THE SPATIAL DIFFUSION OF AGRICULTURAL INNOVATIONS IN KISII DISTRICT, KENYA . , ~ Thesis for the Degree of Pb. D. . . ; MICHIGAN STATE UNIVERSITY RONALD D. GARST °' '1972 ||||IIIIlllllllllllllilllllIIIHIIIIIHIIHIIIIIIIIIHII LIBRARY 3 1293 10488 8403 Michigan State University This is to certify that the thesis entitled The Spatial Diffusion of Agricultural Innovations in Kisii District, Kenya. presented by Mr. Ronald Garst has been accepted towards fulfillment of the requirements for Ph.D. degree 111W QW‘IA I/II 02%us- v Wm Date June 2, 197g 0-7639 ABSTRACT THE SPATIAL DIFFUSION OF AGRICULTURAL INNOVATIONS IN KISII DISTRICT, KENYA BY Ronald D. Garst Due to the failure of the modern sector of the economy to generate sufficient employment opportunities and the rapid growth of its cities, particularly Nairobi, the Kenya government has placed greater emphasis on rural development. It is hOped that greater attention to agri- culture and to the rural areas will increase the production of food, create more jobs in the rural areas and thus slow down rural to urban migration. Previous plans for employ- ment and food production expansion based on opening new lands have been abandoned primarily because of the high cost. Instead, intensification of production on presently used land will be the focus of efforts by the Kenya government. In order to intensify production it is necessary, of course, to change current practices and change will require the farmer to accept new techniques, technical inputs or crops, all of which can be classified as Ronald D. Garst innovations. Therefore, the study of innovation diffusion is intimately related to the problem of agricultural in— tensification. Geographers have generally confined their spatial diffusion research to the developed countries, to the neglect of the developing countries. A gap therefore exists in the literature concerning the develOping countries. This study is an empirical investigation of the spatial diffusion of five new crOps and grade cattle in a densely populated, high agricultural potential area of Kisii District, Kenya. The innovations investigated are coffee, pyrethrum, tea, passion fruit, hybrid maize and grade cattle. The data consist of 1935 short interviews conducted in 93 different sampling areas to determine when farmers first adepted the innovations in question and 485 long interviews ascertaining socio-economic and demographic characteristics of the farmers. A total of 55 computer maps, using the SYMAP technique, were produced to depict the spatial diffusion pattern for each innovation over time. These maps show the distribution of adeption per- centage levels every two years from the time of original introduction to 1970, plus a final map for 1971. The general diffusion pattern is as follows: (1) initially low levels of adoption are found at scattered locations, (2) an outward spread at low levels of adoption occurs, (3) the emergence of peaks of higher Ronald D. Garst levels of adOption, and finally (4) the coalescence of these peaks into broad areas of high percentages of adoption. The forward edge of the diffusion wave moves very rapidly outward to encompass much of the final area cfi adoption in about half the time period. After the initial spread, the gradient between no adoption and the rughest levels becomes progressively steeper. Factor analysis of 57 geographic, socio-economic, demographic and innovation-measuring variables reveals little relationship between the innovation measures and 'Huanon-innovation variables as the latter primarily factor mnzby themselves or with the geographic variables. The pmincipal determinant influencing the location of greatest intensity of adoption and use is the place of original introduction. Thus, the mass media communication channels mulextension services are of minimal importance while Person-to-person communication, as exemplified by the P’ersonal Information Field, is the major moving force twhind the Spatial diffusion process. Recommendations fOr policy planners and suggestions for further research are also given. THE SPATIAL DIFFUSION OF AGRICULTURAL INNOVATIONS IN KISII DISTRICT, KENYA BY Ronald D. Garst A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Geography 1972 ‘” ‘~._ ACKNOWLEDGMENTS The field investigation upon which this study is based was made possible by a Fulbright-Hays graduate fellowship, supplemented by a grant from the Michigan State University office of International Studies and Pro- grams. The Institute for Deve10pment Studies, University of Nairobi, provided part of the field Operation expenses. Computer costs were met in part by a grant from the National Science Foundation. As is typical of any dissertation involving foreign field work, numerous persons were involved. Mr. Malcolm N. H. Milne, Planning Officer, Nyanza Province, provided valuable assistance in securing approval for the field investigation. Dr. E. S. Ole-Mperre and Mr. S. M. Wambua, Kisii District Agricultural Officers, were most COOpera- tive and helpful in providing assistance while in Kisii. Numerous other government officials gave freely of their scarce time to provide background information. Dr. Joseph R. Ascroft, of the Institute for Development Studies, who was instrumental in getting the field in- vestigation underway, assisted greatly in problems of ii field sampling, questionnaire design and data compilation. My colleague on the field investigation, Miss Carolyn Barnes, endured the many physical and emotional hardships of field work with unfailing dedication and determination. The Kisii Department of Agriculture provided three agricultural extension agents to serve as interviewers: Mr. Andrew Atone, Mr. Fred Nyagwaya and Mr. Mathew Onsomu. Without them this study would have been impossible. Mr. Timothy Ahoda, Data Processing Unit, Institute for Develop- ment Studies, supervised the data punching and preliminary computer analysis in Nairobi. My graduate guidance committee provided greatly appreciated stimulation, advice and guidance on this dissertation. Dr. John M. Hunter, chairman of the gui— dance committee, sparked my interest in the problems of developing areas in an excellent seminar that eventually saw five of its participants do foreign area field work. My greatest vote of thanks goes to Dr. Hunter for his un- failing confidence in me; the value of which cannot be overestimated. Dr. Stanley D. Brunn, through his ememplary teaching, rekindled my dormant interest in the quantatitive techniques that are so evident in this dissertation. Dr. Ronald J. Horvath, through personal conversations, helped me to clarify many of my own thoughts about the diffusion process. Dr. Carl K. Eicher brought to life the role of rural develOpment in Africa. Dr. Everett M. Rogers, a iii very exciting teacher, got me more involved with inno- vation diffusion than either of us would have expected back in 1969. Each of these men, in his own way, has been instrumental in formulating my academic interests and this research. Dr. Lawrence M. Sommers, Chairman of the Depart- ment of Geography, Michigan State University, helped finance my graduate program with a Graduate Assistantship, an NDEA Fellowship, a National Science Foundation Trainee- ship and, during the final year, a half-time position as Instructor of Geography. Dr. Robert I. Wittick and Mr. Brian P. Holly assisted greatly with computer mapping techniques. Mr. Sherman Hollander and Mr. Mark Sullivan, respectively, are responsible for the cartography and photographic work. Their skill speaks for itself. Each of the abovementioned peOple have been in— volved in some way with the preparation of this disser- tation. However, there is one group of individuals who must go unnamed. They are the nearly 2,000 Gusii farmers who kindly provided the raw material out of which this document was produced. My hope is that the information they provided will help, in some small way, others like them by furthering our understanding of the spatial diffusion process. iv My final expression of appreciation goes to my wife for persevering through two very difficult years. She accompanied me to Kenya, gave birth to our daughter while there, and after returning to East Lansing managed to complete her own Ph.D. dissertation in music as well as perform the duties of wife and mother. A lesser person could never have accomplished so much. LIST OF LIST OF INTRODU Chapter I. II. TABLE OF CONTENTS TABLES . . . . . . . . . . . FIGURES. . . . . . . . . . . CTION. . . . . . . . . . . . GROWTH OF THE KENYA ECONOMY: 1963 to 1971 . . . . . . . . . . . . Introduction. . . . . . . . . Population Growth and Rural to Urban Migration . . . . . . . . . The Employment Problem . . . . . The Kenya Economy: An Overview . . The Modern Sector . . . . . . Industrialization . . . . . . Tourism. . . . . . . . . . Agriculture . . . . . . . . Principal Exports . . . . . . Principal Objectives of the Kenya Development Plan, 1970-74. . . Conclusion . . . . . . . . . SPATIAL AND TEMPORAL DIFFUSION PROCESSES: A REVIEW AND INTEGRATION . . . . . Introduction. . . . . . . . . Temporal Diffusion. . . . . . . Adopter Categories . . . . . . Characteristics of Innovations . . Sources of Information. . . . . vi Page ix xi ll l6 l6 19 21 22 29 32 35 36 36 36 37 39 45 Chapter Spatial Diffusion . . . . . . . Spatial Diffusion Research by Non- Geographers. . . . . Types of Spatial Diffusion . . . Basic Concepts of Expansion Diffusion Personal Information Field . . . Innovation Waves. . . . . The "S" Shaped Growth Curve . . . Simulation of Diffusion . . . . An Integration . . . . . . . . III. KISII DISTRICT: BIOPHYSICAL AND SOCIO- ECONOMIC BACKGROUND TO THE DIFFUSION PROCESS . . . . . . . . . . . Introduction. . . . . . . . . Physical Geography. . . . . . . Climate. . . . . . . . . Vegetation and Soils . . . . . The Population of Kisii District . . The Gusii. . . . . . . . . . Social Structure. . . . . . . Infrastructural Development. . . . IV. DATA COLLECTION AND METHODS OF ANALYSIS Data Collection. . . . . . . . Selection of the Study Area . . . Selection of the Sample . . . The Field Survey and Data Collection The Interview Schedule. . . . . Methods of Analysis . . . . . . Data Coding and the Raw Variables . Diffusion Maps . . . . . . . Factor Analysis . . . Multiple Regression and Correlation vii Page 49 50 53 55 58 60 62 65 68 73 73 75 81 85 87 92 93 100 105 105 105 106 107 108 110 110 111 115 118 Chapter V. THE SPATIAL ATTRIBUTES OF INNOVATION DIFFUSION IN KISII DISTRICT . . . . . Introduction . . . . . . . . . . The Spatial Diffusion Process. . . . . Spatial Diffusion of Coffee Adoption. Spatial Diffusion of Pyrethrum Adoption . . . . . . . Spatial Diffusion of Tea Adoption. Spatial Diffusion of Passion Fruit Adoption . . . . . . . . Spatial Diffusion of Grade Cattle Adoption . . . . . . Spatial Diffusion of Hybrid Maize Adoption . . . . . . . . . Generalizations on the Diffusion Process. . . . . . . . . . Factor Analysis and Multiple Regression and Correlation. . . . . . . . . Factor Analysis A, 57 Variables . . . Factor Analysis B, 31 Variables . . . Factor Analysis B Related to Innovation Adoption. . . . . . . Some Internal Interrelations . . . The Accelerating Pace of Change . . . . Generalizations About the Factor Analysis and Regression and Correlation Models . VI. SUMMARY, CONCLUSIONS AND RECOMMENDATIONS. Summary . . . . . . . . . . . . Conclusions. . . . . . . . . . . Recommendations for Policy Planners. . . Suggestions for Further Research. . . . BIBLIOGRAPHY O O O O O O O O 0 O O O 0 APPENDIX A. O O O O O O O O O O O O O B. O O O O O O O O O O 0 viii Page 120 120 121 124 141 163 170 178 183 191 193 194 202 206 209 213 217 220 220 229 233 236 238 257 258 Table 10. 11. 12. 13. 14. 15. LIST OF TABLES Page Employment in the "Modern" Sector, 1965’1968 o o o o o o o o o o o 18 Exports of Commodities and Services, 1967 and 1974 . . . . . . . . . . . 30 Kisii District, Population Growth Rates. . 87 Kisii District, Percent of Population by Age Group . . . . . . . . . . . 89 Kisii District, Population, 1948-1962-1969. 90 Variables from In-Depth Interviews . . . 112 Adoption Level Percentages . . . . . . 114 Innovation Intercorrelations Percept Adoption, 1971 . . . . . . . . . 123 Cash CrOp Returns, Kenya and Kisii District 1960-1970, KB. 0 o o o o o o o o 128 Small Holder Credit Scheme . . . . . . 179 Factor Analysis A, 57 Variables (Total Explained Variance: 77.73%). . . . . 195 Factor Analysis B, 31 Variables Used (Total Explained Variance: 73.48%) . . 203 Comparison Between Factor Analysis A and B. 205 Multiple Regression and Correlation Variables Significant at .05. . . . . 208 Internal Intercorrelations, Mean Year of Adoption, Mean Acres per Farm, Percent Adoption 1971. . . . . . . . . . 211 ix Table Page 16. Summary of Diffusion Characteristics . . . 228 LIST OF FIGURES Figure Page 1. Adopter Categories . . . . . . . . . 38 2. Linear Diffusion. . . . . . . . . . 55 3. Personal Information Field . . . . . . 59 4. Innovation Waves I . . . . . . . . . 60 5. Innovation Waves II. . . . . . . . . 62 6. "S" Curve in Time and Space . . . . . . 64 7. Kisii District: Location Map . . . . . 74 8. Kisii District: Farmland. . . . . . . 76 9. Kisii District: Hillside. . . . . . . 77 10. Kisii District: Physiography . . . . . 78 11. Annual Precipitation: Kisii Seed Farm . . 83 12. Kisii District: Annual Precipitation. . . 84 13. Kisii District: Population Density, 1969 . 91 14. Kisii District: Locations and Sub- Locations . . . . . . . . . . . 94 15. Kisii District: Roads and Towns . . . . 101 16. Cumulative Percent AdOption . . . . . . 122 17. Kisii District: Cash Crop Marketing Locations . . . . . . . . . . . 126 xi Figure 18. 19. 20. 21. 22. 23. 24. 25. 26. Spatial Diffusion of Coffee (a) 1940 . . . . . . (b) 1945 . . . . (c) 1950 . . . . . . . . (d) 1952 . . . Spatial Diffusion of Coffee (a) 1954 . . . . . . (b) 1956 . . . . (c) 1958 . . . (d) 1960 . . Spatial Diffusion of Coffee (a) 1962 . . . . . . (b) 1964 . . . (c) 1966 . (d) 1968 . Spatial Diffusion of Coffee (a) 1970 . . . . . . . (b) 1971 . . . . . . . . . (c) Coffee Cooperatives . . . . (d) Pyrethrum Cooperatives . . . Coffee: Adoption Level Percentages Spatial Diffusion of Pyrethrum (a) 1950 . . . . . . . . . (b) 1952 . . . . . . . (c) 1954 . . . . . . . (d) 1956 . . . . . . Spatial Diffusion of Pyrethrum (a) 1958 . . . . . . . . . (b) 1960 . . . . . . . (c) 1962 . . . . . . (d) 1964 . . . . . . Spatial Diffusion of Pyrethrum (a) 1966 . . . . . . . . . (b) 1968 . . . . . (c) 1970 . . . . . (d) 1971 . . . . . Pyrethrum: Adoption Level Percentages xii Page 129 129 129 129 130 130 130 130 131 131 131 131 132 132 132 132 139 145 145 145 145 146 146 146 146 147 147 147 147 149 Figure 27. 28. 25L 3(). 3:1. 322. 323. 344. 353. 3(5. 3'7. 238. 139. 4CL 41. 42, 43. Pyrethrum Pyrethrum Pyrethrum Pyrethrum Pyrethrum Pyrethrum Pyrethrum Pyrethrum Pyrethrum Pyrethrum Pyrethrum Pyrethrum Diffusion, Diffusion, Diffusion, Diffusion, Diffusion, Diffusion, Diffusion, Diffusion, Diffusion, Diffusion, Diffusion, Diffusion, Spatial Diffusion of (a) 1956. (b) 1958. (c) 1960. (d) 1962. 1950 1952 1954 1956 1958 1960 1962 1964 1966 1968 1970 1971 Tea Spatial Diffusion of Tea (a) 1964. (b) 1966. (c) 1968. (d) 1970. Spatial Diffusion of Tea (a) 1971. (b) Collection Sta Tea: Adoption Level Percentages. Spatial Diffusion of Passion (a) 1960. (b) 1962. (c) 1964. (d) 1966. tions. xiii Fruit 0 O O Page 151 152 153 154 155 156 157 158 159 160 161 162 166 166 166 166 167 167 167 167 168 168 171 173 173 173 173 Figure Page 44. 415. 4(5. 4'7. 4&3. 453. so. 51. 52 Spatial Diffusion of Passion Fruit (a) 1968 . . . . . . . . . . . . 174 (b) 1970 o o o o o o o o o o o o 174 (C) 1971 . . . . . . . . . . . . 174 (d) Pick-up Points. . . . . . . . . 174 Passion Fruit: Adoption Level Percentages. 177 Spatial Diffusion of Grade Cattle (a) 1962 . . . . . . . . . . . . 181 (b) 1964 . . . . . . . . . . . . 181 (c) 1966 . . . . . . . . . . . . 181 (d) 1968 . . . . . . . . . . . . 181 Spatial Diffusion of Grade Cattle (a) 1970 . . . . . . . . . . . . 182 (b) 1971 . . . . . . . . . . . . 182 Grade Cattle: AdOption Level Percentages . 184 Spatial Diffusion of Hybrid Maize (a) 1960 . . . . . . . . . . . . 186 (b) 1962 . . . . . . . . . . . . 186 (c) 1964 . . . . . . . . . . . . 186 (d) 1966 o o o o o o o o o o o o 186 Spatial Diffusion of Hybrid Maize (a) 1968 . . . . . . . . . . . . 187 (b) 1970 . . . . . . . . . . . . 187 (c) 1971 . . . . . . . . . . . . 187 Hybrid Maize: Adoption Level Percentages . 190 Composite Adoption Index. . . . . . . 215 xiv INTRODUCTION This study is an empirical investigation of the spatial diffusion process in a densely populated, high agricultural potential, African small-holder agricultural area in Kisii District of western Kenya. The principal objectives are: (1) to determine and map the spatial diffusion for coffee, pyrethrum, tea, passion fruit, grade Cattle and hybrid maize; (2) to determine if the spatial diffusion pattern found in Kisii District corresponds to the pattern found in the more developed countries; (3) to determine, with the use of factor analysis, the inter- nal structure of a series of socio-economic, demographic, locational and innovation-measuring variables; and (4) to determine via multiple regression and correlation the direction and degree of relationship between the spatial distribution of the above-mentioned variables and the SPatial diffusion pattern. In addition to the primary objectives mentioned albove the following will be undertaken. First, the role of rural development in the overall economic structure of Kenya will be evaluated to determine why planners have generally reduced their emphasis on industrialization in favor of agricultural development and why they now favor agricultural intensification over new extensive land settlement schemes. Second, the growth of the Kenya economy from 1963 to 1971 will be examined to determine the contribution of each of the major sectors toward foreign exchange earnings and employment generation. A brief examination of the agricultural sector will evalu- ate the role of each of the six abovementioned innovations towards the foreign exchange and employment problems. Finally, the basic objectives of the 1969-1974 Kenya Deve10pment Plan will be discussed. Third, the literature on aspatial diffusion and spatial diffusion will be re- Viewed and an attempt made to integrate the basic concepts of both into an explanatory model. Fourth, recommen- dations for policy planners and suggestions for further research will be offered. It is normally assumed that the adoption of inno— VEttions is ipso facto beneficial to the adOpter. By beneficial it is meant that the welfare of the farmer and his family is improved. However, this is not always the case. If, for example, a farmer removes a sizable Portion of his land from food production and uses it for dOnImercial crop production the overall quality of the family diet may deteriorate. Cash earned from the sale of commercial cr0ps could be used to purchase an adequate diet, but normally an inferior quality diet is purchased. Clearly, the adoption of commercial crops may not always .1ead to an improvement in living standards. However, tliis dissertation is concerned with the spatial diffusion (of? agricultural innovations and not the consequences of t hose innovations . CHAPTER I GROWTH OF THE KENYA ECONOMY: 1963 TO 1971 Introduction An investigation of the spatial diffusion of six agricultural innovations may, at first glance, seem to be far removed from the more generalized problem of economic development. Spatial diffusion is, however, closely linked to the efforts to achieve a higher level of economic development . Due to problems of population growth, rural to Ilifban migration, inadequate employment generation and in— sufficient foreign exchange earnings the 1970-74 Kepya @elopment Plan has focused greater attention on rural development. The commitment of the Government of Kenya to rural development is no longer questioned. Instead, the questions revolve around the ways to most effectively improve rural levels of living. To help achieve that end the government has established a Special Rural Development Program (SRDP). Fourteen areas, represent- ing differing ecological, agricultural potential and DOpulation density conditions were selected in which to develop programs that will raise rural living standards with a minimum of outside capital or personnel inputs. Techniques devised for implementing change in the SRDP areas must be replicable in other parts of Kenya using only normal government staff and financial resources. Therefore, capital or personnel intensive projects will not suffice. To help develop, test and evaluate the techniques for affecting change the government invited the Institute for Deve10pment Studies of the University of Nairobi to establish an SRDP Research and Evaluation Unit (Heyer, 1969, pp. 1-5; and Kenya, 1969a, pp. 174-178). The process by which change is accepted has both a temporal and a spatial dimension. The temporal dimension has been researched rather thoroughly, but research on the spatial dimension of change in the developing countries is lacking. Therefore, this study focuses on the spatial dimensions of change in a rural area. The six innovations used in this study will serve as vehicles for searching out the influences on the spatial diffusion process. The underlying purpose is to investigate the processes involved rather than the geography of six innovations. For it is only by under- standing the spatial diffusion process that an Optimal spatial structure of a change program can be designed. Thus spatial diffusion research that seeks universals can contribute to the more efficient use of scarce developmental resources. Population Growth and Rural to Urban Migration Until recently East Africa has not been considered as a population problem area. Past rates of annual pOpu- lation increase are as follows: Tanzania 1.8 percent (1948 to 1957), Uganda 2.5 percent (1948 to 1959), and Kenya 3.1 percent (1948 to 1962). Newer figures indicate that the rate of increase has increased. Tanzania (1957 to 1967) is growing at 3.1 percent per year, Uganda (1959 to 1969) at an annual rate of 3.9 percent and Kenya (1962 to 1969) at 3.3 percent annually. Growth rates of this magnitude will cause the populations of these countries to double in 18 to 23 years (United Nations, 1971, pp. 408-409). Annual rates of increase of this size have often made it difficult to provide sufficient food for the growing numbers of peOple. However, the demand for food is much easier to meet than is the demand for employment. Increased food demand first became a problem in the developing countries after World War II, so it has existed for some time, and has been met with reasonable success via the use of new technological inputs that allow greater yields and via the expansion of acreage. There is approximately a 15-year grace period between the birth of a child and the time he enters the labor force. Thus, the current growth of the labor force is a reflection of the birth rate of 15 years ago, and today's birth rate will determine the growth of the labor force 15 years hence. Unfortunately, an expanding body of unemployed poses a greater threat to peace and stability than does famine (Brown, 1970, pp. 121-126). The initial thrust of develOpment activity during the 1950's and the early 1960's was an effort to transform the economic structure of developing countries, that is, to change from a rural-agricultural economy to an urban— industrial economy. Development based on economic trans- formation was abandoned when it became apparent that the industrial sector would be unable to absorb adequate numbers of workers. This failure is most evident in the place that epitomizes the modern sector, the city: for it is here that high urban growth rates converge with the lack of employment generation provided by the industrial sector. Employment generation will be considered below, but first rural to urban migration will be examined. Nairobi, for example, expanded between 1962 and 1969 at an annual rate of 9.5 percent. However, this figure includes all nationality groups; African, Asian, Arab and European. The non-African population in Nairobi actually declined by about (24 percent) 21,000 during this time period. So it seems appropriate to consider the growth of the African pOpulation as an indi- cation of future growth trends for the city. During the time in question the African population expanded from 156,246 to 407,736, or at an annual rate of approximately 14.5 percent (Kenya, 1966a, V. III, p. 21; Kenya, 1966c, V. IV, pp. 7, 41, 58, and 70; Kenya, 1970, V. I, pp. 1 and 70). Theoretically it is possible to slow the migration of job seekers to the urban areas by lowering wage rates and by decreasing urban amenities. This would have the dual effect of making labor an attractive alternative to machinery and would probably reduce the rural to urban migration rate. The latter might not be true, because if more jobs are offered it could in fact attract even more workers, in spite of low pay and poor amenities. Un- fortunately, policies that would tend to restrict wages in any portion of the economy, particularly in urban areas, are politically impractical (Harbison, 1967, pp. 174-193). Rural to urban migration in most develOping countries has been going on for several years at a very high rate. The result is that urban areas, and in particular the largest cities, are growing at two to three times the national rate, as in Kenya. In an effort to better understand why this migration has con- tinued unabated in spite of high urban unemployment rates, Todaro has developed a migration model that involves three basic elements. They are: (l) the urban-rural income differential, (2) the expected probability of getting I . A. .- u- l 1 «I I -.,I 'I a 1 I I p" I .H 'oI w. 'A h a job, and (3) the differences in urban vs. rural ameni- ties that will effect one's "real" income. Urban wage levels are normally several times as high as rural wages, therefore even poor odds of getting a job become favorable considering the potential income if a job is secured. Thus the key element in the Todaro model is a consideration of the probability of finding employment. The probability Of getting a job during time period "t" is equal to the ratio of new modern sector employment Openings in period "t" relative to the number of accumulated job seekers in the urban traditional sector at time "t" (Todaro, 1969, pp. 138—148). The real probabilities of finding employment are not as important as perceived probabilities, for that is what primarily influences the movement Of people into the city. For example, the 1964 tripartite agreement in Kenya between government, management and labor unions to increase the number Of jobs by 15 percent was a failure, because it had the effect of attracting numerous new workers into the urban labor market. A few months later employee attrition, not Offset by new hiring, brought the total employment back to the old levels, while total numerical unemploy- ment increased as a result of the migration induced by the prospect of new jobs. Thus the erroneously perceived probabilities of employment were largely reSponsible for the increased migration (Todaro, 1969, pp. 138-148). 10 The migration process generally involves two steps. First, the unskilled rural worker moves into the urban area and spends a certain amount of time in the urban traditional sector. That is, he is not regularly employed in the modern wage-earning sector, but instead is either overtly unemployed, underemployed, sporadically employed, or earns a minimal existence in petty retail trade and services. The second step is the attainment of a per- manent job in the modern sector (Todaro, 1969, pp. 138— 148). The urban bias in social services that attracts migrants to the city can easily be seen in Kenya. Local government Operations are divided essentially into two groups, the seven municipalities consisting of those cities with over 10,000 people, and the 33 county councils that are largely responsible for services in the rest Of the country. Lower levels of government, urban and area councils, are under the authority of the county councils for both budgetary and administrative matters. During the period 1964-68 the expenditures of the municipalities and the county councils were about equal, but the county councils served about 12 times as many people, with a resultant per capita eXpenditure of about 12 times as great. Per capita expenditures in 1968 were KB 16 (0.8. $38.65) in the municipalities and only KB 1.30 (0.8. $3.30) for the county councils. If social services only (education, health, housing and community development) 11 are considered, then the municipalities averaged about 27 times the per capita rate of expenditure for these services (Kenya, 1969, pp. 179-180). The Employment Problem The structural transformation view of development was based on the fact that most rich countries have only a very small portion of their labor forces in agriculture, and therefore the way to modernize was to withdraw labor from agriculture and move it into the industrial sector as rapidly as possible. A corollary reason for this action was the widespread assumption of extensive disguised un- employment in agriculture. That is, a significant portion of the labor force has a marginal product of labor that is either very low, zero, or even negative. Therefore it would be possible to remove large numbers of workers from agriculture with no decrease in agricultural output and utilize that labor in the modern sector (Kao, 1964, pp. 129-144). This assumed redundant labor was seen as a free form Of capital that needed only to be organized. The free aspect took on two forms; first, the removal Of workers from agriculture would not adversely effect agri- cultural output (in fact, if the marginal product of labor was negative the removal of excess labor would actually increase output) and second, the excess labor could supposedly be organized at little or no cost. It 12 later became apparent that the removal of anymore than a few percent of the labor force would reduce the level of agricultural output, for the marginal product of labor was low, but positive (Kao, 1964, pp. 129-144). The main problem was the gross underestimate of the time, effort and amount Of resources necessary for a structural transformation Of the economy to take place. It also assumes that the develOping countries will be able to industrialize on a massive scale, and that the only way to improve living standards is to duplicate the economic histories of the western developed countries. The basic fact is that the poor countries of today are starting with large populations, high pOpulation growth rates and generally limited resources; problems with which most of today's industrialized countries did not have to contend (Nicholls, 1964, pp. 11-44). The very process of transformation from a tra— ditionally oriented economy to a modern economy seems to be a generator of unemployment. Rather than responding to supply and demand, wages increase due to government policies, trade union pressure and the desire of multi- national companies to make the wages Of local employees comparable to those of expatriate employees. As the rate of production increases the poorer workers are weeded out and those remaining become more experienced, skilled and effective at their jobs, so the number of workers 13 per unit of production decreases. Thus it is not uncommon in the less developed countries to find production in- creasing more than ten times the rate of employment (Eicher, a 11., 1970, pp. 8—9; Harbison, 1967, pp. 174- 19 3) . It is becoming all too obvious to planners that in the modern sector of the less develOped countries, increased production per unit of investment is the goal, the same as in the more developed countries. Workers become more skilled and effective, and the investment per worker increases. Today, for example, a factory of a given level of output employs fewer people than a factory of the same output would have ten years ago. Therefore, as the modern sector of the urban economy grows in output, but not in the size of the labor force, the size of the urban traditional labor force grows. This leads to in- creasing numbers of shoe shine boys, petty sidewalk traders, hand-cart Operators, and hangers—on at the SI“all shop owned by family or friends. In an effort to stem the tide of urban migration planners now are looking to the rural areas with the hope 0f controlling migration. In order to keep peOple out Of the cities, they argue, it will be necessary to create IncDre jobs in agriculture and increase rural living standards to a level where it no longer becomes profitable t0 try the Odds for an urban job. This calls for both increased rural incomes and additional rural amenities. 14 Increased agricultural production can be achieved by clearing new land or by increasing the productivity of presently cultivated land. Opening new land for settle- ment has not been overly successful for a variety Of reasons. In a country of already dense population in relation to the carrying capacity of the land, all of the good land will usually be occupied, leaving only marginal land to be Opened by the settler. Most of the peOple who move to settlement schemes are very poor, thus are unable to make capital investments in the land, and are Often Short on skills. The most common result is for the government to invest much more in the settlement scheme than could ever be justified on purely economic criteria (Lewis, 1964, pp. 299-310). Ruthenberg summarizes problems of the Kenya agri- cnltural schemes as follows: (1) There is seldom a cash cJi‘op capable of providing sufficient income to meet ex- Penses. (2) Unjustifiably large expenditures go to hOusing, feeder roads, water supplies, etc. (3) Settlers are usually either the poorest farmers, uninterested in farming, or the formerly landless who possess few manage- Inent skills. (4) Economic returns are not sufficient to EM:tract the better farmers. (5) Squatters, one group for W1"tom the settlements were designed, often prefer to remain where they are rather than move to a new area and subject themselves to new rules. (6) The agency responsible for 15 the scheme Often tries to do more than is technically, economically and administratively possible. (7) When settlers fail to practice good husbandry and agriculture there is rarely any consistent policy of reprimand or correction. (8) The number and quality of staff the scheme could afford is generally insufficient to meet the needs, while an adequate staff is prohibitively ex- pensive. (9) Little or no continuity of policy is detri- mental tO good Operations. (10) The average cost for thirteen schemes has been about Kb 312 (U.S. $800) per family, but that figure does not include the cost of Settlement Officers' salaries (likely to be the most expensive item), overhead costs of associated government units and the interest charges on the money invested (Ruthenberg, 1966, pp. 55-56). The repeated failures of settlement schemes in 0t-her African countries such as Ghana (Miracle and Seidman, 1968, p. 2), Nigeria (Baldwin, 1957, pp. 166- 171), and Tanzania (de Wilde, 1967, pp. 419-420) has led planners to abandon this method of development. The degree of success seldom justifies the level of expendi- t\lre. Given a limited amount Of money it seems more eRpedient to spend it on intensification of presently 0Qcupied land. Consideration will now be given specifically to the Kenya economy based on the problems of employment 16 generation in both rural and urban areas, and the gener- ation of foreign exchange. The Kenya Economy: An Overview The purpose of this section is to examine the main sectors Of the Kenya economy in order to evaluate their contribution toward reducing the problems of employment and foreign exchange. A three-fold breakdown Of (1) the modern sector, with special emphasis on the role of in- dustrialization; (2) tourism; and (3) agriculture, will be used. Discussion will also be included on the principal exports as well as the principal Objectives of the 1970-74 Egya Development P 1an . EhiModern Sector In 1964 the modern sector of the Kenya economy as<=counted for about 64 percent Of all wage employment. The remainder consisted of employees on small-holder farms outside the settlement schemes, employees of the Settlement schemes and rural non-agricultural activities. In the rural areas the distinction between wage employ- Inent and self-employment is not always clear-cut. A farm oWner may work occasionally for someone else, for the gOvernment, or perhaps in a nearby town. Thus a fair number of peOple probably are counted twice, as wage e11lployees and as self-employeed. Altogether there were about 4,200,000 persons in Kenya engaged in economic 17 activities in 1968, including those counted twice. All wage employees accounted for roughly one-fourth of that total (Kenya, 1968, pp. 119-120). Employment (see Table l) in the private modern sector (wage employment in privately owned enterprises) has actually been declining in recent years. Between 1965 and 1968 there was an overall decrease in employment of 3.7 percent. The decline in agricultural employment can be attributed to the transfer of ownership from the large-scale European farms to private smallholder status, with the former wage employees becoming self-employed. The steady decline in commercial employment is probably due to the departure of non-citizen Asians and Europeans. Also the departure of these high-income peOple and their Purchasing power would have the effect of depressing this industry. The only classifications accounting for a Significant increase in number of employees are manu- facturing and repairs, and building and construction (Kenya, 1969, p. 121). Had it not been for a steady rise in government employment, the overall modern sector would have declined in employment between 1965 and 1968. The public sector, which in 1968 accounted for about 36 percent of the modern Sector grew by 14.9 percent between 1965 and 1968. About three-fourths Of the public sector employment is accounted fcr the Kenya Government and local governments. When the 18 TABLE l.--Employment in the "Modern" Sector, 1965-1968. Number Of Employees ('000) Percent Private Sector Change, 1965 1966 1967 1968 1965-68 Agriculture & Forest 202.4 188.1 172.7 173.0 -l4.5 Mining & Quarrying 2.3 2.3 2.4 2.9 26.0 Manufacture & Repair 52.1 52.4 56.8 58.2 11.7 Building & Con- struction 8.7 10.3 17.4 18.1 108.0 Electricity & Water 2.5 2.7 2.8 2.7 8.0 Commerce 46.5 46.1 43.5 40.1 -13.7 Transport & Com- munication 12.0 14.8 18.1 18.0 50.0 Other Services 75.8 79.3 75.2 73.6 - 2.9 Total Private Sector 402.0 396.0 388.6 386.8 - 3.7 Public Sector 188.2 200.4 212.2 221.9 17.9 Total 590.2 596.4 600.8 608.7 3.0 k Source: Kenya Economic Survey, 1969, pp. 120-121. e1’ltire modern sector, both public and private, is con- Sidered, the increase in employment between 1965 and 1968 arrlounted to 3.2 percent, or about 1 percent per year. C(Drusidering that the modern sector is primarily urban, and that it employs only 14.5 percent of all people en- gaged in economic activity, the growth rates mentioned above are unimpressive. Indeed, the contribution to total employment provided by this sector is minimal (Kenya, 1969, pp. 119-122). It should be noted in examining Table 1 that while most categories showed large percentage increases in employment for the 1965—1968 period, their contribution to employment expansion is 19 not large due to the small base. The two categories showing the greatest decline in employment, agriculture and forestry, and commerce, were also the largest em- ployers. Therefore employment in the private sector declined by over 1 percent per year. Industrialization The potential for manufacturing expansion in Kenya is limited by a small natural resource base that does not show much promise for future expansion, by a limited domestic market, and by rising unit costs that make Kenya products less competitive in the world market. There is no single natural resource in Kenya, such as Petroleum, that could serve to support a major industrial cemplex. The national market, limited in size and pur- czl'lasing power, could be expanded to include all of East Africa; but the worsening political climate between Kerzya, Uganda and Tanzania precludes basing large-scale industrialization on such a market. In spite of the fact t1‘lat Kenya has a vast reservoir of cheap labor the unit costs of manufactured items are high. This is due largely to lack of worker training and skill, caused by the utter neWness of the factory discipline and rules. Unfortu- nately, the net result is a per unit-cost that matches or s‘urpasses that of the developed countries. And consider- ing transport costs and tariff barriers, the overseas market is quite limited (IBRD, 1963, pp. 153-154). “" I ! Jonv .AL 20 The Kenya government remains committed to the process of industrialization primarily because of the need for foreign exchange earnings. The country has not yet exhausted the opportunities for import substitution industries. An additional national goal is to increase the degree of processing on raw materials produced in the country and gradually move to the exportation of processed goods rather than raw materials (Kenya, 1969, pp. 304-305) . As regards the annual increase in manufacturing Output, Kenya has been doing rather well. Overall manu- facturing production rose from a 1964 base of 100 to 105 in 1965, to 112 in 1966, to 116 in 1967, and to 125 in 1968. The average growth was 6.2 percent per year (Kenya, 1969b, p. 83). These figures suggest that while manufacturing output is increasing enough to make it important as a factor in import substitution aimed at S‘aVing foreign exchange, it is not nearly as successful at employment generation. The 1970-1974 Kenya Development Plan calls for an 8' 9 percent annual increase in manufacturing production between 1967 and 1974. Given past performance of manu— facturing output an increase of this order Of magnitude does not seem unreasonable. It also calls for an annual in'lzrease of 3.7 percent in manufacturing employment, ekactly the same as the 1965 to 1968 mean (Kenya, 1969, p~ 314). .. ah- .1! i . 21 No mineral wealth of great consequence has been discovered in Kenya to date, nor are any great discoveries anticipated in the near future. Oil exploration that began in 1960 has almost completely been abandoned. Almost half of the mineral production is accounted for by soda ash. Salt makes up about one-fourth of the total production and gold about one-sixth. Overall, employment in mining and quarrying amounts to less than three thou- sand people, and in 1967 it accounted for only 1.3 per- cent Of the total exports (IBRD, 1963, pp. 146-150; Kenya, 1969, pp. 153-156; Kenya, 1969a, pp. 93-95). Tourism Tourism is the fastest growing segment Of the Kenya economy. In 1963 the International Bank for Re- <2onstruction and Development mission to Kenya suggested that because Of its important contribution to foreign eMohange earnings, tourism be given the highest invest- ment priority (IBRD, 1963, pp. 170-175). Importantly, the annual increase in the number of foreign visitors is clArrently on the order of 25 percent. The total number 015 foreign visitors rose from 50,000 in 1962 to 257,000 in 1968 (Kenya, 1967a, p. 66; Kenya, 1969a, p. 101). In teJi‘ms of foreign exchange earnings tourism ranks third to the general categories of primary agricultural pro- ducts and manufactured products. It is larger than any single agricultural or manufacturing export. In 1968 .Inw 22 foreign exchange earnings from tourism amounted to Kb 15 million (U.S. $38 million) and is expected to increase to Kb 37 million (U.S. $95 million) in 1974. The figure for 1974 represents about 75 percent of the total income de- rived from tourism, the other 25 percent will go for the cost of imported goods used by the industry and repatri- at ion of profits (Kenya, 1969, pp. 427-428). In terms of employment the tourist industry is not important to the Kenya economy, for the current employment amounts to about 20,000 people and the projected employ- mexit for 1974 is roughly 40,000. This amounts to slightly less than 1 percent of the economically active population (Kenya, 1969a, p. 452). Thus tourism justifiably ranks Vary high on the investment priority list because of the foteign exchange it generates, but unfortunately the same a1':'<3ument cannot be used for employment generation. Aar iculture Agriculture has been and will continue to be the maj or sector of the Kenya economy. In terms of export p“doduction for the earning of foreign exchange it is cL‘Ji‘rently the leader. In the realm of employment gener- a‘tion agriculture is also the most important. Largely beeause of the latter reason agriculture has been given the highest priority in the allocation of financial and t QChnical resources (IBRD, 1963, p. 63). w ‘- 9",?!7 L 1--—‘" ran-'21s“ . . 23 Agriculture's contribution to the Gross Domestic Prtaduct of Kenya is the obvious reason for its high priority position. In 1967 non-monetary agriculture marie up about 21 percent of total output and in the morietary sector about 13 percent. About one-third of thee GDP consisted of agricultural products and about 60 pezrcent of the value Of commodity exports are raw or pr<>cessed agricultural products. Of utmost importance 1&3 the fact that about three-quarters Of the population derives its livelihood from the land (Kenya, 1969, p. 191). The following is a brief evaluation of each of the major cash crops and grade cattle treated in this dissertation in terms of the contribution to export eearnings, and employment and the prospects for future Siltowth. Coffee, the leading export cash crOp, faces at ‘hDEESt an uncertain future. The 1962 International Coffee ig‘sgreement set export quotas for producing countries be- 'C=iause of a tendency for world production to amount to Eiloout 130 percent of annual consumption. Therefore, in JLEB64 Kenya imposed a ban on further plantings of coffee tiltees. Old trees could be replaced and new growers could eaInter the market via the division of old plantings among hew farmers. As will be seen later, the ban on further <=us levels. If the North American market Opens up, pr<>duction could be tripled immediately (Interview NO. 5). Hovvever, in anticipation Of a slowly expanding market pro- ducztion is not scheduled to triple until 1974. Even then thee total amount of foreign exchange earned will be on thee order of K5 210,000 (U.S. $540,000), a rather small .fiiggure when compared to the other principal export crOps. ASLrnost all production will continue to be on small farms. III the Kisii area it is anticipated that passion fruit Wigll.replace pyrethrum as the relative price structure begins to favor the former (Kenya, 1969, pp. 250-251; Kearrya, 1969a, p. 69, and Interview NO. 4). Maize production has increased tremendously in Ideecent years, largely due to the introduction of hybrid "Eirieties. In 1965,for example, some 80,000 metric tons <3“15 maize were imported from the United States, and during Itllie first few months of 1966 another 140,000 metric tons were imported (Kenya, 1967a, p. 32). By 1968 production 1Plead increased to the point where the government, through 't111e Maize and Produce Board, was able to export 250,000 InEetric tons, worth Kb 4.8 million (U.S. $12.3 million) (Kenya, 1969b, p. 65). But in order to do this the C3<>vernment subsidized the Maize and Produce Board about Keno.“ (U.S. $0.02) per kilogram. In the future the Isairmer will be paid less for his product as the 28 efficiencies brought about by the introduction of hybrid varieties lower costs. Also, more efficient bulk handling facilities will lower the cost of transporting the maize from farm to dockside by about KSh. 0.07 (U.S. $0.01) per kilogram. The amount of maise exported is expected to rise to 430,000 metric tons, wroth KI: 7.6 million (U.S. $19.5 million) by 1974. However, at that time there will be no government subsidy, thus the value to export earn- ings will be much greater than before. Currently most f armers grow some maize and in a few years nearly all of them will be growing hybrid varieties. Some use will have to be found for excess production. In anticipation of this, maize will increasingly be used for stockfeed (Kenya, 1969, pp. 237-238). Approximately four-fifths of Kenya is too dry for c-‘—'L:Iltivation so the government is looking to cattle as a way to intensify the utliization of this land. Of the eetimated 7 to 7.5 million cattle in the country, only a S"hull portion are being raised for commercial purposes. Most are owned by semi-nomadic herders who do not raise the animals specifically to be sold or to produce milk. Qlzrrently most of the cattle are not of a sufficient gnality to be sold on either the Kenya urban or the international market. Also, they do not produce enough In.1'1k to be of great value to the owner. Today, through- <2tut the country, efforts are underway to increase the gnality Of the cattle. This is done either through the 29 iJitroduction of grade cattle to replace the local cattle or"through a program of upgrading via artificial insemi- nation (Kenya, 1969, pp. 251-268). About 25 percent of the total cattle slaughterings arre marketed through the Kenya Meat Commission and about healf of this es exported while the remainder is sold in trae urban areas. There is a ready market, both local and iriternational, for beef, so the limitations on the growth 01? the industry are on the supply side. As for the dairy ermdustry, there are two trends. One is an effort to in- <21:ease the amount of milk available for the urban areas «341d for the Kenya Cooperative Creamery, Ltd. to process lilito cheese and dry milk. The other is to expand the Ei\nailability of milk for consumption by the rural African ENOpulation. In order to accomplish these goals grade <=ows are replacing the zebu cows that produce only about c>1'1e-tenth as much milk. With higher production per cow ‘lhe number of animals in the more densely populated areas SShould go down. In this way, with virtually no change in 3Land requirements, milk could be made available to the local people (IBRD, 1963, pp. 126-132; Kenya, 1969, pp. 3251-268; Kenya, 1969a, pp. 74—78). Eirincipal Exports The respective roles of agriculture, manufacturing iand tourism in the export economy of Kenya can be ascer- 1:ained from Table 2. As far as commodity exports are a rep—aw.- ”- 30 TABLE 2.--Exports of Commodities and Services, 1967 and 1974. Share of Total, % 1967 1974 A91: icultural Primary Products C2<3ffee 12.0 4.7 T ea 6 . 0 7 . 7 Maize 1 . 1 3 . 7 V011eat 1.2 0.8 Iiice 0.1 0.3 S isal 1 . 6 0. 8 Cotton 0. 5 0. 4 CIther Agricultural Products 3.8 3.2 1.. Total Agricultural Products 26.4 26.7 P recessed Agricultural Products Ddeat Products 2.6 1.9 [Dairy Products 1.7 0.5 (Zanned Fruits & Vegetables 0.9 1.5 Pyrethrum Products 2 . 2 l. 8 VVattle Products 0.7 0.3 .Animal & Vegetable Oils & Fats 0.3 0.3 (Other Processed Agricultural Products 1 2 1.1 , 2. Total Processed Agricultural Products 9.6 7.5 3. Total Primary & Processed Agricultural Products (1&2) 35.9 34.1 4. Forestry, Hunting & Forestry 0.5 0.4 5. Minerals 1.3 0.8 Q ther Manufactured Products IBeverages and Tobacco 0.7 0.4 'Textiles 1.8 1.5 Clothing & Footwear l. 6 2 . 0 Wood Products 1 . O 0 . 9 IPaper and Printing 1.7 1.5 ILeather Products 0.2 0.5 IRubber Products 0.3 1.1 Chemical Products 2 . 8 4 . 3 iPetroleum Products 9.0 7.0 (Other Mineral Products 1.5 1.6 IMetal Products & Machinery 2.3 1.6 IMiscellaneous Products 0.4 0.5 6. Total "Other" Manufactured Products 23. pk N N 0 KO 31 TAB LE 2 . --Continued . Share of Total, % 1967 1974 ‘7 - Total All Manufactured Products (2+6) 33.0 30.3 8 - Total Commodity Exports (1+4+5+7) 61.2 58.2 Exports of Services Freight and Insurance 7.4 7.3 Other Transportation 12.3 11.4 Foreign Travel (Tourism) 11.3 18.0 Other Services 7.9 5.1 9. Total Services 38.8 41.8 10. Total Exports of Commodities & Services (8+9) 100.0 100.0 Source: 1970-74 Kenya Development Plan, pp. 153 and 157. 32 concerned, the relationships between the various cate- §;<:>1::ies will remain substantially the same. Primary and processed agricultural products will continue to dominate vvnj_t:11 about 35 percent of total exports for the period 1967 tzco» £1974. Coffee and tea will continue to dominate agri- c:\J;1:tura1 exports, but tea will gain and coffee will de— CZJLJiJfie in relative position. Maize will climb to 3.7 Percent of all exports by 1974. Manufactured products will decline slightly from 23.4 percent of total exports 5—11 1967 to 22.9 percent in 1974. This seeming stagnation 5953 largely due to more of the products being kept in I(Ganya as import substitutes. By far the biggest gain in the balance of payments account will be made by tourism. ‘N7jsth an anticipated annual gain of 14.1 percent this 'C2Eitegory will bring in fully 18.0 of all foreign exchange ‘j‘rl 1974, or KB 37 million (U.S. $95 million). This com— I53513:an with 9.7 percent and 7.7 percent, respectively, for QQffee and tea. Another service category, simply labeled ..‘:>ther transportation" is expected to bring in Kb 23.5 Inillion (U.S. $60.2 million) in 1974 or 11.4 percent of tgtal exports (Kenya, 1969, pp. 153-156). Principal Objectives of the Kenya Development Plan, 1970-74 The basic objective of the 1970-74 Kenya Develop- M can be summed up in the following quote: '. - . . rural development should not be seen as a special 33 programme but as the underlying strategy of the whole Plan" (Kenya, 1969, p. 2). The reasons for the emphasis on rural development are three. First, is the desire to attain a more uniform and equitable distribution of the national income between economic sectors, individuals and areas of the country. Second, by creating better living conditions, higher incomes and more employment Opportunities in the rural areas it is hoped that the massive movement to the urban areas will be reduced. Third, it is recognized that the modern sector, and particularly the industrial sector, will not be able to generate sufficient employment opportunities, so efforts are now directed to the rural areas to achieve that end (Kenya, 1969a, pp. 1-4, 11-12, 15-16). Due to its small portion of the total economy, wage employment in industry and commerce will not be a S:Lgnificant generator of employment, even at the highest Q0Inceivable growth rates. Therefore the Plan is geared EQWard the creation of self-employment opportunities in the rural areas on large-scale farms, small-holder farms a‘t‘td in rural non-agricultural activities. An important goal of the Plan is free and uni— Qrsal primary education. But the very achievement of that goal can exacerbate problems. An urban, or at least Q hon-agricultural bias in education is a large cause of l‘ban migration. Curriculum changes could do much to H- '7‘('_' v 34 lessen urban and wage employment aspirations and strengthen a rural orientation. Rural development is more than just agricultural development. It is also the improvement of rural amenities such as health care, housing and water supplies. These improvements will not take place in the rural areas them- Selves during the plan period, but will be located in numerous small urban centers and what are classified as major rural centers. Efforts are being made to increase the attractiveness of major urban areas other than Nairobi and Mombasa. The principal method will involve the placing of new factories and other enterprises in designated L1JE‘ban growth centers to act as a stimulus to concomitant growth. Over the Plan period KI: 43 million (U.S. $111 million) will be spent on improving the secondary feeder road system to tie into the already adequate primary road thwork. This will help to alleviate the rural-urban d ‘ . . . . . . {Sparlty in liVing conditions. Import substitution industries and controls on 65 importation of luxury items W111 help to keep the lance of payments defiCit to a minimum. A major com— rlent of the future foreign exchange picture Will be t QLlrism, for it significantly lowers the country's Ge . . Dendence on foreign capital for development. To meet Velopment needs not met by domestic saVings, foreign i hvestment will be encouraged. The government has 35 provided guarantees on the repatriation of profits and capital, and on compensation in the event of nationali- zation (Kenya, 1969, pp. 1-19). Conclusion The above examination of the Kenya economy should make clear the reasons why the government is turning more towards rural development rather than towards accelerated urban development. Problems engendered by urban growth are simply exacerbated by attempts to solve them by the application of resources to the urban areas. New urban j Obs often increase, rather than decrease, urban unemploy- I"Rent as they may attract a greater number of migrants. If the ability of a city to adequately accommodate more pGOpIe is increased that new-found ability is frequently Q"erwhelmed by masses of new migrants. The shift in er“phasis towards rural development is an attempt to 8lee some of the problems of production, employment a‘Iid the provision of social amenities at their source a"bid reduce the concentration of these problems in the Q ities of Kenya. CHAPTER II SPATIAL AND TEMPORAL DIFFUSION PROCESSES: A REVIEW AND INTEGRATION Introduction The purpose of this chapter is to (1) review the relevant literature on the diffusion of agricultural inno- vations as it pertains to social characteristics of a~<1<>pters, innovation characteristics, and sources and tyPes of information; (2) review recent spatial diffusion literature commenting specifically on the relationships be‘cween spatial and temporal diffusion; and (3) attempt to integrate the basic concepts of temporal and spatial diffusion into an explanatory model. Temporal Diffusion The social system in which innovation diffusion takes place is a collection of interacting and function— $1 1y differentiated units. The intensity of interaction e‘1').d the distance over which it takes place are directly thated to the degree of functional differentiation be- Qeluse each spatial unit may tend to perform functions QQ:nsiderably different from neighboring spatial units. 36 e +1- 1- fifl 37 Thus, in a more modernized society individual spatial units have more functional differentiation so there will be more interaction over space. Increased functional and spatial differentiation will cause a lessening in intensity of local interactions and an increase in wide-ranging re- lations. In a more traditional society where each spatial unit performs the same basic functions the intensity of interaction will be most intense with those units in close Pr<>pinquity and less intense with non-adjacent units. For example, in a densely populated agricultural area of 100 SQUare miles where any single square mile is nearly identi- Cal in character to any other [square mile] there will be little need for interaction between non-adjacent units 3 ince they all perform the same basic functions. So the C=haracter of the local milieu strongly influences the Process of diffusion (Timms, 1971, pp. 138-140; and Wilson and Wilson, 1945, pp. 24-44). Ifig‘pter Categories Most diffusion research on agricultural inno- vations starts with the readily observed notion that improved farm practices are not uniformly accepted within a Pepulation. The basic hypothesis posited is that differ- ing personal characteristics lead some peOple to adopt innovations more readily than others. Considerable r3 search has been directed toward dividing a population into groups according to the relative date of adoption as e 38 compared to others in their social system. The most common set of adOpter categories is that postulated by Rogers (Robers, 1971, pp. 183-185). They are: (l) innovators, (2) early adopters, (3) early majority, (4) late majority, and (5) laggards. Figure l ADOPTER CATEGORIES Early Majority Late Majority Early Adapters ”2% Laggards Innovators I l3'/2°/o 340/0 : 340/0 ISO/o TIME OF ADOPTION OF INNOVATION—e Through the use of field surveys and data aflatlysis it was determined that earlier adopters in India belonged to higher casts (Bose, 1961) , were more literate (Rahudkar, 1962; Savale, 1966), had more education, were of a higher social status, had larger farms, and were more wealthy (Shetty, 1966). While most studies seem to ind icate that young people are more prone to adOpt some- thing new, Chaudhari and others (Chaudhari, 1967) found the opposite tendency (see also, Marsh and Coleman, 1955; Grcss, 1949; Graham, 1956). Earlier adopters have greater exposure to both interpersonal and mass media communication channels than later adopters. Thus, earlier adopters will have a wider 39 spatial range of personal contacts, whereas later adopters tend to have highly localized personal contact patterns. Also, earlier adopters rely more heavily on mass media communication channels than later adopters, and later adopters require more interpersonal communication before they will adopt an innovation. The communication pattern of earlier adopters is thus characterized as being more cosmopolite while the later adopter is more localite (Rogers, 1971, pp. 188-191). Characteristics of Innovations Simple logic would suggest that the overt charac- teristics of an innovation should have an influence on tdae rate of adoption, and indeed, such is the case. Most Cfllaracteristics of innovations can be categorized accord- ilmg to the system suggested by Rogers. The five cate- gories are: (1) relative advantage, (2) compatibility, (:3) complexity, (4) trialability and (5) observability (Rogers, 1971, pp. 137-157). Even though different researchers have generally agreed on the major characteristics, there is not agree- rn‘itrt on the relative influence of each. As regards farm IDITactices, one of the most important controversies started wllen Griliches (Griliches, 1957, pp. 501-522) examined reEgional differences in the rates of adOption and use of ‘fllfbrid seed corn in the United States. He argued that ‘:4}1e delay in the develOpment of hybrids suitable for a ["1 t—p'-..“-‘-_mpm. 3m x I 4O particular area and the delay of seed producers in pro- viding seed for those areas was explained by varying profitability. Profitability here was defined as the density of the potential adoption market, costs of the innovation and marketing costs. Differences in the maxi— T mum potential rate of adoption and the rate of adoption over time were explained, in part at least, by differences in profitability between hybrid and open pollenated varieties. Later Havens and Rogers (Havens and Rogers, 1961, pp. 409-414) stated that an innovation must be economically profitable even to be considered by most farmers, but the interaction effect (defined as the informal pressure to adept caused when those who have adopted an innovation irifluence those who have not yet adopted) was the most important variable in explaining the rate of adoption over Every acceptance of the innovation served as a They maintained tlirne. St imulus for the non-adopters to accept. t11€it if the only variable were profitability, the rate the adoption would be as rapid as profitability could be demonstrated . Griliches responded (Griliches, 1962, pp. 327-330) 'V'thh a comment critical of the Havens and Rogers article (Drl the interaction effect because, he claimed, rather tlklan ignoring the interaction effect, his model was based t1 it, and that the differences in the average rate of 41 adoption between different situations could not be ex- plained by interaction. Finally, Rogers and Havens (Rogers and Havens, 1962, pp. 330-332) wrote a rejoinder to Griliches explaining that the argument was over the relative importance of profitability versus interaction. Griliches contended that profitability was the most im- portant factor while Rogers and Havens contended that profitability was necessary for adOption to take place at all, but interaction explained the rate of adoption from year to year, and perhaps even spatially. Unless the innovation in question provides a significant increase in return, (Desai and Sharma, 1966, ‘pph 141-154), no matter what the initial cost, it will IMDt be adopted. Desai agrees with Rogers and Havens in tllat regard but not with Fliegel and Kivlin, and Jones (Ffiliegel and Kivlin, 1962, pp. 364-370; Jones, 1967, pp. 13‘134). In separate statements they contend that high- cOst practices are adopted at least as fast as low-cost innovations. However, four years later Fliegel and Ki-Vlin (Fliegel and Kivlin, 1966, pp. 197-206) reversed thEmselves on the basis of new evidence and maintained t1lat cost was a factor. Cost effects the adoption rate Elczcording to the size of the farm in question primarily kbeicause the medium— and large-scale farmers perceive ex- EDEEnsive items as investments whereas the small-scale 15Eirmer sees only the cost. 42 Day (Day, 1971, pp. 68-76) has further postulated a series of economic factors that will influence the adoption of new techniques. These could be subsumed under the rubric of relative advantage of the innovation in the Rogers schema. Briefly, they are as follows: (1) The magnitude of the capital investment necessary for adoption is directly related to the economic return expected. Often a gradual decline in further adoption is caused by the lowered expected profitability of additional investment because new techniques may lead to increased output which can saturate the market. (2) Uncertainty of economic re- thrn may reduce the adOption rate or the extent to which ari individual firm or person adopts. Also, there may be the desire to save capital to invest in new and improved techniques as they come along, thus reducing the adoption Jtéit:e for the old innovation. (3) The availability of the iI'D-l'iovation may not be sufficient to allow all who want to acitszt to do so. (4) The supply of financial capital may IICDi: be sufficient to allow all desired adoptions to take E33~ace. (5) Given the fact that managerial skills tend to k>€3 normally distributed there will be laggards who refuse t1<> adopt until almost everyone else has done so. (6) A SacDcial system in a state of change induced disequilibrium :nniiy lead to a decline in enthusiasm for a new product or ptocess. (7) Newer technologies may replace the old k3fiefore the old is fully adopted. 43 Innovations that are highly compatible with existing practices tend to be adopted more rapidly (Fliegel and Kivlin, 1962, pp. 346-370). A change that requires a complete reversal from the current practice has a lower probability of adoption than one that requires only a minor change. For example, the replacement of local maize with hybrid maize requires a minimum of change on the part of the farmer and therefore hybrid maize usually has a high rate of adoption. However, if the nenv maize has a taste or consistency that is incompatible IViizh established preferences it will very likely have a Slxow rate of adoption, have a high rate of discontinuance, <=iety is western and highly mechanized a new practice (25111 be quite complex with no reduction in the adoption rei1:e, but in a non-western non-mechanical society the it"Inovation could not be very complex before the rate of Eiéloption was reduced. The possibility of trying an innovation on a 531“all scale generally has the effect of improving the Ebc3ssibilities of adoption, but occasionally such is not tzlle case, as Fliegel and Kivlin found no relationship 1‘” on... I a, 1.3- vmma'll. 44 between the two (Fliegel and Kivlin, 1962, pp. 364-370). An important variable in the adoption rate is the obser- vation of good results obtained by neighbors. Seeing others succeed with the innovation apparently is an in- centive to adopt (Lindstrom, 1958, pp. 171-183). There- fore, it is logical to assume that physical objects will be more readily adopted because of their readily observed attributes. A visible physical innovation leaves little doubt as to its use or potential; however, non-material iJInovations, such as a new farm practice, may not have tlLis favorable characteristic (Burnett, 1967, pp. 351-363). An individual's overall perception of an inno- Vation is the sum of each of these categories of inno- Vation characteristics. Total innovation perception IEiIiges on a continuum from highly positive to highly negative. The way an individual perceives an innovation V'i-J_l determine the character of the message he passes by in~terpersona1 channels. It is hypothesized that the more EDCDESitive the composite of innovation characteristics the n“<313e likely an adopter will communicate that message to at Inon-adopter. Thus, the overall character of an eiélOpter's experience with an innovation is directly Ireilated to the type of message he relays via interpersonal (I‘Dmmunication channels. 45 Sources of Information Information sources on innovations can be divided (1) informal, consisting of friends into two groups: formal, consisting of mass media and neighbors; and (2) sources and agents of formal organizations, such as agri- cultural extension services. As would be expected, farmers in the United States and Europe learn more about inno- vations from formal sources than do farmers in developing countries. In Iowa (Smith, 1958, pp. 51-57), for example, .21 percent of the farmers learned of hybrid seed corn from reuiio and farm journals, while 15 percent learned from nesighbors. All formal sources, such as newspapers, agri- Ctthural extension agents, and salesmen, would push this fiisiure towards 85 percent. By way of contrast, in India 7‘5 percent of all farmers used informal sources to learn ‘alDCJut farm innovations while 24 percent used formal sources. Al so, in the Indian case fully 87 percent of the laggards 11$€3dinformal information sources, illuminating the great 'iI“I?ortance of local sources of information for the laggard group (Dasgupta, 1965, pp. 330-337). In a related study of information sources that J‘J-lustrates differential media exposure Belcher found tllleme was no relationship, contrary to expectations, be- tween the acceptance of Salk polio vaccine and socio- c=‘11tural factors considered important in the acceptance <3 15 farm practices. The finding that rural blacks, who I 46 are almost always low on socio-economic measurements, exhibited a high rate of adoption of Salk polio vaccine seemed surprising. In retrospect the reasons are clear as there was widespread publicity about the vaccine using all media sources and ample time for word-of-mouth communi- cation to be fully operative. A controversy arose about the safety and effectiveness of the Salk vaccine, but as blacks did not have a high level of newspaper readership they were less aware of the controversy and thus presented tfliemselves for innoculations at a higher rate (Becher, 1&358, pp. 158-170). Similarly, studies have revealed tliat among physicians information diffused via friend- Sliip networks that allow word-of-mouth communication (CZCDleman, 1957, pp. 253-270). This illustrates that in VTiittually all situations information transmitted by clirect contact is a very important influence on the aciooption rate, even though exposure to mass media sources nnjLght suggest otherwise. It appears that as farmers approach a decision a11>out whether or not to adOpt an innovation they seek out Inc>re authoritative sources such as an agricultural exten- sion agent or a neighbor with a reputation as a good The important consideration is that these more Thus, the 15Eirmer. E1Klthoritative sources are all interpersonal. JLC>cation of the information source in relation to the IEfirmer becomes important. If a perceived authoritative fl I" 47 information source is in close propinquity to the inde- cisive farmer the probability is high that he will decide to adopt. On the other hand, if an authoritative source is not readily available the skeptical farmer probably will not adopt (Mason, 1967, pp. 40-52; Rogers and Beal, 1958, pp. 329-335). Knowledge can be divided into three types: (1) awareness, that an innovation exists; (2) "how-to" knowl- edge; and (3) principles knowledge. "How-to" knowledge Inefers to the information needed to acquire and utilize arl innovation. Principles knowledge deals with the uruierlying principles of the innovation. Awareness J(Knowledge must exist first, and principles knowledge is Ilcrt necessary for successful adoption. That leaves "how- 't<>" knowledge as the most important when the decision to iiCicapt is made. Two generalizations hold as regards "how- t1<>" knowledge. First, the more complex the innovation the rn<>lre "how-to" knowledge required for a person to adopt. S€3<=ond, earlier adopters require less "how-to" knowledge 6“: the time of the adoption decision than later adopters. 1“: the time an innovator acquires an innovation there is v'irtually no "how-to" knowledge available from other than theiss media sources because few other local adoptions exist. leld mass media sources generally will not carry a great Later <1Gial of highly localized "how-to" knowledge. ElSiopters secure most of their information from ffi‘r fl 48 interpersonal sources, which at the time they adopt would possess abundant "how-to" knowledge because of the high level of adoption nearby. Thus, the types of information Sources available at different times in the adoption Process provides a clue as to the degree of "how-to" knowledge required by earlier or later adopters (Rogers, 1971, pp. 106-107) . In summary, a combination of the four basic com- POnents of the diffusion process discussed here, acting in concert, appear to have a major impact on the spatial aspects of diffusion. The interlocking interrelationships are as follows: (1) Areas with a minimum of functional differentiation between spatial sub-units require little horizontal interaction, therefore short distance communi- cation about an innovation results. (2) Earlier adopters utilize more mass media information sources whereas later adopters make more use of interpersonal communication Channels. (3) Earlier adopters will adopt with a mini- mum of "how-to" knowledge, but later adopters require Considerable "how-to" knowledge. (4) Earlier adopters therefore will adopt with very few local information sources while later adopters require a higher level of information from propinquitous sources. The nature of information channels and the type of information required thus effects the spatial diffusion process. (5) The sum total of all innovation characteristics determines the 49 positive or negative nature of the messages communicated about the innovati on . Spatial Diffusion Spatial diffusion seeks to add a spatial dimension to temporal diffusion in an effort to complete the here- tOfore partial picture of the diffusion process. Until Hagerstrand published his initial work on spatial dif- fusion in 1952 (Hagerstrand, 1952) , geographers generally Considered only the particular item being diffused and not the process involved, nor the degree to which the item was found in an area. It was only with the introduction of quantitative techniques and behavioral interests that the rate of diffusion and the degree of intensity of occur- rence could be considered. Hagerstrand argued that changes in the spatial distribution of cultural elements were not random events, but followed certain identifiable patterns. He also insisted that the principles controlling changes Could be discovered. In the search for those principles it becomes necessary to use certain selected surrogates for measurement, but the surrogates are not the subject Of the investigation. The subject is the process of Spatial change and the influencing cultural processes. Thus, geographic research on spatial diffusion has less to do with the selected objects than with the geography 0f cultural behavior (see also Brown and Moore, 1969, pp. 119-157) . 50 Spatial Diffusion Research by Non-Geographers One of the most important forerunners of modern Synatial diffusion research was Pemberton, a sociologist, ‘WTKD attempted to explain the effects of culture contacts of! the rate of culture diffusion within a spatial frame- WOrk. He found that the probability of contacts resulting 111 the transfer of a culture trait are highest near the Center of diffusion (Pemberton, 1938, pp. 246-251). Even earlier, in 1932, Bossard analyzed five t-housand consecutive marriage licenses where one of the Ewartners was a resident of Philadelphia to determine if tihere was a relationship between a person's place of resi- <1ence and the residence of the marriage partner (Bossard, 3L932, pp. 219-224). This study, and other marriage studies 1'-'-hat followed (Davie, 1939, pp. 510-517; Abrams, 1943, pp. 1288—294; Ellsworth, 1948, pp. 444-448; Clarke, 1952, pp. 17-22) showed that at least one-third of the couples lived ‘within five blocks of each other. The two principle factors associated with prOpinquity in marriage within the urban area were distance and social segregation. All Other things being equal a person was most likely to select a person living nearby for marriage. With vari- ations in socio-economic characteristics being less ‘within a neighborhood than between neighborhoods, it was host likely that marriage would involve a person from Cme's own neighborhood. However, there is a slight 51 'teruiency for an increase in marriage partner selection frcnn greater distances simply because of the increase in the numbers of potential partners. Rural sociologists have been concerned with neighborhood norms rather than the effects of distance ir1 the diffusion of farm practices (see, for example, Iaixanberger and Hassinger, 1954, pp. 378-385; Marsh and COleman, 1954, pp. 385-389). The basic assumption is tllat long-term personal associations among farmers will affect their decision making. Farmers tend to be less mObile than urban members of the population so they will tLherefore interact with their neighbors for an extended Period of time; thus mutual expectations and norms develop that influence the behavior of the persons in- ‘fcflved, and the individual does not act independently of 1these norms and expectations. Young and Coleman have Suggested that neighborhoods differ in the degree to ‘Nhich they use advanced agricultural practices, the influence of neighbors and in the use of information Sources (Young and Coleman, 1959, pp. 372-380). It is apparent that the interface between the spatial distri- bution of socio-economic characteristics and the spatial diffusion process has so far not been investigated to any great extent. Group formation is caused and facilitated by dis- tance, because people in the same group or area tend to I31'4l I nil; ‘ " F I. MI Vw . "2’: 'Ovhu .. .9. (n n; . nu ‘- ~ ““3 3.1:: u ."I Q .- ..‘.. ~ I u A~Hg “v.1”; l 71 '- 5“.“ ‘5 ‘U l!.~ .‘h r r (I) ‘5 I 52 have more cultural traits, experiences, values, opinions in.¢:ommon and contacts with each other than with more diJstant individuals. Thus, for persons living in a given arwea.the mechanisms that would tend to define for them a Personal contact field of limited range are as follows: (:1) the probability of unintentional contact declines with <3i.stance, (2) opportunities for intentional contact de- Cfilines with distance, (3) knowledge about contact Oppor- tllnities declines with distance, (4) group and area I'1<:>mogeneity declines with distance, and (5) possibly tinere are norms against distant personal contacts (Young and Coleman, 1959, pp. 372-380). Biologists and medical researchers have been in— ‘Jolved with spatial communications through their research <>n the spread of disease (Rapaport, 1951, pp. 85-92; ILandahl, 1953, pp. 367-383). The prime effort has been ‘to construct a spatially structured predictive model of epidemic spread where the probability of contact depends On the distance between individuals. Every individual is Considered a point in a random network with the proba- bility that there exists a communication route between any given point in the network and any other point a function of both point density and the distance between the points. Using models, it is possible to predict the Spread of a disease on the basis of its virility, point density and the probability of contact between points. 53 TyBes of Spatial Diffusion Geographers have generally considered spatial diffusion from two VieWpoints. First, differences in cultural traits from place to place and the generalized movements that have caused a particular distribution (Stanislawski, 1946, pp. 105-140; Kniffen, 1946, pp. 549-577; Deshler, 1965, p. 612). Second, the processes involved in the movement are investigated, with the item diffused being relatively unimportant. Later, expansion diffusion will be discussed in C1etail, but first the other types of spatial diffusion Will be briefly introduced. Relocation Diffusion.--The most obvious example of relocation diffusion is migration, where the item being diffused actually transfers location (see, for eJ'cample, Demko and Cassetti, 1970, pp. 533-539; Carol, 1971, pp. 369-373; Villeneuve, 1970, pp. 369-375). Mi- gration research concentrates primarily on the source regions, destination regions, migration routes, incentives to migration, communication channels between source and destination regions, and the characteristics of persons migrating. Diffusion of this type does not lead to an increase in the number of items being diffused, but Il'levement to a new set of locations. Hierarchical Diffusion.--Bowers in 1937 described the movement of inventions from city to city. Even 54 though his approach was descriptive rather than analytical, he was in fact dealing with diffusion downward through a Genitral place hierarchy. In contrast to other types of Ssnatial diffusion, straight line distance is not always true most important consideration in hierarchical dif- fllsion. Instead the most important variable is the move- “Hant downward from the highest level in the central place hHierarchy to the second highest level, and so on to the llvwest level (Hagerstrand, 1966, pp. 27-43; Hudson, 1969, Fflp. 45-58; Brown, 1969, pp. 189-211). Most hierarchical Cliffusion research has been confined to one country, but 1Pederson gave his research an international sc0pe by Einalyzing the movement of innovations between capital or :leading cities of separate nations and between cities \Vithin the nations (Pederson, 1970, pp. 203-254). Linear Diffusion.--In certain instances the item lbeing spread by the expansion diffusion process is con- fined to a linear corridor, such as a highway, and cannot exPand at right angles from that corridor for any appreciable distance (Colenutt, 1969, pp. 106-114). Linear diffusion can be a combination of relocation and expansion diffusion in a highly confined environment, for some items will actually move to a different location on the corridor and in other cases there is an expansion in the number of items along the corridor, as shown in Figure 2. 55 Figure 2 LINEAR DIFFUSION Tl E o o o 0 j EXPANSION T2 1;. o o o o o o o o o o o 0:] T. L o o o 0 j RELOCATION T2 L o o o o o ] B§sic Concepts of Expansion Diffusion Also referred to as contagious spread because of fists similarity to disease diffusion, this type is the most tlhoroughly investigated. Hagerstrand was the first re- ESearcher to systematically investigate the way inno- ‘Iations spread across the landscape (Hagerstrand, 1953). The diffusion pattern is generated principally by the Spread of information from one person to another by interpersonal or word-of-mouth communication channels. It is assumed that a person will adOpt the innovation after learning about it, with the exact time of adOption being a function of his resistance to change. Due to the limited spatial range of personal contacts, caused by the friction of distance, a person adopting an innovation will tend to be located close to a previous adopter. Therefore, expansion diffusion causes the innovation to move gradually across the landscape, modified by differ- ences in population density, ecological conditions, his) 0 'Ip p ...C '4 U . 'UA. .."-1 :.:~E "Iu I ":u: " ‘-¢.‘ 5h ' . (I, p; H‘E 56 receptivity to innovation, infrastructural variations, and political, ethnic, linguistic and physical barriers. Hagerstrand was able to identify four stages in true expansion diffusion process in southern Sweden (Phagerstrand, 1952). The four stages described the n the diffusion nodes and as new diffusion nodes appear. rI'he differences between the adopting and non-adOpting ameas found in the first stage are reduced as adoption llas taken place to at least a limited intensity in nearly all areas. Stage III, the Condensing Stage, the inno- vation has spread to the entire area and all potential adopters know about it. The rate of intensification is about equal in all areas. Stage IV, the Saturation Stage, is a slow but general increase in the percent of the pOpulation that has adopted the innovation, moving toward the maximum possible. The spatial diffusion pattern found in southern Sweden is exemplified by an outward spread at low 57 intensity of adoption, followed by an increase in intensity up to the maximum possible percent of adoption. Assuming that the spatial pattern of innovation diffusion is largely determined by the spatial arrangement of communication channels and modified by the physical and cultural en- vironment, it is logical to assume that differences in these factors will produce a different spatial diffusion pattern. The analytical portion of this dissertation will investigate that proposition. The work of Pyle points up some of the inter- connections between the contagion diffusion spread and diffusion within an urban hierarchy. Because of a virtu- ally nonexistent transportation system and no urban hierarchy, the first United States cholera epidemic of 1832 spread across the country in the contagious fashion. However, by 1849 when the second epidemic struck, the transport system and the urban hierarchy system were both better developed, thus the spread was different. First, the disease filtered down the United States urban hier- archy, beginning with the largest cities first, and later, the smaller cities adjacent to the larger cities contacted the disease after the primary centers. The 1866 epidemic moved even more clearly down the urban hierarchy (see also, Bowden, 1965; Brown, L. A., 1967, p. 783; Redlich, 1953, pp. 301-322; Witthuhn, 1968, pp. 5~20). 58 Personal Information Field Even though early sociological researchers dealt with marriage partner selection they were in fact using the concept of the personal information field, that is, the probability of contact with another person will de- cline with increasing distance (Gould, 1969). Later, others began to deal more directly with the phenomenon of distance (Boalt, 1957, pp. 73-97; Miller, 1947, pp. 276-284). They have defined the probability of contact as that of a "J" shaped curve. The probability of select- ing a person for contact declines with distance from the chooser, while the distance decay gradient varies with personal characteristics, the communication and trans- portation system available, and the density of population. The curve was described as having a "J" shape because the probability of contact increases at long distances due to the greater numbers of potential contacts resulting from the expanded area. In places with high population den— sity the distance decay gradient is steeper as the number of propinquitous potential contacts increases. Geographers have taken the "J" curve concept and made it omnidirectional as opposed to unidirectional. The concept has been set into a stochastic framework and labeled either the Mean Information Field (MIF) or the Personal Information Field (PIF). Mean Information Field refers to the mean probability of contact as distance 59 increases and does not allow for different probabilities of contact according to different personal characteris- tics. The Personal Information Field is more versatile because it allows for more cosmopolite persons to have a wider contact field than more localite individuals. The Personal Information Field graphically depicts (see Figure 3) the declining probability of communication, in all directions, as distance increases. The magnitude of the distance-decay function for any given area can be determined by empirical investigation (see, for example, Hagerstrand, 1953, pp. 165-241; Marble and Nystuen, 1963, pp. 99-109; Morrill, 1963, pp. 75-84; Morrill and Pitts, 1967, pp. 401-422; Warntz, 1966, pp. 47-64). Figure 3 PERSONAL INFORMATION FIELD PROBABILITY OF PASSING A MESSAGE ‘——- DISTANCE ——-> 60 Innovation Waves Morrill, while recognizing that the use of the wave analogy to describe human spatial phenomena is less than completely accurate, considers it an effective tool in the interpretation of the diffusion process (Morrill, 1968, pp. 1-18). The basic idea is that an innovation is introduced in some location by either chance or design, after which the diffusion process causes an outward move- ment in a wave-like pattern such that over a series of time periods places farther away will begin to adOpt the innovation at the same time close-in places are increas- ing the rate of use. Depending on a variety of circum- stances the impulse of the wave will diminish with in- creasing distance, and will eventually disappear entirely (Hagerstrand, 1952). Figure 4 illustrates how the waves decrease in intensity over increasing distance and increase in intensity with time. Figure 4 INNOVATION WAVES I O PERCENT ADOPTION DISTANCE fl..- cus ‘ (Y‘ 'FA . I "vU‘ 5. use, “VF! U.~~ - in l.“ I U (I) “vs. Wu ‘A . ~ ‘~ I :T‘m 61 The structure of the innovation waves suggested by Morrill are based on two assumptions. First, a Mean Information Field as Opposed to a Personal Information Field is assumed so that no differences in the areal extent of personal contacts are allowed for. The Morrill model of innovation waves is based on every person having the same distance-decay function of personal contact probabilities. Second, this model assumes that all indi- viduals require the same type and amount of information before deciding to adopt. The combination of these two assumptions will produce an innovation wave that moves outward and upward (as in Figure 4) at a uniform rate. Morrill's implicit assumptions will now be modi- fied. First, Personal Information Fields are used to provide for a wider range of contacts for earlier adopters and a more narrow range for later ad0pters. Second, differential receptivity is introduced such that earlier adOpters will accept an innovation with fewer tellings than later adopters. These assumptions will produce innovation waves characterized by non-uniform outward and upward movement (see Figure 5). The new wave will tend to move outward rapidly at low levels of adOption as earlier adopters throughout the entire area rapidly acquire the innovation. Then a localized intensification Will occur at the locations of original introduction as the later adopters accept in those places where a 62 sufficient number of tellings are generated. Additional tellings are required because later adopters require more interpersonal communication and more "how-to" knowl- edge than earlier adopters. The conditions required by later adopters are only met in locals with a high inten- sity of adoption. Finally, the more intensive levels of adoption will move outward from the original areas of introduction. At this time the gradient of the forward edge of the diffusion wave will be rather steep. Figure 5 INNOVATION WAVES II Z 5 <2 I; E O 8 2 4 I— I— Z 2 DJ DJ 0 U (I (I UJ LIJ CL CL 0 DISTANCE 0 DISTANCE From Morrill Generalized from Kisii data Ihe "S" Shaped Growth Curve The "S" shaped growth curve can be seen in the Context of both time and space. In the temporal sense the original phase of the curve is associated with earlier adopters, the middle phase with the majority of adOpters, and the final phase with the later adopters. 63 When an innovation is initially introduced into an area the first adOpter will tell a few non-adopters about the innovation, they will tell a few others, and so on, with the result being exponential growth. When half of the potential adOpters have adquired the innovation, the "S" shaped curve begins to level off because each new adopter finds it increasingly more difficult to find a non-adopter to tell about the innovation (Rogers, 1971, pp. 176-179). In the spatial context the first phase of the growth curve is represented by a few scattered adOptions, the second phase is a period of rapid adOption and the third phase is a slow and gradual increase leading to the maximum possible level of adoption (Brown and Cox, 1971, pp. 551-559; Dodd, 1955, pp. 392-401; Hagerstrand, 1966, pp. 27-43). Figure 6 illustrates the "S" shaped growth curve in time and space. The curves are reversed here to allow both the time and distance zero point to be in the same location. The percent of the pOpulation who have adOpted, increases from the time of introduction. Two things happen simultaneously in the spatial sense: the percent Of persons adopting at the original point of introduction increases and the percent of adoption decreases as dis- tance from the innovation center increases. 64 Figure 6 "3" wave In TIME AND SPACE a \s E x .— O. O O < *5 Li; 045‘» 4w 6 Ce Q 0 _ TIME OR DISTANCE If the "S" shaped growth curve plotted over time represents earlier adopters in the first phase and later adOpters in the final phase, then it should be logical to assume that when plotted over distance it represents a similar distribution of adopter categories. Therefore, at a long distance from the point of introduction only the earliest of adOpters have acquired the innovation while at the point of introduction the latest of adOpters have it. An anamolous situation is revealed here. Persons with the personal characteristics of early adOpters, located far from the innovation center, will adOpt the innovation at the same time as individuals near the innovation center but with the personal characteristics of late adopters. The seeming anamoly can be resolved by dividing the area into numerous sub-units. Each sub- unit, if there is a normal distribution of adOpter 65 characteristics, will tend to Operate as a separate entity, such that the movement of a diffusion wave into the sub- unit acts much like the original introduction. So the determining factor is the relationship between the location of the innovation wave and the location of the sub-unit. As regards the spatial distribution of adopters, Cassetti has suggested the following postulates: (1) persons adopt when they are brought under the influence of previous adopters in the course of direct personal contacts, (2) potential adopters exhibit different levels of resistance to adoption, and (3) resistance to change breaks down only when there are a sufficient number of messages about the innovation (Cassetti, 1969, pp. 101- 105). These postulates would suggest that those adopters at the outer edge of a diffusion wave are innovators and early adopters (i.e., persons fitting the socio-economic characteristics associated with innovators and early adOpters), and those in the second phase of the growth curve are early and late majority adopters while those in the third phase are the laggards. Simulation of Diffusion Computer simulation models, both aspatial and spatial, have generally been of the Monte Carlo type, that is, they are constructed within a probabilistic framework and powered by a random numbers table. The 66 aspatial models attempt to simulate the cumulative percent of adoption of the innovation within a population, while the spatial models try to simulate the spatial distri- bution of adOpters for a series of time periods, called generations. Both types of models have a certain heuris- tic value in that the very construction of the model forces the designer to consider the magnitude and direction of a variety of influences affecting the diffusion process. An example of an aspatial computer simulation model is SINDI 1 and 2 (for Simulation of Ignovation Diffusion) developed by Carroll and utilized by Hanneman (Carroll, 1969; Hanneman, 1969, pp. 36-45; Hanneman, 1969). These models incorporate a number of non-spatial con- straints and influences such as communication within and between cliques in a social system, social distance, extension agent influence, cosmopolite influence, the impact of print and electronic mass media information flows, word-of-mouth message transfers, community organi- zation meetings, individual resistance to adOption, and the cumulative influence of repeated messages. As would be expected, considering the number of constraints in- corporated into the models, they can simulate the real world rather accurately. Spatial models attempt to simulate both the cumulative rate of adOption and the distribution of 67 adopters over space. The Hagerstrand model II (Hager- strand, 1953, p. 246) is based on four assumptions: (I) initially only one person in the population is informed about the innovation, (2) as soon as a person hears about the innovation he accepts, (3) information can only be received by a pair-wise telling, and (4) information is passed on only once per time interval (generation) to another person (see also, Gould, 1969, pp. 28-38; Hager- strand, 1965, pp. 43-67; Anderson, 1970, pp. 9-14). Obviously such assumptions are unrealistic, but it does have the advantage of simplification, and later, if the assumptions appear to be inadequate they can be modified. Modifications of the original Hagerstrand models have been made by Hagerstrand and by Pitts (Pitts, 1963, pp. 111-122). These models include new constraints and controls that make the model more sophisticaled and flexi- ble. The most important of the variations is the inclusion of a means for deciding if there is a barrier between the teller and receiver, and if so, to what degree (Yuill, 1964; Misra, 1966, pp. 149-155). Another variation is the inclusion of a psychological resistance factor based on adopter categories. Those highly resistant to the innovation will have to be told a number of times before adopting. Most empirical studies that have used simulation techniques have investigated agricultural innovations. 68 Tiedeman and Van Doren (Tiedeman and Van Doren, 1965) studied the spatial diffusion of hybrid seed corn in Iowa and performed several simulation runs, but could not com- pare their results with reality as they did not have information on the names and addresses of the actual adopters in the study area (Cassetti and Semple, 1969, pp. 254-259). Bowden and Ramachandran examined the spatial diffusion of irrigation wells in eastern Colorado and southern India, respectively, and were able to simulate the process rather well (Bowden, 1965; Ramachandran, 1969). In fact Bowden's simulation was remarkably similar to the real world in terms of overall patterns, intensities, and distributions. De Temple simulated the diffusion of Harvestore Systems in northeastern Iowa, utilizing contact fields and a central place hierarchy, while Johansen simulated the diffusion of strip crOpping in Wisconsin (De Temple, 1970; Johansen, 1971, pp. 671- 683). It can be easily seen that the simple model used by Hagerstrand in 1952 has evolved into a series of highly complex spatial simulation models. An Integration This section will attempt to integrate some of the basic elements of spatial (expansion) diffusion with those of temporal diffusion. The model developed is general with no specific parameters, so it is explanatory II .Au 7. he FI‘ in .L‘ 1'! s . I. M- V“)- .“‘ I‘V i‘: ‘I. 69 rather than predictive. It attempts to give an indication of the probability of adoption for a person residing in one cell in a lattice of "n" cells. There are four basic elements in the proposed model: (1) The characteristics of the adopter. The probability of adOption in any time period is minimum at the laggard end of the continuum and maximum at the innovator end. (2) Combined innovation characteristics generate a level of desirability that ranges from a maximum negative for an undesirable inno- vation through neutral to a maximum positive for a de- sirable innovation. The five characteristics of relative advantage, compatability, complexity, trialibility and observability combine to form the overall characteristics of the innovation. (3) The intensity of adoption in any lattice cell refers to the percent of potential adopters who have acquired the innovation in the time period in question. (4) The type of interpersonal knowledge generally available varies from awareness knowledge to specific "how-to" knowledge and is positively related to the intensity of adOption. Positive innovation characteristics will cause favorable messages to be passed that will facilitate adoption. Since the messages will be passed only a short distance, due to the nature of spatial communication as exemplified by the Personal Information Field, adopters will tend to become concentrated in specific locations. 70 Increased ad0ption intensity will produce a Regional Information Field, that is defined as the probability of an individual receiving a positive (i.e., favorable) message about an innovation that contains specific "how- to" information. Around each adopter is a Personal Information Field that indicates the declining probability of the adopter passing a message to another person as distance increases. Theoretically every person in the world is a potential contact for the adopter, but after a short distance the probability of contact approaches zero. As the number of adOpters in a confined area in- creases the combined probabilities of contact caused by numerous overlapping Personal Information Fields virtually guarantees that the non-adopter receives a message about the innovation. The number of messages successfully passed from an adOpter to a non-adOpter may decrease with increased adOption intensity because non-adopters become increasingly scarce, but the probability of a telling remains high. However, high contact probability may not be sufficient for the later adOpters require more than one message due to their resistance to inno- vation. Resistance, more than unsuccessful tellings is the probable cause of the upper inflection of the "S" shaped growth curve. Another element enters at this juncture. A higher percent of use in an area will insure a high level of 71 specific knowledge. Non-adopters can easily observe the innovation and have ample opportunity to discuss its use with a person using it. In this way "how-to" knowledge becomes common to all persons in an area. Thus, the three elements (innovation characteristics, adoption intensity and information type) acting together produce an intensi- fied Regional Information Field which in turn increases the probability of adoption. AdOpter characteristics will enhance the rate of adOption in the early stages and re- tard it in the later stages. To summarize the model, assume that an inno- vation with desirable characteristics is introduced at a specific location. Those positive features will insure adOption by innovative people, and thus increase the intensity of adOption from zero to a positive value. Slowly the local level of information about the inno- vation intensifies and changes from general to specific. The increase in "how-to" knowledge in turn causes more adOption, forming a positive feedback loop. A positive feedback loop is defined as a return of the effects of a process (adOption) to its source (here, in the form of a more intense Regional Information Field) so as to reinforce or modify the prior condition (produce more adoption) (Carroll, 1968, p. 3; Meadows, 1972, pp. 31-33). The Regional Information Field could be referred to as spatial variation in the diffusion effect (see Rogers, 72 1971, pp. 161-164) or a modified version of the neighbor- hood effect (Hagerstrand, 1953, pp. 158-163). Assume that each lattice cell contains approxi- mately the same distribution of ideal adOpter types. Since innovators are willing to adOpt with less specific "how-to" knowledge, then one could assume that these indi- viduals would adOpt as soon as practicable. A pattern develOps where a few adopters are found over a wide area. Later the intensification of adOption in certain lattice cells and the concomitant intensification of the Regional Information Field results. The reasons for the initial intensification in a given cell may be spatial variations in the advantages of the innovation, specific efforts on the part of a change agency, or random variation. The end result, however, appears to be a varying level of information from one cell to another caused by the generation of a more intense Regional Information Field in that area, resulting in variation in the spatial diffusion of adoption. CHAPTER III KISII DISTRICT: BIOPHYSICAL AND SOCIO-ECONOMIC BACKGROUND TO THE DIFFUSION PROCESS Introduction In southwestern Kenya, about 400 kilometers (250 miles) west of Nairobi and 50 kilometers (30 miles) east of Lake Victoria is Kisii District (see Figure 7), home of the Gusii. The political unit of Kisii District, which does not correspond perfectly with the physical unit of the Kisii highlands, is 2217 square kilometers (856 square miles) and in 1969 had a pOpulation of 675,000. Included is the former EurOpean settlement area, now Borabu Location, of 653 square kilometers (252 square miles) (Kenya, 1970, p. 39). The economy is basically small-holder cash crOp and subsistence agriculture. On the small farms are grown a variety of food crOps, both for home consumption and sale at the 75 local markets, and cash crOps, such as coffee, tea, pyrethrum, and passion fruit. These crOps are the mainstay of the Kisii economy for there is little else besides agriculture to bring in money from outside the district. 73 74 6' X Od a?! 29.203 5.55 :3. h masons ......... \.:::./.I . . , m \J W a _ _ u .. .w. man. je _ .. z —— .— \\l< ’- . bun” / ”m \\I/ // J” I“ 4 “WWW: ”WW”. / u a I, II, 0 EH. ............................ i/ If U 887/ v r /.\\ IIIII) l../ / \\ .N . ,f U / , ., m . 5.1.55 J. x/ // ,. Ads. 0.: rm // / Box—<2 J :3. “RV J t e a . . . z / s I It onP .—/ /, // \\ .r. III . I m x // \ o 9.9.32 I; i ShQO . u // // A\\I\II\\-\anl/ Aafn\\\ .. (\\ 5 ... .h..r....,.. . .. .. .. . I J \ C . M a. .. .u, u .I/ I / J/ \\ C We m .......... .. . . ... . .. . . . .. b\\Q\Q\L .. /ooo«/ /v \\ k M tam 9353.29... .. . . . M b \\ 3 a . .... . .. . . * IV ”baxcuz h (VU \II\ \I (I \. up. . . . . . . . .... .3.“ . . . H ’II‘ . .n. . . ., w ’ul\.ll) I I . , cu . v I \‘l" I a . 0 , .(fisxp/D /. .::aanx n .n / \ \v . u. z a \ . \I/xx no \).(\I|IIIJ n ................... \. ...\\.. ‘1 ..§0 fl’\\\llllll ....... . I. . .. , ....... oo / n/ (\IIIK . In t cannon ' Jill/l . \.|\\ x . . \ (I) (D (H ‘Fw -Ubt v'p. (.3 '1 p., 'l-. -1. r '(7 In I); If 1 I" 75 The district ranges in elevation from about 1525 meters (5,000 feet) to over 2135 meters (7,000 feet) above sea level. The resultant cool temperatures, plus abundant rainfall, give the region an appearance that does not con- form to the stereotype of East Africa (see Figures 8 and 9). The verdant green landscape is more reminiscent of what one would expect in Europe or parts of Asia. Popu- lation pressure has necessitated the utilization of nearly all suitable land for agriculture. Thus one will find very little of the original forest covering remaining. What is left is confined to small woodlots, watercourses, or along roads. Farms average about six acres in size, with the individual fields separated by hedgerows. The highlands are a dissected uplifted peneplain which now has steep sided valleys with broad flat bottoms that are frequently swampy. Physical Geography The physical unit of the Kisii Highlands (see Figure 10) occupies some 3250 square kilometers (1250 square miles) of territory and can be divided into two parts. The eastern two-thirds is a deeply dissected Cretaceous peneplain with steep ridges up to 2160 meters (7,200 feet) separated by deep, often flat-bottomed valleys that are occasionally chocked with swamp grass and papyrus. In this section the main ridges show a westward leping summit level grading from the previously 76 Figure 8 Kisii Farmland In the left foreground is tea and in the right foreground is a passion fruit vine. The light- colored fields in the middleground are planted to pyrethrum. In the background is flat-topped Itumbe Hill, 2,075 meters (6,800 feet elevation, a remnant of the Cretaceous dissected peneplain. 77 Figure 9 Kisii Hillside Located near Keroka, this area has over 575 persons per square kilometer (1,500 per square mile). Farms seldom have over one piece of land and extend from the bottom of the hill to the top. Hybrid maize and tea are the main crops. 78 Figure 10 KISII DISTRICT PHYSIOGRAPHY 7000 \ (2'35M)| \ I 0"" r L .1- (:33. fl I m <5“? 11 bf .4”’ .a*”’ .ac’T’ \ \ 3. . I: ’1' ,. ///2 H ' L \ f I “3%" ---~ ContoursUOOO Intervals) ...\_/,-' -’ --— Rivers --— Kisii District boundary 9 - “2' . £9 SOURCE: SURVEY OF KENYA, "250,000 TOPOGRAPHICU‘ISUMU). I969. ,L 79 mentioned maximum of 2160 meters (7,200 feet) in the east to 1950 meters (6,400 feet) in the west. These ridge t0ps probably represent the remains of an ancient tilted peneplain. The western one-third is lower and more gently undulating country of a sub-Miocene peneplain (Kenya, 1952, pp. 4-5). There are three primary rock systems that make up the Kisii highlands. They are the Nyanzan System, the Kavirondian System and the Bukoban System. The rocks of the Pre-Cambrian Nyanzan System occur mainly in the west- ern portion of the area. Acid volcanic lava material makes up by far the largest portion of the system, along with a few unweathered outcrOps of basalt. The Pre- Cambrian Kavirondian System, made up of massive boulder conglomerates with subordinate sedimentary materials is well developed in several localities in the west-central part of the highlands. Their angular unconformity with the underlying Nyanzan rocks is manifest by strong vari- ations of dip and strike. The thickness of the Kavirond- ian rocks is unknown but it appears to be greater in the north, where it is estimated to be 1500 meters (5,000 feet) or more thick. In the south the thickness is not more than 1050 meters (3,500 feet). Most of the district is overlain by the Bukoban System, Kisii Series. Some 1950 square kilometers (about 750 square miles) of the eastern portion of the highlands 80 is built up of the Kisii Series, a variation of the Bukoban System found in Uganda and Tanzania, which con- sists of almost flat-lying basalts, quartzites and lavas. Pleistocene deposits are all of a superficial nature and nowhere do they attain any great thickness. The composition of the material is sparse terrace gravels, lateritic ironstone cappings, semi-consolidated river alluvium, quartz rubble, recent river alluvium and swamp deposits. Many of the valleys of the Kisii highlands are extremely steep-sided but have wide, flat-bottoms filled with decomposed vegetation and silt. The swamp vegetation effectively reduces the speed of the streams and thus assists in arresting erosion (Kenya, 1952, pp. 6-36). A general east to west drainage pattern of the Kisii highlands is controlled by the westward slope of the ancient tilted peneplain into which the river system is embedded. Most rivers are still actively down-cutting as evidenced, for example, by the rapids that appear in several places on the Kuja River. There is also active cutting back in the headwater areas of most streams. Most streams in the northern part of the district, when debouching from the highlands, swing to a general north- westerly direction towards the Kavirondo Gulf. After leaving the highlands the Kuja River turns south-westward toward Lake Victoria and meanders sluggishly over the gently sloping plains of Luo country. 81 The drainage of the Kisii highlands is accom- plished by three principal river systems: the Sondu River, the Mogusi River, the Kuja (Gucha) River, and their respective tributaries. The drainage of the whole high- land area eventually flows into Lake Victoria. The Sondu River, and one of its principal tributaries, the Kibsonoi River, drain much of extreme northeastern and eastern Kisii District. The Kibsonoi is primarily outside the district but its small tributaries reach into the district. The Mogusii, the smallest of the three rivers, and a tribu- tary of the Awach Tende River, drains approximately 20 percent of the northwest portion of the district. Most of central and southern Kisii District is drained by the Kuja River and its minor tributaries; while the extreme southern edge of the district is tapped by a major tribu- tary of the Kuja, the Migori River (Kenya, 1952, pp. 4-5). Climate Temperature in an equatorial region such as Kisii District is largely a matter of elevation, whether it is the wet or dry season, and the time of day. The mean day- time temperature in Kisii town is 28.9 degrees C. (84 degrees F.) while the mean night temperature is 12.8 degrees C. (55 degrees F.). The highest temperatures of the year occur during the months of January, February and March; the latter part of the dry season before the long rains. No temperature records are available for any 82 place in the district outside of Kisii town, and even there the records are scanty. But due to elevation differences it can be assumed that Keroka, at 2040 meters (6,800 feet) above sea level, will average about 2.8 degrees C. (5 degrees F.) cooler (Kenya, 1970b, map). Precipitation is distributed throughout the year so that no month receives less than 6.4 cm. (2.5 inches). As the rainfall diagram for the Kisii Seed Farm, located at the edge of Kisii town, indicates, there are two distinct wet periods (Figure 11). March, April and May comprise the so-called long rains. July forms a reason- ably well-defined break before the short rains of August and September. Following that period is a rather steady dimunition of rainfall from October to January. The amounts of rain falling in a given month will vary within the district, but the pattern of seasonality is everywhere the same. The distribution of precipitation (Figure 12) within the district corresponds very roughly with the occurrence of the highest elevations. The highest rain- fall stations are found at about 2000 meters (6,000 feet), at the western edge of the highest section of the district. This is due, in part, to the moisture-laden winds coming from Lake Victoria. Most of the central portion of the district receives in excess of 160 cm. (70 inches) annually. Only the extreme northern and southern portions 83 Figure 11 ANNUAL PRECIPITATION KISII SEED FARM (ANNUAL RAINFALL 73 inches) _l_l.70 nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn ............................................................................................... ---------------------------------------------------------- ----------------------------------------------------------------------------------------------- ooooooooooooooooooo oooooooooooooooooooo ooooooooooooooooooo uuuuuuuuuuuuuuuuuuuu nnnnnnnnnnnnnnnnnnn 300— 250_ 200_ I50— I00_ 59. o A SURVEY OF KENYA, KENYA CLIMATE AND VEGETATION MAP, I970 SOURCE 84 Figure 12 KISII DISTRICT ANNUAL PRECIPITATION —— IsohyeI —-— District Boundary Location Boundary 9 . “.m- . 29 SOURCE; SURVEY OF KENYA, KENYA CLIMATE AND VESETATION MAP, I970 85 of the district receive less than 150 cm. (60 inches) annually. The predominant influence on the seasonal distri- bution of rainfall in Kisii District are the two main wind systems; the northeast and the southeast trades. The passage of the intertropical front over the district coincides with the wet seasons. Kisii District has an added advantage in being close to Lake Victoria, for when the prevailing wind is from the southwest it has the opportunity to pick up moisture from the lake which is subsequently received by the highlands (Hichman, 1960, pp. 6-7). Vegetation and Soils Due to population pressure and the resultant in- tense cultivation Kisii long ago lost its original natural vegetation. Nyangweta Forest, in the southern part of South Mugirango Location, indicates that the original vegetative cover was a moist montane forest. Black wattle, along with some cypress and eucalyptus, can be found along watercourses, roads, and in scattered wood- lots. The 2000 meter (6,000 feet) contour is the rough dividing line between the highland Kikuyu grass zone and the lowland Star grass zone. Both of these grasses are indicative of good soils, adequate rainfall and moderate temperatures. Of the two zones, the Kikuyu grass area offers the greater potential; hot because of the grass, 86 but due to the coincidence with the better soils and generally heavier rainfall. The Star grass zone, due to its lower elevation, is marginal for pyrethrum and tea but completely adequate for coffee, maize and bananas (Uchendu, 1969, pp. 7-8). Unfortunately, no detailed soil survey has been undertaken in Kisii District, so only gross generali- zations can be made. Three soil types have been identi- fied, each divided by one of the major ridge formations of the district. First are the Kisii Red Loams that are underlain by red loam sub-soils of inferior quality, which occupy most of Borabu Location and parts of the rest of the district slightly to the west of the ridge that generally runs north from Keroka towards Sondu, just north of the northern edge of Kisii District. The second type is the Kisii Highlands Loam, found over much of the district. Two distinct sub-types, dependent on lepe, can be distinguished. In the flat valley bottoms the soil is reddish brown to red and is quite deep due to deposition. On the hillsides and hilltOps the soil is much more shallow and stoney, with an occasional rock outcrOp interspersed. The third soil type, found largely in the lower western part of the district, is the Kisii Savannah Loam. A very rough location guide would be west of the main road entering Kisii from the north, going through Kisii town and then south through Ogembo to 87 Nayngusu. This soil is usually shallow and variable in color from reddish-brown to gray. Rock outcrOps are common, erosion can be severe, and runoff is quick. The best agriculture is confined to pockets of deep soil on the valley bottoms (Uchendu, 1969, pp. 5-6). There is no relationship between the land-use and land-ownership patterns among the different soil types. POpulation density, however, is related to rainfall and the former settlement area. The Population of Kisii District By far the most striking characteristic of Kisii District is the number of peOple residing there. Within Kenya this district has the second highest level of popu- lation density and is growing considerably faster than the country as a whole. The following table gives the estimated annual growth rates from 1900 to 1948 and the calculated growth rates from 1948 to 1969. TABLE 3.--Kisii District, POpulation Growth Rates. 1900-1930 0.5 percent annually (estimate) 1930-1940 1.2 " " " 1940-1948 3.7 " " " 1948-1962 5.7 " " (author's calculations based on census data) 1962-1969 3.6 " " " " Source: Uchendu, 1969, p. 12; Kenya, 1966, p. 20; and Kenya, 1970, p. 39. 88 The estimated annual growth rates from 1900 to 1948 seem completely plausible, but the rate of 5.7 per- cent annually is open to question. If one can assume that the 1962 census is the more accurate, then it appears that the 1948 census undercounted the Gusii people. The more current growth rate of 3.6 percent annually makes Kisii District one of the fastest growing places in Kenya, if not in Africa. Such a growth rate will cause the pOpu- lation to double in twenty years. But growth rates alone do not present the total picture. Another facet of the population mosaic is age structure. Two main factors emerge; one is the size of the dependent population and the other is an indication of future growth. Table 4 gives a breakdown of the per- cent of the population in each age group. Fully 55.2 percent of the peOple are age 14 or under. The dependent population, that is, the percent of the population under 15 years and over 60 years old, is 58.7 percent. Thus the 41.7 percent in the productive years must support the remainder of the population. And not all of the people in the so-called productive years are contributing to output. The size of the population under age 15 indicates a possible increase in the overall growth rate in the near future as these children reach reproductive age. 89 TABLE 4.--Kisii District, Percent of Population by Age Group. Cumulative Age Male Female (Male and Female) 0- 4 21.4 20.7 21.0 5- 9 18.9 18.5 39.7 10-14 16.0 15.0 55.2 15-19 11.1 10.5 66.0 20-24 7.2 7.5 73.4 25-29 5.3 6.7 79.4 30-39 7.9 9.2 88.0 40-49 5.2 5.0 93.1 50-59 3.1 3.1 96.2 60 & over 3.7 3.5 99.8* *Does not total 100% due to rounding. Source: Kenya, 1970, p. 121. As one would expect, the distribution of popu- lation within Kisii District is far from uniform. In 1969 the mean density for the whole district was 304 per square kilometer (788 per square mile). The range was from 57 per square kilometer (148 per square mile) in Matutu Sub-location of Borabu Location (the former Settle- ment Area) to 557 per square kilometer (1442 per square mile) in Mwogeto Sub-location of Kitutu Central Location. Table 5 shows the distribution of pOpulation by locations for the 1948, 1962 and 1969 censuses. In 1961 Kisii became a district by itself and most of the locations were divided into two or three sections. Borabu Location was added after the 1962 census. Figure 13 shows the distri- bution of population density by sub-location in 1969. 9O .asrmm .aa .omma .mmamm “paa .a .mpma .spamaop Kmm .a .mmma .mmcmx "mopsom pma.m m omo.m omm.s mom.a c309 aamax mm mmm 0mm.a~ apnoea Nmm mam moa.om amv.ms mam.a~ ammpm cam aom omm.mv msm.pm pmo.ma omcmpamsz ausom omm maa mm~.pm sampom mmommz mom aoa mmm.am mmm.mm mpm.m~ maomao mmommz mmm «ma pae.am omm.- amm.ma mummamz mmm oma mmm.om amm.mm «pumps apmpaummz amm oma mma.mm aam.sm apm.ms anomao aampaummz mom mma mmm.om mmm.mp omcmaamsz amoz mmm mom mmm.mm mmm.om sam.am omcmuamaz.aaaoz mas mma amo.mm maa.mp saunas ummm mms pma mom.ap mmm.mp amm.ap sundae amauamo Nam ama omm.a¢ mma.am saunas umoz mc0auwooa mom mama aso.mmp maa.mam oom.mmm uoauumaa aamax as as spammoo mpma .mmpa mpma mama mama .mpmarmmmarmmma .coaumasaom .uoauumao aamamru.m mamas 91 Figure 13 MS” DISTRICT SQUARE KILOMETER Below 200 ZOI - 300 BOI - 4oo PERSONS PER Y T ...u”.u.. .. S ...... m. N ...... E D n N .......... m .............. w J T A m L H U m i P K O Q P oi KENYA POPULATION census. VOL. I, I969. SOURCE: 92 Kisii District is a remarkably homogeneous area ethnically. Kenya Africans account for 99.98 percent of the total population and the Gusii peOple are 97.98 per- cent of the total. The largest non-Gusii groups of Kenya Africans are, in descending order of size: Luo, Kikuyu, Kipsigis, Luhya, Kamba, Kuria, Masai, Nandi, Mijikenda and Meru. Even the Luo are only 0.8 percent of the total population. Europeans made up 0.05 percent of the total and Asians account for 0.1 percent. The result is that the Gusii are the overwhelming majority in their own district, but they are the exact opposite elsewhere. Approximately 5.7 percent of the Gusii people reside outside their home district. They are found in limited numbers in every district in Kenya. The abrupt change in the ethnic composition of the population at the borders of the district is a remarkable feature of the cultural geography of western Kenya (Kenya, 1969, p. 96). The Gusii The Gusii of Kisii District are a Bantu-speaking people surrounded by the non-Bantu Luo, Masai, and Kipsigis. They are related to the linguistically similar Kuria to the southwest and to the Logoli group of the Abaluhyia from north of the Kavirondo Gulf. Traditionally, Gusii life centered around the herding of cattle while cultivation was relegated to a secondary position. How- ever, in the last two decades increasing population 93 pressure has forced the people to devote more efforts to cultivation and consequently less to animal husbandry (LeVine, 1963, pp. 221-255). Social Structure Prior to the advent of British rule in 1907 the Gusii were made up of seven separate but linguistically related groups. All of the Gusii recognize a common an- cestor, Mogusii, the founder of the nation and the one after whom it is named. Despite the common ancestor and history the entire national group was never unified. On occasion they were united for purposes of warfare against the Kipsigis, but they generally fought among themselves (LeVine, 1966, pp. 3-4). Largely due to the separateness of the seven groups, the British structured their administrative sub- divisions along similar lines. The territories of the seven groups then became the original seven locations. They were as follows: the Getutu in Kitutu Location, the Mogirango in Mugirango Location, the Nchare in Wanjare Location, the Bassi in Bassi Location, the Nyaribari in Nyaribari Location, the Majoge in Majoge Location, and the Mogusero in South Mugirango Location (see Figure 14). The slight modifications from group to locational names were made by the British, presumably by error rather than plan. In 1907 the Mugirango split into two sub-groups; the splinter group becoming the Mogusero 94 Figure 14 KISII DISTRICT / : LOCATIONS and SUB-LOCATIONS g/ KIIIIIII West we“... Madam : "-....,_._(}ontraI: it. South Mugirango Maioge Bo'I‘obu /"-/ \. -/ _____..._ M' BOUNDARIES —--- District -— Location ......... Sub-Location 9 - K_m. - 20 '////I Ki sii Town BASE MAP SOURCE: SURVEY OF KENYA 95 of South Mugirango Location. Despite their physical separation they consider themselves as one. In 1962 several of these locations were divided to give the present 12 locations, plus the former settlement area, currently found in Kisii District (Maxon, 1969, pp. 350- 363). As the Gusii were not unified, similarly the indi- vidual groups lacked central authority. The only exception was the Getutu that consisted of only the Nyankundi clan, rather than numerous clans (Maxon, 1969, p. 350). The other groups of the Gusii are made up of numerous clans that are patrilineal, exogamous, segmented into lineages, and without central authority. Clans and lineages were more localized in the past, but today, after years of movement within the district they are not so territorially distinct. Movement has lessened in recent years because of the lack of vacant land (LeVine, 1963, pp. 221-255). Much of Gusii social life is organized around the patrilineal descent group. The male members of several homesteads who are descendants of the same grandfather regard themselves as members of a common mourning group. They share in ritual head shaving and sacrificial meat eating at each other's funerals. Two or more mourning groups, with the same grandfather or great-grandfather, form a Eiigg. Beyond this group little intimate social interaction is possible due to the numbers of people 96 involved. Several riiga lineages form a clan-house and several of these form a sub-clan. The largest of the social organizations to form an independent political unit before the British came was the clan. Clans are also the largest exogamous unit and the maximal group for the use of kinship terms. The next larger unit of social organization is the group, and finally the Gusii nation forms the largest unit (LeVine, 1966, p. 30). Before 1907 each clan was essentially an autonomous unit with its own territory and decision-making power. Forces of unity and disunity were in a constant state of ebb and flow. Clans within the same group would carry on prolonged feuds involving the abduction of women and the stealing of cattle. However, due to the exogamous nature of the clans, they had to have periods of friendly re- lations, or at least a lack of overt hostility, for the purpose of marriage ceremonies. Inter-clan hostilities could be ended by negotiated settlement and the payment of cattle. Participation in warfare also served as a unifying force. Thus, military alliances and the need for wives prevented hostilities from extending for in- definite periods (LeVine, 1966, p. 4). Today, most of the lineage groups of the Gusii are no longer intact but are fragments found in two or more places in the district. During the 1930's and into the 1940's considerable migration within the district took 97 place. Generally, the population pressure of the lowlands forced people to seek new homes in the higher parts of Gusii-land in the east. Also, the British administration forced a reduction in hostilities between the Gusii and the Kipsigis to the east, so this area was no longer needed as a buffer zone. Soon, however, population pres- sure all over the district brought an end to internal migration. Lineage groups that migrate to new areas may become localized under work groups (risaga) that form a geographical unit. A single lineage that has been frag- mented into two or more parts may be identified by the local risaga group, or as part of a risaga, in each area. If several lineages within the same clan have migrated to a new area they will interact more with each other than with others (LeVine, 1963, pp. 221-255). The large risaga work group is based on a small number of homesteads that recognize the reciprocal obli- gation to participate in the trading of work for beer. While all members of the large risaga are not necessarily of the same lineage they are almost always of the same clan. Each small risaga is made up of members of the same lineage who work together more closely with each other than with the other members of the community. Thus the small risaga is used more often than the large risaga. Boundaries of the large risaga are frequently marked by natural features such as streams, but the 98 neighborhoods that make up the small risaga do not have clear boundaries. Since there is no formal organization among these groups, people, particularly those residing on the boundary between two groups, will work with first one then the other (LeVine, 1966, pp. 35-36). The extended family homestead of Gusiiland has both a social and spatial expression. In a monogamous marriage the husband and wife usually reside in the same house and unmarried sons who have been initiated will have their house. In a polygamous homestead each wife will have a house of her own, and the husband will spend time at each. Sometimes the unmarried children will have their own house if numbers demand it. Married sons with their wives and children will have their own houses. These houses will be built closer to their own mother's house than to any other (LeVine, 1966, pp. 26-27). While the head of the household retains title to the land and cattle each wife will maintain her own land and cattle. The husband is responsible for running the homestead but the individual wives do most of the work in their own fields. Currently, men engage more in field work than in the past when they devoted their time almost exclusively to cattle raising and trading. When a homestead head dies his wealth will be distributed accord- ing to his instructions, however, if differing claims are made, a son will usually receive the land that was 99 cultivated by his mother and the cattle associated with her household. Such claims are not totally unheard of because there is frequently friction between co-wives in the homestead and this tends to influence the son's behavior (LeVine, 1966, pp. 26-27). Before the coming of the British colonial adminis- tration in 1907 there was no central political authority among the Gusii. Political integration was at the clan level, and sometimes at a more local level. Each clan and local community had its own authority system in which considerable influence was wielded by the elders and the wealthy. In 1907 the British imposed a new level of govern- ment on the old system rather than replace the old system entirely. They appointed a District Commissioner as the chief representative of the government in the district. Beneath him are the District Officers for the divisions; which are made up of three or four locations. Within each of the seven groups the government appointed a chief who became a man of considerable influence because of his position. Each chief has under him a number of sub- headmen who function over several clan territories in order to reduce clan parochialism (LeVine, 1966, pp. 66-76). Within traditional Gusii society, and within the present day society to a large extent, it is the wealthy who are the most influential. That they were also 100 lineage elders is, to a great extent, a reflection of their wealth. Such men have more cattle wives, sons, and more daughters to allow them to acquire yet more cattle. As they derive much of their wealth and power from their farms they would naturally be the most influ- ential persons when it came to agriculture. Indeed, many of the first coffee growers were either chiefs or govern- ment officials of some sort. In order to maintain their wealth they proved to be the first in the area to take up cash crop farming. The great acquisitiviteness of the Gusii is largely caused by the esteem with which the wealthy are held (LeVine, 1966, pp. 66-76; Uchendu, 1969, p. 21). Infrastructural Development Transport has been a principal problem in the development of cash crop agriculture in Kisii. It is not that the remote areas are inaccessible, but that the cost of transport over very poor roads becomes excessively high. Currently there are only three sections of road that are tarmac surface in the district (see Figure 15). They are the roads leaving Kisii town for Kisumu in the north, Migori to the southwest, and a four-mile section to Kegati on the road to Sotik to the southeast. The other main roads are murram (gravel); they are passable in all weather. However, most of the access roads are 101 Figure 15 KISII DISTRICT ROADS and TOWNS Maranl 0 To TANZANIA , / 9 K9" . 29 SOURCE: SURVEY OF KENYA, “250,000 TOPOGRAPHICIKISUNU), I969. ) . \ . Rigomo T has... .A. Markets Primary Road (Paved) Primary Road (Gravel) Secondary Road I Unsurfacad) Kisi i District Boundary 102 unimproved and very difficult, and therefore expensive, to travel over. In the late 1950's the Kenya Tea Development Authority began constructing "Tea roads" in the areas of high tea production. The KTDA charges all growers a cess on their production, and much of that money is invested in road improvement to insure that the collection trucks will be able to get into all areas regardless of the weather. Obviously these roads make the marketing of all other cash crops more economic (Uchendu, 1969, p. 50). Other transportation improvements are planned. A new tarmac trunk road is under construction from Kisii southward to Nyangusu on the southern border of Kisii District, and subsequently farther south. A new road connecting Kisii directly with Kericho to the northeast is now under construction. The present road connecting Kisii to Sotik and then Kericho will be improved in the next three years. The feasibility of connecting Kisii to the Kenya railroad system is under consideration in Nairobi. Also, the present landing strip for light aircraft will be improved (Kisii, District Development Advisory Committee, minutes, March 1, 1971). Roads are the most important infrastructural development influencing the profitability of cash crop production, but there are numerous other developments aimed at upgrading the rural standard of living that are important in reducing migration to major urban areas. 103 In 1969 Kisii town for the first time had 24-hour electri- city brought in from the generating plant at Homa Bay, 15 miles away. Plans are underway for the construction of two small hospitals in Nyamira and Keroka (see Figure 15). Water systems for these two towns are now under construction and systems for most of the major towns in the district are planned for the next 15 years (Kisii, District Development Advisory Committee, minutes, March 1, 1971). Of major importance in the district is the ex- pected growth of Kisii town. In 1969 the population was 6080 (Kenya, 1969, p. 41), but the Nyanza Province Town Planning Officer estimates a population in the year 2000 of over 60,000 if the present rate of expansion continues. The most severe problem is that of space. Currently the township covers 250 acres and is expected to expand to over 2000 acres at the turn of the century. Given the high degree of local relief this magnitude of growth will present great problems (Kisii, District Development Advisory Committee, minutes, March 1, 1971). These physi- cal improvements in the district should aid greatly in upgrading the local standard of living by making cash crop farming more profitable. Kisii District simultaneously represents the hope and horror that is potential in Kenya's future. If pro- grams of rural development can convert sufficient numbers 104 of Gusii farmers into highly productive cash crop pro- ducers without sacrificing food production the future is hopeful. On the other hand, if these programs fail, and the population of the district continues to grow at the same rate as in the recent past, it will certainly experience the manifold horrors of overpopulation. CHAPTER IV DATA COLLECTION AND METHODS OF ANALYSIS Data Collection Selection of the Study Area In order to cover a variety of ecological con- ditions, population densities and potential agricultural development levels the Special Rural Development Project selected 14 different areas in Kenya. One of the 14 areas was in Kisii District where the author was conducting his own field work. After some initial consultation it was decided that it would be mutually beneficial for the writer to work with the SRDP Research and Evaluation Unit in a joint field survey that would serve as the base line for evaluation purposes and also provide the data base for the author's own spatial diffusion research. The SRDP area for Kisii District covers only Irianyi Division, that is, Nyaribari Masaba, Nyaribari Chache and Bassi Locations. This area was insufficient in size for the proposed spatial diffusion research so it was expanded to include East Kitutu Location, most of Central Kitutu Location and part of North and West 105 106 Mugirango Locations (Figure 14). This additional area was included at the author's request as it is where pyrethrum, tea, coffee and passion fruit were first introduced. Selection of the Sample Due to the lack of farm plot maps for the entire study area the selection of the sample for use in the field survey was completed in two stages. Overlying the 1:50,000 scale topographic maps used in the field is a grid system, each cell of which is one square kilometer. These cells were numbered in serpentine fashion, beginning in the northwest corner of the study area. Every ninth cell was selected, giving a total of 93. Each sample site was a square area that could not readily be utilized in the field, so using the topographic maps, physical fea- tures such as roads, rivers and ridges were used to out- line irregularly shaped sampling areas of about 0.75 square kilometer each. Physical features are a simple, accurate and easy method to delimit areas in a place with as much local relief as Kisii. A total of 1,935 short interviews, or about 5 percent of the estimated 40,000 homesteads, were conducted in these areas from March 15 to May 1, 1971. The short interview consisted simply of asking the name of the homestead head and the dates when he first began to raise hybrid maize, coffee, tea, Pyrethrum, passion fruit and grade cattle. 107 The second stage, conducted during May and June, 1971, was to do a stratified sample of the homestead heads interviewed earlier. This was done by placing the 1,935 names in order of place of residence and selecting every fourth name. A total of 485 in-depth interviews were conducted, representing about 1.25 percent of the home- steads in the study area. The first stage of the sample, the 1,935 short interviews, has a 99 percent confidence level with a reliability level of 13.0 percent. The second stage, the 485 in-depth interviews, has a 95 per- cent confidence level with a reliability of 15.0 percent (Arkin and Colton, 1963, pp. 145 and 152). The Field Survey and Data Collection The field interviews were conducted, under the author's supervision, by a team of three Gusii men who normally are employed by the Ministry of Agriculture as agricultural extension agents. After one week of train- ing a second week was devoted to pre-testing the inter- view schedule. The in-depth interview schedule then underwent modification while the field team began the short interviews. The latter is referred to as the diffusion sample, for these are the data upon which the diffusion maps are based. The diffusion sample was com- pleted in approximately three weeks (see Appendix A for the short interview schedule). The next two months were 108 devoted to the in-depth interviews (see Appendix B for the in-depth interview schedule). In addition to the field survey, local data sources were searched, especially annual reports of the district agricultural officer. These sources are the most valuable because they are the only places where statistics are kept in the same format from year to year. But even here the same tables do not necessarily appear each year. Also, the files of various offices provided additional information. Other data were obtained via interviews with the individual crop officers, managers of the cooperative societies, and the manager of the Sotik passion fruit processing plant. The Interview Schedule The interview schedule for the in-depth interviews was designed to obtain factual information about the head of the homestead and the farm. With the exception of sociometric nominations, all of the questions required only the recall of facts. The schedule provides loca- tional identification for all respondents. The entire schedule contains 22 questions, each with numerous sub- parts. The categories of inquiry for each of the questions are as follows: (1) The name of the homestead head and his residence, or if absent, the name of the respondent and relationship to the head. (2) The land tenure Characteristics of the homestead, such as the number of 109 pieces of land, whether they are owned or rented, how the land was acquired, and the state of legal title. (3) The kinds of cash crops grown, when they were first grown and the practices of weeding, pruning and fertilization used. (4) The type of maize, whether hybrid or local, planted during the last year. (5) The types and numbers of grade livestock kept, and when they were first acquired. (6) The numbers and kinds of local livestock kept on the farm. (7) Animal husbandry practices used for local and grade dairy cows. (8) Sources of water supply (well, raintank, river or spring) used for the home, livestock or irri- gation. (9) The use of various types of traditional labor, family labor and hired labor for cash crops, food crops, and livestock. (10) Names of markets used when selling chickens or eggs, milk, food crops, or livestock. (11) The kinds of farm records kept. (12) The sociometric nominations (already mentioned) of opinion leaders. (13) Types of client and agent initiated contact with change agencies. (14) The types of demonstrations attended and the ones that were found useful. (15) Whether or not the farmer has ever had a demonstration plot on his land, the number of demonstration plots he has visited in the last year and the number of times he, or others from the farm, have attended courses at the Farmers Training Center. (16) The formal organizations family members belong to and offices held in each. (17) The communications behavior 110 of the homestead head, such as newspaper and magazine readership, frequency of radio listening, frequency of visits to Kisii, outside the district and to Nairobi, frequency of church attendance, and length of stay outside the district. (18) The primary and secondary occupations of the homestead head, and the respondent if the head is out of the area. (19) The personal characteristics of 'fi. the homestead head or respondent, including education, literacy, sex, marital status, age, birthplace of the flu-“m homestead head, birthplace of the head's father, and religion. (20) The number of people living on the farm, including the number of adults and children, the number living away, the number in primary school and higher school, and the persons responsible for paying school fees. (21) The society where the farmer markets his cash crops and the income derived from each. (22) Housing characteristics, including the type of house, the toilet, cooking and lighting facilities, and the possessions of the homestead. Methods of Analysis Data Coding and the Raw Variables Due to the multiple use for which the interview schedule was constructed only selected parts were coded for use in the spatial diffusion research. The data from the 485 in-depth interviews were aggregated to the 93 111 sampling areas. An effort was made to construct a number of Guttman Scales, but unfortunately the coefficient of reproducibility was not sufficiently high to justify its use, so indices were constructed using the same variables. The 57 raw variables used in the statistical analysis are listed in Table 6. Diffusion Maps The 1,935 interviews conducted in the 93 sampling areas (a mean of 20.8 interviews per sampling area) yielded dates of adoption for each of the six inno- vations. For each sampling area the percent of the farmers adopting each innovation was calculated for the appropriate time period. Calculations were made for 1940, 1945, and beginning in 1950, every two years up to 1970, plus 1971. Each innovation fit into these time slots according to the date of introduction into the study area. These data were then punched on computer cards and used for input into the SYMAP computer mapping pro- gram for the purpose of constructing a series of contour maps depicting the percent of adoption within different parts of the study area for selected years. The SYMAP program generated maps on which a series of closed curves connect all points having the same numeric value. Each data point is the centroid of one of the 93 sampling areas. Between any two contour lines a continuous variation is assumed. Thus, the maps produced 112 TABLE 6.--Variables from In-Depth Interviews. Aggregates for 93 Sampling Areasa 18. 24. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. North-south coordinates of sampling area East-west coordinates of sampling area Elevation of sampling area above sea level Annual precipitation Index of distance from Kisii town, based on road qualityb number of pieces of land per farm Mean Mean Mean Mean Mean Mean Mean acres per farm year year year year year of adoption, of adoption, of adoption, of adoption, of adoption, hybrid maize coffee tea pyrethrum passion fruit Mean Mean Mean Mean Mean Mean acres acres acres acres acres acres of of of of of of Use of fertilizer index Index based on number of weedings Mean year of adoption, grade cows Mean number of grade animals per farm Mean number of grade cows per farm Mean number of local cows per farm Grade cow practices index Local cow practices index Tick control, number of dippings or sprayings per month, mean per farm Crop progressiveness index, mean number of years from 1972 that each cash crop was adopted Overall progressiveness index, crOp progressiveness the mean number index plus grade cows Percent Percent Percent Percent Percent Percent of of of of of of were adopted farmers farmers farmers farmers farmers farmers raising raising raising raising raising raising hybrid maize per farm local maize per farm coffee per farm tea per farm pyrethrum per farm passion fruit per farm of years from 1972 that hybrid maize in 1971 coffee in 1971 tea in 1971 pyrethrum in 1971 passion fruit in 1971 grade cows in 1971 Mean last Mean year Mean Mean year number year number number number of change agent initiated contacts in of client initiated contacts in last of demonstrations attended in last year of demonstrations found useful in last TABLE 6.--Continued. 113 Aggregates for 93 Sampling Areas 40. Mean number 41. Mean number attended 42. Mean number 43. Mean number of of of of demonstration plots seen in last year Farmer Training Center courses formal memberships offices held in formal organizations 44. Communications index for sampling area 45. Education-literacy index for sampling area heads-of-households 46. Mean 47. Mean 48. Mean 49. Mean 50. Mean farm 51. Mean 52. Mean farm 53. Mean area is located age of number number number number number number income of of of of of of people per farm adults per farm children per farm family members currently living on children in school per farm children in high school or above per from cash crops per farm 54. House type index 55. House facilities index 56. Household possessions index 57. Population density of sub-location where the sampling a . . Mean number of interViews per sample area, 5.2. b Value of 1 for each km. of all-weather road, 2 for dry-weather roads, and 3 for footpaths. 114 will not be completely accurate in every detail because the areas between the sampling areas were not surveyed. Instead the maps will depict the pattern of spatial diffusion in a somewhat more general form. The contour intervals used on the diffusion maps are similar to the percentages used in the adopter cate- gories by Rogers (Rogers, 1971, pp. 183-185). The six intervals, which will be referred to as adoption levels, are as follows. TABLE 7.--Adoption Level Percentages. Level I, No Adoption Level IV, 50.1% to 84.0% Level II, 0.1% to 16.0% Level V, 84.1% to 99.0% Level III, 16.1% to 50.0% Level VI, 99.1% to 100.0% Level I, no adoption, was used to depict those areas where no farmers had accepted the innovations. Level II goes up to 16.0 percent, eliminating the inno- vation category as the sample size was not adequate to allow the use of a category of only 2.5 percent. To use the innovator category would have required interpolation beyond the limits of the data. The remaining deviation from the Rogers adopter categOry percentages is Level VI which represents 100 percent acceptance. Because of the data limitations this study will not be able to deal with the innovator class (the first 7*” “’73- ‘ ti.“ *I' 3‘. v 115 2.5 percent to adopt). Instead, the category of innovator will be included in an expanded early adopter classifi- cation (the first 16.0 percent to adOpt). A true inno- vator in Kisii may well have adopted many different inno- vations besides those investigated here. Since all of these innovations were introduced by change agencies it is logical to assume that the first farmers to adopt them were well integrated into the social system. Thus, the questions asked cannot really address themselves to those in the innovator category, and particularly the first of the innovators. Factor Analysis Factor analysis is a statistical technique used to reduce a large number of raw variables to a more manageable number of conceptual factors. The assumption upon which factor analysis is based is that many variables are intercorrelated with each other, and therefore a new conceptual variable (usually referred to as a factor or sometimes as a dimension) can be constructed that will be highly intercorrelated with one or a cluster of raw vari- ables. Some measure of the relationship the variables have to each other can be obtained by examining the simple intercorrelation matrix. In a factor analytic problem each raw variable will relate to each conceptual factor in a different manner.‘ The closeness of fit between a variable and the v. --o m” 116 factor is referred to as the loading and is measured in the same manner as the coefficient of correlation, that is, the values range from 1.00 to -l.00. A loading approaching plus or minus 1.00 indicates a close corre- lation between the raw variable and the factor. Loadings approaching 0.00 indicate no relationship. The numerical value of the loading indicates the closeness of fit be- tween the raw variable and the factor, but when two vari- ables have opposite signs it indicates how they relate to each other. A positive loading indicates a direct re- lationship with the factor and a negative loading an in- verse relationship. This can be referred to as loading in opposition. Thus, if two variables load in opposition, one approaches its maximum numeric value in those obser- vational units where the other approaches its minimum numeric value. Another way of stating it is that the factor has dichotomized between the two variables. The fact that a given variable has a positive or negative loading on a factor is meaningless in and of itself; it only becomes important when both high positive and high negative loadings occur on the same factor. The naming of factors sometimes presents problems. If all the variables loading highly on a factor are simply measures of the same underlying characteristic the problem is easy for it is only necessary to identify that charac- teristic. Complications arise when the raw variables seem 1’ 5. I‘(. _ _. ~.',.‘_ m: -‘A‘ 117 to be unrelated but load on the same factor. In such cases it may be necessary to use some rather long and unwieldly titles. An important characteristic of factor analysis is that the factors derived are unrelated to each other. Also the first factor explains more than any other factor, and it may be located between independent clusters of interrelated raw variables which will result in numerous moderate loadings and perhaps none that are really high. In order to reduce the number of variables loading moderately on a factor, the matrix is frequently rotated to increase the relationship between the clusters of interrelated variables and the factors. The use of orthogonal rotation does not change the uncorrelated nature of the factors, but it does clarify the variables related to each factor. In the current problem, varimax rotation, a type of orthogonal rotation, was used to pro- duce a better fit between the variables and the factors. The next step in factor analysis is the con- struction of a factor score matrix. This explains the relationship between the observational unit and the factor. The more involved a raw variable is with the factor the higher its weight. Scores are given in standard deviation units, thus if an observational unit contributes little to the factor its score will be very low, approaching 0.00, or the mean for the contribution 118 of all observational units to the factor. Observational units will have high (positive) or low (negative) factor scores in the same manner as their raw variables relate to the factor (see Blalock, 1960, pp. 383-389; Rummel, 1967, pp. 444-479; and Cattell, 1965, pp. 190-215 and pp. 405-435). Multiple Regression and Correlation Multiple regression is used to indicate the amount of total variation of a dependent variable that can be accounted for by a series of independent variables acting together. However, if the independent variables are highly intercorrelated among themselves an erroneous interpretation could result because one of the basic assumptions is that the independent variables are un- related and the values for the observational units are normally distributed. Therefore, factor scores can be used as the independent variables because they are by definition unrelated. Also, an independent variable cannot be simply a surrogate measure of the dependent variable or an erroneously high correlation will be the result (Blalock, 1960, pp. 326-329). The next chapter will examine the diffusion process for each of the six innovations in detail, as well as the results of the factor analysis for the 485 in-depth interviews. Then the adoption of innovations 119 will be related to the conceptual factors through the use of multiple regression and correlation. CHAPTER V THE SPATIAL ATTRIBUTES OF INNOVATION DIFFUSION IN KISII DISTRICT Introduction With few exceptions (Nwala, 1971 and Ramachandran, 1969), spatial diffusion research has focused upon the developed countries, particularly Sweden and the United States. The study offered here, focusing on a developing area, offers an alternative view of the spatial diffusion process. The answers to three general questions are sought here: First, does the spatial diffusion process in a developing country follow the same general pattern found in the more developed countries? Second, what spatial constraints or influences can be identified that effect the Spatial spread of innovation? And, third, what relationships exist among a series of geographic, socio- economic and demographic variables, and the diffusion of innovation. The methods used to answer the questions include mapping of the spatial spread of innovation, factor analysis and multiple regression and correlation. 120 121 The Spatial Diffusion Process The purpose of this section is to analyze the geometry of the spatial diffusion of each innovation. Comments on the introduction of each innovation will pre- cede the diffusion analysis. Levels of adOption, referred to in the previous chapter, will be used to facilitate the discussion. Each of the innovations investigated has its own distinct growth curve (see Figure 16). Coffee presents a classis "S" shaped growth curve, with a slow rate of adoption in the initial period, followed by a rapid, steady growth, and finally another slow adoption rate as the growth curve levels off. Pyrethrum experienced a very rapid acceptance rate until 1960 when an uncertain market caused a slight slow-down in the acceptance rate. The upper inflection preceding the final phase is discernible in the last three years. Hybrid maize has had the fastest acceptance rate in the history of Kisii, but here too the growth curve is beginning to level off as the 100 percent level is approached. Tea has enjoyed a rate of adoption, in the central part of the growth curve, about equal to that of coffee. Passion fruit and grade cattle are both in the initial part of their respective growth curves as they have not completed the upward inflection that pre- cedes the central growth period. The spatial distribution of each innovation in 1971, as exemplified by the percent of farmers using the 122 Figure 16 CUMULATIVE PERCENT ADOPTION Percent -—— COFFEE ~l00 ---- PYRETHRUM TEA —— HYBRID MAISE '80 ----— GRADE CATTLE ’ PASSION FRUIT ’ -70 I I60 I I! / ~50 / I I I, / ~40 I, // l, ’30 F20 I I I I / p ’l "x' /’M IO / /’ a-r’” .................. fi’.” ............. r . . - fling... T 1 1' I940 I950 I960 |970 123 innovation in each sampling area is depicted in Table 8. A high positive correlation would indicate a similar spatial pattern. No correlation indicates an overlap between the areas of adoption but no relationship between the percent of adoption found in each sampling area. A strong negative relationship between two innovations indi- cates that where one innovation is present the other is absent. TABLE 8.--Innovation Intercorrelations Percent Adoption, 1971. Passion Grade Hybrid Coffee Pyrethrum Tea Fruit Cattle Maize Coffee 1.00 Pyrethrum -.62 1.00 Tea -.48 .40 1.00 Passion Fruit -.26 .04 -.03 1.00 Grade Cattle -.19 .12 .55 -.17 1.00 Hybrid Maize -.36 .33 .17 .22 .11 1.00 Three reasons can be offered for the relationships found among the individual innovations. First, the current distribution is the result of a unique set of origins for each innovation that is further modified by a unique diffusion pattern. Second, the differing ecological con- straints selectively modify the spatial distribution of 124 each innovation. And third, the complimentarity or mutual exclusiveness of each innovation will determine whether or not there is spatial overlap. For example, the negative relationship between coffee and pyrethrum is the result of all three forces. First, coffee was introduced on the western side of the study area and pyrethrum on the east. Second, coffee does not do well at the higher altitudes and pyrethrum does poorly at lower altitudes. And third, they are competi- tive in that the farmer who removes land from food pro- duction must then decide which commercial crop to plant. A casual glance at the maps of diffusion can be deceiving because they depict the percent of adoption, not production. Also, they show the percent of farmers in each area who have at some time adopted the innovation in question. A few of those farmers may not currently be growing the crop or raising cattle. In those areas where the adoption rate is highest each farmer will usually be using the innovation very intensively, but on the periphery of the adoption area where the level of adoption is low the farmers usually raise small quantities. Spatial Diffusion of Coffee Adoption Aribica coffee was first planted in Kisii in 1921 but it was not until 1935 that more than a few acres were planted. At that time two group farms were 125 started in Nyaribari Location, but they failed because of the Gusii preference for private ownership, and the long distance each group member had to walk to get to the plot. Kisii District was one of the first areas in Kenya where Africans were allowed to plant coffee. The British administration had previously refused to allow African ownership for fear of plant disease and inadequate quality control. In the late 1930's private planting of coffee was allowed in Kitutu, Nyaribari and Bassi locations. It was not until the 1950's that coffee began to attain significant acceptance as a cash crop. During the 1950's and the early 1960's there was a steady increase in adoption and spread of coffee growing across Gusiiland; while from 1962 the adoption rate was much slower (Uchendu, 1969, pp. 35-37a). Coffee is marketed through the 26 Coffee Societies and initial processing (cleaning and grading) is carried out at the 65 coffee factories. Six additional coffee factories are planned. The original factories were started at Morumba, Kitutu Location in 1947, Mogunga, Bassi Location in 1952 and in Nyaribari location at Nyosia and Nyaturubo in 1953 and 1954, respectively (see Figure 17 for all cash crop marketing locations). Other factories were established at this time outside the study area. The last society was opened in 1960; since that date all of the new factories have been branches (Inter- view No. 2). 126 :ju;:mm /,//’\ N MARKETING LOCATIONS /- [i /‘ I. W /0 r’ +. 4’ S ./ ,-—-—/ + - /O ’7’. . . . \A/_L) + J . x”/ 9 0 +0 + 'i' x. 2 - + r 7 4. NT ) e ' e + N‘ l A + I 0 ’ + A I o O + \ . A A4- i KISII + + / at + A: e - / eI- ' \ ’ . . u ‘ o ‘0'.” \ E/ . . o . ‘ . . O o + o \\ . + A + Kiroka‘ih. \} O O A :‘g / o e / 9+ + ./ e + \/ . d' + 0 + r 0 - 0 + + + izI ,3 \\\-\ i + /’ /‘ A" . /-’A'° .( O . + /' /' \\\ . ’A’./. , . \ , '/ 0 Coffee Cooperative Society 0 \\ /-/ A Pyrethrum Cooperative Society \_ .V' + Tea Collection 8 Pick-up Station "/ a Passion Fruit Pick-up Point "Green Line" ----- "Brown Line" Primary Roads (3 - Igm. £0 —-— Kisii District Boundary 127 Two facts lead to the conclusion that the 1964 ban on additional planting has not been taken seriously in Kisii District. From 1965 to 1971 the marketing societies have constructed 21 new factories and have 6 more planned. Also, membership in the Kisii coffee cooperative societies grew from 41,512 in 1964 to 45,408 in 1966 and to 53,536 in 1968. Since 1968 there has been a slight decline in membership (Kisii, Annual Reports, 1964, pp. 6-7; 1967, pp. 46-51; and 1968, p. 68). The decision on when and where to build a new coffee factory is based on a combination of two factors. First is the distance farmers must transport their coffee to an existing factory and second is the amount of coffee produced in the new area. If there is sufficient pro- duction at a long distance a new factory will be built. Part of the new factory construction since 1966 could have been catching up with the need, but certainly eight years after the ban all needs would have been taken care of, unless additional planting was going on (Interview No. 2). Figure 16 indicates that for the study area the rounding off at the top of the logistic curve began in 1964. In the future there will be either very little new adoption of cOffee, or perhaps none. As regards produc- tion, Kisii District has continually increased its portion of the Kenya total (see Table 9). In the early 1960's only about 2 percent of Kenya coffee was accounted for by 128 .0hoa use .000a .500a .mooa .Mo>u:m oaaacoum .ehcox can .0h0a 0» amen .uuummoa danced .uUwuuuao awoaz loam ceasesou "uuuusam .ne>wo cud uanuu couoncm new nousuwu aowuunuoum unseeded o: :.uanuha II adso vouquu .00auoavoua eacex Havoc uo useouem a no vodka-«0 «deans 0m.«» .m.0 I 00.a axe 0.a «.0 «.0 m.a 0.n v.0 v.0 unuuuom man.0 vao.a vah.a nv«.0a oah.wa vhn.« 0mh.n II II II sauna 000.«ao 000.0a0 000.000 000.mh0 000.amm 000.0nm 000.anm 000.0mh 000.n00 000.0«0 unseen Danna acaueua w.« «.a 0.0 m.0 m.0 m«.0 nu.0 00.0 00.0 «0.0 useouem 005.com 00a.naa «mo.>h mo«.«m II hav.v« oa0.o woo.m mvn.« mm«.a «wean 000.0ma.aa 000.om«.m 000.5«m.0 000.000.o 000.ann.h 000.ooh.h 000.0«0.o 000.voo.0 000.00h.v 000.n«n.m canon one m.av m.0« m.mm v.m« n.0m a.h« n.aa 0.m n.a~ n.m acouuom «00.0«m ama.«mo 00m.«a0 0m0.nv0 00m.mnm was.mv« 0A0.n«a 00a.0«H 000.0h« n05.0«d «Hafiz 000.H0«.H 000.0«0.« 000.mmo.« 000.«mm.« 000.mhv.a 000.000 000.000.H 000.vH«.« 000.05v.« 000.0~v.« excel nahnuumxm n.v v.h 0.n 0.n 0.n 0.« o.« a.« a.a 0.N AUGOONOQ avn.nam env.mnm «n0.00v «om.vmo «00.5mv «00.000 .wma.mm« 0m0.0m« 000.am« n00.vv« damn: 000.00m.0a 000.050.«H 000.m00.na 000.0Ho.0a 000.cvm.«a 000.0«m.na 000.000.0H 000.«0v.0H 000.v«v.0a 000.0Hv.m e>cou wouuou mama 000a nwaa 000a mwma 000a noma «00a aoma 000a o.ax .onmanoma .uoauuuao aauau can ease: .ucusuex mono £IIUII.m mdmtfi 129 - .. - 4 - ......—.......u..-...--.‘..nun-.‘nnu-a-pn-z ;... .. . . . Ia . .I I ........|.. ‘ t . - ' I ‘ I : l ' ' ' I I I . I . . ‘ l I I . I i I ' ' I l . ~ | - I - I I. ‘ ' I II' " - I I l I l l l y . 0 , I l l ' I l I I i I . I I I - 2 - l . T ‘ ‘ I i I . . ' l I ‘ ' V ‘ . I . , I I l | l I ' I ‘ I I I I , ' l I | I I . l l I .. I y l I . I I I I . o 1 : ‘ 4 ‘ ; - I l I . : ' ' I . ‘ l I ' I l . . . l o I I . | l I . i I ‘ l I l ‘ ' . : = - 5 — : I l . ' I I I I I I l ' l . , . ' ' I l . I l l I . Y I ' I . I : I l ‘ O I l _ v .. 6 - I = I ' I , I ' I ‘ l l I . . I - - ' l ; I . .. . i l I V I i l l I v | v - T - n I ' Y I ‘ I I I I I 1 l l : | ‘ . .. . o . l l . : ; b) I945 : ‘ ' l , - n . .. , ._. .. ... .7 ...., . ... ...........I........... . . ‘ I I I I I I I l I I I | I I I l I I l I I I I l . . . . . . u...‘ - - _ : .- - -..._....- ..... ... ...... ......... ...._.... ,... .... ...,-.....» v - , . - II I I I I I I | i : I . ' I I I I I : : .. I .. = l l I I 1 I i I l I I I I- 2 - I I I I I I I I I 'I 'I l l I I V l I I I I i I I I I I I I I I I I i 'I I I I I z - 5 - I I I I I l T I I l I I I I I I x I I O I I I I I I I I :- T - I : i I I I i I I . (. . : g I I d) I952 g I I - SPATIAL DIFFUSION OF LEVEL '- "0 ADOPT'O" COFFEE up”; LEVEL II. 01% to l6.0 % LEVEL III. I611 to 50.0 % LEVEL IV. 50.I s to 84.0% LEVEL v. 841 x to 99.0 x LEVEL VI. 991% to I00% u.....-............,-......,.. ...-...............u......,. A»r<..—~..D.-.».u ‘ not-q— KISII° III I954 :3." .--..-.............;.....-...-.........._..I...."...I...._.I.. ...-..I..,...._...........-.........‘..I,.I......,......yI.. o-u-cu-n- nun-an—o—n-n-o-‘n-I-n-nau-n—Iu-u-o-nnna l23456789|0|||2 l-.--... ........-., V....Ip.. . .I.. . . . . . -. ... ..... . ....... ... ... ... r..u -I.»..-I-- ....A.|.. ....,,.--- ——-o—o -—~_.——n~n~—-—--nuon-uu nun—nun. c) I959 nun-q...-._—..—_——o——-¢——-c—_-_a—_—_u—_-¢l—-—_o_—-—u—_—uu -—--.-—-.-. -—..-.._.._- --—. . ..- -.......-..._.......I..-......u.....-...-.........I.........l......I..|... 130 ' KISII- Z f: 3 -215, .H... I . . .. . . . ...7.. ..I.. ......-..-......n. ._....._._..... d) I960 .........I.,....... ....H I, ., ..... _. ... .. . . .....4.....-...- . I l I I -—-go "w..._‘.....p.... I-—<.<-——‘—-—<0—< SPATIAL DIFFUSION OF COFFEE N (’4 51 G LEVEL I. N0 ADOPTION LEVEL II. on; to ISO x 35:5"; LEVEL III. I6.l % to 50.0 x LEVEL IV. 50.l as to 84.0% LEVEL v. 84.I III to 99.0 VII LEVEL VI. 991% to I00% I I Il'lll... o.....-...|.....-.-- ---.....l.....‘,V.-H.r.. .z ..v..v.v ..---. -..........A..u........-..- “.maH-nn u a. l IIIIIIICI ac. -~.....4..-.... ..... _ I I KB”- I DI IIVIII Inn-Illu- ICC: ' litlllulnlul . . .u IdeOOJ-‘nogol.IQ I l l I 1 Q 'IIUIIIIIIIIII‘C I I I . l III-no I n...uu uh... u ..-..... u. . .. .--.. u... . _ . ..-o—n, - - _._ ._............-...I...... - ...l... .__._—......-- ._—-. .—.<-—-._.——__.__._c_——..-____._‘-q—...-__p a“_._.__.--<...—<—_.-—q_. Ina-no cau—sn-cnfi -w-c—O--.—-.—_— l—-—o-_--. L...——- .———¢-——-n-————c————u <-——-—._-_- __-—- . .<-....._....-._-..¢__ - L.......... n t I n n a . . I n . . ou—«Iu—.._ .— .—._._——-.o——<_.___~o— |—-—I-‘——-m 152 Nwmu zouwauuuo zamzhmmrm 8.6 8 o 860 8.0 2.8 «0.0 mm musmfim Figure 29 a a: a 2 8 O U C O n. D n ' O O C N O ifi4 + ‘r J- 3: 3 8 3 8 o v o o o o- c- u- D D 153 vvvvvvv w .. PYRETHRUH DIFFUSION 1954 154 , : I fry"! g | , w \um / \ / 7::— I | I EV ‘~\\‘ ,..::’/ ‘!(fld I f I I .15 ‘ << I II II': /’ x - ‘|Ill.::“' , I; IN" ' \‘ g” "M WI?!“ \‘7‘ .II "'{5 "M 5“ :I . .'".'.’I1 III" ‘. ghlfI‘iI‘I] I‘jI IIIIIIIIH‘ “II. I‘u'l'if" IH IIII‘I In ‘I I ‘I" ”I‘ll! If I. H " I I I I l l‘ II II I ‘I ‘ ((Iw W «11“«(I L ‘I x ./I ”I \f\\ 3‘“ I} I WI \ x K ,>) \‘x *' m @(7 I \’\‘\K\;‘.l\ I; .III . ' \\\\\< ’(I \ I 777 “I: ”\\ I. 1" q s‘;\ “I! " ” 8.83338 800’... ###### :8 8 8 3 8 ...... N “““““ / j/ / \ ‘ \ / ll [/I\\\\i\\\\\\\x ( . ' [I ‘ { I, \\ I 33 8.8.1:: 88888 ' ' r r v j N d ...... fi a / (TV/j]: Kg \:*l¢[{<((&k\\ ’//I l :‘I I" ' k\\\\ I I I; \\ o r \ " 7’" , ‘ “ 5 I. III . \‘¥ K. ~_ \“\_\ , [/I I ‘ \I I f/ I ‘ r~ \ _ I” ,I/ H ‘ ‘I ‘ K \ A \ ‘ \ I (I/ \\ I; .\\ KW «6‘ \ ~ . ‘\ I \\\ H i; \ \\ / \ {I I] V I ‘ \\ ’ \\, \‘ ,, I \ é \ II E. l ’ I ‘ \ J. I III . ' ‘ I I . l l [:1 | I ”////x ”/1 ‘.\‘ if. 159 8. F 'N a 8 8 ' . on c A A 4* A '—'_ ' ' l ‘ I F ure 36 160 \\ '7 951-37 * —.;.+ , WI «(WW/w; _;—;_ e ’ f; 4 . “// ”“ ;:”“iifii=====§ / ”MK/GRIN“ $>>>>>M~ 3 . ,. Nu :’g§§ «: l ; .,,/” III§ # *' W ll (KKK ‘ ‘3 LR??? ‘ + 5 F .I: & $§§ijEP \r fi@&§» I <;. , s. ‘~‘(‘\\ \l‘ .\‘\\\\ .f/ ’ ‘\\ \\ I, ,. I. \Www _ .. <. \ , §<é> \ ‘5“\ , t ‘. \ r / %“ ;F\ W ’// \ §< \\ "‘1.“ I‘lI‘. . M _ _ \ Mp I \ H .EL W A1», \ 161 ..... \ 77:»‘77'7777; ‘ . ‘ w/Tl‘..7_§';_':_'-:_~:_-:#-‘j_ '“:.';;";";i’—" ,I [If] ‘ ‘ ‘1 \ \ N ' ’/ [I] ‘\ ,‘y‘l. \. 3 ’1‘”: I . l": r/ ‘1i'\“*" 1, f A_ m'.“*f“ ., 167 b) IQGG - SPATIAL ourrusnou or TEA libs $1}! 19 Q '9 S'i'ib 5 lb 6 M LEVEL I. NO ADOPTION LEVEL ll. 01% to l6.0 % LEVEL II. l6.| S to 50.0 i LEVEL IV. 50.I % to 84.0% - LEVEL V. 84.! s to 99.0 s I LEVEL VI. 99.I s to too; A... o. '3” ;" 0) IOTI 0 LEVEL I. N0 ADOPTION ;~_.:;;'- LEVEL II. 0m I0 I6.0 % {ij LEVEL III. I6.I 15 I0 50.0 s LEVEL IV. 50.I % I0 84.0% LEVEL V. 84.I x to 99.0 s LEVEL VI. 991% I0 I00% .- I IIIIHIIO 182 I 2 3 4 5 6 7 8 9 IO II I2 I 2 3 4 5 6 7 8 9 IO II I2 - 7 -' ........... 5 3 0) I970 :iE-“V' b) I97I g SPATIAL DIFFUSION OF fiiéfzi'; LEVEL I. NO ADOPTION GRADE CATTLE Au LEVEL II. 0.I% I0 I60 ‘5 LEVEL III. I6.I 1. IO 50.0 % mm LEVEL Iv. 501% to 84.0% LI 3.5 2. I 9 . *2 L0 Iiiééiiii LEVEL V. 84.I as to 99.0 x 5 3 I Kilometers . WI! LEVEL VI. 991% Io I007; 183 of the category. Thus, the relatively few adoptions have generally been widely scattered with few concentrations. The only intensification is exemplified by the appearance of two areas of Level IV (H8.0-Vl.5 and H6.0-V3.5). In 1970 the strong regional contrasts are most expressed, in contrast to the Hagerstrand's Stage II. By 1971 there was a further reduction of the non- adOpting areas, a slight expansion of Level III and the emergence of another Level IV (H6.0-V0.5). Adoption had expanded to over half of the study area. In the next three or four years most of the non-adopting areas will be eliminated so Stage III can be entered when all areas will increase at approximately the same rate. Stage IV is still many years away. Figure 48 shows how the non-adopting areas have been reduced in number and the lower adOption levels have expanded. Level II should continue to grow, and Level III will probably increase after a short term decline. Level IV is just beginning to develop. Considering the price of a grade cow and, more importantly, the problems in- volved in acquiring a loan to purchase one, the growth rate exhibited here is quite remarkable. Spatial Diffusion of Hybrid Maize AdOption When a Gusii farmer is asked the date he first started raising local maize he will usually answer that 1i34 Figure 48 GRADE CATTLE ADOPHON LEVEL PERCENTAGES Level __ I nO adOption ---- II 0.I - I6.0°/o ......... 111 I6.I — 50.0% __ IV 50.I — 84.0% Percent (H30 I90 ’80 '70 ISO *50 »4o -30 - 20 185 it was there when he was born. Maize has long been a staple in the Gusii diet. Normally it is dried and ground into maize flour, then boiled in water to make a thick porridge called pg§h_. Since the hybrid maize seeds pro- vide yields approximately twice that of local maize, and the crop is similar in taste to the old, it has been adOpted very rapidly. Local maize is grown throughout the district so no one place appeared to be most apprOpri- ate for introduction. Hence, the agricultural department introduced it through extension agents simultaneously throughout the district. A fairly slow rate of adoption was experienced for the first six years, then it became very acceptable and was adopted much faster than any crop in the history of Kisii District. By mid-1971 only about 20 percent of the farmers in the study area had not adopted. Assuming a decline in the adoption rate at the top end of the "S" shaped growth curve we can predict that it will take about five more years to achieve complete adoption. Hybrid maize was first introduced into the study area in 1959 at nine different places simultaneously (see Figure 49). Two of these original nodes were close enough to have merged and three of them had advanced to adoption Level II. In 1960 the pattern could be described as Hagerstrand's Stage I, the primary stage, but there are no strong contrasts in the percent of adoption from one area to another. ...-.... ....--- -. ...... _.,._,.,‘,. .. ........... .....-.......- ,.... .....I.................-......,_.,,_, --.u. . ..., c.- -. .-.:n..... ,,,,_,., ....- ,,.,.,._,,_ ..-........... ......I..-~ ,.... .... ._._ . , , ............. . ,...,, u-. . .... .. .... .......... ... ........ -- ........ . . .. ............... ..... ,...,.... .--«. '.-.........c- ,.,, .. ......... ,... .... ---~~~~------ ......u .. ,__..._ .....,,,._ ............. .__ ,.__, .. . ...... ........,L........ .. » ... .......... -H ----- . ....... ...»... .... .... .....-...... .lllfil . .. .......................... .... ....... .. I-------~----- .. .. . .. .. . ... ..... ..., ... ... ........ ... ------o-I-A---- ..-.......,.... -- .......................... .. ~------- ---~ ...........- . -- - >'- -o-- ..............-.... . ...‘.-»--.. ,..,..,_‘,,,,. --.~o.....I- I90 ....,.........-.. . -»o-------. ............ .. au--. -.- .. ... ...... ......... - '- I .................. ........ ...._ ...-........... ... . _._ ........ ..........-....... ... _,._. ._ , ... .....-................ . ,.... ... .. ........-.............. ............. ..................-.- ,..., ....... out.--...---- ...-...... ..., _,_,. , ,.‘ ...I.v.......~-.... ,..., A. ...-.... ..... ....... ...-.... .. .. ......................... .......,_.. .._.,, .... .. --- ........... ... ...... ...~...... ... ,............. .. .. ........ -- ........ .. ............ ... . ------------ ---- . ..~............ I- .~...... .............. ........ ... ... '0. . on... u. . .......... ...- ...... -Po ........,. .......... ..... ---- .. .. ............ II-oo .... ... roos-on-... - .............. ............ ................... ~I‘v . I '---- .....-.... ...... ...... ...................................... ........-.. .. -------~--.- . ........, . ......-............ . . ., ........r.. .. vsn- - ...... u.o.. ...... ... . ...... ..................,........... .... -H - .... .... . .:y...n.... - ......................................... - .... oou.....a... I'Iu ............ ..-.. ,.‘_,_,,..,,,,_.,.,.‘,,...,_..,,,.... ...... n.......... ...-.I - ......... .... . . ......,.... ............. . ...-... .-..... a... vu. ......... ,...... .............. ----- . v. - . ......... .... .... .. .. ... ... ..... . ... .......... ... ..... ... ...... .. ... .. -- .........n ....... ----'-- ... ....-.. . ..-.. n-oo-n... ...,....... ........ .--.. ....._. . ..---...... ..... .. . .... .-...--|.....-...'.. I- ‘ _ a - u . n y - - - - I < — n < o 4 . n n n - — - — o . a — — a - a — _ a - -._ c - -.- - n - 1 ¢ _ ‘ n n — — - u — — - — n ‘ - - _ - .....,....., .--.-....-u. ...... .. ...” ..... ..., ... ......o. .. ............ .... I ..-..-.-..--.. ....~..,.-.. ... -. .... .. .... ......... ' - ._ . ..... l . .- . . I I I I I I I I I I I I I l I I I I I I I I “ u . ‘ ‘ . ' ‘ . > u I ,_._, .... I Inc-,.....,, 0 cu. I ,......, u-u -- ... ,,_,.,__ ...-ID I __., , .- .... ..... ...... ,,,,._ . ..... l ,,__,,,,,,._ .... ...-.....e o onunu. ,_,.,, .... o . o ... . III-II .,_.,,., .n o no - -- ........ '0 I I -.-..-..... . - ....n .....v.... .... ...- - -. - -......... -- I - ...-.u-nn- .........- no. . . .....nu- .......... up .....-L...... - ...,‘.,_.,, --- -----'- ......-~...... "' """ ----II.--oo-.-. -0- III-I .............-.- o ... ...-I ..........-- I ... .......... .. 0 ' "' .....I...... DID...- - ....... cou-n '0 l' .... ... _,_.,___ 9.... ... .... ... Den... ..... .. ...: ,...... - ..... us... I. . .,,‘ c ...I. ..- ..... ...-9...... .... , n . .... ...-...... .__,_‘ uvo cu... .. a , ...-u. I, ..,,._ ... a... on V .n . ...-...... .,_,_. ... ....- . v - . ...-n. _.,,.__ ._.... , ... .. .. .. . ..... ._ .. .- .- ... 0.00-II ,. .- .. a. . III _ .... ..., o- oo In a n- , n-o ._._.., II o- -. . ,, ... ,,_._,_ .. .- .. . . l .- .... .... Illl Illllll ---- o o . ... on. IIIII lint. loI I .... -- IIIDI even I o. . ..... --. noel: I II... . ... . .......... -. I! II -...... ... . _ ,.....,. , ll n-u.-v. ... .. ............. o II--~ --« n . . .............. I- - -v- -- ......-.... no. ... .o . ...._... . II ... .. ......- Ill ll. I n...oo,,_,.__.._,,,..,,. a a.- II ._ ,... use I II ...-o ._..,,. _ onto-I I. ,.... I-..- . ...-.... ,...... ...-nu... I. ,. ,,_.,. _ llllll ...o.-.-. ,,._,. -. ..nnoo- .- ............. ----I - I-Il-an ...... III-III-I II ‘,,,.._, u .. ..4.-uol ,.... ...-.ovu. I _______,_ .- n. . - I o ,...... --n--... In ......... I .9.- .... ....... .oc- .... .....-... non-I-c-oo- ,......, -..-----... :I. "'°‘ '0 II III-Ill! ' .' 'l'l’I . ... III-I .......... . .... -¢0-- ..............-... ... '- .... ....... .. ......... ......... I ... .............. ............................ I uh. ..n..-.-..-u.-n- ....a-uo..-.............-eon l . ................. .... ...-...... ......... - ..........-.. .........-.................. .........-..-................-......... o- ...-In. It ...-s o .....- ..-.................-...-...... I- .. .. ...................,....,...... " "l " .. . ..............................-... II I"! II ..n ......-......................... -- III' 0- .. . .............-...-......-..... -' I'I . ..... fillko-lpcq ..«un-e-u-n... ‘ 'F'I' --. a I-.- ..a------I........... ‘ l" u.-- u .0 ..... ........-.-....... III! ...... .....................-.... 4 .- .........-.-.............. '-‘ ... ...................... cue. .....-........ ..... .....- .... . ........-........ IloblltlII O u. I I l u v — u v v . . ‘ u— n - o — — - n ‘1- _ - - o a — _ n o — -.— n u — _ _ _ I _ _ _ _ . --— — I — — — — a — — — - _ -— — — . _ _ - — ~ _ _._ — o _ _ — _ - — — — — n _ — - — - ......-._,...._,._._.,._,,_ ., ...... ..... ...... ............. ...._ l- —.- a - -v— _ u q - -- - — —.--.v — -- — I -.-u- o _.-n — o - _ _ - l --.- — o , .. ..................... . .v...... n.....o-u.. ...-u H ‘ ., ........- ........-...-o- ...,,... Flu... ...-u...- ,._ '.:' ,_.Af ...... ...... ..... .... ...... ........... ..... .. . -.. ............. . . - ,...-... A-fioIOIOIII ...... " "j I" .. ....................... _ ' ""'_ : .,: ....... ............., ........,.. _ . .. ...... ...... n _ ~ , ~ , ,_ ._ ....v......---I.a..-.. ...,...,, ,..., nonn-o- -cn-q..-OI--- . ....... , ... I‘ -‘ .. .. ... ..... ........-...... t-' ...--..L... ...-.- .. ...... . .-... « . ... .. .. . -....._.......... .._,.... .. ... . ... , . ..,,_,._ . ..........-...... ......, .. . ..... . __ _‘_ . .................. ...-.... ... ,. ... ... y ., , ... ..........-...... .... ........... ...... .. . .. .... ............. ..... ..... ...... .. .. I ., . .. ..., .. . ...... . ... . ,....,,,.,,,.. . . ..... .. .. , ._ ..... -.... .........u. -. --- ......., ., no . ..... .. .... . ,... . ,yy.,. ...;.................... ,‘_..._,._ ........... ........... .. .....u. . . ... .. . ....~ “ , _ __, .. . .................... ,....,... ...... ..-....-.. ........ __.. .... .. . . .,. .. . .4,. , y‘ , ‘ .............. .... ..._,.... ...-. ............. . .- . ...... ..... .. .. . . ... - .. III~IIItl-.ooholl.. ........... -.-. .....L--..o..-.......Io ._ ....... y}, _,., _ .........-.......- ............ , .......-..-...... _ j . ,.... , . .. . .. .... . .... ...-- ,, .................. - ...-... .. ....y... II ' .. . . . ... .._.,..... , .... . ,,,‘,..,,.._,,,, n...-. .- . n.-...o.- ,_, . . ...... ....... y y, . ., ......o-I a. .........,,-....... III-neuao IIOCOQOIII .. ......... I ....I. . , ,._ ...............«--. .... .............. . ...... ....- . . ,.... ., .. .. .. . .,_ .. .....,. ....... ,...,..... .. ...., .. .. -... . ,. . ,_ y. . ..... ...-,...... ,, --..--.-...... n... ...-- Ill u-....... . . . . ,, ..I.. , .....-..-...... ‘,,,.._,..‘..,_, ......u... IO'IIIIDI ........ . . . ...... ................-.- ......... ..., L.... ... lIolIIIIVDI ....... .. . - . , ..,......I. . ..... ,.........-.-... n.... . n ..... . . . .-. ............ ... .....n. _.... . ....,.... ...... ....... .. ... . ,. .. .... .... .. . ... ....... _.._,j,_.._,..,,,.. ,.. ...... . ... ... ., .. . . ... .. .... ......u- ,.... ........ .IOII‘OQ. .n ., .... . .. ... .. .....I...... . ,......._.,. ...................oa ,, ._ ... ... .7 . ,_ .I. ..... .... . _ _.,_,y_,,._,.___..,. ... ......... - ...... v...., . ... . ..... ...,....,.. .........,,......._ p.~-...... .....v-n-nu ._..., ..... . .... - . .. ,...... ..., ,_ , ,_, _.,_ ...--.--- IIIIIIIII'IIIC ..... . ............ . ... ..---- ... ....... ... , ...-n ... .uu-o .I.n.-.-v.pl ...,..., ... ,....,,._ .— , . . .... ,_,,_, ,, ,_ ..........,. . IIII leul‘rI-I'Il .___y. __ .. .. . - ,,..,,,, ,_ __, ._‘ .....-.........-. -.-.-.---.-~ :- _,.,.,, ,. , _ ... - .. . ... ,. . . ...I. -....-......... ... unn- nun... -. ,.,. ,... , ._ ...... .. I-.- .. .... ... ....... o...............--..I Illllll . ,... ,A . 7 .. . .~. .. ........ ..,,,_y ........ .. .....-.. .. ......‘.ou ....... . , ...-.. ‘~ ... ..... .,_ ,,‘. ..........,................. .........-... _' y _ 77 . .... ...... ..... ,A.,,._.. ....... ...-...................-........-o _ ,, ‘ , , .-...~...I ............ .,,, .., . ...-.... ....«~...................... . ... _ .......... ......,..... ... ,_,__.,, ...-.... o- ..n-¢--............. ...- ... . ,. ...... .. .... ._‘. ,‘_, ,... .... .. ................. ..-.. ._ ... ... ...».......... ,._ ,......‘,, ...,, ... . . - .... . ' . ,,, ......-... --...a-‘. , ‘ .-.. 1 V . ».. ...-<.. ...,..... ...... ' ' ' ... ,. . .. .... I , ... .. ...,...... .--..-. . ... .7 . ... . .... .... ... ... ... ...-... _ . .. . . . . ..... ... . .... . ........ ----- - ...... .. ... _. . .. _;--.o-.~.I I " “' " ‘ .... ..... .. ... ......... ....,.. .........-... .... ... .. .............. .......... .. ---.-~ -~:- ....... . ... ... ...... ........ ----.- .. ...-nu . _. , y A._. ,, , .,. .__ ......... ... ...“ .... ... ......L... ........ I .- llll'lll! I . .. ... .. ........ ... , .-. .......-- . u- . ............. .... . ... .... . ....... .... .... I'll-II ..... ................... ...... ,. ............,. a .. ... - Ill ...-... .,,,.. . .. . ...,.._, .. ...... . ----- . ......... ... . ... ...., .. ..... ... .n-.I-. .- u-o . .. ,..... ...._........, -..-......-. .. .. .. ......... ...... ---- -I-nn - 0 < < ..- I ,... l lu-‘I I . OI I . ,,_, ......I. .. . - .,_._ __ ..... ... . . . I ....... ... -I-'-c-o~ IIIIOO'I ...... ’9' I ' ,._ ,, ,__, ., .... . ... ,...., ..n . . ...,,,_. .,,._ .... ....... ..... ...-.ec. ,... ,.__, .... . .... ....... _,.._,.._ .. .... . ...~..-........... I-------------~- ......uu-u. """ ' ' on .. ..... ....... .... -....o.‘ ,.,,.,,,._, ......L...-os..o. ...... _. ._,.,,_.,._,_ .............a . .. . .... o ..... _......._......._ ---...~.....-. -- ,......... - ...... __..,, ,,,,,,_ ........ ..... .. ,.... ..... --'----- ............‘ ........._..... 0‘ ......................I.... ... .. .... ,.... ...._. - .. .. ....... . . . .......... ........-. -- ....... .....,.... --.I..I....-.. .............. .. .. . ---- - ...... ,.... ---- ov-----I- .... ....--.. .. .. '- - ...-I .- ..._.,..,. ............. ......... , ..- - -. . .. . ,,. ....-....... .,._,,_,,, . ... o 0. I0 .. .L I-n..- ...-... ..... 4. ........ ...n. .... ,...,,_ I. . .. .- ... .... .. ........ ._ . ....... ...-. ..... . ...... .. ... ..... . ......-....... ..... .... ......... . .,_,, . ........... .- .. ..... , ........... . __.,, . ‘ ............. __.‘_,, ....... .... . o -.. -.... ....... . ,.............. ...I, __ ........ cu. .. Io .. _ ............ .--.- ..... . ,.................. ............ ._ Iccnooo .. ...... . . ...... ... I . .- ........ .. ......... ..... ,__,,....._,,___ . ... .u . ... .,,_ . ,.... . I. ~... . . .......y...-.......... ....- y,,,,,_._,..,_, ... . .- .,._ ... - ...... .... .. .- ...... ...I».............. .... ...,,.......,,., ...-n ... ... . . ...-.... ... ... ... , . .»........... . ,,,_,.._,,,,,. ..... uc . . ....... , -.»... ..... ..... .........,,__, ..,,__...,__,,. . 1.....-.- ...... .... .....I- ...... .... ............. ...,.._.,_., ,.,__,,_ .... ........ ...,...,. ...-... _ ............... ... _.,,__.,.... .. o-nu. ... .. . .. ...... .. .. .......n ., ......... ..... ...... ...-..--....... ... I... . ..... ...... I -I. .... .. ......-.cu.u..,-..-...- .,._,.,_.., ., II II‘H up.-~.... -.-. ll - .. . ... . ..... .... ... ... ._..., .., ... IIII o ......-. :~.-- . ~-- -- ... --------- - ........................ .- I-II I ....... ... ........ ... .... .. ...... .... ., ,...,,.. .... ...... ... . . n l . . ,. -- ..-...... ....- ... ... ...-.... ....,..,. ,.,, ,... .. . | ._ ...n~. I-I....... ... .. . . ,,.,,, _ ..n-n-o I l ..,, .. .... ..-~.. - ..II. . .... ...... . ..,.___,._,.,_,,_ u. I . ._,__ _ . .. .......-. ... ..... ... . ............... II-- II I ,.y. .. ....... ..... ....-.. . ,...,.....‘.._, I! II. I In ._. n...........n-..u IIIAUI'III .......,__._,,,_.,,_,. co III! I ll. .»................ .... ...... - .......... III III Q I ... ..-..........-I-..u... . .-... - ... ,...... -‘-- ... 'I 0 ........................ --..-....I-.. ........- ...., n -. . .... .- q-u-----...........-~.,. .. . ..- ~- . ......I- __,_., .... 0 - IIIIIIII llll I o --..-~u..~.-.o -..I.. l ... .... . ....,... ...... III. on. 'IIIIIIOIIIIII o .... .. .. .... .... . ..-...... .. , -- .. ...,._. .... . ..y ......u-. ......- -..-... . ......... ... . .. .... . ... _._ ......... u.....- . ..... . . .. .... ....... -.... ,...-...... ,.._,,_., ...... . .... ..... III N ... ....................n...---_---a- .....,,_,... u...o-.. o -- . ol ........... ............................ ,,..,.....,_... ......u- . .o- ..lI‘I-In , . .................... ...... ... .................-...o . .... ...... u ... .. ...... .. ..... ................-......... ........ ..............- “ .......... ....... .- . ..., ...._,., ... ..... 8.... o OI-I .... ............ ..... .. llIOlI-l - ...,_._,. _ n... o... .. .... _ . _.. ,.... .... ............. ......... ... ....... n....-.... no ... .... ... .. ........,. .--...... .... .. .... .. ........ . ......... ..... . . ,_ ,...,,,,... H ...... .. ............. _. ,..y ,, ........ ........ .... ,.... ......,.... I . ~.. ........... , ,,.._._.. ...-o .- IIlIt-VII - .... . ........ ... ............. .. ,,_ ... _,., ...-... IIIIIIII u ....... ... ...- . .. ... ,...... . .__._.,.... , .... ...... ll . ....... . .... ............. . ._‘,.. ,. ,.... .... nu.n ... . . n..--.... .- .......»...»..... - -- 'I-I 1" . ... .. . .....--.. _._._‘_, ._. III-u. Ila III I .. ..-.. ... _ ,,_,_.. .oa...-o sun. I. III . . ... . . _,, ,,,_.‘ .......... u. u a. . ... ._.. . ,..,«.,,_,,__ .. ,.... , .. u . u.- ...... .. ,. 4, . -- -u ...-.... an. IIII'IIII .- ._‘,‘ ... ... ............-...-... .... .. .. .............-- ...... , .,_ ..., ..... . 7"».Illvil no. "' . I . . v w I - I I SPATIAL DIFFUSION OF HYBRID MAIZE mm 4}; 9 6'3' 6 We 01% I0 I6.0 % I6.I $ I0 50.0 i LEVEL II. LEVEL III. I‘ll-III. IIII‘VIII III nun-.101: LEVEL V. LEVEL VI. LEVEL IV. 50.I % Io 84.0% 84.I S I0 99.0 % 99.I % I0 IOO % KISII- 0) I968 ............. I I I n I l u u ............ ........... c) I97! 187 IO SPATIAL DIFFUSION OF HYBRID MAIZE Milo: :5 1 {gm . § 1.0 3 IO 5 IO I5 Kibmotm KISII. b) l970 I l I I I I I I I I 2 3 4 5 6 7 8 9 LEVEL I. NO ADOPTION LEVEL n. on. IO l6.0 as LEVEL m. I6.I as to 50.0 74. LEVEL IV. 501% to 84.0% :53; LEVEL V. 84.I *x. to 99.0 as LEVEL VI. 99.I as to ICC as 188 By 1962 an eastern and a western section could be identified as a result of the coalescence of most of the Level II areas. In the intervening two years from 1960 to 1962, the Level II adoption areas expanded to seven. The map for 1962 still represents Stage I because the rapid outward movement had not begun. By 1964 an expan- sion of Level III areas took place and three areas of Level IV adoption developed (H7.5-V6.0; H8.S-V4.0; and H2.0—V3.5). By 1966 the eastern and western parts were still separate because only moderate expansion took place at all levels and in all locations. Stage II, the diffusion stage, is presented only in 1968. Major expansion, coalescence and intensifi- cation took place in the preceding two years causing the merger of most areas of Level III adoption into one region. Also, Level IV emerged as an important area, as well as the development of three Level VI positions (H8.0-V4.0; Hll.0-V6.0; and H7.5-V6.0). This time period is classi- fied as Stage II on the basis of the rapid outward move- ment, but regional contrasts are strengthened over Stage I due to the development of areas with complete acceptance of hybrid maize. These contrasts represent the intensi- fication of nodes that existed in earlier time periods. Stage III, the condensing stage, is found in 1970 and 1971. Only one area with no adoption of hybrid maize exists at the latter time (H8.0-Vl.0). Level IV and above 189 occupies most of the study area, and Level V adOption has merged to cover an extensive region. Fully thirteen different places have achieved total acceptance of hybrid maize. In the 1970 to 1971 time interval, the area occu- pied by Level V enlarged considerably and the Level VI places expanded. If the places that are at Level VI adOption continue to expand at the present rate it will only be three or four years before most of the study area has achieved complete adoption. Thus, Stage IV, the saturation stage, will probably be quite noticeable by 1974, and the entire study area should reach near 100 percent use of hybrid maize by 1975 or 1976. Figure 51, showing the percent of sampling areas at each adoption level presents the same basic pattern as the other innovations being examined, only in a more com- pressed form. The area with no adoption forms the "S" curve in reverse due to the slow start of the spatial dif- fusion process, followed by the rapid outward movement of the diffusion wave, and finally a slowdown as the final few areas show a reluctance to adopt. Levels II and III increase to a peak in 1968, the year of greatest expansion, and decline rapidly. Levels IV and V accelerate rapidly to a 1970 peak and begin to fall. Only Level VI continues to show signs of increasing in the future. 190 Figure 51 HYBRID MAIZE ADOPTION LEVEL PERCENTAGES de I no adaption 11 0| - Ill l6.| 1V 50.I V 84.I V1 99.I -|6096 - 50.0% -' 84.0 °/o - 99.0% - IO0.0°/o Percent r IOO F90 ‘80 ~70 *60 >50 »40 ::-30 #20 191 Generalizations on the Diffusion Process The spatial diffusion of agricultural innovations in Kisii District follows the Hagerstrand typology of spatial innovation reasonably well, with two exceptions. Stage I, the primary stage, is characterized in both southern Sweden and Kisii District by a few isolated locations where a small percent of the farmers have accepted the innovation in question. Stage II, the dif- fusion stage, is one of the exceptions. In the Hagerstrand typology this stage exhibits the development of new dif- fusion nodes and the leveling of regional differences. In Kisii new nodes appear but regional differences are strengthened rather than leveled. New and old nodes in- crease in adoption percentages until they become peaks of very high adoption that stand out above the surrounding area. Stage III, the condensing stage and the second exception, is characterized by Hagerstrand as exhibiting equal increases in all areas. In Kisii this stage pri- marily involves infilling between the adoption peaks and the expansion of peaks do form plateaus. Also, the gradient of the diffusion wave's outer edge becomes steeper. Stage IV, the saturation stage, is not found in Kisii District for any of the innovations. For coffee and pyrethrum, the location of market- ing outlets is important in determining the spatial dif- fusion pattern. The construction of a marketing site 192 seems to encourage farmers in the immediate area to adOpt that particular crop. For passion fruit and tea, that have only pick-up points, which can be easily moved, the market location influence cannot be detected. Grade cattle and hybrid maize do not enter the marketing pic- ture in the same manner so c00perative marketing sites are not important. There is an obvious tendency for the peaks of high adoption to merge. Each individual adoption peak will generate its own large-scale regional information field in much the same way an individual will generate his own Personal Information Field. Because of the close juxta- position of the adoption peaks the large-scale regional information fields will tend to overlap, therefore the farmers in between will have more information about the innovation at their disposal and will have a greater probability of adOpting. Hagerstrand's research on innovation diffusion in southern Sweden revealed that the same areas within the region of study were repeatedly the starting point for new innovation waves (Hagerstrand, 1953, p. 293). The area of multiple origin is characterized by a pOpulation with high receptivity to innovation. In the Kisii study such innovation centers were not found. Each of the six innovations had its own set of starting points. Three indicators, pyrethrum, tea and passion fruit, started in 193 the eastern part of the study area, but not in exactly the same locations. The lack of innovation centers would suggest that the important control on the location of original introduction is ecological rather than social. Thus, any innovation could be introduced in any part of an area with homogeneous ecological conditions with an equal chance of success. In a region with a non-uniform distribution of receptivity to innovation it would be important to precisely locate the innovation centers, but in an area of homogeneous receptivity such as Kisii it becomes less important. Factor Analysis and Multiple Regression and Correlation Two separate factor analysis routines were exe- cuted. The first was with all 57 variables measuring innovation, socio-economic, demographic and locational characteristics. The second was with 31 variables, all innovation measures having been removed. The larger group was analyzed to determine the kinds of relationships that exist between the innovation measures and the other socio- economic, demographic and locational variables. The second, with no innovation measures, was used to generate factor scores for inclusion in a series of multiple regression and correlation models. 194 Factor Analysis A, 57 Variables The extraction of 13 factors explained 77.73 percent of total variable (see Table 11). The 13 factors are as follows. Factor I-A, East-West Dichotomy (13.28 percent of variance). This, the most complex of all the factors, is primarily a measure of location, as all of the measures loading on it are notable for their spatial distribution. Loading in conjunction, from the highest in descending order are: east-west coordinate, elevation, the two pyrethrum variables (mean year of adoption and percent of farmers raising) and distance from Kisii town. Load- ing in opposition to the above group of variables are the three coffee variables (percent growing, mean number of acres per farmer and mean year of adoption) and precipi- tation. Thus, as one moves from west to east across the study area the percent of farmers growing coffee declines, precipitation decreases, elevation increases and the percent of the farmers growing pyrethrum increases. Also, the mean acreage per farm planted to coffee declines as the mean acreage per farm planted to pyrethrum increases. Obviously the distance from Kisii increases as does the numerical value of the east-west coordinate. Factor II—A, Grade Cattle (9.49 percent of vari- ance). Because of the low relationship on the intercorre— lation matrix between the grade cattle measures and all 195 TABLE ll.--Factor Analysis A, 57 Variables (Total Explained Variance: 77.73%). Factor (Variance) Highest . Variables Loading Communality I-A. East-West Dichotomy (13.28%) East-West Coordinate -.88 88.90% Elevation -.75 71.70 Annual Precipitation .63 66.64 Distance from Kisii -.67 80.04 Mean Year Coffee Adoption .88 88.07 Mean Coffee Acres per Farm .91 93.13 Percent Coffee Adoption, 1971 .93 93.96 Mean Year Pyrethrum Adoption -.71 82.54 Mean Pyrethrum Acres per Farma -.43 79.02 Percent Pyrethrum AdOption, 1971 -.71 71.42 Weeding Practices Index .44 76.38 Fertilizer Practices Index -.37 68.24 Mean Year Tea Adoptiona -.43 60.44 Percent Tea Ad0ption, 1971 —.45 72.15 II-A. Grade Cattle (9.49%) Mean Year Grade Cattle AdOpted .82 80.51 'Mean Number Grade Cattle per Farm .86 86.70 Mean Number Grade Cows per Farm .87 86.54 Grade Cow Practices Index .88 86.07 Percent Farms with Grade Cattle, 1971 .77 77.95 Local Cow Practices Index .50 71.46 III-A. POpulation Characteristics (8.71%) Mean Total Farm Population .93 95-47 Mean Number of Adults on Farm .84 85-43 Mean Number of Children on Farm .85 83-64 Mean Number of Family on Farm .92 90.54 Mean Number of Children in Primary School .67 79.48 Mean Number of Pieces of Land .53 67.07 IV-A. Cosmopolite Characteristics (8.57%) Mean Number Agent Initiated Contacts .82 83.93 Mean Number Client Initiated Contacts .73 87.23 Mean Number Demonstrations Attended .77 82.32 Mean Number Demonstrations Found Useful .78 83.27 Mean Number Farmer Training Center Courses .36 60.63 Mean Number Formal Memberships per Head .65 70.45 Mean Communication Index .74 85.73 Mean Education-Literacy Index .47 74.47 196 TABLE 11.--Continued. Factor (Variance) Highest Communality Variables Loading V—A. Housing Characteristics (5.72%) House Type Index .66 69.63% Household Facilities Index .71 79.54 Household Possessions Index .73 83.26 Mean Year Hybrid Maize Adoption .53 66.44 VI-A. Intensity Index (5.47%) Mean Acres Hybrid Maize per Farm .82 73.67 Mean Acres Pyrethrum per Farm .52 79.02 Tick Control Practices Index .79 77.01 Income from Cash Crops .56 68.96 VII-A. Keroka Concentration (5.46%) Mean Year Passion Fruit Adoption -.85 78.97 Mean Passion Fruit Acres per Farm -.88 85.00 Percent Passion Fruit Adoption, 1971 -.85 82.08 P0pu1ation Density -.54 67.07 VIII-A. Tea (4.07%) Mean Year Tea Adoption .44 60.44 Mean Tea Acres per Farm .73 77.23 Percent Tea AdOption, 1971a .37 72.15 Mean Age Head-of-Household .58 66.75 IX-A. Progressiveness Indices (4.04%) Crop Progressiveness Index .80 74.99 Overall Progressiveness Index .59 80.27 X-A. North-South Dichotomy (3.77%) North-South Coordinate .57 83.12 Mean Acres per Farm .75 76.88 XI-A. Information Availability (3.71%) Mean Number Demonstration Plots Visited .64 64.13 Mean Number Formal Offices Held .77 77.14 XII-A. Hybrid Maize (2.91%) Percent AdOption Hybrid Maize, 1971 .68 79.96 Mean Acres Local Maize per Farm -.48 73.14 XIII-A. Traditional-Modern Dichotomy (2.52%) Mean Number of Local Cows per Farm .52 61.27 Mean Number Children in Higher School per Household -.56 72.16 aSecond highest loading. 197 other variables the former loaded together on the same factor. Another measure loading on this factor is the local cow practices index, meaning that the farmers who tend to adopt grade cows will also take better care of their local cows. The moderately high loading of the variable measuring the number of children in higher school indicates the relative wealth of those farmers owning grade cows. Factor III—A, Population Characteristics (8.71 percent of variance). Measures of population, that is, number of people on the farm, the number of adults and children, the number of the family members residing on the farm, and the number of children in primary school all load high on this factor. Moderate loadings are achieved for the age of the head and the number of pieces of land. These latter two are logical because the older farm heads would tend to have larger families and also more pieces of land since they probably inherited the land some time ago when more land was available; there- fore some of them would have more than one parcel of land. Factor IV-A, Cosmopolite Characteristics (8.57 per- cent of variance). The cosmopolite factor is a measure of information-seeking behavior. All of the indicators of change agent contract, formal membership behavior, attend- ance at demonstrations, communications behavior, education and literacy, and Farmer Training Center course attendance 198 load highly on Factor VI. The north-south coordinate loading is in opposition, indicating that as one moves southward in the study area there is a slight tendency for the information available to farmers to diminish. Factor V-A, Housing Characteristics (5.72 percent of variance). The three measures of housing--the indices of house type, facilities, and possessions-~load together on this factor due to the relatively small variability from one part of the study area to another. The variable measuring the mean year of hybrid maize adOption attains its highest loading on this factor, indicating a slight tendency for peOple with better housing to adOpt hybrid maize earlier. However, due to little variability in adOption dates from one place to another, this variable is of minimal importance to the makeup of the factor. Factor VI-A, Intensity Index (5.47 percent of variance). This factor is an indicator of intensity of use due to the high loadings of hybrid maize and pyrethrum acreage, and the high use of tick control measures. Farmers who have adopted the planting of hybrid maize in large amounts also tend to own more local cattle and use approved techniques of tick control. Also, there is a slight tendency for these farmers to have higher incomes. Factor VII-A, Keroka Concentration (5.46 percent of variance). All of the variables measuring passion fruit and pOpulation density load very high on this 199 factor. This title was given because there is a very strong concentration of passion fruit growing in the area of Keroka and that same part of the district experiences the highest population density, thus the factor serves as another locational measure. Factor VIII-A, Tea (4.07 percent of variance). The tea acreage variable is most closely related to this factor, with the mean year of tea adOption and the percent of farmers growing tea of lesser importance. Also achieving its highest loading is the mean age of the homestead head, indicating a possible relationship between older farm heads and tea adOption because older farmers have, perhaps, had a greater Opportunity to have worked in the tea estates around Kericho, and therefore, because of that experience, are quicker to adOpt tea as a cash crOp. Factor IX-A, Progressiveness Indices (4.04 percent of variance). The crop progressiveness index, the mean number of years before 1971 that the crop innovations were adOpted, and overall progressiveness index, which is the crop progressiveness index plus the number of years prior to 1971 that grade cattle were adopted, align most closely with this factor. The lack of relationship between this factor and the variables measuring the mean year of adOption for each innovation indicates that each is of relatively small importance to the progressiveness indices. 200 Factor X-A, North-South Dichotomy (3.77 percent of variance). The total number of acres per farm and the north-south coordinate (which increases in numerical value as one travels southward in the study area) are logically related to each other for the pOpulation density decreases somewhat as one travels in that direction. This relation- ship, however, should not be overemphasized because of the low explanation provided by this factor. Factor XI-A, Information Availability (3.71 per- cent of variance). The two variables loading on this factor indicate a relationship between the number of offices held and visits to demonstration plots. The in- effectiveness of demonstration plots in encouraging adoption is evident from the fact that none of the inno-V vation measuring variables load highly on this factor. Factor XII—A, Hybrid Maize (2.91 percent of variance). This variable is not related to either the mean number of acres of hybrid maize grown by each farmer or to the mean date of adOption. Also, and quite signifi- cant, it does not relate strongly with any other variables. This would point to the lack of relationship between the percent of farmers raising hybrid maize and other vari- ables. Factor XIII-A, Traditional-Modern Dichotomy (2.52 percent of variance). This dichotomy indicates, possible, that peOple who raise more local cows are poorer and thus 201 cannot afford to send their children to high school. Or perhaps the greater numbers of local cows refers to tra- ditionalism, so that secondary school is less important to such people. As the explanatory power of this factor is very low, these tendencies should not be regarded as important. There is one principal conclusion that can be drawn from this factor analysis. For the most part the variables measuring the innovations either factored out as independent of everything else, or they loaded up in con- junction with locational variables due to the spatial distribution they take (i.e., Factors I, II, III, IX and XIII). When the innovation variables did load on the same factor as non-innovation or non-locational variables three characteristics stood out. First, the factor pro- vided only about 4 to 6 percent explanation of variance. Second, the loading of the innovation variable was with only one exception, over .52. Third, the relationship is such that one cannot ascribe cause and effect, only a relationship. The distribution of socio-economic and demographic variables therefore bears little relationship to the distribution of innovation adoption as it existed in 1971. If one can assume the present is reasonably similar to the past, then the lack of relationship between the variables measuring the innovations and the 202 other variables becomes important. The distribution of socio-economic and demographic characteristics is sufficiently random so that every area has about an equal number of innovators and laggards, thus no area can be expected to adOpt sooner than others on this basis. We can thus conclude that the primary influences on the present spatial pattern are the random (or planned) placement of the original adopters and the spatial dif- fusion process generated on that basis, modified by physical, cultural and economic (market) features. Factor Analysis B, 31 Variables A second factor analysis routine was performed on the data matrix with all innovation measures removed. From the remaining 31 variables eight factors explaining 73.48 percent of overall variation were extracted. These eight factors correspond with those derived in the first factor analysis routine according to Table 12. Rather than discuss each of the eight factors separately only the ways in which the second analysis differs from the first will be pointed out (see Table 13). The elimination of factors I-A, grade cattle; IX-A, tea; and XIII-A, hybrid maize, is due to the exclusion of the innovation variables from the routine. Factor IV-B (East-West Dichotomy) differs from Factor I-A only in the removal of the variables measuring coffee and pyrethrum. Factor VII-A, the Keroka Concentration, is primarily made 203 TABLE 12.--Factor Analysis B, 31 Variables Used (Total Explained Variance: 73.48%). Factor (Variance) Highest Communality Variables Loading I-B. Population Characteristics (15.15%) Mean Age Head-of-Household .52 59.51 Mean Total Farm Population .92 92.96 Mean Number of Adults on Farm .84 84.60 Mean Number of Children on Farm .87 86.20 Mean Number of Family on Farm .91 87.27 Mean Number of Children in Primary School .68 76.62 II-B. CosmOpolite Characteristics (12.00%) Mean Number Agent Initiated Contacts -.86 83.06 Mean Number Client Initiated Contacts -.88 82.98 Mean Number Demonstrations Attended -.63 83.35 Mean Number Demonstrations Found Useful -.62 82.42 Mean Number Formal Memberships per Head -.67 68.43 Mean Communication Index -.61 84.55 III-B. Information Availability (10.43%) Mean Number Demonstration Plots Seen .77 65.02 Mean Number Formal Offices Held .65 64.22 Mean Education-Literacy Index .63 78.02 IV-B. East-West Dichotomy (9.66%) East-West Coordinate .88 84.95 Elevation .88 88.42 Annual Precipitation —.67 75.77 Distance from Kisii .68 70.58 V-B. Housing Characteristics (9.45%) House Type Index .77 73.22 Household Facilities Index .57 70.66 Household Possessions Index .70 76.49 Mean Number Children in Higher School per Farm .71 56.52 Mean Number Farmer Training Center Courses Attended .45 54.80 204 TABLE 12.-—Continued. Factor (Variance) Highest Variables Loading Communality VI-B. North-South Dichotomizer (7.15%) North-South Coordinate .84 79.54 Annual Precipitationa -.43 75.77 Population Density -.70 54.02 VII-B. Land Holdings (5.08%) Mean Total Acres .72 64.79 Mean Fertilizer Index -.45 59.52 Mean Income .55 63.65 VIII-B. Weeding Practices (5.03%) Weeding Practices Index .71 70.74 aSecond highest loading. 205 TABLE l3.--Comparison Between Factor Analysis A and B. Factor Analysis "A" (77.73%) Factor Analysis "B" (73.48%) I. East-West Dichotomy (13.28%) IV. East-West Dichotomy (9.66%) II. Grade Cattle (9.49%) III. Population Characteristics I. Population Characteristics (8.71%) (15.15%) IV. Cosmopolite Characteristics II. Cosmopolite Characteristics (8.57%) (12.00%) V. Housing Characteristics V. Housing Characteristics (5.72%) (9.45%) VI. Intensity Index (5.47%) VII. Keroka Concentration VI. North-South Dichotomy (5.46%) (7.15%) VIII. Tea (4.07%) IX. Progressiveness Index (4.04%) X. North-South Dichotomy VI. North-South Dichotomy (3.77%) (7.15%) XI. Information Availability VII. Land Holding (5.08%) (3.71%) III. Information Availability (10.43%) XII. Hybrid Maize (2.91%) XIII. Traditionalism-Modernity Dichotomy (2.52%) VIII. Weeding Index (5.03%) 206 up of measures of passion fruit adoption, but also con- sists of measures of population density and north-south measurement. The elimination of passion fruit measures allows the north-south dichotomy to be strengthened on the factor. Total acreage is the main component of Factor X—A, the North-South Dichotomy, and the coordinates are the second most important variable. Therefore, Factors VI-B, North-South Dichotomizer and VII-B, Land Holding Characteristics, correspond to that factor. The Intensity Index, Factor VI-A, does not appear in Factor Analysis B due to the elimination of the main variables, dealing with hybrid maize, pyrethrum and tick control. Likewise, the Progressiveness Indices, Factor IX-A, were removed. Factor XIII-A, the Traditionalism- Modernity Dichotomy, disappeared because the variables load on a new set of factors (such as number of children in high school) or were removed (such as the numbers of local cattle). Lastly, the weeding index, which in Factor Analysis A loaded moderately with several of the innovations, factors out by itself because it no longer has anything to relate to. Factor Analysis B Related to Innovation Adoption The final step in the statistical analysis related the factor scores of the eight factors derived from the 31 non-innovation variables to the percent of adoption in 207 1971 for each of the six innovations. This was done by using a least squares multiple regression deletion routine with each innovation, in turn, as the dependent variable and the factors as the independent variables. The factors relating to the innovation distribution pattern in 1971 at a significance level of .05 or better are shown in Table 14. Beta weights are included to indicate the relative contribution of each variable to the amount of explanation provided by the combination of variables (the sum of squared beta weights is equal to R2). It is significant that the East-West Dichotomy was the most im- portant variable in four out of the six cases. This factor simply is a measure of those items of location, i.e., coordinates, distance from Kisii, rainfall and elevation, that differ from the eastern to western parts of the study area. Thus, no explanation of cause can be offered by this factor. Housing characteristics were most closely related to the spatial distribution of tea and grade cattle adoption. This leads to the question of cause and effect. Because they adopted these innovations are the farmers better able to afford higher quality housing, or due to more wealth, as manifest by their better housing, are the farmers better able to afford to adOpt these relatively expensive innovations? The cosmopolite factor seemingly should correlate with the rate of adoption but it does not. In fact, 208 TABLE 14.--Multip1e Regression and Correlation Variables Significant at .05. Dependent Variable (R2) . Bet Wei hts Independent Variables a g Coffee Adoption, 1971 (R2 = 0.73) Factor II, Population Characteristics -.14 Factor III, East-West Dichotomy -.81 Factor VII, Weeding Index .20 Factor VIII, Housing Characteristics -.12 Pyrethrum Adoption, 1971 (R2 = 0.47) Factor I, Information Availability .19 Factor III, East-West Dichotomy .63 Factor V, North-South Dichotomy .19 Tea Adoption, 1971 (R2 = 0.38) Factor I, Information Availability .24 Factor III, East-West Dichotomy .37 Factor IV, Cosmopolite Characteristics .19 Factor VIII, Housing Characteristics .38 . . . 2 PaSSion Fruit AdOption, 1971 (R = 0.11) Factor III, East-West Dichotomy .24 Factor V, North-South Dichotomy -.23 Grade Cattle AdOption, 1971 (R2 = 0.46) Factor II, Population Characteristics .21 Factor IV, Cosmopolite Characteristics -.29 Factor VIII, Housing Characteristics .58 Hybrid Maize Adoption, 1971 (R2 = 0.20) Factor I, Information Availability .22 Factor III, East-West Dichotomy .39 209 there is either an extremely low correlation, as with passion fruit adoption at .0056, or a negative relation- ship, the highest being with grade cattle at -0.2920. This would seem to indicate that the more cosmopolite farmers are not likely to adOpt grade cattle. But that proposition does not necessarily hold true, for the places where grade cattle were first introduced will have a pro- found influence on the subsequent adOption pattern, and those places are not necessarily related to the distribu- tion of cosmopolite characteristics. Some Internal Interrelations A careful examination of the intercorrelation matrix reveals an interesting relationship for each of the six innovations between the mean year of adoption, the mean number of acres (or cattle) per farm, and the percent of farmers raising the item in 1971. The earlier the mean year of adoption, the greater number of acres (or cattle) each farmer will cultivate (or keep). The earlier an area adopted the greater the percentage of farmers raising the item. The higher the percentage of farmers raising the item in an area the higher the number of acres cultivated or cattle owned. Thus the original nodes will have the highest level of adoption, in terms of the percent of farmers adOpting, and the greatest intensity, measured by the number of acres or the number of cows. 210 This relationship holds very well for coffee (see Table 15), passion fruit and grade cattle. For pyrethrum it holds fairly well, except for the relationship between acres per farm and the mean year of adOption. Tea ex- hibits a modest relationship between all three variables. So, for all of the cash crops and grade cattle the re- lationship holds to varying degrees. But for hybrid maize there is none. This can, perhaps, be explained by the fact that hybrid maize is not normally a cash crop, and as such market considerations do not enter into the farmers' calculations, therefore he will not be as cautious about adopting. Also, it has the most rapid rate of adOption, because it is a relatively simple change from local maize, so the relationships do not hold. In spite of the rapid rate of adOption the correlation of .23 between the percent of adoption in 1971 and the mean year of adoption suggests that the original nodes have the highest levels of adoption. The relationship between the mean year of adoption, and the percent of adoption in a sampling area exists primarily because of the friction of distance. Awareness knowledge, that is, knowledge about the existence of the innovations in question is very wide- spread. But specific "how-to" knowledge is confined to those areas having greater experience with the inno- vation, therefore more farmers have adOpted and each 211 TABLE 15.-~Internal Intercorrelations, Mean Year of Adoption, Mean Acres per Farm, Percent Adoption 1971. Year Mean Percent Year Mean Percent Adopted Acres 1971 Adopted Acres 1971 COFFEE PYRETHRUM Year Adopted —- -- Mean Acres .91** -- .39** -- Percent 1971 .86** .90** -- .61** .55** —- TEA PASSION FRUIT Year AdOpted -- —- Mean Acres .54** -- .79** —- Percent 1971 .51** .60** -- .68** .74** -- GRADE CATTLE HYBRID MAIZE Year AdOpted -- -— Mean Acres or Number Percent 1971 .74** -- .67** .69** -- .08 -- .23* -.O6 -- *Significant at .05 level. **Significant at .01 level. 212 farmer uses it more intensively. Knowledge of exactly how one goes about getting material to start raising the crOp (or how to acquire grade cattle), knowledge about horticultural methods, harvesting techniques and market- ing procedures are progressively more important as one goes from innovator to laggard on the adOpter category continuum (Rogers, 1971, pp. 259-261). Awareness knowl- edge depends on large numbers of farmers raising the item so that non-adOpters can observe and ask questions about specific procedures. Thus the level of "how-to" knowl- edge varies greatly over space. As an indication of the spatial behavior exhibited by Gusii farmers consider the following. The field survey data indicate that 61 percent of the farmers in the study area never travel to a market, for either buying or sell- ing, that is over 4 km. (2.5 miles) away from their farms. Most of the time people travel to the nearest market, so 91 percent of the farmers report that they normally travel to a market located within 4 km. (2.5 miles). Also, the greater the population density of the area the closer the spacing of markets, therefore people travel even shorter distances. For example, in Kitutu East, the most densely populated location in the study area, 61 percent of the farmers travel to markets located within 2 km. (1.25 miles) of their farms. Using distance traveled to markets as a surrogate measure, one could assume that farmers would travel much shorter distances to visit another 213 farm for the purpose of acquiring Specific "how-to" knowledge. Strong distance friction would cause a very small Personal Information Field. Distance friction is, of course, caused by the fact that people walk most of the time. Virtually no farmers in the district have auto- mobiles, few have bicycles, and bus transport is oriented to the larger towns only, the result being that most peOple walk on journeys of less than, say, a one-hour walk. In spite of the high local relief the area oper- ates somewhat like an isotropic plain in that foot travel is nearly omnidirectional. The maze of footpaths is interrupted only by the steepest of relief features. While awareness knowledge can travel from adOpter to non-adOpter and subsequently to other non-adOpters, specific "how-to" knowledge generally moves from adOpter to non-adOpter only. Thus the lower level of knowledge is relatively unaffected by distance, but the higher level is very much retarded by distance, because "how- to" knowledge can diffuse spatially only as fast as adOption spreads. The Accelerating Pace of Change A question that could be asked in the context of this study is: "Has the pace of change increased over the years?" One might ask such a question after examining the graph showing the cumulative percent of adoption 214 for each of the six innovations. It appears that 'there is a steady increase in the adoption rate when one considers first coffee, then pyrethrum, and finally hybrid maize. Thus, over time, the length of the bottom end of the logistic curve shortens and the steepness of the central part of the curve increases. So from this inquiry one would conclude that the pace of change is quickening. But when one considers the three newest inno- vations, tea, grade cattle and passion fruit, the picture becomes murky. For grade cattle and passion fruit, in particular, the lower end of the logistic curve seems to be even less steep than for the innovations introduced earlier. However, there is a fallacy in considering innovations separately, because innovativeness is a combi- nation of all of these innovations, and therefore they should be considered in concert (Figure 52). Simply adding together the percents of adoption for the six innovations during each time period will give a crude measure of the pace of change. This composite adoption index has a theoretical maximum value of 600, assuming that all farmers within the study area were to adopt all six innovations. Considering five-year incre- ments from 1940 we find that there has indeed been a con- stant acceleration in the pace of change. From 1940 to 1950, a ten-year period, there was only a change of 2.7 215 Figure 52 COMPOSITE ADOPTION INDEX (Malfi'iixeoo) r250 ~225 *200 475 H50 fi25 400 '75 “50 '25 l950 I960 I970 216 in the index. Between 1950 and 1955, only five years, the index increased by 17.6, or 6.5 times as much as the previous time period. During the next five years, from 1955 to 1960, the increment was 44.0, or still over twice as much as the previous time period in spite of the in- creased size of the base. From 1960 to 1965 the rate of increase diminished somewhat as the growth was only 42.4. So, in spite of a declining relative rate of change during this time period, the overall change as measured by these innovations is still several times as great as it was during the 1940 to 1950 period. After 1965 the pace of change again quickens. The amount of increase from 1965 to 1970 is 120.1 on the composite adoption index for this five-year period. This is nearly three times the rate of increase for any other comparable five-year period. If the growth in the index from 1970 to 1971 is projected out to 1975 the rate would diminish only slightly from the 1965 to 1970 pace. It is evident that one must consider the alter- native items available for adOption, for given a selection of innovations the rate of adoption for any one item will most likely be slower. So if farmers, for example, did not have the alternatives of grade cattle or passion fruit adOption during recent years but only the adoption of tea, the adoption rate for tea would have been much higher. One can only reach the inescapable conclusion 217 that the rate of change has acceleratec considerably from the decade of the 1940's to the 1960's. The overall rate of change in the future seems to be more dependent on the number of innovations available for adOption rather than the speed with which one single innovation is accepted. The Composite Adoption Index leads back to a question raised in the introduction to this dissertation. That is, does an accelerating pace of change, as measured by innovation adoption, lead automatically to an improve- ment in family welfare? If, for example, a farmer raises commercial crOps to the exclusion of food crops, and then purchases an inferior diet, the family suffers because of the innovation. On the other hand, if the adoption of hybrid maize allows a farmer to produce an adequate amount of food on less land, and then uses the extra land to raise a commercial crop to earn money to purchase high protein supplements to the family diet, then the family is better off. The contribution of innovation adoption to improved family welfare is not measured here. The Composite AdOption Index only examines the pace of change, and not the consequences of change. Generalizations About the Factor Analysis __Efid_fiegression andTCBrrelatibn Models Factor analysis of the 57 variables measuring the innovations plus socio-economic and demographic variables showed that the innovations were quite unrelated to the 218 other non-innovation measures. The innovation measures either factored out by themselves or on the same factor as the locational measures. There was little relationship between innovation measures and the measures of communi- cation behavior. Analysis indicated a strong interrelation- ship between the various measures of communication behavior. Variables measuring pOpulation are quite similar so they logically factored out together. Other indicators such as housing characteristics, and the number of children in high school are measures of wealth, so they emerge on the same factor. The second factor analysis, with the inno- vation measures removed, reveals that the basic dimensions are much the same, minus, of course, the innovations. For example, the factors measuring only an innovation were no longer present, and the geographic factors were minus the innovation variables. The basic reason for the small amount of expla- nation provided by the variables used seems to be one of scale. When considering the study area as a single unit the patterns found match those found in other research. That is, the growth curves of adOption over time are completely normal and understandable and the relationship of the socio-economic and demographic variables to each other is completely logical. A problem, however, is encountered when one tries to relate the socio-economic and demographic variables for the entire study area to 219 the spatial diffusion pattern. An attempt at prediction of the spatial diffusion of adoption based on a knowledge of the spatial pattern of socio-economic characteristics is not substantiated because of a conflict in scale. Throughout the study area there is a relatively uniform distribution of socio-economic characteristics, so no one area stands out as being more likely to adOpt earlier. Therefore the place of original introduction, whether random or planned, is the principal determinant of the early pattern of adOption. So a person residing at the locus of original introduction with the characteristics of a laggard may adopt the innovation at, say, time T plus 3, and another individual, residing a long distance away, with the characteristics of an innovator or at least an early adOpter, may adOpt at time T plus 6. Thus any attempt to predict when a person is likely to adopt must include the fact of propinquity to other adopters. CHAPTER VI SUMMARY, CONCLUSIONS AND RECOMMENDATIONS Summary Only in recent years has the explosive nature of African population growth come to light. Kenya, for example, is increasing in numbers by 3.3 percent annually, a doubling rate of 21 years. An exponentially growing pOpulation such as this puts severe strains on the govern- ment to provide the services needed by the people, but more importantly it generates demand for meaningful employ- ment at a faster rate than the economy can supply. Closely associated to the employment problem is that of migration. Wage rates are several times as high in the urban areas as opposed to rural areas, so large numbers of people migrate to the city to seek employment. Even though the probability of securing a job is low, the dramatic in- crease in income makes such an attempt rational. As a result of such conditions Nairobi has been growing by 9.5 percent annually and the African pOpulation of the city, that continues to make up a larger portion of the total each year, is expanding by 14.5 percent annually. 220 221 During most of the 1950's and into the 1960's develOpment planners generally subscribed to the theory of advancement through industrialization, but unfortunately, it has not generated adequate employment opportunities. Manufacturing as an employment category has actually de- creased over the past decade in Kenya. The so-called modern sector (i.e., in general terms, wage employment) has expanded slightly only because of a major increase in government employment. The output of industry, as mea- sured by the value of the goods produced, is expanding by about 6 percent annually, in spite of declining employ- ment. The reason is that owners seek to maximize returns per worker and per unit of investment. No mineral deposits of consequence have been found in Kenya. Agriculture, therefore, is, and will continue to be, the most important sector for employment and export earnings. About one-third of the Gross Domestic Product consists of agricultural products, about 60 percent of commodity exports are raw or processed agricultural goods, and most importantly, three-fourths of the population derives its livelihood from the land. The fastest growing portion of the Kenya economy is tourism. The annual increase in the number of foreign visitors to Kenya in recent years is approximately 25 percent. The projected foreign exchange earnings for 1974 are KB 37 million (U.S. $95 million) which will 222 make tourism the single most important earner of foreign exchange, surpassed only by export categories such as agriculture and manufacturing. In terms of employment, tourism is not very important as the projected 1974 workforce is only 40,000, or slightly less than 1 percent of the total. In agriculture there are two general ways to attack the dual problem of production and employment; opening new land and intensification of production on currently used land. The former has generally been abandoned due to a variety of reasons, the most important being the high cost of settling each family. Intensifi- cation involves the application of new technology, or simply old technology in a more concentrated form, to increase output per acre. Increased output per unit of labor is important, but since employment generation is a principal goal, efficiency takes a secondary position. In order to intensify the farmer must accept a combination of new methods, technology and crops. The rate at which these innovations are accepted by the farmers largely determines the success of the intensification pro- gram. In order to understand the process of innovation adoption, with the ultimate aim of facilitating the diffusion of innovations, a whole area of academic in- quiry has developed. Several broad categories of investi- gation are undertaken by diffusion researchers. They are: 223 (l) The communication network through which the farmer receives information about the innovations. (2) The personal characteristics of peOple who adOpt at different times. (3) The characteristics of the innovations that influence the rate of adOption. (4) The growth of adoption over time. (5) The impact of organizations Specifically engaged in the introduction of change. (6) The consequences, to the individual and to society, of innovation adoption. Geographers are concerned with the spatial aspects of the diffusion process, that is, the manner in which the innovation spreads across the landscape and the combi- nation of forces that facilitate, retard and modify that spread. Hagerstrand postulates four separate stages in which the innovation is introduced into an area, spreads outward rapidly, intensifies, and finally increased slowly up to the maximum acceptance level. The central feature of the spatial diffusion model is the Personal Information Field. It refers to the de- clining probability of communicating about the innovation to another person as distance increases. Personal Infor- mation Fields can and do overlap with one another as the distance between adopters lessens. Therefore, in areas where the adOption rate is high the numerous overlapping Personal Information Fields will create a more intense Regional Information Field. The localized generation of 224 specific information (i.e., information of the type needed for a person to adOpt) about a potential innovation is an important factor explaining the spatial diffusion process. Kisii District, in southwestern Kenya, is 2217 square kilometers (856 square miles) with a population of about 675,000 people, most of whom are Gusii. The economy is based on a combination of small-holder cash crOp and subsistence agriculture. A variety of food crOps are grown for local consumption and several cash crops, such as coffee, tea, pyrethrum and passion fruit, are grown for export. There is little other economic activity in the district, save the provision of a small quantity of consumer goods. The district ranges in elevation from about 1525 meters (5,000 feet) to over 2135 meters (7,000 feet), is cool year around, receives abundant rainfall that is reasonably well distributed temporally and spatially, and has good soils throughout. The landscape that results from the combination of intense cultivation, dense popu- lation, high altitude and abundant rainfall is not one that fits the typical stereotype of Africa. Perhaps the most serious problem facing Kisii District is population growth. Currently portions of the district have over 580 persons per square kilometer (1,500 per square mile), and even the most sparsely pOpulated areas, with the exception of the former settlement area, 225 have over 135 persons per square kilometer (350 per square mile). Add to this the annual growth rate of 3.6 percent for the district as a whole and one can easily see the pOpulation problem Kisii faces. Six innovations are under investigation in this study. Coffee was first planted in the district in 1921, but it was not until about 1950 that it began to spread. Today a little over half of the study area has some coffee grown on it, but diffusion has essentially stopped, due to a ban imposed on additional planting. The ban, however, has not been strictly enforced. Tea was first introduced in 1957 in the eastern part of the study area, because farmers in that part of the district had experience working with tea on the nearby estates. It spread rapidly to the west, but now the Kenya Tea Deve10pment Authority is attempting to restrict the Spread of adoption, instead opting for more concentrated adoption. In 1950 pyrethrum was first grown in the eastern part of the study area. It has spread to most of the study area, but remains concentrated in the east. The main reason for the lack of spread to the entire area was the prohibition on planting below 1890 meters (6,200 feet), which slowed the adoption rate in the western one- third of the study area. Now that demand has increased substantially pyrethrum is grown below that level. 226 Overall, adoption continues, but the Slowdown at the tOp end of the growth curve is becoming faintly perceptible. Passion fruit was introduced into the Keroka area in 1959, and today remains concentrated in the same place. After a series of difficulties, such as low price, disease and discouragement by the agriculture department, the crOp started to gain acceptance in 1966. Presently the adOption rate is comparable to that for tea. Grade cattle are gradually replacing the local breeds in Kisii. The cattle were introduced in a number of locations at the same time so the diffusion process is primarily characterized by intensification and infilling rather than an outward movement of a diffusion wave. In spite of the high cost of acquiring grade cattle the rate of adOption is quite high. The cost in and of itself is not a factor in the rate of adoption because very few Gusii farmers could afford to adOpt without a loan. The hinderance to adoption therefore is the difficulty of securing a loan with which to buy the cattle. The profitability of grade cattle is obviously sufficient to justify the expenditure involved. Hybrid maize has experienced the fastest adoption rate in the history of Kisii District. From its intro- duction in the early 1960's until 1966 the adOption rate was not very high, but in that year it began to be accepted with amazing speed. In 1971 only about 20 227 percent of the farmers had failed to adopt hybrid maize. The adoption rate has now slowed slightly as the upper part of the growth curve has been reached. Table 16 is a summary of the principal character- istics of the diffusion of the six innovations being investigated here. All of the items in the table are presented in the analysis of the diffusion process in Kisii, but are included in tabular form here for the convenience of the reader. A dissertation should seek to be unique and inno- vative if it is to add to the current body of knowledge. To that end, this dissertation has attempted to do several things that have heretofore seldom appeared in any research. The distinctions are as follows: (1) Innovations were examined as they diffuse through space as well as time. (2) The focus of the research was on an area in the develOping world. (3) The unit of analysis was the Spatial unit rather than the individual farmer. (4) Percent of adoption by area was used rather than individual discreet adoptions. (5) Spatial diffusion was related to socio-economic and demographic characteristics via factor analysis and multiple regression and corre- lation. (6) Computer maps and three-dimensional computer graphics were used to visually portray the spatial diffusion pattern. (7) An attempt was made to integrate the basic concepts of temporal and spatial diffusion. 228 TABLE 16.--Summary of Diffusion Characteristics. Passion Grade Hybrid Characteristic Coffee Pyrethrum Tea Fruit Cattle Maize Year of Introduction 1921 1950 1957 1959 1961 1958 Number of Originating Centers 3 3 2 l 7 9 Hagerstrand Typology (yrs) Stage I 1940-50 1950-52 1957—58 1959-68 1961-64 1958-66 Stage II 1952-62 1954-64 1960-64 1970-71 1966-71 1968 Stage III 1964-71 1966-71 1966-71 NA NA 1970-71 Stage IV NA NA NA NA NA NA Deviations from Stage II Stage II Stage II Stage II Stage II Stage II Hagerstrand and III and III and III and III Mean Year of Adoption vs. % Adoption 1971 r-.86 r-.61 r-.51 r-.68 r-.67 r-.23 Mean Year of Adoption vs. Use Intensity 1971 r-.9l r-.39 r-.54 r-.79 r-.74 r-.08 % Adoption 1971 vs. Use Inten- sity 1971 r-.90 r-.SS r-.60 r-.74 r-.69 r--.O6 Marketing Location Important? Yes Yes No No No No % Adoption 1971 42% 78% 32% 9% 11% 78% Innovation Wave same as Morrill's Different Different Different Different Different Different Socio-economic Variables Load- ing on Factor Population Analysis None None None Density None None 229 Conclusions The spatial diffusion of agricultural innovations in Kisii is characterized by the following. (1) In the initial period there is little contrast between the adopting and the non-adopting areas as the former have achieved only a low percent of acceptance. (2) During the second period new dif- fusion nodes form in isolated locations and peaks of higher adoption appear above the general low level of adoption. (3) The valleys between the adoption peaks begin to fill in as the peaks reach 100 percent acceptance and spread out to form plateaus of saturation level adOption. (4) Finally, the gradient of the diffusion wave becomes progressively steeper as the distance from locations with no adOption to saturation adoption lessens. The stages in the Hagerstrand typology of inno- vation diffusion are Similar to those found in Kisii, with two exceptions. Stage II, the dif— fusion stage in the Hagerstrand typology is denoted by the development of new diffusion nodes and the leveling of regional differences. In Kisii new nodes appear but regional differences become more pronounced. Old and new nodes increase in adoption percent until they become 230 peaks of very high adoption levels that stand out above the surrounding area. Stage III, the con- densing stage, is characterized by Hagerstrand as exhibiting equal increases in all areas. ~In Kisii this stage primarily involves filling in between the adoption peaks and the outward spread of those peaks to form plateaus. There is a strong positive relationship between the mean date of adoption, the percent of farmers adOpting the innovation and the intensity of use within the individual sampling areas. Thus, the first areas to adOpt an innovation also tend to be the ones with the highest percent of the farmers who have adopted and where each farmer raises more acres of the crop (or keeps more grade cattle). The exception to this tendency is hybrid maize. The adoption rate has been much more rapid than for the other innovations because it simply re— places an old crOp. It would tend to have more uniformity of acreage per farm because it is a food crop. Also, everyone raises local maize so the conversion is largely the result of seed availability. The location of marketing sites for coffee and pyrethrum are important in understanding the present Spatial pattern exhibited by these two 231 crops. In order for either a coffee factory or a pyrethrum c00perative marketing society to be built it is necessary to have enough production nearby to justify the expenditure involved. Fre- quently when a coffee factory was built the level of coffee adoption in adjacent areas would quickly increase. In 1959 and 1960 the pyrethrum societies were opened. Those areas located near the new cooperative marketing societies experienced a rapid decrease in acceptance. Today, areas with 100 percent adoption of pyrethrum are all located within a few kilometers of a cooperative marketing society. The Gusii have exhibited a ready response to the profitability of innovations. Price declines, or an uncertain market future, have a noticeable effect on the growth curves of adOption. For example, the inflection in the pyrethrum growth curve that takes place in 1960 was caused by a drOp in price. Passion fruit experienced a very low rate of adoption for several years until it became certain that the cr0p would be profitable to adopt. With the exception of hybrid maize, all the innovations exhibit approximately the same rate of adoption. In the early 1950's there were only 232 two innovations available, but in the 1960's all six became available. The rate at which each individual innovation is accepted has not signifi- cantly increased in recent years. Therefore, the combination of all innovation adoptions added together is a measure of culture change. This is not to say, however, that change necessarily is for the better, for spatial innovation dif- fusion, rather than the consequences of innovation is the subject of this research. Due to the very limited division of labor, both socially and spatially, each part of the study area is much like all other parts. That is, there is little range in the size of farms, Size of families, levels of income, etc. from one place to another. Each sub-area will, therefore, have approximately the same distribution of adopter ideal types, from innovator to laggard. An innovation with no ecological limits or market- ing location constraints could be introduced into any part of the study area with the same proba- bility of success because no area seems to have greater receptivity to innovation adoption. Thus, the most important determinant of when an area accepts an innovation is the time when the innovation wave moves into the area. spatial 233 Factor analysis reveals that the innovation measuring variables are structurally unrelated to the other socio-economic and demographic vari- ables. The exceptions are the geographic vari- ables, such as, distance from Kisii town, locational coordinates, precipitation and ele- vation. Pyrethrum and tea are related positively to each other while coffee is inversely related to both of these. Grade cattle and hybrid maize are not closely related to any of the other variables as they factor out by themselves. Passion fruit is related only to population density. The innovations are related in this manner because of their geographical distribution. Recommendations for Policy Planners The following recommendations will be confined to considerations. To reduce marketing costs individual cash crops should be confined to reasonably defined areas. Within those areas a high percent of the farmers should be raising the crop. Cash crop production areas should overlap, rather than be spatially exclusive, so farmers have the benefit of diversity in the event of a price decline. 234 Demonstration plots appear to be of minor im- portance in Kisii District, but have proved to be quite effective in other parts of the world. Plots should be strategically located where large numbers of people will see them. Also, they need to be prOperly marked and well main- tained, and the number of plots Should be increased. Agricultural innovations should be introduced simultaneously at several places in a small area (about the size of Kisii District) so that numerous innovation waves can be generated. This will allow the build-up of specific "how-to" knowledge that will facilitate adOption. Efforts of change agents should be concentrated around the points of original introduction to generate enough adoption for diffusion nodes to appear. The outward movement of innovation waves from several nodes will cause more rapid adOption than would occur if only one or two nodes existed. After the spatial spread is well develOped only a minimal push from the extension services will be required. Each individual farmer must make decisions as to the rational use of his limited resources. When competing alternative innovations are available for 235 adoption in an area farmers will be forced to decide between them. A farmer may decide to adopt only one, or he may utilize all available alternative innovations to a limited extent. Therefore, planners Should be prepared to accept either a Slower rate of adoption for each inno- vation, or a lower level of utilization. Inten- sive utilization and a high percentage of adoption for crop innovations Simply may not be possible due to the scarcity of land. Farmers at the laggard end of the adopter cate- gory continuum require more specific "how—to" knowledge and word-of-mouth information than others before they are willing to adOpt. This kind of information seems to be more concentrated in areas of highest adoption. Therefore, change agents should suggest to laggards only those innovations that have achieved a very high level of acceptance in the immediate area. If cash crOp marketing locations are established before there is sufficient demand nearby farmers will be encouraged to adopt. The first years of uneconomic operation could be justified on the grounds that it encourages adoption. 236 Suggestions for Further Research Innovations other than crops, such as fertilizers, should be investigated to determine the spatial diffusion pattern. Other spatial diffusion studies should be under— taken in areas of peasant agriculture, but with low pOpulation density, for the purpose of com- paring the diffusion pattern and the nature of the personal information field. The spatial pattern of extension agent activity should be examined to see how they distribute their effort over the area assigned to them, and to evaluate the impact of concentrated versus dispursed efforts. The relationship between the locations where an innovation is available to the farmer and the spatial diffusion pattern needs to be investi- gated. Non-agricultural innovations should be researched to see if they diffuse spatially through a rural population in the same manner as agricultural innovations. The "S" Shaped growth curve should be aggregated at different levels to examine the impact of study area size. 237 7. The consequences of innovation adOption in terms of improved family welfare Should be examined. B IBLIOGRAPHY BIBLIOGRAPHY Abrams, C., 1943. "Residential Propinquity as a Factor in Marriage Selection: Fifty Year Trends in Philadelphia." American Sociological Review, Vol. 8, pp. 288-294. Anderson, David L., 1970. "A Simple Innovation Diffusion Model." Proceed- ings of the Canadian Association of Geographers, Manitoba, pp.’9-l4. Arkin, Herbert and Raymond R. Colton, 1963. Tables for Statisticians. New York: Barnes and Noble, Inc. Baldwin, K. D. S., 1957. The Niger Agricultural Prgject: An Experiment in Africa. Oxford: Basil Blackwell, p. 221. Barnett, H. G., 1964. "Diffusion Rates." In Robert A. Manners, ed., Process and Pattern in Culture. Chicago: Aldine Publishing Co., pp. 351-363. Barnett, Homer G., 1964. "The Acceptance and Rejection of Change." In G. K. Follschan and W. Hirsch, Explorations in Social Change. Boston: Houghton Mifflin, 1964. Belcher, John C., 1958. "Acceptance of the Salk Polio Vaccine." Rural Sociology, Vol. 23, pp. 158-170. Blalock, H. M., 1960. Social Statistics. New York: McGraw-Hill. BOalt, Gunnar and Carl—Gunnar Janson, 1957. "Distance and Social Relations." Acta Sociologica, Vol. 2, pp. 73-97. 238 239 Bose, Sante Priya, 1961. "Characteristics of Farmers Who Adopt Agricul- tural Practices in Indian Villages." Rural Sociology, Vol. 26, pp. 138-145. Bossard, R., 1932. Bowden, "Residential Propinquity as a Factor in Marriage Selection." American Journal of Sociology, Vol. 38, pp. 219-224. Leonard W., 1965. Diffusion of the Decision to Irrigate. Research Paper No. 97, Department of Geography. Chicago: University of Chicago. Bowers, Raymond V., 1937. Brown, Brown, Brown, Brown, Brown, Brown, "The Direction of Intra-Societal Diffusion." American Sociological Review, Vol. 2, pp. 826- 836. Le' 1968e Diffusion Dynamics. Lund, Sweden: C. W. K. Gleerup. Lawrence A., 1969. "Diffusion of Innovation: A Macro View." Economic Deve10pment and Cultural Change, Vol. 17, pp. 189-211. Lawrence A., 1968. Diffusion Processes and Location: A Conceptual Framework and Bibliography. Philadelphia: Regional Science Institute. Lawrence A., 1965. Models for Spatial Diffusion Research--A Review. Technical Report No. 2. Evanston, Ill.: North- western University, Department of Geography. Lawrence A., 1967. "On the Spatial Diffusion of Innovation." Abstract, Annals pf the Association of American Geographers, Vol? 57 (December), p. 783. Lawrence and Eric G. Moore, 1969. "Diffusion Research in Geography: A Perspective." In Progress in Geography: International Reviews of Current Research, Vol. 1. Edited by ChrIStOpher Board, et al. New York: St. Martins Press, pp. 119-157. 240 Brown, Lawrence A. and Kevin R. Cox, 1971. "Empirical Regularities in the Diffusion of Innovation." Annals of the Association of American Geographers, Vol. 61, pp. 5514559. Brown, L. H., 1962. "Agriculture and Land Tenure in East Africa (Kenya)." The Natural Resources of East Africa. Edited by E. W. Russell. Nairobi: E. Africa Literature Bureau. Brown, Lester R., 1970. Seeds of Change: The Green Revolution and Development in the 1970' s. New York: Praeger Publishers. Carol, Hans, 1971. "Diffusion-Acceptance Pattern of Geographic Con- cepts in English and French-Speaking Universities in Africa." The Professional Geographer, Vol. 23, pp. 369-373. Carroll, Tom W., 1969. SINDI 2: Simulation of Innovation Diffusion in a Rural Community of Brazil. East Lansing, Mich.: Technical Report No. 8, Project on the Diffusion of Innovations in Rural Societies. Carroll, Tom W., and R. Vincent Farace, 1968. Systems Analysis, Computer Symulation, and Survey Research. Project on the Diffusion of Innovations in Rural Societies. East Lansing, Mich.: Michigan State University. ' Cassetti, E., 1969. "Why Do Diffusion Processes Conform to Logistic Trends?" Geographical Analysis, Vol. 1, pp. 101- 105. Cassetti, Emilio and George J. Demko, 1969. "A Diffusion Model of Fertility Decline: An Application to Selected Soviet Data: 1940-1965." Discussion Paper No. 5. Department of Geography. Ohio State University, 18 pp. Cassetti, Emilio and R. Keith Semple, 1969. "Concerning the Testing of Spatial Diffusion Hypotheses." Geographical Analysis, Vol. 1, pp. 254-259. 241 Chaudhari, Haider Ali and others, 1967. "Social Characteristics of Agricultural Innovators in Two Punjabi Villages in West Pakistan." Rural Sociology, Vol. 32, pp. 473-486. Clarke, A., 1952. "An Examination of the Operation of Residential PrOpinquity as a Factor in Mate Selection." American Sociological Review, Vol. 17, pp. 17-22. Clayton, Eric, 1961a. "Cash Crops for Smallholders." World Crgps, VOle l3, NOe 8' pp. 295-297e Clayton, Eric, 1961b. "Small-Scale Cash Crop Production in a Developing Economy." Economic Development and Cultural Change, Vol. 9, No. 4, Part I, pp. 618-624. Cliff, Andrew D., 1968. "The Neighborhood Effect in the Diffusion of Innovations." Transactions and Papers of the Igstitute of British Geographers, Vol. 44, pp. 75-84. Cole, S. M., 1950. An Outline of the Geology of Kenya. Nairobi. Coleman, James, Elihu Katz and Herbert Menzel, 1957. "The Diffusion of an Innovation." Sociometry, Vol. 20, pp. 253-270. Colenutt, R., 1969. "Linear Diffusion in an Urban Setting, An Example. Geographical Analysis, Vol. 1, pp. 106-114. Dasgupta, Satadal, 1965. "Communication and Innovation in Indian Villages." Social Forces, Vol. 43, pp. 330-337. Davie, M. R. and R. J. Reeves, 1939. "PrOpinquity of Residence Before Marriage." American Journal of Sociology, Vol. 44, pp. 510-517. Day, Richard H., 1970. "A Theoretical Note on the Spatial Diffusion of Something New." Gquraphical Analysis (January), pp. 68-76. Demko, George J. and Emilio Cassetti, 1970. "A Difquion Model for Selected Demographic Vari— ables: An Application to Soviet Data." Annals of the Association of American Geographers Vol. 60 ppe 533-g39e ' I 242 Desai, D. K. and B. M. Sharma, 1966. "Technological Change and Rate of Diffusion." Indian Journal of Agricultural Economics, Vol. 21, pp. 141-154. Deshler, Walter, 1965. "Man's Mark on the Landscape, A Clue at Culture Differences and Diffusion." Abstract, Annals of the Association of American Geographers, Vol. 55, p. 612. DeTemple, David J., 1970. "A Space Preference Approach to the Determination of Harvesture Systems in Northeast Iowa." Unpub- lished Ph.D. dissertation, Department of Geography. East Lansing: Michigan State University. deWilde, John C., 1967. Experiences with Agricultural Development in Tropical Africa. Volume I: The Synthesis and Volume II: The Case Studies. Baltimore: The Johns Hopkins Press. Vol. I, p. 254, Vol. II, p. 466. Dodd, Stuart C., 1955. "Diffusion is Predictable: Testing Probability Models for Laws of Interaction." American Sociological Review, Vol. 20, pp. 392-401. Eckans, Richard S., 1966. "Notes on Invention and Innovation in Less Developed Countries." American Economic Review, Vol. 56. pp. 98-117. Eicher, Carl, Thomas Zalla, James Kocher, and Fred Winch, 1970. Emplgyment Generation in African Agriculture. Research Report No. 9. East Lansing, Mich.: College of Agriculture and Natural Resources, Institute of International Agriculture. Ellsworth John S., Jr., 1948. "The Relationship of Population Density to Resi- dential PrOpinquity as a Factor in Marriage Selection." American Sociological Review, Vol. 13: Pp. 444-448. Etherington, D. M., 1967. "Projected Changes in Urban and Rural Population in Kenya and the Implications for Development Policy." Education, Employment and Rural Develop— ment. Edited by James R. Sheffield. Nairobi: E. A. Pub. House, pp. 54-74. 243 Fliegel, Frederick C., 1960. "Obstacles to Change for the Low-Income Farmer." Rural Sociology, Vol. 25, pp. 347-351. Fliegel, Frederick C. and Joseph E. Kivlin, 1966. "Attributes of Innovations as Factors in Dif- fusion." American Journal of Sociology, Vol. 72, pp. 235-248. Fliegel, F. C. and J. E. Kivlin, 1966. ‘ "Farmer's Perceptions of Farm Practice Attributes." Rural Sociology, Vol. 31, pp. 197-206. Fliegel, Frederick C. and Joseph E. Kivlin, 1962. "Farm Practice Attributes and Adoption Rates." Social Forces, Vol. 40, pp. 364-370. Goldthrope, J. E., 1958. Outlines of East African Society. Kempala, Uganda: Makere College, The University College of East Africa, Department of Sociology, 277pp. Gould, Peter, 1964. "A Note on Research into the Diffusion of Deve10p- ment." Journal of Modern African Studies, Vol. 2, pp. 123-125. > Gould, Peter R., 1969. Spatial Diffusion. Resource Paper No. 4. Association of American Geographers. Gould, P. R., 1970. "Tanzania: 1920-63: The Spatial Impress of the Modernization Process." World Politics, Vol. 22. Graham, Saxon, 1956. "Class and Conservatism in the Adoption of Innovation." Human Relations, Vol. 9, pp. 91-100. Gross, Neal, 1949. "The Differential Characteristics of Acceptors and Non-Acceptors of an Approved Agricultural Techno- logical Practice." Rural Sociology: Vol. 14, pp. 148-156. Griliches, Zvi, 1957. "Hybrid Corn: An Exploration in the Economics of Technological Change." Econometrica, Vol. 25, pp. 501-522. 244 Griliches, Zvi, 1962. "Profitability versus Interaction: Another False Dichotomy." Rural Sociology, Vol. 27, pp. 327-330. Gulliver, P. H., 1961. "Land Shortage, Social Change and Social Conflict in East Africa." The Journal of Conflict Resolution, Vol. 5, No. 1. Hagerstrand, Torsten, 1965a. "A Monte Carlo Approach to Diffusion." Archives Europeennes de Sociologie, Vol. 6, pp. 43-67e Hagerstrand, T., 1966. "Aspects of the Spatial Structure of Social Communication and the Diffusion of Information." Regional Sciences Association Papers, Vol. 16, pp. 27-43. Hagerstrand, Torsten, 1967a. Innovation Diffusion as a Spatial Process. Chicago: University of Chicago Press. Hagerstrand, Torsten, 1967b. "On Monte Carlo Simulation of Diffusion." In W. L. Garrison and D. F. Marble, eds., Quanti tative Geography, Part I: Economic and_Cultural Topics. Studiés in Geography No. 13. Evanston: Northwestern University, pp. 1-32. Hagerstrand, Torsten, 1952. The PrOpagation of Innovation Waves. Lund Studies in Geography, Series B, Human Geog., No. 4. Hagerstrand, Torsten, 1965. "Quantitative Techniques for Analysis of the Spread of Information and Technology." In C. A. Anderson and M. J. Bowman, eds., Education and Economic Deve10pment. Chicago: Aldine. .Hanneman, Gerhard J., 1969. "A Computer Simulation of Information Diffusion in a Peasant Community." Unpublished Master's thesis, Michigan State University. Halnneman, Gerhard J., et al., 1969. "Computer Simulation on Innovation Diffusion in a Peasant Village." The Amegican Behavioral Scientist, Vol. 12, pp. 36-45. 245 Harbison, Frederick H., 1967. "The Generation of Employment in Newly DevelOping Countries." In James R. Sheffield, ed., Egg- cation, Emplgyment and Rural Development. Nairobi: East African Publishing House, PP. 174-193. Havens, A. Eugene and Everett M. Rogers, 1961. "Adoption of Hybrid Corn: Profitability and the Interaction Effect." Rural Sociology, Vol. 26, pp. 409-414. ‘ Heisel, D. F., 1967. "Demographic Trends and Educational Needs." In James R. Sheffield, ed., Education, Employment and Rural Development. Nairobi: East African Publishing House, pp. 75-83. Hennings, R. 0., 1952. "Some Trends and Problems of African Land Tenure in Kenya." Journal of African Administration, Vol. 4, No. 4 (October), pp. 122-134. Heyer, Judith, Dunstan Ireri and Jon Morris, 1969. Rural Developmgpp in Kenya. Nairobi: Institute for Deve10pment Studies, University College Nairobi. Homan, F. D., 1962. "Consolidation, Enclosure and Registration of Title in Kenya." Journal of African Administration, Vol. 1, No. 1. Homan, F. D., 1963. "Succession to Registered Land in the African Areas of Kenya." Journal of Local Administration Overseas, Vol. 2, No. l (Januaryi, pp. 49-54. Hudson, John C., 1969. "Diffusion in a Central Place System." Geographical Analysis, Vol. 1, pp. 45-58. Hunter, Guy, 1967. "Education, Employment and Rural Development: The Problem in East Africa." In James R. Shef- field, ed., EducationL Employment and Rural Development. Nairobi: East African Publishing House. IBRD (International Bank for Reconstruction and Deve10p- ment), 1963. The Economic Development of Kenya. Baltimore: The Johns HOpkins University Press, p. 380. 246 International Monetary Fund, 1969. Survays of African Economies. Volume II, Kenya, Tanzenia, Uganda, and Somalia. Washington, D.C.: International Monetary Fund. Johansen, Harley E., 1971. "Diffusion of Strip Cropping in Southwestern Wisconsin." AnnalngAssociation of American Geographers, Vol. 61, pp. 671-683. Jones, Gwyn E., 1967. "AdOption and Diffusion of Agricultural Practices." World Agricultural Economics and Rural Sociological Abstracts, Vol. 9, pp. 1-34. Kao, Charles H. C., Kurt R. Anschel and Carl K. Eicher, 1964. "Disguised Unemployment in Agriculture: A Survey." In Carl Eicher and Lawrence Witt, eds., Agriculture in Economic Development. New York: McGraw-Hill Book Company, pp. 129-144. Katz, Elihu, Martin Levin and Herbert Hamilton, 1963. "Traditions of Research on the Diffusion of Innovation." American Sociological Review, Vol. 28, pp. 237-252., Kniffen, Fred, 1965. . "Folk Housing: Key to Difquion." Annals of the Association of American Geographers, V01. 55: PP. 549-577. Knowles, 0. S., 1956. "The Deve10pment of Agricultural Marketing in Kenya." The East African_Economics Review, Vol. 3, No. 1, pp. 191-197. Landahl, H. D., 1953. "On the Spread of Information with Time and Distance." Bulletin of Mathematical Biophysics, V01. 15! Pp. 367-383. Larson, Otto N and Richard J. Hill, 1958. "Structure and Interpersonal Communication." American Journal of Sociology, Vol. 63, pp. T577555. Lerner, Danial and Wilbur Schramm, eds., 1967. Communication and Change in the DeveloEing Countries. Honolulu: East-West Center Press. 247 LeVine, Robert A., 1960. "The Internalization of Political Values in Stateless Societies." Human Organization, Vol. 19, pp. 51-58. LeVine, Robert A., 1962. "Wealth and Power in Gusiiland. " In P. J. Bohannan and G. Dalton, eds., Markets in Africa. Evanston, Ill.: Northwestern UniverSity Press. LeVine, Robert A., 1963. . "Witchcraft and Sorcery in a Gusii Community." {71 In John Middleton and E. H. Winter, eds., Witch- ; craft and Sorcery in East Africa. New York: 3 Frederick A. Praeger, pp. 231-255. ‘ LeVine, Robert A. and Barbara B. LeVine, 1966. L Nyansongo: A Gusii Community in Kenya. New York: John Wiley and Sons, Inc. Lewis, W. Arthur, 1964. "Thoughts on Land Settlement." In Carl Eicher and Lawrence Witt, Agriculture in Economic Development. New York: McGraw-Hill Book Co., pp. 299-310. Lindstrom, David E., 1958. "Diffusion of Agricultural and Home Economics Practices in a Japanese Rural Community." Rural Sociology, Vol. 23, pp. 171-183. ‘ Linge, G. J. R., 1963. "The Diffusion of Manufacturing in Auckland, New Zealand." Economic Geography, Vol. 39, pp. 23-39. Lionberger, Herbert F. and Edward Hassinger, 1954. "Neighborhoods as a Factor in the Diffusion of Farm Information in a Northeast Missouri Farming Com- munity." Rural Sociology, Vol. 19, pp. 378-384. MacArthur, J. D., 1967. "Some Thoughts" on Future Trends in Farm Employ- ment in Kenya. In James R. Sheffield, ed., EducationL Employment and Rural Develppment. Nairobi: East African Publishingfl House, pp. 122- 140. .Madigan, Francis C., 1962. "Predicting Receptivity to Community Development Innovations." Current Anthropplogy, Vol. 3, pp. 207-208. 248 Marble, Duane F. and John D. Nystuen, 1963. "An Approach to the Direct Measurement of Com- munity Mean Information Fields." Papers, Regional Science Association, Vol. 2, pp. 99-109. Marsh, C. P. and A. Lee Coleman, 1954. "The Relation of Neighborhood of Residence to Adoption of Recommended Farm Practices." Rural Sociology, Vol. 19, pp. 385-389. Marsh, C. Paul and A. Lee Coleman, 1955. "The Relation of Farmer Characteristics to the Adoption of Recommended Farm Practices." Rural SociologY, Vol. 20, pp. 289-296. Mason, Robert G., 1964. "The Use of Information Sources in the Process of Adoption." Rural Sociology, Vol. 29, pp. 40-52. Maxon, Robert M., 1969. "The Gusii Uprising of 1908." Proceedings, Fifth Annual Conference, Social Science Council, University of East Africa, Nairobi. Mayer, Phillip, 1950. Gusii Brideswealth, Law and Custom. Rhodes- Livingstone Paper No. 18. Cape Town: Oxford University Press. Meadows, Donella, et al., 1972. The Limits to Growth. New York: Universe Books. Miller, G. A., 1947. "Population, Distance and the Circulation of Information." aggrican Journal of Psychology, Vol. 60, pp. 276-284. .Miracle, Marvin P. and Ann Seidman, 1968. State Farms in Ghana. Madison: Land Tenure Center, University of Wisconsin, p. 54. .Misra, R. P., 1966. "A Preliminary Quantitative Analysis of Spatial Diffusion in Human Geographic Continuum." Th3 National Geographical Journal of India, Vol. 12, pp. 149-155. Misra, R. P., 1969. Diffusion of Agricultural Innovations: A Theoretical and Empirical Study. Prasarangai: University of Mysore. . l7—> .. é ! 249 Morrill, Richard L., 1968. "The Distribution of Migration Distances." Papers of the Regional Science Association, Vol. 11, pp. 75-84. Morrill, Richard L., 1968. . "Waves of Spatial Diffusion." Journal of Regional Science, Vol. 8, pp. 1-18. Morrill, Richard L. and Forrest R. Pitts, 1967. . "Marriage, Migration and the Mean Information a- Field." 'Annals ofythe Association of American : Geographers, Vol. 57, pp. 401-422. Moris, Jon, 1967. "Farmer Training as a Strategy of Rural Develop- ment." In James R. Sheffield, ed., Education, Employment and Rural Development. Nairobi: East African Publishing House, pp. 322-365. Mueller, P. and K. H. Zevering, 1969. "Employment Promotion Through Rural Development, A Pilot Project in Western Nigeria." International Labor Review, Vol. 6, No. 2, pp. 111-129. National Science Foundation, 1961. "Diffusion of Technological Change." In Reviews g£_Data oanesearch and Development. Washington, D.C.: NFS Report #31. Ndegwa, Philip and Ole David Koht Norbye, 1967. "The Strategy of Kenya's Deve10pment Plan 1966- 1970." In James R. Sheffied, ed., Education, Employment ang Rural Development in Kenya. Nairobi: East African Publishing House, pp. 87-98. Nicholls, William H., 1964. "The Place of Agriculture in Economic Develop- ment." In Carl Eicher and Lawrence Witt, eds., Aggiculture in Economic Development. New York: McGraw-Hiil Book Company, pp. 11-44. Nwala, Eze Ogbueri Ajoku, 1971. "Spatial Patterns of Institutional Innovations Within a Modernizing Society." Unpublished Ph.D. dissertation, The Ohio State University. Odingo, Richard S., 1967. "Land Settlement in the Kenya Highlands." In James R. Sheffield, ed., Education, Employment and Rural Development. Nairobi: East African Publishing House, pp. 141-151. 250 Olsson, Gunnar, 1965. Distance and Human Interaction. Bibliography Series No. 2. Philadelphia: Regional Science Research Institute. Ominde, S. H., 1968. Land and POpulation Movements in Kenya. London: Heinemann Educational Books, Ltd. Pearson, Lester B., 1969. , Partners of Development. Report of the Commission on International Development. New York: Praeger Publishers, p. 55. '3 Pemberton, H. Earl, 1938. "Spatial Order of Cultural Diffusion." Sociologyland Social Research, Vol. 22, pp. 246-251. (I J“ ‘1 ‘1.” -G'm! Pitts, Forrest R., 1963. "Problems in Computer Simulation of Diffusion." Papers of the Regional Science Association, Vol. 11, pp. 111-122. Ponsioen, H. Earl, 1962. The Analysis of Social Change Reconsidered: A Sociological Study. S-Gravenhaye, Netherlands: Institute of Social Studies. Poole, W. C., 1928. "Distance in Sociology." American Journal of 'Sociology, Vol. 33, pp. 99-104. Pred, Allan, 1969. Behavior and Location: Foundations for a Geographic and Dynamic Location Theory. Part II. Lund Studies in Geography, Series B, Human Geography, No. 28. Pyle, G., 1969. "The Diffusion of Cholera in the United States in the Nineteenth Century." Geographical Analysis, Vol. 1, pp. 59-75. .Rado, E. R., 1967. "Manpower Planning in E. Africa." The E. African Economic Review, Vol. 3, No. 1, pp. 1-30. lRahudkar, Wasudeo B., 1962. "Farmer Characteristics Associated with the Adoption and Diffusion of Improved Farm Practices." Indian Journal of Agricultural Economics, Vol. 17, pp. 82-85. 251 Ramachandran, Ranganathan, 1969. "Spatial Diffusion of Innovation in Rural India: A Case Study of Irrigation Pumps in the Coimbatore Plateau." Unpublished Ph.D. dissertation, Clark University, Worchester, Mass. Rapoport, Anatol, 1951. "Nets with Distance Bias." Bulletin of Mathe- matical Biophysics, Vol. 13, pp. 85-92. Ray, Robert S., 1967. "The Structure of Employment in Kenya." In James R. Sheffield, ed., Education, Employment and Rural Development. Nairobi: East African Pub- lishing House, pp. 154-213. Redlich, Fritz, 1953. "Ideas, Their Migration in Space and Transmittal Over Time." Kyklos, Vol. 6, pp. 301-322. Riddell, J. Berry, 1970. "Network and Hierarchical Components of Moderni- zation in Developing Regions: Evidence from Sierra Leone." Proceedings, Canadian Association of Geographers, pp. 2 - 02. Riddel, J., 1970. The Spatial Dynamigg of Modernization in Sierra Leone: Structure, Diffusion andCResponse. Evanston, Ill.: Northwestern University Press. Rogers, Everett M., 1962. Diffusion of Innovations. New York: The Free Press. Rogers, Everett M., Joseph R. Ascroft and Niels G. Ruling, 1970. Diffusion of Innovations in Brazil, Nigeria and India. Diffusion of Innovations Research Report No. 24. East Lansing: Michigan State University. Rogers, Everett and G. N. Beal, 1958. "The Importance of Personal Influences in the Adoption of Technological Changes.” Social Forces, Vol. 36. Pp. 329-335. Rogers, Everett M. and Eugene Havens, 1962. "Rejoinder to Griliches' 'Another False Dichotomy.'" Rural Sociology, Vol. 27, pp. 330- 332. 252 Rogers, Everett M. with F. Floyd Shoemaker, 1971. Communication of Innovations: A Cross-Cultural Approach. 2nd Edition. New York: The Free Press. Ruthenberg, Hans, 1966. _ African Agricultural Production Develo ment Policy in Kenya 1952-1965. Berlin: Springer- Verlag, p. 164. Ryan, Bryce and Neal C. Gross, 1943. ?* "The Diffusion of Hybrid Seed Corn in Two Iowa Communities." Rural Sociology, Vol. 8, pp. 15-24. Savale, R. S., 1966. "Technological Change in Agriculture." Indian Journal of Agricultural Economics, Vol. 71: PP. 199-208. Schultz, Theodore W., 1964. '* Transforming Ttagitional Agriculture. New Haven: Yale University Press. Sheffield, James R., ed., 1967. Education, Em lo ent and Rural Development. Nairobi: East A¥rican Publishing House. Shetty, N. S., 1966. "Inter-Farm Rates of Technological Diffusion in 9 Indian Agriculture." Indian Journal of Ag. Econ., Vol. 21. pp. 188-198. Smith, H. T. E., 1958. "The Communication of Ideas to Farmers." Agricultural Progress, Vol. 33, pp. 51-57. Standard Bank, 1970. ' KenyaL Annual Economic Review. London: Economic Department, The Standard’Bank, Ltd., p. 16. Stanislawski, Dan, 1946. "The Origin and Spread of the Grid Pattern Town." Geographical Review, Vol. 36, pp. 105-120. Swynnerton, R. J. M., 1962. "African Farming Transformed." Commonwealth Journal, Vol. 5, No. 3, pp. 137-140. Swynnerton, R. J. M., 1967. "Agricultural Advances in Eastern Africa." African Affairs, Vol. 61, pp. 201-213. 253 Taylor, D. R. F., 1965. "New Tea Growing Areas in Kenya." Geography, Vol. 50, Part 4, pp. 373-375. Tiedeman, C. E. and C. S. Van Doren, 1964. The Diffusion of Hyhrid Cornlin Iowa: h_Spatia1 Simulation Model. Technical Bulletin. East Lansing: Michigan State University, Institute for Community Development and Services, pp. 13-44. Timms, D. W. G., 1971. The Urban Mosaic: ITowards a Theory of Residential Differentiation. Cambridge: Cambiidge UniVersity Press, 277 pp. Todaro, Michael, 1969. "A Model of Labor Migration and Urban Unemploy- ment in Less Developed Countries." American Economic Review, Vol. 59, No. 1, pp. 138-148. Todaro, Michael P., 1968. "The Urban Employment Problem in Less Developed Countries: An Analysis of Demand and Supply." Yale Economic Essays, Vol. 8, pp. 329-402. Uchendu, Victor C. and Kenneth R. M. Anthony, 1969. EconomicL Cultural and Technical Determinants of Agricultural Chan e in Tro ical Africa. Prelimi- nary Report No. 2, Field Study of Agricultural Change: Kisii District, Kenya. Palo Alto: Food Research Institute: Stanford University. United Nations, 1971. Demographic Yearbook 1970. Statistical Office of the United Nations, Department of Economic and Social Affairs, United Nations, New York. Villeneuve, Paul Y., 1970. "Diffusion in Space and Time: The Case of an Ethnic Colony." Proceedings of the Canadian Association of Geographers, Manatoha, pp. 369- 375. Wagner, Gunter, 1949. The Bantu of North Kavirondo, Vol. I. New York: Oford University Press, pp. 132-133. Warntz, William, 1966. . "The Topology of a Socio-Economic Terrain and Spatial Flows." Papers of the Regional Science Association, Vol. , PP- -6 . 254 Wilson, Godfrey and Monica Wilson, 1945. The Analysis of Social Change: Based on Obser- vations in Centrai Africa. Cambridge: Cahhridge University Press. Witthuhn, Burton 0., 1968. "The Spatial Integration of Uganda as Shown by the Diffusion of Postal Agencies, 1900-1965." East Lakes Geographer, pp. 5—20. Wolpert, Julian, 1964. "A Regional Simulation Model of Information Diffusion." Public Opinion Quarterly, Vol. 30, pp. 597-608. Yapa, Lakshman Senarath, 1969. WASpatiometric Model for the Diffusion of Inno- vations: Simulation Experiments." Unpublished Ph.D. dissertation, Syracuse University. Young, J. N. and A. Lee Coleman, 1959. "Neighborhood Norms and the Adoption of Farm Practices." Rural Sociology, Vol. 24, pp. 372-380. Yuill, Robert S., 1964. A Simulation Study of JBarrier Effects in Spatial Difquion Problems. Technical Report No. l. Evanston: NorthWestern University, Department of Geography. Kenya Government Publications Geology Department, 1952. The Geology of the Kisii Highlw wds. Nairobi: Government Printer. Kenya, 1953. African Population of the Kenya Colon and Protectorate, l§T8Z Nairobi: The East African Statistics Department. Kenya, 1966a. Development Plan, 1966-1970. Nairobi. Ministry of Economic Pianning and Development. Kenya, 1969. Development Plan: 1970 to 1974. Nairobi: Ministry of Economic Planning and Development. 255 Kenya, 1964a to 1969a. Kenya Economic Survey, 1964 to 1967. Nairobi: Statistics Division, Ministry of Economic Planning and Development. Kenya, 1964. Kenya Population Census, 1962. Vols. I and II. Nairobi: Economics and Statistics Division, Ministry of Finance and Economic Planning. Kenya, 1966b. Kenya Population Census, 1962. Vol. III. African Population. Nairobi: Statistics Division, Ministry of Economic Planning and Deve10pment. Kenya, 1966c. Kenya Populaplon Census, 1962. Vol. IV. Non-African Population. Nairobi: Statistics Division, Ministry of Economic Planning and Development. Kenya, 1970. Kenya Population Census, 1969. Vol. I. Nairobi: Statistics Division, Ministry of Finance and Economic Planning. Kisii, 1970. Department of Agriculture. District Annual Reports (Kisii District), 1960-1970. KiSii, 19716 Kisii District Development Advisory Committee. District Commissioner's Office, March 1, Minutes. Interviews Mr. N. N. Khaniri. Tea Officer, Kisii District. Location: District Agricultural Office, Kisii. Date: July 1, 1971. Mr. Nathan Migire. Manager, Kisii Farmers Cooperative Union, Ltd. Location: Kisii Farmers Cooperative Union Offices, Kisii. Date: June 28, 1971. Mr. James M. Ombui. Assistant Manager, Masaba Union Farmers Cooperative Society. Location: Masamba Union Farmers Cooperative Union Offices, Keroka, Kisii District. Date: June 27, 1971. 256 Mr. Michael Owen. Manager, Kenya Fruit Processors, Ltd. Location: His home near Sotik. Date: June 30, 1971. Mr. Christoph von Tresckow. Manager, Small-Holder Credit Scheme. Location: District Agriculture Office, Kisii. Date: May 12, 1971. APPENDICES APPENDIX A Location Sub-location Coordinates Grid No. Interviewer Date Year when each innovation was adopted Hybrid Passion Grade Name Maize Coffee Tea Pyrethrum Fruit Cattle -_h____ _i. J lO. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 257 APPENDIX B location 5UP‘1S'99149!11,...__....-.Qrfihl1‘9 - Coordinates ._ ____l_‘arm No . 1 :Llolevat ion Ra int all Dist ance Interviewer Date 3:" Good Day, 1 am one of the officers going around in this Division asking a few u::f farmers like yourself quest ions about farming matters. We need this information 5__. to help us plan better services to farmers. We would be grateful 11 you could 3... help us with your answers. First, who is the head or owner of this farm and 8"“ where is he?(IF mgssgkvlg'tr 63391147143”; 11‘ HEAD LIVES AwAYJgnf [Enjoy] Qla) r- Farm Head Male __f‘em Residence “1}— ‘Hho is responsible for manag- [Nane____ 1.- , 1...--." _ J 2... ing this farm from day-to-day?' Head Wife ffon/Dtr Fthr/Mthr Bro/his Rel Hngemh 3.... How related to the head? ' ,1 2' 3 7 4 5 6 7 8 4 5—»- 'Q.’ Now, about lenianed or rents: by head. Bub-Coo Locihivgl‘list Prov 0th 6"" 02a How many pieces of land are owned/rented in this... 7 Ti 7: 22b How many acres total are owned/rented in (his 3,.“ 720 Is this land l)' nherited 2)Bought 3)Rented u)oth... (?_. ated Is it 1)Adjudctd 2)?»urvyd —3)Registrd I4)'I illed ... ”7”" " 3.... 03 Nomiahout crops grown on all .ybrd Local Passn {- land in this Division only Maize Maize _ fruit 3" 03a Haveyou evergrown... ”'“L ..... _- d----i J-__-1 ______ ‘1: {031: When did you first start growing... ..,ru .-----. ----- b . ..... l---- _ 2.... CBC we u egerftopped growing... _ L if 1 i 7—-» .13d 11' —7Hhen did you f1: at stop growing L ' " 4' 7 ”'1'"- ”" 1 ‘ “7 10-....- '3” . When did you last start again..'.. “"4“ ”J"- _' '" " """ 9: . .. _- ..--.. --- -____, How often have you stopped “growing j NO... O3e Km you still growing ,. 7 1 "" I "’ ' " ’ é... U1 (IF YES)wa many acres/trees are under .. T 1 3..- Hhat spacing do you use ... .. ' " “ ° ’ 1"" ' 'i " ' ‘ ’ " ' ‘f' "‘ .,""' Do you use Chem FergFYHor None . b""'" 5 " " " ” ‘ ' 7'7"“ " 5" Q33 How often(per crop maize) ' ’W'e ed . . '" 7 i i -... 5___ (between lung rains oth crops) Prune .. Li Ziiz: : : :‘L' . '- I. 2.11:: : .- g" doynu . [Just/Spray . 0-.- .‘14 What kind of maize iid you grow during (hi 5 years long rains? Eiyi : 1d Local___m ”T- .31” 1 And what kind did you grow during last 611"" short rains? hrid Loca [1*- 25 ' Now, about grade livestock raised on inws Bulisr Strs Heiflrs [Pigs icks 9‘: land in this Division o_n_l_.y_ Calves * , -..—WI!— 3“ Qba Haveyou ever kept ... .‘_-_- _d--_; -__-_ 3...... QSb When did you first start keeping.. .. U V g--- ‘15“? Have an ever sttpped reeping ... " ’ 7 " h 7 " 7 ' 7 ' " " " "’ 7— w‘wi (IF VIIISTHhen did you First step .. ”77'." ’7'ii'“"1"”-'1"""J J ' ' ""17" 7"" ’ 8 '— When dilyculmt :tart. ’“JF"‘"’ 1" “ " ‘ " F "" 9:“ Howoftendilyou stop... ' ‘ "‘ '““'1“1'“""90__ be How mahy. .do you have now. r """i’" "mi" ' i“- ' ' ’" ' i 5. Q6 And local 1.1 vested: raised on Cows Bulls Strs Heifers Sheep Coat Chip; q--—- land in this Division Calves 1"” How many........do you have now... 5" :7 (Ir EITHER GRADY: on LOCAL tows 731%) Grade -‘~ws Local Co" 6.): ‘37s How many of your cows are now in milk... ... _ ‘ l 2... 07b And how many of them are now dry ... ' 7’ ...; "7“779— r27c: How many pints per day total do you usually milk _ " ‘i""'”'”’_”""‘§o"" r77d And how many of these pints do you usually sell :J‘. M 1"" 7% Do you have l)A Dairy Shed ”Cleaning Stuff? ... ... ... '7'" " " ”‘7 “‘ 2: '37f (Do you own EM P-‘asture‘ 1;. it 'H‘enoed-J‘ 5)Paddocked? ”' “ ' ”“"""‘ 3-.. 07g Do you provide your cows with {)Foduers? 7)Concentrates?‘;_ )- 2"” Q'Ih Do you have your cows served byB 8M. I. 7 ”Bull? """" '7 "Hi-"‘- ‘0—. ‘ ."Ii How often do you dip or spraljour cattle) ‘Fimes pe: Month 7"” ‘23 WW about your water suppiy: “QWWW '71 her (SpecT 3; Where do you get water for . . ' . "“ v9“ -,~8a Home use ... _ 1’ .2 ._:__3. - “. 8"” 08b Livestock ... ... Z -. b ' I... 6,. ..,_..-.7.._,i_. 8 ..-- ,.....-.-.1i 2"” ‘18c __ ....... ixrigation. .. .. ,9 -. 9 1.5.1, y; _,_ 3"” 68¢! When did you achixe yt ur ..a ..—i ”KM-“A 7’ , 1 “.....- 09 Now, about farm labor: What wfiber I‘amly itelf Risaga 7:11p“ Paiu fgesangTo 0th S... did you use between last long rain “"1131"? Help Cntrct Rglv- 7"” 09a Cashcrops... ... ..., l 2 4 u 6 J-____ 12"“ 1911’ Food crops... ..., 7 8. 119.-.? [‘2‘ _ .y-_._ ____‘§:___ fflc Livestock l 2 1 la ‘- 6 80 259 . I . . ——-—-———-—.w _. (210 Where do you sell ' --.--- am???-miTR-éfit].}hbl:5oni.oc. -U1v 'D'i'st ”Pia-L leewhere (S ec 110a. . Chicks/Eggs... .. P i—fii: m _ l _ r2)“;+ 9.+ _5 I slob. «Milk... ..L _ _ l__J_’,2‘.L_3. ”3 ““5 ; _ QlOc :":Food crops .4; 1; r21 3 T 14 1 5 :10e - Meat... ..L " V, Ifj‘f‘fzfjffs'ff‘yfifj'js ' " .lOf :.1vestock.. .. j 1 2 3 i“ u s] 111 What kinds of farm Org. _ ‘ records do you keep’ ' C12 Not counting extension officer',name 2 ta1mers l) in this area wt.) you can trust to give IOU right information 8 advice about farming matters? 2) "" 012a And which two farmers in this area are usually 1 "' ahead of other farmers when it comes to trying H-"*' more mod.‘ngg§ys of cropior animal husbandry? 2) g””' Q13 Which of the following officials visited this Vet Ram C DeviHlth Y“? Comrce "* ‘farm at least once since the long rains lastyr 2 19;, -..‘L- kiwi] ___ '7... - Q13a And which did anyone from this farm visit? 1 2 3 u 1 5 6 -—— Qlu Which 0’? the following meetings did Chie Crop Anim H.Ec Cmrce Family Agric *{é—m anyone irom this farm attend at least Bares Demo Demo Demo Demo Plnin Show i)’ once since the long rains last year?... “£me .-..2. ... 3 , ff“ 5_ .-...61. 4i» _7-__. u '" ‘lua Which were found very useful?... ... .. 1 2 u h 6 7 7-~ 015 Has ‘there evefi‘hhen a demonstratiOn plot in any of’your .and?’ Yes 1) ‘fio g) "B- LLSa Since Xmis last year. how many demo plots lid vou see/visit in Ihis Dist? h- leb How many times have You and other-s from this farm attended FTC courses? “times cf“ 015 Is anyone fromifhis rarm atVIkanu oop Loca Haen- Soh'l HrfiBeifihrch E? ’Uther 5"” present a member of ... boc Counl deleo Board Group ‘luu Club .VW' (Ir MEMBER) P., 2 1—3 ..-..9 .11.... +1.9... .. :1 --.- uléa Is he/she an office-bearer? l 2 3 u 5 6 17 8 i 9 9" [$17 How often do you Daily Few times few times 1"3e1 om ever 35 . Read or have read to you or almost per week .pgr month : 1‘” ’Jl7a Daily Newspapers? _.rluwu .‘1” i_ 2 i l 4 ”U 1?- I.l7b Mthly/Wkly Mjgazines? t” ‘ a _3., ‘2], 2 : l 4 0 3" 917C Listen to Radio. . ... ... u ! 1 ‘ 7 ‘ ‘l i -0 {H- ql7d Visit Ki:;ii 'Iown .. u 1 3 J _ .‘ ?. l U 5‘ 017a Other istrirts in Yenya { M 1 3_ - .1 2 i l 1 i L“ Ql7f Nairoli. . .. ... ... 9 i 3 '""-I . 2 l j 0 17 ” Qi7g Go to Cht ch.. ... u I J 2 l_“1_0 8“ Ql7h How long did you live ,work ,studyjoutside Lisii dist in your life q" Jig What is leads main occup? _ Others?7 “5 ul8a What i: resp main CLCUP? ' Others? 3“ mg EDUCAT.1__N_ 5 _ g 1.133:ch 11 3 MARTY. AES‘ sTflU§ E E ‘EIIim FER! 15? REID PM None U 0 None 0 0 single 1 1 In this Sub-loc____.n‘ Up to Std 3 l l Vermac l l Married (Monog) 2 2 Llsewhere($pec) P— Up to Primary 2 2 iwahili 2 2“ Married (Polyg) 3 3 g‘ Tech Training 3 3_ English 3 3| Widowed u u ,H-B HE ”“525 FATfififi 5" Up to Form 2 ‘4 u rSeparated 5 5 ‘In this Sub-10¢ 7'" UP ‘0 for“ ” 5 5 '., \ . Elsewhere (Spec) “ Prof Training 1, 13 mzx 53!; i "" -.- Up to from f. 'I 7 Eagle 1 l ”f H‘adm“ W wig”? ‘ Cull/Univ H 8 Iemnle 0 0 3f Reap 2*".4-” n‘ ' 020 How many people live on this farm? :Kdults 16+? m :?linors l5-? 9" Q2Ua How many of the Heads' dirett family a1e on this farm? :And [way?_ ___ : tr 020b How many of the Heads children attend ,(nool?m :Primary?“ :Higher'?_ ;{ QZOc Who all help to pay school tees for N, I Head Wife Son/Dtr Fthr ”thr bro/Sis 0th 5” childoen from this farm? . ... ... O 1 2 3 " 5 6 5 021 What is your ... Geffee Pyrethrum ’Tea naize r'?.F1uit ' :fiiIR y__n «ma Co-oo society... ,- - _. "’ 1.1:... ._ ....._ _ _ “ c2111 Co-o;1/KTDA No... I 9 ' - . - . , , ... .. . AP-.. .. .... . , 1 L .. . ... Q2lc 1970 1ncome from , _ ,flrdy._,.‘bm-d g. ““i u, D“ and (on: c ' ' L. - . 4‘ Q22a House type I WALLS moi“ rum; WINDOWS 2' 5 ‘ Fione Wood Mud Tile Iron Tin Thatch. Cement_Wood Earth Glass Wood?” ' .i ..1 3. 3 1+ . 5 6 ' 7 - 8' ‘I I) x y n ‘W- .. . y‘ ‘ ' . . . r d22b House , ‘, TDlLET .UOFIN' , LIGHTING iENCINb a ' taci}'iies W.N. Pit Nii stove Jikn Stones Lleo Prfn 01hr; Wire Hedge ”one L” ‘ ' ___"-_fl __ww~‘”.__‘ t}‘—'——2::-‘TT' i u " R 6 ‘27 8 $1. “‘ 5 ’L.--. V- Q22C House old towibaw _Spr _W¢H1rrow TVQStOPT34K?1L3W2w9199i‘“”4333T'V.Ca”.f¥9nffi- i posse.sions '}W_1 L 2 3 - u H 6 L, 7 8 9 “ x y -.-. —.---.-¢n~. 4- ..-—... “- .. .. ..." ‘F‘--‘nu* ....-- - .. —-—..——-— - ~———.— -- .. ....