Growth Performance of Broilers Using a Phase-Feeding Approach with Diets Switched Every Other Day from Forty-Two to Sixty-Three Days of Age 1

Growth Performance of Broilers Using a Phase-Feeding Approach with Diets Switched Every Other Day from Forty-Two to Sixty-Three Days of Age 1 T. Pope, L. N. Loupe, J. A. Townsend, and J. L. Emmert 2 Department of Poultry Science, University of Arkansas, Fayetteville, Arkansas 72701 ABSTRACT Phase-feeding (PF; Treatment 2) was tested differ (P > 0.05) among birds fed PF treatments and Treatment relative to an NRC regimen (Treatment 1) to evaluate the impact of PF on broilers during the finisher period (42 to 63 d). Two modified PF treatments were also tested during this experiment: Treatment 3 involved lowering the amino acid requirements predicted with PF linear regression equations by 10% (PF10), Treatment 4 involved increasing the slope of the linear regression equations by 15% (PF15). Broilers fed Treatment 1 consumed a single diet throughout the experiment, whereas PF treatments were fed in a series of 11 diets in which feed was switched every other day, resulting in steadily decreasing lysine, SAA, and threonine levels. On Day 63, weight gain, feed intake, feed efficiency, and crude protein intake did not 1. Digestible lysine and threonine intakes were in- creased (P < 0.05) for Treatment 1 relative to the PF treatments, and digestible SAA intake for the NRC treatment was increased (P < 0.05) relative to the PF10 and PF15 treatments. Compared to Treatment 1, no differences (P > 0.05) in percentage breast, wing, leg, or abdominal fat were observed when birds were fed PF, PF10, or PF15 diets. These results indicate that PF supports maximum growth performance during an extended finisher period, even when lysine, SAA, and threonine levels are reduced every other day. Economic analysis indicated that PF regimens may lead to savings ($ per kg weight gain or breast meat). (Key words: broiler, phase-feeding, growth performance, amino acid, feeding program) 2002 Poultry Science 81:466 471 INTRODUCTION The poultry industry is evolving every day, and it is the challenge of poultry nutritionists to keep up with this fast pace. The NRC (1994) provides a single set of recommendations that includes both males and females, with dietary amino acid requirements segregated into three fixed periods including the starter (0 to 3 wk), grower (3 to 6 wk), and finisher phases (6 to 8 wk). Although useful as a reference, these recommendations may be difficult for nutritionists to apply to commercial nutrition programs. Phase-feeding (PF) is a system described by Emmert and Baker (1997) that was designed to meet the needs of diversified poultry production systems. The PF system is based on Illinois Ideal Chick Protein (IICP; Baker and Han, 1994; Baker, 1997) and NRC (1994) amino acid recommendations. For PF, regression equations are used to express digestible amino acid requirements as a function of age; thus, it is possible to predict amino acid requirements for any period of time, even beyond 8 wk 2002 Poultry Science Association, Inc. Received for publication June 21, 2001. Accepted for publication November 5, 2001. 1 Supported by the Arkansas Agricultural Experiment Station, Fayetteville, AR 72701. 2 To whom correspondence should be addressed: jemmert@uark.edu. of age. Because amino acid levels decrease steadily throughout the grow-out period, PF is designed to eliminate excess amino acids, thus potentially reducing dietary costs and nitrogen excretion. Previous research has concluded that PF is a good means of reducing dietary feed costs during the grower and finisher phases (Warren and Emmert, 2000; Pope and Emmert 2001; Pope and Emmert, unpublished data). In these experiments, no differences in growth performance or carcass yield have been detected between birds fed PF- or NRC-based diets, and PF has appeared to reduce production costs ($ per kg weight gain or breast meat). However, in previous studies PF has been evaluated using diets switched weekly. Our objective was to evaluate growth performance of broilers using a PF approach with diets switched every other day during the finisher period (42 to 63 d). MATERIALS AND METHODS All procedures were approved by the University of Arkansas Institutional Animal Care and Use Committee. The experiment was conducted with male broiler chicks Abbreviation Key: HN = high nutrient; IICP = Illinois Ideal Crude Protein; LN = low nutrient; PF = phase-feeding; SAA = sulfur amino acids; SBM = soybean meal. 466

PHASE-FEEDING 467 TABLE 1. Composition of experimental diets fed to broilers from 42 to 63 d of age Ingredient NRC 1 PF HN 2 PF LN 2 PF10 HN 3 PF10 LN 3 PF15 HN 4 PF15 LN 4 Corn 72.87 69.93 77.14 73.07 79.56 69.93 78.22 Soybean meal 20.77 23.73 16.58 20.64 14.21 23.73 15.51 Poultry fat 3.00 3.00 3.00 3.00 3.00 3.00 3.00 Vitamin mix 5 0.20 0.20 0.20 0.20 0.20 0.20 0.20 Mineral mix 5 0.10 0.10 0.10 0.10 0.10 0.10 0.10 Dicalcium phosphate 1.20 1.20 1.20 1.20 1.20 1.20 1.20 Limestone 1.30 1.30 1.30 1.30 1.30 1.30 1.30 NaCl 0.30 0.30 0.30 0.30 0.30 0.30 0.30 Choline Cl (60%) 0.10 0.10 0.10 0.10 0.10 0.10 0.10 DL-Methionine 0.0639 0.1200 0.0603 0.0862 0.0325 0.1206 0.0520 L-Threonine 0.0921 0.0199 0.0178 0.0069 0.0049 0.0195 0.0170 1 NRC diets contained lysine, sulfur amino acid, and threonine levels recommended by NRC (1994). 2 Phase-feeding (PF) diets were formulated to contain lysine, sulfur amino acid, and threonine levels predicted by linear regression equations (Table 2) for 41-to-43-d-old [PF high-nutrient (HN)] or 61-to-63-d-old [PF lownutrient (LN)] broilers. Experimental diets were produced by blending PF HN and PF LN diets in variable quantities (see Materials and Methods). 3 Adjusted phase-feeding (PF10) diets were predicted by linear regression equations and lowered by 10% (Table 2) for 41-to-43-d-old (PF10 HN) or 61-to-63-d-old (PF10 LN) broilers. Experimental diets were produced by blending PF10 HN and PF10 LN diets in variable quantities (see Materials and Methods). 4 Adjusted phase-feeding (PF15) diets were predicted by linear regression equations that were modified to contain 15% more negative slope after the initial week of the experiment (Table 2) for 41-to-43-d-old (PF15 HN) or 61-to-63-d-old (PF15 LN) broilers. Experimental diets were produced by blending PF15 HN and PF15 LN diets in variable quantities (see Materials and Methods). 5 Emmert et al. (1999). (%) (Cobb Cobb 3 ) that were purchased from a local hatchery. Chicks were housed in floor pens containing new pine shavings. A 24-h photoperiod was maintained, and water and experimental diets were freely available. Dietary treatments (Tables 1 and 2) consisted of the following: 1. A single diet, formulated to contain NRC (1994) recommendations for lysine, sulfur amino acids (SAA), and threonine, that was fed for the entire 3 wk experiment. 2. A series of 11 diets (PF), formulated to contain lysine, SAA, and threonine levels predicted by linear regression equations (Emmert and Baker, 1997) with dietary amino acid concentration lowered every other day of the experiment. 3. A series of 11 diets (PF10), formulated to contain 10% less lysine, SAA, and threonine than those contained in PF diets (Treatment 2), with dietary amino acid concentrations lowered every other day of the experiment. 4. A series of 11 diets (PF15), formulated to contain lysine, SAA, and threonine levels predicted by linear regression equations (Emmert and Baker, 1997) and modified to contain a slope that was increased by 15% (after 42 d of age). Regression equations from Emmert and Baker (1997) were modified to reflect male requirements and were used to predict every other day PF (Treatment 2) as follows: digestible lysine, y = 1.22 0.0095x; digestible methionine and cystine, y = (0.88 0.0063x)/2; and digestible threonine, y = 0.8 0.0053x, where y = digestible amino 3 Cobb-Vantress, Inc., Siloam Springs, AR 72761. acid level, and x = midpoint (day) of the desired age range. Amino acid requirements used to derive the equations have been described previously (Emmert and Baker, 1997; Warren and Emmert, 2000). Regression equations used to predict digestible lysine, SAA, and threonine requirements for Treatment 4 were modified to reflect a 15% increased (more negative) slope beginning at Day 42 of the experiment. The slope was modified for the purpose of predicting lower amino acid requirements as the experiment progressed and was not intended for predicting requirements for ages younger than 42 d. In fact, application of the modified equation to birds younger than 42 d would result in amino acid requirement predictions substantially higher than levels known to be efficacious. We were only interested in using the modified equations to predict lysine, SAA, and threonine requirements from Days 42 to 63, during which time predicted requirements were progressively lower for PF15 relative to PF. The modified regression equations were digestible lysine, y = 1.25 0.01093x; digestible methionine and cystine, y = (0.90 0.00725x)/2; and digestible threonine, y = 0.82 0.0061x, where y = digestible amino acid level, and x = midpoint (day) of the desired age range. For all treatments corn and soybean meal (SBM) were added in sufficient quantities to meet the target digestible lysine concentration, and crystalline methionine and threonine were supplemented as needed to meet their requirements. Because NRC (1994) recommendations are based on total dietary amino acid needs, after formulating the NRC diet to meet total amino acid recommendations the digestible lysine, SAA, and threonine content was calculated by applying digestibility coefficients for corn and SBM. Corn was analyzed (as fed) to contain 7.86% CP,

468 POPE ET AL. TABLE 2. Calculated digestible amino acid levels Digestible content (% of diet) 1 CP ME n 2 Diet Days Lysine Methionine Cystine Threonine (%) (kcal/kg) NRC 3 42 to 63 0.74 0.28 0.26 0.61 16.2 3,212 PF HN 4 41 to 43 0.82 0.31 0.31 0.58 17.5 3,186 PF LN 5 61 to 63 0.63 0.24 0.24 0.47 14.4 3,253 PF10 HN 4 41 to 43 0.74 0.28 0.28 0.52 16.2 3,215 PF10 LN 5 61 to 63 0.57 0.22 0.22 0.42 13.4 3,276 PF15 HN 4 41 to 43 0.82 0.31 0.31 0.58 17.5 3,186 PF15 LN 5 61 to 63 0.60 0.24 0.24 0.46 14.0 3,263 1 Digestible amino acid, CP, and dietary ME content calculated from the analytical values for total lysine, sulfur amino acids, and threonine in corn and soybean meal and published digestibility coefficients (Parsons, 1991; see Materials and Methods). 2 Metabolizable energy values for corn, soybean meal, and poultry fat were assumed to be 3,350, 2,440, and 8,800 kcal ME n /kg, respectively. 3 Although the NRC (1994) provides total dietary amino acid recommendations, digestible amino acid levels were calculated after formulation of diets to meet total NRC (1994) recommendations for dietary lysine, sulfur amino acids, and threonine. 4 Phase-feeding high-nutrient (PF HN, PF10 HN, PF15 HN) diets were formulated to contain lysine, sulfur amino acid, and threonine levels predicted by linear regression equations for 41- to 43-d-old broilers, although the experiment did not begin until Day 42. Experimental diets were produced by blending HN and LN diets in variable quantities (see Materials and Methods). 5 Phase-feeding low nutrient (PF LN, PF10 LN, PF15 LN) diets were formulated to contain lysine, sulfur amino acid, and threonine levels predicted by linear regression equations for 61 to 63-d-old broilers. Experimental diets were produced by blending HN and LN diets in variable quantities (see Materials and Methods). 0.26% total lysine, 0.15% total methionine, 0.17% total cystine, and 0.27% total threonine, and the digestibility of lysine, methionine, cystine, and threonine in corn was assumed to be 78, 91, 86, and 84%, respectively (Parsons, 1991). Soybean meal was analyzed (as fed) to contain 50.5% CP, 3.18% total lysine, 0.71% total methionine, 0.79% total cystine, and 1.89% total threonine, and the digestibility of lysine, methionine, cystine, and threonine in SBM was assumed to be 90, 92, 83, and 89%, respectively (Parsons, 1991). The metabolizable energy contents of corn, SBM, and poultry fat were assumed to be 3,350, 2,440, and 8,800 kcal ME n /kg, respectively (NRC, 1994). This experiment was conducted to evaluate the efficacy of PF in supporting growth of chicks during a finisher period (42 to 63 d of age). Although NRC (1994) recommendations for the finisher period only encompass 42 to 56 d of age, no attempt was made to adjust the NRC (1994) recommendations for lysine, SAA, and threonine levels for the extended finisher period for this experiment. Prior to the experimental period, chicks were fed a common starter diet from 0 to 21 d and a common grower diet from 21 to 42 d; both diets were formulated to meet essential nutrient recommendations (NRC, 1994). Treatment 1 consisted of a single NRC diet fed from 42 to 63 d (Tables 1 and 2). In Treatments 2, 3, and 4, an initial high-nutrient (HN; Tables 1 and 2) diet was formulated to contain predicted lysine, SAA, and threonine requirements for broilers from 41 to 43 d of age (although the experiment did not commence until Day 42). A low-nutrient (LN; Tables 1 and 2) diet was also prepared for Treatments 2, 3, and 4, and was formulated to contain predicted lysine, SAA, and threonine requirements for broilers from 61 to 63 d of age. Nine intermittent diets (43 to 45, 45 to 47, 47 to 49, 49 to 51, 51 to 53, 53 to 55, 55 to 57, 57 to 59, and 59 to 61 d) for Treatments 2, 3, and 4 were prepared by blending the respective HN and LN diets (Tables 1 and 2) in variable quantities, with the percentage of HN decreasing by 10% for each subsequent period. Five pens of 20 male chicks each were weighed and assigned to each of the four experimental feeding regimens. Chicks were weighed at 63 d of age for determination of weight gain, feed intake, and feed efficiency. Feeders were removed from experimental pens 10 h prior to processing on Day 63. After being weighed, five birds per pen were randomly selected for processing at the University of Arkansas processing plant. Part weights were recorded for wings, legs (drum and thigh), breast, abdominal fat, and rack, and parts yields were calculated as a percentage of eviscerated weight. Statistical Analysis The experiment was analyzed as a completely randomized design, and the general linear models (GLM) procedure of SAS software (SAS Institute, 1998) was used to conduct ANOVA on all data. Differences among treatment means were established using the least significant difference multiple comparison procedure (Carmer and Walker, 1985). RESULTS AND DISCUSSION Emmert and Baker (1997) described PF as a nutritional program that, by gradually decreasing dietary amino acid levels in accordance with the bird s requirement, could potentially reduce dietary costs without sacrificing growth performance or carcass yield. Phase-feeding has been explored during the starter, grower, and finisher phases (Warren and Emmert, 2000; Pope and Emmert,

PHASE-FEEDING 469 TABLE 3. Growth performance of broilers fed NRC-, PF 1 -, PF10 2 - or PF15 3 -based diets from 42 to 63 d of age 4 Treatment Parameter NRC 5 PF PF10 PF15 SEM Weight gain, g 1,721 1,652 1,737 1,770 96 Feed intake, g 4,190 4,273 4,459 4,149 167 Gain:feed, g:kg 409 386 389 428 14 Crude protein intake, g 680 681 659 653 26.1 Digestible lysine intake, g 35.6 a 30.9 b 29.1 b 29.6 b 1.23 Digestible SAA 6 intake, g 25.1 a 23.6 ab 22.1 b 21.8 b 0.91 Digestible threonine intake, g 28.5 a 22.4 b 21.0 b 21.4 b 0.94 a,b Means within a row lacking a common superscript differ (P < 0.05). 1 Levels of amino acids in phase-feeding (PF) diets were predicted by linear regression equations (Table 2). 2 Levels of amino acids in adjusted phase-feeding (PF10) diets were predicted by linear regression equations and lowered by 10% (Table 2). 3 Levels of amino acids in adjusted phase-feeding (PF15) diets were predicted by linear regression equations that were modified to contain 15% more negative slope after the initial day of the experiment (Table 2). 4 Values are means of five pens of 20 male chicks fed the experimental diets from 42 to 63 d posthatching, average initial weight was 2.10 kg. 5 NRC diets contained lysine, sulfur amino acid, and threonine levels recommended by NRC (1994). 6 Sulfur amino acids. 2001; Pope and Emmert, unpublished data) using a system in which dietary lysine, SAA, and threonine levels were adjusted weekly. During each period of production, PF-fed birds have exhibited equivalent growth performance and carcass yield to broilers fed diets based on NRC (1994) requirements, and PF has resulted in substantial savings ($ per kg gain or kg breast meat). Under commercial conditions it is unrealistic to expect a different diet to be provided each week to a flock of broilers, due to increased feed transportation costs and increased bin capacity required at the feed mill. The feasibility of PF would be enhanced by blending complete high-nutrient and low-nutrient diets in variable quantities, reducing the number of diets for which a feed mill would be responsible. This approach is similar to systems in use in various parts of the world, in which a highnutrient diet is diluted with an intact, low-protein ingredient such as wheat. However, the dilution approach changes the dietary amino acid profile disproportionately to the broiler requirements, thus to achieve maximum growth performance the high-nutrient diet must be overfortified to prevent deficiencies as the low-nutrient ingredient increases as a percentage of the diet. In contrast, our approach allows target amino acid requirements to be met and eliminates the need for over-fortification. Our goal was to test a PF approach in which two complete diets were blended in variable quantities such that the target amino acid levels were achieved. Phase-feeding diets were switched every other day to mimic an on-farm blending system. No differences (P > 0.05) in weight gain, feed intake, or feed efficiency were observed among birds fed the NRC, PF, PF10, or PF15 diets (Table 3). Our experimental design was similar to that of Pope and Emmert (2001), who fed NRC, PF, PF10, and PF15 diets to broilers from 43 to 64 or 71 d of age, but in the previous study diets were switched weekly instead of every other day. In both experiments birds fed PF, PF10, and PF15 diets displayed equivalent weight gain and feed intake to birds fed an NRC-based diet, but Pope and Emmert (2001) noted a decreased feed efficiency of birds fed PF10 and PF15 diets. Carcass yield (as a % of live weight) was reduced (P < 0.05) by PF10 relative to PF and PF15 diets, but none of the PF regimens reduced (P > 0.05) breast, wing, or leg yield relative to Treatment 1 (Table 4). Abdominal fat percentage was not affected (P > 0.05) by dietary treatment. Pope and Emmert (2001) also observed no differences in breast, wing, or leg yield among birds fed PF, PF10, PF15 or NRC diets from 43 to 64 or 71 d, but they did observe an increase in abdominal fat percentage when PF10 and PF15 diets were fed. The observation that broilers fed PF10 and PF15 regimens maintained a rate of growth comparable to that of birds fed an unmodified PF or NRC regimen (Table 3) suggests that lysine, SAA, and threonine requirements on which the linear regression equations are based are overestimated. However, it is possible that the method used by Emmert and Baker (1997) to derive the equations led to overestimated predicted requirements. In deriving the equations, amino acid requirements for the starter (0 to 21 d), grower (21 to 42 d), and finisher (42 to 56 d) periods were assumed to represent the average requirement for the respective period, so midpoints (10.5, 31.5, and 49 d for the starter, grower, and finisher periods, respectively) for each period were used on the x axis. Had the initial day of each period been used instead, y- intercepts would have been reduced and predicted requirements derived from the equations would be lower and more similar to the PF10 requirements. The ability of PF to support growth performance and carcass yield indicates that our PF diets, with progressively lower CP levels (Table 2), contained enough nitrogen to support dispensable amino acid synthesis. Although digestible lysine and threonine intake were reduced (P < 0.05) by each of the PF regimens, CP intake did not differ (P > 0.05) among treatments (Table 3). In

470 POPE ET AL. TABLE 4. Carcass and economics of broilers fed NRC-, PF 1 -, PF10 2 - or PF15 3 -based diets from 42 to 63 d of age 4 Treatment Parameter NRC 5 PF PF10 PF15 SEM Eviscerated carcass, % of live weight 6 69.0 bc 70.1 a 68.9 c 70.0 ab 0.32 Breast, % of chilled carcass 24.2 24.8 24.0 24.4 0.29 Wing, % of chilled carcass 11.9 ab 12.0 a 11.9 ab 11.6 b 0.13 Leg, % of chilled carcass 33.8 ab 34.4 a 33.6 ab 33.4 b 0.33 Abdominal fat, % of eviscerated carcass 2.6 2.5 2.8 2.5 0.18 Value of feed consumed, $ 7 0.478 0.478 0.487 0.463 0.019 $ per kg weight gain 8 0.281 ab 0.292 a 0.281 ab 0.262 b 0.010 $ per kg breast meat 9 0.760 0.711 0.737 0.691 0.029 a-c Means within a row lacking a common superscript differ (P < 0.05). 1 Levels of amino acids in phase-feeding (PF) diets were predicted by linear regression equations (Table 2). 2 Levels of amino acids in adjusted phase-feeding (PF10) diets were predicted by linear regression equations and lowered by 10% (Table 2). 3 Levels of amino acids in adjusted phase-feeding (PF15) diets were predicted by linear regression equations that were modified to contain 15% more negative slope after the initial day of the experiment (Table 2). 4 Values are means of five pens of 20 male chicks fed the experimental diets from 42 to 63 d posthatching, average initial weight was 2.10 kg; five birds from five replicate pens per treatment were processed on Day 63. 5 NRC diets contained lysine, sulfur amino acid, and threonine levels recommended by NRC (1994). 6 Live weights for birds fed NRC, PF, PF10, and PF15 diets were 3.77, 3.89, 4.00, and 3.94 kg, respectively. 7 Feed cost per bird was determined by multiplying the amount of feed consumed by the dietary cost, which was calculated based on values of $0.1036/kg for corn, $0.1631/kg for soybean meal, $1.1023/kg for L-lysine HCl, $2.4251/kg for DL-methionine, and $3.5053/kg for L-threonine. 8 Calculated by dividing the feed cost per bird by the weight gain per bird. 9 Calculated by dividing the feed cost per bird by the amount of breast yield per carcass, as determined by multiplying the breast yield by the average carcass weight. contrast, Pope and Emmert (2001) observed that PF, PF10, and PF15 reduced the CP intake (relative to an NRCbased diet) of broilers fed from 43 to 64 or 71 d of age. Previously observed reductions in CP intake of birds fed PF regimens during the grower and finisher phases (Pope and Emmert, 2001; Pope and Emmert, unpublished data) led to the presumption that dietary nitrogen excretion may be decreased by PF, a theory that has been supported by research with swine (Boisen et al., 1991). Further research is needed to assess the impact of PF regimens on nitrogen excretion in broilers. Using a PF regimen in which diets were switched weekly, Warren and Emmert (2000) and Pope and Emmert (2001) concluded that substantial economic savings may result from the use of PF during the finisher period. In the current experiment, no differences (P > 0.05) in $ per kg weight gain or breast yield were observed between birds fed any of the PF regimens and the NRC diet (Table 4), although numerical reductions in $ per kg breast yield resulted from each of the PF regimens. In calculating $ per kg weight gain or kg breast yield, the same ingredient prices have been used in this (Table 4) and previous experiments (Warren and Emmert, 2000; Pope and Emmert, 2001). Ingredient prices change daily, but it would be very difficult to reflect these changes when reporting results of PF research. Consistency in the use of ingredient prices have allowed us to compare the results of different trials. In examining the dietary changes that occur with the formulation of PF regimens (Tables 1 and 2) it appears that the economic advantages of PF should increase as the prices of crystalline amino acids (lysine, methionine, threonine) increase. Similarly, changes in the ratio of the price of corn to SBM would affect the economic impact of PF, with PF regimens being favored by decreases in this ratio. The finisher period in this experiment was extended beyond 8 wk of age, potentially increasing the advantage of PF regimens. Pope and Emmert (2001) observed greater economic advantages with PF when broilers were fed until 71 d, compared to 64 d. The use of extended finisher periods is justified because it is not uncommon for companies to raise broilers (in some cases sexes separately) beyond 8 wk of age to increase breast meat production for further processing. Phase-feeding may be particularly appropriate for integrators using the aforementioned approach because in this and other experiments (Pope and Emmert, 2001; Pope and Emmert, unpublished data) the greatest economic advantages associated with PF regimens have been measured in $ per kg breast meat. Hurwitz et al. (1978) also used linear regression equations to predict requirements for broiler chickens (White Rock cockerels). Compared to birds fed NRC requirements (from the 1971 edition), birds fed lysine, SAA, or arginine requirements predicted by linear regression equations had equivalent weight gain, increased feed intake and abdominal fat, and reduced feed efficiency. Although a logical model, their equations were complicated, being based on maintenance, weight gain, body weight, proportion of feather protein, and amino acid composition of feathers and carcass. Moreover, assumptions were made with regard to the percentage contribution of feathers to the entire carcass, the amino acid composition of feathers and carcass, and the dietary protein level required for maintenance. This model would be difficult

PHASE-FEEDING 471 to use in a commercial setting, and the assumptions of Hurwitz et al. (1978) would have to be reexamined if they were to be applied to modern broilers raised under current commercial conditions. Our results are encouraging, in that we found no reduction in growth performance or carcass yield, even when diets were switched every other day and lysine, SAA, and threonine levels were lowered more aggressively in the PF10 and PF15 treatments. It also appears that the cost of production ($ per kg breast meat) could be reduced by PF regimens fed during the finisher period. Previous trials using diets switched weekly have outlined advantages of PF during the starter and grower periods (Warren and Emmert, 2000; Pope and Emmert, 2001), but further research is needed to determined if a PF approach in which diets are switched every other day if efficacious during these periods. In addition, further studies should be conducted to assess the interaction of PF with factors such as dietary energy and bird density, and to evaluate the impact of PF on nitrogen excretion. REFERENCES Baker, D. H. 1997. Ideal amino acid profiles for swine and poultry and their application in feed formulation. Pages 1 24 in: Biokyowa Technical Review. No. 9. Biokyowa Press, St. Louis, MO. Baker, D. H., and Y. Han. 1994. Ideal amino acid profile for chicks during the first three weeks posthatching. Poult. Sci. 73:1441 1447. Boisen, S., J. A. Fernandez, and A. Madsen. 1991. Studies on ideal protein requirement of pigs from 20 to 95 kg live weight. Page 299 in: 6th International Symposium on Protein Metabolism and Nutrition, Herning, Denmark. Carmer, S. G., and W. M. Walker. 1985. Pairwise multiple comparisons of treatment means in agronomic research. J. Agron. Educ. 14:19 26. Emmert, J. L., and D. H. Baker. 1997. Use of the ideal protein concept for precision formulation of amino acid levels in broiler diets. J. Appl. Poult. Res. 6:462 470. Emmert, J. L., M. W. Douglas, S. D. Boling, C. M Parsons, and D. H. Baker. 1999. Bioavailability of lysine from a liquid lysine source in chicks. Poult. Sci. 78:383 386. Hurwitz, S., D. Sklan, and I. Bartov. 1978. New formal approaches to the determination of energy and amino acid requirements of chicks. Poult. Sci. 57:197 205. National Research Council. 1994. Nutrient Requirements of Poultry. 9th rev. ed. National Academy Press, Washington, DC. Parsons, C.M. 1991. Amino acid digestibilities for poultry: Feedstuff evaluation and requirements. Pages 1 15 in: Biokyowa Technical Review No. 1. Biokyowa Press, St. Louis, MO. Pope, T., and J. L. Emmert. 2001. Phase-feeding supports maximum growth performance of broiler chicks from forty-three to seventy-one days of age. Poult. Sci. 80:345 352. SAS Institute. 1998. SAS User s Guide: Statistics. Version 7.0 Edition. SAS Institute Inc., Cary, NC. Warren, W. A., and J. L. Emmert. 2000. Efficacy of phase-feeding in supporting growth performance of broiler chicks during the starter and finisher phases. Poult. Sci. 79:764 770.