The Effect of Enzyme Treatment of Soybean Meal on Oligosaccharide Disappearance and Chick Growth Performance

K. K. Graham,* M. S. Kerley,*,1 J. D. Firman,* and G. L. Allee*

*Department of Animal Sciences, CAFNR, University of Missouri, Columbia, Missouri 65211 ABSTRACT Research was conducted to determine the effects of enzymatically hydrolyzing rafnose and stachyose from soybean meal (SBM) on fecal oligosaccharide concentration and growth performance of chicks fed a corn-SBM diet. The -galactosidase treatment was optimized for oligosaccharide degradation. Enzyme treatment degraded rafnose and stachyose in SBM by 69 and 54%, respectively, compared to untreated soybean meal (USBM). Diets containing enzyme-treated soybean meal (ESBM) resulted in excreta rafnose and stachyose concentrations reduced to below measurable levels (<0.1 mg/ g feces). Enzyme treatment increased (P < 0.05) TME from 2,974 to 3,328 kcal/kg. Three chick growth studies were conducted to determine the effect of feeding ESBM on growth performance. There were no statistical differences (P > 0.05) in growth performance among treatments. Chicks fed the ESBM diet had an increased (P < 0.05) fecal neutral detergent ber (NDF) content in one of two studies. A fourth experiment was conducted to determine if heating, used to enhance enzyme treatment, would decrease lysine availability. Heating signicantly (P < 0.05) reduced lysine availability compared to USBM. These experiments demonstrated that feces could be made void of rafnose and stachyose, but chick growth performance was not signicantly (P > 0.05) improved by enzyme treatment.

(Key words: galactosidase, soybean meal, enzyme, chick, oligosaccharide) 2002 Poultry Science 81:10141019

INTRODUCTION
Soybean meal (SBM) contains -galactosides (rafnose and stachyose) that cannot be digested in the small intestine of monogastrics due to the absence of -galactosidase (Gitzelmann and Auricchio, 1965). Rafnose and stachyose, cumulatively, make up approximately 6% of SBM dry matter. Even though these oligosaccharides are attributed, in part, as responsible for lowering the TME value of SBM for poultry (McGinnis, 1983), their removal has not resulted in improved growth performance (Irish et al., 1995). The oligosaccharides also cause feces to have hydroscopic properties, making the litter remain wet (Bedford, 1995). Leske et al. (1991) observed that chicks fed SBM supplemented with rafnose produced excreta that appeared to be more wet than that of chicks fed ESBM. The wet litter, conducive to bacterial growth, can cause footpad lesions and discoloration of the skin on the breast (Chesson, 1993), reducing the performance and value of the bird. These negative effects on production could be alleviated if rafnose and stachyose were not present in the excreta.

This research was conducted to determine if treating SBM with -galactosidase prior to mixing the diet would reduce rafnose and stachyose concentrations in the excreta. Growth performance effects and TME of the treated SBM were also measured in the chicks.

MATERIALS AND METHODS Chick Studies

Experiment 1 consisted of three treatment diets allotted randomly to six pens, with ve birds per pen for each diet. Experiment 2 used the same three treatment diets but had eight pens allotted randomly to each diet with ve birds per pen. All studies followed the experimental guidelines discussed below with diets fed for 14 d. Diets fed to chicks in Experiments 1 and 2 (Table 1) comprised 55% corn and 36% untreated soybean meal (USBM), or water-treated soybean meal (WSBM), or galactosidase-treated soybean meal (ESBM). WSBM was used to determine that effects of enzyme treatment were attributable to the enzyme and not the aqueous solution. Diets fed in Experiment 3 consisted of 49.6% corn and

2002 Poultry Science Association, Inc. Received for publication May 21, 2001. Accepted for publication February 13, 2002. 1 To whom correspondence should be addressed: kerleym@ missouri.edu.

Abbreviation Key: ADFI = average daily feed intake; ESBM = enzyme-treated soybean meal; NDF = neutral detergent ber; SBM = soybean meal; USBM = untreated soybean meal; WSBM = water-treated soybean meal.

1014

ENZYME TREATMENT OF SOYBEAN MEAL

TABLE 1. Ingredient composition of the diets fed in Experiments 1, 2, and 3 Ingredient Ground corn Soybean meal Corn oil Dicalcium phosphate Limestone Salt DL-Methionine Trace mineral1 Vitamin premix2 Choline chloride L-Lysine HCl Selenium premix3 Copper sulfate Crude protein (%)4 Lysine (%)4 Stachyose Rafnose Experiments 1 and 2 (% as fed) 55.20 35.90 5.20 1.90 0.85 0.41 0.21 0.10 0.07 0.07 0.05 0.03 0.01 21.0 1.3 1.4 0.4 Experiment 3 (% as fed) 49.60 39.40 6.00 2.60 1.20 0.50 0.35 0.10 0.10 0.10 ... 0.02 0.01 24.2 1.3 1.5 0.4

1015

1 Trace mineral premix contained 2.5% calcium (calcium carbonate) minimum, 3.5% calcium maximum, 6% iron (ferrous sulfate), 2.68% magnesium (magnesium oxide), 11% manganese (manganese sulfate), 11% zinc (zinc sulfate), and 2,000 ppm iodine (ethylenediamine dihydriodide). 2 Provided per kilogram of premix; 17,637 KIU vitamin A, 7,716 KIU vitamin D3, 27,558 IU vitamin E, 110,231 mg niacin, 33,069 mg D-pantothenic acid, 11,023 mg riboavin, 4,409 mg pyridoxine, 3,307 mg menadione, 2,756 mg folic acid, 2,205 mg thiamine, 441 mg biotin, 22 mg vitamin B12, 2% ow agent (rice hulls). 3 Selenium premix contained 38% calcium (calcium carbonate) and 600 mg/kg selenium. 4 Calculated from NRC (1994).

for that bird by dividing the total feed intake up until the day the bird died by the number of animals in the pen. Gain to feed was calculated by dividing the average daily gain by the average daily feed intake of the pen. Total feces were collected on the last 3 d of Experiments 1, 2, and 3, and stored in ziplock freezer bags at 0 C until analyzed for dry matter, oligosaccharides, and neutral detergent ber (NDF). A fourth chick experiment was conducted to determine if drying (55 C) the ESBM resulted in reduced lysine availability. The same feeding and handling procedures as the rst studies were followed, except that birds (ve birds per pen, six pens per treatment) began receiving experimental diets at 7 d of age. The study was conducted over 14 d. Basal diets in this experiment (Table 2) consisted of three increasing levels of lysine, added as L-lysine HCl. By using a slope ratio technique (Littell et al., 1995), USBM and ESBM were each fed at two dietary levels to determine lysine availability. Lysine percentages of diets were measured.

SBM Preparation
All SBM used in these studies was taken from the same batch load. The WSBM was prepared by spraying tap water 1:5 (vol/wt) on 5,000 g of SBM in a 15-L plastic jug, using a hand-held sprayer. The jug was then rolled for 1 h on a ball mill roller to facilitate complete coverage of the SBM. The -galactosidase treatment was accomplished by spraying a 4% (vol/wt) enzyme solution onto the SBM. The -galactosidase enzyme (49.6 mg protein/ mL) used in these studies released 0.1 M galactose per min/mL from stachyose. After mixing for 1 h, the SBM was spread out on wire screens covered in cheesecloth and air-dried at room temperature for 3 d (88% dry matter). In Experiment 4 only, ESBM was placed into a drying oven (55 C) for 48 h after application of -galactosidase.

39.4% WSBM or ESBM. All diets were formulated to meet or exceed the requirements according to the NRC (1994). Experiment 3 diets were formulated to meet 120% of the NRC amino acid requirement. All procedures involving animals were approved by the University of Missouri Animal Care and Use Committee. Day-old male broiler chicks were obtained from a commercial hatchery. Chicks were fed a commercial chick starter crumble diet,2 until the start of the trial. At 3 d of age, chicks were individually weighed and allotted randomly to pens. Birds were housed in raised-wire oor, stainless steel chick batteries within a temperaturecontrolled room. Temperature was 32.2 C for the rst 5 d and then reduced to 26.7 C over 2 d for the remainder of the study. The chicks were exposed to 24 h of light. Feed and water were offered ad libitum throughout the experiment. At the conclusion of the experiments, birds were weighed individually. Total gain was calculated by subtracting the beginning weight of the chicks from the nal weight. Daily feed intake was calculated by dividing the total pen intake by the number of birds in the pen by the number of days fed experimental diets. If a pen contained a dead bird, feed intake was calculated

Analyses
The TME analysis of the SBM was conducted with adult roosters following the method of Sibbald (1986). The TME value for each SBM treatment was the average of data from six roosters. Fecal samples were dried in a 55 C oven for 3 d and ground in a Wiley Mill ,3 through a 1-mm screen. Rafnose and stachyose were extracted using 85% methanol. A 2.5-g subsample of dried, ground fecal material was extracted in 25 mL of 85% methanol for 2 h. The liquid fractions of the samples were then ltered through glass microber lter paper with a pore size of 1.5 m. The retentate was extracted two additional times. The ltrate was then brought to a total volume of 100 mL in volumetric asks using 85% methanol. One milliliter of the total volume was evaporated to dryness under a stream of nitrogen, and 5 mL of distilled water was then added. One milliliter of the diluted sample was ltered through a 0.2-g Acrodisc ,4 and analyzed using a Beckman HPLC5 equipped with

2 3

MFA, Inc., Columbia, MO 65202. Arthur H. Thomas Co., Philadelphia, PA 19102. 4 Gelman Sciences, Inc., Ann Arbor, MI 48103. 5 Beckman Coulter, Inc., Fullerton, CA 92835.

1016
Basal12 ... 46.8 1.0 46.2 2.3 1.1 0.1 2.5

GRAHAM ET AL.
TABLE 2. Ingredient composition of the diets fed in Experiment 41 Basal Soybean meal Starch Corn oil Peanut meal Dicalcium phosphate Limestone L-Lysine HCl Premix3
1 2

Basal22 ... 46.6 1.0 46.2 2.3 1.1 0.3 2.5

USBM32 3.3 43.7 1.0 46.2 2.2 1.2 ... 2.5

USBM62 6.6 40.4 1.0 46.2 2.1 1.2 ... 2.5

ESBM32 3.3 43.7 1.0 46.2 2.2 1.2 ... 2.5

ESBM62 6.6 40.4 1.0 46.2 2.1 1.2 ... 2.5

... 46.9 1.0 46.2 2.3 1.1 ... 2.5

All numbers expressed on as-fed percentage of the diet. Basal1 and Basal2 = basal diet with 0.1 and 0.2% added lysine, respectively; SBM3 and SBM6 = basal diet with 3.3 and 6.6% SBM added, respectively; ESBM3 and ESBM6 = basal diet with 3.3 and 6.6% ESBM added. 3 Contents of premix based on percentage of the diet; 0.50% methionine, 0.50% NaCl, 0.39% threonine, 0.29% isoleucine, 0.27% valine, 0.14% choline chloride, 0.10% trace mineral premix [2.5% calcium (calcium carbonate) minimum, 3.5% calcium maximum, 6% iron (ferrous sulfate), 2.68% magnesium (magnesium oxide), 11% manganese (manganese sulfate), 11% zinc (zinc sulfate), and 2,000 ppm iodine (ethylenediamine dihydriodide)], 0.10% vitamin premix [provided per kg of premix: 17,637 KIU vitamin A, 7,716 KIU vitamin D3, 27,558 IU vitamin E, 110,231 mg niacin, 33,069 mg D-pantothenic acid, 11,023 mg riboavin, 4,409 mg pyridoxine, 3,307 mg menadione, 2,756 mg folic acid, 2,205 mg thiamine, 441 mg biotin, 22 mg vitamin B12, 2% ow agent (rice hulls)], 0.03% tryptophan, 0.003% selenium premix, 0.0003% copper sulfate, 0.18% lysine HCl.

a Dionex pulsed amperometric detector6 to determine the concentration of fecal rafnose and stachyose. Flow rate was 0.9 mL (50 mM NaOH)/min for 20 min. A Dionex Carbopac,6 column was used to separate oligosaccharides. The NDF (Goering and Van Soest, 1970) was determined in feces by using the ANKOM 200 ber analyzer.7 In Experiment 3, feces were collected daily from each pen on Days 8 through 10 of the experiment. Feed intake during the collection period was measured 24 h in advance of fecal collection for each pen. After collection, feces were immediately dried in a 55 C oven for 3 d and then ground through a Wiley Mill (1-mm screen). Samples were composited by pen and dry matter and calculated by drying feces at 105 C for 24 h. Percentage apparent digestibility was determined by dividing the feed dry matter apparently digested by the feed dry matter intake. The feed dry matter digested was calculated by subtracting the total dry matter excreted during the 3-d collection period from the feed dry matter consumed during that period.

procedure of SAS software (SAS Institute, 1996). The results in Experiment 4 were analyzed as a completely randomized design in a 3 2 factorial plus a basal diet. Regression analysis of basal, SBM, and ESBM was conducted by regressing total gain on added lysine intake. The methods of Littell et al. (1995) were used to determine curvature of the individual slopes, intersection at the y-axis, and differences between slopes. Means (GLM) were separated using the least squared difference procedure of SAS software. Statistical signicance was accepted at P 0.05.

RESULTS AND DISCUSSION Chick Experiments

Maximum hydrolysis of rafnose and stachyose was achieved by a 1:5 (vol/wt) ratio of 0.4% enzyme solution and SBM (data not reported). To ensure maximum removal of the oligosaccharides in SBM fed in these experiments, concentration of the enzyme solution used to treat SBM was increased 10-fold. This level of enzyme solution was shown in preliminary experiments to achieve maximum oligosaccharide hydrolysis. Treatment of SBM with -galactosidase reduced (P < 0.05) rafnose and stachyose by 65 and 50%, respectively. Fecal rafnose and stachyose in the ESBM diets were also signicantly reduced (P < 0.05) in chick feces (Table 3). These data illustrated that -galactosidase treatment of the SBM effectively removed rafnose and stachyose in the feces. Because these galactosides are the primary cause of fecal hydroscopic properties in poultry, their removal should alleviate the consequences associated with sticky droppings. Leske et al. (1991) observed that chicks fed SBM supplemented with rafnose or SBM that had been extracted with ethanol produced excreta that appeared wetter (sticky droppings) than that of chicks fed SBM treated with -galactosidase. Sticky droppings are attributed to

Statistics
The TME data were analyzed as a completely randomized design using the general linear models procedure of SAS software (SAS Institute, 1996). Individual animal data were the experimental unit for TME data. The results in Experiments 1, 2, and 3 were analyzed as a completely randomized design by using the general linear models procedure of SAS software (SAS Institute, 1996). Pens were the experimental units in Experiments 1, 2, and 3, and least squared means are reported. Differences among treatment means were determined by meanwise comparison using least squared difference

6 7

Dionex Corporation, Sunnyvale, CA 94085. ANKOM, 1997; Fairport, NY 14450.

ENZYME TREATMENT OF SOYBEAN MEAL

TABLE 3. Growth performance and fecal oligosaccharide composition of chicks fed enzyme treated and untreated soybean meal (Experiments 1, 2, and 3)1 Diet2 USBM WSBM ESBM SEM Initial wt (g) 75 74 75 0.6 Initial wt (g) USBM WSBM ESBM SEM 67 68 66 0.07 Initial wt (g) WSBM ESBM SEM
ac 1

1017

BW gain (g) 514.8 539.8 512.0 13.4 BW gain (g) 511.9 517.0 525.4 10.6 ADFI (g/d) 44.1 43.4 45.8 1.0 BW gain (g) 502.5 525.4 10.4

ADFI2 (g/d) 45.9 47.8 45.8 0.71

G/F Experiment 1

Fecal rafnose (%) 0.43a 0.48a 0b 0.05 Fecal stachyose (%) 0.33a 0.34a 0b 0.07 Fecal NDF (%) 24.90 25.32 0.003

Fecal stachyose (%) 1.36a 1.51a 0b 0.18 Fecal NDF (%) 28.39b 29.24b 30.92a 0.39

0.80 0.81 0.80 0.01 Experiment 2 Fecal G/F rafnose (%) 0.83 0.86 0.82 0.02 0.15a 0.09b 0c 0.02 Experiment 3 G/F 0.80 0.82 0.01

ADFI (g/d) 45.0 46.0 0.85

APDig (%) 71.39 72.07 0.004

72 73 0.5

Values within columns within experiments with unlike subscripts are signicantly different (P < 0.05). ADFI = Average daily feed intake per bird; G/F = gain to feed ratio; NDF = neutral detergent ber; APDig = apparent digestibility expressed on a dry matter basis as a percentage of intake. 2 Untreated soybean meal (USBM), water-treated soybean meal (WSBM), and enzyme-treated soybean meal (ESBM).

several problems in the poultry industry such as meat discoloration, foot problems, and increased disease. Wet droppings can also cause an increase in ammonia production, reducing air quality in poultry houses and attract ies and rodents (Choct, 1996). A decreased oligosaccharide concentration would potentially reduce viscosity of the digesta, leading to slower digesta passage rate (increased retention time), greater access of digestive enzymes to substrates, and more rapid diffusion of absorbable nutrients to the intes-

tinal mucosa. The result would be an increased digestibility of the diet. Supporting this theory was an increase (dry matter basis) in true metabolizable energy due to -galactosidase treatment. USBM had a TME of 2,974 35.8 kcal/g and ESBM had a TME of 3,328 145.4 kcal/ g. If the metabolizable energy density of the diet is increased, efciency of growth should be increased. The rafnose and stachyose may be the primary cause of lower TME values for SBM in chickens due to the poor digestibility of the complex sugars (Leske et al., 1991).

FIGURE 1. Effect of lysine intake per bird on body weight gain of chicks fed diets containing soybean meal that was heated during -galactosidase treatment. Equation for the chick growth response to the basal, untreated, and treated soybean meal diets was y = 751.34 (lysine intake) + 75.80 2 (r = 0.8576), y = 736.98 (lysine intake) + 74.15 (r2 = 0.7914), and y = 745.91 (lysine intake) + 54.12 (r2 = 0.6993), respectively.

1018

GRAHAM ET AL.
TABLE 4. Chick performance data and lysine intake of Experiment 41 Basal Initial wt (g) BW gain (g) ADG (g/d) ADFI (g/d) G/F DLI (g/d)
ae 1

Basal1 154.8 442.9b,c 36.48 51.58b 0.73a 0.46c

Basal2 155.8 510.3a 42.21 58.80a 0.72a 0.59a

USBM3 154.7 423.2c,d 34.33 51.42b 0.68b,c 0.46c

USBM6 154.3 497.3a,b 40.85 57.68a 0.72a,b 0.58a

ESBM3 156.6 358.5d,e 31.76 49.30b,c 0.66c 0.44c

ESBM6 155.5 439.9a,b 35.76 50.48b,c 0.72a 0.50b

SEM 5.7 20.3 1.18 1.54 0.01 0.01

154.6 341.1e 28.05 46.87c 0.60d 0.37d

Unlike superscripts within rows are signicantly different (P < 0.05). Basal1 and Basal2 = basal diet with 0.1 and 0.2% added lysine, respectively; SBM3 and SBM6 = basal diet with 3 and 6% SBM added, respectively; ESBM3 and ESBM6 = basal diet with 3 and 6% ESBM added; ADG = average daily gain; ADFI = average daily feed intake; G/F = gain to feed ratio; DLI = daily lysine intake.

Addition of rafnose and stachyose to a broiler diet signicantly reduced TME values compared to controls (Leske et al., 1993), but removal of rafnose and stachyose from SBM by using 80% ethanol extraction resulted in an increased TME for broiler chicks and roosters (Coon et al., 1990; Leske et al., 1991; Leske and Coon, 1999). These results were similar to results of our studies using -galactosidase to hydrolyze the oligosaccharides. The growth parameters gain, ADFI, and gain/feed among diets of Experiments 1, 2, and 3 (Table 3) did not differ signicantly (P > 0.05) due to -galactosidase treatments, although in Experiments 2 and 3, there was a trend (P < 0.15) for the ESBM diet to support a greater body weight gain over 14 d than diets containing nonenzyme-treated SBM and WSBM (Experiment 3). Similar to our data, chick performance was not improved with the addition of -galactosidase to a corn-soy diet in an experiment done by Irish et al. (1995). The higher average daily gain in Experiment 2 could be attributed to an increased feed intake. The bird has more potential to grow if it takes in more nutrients to support that growth. Even though a numerically improved growth performance was seen in Experiments 2 and 3, it was not concluded that the increased growth response was consistent. Fiber analysis performed on the feces of chicks in Experiment 2 and 3 showed an increased NDF percent for the ESBM diet. The increase in NDF was due to the removal of rafnose and stachyose from the feces (nonNDF constituents) causing the percentage of hemicellulose, cellulose and lignin (NDF constituents) to be elevated when compared to the other two diets. Increasing the digestibility of a diet should increase the amount of nutrients the bird can utilize for growth. Reduction of rafnose and stachyose using -galactosidase could improve digestibility by reducing viscosity of the digesta, or hydrolyzing the oligosaccharides into monosaccharides the bird can utilize. In these experiments, there was no signicant difference (P > 0.05) in apparent dry matter digestibilities between treatments. Based upon the reduction of rafnose and stachyose in the feces, an increased dry matter digestibility was expected. However, the lack of a response in digestibility explains why growth parameters were not increased due to -galactosidase treatment. The lack of growth response due to

-galactosidase treatment is in agreement with Irish et al. (1995). Heating (55 C) compared to not heating SBM after -galactosidase application further decreased rafnose and stachyose concentrations from 35 and 50% to 7 and 16%, respectively. This result may have occurred because heating increased enzyme activity. Experiment 4 was conducted to determine if heating reduced lysine availability of -galactosidase treated SBM. Based upon statistical differences in slopes at a 95% condence interval, heating SBM in Experiment 4 produced a lower gain for chicks fed ESBM compared to chicks fed the USBM and basal diets (Table 4). Daily lysine intake and ADFI was also lower in the ESBM diet. Lysine bioavailability of USBM was 98.7% of the basal diet measured by regressing lysine intake on body weight gain (r2 = 0.7914). Lysine bioavailability of -galactosidase treated SBM, determined using the same procedure, was 75.9% of the basal diet (r2 = 0.6993; Figure 1). The birds fed the ESBM diets did not consume enough lysine to meet their growth needs due to the decreased bioavailability, which explains the smaller average daily gain. Heating ESBM decreased further the oligosaccharide concentration in SBM and increased the concentration of free sugars. These sugars contain a carbonyl group that can bind to free amino acids or to the amino group of lysine, forming a Maillard product that is intestinally unavailable. This conversion may have been the reason that lysine availability was reduced in heat-treated compared to nonheat-treated ESBM in this study. These studies indicated that addition of the enzyme -galactosidase to poultry diets without heat treatment could eliminate rafnose and stachyose in feces, which may help improve growth parameters in young chicks and may also reduce digesta viscosity and, therefore, reduce occurrence of sticky droppings. If heat treatment is used to enhance oligosaccharide removal from SBM, consideration should be given to effects on lysine availability.

REFERENCES
Bedford, M. R. 1995. Mechanism of action and potential environmental benets from the use of feed enzymes. Anim. Feed Sci. Technol. 53:145155.

ENZYME TREATMENT OF SOYBEAN MEAL Chesson, A. 1993. Feed enzymes. Anim. Feed Sci. Technol. 45:6579. Choct, M. 1996. Enzymes in animal nutrition: The unseen benets. Proceedings of the 1st Chinese Symposium on Feed Enzymes, Nanjing, Peoples Republic of China. Coon, C. N., K. L. Leske, O. Akavanichan, and T. K. Cheng. 1990. Effect of oligosaccharide-free soybean meal on true metabolizable energy and ber digestion in adult roosters. Poult. Sci. 69:787793. Gitzelmann, R., and S. Auricchio. 1965. The handling of soy alpha-galactosidase by a normal and galactosemic child. Pediatrics 36:231232. Goering, H. K., and P. J. Van Soest. 1970. Forage ber analysis (apparatus, reagents, procedures, and some applications). Agricultural Handbook 379. ARS, USDA, Washington, DC. Irish, G. G., G. W. Barbour, H. L. Classen, R. T. Tyler, and M. R. Bedford. 1995. Removal of the -galactosides of sucrose from soybean meal using either ethanol extraction or exogenous -galactosidase and broiler performance. Poult. Sci. 74:14841494. Leske, K. L., and C. N. Coon. 1999. Nutrient content and protein and energy digestibilities of ethanol-extracted, low -galactoside soybean meal as compared to intact soybean meal. Poult. Sci. 78:11771183.

1019

Leske, K. L., O. Akavanichan, T. K. Chen, and C. N. Coon. 1991. Effect of ethanol extract on nitrogen-corrected true metabolizable energy for soybean meal with broilers and roosters. Poult. Sci. 70:892895. Leske, K. L., C. J. Jevne, and C. N. Coon. 1993. Effect of oligosaccharide additions on nitrogen-corrected metabolizable energy of soy protein concentrate. Poult. Sci. 72:664668. Littell, R. C., A. J. Lewis, and P. R. Henry. 1995. Statistical evaluation of bioavailability assays. Pages 533 in Bioavailability of Nutrients for Animals. C. B. Ammerman, D. H. Baker, and A. J. Lewis, ed. Academic Press, New York. McGinnis, J. 1983. Carbohydrate utilization in feedstuffs. Pages 106107 in Proceedings of the Minnesota Nutrition Conference, St. Paul, MN. National Research Council. 1994. Nutrient requirements of poultry. 9th rev. ed. National Academy Press, Washington, DC. SAS Institute. 1996. SAS/STAT Users Guide. Software Release 6.12. SAS Institute Inc., Cary, NC. Sibbald, I. R. 1986. The T.M.E. system of feed evaluation: Methodology, feed composition data and bibliography. Agriculture Canada Technical Bulletin 1986-4E. Agriculture Research Center, Agriculture Canada, Ottawa, ON, Canada.