Comparison of Sample Source (Excreta or Ileal Digesta) and Age of Broiler Chick on Measurement of Apparent Digestible Energy of Wheat and Barley 1

Transcription

1 Comparison of Sample Source (Excreta or Ileal Digesta) and Age of Broiler Chick on Measurement of Apparent Digestible Energy of Wheat and Barley 1 T. A. SCOTT,*,2 F. G. SILVERSIDES, H. L. CLASSEN, M. L. SWIFT, and M. R. BEDFORD *Pacific Agri-Food Research Centre, P.O. Box 1000, Agassiz, British Columbia, Canada, V0M 1A0, 426 St. Andrew s Street, Nanaimo, British Columbia, Canada V9S 1S2, Department of Animal and Poultry Science, University of Saskatchewan, Saskatoon, Saskatchewan, Canada, S7N 5B5, Pro-Form Feeds, P.O. Box 1000, Chilliwack, British Columbia, Canada, V2P 6J6, and Finnfeeds International, Box 777, Marlborough, Wiltshire, United Kingdom SN8 1XN ABSTRACT The broiler chick bioassay measures AME of wheat- or barley-based diets, with or without an enzyme, from excreta (24-h collections at 8 or 16 d) and ileal digesta (17 d). The objective was to discuss the merits and accuracy of sample source (excreta vs ileal digesta) and bird age for determining the feeding value of wheat and barley. The bioassay utilized 80% of a test cereal grain, 20% basal diet containing 1.1% acid insoluble ash marker, and fed with or without an enzyme to four pens of six male broilers from 4 to 17 d. A total of 138 wheat and 97 barley samples (with and without an enzyme) were tested in 15 and five bioassays, respectively. Within each wheat or barley bioassay two control wheat and barley samples were measured. The among-pens and betweenassays CV for AME were calculated for these control samples, and correlation coefficients between the measures were calculated for the controls and for all of the 138 wheat and 97 barley samples included in the assays. For wheat samples, values for AME were lowest for excreta samples collected at 8 d, and similar for excreta and ileal digesta samples collected at 16 and 17 d, respectively. For barley samples, the three values were significantly different. The among-pens and betweenassay CV were low for AME among both wheat and barley samples. Correlation coefficients between several measures of AME at 8 and 16 d were significant for the control samples with enzyme supplementation. When all samples were included in the analysis, correlation coefficients between AME measures were moderate to high. On the basis of accuracy, precision, and cost, these data favor measuring AME on excreta samples at 16 d of age. Comparisons of number of pens of broilers used to determine AME would suggest that much of the variability predicted with four pens of six broilers each could be achieved with three, and possibly two pens of six broilers each, thereby greatly increasing the capacity of the assay to screen large numbers of samples (Key words: metabolizable energy, digestible protein, wheat, barley, enzyme) INTRODUCTION Measuring the feeding value (i.e., available energy and protein) of feed ingredients has and continues to be important to the broiler industry. The chemical energy contained in a feedstuff are relatively easy to measure. However, these estimates do not provide an accurate measure of the nutrients that the bird is able to extract, and bioassays must be used to measure ME. Many bioassays for ME have been used, and the relative merits of these have been debated extensively in the scientific literature (Farrell, 1981; Sibbald, 1985; Härtel, 1986; Askbrant, 1990; McNab, 1996). The AME bioassays can 1998 Poultry Science 77: also facilitate measurements of nitrogen retention; however, not all AME bioassays allow collection of ileal digesta to measure amino acid digestibilities (Bragg et al., 1969; Likuski and Dorrell, 1978; Papadopoulos, 1985). The energy extracted from a feedstuff is dependent in part on the strain (March and Biely, 1971; Shires et al., 1987), and the age of the bird used in the assay. Higher ME values were found in adults than in young birds (Lodhi et al., 1969; March et al., 1973; Peterson et al., 1976; Shires et al., 1980). Yolk absorption by very young chicks can result in high and variable ME values. These values decline to 8 d as the yolk contents were exhausted and then increase to 14 d as the chick s ability to digest feed Received for publication June 12, Accepted for publication October 20, AAFC Station Contribution Number To whom correspondence should be addressed: ScottTA@.em.agr.ca Abbreviation Key: AME8 = AME at 8 d of age; AME16 = AME at 16 d of age; AMEIL = AME determined by ileal digesta after necropsy; CPS = Canadian Prairie Spring wheat; HRS = Hard Red Spring wheat; NSP = nonstarch polysaccharides. 456

2 VARIABILITY IN FEEDING VALUE MEASUREMENTS 457 TABLE 1. Composition of the diets containing Hard Red Spring (HRS) and Canadian Prairie Spring (CPS) wheat and hulled and hulless barley 1994 Wheat 1995 Wheat/1994 and 1995 Barley HRS CPS HRS CPS Hulled Hulless Ingredient wheat wheat wheat wheat barley barley (%) Cereal Tallow Isolated soy protein Corn gluten meal L-Lysine DL-Methionine Vitamins 1 and minerals Celite Chromic oxide Enzyme 3 Analyzed composition n DM Gross energy, kcal/kg 4,660 4,600 4,570 4,451 4,440 4,460 Crude protein (N 6.25) Acid insoluble ash AME + enzyme, kcal/kg 3,200 3,200 3,140 3,150 3,130 3,230 AME, no enzyme, kcal/kg 3,060 3,040 3,020 3,020 2,980 2,910 1Supplies per kilogram of diet: vitamin A, 9,000 IU; cholecalciferol, 1,500 IU; vitamin E, 10 IU; vitamin K, 0.5 mg; vitamin B 12, mg; thiamin, 0.4 mg; riboflavin, 6 mg; folic acid, 1 mg; biotin, 0.15 mg; niacin, 35 mg; pyridoxine, 4 mg; choline chloride, 1,000 mg; DL-methionine, 1,184 mg; ethoxyquine, mg. 2Supplies per kilogram of diet: salt (NaCl), 2 g; manganese SO 4, 60 mg; copper SO 4, 5 mg; selenium (sodium selenate), 0.1 mg; iodine, (EEI), 0.35 mg; zinc SO 4, 50 mg % enzyme was added to the mixed diet, therefore actual levels of ingredients in this diet portion should be factored by Care of the birds follows the guidelines of the Canadian Council of Animal Care (1993) and the protocol has been approved by the Animal Care Committee at the Pacific Agriculture Research Centre at Agassiz, BC, Canada. increases (Zelenka, 1968). The digestive transit time also changes with age, and is also dependent on the form of the material (Vergara et al., 1989). The effect of the bird age is particularly important for feeds containing antinutritive substances (March et al., 1973) such as soluble nonstarch polysaccharides (NSP). The growth depressing effects of soluble NSP in young chicks can be diminished by the inclusion of fungal or bacterial enzymes in the diet (Campbell and Bedord, 1992). Generally, the antinutritive effects of NSP diminish after 3 wk of age, and the positive effect of enzymes is reduced. However, Bedford (1996) states that for broilers started on diets with enzymes there is a continued requirement for enzymes in the diet, at least until the birds reach market age. Adaptation by the microflora of the gut to available substrates (undigested portions of the diet, i.e., NSP) or unique environmental conditions created by ingredients of the diet would be expected over time (Choct et al., 1992; Smits and Annison, 1996). Scott et al. (1997) presented a broiler chick bioassay with highly repeatable AME measurements of common control samples of wheat and barley in successive bioassays. In addition to the concerns regarding repeatability of the AME measurements were: 1) having a bioassay that allowed screening of large numbers of test samples of barley and wheat quickly and economically; 2) limitations in the amount of ingredient required for testing; 3) having a bioassay that measured the bird s or its gut microflora s ability to adapt to a test ingredient; and 4) the minimum number of replicate groups of birds required for accurate measurement of feeding value variability between samples. The aim of this publication was to compare three measures of AME, the random variation associated with each, and the statistical relationship between them. The present study also assesses the number of pens of six male broilers that would provide an accurate measurement of feeding value. MATERIALS AND METHODS The procedures used for the AME broiler chick bioassay were described by Scott et al. (1997). Briefly, the test diets were formulated to contain 80% of the ground cereal grain sample and 20% of a basal mixture containing isolated soy protein, amino acids, vitamins, minerals, and an acid insoluble ash marker (Table 1). The test diet was then split equally, one portion fed as is, and the other portion supplemented with an appropriate NSP enzyme. Each diet, with and without enzyme supplementation, was consumed ad libitum by four pens of six male broiler chicks each from 4 to 17 d of age.3 Excreta samples, from each pen of broilers were

3 458 SCOTT ET AL. TABLE 2. The number of control and test cereal grain-based diets (with or without enzyme) and the respective number of excreta (8 and 16 d) and ileal digesta samples per diet used to determine AME levels of diets Bioassay series Wheat Variable Barley Number of bioassays used Control diets 1 measured in each bioassay (Total number of control diets tested) (44) (16) (20) 8 d excreta samples collected per diet d excreta samples collected per diet d ileal digesta samples collected per diet Total test cereal grain-based diets d excreta samples collected per diet d excreta samples collected per diet d ileal digesta samples collected per diet Total control and test cereal grain-based diets All ileal digesta samples, except for wheat 1994, were based on the combined digest from two replicate pens of broilers (total 12 broilers). For Wheat 1994 only two pens of birds were used for each control diet, and the ileal digesta samples tested represent the digesta collected from the six birds within each pen. 2Each control and test cereal grain-based diet was fed with and without an enzyme. collected for 24 h to measure AME at 8 (AME8) and 16 (AME16) d of age. At 17 d of age, the chicks were killed by cervical dislocation and the ileal digesta collected and used to determine AME (AMEIL). The diet, excreta, and digesta samples were analyzed for DM, gross energy, and insoluble ash, using standard laboratory procedures (Association of Official Analytical Chemists, 1990). A total of three bioassay series were required to evaluate different wheat (from 1994 and 1995) and barley samples. The number of control and test cereal grain-based diets, and the respective excreta and ileal collections per diet within each bioassay series are presented in Table 2. Common control samples of wheat [Hard Red Spring (HRS) and Canadian Prairie Spring (CPS)] and barley (hulled and hulless) were tested, respectively, in the wheat (1994 and 1995) and barley bioassays. The control samples were employed to measure bird and environment variation between bioassay runs. The AME was calculated using the following formula: AME = GE diet ( / GE excreta or digesta Marker diet Marker excreta or digesta). The three AME measures obtained from the control wheat-based diets in 1994 and 1995, and from the control barley-based diets were subjected to analysis of variance using the General Linear Models (GLM) procedure of SAS (Littell et al., 1991). In each of the three analyses, the assay and control diets were included as fixed, main effects along with the assay by diet interaction. The error term was used for all F tests, and least squares means (LSM) and their SE was calculated by the GLM program. The three AME measures were also analyzed using models including age (8, 16, or 17 d), diet, and assay, and the two- and three-way interactions of these effects. The data were sorted by control diet within bioassay series and analyzed with the individual assays as the only effect. Using this model, the error CV was the among pens CV, and the MS for the assay can be used to calculate the between-assay CV. Product-moment correlation coefficients were calculated between the three AME measures for the control diets within each bioassay series. To assess the statistical relationships among these variables in a more varied group of cereal grains, correlation coefficients were also calculated using all of the samples of wheat (276) and barley (194) (Table 2). This provided a range of AME values in which the variation common to the measures should be related to the true ME values of the sample. The accuracy of incremental numbers of replicate pens of six male broilers used for determining AME8, AME16, and AMEIL were measured for all wheat and barley samples. Mean AME8 and AME16 values based on one, two, three or four pens of 6 broilers each and AMEIL based on one or two groups of ileal digesta (6 or 12 broilers per group) were correlated and the correlation coefficient used to assess the accuracy determined with different replicate numbers. RESULTS For both control wheat and barley samples, AME measured at 8 d was lower (P 0.01) than that measured at 16 or 17 d. In both wheat bioassay series, the latter two measures were statistically similar (Table 3), whereas for the control barley samples, the three measures of AME were significantly different. For each

4 VARIABILITY IN FEEDING VALUE MEASUREMENTS 459 TABLE 3. Apparent metabolizable energy values obtained from excreta at Days 8 (AME8) and 16 (AME16), and from ileal contents at Day 17 (AMEIL) for samples of Hard Red Spring (HRS) and Canadian Prairie Spring (CPS) wheat and hulled and hulless barley with and without enzyme supplementation (AME16- (AMEIL- AME8) AME16) Diet n AME8 AME16 100/AME16 AMEIL 100/AMEIL (kcal/kg) (%) (kcal/kg) (%) Wheat, 1994 HRS + enzyme 22 3,450 a 3,650 a 5.5 3,240 a (21) HRS 22 3,330 b 3,440 c 3.2 3,510 b (19) CPS + enzyme 22 3,360 b 3,560 b 5.6 3,550 b 0.3 CPS 22 3,200 c 3,340 d 4.2 3,320 c 0.6 SE Overall mean 88 3,330 y 3,500 x 4.9 3,530 x (84) 0.9 SE Wheat, 1995 HRS + enzyme 16 3,600 a 3,650 a 1.4 3,730 a 2.1 HRS 16 3,380 bc 3,480 c 2.9 3,480 c 0.0 CPS + enzyme 16 3,410 b 3,580 b 4.8 3,600 b 0.6 CPS 16 3,310 c 3,400 d 2.7 3,410 c 0.3 SE Overall mean 64 3,420 y 3,530 x 3.1 3,560 x (32) 0.8 SE Barley Hulled + enzyme 20 3,140 b 3,160 b 0.6 3,250 b 2.8 Hulled 20 2,980 c 2,980 c 0.0 3,040 c 2.0 Hulless + enzyme 20 3,280 a 3,290 a 0.3 3,440 a 4.4 Hulless 20 2,790 d 2,890 d 3.5 2,820 d 2.5 SE Overall mean 80 3,050 z 3,080 y 1.0 3,140 x (40) 1.9 SE a dmeans within columns, cereal grains, and years with no common superscript differ significantly (P 0.05). x ymeans for the three AME measures with no common superscript differ significantly (P 0.05). 1When n is not equal to that given, it is given in parentheses, n is 8 for AMEIL for 1995 wheat assays, and 10 for AMEIL for barley assays. 2The SE of HRS + enzyme is 29 kcal/kg and that of HRS is 31 kcal/kg. measure, HRS wheat had higher AME values than did CPS wheat, and enzyme addition increased AME significantly. For barley, hulless barley with an enzyme had the highest AME values and hulless barley without an enzyme had the lowest values. Enzyme addition to the hulled barley sample increased the AME value in a manner similar to that in wheat, but the improvement to hulless barley was much greater (i.e., 20% for AMEIL). It is evident from the data (Table 3) that there were greater improvements in digestibility of the wheat-based diets at 16 d of age compared to 8 d of age, than were observed for barley-based diets. It was also evident that for the respective control diets the AME levels as determined with ileal digesta were not always consistently higher than AME levels determined from excreta, ranging from 2.5% for hulless barley (without enzyme) to 4.4% for hulless barley with enzyme. Overall, the difference between AME values determined with ileal digesta as compared to excreta (collected at 16 d) was greater for barley diets than for wheat-based diets. Generally, the difference in AME values between ileal and excreta (16 d) was also higher for diets with as compared to without an enzyme. The CV among pens and between assays for AME measures are shown in Table 4. All of the among-pens CV were low for both wheat and barley (between 1.2 and 6.7%). Of the three measures, the among pens CV at 8 d were most often greatest and those at 16 d were most often the smallest. Diets supplemented with an enzyme had smaller among pens CV than comparable diets without an enzyme at 16 and 17 d. For the wheat samples, AME measured at 8 d also had the highest between-assays CV, but for barley samples, the betweenassay CV was highest at 16 d. Lowest between-assays CV were at 16 d for wheat in 1994, at 17 d for wheat in 1995, and at 16 or 17 d for barley. The consistent decrease in among-pens CV brought about by enzyme supplementation was not seen for the between-assays CV. The between-assays CV for wheat diets were very similar to the among-pens CV, and those for barley diets were only slightly higher. The correlation coefficients (Table 5) between AME8 and AME16 were all significant (HRS in 1994, P 0.10) for enzyme-supplemented wheat and for all barley samples. The correlation coefficients between AME16 and AMEIL with enzyme supplementation were also significant for 1994 wheat. As would be expected, the correlation coefficients between measures at 8 and 17 d were lower. Without enzyme supplementation, the

5 460 SCOTT ET AL. TABLE 4. Coefficients of variation among pens and between assays for AME values obtained from excreta at Days 8 (AME8) and 16 (AME16) and from ileal contents (AMEIL) for samples of Hard Red Spring (HRS) and Canadian Prairie Spring (CPS) wheat and hulled and hulless barley with and without enzyme supplementation CV among pens CV between assays Diet n 1 AME8 AME16 AMEIL AME8 AME16 AMEIL (%) Wheat, 1994 HRS + enzyme (21) HRS (19) CPS + enzyme CPS Wheat, 1995 HRS + enzyme HRS CPS + enzyme CPS Barley Hulled + enzyme Hulled Hulless + enzyme Hulless There were 11 assays of wheat in 1994, each with two repetitions, four assays of wheat in 1995, each with four repetitions, and five assays of barley, each with four repetitions. When n is not equal to that given, it is given in parentheses, n is 8 for AMEIL for 1995 wheat assays, and 14 for AMEIL for barley assays. correlation coefficients between these three variables were lower, and only a few were significantly different from zero. All of the correlation coefficients among the three AME values (Table 6) when calculated with all of the samples included in these assays were significant and TABLE 5. Correlation coefficients (r) among the AME measures obtained from excreta at Days 8 (AME8) and 16 (AME16) and ileal contents (AMEIL) for samples of Hard Red Spring (HRS) and Canadian Prairie Spring (CPS) wheat and hulled and hulless barley with and without enzyme supplementation With enzyme there was no clear effect of enzyme supplementation. The accuracy in measuring AME8, AME16, and AMEIL using one to four replicate pens of six male broilers each is presented in Tables 7 and 8, respectively, for wheat- and barley-based diets. Mean AME values of the 276 wheat and 194 barley diets were relatively Without enzyme Sample n 1 AME8 AME16 AME8 AME16 HRS, 1994 AME NS 0.16 NS AMEIL NS 0.72** 0.37 NS (19) 0.14 (19) CPS, 1994 AME ** 0.23 NS AMEIL NS 0.68** 0.07 NS 0.45* HRS, 1995 AME ** 0.10 NS AMEIL NS 0.13 NS 0.58 NS 0.46 CPS, 1995 AME * 0.19 NS AMEIL NS 0.62 NS 0.72* 0.40 Hulled barley AME ** 0.10 NS AMEIL NS 0.53 NS 0.16 NS 0.53 Hulless barley AME * 0.37 NS AMEIL NS 0.49 NS 0.12 NS 0.75* 1When n differs, it is given in parentheses following the correlation coefficient. *P **P 0.01.

6 VARIABILITY IN FEEDING VALUE MEASUREMENTS 461 TABLE 6. Correlation coefficients (r) between AME measures obtained from excreta at Days 8 (AME8) and 16 (AME16), and from ileal contents (AMEIL) in wheat and barley samples with and without enzyme supplementation With enzyme Without enzyme Sample n AME8 AME16 AME8 AME16 Wheat AME ** 0.77** AMEIL ** 0.79** 0.65** 0.80** Barley AME ** 0.88** AMEIL ** 0.67** 0.59** 0.55** **P consistent whether one or more pens of six male broilers were used. However, the SD of the mean, as expected, decreased with each extra pen of birds used in the determination of the three AME values. It is evident, based on correlations, that mean AME values for the respective diets would be nearly as accurately predicted if we used three, and possibly two replicate pens per diet, rather than the four pens currently used. Based on these determinations, we gain more confidence in using two replicate groups for ileal measurements. Ileal digesta was pooled from two pens of six broilers to ensure adequate amounts of sample were available for all the laboratory analyses. DISCUSSION Differences between the three AME values could be related to differences in age of the chicks or to the sampling location. Digestive system maturation increases the AME of a feed (March et al., 1973; Peterson et al., 1976; Shires et al., 1980) and this likely explains the difference between 8 and 16 d of age. The difference between excreta samples at 16 d and ileal samples at 17 d were expected to be largely due to the action of microflora posterior to the ileum. Differences in AME determined using excreta collected at 16 d and ileal digesta collected at 17 d was not significant for wheatbased diets, but was significant for barley-based diets. This latter observation may relate to the higher fiber content of hulled barley and the higher NSP content of hulless barley as compared to wheat. The undigested amount of fiber providing a significant food source for an adaptable gut microflora population. The question still remains, is the greater variability in AME values measured at 8 d and its lower relationship (i.e., correlation) with AME determined at 16 or 17 d TABLE 7. The effect of number of replicate groups of broilers (6 per group for AME8 and AME16, 12 per group for AMEIL) on accuracy of determining AME levels of wheat (n = 276) with and without enzyme Values determined with 4, 3, 2 or 1 replicate pens of 6 broilers Variable 4 pens 3 pens 2 pens 1 pen AME8 Mean 3,396 3,394 3,387 3,382 SD Range 2,436 to 3,828 2,395 to 3,328 2,341 to 3,845 2,267 to 3,814 r 2 with 4 pens AME16 Mean 3,564 3,563 3,553 3,553 SD Range 2,869 to 3,915 2,832 to 3,925 2,876 to 3,908 2,582 to 3,960 r 2 with 4 pens Values determined with 2 or 1 replicate pens of 6 or 12 broilers 2 pens 1 pen AMEIL Mean 3,617 3,623 SD Range 2,798 4,701 2,798 4,701 r 2 with 2 pens The r 2 values represent the correlation coefficients between measurements of different number of groups of broilers with the means determined with four groups for AME8 and AME16 and two groups for AMEIL.

7 462 SCOTT ET AL. TABLE 8. The effect of number of replicate groups of broilers (six per group for AME8 and AME16, 12 per group for AMEIL) on accuracy of determining AME levels of barley (n = 174) with and without enzyme Values determined with 4, 3, 2 or 1 replicate pens of 6 broilers 4 pens 3 pens 2 pens 1 pen AME8 Mean 3,160 3,154 3,149 3,146 SD Range 2,493 to 3,751 2,473 to 3,754 2,431 to 3,742 2,375 to 3,756 r 2 with 4 pens AME16 Mean 3,207 3,208 3,206 3,204 SD Range 2,682 to 3,711 2,674 to 3,736 2,611 to 3,723 2,475 to 3,710 r 2 with 4 pens Values determined with 2 or 1 replicate pens of 12 broilers 2 pens 1 pen AMEIL Mean 3,179 3,185 SD Range 2,442 to 3,692 2,249 to 3,785 r 2 with 2 pens The r 2 values represent the correlation coefficients between measurements of different number of groups of broilers with the means determined with four groups for AME8 and AME16 and two groups for AMEIL. (using excreta or ileal digesta, respectively), an anomaly of the bird (i.e., presence of unabsorbed yolk or variable endogenous enzyme capacity) or gut microflora variability. Poor digestibility of a cereal grain-based diet during the very early growth period could significantly impact the production and efficiency of the flock. Therefore, there is some impetus for knowing whether this poor digestibility relates to diet, bird, or gut microflora variability. The among pens CV for all measures were low, and lower at 16 and 17 d than at 8 d, reflected reduced variation with maturation of the digestive system. For both wheat and barley, the between-assays CV for AME were only slightly higher than the among-pens CV, suggesting that the AME measures were repeatable both within and between assays. Enzyme addition consistently reduced the between-pens CV, representing random variation, for AME, but not the between-assays CV, which represents both random variation and real differences between assays. Correlation coefficients between measures on the control samples will be low if the only source of common variation is random. However, several correlation coefficients for enzyme-supplemented diets, and some for the unsupplemented diets were significant for AME measurements, suggesting that the variation between assays was not entirely random, and the three were measuring an underlying difference between the control samples in different assays. Correlation coefficients were calculated for all of the cereal grain samples tested in these assays because the common variation measured represents nonrandom variation. The coefficients between AME measures were all significant. The measure at 16 d provided a compromise between the earlier measure at 8 d and the ileal measure, and enzyme addition had no clear effect. The data comparing accuracy of AME determinations over large numbers of diets with one to four groups of six male broilers would indicate that good measurements of variability between cereals may be achieved with three, and possibly two, as compared to four replicate groups. This is encouraging because it could greatly reduce the time and cost of measuring large numbers of cereal samples. Accurate estimates of feeding value on large numbers of samples is imperative for validation of lab-bench prediction (i.e., near infrared spectrometry, and carbohydrate profiling) of bioassay measurements. Scott et al. (1997) described a bioassay for broiler chicks, and the variation associated with the measure of AME taken at 16 d. Table 9 allows comparison of the three AME measures recorded in this assay on the basis of accuracy, repeatability, and cost. Measurements of AME taken at 8 d are rapid, and require less test cereal grain sample, but have lower values and higher CV than those taken later. Measuring AME from excreta and ileal samples provide similar values and precision, but obtaining the ileal contents was time consuming and requires euthanasia of the bird, leading to the conclusion that measurement of AME was most cost effective from excreta at 16 d of age. Measurements of the ileal TABLE 9. Relative merits of the three AME values 1 Variable AME8 AME16 AMEIL Accuracy Repeatability Time and expense indicates low merit, ++indicates moderate merit, +++indicates high merit.

8 VARIABILITY IN FEEDING VALUE MEASUREMENTS 463 contents were, however, still deemed necessary to determine amino acid digestibility. Wheat and barley samples can contain NSP, which have antinutritive effects that were reduced or eliminated by supplementing the feed with enzymes. Enzymes have a significant effect on energy digestibilities and experimental estimation of ME must be made with enzyme supplementation if it will eventually be used in the feeding of commercial chicks. ACKNOWLEDGMENTS We gratefully acknowledge the financial support of the Alberta Barley Commission, Canadian Wheat Board, Finnfeeds, Int., IRAP/NSERC, BC Broiler Chicken Marketing Board, and Agriculture and Agri-Food Canada. We also thank J. Helm, B. Rossnagel and P. Hucl for growing and shipping the wheat and barley samples. The poultry staff of the Pacific Agri-Food Research Centre are also thanked for their dedication and hard work in maintaining the birds, and collecting and analyzing samples. REFERENCES Askbrant, S., The concept of metabolisable energy for poultry. Swedish University of Agricultural Sciences, Department of Animal Nutrition and Management, Report 194, Uppsala, Sweden. Association of Official Analytical Chemists, Official Methods of Analysis. Association of Official Analytical Chemists, Washington, DC. Bedford, M. R., The effect of enzymes on digestion. J. Appl. Poult. Res. 5: Bragg, D. B., C. A. Ivy, and E. L. Stephenson, Methods for determining amino acid availability of feeds. Poultry Sci. 48: Campbell, G. L., and M. R. Bedford, Enzyme applications for monogastric feeds: a review. Can. J. Anim. Sci. 72: Choct, M., G. Annison, and R. P. Trimble, Soluble wheat pentosans exhibit different anti-nutritive activities in intact and cecectomized broiler chickens. J. Nutr. 122: Farrell, D. J., An assessment of quick bioassays for determining the true metabolizable energy and apparent metabolizable energy of poultry feedstuffs. World s Poult. Sci. J. 37: Härtel, H., Influence of food input and procedure of determination on metabolizable energy and digestibility of a diet measured with young and adult birds. Br. Poult. Sci. 27: Likuski, H. J., and H. G. Dorrell, A bioassay for rapid determination of amino acid availability values. Poultry Sci. 57: Littell, R. C., R. J. Freund, and P. C. Spector, SAS System for Linear Models. 3rd ed. SAS Series in Statistical Applications. SAS Institute Inc., Cary, NC. Lodhi, G. N., R. Renner, and D. R. Clandinin, Studies on the metabolizable energy of rapeseed meal for growing chickens and laying hens. Poultry Sci. 48: March, B. E., and J. Biely, Factors affecting the response of chicks to diets of different protein value, breed and age. Poultry Sci. 50: March, B. E., T. Smith, and S. El-Lakany, Variation in estimates of the metabolizable energy value of rapeseed meal determined with chickens of different ages. Poultry Sci. 52: McNab, J. M., Factors affecting the energy value of wheat for poultry. World s Poult. Sci. J. 52: Papadopoulos, M. C., Estimations of amino acid digestibility and availability in feedstuffs for poultry. World s Poult. Sci. J. 41: Peterson, C. F., G. B. Meyer, and E. A. Sauter, Comparison of metabolizable energy values of feed ingredients for chicks and hens. Poultry Sci. 55: Scott, T. A., F. G. Silversides, H. L. Classen, M. L. Swift, M. R. Bedford, and J. W. Hall, A broiler chick bioassay for measuring the feeding value of wheat and barley in complete diets. Poultry Sci. 77: Shires, A., A. R. Robblee, R. T. Hardin, and D. R. Clandinin, Effect of the age of chickens on the true metabolizable energy values of feed ingredients. Poultry Sci. 59: Shires, A., J. R. Thompson, B. V. Turner, P. M. Kennedy, and Y. K. Goh, Rate of passage of corn-canola and cornsoybean meal diets through the gastrointestinal tract of broiler and White Leghorn chickens. Poultry Sci. 66: Sibbald, I. R., The true metabolizable energy bioassay as a method for estimating bioavailable energy in poultry feeding stuffs. World s Poult. Sci. J. 41: Smits, C. H., and G. Annison, Non-starch plant polysaccharides in broiler nutrition towards a physiologically valid approach to their determination. World s Poult. Sci. J. 52: Vergara, P., M. Jiminez, C. Ferrando, E. Fernandez, and E. Gonalons, Age influence on digestive transit time of particulate and soluble markers in broiler chickens. Poultry Sci. 68: Zelenka, J., Influence of the age of the chicken on the metabolizable energy values of poultry diets. Br. Poult. Sci. 9: