+Bioline International Official Site (site up-dated regularly)

Opportunities for improving integrated nutrient management by smallhold farmers in the Central Highlands of Kenya

P. L. WOOMER , N. K. KARANJA and J. R. OKALEBO1
Department of Soil Science, University of Nairobi, P. O. Box 30197, Nairobi, Kenya
₁Department of Soil Science, Moi University, P. O. Box 1125, Eldoret, Kenya

Integrated Nutrient Management (INM) is an important strategy for soil improvement in smallholder systems of East Africa but successful implementation depends upon identification of opportunities for intervention, which are in turn dependent upon a practical knowledge of farm enterprises and decision making. Surveys addressing soil fertility management were conducted among 190 households in the Central Kenyan Highlands. Farmer-available organic resources were collected during these interviews and later analysed for plant nutrients. The business strategies of 139 retailers of agricultural supplies were assessed through a 23-question survey conducted in 17 districts in the Kenyan highlands. These surveys provided background for identifying promising interventions by farmers to improve soil fertility and for retailers to meet farmers’ demands. Soil fertility depletion results from both internal and external flows of nutrients and is often expressed as nutrient deficient patches in farm "outfields". An approach to improving INM separates management practices into four general categories; nutrient recycling, livestock-crop interactions, biological nitrogen fixation and fertiliser use. Fifteen specific farmer practices as candidates for strengthening INM were identified within these four categories. Our findings suggest that farmers in the Central Highlands of Kenya are practising the fundamentals of INM but their efforts are restricted by the availability of organic resources and investment capital.

Key Words: Biological nitrogen fixation, crop-livestock interactions, East African Highlands, fertiliser, organic-inorganic interactions, nutrient replenishment

La gestion intégrée des élements nutritifs (INM) est une stratégie importante pour l’amélioration des sols dans les systèmes des petits exploitants de l’Afrique de l’Est mais le succès de la mise en application depend de l’identification et opportunités d’intervention, qui de retour dépendent aussi de la connaissance pratique des entreprises des fermes et de la prise de décision. Des enquêtes adressant la gestion de la fertilié du sol ont été conduites chez 190 exploitants dans la partie Mantagneuse Centrale du Kenya. Les resources organiques disponibles chez le fermier ont été collectées pendants deux interviews et plus tard analysées pour les élements nutritifs des plantes. Les stratégies de commerce de 139 détaillants d’ intrants agricoles ont été évaluées à travers une enquète de 23 questions conduites dans 17 districts dans les régions Mantagneases du Kenya. Ces enquêtes ont fourni de données de base pour l’identification des interventions promettantes pour les fermiers pour améliorer la fertilité du sol et pour les détaillants rencontrer les demandes des fermiers. La dégradation de la fertilité de sol résulte des mouvements internes et externes des éléments nutritifs et est exprimée comme spotte d’élements déficients en champs "outfields". Une approche d’améliorer INM sépare les pratiques de la gestion en 4 catégories principales: recyclement d’éléments nutritifs; interaction entre elevage et cultures; la fixation biologique de l’azote et l’utilisation des engrais. Quinze pratiques spécifiques des fermiers ont été identifiées comme candidats pour renforcer l’INM à l’interieur de ces quatre catégories. Nos résultats indiquent que les fermiers de la partie montagneuse Centrale du Kenya pratiquent l’essentiel de l’INM mais leurs efforts sont limités par la disponibilité de resources organiques et du capital d’investissement.

Mots Clés: Fixation biologique de l’azote, intéraction elevage-culture, zones montagneuses de l’Afrique de l’Est, engrais, intéraction organique et inorganique

Integrated Nutrient Management (INM) seeks to maximise the complementarity of mineral and organic nutrient sources (Jenssen, 1993), and is emerging as an important approach in improving the productivity of soils in smallholder farming systems of Kenya. The role of mineral fertilisers requires special attention within INM strategies of these systems because of farmers’ strong orientation towards optimising available organic resources. Fertilisers are not considered the principal source of farm nutrients but rather as one of a "menu" of interventions to be employed and combined given the changing circumstances of available organic and financial resources.

Soil fertility depletion was described by Sanchez et al. (1996) as "the fundamental biophysical root-cause of declining per capita food production in Africa". Decline in per capita food production of 7% was observed between 1980 and 1995 in Africa compared to a 25% increase in Asia over the same period (World Bank, 1996). The average annual loss in soil nutrients of 42 kg N, 3 kg P and 29 kg K ha^-1 in Kenya is among the greatest in Africa (Smaling, 1993). Nutrient depletion at the farm-scale results from an imbalance of nutrient inputs and losses over time and reaches critical proportions when land is continuously cultivated without the addition of adequate external nutrient inputs. Nutrient depletion at the field-scale is also caused by the internal flows of organic resources, particularly crop residue harvested as livestock feed and subsequent allocation of manures to higher-value crops and/or kitchen gardens (Fig. 1).

During the past few years, our research team at the University of Nairobi has conducted several different types of investigations on long-term soil management (Kapkiyai et al., 1999), fertiliser retailing (Mwaura and Woomer, 1999), on-farm limiting nutrient field trials (Woomer et al., 1997a), farmer awareness of legume nodulation and inoculants (Woomer et al., 1997b), nutrient depletion in farmers’ fields (Murage, 1998) and household decision making on nutrient resource use (Kapkiyai et al., 1998; Woomer et al., 1998a) relevent to smallhold farmer conditions in the Kenyan Highlands. In this paper, we gather this diverse information, combine it with recent chemical analyses of farmer-available organic resources, and interpret these findings within the context of identifying realistic opportunities for improving nutrient management in smallhold farming systems of the Central Highlands of Kenya.

Household survey. A 22-question, formal survey was designed to characterise patterns of resource management within smallhold farming systems of the Central Kenyan Highlands. Farmers selected at random from district-level agricultural extension records were queried about the size and age of their farm, the number and gender of occupants and farm workers, their principle crops and yields, number of livestock, what those livestock are fed and whether or not livestock are raised in confinement, use of fertilisers, animal manures and crop residues, preparation of composts, and combination of organic and inorganic material as soil amendments (Kapkiyai et al., 1998). The survey was conducted among 190 farm families between June and September, 1995, and covered four adjacent districts in central Kenya; Machakos, Thika, Kiambu and Nyandarua. These districts represent an elevational transect of 1100 to 3000 m from east to northwest of Nairobi. Data were compiled within a computer spreadsheet, with individual farms represented as rows and responses as columns, inspected and then imported into a statistics programme where summary analyses were performed.

Retailer survey. The business strategies of 139 retailers of agricultural supplies were assessed through a 23-question, formal survey conducted in 17 districts in the Kenyan Highlands between June 1996 and February 1997 (Mwaura and Woomer, 1999). An enumerator visited 74 business centres of larger villages and towns. Between one and six retailers were surveyed at each business centre. Shop operators were queried concerning the number of years in business, types and prices of fertilisers marketed, whether or not they repackaged fertilisers, the relative importance of pesticide, seed, livestock supplies and fertilisers, sources of new products, knowledge of local extension agents, availability and distribution of written information and the availability and extension of credit. During the interview, the enumerator estimated the size of the retail outlet and the proportion displaying fertilisers. Survey results were compiled into a computer spreadsheet with cases (interviews) entered as rows and responses as columns. The spreadsheet was inspected for accuracy and then imported into a computer statistics programme for sorting and calculation of summary statistics.

Organic resource collection and chemical analyses. Organic resources available to and in use by farmers were collected during the farmer survey, transported to the labotratory in a cool box, dried at 65o for 24 hours, ground to pass through a 1.0 mm sieve and analysed for total N, P, K, Ca and Mg as described by Okalebo et al. (1993). Ash contents were determined by slow ignition of samples in a muffle furnace brought to a final temperature of 550oC (Okalebo et al., 1993). The results were compiled into a spreadsheet data base with individual organic resource collections entered as rows and nutrient contents as columns, inspected for accuracy, imported into a computer statistics programme and summary statistics and linear regressions performed.

Integrated Nutrient Management (INM) recognises that nutrient inputs may be applied as both mineral fertilisers and organic resources and that different nutrient sources may be substituted for and complement one another. In principle, a requirement for nutrient inputs is established based upon yield targets and an estimate of losses within a farm and are replaced on an annual or cropping season basis (Jenssen, 1993). The approach seeks to maximise nutrient recycling, maintain soil organic matter and protect beneficial soil biota (Palm et al., 1997). In practice, smallholders rely upon integrated approaches to soil fertility management by taking advantage of nutrient recycling, particularly by combining crop and livestock activities. However, crop yields decline due to inadequate addition of fertilisers and nutrient-rich organics obtained from outside the farm fail to offset losses (Table 1).

Table 1. Land management practised in maize-bean cropping systems of the Central Highlands of Kenya (from Kapkiyai et al., 1998b)¹

Land Management	% of respondents

Apply animal manure to field crops	98%
Feed maize stover to livestock	95%
Raise cattle under confinement	91%
Apply inorganic fertilisers to field crops	81%
Combine organic and inorganic inputs	79%
Produce domestic composts (including manure compost)	56%
Consider soil conditions a major farm constraint	33%
Sell organic residues to others	7%
Apply no external inputs and remove all maize residues	0%

¹based on household ranking of cropping priorities, 79% of 190 farms were classified as maize-bean cropping systems (n=168).

Mineral fertilisers offer several advantages to the smallhold farming sector (Table 2). Fertilisers are the most concentrated form of plant nutrients, offering both cost and labour efficiency. Most fertilisers rapidly dissolve into plant-available forms upon application. Well-informed buyers know the nutrient contents of purchased fertilisers which facilitates calculation of application rates and timing. However, relatively few formulations are available through imperfect markets and information networks that limit their use (Table 3). Fertiliser application may pose additional risk to cash poor farmers because investment returns depend upon other growth conditions (e.g. drought) and changing commodity prices. Furthermore, greater knowledge of fertilisers is required among smallhold farmers because many cannot identify optimal formulations and application rates or timing and there may be negative environmental consequences resulting from fertiliser misuse.

Feature	Nutrient source
	Mineral fertiliser	Organic resource
Nutrient concentration	higher and known based upon labeled nutrient contents	lower, bulky, unknown and variable between "batches"
Nutrient availability	rapid chemical dissolution, subject to loss through leaching and sorption	slower release, regulated and protected by soil biological processes

Acquisition and cost	costly, purchased in imperfect markets with few credit opportunities	locally produced or gathered, often in short supply and with competing uses

Labour requirements	easily applied and compatable with other field operations	higher recovery and handling efforts, may interfere with other field operations
Environmental impacts	negative at excessive rates, pollution of groundwater and aquatic systems	positive, favours carbon sequestration and soil biodiversity

Table 3. Business activities of fertiliser retailers in Kenya (after Mwaura and Woomer, 1998)

Business practice	% of respondents

Recommend products to customers	81%
Repackage fertilisers into smaller sizes	79%
Report periodic shortages	61%
Provide written information to customers	39%
Extend credit to customers	31%
Consider fertilisers their most important commodity	29%
Able to identify their local extension agent	19%
Receive credit from suppliers	14%

Based on interviews of 139 fertiliser retailers in 74 business centres from 17 Districts of Kenya

The application of organic resources as soil amendments offer a different set of advantages and disadvantages. Nutrient ratios of organic resources are often present in the proportion required by plants. The chemical composition of organic inputs regulates the rate of nutrient release and may be added to soils in a manner which allows a steady supply to plants (Myers et al., 1994). Decay products of organic additions reduces the activity of toxic soil cations such as aluminum (Woomer et al., 1994). Furthermore, environmental benefits are realised from the addition of organic resources through C sequestration as soil organic matter (SOM) (Woomer et al., 1998b), and as an energy source to beneficial soil biota (Lavelle et al., 1994; Palm et al., 1997). However, organic materials tend to be bulky and have reduced nutrient contents compared to fertilisers (Table 4), even manures and composts (Table 5), resulting in high labour requirements. Furthermore, the nutrient contents of organic resources vary greatly between farms and "batches" of manures and composts (Fig. 2). Nutrient release occurs through soil biological activities at rates less rapid or predictable than that of fertilisers (Myers et al., 1994). At the farm level, organics have unknown compositions with much variation in nutrient contents between resources and with handling and storage. The decay of low quality organic inputs immobilises soil nutrients and temporarily reduces soil fertility and many crop residues harbour pests and diseases, particularly with continuous, non-rotational cropping. When the advantages and disadvantages of fertilisers and organic resources are compared (Table 2), their complementary natures emerge.

Table 4. Frequency of use and nutrient contents of some commonly available organic resources among smallhold farmers in the Central Kenyan Highlands

Resource	Household use¹		Nutrient content
	Soil input	Animal feed	N	P	K	Ca	Mg

	%

Napier grass	na²	98	1.02	0.11	2.63	0.35	0.06
Maize stover	97	92	0.89	0.08	2.78	0.41	0.18
Bean trash	88	81	1.20	0.13	2.06	0.89	0.16
Banana trash	59	40	0.83	0.06	4.54	1.22	0.22
Wood chips	na	55³	0.43	0.04	0.12	0.23	0.09
Sweet potato vines	na	50	2.27	0.14	3.05	1.32	0.53
Coffee husks	21	nf⁴	1.01	0.05	1.06	0.77	0.02
Pigeon pea prunings	15	16	1.33	0.10	1.02	0.37	0.09
Cowpea trash	9	12	0.57	0.05	1.79	0.81	0.08

¹ Based on interviews with 190 households in central Kenya; ² infrequently applied to soils; ³ approximation based on the frequency of households raising confined poultry, commonly applied as poultry bedding material; ⁴ not fed to livestock

Table 5. Frequency of use as a soil input and nutrient contents of composts and animal manures in the Central Kenyan Highlands

Resource	Household use as soil input¹	Minerals (ash)	Nutrient content
			N	P	K	Ca	Mg
			%

Cattle boma² manure	92	53	1.40	0.20	2.38	0.39	0.27
Poultry manure	65	35	3.11	0.42	2.40	0.82	0.42
Goat/sheep manure	65	41	1.48	0.20	3.31	0.94	0.42
Domestic compost	58	70	1.34	0.20	1.82	0.39	0.22
Swine manure	6	63	1.40	0.23	2.02	0.99	0.19

A "menu" for farmer intervention to correct declining soil fertility is under development by university-based scientists in Kenya and Uganda (Woomer et al., 1998a) that is based on principles of integrated nutrient management (Fig. 3). The approach focuses upon improvements in nutrient recycling, livestock-crop interactions, biological nitrogen fixation by legumes and fertiliser use.

Improved nutrient recycling. Nutrients are recycled when crop residues, animal manures, tree prunings and other non-harvest organic materials are returned to the soil. Strategic application of farm organic resources and their supplimentation with externally derived nutrients are fundamental principles of INM. The results from the household survey (Table 1) suggest that smallhold farmers are concientious in their collection and use of farm resources and frequently combine these with external nutrient inputs. However, the quantity of organic resources available to farmers appears constraining as evidenced by the paucity of under-exploited materials. Impressions gained during the course of farm visits suggest several possibilities for improvement.

Trashlines in croplands (1). Coarse crop residues, particularly maize stalks and bean stems, are recovered by cutting at ground level following harvest and offered to livestock as feeds, only to be rejected. An alternative use of coarse residues is their placement as "trashlines" along slope contours. These trashlines reduce runoff resulting from seasonal and heavy early rains and during the following season decompose sufficiently for ready incorporation into the soil. In this way, the fraction of crop residues least valuable to livestock operations is retained in fields devoted to food crops and offers both short-term (erosion control) and longer-term (input to SOM) benefits.

Improved external (2) and internal (3) farm boundaries. Another means of increasing the availability of organic resources produced on-farm may result from improved use of farm and field boundaries. Often, farm boundaries are marked by unmanaged stands of indigenous and naturalised weedy shrubs that are viewed as satisfactory by farmers provided that hedgerows are sufficiently dense to exclude neighbouring livestock and wildlife. Examples of persistent, difficult to manage, wild hedgerow species are the indigenous Euphorbia trucalli and naturalised Lantana camara. Native trees are often absent from farm boundaries, both because of their shading effect on small fields and household demands for wood fuel. Farmers facing shortages of soil organic inputs or feeds should better plan their management of farm and field boundaries. Multi-purpose, fast growing trees may be planted along outside boundaries such as indigenous Croton meglacarpa, naturalised Grevillea robusta or recently introduced Calliandra calothrysus. Interior boundaries, when present, tend to be better planned and managed, and are frequently oriented along slope contours as erosion control strips between annual crop outfields. In many cases, these interior boundaries are also exploited for fodder, particularly napier grass (Pennisetum purpureum), and as such may contribute to enternal nutrient flows resulting in fertility-depleted annual crop outfields. An alternative approach is to establish plant species that may be periodically prunned for use as soil inputs within the annual crop outfields and to establish fodder plants as hedgerows closer to livestock operations. In this way, one of the pathways to nutrient depletion are disrupted and the distance between fodder production and livestock shortened. Tithonia diversifolia is currently under investigation as a boundary shrub suited to providing organic inputs to soils (Sanchez et al., 1996).

Pasture or trees in degraded lands (4). As nutrient depletion continues through both an imbalance of external flows and internal redistribution, farmers are confronted with non-productive annual crop outfields, plots and patches expressing severe deficiency symptoms (Fig. 1). Too often, farmers accept this condition and respond by continuing to cultivate low yielding cereal-legume intercrops. An alternative is to allocate degraded lands to ley pastures, fodder plots, orchards or woodlots, uses which are documented to restore soil fertility while providing household products (Young, 1989; Boonman, 1993). Another alternative is planting short-term improved fallows (Buresh and Cooper, 1999), usually with leguminous shrubs and trees such as Tephrosia vogelii or Sesbania sesban (Kwesiga and Coe, 1994).

Strengthened livestock-crop interactions. Traditional highland farmers value livestock as buffers against food shortages and sources of appreciating wealth and manure and to a lesser extent, draft power. Crop and livestock enterprises coexist as functionally-linked enterprises where crop residues are utilised as feeds and resultant manures provide organic inputs to soils (Fig. 1). An evolution toward maintaining livestock within stalls is inevitable as population pressure reduces land availability but simultaneously increases demand for animal products, including manure (Powell and Williams, 1994). As dependance upon manure as a crucial input to maintenence of soil fertility increases, farmers then seek means to improve the quantity, recoverability and quality of manures and to integrate them more effectively with other nutrient sources. The importance of livestock in highland maize-based smallholdings is highlighted in a long-term (18 year) experiment where stover retention resulted in reduced yields but application of farmyard manure improved crop productivity (Kapkiyai et al., 1998).

Increasing livestock number and quality (5 and 6). Increasing the number of livestock within the farm correspondingly increases manure supply, but this becomes a less viable option as land use intensifies. Improved breeds, particularly dairy cattle, increase the profitability of livestock enterprises and provide greater incentive to raising more animals on a larger proportion of purchased feeds but this opportunity is constrained by market access and availability of investment capital. An alternative to investment in livestock operations is the initiation or expansion of market-oriented poultry enterprise where chicken and egg sales offer short term returns through strong local and urban markets. Poultry manures are exceptionally rich in plant nutrients (Table 6).

Table 6. Forms, frequency, Kenya Shillings per 50 kg bag, nutrient contents and unit costs per nutrients of selected fertilisers marketed by agricultural retailers in Kenya (Mwaura and Woomer, 1999)

Form	Frequency¹	Price	Nutrients	Unit cost²
	(%)	(KSh 50 kg-1)	(NPK 50 kg-1)	(KSh kg-1)
DAP	94	1555	9 kg N,10 kg P	81
CAN	81	1170	13 kg N	90
Urea	43	1335	23 kg N	58
Triple 17	28	1225	9 kg N,4 kg P,7 kg K	64
TSP	11	1294	10 kg P	129

¹ Frequency (%) of sales by 139 local retailers; ² equal values for N, P and K in fertilisers bearing multiple nutrients

Improved recovery of manures (7). In poorly-managed, confined livestock situations, bedding material in the stall accumulates from rejected stalks and stems offered as feeds and tends to be unevenly distributed and poorly sorptive of excretia. Recovery of manures from the floors of stalls may be greatly improved by either applying finer, nutrient-rich bedding material, such as tree prunings, non-palatable crop residues or wood chips, or by "paving" stalls with coarse-surfaced mixtures of cement and gravel. Smooth surfaces facilitate collection of excretia and better allow for separation of manure and urea but pose hazards to livestock when wet and slippery. The mixture of excretia, bedding and rejected feeds are then pushed to one side of the pen until their accumulation hampers further recovery operations and then moved to storage areas for "curing" or mixed with other organic materials and composted.

Improved handling and storage of manures and composts (8). During storage and composting, manures are exposed to nutrient losses through leaching and gaseous emission of nitrogen. Simply covering manure piles with plastic sheeting or thatch reduces these losses as well as retains moisture of composts. Drainage and aeration of manure piles and composts further reduces gaseous loss and accelerates the composting process. Many farmers do not avoid contamination of manures and composts with mineral soil and some layer soil and organic resources within composts as a means of conserving moisture. Higher mineral contents result in lower nutrient concentrations (Fig. 3, Table 6) that in turn raise future labour requirements to deploy manures and composts and provide disincentive to transport them to more distal areas of the farm. The resulting "necessity" of applying manures and composts closest to livestock operations may then lead to unbalanced internal nutrient flows and accelerated nutrient depletion of more distal croplands (Fig. 1).

Tether grazing (9). An option available to livestock managers is to graze livestock on crop residues for a short time following harvest. Deposition of manure may be targeted by the location and length of tethers. Some evidence suggests that crop residues grazed in the field are not utilised as efficiently compared to stall feeding because cattle are reluctant to browze near their recent excretia. On the other hand, less labour is required to lead cattle to harvested fields than to carry crop residues for stall feeding. Two disadvantages to tether grazing, however, are the increased risk of livestock theft and a reduced feed use efficiency.

Increased biological nitrogen fixation. A practical and inexpensive option for increasing nitrogen stocks available to farmers is through biological nitrogen fixation (BNF). The atmosphere contains 70% dinitrogen (N₂) but this form is not metabolised by plants. Bacteria are able to reduce N₂ into assimilatable nitrogenous compounds either as free living chemoautotrophs or as microsymbionts (Paul and Clark, 1989). The most readily managed BNF system available to farmers is through the legume-Rhizobium symbiosis but full benefits are not being realised for several reasons.

Increasing BNF by legume intercrops (10). A predominant cropping pattern is a continuous maize-bean intercrop. Beans often demonstrate reduced physiological potential for symbiotic BNF (Giller and Wilson, 1991), a powerful reproductive sink for translocated nitrogen during maturation and are readily nodulated by Rhizobium tropici, which is indigenous to East Africa and an ineffective microsymbiont of Phaseolus vulgaris (Giller et al., 1994). There is no lack of alternative legume intercrops suitable for tropical highlands (Duke, 1981), however, beans are prefered by households for their quick maturity, tolerance of short-term drought, ease in harvesting, rapid cooking and taste and many farmers are reluctant to consider other legumes. Pigeon pea (Cajanus cajan) and cowpea (Vigna unguiculata) have greater potential for BNF than beans in the middle uplands (1200-1600 m). Pea (Pisum sativum) and faba bean (Vicia faba) offer similar advantage at higher elevations. Other pulses with potential to increase BNF through replacement of intercropped beans include dolichos bean (Lablab purpureus), lima bean (Phaseolus lunatus), groundnut (Arachis hypogaea) and soybean (Glycine max) although these species may require greater management adjustments due to either aggressive climbing behaviour or reduced shade tolerance. Locally-adapted varieties of these alternative legumes are available through the Kenyan commercial sector and a wide variety of rhizobial inoculants are distributed by the University of Nairobi Microbial Resource Centre.

BNF by live mulches (11) and boundary plantings (2). Potential to increase BNF within smallhold farms also exists by establishing creeping and weakly trailing leguminous groundcovers as live mulch (e.g. Trifolium spp., Arachis pintoi) in perennial croplands, by including legume as a component of paddock improvements (e.g. Desmodium intortum, D. uncinatum) and by planting N-fixing leguminous shrubs and trees along farm boundaries and in woodlots (e.g. Calliandra calothrysus, Sesbania sesban). Woomer et al. (1997b) provide evidence that the limitation to improving BNF in East and Southern Africa is not lack of tools and germplasm, but rather farmers’ understanding the legume-Rhizobium symbiosis, recognition of effective nodulation and management of microsymbionts through inoculation (Fig. 4).

Strategic fertiliser application. The inability of agricultural planners, scientists and extension to translate the repeated experimental success of fertiliser application into routine farmer practice is the major constraint to combatting nutrient depletion in the Kenyan Highlands. In large part, this failure is due to scientists not taking economic conditions and commodity prices into account during their analyses. While it is widely accepted that agricultural technologies cannot be laterally transferred between high technology, commercial farmers in developed countries to traditional farmers in the tropics, methods of evaluating and promoting fertilisers bear tremendous resemblence. Field trials seek to develop recommendations based upon identified production optima for various broadly represe-ntitive soil-crop combinations, often with lowest application rates beyond what is affordable to smallholders. Furthermore, refinement of recommendations is designed to narrow geographic coverage rather than discovering ways to increase the use efficiency of affordable amounts of fertiliser through its combined use with other nutrient inputs. Blanket fertiliser recomme-ndations have their place, especially among commercial farmers. At the same time, smallhold farmers must be assisted in making better use of fertilisers to solve their production problems resulting from many years of soil degradation. These farmers are often reluctant to invest in fertilisers because of the risk of periodic crop failure or competing demands upon limited household incomes. Credit is not widely available for purchases of agricultural supplies (Table 2). Fertilisers may be marketed in quantities regarded as too large and expensive, and even when more suitably packaged, a limited range of products may force retailers to recommend improper fertiliser forms (Mwaura and Woomer, 1999). Smallhold farmers are less willing to innovate than many institutional agriculturalists imagine because the consequences of failure are often severe.

There are four qualitatively different approaches to fertiliser use. Replacement fertilisation regards mineral inputs as the sole external input and is intended for routine application at recommended levels. Nutrient replenishment seeks to recapitalise soils after several years or decades of losses as a means of restoring agricultural productivity and community prosperity (Buresh et al., 1997). Patch amelioration is designed to provide small amounts of fertiliser to farmers so that they may experiment with soil fertility problem solving on their least productive lands. Organic resource backstopping involves addition of small amounts of minerals to organic resources as a means of improving their nutrient contents and ratios, and to facilitate key biological processes such as composting. Further explanation on these approaches follow.

Replacement fertilisation (12). Strategies for replacement fertilisation are well developed in Kenya based upon the Kenya Agricultural Research Institute district-level recommendations resulting from the Fertiliser Use Recommendation Project. Multi-site field trials were conducted over seven years leading to specific crop recommendations, which in turn strengthens farmers’ capacity to invest in fertilisers. The recommendations are not absolute but rather provide a starting point for farmer innovation to optimise yields within market-oriented settings. The survey of fertiliser stockists suggests that fertilisers are available to farmers in remote rural settings (Table 4), particularly Diammonium Phosphate and Calcium Ammonium Nitrate, the formulations most widely recommended in Kenya.

Nutrient replenishment (13). Nutrient replenishment is a recent approach that views nutrient depletion as a loss of national resource capital (Buresh et al., 1997). It is directed toward nutrient-depleted smallholder farming communities and requires a large initial investment to replace nutrient losses that have resulted from many years of past land management. Benefits are viewed as shared by society as a whole, through revitalisation of food security, and farm communities, which then enjoy increased productivity and incomes. Plans are being formulated to assure that those who benefit share costs in an equitable manner and the approach is attracting the attention of donors (World Bank, 1996). In general, phosphorus replenishment of soils requires investment in phosphorus fertilisers while nitrogen status is improved through biological nitrogen fixation (Woomer et al., 1997a). The high profile emergence of nutrient replenishment onto the agricultural development agendas in Africa is necessary and welcome but confusion surrounds the details of various operational strategies and their possible conflict with recently-initiated structural reforms in fertiliser markets (Kimuyu, 1994; Alexandratos, 1997).

Patch amelioration (14). Patch amelioration focuses upon the low fertility patches which develop from smallholder practice in fields cultivated for household food. There are distinct advantages to management directed toward low fertility patches in that externally-supplied nutrients are being positioned farthest "upstream" in the farm nutrient recycling process (Fig. 3) resulting in increased food production for the household as well as more and higher quality livestock feed and manures. These patches are viewed as requiring specific management which is best provided as affordable products from local agricultural suppliers. One such product, PREP-PAC, combines fertiliser, legume seed, inoculant and information, is targeted toward a 25-50 m² low fertility problem area and is assembled for a cost of US $0.67 (Fig. 5). This approach is currently being tested in western Kenya through university-NGO-farmer collaboration and has rapidly attracted the attention of the private sector (Woomer et al., 1998a).

Organic resource backstopping (15). Organic resource backstopping is at a less developed stage of development and is a focus of research by many research institutes and local non-governmental organisations. It relies on the strategic combination of organic resources with mineral fertilisers such as composts fortified with rock phosphate or lime, mineral supplements to improve livestock gain and the manures they produce, combining nitrogen fertiliser with low quality crop residues to reduce nutrient immobilisation and improved legume seed inoculant delivery systems by pelleting with phosphates and lime. One of the key principles of this approach is to rely primarily on locally-available resources and to improve upon them with mineral fertilisers. In this way, the approach has a great potential for improving nutrient use efficiency and for rapid adoption by farmers and private sector.

Integrated Nutrient Management: from promises to products. Many farmer practices are consistent with the principles of Integrated Nutrient Management but nutrient inputs continue to limit yields. New strategies are being formulated to increase the availability of nutrient inputs but additional effort is required to formalise know-ledge into research products understood by, and affordable to farmers. Retailers provide a promising entry point to improved fertiliser availability and use but require better credit facilities and sources of information. Different strategies of fertiliser use improvement are required for different geographic areas, soils and scale of production. These strategies are either in place or under development in Kenya. Finally, maximising fertiliser use efficiency through combined use of organic and mineral resources is a promising entry point into smallholder systems and the fertiliser industry must view this approach as consistent with their interests for rapid advances to be made.

The useful discussions with and assistance of Dr. Paul Smithson of ICRAF, Mr. Francis Mwaura of Moi University and Mr. Kenneth Gathua of KARI during the formulation of this research and preparation of this manuscript are greatly appreciated. We thank two smallhold farmers in Kiambu District, Mrs. Itotia and Mr. Wagacha, for providing suggestions on approaches to farmer "menus" and the many other farmers and retailers who contributed to our surveys. We also thank the Rockefeller Foundation Forum for Agricultural Resource Husbandry for providing financial assistance to this study.

The following images related to this document are available:

Photo images