GRAZING LAND TENURE AND LIVELIHOOD SECURITY:

A STUDY OF TWO CLUSTERS OF VILLAGES IN THE GAMBIA

[table of contents] [previous] [next]

4.0 Research Methods

4.1 Overview of Qualitative Research Methodology and PRA

The primary goal of this research was to construct a model (or models) of livelihood security strategies for rural Gambian households with special attention to how those strategies relate to livestock and grazing land tenure. The second goal of the research was to use any model that was constructed to examine potential modifications to the tenure system as it applies to grazing land, to examine how those modifications might affect livelihood security, and to explore policy implications arising from that examination. At the heart of the first goal of the research is a desire for a holistic understanding of some complex phenomena: people's survival decisions and coping behaviors, with all of the diverse influences that those decisions and behaviors have. In formulating this goal it was stated that in constructing the model special attention was to be given to how security strategies relate to livestock and grazing land tenure. In Section 3.2 it is argued that land tenure systems must be seen as systems -- complex systems of various degrees and types of rights set within a wider social system. This adds to the complexity of the subject matter.

Because little is known about the details of the relationships between the various degrees and types of rights to land and people's security strategies, this research was inductive and exploratory. The primary goal of the research, in other words, was to induce a model of security strategies. Qualitative research methods, therefore, seemed appropriate. Marshall and Rossman (1989:46) state that qualitative research is appropriate for several kinds of research including research that is exploratory and research that "delves in depth into complexities and processes." According to McCracken, "Qualitative research does not survey the terrain, it mines it. It is, in other words, much more intensive than extensive in its objectives." (McCracken, 1988:17)

Techniques taken from Participatory Rural Appraisal (PRA) can be an effective part of an exploratory research design. PRA has had a number of influences. According to Chambers (1992:2), five traditions have been sources and parallels to PRA: activist participatory research, agroecosystem analysis, applied anthropology, field research on farming systems, and Rapid Rural Appraisal (RRA). PRA is "a family of approaches and methods to enable rural people to share, enhance, and analyse their knowledge of life and conditions, to plan and to act." (Chambers, 1992:1) Although it involves research, it is more than that. "PRA is both an attitude and a methodology." (Joseph, 1991:132)

Furthermore, the techniques of PRA, most of them research techniques, are not, in themselves, PRA. Some of the techniques were already being used in Rapid Rural Appraisal before an emphasis on participation produced PRA. With these techniques, RRA

was, and remains, less exploitative than extractive questionnaire surveys where much is taken by the outsider, and little or nothing is given back. All the same, like most past farming systems research, its normal mode entails outsiders obtaining information, taking it away, and analysing it.

(Chambers, 1992:9)

Unless the rural people themselves have control over the entire process, including analysis and final use of the information, the process should not be called "Participatory Rural Appraisal." Nevertheless, the distinction between PRA and RRA may not always be easy to draw. Chambers (1992:13) suggests that "In practice there is a continuum between an RRA and a PRA." In this thesis I use the term "PRA" in describing some of the research techniques used. However, it should be emphasized that this research is not an example of PRA; I have merely used some techniques from PRA. I do not describe them as "RRA techniques" simply because most were developed and popularized as parts of the PRA methodology.

Techniques that were used include participatory mapping of villages and of village territory, dispute matrices, matrix scoring, transect walks, timelines, Venn diagrams of institutions, and pie charts. These techniques are widely used and are recognized as effective ways of getting valid, detailed information from "local experts." Discussions of these and other PRA techniques can be found in ActionAid The Gambia (1992), Bruce (1989), Government of Kenya National Environment Secretariat (1990), Grandin (1983), Njiforti, et.al. (1991), RRA Notes (all issues), and Swift (1981).

Chambers names several important discoveries or, more accurately, rediscoveries which have shaped PRA. The first four of these proved especially relevant to this research:

o The first discovery is that villagers have a greater capacity to map, model, quantify and estimate, rank, score and diagram than outsiders have generally supposed them capable of. . . .

o The second discovery is that relaxed rapport between outsiders and rural people can and should be established early in the process. . . .

o The third discovery is the popularity and power of participatory diagramming and visual sharing. . . .

o The fourth discovery is the power and popularity of sequences of participatory methods.

(Chambers, 1992:20-23)

One of the reasons I gave for choosing qualitative research methods for this research was the value of qualitative methods for exploring ideas and defining categories during the research process. Many PRA techniques can be helpful in this regard in that they involve not only producing information but analyzing it. As Chambers notes in the "first discovery," villagers are capable of modelling, drawing, and diagramming. For example, Venn diagrams of institutions allow villagers not only to present the institutions that exist in their village, but also to show relationships between institutions. Furthermore, the group setting in which most PRA techniques are used allows for discussion between villagers, instead of only between the researcher and the villager, and it is often in the discussions rather than in the exercises, themselves, that the most valuable information emerges. The group setting, combined with the analytical nature of many PRA techniques, allows topics to be explored in great depth.

A related value of PRA techniques in the context of a qualitative research program is their rapidity. Analysis happens during the exercises. Furthermore, the sequencing of techniques, Chambers' "fourth discovery," allows for a very efficient way of accomplishing a great deal. In this research, for example, participatory mapping was used as a census of village households; this was followed by wealth ranking of these households; which was later followed by purposive sampling of households of various levels of wealth for interviews with household members. The efficiency that comes with a well-structured research process with PRA techniques allows more time for exploring issues in depth, thereby moving the enquiry closer to accomplishing its goal of achieving a holistic understanding of livelihood security strategies. If planners and policymakers cannot achieve this holistic understanding there is a danger that their actions will do more harm than good.

 

4.2 The General Methodology

The research process began with a review of the literature on food security, livelihood security, and land tenure. From the literature, a model of livelihood security was adopted, developed further, and adapted for an examination of land tenure systems (see Chapter Three). This adapted model served as a preliminary, general model of livelihood security and land tenure and guided what information would be needed and how that information would be interpreted.

Information relating to three broad areas -- the land tenure system, livelihood security strategies, and policy -- was required. Information requirements related to the land tenure system included information on herd mobility, land tenure rules and access rights, and patterns of decision-making for communal pastures at the village and inter-village level. Information on security strategies included details about household wealth and livestock ownership, about what livelihood assets households rely upon under what conditions, and about how households invest in livelihood assets. In the area of policy, information was needed about likely policy interventions in the area of grazing land tenure (if any), about the legal and bureaucratic provisions that would be required for likely policy interventions, and about any administrative boundaries at the inter-village level.

To achieve the goals that were set out, a holistic understanding of security strategies and their relationship to land tenure was required. Therefore, depth of understanding was more important than breadth of understanding. Also, because no single hypothesis guided the research, the methods were inductive and exploratory. To allow for this exploratory approach, it was decided to spend a relatively long period of time in a small number of sites. Two village sites were selected to represent some of the diversity of the country in terms of ethnicity, physical environment, land availability, and farming systems. Despite conducting most of the field research in only two sites, attention was given to the eventual generalizability of the results. In terms of several important criteria the two sites can be seen as extremes with the rest of the country falling in between (see Section 4.2).

I relied upon contacts I had developed in the Gambian Ministry of Agriculture to provide me with introductions to local leaders at each site. These local leaders then assisted me with introductions within their respective villages. In addition, in each village was posted an extension agent from the Department of Livestock Services, Ministry of Agriculture. They also helped with introductions. At each site, within a few days of my arrival I held a large meeting attended by the alkalo (village head) and other interested men to discuss what I would be doing in the village and to ask some initial questions. A similar meeting was held a day or two later with the village women's committee (the second village, Sutukunding, had two separate women's committees, so two introductory meetings were held). At each cluster of villages six weeks were spent conducting the research.

One of the primary methods used for the field research at each site was semi-structured interviews with heads of households, senior wives, and other livestock owners. The interviews, not being guided by a rigid questionnaire, allowed various aspects of security strategies, herding practices, and land tenure to be explored as the topics arose. Group sessions using Participatory Rural Appraisal techniques allowed for the same sort of open-ended exploration of issues (see Section 4.3).

In conducting the interviews and group sessions I was assisted by a translator. This person would translate while I took notes, then after the interview or group session I would confer with the translator to ensure that I had recorded the information accurately and that no important details had been missed. Over the course of the research I worked with two different translators, both of whom were fluent in Wolof, Manding, and English and were experienced local language teachers for a foreign volunteer organization. One was also fluent in Fulani. Some of the maps, matrices, and diagrams that were produced with the PRA techniques were drawn on the ground and some were done directly onto large sheets of paper. Those that were drawn on the ground were transcribed by me.

The data that was collected was then applied to the goal of constructing a model and achieving a holistic understanding. In doing so, as much attention was paid to synthesizing the data as to analyzing it. The general model of livelihoods and land tenure that was derived from the literature guided the understanding of the information and a model of Gambian livelihood security strategies emerged. This approach allowed for the second goal for the research -- analysis of policy issues -- to be approached with a deeper understanding of the needs of the rural poor.

 

4.3 The Two Sites for Village Research

4.3.1 SELECTION OF THE SITES

Information related to land tenure systems and livelihood security strategies was collected through field research that was conducted in two clusters of villages, with most of the research being done in one "core" village in each cluster. The research was conducted in clusters of villages, rather than in several individual, dispersed villages, to allow for the study of inter-community processes. Six weeks were spent at each of two sites, which were selected to act as two case studies that would represent, at least partially, the variety of social and physical conditions that exist in The Gambia. The particular conditions that were used to differentiate the two sites were those thought likely to influence livelihood security strategies and land tenure systems -- conditions such as physical environment, farming systems, land availability, and ethnicity. The exact criteria used to compare sites, drawn from secondary sources, included annual rainfall levels, type and mix of crops grown, and area of land suitable for agriculture per person. The two clusters of villages were also selected to contain a mix of ethnic groups.

Comparative information on farming systems was available at the division level (See Table 4.1). The Gambia Agricultural Research and Diversification Program (Torrence, 1991:23) had found that households in Upper River Division (URD) are more likely to grow maize and sorghum than households anywhere else in the country, with 97% percent of URD households growing maize and 66% of them growing sorghum; in contrast, households in North Bank Division (NBD) are least likely to grow maize -- 33% of NBD households -- and are very unlikely to grow sorghum -- 4% of NBD households. Households in NBD are more likely to grow early millet than households anywhere else in The Gambia, with 94% of households growing that crop; only 1% of households in URD grow early millet (Torrence, 1991:23). For the purposes of studying land tenure there is an even more important difference between the two divisions. For land classified as suitable for agriculture, in 1983 NBD had the highest suitable land to person ratio in the country, 0.8 hectares/person; URD had the lowest ratio, 0.4 hectares/person (Torrence, 1991:19). In terms of these criteria, these two divisions can be seen as extremes with the rest of the country falling somewhere in between. By choosing to have one research site in each of these two divisions, I hoped to enhance the generalizability of the results.

Finding different locations with significant rainfall differences was more difficult, as rainfall is relatively uniform throughout the country with the exception of areas near the coast. However, there is some difference in mean annual rainfall at the two sites that were finally chosen, with the NBD site receiving 870.2 mm. per year and the URD site receiving 1001.4 mm per year.¹

Table 4.1: Percentage of Households Growing Major
Crops By Division, The Gambia, 1991

Division	Maize	Early Millet	Late Millet	Sorghum	Upland Rice	Swamp Rice	Groundnuts
Western	57	11	69	19	37	41	70
North Bank	33	94	1	4	30	42	83
Lower River	44	82	6	3	27	70	82
MacCarthy Island-north	72	71	2	26	4	58	91
MacCarthy Island-south	67	79	10	11	0	46	90
Upper River	97	1	60	66	22	0	98
The Gambia	59	54	28	20	23	39	84

(Torrence, 1991:23)

 

I relied upon the opinions of local contacts and experts to suggest villages in NBD and URD that would be receptive. The first of the two research sites that was chosen was comprised of the village of Ngeyen Sanjal, Upper Baddibu District, North Bank Division and several neighboring villages and the second was comprised of the village of Sutukunding, Wuli District, Upper River Division and several neighboring villages (see map on foldout page). For example, in NBD, Ngeyen Sanjal became the core village for the research, but the neighboring villages of Balo Ibra, Balo Omar, Palen Fula, Palen Wolof, and Dafa were also visited. In URD, Sutukunding was the core village, but the neighboring villages of Bani, Farato, Taibatu, and Madina Koto were also visited. The sites differed in ethnicity, Ngeyen Sanjal being a Wolof village and Sutukunding being Mandinka. In addition, both Ngeyen Sanjal and Sutukunding had at least one neighboring Fula village that was included among the villages visited. In this manner, villages of the three largest ethnic groups in The Gambia (Fula, Mandingo, and Wolof) were studied.

The findings of the research were influenced by the social, physical, and economic conditions that exist at each site. Therefore, before presenting the content of the research, a description of the context is in order. What follows are descriptions of the two core villages of each cluster.

4.3.2 NGEYEN SANJAL

I first arrived in Ngeyen Sanjal in early December, 1993, when the harvest was almost complete. Ngeyen Sanjal (see Figure 4.1) is just over one hundred kilometres inland from the coast and is nineteen kilometres from the large town of Farafenni. It is a Wolof village to the north of the River Gambia, seven kilometres from the river and two kilometres from the Senegalese border. It sits at the junction of the main east-west road of the north bank and a lesser road that serves a number of villages closer to the river. Perhaps because of this crossroads location the village is the focal point for a number of activities such as a lumo, a weekly market that attracts buyers and sellers from several nearby villages. It is the site of some government services for the surrounding villages, including a health clinic, the extension agent for the Ministry of Agriculture Department of Livestock Services, and a primary school. The seyfo (chief) of the district lives in Ngeyen Sanjal and the District Tribunal, which the seyfo chairs, sits in the village.

The village is quite large and somewhat extensive, with a population of approximately 1,500 spread over an area over one and a half kilometres long and almost one kilometre wide. It is also divided into four distinct neighborhoods. About two and a half kilometres from the village is a large area of uncultivated land in and surrounding a government forest park. Nearer to the village, fields are planted with millet and groundnuts. Rice is grown some distance away closer to the river, although the rice fields are losing their usefulness because of salinization. Small amounts of other cereals and vegetables are grown by some.

4.3.3 SUTUKUNDING

My first visit to Sutukunding was in January, 1994 although this was a very brief stay and I returned to actually begin research in mid-February. Sutukunding is a Mandinka village² located at the eastern end of the country, also on the north side of the river. It is three kilometres from the river at its nearest point and eight kilometres from the town of Basse on the south bank of the river (see Figure 4.2). From some elders in the village, I learned that it was settled approximately one hundred years earlier on the site of an abandoned Fula village. Like Ngeyen Sanjal, Sutukunding is relatively large, with a population of about 1,000, and is at a crossroads; however, except for having a number of extension agents located in the village, Sutukunding is not the centre of activity that Ngeyen Sanjal is. Most marketing activities, for example, are taken across the river to Basse.

The village is also different from Ngeyen Sanjal in that it is quite compact with very little open space within the village-proper. Sutukunding is the largest of a group of four Mandinka and Jahanka villages located within a two kilometre radius. Most of the uncultivated land that villagers use to pasture livestock is located in a rocky, hilly area east of the village, beside the river. The crops that are grown on the fields surrounding the village include millet and groundnuts and lesser amounts of maize, sorghum, and beans.

 

4.4 Methods

4.4.1 INFORMATION ON THE LAND TENURE SYSTEM

At each of the two research sites information was required about land as an asset: the intangible asset of access rights and other land tenure rules, and the tangible asset of land as a resource, including information about investment in land. Other information related to land tenure included information on herd mobility and on ways decisions are made about communal pastures at the village and inter-village level. The methods included group sessions with village elders and decision makers from several villages in each cluster, with organizations such as women's and youth committees, and with impromptu groups of household heads and other livestock owners.

Participatory Rural Appraisal techniques were used, including activities such as mapping of village territory, transect walks, timelines, venn diagrams of institutions, dispute matrices, and matrix scoring. For the most part, the value of these techniques lay more in the discussion that they generated than in the particular map or diagram that was produced. The particular techniques are described below, as they are presented. Information was also gathered through interviews with various local leaders such as seyfolu (district chiefs), alkalolu (village heads), and members or leaders of various organizations such as the Kunjata-Bambali Rangeland Project Management Committee, the Livestock Owners Association of Upper Baddibu District, the Ngeyen Sanjal Village Development Committee, the Sutukunding Village Development Committee, and the Sabach-Sanjal Women's Association. Interviews and walkabouts were also done with herdsmen.

In the various sessions in which PRA techniques were used and in the interviews with local leaders, livestock owners, and herdsmen, similar questions were asked a number of times. Analysis of the information that was gained was guided by a desire for triangulation in order to develop a consensual picture of the various rights of access as well as the actual practices of access that make up the land tenure system. An attempt was made to understand what rights, responsibilities, and rules people agree upon and to understand where there is disagreement or conflict.

4.4.2 INFORMATION ON LIVELIHOOD SECURITY STRATEGIES

Information that was needed on security strategies included details about what livelihood assets households rely upon under what conditions and about how households invest in livelihood assets. Another objective was to relate this information to level of wealth and to any other factors that may distinguish the strategies of one household from the strategies of another. One of the difficulties in investigating livelihood security at the level of the household was deciding upon a definition for "household," a task which for many African societies can be problematic. The most obvious definition equates "household" with "compound" -- the extended family members living in a group of huts or houses usually surrounded by a single fence or wall. However, compounds are sometimes subdivided into more than one production unit and/or more than one consumption unit.

In the Wolof village, Ngeyen Sanjal, choosing a definition for "household" was relatively easy since few of the compounds had these sorts of subdivisions. The few that did were quite neatly divided and seemed to function as two completely separate social compounds within one physical compound. Therefore, in Ngeyen Sanjal, "household" was defined as the compound. The compounds that had two social compounds were considered to be two households. In the Mandinka village, Sutukunding, defining "the household" was not as easy. Many compounds had more than one sinkiro (lit. "bowl") or consumption unit and a few had more than one dabada (lit. "the mouth of the hoe") or production unit. Enumerating every sinkiro or every dabada would have been extremely difficult, therefore in Sutukunding the same definition for "household" was used, equating it with the compound.

Sampling of households within each village was purposive rather than random and was designed to ensure that households of varying degrees of wealth were studied. Because rural livelihoods in The Gambia are diverse, no single criteria of wealth stands out. The findings of a Gambia Mixed Farming Project technical report (Haydu, et.al., 1986) suggest that the correlation between number of livestock owned and other measures of wealth is not strong enough to be a reliable measure of general household wealth. The same applies to size of landholding. Furthermore, because households are not engaged in the cash economy to the same degree, a monetary measure of wealth would not likely be valid.

Therefore, wealth was assessed in aggregate using PRA wealth ranking exercises based on the method developed by Barbara Grandin (1983). The first step in each core village was an enumeration of the households, which was done by means of drawing a village map. Small groups of villagers drew the maps and then named the head of each compound shown on the map. These names were written on index cards. Ngeyen Sanjal is a large village made up of four distinct neighborhoods. Maps, as well as the subsequent wealth ranking, were done for only two of those neighborhoods. In Sutukunding, all compounds were put onto a single map. The cards with the names of compound heads were then divided into rank categories by key informants from the village. It was made clear to informants that they were to rank the cards based upon the wealth of the compound as a whole, not the individual compound head. For each set of cards, the exercise was done three times with different key informants to triangulate the results. That there were very few cases of compounds being placed in radically different categories, testifies to the legitimacy of the method and suggests that the definition used for "household" was meaningful. Other criteria, such as ownership or non-ownership of draft animals and numbers and types of livestock owned, were used as a secondary check on the wealth ranking results.

Some information on livelihood security, not related to any specific households was collected through group sessions using PRA techniques, such as matrix ranking of various ways of saving and pie charts showing livestock ownership by gender. Household-specific information on security strategies was gained through semi-structured interviews with compound heads, heads of dabadas (a type of subdivision within some Mandinka compounds), senior wives, and other livestock owners. In Ngeyen Sanjal interviews were conducted with members of thirty-three households, and in Sutukunding with members of thirty-seven households. The households were of varying levels of wealth. Eight interviews were done in households of nearby villages, primarily Fula villages, thereby including a third ethnic group in the study of security strategies, although no wealth ranking was done for these households.

The semi-structured interviews were based on the following questions:

o Because of the rains or for other reasons you can have good years and bad years. When was the last year for your compound that was relatively bad?

o What made it a bad year?

o What did you have to do to get through that year?

o Did you or anyone in the compound have to sell any possessions to get money?

o Did you have to sell or slaughter any livestock? What types and how many?

o If the compound needs to sell an animal to buy food, whose animal is sold? Husband, wife, other person in compound?

o Did you have to borrow from anyone? Who did you borrow from -- friends, relatives, shopkeepers, wealthy neighbors?

o If you had a bad year and had to choose between selling an animal or borrowing from someone, which would you do and why?

o In this compound, what do you do to protect yourselves in case the next year is a bad one?

o When was the last year for your compound that was relatively good?

o What made it a good year?

o Were you able to save that year? What did you invest in?

o If you had two thousand dalasis to save (approximately enough to buy one cow) what would you do with the money? What makes your answer a better investment than [alternative investments: a cow, sheep, goats, bags of rice, etc.]?

o What if you had only five hundred dalasis to save?

o In this compound, do you rely on the bush in any way for food or income?

o How many people live in this compound?

o What types of livestock do you have in this compound? How many of each? Do you have any draft animals?

o Why do you have this type of animal and not some other?

Analysis of the information obtained was guided by a concern for what needs, rules, priorities, and values influence people's decisions about security under various conditions. The information was also related back to the model of livelihood security that was presented and adapted in Chapter Three. For example, the interviews revealed what actions people would take when faced with a shortfall in production. One question asked about that information was, "Which livelihood asset or assets are people relying upon?"

4.4.3 INFORMATION ON POLICY CONSIDERATIONS

In the area of policy, information was needed about possible government interventions that would affect grazing land tenure (if any) and about the legal and bureaucratic provisions that would be required for possible interventions. The information was obtained through secondary sources (government documents, project documents, previous studies, etc.) and through interviews with officials in the Gambian bureau of the UNDP and in various Gambian government bodies -- the Ministry of Agriculture (Departments of Planning and of Livestock Services), the Ministry of Natural Resources and Environment (Environment Unit and Planning Unit), the Law Reform Commission, the Working Group on Resource Tenure and Land Use Planning.

Rangeland project sites at Dankunku, MacCarthy Island Division and Kunjata-Bambali, North Bank Division (near Ngeyen Sanjal) were also visited. The project at Dankunku has been operating for some time and subsequent rangeland and dairy projects in The Gambia have been modelled after it, with some revisions. I spent two days at Dankunku interviewing villagers, including members of the project's management committee, interviewing the Department of Livestock Services staff member posted in the village, and touring the rangeland site and dairy. I also participated in a two-week case study conducted by the Working Group on Resource Tenure and Land Use Planning. This body is an interministerial committee of the Gambian government that is being supported by the Land Tenure Centre at the University of Wisconsin. The case study was conducted by a team of nine researchers in the village of Darsilami, Sandu District, Upper River Division, approximately twenty kilometres from Sutukunding. PRA techniques were used to examine a wide variety of issues related to land use and land and resource tenure.

 

4.5 Critique of The Methodology

The primary goal of the research, to construct a model of livelihood security strategies for rural Gambian households, is accompanied by a danger that is inherent in any model-building approach: the danger of oversimplification. A model cannot describe or predict all behavior under all conditions; however, it can make clear some of the major influences on behavior. It was hoped, as well, that by restricting the number of sites to only two and exploring issues in depth at each site, the model produced would be well grounded, reflecting the most important components and influences of livelihood security strategies.

A related weakness of the methodology that was followed in this research is that definitive generalizations cannot be made. The two sites were selected for contrast, but two sites cannot represent the entire country and there were other potential locations that are different from both sites. For example, some areas of MacCarthy Island Division experience an influx of cattle herds every dry season. This would likely affect grazing land tenure in those areas. Furthermore, although information was available that made it possible to differentiate North Bank Division from Upper River Division, there was no way to ensure that the particular sites chosen were typical of these two divisions.

The PRA techniques that were used were found to be quite effective. One technique, Venn diagrams of institutions operating in the village, yielded no information of direct relevance to the information that was needed. However, I found it to be an excellent introduction to the village, as it identified formal and informal power structures within the village and identified people or groups worth interviewing. The problematic issue for the PRA methods was not the techniques, themselves, but the composition of the discussion groups. I was not able to select participants for those groups randomly. I dealt with the problem of representativeness of these sessions by trying to meet with different types of groups. In some cases, meetings/discussions were announced and people who were interested would come; in some cases, I met with existing committees or groups; in some cases I held impromptu sessions with whomever was sitting at the bantaba or central village square. I also paid attention to who was at each meeting to be sure that I was including people I had not included before.

As this research relied on information from informants, the reliability of the informants was a concern. I attempting to check the accuracy of what I was told by approaching important questions in different ways and by seeking opinions from a variety of people with a variety of personal interests. More importantly, I spent several weeks in each of the two core villages and was able to build a good rapport with the people there.

---

1.  Rainfall levels are based on data collected at the nearest meteorological stations, Jenoi for the NBD site and Basse for the URD site (GOTG, 1982:13,21).  Jenoi is not in NBD but it is the closest meteorological station to the site that was finally chosen.

2.  Approximately half of the residents of Sutukunding are Jahankas, a sub-group within the larger Mandinka ethnic group.

[table of contents] [previous] [next]