Pasture based livestock production with domestic ruminants, dromedaries and donkeys is the dominant economic activity in the dry lowlands of Kenya. These areas are characterized by low and erratic rainfall, a high probability of droughts, as well as by a general scarcity of permanent surface water and are therefore generally unsuited to cultivation. Migratory pastoralism has evolved under these marginal production conditions and remains the most important land use system in these regions. Pastoral production systems share a number of unique characteristics that are aimed at minimizing production shortfalls caused by large variations in forage productivity. One of the most prominent risk minimising system attributes relates to the mobility of pastoral livestock herds and households, allowing full exploitation of forage resources that are unequally distributed in space and time (Coughenour et al., 1985; Ellis et al., 1988). Orientation towards subsistence production, reliance upon a combination of different animal species, considerable sharing of resources and products within small groups, and an emphasis on milk rather than meat production are additional features contributing to risk avoidance in pastoral economies (Swift, 1982; Western, 1982).
Change in migratory pastoralism: sedentarisation and commercialisation
These features have helped pastoral production systems to remain remarkably stable for a long time. However, fundamental demographic, economic, and political changes have disrupted the delicate balances between human populations, livestock numbers, and rangeland resources in these socionatural systems. These changes originated partly in programmes of the colonial administration aimed at raising productivity in the pastoral sector, which, subsequently, were taken over almost without interruption by the Kenyan government after independence (Oxby, 1975; Bennett, 1988). Attempts at restructuring pastoral production systems to increase their economic self-sufficiency and contribution to the national economy have almost always failed. Instead, such attempts have affected traditional land-use practices, have led to permanent differential access to basic productive resources, and to a substantial increase in income disparities among pastoral households (Hogg, 1986; Mayer et al., 1986). However, a more serious influence on pastoral production systems relates to the constantly increasing human population in the semi-arid rangeland areas of Kenya, and to the accompanying conflict over land resources. The rapid population growth in neighbouring agricultural communities has led to an expansion of cultivation onto semi-arid rangelands. In particular the small pockets of high potential pastoral land which formerly served as grazing reserves, are increasingly being occupied by agriculturalists. These losses have been aggravated by the establishment of commercial ranches and national parks on pastoral land (Schwartz & Schwartz, 1985). The increased competition for resources among different land use systems was also paralleled by a steady growth of the pastoral population, although at a slower rate than in other groups in Kenya, which further decreased the per capita availability of land for grazing (Swift, 1982).
The combined forces of demographic pressure, steady loss of rangeland to other sectors, and development interventions in pastoral economies have contributed to a rapid decrease in the mobility of pastoral herds and settlements throughout Kenyan rangelands (Grandin, 1988, 1991; Fratkin, 1992; Schwartz et al., 1995; Roth, 1996). Sedentary pastoralism in the vicinity of small towns, trade centres, famine relief stations, and mechanized water sources has become a widespread practice, especially among impoverished pastoral households (O’Leary, 1990). The transition to sedentarism, for long a primary objective of development policies aimed at pastoralists in Kenya, is now inevitable given the major alterations in system functions; this process is likely to cause substantial ecological and economical problems. Schwartz et al. (1995), for instance, note that concentration areas are marked by severe and spreading degradation of vegetation and soils. This, in turn, lowers herd productivity, increases herd sizes required to meet household needs, and thus further accelerates environmental degradation and the likelihood of destitution.
The sedentarisation of pastoral households is intrinsically related to the incorporation of pastoral economies into regional and national markets. Whether the observed emergence of market oriented production over traditional subsistence production is an additional cause or merely a consequence of sedentarisation remains ambiguous; both processes are generally so intertwined that it is difficult to distinguish cause and effect [page 10↓](Sikina et al., 1993). Nevertheless, both processes entail a gradual change in pastoral herd management and species composition, and, ultimately, a redefinition of production goals. Several studies conducted in northern Kenya have documented changes in species composition in pastoral livestock herds arising from population increase, reduced mobility, and commercialisation of production. Roth (1990) and Roth and Fratkin (1991) observed a recent emphasis on cattle as opposed to camels in Rendille livestock herds, which is a reflection of an increasing interaction with a cash economy. The transition from camels to cattle as the predominant species is a strategy adopted by livestock wealthy Rendille households who attempt to make the shift to commercial production and to minimise losses in a changing socioeconomic environment. Although less ecologically adapted, cattle have a retail market in northern Kenya which camels lack. Due to their higher market value, cattle tend to replace camels in their function as a store of wealth and as a means for exchange purposes (Roth, 1990).
Poor Rendille households, in contrast, concentrate on smallstock. The rapid rate of reproduction of smallstock makes them a major means of post-drought recovery. Moreover, accumulation of smallstock is a sensible strategy for poor households since they are easily converted to cash for household needs, and can be utilized as a means of acquiring large stock (Roth & Fratkin, 1991). Mace and Houston (1989) arrive at similar conclusions concerning the importance of smallstock to maximising household survival chances at low herd sizes. According to Schwartz (1987), the goat has also become the main meat supplier for most pastoral consumers in northern Kenya, a fact which strongly indicates a drastic change in socioeconomic patterns in pastoral societies that have to deal pragmatically with recurring emergencies. Likewise, data presented by McCabe (1987) for the Ngisonyoka Turkana suggest a shift from a pastoral subsistence based on cattle to one based on smallstock in a situation where livestock holdings per capita are decreasing. In contrast to sheep and goats, cattle have the disadvantage of being large indivisible units, such that substantial amount of the herder’s wealth is stored in only a few animals. Poor households are therefore less vulnerable to livestock losses when concentrating on smallstock, since here, the capital accumulated in each animal is minimal.
Similar patterns of change have also been observed on Maasai group ranches in southeastern Kajiado District (King et al., 1984). High population growth was accompanied by an increase in labour availability and food requirements and hence a need for a more intensive exploitation of fodder resources through species diversification. Maasai began to rear smallstock in response to adverse climatic conditions and an increase in population pressure. Sheep and goats played a critical role in maintaining domestic food supplies during the disastrous droughts in the early sixties and in the mid-seventies. Since then, a substantial increase in the smallstock population and in the importance of meat and milk from small ruminants in the Maasai diet has been observed (Njoka, 1979). However, the overall number of livestock units per capita decreased, and this induced higher levels of exchange of pastoral for agricultural foodstuffs, inputs for livestock production, and other essential goods (Grandin & Bekure, 1982; Grandin, 1988).
Patterns of smallstock and cattle utilization vary according to the scale of the operation unit. Poor Maasai producers have much greater rates of utilization of their smallstock in terms of milk, meat and especially live animal sales than rich producers. Due to their low livestock to human ratio and pressing consumption needs, poor households have to engage in market exchange in order to convert their livestock products to foodstuffs with higher energetic value (Ensminger, 1987; Grandin, 1988; Sikina et al., 1993). With respect to cattle, poor herd owners tend to rely on an intensive extraction of milk from their herds, whereas richer herders deliberately forego some of the potential milk output in favour of calves and derive higher levels of income from live animal trade. In general, poor producers attempt to offset diseconomies of scale by intensive methods of extracting value from animals (Behnke, 1984).
The case of the Maasai group ranches suggests that the growth in smallstock holdings may also be linked to a change in rangeland vegetation induced by an increase in land use pressure (Njoka, 1979). Goats in particular have a wider dietary range and lower water requirements than cattle, and are better adapted to cope with drought and poor grazing conditions such as often occurs in the vicinity of permanent settlements. King et al. (1984) report that reduced herd mobility combined with increasing livestock populations has led to significant changes in plant composition of the rangeland vegetation in Maasai group ranches, and that this change has favoured the shift towards smallstock. Likewise, Bekure and de Leeuw (1991) state that the growing grazing pressure on rangeland in southeastern Kajiado is likely to lead to the replacement of perennial grasses by bushes, dwarf shrubs, forbs and ephemeral annual grasses and encourage Maasai producers to keep more smallstock to effectively exploit forage resources. Thus, the increase of smallstock holdings especially among stock-poor households could also be interpreted as an adaptation to a degrading habitat. In general, it can be expected that those households that diversify their herds by keeping a significant proportion of sheep and goats may likely adapt more readily and securely to a sedentary lifestyle (DeVries & Pelant, 1987).
[page 11↓]Consequences for the future of pastoral economies
From the foregoing, a number of consequences for the future of pastoral economies in Kenya can be anticipated. First, the rapid integration of pastoral economies in regional and national markets will continue to impact on management strategies and production goals of pastoral producers. The process of commercialization is, however, not neutral to scale, and will therefore affect large and small subsistence operations differently. Large herd owners can more readily accommodate the shift from in-kind milk and meat production to market oriented meat production, since the reduction in overall biological productivity implied by this shift is potentially more than offset by the higher economic profitability achieved through live animal sales (Behnke, 1984). Small operation units, in contrast, can not afford to abandon traditional subsistence forms of animal use. The potential offtake rate of live animals for sale from their herds is too low to meet the household expenses required to substitute the majority of subsistence products for non-pastoral foodstuffs. Small herd owners will therefore have to retain a predominant subsistence oriented mode of production in order to increase herd sizes for an eventual shift towards commercial production. On the other hand, the process of commercialisation has been observed to entail a reduction in the number of surplus livestock from wealthier households available for redistribution through animal loans, gifts and other transfers, thus depriving small herd owners of an important source of livestock to build up their herds (Behnke, 1983; Grandin, 1991; Moris, 1991; Sikina & Kerven, 1991). In addition, those producers who successfully make the shift to commercial production will attempt to reinforce their economic superiority by acquiring private use rights to land, a phenomenon which is already common among Maasai pastoralists in Kajiado District (Grandin, 1989; de Leeuw et al., 1984). Together, these developments will further undermine the viability of small production units and increase income disparities in pastoral societies (Bennett, 1988).
A second concern relates to the tendency of livestock herds to grow in commercialising pastoral economies, a trend which will aggravate the pressure on natural resources already exerted by the growing and sedentarising pastoral population. This is because the decline in biological productivity per animal unit in commercialising operations may force herd owners to increase their herd sizes in order to maintain overall household income levels (Behnke, 1984). On the other hand, herd owners can also benefit from economies of scale by increasing their herds, since larger herds tend to have lower per unit operating costs. Whereas under the traditional, labour intensive subsistence mode of production, herd growth is restricted by labour availability, this limitation is relaxed in a commercialising operation, so encouraging producers to increase their operations.
Lastly, the economic polarisation of pastoral societies will promote the exclusion of poor, subsistence oriented households from the pastoral sector (Swift, 1982). Households falling below the minimum livestock per capita ratio required to insure self-sufficiency have a limited set of alternative strategies to choose from in order to complement household incomes. They may either realise additional income from herding for other, wealthier households, seek for wage labour in nonpastoral activities, or abandon the pastoral economy altogether and migrate to small towns or large cities in search for employment (Fratkin & Roth, 1990). Yet, the capacity of pastoralism to absorb surplus labour from households which lack sufficient stock to support themselves independently is very limited. Additionally, in Kenya, employment opportunities outside the pastoral sector are few and pastoralists are at a comparative disadvantage in competing against the often better-educated agriculturalists (Grandin, 1989). Under these conditions, the likelihood of an increase in the number of displaced and destitute pastoralists is high.
In light of the complexity of the problems with which pastoral economies are confronted today, it would seem difficult to isolate a single development strategy that is capable of simultaneously alleviating the above mentioned social, economic and environmental concerns. The present demographic and macroeconomic conditions in Kenya prevent the relocation of large parts of the pastoral population to other sectors of the national economy. However, the exclusion of marginal operation units from the pastoral economy is unavoidable when the process of commercialisation is left unattended. Under these circumstances it is important to identify interventions that will allow stock-poor pastoralists to continue to support themselves largely through livestock-related activities without further increasing the pressure on natural resources in the rangeland areas (Grandin, 1989). Although a focus on technical alterations alone is clearly not sufficient to achieve this goal, improvements in the productivity of livestock herds are a necessary feature of any effort to (1) increase total output of operation units; (2) enhance food security of pastoral households, and (3) reduce the risk of environmental degradation.
Traditional pastoral production strategies are geared towards a balanced supply of food for human needs throughout the year without major shortages. This is best achieved through a combined milk and meat production, where milk is the mainstay of the diet, complemented by voluntary slaughters of animals and, to a limited extent, market exchange of animals for agricultural foodstuffs. Commercial production systems, on the other hand, adapt their production cycle to seasonal forage availability and overcome income shortages during dry seasons by using capital that is accumulated through sales in wet seasons (Gill, 1991). In the previous section it has been shown that livestock-poor households lack sufficient animals to successfully adopt the latter mode of production. Due to the inadequate food supply from their herds, poor households may nevertheless be forced to increasingly rely on livestock sales in order to cover subsistence requirements. Situations may occur where producers find themselves selling off their breeding stock simply to feed their families, thus seriously undermining household viability (Mace & Houston, 1989; Moris, 1991). Furthermore, livestock losses in recurrent dry seasons or disease outbreaks exacerbate the marginality of small-scale operations (Roth & Fratkin, 1991). Such operations may get trapped in a vicious cycle by declining herds that are too small to meet household requirements.
An important question to ask, therefore, is how could small-scale producers increase the output of milk and surplus animals for home consumption, trade, and exchange in order to enhance the viability of their operations. One approach to improving household self-sustenance is to support small-scale producers in their attempt to rely on smallstock. In the past, a variant of this policy has been adopted by Oxfam for herd reconstitution among pastoralists in northern Kenya (Moris, 1991). The approach was to supply each recipient household with a nucleus herd of small ruminants presumed to be sufficient for each family’s continued livelihood.
The case for concentrating on sheep and goats to support poor households under the changing socioeconomic and environmental circumstances rests upon several arguments. Firstly, the material reviewed in the previous section suggests that the "smallstock strategy" (Grandin et al., 1994) observed among Rendille and Maasai is probably one of the most efficient strategies for poor households to generate cash-income, rebuild their herds, and as a means to acquire large stock. The findings reported by Mace (1990) are a strong evidence that investing in smallstock is indeed the most appropriate strategy to maximize household survival chances at low herd sizes. Although, living off smallstock alone is considered to be a very risky 'boom or bust' short-term strategy that should only be followed as long as smallstock herd sizes are below a critical threshold above which an exchange of smallstock for camels becomes feasible (Mace & Houston, 1989; Mace, 1990).
Secondly, sheep and goats confer an advantage in terms of food security to small-scale producers. In sheep and goats, drought mortality rates are lower than in cattle, and post-drought herd recovery is much faster than in cattle and camels. Particularly goats may also play a crucial role in balancing year-round supply with milk for subsistence due to their ability to lactate during dry seasons, when for instance cattle produce little or no milk at all (Schwartz & Schwartz, 1985).
Thirdly, poor households tend to curtail their seasonal movements, and choose to settle around permanent water sources or trading centres. As has already been mentioned, the increasing concentration of people and stock leads to a reduction and degradation in pasture in the vicinity of permanent settlements. Due to their feeding behaviour, goats, and to lesser extent, sheep, are well suited to exploit the feed resources available under these circumstances (Lancelot et al., 1995).
A final reason why the focus on smallstock seems to be appropriate for poor households relates to the production economics of pastoral operations. Traditional pastoral herd management methods: herding, watering livestock, and milking are all labour intensive (Behnke, 1983; Grandin, 1989). Sheep and goats, however, require less labour than do cattle and camels per livestock unit (DeVries & Pelant, 1987). Smallstock holdings thus do not compromise the labour force required for the management of large stock. Also, the capital cost of the animals themselves, and consequently the economic loss incurred when an animal dies, is relatively small with sheep and goats (Johnson et al., 1986).
In order to enhance the viability of poor pastoral households, simply increasing the size of small ruminant flocks through restocking projects does not seem to be a satisfactory approach. Although they have an important redistributional advantage, herd reconstitution through donations alone will not alleviate the constraints which undermine the long-term viability of small-scale operations. In the long-run, self-sustenance of households will depend upon achieving levels of herd growth adequate to provide enough surplus animals for subsistence, market exchange, and herd accumulation. This calls for an improvement of two main traits of flock performance: reproduction and survival. These traits determine the rate of increase [page 13↓]and hence the sustainable offtake from the flock (Upton, 1989).
On the other hand, land, grazing and water resources are replacing labour as the limiting factor on economic activity in pastoral systems, due to population growth and increased competition for resources from other land uses (Swift, 1982). Development interventions aimed at improving (or merely maintaining) the livelihood of pastoral people and to checking the risk of environmental degradation must therefore seek to promote an increase in the physical output per animal in terms of milk, meat and other goods. In summary, with respect to the advocated smallstock strategy, an ideal intervention would, at the same time, contribute to significant improvements in herd growth and in the productive performance of pastoral sheep and goat flocks.
Before discussing potential interventions towards these goals, an important consideration needs to be kept in mind. It relates to the existing trade-off between the production of replenishable goods, i.e. milk, and surplus animals for sale, consumption, exchange and accumulation. The production of surplus animals in good condition requires that sufficient milk is allocated to lambs and kids (Ruvuna et al., 1988). Survival rates and growth performance of young stock are generally depressed under traditional pastoral management when substantial amounts of the milk produced are extracted for human consumption (Field et al., 1984). In order to maximize the net reproductive rate of pastoral sheep and goat herds and, ultimately, the potential offtake rate, most, if not all, of the subsistence milk production would have to be foregone. This might in turn affect the distinctive property of smallstock as a buffer against insecurities and imbalances in food supply, and must therefore be taken into account in evaluating specific recommendations.
Measures to improve small ruminant flock productivity
Under semi-arid environmental conditions, the main factor influencing the productivity of small ruminant flocks is climatic seasonality. Seasonal changes in temperature and moisture availability affect animal production not only by influencing the quantity and quality of available nutrients, but also by affecting the dynamics of parasitic organisms and the incidence of infectious diseases (Carles, 1985). Several studies of pastoral production systems in Kenya have found high pre-weaning mortality in sheep and goats caused by infectious diseases to be a major constraint on flock productivity in traditional management systems (Carles, 1986; de Leeuw et al., 1991; Wilson et al., 1985). The diseases afflicting small ruminants are respiratory, such as contagious caprine pleuropneumonia (CCPP), tick-borne, and brucellosis. Gastro-intestinal parasites appear to be a major health problem affecting productive performance in sheep and goats under the semi-arid conditions in northern Kenya (Field et al., 1984). Excessive helminth parasite loads are critical particularly at stages in the productive cycle when physiological requirements of animals are highest, such as during late pregnancy, lactation, and early growth period of weaned kids and lambs.
Principally, significant improvements in survival rates, productive performance (milking and growth performance), and thus flock productivity can be achieved by applying a health programme which includes vaccinations against infectious diseases, anthelmintic treatments, tick control and mineral supplementation (Wilson, 1984). Such interventions, however, are costly and need close and constant supervision. The effectiveness of health care measures in improving productivity is also likely to be reduced by the restricted and erratic supply of nutrients. In addition to its immediate effect upon survival, growth, and lactation, insufficient nutrition lowers the resistance of animals to pathological agents (Carles, 1983; Lebbie et al., 1996).
In a study conducted under semi-arid conditions in Mali, Ba et al. (1996) found that inspite of vaccinations against infectious diseases (Pasteurellosis, Anthrax, Petite Peste des Ruminants) and deworming against nematodes 50% of the kids died due to infections. The treatments had no effect on kid survival or growth until after weaning at 5 months of age. Since about 40% of kids died of malnutrition, injuries and predation losses, this study shows that first priority in reducing kid mortality should be given to improving management practices. The authors concluded that the impact of the veterinary package was limited to a small increase in flock sizes, due to a positive effect of deworming on adult goat survival. However, the costs of anthelmintics prohibit their introduction as regular deworming treatments.
Research on nomadic pastoral goat herds in Marsabit District of northern Kenya, provide some evidence that the impact of eliminating pathogenic effects on the incidence of morbidity and mortality is limited and highly dependent upon environmental conditions (Carles et al., 1984; Carles, 1986; Schwartz, 1988). In a year with above-average rainfall, treating goats with anthelmintics led to a slight increase in kid survival, but had no appreciable effect on mortality of adult animals. The anthelmintic treatment increased milk yields during and shortly after the rainy season; due to the low availability and quality of feed, however, this effect vanished during the dry season (Schwartz, 1988). By contrast, in the second year of the experiment, when rainfall was below average and the health care programme was extended to include vaccinations against CCPP and [page 14↓]Brucellosis, tick control measures as well as mineral supplementation, herd productivity fell drastically in both the control and the treatment flocks due to inadequate nutrition. Yet, a substantial interaction occurred between health care measures and feed availability such that the reduction in productivity in the treated herd was more than twice as high as that in the control herd. These results suggested that health care measures, especially anthelmintic treatments, improved pastoral goat herd productivity only when feed availability was not constraining animal performance (Schwartz, 1988).
The scope for alleviating restrictions imposed on animal performance by the seasonally and annually varying supply of nutrients is very limited, especially under the changing socioeconomic and environmental conditions outlined above. Given the erratic nature of primary production in semi-arid areas, the pastoral producer lacks the ability to exert direct control over future biological states of the grazing system, and thus to manipulate his forage supplies. The basic problem of matching feed supply and demand is overcome by adaptive management, of which the most prominent form is the movement of animals onto pastures with adequate forage (Behnke, 1994). However, the distinctive advantage derived from herd mobility for balancing nutrient shortages is lost in the transition to sedentariness, and this introduces an additional constraint on animal productivity. In this situation, intensifying land use by adopting forage production and/or forage conservation are potential options for synchronising feed supply and demand, especially during the dry season. Successful forage cultivation is feasible only on high potential land within the grazing system, particularly on the scarce dry season pastures. Production and conservation of forage are very labour intensive, and this represents a serious limitation to pastoral households considering these activities as a way of improving their feed resources. Also, forage interventions may be difficult to implement under communal grazing tenure (Bekure & de Leeuw, 1991; Coppock, 1994).
Alternatives to herd mobility as the most efficient form of adaptive management to overcome nutritional deficits are few. One strategy is to manipulate the total seasonal nutrient requirements of the herd, instead of attempting to improve the supply side of the nutrient balance. Adapting total forage demand of livestock herds to feed availability is essentially what Sandford (1982) has called an "opportunistic strategy". The main idea is to adjust the number of stock up or down in response to variation in forage resources. In his paper, Sandford discussed opportunism as a management strategy to cope with inter-year variability in primary productivity of the grazing system, but opportunism could as well be used as an adjustment to inter-seasonal fluctuations in forage supply. Irrespective of time scale, the practicability of the opportunistic strategy hinges upon access to livestock markets to dispose of surplus stock, and on sufficient market demand to capture the implied increases in flow without a collapse in prices (Sandford, 1982). Difficulties relating to both factors have been shown to hamper market sales by pastoralists during dry seasons in northern Kenya (Schwartz, 1986), Maasailand (Bekure & de Leeuw, 1991), and southern Ethiopia (Coppock, 1994). Although the same studies found that forced and voluntary slaughters of stock generally peaked during dry seasons or droughts, the authors argue that substantial increases in market sales would be needed to achieve a balance between the nutrient requirements of the herds and pasture forage production.
Nevertheless, the adjustment of stock numbers could be complemented with an additional measure geared towards manipulating total seasonal nutrient requirements of the herd, namely the confinement of breeding to one or several selected periods of time in a year. Continuous breeding throughout the year is typical for pastoral sheep and goat flocks in Kenya, an important exception being Maasai producers in Kajiado District who attempt to control breeding of their smallstock using breeding aprons (de Leeuw & Peacock, 1982; de Leeuw et al., 1991). Continuous breeding is possible due to the ability of local sheep and goat breeds to reproduce year round. The principal advantage of an aseasonal breeding regime is a continuous supply (although in seasonally varying proportions) of milk, meat and surplus animals to pastoral households, and the low input of labour and management required. Yet, this strategy produces low milk yields, kid survival and growth whenever late pregnancy and birth fall into periods with suboptimal forage availability (Mellado, et al., 1996; Wilson et al., 1985). Moreover conception rate, prolificacy and kidding rates are compromised whenever mating occurs under poor nutritional conditions (Delgadillo & Malpaux, 1996; Walkden-Brown & Restall, 1996). In sum, an aseasonal breeding regime implies that nutrient requirements of the flock remain relatively constant over the year; due to variation in pasture growth, this leads to both wastage and deficiencies (Carles, 1983).
The control of reproduction might allow the grouping of stages in the productive cycle with critical nutritional requirements during the limited periods in which these requirements can be covered by the forage produced. Although it would be necessary to split up herds in several management units, and thus increase labour demands for herding, watering, and breeding management, significant economies of scale are likely to be achieved from creating more homogenous groups of animals. For example, guarding lambs and kids around the homestead would be limited to certain periods in a year, thus saving labour for other tasks (de [page 15↓]Leeuw et al., 1991). In addition, management practices such as castration and vaccination would be carried out more efficiently, since they would be performed on a larger number of animals on a few occasions only.
However, past experience with development interventions aimed at altering existing pastoral techniques of livestock and rangeland management indicate that due attention has to be payed to key features of these systems, among which uncertainty in production conditions and its effects on pastoral households is probably the most pervasive and serious one (Ellis & Swift, 1988; Mace & Houston, 1989; Southey, 1992). Intra-seasonal and inter-year variations of climate result in uneveness in resource requirements and the flow of output of livestock herds. Decision rationality dictates that pastoral producer's behaviour and decision making are not invariant to the perception of this production risk. Consequently, situations may arise in which some of the various shortfall minimizing pratices developed by traditional herdsmen are incompatible with the adoption of a seasonal breeding strategy. For instance, limiting reproductive activity in smallstock herds to a short period in a year implies a once a year reproductive schedule, such that all outputs will also tend to be produced over a short period of time. Inadequate rainfall during stages of the production cycle with high nutrient requirements, i.e., from late gestation until weaning, may precipitate high rates of reproductive wastage, poor survival of both breeding animals and youngstock, and low milk output from the herd. Hence, such uncertain events may result in a considerable loss of production and may undermine the survivability of pastoral households and their smallstock herds. Also, the once-a-year pulse of output asssociated with a seasonal breeding strategy may conflict with the role of smallstock as a buffer against insecurities and imbalances in food supply. It could be argued that an aseasonal breeding strategy is better suited to ameliorate the effects of uncertainty and to produce a more constant flow of goods, although at the cost of a somewhat lower long-term average herd productivity.
Current state of knowledge
The restriction of breeding as a management strategy to match nutrient demands in pastoral sheep and goat flocks with seasonal feed supplies has so far received little attention in research, although it is often considered to be an important step in improving small ruminant productivity ( Carles, 1983; Field, 1984; Bradford & Berger, 1988; Delgadillo, 1996). From observations made on Rendille goat herds, Carles (1986) concluded that birth, just before or just after a rainy period (defined as a six-week period receiving measurable rainfall), resulted in the highest growth rates in kids. However, no account of environmental influences on conception and birth rate, litter size, kid survival or lactation yields was given. In a simulation study, Smith et al. (1982) used the same data to determine the optimum seasons for kidding with regard to milk yield and kid survival. Based on a single-animal simulation of a two-year old Small East African doe, kidding in October at the onset of the short rainy season was found to result in the highest milk yield, kid survival and growth rate.
The most detailed assessment of the impact of seasonality on pastoral sheep and goat flocks is that of Wilson et al. (1983, 1984, 1985). This study was conducted on a Maasai group ranch in Kajiado District, Kenya, and revealed a marked effect of the season of birth (short rains, short dry, long rains, long dry) on survival and growth rates in lambs and kids until weaning. Although some differences in reproductive performance, as determined by litter size and parturition interval, were observed between seasons of birth, the results were inconclusive with regard to the factors which may have influenced these two traits. This was due to the observational nature of the study: breeding was actually not controlled experimentally. Since only the date of birth and not the time at which females were re-bred after giving birth was recorded, the reported estimates of parturition intervals are unlikely to be free of error (see Upton (1989) for an account of problems associated with defining and estimating parturition intervals).
Due to the non-experimental nature of the study, inferences concerning the effect of environmental conditions at the time of mating upon key parameters determining reproductive performance such as conception, birth and prolificacy rates could not be drawn. Indeed, the finding that the highest productivity in Maasai sheep and goat flocks (measured as total live weight (kg) weaned per dam post-partum weight (kg) and year) is achieved when births occur during the short dry season (see Wilson et al., 1985) conflicts with that of de Leeuw et al. (1991), which suggest that restricting breeding might have caused poor reproductive performances in Maasai group ranches. The latter authors argue that restricting breeding to a 3-4 month period beginning at the onset of the long dry season leads to high flock productivities only if, prior to the mating season, good rains, or the movement of flocks to pastures with adequate forage ensure high conception rates.
It is difficult to assess from the above information if controlled breeding is indeed an effective strategy for alleviating the constraints imposed by seasonal nutrient shortages. And in the light of the limited options available for improving livestock production in pastoral systems, it is important to increase our understanding of the potential benefits and pitfalls associated with controlled breeding as a management strategy.
[page 16↓]Methodological issues
In studying the effects of seasonal breeding on biological productivity of pastoral goat herds two aspects deserve special attention. Firstly, seasonality is a composite of a number of environmental effects that have not only an impact upon, in the short term, performance parameters of goat herds, but also exert a sequential influence over the entire production cycle. Therefore, in studying the effect of a restricted breeding regime on herd productivity, it is important to extend herd recording, at least from the time at mating, until the time of disposal of young stock. In the presence of large, climatically induced variations in production conditions between both years and seasons, studies of controlled breeding should be designed to cover at least two complete productive cycles in different years, whereas the treatments should consist of several consecutive seasonal "breeding groups" in which breeding is confined to a fixed period of time. Both aspects of experimental design require setting up several experimental groups of animals which need close supervision and management. These conditions are essential to achieving an acceptable degree of comparability and are generally difficult to meet in traditionally-owned pastoral herds. The particular role of experimentation is emphasized here because a large proportion of previous reports on the effects of seasonality on African pastoral goat herds were purely retrospective in nature, such as that of Wilson et al. (1983, 1984, 1985) mentioned above. Testing specific hypothesis with data that were collected in observational or survey-type studies tends to be problematic, because, often, the set of data to be analysed was not, or not optimally designed for this purpose. With such data, influences that are not under the control of the investigator are almost always present and acting on the study objects. The main problem is that in the presence of confounding influences, it is logically not possible to attribute causality to perceived differences between measured variables (Underwood, 1996). This is confirmed by the study of Wilson et al. (1983, 1984, 1985) who found that the most general influence on productivity of sheep and goats was exerted by flock ownership. Although differences and relations between measured variables may be detected in observational studies, their interpretation is problematic and requires caution (Hurlbert, 1984; Jager & Looman, 1995).
Secondly, productivity assessments at the flock level that incorporate all outputs (meat, milk etc.) should be used to compare levels of technical efficiency achieved in flocks mated in different seasons. Available efficiency indices and their suitability for productivity assessments in small ruminant production systems have recently been reviewed by Bosman et al. (1997). These indices typically relate gross output in terms of monetary value, mass or energy to the number of animals (or their total liveweight, or metabolic liveweight) required to produce it. The majority of the available measures of efficiency is geared towards single output meat systems, and therefore are, at best, only suited for evaluating isolated factors contributing to overall flock productivity, such as reproductive performance of breeding females. Although Bosman et al. (1997) discuss various fallacies and drawbacks associated with the individual approaches, they fail to address two more fundamental aspects which, when ignored in computing efficiency measures, can lead to questionable results in livestock herd productivity assessments. These relate to the effect of population structure and population dynamics upon the calculation of efficiency measures.
Neglecting for the moment effects of genetic and environmental origin, performance traits such as the probability of conceiving, survival, litter size, milk yield and liveweight gain are undoubtedly a function of the stage (e.g., age, parity number, age within parity number, or another biologically meaningfull categorization) an animal has reached in its species specific life-cycle. For example, it is well established that fecundity in breeding does increases almost linearly with parity number, whereas the amount of milk produced increases curvilinearly with the number of lactations (Raats et al., 1982; Devendra & Burns, 1983). Consequently, these changes in performance over successive life-cycle stages must affect the computation of productivity measures whenever the stage classes are not represented in equal proportions in the sample data. Irrespective of the type of index used, i.e. reproductive, productive or flock performance, the bias introduced by unequal stage abundances in a herd can be controlled for by requiring the underlying distribution to be stationary, i.e. relative numbers of individuals in each stage do not change over time (Putt et al., 1987). If no adjustment is made for such a stable herd or flock structure, comparisons between different breeds or between flocks of the same breed under different environments, treatments, or management systems made on the basis of productivity indices are likely to lead to invalid inferences. Also, as has been pointed out by Upton (1989,1993) and Baptist (1988, 1992a), evaluations of this kind are only valid if carried out under 'standard conditions' with respect to herd growth and size. Allowance has to be made for breeding flock replacements in order to avoid comparing flocks which grow at different rates. This problem may be resolved by keeping herd size constant and considering the differences in productivity between each case that could be achieved subject to this constraint (Upton, 1989).
Available efficiency measures for productivity assessments at the flock or herd level do not correct for the bias introduced by differences in population structure. The denominator of the respective output/input ratios are usually calculated as averages over all life-cycle stages, such that all individuals are treated as identical. Peacock (1987) argues that using the mean weight of the flock as denominator can serve as a representation [page 17↓]of the demographic structure of the flock. However, whether liveweight alone can serve as a predictor of an individual's future survival, growth and fecundity is questionable, and the problem that demographic stability is implicitly assumed still prevails with this approach. Similar arguments hold with respect to the effect of herd growth upon the calculation of productivity indices. Although attempts have been made to allow for overall net inventory changes (e.g. Peacock's (1987) flock productivity index), this cannot be regarded as an appropriate representation of the underlying population dynamics, since accurate assessments of inventory changes must be based on estimates of changes in animal numbers in each stage class and sex cohort (Upton, 1989).
Several approaches based on herd growth modelling techniques have recently been developed to overcome the shortcomings of conventional productivity measures. The use of a so-called steady-state herd model along with an appropriate measure of technical efficiency, e.g., total output produced (measured in monetary or energetic terms) per unit feed energy requirement when the population is in a steady-state with respect to population dynamics, has been advocated by Baptist (1992a), James and Carles (1996), and Upton (1989, 1993) for making standardised comparisons. With a steady-state herd model, the stable structure of a livestock herd is computed for a specified set of stage and sex specific survival and fecundity parameters. In order to maintain herd size constant over time, a culling regime has to be imposed, which in its simplest form consists of culling at a pre-specified age all female and male surplus animals not needed as breeding stock replacements. The productivity index itself is then, for example, computed by dividing the sum of the value of all outputs (milk, meat, etc.) by total feed-energy requirements of the herd at equilibrium.
The techniques that have been developed so far for assessing herd productivity at the steady-state are based on an age-classified approach of modelling herd dynamics over time. Although classifying individuals in a population by age cohorts is the the most straightforward and widely used classification, age-structured population models implicitly assume that properties other than age are irrelevant to an individual's demographic fate. From this is follows that if vital rates (i.e. rates of survival, growth, and reproduction) also depend on factors other than chronological age, these must either be highly correlated with age or the distribution of individuals among the relevant categories must be stable. However, as indicated above the demography of domestic livestock species may depend on size or developmental stage much more than on age (e.g., parity or size), and often these variables are only weakly correlated with age. Situations may also occur in which the age of individual animals is difficult to determine accurately, but other characteristics such as body size, reproductive status, or parity number may be more convenient to measure and more pertinent to questions relating to population dynamics.
A further restriction of currently available methods for steady-state herd productivity assessments (Baptist, 1992a; James & Carles, 1996) are that they allow for only a broad categorization of the different types of animals in a herd into a rather limited number of classes. Additionally, within these categories performance traits cannot be made dependent upon the age an animal has reached in its species-specific life-cycle. Especially vital parameters and performance traits pertaining to breeding females such as fecundity, litter size, milk yield, liveweight and liveweight gain are assumed to remain constant throughout their productive lifespan. The methods proposed by Baptist (1992a) and James and Carles (1996) also impose restrictions, to varying degrees, upon the specification of survival rates. The approach of James and Carles (1996), which has been implemented in a computer program, LPEC (PAN Livestock Services, 1991), assumes constant survival rates in each animal class, irrespective of age. The computer program (PRY) developed by Baptist (1992a) is somewhat more flexible in this respect, but it still assumes, for instance, that the same survival rates apply to surplus and breeding females in a herd, as well as to reproductively active and (temporarily) infertile breeding females.
These limitations can have profound impacts upon the determination of herd dynamics, herd structure at the steady-state, and hence optimal offtake rates and productivity. In order to refine the concept of steady-state herd productivity assessment and improve its flexibility and applicability, it therefore appears worthwhile to explore alternative methods of modelling herd dynamics over time and of deriving optimal steady-state herd structure and offtake rates.
In view of the limited information available on the effect of seasonality upon small ruminant productivity in semi-arid areas, as well as on the advantages and disadvantages associated with a controlled breeding regime, a systematic breeding programme in a herd of Small East African (SEA) goats was initiated for a period of three years (1984-1987) at the Ngare Ndare research station of the University of Nairobi in Isiolo District, northern Kenya (Schwartz & Carles, 1987). Part of the data collected in this experiment has already been published in a series of papers reporting either preliminary results (Rutagwenda et al., 1985; Schwartz & [page 18↓]Carles, 1987), or investigating factors affecting kid mortality (Gachuiri et al., 1986) and milk yields (Wahome et al., 1994) of SEA goats under semi-arid conditions in northern Kenya. However, a detailed statistical analysis of these data covering all parameters relevant to assessing overall flock productivity has not yet been undertaken. Similarly, a systematic investigation of the effects of the implemented controlled seasonal breeding regime on overall performance of experimental flocks is not yet available.
Using this set of data, the present study was undertaken to (1) assess the effect of a seasonal breeding regime on various traits determining biological performance of pastoral goat flocks, and (2), to test the hypothesis that there is an optimal period in a year to which breeding can be restricted to improve overall biological flock productivity. However, the foregoing section made clear that the adoption of the latter hypothesis may be considered a necessary, but not sufficient condition to prove the superiority of controlled versus uncontrolled breeding. In particular, it will be necessary to address the question whether, in biological terms, restricting reproductive activity in pastoral goat herds to a single short mating season at a particular time of year entails a significant increase in production risk. Also, the distinctive role of goats in buffering insecurities and imbalances in food supply to pastoral households needs to be taken into account. In this end, goats traditionally are reared for dual-purpose meat and milk production. Nevertheless, the transition towards a more sedentary life-style is likely to increase the pressure on pastoral producers to specialise and commercialise their mode of production. It is therefore interesting to test the hypothesis whether different optimum breeding periods can be identified for a dual purpose meat and milk production and for a single purpose meat production.
This thesis is structured in 8 chapters. With the exception of the present and the final chapter, all other parts (Chapter 2 through 7) were written with the intention to be self-contained and to stand on their own as independent papers. Chapters 2 to 5 will be devoted to the statistical analysis and interpretation of the effects of six different, seasonally restricted breeding regimes on, respectively: reproductive performance traits; survival rates of kids and does; growth performance of kids and body weight development of does; and milk production. As the first paper in this series, the materials and methods section of Chapter 2 provides a detailed description of the study area and study design
In Chapter 6 an extension to available methods of steady-state herd productivity assessment and modelling will be developed. It will be based on a stage-structured matrix population model and use non-linear mathematical programming to derive optimal herd structure and stage-specific offtake rates. Estimates of traits of reproductive and productive performance, as well as of survival obtained from the statistical analyses of the experimental data presented in preceeding chapters will be used in Chapter 7 to parameterise the proposed productivity assessment model for six different breeding seasons and an aseasonal reference group. This modelling exercise is carried out in order to test the formulated hypotheses concerning the effects of seasonal breeding on the biological performance of pastoral goat flocks. Finally, Chapter 8 provides a summary and a general discussion of the main themes and results of this work.
|© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.|
|DiML DTD Version 3.0||Zertifizierter Dokumentenserver|
der Humboldt-Universität zu Berlin