African Crop Science Journal, Vol. 7. No. 4, pp. 569-583, 1999
Prospects and constraints of finger millet production in Eastern Uganda
J. S. Tenywa, P. Nyende, M. Kidoido1, V. Kasenge1,
J. Oryokot2 and S. Mbowa1
Code number: CS99048
Finger millet (Eleusine coracana L.) is a major staple and cash crop in northern, eastern, western and southwestern Uganda. However, research on the crop has been limited. As such, a survey was conducted in eastern Uganda (Kumi, Pallisa and Kamuli districts) to establish the status of the crop, its production constraints and prospects for its development. A semi-structured questionnaire was administered to fifty households per district, during the period October-December, 1998. Land productivity has declined considerably in the study districts and low soil fertility is a major factor in this respect. Production of the crop is hampered by many constraints, the major ones being inadequate labour for weeding and harvesting, frequent drought, pests and diseases, and soil exhaustion. The region is highly dependent on oxen for land preparation. Finger millet production activities are generally gender neutral. The crop is grown almost exclusively once a year. Inter-cropping and crop rotation are common practices in finger millet production, and the role of finger millet as a cash crop is on the increase. Prospects for development of the crop lie in increasing yield through generation of labour-saving technologies, particularly for weeding and harvesting, and development of farmer-friendly packages for pests and disease management, drought mitigation and soil fertility management. A follow-up on-farm soil fertility management study was conducted to investigate the response of finger millet sown in rows or broadcasted, to various soil fertility management regimes. Row planting resulted in significantly (P<0.05) better finger millet growth and yield than broadcasting. Combined application of P and N, or P and manure gave the highest yields. Application of N alone delayed flowering and physiological maturity of the crop by 1 and 2 weeks, respectively. There was no significant (P<0.05) effect of planting pattern or fertiliser treatment on threshing percentage.
Key Words: Drought, Eleusine coracana, gender, labour, manure, soil fertility, weeds
Leleusine (Eleusine coracana L.) est une dénrée de base et une culture de revenu dans lEst, Nord, Nord-West et Soud-Ouest de lUganda. Cependant, la recherche sur cette culture a été très limitée et ainsi une enquête a été conduite dans lEst de lUganda (Kumi, Pallisa et Kamuli districts) pour établir la situation de la culture, les contraintes de production et les perspectives pour son développement. Un questionnaire semi-structuré a été administré sur 50 ménages par district, durant la période Octobre-Décembre 1998. La productivité de la terre a baissé considérablement dans les districts étudiés et la faible fertilité du sol constitute un facteur principal à cet égard. La production de cette culture est handicapée par plusieurs contraintes, les principales étant le travail unadéquat du sarchage et de la recolte, fréquente sécheresse, pestes et maladies et lépuisement du sol. Le région est hautement dépendante de lox pour la préparation de la terre. Le activités de production de léleusine sont générallement de sexe balancé. La culture est faite presque exclusivement une fois lannée. Lassociation et la rotation des cultures sont pratiques communes dans la production de leleusine et son importance en tant que culture de revenu saccroit. Les perspectives de développement se trovent dans laugmentation des rendements par la génération des technologies qui allègent le travail, en particulier pour le sarclage, la recolte et le développement des paquets sains pour les agriculteurs dans la gestion de pestes et maladies, la mitigation de la sécheresse et la gestion de la fertilité de sol. Un suivi dessai en milieu réel de gestion de la fertilité du sol a été conduit pour étudier la réponse déleusine semée en lignes ou à la volée dans differents régimes de gestion de la fertilité du sol. La plantation en lignes a favorisé significativement (P<0.05) une meilleure croissance déleusine et du rendement plus que la plantation à la volée. Lapplication combinée du P et dN ou du P et de la fumure a donné des rendements plus élevés. Lapplication dN seul a retardé la floraison et la maturité physiologique de la culture dune et deux semaines respectivement. Il ny avait pas deffect significatif (P<0.05) de mode de plantation ou du traitement dengrais sur le pourcentage de battage.
Mots Clés: Sécheresse, eleusine coracana, sexe, travail, fertilité du sol, manure, mauvaises herbes
Finger millet (Eleusine coracana L.) is a major staple crop in Uganda and is rated second to maize (Zea mays L.) in importance among the cereals (Esele, 1995; Wasswa and Odelle, 1995). According to the National Agricultural Research Organisation, finger millet is a high priority food commodity research crop, ranking second only after bananas (NARO, 1991; Kiwuwa and Nabasirye, 1997). The crop is grown on an estimated annual area of 420,000 ha which provides grain harvests of up to 643,000 metric tonnes (FAO, 1995; Wasswa and Odelle, 1995). Production of the crop is largely in the northern, eastern and western regions of the country.
Besides its importance as a staple food crop, finger millet contributes greatly to the incomes of rural households, particularly women. It is brewed into local beer for sale or is sold directly as grain in local markets where there is ready demand. Furthermore, finger millet plays a major role in providing for the dietary needs of the rural people who constitute more than 80% of the Ugandan population. It is a major preventative food against malnutrition, owing to its high content of essential amino acids, namely, tryptophan, cystine, methionine, and total aromatic amino acids (phenylamine and trysone). Finger millet is a fairly resilient crop; it is drought tolerant and its grain has an extended shelf life of several years without significant damage by storage pests. Finger millet, therefore, offers great food security opportunities for the country.
Despite the great value associated with this crop by the population, its productivity (yield per unit area) has remained low. Production figures consistently show an increase in area under finger millet production, but paralleled by a decline in land productivity (Statistics Department, Uganda Ministry of Agriculture, 1995, unpublished report). Several causes have been advanced to explain this decline and the major ones are thought to be low soil fertility and weed proliferation (Koma-Alimu, 1985). These inferences are, however, largely drawn from data of more than twenty years ago. For any meaningful management intervention to be executed, there is need for a thorough update of the resource base as well as the management structure of the crop within the existing farming systems. There is also a need for quantitative re-assessment of the constraints profile on the ground, vis-a-vis the recent changes in economic focus of the country as well as the global climatic and general environmental changes.
A study was, therefore, undertaken in eastern Uganda to assess the status of finger millet at household level and its production constraints and prospects, with a view to establishing a framework for meaningful research and development of the crop.
Socioeconomic and biophysical survey. A rapid rural appraisal was carried out in the districts of Kumi, Pallisa and Kamuli in Eastern Uganda during October-December, 1998. One major finger millet growing sub-county was selected from each district for the study, namely, Atutur in Kumi, Putiputi in Pallisa and Namwiwa in Kamuli. Sub-county selection was done with the help of grass-root extension agents and local admini-strators. Selection was done purposively using farmer lists.
A semi-structured questionnaire was used backed by a checklist to permit eventual statistical analysis of the data. The questionnaire was pre-tested to ensure appropriateness for the purpose. The major aspects considered in the questionnaire included socio-economic and biophysical issues at household level, in relation to finger millet production, as well as other enterprises on the farm. Farm resource base and flow among enterprises, labour structure and adequacy, general management related issues, post-harvest, marketing, gender roles and production constraints were also among the key questions asked during the survey. Respondents were asked to rank constraints affecting the productivity of their farms.
In each subcountry, fifty households selected purposively using farmer-lists provided by local extension agents constituted the study units. Respondents consisted of men and women. As much as possible, the interview session was kept informal to ensure a relaxed atmosphere for the interviewees. For each household, at the end of the session, interviewers were conducted around the farm setting to observe directly the nature of finger millet-grown fields. Soil samples were taken from ten farmers per sub-county and analysed for soil fertility parameters using procedures outlined by Okalebo et al. (1993). Coded data were entered and processed using the Statistical Package for Social Scientists (SPSS).
Soil fertility management study. A researcher-designed and farmer-researcher managed experiment was conducted during the period of March to July 1999 in the districts of Kumi (Atutur subcounty), Pallisa (Putiputi subcounty) and Kamuli (Namwiwa subcounty) in Eastern Uganda, on six farmers fields. Eastern Uganda receives an average annual rainfall of 1000 mm, and generally has soils low in N, P, but relatively adequate amounts of exchangeable bases (Table 1). The typical farming system of the region is the cotton-finger millet system (Yost and Eswaran, 1990), with cowpeas (Vigna unguiculata L. Walp), maize (Zea mays L.), cassava (Manihot esculenta Crantz), beans (Phaseolus vulgaris) and sweet potato (Ipomea batatus) as rotational or mixture crops.
Table 1. Soil fertility status (ranges) for Kumi, Pallisa and
Kamuli districts (Soil depth = 0 - 15 cm)
Minimum requirement: pH=5.2; OM=3%; Total N=0.2%; Bray 1P=15 mg kg-1;
Exch. Ca2+=4.0 cmol kg-1; Mg2+=0.5 cmol kg-1;
K+=0.2 cmol kg-1 (Source: Okalebo et al., 1993)
Treatments included a control (0 N, P or manure kg ha-1); 45 kg N ha-1; 17 kg P (or 40 kg P2O5) ha-1; 22.5 kg N + 8.5 kg P ha-1, and 5 t kraal manure ha-1. Urea (46% N) and TSP (19.5% P2O5) were the inorganic fertilisers used. The other treatments included row planting at a spacing of 30 cm by 6 cm, and broadcasting according to the farmers own practice. The experiment was laid out in a split plot design with row or broadcast planting as the main plots and fertiliser treatments as sub-plots. An improved finger millet variety, PESE I, was used as the test crop. Finger millet was row- planted using a tin with a hole perforated at the base to allow uniform dropping of seed in lines. This technique was introduced by the extension staff to ease on the tedious direct hand-sowing. Broadcasting was done by the farmers themselves as they normally do in their fields. Fertiliser treatments were hand-applied and hoed in at planting. The treatments were replicated thrice, with each farmer constituting a replicate. The main plots were 5 m by 10 m, while subplots were 5 m by 5 m.
Finger millet growth was monitored by measuring plant height at 4 and 8 weeks after crop emergence, at 50% flowering and at physiological maturity (14 weeks). At physiological maturity, plant population and total above ground biomass were determined from each plot. Thereafter, sub-samples from each plot were oven-dried at 65 0C for 48 hours, for dry matter and grain yield determination. Finger millet heads were oven-dried and threshed using a mortar and a winnow to determine threshing index. The data were analysed using Genstat statistical package; mean separation was done using Fishers Least Significant Difference (LSD) test.
Results and discussion
Socioeconomic and biophysical survey. Resource base and access are among the factors considered in assessing agricultural potential and sustainability levels of cropping systems on a farm. Table 2 presents data of household land size, finger millet hectarage and yield; as well as livestock numbers per household in each study district. Household land size was marginally greater in Kumi and Pallisa than in Kamuli district. Finger millet hectarage, however, was smallest in Kamuli district. Land productivity (yield per unit area), on the other hand, was much greater in Kamuli and Pallisa districts than in Kumi district. The average number of animals per household was greatest in Kumi district, with Kamuli district having less than half of the former.
Table 2. Household resource base data for Kumi, Pallisa and Kamuli
districts as of 1998
Animals include cattle, goat, sheep and poultry. Sample size: 50 households per district
Generally, average household land size was not a factor limiting crop production, compared to other highly populated districts in Uganda such as Kabale, where household land size ranges from 0.2-2 ha (Lindblade et al., 1998). The hectarage occupied by finger millet ranged from 12% (Kamuli) to 24% (Kumi) of land owned by each household. Over all, up to 48% of the land under crops was occupied by finger millet, an indication of the significance of the crop at household level in the eastern region.
Despite the large hectarage devoted to finger millet production in Kumi district, grain yield per unit area was rather low compared to Pallisa and Kamuli districts. Moreover, it is barely 25% of the research station average (2,500 kg ha-1, Wasswa and Odelle, 1995). Yields in Pallisa and Kamuli districts were also about 30% of station yield. These figures attest to earlier reports of low and declining grain yields in eastern Uganda (Esele, 1995; Wasswa and Odelle, 1995; Department of Statistics, Uganda Ministry of Agriculture, Animal Industry and Fisheries, unpublished report). The study farmers also echoed the serious decline in finger millet yields.
Farm animals provide a reflection on farmer access to manure, draught power, alternative sources of high quality protein and income from non-crop products. Kumi district had the greatest number of farm animals (Table 2) and was the only district in which manure was applied in fields deliberately to address the low fertility problem. In fact, there is overwhelming belief by Kumi farmers that crop performance is invariably poor without manure use. Some farmers rotate kraals on their farms along with cropping fields as a means of soil fertility management. The source of manure was exclusively the cattle herds. Use of manure from other animals was not reported, neither was buying manure from other farms practiced. The amount of manure applied per unit area was not known, but was dependent on the quantities provided by the animals. In contrast, none of the respondents in Pallisa and Kamuli indicated use of manure as an input for crop production. In fact, manure was either set ablaze close to the kraals or dumped away as garbage. All these could be attributed to the fact that Pallisa and Kamuli soils were relatively more fertile (Table 1) and productive (Table 2) than Kumi soil. Nevertheless, there was unawareness of the value of manure as a farm input both in Pallisa and Kamuli districts; extension efforts are definitely needed in this direction.
All respondents had access to draught power for primary and secondary land preparation, through ownership or hire (Fig. 1). In Kumi district, majority of the farmers (80%) owned at least a pair of oxen and an ox-plough. Pallisa district too had a similar scenario, though less in magnitude (60% respondents). In contrast, Kamuli had the fewest households (25% respondents) with such tillage facilities and majority of the farmers (75% respondents) depended on hiring them from other farmers.
Weed management is, by far, the most important pre-occupation in finger millet production in the study region. This is even more serious in Kumi (100% respondents) and Pallisa (50% respo-ndents), where weeding is done twice per season (Fig. 2). A variety of weeds, mainly grasses (Eleusine indica, Digitaria spp., Cynodon dactylon, Cyperus spp., Panicum maximum) were reported to be the most obnoxious in finger millet fields because of their resemblance to the crop, which makes weeding, particularly in broadcast millet, cumbersome. Earlier reports also alluded to this weed problem (Koma-Alimu, 1985). Major broad-leaf weeds included Euphorbia spp., Commelina spp., and Bidens pilosa. Weed species composition was similar in the three study districts. The only exception was the relatively more recent weed in the region, Striga spp., especially in Kumi and Pallisa districts. It is recognised as a highly debilitating weed on cereals. Farmers lacked experience with Striga management, yet they associated its proliferation with soil fertility decline. This is consistent with research findings on the weed done elsewhere (Ayensu et al., 1984; Packer, 1991). Striga parastises the crop mainly through root physical and physiological interactions, and sucks out resources captured or manufactured by the crop. Consequently, the crop becomes stunted and yields poorly (Ransom et al., 1995).
Weed control is exclusively manual, assisted by small short-handled hoes. The efficiency of this method is very low, especially in broadcast finger millet . Posture while weeding is extremely strenuous; some farmers weed while bending and others squatting. Family labour is invariably inadequate and every household interviewed hired labour for weeding. Owing to the field sizes being unmatched by the required labour force, timely weeding is often unattainable. It is, therefore, likely that the crop is severely stressed by weed competition resulting in low yields. There is need for research to assess the impact of weeds and weed management in general on finger millet yield losses in this region.
The structure of the labour force used in weeding finger millet is presented in Table 3. Weeding is done by all gender groups including children. The labour force, however, comprises largely of adults above 18 years (75% of the labour force). It was generally expressed that the introduction of Universal Primary Education (UPE) Programme by the Uganda government, which offers free education to children, had reduced child involvement in weeding and other farm activities. Gender was not a factor in weed management since all gender groups were equally represented. Generally, hired labour predominated family labour and more men than women were hired for this purpose.
Table 3. Seasonal labour structure and availability for weeding,
by gender and age in Kumi, Pallisa and Kamuli districts
Finger millet is generally grown once a year in the long rains of March to July (Fig. 3). There were a few cases of farmers who grow the crop twice a year (March-July and September-November). The single season growers reasoned that the crop was labour intensive and two crops a year would be a burden to the family. Additionally, the second season is characterised by unreliable rainfall in the second season. Over all, yields in the second season were reported to be too meagre to warrant family re-investment.
Crop rotation is an integral practice in finger millet production in eastern Uganda (Table 4). Various crop types are used as rotation crops, including legumes. The sequence of rotation does not follow a particular crop pattern; it is entirely dependent on the farmers decision. Reasons for crop rotation were given as (i) breaking pest and disease cycles, and (ii) minimising decline in soil productivity. Furthermore, rotation with cotton was widely believed to maintain the field clean for the next finger millet crop, since cotton is weeded thrice during a cropping season. Knowledge of the value of legumes in soil fertility improvement was not evident among the farmers.
Table 4. Major rotation crops in finger millet cropping systems
in Kumi, Pallisa and Kamuli districts
Finger millet grain storage is a corner-stone in ensuring extended quality food shelf-life for households. Grain storage in Pallisa is dominantly (72% respondents) in granaries constructed from locally available materials. Only 50% of the respondents in Kumi used granaries as storage facilities, the rest stored it in their houses in gunny or polythene bags. In Kamuli, too, the majority (65%) kept finger millet grain in bags within their houses. Granaries were reported to be the traditional storage facilities in the study districts, but a general rise in theft among communities has forced farmers to keep their produce in more secure places.
Finger millet was rated highly both as a food and cash crop in Kamuli (82% respondents) and modestly in Pallisa district (56% respondents) (Table 5). The crop is increasingly becoming a cash crop, in part due to the shift in national economic policy, which entails transformation of food crops into cash crops. Additionally, the resource base that usually provides farmers with alternative income generating crops is on the decline. For instance, the frequency of drought conditions is on the increase, to the extent that only drought tolerant crops such as finger millet can be depended upon for both nutrition and income. Finger millet is sold as local beer or directly as grain. The impact of transformation of the crop into a cash crop status, particularly in Kamuli district, needs to be evaluated since this could indirectly induce food insecurity within rural households.
Table 5. Uses of finger millet in Kumi, Pallisa and Kamuli districts
Sample size: 50 households per district; NA=data not available
The roles of men and women in finger millet production activities in the study districts are presented in Table 6. Generally, bush clearing is done by men in the three districts. Seed sowing is done mostly by men in Kumi, but by women in Kamuli. In contrast, both gender groups participated equally in sowing in Pallisa district. Weed management and harvesting are done by both sexes. However, while both sexes are involved, this is largely done by men in Kamuli and Pallisa.
Table 6. Gender profile in finger millet production activities
in Kumi, Pallisa and Kamuli districts
Sample size: 50 households per district; F=female; M=male
Farmer ranking of finger millet production constraints. The major finger millet production constraints as identified and ranked by farmers in Kumi, Pallisa and Kamuli are shown in Figure 4. The major constraints included pests and diseases, persistent drought, exhausted soils, shortage of labour for harvesting and weeding and high weed infestation. The rankings differed among the three districts, but overall, weeding and harvesting were considered the most important. This was followed by drought across the districts. Soil exhaustion was ranked the least serious problem in Kumi and Pallisa districts. It was, however, apparent that farmers were not familiar with soil related issues. Nevertheless, it was reported in Kumi that manure application was becoming mandatory for viable crop production. This is consistent with our soil analysis results (Table 1) which show that N, available P and some exchangeable bases were extremely sub-optimal for most crops based on the scale outlined by Okalebo et al. (1993). It also conforms to the results of Madiagne et al. (1999) in Senegal where application of manure generally resulted in higher pearl millet and peanut yields.
Prospects of finger millet production in eastern Uganda. Despite the gravity of the production constraints expressed by finger millet farmers (Fig. 4), households remain committed to production of the crop due to the socio-economic importance attached to it. Farmers are eager for any form of assistance that would alleviate finger millet production constraints. Technologies that are labour saving, particularly during weeding and harvesting, are urgently needed. These could include transformation of weeding operations from hand-hoe to simple mechanisation, especially where draught power is involved. This will definitely require that the planting pattern is changed from broadcast to row-planting, which has been shown to save as much as 80% labour requirement for finger millet weeding.
To address the problem of drought, development of early maturing finger millet varieties would be most appropriate. With regard to soil fertility management, farmers were very receptive to ideas that augment their traditional soil improvement approaches. In Kumi district, some farmers maintained their kraals in rotation with cropping fields and in the process input manure into fields without incurring labour expenses for application. Furthermore, farmers in the three districts practice crop rotation with leguminous crops (cowpea, Phaseolus beans, green grams and groundnuts), which contribute towards soil fertility improvement through N fixation. There is certainly a need to develop a farmer-friendly soil improvement package for finger millet in the region to prevent the extinction of this community-valued crop.
Soil fertility management study
Soil properties prior to the study. Data for soil properties of the study districts (sub-counties) are presented in Table 1. Each value is an average over 3 farms per sub-county. Based on Okalebo et al. (1993) recommendations, soil pH is generally favourable for most East African crops. Total (Kjeldahl) N, soil organic matter (SOM) and available P are extremely low. Exchangeable bases, on the other hand, are relatively adequate. This might be the reason why soil pH values are favourable for most crop production. These results confirm the farmers inferences that low soil fertility is an impediment to crop production and that cow manure gives positive yields. The extremely low SOM content has great implications on the sustainable productivity of the regions soils, not only in terms of nutrient supply, but also moisture retention and availability to crops. The soils of the region are very light with texture ranging from sandy-loam to sandy clay-loam. In these soils, therefore, organic matter fraction is crucial both as a reservoir of nutrients and as a conservator of moisture and, in turn, a protector of crops against instantaneous droughts. The organic fraction also helps to retain the still fair amount of exchangeable bases present in the soil, by providing extra cation exchange capacity (CEC). Hence, for better and sustainable productivity of these soils, it is important that all available forms of organic inputs are applied to the soil.
Nutrient content in kraal manure used. Kraal manure consisted of 1.7, 0.1and 0.5% N, P and K respectively. Similar low quality manure has been reported in the communal areas of Zimbabwe (Mugwira and Mukurumbira, 1984; Tanner and Murwira, 1984). The low content of nutrients in the kraal manure implies that a substantial amount will be required to make a fertility impact on the light and nutrient depleted soils of this region. Nevertheless, the non-nutrient attributes of manure, e.g., soil structure improvement and water-holding capacity could be greatly enhanced by application of even small doses.
Effect of planting pattern on finger millet plant height at various physiological stages. Finger millet plant heights measured at three physiological stages of growth are shown in Table 7. Plant height remained uninfluenced by planting pattern either in rows or broadcast, up to flowering (8 WAE). However, by physiological maturity, row planting phenomenally boosted plant height by tenfold over broadcasting. This observation has great implications, especially in the study soils where considerable amounts of organic matter are direly needed. Increased biomass yield could be recycled to build up organic matter in the soil. Plant height is also important in cases where this growth attribute is directly or indirectly related to grain yield such as in this particular case (Table 7). Row planting consistently out-yielded broadcasting (P<0.05) for all plant population densities. This indicates that planting pattern influenced both growth and yield of the crop.
Table 7. Effect of planting pattern on finger millet plant height
The cause of the superior performance of the row planted crop over its broadcast counterpart could be due to the nearly optimal utilisation of resources by the latter than the former. Row planting was done following judicious research recommendations for the study finger millet variety, PESE 1. On the other hand, broadcasting was done following the traditional farmers practice. It was clear in the field that the row planted crop was spatially uniformly distributed (at about 6 cm x 30 cm). In contrast, the broadcast crop was spatially non-uniform in the field and was characterised by over-crowding in some parts and sparseness in others. This distribution pattern could have adversely affected resource utilisation by plants. Overcrowding could have caused undue competition for resources (nutrients, water, light, etc.), while sparseness could have led to resource wastage.
Effect of planting pattern and fertiliser application on finger millet attainment of flowering. Application of N as urea at a rate of 40 kg ha-1 clearly delayed flowering by nearly 10 days for both row- and broadcast-planted crops (Fig. 5). Application of P alone, manure alone, manure+P or P+N had no significant effect on time to flowering. The delaying effect of N also extended the physiological maturation of the crop by up to two weeks in the three study districts. This N effect is a widely known phenomenon, especially when N is applied in excessive amounts. Excess of N not only stimulates vegetative growth and delays senescence, but also changes plant morphology, particularly if N availability is high in the rooting medium during early growth of the crop (Klemm, 1966; Marschner, 1995). However, its occurrence in areas where doughtiness is becoming lengthy and unpredictable such as in the eastern region of Uganda, is a cause for concern. Although such N-fed plants looked greener and more vigorous than the rest treated differently, delayed maturation not only exposes the crop to a risk of drought stress, but also increases the farmers labour needs, especially for control of vermin particularly birds.
Effect of planting pattern and fertiliser application on finger millet dry matter and grain yield. Data for dry matter and grain yield as influenced by planting pattern and fertiliser application, are presented in Figures 6 and 7, respectively. Under both planting patterns, fertiliser application increased dry matter yield significantly (P<0.05) only in a few cases. Evidently, only the combined application of manure and P or N and P increased dry matter yield significantly (P<0.05) under both row and broadcast-planting pattern. Under row planting, however, application of P alone, P+manure or P+N significantly (P<0.05) increased dry matter yields. The row-planted crop generally responded better to nutritional management than the broadcast one. Grain yield responded similarly, with P alone, P+manure or P+N significantly (P<0.05) increasing yields under both planting patterns. Similarly, under row planting with fertiliser treatments were generally superior to their counter parts under broadcasting. Application of N or manure alone had no significant effects on both yield components and for the two planting patterns.
The better yields under fertilisation, irrespective of planting pattern, confirm that these soils have sub-optimal fertility for crop production. The lack of significant dry matter or grain yield increases under the N alone treatment implies that N utilisation efficiency was poor. In effect, this implies that N application alone is wasteful and should be avoided. In the case of P treated plots, P is a key limiting nutrient in these soils and must be applied for improved finger millet productivity. The superior performance of manure+P overall suggests that the soil needs additions of other nutrients apart from P. These results support previous work by Yadvinder-Singh et al. (1988) and Bationo et al. (1993) which showed that mixing inorganic P or rock phosphate with organic fertilisers can increased P availability and uptake by plants. Similar responses have been obtained by Coaldrake and Pearson (1985) and Paynet et al. (1995). It is thought that phosphorus availability is enhanced in the presence of decomposing organic residues (Yadvinder-Singh et al., 1988), which cause formation of organic acids and other compounds that reduce P sorption. Our study is also consistent with the findings of Madiagne et al. (1999) who found that animal manure/compost together with modest amount of mineral fertiliser, maximised yields of pearl millet in semi-arid region of Senegal and Niger (Maman et al., 2000 a, b).
On the other hand, the poor performance of manure alone could be attributed to the fact that manure, like all other organic materials, are naturally poor in P content (J.Y.K. Zake, 1999, Department of Soil, Science, Makerere University, Uganda pers. comm.). The manure used in this study contained barely 0.1% P and at the highest rate of 5 t dry manure ha-1, the material potentially supplied a maximum of 5 kg of P ha-1. This is extremely low considering that the crop needs at least 40-50 kg P ha-1. However, the quality of manure used in this study, though typical of the communal manures in Africa which are handled poorly (Murwira and Kirchmann, 1993), is very low compared to manure collected under intensive systems such as zero-grazing. Zero-grazing manure has been reported to contain 2.25, 0.3 and 2.0% N, P and K, respectively (Nzuma et al., 1997). Under zero-grazing, both solid and liquid manure fractions are preserved. This is not true under the traditional free-range kraal system.
The equally good performance of N + P as that of manure + P, and the relatively better performance of the former than that of P alone implies that N equally limits plant growth and yields in the regions soils. However, both fertiliser combinations suggest that manure was adequately endowed with N to the extent that supplementary application of this nutrient was not necessary to achieve the highest yield levels in the study. Hence, both N and P are confirmed limiting nutrients to finger millet production in the soils of eastern Uganda. Their combined application under row planting can raise grain yields to three-fold that of the present farmer no-fertility management practice. Even under the traditional broadcast-planting pattern, grain yield can be increased by up to 140% as a result of N and P or manure and P application. It should be noted that row-planting pattern clearly out-yielded broadcast planting by about 13% for the best performing fertiliser treatments. There was no significant (P<0.05) effect of planting pattern or fertiliser treatment on threshing percentage of finger millet, which ranged from 76 to 85%.
In conclusion, soil fertility is very low in the soils of eastern Uganda. This is, especially so in terms of N, P and organic matter. The soils are highly sandy and need a substantial amount of organic matter to enhance their productivity. Application of N alone as urea at a rate of 45 kg N ha-1 delays finger millet maturation by up to 2 weeks. This is risky in this drought prone region. Also, application of N alone has no dry matter or grain yield advantage over the no N-fertiliser application. Application of P2O5 (40 kg ha-1), P2O5 (20 kg ha-1) + manure or N (22.5 kg ha-1) + P2O5 (20 kg ha-1) result in substantial grain yield increases. The best performing treatments, however, are the latter two with yields up to 2.8 t ha-1, which is greater than the 2.5 t ha-1 on-station yield earlier reported as the highest for Uganda (Wasswa and Odelle, 1995). Row-planting performs better than broadcasting with or without fertiliser application. Kraal manure available to farmers in eastern Uganda is generally of low quality. It contains a reasonable amount of N but is very inadequate in P.
Further studies on nutrient mineralisation, and crop nutrient uptake during the growing season are required to predict the soil biological processes that regulate soil fertility and examine the nutrient use efficiency resulting from combination of organic and inorganic fertilisers. There is also need to develop manure management options that will minimise nutrient losses from kraal manure and enhance manure quality and, hence, increase crop productivity.
Authors acknowledge the financial support provided by The Rockefeller Foundations Forum on Agricultural Resource Husbandry, towards this study. The district agricultural officers of Kumi, Pallisa and Kamuli are also thanked for their cooperation.
©1999, African Crop Science Society
The following images related to this document are available:
Photo images[cs99048d.jpg] [cs99048b.jpg] [cs99048a.jpg] [cs99048g.jpg] [cs99048c.jpg] [cs99048f.jpg] [cs99048e.jpg]