Livestock Research for Rural Development 23 (5) 2011 Notes to Authors LRRD Newsletter

Citation of this paper

In sacco and in vivo evaluation of marula (Sclerocarya birrea) seed cake as a protein source in commercial cattle fattening diets

V Mlamboa, B J Dlaminib, M D Ngwenyab, N Mhazob, S T Beyenec and J L N Sikosanad

aDepartment of Food Production, Faculty of Science and Agriculture, University of the West Indies, St Augustine, Trinidad & Tobago.
bFaculty of Agriculture, University of Swaziland, P. O. Luyengo, M205, Swaziland,
cFaculty of Agriculture and Natural Resource Management, University of Namibia, Neudamm Campus, P/Bag 13301, Windheok, Namibia.
dMatopos Research Station, Department of Agricultural Research and Extension, P. Bag K5137, Bulawayo, Zimbabwe.   /


The deciduous marula (Sclerocarya birrea subspecies caffra) trees are abundant in the lowveld of Swaziland where the ripe fruits are used to make a local alcoholic beverage (buganu). Marula seed cake (MSC) is a residue remaining after the extraction of oil from marula kernels. Although it is generally accepted that MSC is rich in protein, the quality of this protein and hence its nutritional value to ruminant animals is largely unknown. This study, therefore, evaluated the in sacco degradability of MSC protein and its potential as a protein source in fattening rations for feedlot cattle.

In the first experiment the in sacco degradability of MSC protein was evaluated by incubating MSC-containing nylon bags in the rumen of goats. The effect of particle size on degradability of MSC protein was also assessed by incubating unmilled and milled MSC. An increase in the immediately degradable fraction (a) was observed in milled MSC. However, milling reduced (P < 0.05) the slowly degradable (b), potential degradable (PD) and effectively degradable (ED) fractions of dry matter and nitrogen in MSC. In the second experiment, three iso-nitrogenous diets were evaluated in vivo using dairy weaners reared in feedlots. Weaners were offered a commercial fattening ration (with urea as a N source) (CFR-U), commercial fattening ration in which urea was wholly replaced with MSC (CFR-MSC) or  commercial fattening ration in which MSC and urea contributed equal amounts of N (CFR-MSC+U). The growth rate, total feed intake and feed conversion efficiency of dairy weaners did not differ (P > 0.05) between the 3 diets. CFR-U, CFR-MSC and CFR-MSC+U promoted similar average daily gains of 1.62, 1.75 and 1.82 kg, respectively. It was concluded that replacing urea, as a source of N, with MSC did not reduce the feed value of commercial fattening ration, therefore MSC can be used in place of urea in resource-poor communities to reduce feed costs and possible incidences of urea poisoning in feedlot animals. 

Key words: dairy weaners, feedlot cattle, growth performance, particle size, rumen degradability


Marula (Sclerocarya birrea subspecies caffra) trees are abundant in the lowveld of Swaziland where rural communities make extensive use of its fruits within their localities. Marula has multiple uses: fruits are eaten fresh or fermented to make an alcoholic beverage (buganu), kernels are eaten or oil is extracted from them, leaves are browsed by livestock and have medicinal uses, as does its barks (Shackleton 2002). Recognizing the importance of the marula tree to the livelihood of people living in the communal areas known as the Swazi Nation Land (SNL), the government has started trials to establish commercial plantations of marula trees.  

In 2003, Swaziland’s cattle population stood at 600 252 of which 87% were owned by resource-poor smallholder on the SNL. Although Swaziland has a beef quota to supply European markets, this quota has never been met due to the low take-off from the SNL (MOAC 2004). In order to boost beef production, communal farmers are being encouraged to venture into feedlotting of beef cattle and thus improve the contribution of the beef industry to the country’s economy. The major constraint to rearing feedlot cattle for resource-poor farmers on the SNL is the cost of feedstuffs especially the protein/nitrogen component of the fattening ration. In an effort to cut down on feed cost in feedlots, farmers are encouraged to formulate their own diets using readily available feed resources. Currently, commercial fattening rations are formulated with urea as a nitrogen supplement. For communal farmers, urea is not readily available, unaffordable and requires technical know-how to reduce incidences of urea poisoning. Clearly, any locally available, low-cost feed resource that can supply nitrogen in place of urea would allow farmers to formulate cheaper rations and improve the profitability of feedlot cattle production.  

One such feed resource is marula seed cake, a residue from oil extraction of marula kernels. Marula nuts are collected after the use of fruit pulp in processing local alcoholic brew which usually occurs between January and April each year. The nuts are sun dried in yards and mechanical decortication yields the kernels. Kernels are roasted at temperatures between 45oC and 47oC to make oil extraction easier. Oil from marula kernels is extracted mechanically by applying pressure on the seed kernel using manual ram presses. Marula seed cake is the residue that remains after oil extraction. By virtue of it being a seed residue, MSC contains appreciable quantities of crude protein. However, little is known about the possible utility of MSC protein for ruminant animals. This study was, therefore designed to investigate the quality of protein from MSC using an in sacco DM and nitrogen degradability technique as well as establishing the comparative nutritional value of the seed cake when offered to feedlot cattle in place of the conventional urea. 

Materials and methods

Study Area 

The feedlot trial was conducted at the University of Swaziland Farm located in the upper Middleveld of Swaziland (coordinates: 260 32’ S - 310 14’ E, altitude 738 m). Annual rainfall for this site ranges from 850 mm to 1000 mm (Monadjem and David 2005). Chemical analyses were carried out in the nutrition laboratory at the University of Swaziland’s Faculty of Agriculture. The in sacco degradability experiment was carried out at Matopos Research Station, Bulawayo, Zimbabwe (coordinates: 20░ 23' S – 28░ 28' E, altitude 1340 m) using rumen cannulated goats.  

Chemical analysis 

Marula seed cake was obtained from Swazi Secrets, an oil extraction company situated at Mpaka, Swaziland. Part of the MSC samples was milled to pass through a 2 mm screen for chemical analysis. Ether extract (EE) was determined using a Soxhlet apparatus with petroleum ether as a solvent according to the method of AOAC (1999). Ether extract residues were analysed for neutral detergent fibre (aNDFom) and acid detergent fibre (ADFom) according to Van Soest et al (1991). Neutral detergent fibre was assayed without sodium sulphite but with a heat-stable α-amylase to assist with starch hydrolysis. Both aNDFom and ADFom were expressed without residual ash. Crude protein content was estimated using the standard macro-Kjeldahl method (AOAC 1999, method no. 976.06). Dry matter (DM) was estimated by drying samples in an oven at 105░C for 12 hours, while organic matter (OM) was estimated by ashing samples in a muffle furnace at 550░C for 6 hours. 

In sacco degradability  

Two castrated Matebele goats with surgically fitted rumen cannulae were used to assess the in sacco DM and N disappearance from MSC. The goats were offered an ad libitum supply of mixed grass hay as a basal diet and 200 g daily supplement of commercial goat meal (National Foods, Bulawayo, Zimbabwe). Fresh water was always available in troughs. The MSC was evaluated in two forms, milled (particle size ≤ 2 mm) and unmilled (particle size as obtained from the oil company’s presses). Three grams of milled marula seed cake (MMSC) and unmilled marula seed cake (UMSC) were weighed into nylon bags measuring 6 x 12 cm with a pore space of 40 Ám (Lockertex, Cheshire, England). Each nylon bag was then folded in half and heat sealed on two sides before being incubated intraruminally in duplicate per incubation period. Ten samples of each feed sample were suspended in rumen. The incubation used a 2 x 2 Latin square design with two periods of 3 days and a changeover period of 24 h. All bags in each period were inserted at the same time, before the morning feeding. Two bags per treatment were sequentially withdrawn at 6, 12, 24, 48 and 72 h post-incubation. At the end of each of the 5 incubation periods, samples were washed under running tap water until rinse water was clear and dried in an oven for 48 hours at 60oC to determine DM disappearance. Residual DM in nylon bags was analysed for N using the standard macro-Kjeldahl method (AOAC 1999, method no. 976.06). The rapidly degradable fraction a (at time zero) of the samples was determined by washing nylon bags with warm water (39oC) without incubation in the goats’ rumen. The in sacco disappearance of DM and N fraction was estimated by fitting degradation data using NEWAY Excel Version 5.0 package (Chen 1997) to the ěrskov and McDonald (1979) non-linear model, p = a + b [1-e-ct], where p = disappearance of DM and N; a = soluble fraction; b = slowly degradable fraction; c = rate of degradation of b; t = time of incubation; e = exponential constant. Potential degradability (PD) was calculated as the sum of fractions a and b. Effective degradability (ED) was calculated after assuming a 5 %/h outflow rate (k) according to the equation: ED = , where a, b, and c are the constants from the ěrskov and McDonald (1979) equation above and k is the outflow rate.

Feedlot trial

Animals and housing 

Twelve 18 - 24 month-old dairy male weaners (185.9 ▒ 16.9 kg mean live weight) were placed in 4 groups of 3 according to initial live weight and three diets were randomly assigned within each weight group. Weaners were put in individual pens measuring 2 x 2.5 m equipped with feeding and drinking troughs. The animals were dewormed using Tramizole (Intervet, South Africa) at a rate of 60 ml/animal. For the duration of the trial (58 days) animals were weighed fortnightly to measure performance in terms live weight gain. Total feed intake per animal was also measured to enable estimation of feed conversion efficiency (FCE).

 Diets and feeding 

In this trial N source was the variable under investigation.  Therefore, three iso-nitrogenous diets were formulated as follows:

  1. Commercial fattening ration (urea is the main N source) (CFR-U).
  2. Commercial fattening ration in which all urea was substituted with marula seed cake (CFR-MSC). Quantity of seed cake added was enough to supply same amount of N as the substituted urea.
  3. Commercial fattening ration in which urea and marula seed cake supplied equal amounts of N (CFR-MSC+U) with the total N being equal to the amount of N supplied by urea in CFR-U diet.

A daily allocation of 10 kg of feed was offered to each animal in two equal rations in the morning (0800 h) and afternoon (1400 h). Refusals were weighed to calculate feed intake the following day before offering fresh feed. Daily feed intake was estimated by subtracting amount of refusals from amount offered. Animals had unlimited access to fresh water. Samples of diets offered to dairy weaners were milled using a rotor mill (Fritsch Pulverisette 14, Glen Creston Ltd, Middlesex, UK) to pass a 1 mm screen for chemical analyses. Dry matter, ash, EE, aNDFom, ADFom and CP were chemically analysed as described for MSC samples above.

Statistical analysis

In sacco DM and N degradability data was analysed using the general linear models (GLM) procedure of the statistical analysis system (SAS 2001). Particle size (milled and unmilled MSC) means were separated using the probability of difference (PDIFF) option of the least squares means statement of the GLM procedure of SAS. The GLM procedure was also used for a one-way analysis of variance on feedlot cattle performance data with diet as the main effect. Diet (CFR-U, CFR-MSC, and CFR-MSC+U) means were also separated using the PDIFF option (SAS 2001).


The chemical composition of MSC is presented in Table 1. The CP content of MSC was 470 g/kg DM while EE content was 394 g/kg DM.  

Table 1. Chemical composition (g/kg DM) of marula seed cake used for in sacco degradability and feedlot trials.

Chemical constituent


Dry matter


Crude protein




Ether extract


Neutral detergent fibre


Acid detergent fibre


Table 2 presents the ruminal degradation kinetics for milled and unmilled MSC. The instantly soluble DM fraction a, was 0.1 % and 2.5 % in unmilled and milled MSC, respectively. Milling MSC (i.e., reducing particle size) significantly (P < 0.05) increased the soluble fraction (a) of DM but reduced (P < 0.05) the slowly degradable fraction (b), potentially degradable fraction (PD) and effectively degradable fraction (ED). Milling MSC had no effect (P > 0.05) on the rate of degradation (c) of the slowly degradable fraction which was 5 %/h. Reducing particle size of MSC affected (P < 0.05) degradability of N in a similar manner to DM with the exception that higher degradability values were observed as expected. 

Table 2. Ruminal degradation kinetics, potential degradability, and effective degradability (ED) of dry matter and nitrogen of marula seed cake



Marula seed cake







Dry matter

a (%)





b (%)





c (%/h)





PD (%)





ED (%)





a (%)





b (%)





c (%/h)





PD (a + b)





ED (%)




a,bWithin rows means with similar superscripts do not differ (P > 0.01). 1For dry matter; a, b, PD, and ED are measured as % of DM incubated, while for Nitrogen a, b, PD and ED are measured as % of N incubated.

In vivo evaluation 

The chemical composition of the three diets offered to dairy weaners in feedlots is presented in Table 3. Inclusion of MSC appears to increase NDF, ADF and EE content of the diets. The 3 diets had similar levels of N as per design of the experiment (iso-nitrogenous) as well as similar ash content. The average initial live weight of the weaners was 186 ▒ 16.9 kg. Total feed intake for the entire experimental period was 367, 397 and 410 kg /animal for animals offered CFR-U, CFR-MSC, and CFR-MSC+U, respectively, however these intake values did not differ (P > 0.05). The average daily weight gain per animal was 1.62 kg, 1.75 kg, and 1.82 kg in animals offered CFR-U, CFR-MSC, and CFR-MSC+U, respectively but these were not significantly different (P > 0.05). 

Table 3. Chemical composition (g/kg DM) of experimental fattening diets





Dry matter




Crude protein








Ether extract




Neutral detergent fibre




Acid detergent fibre




1CFR-U - commercial fattening ration (contains urea as a main N source).

2CRF-SC+U - commercial fattening ration in which urea and marula seed cake contribute equal amounts of N.

3CFR-MSC - commercial fattening ration in which urea was wholly substituted with marula seed cake.

Likewise, partial or whole substitution of urea N in the CFR-U diet did not affect (P > 0.05) feed conversion efficiency whose values were 0.263, 0.255 and 0.258 for animals offered CFR-U, CFR-MSC, and CFR-MSC+U, respectively.

Table 4. Performance of feedlot cattle offered experimental fattening rations






Fortnightly live weight (kg)





Day 14





Day 28





Day 45





Day 58





Weight gain (kg) after 58 days





Average weight gain (kg/day)





Total feed intake (kg)





Feed conversion efficiency





aIn a row means with the same superscripts do not differ (P > 0.05).

1CFR-U - commercial fattening ration (contains urea as a main N source).

2CFR-SC+U - commercial fattening ration in which urea and marula seed cake contribute equal amounts of N.

3CFR-MSC - commercial fattening ration in which urea was wholly substituted with marula seed cake.

Although statistically insignificant (P > 0.05), observed measurements show that the CFR-MSC and CFR-MSC+U diets promoted slightly higher average live weight gains and total feed intake compared to the CFR-U diet.  


Chemical composition of marula seed cake 

As far as we know, this is the first study to characterize MSC in terms of chemical composition, in sacco degradability and in vivo utilization by ruminant animals. Being an oil seed residue, MSC is rich in crude protein content (470 g/kg DM) (Table 1). A related report on CP content of marula products was authored by Aganga and Mosase (2001) who reported CP content to be 61.7 g/kg DM for the whole nut without decortications. Clearly, this CP value is of not much use to domesticated ruminant animals that lack the digestive ability to breakdown the whole marula nuts. The low CP value obtained by Aganga and Mosase (2001) study is to be expected due to the dilution of protein concentration by oil and the hard shell of the kernel. Thus MSC has higher CP content compared to groundnut meal, sunflower cake, soybean meal, and cotton seed cake suggesting it can be used to supply much needed protein in resource-poor farming systems. Brand and van der Merwe (1993) found cotton seed cake to have CP content of 380 g/kg DM while McDonald et al (2002) report that the CP content of groundnut meal vary between 250 and 300 g/kg DM. Crude protein concentration in MSC will vary depending on the oil extraction techniques employed. Those techniques efficient enough to extract the highest proportion of oil produce MSC with higher CP concentration. The oil extraction process at Mpaka may need to be improved in light of MSC lipid content results reported here. A residual lipid content of 38% is probably too high and points to inefficiency in the oil extraction techniques employed by the company. However, the high lipid content of MSC may indicate potential for supplying additional energy in fattening rations. On the other hand high lipid content can negatively impact on rumen function by reducing the hydrophilicity of feed particles thus inhibiting microbial adhesion and digestibility of diet cell wall components. Inclusion of MSC altered the total EE content of the commercial fattening rations. As expected, addition of MSC in the commercial feed ration increased both aNDFom and ADFom content (Table 3). The fibre content assayed for in MSC is mostly made up of remnants of the hard shells surrounding the kernel that is crushed during the decortication process. 

In sacco degradability of marula seed cake 

In the absence of fistulated cattle to use in the in sacco experiment, the authors felt that using fistulated goats would assist in the characterization of marula seed protein and thus provide applicable background data required to explain the performance of feedlot cattle offered this protein resource. Despite the nylon bag technique’s usefulness in evaluating protein quality for ruminant animals, its accuracy is affected by many factors such as sample size, bag size, pore size and sample particle size (Kempton 1980). Perhaps with the exception of particle size, the other factors can easily be standardized. For non-conventional feedstuffs, it is therefore, prudent to measure degradability kinetics at different particle sizes. Milling MSC significantly increased the immediately degradable fraction (a) of DM and N. At the same time, the proportion of the slowly degradable DM and N fractions declined due to milling. This confirms that finer particle size had an effect on DM and N in sacco degradability. There are two possible explanations for this observation. Firstly, milling feed samples increases the surface area by reducing particle size and thus allows rapid microbial invasion and enhances degradation. This would cause an increase in the immediately degradable DM and N fractions as observed. Indeed, Bengochea et al (2005) reported that decreasing screen size increases the surface area exposed resulting in higher degradability values. Secondly, the reduction in particle size means the finer particles can escape through the nylon bag’s pores without necessarily being degraded. This can lead to erroneous degradability estimates. Bengochea et al (2005) found that DM disappearance rate increased with decreased particle size of barley. This has implications on the particle size of substrates evaluated using the nylon bag technique. The closer the particle size is to the chewed product, the more useful will be the estimates obtained from the nylon bag. In this regard, it is likely the MSC will be offered to animals in its original form to feedlot cattle to reduce milling costs thus the degradability values observed for unmilled MSC will be more applicable in practical feeding. With low values for immediately degradable fraction (Table 2), MSC does not readily supply soluble N for rumen microbes thus will be more useful if offered in conjunction with feedstuffs that supply soluble N for rumen microbial function. With 85.6% slowly degradable DM, unmilled MSC is a slowly degradable feedstuff possibly due to the presence of appreciable levels of residual lipids. Indeed, Lindberg (1985) concluded that lipid content had a negative effect on degradability of feed by reducing hydrophilicity of feed particles thus preventing microbial adhesion and digestion. Excessive lipids can have a profound negative effect on rumen microbial balance through the suppression of methane bacteria (Hills and West 1991).  

Feedlot trial 

The objective for substituting urea with MSC was to investigate the effect of MSC as a N source on diet intake and growth performance of feedlot animals. There were no statistical differences in the growth performance between treatments hence inclusion of MSC in the commercial fattening ration did not alter the performance of feedlot cattle. These results are in agreement with Brand and van de Merwe (1993) who found no significant difference in performance of lambs, which received iso-nitrogenous diets with cotton seed cake, sunflower seed, sunflower seed cake, soybean meal or urea as source of N in grain sorghum basal diet. At an average daily gain of 1.75 kg, MSC supported a slightly high growth rate comparable to that of urea (1.62 kg) in commercial fattening ration. However, the CFR-MSC+U diet promoted an average daily weight gain of 1.82 which was slightly higher than observations in the other 2 diets. The combination of these two different sources of N also seems to have boosted total feed intake (410 kg vs 397 and 367, for CFR-MSC and CFR-U, respectively). It should be recalled, from the in sacco degradability study, that MSC supplies very little immediately soluble N thus MSC is likely to escape rumen fermentation. The synergy between MSC and urea is expected since the soluble N from urea provides rumen microbes with sufficient N for efficient capture of energy from the fattening ration. This allows the MSC protein to be used with higher efficiency postruminally. Thus energy-protein synchronization is likely to have been optimum in the rumen of animals offered CFR-MSC+U diet than in those offered CFR-U and CFR-MSC diets. 

The non significant increase in intake for CFR-MSC and CFR-MSC+U diets showed that inclusion of MSC may actually improve the palatability of the fattening ration. In addition, Donaldson et al (1991) found increased intake of diets supplemented with protein sources that supply by-pass protein to steers. Due to the presence of a high proportion of slowly degradable N fraction, MSC is a good source of by-pass protein to ruminant animals and thus is also likely to boost feed intake. Unlike urea, MSC also provides energy in addition to rumen degradable N and by-pass protein which might have a positive effect on finishing weight depending on level of lipids consumed. The use of urea in fattening diets is preferred to other N sources, such as soybean meal, primarily due to cost considerations. However, supplementing ruminants with rumen undegradable protein can increase the flow of N and amino acid to the small intestine and results in improved growth and efficiency of N utilization. Feeding ruminal undegradable protein can decrease the efficiency of microbial protein synthesis and the flow of amino acids to the small intestine (Siddons et al 1985; Cecava et al 1991) compared with more degradable protein sources. Compared to urea, marula seed cake provides both rumen degradable N and rumen undegradable N thus the efficiency of rumen microbial protein synthesis is enhanced. However, offering MSC together with feedstuffs that supply immediately degradable N such as urea may improve animal performance. Indeed, the performance of cattle offered MSC together with urea seem to corroborate this view. 


Marula seed cake is rich in protein but most of it is slowly degradable in the rumen. Wholly substituting urea with MSC as a source of N in fattening rations had no negative effect on voluntary feed intake and growth rate of feedlot cattle. It is concluded that MSC can be used in place of urea in the formulation of fattening rations for the benefit of farmers in resource-poor farming systems. Marula seed cake has potential to be used as a protein supplement for other classes of ruminant livestock to improve their performance during the long dry seasons of Swaziland.  


The authors are grateful to Swaziland Water and Agricultural Development Enterprise (SWADE) for providing funds to support this study. The views expressed are not necessarily those of SWADE. 

Literature cited

Aganga A A and Mosase K W 2001 Tannin content, nutritive value and dry matter digestibility of Lonchocarpus capassa, Zizyphus mucronata, Sclerocarya birrea, Kirkia acuminata and Rhus lancea seeds.  Animal Feed Science and Technology 91: 107 – 113.


AOAC 1999 Official Methods of Analysis of AOAC International. 16th Edition. Association of Official Analytical Chemists, Arlington, VA, USA.  


Bengochea W L, Lardy G P, Bauer M L and Soto-Navarro S A 2005 Effect of grain processing degree on intake, digestion, ruminal fermentation, and performance characteristics of steers fed medium-concentrate growing diets. Journal of Animal Science 83: 2815 – 2825.  


Brand T S and van de Merwe G D 1993 Comparison of different protein source in enriched grain mixture for fattening lambs. South African Journal of Animal Science 23:13 – 17.  


Cecava M J, Merchen N R, Berger L L and Fahey G C 1991 Effect of dietary energy level and protein source on site of digestion and duodenal nitrogen and amino acid flow in steers. Journal of Animal Science 66: 961 – 974.  


Chen X B 1997 A utility for processing data of feed degradability and in vitro gas production. NEWAY Excel Version 5.0 XBC Laboratory, Rowett Research Institute, Aberdeen, Scotland. 


Donaldson R A, McCann M A, Amos H E and Hoveld C S 1991 Protein and fibre digestion by steers grazing winter and supplemented with ruminal escape protein. Journal of Animal Science 69: 3067 – 3071. 


Hills G M and West J W 1991 Rumen protected fat Kline barley or corn diets for beef cattle. Journal of Animal Science 69: 3376 – 3388. 


Kempton T J 1980 The use of nylon bags to characterize the potential degradability of feeds for ruminants. Tropical Animal Production 5: 107 – 116. 


Lindberg J E 1985 Estimation of rumen degradability of feed proteins with the in sacco and various in vivo methods: A Review. Acta Agriculturae Scandanavica Supplementum 25: 64 – 94. 


McDonald P, Edwards R A, Greenhalgh J D F and Morgan C A 2002 Animal Nutrition. Prentice Hall, London, United Kingdom. 


Ministry of Agriculture and Co-operatives 2004 Livestock Census, Annual Report, Mbabane, Swaziland. 


Monadjem A and David K G 2005 Nesting distribution of vultures in relation to land use in Swaziland. Biodiversity and Conservation 14: 2079–2093. 


ěrskov E R and McDonald J 1979 The estimation of protein degradability in the rumen from incubation measurements weighted according to rate of passage. Journal of Agricultural Science (Cambridge) 92: 499 – 503. 


SAS  2001 User’s guide: Statistics, Version 6.12. SAS Institute, Inc. Cary, USA. 


Shackleton C M  2004 Use and selection of Sclerocarya birrea (Marula) in the Bushbuckridge lowveld, South Africa. In: Rao M.R. and Kwesiga F.R. (eds.). Proceedings of Regional Agroforestry Conference on Agroforestry Impacts on Livelihooods in Southern Africa: Putting Research into Practice, World Agroforestry Centre, Nairobi, Kenya. pp. 77– 92. 


Siddons R C, Paradine J, Gale D L and Evans R T 1985 Estimation of the degradability of dietary protein in the sheep rumen by in vivo procedures.  British Journal of Nutrition 54: 509 – 520. 


Van Soest P J, Robertson J B and Lewis B A 1991 Methods of dietary fibre, neutral detergent fibre and non-starch polysaccharides in relation to animal nutrition. Journal of Dairy Science 74: 3583 – 3597.

Received 2 November 2009; Accepted 11 August 2010; Published 1 May 2011

Go to top