African Crop Science Journal African Crop Science Society ISSN: 1021-9730 EISSN: 2072-6589 Vol. 7, Num. 4, 1999, pp. 591-598

African Crop Science Journal, Vol. 7. No. 4,  pp. 591-598, 1999                                                        

Determinants and Impact of Integration of Forage Legumes in Crop/Livestock Systems in Peri-urban areas of Central Uganda

T. K. Mugisa, P. K. Ngategize¹  and  E. N. Sabiiti²
Department of Agricultural Economics, Makerere University, P. O. Box 7062, Kampala, Uganda
¹National Agricultural Research Organisation, P. O. Box  295, Entebbe, Uganda
²Department of Crop Science, Makerere University, P.O. Box 7062, Kampala, Uganda

Code Number CS99050

ABSTRACT

Majority of the intensive smallholder crop/livestock systems in peri-urban areas of central Uganda are characterised by low productivity.  This is probably due to several factors such as poor management and inadequate feeds in terms of quantity and quality.  A study was therefore undertaken to ascertain determinants and impact of integration of forage legumes on productivity of the systems. Data were gathered using an interview schedule with 90 smallholder milk producers.  An econometric model was then used to quantitatively evaluate socio-economic factors impacting on the integration of forage legumes. Findings show that integration is more likely to be practised by farmers who have less farmland and/or are close to milk and inputs markets. Farmers who integrate legumes into elephant grass (Pennisetum purpureum) obtain more herbage that remains greener into the dry season, due to the ability of legumes to fix nitrogen in the soil, compared to non integrators. They spend less money on artificial insemination services and their animals have lower incidence of disease.

Key Words: Eastern Africa, logistic regression model, livestock feed resources, milk producers, peri-urban agriculture

RÉSUMÉ

La majorité des systèmes interisifs des cultures/bétail des petits exploitants dans les zones urbaines périphériques de l’Uganda centrale sont caracterisés par une faible productivité. Ceci est probablement dé aux multiples facteurs tels que la pauvre gestion et l’alimentation inadéquate en termes de quantité et de qualité.  Ainsi une étude a été entreprise pour établir les déterminants et l’impact d’intégration des légumeuses fourragères sur la productivité des systèmes.  Les données ont été assemblées par un interview programme avec 90 petits exploitants producteurs du lait.  Un modèle économétrique a été utilisé pour évaluer quantitativement les facteurs socio-économiques ayant impact sur l’intégration des légumes fourragères.  Les découvertes montrent que l’intégration est  probalement à être pratiquée par les agriculteurs ayant des petites parcelles et/ou sont proches du lait et des marchés d’intrants.  Les agriculteurs qui intègrent les légumuneuses dans l’herbe d’éléphant (Pennisetum purpureum) obtiennent plus d’herbage qui reste vert pendant la saison sèche suite à l’aptitude des légumineuses de fixer l’azote dans le sol en comparaison de ceux qui ne pratiquent pas l’intégration.  Ils font moins de dépenses sur les services d’insémination artificielle et leurs animaux ont la faible incidence de maladie.

Mots Clés: Afrique de l’Est, modèle de regression logistique, sources d’aliment du bétail, producteurs de lait, agriculture des zones urbaines périphériques

INTRODUCTION

In order to improve household nutrition and income, Uganda government and some Non-governmental organisations (NGOs) have established heifer projects that target the poor, especially women farmers. Most project beneficiaries confine their animals due to shortage of land. Recent studies have found zero grazing system to be the most profitable dairy production system in peri-urban areas (Staal and Shapiro, 1996; Tumutegyereize et al., 1998). Presently, however, poor animal nutrition in general, and particularly during the dry season, is a major factor limiting increase and sustainable milk production (Mpairwe, 1998). The average milk production is about 10 litres per cow per day compared to 25 litres per cow per day under improved management and adequate nutrition (Nsubuga,1993). The majority of the farmers rely on elephant grass (Pennisetum purpureum) as the basal diet (Mpairwe et al., 1998) but less than 30% integrate it with forage legumes (Muwanga, 1994). This is a major problem among milk producers because elephant grass alone is deficient in nutrients that are required to sustain high milk production (Boonman, 1993). Its quantity is further limited by the fact that the farmers own between 0.5 to 5 ha and use the same land to grow food crops such as maize (Zea mays L.), beans (Phaseolus vulgaris) and a variety of vegetables. Hence, there is declining productivity of elephant grass in terms of fodder with repeated defoliation. 

On average, elephant grass can deprive one hectare of land some 150 kg N, 75 kg P₂O₅ and 450 kg K₂O annually (Boonman, 1993). This alarming depletion of soil can be restored by nitrogen fertiliser (Boonman, 1993) and/or legumes (Moog, 1991; Sabiiti, 1993; Muhr et al., 1997).  The nutritive value of pasture grasses and crop residues can be improved by incorporation of herbaceous forage legumes in rotation, intercropping or undersowing (Dzowela, 1986; Nsubuga, 1993; Mpairwe, 1998).  Furthermore, farmer’s circumstances directly influence the acceptance and adoption of new technologies. It is not easy to introduce technological innovations in livestock production, at the level of the smallholder producers, without adequate knowledge of the socio-economic characteristics of the target communities (Preston, 1986).

The objective of the study reported in this paper was to ascertain the determinants of integration of forage legumes into peri-urban crop/livestock systems and their impact on productivity.

MATERIALS AND METHODS

Three peri-urban areas of Entebbe, Mukono and Kampala were selected for the study because they have the highest number of zero grazing farmers. The areas are located in Central Uganda astride the equator in the Lake Victoria Crescent. The Lake Victoria Crescent, a robusta coffee/banana zone, is 14,797 km₂ and 1,174 metres above sea level. It experiences moderate temperatures slightly above 200 ºC and an annual bimodal rainfall­of 1200 mm (Wortmann and Eledu, 1999).

The study was restricted to smallholder milk producers who had planted elephant grass (Pennisetum purpureum). They practice zero grazing or cut and carry system, where between 1-9 milking cows are continuously confined on 0.5-5 hectares of farmland.  Data were gathered from 90 smallholder milk producers during May and June, 1999. A quick reconnaissance survey was carried out to identify potential milk colonies (concentration of producers) in the three target peri-urban areas. A sample frame was then obtained  using a list of all zero-grazing farmers in the target milk colonies, provided by district veterinary officers. A stratified sampling method was used to obtain two mutually exclusive groups of zero-grazing farmers: those who integrate (users) forage legumes (Desmodium intortum, D. uncinatum, Centrosema pubescens and Macroptilium atropurpureum) into elephant grass and those who do not (non-users). Subsequently, a semi-structured questionnaire was used for documenting qualitative and descriptive data.  It was administered through face-to-face interviews supplemented by on-site observations. 

Data processing and analysis. The logistic regression model of qualitative choice was used to determine various socio-economic factors that have a significant relationship with integration of forage legumes into crop/livestock systems. The independent variables studied included: credit, gender, education, land and distance of farm from market. In addition, any relationship existing between the independent and the dependant variables were established using the chi-square.  

Soil samples were analysed after a 6 months season using the kjeldahl method (Landon, 1991) to find out if legumes contributed nitrogen to soil. The ANOVA and cross-tabulation were used to analyse number of services per conception and incidence of diseases respectively, using the Statistical Package for the Social Sciences, v. 8.0 (SPSS,1994).

Theoretical model. The logistic regression analysis model was adopted for this study mainly because errors in the variables that may bias the estimate of the parameters are reduced by logistic regression. Results can easily be interpreted and indicate how the probability of integration of forage legumes into crop/livestock systems is related to the independent variables such as gender, education, age, etc. These opinion variables were grouped dichotomously into ‘no’ and ‘yes’ responses. The model can be written (Pindyck and Rubinfeld, 1991; Gujarati, 1995) as;

Prob. (event), P_i = 1/ 1+ e^-z ...................................................................(1)

Where:

 P_i   = the probability that an event will occur.

    e = is the base of the natural logarithm,                   (approximately 2.718).

Z =      the linear combination or relationship of the socio-economic factors (x_i), i.e.,

Z = B_o + B₁ X₁ + B₂ X₂ + .. + B _n X _n + U₁ .......................................................(2)

Where:

 B_o .... B _n = are coefficients to be estimated from the data.

X₁ ... X n = are the explanatory or independent variables.

 U₁          =  is the error term.

The probability of the event not occurring was estimated as: Prob. (no event) = 1 – Prob (event). Equation (1) represents the (cumulative) logistic distribution function. Parameters of the model were estimated using the maximum likelihood method, i.e., the coefficients that make the observed results most likely were selected. Since the logistic regression model is non-linear, an iterative algorithm was necessary for parameter estimation (Hosmer and Lemeshow, 1989; Gujarati, 1995) as;    

 Ln P - 1 = ñ_i = eß_o + ß₁X₁ + .... + ß_nX_n.......... (3)  

     1 - P_i

Empirical model. Taking Ir (equation 4) to represent the practice of integration as observed on respondent’s farms and using equation 3, the following model was fitted into the data and regressed to determine the coefficients in the logistic regression model.

I r = b_o + b₁ AC + b₂ Ag + b ₃ Dm + b₄ Fs + b₅ Led + b₆ Pf + b₇ Pm+ b₈ Sx +u ............................... (4)

Where:

      AC = Access to production credit,

      Ag  = Age of the farmer,

      Dm = Distance from farm to the market,

      Fs  = Farm size,

      Led = Level of education,

      Pf  = Price of forage seed in the market,

      Pm = Price of milk,

      Sx  = Sex,

      b₀ , b₁ .... b₈ = are coefficients, and

      u = Error term.

RESULTS AND DISCUSSION

Site soil analysis. Table 1 shows total nitrogen added to soil after a 6 months season of integration of forage legumes into the elephant grass system. There was a rise in nitrogen in the soil over the control. Forage legumes improved soil fertility through their ability to fix nitrogen naturally, at both soil depths. Since nitrogen is the most limiting nutrient in Ugandan soils (Zake, 1993), the additional nitrogen resulting from integration  would benefit the elephant grass and any subsequent crop, both deep and shallow rooted.  Based on earlier studies, Sabiiti (1993) also concluded that farmers benefit by incorporation forage legumes in their production systems. Indeed, integration of legume is  a way to improve elephant grass productivity and of similar grasses (Boonman, 1993).      

TABLE 1.  Total amount of nitrogen contributed by legumes in peri-urban Kampala

 (+) = Soil samples from plots under forage legumes integrated into elephant grass.
 (- ) = Soil samples from control plots under elephant grass alone, without legumes.

Factors of forage legume integration and  impact of socio-demographic characteristics on integration. Table 2 indicates that the logistic regression model correctly predicted 32 respondents who integrate forage legumes into their elephant grass system. Thirty-six respondents who do not integrate forage legumes into their elephant grass system were also correctly predicted. That is, the model correctly predicted 86% non-integrators.  Overall, the model correctly predicted 81.93 % of the respondents. The off-diagonal entries in Table 2 indicates that 15 respondents were incorrectly classified, i. e., 9 integrators and 6 non-integrators.

Table 2. Classification for the logistic regression output

U = Integrators;   n = Non-integrators

Table 3 shows the logistic regression output for the dependent variable integration of forage legumes.  The two most significant factors in the integration of forage legumes into elephant grass systems are the total land devoted to livestock and the distance of farm from the market for milk. The two coefficients are significant at 99 and 95%, respectively. The negative sign of both coefficients implies that they are negatively related to integration. That is, the more land a milk producer has devoted to livestock production, the less likely that he/she will integrate forage legumes into his/her elephant grass system.  Contrastingly, the closer a milk producer is to a milk market and farm inputs, the more likely that he/she will integrate forage legumes into the elephant grass systems.

TABLE 3. Logistic regression model for the dependent variable, integration of forage legumes

Exp (B) = odds ratio.  NS = not significant at 5%;  *  and  ** = significant (0.05) and highly  significant (0.001), respectively.  Values in parenthesis are standard errors.

The odds ratio for size of land devoted to livestock and distance of farm from market for milk are both less than one, implying that as these two factors increase, the probability of integration of forage legumes decreases. Three factors, namely, age, sex and formal education of respondents have partial correlation of 0.00. This implies lack of effect on the dependent variable.  However, this does not mean that the three factors are not important.  They are important in a sense that most non-government organisations (NGOs) select ­beneficiaries of heifers on the basis of these factors.                       .                 

Table 4 shows cost and number of artificial insemination services per conception in dairy cattle. Milk producers who integrate spend on the average less than half on artificial insemination compared to the milk producers who do not integrate forage legumes into their elephant grass systems. This finding may be explained by the fact that most milk producers who integrated fed their cows better quality feeds compared to the non-integrators. Past studies (FAO, 1982; Mukasa-Mugerwa, 1989) have established that feed deficiency is a main causal factor in lengthening the reproductive cycle in cattle. Dindorkar et al. (1982) reported that cows kept on a low plane diet neither cycled nor ovulated. In Randel’s (1990) study, inadequate protein intake during both the prepartum and postpartum periods resulted in a pregnancy rate of 32% in cows with low protein intake, compared with 74% in cows with higher protein intake.

TABLE 4.  Cost and number of artificial insemination services per conception

¹Mean cost of one artificial insemination service is USh. 12,000 /= (US $ 1 = USh. 1,400/= as of May,1999); USh. = Uganda shilling; N=Number of respondents (sample size=90)

Incidence of disease. Table 5 shows the incidence of dairy cattle diseases as reported by the farmers. The incidence of disease is much higher among non-integrators (non-legume users) than integrators (legume users). The number of farms reporting no incidence of disease is about five times among legume users compared to non-legume users. The most commonly reported diseases among non-integrators included reproductive tract diseases, mastitis, internal parasites and east coast fever. Winrock (1992) describes most of these diseases as both infectious and non-infectious.  Their prevalence and severity are greatly influenced by nutritional status of the animals, management practices and genotype.  

TABLE  5.  Incidence of dairy cattle diseases in Central Uganda

¹Total number of respondents = 90

Table 6 gives calculated values of the chi- squared distribution. The tabulated value at 0.1 % (13.8) is less than the calculated value, hence there is a significant relationship between integration and incidence of disease. A study on the extent to which integration affects specific infections and or diseases of cattle under zero grazing in peri-urban areas may be necessary in order to assist farmers to avoid potential depression in dairy production.

TABLE  6.  Chi-square test on integration and incidence of disease

Increasing the use of forage legumes can reduce some fertiliser and commercial feed costs, and enable peri-urban milk producers to improve on the productivity of their dairy cattle (Sabiiti, 1993). Higher animal productivity would be achieved through, among others, more milk production, lower incidence of disease and affordable artificial insemination costs. Block (1994) observed that increased farm productivity provides an incentive for further adoption of existing technologies.

CONCLUSION

This study has revealed that forage legume integrators live in close proximity to better markets for milk and inputs, and generally own small farmland. They spend less on artificial insemination services and have lower incidence of animal diseases than the non-integrators. Integration of forage legumes into elephant grass also increases the level of nitrogen in the soil. The nitrogen fixed by legumes may improve soil fertility and enhance herbage yield. Further integration of forage legumes into peri-urban crop/livestock systems may be enhanced by targeting dairy farmers who are more land constrained and close to milk markets. The influence of age, sex and formal education on integration of forage legumes into crop/livestock systems in Central Uganda’s peri-urban areas was not significant. 

ACKNOWLEDGEMENT

The Rockefeller Foundation funded the study through the Forum on Agricultural Resource Husbandry (Grant RF 96008 # 82).

REFERENCES

Boonman, J. G. 1993.  East Africa’s grasses and fodders: their ecology and husbandry. Kluwer Academic Publishers, Dordrecht, The Netherlands. 343 pp.

Block, S. A. 1994.  A new view of agricultural productivity in sub-Saharan Africa. American Journal of Agricultural Economics 76: 619 – 624.

Dindorkar, C.V., Muzaffar, S.J. and Kaikini, A.S.  1982.  Note on the effect of feeding extra energy diets to anoestrous cattle. Indian Journal of Animal Science 52:1228-1229.

Dzowela, B. H. 1986.  Value of a forage legume component in summer beef fattening systems in Malawi. In: Potentials of forage legumes in farming systems of Sub-Saharan Africa. Proceedings, ILCA workshop.  Haque, I., Jutzi, S. and Neate, P. J. H. (Eds.), pp. 540-546.Addis Ababa, Ethiopia, 16 – 19 Sept. 1985. 

Food and Agricultural Organisation of the United Nations (FAO), 1982. Reproductive efficiency in cattle: A guideline for projects in developing countries.  Vandeplassche, M. (Ed.). FAO Animal Production and Health Paper No. 25. Rome, Italy.118 pp.

Gujarati, D. N. 1995.  Basic Econometrics. 3rd edition. McGraw-Hill Inc., New York, USA. 838 pp.

Hosmer, W. D. and Lemeshow, S. 1989.  Applied Logistic Regression.  John Wiley & Sons, New York. 245 pp.

Moog, F. A.  1991.  Forage and legumes as protein supplements for pasture based systems. In: Feeding dairy cows in the tropics.  Proceedings of the FAO expert consultation held in Bangkok, Thailand, 7 – 11 July 1989. Speedy, A. and Sansoucy, R.  (Eds.), pp. 142-148. FAO Animal Production and Health Paper No. 86, FAO, Rome.

Mpairwe, D. R. 1998.  Integration of forage legumes with cereal crops for improved grain yield, forage production and utilisation for smallholder dairy production systems.  Ph.D. Thesis, Makerere University, Kampala. 240pp.

Mpairwe, D. R., Sabiiti, E. N. and Mugerwa, J. S. 1998.  Effect of dried Gliricidia sepium leaf supplement on feed intake, digestibility and nitrogen retention in sheep fed dried KW4 elephant grass (Penisetum purpureum) ad libitum. Agroforestry Systems 41:139-150.

Muhr, L., Tarawali, S. A., Peters, M., Merkel, U., Schultze-Kraft, R. and Berner, D. 1997.  Multiple uses of tropical forage legumes for sustainable farming in the moist savannahs of Africa. In: Proceedings of XVIII International Grassland Congress, June 8-19, 1997. Winnipeg, Manitoba, Saskatoon, Saskat-chewan, Canada.  2:19-50.

Mukasa-Mugerwa, E. 1989. A review of reproductive performance of female Bos indicus (zebu) cattle. ILCA Monograph 6.  ILCA, Addis Ababa, Ethiopia. 134 pp.

Muwanga, J. W. 1994. An economic evaluation of zero grazing dairy production systems in Uganda: A case study of Mpigi and Mukono Districts. M.Sc. Thesis, Makerere University, Kampala. 97 pp. 

Nsubuga, H. S. K. 1993.  Pastures in the feeding of dairy cattle in Uganda, with particular reference to zero grazing.  In: Pasture and feeding systems for sustainable livestock production and environment. Proceedings of the second Uganda Pasture Network Workshop held at Makerere University, 14-16 December 1992. Sabiiti, E. N., Bareeba, F. B. and Mwebaze, S. (Eds.), pp. 64-68. Kampala, Uganda.

Pindyck, S. R. and Ruben, D. L. 1991.  Econometric models and economic forecasts. 3rd edition. Mcgraw-Hill Inc., New York. 596 pp.

Preston, T. R. 1986.­  Matching livestock systems with available feed resources in tropical countries. CTA, The Netherlands. 10 pp.

Preston, T. R. 1995. Tropical animal feeding.  A manual for research workers. FAO Animal Production and Health paper No. 126. FAO, Rome, Italy.

Randel, R.D. 1990.  Nutrition and postpartum re-breeding in cattle.  Journal of Animal Science  68:853-862.

Sabiiti, E. N. 1993. Integration of forage legumes in crop/livestock production systems.  In: Pasture and feeding systems for sustainable livestock production and environment. Proceedings of the second Uganda Pasture Network Workshop held at Makerere University 14-16 December 1992. Sabiiti, E. N., Bareeba, F. B. and Mwebaze, S. (Eds.), pp.19-26. Kampala, Uganda.

SPSS, 1994. SPSS Advanced statistics 6.1.  Chicago: SPSS Inc.

Staal, S. J. and Shapiro, B. I. 1996.  The impact of policy liberalisation on smallholder peri-urban dairy producers at Nyeri, Kenya.  In: Agricultural competitiveness, market forces and policy choice: Eastern and Southern Africa perspectives and case studies. Mudimu, G. D. and Mabeza-Chimedze, R. (Eds.), pp. 165-179. Studien Zu, 52.  Lit Verlag Miinster, Hamburg. 

Tumutegyereize, K., Hyuha, T. and Sabiiti, E. N. 1999.  Factors affecting dairy production in the peri-urban areas of Kampala. Uganda Journal of Agricultural Sciences 4:7-11.

Winrock International, 1992.  Assessment of Animal Agriculture in Sub-Saharan Africa.  Institute for Agricultural Development.  Arkansas, USA. 125 pp.

Wortmann, C. S. and Eledu, C. A.  1999.  Uganda’s agro-ecological zones: A guide for planners and policy makers. CIAT, Kampala, Uganda. 54 pp.

Zake, J.Y.K. 1993.  Soil fertility in relation to mineral nutrition of pastures and environment protection.  In: Pasture and feeding systems for sustainable livestock production and environment. Proceedings of the second Uganda Pasture Network Workshop held at Makerere University 14-16 December, 1992. Sabiiti, E. N., Bareeba, F. B. and Mwebaze, S. (Eds.), pp. 96-105.  Kampala, Uganda.

 ©1999, African Crop Science Society