The determination of amino acid digestibility in poultry feedstuffs.

Abstract

THE DETERMINATION OF AMINO ACID DIGESTIBILITY IN POULTRY FEEDSTUFFS Y. RAHARJO* and D.J. FARRELL* SUMMARY Measurements were made on protein sup,plcments in.a basal diet, of dry matter, nitrogen and amino acid digestibilities using mature and' adult cockerels. Birds were killed 5h after receiving their diet in a single feed. Analyses were made on digesta from (i) the ileum, (ii) the terminal ileum, (iii) post caeca, and on excreta. In a second exDeriment -adult cockerels were fitted with a sitiple cannula in the terminal ileum. Cr203. was used as an indicator in the diet. Nitrogen-free diets were also fed. These contained different amounts of rice hulls. The results showed that digestibility measurements of nutrients based on ileal contents underestimate the true values. 'Consistently' higher values were found at ,the terminal ileum for the same diets. Considerable fermentation occurred in the caeca, consequently digestibilities measured on excreta did not always agree with those at the terminal ileum. Nitrogen digestibility of six meat meals va.ried from 0.79 to O-67 at the terminal ileum. A similar range was observed for most of the amino acids measured at this site. Cannulation of the terminal ileum was satisfactory and cannulae remained patent for several months. Lysine digestibilities for the four fish meals varied from 0.85 to 0.94. Corresponding values in excreta were 0 .95 to 0.86. Withincreasing amounts of fibre in the diet there was a significant increase in endogenous nitrogen excreta., Generally there was a corresponding increase in the concentration of several amino acids in excreta. of It was conc,luded that ileal cannulation was a practical method measuring amino acid and nutrient digestibilities in fowl. Excreta ' analysis may be useful for many high quality protein feedstuffs but those which result in large amounts of residue in the hind gut may undergo substantial change in amino acid profile. There is still the possibility of using amino acids in .excreta as a means of ranking p,rotein supplements for their digestibility. : INTRODUCTION The utilisation of amino acid acids by.liv&tock varies markedly for many reasons. Of most interest and of central importance to work reported in this paper is the feed ingredient. Utilisation, as defined here, embraces digestibility, absorption and metabolism. In this paper major emphasis will be placed on the digestibility or disappearance of amino acids from the small intestine of the fowl when offered .a range of feedstuffs. The concept of amino acid availability was by McNab (1979a, b) and Thomas (1980). There is chemical methods used to estimate availability of there is frequently, for many feedstuffs, a poor * reviewed most recently some uncertainty about some amino acids since correlation between Department of Biochemistry and Nutrition,, University of New England, Armidale. N.S.W. 2351 values determined in vitro and in vivo (Batterham 1979; Batterham et al. - -- ---__I_ 1981; Taverner and Farrell 198.1). Dye-binding techniques using acidazo dyes show some promise (Hurrell - al. 1979). The slope-ratio method is et probably the method that has the most direct application (Carpenter and Booth 1973) since it measures incremental growth and feed intake responses to additions to a diet of a test amino acid in a feed ingredient. However, there are still several potential problems with this assay (McNab 197923) and values for utilization of mine acids may exceed 100% (Varnish and Carpenter 1970). Maintenance of balance of the important essential amino acids when the test amino acid is added in incremental amounts is not always considered. Growth depression due to amino acid imbalance may therefore occur. The availability of free amino acids, when added to test diets is assumed to be 100% and that absorption occurs at the same rate as amino acids in the intact protein. Recently Sibbald (1979a)described a method of determining amino acid digestibility from.concentration in feed and corresponding excreta. This general approach was used previously by Kuiken and Lyman (1948) and Bragg et al. (1969) and is the basis of amino acid digestibilities reportedinfeed tables (Janssen et al. 1979). InSibbald's (1979a)method a correction is made for endosenous amino acids voided in excreta, and a true digestibilitv of the amino acids in a feedstuff is calculated. The basis of this correction has been questioned (Farrell 1981) since starved birds are used to provide endosenous amino acid output. Moreover it was shown bv Farrell (*1981) that dietary fibre can increase the output of endoaenous excreta bv coc`kerels. N and amino acid contents mav be similarlv increased. Clearly the easiest and most attractive method of determination of amino acid digestibility is that of excreta analysis (Janssen et al. 1979). This method is based-on the unproven assumption that *microbial changes to the amino acids in digesta during passage along the large intestine are minimal and do not mask changes in amino acid profile between the terminal ileum and the cloaca* Although the alimentary tract of the fowl has an active microflora (Javne-Williams and Coates 1969) it is assumed that feed residues pass rapidly through the large intestine thereby reducing the opportunity for major changes in amino acid profile of residues to occur. Observations on pigs, rats and cockerels (Slump et al. 1977) on the same diet would tend to support the concept that changes in amino acid profile of feed residues in cockerels are less than for other species. Salter and Coates (1971) working,with germ-free and conventional chicks showed that good and poor quality proteins were digested to the same extent by both qroups. The removal of the ileal contents of birds previously fed a test diet and the determination of amino acids in the feed and corresponding ileal contents was used by Payne et al. (1968), Soares and Kifer (1971), Varnish and Carpenter (1975) and GhKewhu andHewitt (1979) to estimate disappearance of amino acids from a knowninput of feed. 'It is necessary to incorporate chromic oxide (Cr203) in the diet and to recover the marker quantitatively in ileal digesta in order to estimate total contents from the feed consumed. The main objective of this paper is to describe research undertaken to evaluate biological methods of measuring amino acid disappearance from the small intestine of poultry. The method used was tedious and involved In an attempt to overcome this slaughter of large numbers of birds. problem a permanent T piece cannula were inserted in the terminal ileum of adult cockerels. An important aspect of the research described here was to examine changes that occurred in amino acid profile of digesta during transit along the large intestine. The purpose was to determine with poultry if excreta were consistently suitable material for determining amino acid digestibility of a range of feedstuffs. The influence of increments o,f dietary fibre on N and on amino acids in endogenous excreta was also studied to obtain estimates of true digestibility values. MATERIALS AND METHODS Slaughter experiments .One hundred and four White'Leghorn x Black Experiment 1 Australorp male chicks, 4 weeks of age, used in a previous experiment, were held in suitable group cages until 8 weeks of age. They were then placed in individual mesh-wire cages and trained to eat their daily feed allowance (about 100 g) in one hour (Farrell 1978). Birds were fed a 'commercial chick-grower diet adequate in all nutrients to allow maximum growth rate. The basal test diet consisted of corn and bone meal (2%). supplemented with minerals and vitamins. The 'basal diet was combined with 50% of a protein supplement from one of the following sources: 6 different meat meals, blood meal, fish meal, sweet lupinseed meal, soybean meal, maize meal, copra meal, cottonseed meal, peanut meal and sunflower meal. A nitrogen-free diet based on starch, glucose, oil bone meal and cellulose was used. Cr203 was added to all diets (0.1%) which were cold pelleted. Each dietary treatment had 3 replicates each of 3 birds. The test diet was given for 2 d; 3 birds were starved for 48 h and then offered 100 g of the test diet forlh. Five hours later the birds were killed by cervical dislocation and digesta collected from (i) the ileum - from Mechel's diverticulum to 2.5 cm anterior to the ileo-caecal junction (Sl); (ii) the region immediately post 'caeca to 1 cm from the cloaca (S3), and (iii) excreta (S4). Excreta were collected from a tray placed beneath each caqe. Experiment 2 (ii) One hundred and twenty six White Leghorn x Black Australorp male chicks were housed and treated similar to that described in experiment 1, except that the fish meal, blood meal and peanut meal-based diets were omitted due toI contamination with mould. When the birds were killed, digesta were collected only from the last 10 cm of the ileum (S2). Diaesta and excreta These were combined for 3 birds for each site and collected in plastic containers and quickly cooled, frozen and freeze-dried to constant weight. The dry material was weighed and milled then stored for analysis. Ileal cannula experiments A glass T piece cannula was inserted permanently in the terminal ileum about 2.5 cm anterior to the ileocaecal junction of adult cockerels (2-3 kg) using standard surgical procedures. When collection of digesta was required the screw cap and plug were replaced by a small plastic vial which was attached to the thread on the cannual stem. Digesta slowly entered the vial for about, 30 min. and was removed two or three times . (0 TABLE 1 Apparent dry matter (DM) and nitrogen digestibility coefficients of protein concentrates in the basal diet and measured in ileal contents (Sl), at the terminal ileum (S2), in excreta (S4) of three groups of birds killed 5 h post-feeding. 202 during the collection period. single meal (100 g) in L h to 5 rapidly and treated as already total collection was made from This commenced about 5 h after feeding a trained birds. Digesta were cooled described. For analysis of excreta, a 5 birds for 32 h. 'Diets were the same as for experiment 1, with three additional fish meals. The N-free diet was diluted with additions of milled rice, hulls to give incremental levels of ADF. Pure cellulose (Solkafloc) was included in one N-free diet. . . Analytical and statistical methods Chemical analyses were undertaken on finely milled representative samples of diet, digesta and excreta using the following procedures. Dry matter was determined by heating ground samples at 105 C for 24 hoursNitrogen was determined using the autoanalyser method of Clare and Stevenson (1964). .Acid detergent fibre (ADF) was measured according to the method of Van Soest (1963). Amino acid ,analysis was determined on the acid hydrolysates (Spackman et al. 1958) using a T.S.M. Technicon amino acid analysed. -m Cr2O3 was determined using an atomic absorption spectrophotometer (Perkin-Elmer, Model 360) and following the procedures outlined by Williams et al. (1962) except that an acetylene-nitrous oxide flame was used. -Data were comb ined for experiemnts 1 and 2 of variance was under taken . Differences b e twee n using Duncan's multiple range test (Duncan 1955). critically differences between measurements made terminal ileum and the entire ileum (experiments analysis of variance was made on these data. RESULTS Slaughter experiments The apparent digestibility of dry matter and N at the different sites are given for the experimental diets in Table 1. No correction was made for the contribution of the basal diet to these data or those in Table 2. Although dry matter digestibility differed significantly (P < 0.05) betw een sampling sites for individual feedstuffs, the overall mean's showed no differences (P > 0.05) between total ileal (Sl) and terminal ileal (S2) contents; differences (P < .0.05) were observed between values combined for these two sites and those post-caeca (S3) and excreta (S4). However more (P < 0.05,) N was digested at S2 than Sl, but at's4 there was a large depression in N digestibility due. to the addition of urinary N. For the majority of the meat meal diets dry matter and N digestibilities were lower at Sl than S2. The range of overall apparent digestibilities combined for the four sites for dry matter and for N ranged from 56 and 50 for copra meal, to 79 and 83 for maize gluten respectively. The N-free diet showed a decline (P < O-05) in dry matter digestibility in digesta post caeca; while for N, there were higher concentrations (g/100 g) at Sl than at the other three sites. and a one-way analy sis means were examined To examine more digesta'from the on 1 and 2), a separate 0 204 Fig..2. Mean apparent digestibility of in digesta of birds fed diets me& (SBM), maize gluten (MG) in the ileum and the terminal feeding. lysine, threonine and isoleucine _ based on meat meals (MM); soybean and cottonseed meal (C:;rj) measurc(l ileum of birds killed 5 h pbst 208 The mean disappearance of amino acids for the 8 plant-protein based diets, and the 8 animal protein diets at the 4 sites is given in Table 2. The mean of the 9 amino acids measured showed differences (P < 0.05) for both groups between Sl and S2. Amino acid disappearance increased from Sl through to S4. although this was not always significant. At the terminal ileum (S2) mean disappearance of amino acids for the animal protein-based diet was 0.75 compared with 0.81 for the plant tirotein-based diets. The.re was a wide ranse of values for disappearance among individual amino acids at the four sites. For lysine and threonine disappearance from the important site the terminal ileum (S2) I mean values were 0.77 and 0.68 for the animal proteins and 0.68 `and 0.76 for the plant proteins. Values tended to be greater than 0.80 for all amino acid digestibilities using excreta irrespective of the protein sourceat The mean apparent'digestibility of 9 amino acids in the 8 diets based on animal proteins, and those based on plant proteins at the four sites is shown in Figure 1. The major increases occurred between Sl and s2. With the exception of lysine in plant proteins, and glycine in animal proteins, no marked changes occurred in amino acid digestibilities in samples taken post-caeca and from excreta. Mean disestibilitv of three essential am&no acids, often limiting in poultry diets, .for the individual protein sources at .the four sites,is given in Tables 3A & B.Forthe 6 meat meal diets, considerable variation. was observed both between and within sites for all amino acids. Blood meal showed constant and consistently high values at all threexites ' measured. Within the group of plant proteins, maize gluten and to a lesser extent soybean meal had consistently high values for amino'acid digestibilities at the four sites for the.three amino acids. Copra meal and cottonseed m.eal on the other hand gave low values. Shown in a histogram (Fig. 2). is the apparent digestibility of three amino acids at Sl and S2 for two samples of meat meal, soybean meal, mdize qlu.ten and cottonseed meal. In all there was a considerable increase in digestibility for all feedstuffs at S2 compared with Sl. cases The in the 16 there was Sl to s4, (P < 0.05) mean apparent digestibility of the three important amino acids feedstuffs at the four sites is shown in Figure 3. In all cases a stepwise increase in digestibility of all amino acids from Only for lysine and threonine were differences significant .. between Sl and S2. Ileal cannula experiments Birds recovered rapidly from surgery and normally resumed eating a special high-energy, low residue diet within 36 h. Maintenance of the cannula (P&e, 1) was minimal. It remained patent for several months. Apparent digestibility .of 'dry matter, N and four amino acids at Sl and S2 are shown in Table 4. Meandry matter digestibility was lower (P < 0.05) in excretd than at the terminal'ileum. Cr203 estimates of excreta compared favourably with those determined by total collection; . mean values were 0.610 and 0.614 respectively. Apparent digestibility of amino acids at the two sites varied considerably depending on the protein source. Within the 5 meat meals lysine digestibility varied from 0.71(MM 16) to 0.85 (MM 13); corresponding values for threonine were 0.70 and 0.73. Interestingly for arginine 209 Fig. 4. The relationship between N in excreta (g]lOO g feed intake, y) and acid detergent fibre (ADF) in a N-free diet (g/100 g, X) fed to 5 adult cockerels at each level of ADF intake. Plate 1. A simple 'T' piece glass cannula with cap and stem used to collect digesta from the terminal ileum of adult .cockerels. 212 cottonseed meal had a similar ileal digestibility to corn of over 0.90 qbut for lysine,the values were 0.68 and 0.82 respectively. Mean values ,,between the two sites S2 and S4 were generally similar. Exceptions were lysine 0.84 (S2) and 0.75 (S4), and glycine 0.74 and 0.58 respectively. There was a significant P < 0.01) increase in N output in ileal digesta (S2) and in excreta (S4) with increasing increments of ADF in the 'N-free diet (Fig. 4). At the two highest levels (110 and 150 ADF g/kg) I Lmeasurements were made on excreta only. Data for excreta are shown for '4 and 6 observations. The two additional levels of ADF were fed in a idifferent experiment when it was realized that the first four levels did not cover the range of fibre contents`of the test diets containing protein supplements. Similar data for lysine, threonine and isoleucine in digesta and excreta of birds on the N-free diets are shown iti Figure 5. Changes in content did not always correspond to changes in dietary ADF. Dry matter digestibility, estimated at the terminal ileum using Cr203 in digesta, and in excreta, where total collection values were Comparisons between total also calculated, are not given here. collection of excreta and its estimation using Cr203 were excellent. Except for the diets containing added cellulose (59g/kg) ileal and excreta digestibilities were similar. For the diet containing cellulose, dry matter digestibility declined from 0.88 to 0.74. DISCUSSION The,justification for the inclusion of protein supplements at 5,0% of the basal diet was firstly to attempt to accentuate differences in digestibility between diets, and secondly to minimise the influence of endogenous excreta on the apparent digestibility valuesIt is clear from the results' in Tables 1 and 2 that the use of ileal contents, taken below Meckel's diverticulum, underestimates the digestibility of several diets for the component under examination. Thus the report by Payne et al.. (1968) that essentially all of the protein disappears from the gulf the fowl at a point two-thirds of the way along the jejunum is for many proteins incorrect. There was a marked increase in the apparent digestibility of nitrogen (14%) and of amino acids (12%) between the ileum (Sl) and the terminal ileum (S2)'. It should be pointed out that these comparisons were made in two separate experiments sin,ce it was necessary to collect all digesta from both sites. This would be impossible in a single experiment using the same birds. For dry matter, digestibility differences between the two sites may .be due to components other than those containing N, and clearly starch and some minerals may be absorbed from the lower part of the ileum Examination of N digestibilities (Table 1) showed that N was the major component explaining differences between the two sites (0.63 vs. 0.75). For all amino acids.measured, there was an increase in digestibility between Sl and S2 (Fig. 2), although this difference was not always significant. The low value for lysine digestibility of 0.48 found at Sl for cottonseed meal is identical to that of Soares and Kifer (1971) using the same technique; In the present study no correction was made for the 214 contribution of the basal diet (50%) although this diet contributed only 4.5% crude protein. There is little doubt that differences between amino, acid digestibility among protein supplements can be identified using ileal analysis (see Table 2). Achinewhu and Hewit (1979) showed differences between heat-damaged and untreated soybean protein to be 0.48 and 0.92 respectively. Values may have been even lower at the terminal ileum. It is clear that further modification occurred to the drv matter in most diets as digesta passed along the large intestine. Thornburn and Wilcox (1965) showed that with mature cockerels caecal digestion occurred, while Payne et al (1968) observed that caecetomised birds on fish .mealbased diets gave lower amino acid digestibilities than those with the caeca intact. However in the present study, for many diets there was a decrease in N digestibility between S2 and $3, but not for many of the individual amino acids (Table 2,.Fig. 1, 2 and 3) in these same diets. It would appear that digesta post caeca contained additions of nonprotein N contributed probably through microbial. fermentation (Okumura e t al. 1978).. This would suggest that deamination of some amino acids (Table 2) h as occurred and that ammonia was released and retained in digesta. The extent of the ammonia liberated ,and its retention in digesta would depend on the pH of digesta. l It would be expected that N digestibility based b'n excreta would d,ecline due to the contribution of N in urine (Table 1). This would not be expected for individual amino acids since urine contains only trace amounts of these (Bragg et al. 1969). Further increases in the apparent digestibility of several== acids occurred between S3 and S4 (Table 2, Fig. 1) indicating further microbial activity. The amino acid content o.f the N-free diet is shown in Table 5, and the values (mg/lOOg dry matter) compare well with published d,ata also given in Table 5. In contrast to the changes in N digestibility between S2 and S3 for the protein-containing diets,. no change occurred on this N-free diet (Table 1). Payne et al. (1968) suggested' that endoqenous amino acids may'be absorbed from the caeca. Not only is this highly unlikely, but there is no evidence of. this occurring from the data presented here. On the other hand dry matter digestibility did increase at S3 suggesting that so,me fermentation had occurred. Because N digestibility was lower at S3 (0.72) than at S2 (O-75), there was probably removal of amino acids through deamination and absorption of the fermented residues. Although the correction for endogenous amino acids is relatively unimportant in the pre,sent study because of the high daily intake and, high inclusion of the protein supplements in the basal diet such a corre,ction is essential,when small amounts of diet are fed. Sibbald (1980) was unable to show any effect of increasing amounts of dietary cellulose and sand on endogenous RA output. This study was criticised by Farrell (1981) because of the unlikely possibility of all the sand. reaching the small intestine from a single input of feed.. Our results here would indicate that pure cellulose may not be a useful'source of fibre in adult birds. It behaved in a different manner to that of milled rice hulls. There was an unexplained decline in dry matter digestibility between S2 and S4 from O-88 to 0.74, 'This large decline was not seen on the diets with rice hulls. But there was little difference in the N digestibility or amino acid contents of excreta from 215 this diet with added cellulose compared with diets containing similar additions of AW from hulls. The effects of dietary fibre (ADF) on endogenous excreta are in agreement with the increased output of energy in endogenous excreta of cockerels (Farrell 1981) and increased endogen-, et ous amino acid output in pigs (Taverner - al. 1981). The use of ileal cannulae in adult cockerels is a viable and simple method of determining amino acid digestibility and that of other nutrients in feedstuffs. The birds are easy to maintain provided they are housed individually in suitable cages. The changes that occur in the amino-acid content of digesta during transit through the hind gut, may mask`differences in digestibility when measured in excreta.' This has already been discussed. There is some uncertainty about the ability of CrzO3 to measure digesta output. Where comparisons were . made excreta dry matter output measured by total collection agreed favourably with that estimated using the indicator Crz03 (Table 4). This would suggest that ileal digesta output can be estimated using this indicator with sufficient precision to qive reliable values for calculating nutrient digestibility at this site. A comparison wasmade of N and amino acid digestibilities in the 16 feedstuffs at the terminal ileum and excreta of slaughtered birds, and of birds prepared with simple: cannulae in Table 6. Although there was a significant (P < 0.01) relationship for each amino acid between the two methods, there were significant differences (P < 0.05) for some values obtained by the two methods both in digesta at the terfiinal ileum and in excreta. At the terminal ileum differences were observed only for lysine and methionine but not for N. In excreta, amino acid digestibilities were consistently higher for killed than cannulated birds. In the latter group excreta were on trays for 32 h while for killed birds excreta were collected within about 5 h. Differences.may also be explained by age of bird and method used- In the slaughter experiment birds were at least 10 weeks old, while birds prepared with cannulae were at least one year old. Perhaps of greater significance is the shedding of material from the mucosal cells in the gut lumen that occurs when birds are killed and gut contents removed (Badawy 1964; Horzczaruk 1971). This would tend to increase the concentration of some amino acids and N in digesta. As a consequence there would be reduced digestibility but this was generally not the case for selected amino acids in Table 6. ' Like many biological techniques, both methods used here to estimate amino acid disappearance have their deficiencies and may therefore be criticised.' Removal of contents from killed birds, whether these be from the entire ileum or the terminal portion, may result in contamination as mentioned. Time of killing in relation to feeding time could also influence digesta concentration in different ' areas of the gut. Although frequently used, five hours after feeding may not be the most appropriate killing time of birds given a single meal- For cannulated cockerels, flow of material past the cannula mav not be identical to other areas in the gut, and digesta may not there-i fore be representative of gut contents- The role of Cr203 as a suitable marker has been discussed. In order that data can be presented in a form useful to feed formulators, it is essential to distinguish between' those amino acids that disappeared from the entire diet from those in the protein 216 supplement in the diet. Corrected estimates of some amino acids disappearing at the terminal ileum and through the entire gut, using is excreta, are given in Table 7 for a six protein supplements. I t apparent that there is a ranqe of values between amino acids and between protein sources at both sites (S2 and S4). Despite the uncertaintv of the method of amino acid analvsis of excreta to determine amino acid digestibility of feedstuffs, this technique was used to provide data on amino acid diqestibilities of a range of feedstuff's bv Janssen - al. (1979).. et TABLE 7 Apparent digestibility coefficients of lysine, threonine, and isoleucine in some protein supplements (corrected for the. basal diet contribution) determined at the terminal ileum (S2) and in excreta (S4) of cockerels prepared with simple 217 On the basis of the experiments reported here it would seem that changes occur in the amino acid profile in digesta as it proceeds from the ileum to the anus of the bird- The extent of these chanqes appear to depend to some extent on the quality of the protein and on the particular amino acid. For example maize gluten, soybean meal and blood meal were readily digested at Sl, and the amino acids in these sources were generally highly digestible. For many of the other proteins which were of poor quality including fish meal with crude protein of 446 g/kg (Safcol, S.A.), considerable differences were observed between Sl and S2 and S4. It follows that if large amounts of protein are entering the large intestine there is opportunity for substantial proteolytic activity to occur- A similar conclusion was made by Varnish and Carpenter (1975). The use of a T piece cannula sited at the terminal ileum appears to be 'a reasonable solution to a different problem. ACKNOWLEDGEMENTS . We thank the Australian Development Assistance Bureau for a scholarship for.Y.R., the Australian Chicken Meat Research Committee and the Poultry Research Advisory Committee for financial assistance, and Associate Professor R.B. Cumminq for his interest in this project. ._ REFERENCES ACHINEWHU, S.C. and HEWITT, 'D. (1979). 'Br. J. Nutr. 41:557. BATTERHAM, E.S. (1979). In 'Recent Advances in AnimalNutrition - 1979', P-11, editors W. Haresign and D. Lewis' (Butterworths: London). BATTERHAM, E.Sr MURISON, R-D. AND LOWE, R-F. (1981). Br. J. Nutr. 45: . C 401. BADAiiy, A.M. (1964). In 'The Role of Gastrointestinal Tract in Protein Metabolism', p-175, editor H. Munro (Blackwell: Oxford). BRAGG, D-B., IVY, C.A. and STEPHENSON, E.L. (1969). Poult. Sci. 48:'2135. CARPENTER, K.J. and BOOTH, V.H. (1973). Nutr. Abstr. Rev. 43:423. CLARE, N.T. and STEVENSON, A.E. (1964). N.Z. 51 agric. Res. =7:198. = DUNCAN, D-B. (1955). Biometrics. 11:l. FARRELL, D.J. (1978). Br. Poult. Sz. 19:303, FARRELL, D.J. (1981). Wld's Poult. Sci-7. 37:72. HURRELL, R.F., LERMON, P. and CARPENTER, K.J.71979). J. Fd. Sci. 44:122. HORSZCZARUK, F. (1971). Biul. Ins. Genet. Hodow. Zurerik. Pal. Akatt. Nauk. 21:117 -.JANSSEN, W.M.M., TERPSTRA, K., BEEKING, F.F.E. and BISALSKY, A.J.N. (1979). 'Feeding Values for Pountry' (Spelderholt Mededeling 303). JAYNE-WILLIAMS, D.J. and COATES, M.E. (1969). In 'Nutrition of Animals of Agricultural Importance', p-241, editor D. Cuthbertson, (Pergamon Press: London). KUIKEN, K.A. and LYMAN, C.M. (1948). J. Nutr. 46:13. McNAB, J.M. (1979a). In 'Recent Advances in Animal Nutrition - 1979', p-1, editors W. Haresign .and D. Lewis (Butterworths: London). McNAB, J.M. (1979b). In 'Proc. 2nd European Symposium on Poultry Nutrition', p-102, editors C.A. Kan and P.C.M. Simons (PUDOC: Wageningen). OKUMURA, J., HEWITT, D. and COATES, M-E. (1978). Br. J. Nutr, 39:99 PAYNE, W.L., COMBS, G.F., KIFER, R.R. and SYNDER, D.G. (1968). gd. Proc. 27:1199. . SALTER, D.N. a= COATES, M.E. (1971). Br. J. Nutr. 26:55. SIBBALD, 1-R. (1980). Poult. Sci. 59:836. Poult. Sci. T&668. SIBBALD, I.R. (1979a). SIBBALD, I.R. (1979b). Poult. Sci. 59:934. SIBBALD, 1-R. (1980). Poult. Sci. 59:836. SLUMP, P., van BECK, L., JANSSEN, W.KM.A., TERPSTRA, K., LEWIS, N.P. In 'Proc. 2nd 3jlt. Symp. on Protein and SMXTS, B. (1977). (PUDK: Wageningen). rcletabolism and Nutrition', p.70. SOARES, J.H. and KIFER, R.R. (1971). Poult. Sci. !jo:41. SPACKMAN, D-H., STEIN, W .H. and MOORE, S. (1958). Anal. Chem. 3o:`i190. TABERNER, M.R. and FARRELL, D.J. (1981). Br. J. Nutr. (in press). TERPSTRA, K. (1979). In 'Proc. 2nd European Symposium on Poultry Nutrition', p-97, editors C.A. Kan and P.C.M. Simons (PUCOD: Wageningen). TAVERNER, M.R., HUM& I.D. and FARRELL, D.J. (1981). Br. J. Nutr. (in press). THOMAS, 0-P. (1980). In 'Proc. 1980 Arkansas Nutrition Converence', p-101 (University of Ark