This wo rk is lic e ns e d unde r a Cre a tive Co m m o n s Attrib u tio n -N o n Co m m e rc ia lNo De rivs 3 .0 Lic e nc e . To vie w a c o py o f the lic e nc e ple a s e s e e : h ttp : / / c re a tive c o m m o n s . 0 rg / iic e n s e s / b y-n c -n d / 3 . 0 / THE EFFECTS OF INCREASED CROP PRODUCTION ON LIVESTOCK INVESTMENTS IN A SEMI-ARID AREA: SOME EXAMPLES FROM BARINGO DISTRICT, KENYA. By Peter D. Little WORKING PAPER NO. 386 : . . . . INSTITUTE FOR DEVELOPMENT STUDIES ^ U N I V E R S I T Y OF NAIROBI .P.O. BOX. 30197 NAIROBI s Kenya August 19.81 Views expressed in this paper are those of the author. They should not be interpreted as reflecting the views of the Institute for Development Studies or of the University of Nairobi. This paper has protection under the Copyright Act, Cap. 130 of the Laws of Kenya. -4'- IDS/WP 386 The Effects of Increased Crop Production on Livestock Investments in a Semi-Arid Area : Some Examples from Baringo District,, Kenya By Peter D. Little Abstract The paper addresses the major premise of the arid and semi-arid lands (ASAL) program in Kenya which states that increased crop production in the semi-arid areas will lead to a reduction in livestock numbers since dependence on livestock for subsistence needs will be reduced. Theoretical support for this position derives from a "school of thought" in pastoral studies which posits that indigenous livestock production systems in Eastern Africa are maintained mainly for the purpose of supplying a steady source of food (i.e., pastoral products) for the pastoral population. Evidence from Njemps Location, Baringo District suggests that while increased crop production has substituted grain for pastoral products in the diet, it has not undermined the role of livestock as a form of investment. It is further argued that a focus on "statistical" averages in livestock holdings, rather than the range and variation in property ownership, informs development planners little about which group or groups of livestock owners will respond to increased opportunities for agriculture. Capital and labour requirements for irrigated agriculture in Njemps are such that the wealthier livestock owners are the individuals which are most capable of taking advantage of crop production aopportunities. The dilemma in Njemps Location arises due to the fact that groups of livestock owners are responding to irrigated agriculture for different reasons: the rich pastoralists are investing in agriculture as a mechanism for supporting their livestock holdings, while the poorer families are being forced into agriculture out of necessity to meet subsistence needs and these latter households have only a minimal impact on the livestock sector. Because of this second group of households it is argued that crop production, especially low cost methods of dryland farming, should be encouraged in Njemps Location, but should not be perceived as a mechanism for supporting range management programs. 5 — IDS/WP 386 INTRODUCTION A commonly held belief among policy planners and. development specialists in regard to the; development of Kenya's arid and semi-arid lands .(AjlAL) is that an increase in.food crop production in these areas will lead to a reduction in livestock numbers:-(.of'* Republic af-Kenya,.'1978; Ottley, Wanjaiya and Martin, t978jb .This' premise.,in faot? forms the basis for Kenya's present development, efforts in the arid and semi—arid landsand it is believed that through inoreased, crop production, soil conservation and livestock development needs will be served beoause grains will be • substituted.for pastoral products in the diet, and thus livestook numbers reduced* To quote one. source (Republic of Kenya, 1980:- 124): "The food crop development relates closely with livestock development in that when a fg^mer has sufficient food, his relianoe on livestook is reduced hence indirect destooking," Increased foSd production in the semi^-arid areas, then, is -viewed as an important mechanism for solving the major perceived development.'obstacle in the pastoral areas, inadequate resource management practices often characterized by overgrazing and range degradation. This particular perspective emphasizing the subsistence aspect of indigenous livestock derives theoretical support from numerous scholars which argue that Eastern .African pastoralism is a system which functions .• mainly- to insure adequate food production for human consumption. Some very valuable works have resulted from this "school of thought, most noteworthy is Dahl and Hjort's (.1976) b&ok which represents the most ambitious work to date on African livestock production. Another is Brown's (1971) article relating human diet requirements to pastoral herd sizes. In the. latter work, the author simulates minimum cattle herd sizes necessary to. :• keep a pastoral population alive on a diet of varying proportions of milk and meat0 Both of these works, as. well as others.(of, Pratt and Gwynne^ 1977) j. emphasize only, one of the two roles of indigenous livestock-in Kenya; that is, the food commodity aspect of livestock. Another, and in. some areas of Kenya,more important dimension of livestock is that as a form of investment, .or capital, (cf. Schneider, .1981), In this respect, investing in livestock is often achieved at low costs and earns 'lint'creSt" in<the form of reproduction (oalves' or kids). As Dahl (1979; 54) points out in. a more reCcJnt Work, the two dimensions of indigenous livestock ace not contradictory in that both aspects emphasize reproduction and hehce herd ' structures" which ore biased toward adult females. The-degree to which .IDS/WP386 - 2 - either dimension of livestock dominates the other depends partially on the extend to whioh the pastoral population- can.be food self-sufficient on pastoral.pro ducts alone. In areas where agricultural markets are readily accessible and alternative foodstuffs have become an important part of the pastoralists' diet, livestock's role as a food commodity will be minimized. The capital aspect of livestock, in turn, will be less important in regions of Kenya where distribution networks of alternative foodstuffs are less developed and pastoral populations thus are compelled to be more self-sufficient. In the latter case, such areas would include the Northeast and parts of Turkana District, It would also be expected in these .areas., that full utilization of pastoral products such as blood, sour milk and ghee would be greater than in regions where market integration is more advanced. This is well illustrated in Baringo District where the, II Ch.am.us (Njemps) that have greater access to grain markets do not as fully utilize ghee and animal bleeding potential as those II Chamus residing in the more remote areas where grain distribution networks are less developed. As will be dis- cussed in the paper, it is actually.the former group of II Chamus which are the wealthier . livestock owners ind"consequently more capable of subsisting only on pastoral products. Whether any pastoral groups in Kenya were able to ever subsist throughout the year on pastoral products alone is questionable. Nevertheless, the present trend in most arid and semi-arid areas of Kenya is toward increasing diversification of the local diet. In regard tc Kenya's current development efforts in the ASAL, a program which encourages the substitution of grains for pastoral products in the diet through increased production of food crops confronts only the food commodity Cu aspect of livestock and does not undermine its role as a form of investment. In areas where livestock is often sold in order to purchase grain, crop production may actually support the buildup of live-stock since it will reduce cash demands tit-.the household levels ...Moreover, if land for crop production does not directly compete with land for pastoral use and if livestock are grazed Qn harvested fields, then crop production may further enhance the livestock sector by increasing available animal fodder. ,A A. * '. •.-.. b v i ». The present paper examines the impact of increased crop production on livestock holdings in a semi-arid area of Kenya, focuses More specifically., it. onrthe II Chamus area (Njemps location) of Baringo District, a re- gion which is currently receiving substantial development inputs. argued in the paper that distinctions between agriculturalists and It is - 3 - . IDS/WP 386 pastoralists or the "agricultural sector and the pastoral sector which are frequently made in the literature often "mislead people into believing that crop production in pastoral "areas will only attract PQ°r stockowners. Movements from the pastoral sector to the agricultural sector are usually described as being' limited to the poor stockowners, especially those who have recently lost their herds, often perceived ouch transformations to agriculture are as a mechanism for helping the destitute stockowner to build tip sufficient surplus in order to reinvest! it into th<? pastoral sector. That shifts of this type may also be a strategy among wealthier stockowners as a means for supporting their present livestock holdings is not usually considered. In addition, there has also been.a tendency in studies of pastoral, and semi-pastoral areas to focus on "statistical averages" rather than the variation and range in livestock holdings which is more valuable information in regard to development planning. far The failure to recognize the great disparity in livestock ownership in pastoral areas has led some scholars to overemphasize the egalitarian nature of pastoral societies which in the political sense may be true, but in relation to equal access to property is not the case. The findings and suggestions presented in this paper are not assumed to be applicable to all other serai-arid areas of Kenya, 2 to other pastoral ways. . . areas of Kenya N^emps is unique m In fact, relative • a numher^of different The II Chamus area of Baringo District, for example, has always had more crop production than most other pastoral areas cuad also has greater •J;# The fact that opportunities for economic advancement ("rags to riches syndrome) are.: available in livestock areas ' * v. ; ^ relates, in part, to the egalitarian political structures of these societies (cf, Schneider, •J980), Such social mobility in these systems is: becoming increasingly restricted,, •t 2, The term pastoral in this context refers to a system where livestock production is the predominant source of income or livelihood for the majority of people. Finer distinctions can easily be made, and a term such as agro-pastoral or semi-paStoral may more accurately define . contemporary II Chamus economy where: approximately one-third of the families now have access to.farms,- However, as will be discussed later in the paper, most of the.renewed interest in agriculture has occurred in the past fifteen years, .... ID3/WP 336 irrigated agriculture potential than raost semi-arid regions of. Kenya, Eon this reason, as well as the fact-that the area lies in close proximity to high potential agricultural regions and markets, grain has always been an important part of the II Chamus diet, especially in regard to the dry season. In addition, Njemps is somewhat unique in that the pastoral economy revolves around a highly productive seasonal swamp area, of approximately twenty square kilometers. This dry season grazing area circumvents the shores of lake Baringo and extends from the Pekerra-Molo basin south of Lake Baringo to an area approximately two kilometers north of Lake Bogorio . The presence of this large swamp allows the II Chamus, in most parts of Njemps Location, to be relatively sedentary; perrenial water sources such as the Perkerra River and Lake Baringo means that stock watering distances in the dry season in most areas of Njemps do not exceed seven kilometers. Nevertheless, that Njemps Location represents a region where crop production has increased considerably in the past fifteen years may provide some lessons and guidelines fox- other semi-arid areas of Kenya where development assistance in agriculture is planned. The data presented in this paper are drawn from eighteen months of fifeldwork conducted mainly in the Ngambo, Salabani and Eldume sub3 locations of Njemps, Two sample sets of twentyApight randomly selected households were monitored for labour use, consumption, cash expenditures, production and marketing.for periods of seven months in both 1930 and 1981. Households represented art.-as where crop production is a reCent phenomenon and regions where it.has a relatively long history (over twenty- years).4 Numerous surveys concerning such topics as residence histories and kinship organization were also conducted in these three sub-locations as well as the other sub-locations of Njemps, Mukutani and Loininange. In addition, marketing surveys were carried out in the Tugen areas of Ravine Division (south Baringo) and in Loboi, Chapchap, Ewalel and Saimo Locations of Marigat and Kabartonjo Divisions, This was done to determine ther- •amount of food commodity trade between these more agricultural regions and Njumps Location. Sample sizes in the extensive surveys were often over 100 housuholds and in some cases as large as 250 households. Much of the information 3. This study was funded by a doctoral fellowship from the Social Science Research Council, UoA. Their support is gratefully acknowledged. The views expressed in the paper are those of the author and do not reflect the views of the funding institution. 4. Until the beginning of the twentieth century, the II Chamus were predominantly agriculturalists due mainly to the fact that frequent stock raiding by larger groups such as the Turkana at the time made pastoralism very hazzardous. Shifts to cultivation here refer only to the last century. IDS/vTP 386 -zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 5on such important issues as land tenure and family cooperation in agriculture could not be obtained through questionnaires. Thus, the researcher often relied for data collection on what the anthroplogists call the 11 participant-observation" method. The following paper is organized around the themo of increased crop production in Njemps Location and throe main issues will be addressed. These are: 1) The economic, ecological and demographic pressures which have resulted in the past fifteen years in a greater incorporation of grain production inftj craps, 2) The question of who has talc en most advantage of the increased opportunities for cultivation, especially irrigated agriculture: the rich or the poor stockowners, the young or the old household heads. 3) What are the implications of increased crop production for rural development in Baringo District. Background information will be presented in the paper on demography, livestock and grain marketing, household expenditures, coats. of"crop production and other topics which are critical to an understanding of the broader issues outlined above, 1. General Ecology and Demography of l-Tjemps Location The Maa-speaking II Chamus reside in Kenya's Baringo-Bogorio basin and in the hill region extending east of Lake Baringo to the Laikipia escarpment. Administratively, they occupy most of Njemps . Location, Baringo District, an area of approximately 600 square kilometers. The II Chamus homeland is located in Kenya's central Rift Valley and altitude in the region ranges from approximately 3100 to 4500 feet. The ecology of the area is characterized by brush Acacia, especially species of Acacia reficiens, Acacia mellifcra, Acacia Senegal and Acacia tortillas (UNDP, 1967), In some of the lowland areas ("flats"), severe gully erosion has occurred and grass cover is sparse (not including the swamp areas) throughout the year. Common annual grasses found in this region are Eragrostis spp.;, and Tetrapogon spathecos and- forbs such as Jeliotrplum spp.(highly un• palatable) and Portulaca spp. are becoming increasingly dominant as a percentage of total ground cover. The only region in the "flats" area where perrenial grasses are found is in the seasonal swanps-. Dominant grass species there ar^ Cynodon dactyl on, Cynod on p lb c t q»11 achy us, Bchinochloa haplocada and Digitaria velutina. IDS/WP 386 -6'- Average annual rainfall at Marigat (based on the past sixteen years) is 640 mm. and is unixiodally distributed with the largest concentration of annual rainfall falling in the April-August months. Monthly rainfall sta- tistics, however, are misleading in that in any given month rainfall is often concentrated in only a few days of the month. Rainfall patterns in Baringo District are highly localized end monthly variations in any given year may differ considerably from the statistical norm. In Marigat, for example, nine out of the twelve months of the year have been leading rainfall months in at least one of the past sixteen years. Based on Kenya's 1979 population census, the II Chamus population is 6800 and the population of Njemps Location isO8 6 4 2 , The twenty-one percent of norir-Il Ghamus who live in Njemps Location are mainly Tugen* II Chamus population increased more than tliree-hundred percent from 1929 to 1979, an annual increase of approximately two percent (Golony and Protectorate of Kenya, 1930: 16), This was due to a large extent to the incorporation of Tugens as they moved out of the Tugen Hills at the beginning of this century. ten years. This process seems to have slowed considerably in the past Prom 1969 to 1979y the II Chamus population only increased from 5933 to 6800, an annual growth rate of less than one percent. This period correlates with a time that more II Chamus went outside of the District seeking employment and at census time may have been identified as Maasal or Samburu, Approximately, 300 II Chamus are presently employed as unskilled farm labourers outside of the District, The majority of these are working on farms or ranches in Laikipia and Nakuru Districts, Within Njemps Location, regional population densities vary considerably according to economic and ecological criteria-. Human population and population densities are presented in Table 1. for each of the five sublocations of Njemps, Note is also made of the percentage increase in popu- lation for each of the sub-locations since the 1962 census. Table 1. Human Population, Njemps Location 1979 Pop, Sub-Location Ngambo Loiminange Eldume Mukutani Salabani 1979 Population % Increase Since 1962 3116 101 % . 1519 62^b" 810 57oji 'jo.. 17O8 . .46^ 1789 8642 56/0 * Based on 1962 and' 19t9 population censuses. per square km, 66 8 19 8 16 _ 7 IIS/WP 386 - The largest population concentration and greatest increase in human popuv lation in the past eighteen years has been in Ngambo sub-location. Of all sub-locations in Njemps, population has always been highest in Ngambo nainly because of its easy accessibility to swamp grazing. More than forty percent of the twenty square- kilometers of swamp in Njemps is located in Ngambo sublocation. This includes most of the Molo-Porkerra basin, swamp, the main dry season grazing area for the II Chamus. Moreover, water is also readilyavailable throughout the year in the Ngambo region and distances to water sources in the dry season never exceed four kilometers, Factors accounting for the relatively recent growth (1962-1979) in population in Nganbo arerelated to the influence of the Perlcerra Irrigation Scheme and the beginning of smallL-scale irrigated agriculture in the area. Families which have been forced out of the pastoral sector due to livestock losses and do not have the necessary resources to engage in agriculture themselves have settled in the Ngambo area in order to engage in wage labour on the nearby Perkerra Irrigation Scheme, Other, families moved to the Ngambo region in the past ten to twelve years because of.-the initiation of a small-scale, indigenous irrigation scheme which was staried in 1963 and began to operate on a, large scale in the late 1960s. This latter scheme was begun under local initiative and irrigates from the excess, runoff water of the Perkerra Irrigation Scheme. Of all, the II Chamus areas, probability of adequate water for irrigated agriculture in any given year is highest in J [Ngambo, 2_. The Lives took. Dimension, of 11 Chamus Economy 2.1 Herd Sizes.., Herd. Structures- and Herd Ownership Cattle numbers in Njemps were greatly reduced during the 1979—1980 drought. Estimates based both on hides and skins sales records... for the drought period and interviews with individual livestock owners indicate that the II Chauus lost approximately sixty percent of their cattle. The majority of cattle mortalities occurred in the last quarter, of 1979, All sub-locations of Njemps except Eldume, lost at least fifty percent of their cattle, Regional variations in herd depletion-ranged from forty percent to seventy percent of total herds. Causes for cattle mortalities were usually related to one of two factors, East Coast Fever or drought (lack of grazing), the more important of the two varying in different areas. In Mukutani s,ub—loca- tion, for example, the main cause of cattle mortalities identified by the local people was East Coast Fever (ECF). This was followed in order of importance by trypanosomiasis and lack of grazing. In Ngambo sub-location, on the other hand, the. main cause of cattle deaths was-said to be lack of zwvusrponjigebaM grazingj"ECF was identified as the main reason for mortalities by only twenty- _ 8 four percent of the infomants. irs/wp 386 It should be renenbered that it is sonetimes very difficult to analytically distinguish between drought and disease as seperate causes of cattle deaths when in fact in a drought year they are often closely related. That is to say, poor nutrition will exaggerate an animal's susceptibility to disease and there will be a sig5 nicantly higher mortality rate if grazing is ina.deq.uate. In total, causes of cattle mortalities identified by II Chamus stockowners indicate that forty-nine percent of stock deaths in the 1979-1980 period were related to East Coast Sever, while forty percent w e re caused by lack of grazing and a further eleven percent were predominantly a result of Trypanosomiasis and other stock diseases. Present herd composition and structures in Njemps reflect the recent occurence of drought and stock disease outbreaks in the region. In the sample areas of Eldume, Salabani and Ngambo sub-locations, average herd sizes per family of seven (rt= 71) are nineteen cattle, twenty-nine goats and seventy-three sheep. Standard deviations for all stoek categories are very high, e.g., twenty-six feozteattle, thirty-six for goats and one hundred and eighty-five for sheep, and indicate a highly distorted distribution pattern.Of the three sub-locations, the largest average cattle herds are found in Ngambo with an average of twenty-one cattle per average sized family (seven members), while the largest average sheep holdings, 109 per family, are in Salabani sub-location. The high proportion of sheep in herd compositions are indicative of the recent drought and stock disease outbreak which only affected sheep minimally. An estimated thirty-five'percent-of total goats, however, died during 1979 due to an outbreak of Contagious Caprine pleuropneumonia (CCPP). The. large numbers of small stock presently in Njemps may also be a function of a common post-drought strategy for-j?obuilding herds among the II Chamus;. that is, investing in small stock initially because of their fast rate of reproduction and then selling them and buying cattle or exchanging the small stock directly for heifers or young bulls. 5. Periodic depletion of herds on a large-scale are not unusual in Njemps and in each of the 1960, 1965 and 1973 droughts an estimated thirty to thirty-five percent of total cattle were lost, 6. In addition to the factors discussed above, average stockholdings in the II Chamus area are low because of the fact that many very poor households(in terms of livestock) are able to remain'in the area due to the opportunities., either to engage in wage employment at the Perkerra Scheme or to work on the large irrigated farms of the wealthy stock owners. The presence of these families greatly reduces the average livestock holding figure. - 9 - . IDS/WP 386 Present herd structures for II Chanus cattle are also indicative of the 1979 livestock disasters in lljenps, This is especially in regard to the low percentage of adult cows represented in II Chanus herds.. the aniaals nost affected-by the drought and disease outbreaks. These Y/ere Table 2, depicts present herd structures for II Chanus cattle? sheep and goats. It shoiild be n£ted that castrated goacs and sheep were not distinguished froa other nalu natures because of the difficulties in dealing with the large flocks of snail stock in igenps, The two flocks which I did identify revealed that seventy percent of nale nature-Sheep and goats were castratest 7, Based on data from Range Management Piles, Ministry of Livestock Development,. Marigat, Baringo District, _ 10zwvusrponjigebaM - IIS/WP 386 The large standard deviations cited earlier for livestock holdings in Nj.enps are evidence of a highly; unequal distribution of livestock ownership. Such p pattern makes it extremely difficult to discuss "averages" in regard to II Chamus livestock holdings. Table 3. below demonstrates just how polarized livestock ownership is in Hjemgs, figure 1. illustrates the relationship on a graph, For calculation purposes, a stock unit (S,U,) is assumed .to be six small stock or one head of cattle. This best approxi- mates the market value of the animals in Nj emps, AS; the above table indicates, tlie richest twenty percent of livestock owners in the sample control sixty-five percent of total stock units. these families have average herd sizes of over 100 stock units. And The poorest fifty percent of households, in turn, have access to-only-ten percent ofO -7total stock units in Njemps, Families in these percentiles have average herd sizes of under 10 stock units. At present, five percent of II Chamus families hav,e no stock at all and are mainly employed as casual labourers for the-rich stockowners or for the Perkerra Irrigation Scheme, 8, As will be shown later in the paper, livestock ownership is slightly' more polarized in areas where crop production is important than it is in regions which are more strictly focused on livestock production. In Table 3,, the sample was drawn from areas where cultivation is important, i.e., the Ngambo- and Eldume areas. * . IDS/WP 386 - 11 2.2 Livestock Marketing •Despite the stratification which exists in terns of livestock ownership, incone iron the sale of livestock accounts for the najority of nost household incones in Njemps and is responsible for approximately sixty/.. percent of total ce.sh incone in Njemps, Only in the botton three percentiles of households (botton thirty percent in Table 3.), does non-pastoral inccaae, in this case off-farm employment, account for a greater proportion of average household incone than cash earned from livestock sales. In terns of all livestock owners represented in Table 3.? average annual incone per household fron the sale of livestock is 1298 Kenya shillings. As would be expected, however, the range in incone frco stock sales varies considerably between the rich and poor families. In the sample of seventy-one families, household cash earned frco the sale of stock ranged from 0 to over 9000 Kenya shillings p_r year. Por reas dns to bo discussed later in the paper, cash incomes per household from tile sale of livestock do not increase proportionate to the size of herd. At present, small stock sales in Njomps account for approximately two tines as nuch cash incone as that earned fron cattle sales. How tiie II Chamus respond to livestock marketing is a very complex issue and concerns a number of factors such as the condition of range and livestock at the time of sales, relevant social factors such as brideprice demands and livestock loaning, size of the herd, cash needs and the price being offered at the market or by the trader (cf. White and Maedows, 1979). In a survey of ninety-four herdowners, fifty-one percent of the sample said that price was the major factor influencing whether or not they marketed an animal, while forty-nine percent cited cash needs as tile most important criterion for determining stock sales'.' Actual marketing patterns in Njenps fortthe past twenty-five years reflect these findings in that the dry season month of December, when cash needs are highest because of -cJao decline in milk production and increased dependence on grains, is also the largest volume month in terms of stock marketing (cf. Utile, 1981: 83). zwvusrponjigebaM Moreover, this is also the month when livestock prices on average are highest in Baringo because of the increased demand in other ports of Kenya for meat during the Christmas period. This is well illustrated in the Baringo case where, for example, prices for small stock at the Baringo , County Council Auction (Marigat) in December, 1980 were approximately twentyfive percent higher than for any other month during 1980, Months, in turn, - 12 . IDS/WP 386 - when volume of marketed stock has been lowest in.the past twenty-five years are the wet season months of April and May. This is the case" in-spite of the fact that there is not a significant drop in price from other months with the exception of December. needs begin to April and May are the months when aash ocline due to-the availabilityof milk and alternative food- stuffs such as wild vegetables.. It is- also a time during the year when range conditions are especially good and II Chaaus herdowners are oft.O .11 • 9 purchasing livestock during these months. The formal livestock marketing system (Baringo County Council) of Baringo District is responsible for the marketing of the majority of j cattle in Njemps Location, but less than twenty percent of marketed small stock. The largest percentage of marketed small stock are sold to . traders from south Baringo and a few from Nakuru District. Outside of Mari- gatsthere are no butcher shops within Njemps Location; however it is cccmon for a local herdowner to slaughter a shoat in the bush and sell the meat at approximately eight Kenya shillings per kilogram (bene included). It is also a standard procedure in some areas of Nj^mpe ±ur small stock to be exchanged directly for maize at the price of two ''debes"(32 kg.) of maize for one small stock of approximately seventy shillings value. changes, the price of maize remains constant In such ex- at 2 / 2 5 shillings per kilogram and if three "debes" of maize are exchanged then the goat exchanged must'bo worth approximately one hundred shillings, Table 4 . depicts the percentage of stock sold tirrough the various available marketing channels in Baringo District and the average prions paid for the stock. The table is based on data collected monthly from fifty- eight II Chomus households during the period June, 1980 to February, 1981, 9. This is obviously an oversimplified discussion of II Chamus responses to livestock marketing. A more detailed, quantitative analysis of livestock marketing dealing with price responsiveness and other relevant factors will be included in my dissertation thesis. IDS/WP 386 -13'TableO4 . Livestock Marketing, H.iomps Location Market Outlet """ " " " %Total Cattle Total Shoats County Council 73/o 18% Private Traders 27% 52% AVg. "Catt. Price(shs< ) Avg. Shoat Price ,(shs. 492 (280-1180) 92 (30-170) 429 (20 0 -680 ) Slaughtered and Sold (Bush) 16% Exchanged for Maize 14% 82 (45-145) 32 kg. naize 448 (200-1180) Averages 81/55 ( 5 0 - 17 0 ) 85 (30-170) *G-oat; prices tend to be twenty percent higher than prices for sheep. As table 4t indicates,.average prices per head of cattle and snail stock are respectively 448 shillings and 85 shillings, low for nost livestock producing regions of Kenya (cf. Matth.es, 1979). Cenparitively, livestock prices in Kitui District, for example, are almost two tines the price of those in the II Ciianus area (personal communication, Janes Ayres). On a per kilogram basis (liveweight), average producer price for cattle at Marigat auctions is 2/20 per kilogram. This explains, in part, why the II Chamus, unlike most other ^roups in Kenya, very much favour selling cattle to the Livestock Marketing Division (IMD) with its present producer price of 3/20 per liveweight kilogram. The low stock prices paid at the Marigat auctions seen to be related to nonopsonistic buying at the next stage in the marketing hierarchy when the cattle are sold in Nakuru District, Although present prices for livestock are low, there seems to have only been minimal changes in stock prices at Marigat in the past eight years. This is surprising given the general inflationary trend in recent years for other food products in Baringo District. Maximum and minimum livestock prices for the T974-198I period are given in Table 5} average prices are not available for this period. - 14 - . IDS/WP 386 As the table indicates, there has been very little change in maximum prices at Marigat in the past eight'years, indicating perhaps a general deterioration in the quality of livestock during this period (personal communication, Jeffrey Lewis). an increase since 1974. .— Minimum cattle prices " V .) lliWe, however, shown . . * * _.,.. lii =ts.rms of small'.Stock, tnere has been very little increase in either minimum or maximum producer prices since 1974. It should be noted that the new tarmac road to Baringo District from Nakuru should increase livestock prices in the area in the near future. During this same time period (1974-1981), institutional factors related to marketing have also adversely affected livestock producers in Njemps. These are mainly considerations which are exogenous to the realm of control of the individual stockowner. For example, on an average of five months in each year since 1975, livestock auctions at Marigjat have been 0 cancelled due in most cases to Foot and Mouth disease outbreaks. At the producer-level such closures make it very difficult to plan marketing strategies; most marketing bans occur without prior warning and are usually lifted in the Same fashion. This is especially problematic when the producer out of necessity is linked to a cash market, especially in regard to the purchase of foodstuffs such as poslio (maize meal). Not surprisingly, the II Chamus view livestock marketing as a major problem and a survey conducted in 1980 (n = 70) revealed that eighty-nine percent of them wished that cattle auctions were held more frequently. ire/wp 3. Crop Production Dimension of 11 Chanus Economy 3,. 1 Increased Crop Production In addition to livestock marketing, a discussion of the regional grain trade in Baringo District is also pertinent for under-standing, in part, the pressures which have led to increased crop production in Njenps Location, As economists and other social scientists (cf. Lees and Bates, O 19 77; Little, 1980) have pointed out, household production strategies cannot be viewed in isolation fron the marketing aspect, especially in regard to a lives'cock producer1 s decision to diversify his/her household economy (i*e», incorporate more grain production). This is especially relevant in the II Chanus case where, although the prominence of grain in the diet has increased in the past thirty years due to soue extent to increased hunan population resulting in reduced livestock/person ratio, imported grains have always assumed sane importance in the local diet. To quote a colonial source fron the 1940s (Colony and Protectorate of Kenya, 1940: 9): "The supply of posho to the trading centers serving the plains people (il Chanus, Tugen and Pokot) necessitates heavy imports during the dry season and constitutes the bulk of the trade of the district." That the II Chanus, as well as their Baringo neighbours, have always relied on non-pastoral products for consumption, even in the 1930s and 1940s when average stock holdings wore over fifty stock units per family of seven, is probably related to the uninodal rainfall pattern of Baringo District. Such a rainfall pattern results in , an annual dry season of seven to eight months. At this tine of the year, 10 mi 11c production fron cattle is reduced to nil, Posho consunption during these tines increases to .7 kilograms per adult unit, an approximate increase in grain consumption of over four hundred percent of grain consumed in the wet season. At present, the average II Chcaus family of seven consumes 779 lcilograns of posho per year. Translated to the locational level, Njenps Location imports over 8000 sacks (90 kg. each) of naize per annum, the equivalent of 1,440,000 Kenya shillings. While dependence oil posho in the II Chamus diet has increased in recent years, regional changes in grain marketing and landuse patterns in Baringo have adversely affected households specialized in livestock 10. Hunan consumption of milk (cattle, goat and sheep) per adult unit varies from .58 liters t-t 1.75 liters in the wot season to ,49 liters to under ,1 liters in the Aty season. At the end of the dxy season, usually only goats and sheep are rallied and milk production per animal is respectively ,16 liters and ,05 liters. - 16 - . IDS/WP 386 production and thus dependent on markets to buy grain0 Thi most important regional changes outside of -Nj^mps. that .have trJcen place in Baringo in the last ten to fifteen years are: 1) a shift, in South and North Bfiringo's high potential areas to crops such as pyretheun, coffee and wheat; and 2) a decline in private grain trade "and a more oninont role for government parastatalsin the marketing of grain in Baringo. While tne farmer change has reduced the amount of agricultural land devoted to such, food crops as. i maize and finger millet, the latter has created inefficiencies in the present grain marketing system in Baringo, Of the two changes stated above, the most relevant to the II Chamus producer is the recent transformation in Baringo's grain marketing system. retail prices for maize in Njemps, In part, this has resulted in higher The maize producing region of Baringo which has been of most importance to the II Chamus in the past is the Eldama Ravine area., South Baringo. Proa the late 1240s until the 1970s,: most of the maize consumed in Njemps came from this region, The grain was usually transported to the grain deficient areas of Njemps by a network of traders who either bought the maize directly from the farmers or from middlemen. To a lesser extent, such a free trade network also existed in the Kabartonjo area of the northern Tugen Hills„ The first serious market interventions in this intra-uistrict trade occurred in the mid-1960s when the government attempted by establishing government buyers in the area to control surplus maize marketed from South Baringo„ Among some Baringo District officials,such,an initiative by t^e government was viewed as potentially detrimental to the residents of North and East Baringo,, I quote (Republic of Kenya, 1968: l): Formerly, till 1965, lembus Location (South Baringo) , 1/20 the size of Baringo, was considered the gronnary of Baringo District and maize produced there was sold gristed to posho at 45 shillings per bag within the district only0 In order to increase the amount' uf home-grown maize, Perkerra Irrigation Scheme was established during the emergency ^ period with a view to. growing more maize for feeding the famine-stricken community in North and East Baringo. During the said era, maize could not be exported from the district by any means av liable,....The prevailing and expanding price of maize ore artificially induced by Maize and Produce Board which exports maize from Lembus to Nakuru, and then selLs it back to residents of Baringo at approximately 200^ increased -price for posho even before it enters the trader's h a n d s . N e v e r t h e l e s s with maize 22/60 a bag to the grower in South'Baringo and 70/ the same bag, retail in Marigat; the existing state of affairs is most unsatisfactory. 11, Although the Perkerra Scheme was initially devoted to maize, at present over ninety percent of hectarage on the Scheme is devoted to the production of cash crops, chillies and onions„ . '' - 17 - . IDS/WP 386 The astute insights of this "free narlcot" oriented official quoted above did not stop the trend of government control of maize marketing in Baringo, Now in almost all the surplus maize producing areas of the dis- trict government control of marketing is the norm. Traders, who as late as 1978:purchased maize within Baringo District, must now travel to Nakuru in order to buy maize. In terms of transport alone, extra costs at present are /15 per kilogram by the time it arrives in Marigat, Not only have the marketing changes resulted in higher prices for grains in Njemps, the II Chamus have also lost access to the informal maize distribution system which 12 in many cases was more dependable than the present system. The market and demographic pressures cited earlier have, as would be expected, have led to a greater diversification of the II Chanus economy. This has meant a steady growth in crop production in Njemps in the past fifteen years. An indication of the present seriousness,on the part of the II Chamus, is their renewed interest in irrigated agriculture. As opposed to dryland cropping, irrigation for the II Chamus has a veiy different connotation. It implies a sense of permanency which is not perceived with dryland cropping,. Irrigated agriculture also involves a far greater investment in effort and orgrhization than is the case with rainfed agriculture. More importantly, unlike rainfed farming areas , irrigated faros can be considered to be individually owned and in some areas is now being bought and sold. Increases in II Chamus irrigated agriculture since 1970 can be documented as follows: Table 6, Estimated Increases in Irrigated Agriculture, Njemps Location: 1970-1981 1970 Area no. farmers 1981 ha. no. fainers ha, Ngambo 82O 49 157 151 LamelOk — — 63 19 Eldune 34 16 38 20 Loininange — — 26 24 iiukus — — 21 22,8 Mukutani 40? 20 52 25 12. Translated to the livestock-producer level, a goat in Njenps was equivalent on the narket in 1969 to 3-4 debes of naize. At present, the exchange rate is two debes of naize per goat, oimilarily, prices for finger millet in relation to livestock have also incroase'd; the amount of finger millet the II Chanus could buy for one goat in 1969 now must be exchanged for two goats. IDS/vTP 386 - 18 Table 6 (cont.) Totals 156 85 357* 251 * Represents twenty-nine percent of all 11 Ciiauus families. Based on ray own data and Baringo Pilot Seni-arid Area Project Crop Survey, 1981. As the table indicates, total irrigated land lias increased over 30C% since 1970, The largest single increase in total irrigated land during this tine period took place in the post-drought years (1980-1981) when cultivated land increased approximately thirty percent. This was especially the case in the Loininange and Rukus areas where there were very heavy livestock losses. During the tine period represented in Table 6,, Hjenps Location has 'gone fron being ninety percent dependent on imported grain to now having to import only seventy percent of its grain needs. Increased opportunities for irrigated agriculture have not been distributed evenly anong households xn Mjenps due to the fact that only the wealthier families who have access to nuch family (i.e., nore than two wives) labour and capital can cultivate large, irrigated faxas. Labour requirements f irrures and production / per . 5 hectare irrigated farm fire listed in Table 7. Table 7. Labour Requii-enonts and Production for Irrigated Agriculture (oer .5 ha.) Field Preparation (pre-plantin^) Mondays Fencing and Clearing 65 Cultivation and Basin 'Construction "' 64 129 Planting, Watering and Weeding Planting 9 Watering and Canal Maintenance 18 Weeding 40 67 Harvesting and Threshiiy; (per 450 kg. unit) Harvesting and Threshing—Maize " " " » « " • . 19 —Finger Millet 76 —Beans* 50 . IDS/WP 386 - 19 Table 7 (cent.) Prod, (kg.) Total = Maize # 2 1 5 : .'642 P.Millet * BeansO 272 364 246 250 *Local varieties ** Manday = 6 hours u Malq/Penale Age 7-14 = .75 Monday Male/Female Age 15-60 = 1 Manday Male/Pciiale Age 61-75 = .5 Manday Most irrigation schemes in Njenps Location are similarily constructed so that laboui inputs per *5 hectare for the different schemes are relatively consistent. All the irrigated farms use a rectangular basin system whereby rows of basins are constructed, each basin being approximately 2 4 5 meters x 4 meters. Between each row of basins, there is a canal where water passes and is directed into each basin by entrances made on the toj;side of the basin. Variations in labour requirements for irrigated agriculture are mainly determined by four criteria: l) the amount of clearing needed it the beginning of the season; 2.) the amount of fencing needed; 3) the length of the main -canal; and 4 ) weed growth, 3.2 Economic Costs and Returns of Irrigated Agriculture labour costs are the main source of financial committments to irrigated cropping in any given year in N jumps and the use of hired labour is widespread. In all II Chamus areas, labour is completely monetized and even when traditional beer is brewed for work parties, the emplcyee has the choice of either drinking a certain amount of beer or being paid the cash equivalent. Labour is usually paid for on a piece-work fashion as follows: basin construction is 1/ to 1/50 oach; fencing is 400/ per hectare; weeding is 1/ to 1/50 per basin;.,..and planting /.50_.per..,-basin,. , The average agricultural wage for a six hour day is seven shillings and only varies in the Ngambo area where labour costs are nine shillings per day. In this region, the proximity'to the Perkerra Scheme necessitates that wages for casual 1 labour in Ngrmbo must be competitive with the Scheme's wage of 6/90 per day* -20'- IDS/WP 386 • In- the sample of fifty-eight households stated previously,families of seven were able to cultivate irrigated farms of ,5 hectare or less without the use of hired labour. Richer households with two wives or more and average family size of ten members were capable of cultivating up to o7 hectares without cash labour costs. Families with only four members, which represents mainly the poorer families, were able to only cultivate .3 hectares of irrigated land utilizing only household labour. Eighty- three percent o^he farmers in the sample with farms of ,8 hectare or 13 more employed non-family members to work on the farm. Costs of production for irrigated agriculture are such that given the average II Chamus cropping pattern of8/10 maize, 1/10 finger millet and 1/10 bc.ans, marginal returns per hectare after a threshhold of three hectares do not offset one's labour costs for most families. For a family of ten with six members over the age of ten years, it is necessary to hire labour for almost all agricultural tasks for any increments in farm size after two hectares. Marginal returns per hectare in the sample of fifty-eight households wore approximately 2300 shillings for the first hectare, 1900 shillings for the second hectare, and loss than 50 shillings for the third hectare cultivated» ( The last figure is derived losing production costs for smaller farras, since no farm in my sample using only manual labour was larger than two hectares). These calculations are based on the assumption of zero opportunity costs for family labour on the farm (mainly women), Moreover, it should be noted that cash returns per hectare do not actually represent earned cash income, but rather the monetary value of the crops produced which in some cases were consumed totally' at home by family members. 13, A major topic of my dissertation will be the role of kinship and other network mechanisms in mobilizing non-household labour, . In the main area under consideration lierg, Ngambo, such extra-domestic mechanisms are not important. 21 IDS/WP 386 The use of hired tractors from the Perkerra Irrigation Scheme h a s , for very good reason, been especially popular among the richer Ilchamus farmers in the irrigated areas. For a cost of 150 shillings per acre, much of the labour in the pre-planting stage is minimized at a time when labour bottlenecks tend to be most severe because livestock are receiving much attention at this time (i.e. February and March). In a .8 hectare plot, production costs for cultivation and basin construction are reduced by sixtytwo percent when one uses a tractor; basins must still be constructed but price per basin when tractors are used drops 1/50 to -/50 per basin. The profitability of cultivating farms has meant that overall hectarage in areas where from large tractors are used has increased over areas where farms are still cultivated by hand. In the only Ilchamus area where tractors have been used for the past two years, Ngambo, average farm size for farmers who hired a tractor is 1.89 hectares. This is more than two times the average irrigated farm holding in Njemps. Since access to capital (livestock) and large inputs of family labour is restricted to the wealthier families, it is not surprising that one finds ownership of irrigated landholdings pola14 . . . rized. Concentration of land ownership tends to be highest in the Ngambo area where, as stated above, some farmers use tractors. In this area, ten percent of the farmers 'control thirty-nine percent of the irrigated land and twenty percent of the farmers own sixtyone percent of irrigated land. The bottom fifty percent of farmers in terms of land ownership control only eleven percent of the irrigated hectarage in Ngambo. In Eldume, on the other hand, where tractors are riot available land is more equitably ted. distribu- This is illustrated by the fact that ten percent of the farmers in Eldume control only twenty six percent of the land (20% of the landholders own forty-eight percent of irrigated hectarage), while the bottom fifty percent of landholders have access to twenty percent of irrigated land. Reasons for the more 14. Land ownership in the area tends to be more equitably distributed than stock ownership because labour is not such a constraint in the accumulation and production of livestock as it is in crop production. 22 IDS/WP 386 equitable tenure pattern in Eldume relate to the non-profitability of cultivating large farms using hired labour as the main input. 4-. Mixed Livestock Enterprises: Some Factors to be Considered 4.1 Crop Production and Its Effect on Livestock Holdings That increased food crop production will lead to a decline in livestock numbers, as had been predicted by planners, is not the case in Njemps Location 15 Kenya, 19 80). development (cf Government of In fact, it is the area of Njemps> Ngambo, where crop hectorage is greatest that one also finds the largest number of livestock. That crop production will appeal mainly to the poor stockowners is alsb invalidated in the II Chamus ease, Unlike Turkana and other pastoral area's (cf. Broch-Due and Storas, 1980:14), wealthy II Chamus households of Ngambo and neighbouring areas are able to maintain a viable production unit with investments in both the pastoral and agricultural sectors. This is achieved mainly because of the sedentary nature of pastoralism in the area and 16 the region's readily available supply of hired labour. In this regard, it is the wealthy stockowners of Njemps which are more capable of managing a mixed livestock cropping enterprise. 4 indicates the relationship between irrigated farm size a n d Table live- stock holdings among 90 randomly selected households in the Ngambo, Loiminange and Eldume areas. Figure 2 illustrates this on a graph. 15. The planned irrigation scheme in Sandai, Loboi Location, an area which borders the II Chamus region, is also based on the premise that increased grain production will lead to a reduction in Sandai livestock numbers (cf. Republic of Kenya, 1981). 16. As mentioned earlier, the paper focuses on the area of Njemps where the pastoral system centers on the swamps. This represents more than seventy percent of total Ilchamus families. In eastern Njemps, where dry season livestock movements are to the hills toward Laikipia, households have more difficulties in establishing stakes in both pastoralism and agriculture. 23 IDS/WP 386 In the sample, people with under 5 S.U. represented 18^ of the total households and had an average farm size of .08 ha, IDS/WP 386 24 Table 8 and Figure 2 illustrate that stock units tend to increase in relation to size of irrigated holding. However, ranges in farm sizes within each wealth category tend to be siderable. That some wealthy stockowners do not have con- irrigated farms is. probably more a function of age than anything else. Older, rich stockowners over the age of sixty years tend not to invest in cultivation in many cases because their grain needs are met their sons' farms. from 1 Some elderly livestock owners at the time when irrigated agriculture opportunities became' available were heads of families which were at what Fortes (1958:5) calls stage" in the developmental cycle of the family. the "dispersion This is the phase in family development when sons and daughters begin to marry and leave their household in order to establish their own homesteads. The loss of family labour as well as perhaps a lack of ambition among these households to diversify the household economy at a late age may also be reasons for the lack of irrigated among some wealthy stockowners. As Figure 2 indicates, it is not Until after that stock units begin to increase exponentially. sufficiency farms .1.0 hectare Grain self- for a family of seven in the irrigated areas is achieved at .7 hectare, if .6 hectare are devoted to maize and .05 hectare each to beans and finger millet. If we use the average family size of ten with six members over'the age of ten years, which approxi- mates the family unit of stockowners with stock unit holdings seventy, then grain self-sufficiency approximately 1.0 hectare. over for the family is reached at Production above costs after this point can be sold or exchanged directly for livestock as is quite often the case. Moreover, grain self-sufficiency at the household level greatly reduces family expenses since purchase of maize meal accounts for over sixty percent of average household That savings in cash allocations for maize meal expenditures. will be diverted to other consumption items does not seem to be the case. Other than perhaps additional purchases by the wealthy households of clothes, cooking utensils and other "luxury" articles, consumption activities between the rich and poor households do not vary as much as the differences in their economic statuses would imply. Stated differently, II Chamus household consumption patterns do not increase in proportion to the amount of livestock or irrigated land a family IDS/WP 386 25 In the irrigated areas, grain production actually serves to support the livestock sector since it reduces the need to deplete herds to buy grains. It also allows surplus from agri17 culture to be invested in livestock. This is best illustrated by examining livestock investment patterns among households two different areas of Njemps. These are: in 1) Ngambo where, as mentioned earlier, irrigated agriculture is widespread and grain self-sufficiency in the region is approximately fifty percent; and 2) Salabani, a region where there is virtually no irrigated agriculture at present and ninety-nine percent of local grain needs must be imported. the former In a survey of thirty-two households from area with average farm size over .8 hectare, eighty four percent of the sample had bought livestock in the past year. Average stock purchased per herd-owner was and 5 small stock (range 0-40). 3.6 cattle (range 0 - 3 5 ) In regard to cattle, the majority was purchased from South Baringo livestock traders often at an exchange rate of six to seven small stock for a heifer or a young bull. Those cattle purchased main II Chamus grain deficient locally in Njemps came from the areas, Salabani and Loiminange. In these two sub-locations, herdowners are inclined to sell livestock in order to buy grain. Of the small stock purchased by the Ngambo households, over seventy percent were purchased locally in Njemps and twenty-two percent of the small stock purchased was paid for in exchange with maize at the rate of thirty-two kilograms of grain per small stock. between households Much of this latter trade took place from Ngambo and Salabani. Livestock purchasing patterns in Salabani, on the other hand, differ from those described above. wealthy herdowners In a survey of twenty (average holding of over fifty stock units) representing some of the most active marketing (both buying and selling) households in the area, only forty-five percent of the stockowners had bought livestock in the past year. Purchases per 17. At present, land for livestock only competes minimally with cultivated land. Stock are mainly grazed in the swamp areas where farms are not found and animals are always grazed cn harvested fields. Trespass of stock in cultivated fields is common, but serious disputes tend to be minimal because most rich farmers also have large investments in livestock. IDS/WP 386 26 household were on average 1 cattle small stock (range 0 - 5 ) and less than one (range 0 - 3 ) , approximately investments in Ngambo. J of average livestock Moreover, in terms of livestock sales, seasonal livestock marketing patterns in Salabani differ from those occurring in Ngambo. This is especially in regard to the immediate post-harvest months (October-December) when home-grown grain is available and marketing of stock per household in Ngambo declines as much as thirty percent. Such a radical change in household marketing behavior does not occur in Salabani in these months of the year. 4.2. Irrigated Agriculture and Its Impact on Equity Differentials in property ownership in Njemps are most salient in the Ngambo area where polarization of livestock holdings is fifteen percent greater than patterns revealed in other II Chamus areas. Explanations fo» the greater inequity in Ngambo relate to the fact that unlike other areas of Njemps, possibilities for extension of irrigated agriculture are minimal and the present system of land ownership is now fixed. For one to acquire an irrigated plot in Ngambo, he/she must either borrow land from a friend or kinsmen or purchase the land outright. The latter is difficult since there are few people in the area who are willing to sell their irrigated farms. The cycle of grain production and livestock investments described previously serve to promote the interests of those individuals who had earlier staked out claims to irrigated farms. The lacunae b e t w e e n the rich and poor households in Ngambo seems to be increasing as livestock investments, as w e l l as investments in retail stores and other enterprises, tend to be concentrated among those households with both irrigated and large livestock holding. which farms Natural factors such as drought in most pastoral areas tend to periodically reduce the economic differences between the rich and poor livestock is not an effective "levelling" mechanism in Ngambo. owners In fact, those, stockowners with a stake in irrigated agriculture can more easily recover from heavy livestock losses because of the potential for converting agricultural surplus into livestock holdings.' IDS/WP 3 86 27 4.3 Crop Production and Rural Development in Baringo District Given the information presented in the paper relating increased grain production to livestock investments, what are the implications for a cropping policy in regard to rural in Baringo District. development In other words, is it desireable to increase crop production if it serves to support the buildup of livestock numbers in the area? To answer the question, it should first be pointed out that for most II Chamus households investing in livestock is an option that they do not have the resources to undertake i Unlike the wealthy households (percentiles 1 and 2 in Table 8) who perceive agriculture as a support mechanism for their livestock holdings and investments, the poor households are farming to meet subsistence needs and their farming activities have little relation to livestock buildup. Referring again to Table 8, irrigated farm owners in the fifth percentile and below are not even fifty percent grain self-sufficient and those in the third and fourth percentiles are only borderline grain self-sufficient. The poorer II Chamus households are actively involved in agriculture and because of the limited number of irrigated farms many are investing in dryland agriculture, an enterprise which only \ the labour costs of irrigated agriculture. packages entails Suitable crop (i.e. Sorghum and finger millet) for the dryland areas should be encouraged among those households, mainly the poor ones, 18 presently engaged in rainfed farming. If one wishes to control livestock numbers in Njemps Location by such means as establishing grazing fees, a policy which should be made only after a more careful consideration of the productivity of the swamps, the program should not be undertaken at the expense of encouraging crop production. especially in regard to the poorer households. This is Recognition should also be made of the fact that while a reduction in livestock 18. This is the present crop policy of the Baringo Pilot Semi-Arid A r e a Project. IDS/WP 3 86 28 numbers may serve soil conservation and range management purposes, it also has the potential of reducing household incomes among the seventy percent of II Chamus households which presently derive the majority of their cash income from the sale of livestock. Under these circumstances, such a decline in per capita income will occur unless alternative sources of income are made available the area. in Finally, the establishment of a program based on grazing fees may incur costs to the producer, which cannot be made up from the added market value of the animal maintained under a more able stock unit to rangeland ratio. favour- That is to say, the increased value of better managed indigenous cattle, Zebu, may not offset the costs of grazing fees. Another factor which may reduce household incomes in Njemps. serve•to IDS/WP 3 86 29 References Cited: Baringo Pilot Semi-Arid Area Project 1981 Crop Survey, 1981. Marigat, Baringo District, Kenya. Broch-Due, V. and F. Storas 19 80 The Relationship Between the Pastoral and the Agricultural Sectors in Turkana. Unpublished paper. Brown, Leslie H. 1971 The Biology of Pastoral Man as a Factor in Conservation. Biological Conservation 3:93-100. Colony and Protectorate of Kenya 19 30 District Annual Report 4 Government Printers. 1940 Baringo District.Nairobi: District Annual Report, Baringo District. Government Printers. Nairobi; Dahl, Gudrun 1979 Suffering Grass. Stockholm: Stockholm University Press. Dahl, Gudrun and Anders Hjort 1976 Having Herds. Stockholm: Liber Tryck. Fortes, Meyer 195 8 Introduction. In Jack Goody, ed. The Developmental Cycle in Domestic Groups. Cambridge: Cambridge University Press. Little, Peter D. 19 81 A Sociological Report on the II Chamus of Baringo District: A Report Prepared for the Baringo Pilot Semi-Arid Area Project and the World Bank. Nairobi, Ke ny a. Matthes, M.C. 19 79 Livestock Marketing in Kenya, United Nations Development Program, Nairobi, Kenya. Ministry of Agriculture 19 5 7 Range Management Files, Marigat, Baringo District, Kenya Ministry of Agriculture 19 81 Sandai Irrigation Scheme: Results of Apro-Economic Survey. Rift Valley Provincial Irrigation Unit, Nakuru, Kenya. Ministry of Livestock Development 19 81 Monthly Veterinary Reports: 19 74-19 81, Marigat, Baringo District, Kenya. IDS/WP 3 86 30 Ottley, R.A., J.K. Wanjaiya and Niels L. Martin 197 8 Livestock and Range Management. Marginal/SemiArid Lands Pre-Investment Inventory Report No. 8, Nairobi, Kenya. Republic of Kenya 1968 District Files, Baringo District,. Kenya. 19 7 8 Arid and Semi-Arid Lands Development in Kenya: The Framework for Implementation, Programme Planning, and Evaluation, Nairobi, Kenya. 1981 District Development Plan, Baringo: Nairobi, Kenya. 1979-83. Schneider, Harold K. 1979 Livestock and Equality in East Africa: The Economic Basis for Social Structure, Bloomington: Indiana University Press. 1980 Livestock as Food and Money, Paper presented at the International Conference on the Future of Pastoral Peoples sponsored by the Nomadic Commission, Nairob Kenya. United Nations Development Programme 1967 Rangeland Survey of the Njemps Territory, Baringo District, Nairobi, Kenya. White, J.M. and S.J. Meadows Structure of the Herd and Determinants of Offtake 1979 Rates in Kajiado District, 1962-1977, Pastoral Network Paper 7d, Overseas Development Institute, London.