths as measured from the
base of the frame provides an index of soil surface roughness. Measurements were replicated
at three locations per observation date and in each plot. The first observations were taken
immediately after ridging and subsequently after each rainfall event. Runoff was measured
from 1 m2 plots bordered by vertical tin sheets. Water from the runoff plots was channeled
through a plastic pipe into an open barrel. Water depth in the barrel was measured after every
rainfall event and runoff depth was calculated. Rainfall amount was measured in rain gauges
adjacent to each plot. Runoff water depth was calculated as follows:
(Volume in barrel Volume of rainfall directly into barrel)/1m2
Measurement Methods for Chemical Properties.. Soil pH (in water and IN KC1) was measured
with a glass electrode in a 1:2.5 (by weight) suspension of soil and solution. The pH-meter
has glass electrodes. Total carbon was obtained by the Walkley-Black method (Walkley and
Black, 1934). Nitrogen was measured by the Kjeldahl method (Jackson, 1958). Total
phosphorus was determined after Olsen and Dean (1965). Available P was obtained by the
Bray I method (Dickman and Bray, 1941). Exchangeable bases and CEC were measured
using the IITA Playsier (1978) method, in which exchangeable cations are displaced from
exchange sites with a silver thioure (AgTu) solution. AgTu provides total soil saturation due to
its preferential adsorbtion. Silver content in the extract is determined. The amount of silver
retained in the soil, which indicates CEC, is deduced through comparison with a standard soil
The Evolution of Soil Physical Properties. Table 1 shows that surface horizons in almost all
plots were of sandy-loam texture according to the USDA classification. Plots of ages 0 and 17
years were of loam to sandy loam texture. Bulk density was intermediate in surface horizons,
increasing with depth on all plots.
The decrease in soil surface roughness is the difference between roughness after ridging before
planting and roughness at the end of the season. This difference increased exponentially and
reached a plateau of 15% reduction at around 7 years of field age (Figure 1).
Figure 2 shows that for plots of ages 0 and 17 years, the greater the rainfall event, the greater
the depth of runoff from the plot. This figure also illustrates that the difference in runoff depth
between old and young plots was large for large events. However, old and young fields lost a
relatively similar water volume during small storms. The slopes of the first order polynomials
relating runoff water depth to the size of rainfall event, shown in Figure 2, was positively
related to field age when considered for all of the fields in the study (Figure 3). The slopes
approached a constant value of 0.6 at a field age of 9 years.
Table 1. Physical and chemical soil properties of the chronosequence in Thiougou
Age Depth Silt Silt 20 Sand 50 Sand 250 N Texture Sand Silt Clay ND.A. D.A. Moisture Moisture Moisture Available
to 20 to 50 to 250 to 2mm at at at moisture
PF 2.5 PF 2.5 PF 4.2
(years) (cm) (%) (%) (%) (%) (No.) (USDA) (%) 950 950 (No.) (g/ml) (%) (%) (%) (%)
0 0-20 7 32 14 37 1 L 51 39 10 6 1.56 11.1 6.9 3.9 7.2
20-40 7 28 13 41 1 LS 55 35 11 6 1.55 10.2 7.4 3.9 6.3
40-60 0 5 1.60 *
1 0-20 8 29 16 38 2 LS 53 37 10 3 1.50 12.1 7.0 4.1 8.0
20-40 7 28 16 38 2 LS 54 35 11 3 1.53 11.2 6.8 3.9 7.3
40-60 9 24 12 37 1 LS 49 34 17 3 1.60 14.9 18.0 6.4 8.5
2 0-20 7 27 19 40 2 L 59 34 7 3 1.44 10.3 4.8 2.6 7.7
20-40 8 25 17 40 2 LS 58 33 9 3 1.45 10.0 5.6 3.4 6.6
40-60 8 23 15 390 2 LS 54 31 15 3 1.57 13.4 7.9 5.4 8.0
4 0-20 9 27 18 39 1 LS 57 36 7 3 1.47 9.9 5.7 3.3 6.6
20-40 8 25 16 42 1 LS 58 34 9 3 1.57 11.5 5.8 3.2 8.3
40-60 11 23 13 44 1 LS 57 34 9 0 11.7 5.2 3.1 8.6
5 0-20 7 27 19 39 3 LS 58 34 7 3 1.53 11.8 5.4 3.1 14.5
20-40 8 26 15 41 3 LS 57 34 9 3 1.58 11.3 6.7 3.7 13.8
40-60 9 23 12 38 31 L 50 32 17 0 14.4 8.8 5.7 14.1
7 0-20 10 27 16 39 1 LS 55 37 8 3 1.57 10.5 6.4 4.0 6.5
20-40 9 25 14 41 1 LS 55 34 11 3 1.67 11.3 6.5 3.9 7.4
40-60 9 20 10 39 I IL 49 29 23 0 -16.1 10.7 6.9 9.2
9 0-20 8 21 18 46 1 LS 63 29 8 3 1.50 9.0 5.5 2.7 6.3
20-40 9 20 16 45 LS 61 28 11 3 1.56 11.1 7.7 5.3 5.8
40-60 9 18 10 38 LAS 48 27 25 3 1.61 18.3 14.6 9.2 9.1
17 0-20 8 32 20 31 1 L 51 40 9 3 1.56 9.9 6.2 3.3 32.5
20-40 11 29 17 31 1 L 48 40 13 3 1.56 13.9 7.6 4.6 28.6
40-60 12 25 15 30 1 L 45 37 18 3 1.65 14.9 9.1 6.5 8.5
Table 1 (Cont.)
Age Depth pH water pH KCI Org. C.O. N C/N P total P Ca Mg K Na Bases CEC Sat. Pse
M. total avail. Base
(years) (cm) p(molar) p(molar) (%) (%) (%) (-) (mg/kg) (mg/kg) (meq/ (meq/ (meq/ (meq/ (meq/ (meq/ (%) (%)
100g)' 100g) 100g) 100g) 100g) l00g)
0 0-20 6.7 6.3 1.09 0.63 0.37 17 49 3.0 2.03 0.49 0.15 2.67 3.73 72 *
20-40 5.7 4.8 0.50 0.29 0.17 17 40 0.9 0.68 0.26 0.02 0.96 1.71 56 *
1 0-20 6.3 6.1 0.80 0.47 0.41 13 60 2.4 1.98 0.52 0.16 0.07 2.77 3.98 66 1
20-40 5.4 4.9 0.42 0.24 0.21 12 54 1.1 0.98 0.27 0.08 0.07 1.36 2.10 64 -2
40-60 4.9 4.7 0.43 0:25 0.21 12 56 0.9 1.46 0.38 0.07 0.07 1.98 2.51 79 3
2 0-20 6.2 5.8 0.56 0.32 0.27 12 75 2.0 1.31 0.33 0.14 0.05 1.83 3.36 54 2
20-40 5.3 4.6 0.38 0.22 0.16 14 57 1.1 0.92 0.29 0.07 0.04 1.30 3.22 41 1
40-60 5.2 4.3 0.23 0.16 0.12 13 64 0.8 1.14 0.35 0.10 0.07 1.63 3.52 49 2
4 0-20 5.8 5.8 0.86 0.50 0.15 33 63 0.5 1.80 0.51 0.07 0.04 2.42 3.92 62 I
20-40 4.9 4.7 0.43 0.25 0.16 15 49 1.4 1.76 0.36 0.02 0.04 2.18 4.89 45 1
40-60 4.4 4.2 0.22 0.13 0.13 10 64 0.5 0.96 0.30 0.05 0.04 1.34 1.49 90 3
5 0-20 57 55 0.45 0.26 027 10 66 I 5 1 09 0.32 009 004 1 52 3.46 44 I
20-40 5.0 4.4 0.40 0.23 0.18 49 67 0.9 1.18 0.36 0.08 0.04 1.63 3.52 48 1
40-60 4.9 4.4 0.51 0.32 0.17 22 67 0.6 0.91 0.28 0.07 0.05 1.28 3.12 41 2
7 0-20 5.3 5.4 0.64 0.37 0.29 13 60 1.6 1.01 0.39 0.10 0.04 1.53 1.78 86 2
20-40 4.2 4.1 0.33 0.19 0.17 11 49 0.9 0.98; 0.21 0.05 0.10 1.34 2.67 50 4
40-60 4.0 4.0 0.29 0.17 0.17 10 49 0.3 1.18 0.33 0.07 0.07 1.65 2.92 56 2
9 0-20 6.4 5.7 0.40 0.23 0.22. 10 71 1.4 1.27 0.32 0.12 0.07 1.78 3.64 49 2
20-40 5.7 4.3 0.40 0.23 0.16 15 71 0.9 0.76 0.30 0.12 0.07 1.24 1.78 70 4
40-60 4.8 3.9 0.62 0.36 0.19 19 85 0.5 0.83 0.19 0.15 0.10 1.27 2.37 54 4
17 0-20 6.4 5.7 0.59 0.34 0.29 12 78 2.3 1.49 0.36 0.15 0.07 2.07 4.28 48 2
20-40 5.0 5.0 0.40 0.23 0.26 9 64 0.9 1.57 0.45 0.07 0.07 2.14 3.85 55 2
40-60 6.6 5.1 0.71 0.41 0.16 26, 71 0.7 0.95 0.39 0.05 0.04 1.42 3.46 41 1
Figure 4 shows that water infiltration decreased significantly after first cultivation, and total
seasonal infiltration was drastically reduced at a field age of 17 years. Soil moisture profiles
recorded on several plots during the season (Figure 5) illustrate that surface horizons under
fallow were rapidly saturated at the time of the first measurements early in the season. Fields
cultivated for a long time, such as the 17-year-old field, only reached saturation toward the end
of the season in mid-September. The trend was not clear for fields of intermediate age. Texture
and plant extraction differences among fields may account for this.
15 --- -........ ....-... ...........- --
1 0 ... .. .............. .. ................ ... .......... .......
0 4 8 12 16
Age of field (years)
Figure 1. The decrease in soil
roughness over the season for
fields of various ages.
0 20 40 60
Rainfall event (mm)
- Fallow field
-.--- 17-year-old field
Figure 2. Runoff/rainfall per
storm event for fields of ages
0 and 17 years.
---- -- ---. ------
....... ...... .. ..
June July Aug.
......... ....... ...... .. ...... +.. ...... ............
June July Aug.
I I I
| 0 .5 ........ ..... ... ........................... .................. .......
,5 0 .2 .. ... ...............;..........!............... -- --- -----
- O 0 4 8 12 16
Age of field (years)
Figure 3. The increase of the
slope of the runoff/rainfall
curve (see Figure 2) with
June July Aug.---
Jn Jly u
June July Aug.
Figure 4. Distribution of infiltration and runoff for fields of 0, 1, 9, and 17 year of age.
Soil moisture (g H20/g soil)
3 6 9 12 15
4 0 ....... ........ .. ...... ............
8 0 ...-. ........... ......... .. .............. --..........
1 Fallow field (0 9rs)
E -1 oo
3 6 9 12 15
--- 7 August
....... 17 Sept.
--o pF.2.5 (F.C.)
0 pF 4.2 (P.W.P)
Age of field noted on
graph next to each profile,
all 17 Sept.
3 6 9 12 15
Figure 5. Moisture profiles for fields of various ages.
The Evolution of Soil Chemical Properties. Organic matter in these soils was generally
concentrated in the surface soil. Organic matter content was less than 1% in the surface 20 cm
in all fields sampled, except in the fallow field (1.09%). Organic matter content declined
exponentially (Figure 6), and at 4 to 5 years of age equaled half of the fallow field's level.
Later on, the decline in organic matter content slowed and became somewhat constant at about
Table 1 shows that all field ages under study had a low (<1%) N content. Table 1 also shows
that the evolution of nitrogen content was similar in shape to that the organic matter (Figure 6),
but the exponential model did not explain as much of the observed variability (18% versus
53%). All soils studied were low in total P. No pattern of total P with field age could be
detected (Figure 6). In contrast and more significantly for plant nutrition, soils were generally
deficient in available P (Figure 6). In the 0-20 cm layer, from a value of 3 mg/kg in fallow,
there was a decrease of 50% in the level of available phosphorus after 4 to 5 years of field age
(Figure 6). Thus, the trend of available P with field age was also similar to that for organic
Plots can be grouped in two categories regarding pH, CEC, and exchangeable (Ca, Mg, K,
total) bases (Table 1). The first group includes fallow (0 years) and newly cleared fields (1
year). In this group, there was a drastic difference between the 0-20 cm layer and the 20-60 cm
layers, which matches the pattern for organic matter. The second group contains all the older
fields of the chronosequence. Bases and CEC, on the other hand, were largely similar between
the surface and bottom layers.
The ratio Na: CEC (exchangeable sodium percentage) was an average of 3%. Alkalinity is
therefore not a problem.
0 4 8 12 16
.0 ................. ........... ............
............................ ......................... ......... .
0.20 I i ,
0 4 8 12 16
I I I I
3 ,.'--...-.-..... ,..............- ...-....:......i.......... -- ..
0 4 8 12 16
0 4 8 12 16
Age of field (years)
Figure 6. Change in several chemical properties (organic matter, N, K, and available P).
The Methodology. Several studies have been undertaken to describe the evolution of soil
physical and chemical properties on cultivated, tropical soils. Most were conducted on research
stations. Experimental designs were often difficult to execute, complex, multi-annual, and very
costly. Unfortunately, this high level of investment did not ensure that conditions and results
encountered are those which farmers face. In order to succeed, three challenges had to be
addressed in the on-farm research work: environmental variability, study complexity (and
therefore the cost of execution), and duration of the study. In order to reduce environmental
variability, a single soil mapping unit (modal leached tropical ferruginous soils) and a single
group of cultural practices (farming with animal traction) were considered. In order to
minimize design complexity and cost, field devices were designed to consume recycled
materials (e.g., old tin sheets, plastic pipes, motor oil barrels). The duration of the study, which
would have necessarily been long in order to track individual fields over time, was shortened
through the selection of a "chronosequence", or a series of field ages since clearing.
Nevertheless, the number of plots was limited by two major constraints. First, researchers had
to monitor plot management over a large geographic area Second, observations for a variety of
field ages were conducted on the different fields during a single agricultural season, effectively
concentrating the workload into this season. Despite these constraints, the data analysis
revealed significant trends and results can be compared to those obtained on research stations
and those resulting from monitoring of individual plots over time (Nye and Greenland 1960;
S6dogo, 1993). Therefore, this approach is well suited to studies of soil degradation and could
be employed in other regions.
The Soils. Based on field observation, the soils became progressively sandier as field age
increased. Farmers also call soils on the older fields "bisri", meaning sandy. A textural
analysis of the 0-5 cm layer probably would have revealed this increase in the percentage of
sand, but was not carried out.
The rapid decline of organic matter is consistent with Hien (1990), who found a 60% decrease
of organic matter content after 15 years of continuous cultivation in the more humid cotton-
producing zone of Burkina Faso. The rate of mineralization is remarkably fast during the
weeks after clearing, when 33% of the organic matter is lost. (Soil samples in the fallow 0-
year-old field were taken at the end of the dry season 10 weeks before sampling of the same
and older fields after tillage and planting.) Since part of the fallow was cleared before the first
rains to become the one-year-old field, the fallow and the one-year- old field represent a
monitoring of the same plot over time. The organic matter in the fallow is rapidly mineralized
because of a high level of microbiological activity, which results from the combination of
moisture from the first rains, high soil temperatures (averaging 28"C), relatively coarse soil
textures, and the use of animal traction to mix and aerate the soil to 8 to 12 cm of depth. This
is consistent with S6dogo's (1993) statement that management of sandy soils with no or low
organic amendment eliminates the soil's pool of organic matter through mineralization.
Nitrogen levels are comparable to those reported by S6dogo (1989), Sohoro (1992), and
Bambara (1993). Table 1 shows that N content in the 0-20 soil layer in the fallow (0-year-old)
and new (1-year-old) fields is twice that of the 20-40 cm layer. After two years of cultivation,
this ratio declines to about 1. The same observation is valid for the organic matter content, the
principal source of N in these soils. Thus, rapid mineralization of organic matter in the 0-20 cm
layer after land clearing is concurrent with rapid N loss. This N is either leached, volatilized as
ammonia, denitrified, or taken up by crops and weeds. Since N is not typically added in
organic or mineral form, this loss continues each year.
The trend for available P is very similar. The ratios of available P in the 0-20 cm layer to
available P in the 20-40 cm layer are 3.3, 2.2, 1.7 for the fallow, 1-year-old and 4-year-old
plots respectively (Table 1). The presence and decay of available P in the surface 20 cm is
therefore also correlated with organic matter content. After organic matter is mineralized,
available P content in surface layers approaches levels of available P in deeper layers, where
organic matter, N, and available P are low regardless of field age.
Similar comments can be made for pH, CEC, and exchangeable bases. The high correlation
between these factors and organic matter can be explained by the high percentage of CEC on
the organic matter in these soils (Jones and Wild, 1975).
Organic matter plays an important role in the soil is structural stability. The relatively high
level of organic matter in years 0 and 1, possibly including dead roots and soil flora and fauna,
probably stabilizes structure, maintaining a more constant soil roughness. Consequently, the
decline in soil roughness over the season for young fields was less significant. On older plots,
the degradation of plant and microbial biomass and lower levels of organic matter affect soil
structural stability. Any sort of roughness created by tillage cannot withstand substantial
raindrop impact. This is why initial soil roughness in fields cultivated for over 4 years
decreases by 15% during the agricultural season. The causal relationship between soil
roughness and runoff, whereby soil roughness slows water runoff, is well established (Stallings,
1957). Soil roughness also allows for more infiltration of rainwater, which is then available to
plants. The rapid smoothing of older fields, therefore, leads to more runoff and less infiltration.
To summarize, the change in organic matter content governs the evolutionary trend of soil
chemical and physical characteristics, at least in the 0-20 cm layer. The trend is negative, that
is, it corresponds to a rapid decline in fertility of leached tropical soils. In a span of 10 years
of cultivation, the land is no longer productive. Beyond this time, cereal production and
farmers' livelihoods are jeopardized. Taonda et al. (1995) show that sorghum productivity on
these soils, which was 1300 kg/ha on newly cleared plots, declines to 250 kg/ha in 5 to 6 years,
amounting to an 80% decrease.
Migrants have settled in the frontier zone in search of land with fertile soil, only to find fragile
soils. The vicious cycle of immigration, continuous cultivation, land degradation and
emigration repeats itself in the frontier zone. Should this process continue at its current pace,
the whole region of the southern part of the Central Plateau may lose much of its agricultural
production potential. This would be disastrous, considering the fundamental role this region
plays in the nation's food supply. It is therefore imperative to stop or even reverse this
Any successful soil restoration technology for older fields must include a combination of
improved organic matter management and erosion control practices. Indigenous farmers, aware
of the potential for degradation, have begun fertilizing village fields with various local organic
amendments, such as household refuse, farmyard manure, or compost. In addition, they have
constructed several types of erosion control structures in village fields, such as earth dikes, rock
bunds, and vegetated strips, the latter being the least costly and perhaps the most suitedto the
frontier regions. Bush fields, on the other hand, are cultivated without the benefit of either
organic amendments or erosion control structures. However, Taonda et al. (1995) showed that
the application of a low rate of farmyard manure (2.5 t/ha) on a bush field protected by erosion
control structures increased the productivity of the 17-year- old plot (from this study). It rose
from 13% to about 70% of the sorghum productivity (1300 kg/ha) obtained on 1- and 2-year-
Several constraints restrict restoration of degraded soils. First, individual farmers have to
produce a sufficient quantity of organic amendment. :This, in turn, requires labor, farm animals,
and an integration of livestock and crop production. Second, there are many thousands of
hectares of bush fields. The treatment of whole watersheds with erosion control measures
cannot be achieved at the level of individual farm families. Therefore, construction of such
structures should be handled by whole village communities.
Bambara, D. 1993. Dynamique de la matiere organique selon les systemes de culture dans les
sols agricoles du finage de Thiougou: Approche quantitative. M6moire de Fin d'Etudes
IDR. 100 pages + appendices.
Dickman, S.R., and R.H. Bray. 1941. Replacement of adsorbed phosphate from kaolinite by
fluoride. Soil Sci. 52: 263-273.
Jackson, M.L. 1958. Soil Chemical Analysis. Prentice-Hall, Inc., Englewood Cliffs, N.J.
Jones, M.J. and A. Wild. 1975. Soils of the West African Savanna. Commonwealth Agricultural
Bureau Technical Communication No.55. Farnham Royal, Slough, England.
Nye, P.H. and D.J. Greenland. 1960. The Soil Under ShiftingCultivation. Commonwealth
Agricultural Bureau Technical Communication No.51. Farnham Royal, Slough, England.
Olsen, S.R. and L.A. Dean. 1965. Phosphorus. In: Methods of Soil Analysis, Part 2: Chemical
and Microbiological Properties. American Society of Agronomy. Madison, WI. pp. 1035-
Playsier. 1978. Personal communication.
S6dogo, M. 1993.
Siband, P.1972. Etude de l'6volution des sols sous culture traditionnelle en Haute-Casamance.
Principaux r6sultats. Agron. Trop. 27(5): 574-591.
Stallings, J.H. 1957. Soil Conservation. Prentice-Hall, Inc., Englewood Cliffs, N.J. p.51.
Taonda, S.J.B., J.B. Dickey, and P. S6dogo and K. Sanon. 1995. Characterization of the soil-
plant system in bush fields in the tropical North Sudanian region of Burkina Faso. Part II:
Evolution of field productivity on continuously cultivated soils. In manuscript.
Walkley, A., and I. Black. 1934. An examination of the Degtjareff method for determining soil
organic matter and a proposed modification of the chromic acid titration method. Soil
US Dept. of Agriculture, Soil Conservation Service, Soil Survey Staff. 1975.
Soil Taxonomy: A Basic System of Soil Classification for Making and Interpreting Soil
Surveys. Agricultural Handbook No. 436.
Economics of Rock Bunds, Mulching and Zai
in the Northern Central Plateau of Burkina Faso,
A Preliminary Perspective
D. Kabor6, F. Kambou, J. Dickey, and J. Lowenberg-DeBoer
In response to national policy and farmer requests, the Farming Systems Program (RSP) of the
Institute of Agricultural Research and Studies (INERA) of Burkina Faso is conducting research
on soil conservation and fertility restoration techniques. Soil erosion, compaction, crusting and
declining fertility are common problems in the Central Plateau of Burkina Faso. These soil
problems are particularly acute at the Donsin RSP site in the province of Namentanga, about
103 km northeast of Ouagadougou (but 230 km by road). Rock bunds, zai and mulching are
being tested. The objective of this report is to provide a preliminary perspective on the cost of
implementing these techniques and on yield effects. The data presented here are necessarily
incomplete, but it is hoped that they will help guide future research.
Rock bunds (cordons pierreux) are low rock walls that are laid on the contour. The purpose of
these bunds is to slow runoff, increase infiltration and reduce soil erosion. Rock bunds are a
proven technology and have been constructed in many parts of the central plateau. They are
included in this analysis for the benefit of comparison.
Mulching is a traditional soil conservation and fertility restoration technique in the central
plateau. During the dry season grass cut from uncultivated lands is spread on crop land that
suffers from slow water infiltration and consequently from drought. When rains start, the
mulch reduces runoff and helps hold soil moisture. Termites colonize the mulch and soil
beneath it, increasing porosity and speeding decomposition. Decomposition of the grass adds
to soil fertility.
Zai is an intensive manure management and water conservation technique. During the dry
season a shallow hole is dug for each hill of sorghum or millet that will be planted when rains
start. Each hole is partially filed with manure or compost which is less prone to crusting and
compaction than soil. The holes trap rain water and reduce runoff. The organic matter helps
retain moisture and has a "starter fertilizer" effect on plant growth. Zai has long been practiced
in Yatenga, but is a new technique at Donsin. The primary purpose of this report is to compare
the economics of zai to the better known techniques of building rock bunds and mulching.
The data collection methodology is presented in section 2, a presentation of the data and results
in section 3, discussion in section 4 and researchable issues in section 5.
Primary data was collected in interviews with farmers who use the techniques under study.
This interview data was supplemented by information from the Soil Management and Resource
Conservation Project for the Central Plateau (PATECORE), the Soil and Water Conservation
and Agroforestry Project (CES/AGF) and Foster Parents Plan (PPI).
The farmer interviews were conducted on Sept. 10 and 11, 1993, at Donsin. The interview team
was composed of Daniel Kabore, Economist and Coordinator of the RSP central team, John
Dickey, agronomist on the Agricultural Research and Training Support (ARTS) project
technical assistance team, Salif Boena, RSP technician at Donsin, and J. Lowenberg-DeBoer,
economist and ARTS campus coordinator. The interviews were conducted in Moor6, with
translation by Boena into French for the benefit of Lowenberg-DeBoer. The interviews were
free form. No questionnaire was used because the initial information was inadequate to
formulate appropriate questions.
Three farmers provided most of the information, but 5 other farmers added certain details.
These three farmers were chosen because they employ at least two of the practices studied and
because they cooperated with RSP at Donsin. In addition to the interview, planting density was
measured in the fields of each farmer.
The information from PATECORE was supplied primarily by Robert Delma, head of the
Development and Follow-up Unit, in an interview on Sept. 14, 1993, at the headquarters in
Kongoussi. Frederic Kambou, researcher in the INERA fertilization, soil, water and
mechanization program (ESFIMA), and Lowenberg-DeBoer participated in this discussion. In
addition, PATECORE reports for the years 88 to 92 were consulted. PATECORE is a German
financed project. It does not participate in bund construction directly, but offers technical and
material assistance to government agencies (primarily the Regional Centers for Agricultural
Production (CRPA)) and non- governmental organizations (NGO).
The information from CES/AGF was supplied by Gouyahali Son, project coordinator in
INERA/ESFIMA on Sept. 16, 1993, at the research station at Kamboins6. Kambou and
Lowenberg-DeBoer participated in this discussion. The International Fund for Agricultural
Development (FIDA) finances CES/AGF.
The information from PPI was supplied by their agent at Donsin. In addition, the Sanou and
Ouedraogo report on soil management at Donsin was consulted.
I. RESULTS AND DISCUSSION
Rock Bund Background Many techniques for the construction of rock bunds have been
developed in Burkina Faso. They differ in the way contours are laid, preparation of the soil
where the bunds are to be built, the arrangement of the stones and management of construction
labor. Three basic types outlined by Son are:
I) The three stone system also called the FEER system for the Rural Water and Equipment
Fund (FEER) which did much of the development of this system. In its classic form the FEER
system had 6 steps:
1) Tracing of contours by a topographical team composed of 3 persons using surveying
2) One pass by a tractor and disk plow with soil thrown up slope.
3) Collection of stones by farmers in piles of about one truck load, roughly four cubic meters.
Tools used are pick, digging bar and wheel barrow.
4) Stones are manually loaded on trucks by farmers, hauled to the construction site and
5) Farmers lay the stones on the contour in two layers. On the soil surface two parallel rows
are laid next to each other. A single row of stones is laid in the second layer overlapping
the middle half of each first layer stone. A cross section of the bund would show three
stones (Figure 1).
Cross section oc
Figure 1. Top and cross section views of a rock bund built according to the three stone system.
6) Upslope from the bund farmers pile soil loosened by the plow against the foot of the bund.
This procedure has been modified by some organizations, in particular NGOs, to reduce costs.
NGOs often use water levels made of clear, flexible plastic hose with 2m at each end attached
to graduated poles to trace the contours. They train a few farmers per villages to use;the levels.
These farmers then help their neighbors trace contours. This eliminates the use of trained
surveyors and their travel costs. In addition, the NGOs often eliminate the preparation of soil
where the bund is to be built and manually earthing-up the upslope side of the bund.
According to Son the disadvantages of the three stone system are:
1) large holes between the stones allow a substantial flow of water between the stones, thus
permitting some soil to be transported beyond the bund.
2) if the stones are badly aligned, channels can be created that turn into ravines.
3) soil deposits between the two rows of stones can reduce infiltration and thus cause flooding
upslope from the bund.
Soil conservation efforts in Burkina Faso have largely switched from earth bunds to rock bunds.
Initial work in the late 1970s focused on earth bunds. The soils of the Central Plateau often
have slow rates of water infiltration. Thus water was retained behind the bunds and crops
upslope were flooded. In some cases, farmers intentionally opened the bunds to save their
crops from flooding. In other cases the water overtopped the bunds. In either case gullies were
created. Properly constructed rock bunds allow water to pass, but in a diffuse manner which
does not create gullies.
Earth bunds are still used in some areas where stone is not available within a reasonable
hauling distance (about 15 km) to the fields. In those cases, there is an effort to include a
section of rock every 30 to 50 meters, to allow filtration.
HI) Erect stones with subsoiling (PDS) This system was developed by CES/AGF in response
to problems identified with the three stone system. The procedure follows 6 steps:
1) Tracing of contours by a topographical team composed of 3 persons using surveying
2) Two passes by a tractor and two shank subsoiler. Three furrows are created because one
subsoiler shank passes through the middle furrow twice.
3) Collection of stones by farmers in piles of about one truck load, about 4 cubic meters. Tools
used are picks, digging bars and wheel barrows.
4) Stones are hand loaded on trucks by farmers, hauled to the construction site and dumped.
5) Stones are laid by farmers. Large stones are set upright in the middle furrow with the
smallest end down and the flattest side upslope. Small stones are fitted between the erect
stones from the upslope side so that water forces the small stones into the gaps between the
upright stones. If the small stones were placed below the erect stones, water would tend to
scatter them. Downslope of the erect stones, large stones are laid flat at the foot of the bund
to help hold the stones erect (Figure 2).
Figure 2. Top and cross section views of a rock bund build according to the PDS system.
According to Son, advantages of the PDS method include permeability maintained over a
longer period, improved establishment of wild grasses and lower stone requirement. The classic
three stone method requires 20 to 22 cubic meters of stone per hectare. The PDS method
requires about 12 cubic meters per hectare. This is important because wide spread installation
of rock bunds is exhausting stone supplies close to cultivated areas. According to Son, hauling
distances were commonly 5 to 7 km. for the CES/AGF in 1989.
IV) Aligned stones The simplest rock bund system consists of large stones aligned along the
contour (Figure 3). Small stones are often placed on the upslope in the gaps between the large
Figure 3. Top view of a rock bond built according to the aligned stones system.
Of the three systems outlined, the aligned stones bund requires the least stone, but it also allows
the greatest runoff. Stones are vulnerable to being dislodged by water or animal traffic.
Any of the rock bund systems can be built by community or individual action. Burkinab6
government agencies tend to favor community action. The CES/AGF and the CRPAs usually
work with communities to protect a designated slope or watershed. Relatively large groups of
village residents are mobilized on designated work days.
The government approach may include protection of bushlands upslope from fields. According
to Son, the benefits of anti-erosion work in upslope bushlands include protection of lower
toposequence structures and pasturage improvement. Heavy runoff from unprotected upper
toposequence areas can overwhelm lower slope bunds. In addition, improved water infiltration
in the bushlands provides moisture for plant growth and can help replenish ground water.
Many NGOs work with individual farmers to control erosion on their fields. They often
concentrate on lower toposequence (and thus more productive) fields.
Rock bund labor times Labor times were collected for bund construction by an individual
farmer at Donsin and for PATECORE community work during the 1993 season (Tables 1 and
2). It should be noted that these times relate to specific construction approaches. Labor times
for other systems may differ. The Donsin farmer used a modified three stone system without
preparation of the soil. Most PATECORE sites use a modified three stone system without soil
preparation, but a few use the PDS method.
Table 1. Donsin Farmer Labor Times for Construction of Rock Bunds, hr/ha.
Item Men Women Children Total
Gathering stones 25 25
Hauling stones 18 ,18
Tracing contours 4 4
Laying stones 13 38 50
Total labor 60 38 97
Table 2. PATECORE Labor Times for Construction of Rock Bunds, hr/ha.
Item Men Women Children Total
Labor time ,_
Gathering stones 70 56 47 173
Loading trucks 70 56 47 173
Tracing contours 8 8
Laying stones 130 104 86 320
Total labor 278 216 180 673
According to Delma, the average PATECORE bund density is 500 m/ha, which translates to
bunds spaced about 20 m apart. At the time of the writing of this report the bund density for
the Donsin farmer was not available. The farmer labor times assume the 500 m/ha.
Rock bund construction on a large scale is a relatively new development in Donsin, thus rock is
available relatively close to fields. The farmer in Table 1 benefited from the availability of
stone next to his field. Some stone was broken with a digging bar. He hauled the stone with a
The family labor available in the farmer's case was three men and six women. Men gathered
and hauled stones. The women helped in laying the stones. The work was done during the hot
season. Because of the heat, the usual work day was from 6 a.m. to about 9 or 10 a.m. Thus
the 60 hours of work by the three men represents about one week of activity.
The farmer received the help of a PPI trained neighbor in tracing the contours. A mason's level
supplied by PPI was used.
The PATECORE community work (Table 2) was usually in about 7 hour days. According to
Delma, a typical workday would start at 8 a.m. and continue until 4 or 5 p.m., with a break a
midday. Their construction season starts in October and continues until therains come (usually
in June), but because of the heat and fasting during Ramadan, it becomes increasingly difficult
to organize workdays after January. The seven hour workday is important for efficiency in use
of the truck and project personnel time.
According to Delma, the average workforce composition in 1993 was 30 men, 24 women and
20 children. Women helped in all activities except tracing the contours. The work time of truck,
drivers and project personnel are not included in Table 2.
The labor times estimated by Sanou and Ouedraogo for PPI activities at Donsin are
intermediate between the farmer times and PATECORE averages. Their total is about 432 hr/ha
(assuming 8 hour workdays), with 240 hr/ha for gathering stones, 96 hr/ha for laying out the
contours and 96 hr/ha for laying the stones. The PPI activities at Donsin are concentrated on
individual farmer fields, but some coordination at the community level is needed to insure
efficient use of the truck. The truck is sent to the village only if a sufficient amount of stone
has been collected for a day's hauling.
The value of labor used for rock bund construction depends on other labor opportunities
available. Because construction occurs in the dry season, agricultural labor rates are not
relevant. Dry season wage labor is rare in rural areas and almost nonexistent for women and
children. Study of nonagricultural activities at Donsin (Lowenberg-DeBoer, 1993) suggests that
the opportunity cost of labor is 50 FCFA/hr or less during the dry season. At 50 FCFA/hr the
opportunity cost of labor for the work time estimates are: the farmer in Table 1, 4,850
FCFA/ha; PPI labor times estimated by Sanou and Ouedrago, 21,600 FCFA/ha; and
PATECORE average labor times, 33,650 FCFA/ha.
Labor time Differences A comparison of Tables 1 and 2 indicates that PATECORE labor times
are substantially higher than those of the Donsin farmer. There are several hypotheses which
might explain the labor time difference:
1) The amount of stone used by the farmer was less than the 20 to 22 cubic meters/ha used by
2) The individual farm family was more motivated than the workers mobilized in a community
workday. The farm family benefits directly if their soil is improved. The community work
benefits some people only indirectly.
3) Because the farmer did not use a truck, he was able to concentrate the heavy work of
gathering, moving and laying rock in the relatively cool mornings.
It is hypothesized that the labor times collected by Sanou and Ouedraogo are intermediate
between the two other cases because individual fields were being protected, but truck use
efficiency forced some changes in the work schedule. The key changes are a latter starting time
and work in the afternoons. Because of travel time to the work site, work days for truck
hauling start about 8:00 a.m. Farmers often start work at dawn (about 6:00 a.m.). Efficiency in
truck use requires work to continue into the afternoons in spite of the heat.
Hauling Cost The truck use costs reported by PPI and PATECORE (Table 3) are comparable.
Both projects rent trucks. According to Delma, PATECORE also has some of its own trucks,
but finds rental less expensive. Diesel fuel cost is 245 FCFA/1.
Table 3. Truck Cost for PATECORE and PPI.
Item PATECORE PPI
Area treated per day, ha 2.5 2.5
Rental, FCFA/day 27500 27500
Fuel, FCFA/day 18865 19600
Cost per day, FCFA 46365 47100
Cost per hectare, FCFA 18546 18840
The importance of using trucks for full days can be seen in comparing per hectare hauling cost
for full days (2.5 ha/day) and half days (1.25 ha/day). For example, the PATECORE hauling
cost per hectare rises from 18,546 FCFA/day to 32,976 FCFA with half day work, a 78%
increase. Fuel cost are reduced by half day work, but travel to and from the work site is
Rock hauling with animal or human powered carts has been tried with limited success. The
volume of rock that can be hauled in a cart is small and hauling time is long for the 5 to 10 km
distances often involved.
Rock Bund Subsidy Most rock bund construction in Burkina Faso is subsidized by donors.
The case of the farmer in Table 1 is unsubsidized, except for the water level and the PPI training
of the neighbor in tracing contours. The farmer's case is useful as a pure case of estimating
labor times, but it is relatively rare because stone is usually not available close to fields.
Hauling costs are an important part of the subsidy. The PATECORE and PPI data indicate that
this subsidy is about 19,000 FCFA/ha.
Other subsidies come in the form of project personnel time and small equipment. PATECORE
provides about 200,000 FCFA per village in the form of small tools: shovels, picks, digging
bars, wheel barrows. The bund construction systems which use soil preparation on the bund
construction site have another subsidized element. Many of the projects involved in rock bund
construction are multifaceted. Separation of rock bund costs from other activities is not easy.
Mulching Most farmers at Donsin mulch at least a small area of cereal crops with grass
collected from uncultivated fields. Mature grass is collected in the dry season with a rake. In
most cases, a sickle is not used to cut the grass because it requires bending over in the dust
raised by the cutting. Use of a rake allows the individual to stand erect and above much of the
dust. The rake breaks the brittle grass stems, leaving the root crowns in place.
The grass is collected in bundles and transported to the field. The farmers all mentioned the
competition for grass close to the fields. Early in the dry season farmers collect the grass in
bundles, leaving the bundles in the bush for later transport. Collecting the grass in bundles
establishes an individual's claim to the grass. Later in the season bundles are collected and
transported in a single activity.
Near the end of the dry season, farmers walk up to 2 hours from the village to find mulch grass.
Farmers state that their interest in za' is partially due to the fact that mulch sources have been
The experience of Farmer 2 (Table 4) suggests that availability of a cart can substantially
reduce labor times, though interpretation needs to be made with care because of differences in
coverage per bundle.
The density of mulch application appears to vary from farmer to farmer (Table 4). Because the
interviews took place in the rainy season it was not possible to weigh bundles. Bundle size may
be a factor in the coverage area differences. Farmer 2 uses two coverage rates, one for average
soils and one for small depressions and other more humid spots. For average soils the mulch
application rate varies from 15 to 44 m2 per bundle (Table 4). The rate for humid soils is much
lower, one bundle covering 77 m2.
Table 4. Mulching Experience at Donsin, 1993.
Item Farmer 1 Farmer 2 Farmer 3
Density, hill/ha 30769 37104 26667
Area per bundle, m sq 15 44 30
Mulch, bundles/ha 686 227 333
Transport Carry Cart Carry
by bundle 12 bundles by bundle
Max Walking Time, min: 120 120 4*
Dry season labor time
Gather mulch, hr/ha 69 166 56
Transport, hr/ha 412 Combined Combined
Distribution, hr/ha 137 35 ** 37
Dry season total 618 201 93
Rainy season labor time relative to unmulched, flat planting
Planting x2 x3 x3
1st Weeding x3 x5 x3
2nd Weeding ? ? x2
Expected yield effect x2 ? xl.75
relative to unmulched,
Estimated based on farmer statement that mulch was cut close to field and by subtracting the gathering time of other
respondents from total time given.
** Farmer unable to quantify time. Average time per bundle of other respondents used to complete labor estimate.
Planting density in mulch appears to be the same as in unmulched crops (non-zai), planting
time is increased because mulch must be moved aside to make the holes for planting. Farmers
commented that all mulch must be kept from the planting hole to avoid termite damage to seed.
First weeding work time is from 3 to 5 times higher in mulched than in unmulched fields
because mulch must be moved aside to hoe the soil. Second weeding work time is also
Farmers expect yields on mulched soil to be approximately double those on unmulched fields.
Za' Many farmers in Donsin used the zai technique for the first time during the 1993 rainy
season. Some farmers had seen zai on a trip to Yatenga organized by RSP. Many of them had
seen the single field of zai in their area in 1992, near Bonem (about 15 km).
The willingness to try zai should be seen in the context of rapidly declining soil productivity at
Donsin. The most extreme examples of erosion are found on the zipel6, barren areas of baked
subsoil on which even the hardiest plants have trouble gaining a foothold. Based on aerial
photos, Sanou and Ouedraogo (1993) estimate that between 1990 and 1992, the zipel6 area at
Donsin grew by 1% annually.
But declining soil productivity by itself is not enough to convince farmers to adopt a new
practice. For example, farmers in the Donsin area have not adopted chemical fertilizers and
tied ridges, though these practices have been shown to improve yields. The innovations must
use available land, labor and capital more efficiently than alternative practices.
PATECORE has tried since 1990 to introduce zai as an accompanying practice in fields
protected by rock bunds. Relatively few farmers have adopted the practice in the areas where
PATECORE works. Delmar indicated that the low adoption rate could be linked to the level of
extension effort. Another hypothesis is that farmers are not convinced that zai is a good use of
their limited resources.
Zai production practices varied from farmer to farmer (Table 5). Farmers reported trying
several variations of za'i, consciously experimenting in an effort to find the right technique for
Table 5. Zai Experience at Donsin, 1993.
Item Farmer 1 Farmer 2 Farmer 3
Density regular, hill/ha 30769 37104 26667
Density zai hills/ha 25641 23050 31888
Manure per hole, kg 0.6 0.6 0.6
Manure per ha, kg 15385 13830 19133
Transport Type Bicycle Donkey Cart Bicycle
Dry season labor time
Dig holes, hr/ha 206 162 191
Transport Manure, hr/ha 1923 869 3587
Distribute manure, hr/ha 427 Combined 266
Dry season total 2556 1031 4044
Rainy season labor time relative to unmulched, flat planting
Planting slower faster slower
1st Weeding faster x3 x4
2nd Weeding ? faster x0.66
Zai seed placement in soil in manure in soil
Zai reseeding relative to non- reduced none less than
zal if planted in zai non-zai
Expected yield effect with x2 no effect x3
manure broadcast relative to
unmulched, flat planting
* Farmer unable to quantify time. Figure based on average time per hill of other respondents.
their situation. The variations included differences in hole size and shape, the amount and type
of manure, density and seed placement. The variations were based on information that they had
gathered in Yatenga, what they had seen in the field near Bolza and their understanding of
Holes for zai may be round or rectangular. Farmer 1 dug holes that were roughly the shape of
an inverted cone about 25 cm in diameter at the surface and 10 cm deep in the center. Dirt
from the hole was place just down slope from the hole. Farmer 3 used a similar technique, but
with slightly larger holes, about 36 cm in diameter. Farmer 2 used a roughly wedge shaped,
hole, about 17 cm long across the slope, 20 cm in the down slope direction and nine cm deep
along the down slope edge. Farmer 2 piled the dirt from the holes in the space between two
row of holes to form a low earth bund (Figure 4):
dirt piled between
Figure 4. Top view of zai arrangement used by farmer 2.
Factors in the choice of hole size and shape seem to be the effort required to dig it and the
ability of the hole to trap rain water.
All the farmers interviewed said that they thought a double handful of manure (about 0.6 kg)
per hole was best. Farmer 3 had run short of manure and had used a single handful in some zai.
Farmer 2 had intentionally left some holes without manure to determine the relative impacts of
the soil preparation and the organic matter. At the time of the interview sorghum planted by
Farmer 2 in zai was uniformly about 3 m tall and forming grain. Sorghum planted in zai holes
without manure was about 1.5 m tall, somewhat uneven and only beginning to flower. In the
flat planted fields sorghum was highly uneven, with the tallest plants about 1.5 m and
Estimates of total manure use ranged from 14,000 to 19,000 kg/ha. Mark Powell, International
Livestock Center for Africa (ILCA), has conducted research on manure production by corralled
livestock grazing bush and crop residue in Niger. He estimates that the average daily
production of fecal dry matter per animal is: cattle, 1.28 kg; sheep, 0.37 kg; and goats, 0.2 kg
(Personal communication). Using Powell's estimates, the midpoint application of 16,500 kg/ha
would represent the full year's production (365 days) for 35 cattle, 122 sheep or 226 goats. Few
farms at Donsin would have enough livestock to do zai on one hectare with pure manure. These
estimates suggest that lack of manure would limit the zai area per farm to less than one hectare,
unless ways are discovered to increase manure supply or extend it by using other organic matter
sources. for example by composting.
Farmer 2 had a compost pit (4mx2mxlm) near his compound. He said that it required about 56
hours to dig. He composted livestock manure, wild grass (like that used for mulch), compound
sweepings and wood ashes. Keeping the compost pit moist requires a major commitment of
labor. He said that pumping and hauling twelve 50 liter containers each day required about 3
hours per day (1 hour by three men using a donkey cart) from January to May. Farmer 2 had
learned about composting during a visit to Zoundweogo Province organized by RSP.
Plant density measured in the fields showed that farmers 1 and 2 reduced plant population in
zai. Farmer 3 increased population slightly. The amount of labor per zai hole is probably about
constant, thus one way to reduce per hectare labor requirements is to reduce density.
Farmer opinion was divided on the best seed placement. Farmers 1 and 3 hoed through the
manure to plant the seed in the soil. The seed was covered with a mixture of soil and manure.
They said that seed planted in the manure was subject to insect damage. Farmer 2 planted
directly in the manure. He said that soil covering the seed tended to crust and to interfere with
In general, farmers said that seedling survival was improved by zai. In the 1993 season, many
flat planted fields were replanted 3 times. Given the,short growing season, reseeding has a
major impact on yield potential. Second and third plantings are often not sufficiently mature at
the end of the rains to make grain. Zai appears to reduce the need for reseeding and to have
major implications for yield distributions, and hence the risk characteristics of the cropping
Za' requires large amounts of dry season labor. Estimates based on the interviews range from
1000 to 4000 hours per hectare. Digging requires substantial labor, but transporting manure is
the major element in the labor time. Farmer 2 was able to reduce his labor time by using a
donkey cart and because the corral from which the manure was hauled was near the field.
Manure transported by bicycle is put in a sack.
Farmers are divided on the impact of zai on planting time. Part of the disagreement may be
related to the lack of experience with the new technique. Two of the three respondents said that
planting in zai as slower. Farmer 1 said that in zai he could only plant one row at a time,
compared to 3 rows in flat planting. Farmer 2 said that planting in za'iwas easier because the
soil was already prepared.
They all said that the first weeding in zai required extra time, mainly because weeds grew
vigorously in the zai hole and had to pulled out by hand for fear of damaging the young
sorghum plants. The second weeding was faster in zai because the sorghum grew quickly and
Because this is the first year of zai in the village, the farmers were not able to provide yield
estimates. Their estimates of yield effects of manure broadcast on flat planted fields vary from
no effect to 3 times as high as yields from unmanured fields. Farmer 2 remarked that heavy
rains often washed away manure spread on flat planted fields.
Zai also has potential as a soil reclamation technique for zipel6 (denuded lands). RSP had a
demonstration plot in the center of a zipel6 area. Only a few farmers have used zai on zipel6
and then only next to existing fields to reduce damage from wandering livestock. Most Donsin
farmers have applied zai to improve productivity on existing fields.
The economic reasons for farmer interest in zai as a soil management technique at Donsin are
1) higher yields with Zai.
2) potential for risk reduction zai appears to favor early establishment and reduce the need to
repeatedly reseed crops.
3) low capital investment in general thereis no initial cash outlay. Tools already available
on the farm can be used initially. Experience on Yatenga indicates that farmers eventually
invest on sturdier pick axes for digging.
4) most of the labor occurs in the dry season alternative labor opportunities in the area are
scarce during this period.
Farmers are interested in zai for many of the same reasons that they have already adopted
mulching. They say that they would expand mulch area if mulch materials were available. Two
hours walking distance is about the practical limit of mulch collection. Beyond that travel
times become too long. They recognize that zai and mulching could be complementary
practices. Farmer 2 planted a small area of combined mulch and zai. But making the most of
limited mulch and manure will probably force farmers to apply either one or the other on a
Rock bunds, mulching and zai are complementary practices. Bunds are somewhat effective in
keeping soil in place and increasing infiltration, but without accompanying improvements in
soil fertility and physical properties, the impact on yield is limited about 10% to 30%.
Combination of these practices can increase yields by up to 300%.
Rock for bund construction is becoming a scarce resource. The use of trucks has increased the
practical hauling distance for rock, but it appears to force some inefficiencies in the use of
labor. Other strategies for increasing the supply of usable rock include the use of explosives.
Zai has a much more immediate effect on yields than rock bund construction. The main effect
of rock bund construction in the first year is reducing rainfall runoff and increased infiltration.
Yield effects from the build up of sediment behind the bund show up in the second or third
year. Zai also has a carry over effect. The manure does not decompose entirely the first year
and crop residues produced in the first year can act as mulch if left on the field. Soil physical
and chemical properties may be improved for several years after application. Farmers will
probably rotate zai over their fields.
Rock bunds have a long, but probably limited useful life. Their ability to hold soil and water
may be reduced if sediment deposits behind the bunds build up near the level of the top of the
bund though the terrace thus farmed is less erodible than the original slope. Clay deposits
between the rocks on some types of bunds (Three stone system) may reduce water filtration
through the bund and create flooding. The growth of grasses and trees along the bund may turn
it into a vegetative strip and effectively lengthen its useful life.
Both zai and the construction of rock bunds use large amounts of dry season labor. The farmer
may be faced with a difficult choice between the two techniques. In general, the labor
requirement of bund construction is relatively low compared to zai. But because zai benefits
are more immediate, they may be much more important to the farmer. However, the possibility
of covering relatively more area with rock bunds given the same labor may continue to make
bunds an attractive investment.
At 50 FCFA/hr, the midpoint value of dry season labor time for zai, the required 2500 hr/ha, is
valued at 125,000 FCFA/ha. A crucial parameter in the decision to use zai is the opportunity
cost of labor. A farmer who could earn 50 FCFA/hr in another dry season activity may be hard
pressed to justify zai. At a harvest time price of 50 FCFA/kg of grain, a yield increase of 2.5
metric tons per hectare would be required to breakeven. In addition, dry season earnings from
commerce or artisanal activities would probably be less risky and provide a more even
cashflow than zai. Therefore, to the extent that other renumerative activities are available, they
may be preferred, but zai offers a potentially profitable use of surplus labor.
Constraints The primary constraints to zai use are the supply organic material and labor. The
manure available on most Donsin farms would be enough for only a fraction of a hectare of zai.
There is already evidence of a limiting supply. Farmer 2 said that several people have already
asked for a sack of manure. He commented that before the 1994 crop season people will start
to steal manure. Farmers commented that they were beginning to manage their animals to
increase the amount of recovered manure. For instance, animals were being tethered at night
during the dry season, instead of being allowed to roam at will.
Use of compost may help increase the organic matter supply for zal. All types of organic
materials can be composted, including crop residues. The evidence gathered here suggests that
water hauling is a major constraint to use of compost pits in the Donsin area.
The increased value of manure in zai may have wide ranging effects on the whole farming
system. Livestock production, especially small ruminants, is a well developed enterprise in the
Donsin area. Many of the Donsin animals supply the Ouagadougou market. Increasing the
value of manure may make livestock production more profitable. But more ruminant livestock
means more forage. RSP tried dual purpose (grain/forage) cowpeas in Donsin in 1992, but
farmers favored varieties bred for grain. If zai can stabilize cereal yields and increase the
profitability of livestock production, the forage question may have to be reconsidered.
In Yatenga, where there is a longer history of zai, manure has been extended in some cases by
reducing the amount per hole and adding a small amount of commercial fertilizer. This retains
the physical benefits of organic matter application (for instance in improving water holding
capacity and reduced crusting around the young plant), but allows part of plant nutrition
function to be taken over by mineral based products. The disadvantage of commercial fertilizer
is the capital required for purchasing the product.
The combination of manure (or compost) and rock phosphate has some advantages. The acid in
the manure increases the solubility of the phosphate. In the manure mix, the dustiness and
handling problems associated with rock phosphate are reduced. As with commercial fertilizer,
the capital required for rock phosphate purchase is a constraint.
For farms without animal traction and a cart, labor may be more limiting than manure in some
cases. If zai work is done during the period from January to May period, 6 days per week, with
a work day of 6 to 10 a.m. because of the heat, the total work time per person is 512 hrs. The
farmers interviewed said that most of the zai work was done by men. Women and children
helped mainly in distribution the manure to the zai holes. If women and children did all of the
manure distribution, a farm with one man and no cart could have from 0.14 to 0.25 hectare of
zai (using labor times from farmers 1 and 3). In a similar circumstance, a farm with a cart
(and/or a corral close to the field) could provide the labor for 0.75 hectare of zai.
In the current system, rainy season labor is fully utilized, especially during first weeding. The
higher labor requirements for zai implies that something else will go undone. The lower
replanting rate for zai may offset some of the higher first weeding time. Composting may also
have an impact on first weeding, because weed seeds are killed by the heat in properly prepared
V. RESEARCHABLE ISSUES
Farmers and researchers have much to learn about zai and how it fits into the farming system of
the central plateau. Most of the issues involve both agronomic and economic factors (Table 6).
Table 6. Economic Research Issues Related to Zai'
Problem Economic Aspect*
Plant density Labor time
Zai hole, shape and depth Labor time
Choice between zai and rock bunds Labor time
Amount of manure per hole Link to capital investment in livestock
production, also transport time
Benefits of manure vs compost Labor time for composting, water hauling,
Mechanization of zai including soil Labor time and capital investment in
preparation and transporting manure equipment and draft animals
Animal production techniques that Labor for extra livestock handling, capital
increase usable manure for corrals, etc.
Forage production Profitability of livestock
Composting techniques that reduce water use Labor time, capital investments needed
Manure extending techniques Labor time and capital
Manure fertilizer mixes Capital and/or cashflow
All of the research issues here potentially affect the distribution of yields and profits, and hence risk
For example, the optimal plant density is a function of both plant response and labor
availability. The appropriate quantity of organic matter depends on soil characteristics, plant
nutrition, labor times, capital costs and the profitability of livestock production.
This is a case where a systems perspective is essential. The current system is geared toward
grain production for home consumption. Cash inputs are limited because the opportunity costs
of capital are high and because the crops do not usually produce a cashflow to help pay for the
purchases. Livestock are raised extensively and only a small proportion of animal waste is
recovered for use as a soil amendment.
Zai' has the potential for creating a chain reaction within this extensive system. If zai improves
yields and increases yield stability, farmers can starting thinking about other crop products,
such as forage. The need for manure motivates farmers to confine livestock, but confined
livestock need to be fed, increasing the demand for forage. If livestock production is made
more profitable because of its link to zai, cashflow generated by animal sales might be used to
pay for fertilizer to extend manure availability. The interest in composting to extend manure
supply and reduce weed pressure is also likely to increase.
Research requires foresight. RSP needs to be ready not only to address current problems, but
also to be preparing itself for the problems that farmers will have in a few years. Research that
addresses current problems will always be behind. In the case of zai a forward looking research
strategy requires information on the direction of system evolution. The simplest tools for
analyzing changes in a cropping system are farmer interviews and representative farm linear
Lowenberg-DeBoer, J., Daniel Kabor6 and Souleymane Ouedraogo, "Rapport de Mission A
Donsin, 10 Avril 1993," Projet ARTS, Universit6 Purdue, West Lafayette, Indiana.
PATECORE, "Rapport d'Activit6s Campagne 1988-1989," Kongoussi, Burkina Faso, Nov.
PATECORE, "Rapport d'Activit6s Campagne 1989-1990," Kongoussi, Burkina Faso, Nov.
PATECORE, "Rapport d'Activit6s Campagne 1990-1991," Kongoussi, Burkina Faso, Dec.
PATECORE, "Rapport d'Activit6s Campagne 1991-1992," Kongoussi, Burkina Faso, Jan.
Sanou, Patrice and Alfred Ouedraogo, "Les strategies de conservation du milieu", INERA/RSP,
The Value of Research on Indigenous Knowledge:
Preliminary Evidence from the Case of Zai in Burkina Faso
M. Bertelsen and S. Ouedraogo
Indigenous knowledge (IK) refers to the institutionalized but usually informal practice or practices
which have evolved within a community in response to a particular problem or constraint. In the
context of developing-country agriculture, whether IK is or is not anything more that just a new
name for farmer practice (Erinle, 1993), it is clear that IK has achieved scientific respectability1.
There are good, practical reasons for this. IK practices are virtually always more "low-tech"
using relatively plentiful and cheap labor more intensively than relatively scarce and expensive
capital. Also, IK practices have evolved over many years and are therefore implicitly
"sustainable": they are well adapted to the long-run local social, economic and environmental
conditions. For this reason IK has attracted much attention in the natural resource management
literature (Erinle 1994; Ou6danou 1994).
Two aspects make the study of IK especially interesting in a developing-country agriculture
context. First of all, IK practices can be very localized reflecting evolved solutions to constraints
imposed by conditions in a small area. Due to the lack of infrastructure and other factors which
hinder information flows, the practices may not be adopted even in similar, nearby areas.
Secondly, because IK practices were developed by farmers, they inspire confidence in other
farmers and are therefore relatively easy to extend. Thus, the study of IK practices may provide
large potential net social benefits: IK practices can have much to offer and can be extended
relatively easily without incurring large costs.
The great majority of the scientific effort involving IK has so far been concentrated on the
documentation of practices (see for example, Akinlosotu 1993, Ojeniyi 1993, Obi 1994, Maigida
1994). This necessary first step has been followed by attempts toadapt IK practices to new
technologies with the intention of increasing production and returns to farmers (see for example,
Otegbeye 1994, Gefu 1993). Until now, little effort has been directed toward estimating the value
of IK practices themselves and, implicitly, the research which has brought these practices to light.
This is what we propose.to do in this paper; evaluate the potential impact of a particular IK
practice in Burkina Faso with emphasis on the social value of the research which has made
broader adoption possible.
Background. Zat is an agricultural practice which evolved in the central Sahelian Region of West
Africa sometime during the first half of the 20th century., While its exact origins are unknown, it
1Theodore Schultz (Schultz, 1969) first pointed out the relative efficiency of peasant agriculture. According to
Schultz, the peasant farmer was '... relatively efficient in using the factors of production at its disposal" and that the
key variable in explaining the differences in agricultural production was the level of acquired capabilities of farm
first appeared in the Yatenga Province of Northern Burkina Faso before 1950. Zaf is a technique
for improving the viability of sorghum and millet plantings. It involves the preparation of a small
hole in the ground of about 25 30 cm in diameter and 10 cm deep (Kabor6 et al., 1993). Into
this hole approximately 0.2 0.3 kg of manure or compost is placed before planting. Planting
occurs in the hole after the first viable rains of the season. The effects of both the hole, which
serves as a small reservoir for the plants, and the organic matter apparently combine to enable Zai
plantings to vigorously establish themselves. As a result, re-plantings and the consequences of
early, mini-droughts are often avoided. Increases in yield are dramatic. And although Zai
demands the use of a great deal of additional labor, these demands occur during the dry season
when the opportunity cost of labor is low relative to other times of the year.
Although Zai has been practiced in Yatenga Province since at least 1950, the practice has not
spread beyond this limited area. Farming Systems Researchers (RSP) of the BurkinabB National
Agricultural Research Institute (INERA) became interested in the practice and decided to test the
innovation in the similar RSP village site ofDonsin. RSP transported a small group of Donsin
farmers to Yatenga Province early in 1993 to see and discuss Zai with local farmers. Upon their
return, this group very effectively extended the innovation to their neighbors. Approximate 70%
of Donsin farmers tried Zai during the 1993 growing season (Robins et. al. 1994). RSP
monitored the results of a random sample of these farmers during the season. RSP also
established a Zai demonstration plot on highly degraded land to demonstrate the potential of this
technique to reclaim land.
The methodology adopted for this study follows closely that developed by Akino and Hayami
(1975). In the following figure, the adoption of an innovation causes the supply curve to shift out
from aoSo to a'S1. As a result of this supply shift along the same demand curve Do, changes in
social welfare occur. These changes are measured by changes in 'economic surplus'. Economic
surplus is composed of two elements; consumers' surplus and producers' surplus. Consumer's
surplus arises from the fact that a consumer pays only the marginal value of the last product sold
in the market (market price) even though it is known from the law of diminishing marginal utility
that previous units purchased were worth more (see any introductory economics text such as
Samuelson, pp 448-449). Thus, at the initial market price of Po, consumers' surplus is equal to the
area of triangle P0dp in Figure 1. Similarly, producers'surplus arises from the fact that a producer
receives the marginal cost (market price) for the last unit produced despite the fact that earlier
units cost less to produce (law of diminishing returns). This corresponds to the area of triangle
Popa0 in Figure 1.
Figure 1. Ex-Ante Analysis of Impact
The total net change in economic surplus occasioned by the change in supply is represented in the
figure by the lightly shaded polygon. This area represents the total, ex-ante increase in social
benefits (without partitioning between consumers and producers) and is equivalent to the
theoretical measure described by Akinoand Hayami.
Data. In order to estimate these social benefits, a great deal of data must be obtained.
Specifically, data on price indices, market prices, quantities and estimates of the elasticities of
demand and supply must be obtained. These are combined with estimates of adoption rates and
yield increases in order to estimate probable changes in prices and quantities. Changes in costs of
production and the costs of the research and extension programs must also be estimated. These
baseline estimates are presented in Table 1 below.
Table 1. Baseline Data for the Analysis
1. Price of sorghum/millet 65
2. Price of soghum/millet stover/Kg3 10
3. National production of soghum & millet/yr (000 t)4 1,505,143
4. Potential Zai area in the North Zone
a) Cattle (15/Ha/yr) 68,000
b) Sheep (53/Ha/yr) 44,000
c) Goats (97/Ha/yr) 30,000
d) Total hectarage in North Zone (Ha) 142,000
Assumption: Utilization of'10% total 14,200
e) Assumed adoption (Ha)/yr
5. Estimated cost of zai research program 1990-19955 5,000,000
6. Estimated cost of extention program 1995- 19996 10,000,000
7. Ave. Percentage increase in production/Ha7 84%
8. Elasticity of supply (sorghum/millet) 0.8
9. Elasticity of demand (sorghum/millet) 0.4'
Additionally, we assume that 'off-season' household labor costs provide the only relevant costs
differences for this analysis. The average opportunity cost of labor assumed for this period of
time is 25 F CFA/hour. Any down-stream labor differences such as differences in weeding time
stemming from the use ofzai are ignored. No potential external costs are examined. Finally, we
assume the productivity effects of zai last for two years and remain constant over this time period.
Some items in Table 1 require additional clarification. One limitation of the zai technology is the
availability of manure (4). Using a conservative scenario, RSP researchers have estimated that
some 15 cattle, 53 sheep or 97 goats are required to provide the manure sufficient for an average
zai application (>8 tons/Ha). The GIS database indicates that there are sufficient livestock in the
North Zone to provide manure for some 142000 hectares at this application rate. We suppose,
however, that only about 10% (14,200) of this potential will be realized. This translates to about
6 hectares per village or about 0.04 hectares per exploitation in the North Zone.
2Estimated average post-devaluation farm-gate price for soghum and millet.
3 Average estimated post-evaluation farm-gate price of soghum and millet stover.
4 Average of last nine years of sorghum and millet production. Source: Ministry of Agriculture.
5 Source: Estimate provided by RSP Program.
6 Source: Estimated provided by RSP Program.
7 Source: RSP monitoring survey, 1993-1994.
The adoption rate specified in (4e) assumes a doubling of hectarage in the North Zone every year
for five years as a result of the extension program. The enthusiastic response of farmers in the
RSP village site indicates this may also be a conservative assumption.
The most problematic estimations for the analysis are those for the elasticities of demand and
supply (items 8 and 9). There are simply not any reliable and consistent estimates available.. In
the case of demand, it is highly likely that the real, "average" estimate we seek is less than one
(inelastic) and probably much less than one. We accept 0.4 as a realistic estimate. Similarly, the
supply elasticity is expected to be somewhat inelastic. We accept 0.8 as an reasonable estimate.
Both estimates have been used in similar, regional analyses of impacts (Masters and Sanders,
1994). The results indicated below are much more sensitive to changes in the elasticity of supply
than the elasticity of demand. As the elasticity of supply increases, the internal rate of return falls.
III. RESULTS: POTENTIAL VALUE OF ZAI AND IK RESEARCH
Table 2 presents an evaluation of the potential value of the zaY technique. It is assumed that the
zai technology can be extended throughout the RSP North Zone, a homogeneous zone identified
by the RSP Geographic Information System (GIS) unit. It was in this zone that the technology
was developed and tested. The zone includes 10 provinces with a total area of over 7 million
hectares (see zone map).
Table 2. Economic evaluation of Zai'
1. Average Net Profit (Zai over control, F CFA/Ha) 31,125
a) Total increase in profit per year by the 6th year 441.98 mil
b) Value of the Zaitechnology .944.87 mil
1. Adoption rate doubles each year for 5 years
2. Present value discount rate = 20%
2. Increase in National Agricultural GDP (6th year) 503.04 mil
Percentage of Total AG GDP 0.07%
Percentage of total value of Sorghum production 1.10%
The average returns to zar have been estimated by RSP at 31125 F CFA/Ha (assuming an
opportunity cost of off-season family labor at 25 F CFA/Hr). Accepting this value and the
manure availability limitation, the regional net increase in farmer income is estimated to be about
442 million per year by the end of the 6th year (la). The stream of income into perpetuity has a
present value of some 945 million (lb) if one assumes a discount rate of 20% per year and the
adoption rate assumed in Table 1 above.
Another way to express the potential benefits of za is to consider the potential impact on
Burkina's agricultural gross domestic product (GDP). Assuming the same levels of adoption in
the same zone, agricultural GDP would rise some 503 million or by 0.07% of agriculture GDP by
the end of the 6th year (2a). This figure represents an increase of 1.10% in the total value of
national sorghum production (2b).
Rate of return to research. While the above figures indicate that zar has important potential, even
with the conservative assumptions we have used, they say nothing about the social value of the
RSP research which has made these results possible. Given the integrated nature of the RSP
program and the large number of research themes addressed by RSP researchers during any given
year, any allocation of the RSP budget to a specific theme is somewhat arbitrary. However, if one
allocated a 5 million F CFA portion of the RSP research budget to zai each year from 1990 and
through 1995 and one assumed a 10 million/year extension program to extend zai beginning in
1995 and lasting though 1999, the resulting internal rate of return to research would exceed 50%
for the 15 year period 1990 2004. This is illustrated in Table 3 below where estimated
net social gains are also presented. If twice as much land were put into zai due to (say) collecting
20% of the manure instead of 10%, the returns would rise to almost 70% and the value estimates
in Table 2 would double.
Our analysis of zar indicates that there may be very high potential social benefits stemming from
the current trend towards studying
Table 3. Estimated program costs and net social gains*
Research Year Costs Total Extension Real Cost Net Cost Social Gain
(nominal) (1994 CFA) (1994 CFA) (1994 CFA)
1990 5.0 0.0 5.0 6.26 -6.26
1991 5.0 0.0 5.0 6.19 -6.19
1992 5.0 0.0 5.0 6.26 -6.26
1993 5.0 0.0 5.0 6.19 -5.95
1994 5.0 0 5.0 5.00 -4.61
1995 5.0 10 15.0 15.00 -6.30
1996 0 10 10.0 10.00 7.39
1997 0 10 10.0 10.00 24.76
1998 0 10 10.0 10.00 59.53
1999 0 10 10.0 10.00 129.08
2000 0 0 0.0, 0.00 278.20
2001 0 0 0.0 0.00 278.20
2002 0 0 0.0 0.00 278.20
2003 0 0 0.0 0.00 278.20
Internal Rate of Return: 52.71%
*Real Cost = nominal cost/price index
Net Social Gain = Gross Social Gain Real Cost
The RSP group identified zar as a potentially important IK practice, facilitated its transfer to
another village and studied the impact. This relatively small investment has born very promising
results. Using conservative assumptions, we estimate an internal rate of return to research on zai
The need to find sustainable solutions to agricultural and natural resource management problems
compels development-related scientists to include IK in their research programs. In Burkina Faso
as elsewhere, there are almost certainly other IK practices which are as promising as zai but which
need to be identified and extended to wider user groups. The RSP group in Burkina Faso plans to
continue this important work.
Akinlosotu, J.A.."Indigenous Knowledge on Diseases and PestsManagement in Crop Production".
National Farming Systems Research Network Newsletter 11: 9-11. Zaria, Nigeria, 1993.
Erinle, I.D.. "An overview and critique of indigenous knowledge and farmer participatory
methodology". National Farming Systems Research Network Newsletter 11: 29-33.
Zaria, Nigeria, 1993.
Erinle, I.D.. "Indigenous Knowledge Systems and Sustainable Crop Production Systems in West
Africa". Paper presented at the WAFSRN General Symposium on Indigenous Knowledge
Systems and Participatory Research for Natural Resource Management and Sustainable
Agriculture in West Africa, (GSIK&PR), Cotonou, Benin Republic. June 20-22, 1994.
Gefu, Jerome 0.. "Indigenous Knowledge: The Case of Ethnoveterinary Practices". National
Farming Systems Research Network Newsletter 11: 20-23. Zaria, Nigeria, 1993.
Kabor6, Daniel, Freddric Kambou, John Dickey and J. Lowenberg- DeBoer. "Economie
des Cordons Pierreux, du Paillage et du Zai dans le Nord du Plateau Central du
Burkina Faso, une Perspective Prdliminaire". Unpublished mimeo, Dec. 1993.
Maigida, D.N.. "Indigenous Knowledge Systems among Women Farmers in Nigeria".
GSIK&PR, Cotonou, Benin Republic. June 20-22, 1994.
Masters, William, and John Sanders, "L'Impact Economique de la Recherche Agricole,"
Training Matierals for a Workshop held in Bamako, Mali, June, 1994.
Obi, Alolisa. "Indigenous Knowledge Systems and Sustainable Fish Production in West Africa".
GSIK&PR, Cotonou, Benin Republic. June 20-22, 1994.
Ojeniyi, Elizabeth Tola. "Indigenous Knowledge in Food Processing and Preservation". National
Farming Systems Research Network Newsletter 11: 12-15. Zaria, Nigeria, 1993..
Otegbeye, Gbadegesin 0.. 'Indigenous Knowledge Systems and Sustainable Agroforestry
Development in West Africa". Invited paper. GSIK&PR, Cotonou, Benin Republic, June
Ouedanou, Eugene K.. "Recherche Participative et Gestion des Ressources Naturelles pour une
Agriculture Durable: Experiences du Benin". GSIK&PR, Cotonou, Benin Republic, June
Robins, E. and M.C. Sorgho, "Evaluation Paysanne de la Confection du Zal". RSP Report.
INERA. Ouagadougou, Burkina Faso. Feb. 1994.
Samuelson, Paul A.. Economics, 10th Edition. McGraw-Hill Company, New York. 1976.
Schultz, Theodore W.. Transforming Traditional Agriculture. Yale University Press, New Haven
and London, 1969.