Dry matter intake, apparent digestibility and excretion of purine
derivatives in sheep fed tropical legume hay
J.F. Mupangwa
a,*, N.T. Ngongoni
b, J.H. Topps
c, T. Acamovic
d,
H. Hamudikuwanda
b, L.R. Ndlovu
baDepartment of Agritex, Ministry of Agriculture, P.O. Box CY 639, Causeway, Harare, Zimbabwe bDepartment of Animal Science, University of Zimbabwe, P.O. Box MP 167, Mt. Pleasant, Harare, Zimbabwe
cDepartment of Agriculture, University of Aberdeen, 581 King Street, Aberdeen, AB24 5UA, Scotland, UK dDepartment of Biochemistry and Nutrition, Scottish Agricultural College, Auchincruive, Ayr, KA6 5HW, Scotland, UK
Accepted 27 September 1999
Abstract
Four ruminally cannulated wethers (311.3 kg) were used in an experiment with a 44 Latin square design to estimate the DM intake, apparent digestibility, nitrogen balance, rumen ammonia and microbial protein production. The sheep had ad libitum access to eitherCassia rotundifolia(Cassia),Lablab purpureus(Lablab),Macroptilium atropurpureum(Siratro) orStylosanthes guianensis(Stylo). Dry matter intake of cassia was lower (P< 0.001) than that of lablab, siratro and stylo hays. Organic matter intake was greater (P< 0.001) for lablab, siratro and stylo hays than that of cassia. Dry matter digestibility was higher (P< 0.05) for lablab hay, than that of cassia, siratro and stylo hays. The organic matter digestibility ranged from 0.579 for cassia hay to 0.617 for stylo hay and there were no differences (P> 0.05) among the legume hays. Nitrogen intake was highest (P< 0.05) in sheep given stylo hay and least in sheep fed cassia hay. Animals given lablab, siratro and stylo hays had higher (P< 0.05) faecal and urinary N compared to those on cassia hay. Rumen ammonia N concentration was highest (P< 0.05) in sheep given lablab while sheep offered siratro and stylo had intermediate values, and least in animals fed cassia hay. The ammonia levels were above the recommended optimal level of 50 mg N/l. The total purine derivative excretion in the urine and microbial N supply was not different (P> 0.05) among treatments. From the presented ®ndings it is concluded that the intake and digestibility in sheep of the four legume hays are variable and provide adequate rumen ammonia N for maximum rumen microbial growth making then ideal protein supplements to ruminants fed low quality roughages.#2000 Elsevier Science B.V. All rights reserved.
Keywords:Legumes; DM intake; Digestibility; Purine derivatives
1. Introduction
Tropical forages are low in protein and have high cell wall contents resulting in low digestibilities
(Leng, 1990). Supplementation of tropical grasses with legumes has been reported to result in increased DM intake and to improve DM (Ndlovu and
Bucha-nan-Smith, 1985). Herbaceous tropical forage
legumes such as Cassia rotundifolia(Cassia),Lablab
purpureus (Lablab), Macroptilium atropurpureum
(Siratro) and Stylosanthes guianensis (Stylo) have
*Corresponding author.
E-mail address: agriani@africaonline.co.zw (J.F. Mupangwa)
great potential as protein supplements to low quality roughages (Topps and Oliver, 1993; D'Mello and Devendra, 1995). Studies on DM intake and digest-ibility of tropical legumes offered as sole diets are few (Vera et al., 1989). Very little is known about cassia, lablab, siratro and stylo regarding their intakes by ruminants and the rumen microbial protein production and levels of animal response they can elicit when given to ruminants as sole diets.
The objective of this experiment was to investigate the effect of feeding four legume hays as sole diets to mature sheep on DM intake, apparent digestibility, rumen ammonia levels, urinary excretion of purine derivatives and rumen microbial protein production.
2. Materials and methods
2.1. Animals and diets
Four mature Dorper wethers (311.3 kg) were
used in this experiment. Each animal was surgically ®tted with a rubber rumen cannula (3 cm internal diameter; Piggot and Maskew, Bulawayo, Zimbabwe). The animals were housed in individual metabolism
crates, measuring 0.60.751.0 m and raised
0.5 m above the ¯oor in the Bioassay Laboratory, Department of Animal Science at the University of Zimbabwe. Each crate was equipped with a feed trough and water was available all the time.The legumes were grown each in rows 0.45 m apart in
plots measuring 1550 m in the Marirangwe Small
Scale Commercial farming area near Harare,
Zim-babwe, on sandy soils (pH 5.5 on CaCl2 scale),
receiving a rainfall of 750±1000 mm per year. The legumes were harvested at 20 weeks after germination and sun dried in the ®eld. During sun drying the hay was turned twice a day for 4 days to ensure even drying. Four diets comprising of hays only of cassia, lablab, siratro and stylo were used in the experiment. The animals were randomly allocated to the four
dietary treatments in a 44 Latin square design.
2.2. DM intake measurement
Before each experimental period lasted 21 days, the animals were weighed and then adapted to the diets for 14 d followed by a 7-day of total collection period.
Animals had ad libitum access to legume hays. The daily allocation of hay was offered in two equal proportions at 08:00 and 16:00 h. A mineral and vitamin mix (Hamish Cameron, Harare, Zimbabwe) was added to the diet to ensure an adequate supply of all minerals and vitamins. Water was always available. Feed refusals were collected, weighed every morning and bulked per treatment in each period before being sub-sampled for laboratory analysis. Faeces were collected daily, weighed and stored frozen while a representative sub-sample was taken for dry matter determination. Urine was collected daily in plastic buckets that had 10 ml of concentrated sulphuric acid as a preservative (Chen and Gomez, 1992) to maintain a ®nal pH of 3 or below. A 10% sample of urine was diluted by a factor of ®ve with distilled water to prevent precipitation of purine derivatives (PD) and
was stored at ÿ208C for analysis of PD. A second
sample of urine was collected and kept frozen pending total N analysis.
2.3. Rumen ammonia concentration
On the last day of each measurement period, 200± 250 ml of rumen liquor were collected through the rumen cannula by a suction pump. The rumen liquor samples were taken before feeding at 08:00 h and at 09:00, 11:00, 13:00, 15:00 and 17:00 h post-feeding. Rumen liquor was strained through two layers of cheese cloth and 200 ml acidi®ed with 1 ml of 25%
sulphuric acid and stored at ÿ208C for ammonia
analysis.
3. Laboratory analysis
furnace at 5508C for 24 h. Urine was analysed for N (AOAC, 1984) and purine derivatives according to the method of Chen and Gomez (1992).
4. Calculations and statistical analysis
From the PD excreted, the corresponding amount of microbial purines (P mmol/day) absorbed by the animal was estimated using the model described by Chen and Gomez (1992). The supply of microbial N was then calculated from P using the following fac-tors: digestibility of microbial purines 0.83 and pur-ine-N:total microbial N ratio of 0.116 : 1.00 (Chen and Gomez, 1992):
MicrobialNsupply g=day P70
0:830:1161000
0:727P
The model (1) for DM, OM, N intakes and microbial protein production tested the effects of period, animal and legume type. The model (2) for rumen ammonia concentration tested the effects of period, animal, sampling time and legume type. Animal and period were used as blocks. The models used for the analysis were:
YijkmLiPjAkeijk (1)
YijklmLiPjAkT1eijkl (2)
Where: Yijk or Yijkl is the independent variable (e.g.
DM intake or ammonia concentration, respectively);
Li(i1, 2, 3,4) is the ®xed effect of legume species,
Pj(j1, 2, 3,4) is the effect of period; Akis the effect
of the animal (k1, 2, 3, 4), Tl is the effect of
sampling time (l1, 2, 3, 4, 5, 6) and eijk is the
random error. The differences between treatment means were assessed by the Tukey Studentized Range test (SAS, 1990).
5. Results
5.1. Intake and apparent digestibility
The chemical composition of the four legume hays used in the experiment is shown in Table 1. Among the hays, siratro and stylo were higher in CP, NDF and ADF than cassia and lablab. The intake and apparent digestibility of the legume hays are given in Table 2.
Table 1
Chemical composition (g/kg DM) of forage legume hays
Cassia Lablab Siratro Stylo
DM 956 918 930 924
OM 889 841 800 860
CP 182 162 229 253
NDF 419 473 512 593
ADF 275 294 355 416
ADL 75.7 64.2 72.3 116
NDIN (g/kgN) 91.6 146 139 246
ADIN (g/kg N) 72.5 48.7 59.5 102
Ca 16.8 20.8 22.2 22.1
P 1.40 1.10 1.30 1.30
Condensed tannin 29.5 16.9 12.4 15.6
Table 2
The intake and apparent digestibility of legume hays given to sheep as sole diets
Cassia Lablab Siratro Stylo SED
DM intake
g/kgW0.75 12.1b 48.2a 52.6a 50.9a 2.94
Organic matter intake
g/kgW0.75 10.9b 44.1a 43.3a 47.8a 7.11
DOM intake (g/day) 82.9b 377a 329a 384a 30.8
Apparent Digestibility
DM 0.550b 0.638a 0.581ab 0.577ab 0.034
OM 0.579 0.653 0.588 0.617 0.037
a,bMeans in the same row with different letters differ (P< 0.001).
The DM and organic matter (OM) intakes of lablab,
siratro and stylo hays were greater (P< 0.001) than
that of cassia hay. The digestible organic matter intake (DOMI) followed a trend similar to that of OM intake
but OM apparent digestibility was not (P> 0.05)
different among the four hays. There were signi®cant differences in apparent digestibility of DM. Lablab
hay had a higher (P< 0.05) DM digestibility than
cassia hay. Siratro and stylo had DM digestibilities that were not statistically different from either cassia or lablab.
5.2. Nitrogen balance and purine derivative excretion
Table 3 shows the results of nitrogen balance, daily excretion of purine derivatives and microbial N supply measurements. Nitrogen intake was highest
(P< 0.001) in animals given stylo hay followed by
animals given siratro hay then lablab hay and least for cassia hay. Although there were signi®cant differences in N intake between animals given lablab, siratro and stylo diets, the faecal nitrogen outputs were similar
between the three diets and higher (P< 0.001) than
that of animals on cassia hay. As a percentage of total
N intake, faecal N excretion was greater (P< 0.05) in
animals offered cassia (26%) and lablab (29%) hays than in those given siratro (19%) and stylo (15%) hays.Animals given lablab, siratro and stylo hays had a
greater (P< 0.01) urinary nitrogen excretion than
those on cassia hay. Urinary N excretion as a multiple
of N intake was greater (P< 0.05) in animals offered
cassia (2.55), lablab (1.41) and siratro (1.15) hays compared to those on stylo hay which had a value of 0.863. Animals on all treatments exhibited negative nitrogen balance with those on lablab hay having a
greater (P< 0.05) negative nitrogen balance than
sheep on the other treatments (Table 3). Apparent N
digestibility of stylo hay was higher (P< 0.05) than
that of the other hays and the difference reached signi®cance with cassia and lablab hays but not with siratro hay.
Rumen ammonia-N concentration was highest
(P< 0.01) in rumen of sheep given lablab,
intermedi-ate (P< 0.05) in sheep offered siratro and stylo hays
and least in animals on cassia hay. The daily PD excretion, the estimated purine absorption and the calculated microbial N supply (g/d) were higher in animals which had high DOMI but there was no
signi®cant (P> 0.05) difference between the four
Table 3
N balance (g/day), rumen ammonia (mg N/l), excretion of purine derivatives (mmol/day) and microbial nitrogen supply in sheep given forage legume hays as sole dietsa
Cassia Lablab Siratro Stylo SED
N intake 5.75d 16.5c 25.1b 27.7a 0.83
Faecal N 1.08b 4.77a 4.81a 4.20a 0.55
Urinary N 10.2b 23.1a 29.1a 23.9a 3.54
N retention ÿ5.53ab ÿ11.4b ÿ8.68ab ÿ0.53a 4.04
Apparent N digestibility 0.735bc 0.706c 0.809ab 0.848a 0.05
Rumen NH3-N 150b 225a 159b 161b 15.9
Purine derivatives:
Allantoin 2.89 3.29 3.58 4.73 1.16
Uric acid 0.04 0.12 0.07 0.25 0.13
Xanthine plus Hypoxanthine 0.23 0.29 0.33 0.24 0.06
Total PD 3.12 3.64 3.96 5.23 1.14
Absorbed purine 3.71 4.33 4.72 6.22 1.36
Microbial N supply
g/day 2.70 3.15 3.43 4.53 0.98
g/kg DOMRe 50.1a 12.9b 16.0b 18.1b 8.69
a,b,c,dValues in the same row with different superscripts differ (P< 0.05).
SED±Standard error of difference.
legume hay diets. When expressed on the basis of organic matter apparently digested in the rumen
(DOMR0.65DOMI), the calculated microbial
N supply ranged from 12.9 to 50.1 g N/kg DOMR with no signi®cant differences between lablab, siratro
and stylo hay diets all of which gave lower (p< 0.001)
values than that for cassia hay diet.
6. Discussion
Siratro and stylo hays used in this experiment were higher in CP and had lower NDF and ADF content compared with the legume hays used by Chikumba (1990). He reported CP content of 159 and 114 g/kg DM, NDF values of 640 and 630 g/kg DM and ADF content of 550 and 540 g/kg DM for siratro and stylo, respectively. Bengaly (1996) reported CP content of 122 g/kg DM for lablab hay. The NDF and ADF content of lablab hay used in this study were lower than values reported by Bengaly (1996). The cassia hay had higher CP and lower ADF content compared to the cassia hay used by Ahn et al. (1988) in their study. The difference between the hays used in this study and those reported in other studies could be due to variation in leaf content of the hays and stage of growth at which the hay was harvested. In this study the legumes were harvested at 20 weeks of growth and ®eld cured which could have resulted in some loss of leaf material from the hay.
The DM intakes of lablab, siratro and stylo hays by sheep were higher than for cassia hay. This difference could be attributed to higher apparent DM digestibility which may have resulted in lower rumen retention time and greater turnover rate of particulate matter from the rumen of sheep offered the three legume hays compared to cassia hay. Cassia has been reported to have a relatively low acceptability compared to that of siratro in earlier studies (Clements, 1989). The low intake of cassia in this study could be due to the presence of anti-nutritive factors such as condensed tannins. The hays used in this study had total con-densed tannin content of 29.5 for cassia, 16.9 for lablab, 12.4 for siratro and 15.6 g/kg DM for stylo. Condensed tannins are reported to reduce voluntary intake through astringency, an unpleasant puckering sensation in the mouth brought about by complexing of tannins with salivary glycoproteins (Kumar and
D'Mello, 1995). The extractable CT content of the cassia hay was lower than values of 50±100 g/kg DM (Barry and Duncan, 1984) and 65±90 g/kg DM (Barry and Manley, 1986) reported to cause a reduction in voluntary intake in sheep. With the reduced cassia hay intake a reduced ruminal turnover and rate of digestion are likely as suggested by Manyuchi, 1994. Any reduction in ruminal turnover and rate of digestion may be caused by an inhibition in microbial activity (Salawu, 1997) and inhibition of microbial enzymes (Muhammed et al., 1994) brought about by tannins in the feed.
The intake of lablab hay of 0.624 kg/day was lower than intake levels of 0.91 and 1.57 kg/day of leaf and
stem fractions, respectively, forL. purpureuscv
Ron-gai reported by Hendricksen et al. (1981) with sheep of 47 kg live mass. However, when expressed per metabolic body mass, the intakes of lablab from this study are similar to those obtained by Hendricksen et al. (1981). The intake of cassia hay was lower than values reported in other studies (Ahn et al., 1988) of 1.09 kg/day using sheep of similar weight to those used in this study. The intakes of siratro and stylo hays used in this study were lower than values reported for similar hays made from younger material (Wanapat, 1987). The intakes of lablab, siratro and stylo were similar to that of Neonotonia wightii hay of 50.2
g/kgW0.75/day reported by Vera et al. (1989) and
Minson (1977) reported a general, relationship between intake and digestibility of a range of tropical legumes to be as;
DM intake g=kg W0:75=day 1:76 DM digestibility
ÿ44:5 r0:86
Using this relationship, the potential intake of the legume hays used in this study would be 52.3, 68.1,
58.0 and 57.7 g/kg W0.75/day for cassia, lablab, siratro
and stylo hays. The values obtained in this study are below these intakes. Also the intakes of the hays fell
below the standard intake of 80 g/kgW0.75/day
sug-gested by Crampton et al. (1960) for forages with 0.70 digestibility possibly due to the effect of condensed tannins.
The apparent DM digestibility coef®cients of the
four legume hay diets are comparable to that of N.
wightiihay which had an apparent DM digestibility of
digestibility than that reported in earlier studies of 0.504±0.535 (Skerman et al., 1988) while that of stylo fell within the range of reported values (Wanapat, 1987; Skerman et al., 1988). The DM digestibility of cassia , 0.596, is consistent with the result of 0.555 reported by Ahn et al. (1988) and by Ahn, 1990 as cited by Norton and Poppi, (1995) of 0.555±0.640. The DM digestibility of lablab, 0.638, is higher than values reported in previous studies (Hendricksen et al., 1981; Skerman et al., 1988). Hendricksen et al. (1981) reported DM digestibilities of 0.558 and 0.495,
respectively, for leaf and stem fractions ofL.
purpur-euscv Rongai. Similarly apparent OM digestibility of
lablab was higher than reported in the literature (Hen-dricksen et al., 1981). The lack of signi®cant differ-ences between cassia and the other hays in OM digestibility may indicate that the high condensed tannin content in cassia did not depress microbial activity in the rumen and therefore its observed lower intake could be due to an astringency effect in the mouth.
According to the metabolisable energy (ME MJ/ day) requirements published by the MAFF (1984) the maintenance requirements for sheep used in this study, 30 kg live mass, is 5.1 MJ/day. Calculations based on the AFRC (1993) equation (ME intake
MJ/day0.0157DOMI), the metabolisable energy
intakes of the sheep given the legume hays were, respectively, 1.30, 5.92, 5.17 and 6.03 MJ/day for animals on the cassia, lablab, siratro and stylo hay diets. Animals on the cassia hay diet were clearly in negative energy balance and needed to meet the de®cit by mobilising body reserves. Animals on the other treatments had suf®ciently high intakes to meet ME requirements for maintenance and some gain but showed negative N balance. This observation can be attributed to the type of protein and its degrad-ability and also possibly due to lack of synchrony between N release through protein degradation and energy availability as reported by Stern et al. (1994). This results in reduced ef®ciency of rumen ammonia utilisation by the microbes leading to excess ammonia being absorbed into the blood stream and lost in urine as urea (Robinson, 1997) which is consistent with the increased urinary N excretions by animals on these diets.
The rumen ammonia N concentrations were above the recommended optimal level of 50 mg N/l (Satter
and Slyter, 1974) for maximum microbial growth. Animals in this study given cassia hay although having lower intakes than those reported by Ahn et al. (1988) had similar rumen ammonia levels of 150 mg N/l to those reported by Ahn et al. (1988) which had rumen ammonia levels of 140 mg N/l. Although lablab hay had a lower CP content than the other legume hays, it resulted in the highest rumen ammonia concentration indicating that its protein was more degradable than that of the other legumes. The ®ndings from this study are in agreement with those of Higgins et al. (1992) and of Ahn et al. (1988) who reported rumen ammonia levels of 307 and 140 mg N/l in rumen of sheep given
Aeschynomene americanaandC. rotundifoliaas sole
diets. The high rumen ammonia concentrations in the absence of readily fermentable energy source, which is likely with these diets, could result in energy limit-ing microbial growth and a signi®cant loss of legume protein N in net transfer to the small intestines. Losses
of up to 43% ofA. americanaprotein N in the rumen
have been reported and were found to be associated with a decline in non-ammonia-nitrogen ¯ow to the intestines (Higgins et al., 1992). Results from this study indicate that similar losses might occur when the legume hays are used as sole diets resulting in increased urinary N excretion and reduced N retention by the animals.
The PD excretions indicate that microbial protein supply increased when dietary protein intake was increased from 5.75 to 27.7 g/day for sheep on cassia and siratro hays. This was probably due to an increased intake of degradable N (Kolade and Ter-nouth, 1996) and more fermentable organic matter being made available for microbial fermentation in the rumen of sheep on the other hay diets compared to those on the cassia hay diet. Increased OM intake and of OM apparently digested in the rumen have been reported to result in an increase in microbial produc-tion (Clark et al, 1992). The increased microbial production can be attributed at least partially to the larger amount of energy supplied by the larger quan-tity of OM fermented in the rumen (Chen et al., 1992; Clark et al., 1992).
respectively, were obtained. This result was unex-pected and could indicate a variable contribution of endogenous tissue nucleic acids. Animals on this diet had on average total PD excretion of 0.49 mmol/kg
W0.75/day which was lower than the value of
0.6 mmol/kg W0.75/day below which large net
endo-genous contribution are evident (Dewhurst and Web-ster, 1992). Since DM and N intake were low on the cassia hay diet, animals on this treatment would have been mobilising tissue protein, as shown by a large negative N retention, to meet their energy and protein requirements. This may have resulted in an increased endogenous purine contribution in the urine and an overestimation of rumen microbial protein produc-tion. In view of the likelihood of an increased endo-genous nucleic acid contribution in animals given cassia hay, the derivation of microbial protein produc-tion based on purine derivatives described by Chen and Gomez (1992) has to include endogenous con-tributions especially in animals given similar tropical feeds.
The microbial N supply was lower while ef®ciency of microbial N production was higher from cassia hay, 2.70 g/day and 47.9 g/kg DOMR, than the values of 12.0 g/day and 40 g/kg DOMR reported by Ahn et al. (1988) and AFRC (1993) has reviewed data on the ef®ciency of microbial protein production of different feeds. The values for the ef®ciency vary widely ran-ging from 14 to 49 g/kg DOMR. The legume hays other than cassia gave an ef®ciency at the lower end of the range and this could be due to lack of suf®cient energy to support microbial protein synthesis. How-ever, caution should be used when microbial growth is expressed on an ef®ciency basis because interpretation of the data could be much different when total micro-bial N supply to the duodenum is evaluated (Stern et al., 1994).
Using the ARC (1984) relationship between ME intake (MJ/day) and microbial N production
(Micro-bial N (g/day)ME intake1.34), the calculated
microbial N supply (g/day), from the diets was 1.74, 7.93, 6.93 and 8.08 g/day for cassia, lablab, siratro and stylo hays, respectively. The values obtained in this study for lablab, siratro and stylo hays are lower, while that for cassia hay is higher than these predicted values. The true amino acid N in microbial N is 80% since about 15% consists of nucleic acid N and 5% is composed of other non-protein nitrogen
forms (érskov and Miller, 1988). Microbial protein has a digestibility value of 0.85 (AFRC, 1993), so the microbial protein digested and absorbed from the intestines was 11.5, 13.4, 14.6 and 19.3 g/day for cassia, lablab, siratro and stylo diets, respectively. Based on these calculations, lablab, siratro and stylo contribute greater amounts of microbial amino acids for tissue protein synthesis than does cassia.
7. Conclusions
This study demonstrates that an extensive loss of legume protein as ammonia in the rumen occurs in the absence of a readily fermentable energy source, which can result in a reduction of undegraded dietary protein ¯owing to post-ruminal sites. The ef®ciency of microbial production from rations based on forage legumes can be limited by a lack of readily available energy source and the addition of such energy sources may assist in giving increased ammo-nia utilisation and microbial protein production. The legumes can be fed as protein supplements to rumi-nants consuming low quality tropical grasses and crop residues.
Acknowledgements
The authors are indebted to the European Union (Project Number ERBTS*CT930211) for funding this research. We also wish to thank the laboratory staff of the Department of Animal Science, University of Zimbabwe for their technical assistance.
References
A.O.A.C., 1984. Association of Of®cial Analytical Chemists. Of®cial Methods of Analysis, AOAC, Washington, DC, USA. AFRC Agricultural Food and Research Council, 1993. Energy and
Protein Requirements Of Ruminants. An advisory manual prepared by the AFRC Technical Committee on Responses to Nutrients. CAB International, Wallingford, UK.
Ahn, J.H., Elliot, R., Minson, D.J., 1988. Quality assessment of the fodder legumeCassia rotundifolia. Tropical Grasslands 22 (2), 63±67.
Barry, T.N., Manley, T.R., 1986. Interrelationships between the concentration of total condensed tannin, free condensed tannin and lignin in lotus species and their possible consequences in ruminant nutrition. J. Sci. Food and Agric. 37, 248±254. Bengaly, K., 1996. The effect on intake and digestion of maize
stover when supplemented with urea and/or lablab (Lablab purpureus) hay and given to native cattle in southern Mali. M.Sc. Thesis, University of Aberdeen.
Chen, X.B., Gomez, M.J., 1992. Estimation of microbial protein supply to sheep and cattle based on urinary excretion of purine derivatives an overview of the technical details. International Feed Resource Unit, Rowett Research Institute. Occasional Publication, Aberdeen, pp. 2±20.
Chen, X.B., Abdulrazak, S.A., Shand, W.J., érskov, E.R., 1992. The effect of supplementing straw with barley or unmolassed sugar-beet pulp on microbial protein supply in sheep estimated from urinary purine derivative excretion. Anim. Prod. 55, 413± 417.
Chikumba, N., 1990. Rumen fermentation patterns in sheep given pure diets of Stylosanthes guianensis cv Oxley (Fine Stem Stylo), Macroptilium atropurpureumcv Siratro (Siratro) and Desmodium uncinatum(Silverleaf Desmodium). M.Sc. Thesis, University of Zimbabwe.
Clark, J.H., Klusmeyer, T.H., Cameron, M.R., 1992. Microbial protein synthesis and ¯ows of nitrogen fractions to the duodenum of dairy cows. J. Dairy Sci. 75, 2304.
Clements, R.J., 1989. Rates of destruction of growing points of pasture legumes by grazing cattle. Proc. XVI Int. Grassl. Congr., Nice, France, pp. 1027±1028.
Crampton, E.W., Donefer, E., Lloyd, L.E., 1960. A nutritive value index for forages. J. Anim. Sci. 19, 538±544.
D'Mello, J.P.F., Devendra, C., 1995. (Eds.), Tropical Legumes in Animal Nutrition. CAB International, Wallingford, U.K. Dewhurst, R.J., Webster, A.J.F., 1992. Effects of diet, level of
intake, sodium bicarbonate and monensin on urinary allantoin excretion in sheep. Br. J. Nutr. 67, 345±353.
Goering, H.K., Van Soest, P.J., 1970. Forage Fibre Analysis. Agricultural Handbook No. 379, US Department of Agricul-ture.
Hendricksen, R.E., Poppi, D.P., Minson, D.J., 1981. The voluntary intake, digestibility and retention time by cattle and sheep of the stem and leaf fractions of a tropical legume (Lablab purpureus). Austr. J. Agric. Res. 32, 389±398.
Higgins, S.J., Tabrett, S.J., Poppi, D.P., Norton, B.W., 1992. Duodenal protein supply in cattle grazing signal grass (Brachiaria decumbens) and Glenn joint vetch (Aeschynomene americana). Proc. Austr. Soc. Anim. Prod. 19, 52.
Kumar, R., D'Mello, J.P.F., 1995. Antinutritional factors in forage legumes. In: D'Mello, J.P.F., Devendra, C. (Eds.), Tropical Legumes In Animal Nutrition. CAB International, Wallingford, UK, pp. 95±134.
Leng, R.A., 1990. Factors affecting the utilisation of `poor quality' forages by ruminants particularly under tropical conditions. Nutr. Res. Rev. 3, 277±303.
Energy allowances and feeding systems for ruminants, Reference Book 433. Her Majesty's Stationery, London, UK.
Minson, D.J., 1977. The chemical composition and nutritive value of tropical legumes. In: Skerman, P.J. (Ed.), Tropical Forage Legumes. FAO, Rome, pp. 186±194.
Muhammed, S., Stewart, C.S., Acamovic, T., 1994. Effects of tannic acid on cellulose degradation, adhesion and enzymatic activity of rumen microorganisms. Proc. Soc. Nutr. Physiol. 3, 174 Abstr.
Ndlovu, L.R., Buchanan-Smith, J.G., 1985. Utilisation of poor quality roughage by sheep. Effect of alfalfa supplementation on ruminal parameters, ®bre digestion and rate of passage from the rumen. Can. J. Anim. Sci. 65, 693±703.
Norton, B.W., Poppi, D.P., 1995. Composition and Nutritional Attributes of Pasture Legumes. In: D'Mello, J.P.F., Devendra, C. (Eds.), Tropical Legumes in Animal Nutrition. CAB International, Wallingford, UK, pp. 23±46.
érskov, E.R., Miller, E.L., 1988. Protein evaluation in ruminants. In : érskov, E.R. (Ed.), Feed Science-World Animal Science. Elsevier, Amsterdam, pp. 103±124.
Robinson, P.H., 1997. Modi®ed duodenal ¯ow of amino acids by manipulation of dietary protein sources. Can. J. Anim. Sci. 77, 241±251.
Salawu, M.B., 1997. The nutritive value of the leguminous browse Calliandra calothyrsus andthe role of condensed tannins in ruminant feeds. Ph.D. Thesis, University of Aberdeen. SAS, 1990. Statistical Analysis System Institute. SAS users guide:
Statistics, version 6, 3rd ed. SAS Institute, Cary, NC, USA. Satter, L.D., Slyter, L.L., 1974. Effect of ammonia concentration
on microbial protein production in vitro. Br. J. Nutr. 32, 199. Skerman, P.J., Cameron, D.G., Riveros, F., 1988. Tropical Forage
Legumes, 2 ed. FAO Plant Production and Protection series No 2. FAO, Rome, pp. 1590±1593.
Stern, M.D., Varga, G.A., Clark, J.H., Firkins, J.L., Huber, J.T., Palmquist, D.L., 1994. Evaluation of chemical and physical properties of feeds that affect protein metabolism in the rumen. J. Dairy Sci. 77, 2762±2786.
Topps, J.H., Oliver, J., 1993. Animal Foods of Central Africa. Zimbabwe Agricultural Journal Technical Handbook No. 2. Modern Farming Publications, Harare, Zimbabwe, pp. 76±105. Vera, R.R., Pizarro, E.A., Toscano, A., Rodriguez, N.M., 1989. Sites of digestion of nitrogen of hays made with the tropical legumeNeonotonia wightii. Proc. XVI International Grassland Congress, Nice, France, pp. 921±922.
