Livestock Research for Rural Development 15 (2) 2003

Citation of this paper

Agro-industrial by-products, cowpea residues and urea-treatment of hay for supplementary dry season feeding of mature zebu oxen in Mali 


P W  Bartholomew**, R  Ly*, M  Doumbia*, T  Khibe*, N'g  Kone*, B  Traore* and S  Ba* 


USDA/Agricultural Research Service, Langston University
P.O. Box 730, Langston, OK 73050, USA.  (addressee for correspondence)
Institut National de Recherche Zootechnique, Forestière et Hydrobiologique (INRZFH)
BP 1704, Bamako, Mali
**Formerly: International Livestock Centre for Africa (ILCA), Programme du Sahel
BP 60 Bamako, Mali




During short-term (9 to 11 weeks) dry-season feeding in central Mali, liveweight changes in mature zebu oxen (Bos indicus) were measured to evaluate diets based on natural pasture hay and supplemented with a range of levels and types of feeds available to smallholder farmers. 


Among the feeds tested cottonseed residues had the greatest effect on cattle daily liveweight change (DLWC) which was described as a curvilinear function of the amount of cottonseed residue (CR) offered over levels ranging from 0 to 4 kg.hd-1.d-1; DLWC=1.34-1.42(0.674CR). Each 1 kg.d-1 increment in offer of cowpea haulm increased DLWC by between 0.04 and 0.15 kg.hd-1.d-1. A mixture of 60% hay and 40% molasses with urea added at 10 increased average daily weight gain by 0.19 kg.hd-1.d-1, compared with diets based on untreated hay.  In on-farm trials, supplementation of intake from grazed natural pasture with hay ensiled with urea reduced weight loss from an average of 0.18 kg.hd-1.d-1 to 0.03 kg.hd-1.d-1 over a 2-month period at the end of the dry season. 


The benefit derived from limited quantities of cottonseed residue can be increased by restricting individual animal intake and feeding more animals. Improvement of feed self-sufficiency on smallholder farms should focus on production of legume feeds with increased nitrogen content.  


Keywords:  By-product, crop-residue, dry-season, mali, oxen, supplementation, Zebu 



Livestock production in many tropical environments is constrained by low feed availability and low feed quality during a prolonged dry season (Leng 1984).  In the semi-arid zone of Mali such problems are illustrated by weight losses of 15-25% in mature cattle during the dry season (Wilson 1986; Bartholomew et al 1995).  Dry season feed constraints are further manifest in a reduction in the number and size of cattle offered for slaughter between April and June each year (Direction Nationale de l’Elevage 1989) and, as supply declines, through an increase in the unit price of beef (Delgado and Staatz 1980).  Given adequate feed resources it is possible to avoid seasonal weight loss in cattle. In short-term feeding of mature zebu castrates, weight gains of between 0.5 and 0.9 kg.d- 1 have been shown to be possible, both in experimental conditions (Calvet et al 1969; Calvet and Valenza 1973; Traore et al 1973) and on smallholder farms (Bartholomew et al 1992).  However, such weight gains have been heavily dependent on the use of purchased concentrate feedstuffs or agro-industrial by-products.  Supplies of these are limited and high demand elevates prices (Baur et al 1987).


Identification of feeding regimes that will allow small farmers to reduce their dependence on external supplies of concentrate feedstuffs has the potential to increase the viability of livestock production enterprises, to increase dry season beef supplies and increase household income.  However, little information exists on livestock performance response to a range of supplementary feeds available to smallholder farmers. The experiments described in this paper were undertaken to provide information on productivity of cattle fed a basal diet of natural pasture hay, supplemented with varying levels of agro-industrial by-products and home-grown legume residues, and to identify more efficient dry-season feeding practices for small farmers.



Materials and methods


The effects of supplementation of a basal diet of natural pasture hay with different levels of agro-industrial by-products, cowpea haulm (the dry residue remaining after hand-harvest of cowpea grains) and urea-treatment of hay on the liveweight changes of mature male zebu castrates were assessed in feeding trials carried out in four successive hot dry-seasons in the semi-arid zone of Mali. Experiments 1 and 2, each of 77 days duration, were researcher-managed and carried out on-station.  Experiments 3 and 4 were conducted on smallholder farms with day-to-day management of livestock carried out by cooperating farmers, over feeding periods of 65 days in each case.


On-station trials


The animals used for the experiments were mature male castrate zebu cattle (Bos indicus), typically oxen at the end of their working career. Eighteen of each experimental group were zebu Maure, of provenance northern Mali or Mauritania, and eighteen were zebu Peul (white Fulani), originating from the Niger delta area of Mali.  All the animals were purchased in markets within their respective zones of origin and were trekked to the experimental site (between 100 and 200 km). For up to two weeks after arrival, until the beginning of the trial, the cattle were grazed on dry natural pasture during the day and enclosed at night. During this pre-experimental period all animals were weighed and ear-tagged, vaccinated against rinderpest and contagious bovine pleuro-pneumonia and subject to a deworming treatment with "Dovenix"(a) (Nitroxinil) and injected with "Berenil"(a)  (Diminazene) for control of any trypanosome infestation. Subsequent veterinary attention was given to individual animals as necessary, for external parasite control and for some cases of jaw abscesses. Average liveweights at the beginning of experiment 1 were 269 (±21.8 SD) kg for Maure and 263 (±21.7 SD) kg for Peul animals. The average liveweight of each group at the beginning of experiment 2 was 302 (± 24.7 SD) kg and 277 (±22.6 SD) kg for Maure and Peul, respectively.  Throughout each trial the animals were housed in individual pens covered by a straw shade. All feeding and watering was made on an individual basis and drinking water was available ad libitum throughout the experiments.


In both trials the basal roughage offered was a mixture of grasses from natural pasture harvested as dry standing hay in February-April. The predominant species in the hay were Pennisetum pedicellatum, Schoenfeldia gracilis, Loudetia togoensis and Cenchrus biflorus.  This hay was fed in rations that included  molasses and urea or dried cowpea haulms and cottonseed residues in pellet form containing 45% cotton seed cake, 50% cotton seed husks and 5% salt.  In experiment 1, three “hay presentation” treatments were supplemented with the cottonseed residue supplement offered at 2 or 4 kg.d-1.  These main-treatment feeding levels were arranged in a 3 x 2 factorial layout to give a total of six feeding regimes.  The “hay presentation” treatments consisted of unchopped hay offered either alone, with 2 kg.d-1 of cowpea haulm, or mixed with molasses and urea in proportions of 40% molasses and 60% hay, with urea added at 10 of the mixture.  In all treatments hay, or hay mixed with molasses and urea, was offered ad libitum to an average refusal level of 15%.  In experiment 2, feeding treatments were cowpea haulm offered at 1 or 2 kg.d-1 and cottonseed residue offered at 0, 1 or 2 kg.d-1.  Cowpea and cottonseed feeding treatments were combined factorially to provide six levels of supplementation to natural pasture hay offered unchopped and ad libitum to a refusal of a mean 25%. 


Feeds were provided for each animal in an individual feed trough, consisting of one half of a 225 Liter barrel cut longitudinally, partitioned to allow separate estimation of intake of hay and one supplementary feed.  In treatments that included cowpea haulm in addition to cottonseed residue, the latter was offered in a bucket adjacent to the main feed trough. The daily allowance of each feed, according to treatment, was provided by mid-morning each day throughout the experiments. In both trials levels of offer were based on air-dry feed weights. Feed intakes were estimated daily for each animal throughout the feeding period by measure of the amount of feed offered to each animal and of the feed refusal from the previous day.  Samples of refused and offered feeds were collected on four days each week throughout the experiments, and were dried in a forced-air oven at 100 °C to allow estimation of oven-dry weights of feed intakes. A sample of each offered feed was taken on one day each week for chemical analysis. Samples were bulked by treatment and sub-sampled for chemical analysis for estimation of nitrogen content of feeds by micro-Kjeldahl digestion.  For experiment 1, estimates of metabolizable energy (ME) intake were derived from measured DM intakes of each ration component and from standard figures for the materials used (Boudet 1978; MAFF 1975) as follows; 2.1, 8.2, 11.2 and 12.7 feed for hay, cowpea haulm, cottonseed residues and molasses respectively. 


In on-station trials animals were ranked by weight within breed and blocked in groups of six. Feeding treatments were allocated at random within each block. Each ration treatment was thus replicated six times to allow analysis of the effects of treatments on intake and liveweight change in a randomized block layout.  In both experiments the animals were weighed weekly and animal mean daily liveweight change was calculated by linear regression of the weight of individual animals at successive weighings against time. The calculated regression coefficients were subject to analysis of variance for estimation of the effect of different supplementation treatments on mean daily liveweight change. An additional comparison of performance of the different breeds in experiments 1 and 2 was made by "t" test of the mean daily liveweight gain of each breed, averaged over feeding treatments and replications.


In experiment 2 the body condition of each animal was assessed at the beginning and end of the trial using the 9-point scale described by Nicholson and Butterworth (1986).


On-farm trials


In an on-farm evaluation (experiment 3) 14 cooperating farmers constructed above-ground mud-brick silos to ensile urea-treated natural pasture hay and used the material produced as supplementary feed for their work oxen during the late dry season (April-June).  Each farmer ensiled approximately 600 kg of natural pasture hay, harvested as dry standing-crop. This material was similar in species composition and quality to the hay used in experiments 1 and 2.  The hay was mixed 1:1 (w/w) with water in which fertilizer-grade urea had been dissolved at 40 g urea per kg of water to give a urea inclusion rate of 4% of dry hay.  Following treatment the hay was covered with grass matting and the silo was sealed with a 20 cm layer of mud and fermented for approximately 14 days. A total of 30 work oxen (ten per treatment) on 14 smallholdings were subsequently fed over a period of 65 days in one of the following ways; (1) with grazed pasture (in the form of dry standing natural pasture hay) alone, (2) with grazed pasture and urea-treated hay offered at 6 kg.hd-1.d-1 or, (3) with grazed pasture supplemented with 6 kg.hd-1.d-1 of urea-treated hay and 0.5 kg.hd-1.d-1 of cottonseed residue. All animals were grazed on village common-land and had similar access to grazed feed.  In the following dry season these procedures were repeated (experiment 4), except that cottonseed residue allowance was increased from 0.5 to 1.0 kg.hd-1.d-1.  Average liveweight of oxen at the beginning of each feeding period was 281 (±58.5 SD) kg in experiment 3 and 289 (±61.7 SD) kg in experiment 4.


Feeding treatments were allocated to specific animals after discussion with the group of farmers and it was then left to each farmer-owner to apply the designated treatments, with periodic oversight from technical staff.  Measurement of treatment effect was limited to twice-monthly weighing of animals.  In on-farm trials feeding treatments were distributed across farms and were replicated ten times.  Estimation of mean daily liveweight change was made using the regression procedure adopted for on-station experiments. Comparison of the effects of feeding treatment on weight change was made by analysis of the calculated regression coefficients of liveweight against time for each animal in a completely randomized design of ten replicates.


On-station trials


Feeding treatment effects on feed intakes and liveweight changes are summarized in Tables 1 and 2 for experiments 1 and 2 respectively.  Since there was no significant interaction effect (P>0.05) between feeding treatments in either experiment only main treatment means are presented.

Table 1.  Effect of form of presentation of hay and of level of supplementation with cottonseed residue on main-treatment mean daily feed intakes and liveweight changes in zebu castrates (Experiment 1).


Hay DM intake, kg.d-1

Total DM intake, kg.d-1

ME intake, MJ.d-1

Nitrogen intake, g.d-1

Weight change, kg.d-1

Presentation of hay






Hay alone






Hay + molasses + urea






Hay + 2 kg.d-1 cowpea haulm












Level of supplementation






2 kg.d-1 cottonseed residue






4 kg.d-1 cottonseed residue












Asterisks indicate levels of F probability derived from analysis of variance tables and follow the convention (***) P<0.001, (**) P<0.01 and (*) P<0.05


For experiment 1, average nitrogen contents of the ration components were 4.5, 9.5, 16.9 and 34.4 feed, for hay, molasses-urea treated hay, cowpea haulm, and cottonseed residues respectively. Dry matter content of the materials fed was 974 for hay and cowpea haulm, 963 for cottonseed residues, and 822 for molasses treated hay.  Mean DM intakes of cottonseed supplement were 3.67 kg.hd-1.d-1 and 1.92 kg.hd-1.d-1 respectively for the 4 and 2 kg.hd-1.d-1 levels of offer. Differences in cottonseed residue intake according to the form of presentation of hay were not significantly different (P>0.05).  Cowpea haulm was readily consumed by the animals to which it was offered and its intake was not significantly (P>0.05) affected by level of cottonseed residue feeding.  Mean daily intake of cowpea was 1.90 kg DM.hd-1, from 2 kg.hd-1.d-1 of air-dry (1.95 kg DM) material offered. The amount of hay consumed was influenced by its form of presentation and by level of intake of cottonseed supplement (Table 1). An average intake of 1.90 kg DM.hd-1.d-1 of cowpea haulm reduced daily hay intake by a mean of 0.93 kg.hd-1 (P<0.001). In contrast, an increase in mean DM intake of cottonseed supplement from 1.92 kg.hd-1.d-1 to 3.67 kg.hd-1.d-1, decreased mean hay intake by only 0.60 kg.hd-1.d-1 (P<0.05).   Increase in offer of cottonseed supplement from 2 kg.hd-1.d-1 to 4 kg.hd-1.d-1 decreased hay intake by averages of 0.34, 0.50 and 0.99 kg.hd-1.d-1 on hay alone, hay mixed with molasses and urea, and hay plus cowpea haulm treatments respectively. This interaction between treatments was not significant (P>0.05).


Supplementation with cowpea haulm resulted in an average increase in total intake of 1.03 kg DM.hd-1.d-1, compared with hay supplemented with cottonseed residues only. The advantage of the additional cowpea was more apparent with the 2 kg.hd-1.d-1 level of cottonseed input, by an increase of 1.36 kg DM.hd-1.d-1 in total intake, compared with the higher level of cottonseed supplementation which showed only a 0.71 kg DM.hd-1.d-1 increase in intake attributable to the additional cowpea input.  Total nitrogen intake was highest with diets based on molasses-urea treated hay, but the dietary nitrogen concentration was greatest with diets supplemented with cowpea haulm. The lowest nitrogen concentration was observed with offered rations of hay and 2 kg.hd-1.d-1 of cottonseed residue, at 14.4 feed and the highest with hay, 4 kg.hd-1.d-1 of cottonseed residue and 2 kg.hd-1.d-1 of cowpea haulm, at 28.1 feed.


In experiment 2 average nitrogen contents of the ration components were 5.5, 15.9 and 38.0 feed, for natural pasture hay, cowpea haulm and cottonseed residues respectively. Dry matter content of feeds was 935 for natural pasture hay, 933 for cowpea haulm and 915 for cottonseed residues.  Levels of refusal of cowpea haulm and cottonseed residues were on average less than 30 g DM.d-1, and intakes therefore reflect closely the amounts of these feeds offered. Average daily DM intake of cowpea haulm was 0.93 and 1.81 kg.hd-1 at levels of offer of 1 and 2 kg.hd-1.d-1 respectively. For cottonseed residues DM intakes were 0.88 and 1.81 kg.hd-1.d-1 at feeding levels of 1 and 2 kg.hd-1.d-1.  Addition of cottonseed residue to the diet increased hay intake above the level obtained without residue feeding by a mean 35.7% with 0.88 kg.hd-1.d-1 of cottonseed intake, and by 38.3% with an average daily intake of 1.81 kg DM.hd-1 of cottonseed residue (Table 2). On average over cowpea feeding treatments, the addition of 1 kg.hd-1 of cottonseed residue each day increased daily feed intake by 45.8%, and 2 kg.hd-1 of cottonseed residue increased daily feed intake by 71.9%, in comparison with rations without cottonseed residue supplementation (P<0.001). An increase in the amount of cowpea offered, from 1 to 2 kg.hd-1d-1, increased mean total daily feed intake by only 10.7% (P<0.01). Through a combined effect of increased DM intake and increased nitrogen content in diets, the mean daily nitrogen intake was doubled by addition of 1 kg.hd-1 of cottonseed residue, and tripled with 2 kg.hd-1, compared with the average nitrogen intake on rations without cottonseed residue. An increase in cowpea intake from 0.93 to1.81 kg DM.hd-1.d-1, in contrast, resulted in a mean increase of 18% in daily nitrogen intake. (Table 2.) 

Table 2.  Effect of level of supplementation of natural pasture hay with cottonseed residue and cowpea haulm on main-treatment mean dietary nitrogen concentration, daily feed intakes and liveweight changes in zebu castrates (Experiment 2).


Hay DM intake, kg.d-1

Total DM intake, kg.d-1

Nitrogen intake, g.d-1

Nitrogen in feed,

Weight change, kg.d-1

Cottonseed residue offered

0 kg.d-1






1 kg.d-1






2 kg.d-1












Cowpea haulm offered

1 kg.d-1






2 kg.d-1







        0.189 NS






Liveweight changes


In experiment 1 all the diets produced increases in liveweight (Table 1). The highest mean daily liveweight gain (DLWG) of 1.12 kg.hd-1.d -1 was obtained with animals fed on molassed hay and 4 kg.hd-1.d -1 of cottonseed residue and the lowest (0.72 kg. hd-1.d -1) with those fed on untreated hay and only 2 kg.d-1 of cottonseed supplement. Addition of the molasses-urea mix to feeds increased daily liveweight gains by means of 18.4 and 10.8%, compared with diets consisting of hay and cottonseed alone or of hay, cowpea haulm and cottonseed residues.  The marginal benefit of supplementation with cowpea haulm in addition to 2 kg.d-1 of cottonseed residue was an increase in liveweight gain of 0.07 kg.d-1 for each 1 kg of cowpea haulm.


In experiment 2 animals fed on diets that did not include some supplementation with cottonseed residue lost weight throughout the trial.  With 2 kg.d-1 of cowpea haulm supplementation the daily weight loss was 0.05 kg, equivalent to a total loss of just under 4 kg in 77 days, but with only 1 kg of cowpea haulm per day the mean total loss during the feeding period was just over 19 kg, at an average loss of 0.25 kg.d-1. In contrast the inclusion of 1 kg.d-1 of cottonseed residue produced an average gain of 33 kg over the liveweight at the start of the trial, representing a net gain of 45 kg in comparison with the treatments without cottonseed residue supplementation. The average gain from use of a further 1 kg.d-1 of cottonseed residue, to a total of 2 kg.d-1, was approximately 18 kg, or 0.24 kg.d-1 (Table 2).


Breed effects


Differences in breed performance were small in experiment 1 with an average of 0.94 kg.d-1 for Maure animals and 0.97 kg.d-1 for Peul (P>0.05).  In experiment 2 the average daily weight change of the two breeds over 77 days of feeding in experiment 1 was 0.38 kg.d-1 for Maure and 0.25 kg.d-1 for Peul animals, but again the difference was not significant (P>0.05).  

Intake and weight change relationships   

In experiment 1 differences in DLWG were almost equally well explained by variations in estimated total DM intake (DMI) as by variations in total metabolizable energy intake (MEI). The relationships of total DM intake and ME intake to DLWG were as follows:


            DLWG =  0.081 (±0.0139) DMI + 0.306 (r2 = 0.50, P<0.001)


            DLWG =  0.008  (±0.0014) MEI + 0.505 (r2 = 0.51, P<0.001)


There was no evidence of a consistent relationship between dietary N concentration or total N intake and DLWG.  By contrast, in experiment 2 variation in daily liveweight change (DLWC) was better described by change in total daily nitrogen intake (NI) than by levels of total feed intake (DMI).


            DLWC = 0.0104 ("±0.0008) NI - 0.511   r2 = 0.84  (P<0.001)


            DLWC = 0.226 (±"0.0221) DMI - 0.979   r2 = 0.76  (P<0.001)


Body condition changes   

In the mature oxen studied one unit change in body condition was associated with a weight change of approximately 21 kg.  Within the range of body condition observed (points 2-7 on the 0-9 scale), total weight change (WC) and body condition change (C) approximated a linear relation;


WC = 21.3 ("±3.66) C + 8.21    r2 = 0.50  (P<0.001)


On-farm trials


The effect of treatments on weight change is shown in Table 3.  Feeding of urea-treated hay to supplement intake from grazed pasture produced, at best, a very slight increase in weight over the feeding period and on average reduced the rate of daily weight loss, compared with pasture alone, by 0.15 kg.hd-1.  As in experiments 1 and 2, addition of cottonseed residue to the diet had the greatest impact on weight change. 

Table 3. Effect of 65 days supplementation with 4% urea-treated hay and with cottonseed residue on mean daily weight change (kg) of zebu oxen in village receiving a basal diet of natural pasture by grazing

Supplementary feed      

Weight change

(Expt. 3)

Weight change

(Expt. 4)

No supplement              



Urea-treated hay            



Urea-treated hay + 0.5 kg cottonseed residue



Urea-treated hay + 1 kg cottonseed residue









Animal performance and feed intake   

In experiment 1 zebu castrates were shown to be able to gain weight at a rate of 0.72 kg.d-1 on a diet based on natural pasture hay supplemented with 2 kg.d-1 of cottonseed residue. Addition of dry cowpea haulm to this diet increased daily weight gain, but not sufficiently to justify its use economically; partial-budget analysis of the results showed that the marginal value of cowpea hay in weight gain was approximately half of its value in direct sale (Baur et al 1987).  Feeding levels for experiment 2 were selected in order to investigate the effect of reduction in cottonseed residue supplementation and its substitution by cowpea, and thus to evaluate feeding practices adoptable by smallholder farmers using home-grown forages as a primary source of supplementation to low quality roughages. As expected therefore, the absolute performance levels observed in experiment 2 were lower than in the earlier trial. In similar treatments (hay supplemented with 2 kg.d-1 of cowpea haulm and 2 kg.d-1 of cottonseed residue) weight gain (0.74 v 0.85 kg.d-1) and feed conversion efficiency (9.7 v 8.6 kg gain) were lower than the earlier trial, but still within norms reported in other work with similar animals (Calvet and Valenza 1973).


There was a marked contrast between the trials in weight change response to level of nitrogen intake. This effect may reflect a response to increase in nitrogen concentration from sub-optimal levels, and an associated increase in N intake in experiment 2.  With diets based on molasses-urea treated hay, productivity in relation to estimated energy intake was consistent with expectations based on feeding standards (MAFF 1975). With other feeding treatments, particularly those consisting of hay and cottonseed residue alone, performance was higher than apparent energy intakes would predict. The inconsistency may serve to emphasize comments by Preston (1985) and Chenost and Kayouli (1997) that feeding standards based on assigned nutritive values may be misleading when applied to unconventional or poor quality feedstuffs. Under such circumstances there is little substitute for in vivo assessment of their value in a ration (Orskov 1986).  These observations should have particular resonance in sub-sahelian Africa, where poor quality hay or crop residue is often the principal feed resource and where facilities for feed analyses are often either not available or are prohibitively expensive.  The results of the experiments reported show that DM intake may be as good an indicator of animal performance as estimated feed value. In resource-limited research systems simple intake trials may therefore serve as an effective means of assessing the usefulness of novel feeds or ration combinations.  Combining results from the two trials suggests that the relation of liveweight change to total feed DM intake follows a curvilinear form;


DLWC = 0.53 (±"0.069) DMI - 0.02 (±"0.005) DMI2 - 1.79 (r2 = 0.80)


Breed effects   

In a comparison of Maure and Peul animals up to two years of age, under experimental conditions, no difference between performance of the two breeds was observed (ILCA/IER 1978). In contrast, observations made in-village with mature animals of both breeds, showed a mean liveweight gain of Maure cattle 62% greater (P<0.01) than that of Peul animals (Bartholomew et al 1992).  Breed performance in experiment 1 showed no significant or substantial difference between Maure and Peul, but it is possible that any difference in productivity between breeds was obscured by confounding of breed and age in this trial.  Results from experiment 2 reflected the in-village observations with mean daily weight change of Maure animals showing a 52% advantage over the Peul, although this difference was not statistically significant (P>0.05). In view of its potential practical and economic significance, further research into relative breed performance, particularly with older animals of the type used for fattening, appears to be justified.


Response to supplementation with cottonseed residue


Cottonseed residue (CR) was fed at levels ranging from 0 to 4 kg.d-1 in experiments carried out both on-station and on-farm.  In spite of differences in levels of cowpea supplementation and treatment of roughage with urea, the impact of cottonseed supplementation on weight gains was quite consistent across experiments. The following relation between cottonseed residue intake and daily weight gain was derived from the combined results;


DLWC = 1.339-1.423(0.674CR)  r2=0.90


The diminishing-response nature of the relation shows the possibility for increasing the benefit derived from limited supplies of cottonseed residue; for example, a reduction in supplementation level from 4 to 2 kg.d-1, would reduce the average growth rate of an animal from 1.04 to 0.69 kg.d-1 but with a fixed supply would allow supplementation of twice the number of animals and offer a potential increase in daily gain of over 32%, without unduly compromising individual animal performance.


On-farm trials with “no supplementation” treatments, served to emphasize that the real benefit of dry season feed supplementation for cattle is compounded of the weight gained during the experimental period and of the weight that would have been lost during the same period in the absence of supplementation.  Over a ten-week feeding period the average net weight increase from supplementation with 1 kg.d-1 of cottonseed residue is therefore estimated at 41 kg, and at 64 kg with supplementation at 2 kg.d-1, compared with animals maintained on natural pasture hay alone.


Use of legume forages for feed supplementation


The results demonstrate the inadequacy of traditional supplementation practice using dry cowpea residues (average 16.4 g feed) for fattening livestock. In contrast, feeding cottonseed residue supplement with high nitrogen content (38.0 g feed), even in relatively small quantity, was effective in increasing overall ration nitrogen content and liveweight gain. This illustrates the potential value of a legume forage of high nitrogen content and suggests that forage production research targeted at smallholders should focus on identification of legumes, or on  improved harvest or storage practices, that will maintain a high nitrogen content up to harvest and through dry-season storage. If farmers can be encouraged to produce even limited quantities (on the order of 200 kg DM.hd-1) of high quality (N content of 25 legume forage for supplementary feeding they should be able to increase livestock production while minimizing their burden of forage production.


Urea-treatment of hay


Urea-treated hay as a sole source of supplement may be useful for reduction in dry-season weight loss but is of little value for fattening cattle.  However, the technique merits further investigation in conjunction with other feed sources as a possible means of maximizing the benefit of small quantities of legume hay or agro-industrial by-product.




This work was carried out by a joint ILCA-INRZFH team working under the auspices of Projet Sectoriel de l'Elevage 688-0218, financed jointly by USAID and the Government of the Republic of Mali.


a Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. 


Bartholomew P W, Ly R, Nantoume H, Kone N'g, Traore B, Khibe T, Sissoko K and Baur H 1992 Dry season cattle fattening by smallholder farmers in the semi-arid zone of Mali. African Livestock Research, 1, 31-38


Bartholomew P W, Khibe T and Ly R 1995  In-village studies of the use of work oxen in central Mali.  Tropical Animal Health and Production, 27, 241-248


Baur H, Sissoko K and Debrah S 1987 The economics of peasant cattle feeding in Mali. In: Said, A.N. and Dzowela, B.H. (eds) Overcoming constraints to the efficient utilization of agricultural by-products as animal feed. Proceedings of the fourth annual workshop held at the Institute of Animal Research, Mankon Station, Bamenda, Cameroun, 20-27 October 1987. African Research Network for Agricultural By-Products (ARNAB) 1989. ILCA, P.O. Box 5689, Addis Ababa, Ethiopia. pp. 49-60.


Boudet G 1978 Manuel sur les paturages tropicaux et les cultures fourragères. Manuels et Précis d'élevage 4. Institut d'Elevage et de Médecine Vétérinaire des Pays Tropicaux. Ministère de la Coopération, République Francaise. 258 pp.


Calvet H and Valenza J 1973 Deux années d'expérimentation d'embouche intensive de zébus au Sénégal. Documents présentés à la conférence sur la recherche et la production agricoles en Afrique. Septembre 1971. Supplément du Journal de l'Association pour l'Avancement en Afrique des Sciences de l'Agriculture 32-36.


Calvet H, Valenza J, Orue J and Wane A M 1969 Engraissement intensif de zébus peulh Sénégalais (Gobra). (Etude complémentaire). Institut d'Elevage et de Médecine Vétérinaire des Pays Tropicaux. Laboratoire National de l'Elevage et de Recherches Vétérinaires, Dakar-Hann. Ministère du Développement Rural. République du Sénégal. Travail exécuté hors Convention  no. 274/N/68/N-115/CD/68/VI/N/5. Décembre 1969.


Chenost M and Kayouli C 1997 Roughage utilization in warm climates.  FAO animal production and health paper 135.  Food and Agriculture Organization of the United Nations, Rome 1997. 226p.


Delgado C L and Staatz J 1980 Livestock and meat marketing in West Africa. Volume III. Ivory Coast and Mali. Center for Research on Economic Development. The University of Michigan. February 1980.


Direction Nationale de l'Elevage 1989 Rapport Annuel 1988. Ministère des Ressources Naturelles et de l'Elevage, République du Mali. 110 pp.


ILCA/IER 1978 Evaluation of the productivity of the Maure and Peul breeds of cattle at the Station du Sahel, Niono. ILCA Monograph No 1. ILCA, Addis Ababa, Ethiopia. 109 pp.


Leng R A 1984  Supplementation of tropical and subtropical pastures for ruminant production.  In:  Herbivore nutrition in the subtropics and tropics. editors. Gilchrist F M C and Mackie R I pp.129-144.


Ministry of Agriculture Fisheries and Food (MAFF) 1975 Tables of Feed Composition and Energy Allowances for Ruminants. Ministry of Agriculture Fisheries and Food (UK) February 1975. 36 pp.


Nicholson M J and Butterworth M 1986 A guide to condition scoring of zebu cattle. International Livestock Centre for Africa (ILCA), Addis Ababa, Ethiopia. 31 pp.


Orskov E R 1986 Methods of estimating nutritive value of fibrous residues. In: Little D A and Said A N (editors) Utilization of Agricultural By-Products as Animal Feeds in Africa. Proceedings of a Workshop held at Ryall's Hotel, Blantyre, Malawi, September 1986. African Research Network for Agricultural By-Products (ARNAB) 1987. ILCA, P.O. Box 5689, Addis Ababa, Ethiopia. pp. 1-6.


Preston T R 1985 Strategies for optimizing the utilization of crop residues and agro-industrial by-products for livestock feeding in the tropics. In: Preston T R and Nuwanyakpa M Y (editors) Towards optimal feeding of agricultural by-products to livestock in Africa. Proceedings of a workshop held at the University of Alexandria, Egypt, October 1985. African Research Network for Agricultural By-Products (ARNAB) 1986. ILCA, P.O. Box 5689, Addis Ababa, Ethiopia. pp. 145-166.


Traore N'g, Aw D and Tamboura T A 1973 Résultats des expériences d'engraissement intensif de zébus à la sucrerie de Dougabougou. Utilisation d'une ration à base d'urée. Station d'Elevage et Recherche Zootechnique du Sahel, Niono. Institut d'Economie Rurale, Ministère de la Production, République du Mali.


Wilson R T 1986  Livestock Production in Central Mali. Research Report no 14. International Livestock Centre for Africa, Addis Ababa, Ethiopia. 111 pp.



Received  27 January 2003,  Accepted  19 February 2003


Go to top