A Net Carbohydrate and Protein System for Evaluating Cattle Diets: IV. Predicting Amino Acid Adequacy

Transcription

1 A Net Carbohydrate and Protein System for Evaluating Cattle Diets: IV. Predicting Amino Acid Adequacy J. D. O'Connorl, C. J. Sniffen2, D. G. Fox3, and W. Chalupa4 Department of Animal Science, Cornell University, Ithaca, NY ABSTRACT The Cornell Net Carbohydrate and Protein System was modified to include an amino acid submodel for predicting the adequacy of absorbed essential amino acids in cattle diets. Equations for predicting the supply of and requirements for absorbed essential amino acids are described and presented. The model was evaluated for its ability to predict observed duodenal flows of nitrogen, nonammonia nitrogen, bacterial nitrogen, dietary nonammonia nitrogen, and individual essential amino acids. Model-predicted nitrogen, nonammonia nitrogen, bacterial nitrogen, and dietary nonammonia nitrogen explained 93.2, 94.6, 76.4, and 79.3% of the observed duodenal flows, respectively, based on R2 values from predicted vs observed regression analysis. Based on slopes of regression lines, model-predicted duodenal nitrogen and nonammonia nitrogen were different from observed duodenal flows (P <.05), whereas model-predicted bacterial nitrogen and dietary nonammonia nitrogen were not different from observed duodenal flows ( P <.05). Model-predicted duodenal flows of individual essential amino acids explained 81 to 90% of variation in observed duodenal amino acid flows. Based on slopes of regression lines, modelpredicted duodenal threonine, leucine, and arginine were the only amino acids different from observed duodenal flows (P <.05). Ideas for further model improvements and research in amino acid metabolism were also presented. Key Words: Cattle, Nutrition, Models, Amino Acids J. Anim. Sci : Introduction Amino acid requirements have been determined primarily using breakpoint estimates of animal response to dietary amino acid supplementation. However, extrapolating such estimates to dietary, animal, and environmental conditions that are different from the experimental conditions under which the requirements were estimated has little scientific basis (Owens and Pettigrew, 1989). An alternative is the use of dynamic, factorial mathematical models, which allows prediction of amino acid demand and supply under a wide variety of dietary, animal, and environmental conditions (Owens and Pettigrew, 1989). Russell et al. (1992), Sniffen et al. (19921, and Fox et al. (1992) have developed a carbohydrate and protein simulation model that determines ME and metabolizable protein ( IMP) requirements and supply under widely varying dietary, animal, and environmental conditions. Included are submodels that predict the supply of and the requirements for absorbed amino acids, using a combination of empirical and mechanistic approaches. The objectives of this research were 1) to develop an amino acid submodel for the Cornell Net Carbohydrate and Protein System (CNCPS; Fox et al., 1992; Russell et al., 1992; Sniffen et al., 1992) so that it that can be used to predict the supply of and requirements for absorbed essential amino acids in cattle diets and 2) to validate the model for predicted duodenal flows of nitrogen, nonammonia nitrogen, bacterial nitrogen, dietary nonammonia nitrogen, and individual essential amino acids for a variety of cattle diets. Methods and Data 'Present address: P.O. Box 2077, Corvallis, OR 'Present address: Miner Agricultural Institute, Chazy, NY 3To whom correspondence should be addressed. 4Dept. of Clinical Studies, School of Vet. Med., Univ. of Pennsylvania, Kennett Square Received September 10, Accepted December 18, Model Development The CNCPS was modified to include an amino acid submodel that predicts the daily supply of and requirements for absorbed essential amino acids. The amino acid submodel uses a factorial approach for determining the daily supply of and requirements for 1298

2 PREDICTING AMINO ACID ADEQUACY OF CATTLE DIETS 1299 metabolizable amino acids. The daily supply of absorbed essential amino acids is based on the CNCPS submodels described by Russell et al. (1992) and Sniffen et al. (1992). The daily requirements for absorbed essential amino acids are based on the CNCPS submodel described by Fox et al. (1992) and parameter estimates from Evans and Patterson (19851, NRC (19851, Mantysaari et al. (19891, and Ainslie et al. (1993). Amino Acid Supply Amino acids available for absorption and ultimately for cattle production are supplied by microbial protein synthesized in the rumen, dietary protein escaping ruminal degradation, and endogenous secretions into the digestive tract (Richardson and Hatfield, 1978). The models of Russell et al. (1992) and Sniffen et al. (1992) were used as the basis of the model presented here for estimating the supply of amino acids available from microbial protein synthesized in the rumen and dietary protein escaping ruminal degradation. Amino acids available from endogenous secretions were not included in the model because of limited quantitative information describing the supply and availability of amino acids from endogenous secretions of cattle (Mantysaari et al., 1989). Thus, in this model, the prediction of amino acids available from specific cattle diets was dependent on the estimation of amino acids available from ruminal microbial protein and dietary protein escaping ruminal degradation. Accurate estimation of the amino acids available from microbial protein synthesized in the rumen requires reliable estimation of ruminal microbial yield, microbial protein composition, amino acid composition of microbial protein, and the digestibility of microbial amino acids. The microbial growth model described by Russell et al. (1992) in a previous paper demonstrated that ruminal microbial yield may be reliably estimated using a Michaelis-Menton kinetic model; therefore, the microbial yield model described by Russell et al. (1992) was used as the basis for estimating ruminal microbial yield in this model. In addition to a prediction of ruminal microbial yield, estimates of the protein composition of ruminal bacteria were necessary to determine the amounts of microbial protein produced in the rumen. Ruminal bacteria were assumed to be 10% N or 62.5% CP (Isaacson et al., 1975). However, the distribution of nitrogen in bacterial protein is variable, depending on species and conditions of growth (Van Soest, 1982). Ruminal bacterial protein was assumed to be 25% cell well nitrogen (Bergen et al., 1967) and 15% nucleic acid nitrogen (Purser and Buechler, 1966). The remaining 60% of bacterial nitrogen was assumed to be true protein (Van Soest, 1982). The amino acid compositions of ruminal microbial cell wall and non-cell wall protein fractions need to be Amino acid Table 1. Amino acid composition of ruminal microbial cell wall and non-cell wall protein (g/100 g of protein) Methionine Lysine Histidine Phenylalanine Tryptophan Threonine Leucine Isoleucine Valine Areinine Cell wall b Non-cell wall Ruminal bacteriaa aaverage composition of 441 bacterial samples from animals fed 61 dietary treatments in 35 experiments (Clark et al., 1992). bdata were not available; therefore, content of cell wall protein was assumed to be the same as that of non-cell wall protein. estimated to determine bacterial amino acids appearing at the duodenum and amino acids available from bacterial protein. Amino acids are present in both the cell walls and cytoplasm of bacteria; however, most of the amino acids are contained in non-cell wall material (Chalupa, 1972). The amino acid content of microbial cell wall protein was estimated based on reports by Hoogenraad and Hird (1970) and Hoogenraad et al. (1970). The amino acid content of non-cell wall microbial protein was estimated based on a summary of bulk ruminal microbial protein by Mantysaari et al. (1989). Estimates of the amino acid content of ruminal microbial cell wall and non-cell wall protein used in the model are given in Table 1. These values represent the amino acid content of freefloating bacteria recovered by differential centrifugation but are in good agreement with ruminal bacteria composition values reported by Clark et al. (1992). The amount of bacterial amino acids appearing at the duodenum was determined by multiplying the amino acid content of each bacterial protein fraction by the quantity of each bacterial protein fraction produced in the rumen. The quantity of each bacterial protein fraction appearing at the duodenum was predicted using the previously described model of Russell et al. (1992). The amounts of bacterial amino acids appearing at the duodenum were determined using the following equation: n REBAAi = C (AABCWi *.01 * REBCWj) j=l + (AABNCWi *.01 * REBTPj), where AABCWi = ith amino acid content of ruminal bacteria cell wall protein, g/loo g; AABNCW; = ith amino acid content of ruminal bacterial non-cell wall protein g/loo g; REBCWj = bacterial cell wall protein appearing at the duodenum as a result of fermentation

3 1300 O CONNOR ET AL. of the jth feedstuff, g/d; REBTPj = bacterial non-cell wall protein appearing at the duodenum as a result of fermentation of the jth feedstuff, g/d; and REBAAi = amount of the ith bacterial amino acid appearing at the duodenum, gid. In cattle, the small intestine is the active site of absorption of amino acids (NRC, 1985); however, few estimates exist of the true digestibility of ruminal bacterial cell wall and non-cell wall protein fractions in the small intestine of cattle. Tas et al. (1981) estimated the true absorption of microbial amino acids in the small intestine of sheep to be.87 by regression analysis. Storm et al. (1983) found that the true digestibility of amino acid nitrogen varied little and estimated the average true digestibility of amino acid nitrogen in the small intestine of sheep to be by regression analysis. Microbial cell contents are probably completely digestible (Van Soest, 1982); therefore, amino acids contained in non-cell wall bacterial protein were assumed to be completely digested in the intestines. Amino acids contained in the cell walls of bacteria may not be released by proteolytic enzymes in the abomasum and small intestine and may be of limited value to the animal (Allison, 1970; Mason and White, 1971). Therefore, it was assumed that amino acids contained in the bacterial cell wall protein were completely unavailable for digestion in the intestines. Amino acids available from microbial protein were estimated by multiplying the quantity of amino acids in the cell wall and non-cell wall bacterial protein delivered to the duodenum by the intestinal true digestibility of each bacterial protein fraction. However, because microbial cell wall protein and amino acids were considered to be indigestible, available microbial amino acids were only provided by the noncell wall bacterial protein, which was considered to be completely digestible. Available or absorbed bacterial amino acids were determined using the following equation: escaping ruminal degradation. Feed physical characteristics, chemical composition, and dry matter intake have been identified as primary factors affecting the amount of protein degraded in the rumen (Sniffen et al., 1992). Dietary protein escaping ruminal degradation was predicted using the model of Sniffen et al. (1992) described in a previous paper. The amino acid composition of the dietary protein escaping ruminal degradation also affects the amounts of dietary amino acids appearing at the duodenum. Because a significant portion of the protein contained in many feedstuffs is degraded in the rumen, Mac- Gregor et al. (1978) hypothesized that the amino acid composition of dietary protein escaping ruminal degradation may be better represented by the amino acid composition of the insoluble protein rather than by the amino acid composition of the total protein. Other researchers have also demonstrated that the amino acid profile of the insoluble protein is different from the amino acid profile of the total protein in feedstuffs (Muscato et al., 1983; Bozak et al., 1986; Schwab et al., 1986; Crooker et al., 1987; Mantysaari et al., 1989). Estimates of the amino acid content of the insoluble protein from dietary protein sources appearing at the duodenum for a limited number of feeds are shown in Table 2, based on the data of MacGregor et al. (1978), Mantysaari et al. (19891, and Muscato et al. (1983). Amino acids appearing at the duodenum from dietary protein sources were calculated by multiplying the amino acid content of the insoluble protein by the quantity of each protein fraction escaping ruminal degradation. Amounts of protein fractions escaping ruminal degradation were predicted using the model of Sniffen et al. (1992). The following equation was used to determine dietary amino acids escaping ruminal degradation and appearing at the duodenum: n DIGBAAi = C AABNCWi * j=1.01 * REBTPj, where DIGBAAi = amount of the ith absorbed bacterial amino acid, gid. In addition to an accurate estimate of amino acids available from microbial protein, an accurate assessment of the amino acids available from dietary protein escaping ruminal degradation is needed to determine total amino acids available to the animal. Amino acids available from dietary protein escaping ruminal degradation can be estimated from the amount of dietary protein escaping degradation, the amino acid composition of insoluble protein, and the true digestibility of dietary amino acids in the intestines. The amount of dietary protein escaping ruminal degradation is a primary factor in determining dietary amino acids appearing at the duodenum. A number of factors may affect the amount of dietary protein * (REPBlj + REPB2j + REPB3j + REPCj), where AAINSPij = ith amino acid content of the insoluble protein for the jth feedstuff, gil00 g; REPBlj = B1 protein from the jth feedstuff that escaped from the rumen, g/d; REPBZj = B2 protein from the jth feedstuff that escaped from the rumen, g/d; REPB3j = B3 protein from the jth feedstuff that escaped from the rumen, g/d; REPCj = C protein from the jth feedstuff that escaped from the rumen, gid; and REFAAi = amount of ith dietary amino acid that appeared at the duodenum, g/d. Amounts of amino acids appearing at the duodenum were calculated as the sum of amino acids from bacterial protein and dietary protein escaping ruminal degradation and were determined using the following equation:

4 PREDICTING AMINO ACID ADEQUACY OF CATTLE DIETS 1301 Table 2. Amino acid content of the insoluble protein in several common feeds (g/100 g of insoluble protein)a Essential amino acid Feed name MET LYS ARG THR LEU ILE VAL HIS PHE Source Dry corn grain Corn hominy Ground milo Heavy barley Ground wheat Wheat middlings Oats Alfalfa meal Brewers dry grain Corn distillers dryholuble Corn gluten feed Corn gluten meal Whole cottonseed Cottonseed meal Soybean meal (44% CP) Roasted soybeans Feather meal Fish meal Blood meal Legume hay Legume silage Grass hay Corn silage Beet Dulu b,d 4.88 e 4.95 e 5.60 e 4.84 e 4.74 e 5.86 d 4.68 e 4.80 c,d 4.20 d 1.68 d 6.48 b,d 5 e 5.47 e 4.94 c,d 4.81 e 5.21 b 4.33 b 7.86 b 6.32 e 4.18 c,d 3.50 C 2.94 b,d 2.80 d adata of MacGregor et al. (19783 indicate that the amino acids in the insoluble protein are similar to those in the protein escaping ruminal degradation. bmantysaari et al. (1989). CMacGregor et al. (1978). dmuscato et al. (1983). W. Chalupa, personal communication. Univ. of Pennsylvania. REAAi = REBAAi + REFAAi where REAAi = total amount of the ith amino acid appearing at the duodenum, g/d. After amino acids from dietary protein escaping ruminal degradation have been determined, absorbed amino acids from dietary protein escaping ruminal degradation can be determined. Absorbed amino acids from dietary protein escaping ruminal degradation were calculated by multiplying the amino acid content of insoluble protein for each feedstuff by the amounts of each protein fraction that were digested in the intestines for each feedstuff. The amounts of each protein fraction that were digested in the intestines for each feedstuff were predicted using the model of Sniffen et al. (1992). The model of Sniffen et al. (1992) assumed that the bound protein fraction is not digested and the B1 and B2 fractions are completely digested in the intestines. The B3 protein fraction is assumed to be less than completely digested. The B1, B2, and B3 fractions are assigned a true digestibility of 100, 100, and 8096, respectively, based on the types of proteins in each fraction as described by Sniffen et al. (1992). The following equation was used to calculate absorbed amino acids from dietary protein escaping ruminal degradation: n DIGFAAi = C AAINSPij *.01 j=l * (REPBlj + REPB2j +.8 * REPB3j), where DIGFAAi = amount of the ith absorbed amino acid from dietary protein escaping rumen degradation, gld. The total supply of absorbed amino acids was calculated by summing intestinally digested microbial amino acids and intestinally digested amino acids from dietary protein escaping ruminal degradation. The following equation was used to determine the total supply of absorbed amino acids: w i = DIGBAAi + DIGFAAi, where w i = total amount of the ith absorbed amino acid supplied by dietary and bacterial sources, g/d. Amino Acid Requirements Few quantitative models of amino acid requirements for cattle have been proposed; however, recent models (Oldham, 1980; Evans and Patterson, 1985; Mantysaari et al., 1989) have attempted to quantify amino acid requirements by using a factorial method

5 1302 O CONNOR ET AL. that parallels the classical approach of Mitchell ( The factorial method calculates absorbed amino acid requirements from net nitrogen requirements, amino acid composition of products, and efficiencies of utilization of absorbed amino acids for product formation. Net protein requirements can be calculated as protein deposited in tissue and conceptus and secreted in milk plus protein used for maintenance, which was estimated from endogenous urinary, scurf, and metabolic fecal protein (Mantysaari et al., 1989). The net protein requirements used in this model were based on requirement equations presented in the model described by Fox et al. (1992) in a previous paper. Absorbed amino acid requirements were calculated by multiplying net protein requirements by the amino acid composition of the product formed and dividing by the efficiencies of use of individual absorbed amino acids for product formation. Absorbed amino acids required for scurf protein were determined from the net scurf protein requirement, the amino acid content of scurf protein, and the efficiencies of use of individual amino acids for scurf protein formation. Net requirement for scurf protein represents nitrogen loss due to skin, hair, horn, and detritus (NRC, 1985). The net requirement for scurf protein was calculated from the model described by Fox et al. (1992). Previous models (Evans and Patterson, 1985; Mantysaari et al., 1989) used the amino acid content of tissue protein to determine net amino acid requirements for growth; however, net scurf protein requirements may be more closely related to the amino acid composition of keratin than of muscle tissue because scurf losses include skin, hair, and nails (Owens and Pettigrew, 1989). Estimates of the amino acid content of keratin protein may be obtained from several sources (Mitchell, 1950; Block and Bolling, 1951; Block and Weiss, 1956). The model presented here used estimates based on the average keratin composition of cattle hair, horn, hooves, and skin (Table 3) based on the data of Block and Bolling (1951). The efficiencies of use of individual absorbed amino acids for scurf protein formation were based on estimates of the efficiencies of use of individual amino acids for maintenance given by Evans and Patterson (1985) and are listed in Table 4. The following equations were used for calculating the ith absorbed amino acid requirement for scurf protein: SPN =.2 W.6 SPAAi = AAKERAi *.01 * SPN/EAAMi, where W = live weight, kg/d; SPN = net scurf protein requirement, g/d; AAKERAi = amino acid content of the ith amino acid in keratin protein, g/loo g; EAAMi = efficiency of use of the ith amino acid for maintenance; Table 3. Amino acid composition of tissue, milk, and keratin protein (g/100 g of protein] Amino acid Tissuea Milkb KeratinC Methionine Lysine Histidine Phenylalanine Tryptophan Threonine Leucine Isoleucine Valine Arginine aaverage of three studies summarized by whole empty body values of Ainslie et al. (1993). baghorn and Baldwin (1984). %lock and Bolling (1951). and SPAAi = scurf requirement for the ith absorbed amino acid, g/d. The net requirement for endogenous urinary protein is the nitrogen (protein equivalent) lost in the urine when cattle are fed nitrogen-free diets (NRC, 1985). Endogenous urinary protein components include creatine, urea, ammonia, allantoin plus bilirubin, nucleic acids, hippuric acid, and small quantities of some amino acids such as N7-methyl histidine (NRC, 1985; Owens and Pettigrew, 1989); therefore, urinary nitrogen is probably only a partial amino acid expense (Owens and Pettigrew, 1989). The net requirement for urinary protein was calculated based on the model by Fox et al. (1992). On protein-deficient diets, urinary nitrogen loss is primarily derived from skeletal muscle (Uezu et al., 1985); therefore, the amino acid content of tissue protein as summarized by Mantysaari et al. (1989) was used to determine net amino acids required for urinary nitrogen loss. The efficiencies of use of individual Table 4. Utilization of individual absorbed amino acids for physiological functions (g/g) Amino acid Maintenance Gestation Lactation Methionine Lysine Histidine Phenylalanine Tryptophan Threonine Leucine Isoleucine Valine Arginine oo arequirement for growth varies with stage of growth as determined by Ainslie et al. (1993); EF =.83 - (.00114EBW); EF is efficiency factor and EBW is equivalent body weight as described by Fox et al. (1992). Other values are from Evans and Patterson (1985).

6 PREDICTING AMINO ACID ADEQUACY OF CATTLE DIETS 1303 absorbed amino acids for urinary protein formation were assumed to be the efficiencies of use of individual amino acids for maintenance given by Evans and Patterson (1985). Estimates of the efficiencies of use of individual absorbed amino acids for maintenance are given in Table 4. The following equations were used to calculate the ith absorbed amino acid required for urinary protein: UPN = 2.75 W.5 UPAAi = AATISSi *.01 * UPN/EAAMi, especially for rapidly growing animals (Buttery, 1979; Owens and Pettigrew, 1989). The model presented here uses values from an average of three studies of whole-body tissue composition summarized by Ainslie et al. (1993) in a companion paper. Estimates of the amino acid content of tissue protein are given in Table 3. Estimates of the efficiencies of use of absorbed amino acids for tissue deposition are based on values determined by Ainslie et al. (1993) and are given in Table 4. The following equations were used to calculate the ith absorbed amino acids required for growth: where UPN = urinary protein required, g/d; and UPAAi = urinary requirement for the ith absorbed amino acid, gld. Absorbed amino acids required for metabolic fecal protein ( MFP) of animal origin are those required for replacement of sloughed epithelial cells, mucus secretions, and bile pigments (CSIRO, 1990). After an extensive literature review, CSIRO (1990) adopted a value of 15.2 g of absorbed protein requiredkg of DMI. The NRC (1985) calculated the absorbed protein requirement for metabolic fecal protein to be 90 g of absorbed proteinkg of indigestible DM, which agrees with the CSIRO value at 83% diet digestibility. The German system (Rohr and Lebzien, 1991) adopted an MFP requirement of 18.2 g of net proteinkg of DMI, which agrees with the NRC (1985) value at 70% digestibility, assuming a 67% effkiency of use of absorbed protein for maintenance. We adopted the NRC (1985) requirement for MFP; it seems to agree with these other independently determined values. The amino acid content of tissue listed in Table 1 is used to determine amino acids required for metabolic fecal protein. The following equations were used to calculate the ith absorbed amino acid required for metabolic fecal protein requirement: FPA = -09 * IDM FPAAi = AATISSi *.01 * FPA, where IDM = indigestible dry matter, gid; FPA = requirement for absorbed metabolic fecal protein, gld; AATISS = tissue content of amino acids (Table 3); and FPAAi = metabolic fecal requirement for the ith absorbed amino acid, g/d. Absorbed amino acids required for tissue growth were determined from the net protein required for growth, the amino acid content of tissue protein, and the efficiencies of use of absorbed amino acids for tissue protein formation. Net protein required for tissue growth has been determined by body composition of growing animals and is a multiple of weight gain and composition of the gain (NRC, 1985). Net protein required for growth was predicted using the equations of Fox et al. (1992) that were presented in a previous paper. The amino acid composition of muscle and carcass tissues seems to be similar across animal species (Williams et al., 1954; Smith, 19801, RPN = PB *.01 * EG RPAAi = AATISSi * RPN/EAAGi, where PB = protein content of empty body gain, g/loo g; EG = empty body gain, gid; RPN = net protein required for growth, g/d; EAAGi = efficiency of use of the ith amino acid for growth, glg; and RPAAi = growth requirement for the ith absorbed amino acid, g/d. Absorbed amino acids required for lactation were determined from the net protein required for lactation, the amino acid content of milk true protein, and the efficiencies of use of individual amino acids for lactation. Net protein required for lactation was calculated from the true protein content of milk and the expected daily milk production level according to the requirements model of Fox et al. (1992). Estimates of the amino acid composition of milk protein were necessary to determine net and absorbed amino acids for lactation. The average amino acid composition of milk true protein for lactating dairy cows has been reported by various researchers (Block and Bolling, 1951; Featherston et al., 1964; Jacobson et al., 1970; Broderick et al., 1974; Burroughs et al., 1974; Hogan, 1975; Lampert, 1975; McCance and Widdowson, 1978; Evans and Patterson, 1985). The amino acid composition of milk protein seems to be remarkably constant for different animal species (Reeds, 1988) and seems to be independent of diet and stage of lactation (Featherston et al., 1964). The model described here used the average amino acid composition of milk true protein, as summarized by Waghorn and Baldwin (19841, given in Table 3. The efficiencies of use of individual absorbed amino acids for lactation were based on the data of Evans and Patterson (19851, except that values for leucine, isoleucine, valine, and arginine were based on those suggested by Oldham (1980). Efficiencies of use of individual amino acids for milk production are given in Table 4. The requirement of the ith absorbed amino acid for lactation was calculated using the following equations: LPN = PP * MM * 10 where PP = milk true protein content, g/100 g; MM = expected daily milk production level, kg/d; AALACTi =

7 1304 O'CONNOR ET AL ith amino acid content of milk true protein, g/loo g; EAALi = efficiency of use of the ith amino acid for milk protein formation, g/g; and LPAAi = lactation requirement for the ith absorbed amino acid, g/d. Absorbed amino acids required for gestation can be determined from the net protein requirement for gestation, the amino acid content of tissue protein, and the efficiencies of use of individual amino acids for tissue protein formation. Net protein requirements for gestation were determined using equations from Fox et al. (1992) presented in a previous paper. The amino acid composition of conceptus products is required to calculate amino acids required for gestation. No estimates were found in the literature for the various components of the conceptus; therefore, the amino acid contents of the products of conception were assumed to be similar to the average amino acid composition of tissue. The estimates of amino acid content of tissue from Mantysaari et al. (1989) that are listed in Table 3 were used as estimates of the amino acid content of conceptus products. The effciencies of use of individual amino acids for gestation were based on the data of Evans and Patterson (1985) and are given in Table 4. The following equation was used to calculate the ith absorbed amino acid required for gestation: YPAAi = AATISSi *.01 * YPNEAAPi, where YPN = net protein required for gestation, gid; EAAPi = efficiency of use of the ith amino acid for gestation, gig; and YPAAi = gestation requirement for the ith absorbed amino acid, gid. The total absorbed requirement for an amino acid can be determined as the sum of the absorbed amino acid requirements for physiological functions and were calculated according to the following equation: AAAri = SPAAi + UPAAi + FPAAi + RPAAi + YPAAi + LPAAi, where AAA,.i = total absorbed requirement for the ith amino acid, gid. Amino Acid Balance After the supply of, and requirements for, absorbed amino acids have been determined, an assessment of the limiting amino acids can be made. Metabolizable protein supplies essential and nonessential amino acids. However, a comparison of absorbed essential amino acids supplied and required is necessary; metabolizable protein requirements may seem to have been met by the supply of metabolizable protein when essential amino acid requirements have not been met. Model Validation The model was validated by comparing modelpredicted duodenal flows of individual amino acids with observed experimental duodenal flows. Predicted and observed flows were determined from the data reported by Stern et al. (1983, 19851, Prange et al. (19841, Santos et al. (19841, Pena et al. (19861, Windschitl and Stern (19881, Zerbini et al. (19881, McCarthy et al. (19891, Waltz et al. (19891, King et al. (19901, Klusmeyer et al. (19901, and Cameron et al. (1991) for lactating cattle and reported by Williams and Smith (19741, Fenderson and Bergen (19751, Cottrill et al. (19821, Garrett et al. (19871, Rooke and Armstrong (19871, Bernard et al. (19881, Cecava et al. (1988, 19901, and Titgemeyer et al. (1988) for nonlactating cattle. Reported feed ingredient composition was used in model evaluations when information was available. When feed composition was not available, estimates were obtained from Sniffen et al. (1992). This was necessary because ruminal rates of carbohydrate fermentation, protein degradation, and passage were assigned based on values reported by Sniffen et al. (1992). Amino acid composition of the undegraded feed protein content of ingredients was obtained from Crooker et al. (1987), MacGregor et al. (1978), Mantysaari et al. (19891, and Muscat0 et al. (1983). Lactating cattle experimental reports included diets for early-, mid-, and late-lactation cows; however, most lactating cattle experiments investigated duodenal flows for mid- to late-lactation Holstein cows. Characteristics of lactating cattle experiments used in the model evaluation are summarized in Table 5. Nonlactating cattle experimental reports included diets for Holstein, Friesian, Simmental, Jersey, Jersey and Holstein crossbred, and Simmental and Holstein crossbred cattle ranging in body weight from 125 to 424 kg. Characteristics of nonlactating cattle experiments are summarized in Table 6. Supplements fed in the experiments included blood meal, soybean meal, whole soybeans, roasted soybeans, corn gluten feed, corn gluten meal, cottonseed meal, whole raw cottonseeds, extruded cottonseeds, roasted cottonseeds, wet brewers grains, dried distillers grain, feather meal, fish meal, linseed meal, soyhulls, wheat midds, and urea. Comparisons were made by regressing observed duodenal flows of individual amino acids against model-predicted duodenal flows of individual amino acids. Regression analysis was performed with Quattro Pro version 3.0. Intercepts were computed to interpret how well the model-simulated data fitted the actual data. A regression slope of 1 and an intercept of 0 for the regression equation describing the relationship between observed duodenal flow and modelpredicted duodenal flow would indicate perfect agreement. Differences in the regression equations from that indicating perfect agreement were evaluated by testing for difference from a slope of 1 by using a twotailed Student's t-test. The model predicted duodenal flow for an individual amino acid was characterized as

8 ~ ~~ ~ PREDICTING AMINO ACID ADEQUACY OF CATTLE DIETS Table 5. Characteristics of lactating cattle experiments used in model validation 1305 Item n Mean Range SD ~ No. of studies 12 ~ - ~~ - - Live wt, kg Dry matter intake, kg/d Milk yield, kg/d N intake, g/d Duodenal N, g/d Duodenal nonammonia N, g/d Duodenal bacterial N, g/d Duodenal diet nonammonia N, g/d Duodenal methionine, g/d Duodenal lysine, g'd Duodenal threonine, g/d Duodenal isoleucine, g/d Duodenal leucine, g/d Duodenal valine, g/d Duodenal histidine, g/d Duodenal phenylalanine, gi'd Duodenal arginine, g/d Microbial efficiency, gkg of RFOMa aruminal fermentable organic matter. different from the observed duodenal flow when the regression slope between the observed and predicted values differed from 1 (P c.05). Figure 1 depicts observed and model-predicted duodenal nitrogen flows. The line through the origin with a slope of 1 and an intercept of 0 indicates perfect agreement between observed and model-predicted values. The second line depicts the regression equation describing the relationship between observed and model-predicted duodenal flow. Upon visual inspection, there seems to be good agreement between observed and model-predicted duodenal nitrogen flows. The model explained approximately 93% of the variation in observed duodenal nitrogen flow. However, statistically, the model predicted duodenal nitrogen flow was different from the observed nitrogen flows. Based on the slope of the regression line, the model seemed to underpredict observed duodenal nitrogen flow; however, based on visual inspection of Figure 1, the graph shows that the model overpredicted at low duodenal nitrogen flows and underpredicted at high duodenal nitrogen flows. The Table 6. Characteristics of nonlactating cattle experiments used in model validation Item n Mean Range SD No. of studies Live wt, kg Live wt gain, kg/d Dry matter intake, kg'd N intake, g/d Duodenal N, g/d Duodenal nonammonia N, g'd Duodenal bacterial N, g/d Duodenal diet nonammonia N, gid Duodenal methionine, g/d Duodenal lysine, g/d Duodenal threonine, g/d Duodenal isoleucine, g/d Duodenal leucine, g/d Duodenal valine, g/d Duodenal histidine, g/d Duodenal phenylalanine, g/d Duodenal arginine, g/d Microbial efficiency, gkg of RFOMa aruminal fermentable organic matter.

9 ~ O'CONNOR ET AL. 0 -roo4 NITROGEN Y = l.o88x, S E =5 1.6, R2 = PREDICTED DUODENAL FLOW (&a I BACTERIAL NITROGEN Y = l.l31x, SE45.3, R2 = = " g 350- i l:w% Lgi 200- o 2 I I00 # 5 0 W... B o I Figure 1. Relationship between observed and predicted duodenal nitrogen flow. Figure 3. Relationship between observed and predicted duodenal bacterial nitrogen flow. negative intercept in the regression equation implies an overprediction of observed flows at low duodenal nitrogen flows. Figure 2 depicts observed and model-predicted duodenal nonammonia nitrogen flows. Upon visual inspection, there seems to be good agreement between observed and model-predicted duodenal nonammonia nitrogen flows. The model explained approximately 95% of the variation in observed duodenal nonammonia nitrogen flow. However, statistically, the modelpredicted duodenal nonammonia nitrogen flow was different from the observed duodenal nonammonia nitrogen flow. Based on the slope of the regression line, the model seemed to underpredict observed duodenal nonammonia nitrogen flow; however, based on visual inspection of Figure 2, the model overpredicted at low duodenal nonammonia nitrogen flows and underpredicted at high duodenal nonammonia nitrogen flows. The negative intercept in the regression equation implies an overprediction of observed flows at low duodenal nonammonia nitrogen flows. Figure 3 depicts observed and model-predicted duodenal bacterial nitrogen flows. Upon visual inspection, there seems to be good agreement between observed and model-predicted duodenal bacterial nitrogen flows. The model explained approximately 76% of the variation in observed duodenal bacterial nitrogen flow. The model-predicted duodenal bacterial nitrogen flow was not different from the observed duodenal bacterial nitrogen flow. Based on the slope of the regression line, the model seemed to underpredict observed duodenal bacterial nitrogen flow; however, based on visual inspection of Figure 3, the model overpredicted at low duodenal bacterial nitrogen flows NON-AMMONIA NITROGEN Y = X, S E =44.6, R2 =.946 DIETARY NON-AMMONIA NITROGEN Y= X, SEz44.8, R2z g 350 ii 300 a 250 I Figure 2. Relationship between observed and predicted duodenal nonammonia nitrogen flow. E w 250 I o Figure - 4. Relationship between observed and predicted duodenal dietary nonammonia nitrogen flow. LO

10 PREDICTING AMINO ACID ADEQUACY OF CATTLE DIETS 1307 METHIONINE Y = X, SE=8.4, R2= 2 HISTIDINE Y= X, SEz9.8, R2=.830 a -2oJ io b 60 io io 90 id0 :0 Figure 5. Relationship between observed and predicted duodenal methionine flow. Figure 7. Relationship between observed and predicted duodenal histidine flow. and underpredicted at high duodenal bacterial nitrogen flows. The negative intercept in the regression equation implies an overprediction of observed flows at low duodenal bacterial nitrogen flows. Figure 4 depicts observed and model-predicted duodenal dietary nonammonia nitrogen flows. Upon visual inspection, there seems to be good agreement between observed and model-predicted duodenal dietary nonammonia nitrogen flows. The model explained approximately 79% of the variation in observed duodenal dietary nonammonia nitrogen flow. The model-predicted duodenal dietary nonammonia nitrogen flow was not different from the observed duodenal bacterial nitrogen flow. Based on the slope of the regression line, the model seemed to underpredict observed duodenal dietary nonammonia nitrogen flow; however, based on visual inspection of Figure 4, the model overpredicted at low duodenal dietary nonammonia nitrogen flows and underpredicted at high duodenal dietary nonammonia nitrogen flows. The negative intercept in the regression equation implies an overprediction of observed flows at low duodenal dietary nonammonia nitrogen flows. Figures 5 to 13 depict observed and model-predicted duodenal essential amino acid flows. Upon visual inspection, model-predicted duodenal flows of individual amino acids seem to agree with observed flows. Regression equation coefficients and R2-values of observed vs predicted individual amino acid duodenal flows are presented in Table 7. The R2-values for amino acids indicated that the model explained 81 to 90% of the variation in observed duodenal amino acid flows.. Model-predicted duodenal flows of threonine, leucine, and arginine were the only amino acids LYSINE Y= %, SE=21.5, R2=.892 " -50! I oo i5 lai! n P Q 60- z 40- w 8 2;- PHENYLALANINE Y = X, SE = 1 7.9, R2 =.885 o io iao 2 IO Figure 6. Relationship between predicted duodenal lysine flow. observed and Figure 8. Relationship between observed and predicted duodenal phenylalanine flow.

11 1308 O'CONNOR ET AL. THREONINE Y= X, SE=15.5, R2=.898 ISOLEUCINE Y= X, SEz18.2, R2=.877 B i $ B 40- w g 2; ~~ PREDICTH) DUODENAL FLOW (ad) IO Figure 9. Relationship between observed and predicted duodenal threonine flow. Figure 11. Relationship between observed and predicted duodenal isoleucine flow. different from observed duodenal flows. The slopes of the regression lines for threonine and leucine were > 1, indicating that the model underpredicted the duodenal flows of these two amino acids, whereas the slope of the regression line for arginine was < 1, indicating that the model overpredicted the duodenal flow of this amino acid. Most of the amino acid regression equations included a negative intercept, implying that the model tended to overpredict slightly duodenal amino acids at low duodenal flows. Based on the slopes of the regression lines, the following order of accuracy in predicting observed duodenal flows (from most accurate to least accurate) was established for the model: histidine, phenylalanine, valine, methionine, isoleucine, lysine, threonine, leucine, and arginine. Discussion Overall, the high R2-values and slopes near unity for regression equations imply that the model performed adequately in predicting observed duodenal nitrogen, nonammonia nitrogen, bacterial nitrogen, dietary nonammonia nitrogen, and essential amino acids. Given that most of the regression equations had slopes > 1 and negative intercepts (Table 71, this implied that the model had a tendency to underpredict most duodenal quantities at high duodenal flows, which represented lactating cattle diets, and overpredict duodenal quantities at low duodenal flows, which represented nonlactating cattle diets. The model presented here is one attempt to quantify the supply of absorbed amino acids from dietary protein escaping ruminal degradation and LEUCINE Y= X, SE=33.3, R2= VALINE Y= X, SEz22.7, R2= PREDICTED DUODENAL FLOW (g/d) Figure 10. Relationship between observed and predicted duodenal leucine flow. -50! Figure 12. Relationship between observed and predicted duodenal valine flow.

12 8 PREDICTING AMINO ACID ADEQUACY OF CATTLE DIETS 1309 ARGININE Y= %, SEz21.8, R2= I PREDICTED DUODENAL FLOW (g/d) Figure 13. Relationship between observed and predicted duodenal arginine flow. microbial protein synthesized in the rumen and the daily requirements for absorbed amino acids by cattle in various productive states. The strengths of the model are probably in the following areas: 1) estimates of the amino acid composition of milk, tissue, keratin, and bacterial proteins seem to be relatively constant under different dietary regimens and are based on numerous research reports, and 2) model predictions of amino acids supplied by dietary protein that escaped from the rumen and microbial protein are based on theoretical, mechanistic principles of ruminal carbohydrate fermentation, microbial yield, protein degradation, and passage. However, the model presented here may be improved by further research in the areas listed below that affect the supply of absorbed amino acids by specific cattle diets and the daily requirements for absorbed amino acids for various stages of production More research is needed to define quantitatively the components of metabolic fecal protein and the amino acid composition of these components to predict more accurately the amino acids required for metabolic fecal protein losses. The amino acid content of tissue protein may need to be further defined based on the various components of empty body protein. Research is needed to determine the amino acid content of the various products of the conceptus to describe more accurately amino acids required for gestation. Additional research is needed to determine more accurate estimates of efficiencies of utilization of absorbed amino acids for specific physiological functions, especially because these efficiencies have a larger effect on requirements for absorbed amino acids (NRC, 1985). Further research on the amino acid content of soluble and insoluble available dietary protein escaping ruminal degradation for various natural and by-product feedstuffs is required to increase the accuracy of predicting available amino acids supplied by specific cattle diets. Additional research is necessary to determine accurate estimates of the amino acid composition of bacterial cell wall and non-cell wall protein fractions. More research needs to be conducted concerning the true digestibility of bacterial cell wall and noncell wall protein fractions. Implications A model has been presented for predicting the demand for, and supply of, absorbed essential amino acids in cattle diets. Identification and effective supplementation of limiting amino acids may result in Table 7. Regression statistics for observed vs model predicted duodenal flows (g/d) Quantity Intercept X coeff. SEa R2 n SEb Nitrogen *.036, Nonammonia N *,033, Bacterial N , Dietary nonammonia N Methionine Lysine Histidine Phenylalanine Threonine *.043, Leucine * Isoleucine , Valine , Arginine *,051, astandard error of the X coefficient. bstandard error of the Y estimate. *Model-predicted duodenal flows different from observed flows (P <.05).

13 1310 O CONNOR ET AL. the following benefits to the cattle industry: 1) increased weight gains for growing cattle, 2 improved milk and milk protein production for lactating cattle, 3) increases in the efficiency of protein utilization, 4) maintenance of production with less protein consumption, and 5 ) lowered feed costs lower protein contents of diets. Further improvements will require more research to determine accurate estimates of parameters that affect amino acid utilization and requirements of cattle in various productive states. Literature Cited Ainslie, S. J., D. G. Fox, T. C. Perry, D. J. Ketchen, and M. C. Barry Predicting amino acid adequacy of diets fed to Holstein steers. J. him. Sci. 71:1312. Allison, M. J In: A. T. Phillipson (Ed.) Physiology of Digestion and Metabolism in the Ruminant. p 456. Oriel, Newcastle upon Tyne, U.K. Bergen, W. G., D. B. Purser, and J. H. Cline Enzymatic determination of the protein quality of individual rumen bacteria. J. Nutr. 92:357. Bernard, J. K., H. E. Amos, M. A. Froetschel, and J. J. Evans Influence of supplemental energy and protein on protein synthesis and crude protein reaching the abomasum. J. Dairy Sci. 71:2658. Block, R. J., and D. Bolling The Amino Acid Composition of Proteins and Foods (2nd Ed.) Thomas, Springfield, IL. Block, R. J., and K. W. Weiss Amino Acid Handbook. Thomas, Springfield, IL. Bozak, C. K., C. G. Schwab, and J. E. Nocek Change in amino acid pattern of feed protein upon exposure to rumen fermentation using the in situ bag technique. J. Dairy Sci. 69(Suppl. 1): 108 (Abstr.), Broderick, G. A., L. D. Satter, and A. E. Harper Use of plasma amino acid concentration to identify limiting amino acids for milk production. J. Dairy Sci. 57:1015. Burroughs, W., A. TrenMe, and R. L. Vetter A system of protein evaluation for cattle and sheep involving metabolizable protein (amino acids) and urea fermentation potential of feedstuffs. Vet. Med. Small Anim. Clin. 69:713. Buttery, P. J Amino acids and other nitrogenous compounds. In: M. Recheigl (Ed.) Comparative Animal Nutrition. Vol. 3. p 34. S. Karger, Basel, Switzerland. Cameron, M. R., T. H. Klusmeyer, G. L. Lynch, J. H. Clark, and D. R. Nelson Effects of urea and starch on rumen fermentation, nutrient passage to the duodenum, and performance of cows. J. Dairy Sci. 74:1321. Cecava, M. J., N. R. Merchen, L. L. Berger, and G. C. Fahey, Jr Effects of dietary energy level and protein source on site of digestion and duodenal nitrogen and amino acid flows in steers. J. Anim. Sci Cecava, M. J., N. R. Merchen, L. L. Berger, and D. R. Nelson Effect of energy level and feeding frequency on site of digestion and post-ruminal nitrogen flows in steers. J. Dairy Sci. 73:2470. Chalupa, W Metabolic aspects of nonprotein nitrogen utilization in ruminant animals. Fed. Proc. 31:1152. Clark, J. H., T. H. Klusmeyer, and M. R. Cameron Microbial protein synthesis and flows of nitrogen fractions to the duodenum of dairy cows. J. Dairy Sci Cottrill, B. R., D. E. Beever, D. R. Austin, and D. F. Osbourn The effect of protein and non-protein nitrogen supplements to maize silage on total amino acid supply in young cattle. Br. J. Nutr. 48:527. Crooker, B. A,, J. H. Clark, R. D. Shanks, and G. C. Fahey, Jr Effects of ruminal exposure on the amino acid profile of feeds. Can. J. Anim. Sci. 67:1143. CSIRO Feeding Standards for Australian Livestock. Ruminants. CSIRO publications, East Melbourne, Victoria, Australia. Evans, E. H., and R. J. Patterson Use of dynamic modelling seen as good way to formulate crude protein, amino acid requirements for cattle diets. Feedstuffs 57(42):24. Featherston, W. R., D. R. Frazeur, D. L. Hill, C. H. Noller, and C. E. Parmelee Constancy of amino acid composition of cow s milk protein under changing ration. J. Dairy Sci. 47:1417. Fenderson, C. L., and W. G. Bergen An assessment of essential amino acid requirements of growing steers. J. him. Sci. 41:1759. Fox, D. G., C. J. Sniffen, J. D. O Connor, J. B. Russell, and P. J. Van Soest A net carbohydrate and protein system for evaluating cattle diets Cattle requirements and diet adequacy. J. Anim. Sci. 70:3578. Garrett, J. E., R. D. Goodrich, J. C. Meiske, and M. D. Stern Influence of supplemental nitrogen source on digestion of nitrogen, dry matter and organic matter and on in vivo rate of ruminal protein degradation. J. Anim. Sci. 64:1801. Hogan, J. P Quantitative aspects of nitrogen utilization in ruminants. J. Dairy Sci. 58:1164. Hoogenraad, N. J., and F. J. Hird The chemical composition of rumen bacteria and cell walls from rumen bacteria. Br. J. Nutr. 24:119. Hoogenraad, N. J., F. J. Hird, R. G. White, and R. A. Leng Utilization of I4C-labelled Bacillus subtilis and Escherichi coli by sheep. Br. J. Nutr. 24:129. Isaacson, H. R., F. C. Hinds, M. P. Bryant, and F. N. Owens Efficiency of energy utilization by mixed rumen bacteria in continuous culture. J. Dairy Sci. 58:1645. Jacobson, D. R., H. H. Van Horn, and C. J. Sniffen Lactating ruminants. Fed. Proc. 29:35. King, K. J., J. T. Huber, M. Sadik, W. G. Bergen, A. L. Grant, and V. L. King Influence of dietary protein sources on the amino acid profiles available for digestion and metabolism in lactating cows. J. Dairy Sci. 73:3208. Klusmeyer, T. H., R. D. McCarthy, Jr., J. H. Clark, and D. R. Nelson Effects of source and amount of protein on ruminal fermentation and passage of nutrients to the small intestine of lactating cows. J. Dairy Sci. 73:3526. Lampert, L. M Modern Dairy Products. Chem. Publ. Comp. Inc., New York. MacGregor, C. A,, C. J. Sniffen, and W. H. Hoover Amino acid profiles of total and soluble protein in feedstuffs commonly fed to ruminants. J. Dairy Sci. 61:566. Mantysaari, P. E., C. J. Sniffen, and J. D. O Connor An application model to balance amino acids for dairy cattle. Feedstuffs 61(20):13. Mason, V. C., and F. White The digestion of bacterial mucopeptide constituents in sheep. I. The metabolism of 2,6 diaminopimelic acid. J. Agric. Sci. (Camb.) 77:91. McCance, R. A,, and E. Widdowson In: D.A.T. Southgate and A. A. Paul (Ed.) The Composition of Foods (4th Ed.). HMSO, London. McCarthy, R. D., Jr., T. H. Klusmeyer, J. L. Vicini, J. H. Clark, and D. R. Nelson Effects of source of protein and carbohydrate on ruminal fermentation and passage of nutrients to the small intestine of lactating cows. J. Dairy Sci Mitchell, H. H In: A. A. Albanese (Ed.) Protein and Amino Acid Requirements of Mammals. Academic Press, New York. Muscato, T. V., C. J. Sniffen, U. Krishnamoorthy, and P. J. Van Soest Amino acid content of nonce11 and cell wall fractions in feedstuffs. J. Dairy Sci. 66:2198. NRC Ruminant Nitrogen Usage. National Academy Press, Washington, DC. Oldham, J. D Amino acid requirements for lactation in highyielding dairy cows. In: W. Haresign (Ed.) Nutrition. Butterworths, London.

14 PREDICTING AMINO ACID ADEQUACY OF CATTLE DIETS 1311 Owens, F. N., and J. E. Pettigrew Subdividing amino acid requirements into portions for maintenance and growth. In: M. Friedman (Ed.) Absorption and Utilization of Amino Acids. Vol. I. CRC Press, Boca Raton, FL. Pena, F., H. Tagari, and L. D. Satter The effect of heat treatment of whole cottonseed on site and extent of protein digestion in dairy cows. J. Anim. Sci. 62:1423. Prange, R. W., M. D. Stern, N. A. Jorgensen, and L. D. Satter Site and extent of protein digestion in lactating cows fed alfalfa silage or baled alfalfa hay. J. Dairy Sci Purser, D. B., and S. M. Buechler Amino acid composition of rumen organisms. J. Dairy Sci. 49:81. Reeds, P. J Nitrogen metabolism and protein requirements. In: K. Blaxter and I. MacDonald (Ed.) Comparative Nutrition. p 59. John Libbey, London. Richardson, C. R., and E. E. Hatfield, The limiting amino acids in growing cattle. J. Anim. Sci. 46:740. Rohr, K., and P. Lebzien Present knowledge of amino acid requirements for maintenance and production. Proc. 6th Int. Symp. Protein Metabolism and Nutrition. Herning, Denmark. Rooke, J. A., and D. G. Armstrong The digestion by cattle of silage and barley diets containing increasing quantities of fishmeal. J. Agric. Sci. (Camb.) 109:261. Russell, J. B., J. D. O Connor, D. G. Fox, P. J. Van Soest, and C. J. Sniffen A net carbohydrate and protein system for evaluating cattle diets: I. Ruminal fermentation. J. him. Sci. 70: Santos, K. A., M. D. Stern, and L. D. Satter Protein degradation in the rumen and amino acid absorption in the small intestine of lactating dairy cattle fed various protein sources. J. Anim. Sci Schwab, C. G., C. K. Bozak, and J. E. Nocek Change in amino acid pattern of soybean meal and corn gluten meal upon exposure to rumen fermentation. J. him. Sci. 63 (Suppl. 1):158 (Abstr.). Smith, R. H Comparative amino acid requirements. Proc. Nut,r. SOC. 39:71. Sniffen, C. J., J. D. O Connor, P. J. Van Soest, D. G. Fox, and J. B. Russell A net carbohydrate and protein system for evaluating cattle diets: 11. Carbohydrate and protein availability. J. Anim. Sci. 70:3562. Stern, M. D., L. M. Rode, R. W. Prange, R. H. Stauffacher, and L. D. Satter Ruminal protein degradation of corn gluten meal in lactating dairy cattle fitted with duodenal T-type cannulae. J. h im. Sci. 56:194. Stern, M. D., K. A. Santos, and L. D. Satter Protein degradation in rumen and amino acid absorption in small intestine of lactating dairy cattle fed heat-treated whole soybeans. J. Dairy Sci. 68:45. Storm, E., D. S. Brown, and E. R. Orskov The nutritive value of rumen-microorganisms in ruminants. 3. The digestion of microbial and nucleic acids in, and losses of endogenous nitrogen from, the small intestine of sheep. Br. J. Nutr. 50:479. Tas, M. V., R. A. Evans, and R.F.E. Exford The digestibility of amino acids in the small intestine of the sheep. Br. J. Nutr. 45: 167. Titgemeyer, E. C., N. R. Merchen, L. L. Berger, and L. E. Deetz Estimation of lysine and methionine requirements of growing steers fed corn silage-based or corn-based diets. J. Dairy Sci. 71:421. Uezu, N., S. Yamamoto, T. Rikimaru, K. Kishi, and G. Ionoue Contributions of individual body tissues of nitrogen excretion in adult rats fed protein-deficient diets. J. Nutr. 113:105. Van Soest, P. J The Nutritional Ecologv of the Ruminant. O&B Books, Corvallis, OR. Waghorn, G. C., and R. L. Baldwin Model of metabolite flux with mammary gland of the lactating cow. J. Dairy Sei Waltz, D. M., M. D. Stern, and D. J. Illg Effect of ruminal protein degradation of blood meal and feather meal on the intestinal amino acid supply to lactating cows. J. Dairy Sci. 72: Williams, A. P., and R. H. Smith Concentrations of amino acids and urea in the plasma of the ruminating calf and estimation of the amino acid requirements. Br. J. Nutr Williams, H. H., L. V. Curtin, J. Abraham, J. K. Loosli, and L. A. Maynard Estimation of growth requirements for amino acids by assay of the carcass. J. Biol. Chem. 208:277. Windschitl, P. M., and M. D. Stern Evaluation of calcium lignosulfontate-treated soybean meal as a source of rumenprotected protein for dairy cattle. J. Dairy Sci. 71:3310. Zerbini, E., C. E. Polan, and J. H. Herbein Effect of dietary soybean meal and fish meal on protein digesta flow in Holstein cows during early and midlactation. J. Dairy Sci. 71:1248.