Contents - Previous - Next

This is the old United Nations University website. Visit the new site at http://unu.edu

The history of settlement, land use, and environment (1890-1990)

The 100-year story of environmental and economic changes in Ukambani places current concerns and conditions within a broader context, informed by history as well as multiple perspectives on singular places and events. During this period, both land use and cropping systems changed rapidly in response to new markets, new property relations, population growth, and large-scale migration and resettlement. Many of the conditions that outsiders see as intrinsic to the region or its people are, in fact, results of specific colonial, national, and international policies and interventions. Outsiders noted famines and disease in the 1890s, severe land degradation in the 1920s and 1930s, low productivity and poverty from the 1950s to the 1970s, fuelwood shortages and land degradation in the 1980s, and threats to wildlife and biodiversity in the 1990s. All of these phenomena, however, also reflect the encounter of agricultural and pastoral peoples with international and national forces tied to the continuing transformation of the global economy, including the widespread commercialization of agriculture and its linkage to international markets. The conditions themselves, as well as their official interpretation, have often derived from this encounter. As in the case of the Nepal Himalayas (chapter 4), direct state intervention has constituted both a driving force and a response to environmental change, with both positive and negative consequences.

Fig. 5.4 Causal relations between agricultural production and environmental degradation at the farm level (Source: adapted from Raintree, 1984)

Pre-colonial history and traditional land use

According to most sources, the Akamba, who were hunters at the time, arrived in present-day Machakos from south of Mt Kilimanjaro around 1600 and eventually settled permanently in the Mbooni Hills (Hobley 1910; Krapf 1860; Lambert 1947a,b; Oliver 1965; Porter 1965, in prep.). Here they first became consolidated as a separate people and turned increasingly to agriculture (Lambert 1947a,b; Owako 1971). Eventually, overcrowding forced a move into the bush, and the traditional land-use system of integrating highland agriculture with lowland cattle-grazing came into being (Lambert 1947a,b). The land-tenure system of ng'undu (nearby grazing lands), utui (small clanbased settlements or small villages with permanent household cultivated plots and fallows), kisese (household grazing plots and paddocks), and weu (large tracts of common pastureland) developed in conjunction with this new subsistence system (Lambert 1947a,b; Wamalwa 1989).

Cattle owners led the settlement of dry frontier lands, attracted to the superior grazing on the plains. They began these settlements as cattle posts (with permanent residences retained in the hills) but later established permanent villages (Owako, 1971). From Mbooni, the first major move, occurring about 150-200 years ago, was across the Athi River to Kitui (Lindblom 1920,15). In this and other moves, the Akamba retained integrated highland/lowland, crop/livestock systems of land use. The chief aim of the land-tenure system was to spread risk and ensure group survival. The system was flexible, equitable, and geared to the community as a whole (Wamalwa 1989).

The mobility of the Akamba can be taken as an indicator of a frontier culture (Lambert 1947a,b; Oliver 1965; Owako 1971), although their successful adaptations in place at specific sites of sedentary farming and intensive land use suggest otherwise (Porter 1965, in prep.). Some authors depict an optimally adapted pre-colonial land-use system disrupted by external forces (Muthiani 1973; Wamalwa 1989; for adjacent areas, see also Deacon and Darkoh 1987; Sindiga 1984). On one point- that the traditional land-use system was well adapted to the vagaries of the physical environment - most authors agree. Integrated crop/livestock systems, spatially separated holdings, and mutual reciprocity arrangements served to spread risk and to provide mechanisms for coping with drought (Bernard and Thom 1981; Mutiso 1975; Porter 1965; Silberfein 1989; Waller 1985; Wamalwa 1989).

Akamba adaptability can also be seen as the very core of a stable nature-society relationship based on flexibility of movement and technology change. The traditional land-use system may be conceptualized as a coherent repertoire of diverse strategies, including both expansion and intensification of settlement, agriculture, and livestock production. Akamba farmers and agropastoralists cultivated a readiness to expand, intensify, relocate, or supplement their farming and livestock production activities in response to the changing economic and ecological conditions at local and national level (Gupta 1973).

Colonial contact (1890s)

In the 1890s, the Akamba experienced the first significant contact with colonial settlers and administration. The British Crown took over the administration of East Africa from the private International British East Africa company. The high cost of the Uganda Railroad was one impetus for this change, and making the railroad pay was the major reason behind the decision to try to entice white settlers to the Kenyan highlands (Bates 1987; Bradshaw 1990). Unfortunately, the railroad also seems to have been the means by which rinderpest invaded Ukambani in 1898 (Hardinge 1899). Also in 18981899, foreigners brought smallpox to the region and a severe drought occurred (Ambler 1988; Wamalwa 1989). These forces coalesced into the great famine of 1897-1901 (Ambler 1988). Since cattle normally served as the main drought insurance, the results were devastating. Kitui was especially hard hit (Hardinge 1899). Tate (1904) and Lindblom (1920) estimated the mortality rate in Ukambani during the famine at 50 per cent.

The devastation brought by this famine had lasting social effects (Ambler 1988). Depopulation weakened community bonds, differential survival rates of rich and poor caused tension, and in some places a total social breakdown occurred. Refugee urban migration brought growth to the "Swahili towns." Relief camps located at missions and government posts drew people into wage labour and served to "entrench the increasingly dominant position of these centers in commercial, political, and ideological terms" (Ambler 1988, 139). These trends also weakened the Akamba people's ecological control over their environment (Wamalwa 1989).

Land alienation (1900-World War I)

Among the various regional case-studies in this volume, Ukambani was influenced more than most (except perhaps the Southern Plains of the United States; see chap. 6) by the extent of land alienation for settlement of European farmers and ranchers. This led to large-scale displacement and relocation of rural people and dramatic changes in land-tenure systems, a legacy that continues to affect land use and migration in the region.

The first settler came to Machakos in 1894, but most of the district's land alienation and settlement occurred between 1901 and 1914 (Porter in prep.; Wisner 1977). Ignorance of indigenous land-use practice, combined with a rinderpest-depressed cattle population, caused an underestimation of Akamba land needs (Morgan 1963; Munro 1975; Silberfein 1984; Spencer 1983). Approximately 1 million acres were alienated in Ukambani between 1901 and 1914, and large tracts were placed "off limits" (Gupta 1973). The Akamba lost effective access to about two-thirds of the land that they had formerly controlled (Porter in prep.), including their most fertile lands, half of all their pasture (including their best grazing land), and their freedom to migrate (Munro 1975; Spencer 1983; Wisner 1977).

The Akamba, and their animals, were confined to "Native Reserves" in Machakos and Kitui. Colonial policy was directed at maximizing export crops, which Africans were forbidden to grow in order not to threaten the white monopoly. The state generally ignored "native" agriculture, while livestock activities were actively curtailed. The veterinary service was preoccupied with quarantining African cattle ostensibly to avoid contaminating European cattle, and an almost continuous quarantine existed from 1901 onward. These quarantines inhibited movement (perhaps exacerbating health problems), contributed to overcrowding of cattle on the reserves, and made it very difficult to sell cattle (Spencer 1983).

From the very beginning, the Akamba focused their dispute with the colonial government on the loss of land (Tignor 1976). Their entire land-tenure system had depended on seasonal and periodic access to large tracts of grazing land (Wamalwa 1989). Like many other Kenyan societies, Akamba social organization was based on conditions of abundant land and scarce labour and freedom of movement. Colonialism, by curtailing movement, closing the frontier, taking over existing settled lands, and forestalling the option of opening new lands, reversed these conditions. Land became scarce and labour abundant within a reduced, rigidly bounded area (Bates 1987), a situation that led many Akamba men to serve in the Carrier Corps of the British Colonial Army in World War I. Some recruits were coerced directly (Lelo 1994), whereas others responded to the new relations of land and labour established by land and livestock policy.

Land degradation in the 1920s and 1930s

By 1920, European expansion in Machakos had ended, and the colonial government continued to facilitate the transfer of resources from the reserves to the settlers. On the reserves, population increased, and the economy became more commercial. Land scarcity stimulated settlement in the plains, but poor land, tsetse, lack of water, and the 1924-1925 famine forced an exodus back to the hills, resulting in the virtual depopulation of parts of the plains and increased crowding in the hills (Owako 1971; Wisner 1977).

Relatively few Akamba engaged in wage labour through the 1920s, but this began to change by the end of the decade (Tignor 1976). The retrenchment of population from the plains, coupled with government hostility toward the pastoral sector, led the Akamba to pursue several agricultural innovations on the Machakos Reserve. These included the use of teams of oxen for ploughing and cart transport (allowing acreage increases), new crops, and intensive market gardening, all initiated without government aid (Munro 1975). Livestock production did not fare well. Disease outbreaks and fear of Akamba pressure for a reinstatement of grazing rights in Machakos and Yatta inspired continued official repression of the pastoral sector, including strict quarantines (Spencer 1983).

During this period, colonial officials and the settlers became increasingly aware of land degradation on the Native Reserves and linked this to cattle. State policy focused on conserving grazing land rather than developing African cattle resources (Spencer 1983). Although a few outlets for Akamba pastoral expansion existed in the 1920s, in the 1930s all external options were closed, and any hope for cooperation gave way to increasing regulation, compulsion, and monopoly (Munro 1975, 178; Spencer 1983).

In 1929, the highly publicized Hall Commission report elevated erosion in Machakos to the status of one of the most serious problems in all of Kenya, defining it as a major hazard (Myrick 1975; Spencer 1983). Contemporary accounts suggested that parts of Kitui were barren and depopulated (Champion 1933), that Machakos was eroded down to subsoil over 37 per cent of its area (Harroy 1949), and that over 1 million acres of the Akamba reserve were damaged and might never recover (Stockdale 1937, 290). The actual extent of damage is difficult to estimate because of official tendencies to exaggerate and the lack of systematic analyses (Munro 1975). The visibility of erosion was increased by the effects of locust invasions in 1929-1931 and a drought from 1931 to 1935 (Jacks and Whyte 1939).

Various agricultural changes, including intensification, contributed to land degradation (Munro 1975). Land scarcity led to decreased fallow and more continuous cropping, which, combined with the concurrent replacement of millet with maize, resulted in soil exhaustion and reduced yields. The remaining commonage (in both ng'undu and weu) was under severe strain as a source of forage and wood, particularly because many farmers began cash-cropping, ploughing up their grazing land, and exerting increased pressure on the weu and the shared grazing and fallow lands nearer to settlements. Erosion of pastureland and cropland resulted from overuse as well as expansion to steeper lands and more marginal soils. Colonial officers reported overgrazing (by goats on deforested slopes); grass fires; shifting cultivation on steep slopes without conservation measures; extensive deforestation to plant maize, coffee, and other cash crops; and parallel ploughing on 70 per cent slopes (Harroy 1949; Stockdale 1937).

In the densely populated Machakos Hills, land scarcity resulted, by the early 1930s, in permanent use of land and enclosure with fences or sisal hedges (Munro 1975). Land sales, tenancy, and, for the first time, landlessness became features of Akamba life (Silberfein 1984). Ready access to common lands had provided the basis of egalitarian Akamba society, but land scarcity and a market in land prompted the emergence of social stratification based on landholdings, and land disputes became common (Munro 1975; Mutiso 1975). Earlier traditions of multiple, functionally based tenure rights were replaced with individual-landowner rights, and formerly temporary usufructs were solidified into permanent, exclusive rights (Lambert 1947a,b; Wamalwa 1989), assimilated into "informed" reciprocity arrangements, or lost. Scarcity of pastureland along with other financial and administrative pressures resulted in an overall decrease in Akamba cattle wealth and a drastic decline, for most households, in livestock holdings. By the beginning of World War II, in contrast with the period just after World War I, significant numbers of Akamba were peasant cultivators instead of pastoralists, and crops had become a significant source of income (Munro 1975).

In spite of their own substantial soil degradation problems, the settlers waged a successful campaign to portray African agriculture as a kind of infectious disease, spreading insidiously across the landscape and contaminating productive European lands. Colonial policy throughout the 1930s was one of direct intervention in African land use. For the Akamba, the resulting policies served "to defend the European community by throwing a cordon sanitaire around the Akamba inhabited areas and intervening directly in the management of their land" (Munro 1975, 214).

In Machakos, the poor condition of the reserves, political fears of settlers, the global anti-erosion movement, and the new professionalism of the colonial agricultural bureaucracy (McCracken 1982) all coalesced in 1935 in the policies that emanated from the Kenya Land Commission. Aside from the minor changes in reserve boundaries, the Commission's major recommendations for Machakos were "destocking" and "reconditioning." When destocking began in 1938 (Moseley 1983), resistance erupted (Myrick 1975; Tignor 1971). More than 2,000 Akamba marched to Nairobi and petitioned the governor (Munro 1975), who returned the cattle and ended destocking (Moseley 1983).

Concurrently, agricultural officials focused with renewed vigour on reconditioning, actively encouraging the enclosure and seeding of grazing lands as well as the enclosure of homestead lands. By the end of 1939, 407,000 acres had been enclosed. In 1939-1940, the organizations created to oppose destocking also worked against enclosure, which they cited as a means to create cheap labour for Europeans by causing landlessness. Resistance included the sporadic destruction of hedges and continued challenges to several restrictive policies and practices. With the outbreak of war, this resistance was suppressed, but the call-up of agricultural officers also brought a virtual halt to reconditioning.

Whether all the activity that followed the Commission's recommendations was effective in improving the condition of the land is questionable. In 1944, a district agricultural officer concluded that the campaign of the previous six years had been ineffective and that only the compulsory closure of grazing lands, not individual fencing or ownership, had made any difference to the condition of the range (Lambert 1947a,b). Agricultural intensification and innovation did occur on the reserves during this time-period, but the agricultural services did little to assist innovative farmers (Munro 1975).

World War II and Kenyan independence

With the advent of World War II, colonial officials all but abandoned the quest for soil conservation in favour of campaigns to recruit Akamba men, both voluntarily and under coercion (Lelo 1994), into the army and to increase wartime production of grain crops. Nairobi and London officials sought to transform the rural landscape of Ukambani into an efficient and specialized production unit for grain crops during the war and for cash crops in the post-war period. The resulting soil depletion, combined with grain exports and unfavourable weather, contributed to severe food shortages and famines on the reserves in 1946, 1951-1952, and 1961. The colonial administration and the 2,000 settler families worked throughout the war years to promote commercial, settler-run enterprises with African wage labourers (Spencer 1980, 216).

Initially, the settlers were to intensify, specialize, and commercialize agriculture with Africans as wage labourers on their farms throughout Kenya. The 2,000 settler families took advantage of the war to recover from the depression and to gain control over the economic planning machinery of the colony (Spencer 1980, 513). Unlike the colonists, the reserves did not profit much from the war.

After the war, the government assisted demobilized British officers to "claim" or to buy land in Kenya (Lelo 1994), which intensified land use and increased pressure to evict "squatters," including Kenyan war veterans (Bates 1987). Combined with other pressures, this eventually led to the "MauMau rebellion" of the early 1950s (Bates 1987). The persistent low productivity on the reserves and a growing political insurrection in some parts of Kenya led to a liberal proposal for land-tenure reform, which continues to shape the evolving landscape in Ukambani and all of Kenya.

The Swynnerton Plan of 1954 was supposed to address African land problems by reforming land tenure, consolidating fragmented holdings, issuing freehold title, intensifying and developing African agriculture, providing access to credit, and removing restrictions on growing crops for export (Bradshaw 1990). It consisted of a three-phase programme: (1) land adjudication to "phase out" customary tenure; (2) land consolidation into one block per household to eliminate small, dispersed parcels, to allow greater specialization, and to realize economies of scale in cash crop production; and (3) land registration to provide for security of ownership and to establish a land market. Overall, the aim was to facilitate increased investment and employment in agriculture and to increase rural incomes and the "productivity" of land (Okoth-Ogendo 1976, 1981, 1991; Wangari 1991). The plan was predicated on an assumption that explicitly "successful" or wealthy African farmers would "be able to acquire more land and bad or poor farmers less, creating a landed and a landless class" (Swynnerton 1955, 10, cited in Wangari 1991).

The Swynnerton Plan created the basis for a market-oriented class of African farmers to work within a commercial-farming export sector and was credited with tripling agricultural output between 1955 and 1964 (Bradshaw 1990; Shipton 1988). It also succeeded in fostering land concentration (Shipton 1988) and social stratification, as foreseen by those who drafted the policy.

The simultaneous creation of a successful largeholder class and a landless and near-landless class in the highlands did not have the economic effect desired by Swynnerton and caused unforeseen environmental problems. The people who found themselves pushed off the best lands in their home areas did not all stay to work on the farms of others or to establish non-farm enterprises. Many went to work in the cities, the army, and the police force, and a large number went to drier frontier areas in search of new settlement opportunities.

Some migrants moved to resettlement areas opened up by the state in the late 1940s and the 1950s to alleviate the land shortage. Once independence was nearly at hand, many people prepared to move or risked illegal squatter settlements in anticipation of the reallocation of Crown Lands and large private holdings underutilized or abandoned by settlers (Mbithi and Barnes 1975). Thus, land hunger was often displaced to another more fragile area, rather than diverted into pursuit of wage labour in the same area. The simple equation envisioned by the Swynnerton Plan to describe a local zero-sum game was multiplied into a cascading effect that shook all of Ukambani and continues to the present.

Class formation and stratification, fuelled by the land-tenure reform, also affected livestock holdings, grazing practices, and land degradation. Colonial officials blamed overgrazing on the people most dependent on herding, yet the most overstocked areas of Ukambani in 1961 were the densely populated commercial farming areas in the hills (Porter in prep.). Although fodder grasses grown on farm may have altered carrying capacity somewhat, the degree of land degradation suggests that animals were overstocked relative to technology and management practice at the time. The wealthiest landholders often accumulated large herds and sent them to graze on whatever land was available - that is, "no man's land" under the new property regime. The overstocking problem attributed to the "tragedy of the commons" seems, in retrospect, to have followed the destruction of the commons and communal control and the concentration of private property.

Throughout this period, the conservation programmes of the 1930s continued with less visibility and lower priority than before. Considerable terracing and grass planting were carried out on private land during the period 1937-1944 (DeWilde 1967). Enclosure of grazing land and croplands gave way to the promotion of communal terracing (Munro 1975), and afforestation programmes fell off markedly (Harroy 1949).

The Akamba continued to view "reconditioning" with suspicion; they anticipated that these lands would be turned over to Europeans and that forced labour on settler farms and stock reductions were imminent (DeWilde 1967). In response to the Matungulu Betterment Scheme of 1946, which entailed the introduction of tractors to make terraces on broad tracts of land in a large-scale project, people threw themselves in front of the tractors (Bernard, Campbell, and Thom 1989; DeWilde 1967). Ironically, private soil-conservation efforts in Machakos, in spite of less government support than before, peaked in the mid-1950s amidst many reports of "startling recoveries" (DeWilde 1967, 94).

The response of Akamba farming and agro-pastoral communities to declining yields, the redefinition of property, and the continuing concern over land degradation was not uniform. Negotiation of specific responses played out in a middle region somewhere between the dramatic resistance against the Matungulu Betterment Scheme and the apparent success of government terracing programmes on private cropland. The widespread adoption of terracing in the 1950s did not reflect eventual acquiescence to government conservation interests, nor did resistance constitute rejection of terraces as such. Rather, farmers responded to the perceived risks and opportunities of specific programmes within their own production systems and political context.

For example, the increasing shift from sorghum and millet to maize as a staple crop throughout much of Ukambani accounts in large part for the willingness to terrace, since maize requires more water and responds well to the moister soil profile created by improved water storage on terraced croplands. Farmers switched to maize in response to strong market incentives and the removal of children's labour from the fields (a strong component of sorghum production). That maize was less drought resistant made farmers in the region far more vulnerable to drought. Terraces allowed them to maintain their link to the grain market, while also ensuring a greater supply of available water to the staple crop. On balance, the shift to maize still placed farmers at greater risk of losing their grain crop and reduced the food-production capacity of Machakos and Kitui districts over the long term (Bernard and Thom 1981).

By the time Kenya won independence, the people of Ukambani had experienced a restructuring of their livelihoods as well as their landscape. They had also witnessed the transformation of the property regime that had governed the ecological and spatial order of the* homeland (Bernard and Thom 1981; Mutiso 1975; see also Sindiga 1984). Frequent crop loss due to drought, like sedentary life and land registration, constrained the livelihoods of most residents of Machakos and many Kitui farmers as well (Porter 1965, in prep.). Akamba farmers anticipated independence and perhaps restoration of the grazing lands and the freedom of movement lost during the colonial period.

Independence and development (1961-1978)

Independence brought new hope and a change in the definition of national and regional problems, if not in the proposed solutions. In spite of dramatic increases in national agricultural productivity, particularly of export crops, Ukambani continued to experience food shortages and land degradation. Independence came hard on the heels of a severe drought followed by flooding, which produced a major famine in the region in 1961. The new nation identified poverty and hunger as the two most pressing problems of the rural majority for the country as a whole and for the region in particular. Government sources reported that 46 per cent of small-farm families in Kitui and 38 per cent of those in Machakos fell below minimum food consumption standards during the late 1970s (Republic of Kenya 1979, 66). Environmental concerns were identified as a legacy of the colonial past. Economic development remained the first-priority response of the national government, but the control and distribution of benefits changed.

Redistribution of land, coupled with increasing commercialization of smallholder agriculture, was seen by many as the solution to both poverty and hunger at local and national levels. In the drylands, many smallholders owned plots too degraded or too small to constitute viable production units under prevailing technology and cropping systems. Akamba demands for land were tied to the decades-old battle with the colonial government for the restoration of their grazing lands. Kenyan government policy from 1961 onward, however, extended the logic of the Swynnerton Plan (Migot-Adholla 1984; Wangari 1991). The National Development Plans of 1970, 1974, and 1978 referred explicitly to the need to extend the benefits of this programme to the semi-arid lands (Johnston 1989; Roe and Lewis 1989). Agricultural research followed a parallel route, focusing on the development of a high-yielding, fertilizer-responsive variety of hybrid maize with a short growing season for Akamba farmers (John Gerhart, personal communication). The dry farming areas of agro-ecological zones 4 and 5 were largely left out of this process until the 1970s. Most of zone 5 in Machakos and Kitui districts was not surveyed until much later, although the anticipation of legal demarcation and land title set in motion a process of land sales, land clearing, and consolidation of claims.

The land-tenure reform, as enacted in Ukambani during this period, bought off land hunger for some at the expense of others. It strengthened the private property of some men but left many men and most women with a shrinking commons, a frontier closed to free settlement, and less secure terms of access to resources in the communities in which they lived.

Widespread intensive cultivation increasingly displaced common grazing and gathering areas essential to poor smallholders. Those without capital, savings, or investments other than livestock were often not able to maintain medium-sized or even small farms under such conditions, and they sold out or abandoned their holdings. Women lost many of their rights of use and access, and most retained no legal authority over the land they occupied and even managed. They lost rather than gained, relative to men and even relative to their prior terms of access to resources under customary law.

The new government also promoted cash crops, increasing specialization, and the opening of government rangelands as frontiers for settlement. These were viewed as the main engines of growth and as sources of income, including income for purchasing food. This led to simultaneous, separate, and often conflicting policies to increase the land area planted to export cash crops and to increase food security and income among the rural poor.

Between 1961 and 1980, Machakos farmers doubled the land area under cultivation, while keeping more than 80 per cent of this area in food crop production, especially maize, beans, and peas (Mbogoh 1991). From independence to 1988, food crop area per capita stayed nearly constant (although it has declined slightly since 1982) and cash crop area per capita more than doubled (Mbogoh 1991), while noncultivated area per capita decreased dramatically. The cash crop area represented primarily the commercial expansion of prosperous largeholders. The decrease in non-cultivated area selectively affected those most dependent on off-farm resources (wood, fodder, food, and herbs).

As a counterpoint to its substantial support for privatization of land and the consequent polarization of rich and poor, the state promoted "harambee" or community and group efforts to improve public infrastructure and social welfare, and to provide relief under special circumstances to particular regions or groups of people in need (Hill 1991). This approach appealed to traditional values and practices of self-help and communal solidarity and was widely adopted in Ukambani (Bahemuka and Tiffen 1992; Hill 1991; Mbithi and Rasmusson 1977; Mutiso 1975).

By the 1960s, land hunger was acute throughout much of Ukambani. The dramatic migration of people within the region at this time has been described in part as a response to increased population density and overtaxed carrying capacity (Bernard, Campbell, and Thom 1989; Bernard and Thom 1981). But the impetus for migration had built up over many decades through land alienation, crowding, and reduced mobility. The timing and direction of the 1960s migrations were tied to Independence and the opening, through both formal and informal initiatives, of vast tracts of state
land and former estates to individual and cooperative settlement. For example, the government developed planned irrigation settlements and group ranches and also declared some areas, such as the Yatta Plateau and North Yatta, to be "open" for independent settlement and subsequent land claims. Landless and near-landless farmers from crowded high- and medium-potential areas hastened to occupy other well-situated tracts near water and roads on "unused" state or private lands (Mbithi and Barnes 1975; Migot-Adholla 1984; Okoth-Ogendo 1991; Tiffen 1992b; Wangari 1991). Although much former settler land was turned over to African small farmers (Ojany and Ogendo 1973) under a variety of government programmes, land concentration continued. Even cooperative ranches established on former private and public lands in Machakos and Kitui were dominated by wealthy absentee landowners (DeWilde 1967; Mutiso 1975, 1977). For Kenya as a whole, by the end of the 1980s, less than 0.2 per cent of farms (large farms, plantations, and ranches) controlled more than 40 per cent of farmland (Bradshaw 1990). Some scholars (Hunt 1984; Wasserman 1976) have viewed this policy of dryland resettlement specifically as a means of controlling the force of land hunger in the better sites.

Environment and development (1978-1990)

During the late 1970s, land degradation and resource management in Kenya once again attracted national and international attention. The major environmental concerns of the Kenyan government were sedimentation of dams at major hydroelectric facilities, soil erosion in cropland, declining crop yields, deforestation, and fuelwood shortages. The state focused on the national energy, water, and soil resources needed to increase agricultural production and to support the further development of urban services and industrial production in a developing economy. The national government, as well as its international advisers, identified the rural poor, particularly farmers and herders in semi-arid lands, as the major cause of the problem. This was often explicitly linked to Kenya's population growth rate at that time approaching 4 per cent per annum - one of the highest in the world.

Over the course of the 1980s, environmental concern shifted from soil erosion to deforestation, and the related concerns about rural poverty shifted toward hunger in response to the famine of 19841985. The national agencies did, however, maintain a continuity of concern over issues of energy, food, and water for national needs and rural subsistence. By the close of the 1980s, development and conservation priorities had once again splintered. Development agencies focused on food and cash crop production in the well-watered farmlands at higher elevations. Conservation agencies reverted to the preservation of savannas and wildlife in national parks, as the mainstay of the country's new main industry, tourism. Special agencies focused on the livelihoods of the rural poor in arid and semiarid lands. The perception that the rural poor, their increasing numbers, and their land-use practices were the causes of environmental degradation and food deficits remained constant, albeit with a change in tone to accommodate the recognition of the pressures and constraints propelling them toward destructive use of natural resources.

Soil erosion and watershed degradation

A new generation of soil-conservation programmes in Ukambani began with research on the erosive effects of rainfall, soil erodibility, stream sediment levels, erosion rates, and sediment delivery rates in small watersheds. Researchers in Machakos measured erosion rates in cropland ranging from 2 to 109 tons per hectare per year (Moore 1979a) and sediment delivery rates of 10-20,000 tons per km² per year (Dunne 1979; Dunne, Dietrich, and Brunengo 1978). With no historical comparison of erosion and production at the same sites and without a large sample of different sites at any given time, however, they could not predict the impact of erosion on current and future crop yields. The conservation efforts were driven not so much by concern over soil loss as by worries about the sedimentation of dams downstream.

The national and international agencies in Kenya returned to the colonial strategy of erecting physical structures, somewhat modified from the drainage works and terraces of the colonial period. Smallholder farmers in Ukambani, motivated by the need to intensify production on permanent sites, engaged in widespread construction and repair of modified terraces on their cropland. The new, less labour-intensive terraces were cheaper to build than the bench terraces of earlier campaigns, yet still harvested and stored runoff water, increasing soil moisture and crop yields. They also reduced nutrient loss and facilitated concentration of organic debris, compost, and water in new niches within the cropland.

The agricultural extension service successfully spread the practice of planting bananas in deep, composted holes along the base of terrace risers. Independently, farmers dramatically increased the acreage in other horticultural crops, particularly citrus, papaya, mango, and guava. They also expanded production of vegetables tenfold from the 1970s to the 1980s (Mbogah 1991, 9). Thus, farmers' interests in increasing soil moisture and fertility for more intensive maize and horticulture production on fixed plots coincided with national soil-conservation interests.

Meanwhile, the state also took steps to integrate conservation, food production, and other elements of rural development in semi-arid farmlands. The Machakos Integrated Development Project (MIDP) was one of the first major national development efforts of this type. The initiative was limited to Machakos District but was to serve as a pilot programme to guide government policy in Kitui District and other semiarid farming districts in eastern Kenya. Production farming and conservation were both included but were treated as separate activities.

Although MIDP and related soil-conservation programmes focused nearly all of their efforts on croplands (including already terraced croplands), researchers found that the overwhelming majority of the sediments in Machakos streams originated in grazing lands (table 5.6; Dunne 1979; Moore 1979a). The proportional contribution of grazing versus croplands reflected the much larger area in grazing land as compared with cropland, as well as a higher rate of soil loss per unit area for grazing land (Moore 1979a). Absentee landowners, large herds of livestock, and poorly maintained tracts of grazing land caused much of the problem. In addition, rural roads were found to contribute 1.2-2.2 times as much sediment as croplands to the total sediment yield at district level (EcoSystems Ltd. 1981, 1986; see also Dunne 1979). Although the need for preventive, protective, and rehabilitative measures was greatest for grazing lands and roads, bureaucratic perceptions of smallholder farmers as the causal agents of soil erosion prompted a response directed almost exclusively toward croplands.

Table 5.6 Estimated annual soil-erosion losses^a from Machakos area soils under different land-use systems

Source: Moore (1979a, 425).
a. Calculated using Universal Soil Loss Equation.

This is not to say that cropland conservation was not useful or necessary. As conversion to cropland proceeds, the expected erosion rate from "marginal" pasture plots, once under cultivation, may be substantially higher than the previous rates under grazing or the average cropland rates on better-quality land. The initial procedure of clearing and tillage, and the first season's rains on the newly exposed topsoil, may also result in substantially higher (10 to 100 times higher) soil losses on new croplands than the average rate for cropland (Lal 1976; Rocheleau 1984).

Although this would presumably indicate the need for proportionately more effort to avoid and slow the conversion of forest and pasture to cropland, or to terrace new fields, MIDP focused soil-conservation activities on the terracing of established croplands, the repair of active gullies, and the physical rehabilitation of severely degraded grazing lands. MIDP paid community work groups to construct hillside drains and to check dams and other watershed rehabilitation structures. On-farm terracing was left mostly to individual farmers or their own self-help groups with technical assistance (terrace contour markings) from MIDP soil-conservation extension agents.

Unfortunately, private terracing of cropland tended to proceed from the best lands to the worst, following the progress of the land survey and land adjudication teams from higher-potential, densely populated, and valuable lands outward into more marginal lands. New croplands were often terraced well after the heavy losses of topsoil in the initial cropping season. Thus a wave of intensified conservation practice lagged behind the growing wave of land conversion to cropland and soil erosion in new plots.

As soil erosion preoccupied district, national, and international conservation and development agencies, soil harvesting became a serious problem in "sand rivers" (seasonally dry stream-beds) downstream. Deforestation and erosion in semi-arid watersheds often change streamflow from permanent to seasonal regimes, but the sand deposits in stream-beds partially compensate for the disruption of streamflow, helping to maintain water points well into and even through the dry season. During the 1980s, highway and building contractors sent large trucks and work crews to mine sand from dry riverbeds in rural communities (Diang'a 1991; Thomas 1988). Although the practice of "sand scooping" attained some notoriety in the national press, it was treated as a law-enforcement problem rather than as an environmental issue.

In many communities with no piped water supplies these sand deposits store and filter water and often constitute the closest, cleanest, or only sources of water during the dry season. The destruction of one of these water points may add 4-10 km per day to the water-gathering journeys of rural women and children, which is especially significant for those who carry the water on their backs. It also increases the daily journey to water for cattle and goats, increasing the stress on the animals themselves as well as producing physical damage to roads, paths, river banks, and drainage systems.

Although some legal provisions exist for regulation at district and national level, the scale, timing, and nomadic nature of sand-scooping activity require enforcement at the local level. The community-level institutions that formerly governed access to land and other resources, however, have lost much of their former authority and enjoy no binding legal powers of enforcement, even where local community leadership does take a stand on such issues. Although national and district soil-conservation programmes did support infrastructural development at selected sand river water points in Ukambani, they did not address the protection of existing sites from outside, commercial interests. This partial treatment of the sand-scooping issue derived from the poor definition of rights and responsibilities with respect to shared resources and the segmentation of expertise and authority between local and national institutions.

Meanwhile, beyond the confines of Ukambani, the national development plan for 1980-1985 placed a high priority on Kenya's arid and semi-arid lands, with a view toward expanding land under cultivation further into the dry frontier zones and intensifying existing dryland cultivation. The same five-year plan sought to protect the forest, soil, and water resources of national importance from further damage. The impetus for this national initiative owed much to a decade of intensive research on poverty and environmental issues in Machakos District (Mbithi and Barnes 1975; Mbithi and Rasmusson 1977; Migot-Adholla 1984; Mutiso 1975; Muzaale and Leonard 1985) as well as to intensive research efforts on soil erosion (Barber, Thomas, and Moore 1981; Dunne 1979; Moore 1979a; Thomas, Barber, and Moore 1981) and agricultural technology (Collinson 1979; Gielen 1982; Lynam Gerhart, personal communication). Arid and semiarid land programmes in Kitui (Louis Berger Inc. 1985), Baringo (Thomas and Barber 1983; Thomas et al. 1982), and other districts also emphasized the rehabilitation of degraded hillslopes and water sources. These new initiatives all included soil-conservation activities, with the addition of a strong concern for water supply, dry forests, savanna trees, fuelwood supplies, and charcoal.

Deforestation and energy

During the 1970s, energy supply emerged as a major issue in development planning and environmental management. Surveys of energy use and sources in Africa revealed that most people depended on wood fuel (firewood or charcoal) for cooking and other domestic uses, as well as for many commercial uses. The magnitude of fuelwood demand in both urban and rural households raised serious concerns over future energy supplies in countries in which the annual harvest of wood far outstripped the rate of replacement in forests and savannas. As of 1979, the forest reserves of Kenya produced enough wood to meet 4.4 per cent of national energy needs, leaving trees on private lands and uncontrolled woodlands and scrubland to meet the remainder of the demand (Adams 1979). As with soil conservation, Machakos District served as a research laboratory for the dryland farming districts of the nation and the larger region.

Energy consumption, fuelwood sales, charcoal trade, and agro-forestry surveys identified Machakos District as a wood-fuel deficit area (Barnes, Ensminger, and O'Keefe 1984; Bradley 1991; Buck 1981; Mung'ala and Openshaw 1984; O'Keefe, Raskin, and Bernow 1984; Raintree 1984; Rocheleau 1985; Vonk 1983a). The district also exported charcoal to Nairobi, particularly during the drought of 1984, though the trade was illegal and mostly undocumented. Charcoalselling was ubiquitous along roadsides as well as at markets (Van Buren 1983). Mung'ala and Openshaw (1984) predicted, on the basis of local consumption alone, that fuelwood harvesting would deplete the stock of wood in plantations, dry forests, and savannas in Machakos District between 1986 and 2000.

Many rural people in both districts cited charcoal production and sale as the alternative of first resort when crops failed. It was equated with "starvation" or "reserve" crops (such as cassava) grown as a hedge against failure of more vulnerable crops (such as maize) during drought (Van Buren 1983; J. Kyengo and R. Vonk, personal communication). The deforestation problems in the drier parts of Ukambani resulted more from the widespread reliance on charcoal from dry forest trees as a "cash crop" than from the harvesting of fuelwood for local use (Wiener 1985).

Many research and rural development programmes at the time assumed that local use of fuelwood and urban charcoal markets (e.g. increases in local population and the commercialization of charcoal) caused deforestation in dry forests. Field research conducted in some districts (DeWees 1989) contradicted this prevailing assumption and suggested that, over the long term, land-clearing for agriculture was often the main "engine" of deforestation, with charcoal as a by-product. Even in the case of agricultural expansion and land conversion, however, the existence of the urban charcoal market facilitated more and larger land-conversion efforts by many farmers in dry forests in Machakos and Kitui. Some farmers harvested the standing trees on their land as assets that they "liquidated" (sold off) to subsidize the establishment of new croplands and home compounds. One farmer in a frontier zone in Machakos reported clearing over 10 ha of dry forest for charcoal in order to subsidize the establishment of 2 ha of cropland and a home compound (field interviews, 1991).

Frontier expansion into the dry forests was driven by a combination of population growth, commercial agriculture, and land markets in the densely populated farmlands in the region. The depletion of trees and forests in Ukambani, in both crowded and frontier zones, was driven by more than a simple process of population growth and escalating per capita energy demand. The rise in landlessness, owing to land consolidation and increased local and national market demand for food and cash crops, led to an even higher opportunity cost for keeping land in dry forest (Adams 1979).

Although many assumed that the depletion of forests and trees was a classic case of the tragedy of the commons, increasing privatization of land during the 1970s and 1980s may actually have accelerated the process of deforestation on the remaining common and state lands. In dry-forest frontiers, individual landholders hastened to clear-cut and cultivate as much land as possible in order to legitimize their claims through recognized use and "improvement," which did not include management of woodland and savanna. As noted above, the existence of a commercial charcoal market further encouraged deforestation of entire holdings as farmers "liquidated" wood-fuel assets to subsidize investments in new homesteads and cropland. The remaining commons came under ever greater pressure from the landless and smallholders, as well as from the livestock herds of large-holder farmers.

Not only the cause, but the very nature and awareness of the fuelwood shortage, proved to be quite complex. In Ukambani the shortage was perceived relative to prior conditions, preferred species, and existing practices of brick-making, charcoal production, and home cooking. Perceptions of the nature and importance of the shortage also varied between men and women, since women were responsible for domestic fuelwood collection and use in most households, whereas men controlled charcoal production and trade. The experience and perception of shortage at the household level varied even more substantially according to the terms of access rather than supply in the surrounding landscape (B. Wisner, personal communication). Poor women reported severe hardship imposed by dwindling fuelwood supply in landscapes in which some households enjoyed an abundant supply of wood from their own property or purchased charcoal from local producers (Rhoda Kisusu and others; self-help groups in Mbinni and Katangi locations, personal communication, 1984).

Regardless of the severity and distribution of the fuelwood shortage, in-depth field studies revealed it to be one small part of a far more complex situation, which included a diversity of local-level responses that were "invisible" to national sectoral surveys. It seems that forest and energy planners may well have been looking in all of the wrong places for trees. Using satellite imagery, land-use surveys, and thematic maps of land cover, they could not see the growing numbers of trees and other woody plants on cropland, grazing land, boundaries, and homesteads, for the shrinking forests on state and county council land. Yet, whereas forests accounted for only 1 per cent of the area of Machakos District in 1985, hedgerows and boundary plantings in the landscape covered 1.7 per cent of the land area in some parts of the district and were major sources of fuelwood for household use (EcoSystems Ltd. 1981, 1986; Mortimore 1992; Rocheleau and Hoek 1984).

Not only were many existing woody plants invisible to wood-fuel surveys, but the potential for new planting sites on farms had escaped the notice of many energy programmes. On-farm tree-planting efforts far surpassed any of the plantations on public sites and state reserves (Mortimore 1992). In fact, the farmers of Ukambani had already planted and selectively protected several woody species in cropland, grazing land, home compounds, and boundary lines (Fliervoet 1981; Gielen 1982; KENGO 1985; Poulsen 1981; Rocheleau 1985). They responded rapidly and enthusiastically to the availability of seed and seedlings for a diversity of tree species suited to multiple uses in a variety of niches in the landscape (Getahun 1989; Hoekstra 1984; Hoekstra and Kuguru 1983; KENGO 1989a,b; Rocheleau 1985, 1992).

As with soil conservation, the most successful programmes to protect and replace trees and forests were those that addressed a wide range of environmental and economic concerns, including food, water, fodder, small timber, and cash, as well as fuelwood supplies (Getahun 1989; Mortimore 1992; Rocheleau 1985). Independently, and with the help of the more flexible sources of technical assistance, many Akamba farmers began to integrate fuelwood planting into the land-use system over a much longer time-horizon. Tree-planting on farm resulted in a higher density of trees in the densely settled uplands (including very small farms) than in the more sparsely settled areas downslope. Farmers in the drier parts of the region were still depleting the wood stocks in the surrounding savanna and dry forest, while planting new fruit and timber and fencing trees on their own farms.

The national- and district-scale sectoral focus on energy also masked the multiple uses and values of trees and forests in Ukambani and elsewhere. The Akamba perceived fuelwood largely as a by-product of trees and shrubs serving other "higher" purposes (timber, poles, fodder, food, medicine, boundary markers, fencing), or looked to weedy species growing in fencerows and grazing lands as a source (Raintree 1984; Rocheleau 1985; Vonk 1983a,b). The loss of trees and dry forests as fodder reserves and sources of timber, poles, food, and medicine mattered to many Akamba as much as or more than the loss of future fuelwood supply or convenient sources of high-quality fuelwood. Although less direct and obvious, the impacts of deforestation on local water supplies were a significant hardship for rural peoples throughout the region. For most people the quality, reliability, and proximity of water sources were of higher priority than fuelwood. Moreover, water-supply improvements eventually proved to be a crucial prerequisite for tree propagation and on-farm planting, as well as for protection of existing trees from trampling and browsing by livestock en route to distant water points (Rocheleau 1985, 1992; Jama et al. 1992).

During the 1980s, the Ministry of Energy, the Forestry Department, and the Ministry of Agriculture, as well as MIDP and several special projects, provided technical assistance and millions of tree seedlings and seeds each year to farmers in Ukambani. Through the efforts of the Kenya Rural Energy Development Project and other multi-purpose tree-planting efforts, the availability of seedlings expanded from two species of Cypress and Eucalyptus in forest nurseries and a few exotic fruits in agricultural nurseries in 1978 to over 85 species (28 of them indigenous), stocked in large numbers in nurseries and seed centres throughout the country by 1988.

Biodiversity

The preservation of Kenya's wildlife loomed large on the international agenda during the 1980s (Anderson and Grove 1987; Myers 1982). Over the course of the decade, both scientific and activist environmental programmes paid increasing attention to the habitat of endangered animals, and they identified plant species and finally whole ecosystems as worthy of preservation (Anderson and Grove 1987; Yeager and Miller 1986). However, little national or international attention focused on biological impoverishment (loss of biodiversity) and the homogenization of landscape (loss of ecodiversity) in predominantly agrarian and pastoral landscapes. Official policy and practice sought effective exclusion of neighbouring people from the national parks, rather than the integration of local livelihoods with biodiversity, inside and outside the boundaries of the nation's parks.

Meanwhile, local communities in Ukambani experienced a substantial loss of biodiversity in their immediate surroundings and in their crop and livestock production systems. Their concerns reflected subsistence and drought preparedness as well as religious and cultural values. For poor farmers and herders, biodiversity concerns were intimately linked with livelihood and survival. Many Akamba people shared a practical interest in the continuing availability of medicinal herbs, wild and semi-domesticated foods, and a variety of fibres, dyes, wood products, and fodder sources within agrarian landscapes. The diversity of flora and fauna, with their attendant products for human use and their invisible services to the ecosystem (nutrient fixation and cycling, seed dispersal and plant pollination, water retention and regulation), depended, in turn, on the complexity and diversity of the landscape itself. But little research and discussion addressed the price of agricultural land-use intensification from the point of view of landscape complexity and biodiversity and even less took up the relation of biodiversity to the diversity of livelihood options and the preservation of culture (Oldfield and Alcorn 1991; Richards 1985).

In spite of the prevailing separation of production and biodiversity concerns in national and international institutions, the settlers of dry agricultural frontiers dealt with the integration of both concerns daily. Recent migrants noted the persistence of many large mammals as pests or simply as coexisting residents in landscapes characterized by a mosaic of cropland, pasture, and forest (Jama et al. 1992; Malaret 1991). Active hunting and concerted expulsion of wildlife emptied many frontier farming and herding communities of rhinoceroses, elephants, and large carnivores in the 1950s and 1960s. The surviving fauna in many areas in the 1980s included large and diverse populations of antelope (several species), primates, rodents, wild pigs, and warthogs, as well as amphibians, reptiles, birds, a multitude of insects (including pests, food sources, pollinators, and nutrient cyclers), and the less visible, but crucial, microfauna (field observations and interviews at sites throughout Machakos and Kitui, 1983-1991).

Throughout the drier zones of Machakos and Kitui, rural people tolerated, protected, and even fostered the growth of a multitude of wild plants, including many characterized as "weeds" by colonial, national, and international agricultural scientists. These grew in interstitial spaces such as fencerows, field boundaries, and drainage channels, as well as between row crops, in fallow croplands, in grazing lands, and in forest remnants and patches. The maintenance of complex patterns of land use and cover was essential to the survival of many plant and animal communities within agrarian landscapes that had already been largely deforested. Agricultural intensification and expansion directly threatened this landscape complexity.

An invisible (and irreversible) loss of genetic, species, and cultural diversity proceeded as the landscape was increasingly brought under cultivation, yet farmers initiated many counter-efforts to restore diversity by encouraging the survival of specific plants and animals in the remaining "spaces between." Even so, scores of plants and animal species and locally adapted varieties disappeared from the landscape and local land-use systems, diminishing the biological basis of future land-use options and rural livelihoods. As species and varieties vanished from the daily experience of local people, an attendant loss of knowledge and skill occurred (Rocheleau 1992).

Land-conservation and rural-production initiatives both tended to ignore the issue of biodiversity, in spite of the importance of wild plants to the diet, health, economy, and ecology of rural farming communities in Ukambani. Conservation efforts likewise overlooked the importance of both plants and animals in agrarian landscapes in favour of the concentrations of large mammal populations of global scientific and economic (touristic) interest. The 1984-1985 drought and famine brought renewed interest in indigenous plants and "minor" forest products (KENGO 1989a,b; Rocheleau et al. 1985; Wachira 1987), which attracted some national and international recognition. The connection between biodiversity and local livelihoods in agrarian landscapes, however, remained tenuous in official circles beyond the community level, and the issue was relegated to activist groups and indigenous non-government organizations (NGOs) as in many other parts of the globe (Altieri and Merrick 1988; Fowler and Mooney 1990; Oldfield and Alcorn 1991). The maintenance of diverse landscapes with a variety of wild plants and small animals persisted as a matter of local concern (IFPP 1991; Rocheleau 1992; see Juma 1989a on a parallel process in Bungoma district and Richards 1985 on West Africa), but was not yet recognized as a subject of national and global significance.

Drought and famine

In the midst of renewed national and international concern over vanishing wildlife, deforestation, soil erosion, and threatened energy and water supplies, the people of Machakos and Kitui districts experienced one of the most devastating events of the century, the drought and famine of 1984-1985. Over the course of two crop failures in 1983-1984 and the nearly total disappearance of green fodder sources during the dry seasons of 1984-1985, the Akamba suffered up to 60 per cent reductions in livestock and liquidated many of their hard-won assets in order to purchase food. Many farmers lost their draught animals (oxen), which hampered their recovery from the drought and reduced their ability to cultivate their croplands for years after (residents of Machakos and Kitui, personal communications; Kamau 1989). Some people, particularly children and the elderly, died; widespread acute and chronic malnutrition in Machakos, Kitui, and neighbouring districts left lasting effects on infants and young children deprived of adequate nutrition at crucial stages of mental and physical development (S. Saito, personal communication; Ndegwa, 1989; Neumann et al. 1989).

In Ukambani the drought began with the failure of the 1983 October-December rains, which resulted in crop failure as well as stress on livestock during the subsequent dry season. By the failure of the March-June rains (which affected most of the country), the moister zones of Machakos and Kitui districts were experiencing serious and widespread food shortage and income losses, while people in the drylands suffered from food, water, and fodder shortages.

Many communities in the drier parts of Ukambani were accustomed to losing one crop in three, or even one of every two crops. They had normally offset this loss through livestock sales and the purchase of food at "normal" prices from the cities or the neighbouring highlands. In the case of this severe drought (the worst since 1930), however, the livestock were dying for lack of fodder. In eastern and central Kenya (which includes Ukambani), stock reductions in agro-ecological zones 4 and 5 ranged from 26 to 51 per cent of cattle and 23 to 57 per cent of goats, with most of Machakos and Kitui toward the high end of that range since they had already experienced crop failure and poor rains the year before (Anyango et al. 1989). Since the drought was also national in scope, livestock prices were depressed in a nationally glutted market. For the same reason (for a time) virtually no food was available for purchase, except at vastly inflated prices beyond the means of most rural people (Anyango et al. 1989; Borton 1989; Downing et al. 1989).

By August 1984, many normally "permanent" water points went dry, forcing already overburdened and weakened women and children to trek longer distances for water. In the drier parts of Kitui, men and women began to rotate water-collection and livestock-watering journeys, and some people made the long treks at night to conserve energy and to avoid the midday heat. In response to the low prices and lack of transportation for livestock, Akamba men and boys began long livestock drives on foot to deliver large herds of cattle and goats to Nairobi and the nearby Kenya Meat Commission plant at Athi River. Rotting carcasses lay on the roadsides of the major arteries linking Kitui and Machakos villages to Nairobi, attesting to the prior condition of the livestock and the stress of the journey (Weekly Review 1984).

In spite of their experience with periodic drought, the Akamba were hard hit. Many experienced famine because of their distance from the centre, their peripheral status relative to other affected districts, and the erosion of their prior self-sufficiency in responding to drought under extreme conditions. In Machakos, 1984 was known as the famine "I Shall Die with the Money in my Hand" - Nikw'a Ngwete - (Alice Mwau, Japheth Kyengo, and Francis Lelo, personal communication), reflecting the painful irony of changing times. Eventually people were forced to fall back on their traditional practices in local space, or to wait for the state to bring relief. Although a large body of knowledge persisted, the people had lost much of their local knowledge about famine foods and fodder reserves (Rocheleau 1992; Rocheleau et al. 1989).

In 60 per cent of households, women were left (with or without cash remittances) to fend for themselves, their children, and their livestock. Thus in 1984 the feminization of poverty in Ukambani expanded to include the feminization of famine and of response to famine. Women sought the advice of older men and experimented widely to determine which plants could serve as fodder reserves for livestock; this was previously part of the knowledge and responsibility of men (Rocheleau 1992). Even when they discovered the traditional drought-coping strategies, women often learned that the land-tenure reform process had converted special fodder and famine food reserves in the plains to cropland for new settlers. Land-tenure reform had also constrained (but not completely restricted) local and regional mobility for both people and livestock.

In fact, individual migration prompted by drought provided some indication of the relative impact of this event in different agro-ecological zones of the region. In eastern and central Kenya one or more persons moved away owing to the drought in 38 per cent of all households in zone 5, as compared with only 7 per cent in the upland coffee zone (zone 2), 21 per cent in zone 3, and 26 per cent in zone 4 (Anyango et al. 1989). Overall, reliance on migration and remittances was greater in the drier zones, reflecting less wealth and the differential vulnerability of cattle in zone 5 versus tree-crop and non-farm assets in the wetter uplands.

As with the neighbouring pastoral Maasai, the differential impact of the drought among communities was determined by: ecozone; population density; prior conditions of people, livestock, and land; wealth; and the extent and strength of social support networks (Grandin, de Leeuw, and Lembuya 1989). At the household and individual level, the depth of the crisis depended upon the diversity of income streams; the stock and fluidity of assets; the strength, complexity, and geographical range of social networks; and mobility within those networks. Prior to the colonization and sedentarization of the Akamba, the whole community, as a group, might have moved or sent its large herds to distant, well-watered sites. Although that option no longer existed at the community scale, the principle still held at the household level during the 1984 drought. Many farm families in the dry savanna zones sent livestock, and sometimes family members, to stay with relatives in more well-watered areas. Likewise, many rural families sent some of their members to stay with relatives in towns and cities. Others relied heavily, and more than usual, on cash from family members working in urban centres. At a more local scale, many people shifted their focus of activity from cropland to herding, or from both cropland and grazing land to dry woodlands and in-between spaces (Rocheleau et al. 1989). For the more vulnerable members of the population, however, these options were often either unavailable or insufficient to stave off hunger, illness, livestock loss, or distress sales of household and personal assets.

The differential vulnerability has been distinguished as food shortage (less than 66 per cent food self-sufficiency on farm), food poverty (insufficient income to purchase food and/or inability to acquire food through kin and other social networks), and food deprivation (food shortage serious enough to cause wasting, stunting, and vulnerability to disease, or to force migration) (Anyango et al. 1989, 186-187; Kates et al. 1988). According to the figures compiled by Anyango and colleagues (1989, 187 and 208), in the six districts of eastern and central Kenya the proportion of the population that was both food short and food poor was 10 per cent in average years and 30 per cent in drought years. The figures are even higher for Machakos and Kitui owing to the relatively higher proportion of semi-arid land. By January 1985, a survey found that, throughout the six-district region, 50 per cent of children were malnourished, 14 per cent severely. In zone 4, only 14 per cent of children were healthy and 64 per cent were malnourished. In an earlier background survey, 30 per cent of the children in Kitui and 23 per cent in Machakos showed signs of stunted growth, indicating widespread food deprivation (CBS 1983). Overall, the studies point to a high and growing vulnerability to severe malnutrition among children in zones 4 and 5, owing to rising population densities and inadequate diversification of household access to food among the poor (Anyango et al. 1989; Downing 1988).

At least 300,000 people in Kitui and as many in Machakos received food aid from the Government of Kenya relief programme in 19841985 (Deloitte, Haskins, and Sells Management Consultants 1985). District-level figures indicate that 35 per cent of the population in Machakos and 95 per cent of that in Kitui received food aid (Anyango et al. 1989, 184). Many others moved, relied on assistance from family members, or received assistance from church and other nongovernment organizations outside of the main food programme. Imported food was distributed as direct food relief (15 per cent), through market channels (62 per cent), and through a variety of post-relief programmes (23 per cent) (Borton 1989, 27). Much of the relief and post-relief food was dispersed in "food-for-work" programmes, which in Machakos and Kitui often involved construction and rehabilitation of soil- and water-conservation structures as well as tree planting. By the end of 1985 most households and large tracts of barren landscape in Machakos and Kitui were on the way to varying degrees of recovery.

In retrospect, most analysts attributed the avoidance of widespread famine throughout Kenya to the simultaneous use of market channels (primarily in the wetter uplands) and relief programmes (mainly in the drylands) to distribute food imports and to the flexible collaboration of national, international, and non-government organizations in food delivery and distribution. Some analysts noted an unnecessary national vulnerability to drought owing to: (1) overdependence on maize; (2) the extension of maize cropping (and expanding populations of farmers) in areas of uncertain and irregular rainfall; (3) the failure of livestock marketing mechanisms to allow for rapid (and remunerative) offtake at the onset of drought; and (4) lack of infrastructure for small-scale water storage and irrigation (Wyckoff 1989, 366367).

At the local level, rural people in Ukambani attribute their successful survival of the famine to:

• reliance on their indigenous plants (and their knowledge of them) for food, fodder, and medicine;

• reliance on cash remittances from family members already earning wages (and perhaps residing) elsewhere;

• migration of one or more resident household members in search of employment or simply alternative sources of food and support in another household;

• use of group work and group contacts to secure access to official and external sources of relief aid; and

• mobilization of family, clan, church, and other networks of mutual support to secure employment, loans, outside relief, or temporary residence and food for both people and livestock.

Field observations, interviews, and published accounts all support the importance of political and social networks in gaining access to the relief food in direct aid programmes. This was true for church and NGO sources as well as for deliveries channelled through district and local officials (Borton 1989). In many cases, social networks served as crucial information links to connect centres of supply and specific communities or households in serious need of aid.

Self-help responses to environmental crisis

During the 1980s, the people of Ukambani, and women in particular, attracted national attention for the scale and degree of commitment of their (apparently) communal self-help group efforts in soil and water conservation, reforestation, and on-farm tree planting. The government had encouraged local groups (mwethya and other similar groups) of traditional origin (Mbithi and Rasmusson 1977; Mutiso 1975) to register as official organizations and to take responsibility for development and conservation efforts in rural communities (Bahemoka and Tiffen 1992). At the time of registration by the government in the early 1980s, most of the active self-help groups were women's groups that functioned as associations of individuals engaged in reciprocal work for mutual benefit at individual and household level. They were engaged primarily not in communal or public works but rather in shared work on private land or in small group enterprises (Rocheleau 1992).

Most of the groups focused on rotational group labour to weed croplands, thatch houses, make bricks, build and repair terraces, gather fuelwood, and construct and repair fences for members. Some groups specialized in marketing crafts and produce and provided rotating credit to members (Bahemuka and Tiffen 1992; Muzaale and Leonard 1985; Rocheleau 1985; Wijngaarden 1983). During the 1980s, government demands expanded the mandates of these groups to public works, including up to three days per week on road repair, gully repair, land rehabilitation at degraded sites, construction and maintenance of water impoundments, tree propagation and planting, and construction of primary school classrooms (Rocheleau 1985). A recent survey documented 115 groups in Machakos alone engaged in conservation activities and 345 involved in income-generation projects (Bahemuka and Tiffen 1992; Ondiege 1992).

The self-help groups played a crucial role in procuring, maintaining, and protecting resources for the households of members, and they conducted much of the soil-conservation and land-rehabilitation work at community level during the 1980s. During the drought and famine of 1984-1985, the women's self-help groups connected many of the rural poor in the drier parts of Ukambani to local sources of food, fodder, fuelwood, and employment as well as to sources of relief food (both government and NGO). Although the groups provide leverage for women to gain access to resources otherwise beyond their reach, it should be noted that they often do not represent or serve the very poorest women. Lack of time, child care, and member ship dues may keep the poorest and most isolated women from participating as members (Bahemuka and Tiffen 1992; Rocheleau 1985). Owing to the very visible nature of their soil-conservation and tree-planting work and the strong government and international support for those initiatives, many groups secured needed relief assistance for their entire communities as well as for their neediest members (Rocheleau 1992). They also attracted development assistance for many long-term endeavours in forestry, agriculture, water management, and income generation. In the process, women's groups acquired technical resource management and production skills and strengthened their political leverage to gain and maintain rights of use and access to shared resources on both public and private lands (Rocheleau 1992).

In spite of substantial state support for group activity at the community level, these groups were not linked to planning and policy decisions at the regional and national levels, although Tiffen and Mortimore (1992, 381) report that groups in Machakos "have learnt to pull in capital and expertise from national and international sources." Support for specific activities was contingent on the coincidence of state and local objectives, particularly with respect to watershed protection and rehabilitation. National government responses to soil and water degradation of national concern (sedimentation of dams) affected the timing and nature of local conservation activities. Strong incentives (financial reward, food for work, access to government services) promoted group work on gullies and highly degraded hillslopes, rather than on private farms. In contrast, local self-help responses to land degradation integrated farm- and community-level concerns for food security, water supply, fodder, fuelwood, commercial production, and environmental quality. Nevertheless, these priorities did not directly affect national-level programmes. The 100-year history ends, as it began, with a region caught between two worlds, in this case the distinct realms of state and local power, experience and perception.

Contents - Previous - Next