CHANGE IN AFRICA
CENTER FOR AFRICAN STUDIES
UNIVERSITY OF FLORIDA
CARTER LECTURE SERIES
Center for African Studies, University of Florida
Structural Adjustment and African Woman Farmers,
edited by Christina H. Gladwin (1991)
edited by R. Hunt Davis, Jr. (1991)
Human Rights and Governance in Africa,
edited by Ronald Cohen, Goran Hyden, and Winston P. Nagan (1993)
Population Growth and Agricultural Change in Africa,
edited by B. L. Turner II, Goran Hyden, and Robert Kates (1993)
CHANGE IN AFRICA
Edited by B. L. Turner II,
Goran Hyden, and Robert W. Kates
UNIVERSITY PRESS OF FLORIDA
Gainesville Tallahassee Tampa Boca Raton
Pensacola Orlando Miami Jacksonville
Copyright 1993 by the Board of Regents of the State of Florida
Printed in the United States of America on acid-free paper
All rights reserved.
Library of Congress Cataloging-in-Publication Data
Population growth and agricultural change in Africa / edited by B. L. Turner II,
Goran Hyden, and Robert Kates.
"Carter lectures on Africa."
ISBN 0-8130-1219-8 (acid-free paper)
I. Food supply-Africa, Sub-Saharan-Congresses. 2. Agricultural productivity-Africa,
Sub-Saharan-Congresses. 3. Africa, Sub-Saharan-Population-Congresses. I. Turner,
B.L. (Billie Lee), 1945-. II. Hyden, Goran, 1938-. III. Kates, Robert William.
The University Press of Florida is the scholarly publishing agency for the State University
System of Florida, comprised of Florida A & M University, Florida Atlantic University,
Florida International University, Florida State University, University of Central Florida,
University of Florida, University of North Florida, University of South Florida, and University
of West Florida.
University Press of Florida
15 Northwest 15th Street
Gainesville, FL 32611
Foreword by Peter R. Schmidt xiii
1 Theory, Evidence, Study Design 1
Robert W. Kates, Goran Hyden, and B. L. Turner I
2 Agricultural Transformation in the Robusta Coffee/ Banana
Zone of Bushenyi, Uganda 41
3 Increasing Variability in Agricultural Production: Meru
District, Kenya, in the Twentieth Century 80
F. E. Bernard
4 Defending the Promise of Subsistence: Population Growth
and Agriculture in the West Usambara Mountains,
5 Marginal Coping in Extreme Land Pressures: Ruhengeri,
Robert E. Ford
6 Population Growth and Agricultural Change in Kisii District,
Kenya: A Sustained Symbiosis? 187
H. W. O. Okoth-Ogendo and John O. Oucho
7 Agricultural Expansion, Intensification, and Market
Participation among the Kofyar, Jos Plateau, Nigeria 206
Robert McC. Netting, Glenn Davis Stone, and
M. Priscilla Stone
8 Population Growth and Agricultural Change in Imo State,
Southeastern Nigeria 250
vi / Contents
9 From Agricultural Growth to Stagnation: The Case of the
Ngwa, Nigeria, 1900-1980 302
10 Agricultural Stagnation and Economic Diversification:
Awka-Nnewi Region, Nigeria, 1930-1980 324
Francis C. Okafor
11 The Intensification of Peri-Urban Agriculture: The Kano
Close-Settled Zone, 1964-1986 358
12 Beyond Intensification 401
Goran Hyden, Robert W. Kates, and B. L. Turner II
Fig. 1.1. Expectations of the Population-Agricultural Growth
Relationship: Theoretical Traditions 2
Fig. 1.2. Index of Per Capita Food Production 5
Fig. 1.3. Pathways from Agricultural Growth to Improved
Fig. 1.4. Relationships among Measures of Agricultural
Fig. 1.5. Locations of Case Studies 28
Fig. 2.1. Uganda Districts, 1993 42
Fig. 2.2. The Bushenyi Study Area 44
Fig. 2.3. Banana and Pasture in Igara County, Bushenyi District 56
Fig. 3.1. Kenya and Meru District 81
Fig. 3.2. Meru District: Agro-Ecological Zones 85
Fig. 3.3. Tea Production in Highland Meru 95
Fig. 4.1. Usambara Mountains, Tanzania 116
Fig. 5.1. Rwanda 147
Fig. 5.2. Ruhengeri Prefecture 148
Fig. 5.3. Ruhengeri Prefecture: Topography 150
Fig. 5.4. Wetland Raised Fields/Mounding and Adjacent Slope
Cultivation, Ruhengeri Prefecture 152
Fig. 5.5. Land-Use Catena, Central Plateau, Rwanda 156
Fig. 5.6. Population Density, Ruhengeri Prefecture 158
Fig. 5.7. Population Growth, Ruhengeri Prefecture 159
Fig. 5.8. Contoured Bands/Ridges on Steep Slopes in the Haute
Buberuka Zone, Ruhengeri Prefecture 164
Fig. 5.9. Natural Erosion, Ruhengeri Prefecture 167
Fig. 5.10. Theoretical Erosion, Ruhengeri Prefecture 169
Fig. 6.1. Kisii District, Kenya 189
Fig. 6.2. General Soil Fertility, Kisii District 191
viii / Illustrations
Fig. 6.3. Average Annual Rainfall, Kisii District 192
Fig. 6.4. Population Density, Kisii District 195
Fig. 7.1. Jos Plateau and Benue Piedmont, Plateau State,
Fig. 7.2. Namu Bush Neighborhoods 208
Fig. 7.3. Bush Population near Namu, 1963 216
Fig. 7.4. Bush Population near Namu, 1977 217
Fig. 7.5. Change in Household Size with Years Spent as a
Migrant Cash-Crop Farmer on the Settlement Frontier 221
Fig. 7.6. Household Size and Sorghum Production 222
Fig. 7.7. Household Size and Bush Yam Production 223
Fig. 7.8. Intensively Cultivated Homestead Farms in the
Kofyar Homeland, Kwa Village Area 225
Fig. 7.9. Farm of Shangpan Dakwat, Namu Area 231
Fig. 7.10. Weekly Agricultural Labor Inputs 235
Fig. 8.1. Eastern States and Coastal Plain Sands, Nigeria 252
Fig. 8.2. Imo State, Local Government Areas: Population
Fig. 8.3. Location of Survey Areas in Imo State 260
Fig. 8.4. Reported Past and Present Fallow Lengths 266
Fig. 9.1. Ngwa Region: Location and Rainfall 305
Fig. 9.2. Ngwa Region: Divisional Boundaries, 1920s-1930s 308
Fig. 10.1. Imo State, Nigeria 326
Fig. 10.2. Awka-Nnewi Region 327
Fig. 10.3. Soil Types, Awka-Nnewi Region 331
Fig. 11.1. Kano Close-Settled Zone and Soil Groups 360
Fig. 11.2. Vertical Air Photograph of a Part of the Kano Close-
Settled Zone 20 km NE of Metropolitan Kano,
1981 (enlarged from scale 1:25,000) 361
Fig. 11.3. Land Holdings in Ungogo, 1964 382
Table 1.1. Districts with Rural Population Densities in Excess of
Table 1.2. Case Study Protocol 26
Table 2.1. Average Hectarage, Yields, and Number of Times
Planted per Year of Cultivation for Major
Cultigens, 1974 65
Table 3.1. Population and Land in Eastern and Central Kenya 82
Table 3.2. Population Growth in Meru, 1948-88 93
Table 3.3. Population Densities of Selected Meru Locations 99
Table 3.4. Eastern and Central Kenya: Net Migration, 1979 101
Table 3.5. Crop Production in Meru, 1976 and 1986 (selected
cash crops) 101
Table 4.1. Monthly Average Rainfalls (in mm) 119
Table 4.2. Population Density by Division, 1975 121
Table 4.3. Number of Males for 100 Females in Varying Age
Groups, 1978 125
Table 4.4. Percentage of Total Population in Segments of the
Population Defined by Age and Sex 125
Table 5.1. Population Densities by Commune 157
Table 5.2. Selected Population Characteristics of the Rural
Household in Ruhengeri Prefecture (Rwanda) 161
Table 5.3. Evolution of Forest Cover in Ruhengeri, 1980-85 165
Table 5.4. Amount of Land in Hectares in Various Crop
Combinations for Each Growing Season in
Ruhengeri, 1984 171
Table 5.5. Average Field/Plot Types and Surface Areas (in ares)
per Rural Household in Ruhengeri Prefecture,
Rwanda, 1984 173
Table 5.6. Average Production of Selected Crops in Ruhengeri
Prefecture for all of 1984 (kg/ household) 177
x / Tables
Table 5.7. Total Production (tons) and Value for Selected Crops
in Ruhengeri (U.S. dollars and Rwandese francs,
Table 5.8. Indicators of Improvements in Social and Material
Well-Being: Conclusion and Scenario for the
Table 6.1. Population Size and Growth in Kisii District, Kenya,
Table 6.2. Census Population and Projections and Population
Density Projections for Kisii District by Division,
Table 6.3. Age Distribution of Population of Kisii District by
Sex and Age-Specific Sex Ratio, 1979 (%) 197
Table 7.1. Comparative Production of Traditional Homestead
and Migrant Bush Farms 229
Table 8.1. Population and Densities of Survey Regions and
Projections to 1988 255
Table 8.2. Group Fallow Rotation and Communal Land
Table 8.3. Fallow Periods Reported in Previous Studies in Imo
Table 8.4. Current and Past Fallow Periods in the Survey
Table 8.5. Indicators of Resource Pressure on Fallow
Table 8.6. Farming-System Adaptations to Land-Use
Table 8.7. Reported Prevalence of Land-Use Transfers 281
Table 8.8. Reported Fertilizer Adoption 283
Table 8.9. Estimates of Outmigration 289
Table 9.1. Population Densities (per km2), Eastern Nigeria,
Table 9.2. Palm-Produce Exports from Calabar, Opobo, and
Port Harcourt, 1906-48 313
Tables / xi
Table 9.3. Total Palm-Produce Exports from Nigeria, 1906-64
and Barter Terms of Trade, 1910-48 313
Table 9.4. Income Terms of Trade for Palm-Produce Exports
from Calabar, Opobo, and Port Harcourt 315
Table 10.1. Internal Population Change in the Awka-Nnewi
Region, 1953-63 335
Table 10.2. Population Density in the Awka-Nnewi Region 335
Table 10.3. Comparison of Spatial Distribution of Population
by Region 335
Table 10.4. Landownership Pattern among Farmers in the Awka-
Nnewi Region 341
Table 10.5. Distribution of Farming Households According to the
Ownership of Farmland Types 341
Table 10.6. Estimated Monetary Value of Farmers' Annual
Agricultural Income in the Awka-Nnewi Region,
Table 10.7. Relationship between Farm Type and Crop-
Combination Value 345
Table 10.8. Population Density and Division of Labor in Farm
and Nonfarm Activities in the Awka-Nnewi
Table 10.9. Fertilizer Use in the Sample Households of Awka-
Nnewi Region, 1982 349
Table 10.10. Number of Extension Staff in Relation to the
Farming Population Being Served in the Awka-
Nnewi Region, 1982 351
Table 11.1. Population Structure in 1964 and 1977 365
Table 11.2. Frequency of Main Crops in the Inner CSZ, 1962-63
and 1979-80 373
Table 11.3. Monthly Distribution of Rainfall at Kano
(June-September), 1906-85 379
Table 11.4. Sample Budget, Ungogo District, 1964 386
Table 11.5. Some Comparisons between Central and Outlying
Villages in the Kano CSZ, 1964 387
xii / Tables
Table 11.6. Nonfarming Incomes in the Kano CSZ, 1979-80 389
Table 11.7. Trajectories of Change 393
Table 12.1. Case Study Comparison 404
Table 12.2. Summary of Trajectories toward Sustainable
Agriculture by Case Study 425
The question of population growth looms large in the future of Africa.
The editors of Population Growth and Agricultural Change in Africa
observe in their introductory chapter that the estimated population of
530 million people in 1989 will double by the year 2015. This increase
will take place within a context of decreasing per capital food production
over the last several decades. What does the future hold with these two
trends in play? While noting that both technical and political causes ex-
plain decreasing productivity, the editors are more interested in under-
standing the processes that lead to intensification of agriculture, for it is
by intensification that most increases in agricultural productivity will
The contributors examine the relationship between population growth
and agricultural intensification through a case study approach that is
sensitive to historical data. Through an examination of different high-
density populations in Nigeria, Rwanda, Tanzania, Uganda, and Kenya,
it is possible to see changes in technology and productivity over time
and to isolate the conditions under which agricultural intensification fol-
lows population growth and is sustainable in the long term. In this ap-
proach the book offers concrete examples of positive relationships
between population growth and agricultural intensification, examples
with important implications for policy and land-use planning. It is en-
couraging to see that agricultural histories, especially those that capture
processes of rapid population growth that imitate today's trends, can
have a positive impact for planning to meet the future needs of African
The Center for African Studies sponsors the annual Carter Lecture
Series, which are devoted to critical issues facing Africa today. It hosted
and helped to sponsor the workshop/conference at which the papers in
this volume were discussed and reviewed by the authors and several out-
side discussants. These papers address two of the most important and
interrelated problems on the continent today.
The center is grateful for the administrative support provided by
University of Florida Provost Andrew Sorensen. Thanks go as well to
George Bedell and Walda Metcalf at the University Press of Florida,
xiv / Foreword
who encouraged publication of this volume when financial conditions
and reorganization of the press seemed to dim the possibility that it
would be a book. We are also grateful to Billie Turner and Robert Kates
for their willingness to support Goran Hyden's suggestion that the center
undertake this important publication project.
Center for African Studies,
University of Florida
Perhaps nowhere in the world does the relationship between population
growth and agricultural change currently seem so important as in sub-
Saharan Africa. The two trajectories appear to be on a course for disas-
ter in this subcontinental region; population growth is the most rapid in
the world with little sign of abating in the near future, while per capital
food production continues to decline. In both its present-day and its
pending magnitude, the widening gap between population and produc-
tion has been of central concern in policy analysis and academic re-
search alike. Numerous factors have been invoked to explain the problem:
the subcontinent has a disproportionate amount of marginal or fragile
land for cultivation; it has not received sufficient funding directed at en-
hancing food production; it is woefully lacking in rural infrastructure; it
is suffering from underdevelopment, and so on. The evidence supporting
any of these arguments is surprisingly slim, however; few systematic and
comparative treatments have been carried out in Africa. Moreover,
many of these arguments are grounded in the major theoretical perspec-
tives on population-agriculture relationships. Some see the African case
as supporting Malthusian-like principles of rapid population growth
outstripping growth in agriculture. An alternative position, supported by
a considerable body of data drawn from around the world, posits popula-
tion growth as a prerequisite for agricultural growth (measured by the
intensity of cultivation). Interestingly, several studies indicate a positive
relationship between population growth and agricultural intensification
in African farming communities.
The basic question of this study is whether population growth in
densely settled areas of rural Africa has led to the intensification of agricul-
ture. Densely settled areas are examined because they offer examples of
the land-pressure conditions that are to be expected throughout much of
Africa if population growth does not abate. Hence they may offer many
insights into the future of African agriculture. This basic question is
then used to explore those factors that influence the population-agriculture
xvi / Preface
relationship and its consequences for environmental and economic
This question and its implications are addressed through existing
data as interpreted by the experts who provided them, operating under a
common protocol of study. The editors commissioned ten case studies
on the basis of background work defining and identifying areas of high
rural population in sub-Saharan Africa and matched these areas with
researchers who had studied them in relation to the themes of the
volume. The resulting papers were presented at a workshop at the Uni-
versity of Florida, Gainesville, April 30-May 2, 1988. The workshop fo-
cused on lengthy discussion of individual papers and "break-out" groups
assigned to address common themes and lessons that might be drawn
from the papers. Joining the authors at the workshop were the editors
and a select group of commentators, all long and intimately acquainted
with issues of population and agriculture in the subcontinent: Richard
Bilsborrow, University of North Carolina, Chapel Hill; Ronald Cohen,
University of Florida; Gregory Knight, Pennsylvania State University;
Philip Porter, University of Minnesota; and Donald Vermeer, George
Washington University. All papers were redrafted after the workshop.
Full compliance with the data demands of the protocol proved diffi-
cult for each case study. The data gaps that result, as well as the varying
quality of the data that were obtained, inhibit strict quantitative anal-
yses and leave a number of questions unanswered. These problems
noted, the individual chapters provide empirically rich assessments of
the population-agriculture relationship, and comparisons among them
provide important insights to the suite of questions posed about the
population-agriculture relationship and their meaning for the subconti-
nent in general. We believe that they offer a base from which further
studies can be launched and further insights gained.
This effort was sponsored by a grant from the Rockefeller Founda-
tion. Financial assistance for various facets of the project was also pro-
vided by the Center for African Studies, University of Florida, the Alan
Shawn Feinstein World Hunger Program, Brown University, and the
Graduate School of Geography and the George Perkins Marsh Institute,
Clark University. We thank the various individuals in these programs
who assisted the project during its preparation and implementation, es-
pecially David Mazambani, Mark Johnson, and Richard Nolan. Viola
Haarmann deserves special recognition for her assistance during the edit-
Preface / xvii
ing phase of the project. The graphics were prepared by Margaret
Pearce and Anne Gibson of the Clark Cartography Lab.
B. L. Turner II
Robert W. Kates
Providence, Rhode Island
1 / Theory, Evidence, Study Design
Robert W. Kates, Goran Hyden, and
B. L. Turner II
Over the next century, Africa's population may increase fourfold before
stabilizing, creating densities of population unprecedented on the conti-
nent. Many expect that this situation will severely compound the current
agricultural crisis; others believe that it will stimulate agricultural growth
through the intensification of agriculture, leading to improvements in
food availability and to general economic development. Evidence for
both views has been found. Throughout most of the developing world,
the intensity of agriculture and greater land productivity is broadly as-
sociated with higher population densities. Against this generality, how-
ever, are many instances in which the intensification process has not led
to the types of adjustments in agriculture that improve food availability
or general quality of well-being. Indeed, stagnation, involution, and en-
vironmental deterioration may be equally associated with increasing pop-
ulation density and related agricultural practices. Given the seeming
inevitability of rapid population growth in Africa, therefore, it is partic-
ularly important to understand those situations that lead to positive
conditions of intensification-improved food supply, well-being, and
sustainable agriculture-and those that lead to negative conditions of
stagnation and of environmental degradation.
There are at least four theoretical traditions that yield differing ex-
pectations of the relationship between population and agricultural growth.
Broadly interpreted, they view the relationship optimistically or pessi-
mistically and ground much of the argument within the roles played by
population pressures or economic development (fig. 1.1). Seen as a local-
ized process, especially among subsistence-oriented economies, neo-
Malthusians have negative expectations-population has the potential to
outstrip agricultural change, inducing land fragmentation, environmen-
tal deterioration, poverty, and famine (MacDonald 1989). In contrast,
Boserupians (following Boserup [1965 and 1981] and others) have posi-
tive expectations-population growth is a stimulus for an intensification
2 / R. W. Kates, G. Hyden, and B.L. Turner II
Population Growth under Agricultural Conditions of
Fig. 1.1. Expectations of the population-agricultural growth relationship: theoreti-
of agriculture, which is the triggering mechanism for higher levels of
productivity owing to the technological change and division of labor
that accompany the process. This view blends into the larger context of
markets and external relations of the neoliberal economists who posit
outcomes dependent on an appropriate economic structure that provides
rewards to the individual farmer for intensifying production, encourages
specialization, and facilitates entry into "free" markets (see Schultz 1964;
Mellor 1966). Finally, the market context also gives rise to two pessimis-
tic views. Neo-Marxist social scientists argue that the colonial legacy,
the current international reach of capitalism, and class-based national
politics have created political and economic conditions that inhibit the
Theory, Evidence, Study Design / 3
development spinoffs of intensification, even if they were to occur (see
Watts 1989; Bernstein et al. 1990). Alternatively, neo-Malthusians warn
that rapid population growth, even in well-developed market economies,
leads to pressures on world resources that potentially threaten the global
carrying capacity (see Ehrlich and Ehrlich 1990).
Fortunately, we do not need to rely solely on generalized observa-
tions or on speculative theory to ask whether the extraordinary popula-
tion changes currently taking place in Africa offer opportunities for
encouraging production or simply compound existing problems. There
are today, in many parts of Africa, areas of rapid population growth and
high population density that can serve as case studies or "natural exper-
iments" for examining this question. In this volume, we select a set of
such high-density population areas and examine how, over time, the in-
tensity of agriculture changed, how such changes came about, and what
have been the immediate and longer-term consequences of such changes.
No study can provide equal weight to all elements of a theme as
complex as this one, and we make no attempt to do so here. Our pri-
mary objective is to begin at the beginning, so to speak, focusing on the
first issue through two principal queries. Has population growth in
densely settled areas of sub-Saharan Africa led to agricultural intensifica-
tion, and, if so, under what local environmental and socioeconomic
conditions? Issues of the larger national and international socioeconomic
orders are not explored in as much detail, although their importance
emerges in several of the case studies and is discussed in the concluding
section of this book. Our approach thus counters the emphasis of the
general literature on sub-Saharan Africa that has articulated the impact
on agriculture of the sociopolitical history of the subcontinent and has
paid less attention to local and regional comparisons or to the local dy-
namics of agricultural change (see MacDonald 1989: 207).
Africa, south of the Sahara, is composed of 47 countries, whose 1992
population is estimated at 507 million people with an overall density of
23 people per square kilometer, and growing at a rate of 2.9 percent per
annum (Population Reference Bureau 1992). This population is expected
to double by 2010 and to reach 1.7 billion by 2055 before leveling off,
according to World Bank projections that assume a universally declining
4 / R.W. Kates, G. Hyden, and B.L. Turner II
fertility rate beginning in 1995. Using other assumptions, the population
in 2055 may be 25 percent lower or higher (Chen 1989: 2-3). Nonethe-
less all projections agree that, barring major catastrophes, over a period
when the world population will probably double, African population will
grow twice as fast-a rate of change unprecedented for any continent
within human history that places unprecedented demands on the growth
of African agriculture.
By the middle of the next century, six African nations may have pop-
ulations greater than 100 million, larger than any of the major European
powers today, and the population of Nigeria alone may exceed that to-
day of all of North America. Africa's population density, however, will
still be relatively low compared to that of Asia: about 75 people/km2,
about the density of Greece today (Chen 1989). The great growth in Af-
rican population, then, is taking place on a continent that is still rela-
tively sparsely populated.
The Crisis in Agriculture
Between 1965 and 1985, food production grew in sub-Saharan Africa by
54 percent, but per capital food production declined by 12 percent accord-
ing to standard sources (WRI 1988-89). Although the accuracy of such
estimates is questioned, there is a general consensus that the aggregate
data used to estimate continentwide production trends indicate circum-
stances that do not bode well for the future (see Singh 1983; Ho 1985).
The African case, then, stands in stark contrast to the remainder of the
developing world where per capital food production at the international
scale has risen dramatically in Asia, and maintained itself in the Near
East and Latin America (fig. 1.2).
Many causes for this lag in agricultural production have been sug-
gested. They can be divided into two categories, the technical and the
political, dominated by the international development and the academic
communities, respectively. New technologies were emphasized in interna-
tional development circles in Africa during the 1960s and early 1970s in
an attempt to replicate the "green revolution" successes of Asia. This
approach largely failed, and alternatives were sought. One alternative
concentrated on developing a better understanding of Africa's indige-
nous farming systems (see Harwood 1979; Moock 1986; Dommen
1988), including its environmental basis (Porter 1979); another focused
Theory, Evidence, Study Design / 5
Fig. 1.2. Index of per capital food production. Figure adapted from World Re-
sources 1988-89. Copyright 1988 by the World Resources Institute and the Interna-
tional Institute for Environment and Development in collaboration with the United
Nations Environmental Program. Reprinted by permission of Basic Books, a divi-
sion of HarperCollins Publishers, Inc.
on strengthening its physical and economic infrastructure (see Mellor
1984). During the 1980s, these approaches were complemented by explo-
ration of Africa's complicated and multifaceted land-tenure systems (see
Noronha 1988) and by a strong plea for the importance of the long-term
development of the continent's own research system (see Eicher 1988).
Subsequently, interests have shifted toward support of Africa's low-
6 / R.W. Kates, G. Hyden, and B.L. Turner II
resource farmers (Office of Technology Assessment 1988)-the unimo-
dal approach advocated for a long time by Johnston and his colleagues
(see Johnston and Clark 1982)-and this shift has been paralleled by as-
sessments of past practices by the poorest farmers (see Lipton 1989).
These approaches have fed back to farmers and communities through
projects emphasizing low-capital technology and labor-based land im-
provements (see Rocheleau, Weber, and Field-Juma 1988). These ap-
proaches have also been viewed in a population-resource perspective
(Ho 1985; Lele and Stone 1989), leading to recommendations that em-
phasize the targeting of regions with high agricultural potential for the
development of high-yielding and high-valued crops.
Much of academic debate has been dominated by economists and
political economists. The latter have consistently argued that Africa's
agriculture suffers because of the structure of the global economy and
the exploitation that it gives rise to, by both the rich over the poor na-
tions and the privileged over the underprivileged classes (see Dumont
1962; Wisner 1988). In the 1980s, this approach ventured beyond these
concerns, focusing on gender relations, household-labor processes, and
resource management (Watts 1989).
The former, or neoliberal economic perspective, has explained the
situation in African agriculture in terms of the pace of demands on food
production outstripping the ability of farmers to make technological and
procedural adjustments (see Pingali, Bigot, and Binswanger 1987); the
parallel slow response of "induced innovation" owing to the rather late
development of large-scale, intensive pressures on agriculture; or the use
of inappropriate or alienating technologies introduced from abroad and
not adapted to African conditions (Levi and Havinden 1982:16; Mellor
1984; Dommen 1988). Others emphasize the disincentives to African
farmers, such as the complicated and multifaceted land-tenure systems
(see Netting 1968; Noronha 1988) that may inhibit land improvements;
the artificially low prices paid for staples, a policy followed by local gov-
ernments to ensure "stability" in urban areas (see Schultz 1978); the
negative impacts of inadequate infrastructure on agricultural behavior in
general (Lele 1988); and the variable and generally low market value of
many African staples (World Bank 1984).
Cross-cutting these explanations have been the contradictory "les-
sons" derived from the African experience (Huss-Ashmore and Katz
1989). These demonstrate the resilience, flexibility, and "successes" of
some groups of African farmers through adaptation strategies (see Hill
Theory, Evidence, Study Design / 7
1963; Netting 1968; Coulibaly 1978; Mortimore 1989) and the role of lo-
cal institutions (see Harrison 1987) that allow resistance to negative
forces of change (see Scott 1985).
Although we do not reject any of the broad or specific explanations
noted above, our experience suggests a number of population-related
factors impeding the development of agriculture in Africa that recur
across multiple cases. These include the overall low density of popula-
tion and dispersed settlement patterns, the general absence of effective
localized markets for both inputs and outputs, the rudimentary trans-
portation network that connects a diffusely settled population, the in-
adequate incentives to adopt more demanding (in time and risk) tech-
nologies, and the existence of large areas of potential agricultural and
pastoral lands denied by eradicable disease, such as onchocerciasis or
trypanosomiasis, that could be made productive by higher population
through sustained land use. Such factors have not been thoroughly ad-
dressed or developed into a multicausal framework.
In short, we enter the 1990s with no macrotheory or approach to ex-
plain the agricultural crisis. Africa and the world have come to recog-
nize that there is no single solution to the crisis in Africa's agriculture,
just as there is no concise or simple theory to guide that solution. One
avenue that deserves further investigation is the role of the intensifica-
tion of agriculture in African farming and food crisis. Intensification is
fundamental to increase agricultural production, which is a critical
component of the crisis.
Agricultural growth takes place either through expansion-the extension
of land under cultivation-or intensification-the increased utilization
or productivity of land currently under production. About 80 percent of
the growth in African agriculture has come from the extension of the
area under cultivation (Paulino 1987), much of it on lands once consid-
ered too marginal to cultivate. If Africa is to feed its growing population
and retain its much-needed agricultural export trade, a better balance
will be needed between production gains from the expansion and inten-
sification of agriculture; and surely most of future agricultural growth
will come from more intensive use of the prime agricultural lands (see also
Lele and Stone 1989).
8 / R.W. Kates, G. Hyden, and B.L. Turner II
Theories of Intensification
Theories and themes that seek to explain agricultural intensification can
be placed in two broad categories: those that relate production to
household needs and wants (usually population-driven) under conditions
of "subsistence" and those that relate production to demands from the
market. These theories and themes are potentially complementary, as
each emphasizes one part of the reality experienced by most African
farmers; production for both direct consumption and commodity sale
(see Brush and Turner 1987).
Much of contemporary consumption-based or "needs" theory has its
origins in the works of Chayanov (1966) and Boserup (1965, 1981). It
asserts that farmers are responsive primarily to the biological needs of
the immediate population that they must feed, defined, of course, by cul-
tural standards of acceptable consumption. Output is achieved through
the least-effort means perceived by the farmer and is limited by the im-
mediate need. As the population-land ratio increases, farmers are "forced"
to employ greater labor and technical inputs to achieve greater produc-
tion. In this case, output per unit of area land grows (a reverse process is
possible, however). This growth does not necessarily improve per capital
production. For improvement to take place, technological change is
normally required. In the early or extensive phases of a growth trajec-
tory, these changes may involve "traditional" land improvements (for
example, irrigation or drainage) in which land transformation takes
place; in the later phases they involve modern bioscience-derived inputs
(for example, high-yielding varieties). Importantly, these technological
changes do not necessarily take place in a linear or uniform phasing in
response to demands, particularly for those changes that require large in-
vestments in labor. Rather, thresholds of demand must be met before
the investment is made, and this leads to a stair-step growth pattern in
agriculture (see Robinson and Schutjer 1984). Each "stair step" in agri-
cultural growth involves major improvements in land productivity and,
presumably, improvement in per capital production. Continued pressures
on the new agricultural system result in diminishing returns until a new
threshold level is reached, leading to a new technological system.
This segment of the theory-an elaboration of Boserup's original
thesis-focuses on the behavior of farmers in consumption production
and explains why total output is increased in relation to population
stresses placed on the farmer. As such, it can be seen as a subsistence or,
perhaps more accurately, a "need" theme. In this case, need refers to bio-
Theory, Evidence, Study Design / 9
logical and social factors that set, for example, the expected standards
of consumption, combined with satisfying redistribution and survival
insurance responsibilities. The less-developed segment (see Boserup
1965, chap. 12; 1981) of the theory asserts that the intensification pro-
cess leads to specialization in labor and preconditions for "exchange"
and market production. This stage of "growth" presumably moves be-
havior more toward that described in the neoclassical tradition involving
the "market" theme of demand on agriculture.
The market or commodity theory applied to Third World farmers
has its modern origins in the pioneering works of Tax (1953), Schultz
(1964), Wharton (1969), and other agricultural economists and eco-
nomic anthropologists. It asserts that once farmers accept commodity
production, they respond to market demand within the constraints
placed upon them, maximizing production to the level of maximum re-
ward. This theory has been modified through the critiques of Lipton
(1968), Collinson (1972), Scott (1976), and others, who demonstrate that
economic efficiency in the allocation of resources, strictly defined, can-
not be achieved by such farmers. Rather, farmers tend to be risk averse,
at least in regard to some minimum production needed for survival (see
Levi and Havinden 1982) and, therefore, respond in a "proficient"
manner (Schluter and Mount 1976). These behavioral themes are linked
to a large number of broader economic and agricultural models (for
growth stage or conservation models, see Hayami and Ruttan 1985),
leading to specific views about the trajectories that agriculture will take.
These have been merged into a theory of "induced innovation," in which
technological and institutional changes required to develop agriculture
are endogenously derived as a result of changes in resource endowments
and demand (Binswanger and Ruttan 1978; Hayami and Ruttan 1985).
Pure applications of each theory have largely ignored the others.
Those interested in the consumption or needs side have emphasized its
elaboration; many working on the commodity or market side have as-
sumed that even the slightest market engagement makes farming behav-
ior, and hence agricultural development, subject to commodity theory.
Yet many instances have been documented in which household agricul-
tural functions are divided between consumption and commodity crops
and fields, presumably with different motivations involved in each (see
Gray 1988). And this logic can be extended to the multitude of "devia-
tions" recognized between the commodity theory and the reality of the
farming unit-that is, these deviations follow from not only constraints
on farmers but also a behavior that is fully predicated neither on the
10 / R.W. Kates, G. Hyden, and B.L. Turner II
commodity nor on consumption theories. This has led to a consumption-
commodity or "induced-intensification" theory (see Ali and Turner n.d.),
particularly among those outside economics; it asserts that farming be-
havior for most Third World farmers is predicated on a composite of the
consumption and commodity rationales (see Fisk 1964; Grossman 1985;
Turner and Brush 1987) and that agricultural change, including techno-
logical and institutional changes, is driven by the joint demands placed
on it. On the consumption side, intensification is induced largely through
mechanisms endogenous to the farmer and, therefore, is the product of
indigenous experimentation and accumulated wisdom. On the commod-
ity side, however, intensification is induced by mechanisms exogenous to
the farmer and ultimately is largely associated with technological changes
from commercial and public experimental institutions, following the
induced-innovation theme. Drawing upon themes that focus on these at-
tributes of agricultural behavior in the face of increasing demand seems
particularly appropriate for our study; most African smallholders who
produce directly for their own needs and for markets have demonstrated
considerable resilience to constraints and demands (Pingali, Bigot, and
Binswanger 1987) and have innovatedd" changes in agriculture through
landscape transformations as a response to local demands and, more re-
cently, through other types of technological and institutional changes as
a response to market demands.
Measures of Density and Intensification
Characterizing the relationship between population growth and agricul-
tural intensification often involves comparing measures of population
per unit area for a given period of time. Population density is usually ex-
pressed as the number of persons per unit area of land, of potential ara-
ble land, or of land employed in agriculture. The measure of agricultural
intensification has taken on a rather precise meaning as the total produc-
tion per unit area and time (typically per hectare and year). Its obvious
measure, therefore, should be that of total output. Owing to several
complications and to the paucity of data at the local level, surrogate
measures are commonly employed. The most common two are the fre-
quency of cultivation and the type and number of agrotechnologies
(Turner and Doolittle 1978).
Study of the relationship between population density and agricul-
tural intensification is complicated by several problems, most of which
Subsistence Needs Path
Market Demand Path
Fig. 1.3. Pathways from agricultural growth to improved well-being
12 / R.W. Kates, G. Hyden, and B.L. Turner II
involve the pros and cons of the measure used and, for Africa, the qual-
ity of the data on which they may be based. For a measure of popula-
tion density, "raw" density or population per total area in question (for
example, village or district) fails to account for the great range in land
quality and use that may occur within the area. A low "raw" density
may be generated by the presence of much land highly unsuited for cul-
tivation, and therefore, the "pressures" on the cultivated lands are not
adequately captured. "Relative" density, population per selected area
thought relevant to agricultural production (often expressed as arable or
cultivated land), is an improvement but also can be deceiving. What
constitutes arable land (or cultivated land) varies by changes in technol-
ogy and need. Of the measures, data for the calculation of "raw" density
are typically the easiest to obtain, but the "relative" measure per culti-
vated land may best reflect agricultural pressures if cultivation is fun-
damental to the household economy.
The best direct measures of agricultural intensity in terms of total
output per unit area and time also have problems related to the multi-
plicity of products, times, and conversion measures used to equate them.
Data for systems of mixed cropping (numerous species in one plot) and
for multiple functioning fields, orchards, and livestock zones are difficult
to obtain. Various means can be used to create an overall set of output
figures, but these then require reduction to a common denominator for
effective comparisons. Cultigens vary from food to fiber to condiments
and are produced for both direct consumption and sale. Energy mea-
sures kilocaloriess produced) are inappropriate for fiber crops, and
weight or volume measures say little about the value of production.
Root crops, for example, produce much more weight per hectare than
do cereals, but the caloric value of cereals may be much greater. Mone-
tary measures, of course, are strongly affected by the vagaries of the
market, including instances in which major differences exist in the worth
of the same produce depending upon its use: for instance, the amount of
food purchased with the cash obtained from the sale of the harvest is
usually not equivalent to the amount of food sold. This said, an ade-
quate measure may be obtained by standardizing the entire harvest to a
staple-food equivalent (for example, maize, sorghum, or rice).
Because good production data are so difficult to obtain, the conven-
tional surrogate measure of agricultural intensity is the frequency of cul-
tivation (after Boserup 1965) as shown on the top line of figure 1.4.
Calculating the frequency of cultivation-the number of harvests per
plot over a standard time frame-is complicated by the prevalence of
o o o
S o o
o 0) ~~
m m nm
3 & &
E 5 H
14 / R.W. Kates, G. Hyden, and B.L. Turner II
multiple cropping (more than one harvest per plot per year), the simul-
taneous employment of various fallow practices within the same system,
and the use of species that need not be planted and harvested within a
single cultivation cycle or season (for example, manioc/cassava). Al-
though descriptive categories can be used-for instance, long fallow or
multiple cropping-they are imprecise measures and are not well suited
for comparative analysis. Typically, then, frequency is calculated in
terms of a percent of time that a plot of land is under cultivation. De-
spite its problems, the frequency measure is popular because of the
broad correspondence between total land productivity (not yield) and
cultivation frequency, especially if similar levels of technology are
Beyond frequency of cultivation, specific mixes of technologies have
been used as surrogate measures. For example, one study on African
agriculture has compared operations employed in different farming sys-
tems with increasing frequency of cultivation (Pingali, Bigot, and Bins-
wanger 1987). The use of these operations and their implied technolo-
gies as indicators of intensification is complicated by the disputed role
such technologies are seen to play within extensive agriculture. This is
so because high amounts of labor and procedures can be used as substi-
tutes for technology in some systems; some low-tech systems are more
land-intensive than are high-tech ones (see Turner and Brush 1987: 7).
Even such commonly accepted indicators of intensification as the irriga-
tion works and agricultural terracing found in historical and archaeolog-
ical studies do not necessarily imply widespread intensification in the
African context (Sutton 1984). Within "low-resource" agriculture, subtle
procedures and technologies, constituting creative recombinations of ex-
isting productive inputs, are commonly employed (Richards 1985; Dom-
men 1988; Office of Technology Assessment 1988), and these types of
"technology" are not readily accountable in technological measures.
Evidence of Intensification with Increasing Population Density
Consistent relationships between population density and agricultural in-
tensity (measured through either the frequency of cultivation or output
per unit area or time) have been demonstrated for many parts of the
world (Brush and Turner 1987). These include village- or regional level
comparisons in the tropics (Dalrymple 1971; Turner, Hanham, and Por-
tararo 1977) and highland New Guinea (Brookfield 1962; Brown and
Podolefsky 1976), and intra-village comparisons in Central America
Theory, Evidence, Study Design / 15
(Bartlett 1976) and Southeast Asia (Sahlins 1971). Also, although not
addressing the intensification theme per se, other studies have produced
comparative evidence that strongly supports the general trends between
population density and agricultural intensity (see Ruthenberg 1980).
Few comparative studies have found the relationship lacking, and
those that have are controversial because of methodological and other
problems; for example, Metzner's (1982) findings on the Isla de Flores in-
volved perennial tree crops, a situation in which the relationship was
never intended. In an interesting alternative, Richards (1985, 1987) has
shown that loss of population has led to intensification in parts of West
Africa. This situation results from the migration of males-those who
clear the forests for extensive systems-and the increased significance of
female labor, which is employed to increase output on land already
cleared for cultivation.
A number of studies have identified positive relationships between
increased population (or agricultural demand) and agricultural intensifi-
cation in Africa, especially in West Africa, and at least four general
treatments of the subject exist: Cleave (1974) and Pingali, Bigot, Bins-
wanger (1987) for pan-Africa, and Gleave and White (1969) and Net-
ting, Cleveland, and Steir (1980) for West Africa. Examinations in the
Sahelian zones have been made by Haswell (1953), Grove (1961), Mor-
timore (1967), Benneh (1972), and Norman (1974). Other support comes
from Vermeer (1970) for the Tiv of Nigeria and Basehart (1973) for the
Matengo of Tanzania. The most detailed treatment of the subject, how-
ever, is of the Jos Plateau of Nigeria, where Netting (1968) found popula-
tion density to be strongly associated with the intensity of cultivation.
One of the few contrary interpretations has been offered by Datoo
(1976) for agriculture in the Uluguru Mountains of Tanzania. It is not
clear, however, whether his specific argument employs too narrow an in-
terpretation of intensification or is a case of agricultural involution as
discussed below. More recently, Lele and Stone (1989) have examined
agricultural change in six African nations-Cameroon, Tanzania, Sene-
gal, Kenya, Malawi, and Nigeria-and have concluded that population
growth has not led to agricultural growth because it requires evolution-
ary processes under conditions of low population density and slow popu-
lation growth, while these nations, and most of Africa, are experiencing
very rapid population growth. They do not reject the overall Boserup
theme but argue for two conditions of intensification: an evolutionary
condition (the Boserup and other demand themes) and a policy-driven
condition of high population density and growth.
16 / R.W. Kates, G. Hyden, and B.L. Turner II
The general thesis raised by Lele and Stone is an important one, es-
sentially addressing cases of agricultural involution and stagnation (see
Geertz 1966). They capture many of the mediating factors that can
deflect or alter the trajectories of agricultural change in the modified-
demand theme. Several of their conceptual interpretations and the logic
employed are somewhat confusing to us, however, making their findings
unclear and, therefore, questionable. For example, much of their skepti-
cism is based on the original Boserup (1965) argument, not on its many
elaborations (as noted above), which can account for the adjustments in
definitions and measure they use (Lele and Stone 1989: 5).
More important, Lele and Stone must reconcile their interpretive
base-the impact of conditions of high population-density growth rates
on agricultural change-with the fact that most of Africa is sparsely pop-
ulated by world standards. To do this, they interpret population pressure
in the African context in terms of population per unit of arable land, ac-
cepting the premise that sub-Saharan Africa is dominated by marginal
agricultural lands and that, therefore, overall low "raw" population den-
sities are relatively high. We suggest that this and other such environ-
mental premises must be weighed cautiously. Cultivation in the tropics
in general offers both constraints and opportunities, and we are unaware
of any comparative studies documenting that, on a subcontintental ba-
sis, sub-Saharan Africa fares more poorly than any other tropical area
in terms of agricultural environments. Indeed, now, as in the past, nu-
merous cases of "undersettled" prime agricultural lands and densely
settled lands of lesser intrinsic agricultural quality can be found in sub-
Saharan Africa (Dommen 1988: 115-17). What must be recognized is
that the quality of an agricultural environment is as much a product of
its use, including landscape transformations, as of "raw" nature: the
polders of the Netherlands (below sea level), Andean terracing on ex-
treme slopes, the raised fields throughout the wetlands of highland New
Guinea, and the wheat fields on the arid Great Plains of North America
are examples. The case for Africa is commonly biased toward undocu-
mented explanations of environmental constraints to cultivation rather
than understanding why these environments have not been "reshaped"
to be upgraded.
Finally, Lele and Stone ignore the extensive literature supporting
the rudiments of the demand themes for African agriculture, including
cases from the very countries considered in their study. These have been
Theory, Evidence, Study Design / 17
These problems aside, Lele and Stone also confront the second
phase of the intensification theme-the movement or absence of move-
ment from increased food production to increased standard of con-
sumption or well-being-and detail some of the shortcomings of the
evidence for it in the six nations that they examine. Here they recall a
commonly overlooked statement by Boserup (1965: 118) that the inten-
sification process may not work under conditions of high population
densities and rapid population growth, the very conditions that we ad-
dress in this book. Such conclusions have also been implied in recent
studies of South Asia (see Rodgers 1989), although studies by Boyce
(1987) and Ali and Turner (n.d.) have demonstrated that agricultural in-
tensification has kept pace with population growth in parts of this re-
gion. Ho's (1985) interpretation of regional and national data from
sub-Saharan Africa concludes that the frequency of cropping has in-
creased but that the technological changes (substitutes) required to sus-
tain this trend have been inadequate, contributing to environmental
degradation. Like Lele and Stone, Ho suggests that this lag in techno-
logical development is associated with the extremely high rates of in-
crease in population.
At this juncture, we can only speculate on how to weigh the relative
importance of population and market access in the light of the induced
intensification theme (see Pingali, Bigot and Binswanger 1987). There is
clearly some positive density feedback between the two-market access
invites immigrants to an area, and high population density encourages
the development of a marketing infrastructure. Recent studies from
Kenya (Goldman 1987) and Zimbabwe (Gray 1988; Rohrbach 1988)
underscore the positive interplay between the two forces of market and
population, with a developed infrastructure for commodity production
accelerating efforts to intensify agriculture in response to increasing pop-
ulation density. Thus commodity production surely modifies the simple
relationship between population density and agricultural intensification.
Environment also modifies the relationship. Levi (1976), preceding
Lele and Stone, challenged the positive relationship between population
18 / R.W. Kates, G. Hyden, and B.L. Turner II
pressure and agricultural intensity as long as pressure is dependent
solely on population density, not density as mediated by environmental
conditions (land quality). Brookfield (1962, 1984) and Turner, Hanham,
and Portararo (1977) have shown that the density-intensity relationship
is strongest where environmental constraints to land quality are moderate-
perhaps most cultivated lands. Extremely severe constraints either im-
pede agricultural growth or require such high levels of inputs (landscape
changes) to overcome them that the resulting intensity is higher than the
population density warrants (Turner, Hanham, and Portararo 1977;
Pingali 1987). Conversely, optimal, highly productive environments ex-
aggerate intensity, if measured as output or frequency. Pingali, Bigot,
and Binswanger (1987) show that, at the extremes, intensification in
Africa is inversely related to rainfall. Maintaining agriculture in low-
rainfall areas (<750 mm) requires the use of intensive and expensive (in
labor or capital) technologies. Conversely, the lowland high-rainfall
areas (> 1200 mm) suffer from thin soils, leaching, and acidification that
inhibit continuous cultivation. Lele and Stone (1989: 16-17) argue that
both rainfall and soil conditions throughout much of sub-Saharan
Africa are constraints to agricultural improvement. But although environ-
ment undoubtedly affects the demand-agriculture relationship, African
farmers have demonstrated an ability to support substantial populations
in conditions described by some (see Matlon 1987) as especially suscep-
tible to degradation. Indeed, high population densities have been sus-
tained longer in parts of Nigeria and West Africa on soils thought to be
especially constraining to cultivation than almost anywhere else in the
In sum, there is ample evidence to conclude that the seemingly inev-
itable population increases in sub-Saharan Africa can play a forcing role
for the needed intensification of agriculture. But to do so does not nec-
essarily lead to the adoption of the pronatalist stance of Simon (1981)
that sees in population increase the growth of "the ultimate resource."
The ideal conditions on which the demand themes (fig. 1.3) are based
are not the only operating conditions, as we detail below. Therefore, al-
though population growth and increasing density seem inevitable in sub-
Saharan Africa, a beneficial density-intensity development is not.
Theory, Evidence, Study Design / 19
Involution, Diminished Well-Being, and Environmental
Intensification can lead to involution and the diminution of economic
and social well-being or threaten the long-term sustainability of agricul-
ture through environmental deterioration. The theory of involution, pro-
posed by Geertz in 1963, characterized the condition in which increasing
demand is met by output intensification but at the costs of decreasing or
small marginal and average returns to inputs. In addition to the original
studies on Java, conditions that are involution-like have been found in
South Asian systems characterized by extremely high densities, intensi-
ties, and environmental constraints (Ahmad 1985; Ali 1987). In Bangla-
desh, for example, population pressures have induced extremely high
frequencies of cultivation, but the combination of rapid population
growth and extreme environmental problems (for example, deep and pro-
longed flooding) that cannot always be countered by the farmer or vil-
lage has led to "food" stagnation-a case in which output increases may
not match increases in demand (Boyce 1987; Ali and Turner n.d.). For
Africa, Martin's (1987) historical assessment for parts of Nigeria and
Rwanda, where population densities are high and have been so for a
long time, asserts that involution and perhaps stagnation is evident.
Involution-like or stagnating conditions also follow from the recent in-
terpretations of Ho (1985) and Lele and Stone (1989) for sub-Saharan
Africa in general (also Datoo 1976, 1978; Jewswiecki and Chretien
Similarly, intensification can lead to real losses of social, cultural,
and economic well-being (Johnston and Kilby 1975; Heyer, Roberts,
and Williams 1981; Hart 1982; Berry 1984). In Africa, more intensive
systems almost always mean more work for the individual farmer, often
female, in contrast to South Asia, where the higher labor requirements
may offer employment to the rural landless. Compared to other peoples,
Africans, especially in their indigenous systems, have had a high degree
of equitable access to resources, especially the common property re-
source of land. With population growth and new, intensive land uses,
competition and conflict over land are growing rapidly, especially in areas
of high density (see Bassett 1988). Credit, technological inputs, and
market opportunities are not equitably distributed and may lead to sys-
tematic differentiation based on class, ethnic origin, or locale (see Watts
and Bassett 1986; Wisner 1988; Lipton 1989). Women farmers in partic-
20 / R.W. Kates, G. Hyden, and B.L. Turner II
ular seem to be at a disadvantage in access to the new technologies and
inputs (Mackintosh 1989; Stamp 1989). Also added to the traditional
risks of pest and climate are new social and economic risks of price fluc-
tuation, the dependability of input supplies, and the availability and
price of purchased foods. New dependencies and relationships are
created that are far beyond the ken and control of African farmers
(Watts 1983, 1989; Wisner 1988).
Illiffe (1987), in a recent study, argues that African poverty is differ-
ent in land-rich and land-scarce societies. In the former, poverty arises
from a shortage of labor needed to exploit the land, and the incapaci-
tated, elderly, young, and isolated suffer most. In the latter, these groups
also suffer, but added to these are those able-bodied persons without ac-
cess to land resources or the ability to get a good return for their wage
labor. As people in some places experience land scarcity, Illiffe foresees
an increase in poverty, as "only slowly [does] possession of a family,
rather than lack of one, become a cause of structural poverty" (p. 6).
Equally important are the negative impacts of intensive agricultural
activities and technologies on the long-term sustainability of the agricul-
tural resource base (Farvar and Milton 1972; Clark and Munn 1986;
Lewis and Berry 1988). Usually, more frequent cropping and less fallow
reduce soil fertility and increase soil loss. These problems can be coun-
tered by the addition of procedural and technical inputs (such as fertiliz-
ers, ground-covering cultigens, and terraces), but the degree to which
these inputs have compensated for the decrease in fallow has been ques-
tioned by Ho (1985) and Lele and Stone (1989). In some documented
cases, diminished returns to these inputs, nitrogen, for example, appear
after a number of years, requiring additional nutrients to restore produc-
tivity (Lal 1987). Irrigation, poorly designed or managed, can lead to sa-
linization and waterlogging of soils and, even when well managed, to
toxic-material accumulation, to an increase in human disease habitats
such as schistomiasis, and to a decrease in natural soil fertility. The use
of herbicides promises to ease the intense seasonal shortages of labor in
African agricultural systems but at high costs, both economically and en-
vironmentally. Similarly, pesticide usage to control ample tropical pest
populations leaves residues on food, threatens the health of applicators,
and leads to pesticide-resistant species (Tait and Napompeth 1987).
Lee (1986) has provided the conceptual groundwork for understand-
ing the trajectories of the interactions among population, technology,
and environment, with an implied outcome of development or well-
Theory, Evidence, Study Design / 21
being. His conclusion is that the trajectories do not grow forever (in our
case, intensification) but show spurts of acceleration and deceleration un-
til a "high-technological-high-population stable equilibrium" is reached,
unless preventive checks arrest growth at some lower level of the
The lesson from this brief survey of population-intensification rela-
tionships and Lee's broader thesis is simple. Long-term population
growth and economic development usually do not take place without in-
tensification and agricultural growth, although intensification and agri-
cultural growth do not inevitably follow population growth and are not
necessarily beneficial or sustainable Thus it is critical to identify the
conditions under which agricultural intensification does follow popula-
tion growth, benefits its practitioners, and is sustainable for the long
Case Studies of Growth and Intensification
Design of a Natural Experiment
Toward this objective, our basic strategy was to use the existing areas of
high density in sub-Saharan Africa as "natural experiments" through
which to examine in detail the chain of relationships among population
growth, increased density, and the intensification of agriculture and, in
less detail, the consequences for well-being and the sustainability of
agriculture. To explore this strategy, we drew upon previous fieldwork
and knowledge, identifying researchers with field experience in high-
density areas, and commissioning case studies from them, prepared with
a common, mutually agreed upon protocol.
In developing and carrying out the study design, the following steps
were begun in 1986: (1) identifying existing high-density areas in sub-
Saharan Africa, particularly that large expanse of land falling within the
tropics; (2) identifying researchers with field experience in these areas,
preferably those having published on agricultural issues; (3) developing a
draft protocol for case studies; (4) commissioning case studies and con-
vening a workshop to review and compare draft studies in light of the
common protocol; (5) preparing and publishing the case studies and as-
sessments of their implications.
This approach was taken for several important reasons. We sought a
22 / R. W. Kates, G. Hyden, and B.L. Turner II
comparative study of regional cases through the use of a "natural exper-
imental" approach, therefore examining the proposed relationships be-
tween population growth and agricultural intensification across a number
of cases and through the complexity of the situations in which they oc-
cur. Unfortunately, standardized data of the kind needed to make this
assessment for the regions in question are simply unavailable, and time
and funding constraints would not permit extended field studies to pro-
duce them. Therefore, we drew upon commissioned case studies pre-
pared by highly knowledgeable field-workers who were familiar with the
data for particular regions and could also provide a depth of analysis
and understanding of the situations that could not be generated from a
"synthesis" of the data, even if they were available. These experts used a
common protocol, thus facilitating comparisons among the cases not
found in isolated village or regional case studies, and were brought to-
gether to assist in the interpretations of their own and other cases.
Drawing and building on their work, understanding, and guidance
proved to be a highly economical method, bringing together what was
already known without embarking on expensive field study. In this
manner, we sought to illuminate the commonalities and differences
among cases of long-term population growth and agricultural intensifica-
tion in sub-Saharan Africa in order to explore these linkages. We use
the insights gained as lessons for understanding agriculture change and
policies geared to improve well-being in terms of food production and
Existing High-Density "Rural" Areas within
In Africa, population data are of highly variable quality and typically
out-of-date; data for small spatial units are particularly difficult to ob-
tain. This posed a problem for the identification of high-density zones or
regions of "rural" occupation that was fundamental to our undertaking.
The rationale for the rural criterion is simple; we seek populations of
farmers responding to major increases in agricultural demand, particu-
larly that emanating from local population growth. Eliminated from
consideration, then, were districts with high densities created by the
count of the urban (largely nonfarming) populace.
The next issue focused on the size of the rural area in question. The
microspatial scale-village or small clusters of villages-was eliminated;
Theory, Evidence, Study Design / 23
innumerable anomalies can be found at this level, and the small size of
the unit would not necessarily capture the dynamics in question. Macro-
scale spatial units-ecological zones, states, and nations-were elimi-
nated because of the great variability of population densities and other
conditions within them. The midscale was sought through the use of the
administrative unit "district." But it is precisely this unit for which ade-
quate population information is so difficult to obtain. As we were lack-
ing a central data source, Johnson (1987) created one for us. He exam-
ined census data sources in various North American archives and
agencies for district-level data for forty-eight countries in sub-Saharan
Africa, excluding the Republic of South Africa, which was eliminated
from consideration. No data or suspicious data were located for nine of
the countries, although three-Lesotho, Togo, and Zaire-may include
districts with relatively high densities. For the remaining thirty-nine
countries, censuses were located dating back to the 1960s in five cases,
as current as the 1980s in eight cases, and with the remainder from the
1970s. He then crudely projected all data to an assumed 1985 base, us-
ing the current estimate of national growth rate by the Population Ref-
Based on the results, four "high" population-density categories were
established (people/km2): > 100-<200; >200-<300; >300-<400; and
>400. Omitting four small island countries and districts, and excluding
small districts centered on major urban concentrations, 14 countries
were found to contain 112 districts with densities >100 (people/km2),
and 9 contained 55 districts with densities >200 (people/km2).
Attention was focused on those districts with "rural" densities >200
people/ km2, modifying the list in an effort to account for districts whose
ranking in this density category was due primarily to urban populations.
In so doing, thirty-nine districts in eight countries were identified as main-
taining the required density level (table 1.1). Of these, however, twenty-
one are suspected to be influenced significantly by urban populations,
and further study may demonstrate that their true rural population is
not > 200 people/km2.
Identifying Potential Collaborators with Field Experience
After this exercise, we searched the literature, doctoral abstracts, and
agency papers to identify, first, authors of field studies especially related
to the issue at hand in the previously targeted districts and, second, au-
Table 1.1. Districts with Rural Population Densities in Excess of 200/km2
Country (census data) Districta Population density (/km2) influence
Burundi (1976) Ngozi 390 x
Bujumbura 385 x
Muramvya 378 x
Cameroon (1976) Mifi 265 x
Kenya (1979) Kisii 505 x
Kakamega 376 x
Kiambu 358 x
Malawi (1977) Chiradzulu 300 x
Blantyre 262 x
Thyolo 240 x
Nigeria (1963) Uyo 585 x
Umahia 373 x
Enugu 365 x
Kano 268 x
Rwanda (1978) Kibuye 416
Ruhengeri 375 x
Gitarama 335 x
Gisenyi 315 x
Kigali 262 x
Tanzania (1978) Moshi 291 x
Ukerewe 270 x
Mwanza 236 x
Uganda (1980) Mable 261
alsland districts and countries omitted.
bUrban areas probably contribute significantly to the district's overall population density.
Theory, Evidence, Study Design / 25
thors of studies anywhere in sub-Saharan Africa addressing the popula-
tion density-agricultural intensity question. Of the fifty-five densely
settled districts originally identified, we located only a handful of studies
dealing with any of these districts, and these studies were not necessarily
of the type needed for our purposes.
With this extensive search but sparse yield, we sought further to in-
crease the number of potential case study authors based on our own
knowledge of African scholarship and our extensive network of col-
leagues. Overall, fifteen potential collaborators were identified with ex-
tensive field experience in high-density areas, ten of whom were able to
undertake the preparation of case studies.
The Draft Protocolfor Case Studies
The ten collaborators were asked to select a site or area with a current
"rural" population of about 200 people/km2 or above for which they
had primary or "first-hand" microlevel data on the dynamics of agricul-
ture. The site could range in area size from a village to a small region or
district. If a village was selected, household data could be used; if a dis-
trict, village-level data were to be employed. In all cases, authors were
able to address the issue at the district or subdistrict (region within a dis-
trict). In most cases, household-level data were employed, usually based
on samples taken throughout the area in question. The time depth sug-
gested for the studies varied within a ten- to fifty-year range: flexibility
was required, depending on the data source. The key was to capture a pe-
riod in which the dynamics of population could be assessed against the
dynamics of agriculture.
For comparative purposes, case study authors were requested to pro-
vide information, where possible, on seven key items: the site; its envi-
ronmental setting; the nature of population change; the nature of
agricultural change; changes in external forces of production; the conse-
quences of the various changes; and near-future trajectories of change
(table 1.2). In addition to this common core, they were encouraged to
share their own unique data and insights as they related to the broad
Natural experiments or field tests are by definition imperfect, depend-
ing as they do on the happenstance of nature and society rather than on
the careful controls of the laboratory. And when the experiment is un-
dertaken in this collaborative mode, depending on insights and data
26 / R.W. Kates, G. Hyden, and B.L. Turner II
Table 1.2. Case Study Protocol
Agricultural change: inputs and outputs
External forces of production
Consequences of change
Future trajectories of change
Approaching or above 200 people/km2
Basic soil groups
Flooding and drainage problems
Common crop pathogens and pests
Age and sex structure
Frequency of cultivation
Yields (per harvest)
Land productivity (per unit area and time)
Value of production
Nonfarm sources of food, labor, and capital
International agency activity
Quality of life/welfare
drawn from the work of diverse scholars-often collected for other
purposes-the common protocol becomes a statement of aspirations
rather than a table of contents. Thus, none of the studies was able to pro-
vide the complete set of data deemed desirable, although each offered in-
valuable insights into the population-agriculture relationship.
Our natural experiment also omits, as do most such efforts, the im-
portant component of genuine controls-sites similar in most respects to
those studied, except for the high levels of population density. If we
could find "all other things being equal" situations, we would have had
an effective set of controls for our natural experiment. We would also be
able to balance the bias toward "successful" cases, places where large
Theory, Evidence, Study Design / 27
densities, once achieved, are sustained. The great failures, places where
large densities were not sustained, if they existed, are by definition miss-
ing from our density-defined criterion for case studies. Thus although we
found identifying criteria eminently desirable, we had difficulty in estab-
lishing criteria from which to compare "all other things" and in obtain-
ing the information that would allow comparisons between our case
study sites and possible control sites. But beyond these difficulties, we
speculate that such sites may not exist because nature and society inter-
act so as to fill common human-ecological niches. Over time, environ-
ments suitable for high population densities fill up.
Ten Case Studies
The selected case studies explore densities that range in size from less
than 150 km2 to more than 10,000 km2. With one exception, they in-
volve district or district-like units. They cluster in the highlands of East
Africa (Kenya, Rwanda, Tanzania, and Uganda) and in Nigeria (fig.
1.5). This clustering follows from both the population geography of sub-
Saharan Africa and the happenstance of regional expertise. Although
other highly populated rural zones exist in sub-Saharan Africa, field-
workers and field studies devoted to population-agriculture relationships
are undoubtedly concentrated in certain areas.
Throughout their history, the highlands of East Africa have been
densely populated compared with the regions that surround them and
with most of sub-Saharan Africa. The highlands offer a multitude of
habitats, most of which have long been filled by different economic
adaptations-agriculture, pastoralism, and mixes of the two. For the
most part these habitats include some combination of rainfall, soils, or
pest/disease absence that offers less resistance to occupation than do the
surrounding lowlands, although exceptions exist (Lewis and Berry 1988).
The highlands were favored during colonial occupation of Africa, and
more recently the region has experienced rapid population growth, some
of the highest in the world.
BUSHENYI CASE (UGANDA). Bushenyi is an administrative district
in southwestern Uganda, an area of high rainfall and numerous lakes
and wetlands. Bushenyi has experienced the political turmoil and dis-
"\ s ,
*z -a b
I- I 'I
CM C D 0 C
U r 0
Theory, Evidence, Study Design / 29
ruption of warfare that the rest of Uganda has incurred over the last
twenty years. Despite this, population appears to have grown rapidly,
now approaching 150 people/km2. This growth, a series of changes in
land tenure, including the ability to cultivate wetlands that were once
common grazing lands, combined with a major shift to market cultiva-
tion, has been related to the intensification of agriculture and to in-
creased agricultural production in a zone of relatively good soils. This
increase has followed minimal capital inputs but major increases in la-
bor, with most indicators of agricultural growth up (e.g., increased yields
and market value). These changes have been accompanied by economic
MERU CASE (KENYA). Centered on the slopes of Mt. Meru in the
heart of Kenya, Meru exhibits rather diverse environmental conditions
from the temperate and well-watered upper slopes with good soils to the
tropical and arid foothills where drought is a problem. The better agri-
cultural lands in the district, mostly in the temperate zone, have long
sustained relatively dense populations. Recently, densities have ap-
proached 190 people/km2, exacerbating land pressures and leading to
intensification in the uplands and to cultivation of precarious lands in
the dry foothills below. In this respect, Meru may be seen as two cases:
agricultural growth in the temperate uplands, complete with government
inducements for market crops, and "frontier" farming in the foothills.
The upland case is one of near-annual cultivation, significant increases
in market production, and a rise in capital inputs to cultivation, primar-
ily hybrid crops and synthetic pesticides and fertilizers; this has been ac-
companied by major social changes in economic diversification. The
frontier case has been much more problematic.
USAMBARA CASE (TANZANIA). The Usambara case focuses on a
mountainous region in northeastern Tanzania. Historic evidence sug-
gests relatively large populations and intensive agriculture, much of it on
steeply sloping lands with soils that are not exceptional for cultivation.
Its modern population-agriculture history is a puzzle. During a time of
rapid population growth-now approaching 190 people/km2-the inten-
sification of agriculture has apparently been sporadic, perhaps showing
an overall increase but with major deviations from this trend. This situa-
tion seems to be related to a number of social and economic conditions
that have acted as disincentives not only to intensification but to agricul-
30 / R.W. Kates, G. Hyden, and B.L. Turner II
tural expansion as well. Although market cultivation was not important
at the time of the study, the needs pressures were leading to increased
migration and economic diversification in the face of an overall decline
in food availability.
RUHENGERI CASE (RWANDA). An administrative prefectorate with-
in the northern portion of Rwanda, Ruhengeri is a high-altitude, well-
watered mountainous zone of steep slopes and wetlands, with variable
soil suitability for cultivation. Historically, it has been one of the most
densely occupied regions in all of Africa, with current populations over
400/km2. Agriculture has historically kept pace with demand, perhaps
facilitated by policies that have promoted small-scale farming and, in
many instances, without major landscape capital (e.g., terracing, irriga-
tion) that might otherwise be expected from labor-intensive farming.
Recently, land stress has become increasingly evident, and agriculture
has expanded into wetlands. Cropping frequency is rising while yields
have been more or less stable. Market cultivation has become important,
as have outmigration and economic diversification. Overall, food avail-
ability seems to have kept pace roughly with demand, and economic
well-being has been relatively stable.
KIsII CASE (KENYA). Kisii District, southwestern Kenya, is another
case of a high-rainfall upland with good agricultural soils. During co-
lonial and more recent times, it has been a focus of significant effort for
"modern" development, complete with research, planning, and funding.
This effort has been matched by rapid population growth. Densities now
approach 500/km2, not including the population of cities and towns.
Agriculture, much of which is multicropping, has become strongly
market oriented and has shown remarkable growth, in part because of
the flexibility of the tenure system. Nevertheless, recently there has been
evidence of slowing agricultural growth and environmental stress. Out-
migration is significant, and most farming units are involved in some
type of nonfarming enterprise or wage labor.
West Africa has historically been the most populated large-scale re-
gion in tropical Africa, despite the demographic effects of intensive slav-
ing in the past. In contrast to East Africa, much of its population has
been supported along major river drainages within lowland wet-dry
Theory, Evidence, Study Design / 31
forest zones and in pockets along the Sahelian fringe. Therefore, broadly
speaking, it represents high population densities in environments that
are, perhaps, more common throughout sub-Saharan Africa than are
the mountains of East Africa. Nigeria alone has long sustained the larg-
est population of any sub-Saharan country, even on soils thought to be
marginal at best. With some exceptions, agriculture has grown in re-
sponse to demand.
Jos PLATEAU CASE (CENTRAL NIGERIA). The Jos Plateau con-
sumes most of the Plateau District of central Nigeria, a transition zone
between wet and dry climates. High density growth, ranging from 90 to
300 people/km2, has been a recent phenomenon, and agricultural growth
has been a product of both intensification and land expansion. The Kof-
yar people of the plateau, with minimal formal inducements, have re-
sponded to the opening of new lands for major entry into market
cultivation by adapting their long-developed cultivation practices to the
Benue Valley lands. Their success may be unparalleled in most of tropi-
cal Africa. Cropping frequency, yields, and the market value of produce
are rising, through increases both in labor and in capital, primarily bio-
technic inputs. To date, social changes may have been minimal, as is eco-
nomic diversification. Most important, food availability and economic
well-being are increasing, on average.
IMO CASE (EASTERN NIGERIA). Sample districts in Imo State,
Nigeria, long one of the major centers of population concentrations in a
lowland tropical setting, show population densities that range from 200
to 1,000 people/km2 throughout the area, the higher densities affected
by urban populations. It is situated in the heart of the oil palm produc-
tion zone, and long-term intensification has taken the form of intensive
compound gardens with increasingly intensifying outfields. From a
purely agricultural perspective, much of the area may have reached a
condition of involution, and farming units have diversified their internal
economies accordingly. Nevertheless, the multicropping systems and
yields have increased, as has market cultivation. Biotechnic inputs are
on the rise. Despite these changes, land pressures are so great as to lead
to major outmigration.
NGWA-IGRO CASE (EASTERN NIGERIA). The Ngwa region is also
situated in Imo State, and its rural population densities have exceeded
200 people/km2 for some time. In the first half of this century, the re-
32 / R.W. Kates, G. Hyden, and B.L. Turner II
gion witnessed agricultural diversification and commercialization through
oil palm cultivation, despite relatively poor agricultural soils. These
changes, perhaps motivated by desires to increase standards of living
rather than by population pressures per se, then stimulated further
population growth. The latter half of this century has witnessed rapid pop-
ulation growth, land degradation, and switches to cassava production as
a means to sustain ouput in the face of involution or stagnation in agri-
culture. This study adds historical depth to the broader case study of the
AWKA-NNEWI CASE (EASTERN NIGERIA). This region is situated
within Anambra State, Nigeria. Located immediately north of Imo
State, its environment and socioeconomic history share many common-
alities with the previous two case studies. Awka-Nnewi has a long his-
tory of dense occupation, its settlement pattern described by some as an
extended "suburbia," owing to population densities exceeding 500/km2
throughout. Few opportunities exist for agricultural expansion, and
efforts to intensify may have slowed, becoming more difficult. Market
cultivation is significant, but total production is apparently dropping. En-
vironmental degradation is increasing, and household-diversification
strategies have been substituted as alternatives to cultivation. Food
availability may be on the decline.
KANO CASE (NORTHERN NIGERIA). The area studied is not Kano
city, northern Nigeria, but its immediate hinterlands referred to as the
Close-Settled Zone. In this Sahelian fringe land of variable soil quality
and major drought episodes, purely "rural" populations exceed 500/km2.
Agriculture has intensified largely through major infusions of labor into
a market-gardening system marked by diversification through livestock
and agroforestry. Currently cropping frequency and yields are stable,
while the value of production is increasing. With the addition of house-
hold diversification off the farm, food availability and economic well-
being are apparently stable now and into the near future.
Population Growth and Agricultural Change Workshop
The authors of the commissioned case studies were brought together in
the spring of 1988 at the Center for African Studies, University of Flor-
ida, Gainesville. With the assistance of invited commentators (Richard
Theory, Evidence, Study Design / 33
Bilsborrow, Ronald Cohen, Gregory Knight, Phil Porter, and Donald
Vermeer), the workshop reviewed the draft studies, suggesting revisions
and identifying new issues that the authors were able to address in their
revisions; the gathering was also used to brainstorm the subject of popu-
lation growth and agriculture, leading to various revised interpretations
about the critical elements of agricultural change in Africa. Revisions of
the papers were undertaken through 1988-89, and they constitute the
bulk of this volume.
In what follows, we move through individual case studies beginning
with those from East Africa. The sequencing of the cases for each of the
two clusters largely follows from the less to most densely occupied situa-
tion. The final chapter returns to the central questions of this endeavor,
addressing them through a synthesis and comparison of the case studies
and based upon the themes that were identified during the 1988 workshop.
The individual studies provide a wealth of insight in their own right.
They are rooted in the colonial period, and they bridge what a recent
workshop of African scholars (Achebe et al. 1990) has called the "times
of euphoria" of nation-creating and nation-building and the current
"times of troubles" of the sustained agricultural, economic, and welfare
crises. The very presence of these densely occupied farming regions calls
into question the numerous perspectives on these issues, and the case
studies here provide evidence illustrating the rudiments of the apparent
transformations of conditions in such regions now under way. They also
speak to the great variation in African agriculture and indigenous inno-
vation and adaptation. The case assessments, therefore, testify to how
Africans survive, subsist, and occasionally prosper at population densi-
ties comparable to extremes found in Asia and Europe.
Achebe, C.G., G. Hyden, C. Magadza, and A.P.O. Keyo, eds. 1990. Beyond
Hunger in Africa: Conventional Wisdom and an African Vision. Nairobi:
Heinemann Kenya. London: James Currey.
Ahmad, A. 1985. "The Effect of Population Growth on a Peasant Economy:
The Case of Bangladesh." In The Primary Sector in Economic Develop-
ment, ed. M. Lundahl. London: Croom Helm.
Ali, A. M. S. 1987. "Intensive Paddy Agriculture in Shyampur, Bangladesh." In
Comparative Farming Systems, ed. B.L. Turner II and S.B. Brush. New
34 / R.W. Kates, G. Hyden, and B.L. Turner II
Ali, A.M.S., and B.L. Turner II. n.d. Stagnation or Growth in Bangladesh
Agriculture, 1950-1986. Manuscript.
Barlett, P. F. 1976. "Labor Efficiency and the Mechanics of Agricultural Evolu-
tion." Journal of Anthropological Research 32: 124-40.
Basehart, H.W. 1973. "Cultivation Intensity, Settlement Patterns, and Home-
stead Forms among the Matengo of Tanzania." Ethnology 7: 57-73.
Bassett, T. 1988. "The Political Ecology of Peasant-Herder Conflicts." Annals of
the Association of American Geographers 78: 453-72.
Benneh, G. 1972. "Population Growth and Food Supply in Ghana." Peasant
Studies Newsletter 4: 17-22.
Bernstein, H., B. Crow, M. Mackintosh, and C. Martin. 1990. The Food Ques-
tion: Profit versus People? London: Earthscan.
Berry, S. 1984. "The Food Crisis and Agrarian Change in Africa: A Review Es-
say." African Studies Review 27(2): 59-112.
Binswanger, H. P., and V. Ruttan. 1978. Induced Innovation: Technology, In-
stitutions and Development. Baltimore: Johns Hopkins University Press.
Boserup, E. 1965. The Conditions of Agricultural Growth. Chicago: Aldine.
1981. Population and Technological Change. Chicago: University of
Boyce, J. 1987. Agrarian Impasse in Bengal: Institutional Constraints to Tech-
nological Change. Oxford: Oxford University Press.
Brookfield, H.C. 1962. "Local Study and Comparative Method: An Example
from Central New Guinea." Annals of the Association of American Geog-
raphers 52: 242-54.
1984. "Intensification Revisited." Pacific Viewpoint 25: 15-44.
Brown, P., and A. Podolefsky. 1976. "Population Density, Agricultural Inten-
sity, Land Tenure, and Group Size in the New Guinea Highlands." Ethnol-
ogy 15: 211-38.
Brush, S.B., and B.L. Turner II. 1987. "The Nature of Farming Systems and
Views of Their Change." In Comparative Farming Systems, ed. B.L.
Turner II and S.B. Brush, pp. 11-48. New York: Guilford Press.
Chayanov, A. V. 1966. "Peasant Farm Organization." In A. V. Chayanov on the
Theory of Peasant Economy, ed. D. Thorner, B. Kerblay, and R.E.F.
Smith. Homewood, Ill: R. D. Irwin.
Chen, R.S. 1989. Current Perspectives on Africa's Future, 1957-2057. Alan
Shawn Feinstein World Hunger Program Research Report, RR 89-3. Provi-
dence, R.I.: World Hunger Program, Brown University.
Clark, W., and R. E. Munn, eds. 1986. Sustainable Development of the Bio-
sphere. Cambridge: Cambridge University Press.
Cleave, J. 1974. African Farmers: Labor Use in the Development of Small-
holder Agriculture. New York: Praeger.
Collinson, M. 1972. Farm Management in Peasant Agriculture. New York:
Theory, Evidence, Study Design / 35
Coulibaly, S. 1978. Le Paysan Senoufo. Abidjan and Dakar: Les Nouvelles Edi-
Dalrymple, D.G. 1971. Survey of Multiple Cropping in Less Developed Na-
tions. Washington, D.C.: Foreign Economic Development Service, USDA.
Datoo, B.A. 1976. "Relationship between Population Density and Agricultural
Systems in the Uluguru Mountains, Tanzania." Journal of Tropical Geog-
raphy 42: 1-12.
1978. "Toward a Reformation of Boserup's Theory of Agricultural
Change." Economic Geography 54: 133-44.
Dommen, A.J. 1988. Innovation in African Agriculture. Boulder, Colo.: West-
Dumont, R. 1962. LAfrique noire est malpartie. Paris: Editions du Seuil.
Ehrlich, P.R., and A.H. Ehrlich. 1990. The Population Explosion. New York:
Simon & Schuster.
Eicher, C. 1988. Sustainable Institutions for African Agricultural Development.
Paper for seminar on "The Changing Dynamics of Global Agriculture: Re-
search Policy Implications for National Agricultural Research Systems."
ISNAR, CTA and DSE, Feldafing, Germany, September 22-28.
Farvar, M.T., and J.P. Milton, eds. 1972. The Careless Technology: Ecology
and International Development. Garden City, N.Y.: The Natural History
Fisk, E. H. 1964. "Planning in a Primitive Economy: From Pure Subsistence to
the Production of Market Surplus." Economic Record 40: 156-74.
Geertz, C. 1963. Agricultural Involution: The Process of Ecological Change in
Indonesia. Berkeley: University of California Press.
Gleave, M.B., and H.P. White. 1969. "Population Density and Agricultural
Systems in West Africa." In Environment and Land Use, ed. M. F. Thomas
and G. W. Whittington. London: Methuen.
Goldman, A. 1987. Agricultural Innovation in Three Areas of Kenya. Paper
presented at annual meeting of the Association of American Geographers,
Gray, A.L. 1988. "The Relationship of Production Type to Resource Stock
Manipulation in Chipinge District, Zimbabwe: A Micro-Agricultural Geog-
raphy." Ph.D. dissertation, Clark University.
Grossman, L.S. 1985. Peasants, Subsistence Ecology, and Development in the
Highlands of Papua New Guinea. Princeton: Princeton University Press.
Grove, A.T. 1961. "Population and Agriculture in Northern Nigeria." In Essays
on African Population, ed. K.M. Barbour and R.M. Prothero. London:
Routledge and Kegan Paul.
Harrison, P. 1987. The Greening of Africa: Breaking through the Battle for
Land and Food. London: Penguin/IIED.
Hart, K. 1982. The Political Economy of West African Agriculture. Cambridge:
Cambridge University Press.
36 / R.W. Kates, G. Hyden, and B.L. Turner II
Harwood, R.R. 1979. Small Farm Development: Understanding and Improv-
ing Farming Systems in the Tropics. Boulder, Colo.: Westview Press.
Haswell, M.R. 1953. Economics of Agriculture in a Savannah Village. London:
The Colonial Office, Her Majesty's Stationery Office.
Hayami, Y., and V. Ruttan. 1985. Agricultural Development: An International
Perspective. Baltimore: Johns Hopkins University Press.
Heyer, J., P. Roberts, and G. Williams. 1981. Rural Development in Tropical
Africa. London: Macmillan.
Hill, P. 1963. The Migrant Cocoa Farmers in Southern Ghana: A Study in Ru-
ral Capitalism. London: Cambridge University Press.
Ho, T.J. 1985. "Population Growth and Agricultural Productivity in Sub-
Saharan Africa." In Proceedings of the Fifth Agricultural Sector Sympo-
sium: Population and Food, ed. T.J. Davis. Washington, D.C.: World
Huss-Ashmore, R., and S. H. Katz, eds. 1989. African Food Systems in Crisis.
Part One: Microperspectives. Part Two: Contending with Change. New
York: Gordon and Breach.
Illiffe, J. 1987. The African Poor: A History. Cambridge: Cambridge University
Jewswiecki, B., and J.P. Chretien, eds. 1984. Ambiguites de l'innovation. So-
cietes rurales et technologies en Afrique central et occidentale au xxe
siecle. Ste Foy, France: Editions SAFI.
Johnson, M. 1987. Report of Densely Settled Districts of Sub-Saharan Africa.
Manuscript, Providence, R.I.: Alan Shawn Feinstein World Hunger Pro-
gram, Brown University.
Johnston, B.F., and W.C. Clark. 1982. Redesigning Rural Development: An
Adaptive Perspective. Baltimore: Johns Hopkins University Press.
Johnston, B.F., and P. Kilby. 1975. Agricultural and Structural Transforma-
tion: Economic Strategies in Late-Developing Countries. New York: Ox-
ford University Press.
Lal, R. 1987. "Managing the Soils of Sub-Saharan Africa." Science 236:
Lee, R. D. 1986. "Malthus and Boserup: A Dynamic Synthesis." In The State of
Population Theory: Forward from Malthus, ed. D. Coleman and R.
Schofield, pp. 96-130. Oxford: Basil Blackwell.
Lele, U. 1988. "Comparative Advantage and Structural Transformation: A Re-
view of Africa's Economic Development Experience." In The State of De-
velopment Economics: Progress and Perspectives, ed. G. Ranis and P.T.
Schultz. New York: Blackwell.
Lele, U., and S.B. Stone. 1989. Population Pressure, the Environment, and
Agricultural Intensification: Variations on the Boserup Hypothesis. Man-
aging Agricultural Development in Africa (MADIA) Symposium.
Theory, Evidence, Study Design / 37
Levi, J. F. S. 1976. "Population Pressure and Agricultural Change in the Land-
Intensive Economy." Journal of Development Studies 13: 61-78.
Levi, J., and M. Havinden. 1982. Economics of African Agriculture. Harlow,
Lewis, L., and L. Berry. 1988. African Environments and Resources. Boston:
Lipton, M. 1968. "The Theory of the Optimizing Peasant." Journal of Devel-
opment Studies 4: 327-51.
1989. New Seeds and Poor People. Baltimore: Johns Hopkins Univer-
MacDonald, Andrew S. 1989. Nowhere to Go but Down? Peasant Farming and
the International Development Game. London: Unwin Hyman.
Mackintosh, M. 1989. Gender, Class and Rural Transition: Agribusiness and
the Food Crisis in Senegal. London: Zed Books.
Martin, S. D. 1987. "Boserup Revisited: Population and Technology in Tropical
African Agriculture, 1900-1940." Journal of Imperial and Commonwealth
History 16(10): 109-23.
Matlon, P.J. 1987. "The West African Semiarid Tropics." In Accelerating Food
Production in Sub-Saharan Africa, ed. J.W. Mellor, C.L. Delgado, and
M.J. Blackie. Baltimore: Johns Hopkins Univerity Press.
Mellor, J.W. 1966. The Economics of Agricultural Development. Ithaca, N.Y.:
Cornell University Press.
1984. The Economics of Agricultural Development. Ithaca, N.Y.: Cor-
nell University Press.
Metzner, J.K. 1982. Agriculture and Population Pressure in Sikka, Isle de
Flores. Development Studies Center, Monograph No. 28. Canberra: Aus-
tralian National University Press.
Moock, J., ed. 1986. Understanding Africa's Rural Households and Farming
Systems. Boulder, Colo.: Westview Press.
Mortimore, M.J. 1967. "Land and Population Pressure in the Kano Close-
Settled Zone, Northern Nigeria." The Advancement of Science 23:
1989. Adapting to Drought: Farmers, Famines and Desertification in
West Africa. Cambridge: Cambridge University Press.
Netting, R. McC. 1968. Hill Farmers of Nigeria. Cultural Ecology of the Kof-
yar of the Jos Plateau. Seattle: University of Washington Press.
Netting, R. McC., D. Cleveland, and F. Steir. 1980. "The Conditions of Agri-
cultural Intensification in the West African Savannah." In Sahelian Social
Development. Abidjan: USAID, Regional Economic Development Service
Office, West Africa.
Norman, D. W. 1974. "An Economic Study of Three Villages in Zaria Province,
Part 1, Land and Labour Relationships." Samaru Miscellaneous Paper 19.
38 / R.W. Kates, G. Hyden, and B.L. Turner II
Samaru, Nigeria: Institute for Agricultural Research, Ahmadu Bello Uni-
Noronha, R. 1988. Land-Tenure in Sub-Saharan Africa. Baltimore: Johns
Hopkins University Press.
Office of Technology Assessment. 1988. Enhancing Agriculture in Africa.
Washington, D.C.: OTA.
Paulino, L.A. 1987. "The Evolving Food Situation." In Accelerating Food
Production in Sub-Saharan Africa, ed. J.W. Mellor, C.L. Delgado, and
M.J. Blackie. Baltimore: Johns Hopkins University Press.
Pingali, P. 1987. Institutional and Agro-Climatic Constraints to Agricultural
Intensification in Sub-Saharan Africa. Paper presented at Expert Consulta-
tion on Population and Agricultural and Rural Development: Institutions
and Policy, Rome June 29-July 1.
Pingali, P., Y. Bigot, and H.P. Binswanger. 1987. Agricultural Mechanization
and the Evolution of Farming Systems in Sub-Saharan Africa. Baltimore:
Johns Hopkins University Press.
Population Reference Bureau. 1992. 1992 World Population Data Sheet. Wash-
Porter, P.W. 1979. Food and Development in the Semi-Arid Zone of East
Africa. Foreign and Comparative Studies No. 32. Syracuse: Maxwell
School of Citizenship and Public Affairs, Syracuse University.
Richards, P. 1985. Indigenous Agricultural Revolution: Ecology and Food
Production in West Africa. London: Hutchinson.
1987. "Upland and Swamp Rice Farming Systems in Sierra Leone: An
Evolutionary Transition?" In Comparative Farming Systems, ed. B.L.
Turner II and S. B. Brush, pp. 156-87. New York: Guilford Press.
Robinson, W., and W. Schutjer. 1984. "Agricultural Development and Demo-
graphic Change: A Generalization of the Boserup Model." Economic De-
velopment and Cultural Change 32: 355-66.
Rocheleau, D., F. Weber, and A. Field-Juma. 1988. Agroforestry in Dryland
Africa. Nairobi: International Council for Research in Agroforestry.
Rodgers, G., ed. 1989. Population Growth and Poverty in Rural South Asia.
Rohrbach, David D. 1988. "The Growth of Smallholder Maize Production in
Zimbabwe: Causes and Implications for Food Security." Ph.D. dissertation.
Michigan State University.
Ruthenberg, H. 1980. Farming Systems in the Tropics. 3d ed. Oxford: Oxford
Sahlins, M.D. 1971. "The Intensity of Domestic Production in Primitive Soci-
eties: Social Inflections of the Chayanovian Slope." In Studies in Economic
Anthropology, ed. G. Dalton. Washington, D.C.: American Anthropology
Theory, Evidence, Study Design / 39
Schluter, M.G. G., and T. D. Mount. 1976. "Some Management Objectives for
the Peasant Farmer: An Analysis of Risk Aversion in the Choice of Crop-
ping Patterns." Journal of Development Studies 12: 246-67.
Schultz, T.W. 1964. Transforming Traditional Agriculture. New Haven: Yale
ed. 1978. Distortions of Agricultural Incentives. Bloomington: Indiana
Scott, J.C. 1976. The Moral Economy of the Peasant: Rebellion and Subsis-
tence in Southeast Asia. New Haven: Yale University Press.
1985. Weapons of the Weak: Everyday Forms of Peasant Resistance.
New Haven: Yale University Press.
Simon, J. 1981. The Ultimate Resource. Princeton, N.J.: Princeton University
Singh, S. 1983. Sub-Saharan Agriculture-Synthesis and Trade Prospects. World
Bank Staff Working Papers, No. 608. Washington, D.C.: World Bank.
Stamp, Patricia. 1989. Technology, Gender, and Power in Africa. Techni-
cal Study 6e, International Development Research Centre. Ottawa,
Sutton, J. E. G. 1984. "Irrigation and Soil-Conservation in Agricultural History:
With a Reconsideration of the Inyanga Terracing (Zimbabwe) and Engaruka
Irrigation Works (Tanzania)." Journal of African History 25: 25-41.
Tait, J., and B. Napompeth. 1987. Management of Pests and Pesticides:
Farmers' Perceptions and Practices. Boulder, Colo.: Westview Press.
Tax, S. 1953. Penny Capitalism. Smithsonian Institution, Institute of Social
Anthropology, Publ. no. 16. Washington, D.C.: U.S. Government Printing
Thomas-Slatyer, B.P., and R. Ford. 1990. "Water, Soils, Food and Rural De-
velopment: Examining Institutional Frameworks in Kathetka." Canadian
Journal of African Studies, 23: 250-71.
Turner, B. L. II. 1987. "Purpose, Classification, and Organization." In Compar-
ative Farming Systems, ed. B.L. Turner II and S.B. Brush. New York:
Turner, B.L. II, and S.B. Brush, eds. 1987. Comparative Farming Systems.
New York: Guilford Press.
Turner, B. L. II, and W. Doolittle. 1978. "The Concept and Measure of Agricul-
tural Intensity." Professional Geographer 30: 297-301.
Turner, B.L. II, R.Q. Hanham, and A.V. Portararo. 1977. "Population Pres-
sure and Agricultural Intensity." Annals of the Association of American
Geographers 67: 384-96.
Vermeer, D. 1970. "Population Pressure and Crop Rotational Changes among
the Tiv of Nigeria." Annals of the Association of American Geographers
40 / R.W. Kates, G. Hyden, and B.L. Turner II
Watts, M.J. 1983. Silent Violence: Food, Famine, and Peasantry in Northern
Nigeria. Berkeley: University of California Press.
1989. "The Agrarian Crisis in Africa: Debating the Crisis." Progress in
Human Geography 13: 1-41.
Watts, M.J., and T. Bassett. 1986. "Politics, the State and Agrarian Develop-
ment: A Comparative Study of Nigeria and the Ivory Coast." Political
Geography Quarterly 5: 103-25.
Wharton, C.R., Jr., ed. 1969. Subsistence Agriculture and Economic Devel-
opment. Chicago: Aldine.
Wisner, B. 1988. Power and Need in Africa. London: Earthscan.
World Resources 1988-89. New York: Basic Books.
World Bank. 1984. Toward Sustained Development in Sub-Saharan Africa: A
Joint Program of Action. Washington, D.C.: World Bank.
World Resources Institute (WRI). 1989. World Resources 1988-89. New York:
2 / Agricultural Transformation in
the Robusta Coffee/Banana Zone
of Bushenyi, Uganda
Bushenyi District in southwestern Uganda has experienced significant
agricultural change in the past two decades (fig. 2.1). This change must
be understood within the context of the even greater and ongoing trans-
formation in agriculture created by the colonial political economy since
1900. The most important changes have been the closing of the land
frontier due to increasing population growth and changes in land-tenure
policies and the incapacity of the state to stimulate production of export
crops and maintain basic levels of political security and economic cer-
tainty. Together these factors poise the district on the edge of significant
changes in agricultural practices that will require more effective use of
the land in order to feed its inhabitants and to earn them adequate in-
comes. But few postindependence changes in production techniques, use
of labor, or cropping intensity appear to have been widely adopted in
the district thus far.
Population pressure on the land has turned a district considered
lightly populated at independence into one in which the sons of many
poor farmers will not inherit enough land to grow crops sufficient to
feed their families. Nor will they be able to move onto unused public
land, the conventional solution to this problem in both the precolonial
and colonial periods. Since the 1960s, changes in land tenure and
rapidly rising land values have caused all the remaining public land to
be fenced and converted into private farms.
The change in relations between the state and farmers has been
equally dramatic. The state has always attempted to promote and con-
trol the production of export crops. During the colonial and early inde-
pendence periods, government officials maintained this vital function by
creating close contacts with farmers through by-laws, taxes, coopera-
tives, extension advice, provision of seeds, and production subsidies. The
42 / N. Kasfir
300 E 32 E 340 E
} / 4' 'NJ;
/ MoyO ,-
Au Moyo Kitgum
Arua / Kotido /-
Nebbi c a e Moroto -
./ N Moroto
V / Lira
ZAIRE ZA AacI R
/ ,,Masindi Apc oroti
S/ L. Kvoga I \
Hoima / Kumi /Kap-
S~ \ I chorwa
1 - \ Luwero It Kamuli Pallisa'/Mble
0- Kibale Kiboga -Ie
S- Iganga \Tororo
SKabarole / Mubende \ inJ
N / Mukono'
asese\ Lake - Mpigi 1 KENYA
01 George Kampa
CP r Kampala
Mbarara i ) Masaka- Lake - Victoria
">L Edward/ \ 1993 Boundaries
S ushenyi' i International
Rukun- Rakai KaangIa boundary
S Ntn- .
g amo / i I - District
TA N Z A N I A boundary
RWANDA\ I ninety kilometers
Fig. 2.1. Uganda districts, 1993
most important connection was the insistence of the state on setting
producer prices for all important export crops and requiring farmers to
sell them to the state. Subsequently, the state lost the capacity to im-
plement policies to help farmers, even though its coercive capabilities
remained relatively effective and often intrusive.
Casual observers of Uganda are often surprised to discover that
Bushenyi continued to enjoy a period of dynamic commercial agricul-
tural growth after Idi Amin came to power in 1971. This agricultural
growth has spanned the political and economic uncertainties that Amin's
government introduced, the upheavals of two wars fought near this area,
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 43
which disrupted transport and communication, and the continuing de-
cay of the national economy. Agricultural changes have been most no-
ticeable in the increasing shift toward the production of milk and food
crops for domestic markets rather than export crops for overseas sale.
The transition has produced wealth for enterprising farmers and poverty
for others, rapidly increasing the economic differentiation among farmers
in the district. The virtual disappearance of government support and
subsidies for agricultural production and marketing has not hindered
Measuring changes in agricultural practices and investigating their
relation to population and economic trends is difficult anywhere in
Africa. In Bushenyi only those few farmers who hold leases have had
their land surveyed. The rest, who hold by customary tenure, generally
have an approximate idea of how much land they have but not its pre-
cise measurements. Few of these farmers keep accounts. Typically they
account for their output in terms of locally available, nonstandardized
containers, designed for entirely different purposes.
Precise measurement requires so intrusive and time-intensive an in-
quiry that the numbers of farmers examined will necessarily be few, and
the risk of altering their behavior will always be present. In addition, for
the past decade and a half in Uganda the accurate collection and publica-
tion of population, economic, or agricultural statistics have been as diffi-
cult to produce as any other state service. Civil servants and academics
who might have engaged in farm research have had to take second jobs
and trade on the black market in order to survive. That leaves little time
for data collection and analysis. Published statistics are much less reli-
able than they were two decades ago.'
Basic Features of the Study Site
The area selected for inquiry here consists of the robusta coffee/banana
zone (fig. 2.2), including most of the counties of Igara, Sheema, and, to
a lesser extent, Ruhinda in Bushenyi District (Planning Unit of the Min-
istry of Agriculture, Forestry and Cooperatives [PUMAFC] 1972:
17-20).2 This heavily populated area was part of Ankole District until
1974. Warnings about pressure on available land have been sounded peri-
odically since the 1960s. However, it also is a food-surplus area able to
service urban markets as far away as Kampala. In 1964 Ankole farmers
44 / N. Kasfir
< -t. Road
Rwaslamraire MBARARA Sub-county
r ie boundary
RUKUNGIRI Rro I Ntungamo
30 00'E 30o30'
Fig. 2.2. The Bushenyi study area
accounted for only 6.4 percent of Ugandan land devoted to robusta
coffee, but 70 percent of them had already adopted it (MAC 1966: 66,
70). By this time, Ankole already accounted for 12 percent of the ba-
nana hectarage, third in size behind Busoga and Masaka districts (MAC
1966: 99). Coffee production from Ankole increased until the middle
1970s and then stagnated when sales fell. Banana production has stead-
The creation of this robusta coffee/ banana zone was the most pro-
found agricultural transformation to occur in this area during the twen-
tieth century. Robusta coffee was not cultivated when colonial rule
i -) _
3U u't 30 .30' SUDAN .
Goo, KABAROLE \-
Katun gurru <,' UGANDA.
I Parish used as
RUGURU/ .. I site for study
/ / Robusta coffee/
Irtr I%^.rl n^6 n l
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 45
began. Bananas may have been grown for food, but were not widely cul-
tivated and were not produced for the market. Thus, the time frame for
this study begins with the institution of colonial rule early in the twen-
tieth century but focuses more on the period since widespread adoption
of coffee in the 1950s.
On the basis of its precolonial royal heritage, but with an expanded
territory that for the first time included the area now in Bushenyi Dis-
trict, Ankole was established as a kingdom during the protectorate pe-
riod on the basis of an agreement between the king (Omugabe) and the
British authorities. During colonial rule, British colonial authorities con-
tinued to recognize the existing chiefly elite, Bahima by ethnic identifica-
tion, as the legitimate rulers. By custom the Bahima are a pastoral
people and consist of roughly 5 percent of the people living in Ankole.
The Bairu, customarily agriculturalists, amount to about 90 percent
(Doornbos 1977: 53). Bahima and Bairu are related to each other
through a stratified class system that became a subject of political con-
tention as independence neared. With the arrival of democratic elections
in the 1950s, the Bairu united to overturn Bahima dominance and even-
tually to take control of most elective offices (Doornbos 1970: 1096-1109).
In addition, political parties competed on the basis of existing rivalries
between Protestants and Catholics. The complex interactions of reli-
gious and ethnic rivalries caused factionalism to become institutional-
ized in the politics of Ankole (Doornbos 1977: 72). That pattern has
continued into the 1980s in the two new districts.
In the administrative reorganization of 1974, the five western coun-
ties of Ankole District (formerly Kingdom) became Bushenyi District
(then called West Ankole). The area of the new district is 5,395 km2, of
which 4,905 km2 are land. By 1986 the original five counties had been
reconstituted into seven. The census population of this area rose from
299,832 in 1959 to 410,683 in 1969 and then to 524,669 in 1980. The
1980 Population Census (COMPED 1982: 234, 235) thus reported the
overall population-to-land densities as 84/km2 in 1969 and 107/km2 in
1980 as compared to the rise in reported Ugandan population density
from 48/km2 in 1969 to 64/km2 in 1980. However, there are several diffi-
culties with these figures that suggest that the actual ratios are far
Enormous problems of transport result from lack of maintenance of
the roads, another state function that has been performed quite intermit-
tently since the 1960s. The main road to Kampala is paved, although in
46 / N. Kasfir
1989 it had become so pitted with potholes that it drastically shortened
the life of the reduced number of vehicles that continued to provide the
main economic connection to the capital. The interior roads are mostly
murram (unpaved) and impassable during the rainy seasons. Many vil-
lages are sufficiently far from the paved road to make it impossible for
them to enter the banana trade. Indeed, Ugandan agricultural officials dis-
tinguished two different robusta coffee/banana zones for just this reason,
despite similar ecological conditions (PUMAFC 1972: 19-20). A later
study confirms that farmers with the largest banana hectarage in this area
live in villages close to paved roads (Kyeyune-Ssentongo 1985: 31).
Oberg states that there were no markets before the protectorate gov-
ernment was established (1943: 575), but there had been extensive barter
locally and long-distance trade in salt for at least two centuries before
that (Good 1970: 149-69). The first official market in Ankole, located in
Mbarara, was established in 1902 (Good 1970: 180). There are biweekly
markets throughout the area attended by itinerant traders selling textiles
and other consumer goods and small weekly (and in some places daily)
markets in which the chief merchandise is agricultural produce traded
among farmers (Kyeyune-Ssentongo 1985: 64-65).
Using Chenery's productivity ratings, Langlands and Mbakyenga (1974),
following the procedure described in Langlands (1974c: 1-5), judge the
soils of the robusta coffee/ banana zone (fig. 2.2) in Bushenyi as medium
in productivity in Igara County and medium to fair in Sheema County.
In Igara, the soil is characterized as latosols loam, typically a red-clay
loam, occurring from the weathering of schist and phyllites in Precam-
brian rocks (Langlands and Mbakyenga 1974; Langlands 1974c: 13).
The soils in Sheema are a complex mixture of six types. The largest area
(44 percent) contains sandy loams developed on schists and phyllites on
Precambrian rocks (Langlands and Mbakyenga 1974; Langlands 1974c:
16). Another sixth of the area is the same type of more productive soil
found in Igara. Harrop (1967: 26) points out that "the heavier textured
soils are the more fertile" and adds that characteristic soils of Bushenyi
"have a well developed humic topsoil together with a distinct humic ho-
rizon at some depth in the soil profile."
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 47
The topography consists of a plateau at an elevation of about 1,200
m with gently rolling country generally without steep slopes but in-
terspersed with small, highly eroded stony and steep ridges up to 300 m
high and often 25 to 50 km long with narrow and deep river courses at
their base. The Rwizi River is the largest and longest of these. The val-
ley bottoms often contained papyrus swamps, which are now being
drained by their new owners. Though perhaps not well-founded, there
are widespread fears among farmers and local officials that the drainage
of swamps will lower the water table and may even change the weather
Precipitation occurs throughout the year in a bimodal pattern with
the "long" rainy season from March to May and the "short" rains from
September to November. The former season is less reliable, however, so
most annual crops, particularly millet, are grown during the latter. Rain-
fall ranges from 100 to 125 cm per annum with the total greater in the
West. There are typically 90 to 120 days of rain per year (Kyeyune-
Ssentongo 1985: 6-7). Rain is least likely in January and February.
Drought occurs rarely, but part of the area did experience one in 1980.
The mean minimum temperature ranges between 12.50C and 150C,
whereas the mean maximum varies between 200C and 27.50C. The
western areas of Bushenyi are cooler and more humid than the eastern.
The chief pest of the banana (Musa spp.) is the weevil (Cosmopo-
lites sordida) (Ingram 1970). Banana wilt (Fusarium oxysporum var.
cubense) was first found in Ankole in 1952 (McMaster 1962: 43). But
the banana's ecological suitability has been demonstrated by its success
in resisting both. Much research has been done on diseases and pests
affecting robusta coffee (Coffea canephora). The most significant disease
appears to be Red Blister Disease caused by a strain of Cercospora
coffeicola. The coffee-berry borer (Hypothenemus hampei) is a worri-
some pest (Butt et al. 1970: 201-5). Finger millet (Eleusine coracana)
has been relatively free from disease (McMaster 1962: 50). The main
pests are birds and the infrequent invasions of locusts. Cassava (Mani-
hot esculenta) suffers most frequently from mosaic, spread by a virus
(McMaster 1962: 69)-its most common pests are monkeys. Beans
(Phaseolus vulgaris and P. lunatus) are attacked by anthracnose. Sweet
potatoes (Ipomoea batatas) are most affected by a virus disease first
recognized in southwestern Uganda in 1944 but now becoming more se-
rious, as well as by rats and moles (McMaster 1962: 74). Diseases and
48 / N. Kasfir
pests attacking groundnuts (Arachis hypogaea) are not numerous; the
most widespread disease is rosette (McMaster 1962: 77), although
crested cranes and crows sometimes destroy entire gardens.
Changes in Population and Land Tenure
Over the last two decades all unused public land has disappeared in re-
sponse to changes in rural-population density, land-tenure legislation,
and economic insecurity caused by rapid inflation. The three factors
have combined to place farmers in this agricultural zone on the brink of
a very serious situation. In my own surveys during 1984 and 1986, I
found few people who were already landless, although several poor
farmers did not have enough land of their own to supply the subsistence
needs of their families.3
The population densities reported for Bushenyi District in the 1980
census (COMPED 1982: 234) were derived by measuring the land avail-
able by simply subtracting the area of open water. A more useful mea-
sure is the ratio of population to cultivable area defined as total land
area excluding open water, permanent swamps, lands over 2,000 m and
all lands legally designated for a specific nonagricultural use, such as
forest reserves, national parks, and game reserves (Langlands 1974d: 55,
and see Langlands 1974a). On this basis, Langlands (1974d: 63) calcu-
lated the total cultivable land in each of the counties now comprising
Bushenyi. The cultivable area of the district amounts to 3,559 km2,
which yields a districtwide population density of 85/km2 for 1959 and a
rural population density of 115/km2 for 1969 (calculated from Lang-
lands 1974d: 63).4 On this basis the rural population density of Bushenyi
in 1980 rose to 147/km2, about three-eighths higher than the density the
But there are other problems with the 1980 census. Much of its origi-
nal data was lost, although almost all the population records for Bushenyi
survived. In a recent re-analysis of these surviving records, it became
clear that children under one year were almost totally ignored by the
census takers (MPED 1988: 22-24). Correcting for this error increases
Bushenyi's total population by about 4 percent and provides a 1980
rural-population density of 153/km2 of cultivable land. In addition, it is
reasonable to assume that in general there was an undercount by the
1980 census takers, given the extremely difficult and insecure conditions
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 49
of the transitional political regime just following the overthrow of Amin.
Within Bushenyi the robusta coffee/banana zone has a higher popu-
lation than the rest of the district. Unfortunately, Langlands's calcula-
tions of cultivable area are determined at the county level only and
based on those counties existing at the time of the 1969 census. Since
Ruhinda County was created out of subcounties taken from both Igara
and Sheema counties, it is impossible to determine the 1980 rural-
population density based on cultivable areas within each of those three
counties, or within subcounties or parishes in them. Nevertheless, a 1980
rural-population density can be established for Igara and Sheema coun-
ties taken together as they existed in 1969. That figure provides a good
estimate of the population density of the Bushenyi portion of the ro-
busta coffee/banana zone, even though these county boundaries do not
entirely coincide with those of the zone.
The population density of Igara County grew from 78/km2 in 1959
to 110/km2 in 1969, while that of Sheema increased from 95 to 137/km2
over the same period (Langlands, 1974d: 63).5 Combining Igara and
Sheema as they existed in 1969 (i.e., including Ruhinda), the 1980 popu-
lation figures give a rural density of 152/km2.6 When 4 percent is added
to correct for the error in the 1980 census, the rural-density figure be-
comes 158/km2.7 I have no reason to assume that subcounty rural-
density figures are much different from the county figures. Nevertheless,
the hue and cry in the area about land scarcity indicates that access to
land has become much more unequal, which means that rural density is
significantly higher than even this figure in parts of the zone.
Migration patterns-both temporary and permanent-contribute to
the explanation of population density, the availability of adult agricul-
tural labor, and the gender balance on local farms. Data on changes in
age and sex structure for the population of this zone, from which migra-
tion patterns could be inferred, do not exist. The absence of county-level
breakdowns in the 1959 census requires that any comparisons over time
be made at the level of the former Ankole District. The absence of any
breakdown by age in the 1980 census further inhibits analysis.
Nevertheless, some explanation is possible. The agricultural and pop-
ulation census data show that from 1959 to 1969 male to female popula-
tion ratios by age group for Ankole follow the conventional pattern of
more males than females leaving the area during their most productive
years. The necessity to produce a cash income spurred temporary mi-
grants to seek employment on the farms of Buganda or in urban occupa-
50 / N. Kasfir
tions in Kampala and then to return home, sometimes after a season
and sometimes after many years. Thus, Ankole served to some extent as
a labor reserve for Buganda during this period, reducing the agricultural
labor available for production in the district. In addition, there was
greater reliance on female laborers remaining on their farms.
The Report on Uganda Census of Agriculture (MAC 1965: 44)
shows that in 1963 there was a surplus of 106 males for each 100 females
up to the age of 16, 86 males for 100 females between the ages of 16 and
45 and an equal number of males and females over 45. A roughly similar
figure occurs in the analysis of the 1959 population census. There, the
number of males falls below the number of females between the ages of
10 and 14, reaching a low point of 71 males for every 100 females be-
tween the ages of 25 and 29, and rises to 100 for the ages of 45 to 49
(SBMEA 1961: 72). In the 1969 census the number of males falls below
that of females in the 15-19-year category, reaches a low of 75 males for
every 100 females at 25-29 years, and rises to 91 at the ages of 45-49
(SDPO, Statistics Division, President's Office 1976: 2). The changes in
these ratios over this ten-year period suggest a trend for agricultural la-
bor in Ankole to become more balanced by gender in the first decade af-
Over this period there may also have been a reduction in the propor-
tion of the total population that migrated, as agriculture in Bushenyi
became more commercial. It would be reasonable to anticipate such a
trend to accelerate (or else to begin) in the 1970s and 1980s. The sharp
rise of food prices and insecurity in urban areas made life in the coun-
tryside far more attractive. In addition, rising urban food prices pro-
vided a remarkable reversal in income opportunities favoring farmers
over civil-service employees for the first time since the establishment of
a monetary economy.
The 1980 population census, however, suggests that the same level
of gender imbalance in outmigration continued through the 1970s. It pro-
vides a breakdown between the sexes but no division into age categories.
For Sheema it reports 89 males for every 100 females, for Igara 92 males
and for Ruhinda 88 males (COMPED 1982: 32, 33, 39). That compares
to the 1969 figures of 89 males for every 100 females in Sheema and 92
in Igara (SDMPED, Statistics Division, Ministry of Planning and Eco-
nomic Development 1971: 292, 312). Since the differential favoring
agricultural over civil service income increased even more rapidly during
the 1980s, the effect may merely have been delayed. Therefore, one
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 51
would expect the next census to show a closer balance between males
and females living in rural areas. There has probably also been a decline
in temporary outmigration from Bushenyi for the purpose of earning a
cash income, but I have no data to demonstrate it.
Inmigration during the 1950s and 1960s came primarily from the
inhabitants of Rukungiri and Kabale districts where a period of land
scarcity had begun a generation earlier than it did for Bushenyi resi-
dents, and from Rwanda where the overthrow of the Watutsi monarchy
led to a large number of refugees. In the 1969 census Ankole showed a
net increase of 38,000 migrants, about 4 percent of its total population
(SDPO 1976: 37). There are no data on migration per se in the 1980
census. However, the annual rates of population growth from 1969 to
1980 between Bushenyi and Mbarara districts differ considerably, 2.3
percent and 4.2 percent respectively. The most plausible explanation is
net permanent outmigration from Bushenyi where population is more
dense and land is more expensive and net inmigration into Mbarara, al-
though not necessarily from Bushenyi only (Ntozi and Gasana 1980: 3).
Within the zone, chiefs in Bushenyi subcounties where land is ex-
pensive report an excess of people selling their land and leaving over
new arrivals buying farms (Kasfir 1984-86). But where land is cheaper,
chiefs report the reverse. The rapidly rising price of land in the robusta
coffee/banana zone is one of the most frequent reasons farmers give to
explain outmigration from this area (Kasfir 1984-86). In a shameful epi-
sode, with overtones of tensions over land, about 80,000 people, iden-
tified ethnically as Rwandans but including many Ugandan citizens,
were uprooted from their homes, deprived of much of their property,
and forced into Rwanda or make-shift refugee camps near the border by
popular action orchestrated by local political officials (Clay 1984). In
general, migration trends are not so distinct that they can be drawn from
existing data, but the patterns becoming apparent indicate that re-
sponses to overcrowding are beginning to appear in this zone.
The implications of growing rural-population density in Bushenyi
for agricultural change revolve to a very large extent around the recent
disappearance of public land (karandaranda). Only thirty years ago
there were enormous stretches of Crown or public land, particularly in
the eastern part of Ankole, but also in what is now Bushenyi. This ex-
plains why McMaster (1962: 35) estimates that in 1958 only 11 percent
of available land was in use. According to the agricultural census, in
1964 only 18.8 percent of Ankole's land was claimed as part of private
52 / N. Kasfir
holdings.8 Even as late as 1970 the Department of Land and Surveys
still regarded 76 percent of agricultural land in Sheema and 60 percent
in Igara as "unused" (Kyeyune-Ssentongo 1985: 18).
But starting in the late 1960s, "in West and East Ankole, there
[was] . a rush for people to fence their land" (Muwonge 1978: 168).
Now, as Kyeyune-Ssentongo (1985: 18) tartly points out, "in spite of the
abundance of land, there is no land which is free. All land which is not
gazetted as Forest Reserve, National Park or controlled hunting area in
Ankole, is occupied by somebody." In Bushenyi and Mbarara districts,
the land frontier has closed (Kasfir 1988). Long gone are the days when
"as a rule a peasant could cultivate any piece of land he liked and there
were no restrictions on his breaking up new land except previous occupa-
tion" (Roscoe 1923: 95). The grandchildren of many of those "peasants,"
who continue to hold their land in customary tenure, are considerably
more at risk. It is not just a matter of reduction in available land, but of
profound changes in the land-tenure system during both the colonial
and independent periods.
Since customary tenure is based on occupation and use, the avail-
ability of land to which no one has legal title is essential to permit the
system to adapt to population growth. During the colonial period the
contradiction between the European conception that all land must be
owned by someone and the local notion of customary tenure was neatly
solved through the fiction that the state owned all land not held pri-
vately in freehold or leasehold but exercised no incidents of ownership.
In Bushenyi, only a small amount of the land was ever alienated in free-
hold. Early in the colonial period, chiefs were given freehold estates
(mailo) as a reward for agreeing to British overrule, and, in a short-lived
experimental scheme at the end of the colonial period, enterprising
farmers were permitted to apply for "adjudicated freehold" for farms
they already held in customary tenure as an incentive to improve pro-
ductivity (Doornbos 1975). Since government policy was generally op-
posed to freehold, most grants of land to individuals by the Land Board
since independence have been leaseholds. Nevertheless, most farmers
probably still hold their land by customary tenure.
When the Land Reform Decree was passed in 1975, it "national-
ized" all land, vesting freehold in the state alone. Since then, nothing
more than a long-term renewable leasehold can be privately owned in
Uganda. Mailo and adjudicated freehold estates were automatically con-
verted to leasehold, although technically the land had to be surveyed
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 53
and plans for economic development had to be submitted. Since the
leases could be held for up to ninety-nine years and are then renewable,
the owner loses little. But customary tenants on former mailo land as
well as customary holders on public land on which a lease is acquired
lose the security of tenure that the law previously gave them. They may
be dispossessed, although they must be given six months' notice and
adequate compensation. In theory the Land Board must also resettle
them on comparable public land. But this assumes that there remains
such land, and even then there is no guarantee that someone else might
not lease that public land and start eviction proceedings all over again
(Nabudere 1980: 205).
Nonetheless, the lack of government capacity since the Land Re-
form Decree was promulgated to survey lease applications rapidly, or to
review lessee plans for economic development, has caused great uncer-
tainty over whether the decree will actually be applied. In such circum-
stances, uncertainty frequently works in favor of those who have
connections to state notables. Generally speaking, all a powerful politi-
cal figure had to do was convince a court or a land board that he had a
plan to develop the area. As one Lands Officer (DLS, Department of
Lands and Surveys 1980: 18) wryly noted: "Such a situation has not
proved satisfactory as hundreds of customary tenants have had their bi-
banjas [customary holdings] leased to influential persons without reason-
able excuse." Many of those displaced during the Amin regime, he
added, received no compensation at all. Other observers point to the
same practice in the 1980s under the Obote government.
Thus, in the late 1960s when rising coffee income made land a valu-
able commodity in Bushenyi, wealthy people began to invest their
money in it. Unused public land provided a windfall profit, because it
had always been free so long as no one had put it into use. Now merely
for the price of a survey and legal fees, it could be held as a long-term
lease. Nor was evidence of use essential to ownership. Public land soon
Fencing was a reliable method for indicating boundaries to newly
acquired land that no longer had to be occupied and farmed to be
owned. Since the government subsidized fencing materials in the 1960s,
the growth of enclosed farms increased all the more rapidly (Muwonge
1978: 182). Fencing is, of course, important for improving the productiv-
ity of a farm. Good dairy practice requires separating paddocks, and
fences allow cropland to lie fallow without ambiguity over ownership
54 / N. Kasfir
(Muwonge 1978: 180). But for the most part what was fenced was
simply the perimeter.
The combination of increased human and livestock populations
(Muwonge 1978: 178-79), the dizzying rate of monetary inflation, and
the high returns to food and dairy agriculture made investment in land
extraordinarily compelling. For the first time it became an attractive in-
vestment to lease and drain swamps in order to turn them into private
pasture. According to custom, a swamp had always been considered
communal property that could not be held even by customary tenure.
Instead, they were available to all who lived nearby for planting sweet
potatoes (on large dirt mounds), for collecting clay for pots (Roscoe
1923: 103), and for gathering thatch for roofs. Now they are private
property complete with perimeter fences to keep out neighbors. It is not
surprising to discover reports of violent land disputes in the district
commissioner's files (Kasfir 1988: 170-71). In addition, out of 59
farmers in my 1984 sample, 22 reported that their families had expe-
rienced seasonal hunger, and 19 others said they had not but knew of
families that had. Frequently, poor peasants sell their farms and mi-
grate, often to Mbarara District where land prices are lower (Muwonge
1978: 181; Ntozi and Gasana 1980: 3).
Whether or not "fencing out" the rapidly increasing rural population
was the cause, it is clear that by the mid-1970s the average holding of
farmers in the robusta coffee/ banana zone was smaller than might be ex-
pected from earlier reports of vast amounts of unused land and con-
siderably smaller than the fenced farms. The agricultural census (MAC
1966: 20) found that the median holding of farmers in Ankole in 1964
was 1.8 ha (corrected to size of cultivated area).9 The average farmer in
the 1974 sample from the robusta coffee/ banana zone held 2 ha, where-
as only 8 percent had more than 4 ha. Two-fifths of these farmers had
0.8 to 1.6 ha, while two-thirds had between 0.4 and 2.4 ha (Kyeyune-
Ssentongo 1985: 21, 22).
One indication of the coming land squeeze is the surprisingly low pro-
portion of inheritance of the original farmstead in the 1974 survey. Only
57 percent of the respondents reported they inherited their first parcel.
Almost one-quarter said they started out by purchasing their first plot
(Kyeyune-Ssentongo 1985: 19). Of the 28 percent of farmers who added
more land, 82 percent did so by purchasing it. Indeed, for these respon-
dents the purchase of land was the second largest farm expense after
hired labor (Kyeyune-Ssentongo 1985: 68-69). In comparison Muwonge
(1978: 175) discovered for the same counties in 1973 that Sheema con-
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 55
trained 283 farms that had been fenced, averaging 30.43 ha, whereas
Igara had 232, averaging 38.34 ha.
All local observers say the trend to buy and fence large farms
throughout Bushenyi District has continued to accelerate. Knowledge-
able local observers estimate that the increase in land values over the
past two decades has gone up several hundredfold (Kasfir 1988: 166). It
is hard to be sure that the real value of land has risen much, given the
disastrous decline in the shilling. But this is beside the point. Land was
where rational Ugandan investors who could not export their capital
ought to have stored their money. The general consequence has been to
raise effective rural-population densities-that is, the density of accessi-
ble land-to a far higher, if presently unknown, level.
The Nature of Agricultural Change
The typical farming system in the robusta coffee/banana zone (fig. 2.3)
in Bushenyi consists of one or a few nearby plots, on which the home-
stead is located. Villages therefore comprise dispersed settlements. The
farmer uses a few simple tools, particularly the hoe, makes few capital in-
vestments in his land, relies primarily on family labor, although fre-
quently labor also is hired or hired out, and has received virtually no
direct government assistance for the past fifteen years. Almost all
farmers grow a mix of similar cultigens for subsistence and cash income,
but these cultigens have changed significantly over the past eighty years.
Within the general framework proposed by Turner and Brush (1987:
6-8), these farming systems would be characterized by low but probably
increasing output intensity, relatively static paleotechnic rather than
neotechnic inputs, and mixed production for both direct consumption
and, increasingly, market sales.
The most important agricultural changes in this zone have been the
shift from finger millet to bananas and the introduction of robusta
coffee. Widespread ownership of cattle by the nonpastoral Bairu for
dairy and meat production also has been a significant innovation. These
changes were largely the consequence of the spread of the monetary
economy from Buganda following the imposition of colonial rule. Cash
became increasingly indispensable for taxes and for school fees, as well
as for purchasing first imported commodities and eventually even locally
grown food crops. Farmers became increasingly market oriented.
Finger millet was "the main crop" during the earlier colonial period,
56 / N. Kasfir
Fig. 2.3. Farm with three buildings (including granary), Igara County, Bushenyi
District, 1989. Bananas in foreground and background surround open pasture.
Young cassava below granary and pineapples below cassava. Fence in field used to
although it was supplemented with bananas, sweet potatoes, peas,
beans, groundnuts and marrows (Roscoe 1923: 96, 100; Oberg 1943:
573, see Thomas 1970). Williams (1936: 203) calls attention to "the su-
preme importance attached to the performance of all that is humanly
possible to ensure the successful harvesting of the staff of life-millet,
called buro." Its use in sealing blood brotherhood indicates its symbolic
importance (Williams 1934). Oberg (1943: 573) found that "elaborate
magical practices" were used to protect the millet crop, but not others. It
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 57
was usually the first cultigen planted on new land (Taylor 1969: 105).
Even today when it is no longer the subsistence staple of the Bairu, dis-
cussions of millet generally evoke palpable enthusiasm. Despite their
overwhelming importance bananas have not replaced millet as a symbol
of ethnic identification.
Considerable expansion of banana cultivation occurred after 1952
(McMaster 1962: 52; Mukasa 1970: 139). By 1965, 97 percent of Ankole
farmers grew bananas, whereas only 65 percent grew millet (MAC,
1966: 46). In 1972 agricultural officials reported that "finger millet
S. [had] lost its dominant place as the major food crop" in the ro-
busta coffee/banana zone (PUMAFC 1972: 18). In 1974 bananas pro-
vided over one-half of the total available food for farmers in the zone,
whereas millet provided only 4 percent (Kyeyune-Ssentongo 1985: 28-29).
A little later bananas replaced coffee as a cash crop in some areas,
primarily in Sheema county, but more typically they were combined
with coffee and either replaced annual crops or were planted on pre-
viously unused land. Further expansion of banana production occurred
as the urban market price of bananas inflated at a dizzying rate-over
two hundredfold between 1970 and 1984 and 150 times again by 1988.
Good (1970: 178) attributes the change from millet to bananas to
the cultural prestige of the Baganda.10 Practical advantages, however,
may better explain this shift. In comparison to millet, bananas have
lower labor requirements, greater reliability, and year-round availability,
although millet can be stored for a long time. The shift to bananas may
have been further reinforced by the importance of planting a permanent
crop in order to demonstrate ownership in times of increasing land scar-
city. If coffee was considered undesirable for this purpose, bananas would
be the most logical substitute.
Arabica coffee was planted by local Bushenyi farmers in the late
1920s, as the Ugandan experiment with coffee as a plantation cultigen
grown by settlers was coming to an end. The pace of adoption was slow
because arabica was susceptible to disease. In the 1930s arabica trees
were uprooted and replaced with robusta, except in Bunyaruguru County.
Kyeyune-Ssentongo (1985: 91) suggests that poor farmers who worked
on coffee farms in Buganda introduced coffee into Ankole. Land
planted to coffee in Western Province (which includes Bushenyi) has al-
ways lagged behind Buganda-in the 1930s about one-third as much
land was used for coffee.
"Really phenomenal growth in the Uganda coffee industry has oc-
58 / N. Kasfir
curred only since the war" (Uganda Government 1967: 2). Nevertheless,
as late as 1951 production in Ankole remained trivial, amounting to less
than 200 metric tons. A year after independence, the whole kingdom
still produced only a little over 5,000 metric tons a year. Coffee adoption
expanded sharply in the 1950s and after, so much so that the agricul-
tural census reported that by 1965, 70 percent of the farmers in Ankole
had some coffee trees (MAC 1966: 70). In 1974, 92 percent of the
farmers sampled in this zone had some trees (Kyeyune-Ssentongo 1985:
30). Because of the coffee quota, little new planting occurred after the
mid-1960s. Nevertheless, coffee had become the main source of cash in-
come for these farmers in Kyeyune-Ssentongo's sample in 1973-74, pro-
viding 62 percent of total income (1985: 41).
By the 1980s the decline in returns to coffee and the rise in the value
of sales of bananas, other food crops, and milk had again altered
farmers' production preferences. Economic incentives must be quite sig-
nificant before farmers will uproot coffee trees that have taken three to
four years to mature in order to plant bananas or introduce pasture for
livestock. Good (1970: 26) reports that some farmers were already up-
rooting their coffee trees in the 1960s with the encouragement of local
agricultural officials." However, the practice did not become noticeable
until 1975 when farmers were not paid for their coffee by the govern-
ment (Kasfir 1984-86). Uprooting trees seems to have occurred pri-
marily in areas of Sheema in which bananas do particularly well, but it
also took place in Igara. Elsewhere in this agricultural zone coffee was
given little attention or investment from the mid-1970s, but the trees
were left standing. Yields fell, but the reduced crop was usually har-
vested and sold. Energies went instead into clearance of unused land for
the expansion of banana plantations for market production and for pas-
ture for dairy cattle (Kasfir 1984-86).
By traditional preference the Bairu, the predominant inhabitants of
this zone, were agriculturalists, whereas the Bahima, about 5 percent
of the population, were nomadic cattle herders. However, in a cattle cen-
sus conducted in 1937 Bairu owned two-thirds as many head of cattle in
Ankole as the Bahima and had achieved the same rate of increase with
only half the number of breeding cows (Mackintosh 1938: 29). In the
1970s Bairu owned two-thirds of the cattle in Bushenyi (Muwonge 1978:
170). With the startling rise in the price of milk and spread of its con-
sumption in urban areas, profits from dairy herds frequently exceed
coffee and rival bananas. The indigenous long-horned Ankole cow
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 59
(sanga) successfully resists local diseases, but it is a poor milk (and beef)
producer. Exotic dairy cows give far more milk (up to twenty liters a
day) but are highly susceptible to East Coast Fever, unless protected
with acaricides. Exotic cattle, Guernsey and Friesian, were first intro-
duced into Bushenyi in 1965. Despite the extraordinary difficulties in-
volved in importing drugs on a reliable basis, dairy herds of up to sixty
head, generally crossbred but in a few cases entirely purebred, have been
maintained throughout all the difficulties of the Amin and subsequent pe-
riods (Kasfir 1984-86).12
With a few notable exceptions, particularly dairy production, Uganda's
economic and political difficulties have made it impossible for most
farmers to move away from paleotechnic systems using little energy,
emphasizing primarily manual labor, and introducing few chemical or
mechanical improvements into the production process. Precolonial farm-
ing in this zone was based on a hoe culture, and "no source of power
other than the human arm and back was utilized" (Oberg 1943: 573).
Hoes and knives were produced by local blacksmiths. The land was
cleared and prepared by shallow turning of the soil in each of the dry
seasons, followed by deeper cultivation at the beginning of the rains. A
bush fallow system was followed with cultivation for approximately
three years before the land was left fallow. No fertilizer or artificial irriga-
tion was used (Taylor 1969: 105). The situation today remains virtually
Hand labor is still the most important agricultural input; neither
oxen, nor tractors, nor fertilizer, nor pesticides, nor improved seeds are
used regularly (Kyeyune-Ssentongo 1985: 47). Thus, the hoe remains by
far the most important and widely used farming implement. Cheap and
widely available up to the early 1970s, it is not so easily replaced today
because of the production problems in Uganda's factories and of the
disappearance of local blacksmiths a generation ago. Nevertheless, all
households have at least one hoe, three-quarters possess a panga (ma-
chete), and almost half have a long-handled pruning knife (Kasfir 1984-
86). Richer farmers frequently own six to ten hoes, though only one or
two pangas and pruning knives. What is more important, however, is
that these rich farmers rarely own any other tools that are technologically
superior to those of poor farmers (Kasfir 1984-86). Wealth or size of
60 / N. Kasfir
farm, then, seems to have virtually no effect on the kind of tools owned
by these farmers. This is a significant indicator that technological changes
in farming-outside the dairy industry-are not presently occurring.
Ox plowing has not been adopted in Bushenyi, although it has spread
widely over the past seventy-five years in other parts of Uganda, particu-
larly in the cotton-growing areas in the east (Uchendu and Anthony
1975: 36-38).'3 The reasons for not moving from the hoe to ox plowing
in Bushenyi are based on environmental constraints, choice of cultigens,
and plot size, in addition to the difficulties of introducing technological
innovation in the midst of political and economic instabilities. Even
where population density is relatively high, there often is no comparative
advantage of the ox plow over the hoe where plots are small, the terrain
is steep and much of the land is devoted to tree crops (Pingali et al.
1987: 7). All three factors are characteristic of Bushenyi. In addition,
only sixty kilometers south and east of the robusta coffee/banana zone
efforts to eradicate animal trypanosomiasis transmitted by the tsetse fly
(Glossina sp.) that had been relatively effective in the 1960s have been
losing ground since then.
The family continues to be the most important source of farm labor.
The 1964 agricultural census (MAC 1965: 56) found that only 20 per-
cent of land holders in Ankole hired labor, a little more than half the na-
tional average. It also indicated that the percentage of holdings employing
laborers increased with size but only to 28 percent of holdings of more
than 4 ha (MAC 1966: 36). In the robusta coffee/banana zone in 1974
almost half of the farmers hired labor, which provided only 7 percent of
the average monthly input on all farms (Kyeyune-Ssentongo 1985:
47-48, 55).14 Most hired laborers were not permanent, but worked at
times of peak labor demand. Nevertheless, hired labor was the largest
average expenditure for these farmers (Kyeyune-Ssentongo 1985: 68).
A decade later rich farmers in this zone were found employing up to
seven laborers full time, although many were retreating to family labor,
using hired labor only for short-term contracts for specific projects such
as land clearance (Kasfir 1984-86). Rapidly increasing inflation made
hiring permanent labor unprofitable even for most wealthy farmers.
However, in another respect hired labor has greatly increased since in-
dependence, although to what extent is unknown. Population pressure
and fenced property have led to the practice of borrowing land-almost
unheard of a generation earlier. Since the "rent" for such land is not
paid in cash but consists of clearing it and providing a small part (8-15
percent) of the harvest, farmers who own such land usually do not think
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 61
of this labor as "hired." Consequently, there is reason to think that the
amount of hired labor, although still limited, is growing. Population
pressure on the land is likely to cause it to increase more rapidly during
the next generation.
Kyeyune-Ssentongo (1985: 47, 49, 52) found that family-labor use in
this area was lower than in other nearby agricultural zones, with a mo-
dal average of 600 to 800 hours per year per husband and per wife. On
average, women provided 14 percent more hours of labor on the farm-
in addition, of course, to their work in the home.15 The farmer provided
35 percent, his wives 40 percent, and children provided about 10 percent
of average total annual farm labor input (Kyeyune-Ssentongo 1985: 54).
Women missed fewer days and worked longer hours.
However, he adds that "the days of sex specialization by crop or by
type of operation-if they ever existed-are gone in Central Ankole"
(1985: 54).16 Both husbands and wives contribute labor to every task in
growing each crop. Husbands contributed more than half the labor of
pruning and mulching bananas and coffee, and of harvesting coffee, but
provided almost no labor toward sowing and weeding millet (Kyeyune-
Ssentongo 1985: 45, 55). Bananas and coffee received 80 percent of total
farm-labor input. He found that for both crops the marginal value of la-
bor was well above zero, indicating that farmers could significantly raise
their income by intensifying their use of labor (1985: 59). He also found
that the amount of family labor was relatively constant throughout the
year, and that the higher the labor input on a farm, the larger the share
of labor contributed by children and hired labor (1985: 48, 50).
The most important land modification has been swamp drainage
that is the product of the hunger for land described above. As noted,
custom used to treat swamps as public land that could not be privately
held. But postindependence laws permitting leaseholds over formerly
public land did not distinguish between swamps and dry land. Thus, as
land became more desirable, people filed to lease swamps and then
drained them for crops or pasture. Local fears that this might lead to a
fall in the water table and to a growth in tensions between the new
holders and former users led to an informal administrative ban on
further drainage of swamps (BDG 1982). There is very little terracing or
other forms of soil protection on hillsides, unlike the steeper and univer-
sally terraced slopes of neighboring Kabale District. Coffee and particu-
larly bananas are frequently mulched, a practice which protects against
In the 1950s and 1960s the government enthusiastically promoted
62 / N. Kasfir
new, largely imported, technologies to improve agriculture. These in-
puts-the use of fertilizer, pesticides, better seeds and tractors-were
made available to the fortunate few "progressive farmers" in the hope
that their use would create knowledge of their benefits and demand for
further distribution by other farmers. Today, however, availability of
these inputs depends on the spasmodic generosity of international
The few tractors still in working order are more profitably used for
transportation than for farming. Many farmers insist they would use fer-
tilizer and spray pesticides on their coffee trees if the real price of coffee
were more attractive. However, even if there were a demand for fertilizer
and pesticides neither the government nor private channels could pres-
ently supply them in adequate quantities. Throughout the 1980s, hoes
have been in short supply, and expensive when available in markets. Use
of improved seeds has never been widespread. Farmers rely on the pre-
vious year's harvest for their seed (Kyeyune-Ssentongo 1985: 47). In the
1980s, however, demand for new coffee seedlings surfaced intermittently
when the government price stayed ahead of inflation (Kasfir 1984-86).
The question of measuring increased frequency of cultivation in re-
sponse to increased population and land pressures is complicated by the
shift to perennial crops grown partly for subsistence and partly for the
market. Schemes for measuring agricultural intensification by multiply-
ing the number of times a particular crop is cultivated during a specific
time period, such as the scheme proposed by Turner and Doolittle
(1978), cannot account for reliance on perennial crops. Improved tech-
niques can increase yields, but-by definition-not the frequency of
perennial crops. That makes it more difficult to assess agricultural inten-
sification in this zone where bananas, coffee, and pasture account for
over four-fifths of the area of the average farm and most of the typical
farmer's subsistence staple and cash income.
Other cultigens are planted either once or twice a year at the onset
of each rainy season with almost double the area devoted to crops
planted in the second rains. Beans, peas, groundnuts, sorghum, and
sweet potatoes are generally planted twice, whereas millet is usually
planted once during the more reliable second rains (Kyeyune-Ssentongo
1985: 24). Quickly maturing cultigens, such as peas, beans, and drought-
resistant sweet potatoes and sorghum, are more likely to be planted in
the more unpredictable first rains (Kyeyune-Ssentongo 1985: 24-25).
Cassava and to a lesser extent pineapples have become widespread in
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 63
the 1980s. Maize, pumpkins, yams, Irish potatoes, and sugarcane also
are grown. However, the Bushenyi Agricultural Officer reports that the
frequency of cultivation of these crops has not changed for many years
(Jamil Ziyimba, personal communication, 1988).
The shift in crops has not only changed the composition of agricultural
production in this zone but it has led to an expansion in total produc-
tion. In addition, the current value of sales of coffee and bananas (and
milk, a taste for which also has dramatically increased in the past genera-
tion) far outweigh present sales of all cultigens traditionally grown in
Bushenyi before colonial rule. This change was completed in the first
few years after independence, although the expansion in sales of ba-
nanas was a response to the inflation in urban prices for food that oc-
curred a decade later. Thus, the monetary economy introduced during
colonial rule produced a new degree of output intensity. In that sense,
farmers in this zone have passed through a stage of agricultural intensifi-
cation during this century for reasons that have little to do with popula-
Now, for the first time, population pressure is being felt by farmers
in the zone, which poses the question of whether they are now going
through, or are about to go through, another period of agricultural in-
tensification. With no new cultigens being brought onto the market,
output intensity, if it is to be demonstrated as responsive to population
growth and land scarcity, must be measured by increases in production
over a specified time or area (see Turner and Brush 1987: 6). There are
reasons to think that output intensity is probably increasing and is on
the verge of changing substantially now that the opportunities for leav-
ing land fallow have diminished sharply. Unfortunately, farmers do not
measure their production in standardized units, and data that analyze
changes in yields during the past two decades are not available, so a
precise judgment cannot be made.17 However, measures of certain fac-
tors influencing yields provide some indication of change.
Since independence there has been a substantial and probably accel-
erating reduction of land left fallow. In Kyeyune-Ssentongo's 1974 sur-
vey (1985: 22), one-quarter of the farms had no fallow areas at all and
over one-half had less than 0.4 ha.'8 The expansion of bananas and coffee,
both perennial crops with tree lives of several decades, has resulted in
64 / N. Kasfir
much less area devoted to annual food crops. These two perennials
covered 77 percent of farm area in the 1974 survey-about 0.91 ha on
average (Kyeyune-Ssentongo 1985: 48-49). As dairy production ex-
panded, reclaimed land put to pasture in the past decade has also re-
duced available fallow. Present practices concerning fallow for other
subsistence cultigens are not known. But because each occupies only a
small percentage of the area of most farms-generally less than 0.2 ha,
except for millet, which is grown relatively infrequently-the amount left
fallow to restore soil fertility is low (Kyeyune-Ssentongo 1985: 38). Pop-
ulation pressure and removal of land through leases and fencing ensure
that the crop-to-fallow ratio for the vast majority of farmers will con-
tinue to rise through the next generation.
Coffee yields are reduced by intercropping and lack of fertilizer or
proper maintenance. Trials at government research stations in the 1960s
demonstrated that clean weeding could double annual yields per hectare
and that adding fertilizer could increase yield one-third more (Butt et al.
1970: 197). Mulching and some intercropping with legumes can also in-
crease yield, but intercropping with bananas will reduce yields by one-
half or even two-thirds (Butt et al. 1970: 195, 198). Kyeyune-Ssentongo
(1985: 46) found that yields per hectare of coffee interplanted with ba-
nanas were only 35 percent of yields from pure stands of coffee.
In 1964, when many farmers in Ankole were introducing robusta
coffee, the agricultural census (MAC 1966: 66) reported that it was
planted in pure stands on only one-fifth of the hectarage, it was predom-
inant on two-fifths, and not predominant on the remaining two-fifths.19
A decade later pure stands of coffee in the robusta coffee/banana zone
had risen to three-eighths of coffee hectarage, but, because of quota re-
quirements, without significant changes in the total area planted. Kyeyune-
Ssentongo (1985: 42) attributes this rise to the maturing of coffee.
Farmers could now maintain an economic return to land dedicated to
coffee without relying on a second cultigen.
Since then, the problems of the deteriorating economy, particularly
the low government price for coffee, the rapidly increasing cost of labor,
and the difficulty of acquiring fertilizer and pesticides (for the few who
applied them) have caused farmers to put far less effort into coffee. This
has led to far lower standards of maintenance, which, in turn, have un-
doubtedly depressed yields. Thus, one might expect coffee yields in this
zone to have increased modestly during the first decade of independence
and fallen somewhat since then.
As the increasing hectarage figures presented above indicate, a large
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 65
proportion of the banana plantations in Bushenyi are young, which
probably raises their overall yield. In the 1974 survey (Kyeyune-Ssentongo
1985: 27) the average annual yield of bananas was 9.9 metric ton per
hectare. This compares relatively well with the figure of 11.2 metric ton
that is considered a "good" yield for southern Ugandan conditions (Mu-
kasa 1970: 140).20 Kyeyune-Ssentongo (1985: 33) observes that "in
general Ankole has the best attended banana plots in Uganda." Unfor-
tunately, I do not know of any more recent measures of banana yields.
Table 2.1 gives the findings of the 1974 survey for hectarage, yields,
and number of crops per year for cultigens planted in this zone. The low
average crop yield and large variations in yield for most crops-
frequently more than 350 kg/ha-suggest the potential for increasing
food-crop yields on existing land within this zone (Kyeyune-Ssentongo
1985: 40). The monetary implications of production are discussed in the
The Nature of Changes in the Forces of Production
Throughout the twentieth century there has been a dramatic and univer-
sal shift in Bushenyi from reliance primarily on production for subsis-
tence to a far larger proportion of production for sale in the market.
Table 2.1. Average Hectarage, Yields, and Number of Times Planted per Year of
Cultivation for Major Cultigens, 1974
Cultigen Hectarage Yield (kg/ha)a Number p.a.b
Bananas' 0.64 9,911.0
Clean coffee (mixed with banana)d 0.28 349.3
Clean coffee (pure stand)d 0.17 987.3-
Sweet potatoes 0.15 21,774.5 2
Millet 0.15 1,393.6 1
Beans 0.08 961.2 2
Peas 0.04 432.4 2
Groundnuts 0.03 1,141.0 2
Sorghum 0.03 2,137.4 1-2
Source: Adapted from Kyeyune-Ssentongo, 1985: 23, 25, 27, 46.
aAverage yields conceal great variation in actual yields in this zone.
bNumber of times of planting refers to seasons planted, not to number of times planted per year on
the same plot.
'Unclear whether this refers to pure and/or mixed stands.
dClean coffee refers to hulled sun-dried cherry (hulling removes about 54% of the weight of the dried
66 / N. Kasfir
Except for a few commercial producers, all farmers still grow much of
their subsistence requirements-to which they have added new crops for
sale. In the late 1960s, Good (1970: 80) found that farmers in this area
"prefer to produce the bulk of their subsistence needs themselves ..
As a consequence, demand for farm produce among rural producers
themselves is extremely narrow."
On the other hand, Kyeyune-Ssentongo (1985: 27) found a some-
what greater emphasis on food crops marketed. But, like Good, he felt
that the farmer is not entirely driven by commercial criteria, but rather
sells whatever surplus remains after subsistence requirements have been
met.21 The only exceptions to this latter economic strategy that were
found in the mid-1980s survey of farmers in the same zone were one or
two wealthy farmers who bought virtually all of their food requirements
from the market. Everyone else grew all or as much of their staple (al-
most always bananas) as the land to which they had access would yield.
However, they also invariably bought some foodstuffs in the market
(Kasfir 1984-86). At this time, it is difficult to tell whether farmers are
shifting significantly away from a subsistence-first production strategy
for all crops and adopting a more complex strategy in which a propor-
tion of some nonstaple crops are grown specifically for sale.
In either case, most farmers in this area fit the classic definition of
African peasants who produced much of their own food, but who were
inextricably involved in the wider economic system (Saul and Woods
1981; Kasfir 1986). Even those who were wealthy and who, by most
criteria, were commercial farmers produced most of their own food. One
reason for this appears to be the inefficiency of local rural markets.
There was a wide discrepancy despite seasonal variations between farm-
gate prices and local-market prices. Kyeyune-Ssentongo (1985: 66) re-
ports that in 1974 local Bushenyi food markets were highly active in
livestock and seed crops (beans, groundnuts, and peas) but not in sta-
ples like bananas. The open-market prices for foods were never less than
170 percent of the farm-gate price, and for beans it was always more
than 250 percent (Kyeyune-Ssentongo 1985: 25-27, 73).
The two most dynamic changes in the income of these farmers over
the past decade and a half were the consequences of inflation. First,
remittances from family members in wage employment ceased to be im-
portant. Second, bananas frequently replaced coffee as the dominant
source of farmers' cash income. In the 1960s the mostly likely explana-
tion of variations in income of farmers was the presence of a family
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 67
member holding a salaried position in the city.22 Thus, anyone who
could get at least his sons into secondary school hoped they would find
careers outside farming. By the 1980s the real value of salaries, even of
the most highly paid state officials, had fallen to ludicrously low levels.
Rich peasants, not just the group of commercial farmers, were likely to
earn a higher income than the salary paid to the highest-level civil
In addition, the 1974 survey found that 62 percent of farmer cash
income came from coffee, even though a comparison of returns on labor
and inputs demonstrated that bananas paid much better (Kyeyune-
Ssentongo 1985: 41, 59). The 1984 and 1986 surveys indicated that the
proportion of farmer income derived from bananas had grown consider-
ably, in part because the state price for coffee had lagged behind infla-
tion, whereas the open market price for bananas had not (Kasfir 1984-86;
Benner 1988). Dairy farming also became increasingly attractive, because
the price tended to keep up with inflation.
Neither government nor international agencies play an important
positive role in the farming economy. Low coffee prices set by the gov-
ernment, about 20 percent of the world price, and failure to pay farmers
quickly go a long way to explain lack of interest in that cultigen. Most
of the programs developed in the 1950s and 1960s to supply farmers
with subsidized inputs and technical advice have almost disappeared.
Local agricultural officials cannot spend much time in their offices, be-
cause their salaries will neither feed nor clothe their families. They have
no transport to take them into the countryside. Even in the 1960s, ex-
tension agents tended to concentrate their attention on a few "progres-
sive farmers." Occasionally in the 1980s agricultural-development schemes
were launched, for example in Bushenyi an intensive European Eco-
nomic Community rehabilitation program intended to improve the qual-
ity of coffee. But this program reached few farmers. In the early 1980s,
support from international donors enabled the government to distribute
subsidized hoes widely, but these amounted to only a drop in the bucket
and were frequently diverted to the black market.
Consequences of Changes and Prospects for the Future
To call this area the "robusta coffee/banana zone" is to indicate how
thoroughly fundamental changes in farming have occurred during this
68 / N. Kasfir
century, as neither cultigen was widely grown before the colonial period.
These changes are only a few of the complex basic choices farmers in
this area have made and continue to make. By adopting and selling
these cultigens, they have certainly increased output by unit area and by
time as measured in cash return. These changes constitute agricultural
intensification. This transformation, however, was the direct consequence
of the introduction by the colonial government of a monetary economy
and not a response to increases in population. Thus far, despite popula-
tion density far in excess of typical levels in rural Africa, farmers in this
zone have not yet found it necessary to adopt new production techniques.
The most important change in choice of food staples has been the
shift from finger millet to bananas. This has meant a sharp decrease in
labor input per hectare, as millet requires almost twice as much labor
as bananas (Kyeyune-Ssentongo 1985: 32, 38). Because the hills in this
zone frequently suffer from flooding during rainy seasons, which results
in some crop destruction and loss of topsoil, and because well-mulched
banana plantations check soil erosion (McMaster 1962: 39), the shift to
banana plantations, which tend to be well maintained in Bushenyi, has
had an important side benefit.
Sweet potatoes, the third staple in terms of hectarage, require little
weeding though much labor in the removal of vines. They are planted
on mounds on dry land and on ridges (to avoid waterlogging) in swamps
(Kyeyune-Ssentongo 1985: 36-37). The amount of land now devoted to
foods other than bananas, millet, and sweet potatoes rarely exceeds one-
fifth of a hectare each per household (Kyeyune-Ssentongo 1985: 38).
Less than 10 percent of land devoted to foods in 1958 was planted in
cassava, though it was on its way toward becoming a staple (McMaster
1962: 66, 67). Only 17 percent of the 1974 sample grew it (Kyeyune-
Ssentongo 1985: 38). Since then the amount of cassava grown in this
zone appears to have increased considerably, although I know of no re-
liable measures. The reduced labor cassava requires for food preparation
may have been an important reason.
These changes in food crops have important dietary implications.
Millet is much higher in protein and carbohydrate content and, there-
fore, is more nutritious than bananas, its replacement staple (McMaster
1962: 44, 52; Langlands 1974b: 76). Sweet potato and cassava are ex-
cellent protection against famine, the former because it can be success-
fully and quickly grown when weather conditions are poor, and the
latter because it can be stored in the ground for a long time. Although
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 69
both foods can contribute to a balanced diet, as staples they offer rela-
tively little nutrition (McMaster 1962: 71, 74; Langlands 1974b: 76).
The patterns of disease in this area do not indicate that changes in
diet have caused serious problems in recent years that might motivate
farmers to modify their approach to agriculture. The trends since inde-
pendence in nutritional diseases suggest a pattern of slow improvement,
according to the former Bushenyi Health Officer (Tom Mwebesa, per-
sonal communication, 1988). Malnutrition stemming from protein defi-
ciencies and anemia, a moderate problem in the colonial period, are
much reduced. Diseases exacerbated by overcrowding show a mixed pic-
ture of change. Smallpox has disappeared, whooping cough is much re-
duced, pneumonia and tuberculosis still carry rates of high morbidity
and mortality. Measles has become worse-it is now the most important
cause of death among children.
Sexually transmitted diseases also show a mixed record. Yaws,
common during the colonial period, has disappeared. The prevalence of
syphillis has been greatly reduced, whereas cases of gonorrhea are in-
creasing. Cases of AIDS, an increasingly serious problem in parts of
Uganda, are no longer rare in this rural area. The vector of this disease
follows the truck route around Lake Victoria, to the east of this agricul-
tural zone (Hooper 1987: 474). Body stature has increased since inde-
pendence, despite all the political and economic disruptions since 1971.
There has been an increase in height and weight, as well as in life expec-
tancy (Tom Mwebesa, personal communication, 1988). In 1985 both
crude death and infant mortality rates were only a fraction of what they
had been in 1960. Thus far, it is safe to say, neither the epidemiological
pattern nor overall changes in health and body size suggest that Bushen-
yi is now so densely populated that farmers must immediately adopt new
techniques to intensify agricultural production in order to remedy defi-
ciencies in food supply.
With regard to cash crops during the past two decades, the general
response of farmers in this zone has been to make only small invest-
ments toward introducing new cultigens or production techniques. This
risk-avoidance behavior is probably closely related to the virtual absence
of government technical assistance during this period. Farm studies in
Tanzania demonstrate smallholder willingness to adopt new crops and
new means of production but-in the absence of external support-only
when prices or reductions in cost are especially attractive (Ruthenberg
70 / N. Kasfir
The most important change has been the expansion of banana pro-
duction. Because farmers changed their preference in a staple food to
bananas, because the technology needed for successful adoption is cheap
and well understood, and because bananas at present generally yield
higher returns than coffee, this is a conservative strategy-except in the
relatively small number of cases where coffee was uprooted. Farmers
with a lot of coffee can maintain their trees with virtually no further
investment-at the cost of reduced yields. They switched their attention
to bananas without making large investments or taking serious risks.
Thus, for the moment, most of them follow a strategy of low invest-
The main exception to this strategy is dairy farming. Farmers own-
ing purebred imported cattle or crossbreeds must use acaricides on a
regular basis, which requires them to maintain access to either spraying
equipment or dips. The rapidly rising price of milk, however, has sup-
ported these expensive imported inputs. In addition, dairy farmers
generally erect at least perimeter fencing. The acaricides and related
equipment are essential costs requiring a large capital investment that
has produced a return that probably increased faster than the extraordi-
nary rate of inflation over the past decade in Uganda. But it remains a
gamble. In 1987 and 1988, for example, when inflation went through the
roof, the cost of acaricides far outstripped increases in the market price
for milk. In Bushenyi, the dairy industry is a high-risk, high-yield form
of production by comparison to either coffee or bananas.
Finally, farmers also consider their other food crops as sources of
cash income. Here, as with bananas, most smallholders follow a conser-
vative strategy of limiting sales to surplus production after family sub-
sistence requirements have been met. The difference appears to be that
bananas are planted with the intention of selling a portion in the
market, while in most cases other food crops are planted in order to sat-
isfy domestic consumption. Here too, the dominant strategy seems to be
one of low risk, low yield.
What do these challenges in farming indicate about agricultural in-
tensification? The shift to perennials has reduced the necessity for fal-
low, thus permitting some additional increase in rural-population den-
sity. On the other hand, continued expansion of production beyond
domestic consumption will increase the amount of land needed, as
commercial exploitation is added to subsistence use. But the dynamic
between commerce and subsistence is never simply additive. Helped
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 71
along by political connections, advantaged by changes in land law, and
driven by the pressures of inflation, the alert and the wealthy began to
assemble large farms by enclosing public land and buying out small-
holders. As they continue to fence and lease large tracts of land, poor
farmers are being left with plots too small to provide subsistence. The
latter have already begun to borrow land to grow additional food and to
seek agricultural employment in growing numbers. The closure of the
land frontier in this zone suggests that new forms of agricultural inten-
sification specifically in response to population density are probably on
The creation in the twentieth century of the robusta coffee/ banana zone
was a major agricultural transformation. Taking advantage of good
climate, adequate water, and reasonably good soils at an appropriate alti-
tude, the colonial government encouraged farmers in this area to grow
coffee in order to build the state's revenue base. Poll taxes, school fees,
new markets, and new consumer goods created a demand for cash that
made sales of coffee increasingly attractive to farmers. Once the mistaken
policy of encouraging arabica coffee on settler plantations was changed,
smallholder adoption of robusta coffee became virtually universal.
The reasons for substituting bananas for millet as the subsistence
staple are less clear. Bananas did not become an important cash crop
that rivalled coffee until the 1970s. But they did involve less labor and
provided the convenience of maturation throughout the year. Change in
diet and commodification of the economy resulted in growing new culti-
gens with a dramatic increase in production for the market. By indepen-
dence this change had clearly resulted in a dramatic intensification of
agricultural production, even though the hoe culture remained virtually
unchanged, the population-to-land ratio remained low, and public land
continued to be freely available.
A further agricultural transformation may now be starting, but
characteristic changes in production are not yet in evidence. Thus far,
Bushenyi farmers have continued to rely primarily on their own produc-
tion for their families' subsistence and have taken few initiatives to shift
to neotechnic inputs. The change in balance of cultigens produced for the
market from an almost exclusive attention to coffee to greater emphasis
72 / N. Kasfir
on bananas suggests another increase in output intensity, though present
high inflation creates too much volatility to know for sure. The more re-
cent interest in dairy production, which has led to conversion of cropped
areas and wetlands into pasture, also betokens a new rise in output
These changes have been made with little assistance from the gov-
ernment or international sources. Indeed, farmers in this area have had
to cope with government harassment, with two wars fought near this zone
and, above all, with continuing rapid inflation. Not surprisingly, most
farmers, with the significant exception of a small group of aggressive
commercial farmers, have responded to the highly unpredictable politi-
cal and economic environment in which they now live by making small
capital investments in crop maintenance, particularly coffee, and, in
general, by following conservative low-risk strategies. Farmers in the
dairy industry provide the main exception, because they must rely on
high-cost imported inputs.
Other changes are creating the conditions for passing through a new
threshold of agricultural intensification. High and increasing popula-
tion density in this zone and changes in land tenure have resulted in the
disappearance of public land and in rapidly rising land prices. For the
first time, large farms have been created and fenced, in part by draining
swamps that by custom had been available to all. A sharp reduction in
land available for fallow has occurred, but that has been largely masked
by the shift to perennial crops over the last three generations. The prac-
tice of "borrowing land" to grow food crops, and paying rent in kind
and in labor, has become common. The families of some poor and mid-
dle peasants often endure a "hungry season." The sons of these peasants
will not acquire farms as large as those held by their fathers, nor large
enough to provide subsistence for their families. Nevertheless, they are
inextricably involved in the monetary economy.
On the other hand, rich farmers have not yet adopted new produc-
tion technologies, nor found it economic to hire labor on a large scale.
Most still produce their subsistence staple with some contribution of
family labor and use the same tools as those less well-off. Despite mod-
est technological innovations, these farmers can generally still manage
not only to feed their families but often to produce a surplus in food
sold in urban centers. Population demands have not reached the point
where yields must be increased, even though new land is no longer
available. When that time occurs, a new agricultural technology, perhaps
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 73
the ox-drawn plow, may become a necessary replacement for the hoe for
both rich and poor farmers.
Thus, at the moment, the indicators of entry into a new stage of
agricultural intensification remain mixed. This zone in Bushenyi does
not yet provide a case that helps to settle the question of what happens
when a population becomes so dense that the land cannot support its
farmers unless they fundamentally change their techniques for food
production. But population and agricultural challenges in Bushenyi indi-
cate that a test of that question may not be far off in the future.
I am grateful for the help I received from Hakim Kasozi and Jamil Ziyimba,
my research associates in Bushenyi, from Tom Mwebesa who interrupted his
own vacation to gather information in Bushenyi on specific issues raised by this
chapter and later provided useful suggestions on an intermediate draft, and
from Ceda Ogada and, particularly, Earnest Wotring, my research assistants at
Dartmouth College. Helpful suggestions were also made by members of the
Workshop on "Population Growth and Agricultural Change in Sub-Saharan
Africa." I also wish to thank the National Research Council of Uganda and the
Dartmouth College Research Committee for their support of this project. Any
errors are my responsibility.
1. For example, the only data that have been published so far from the
1980 census (COMPED 1982) have been provisional population figures for ad-
ministrative areas, and these have recently been officially questioned (see popula-
tion discussion below). Annual agricultural reports are still being published, but
the department is about ten years behind. Estimates of cropped area and other
agricultural statistics continue to be collected, but haphazardly. A new national
agricultural census intended to update the data collected in 1963-64 was post-
poned for several years and finally published after this chapter was completed
(MAAIF 1992). Outside researchers frequently encounter suspicious local and
national government officials forced to cope with continuing security problems.
Nevertheless, officials just as frequently go out of their way to assist researchers.
2. The zone also extends to the East into neighboring Mbarara District.
3. Interviews were recorded with 149 heads of rural households in 1984 and
1986 in 4 parishes in Buhweju, Igara, Ruhinda, and Sheema counties (Kasfir
1984, 1986). The Buhweju interviews concerned smallholder tea production that
is part of another zone and will not be considered here. Discussions focused on
the extent of production for subsistence and for the market; other
74 / N. Kasfir
market connections, largely through the purchase and sale of land and live-
stock; the consumption of purchased goods, including food; and the payment of
taxes and children's school fees. The interviews emphasized the present situation
of the farmer but contained a historical dimension through inquiry into the his-
tory of land acquired and crops planted and the pattern by which the first
generation in the family entered the monetary economy.
4. Urban land was not excluded but is trivial. Land kept idle to meet fal-
low requirements was also not excluded. The exclusion of "permanent" swamps
raises difficulties as some swamps are arable and many swamps have been
drained since Langlands made his calculations. Assuming that they were in-
cluded in his figure for permanent swamps, both factors would increase cultiva-
ble area. The impact on population density is problematic, however, as it
ultimately depends on how many people have access to each drained swamp. It
is not possible to determine how many of the swamps have been drained-nor
for that matter how much of the forest reserves has been illegally converted to
agricultural use. For the specific calculation of the land uses needed to deter-
mine cultivable area for each county in Bushenyi District, see Langlands (1971:
171, 175-77, 179).
5. Muwonge (1978: 176) reports much higher population-density figures for
Igara and Sheema counties for 1959, 156/km2 each and for 1969, 222 and
226/km2 respectively. He says he extracted these figures from the Report on the
1969 Population Census (SDMPED 1971; SDPO 1976), but does not give an
explanation or a page reference.
6. At the county level the density figures for 1980 are also somewhat over-
stated because they make no provision for drainage of swamps and reduction of
forest land since Langlands made his calculations. Swamps composed 3.4 per-
cent and 5.8 percent of the total land areas of Igara and Sheema counties respec-
tively in 1971. Forest below 2,000 m amounted to 2.9 percent of Igara. It is
impossible to find out how much either swamps or forest have since shrunk in
these counties, but, as discussed below for swamps, the amount has been consid-
erable. At the same time the fencing of land (including many drained swamps)
for commercial purposes, or just as protection against inflation, has increased
population density on unfenced land by a far greater amount.
7. The analysis of the error in the Background to the Budget 1988-1989
(MPED 1988) is presented only at the district level. Taking 4 percent as the er-
ror factor for particular counties is a reasonable inference. It assumes that the
age structure, at least for infants, is similar throughout the counties of Bushen-
yi. Since virtually no infants were reported for the whole district, the error must
have occurred in all counties.
8. This figure excludes swamps and open water but not forest reserves, na-
tional parks, or urban areas (MAC 1966: 16-17).
9. There was comparatively little fragmentation of plots in Bushenyi in
1964. The agricultural census (MAC 1966: 29) found that on average there were
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 75
only 1.5 blocks per holding regardless of the size of the holding-less than in
several other districts. Over 71.7 percent of the holdings were not fragmented at
all (MAC 1965: 54). A "block" is an area of claimed land that is not interrupted
by someone else's holding or unheld land, regardless of how many plots the
owner has created on it.
10. It is not clear when bananas became an important subsistence food for
Bairu. Good (1970: 178) suggests that it occurred with the arrival of Ganda
agents about 1900, but the evidence he provides is not convincing. By 1920 Ros-
coe (1923: 100-101) regarded bananas as a characteristic, though not primary,
food and source for beer for the Bairu, which suggests that bananas may have
been adopted earlier.
11. Local farmers say that agricultural officials in Bushenyi never encour-
aged them to uproot their coffee (Tom Mwebesa, personal communication,
1988). At the same time, though, Uganda's planners called for replacement of
some robusta with "tea, vegetables, dairy cattle and cocoa" (Uganda Govern-
ment 1966: 65).
12. Even larger crossbred dairy herds up to 300 head were reported in
Bushenyi for the early and middle 1970s, but these could not be sustained pri-
marily because of problems of insecurity.
13. A small ox-plowing scheme to nose-punch oxen and train farmers in
plowing was initiated by Euro Action ACORD in the Oruchinga Valley in
neighboring Mbarara District in 1988.
14. This seems too low, particularly since in two of the villages hired labor-
ers accounted for 13 percent and 26 percent of total hours-and less than 4 per-
cent in all six of the others for which he had data. However, these two villages
were among the three containing farms over 4 ha, and thus more likely to
employ labor more extensively (Kyeyune-Ssentongo 1985: 21, 49). However,
another 8 percent of total labor was contributed from outside the family by
"friends or visitors." Some of these may have been paid, others may have been
15. In this sample, however, there was an average of 1.1 wives per husband
(Kyeyune-Ssentongo 1985: 51).
16. Roscoe (1923: 97-98) notes some gender specialization in the produc-
tion of millet in the early years of the colonial period. A generation later, Oberg
(1943: 581) reports a sharp division of labor between commercial crops grown
by men (though presumably weeded by women) and food crops grown by women.
17. According to the Bushenyi Agricultural Officer, data on yields have not
been collected by government officials since the early 1970s (Jamil Ziyimba,
personal communication, 1988).
18. The amount of fallow varied widely from one village to another
(Kyeyune-Ssentongo 1985: 21, 22).
19. In seven-eighths of the mixed plots, the other cultigen was bananas
(MAC 1966: 69).
76 / N. Kasfir
20. Mukasa (1970: 140) adds that, in general, the management of banana
gardens has "seriously declined" with the development of a cash economy and
with interplanting of coffee, sweet potatoes, and cassava.
21. He suggests that "commercial" farmers plan to grow this surplus,
whereas "subsistence" farmers merely take advantage of good weather. This dis-
tinction fails to grapple successfully with the fact that every farmer is involved
in the monetary economy. Kyeyune's notion is an attempt to reconcile the older
argument that African farmers grow a "normal" surplus in good years because
they follow a famine risk-avoidance strategy (Allan 1965: 38) with the evident
reliance on a cash income by every Bushenyi farmer. There are obvious difficul-
ties in using the concept of "subsistence" farmers as one side of a dichotomy
concerning the presence of commercial intentions when both groups sell on the
22. For example, see Hunt's (1979: 263-64) analysis of Mbere, an ex-
tremely poor farming community in Kenya. This is a particularly striking con-
firmation, because her evidence of the overwhelming importance to farmers of
urban-wage remittance swamped her application of Chayanov's argument that
number and proportion of family laborers would explain peasant income.
Allan, William. 1965. The African Husbandman. Edinburgh and London:
Oliver and Boyd.
BDG (Bushenyi District Government, Republic of Uganda). 1982. Office of the
District Commissioner, Files.
Benner, Christopher. 1988. Commodity Production in Bushenyi. Seminar paper,
Dartmouth College, Hanover, N.H.
Butt, J.D., B. Butters, J.W. Dancer, and D.N. McNutt. 1970. "Coffee: Local
Aspects." In Agriculture in Uganda, 2d ed., ed. J.D. Jameson. London:
Oxford University Press.
Clay, Jason. 1984. Expulsion of the Banyarwanda from Uganda. Boston: Cul-
COMPED (Census Office, Ministry of Planning and Economic Development,
Republic of Uganda). 1982. Report on the 1980 Population Census, Vol. I,
The Provisional Results by Administrative Areas. Kampala: Government
DLS (Department of Lands and Surveys, Uganda Government). 1980. Guide-
lines to a New Land Policy. Mimeo. Kampala.
Doornbos, Martin R. 1970. "Kumanyana and Rwenzururu: Two Responses to
Ethnic Inequality." In Protest and Power in Black Africa, ed. Robert I.
Rotberg and Ali A. Mazrui. New York: Oxford University Press.
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 77
1975. "Land Tenure and Political Conflict in Ankole, Uganda." Journal
of Development Studies 12(1).
1977. "Ankole." In Uganda District Government and Politics 1947-
1967. Madison, University of Wisconsin, African Studies Program and
Uganda Institute of Public Administration.
Good, Charles M. 1970. Rural Markets and Trade in East Africa. Chicago:
University of Chicago, Department of Geography.
Harrop, J. F. 1967. "Soils." Atlas of Uganda. 2d ed. Kampala: Department of
Lands and Surveys.
Hooper, Edward. 1987. "AIDS in Uganda." African Affairs 86(345).
Hunt, Diana. 1979. "Chayanov's Model of Peasant Household Resource Alloca-
tion." Journal of Peasant Studies 6(3).
Ingram, W. R. 1970. "Pests of Bananas." In Agriculture in Uganda, 2d ed., ed.
J. D. Jameson. London: Oxford University Press.
Kasfir, Nelson. 1984-86. "Farm Surveys in Bushenyi and Mbarara Districts."
Unpublished manuscript in possession of author.
1986. "Are African Peasants Self-Sufficient?" Development and Change
1988. "Land and Peasants in Western Uganda: Bushenyi and Mbarara
Districts." In Uganda Now: Between Decay and Development, ed. H.B.
Hansen and M. Twaddle. London: James Currey.
Kyeyune-Ssentongo, L. L. 1985. An Economic Study of the Farming System of
Central and South Ankole. Kampala: Makerere Institute of Social Research.
Langlands, B.W. 1971. The Population Geography of Ankole District. Occa-
sional Paper No. 42. Kampala: Makerere University, Department of Geog-
1 974a. Atlas of Population Census 1969 in Uganda. Occasional Paper
No. 48. Kampala: Makerere University, Department of Geography.
1974b. "Manpower and Nutritional Energy Resources in Uganda." In
Patterns of Food Crop Production and Nutrition in Uganda, ed. R.J.
Hyde and B. W. Langlands. Occasional Paper No. 58. Kampala: Makerere
University, Department of Geography.
1974c. "Soil Productivity Map of Uganda." In Soil Productivity and
Land Availability Studies for Uganda, ed. B.W. Langlands. Occasional
Paper No. 54. Makerere University, Department of Geography.
1974d. "Cultivable Land: Cultivable Land/Population Ratios 1959 and
1969." In Soil Productivity and Land Availability Studies for Uganda, ed.
B. W. Langlands. Occasional Paper No. 54. Kampala: Makerere University,
Department of Geography.
Langlands, B.W., and S.K. Mbakyenga. 1974. "Soil Classification by Coun-
ties." In Soil Productivity and Land Availability Studies for Uganda, ed.
B.W. Langlands. Occasional Paper No. 54. Kampala: Makerere University,
Department of Geography.
78 / N. Kasfir
Mackintosh, W. L. S. 1938. Some Notes on the Abahima and the Cattle Indus-
try of Ankole. Entebbe: Government Printer.
McMaster, D.N. 1962. A Subsistence Crop Geography of Uganda. London:
MAC (Ministry of Agriculture and Co-operatives, Uganda Government). 1965.
Report on the Uganda Census of Agriculture. Vol. I. Entebbe: Government
1966. Report on the Uganda Census of Agriculture, Vol. III. Entebbe:
MAAIF (Ministry of Agriculture, Animal Industry and Fisheries, Uganda
Government). 1992. Report on Uganda National Census of Agriculture
and Livestock. Vols. 1-3. Entebbe: MAAIF.
MPED (Ministry of Planning and Economic Development, Republic of Uganda).
1988. Background to the Budget, 1988-1989. Kampala:MPED.
Mukasa, S. K. 1970. "Bananas (Musa spp.)." In Agriculture in Uganda, 2d ed.,
ed. J.D. Jameson. London: Oxford University Press.
Muwonge, J.W. 1978. "Population Growth and the Enclosure Movement in
Ankole, Uganda." Eastern Africa Journal of Rural Development 2(1-2).
Nabudere, D. 1980. Imperialism and Revolution in Uganda. Dar es Salaam:
Tanzania Publishing House.
Ntozi, J., and G. R. Gasana. 1980. The Provisional Results of the Uganda Pop-
ulation Census 1980: An Explanation of the Growth Rates. Paper delivered
at the Workshop on Development of an Integrated Statistical System for a
Developing Country, location not given.
Oberg, Kalervo. 1943. "A Comparison of Three Systems of Primitive Economic
Organization." American Anthropologist New Series 45(40).
Pingali, Prabhu, Yves Bigot, and Hans P. Binswanger. 1987. Agricultural Mech-
anization and the Evolution of Farming Systems in Sub-Saharan Africa.
Baltimore, Md.: Johns Hopkins University Press.
PUMAFC (Planning Unit of the Ministry of Agriculture, Forestry and Coop-
eratives, Republic of Uganda). 1972. The Agricultural Zones of Kigezi,
Ankole, Bunyoro, and Toro, and Their Potential for Development. Mimeo.
Roscoe, John. 1923. The Banyankole. Cambridge: Cambridge University Press.
Ruthenberg, Hans. 1968. "Some Characteristics of Smallholding Farming in
Tanzania." In Smallholder Farming and Smallholder Development in Tan-
zania, ed. H. Ruthenberg. Munich: Weltforum Verlag.
Saul, John, and Roger Woods. 1981. "African Peasantries." In Political Econ-
omy of Africa, ed. Dennis L. Cohen and John Daniel. London: Longman.
SBMEA (Statistics Branch, Ministry of Economic Affairs, Uganda Protector-
ate). 1961. Uganda Census 1959: African Population. Entebbe:SBMEA.
SDMPED (Statistics Division, Ministry of Planning and Economic Develop-
The Robusta Coffee/Banana Zone of Bushenyi, Uganda / 79
ment, Republic of Uganda). 1971. Report on the 1969 Population Census.
Vol. I: The Population of Administrative Areas. Entebbe: Government
SDPO (Statistics Division, President's Office, Republic of Uganda). 1976. Re-
port on the 1969 Population Census. Vol. IV: The Analytical Report. En-
tebbe: Government Printer.
Taylor, Brian K. 1969. "The Western Lacustrine Bantu (Nyoro, Toro, Nyan-
kore, Kiga, Haya, and Zinaz, with sections on the Amba and Konjo)." In
East Central Africa. Part 13. Ethnographic Survey of Africa, ed. Daryll
Forde. London: International African Institute.
Thomas, D.G. 1970. "Finger Millet (Eleusine coracana [L.] Gaertn.)." In Agri-
culture in Uganda, 2d ed., ed. J.D. Jameson. London: Oxford University
Turner B.L. II, and Stephen B. Brush. 1987. "Purpose, Classification and Or-
ganization." In Comparative Farming Systems, ed. B. L. Turner II and S. B.
Brush. New York: Guildford Press.
and William E. Doolittle. 1978. "The Concept and Measure of Agricul-
tural Intensity." The Professional Geographer 30(3).
Uchendu, Victor C., and Kenneth R. M. Anthony. 1975. Agricultural Change in
Teso District, Uganda. Nairobi: East African Literature Bureau.
Uganda Government. 1966. Work for Progress: Uganda's Second Five- Year
Plan, 1966-1971. Entebbe: Government Printer.
1967. Report of the Committee of Inquiry into the Coffee Industry,
Entebbe: Government Printer.
Williams, F. Lukyn. 1934. "Blood-brotherhood in Ankole (Omukago)." Uganda
1936. "Sowing and Harvesting in Ankole." Uganda Journal 3(3).
3 / Increasing Variability in
Agricultural Production: Meru
District, Kenya, in the
F. E. Bernard
Kenya's current rate of population growth, 3.7 percent per year, is one
of Africa's and the world's highest (Population Reference Bureau 1987).
At this rate of growth, Kenya's national population of 26.2 million will
expand to almost 45 million by the year 2000 (Population Reference Bu-
reau 1992). Although rural-to-urban migration is causing Kenya's towns
and cities to grow at twice the rate of natural increase, about 85 percent
of its people are still rural (Kenya 1986a). Only 17 percent of rural
Kenya is of medium or high agricultural potential.' Population densities
and population pressure on this small island of good land, primarily in
the highlands east and west of the Rift valley, have been building rap-
idly over the past generation, now reaching several hundred persons per
square kilometer-some of the highest rural densities in Africa. Techno-
logical and structural innovations and periodic intensification have also
characterized agricultural change in this region over the past forty years.
Among the most pressured regions in this island of heavy density are
the nine districts of the Eastern and Central Provinces (fig. 3.1) clustered
around Mount Kenya. Population-land relationships here are high (table
3.1). Raw densities in the region range from 29 persons/km2 in vast and
mostly semi-arid Kitui, which has a low capacity to support rain-fed agri-
culture, to the high-potential Kikuyu districts of Nyeri, Murang'a,
Kirinyaga, and Kiambu, all with raw densities of over 200 persons/ km2.
Parts of Kiambu District, which is at Nairobi's doorstep, possess rural
densities in the 600 to 1,100 persons/km2 range (Kenya 1981).
Physiologic densities, perhaps a more realistic measure, surpass the
200 persons/km2 mark, with the exception of Nyandarua, yielding aver-
age amounts of land per person significantly below Kenya's estimated
Meru District, Kenya, in the Twentieth Century / 81
S "I I \
.:. ':*. ,: .: .
*r /;- ^ C v ^'
S '' / District. -r -
< ." \
Nairobi PROVINCE I
" "^; ^ ^ ; ," *': : ..... .
o,. *"' ; ;:' :
360E 40. E
L International boundary
Fig. 3.1. Kenya and Meru District
1988 average of 0.49 ha.2 With a raw overall density of 188 persons/km2
and a physiologic density of 428 persons/km2, Meru District is in the
"less-dense" half of the entries in the table. This is a relative position in
a densely packed subset. The internal structure and processes of change
in Meru have created immense population pressure in medium-and high-
0% 0 C ) -
s0 cqrW \O F
CO -S 4 cOs
0a ca s g
CO S 5 COz z S
r'N 0% % r e O0 T
O N ON mN Cm IRT Ql o
-- !000 '00 00
C- 0 00 C) 00 0 0 0
cr C" (M C10 00 -n C1
00 -W) -C n C -
r mr t 0\ 0\ 00 -0 \ 11
CN 1 4r \o o ". 0- cC= ON
tn ^ ^ f- -4
ON r- C) c) f) 00
v^ \O (N O\ r^ o rn oo0-
00 00 00 1o o r- oa Qu o
IT Q "It qt Q 1o C14 wl
C1 00 r 0% O 00 (O 0% %O
f \O CN r- It %o tn CN 00
C# "O (N 00 ( 0 C
O- rcll 00 en
CV oT (N Cc m' Ccd (^ r^ r
, 00 -
0h 0% f