ENVIRONMENT AND HEALTH. Meal Feeding Is More Effective than Early Feed Restriction at Reducing the Prevalence of Leg Weakness in Broiler Chickens 1

Transcription

1 ENVIRONMENT AND HEALTH Meal Feeding Is More Effective than Early Feed Restriction at Reducing the Prevalence of Leg Weakness in Broiler Chickens 1 G. SU,* P. SØRENSEN,*,2 and S. C. KESTIN *Department of Animal Breeding and Genetics, Danish Institute of Agricultural Science, Foulum, DK-8830 Tjele, Denmark, and Department of Clinical Veterinary Science, University of Bristol, Langford, Bristol BS40 5DU, United Kingdom ABSTRACT Two trials were conducted to investigate whether manipulation of feeding pattern or early feed intake affected the prevalence of leg weakness in broiler chickens. In Trial 1, the birds were offered two, three, or four meals per day or consumed feed ad libitum. In Trial 2, a multifactorial design was used with age at start, duration of restriction, and severity of restriction as factors. The start of restrictions were at 5, 7, or 9 d, duration of restriction was 5 or 7 d, and feed was restricted to achieve 25, 50, and 75% of predicted growth during the restriction period. Ad libitum birds served as controls. Leg weakness was assessed by gait scoring (GS) and tibial dyschondroplasia (TD) by radiography. Foot burn, hock burn, angulation of the hock joint, feed consumption, and body weight gain were also assessed. The response of the birds to meal feeding was clear. Fewer meals per day was associated with less TD, less hock burn, better walking ability, lower body weight, and better feed conversion. The response of the birds to feed restriction was also clear. Earlier restriction, longer duration, and more severe level of restriction were all associated with lower prevalence of TD, better walking ability, lower body weight, and better relative growth rates and feed efficiency. However, adjusting the observations for differences in body weight removed many of the significant differences; only birds that started feed restriction earlier had less TD. From these trials, it was concluded that meal feeding can beneficially affect the prevalence of leg weakness, and that the major part of this effect is independent of changes in body weight. It was also concluded that early feed restriction reduced many aspects of leg weakness, but that these effects were mainly a result of reduced body weight. Meal feeding and early feed restriction improved feed efficiency. (Key words: broiler, leg weakness, tibial dyschondroplasia, feed restriction, meal feeding) 1999 Poultry Science 78: INTRODUCTION Leg weakness is a serious problem in fast-growing broilers, giving rise to economic losses as well as reduced animal well being. In Sørensen et al. (1999) we discussed the need for control strategies based on husbandry manipulations because reduction of the causes of leg weakness by genetic means was likely to be a long-term strategy. We also presented results from a series of trials in which photoperiod was manipulated. In these trials, although some indices of leg weakness were reduced by the application of moderate photoperiods, body weight was also affected. When the data were adjusted for differences in body weight, no significant Received for publication July 13, Accepted for publication February 2, This program was supported partly by the Danish Ministry of Agriculture, The Danish Poultry Council and the U.K. Ministry of Agriculture Fisheries and Food. 2To whom correspondence should be addressed: Poul.Sorensen@ agrsci.dk. reductions of functional indices of leg weakness were found. If husbandry control of leg weakness is ever to be acceptable, strategies that do not affect the economics of production are desirable. The improvements in feed efficiency and carcass composition resulting from early feed restriction have been reported by several authors (Plavnik and Hurwitz, 1985; Jones and Farrell, 1992a,b; Roth et al., 1993; Cristofori et al., 1997), but the effect on leg weakness is less well documented. Some skeletal disorders can be reduced by feed restriction (Duff and Thorp, 1985; Classen and Riddell, 1990; Robinson et al., 1992) and a large decrease in growth rate can virtually abolish skeletal disease (Riddell, 1983). However, uncertainty exists as to whether beneficial reductions in skeletal disorders affect overall walking ability and whether these are linked to changes in body weight or are independent. It is also unclear what age, level, and Abbreviation Key: FCR = feed conversion ratio; GS = gait score; GSA = gait score excluding score 4 and 5; TD = tibial dyschondroplasia. 949

2 950 Trial TABLE 1. Summary of experimental treatment details Replicate Birds Bird per per per square treatment replicate meter Meal feeding Early feed restriction Lixiscope, Lixi Inc., Downers Grove, IL Feeder space (cm/bird) duration of restriction is most effective in reducing leg weakness. Meal feeding of broilers is also believed to reduce leg weakness and improve feed efficiency and is currently practiced by some parts of the industry in the U.K. (M. Potter, RSPCA, Horesham, West Sussex RH1 2HG, U.K., personal communication). However, the effectiveness of this strategy has not been evaluated. This paper reports and discusses the results of two trials. In one trial, the effect of feeding birds two, three, and four meals per day for the entire production cycle was compared with those that consumed feed ad libitum. A second trial was undertaken to determine the effect of different aspects of early feed restriction. Age at start and the duration and severity of restriction were all tested in a multifactorial design. The effect of the treatments on leg weakness was assessed using the same traits as before (Sørensen et al., 1999). MATERIALS AND METHODS Experimental Methods In both trials, Ross 208 chicks were obtained from a commercial hatchery sexed and tagged at 1-d-old, and 30 assigned at random to each of 96 treatment pens, but with each pen balanced for sex. See Table 1 for summary of individual trial details. Bird husbandry was as described in Sørensen et al. (1999) unless otherwise stated. In both trials, a photoperiod of 21 h was provided with a light intensity of approximately 15 lx at bird height with less than 0.3 lx during the dark period. Thirty minutes dusk (approximately 4 lx) was provided towards the end of the photoperiod. Birds on both trials were fed broiler starter ration from 1 to 14 d of age (ME = 3,105 kcal/kg, 21.8% protein) and broiler grower ration (ME = 3,150 kcal/kg, 20.2% protein) thereafter. The starter diet contained a coccidiostat. During both trials, records of feed consumed by each pen were kept and birds were weighed on a pen basis on the age of 21 d and 30 d so that feed conversion ratios (FCR) and relative growth rates could be calculated. In both trials, water was available for ad libitum intake from nipple drinkers. SU ET AL. Assessment of Traits In both studies, the main indices of leg weakness adopted, gait score (GS) and tibial dyschondroplasia (TD), were assessed in all birds according to the methods of Kestin et al. (1992) and Ducro and Sørensen (1992). Birds were gait scored individually by an experienced assessor who assessed the walking ability of the birds when moving spontaneously in the rearing environment, assigning the bird a score from 0 for a perfectly normal bird to 5 for a bird that could not walk at all. To estimate the prevalence of TD, the left leg of each bird was assessed using an x-ray fluorescence device3 as described by Bartels et al. (1989). The size of the cartilage plug present was scored on a scale from 0 to 3, in which 0 was given for no occurrence of TD and 3 for cases in which the cartilage almost filled the proximal head of the tibio-tarsus (Edwards and Veltmann, 1983). At gait scoring and TD assessment, birds were individually weighed. Postmortem, carcasses were assessed for prevalence of foot pad burn, hock burn, and valgus/varus angulation. The foot pads and hocks of each bird were evaluated as the carcasses passed on the evisceration line and a combined score for both legs given from 0 to 3, with 0 for no sign of damage and 3 for extended burn and inflammation. The valgus/varus angulation was subjectively evaluated and scored at the same time on a scale of 0 to 3 with 0 for no angulation of the hock (less than 5 degrees) and 3 for severe angulation (greater than 40 degrees). The assessment of TD was performed at 28 d of age when the birds were housed in the trial pens. However, because the pens were too small to allow adequate observation of walking ability, 24 h before assessment of walking ability at 35 d of age, all pen divisions were removed, allowing birds from all treatments to mix. Birds were therefore assessed at random and assessors were unaware of treatment. At GS and TD assessment, the birds were individually weighed. In total, 2,686 birds were assessed in Trial 1 and 2,864 birds in Trial 2. Trial 1. Meal Feeding See Figure 1 for details of meal feeding regimens used. Each replicate pen contained a hopper feeder, which was normally kept covered to deprive the birds of access to feed. However, at the preset times each day (see Figure 1), the covers were removed to allow birds to access feed, and replaced when the designated time had elapsed. During meal time, feed was available for ad libitum consumption. The total time the feeders were exposed was the same in all meal feeding treatments (240 min/d) and was calculated to be approximately 50% more than the average time birds spend feeding each day at the end of the production cycle (Weeks and Davies, 1996). Sufficient space for all birds to feed simultaneously was provided. Meal feeding commenced when the birds were 5 d old and continued until 1 d before GS at 35 d of age. Behavioral observations were

3 MEAL FEEDING, FEED RESTRICTION, AND LEG WEAKNESS 951 the predicted growth of ad libitum birds during the restriction period. To achieve this, the expected daily growth and feed intake during the restriction period was calculated from a standard table recording these attributes for Ross 208 birds generated from commercial practice. Maintenance requirements were calculated using the formula: feed (grams per day) = 42 BW0.75. To obtain the feed allowances for the different levels of restriction, the difference between the calculated maintenance requirements and the predicted ad libitum feed intake was multiplied by the appropriate factor. Feed allowances were provided to the birds twice per day in feed hoppers. The hoppers provided sufficient space for all birds to feed simultaneously. After the restriction program ended, all birds were offered ad libitum access to feed. Statistical Analysis FIGURE 1. Diagram of meal feeding regimens and relationship to photoperiod. made on a representative sample of pens from each treatment at preset times each day and will be reported elsewhere. Trial 2. Feed Restriction In Trial 2, birds that consumed feed ad libitum were compared with birds fed on various restriction programs in a multifactorial design. The age at start of restriction was either 5, 7, or 9 d, with the feed restriction lasting 5 or 7 d except that the combination of latest start day and longest duration was omitted. Three levels of restriction were evaluated, with birds being fed predetermined amounts of feed calculated to achieve 25, 50, and 75% of For both trials, the data of the five traits to describe leg disorders were analyzed by an ANOVA model with a factorial arrangement of treatment and sex with pen nested within treatment. For Trial 1, the treatment was meal treatment, for Trial 2, the treatment was one of start age, restriction level, or duration, respectively. As a second model for statistical analysis, body weight at the time of measurement was used as covariable for TD and GS. In order to examine the effect of body weight on TD and gait score in detail, a regression analysis was conducted using a simple regression model. Because previous studies have indicated that most of the birds with a GS of 4 or 5 are atypical and have infections in their joints or bones (Sørensen et al., 1999), in most analyses, GS was also analyzed with these birds excluded from the data set, and is presented in the tables as GS adjusted (GSA). The ANOVA and the regression analyses as well as the least squares means were calculated by using the General Linear Models procedure of SAS (SAS Institute, 1994). TABLE 2. Least squares means for leg weakness traits as an effect of four meal feeding treatments, Trial 1 Model 1 Traits 2 Two meals Three meals Four meals Ad libitum 1 TD b b b a GS c c b a GSA c c b a Foot a a a a Hock c bc b a Angul ab a ab b 2 TD b b b a GS c bc b a GSA b b b a Foot a ab ab b Hock b ab ab a Angul ab a ab b a cestimates in a row with no common superscript differ significantly (P < 0.05). 1Model 1 excluding and Model 2 including body weight as covariable in the analysis of leg weakness traits. 2Abbreviations for traits: TD = tibial dyschondroplasia; GS = gait score; GSA = gait score excluding score 4 and 5; Foot = foot pad burn; Hock = hock burn; Angul = valgus/varus angulation.

4 952 SU ET AL. Traits 1 TABLE 3. Least squares means for growth traits and feed conversion ratio as an effect of four meal feeding treatments, Trial 1 28-d BW 1,198 d 1,220 c 1,260 b 1,297 a 35-d BW 1,592 d 1,633 c 1,677 b 1,701 a Gain 21 to 30 d 563 c 582 b 592 b 623 a RGR 21 to 30 d a a a a FCR 0 to 21 d c bc b a 0 to 30 d c bc b a a destimates in each row with no common superscript differ significantly (P < 0.05). 1RGR = relative growth rate; FCR = food conversion ratio. Meal Feeding Two meals Three meals RESULTS Four meals Ad libitum The least squares means for the leg weakness traits measured in Trial 1 are shown in Table 2. The responses of the birds to the four levels of meal feeding were clear. Fewer meals per day were associated with better walking ability and less hock burn. Birds that consumed feed ad libitum had more TD than those that consumed meals. Footpad burns were unaffected by meal feeding, and the effects on valgus/varus angulation of the hock inconsistent. Adjusting the observations to remove birds with a GS of 4 and 5 gave the same pattern. The proportion of birds assigned a GS of 4 and 5 did not exceed 1.3% in any treatment. Adjusting the observations for differences in body weight did not substantially alter the findings (Table 2). Fewer meals per day were still associated with better walking ability and less hock burn. Similarly, birds that consumed feed ad libitum still had more TD than birds that ate meals. Footpad burns were affected by meal feeding, and the inconsistent effects on angulation of the hock remained. Adjusting the observations to remove birds with a GS of 4 and 5 still did not alter the findings. The body weight at 28 and 35 d of age, the relative growth of the birds from 21 to 30 d of age, and the FCR from 1 to 30 d of age are shown in Table 3. Fewer meals per day were associated with significantly lower body weight at all the ages. No differences in relative growth of the birds between treatments were found. However, birds fed fewer meals had better FCR than birds fed more meals, or those that consumed feed ad libitum. Early Feed Restriction The least squares means for the leg weakness traits measured in Trial 2 are shown in Table 4. The age at start, the duration, and the level of restriction all had clear effects. Earlier restriction, longer durations of restriction, and more severe levels of restriction were all associated with a lower prevalence of TD and better walking ability, both overall and when birds with a GS of 4 and 5 were removed. The proportion of birds assigned a GS of 4 and 5 did not exceed 1%. Adjusting the observations for differences in body weight removed many of the significant differences. The age at start of restriction was still found to affect the prevalence of TD, with birds that started feed restriction earlier having less TD. Similarly, the age at which restriction started significantly affected walking ability, with birds starting restriction later having better walking ability. However, the effects of duration and the level of restriction became not significant. The least squares means for the body weight, relative growth rate, and FCR are shown in Table 5. The age at start, the duration, and the level of restriction all had consistent effects on BW. Earlier restriction, longer durations of restriction and more severe levels of restriction were all associated with lower body weight at all measurement ages. The growth and FCR of the birds was analyzed in two phases, from 0 to 21 d of age, the period when restriction was carried out and from 21 up to 34 d, the age when birds were mixed prior to GS. The gain in body weight between 21 d, after restriction had ended, and 34 d was the same in almost all treatments. Severity of restriction did affect the growth of the birds in this period, but the effects were inconsistent. Because the body weight of the birds at 21 d was different (see Table 5), significant differences in relative growth rate were found. Earlier restriction, longer durations of restriction, and more TABLE 4. Least squares means for leg weakness traits as an effect of early feed restriction, Trial 2 Starting age Expected growth Duration Model 1 Trait 2 5 d 7 d 9 d Control 25% 50% 75% Control 7 d 5 d Control 1 TD b b a a c bc b a b b a GS b b b a b b a a c b a GSA b b b a b b a a c b a 2 TD b b a ab a a a a a a a GS a ab b a a a a a a a a GSA a ab b ab a a a a a a a a cestimates in each row within measurement with no common superscript differ significantly (P < 0.05). 1Model 1 excluding and Model 2 including body weight as covariable. 2TD = tibial dyschondroplasia; GS = gait score; GSA = gait score excluding Score 4 and 5.

5 MEAL FEEDING, FEED RESTRICTION, AND LEG WEAKNESS 953 TABLE 5. Least squares means for growth traits and feed conversion ratio as an effect of early feed restriction, Trial 2 Start age Expected growth Duration Trait 1 5 d 7 d 9 d Control 25% 50% 75% Control 7 d 5 d Control BW 21-d 688 c 714 b 729 b 818 a 664 d 703 c 753 b 818 a 678 c 726 b 818 a 28-d 1,172 d 1,198 c 1,216 b 1,292 a 1,139 d 1,188 c 1,247 b 1,292 a 1,157 c 1,214 b 1,292 a 35-d 1,560 c 1,581 b 1,596 b 1,672 a 1,523 d 1,572 c 1,631 b 1,672 a 1,547 c 1,594 b 1,672 a 42-d 2,108 b 2,129 b 2,126 b 2,248 a 2,051 c 2,135 b 2,176 b 2,248 a 2,098 b 2,132 b 2,248 a Gain 21 to 34 d 841 a 851 a 843 a 858 a 834 b 858 a 846 ab 858 a 849 a 844 a 858 a RGR 21 to 34 d a ab b c a b c d a b c FCR 0 to 21 d c c b a bc c b a c b a 0 to 34 d b b ab a bc c ab a b a a a destimates in each row within measurement with no common superscript differ significantly (P < 0.05). 1RGR = relative growth rate; FCR = feed conversion rate. severe levels of restriction were all associated with higher relative growth rates during the 21 to 34 d period. These differences were reflected in changes in FCR, with earlier restriction, longer duration of restriction, and more severe levels of restriction all associated with better FCR. The important correlation coefficients for the main traits measured in Trials 1 and 2 are shown in Table 6. The strongest correlation was found between GS and body weight. This correlation was significant, with the coefficient increasing if the data set was adjusted to remove the birds with GS of 4 and 5. Tibial dyschondroplasia was also positively correlated with body weight, but this was not as strong as for GS. In Trial 1, GS was also positively correlated with hock burn and TD. These findings were similar to those reported in Sørensen et al. (1999). The regressions of leg weakness traits on body weight were consistent among the groups of each treatment with very few exceptions. Across treatment in each trial, the linear regression coefficients for the main variables measured are shown in Table 7. The regression for TD on body weight was weak but significant. The regression for GS on body weight was much stronger, and more so if the data were adjusted to remove the birds with GS of 4 and 5. DISCUSSION The meal feeding trial produced some interesting results. Despite having access to feed for a similar number of hours each day, birds fed fewer meals each day had lower body weight than those fed more meals or that consumed feed ad libitum. Birds fed fewer meals also had better FCR than birds fed more meals, or fed ad libitum. Clearly, the birds were unable to fully adapt their feed intake to the feeding regimens applied. Whether this effect could have been abolished and whether the improvements in FCR would be reduced if more time had been allowed at each meal time are not known and require further study. Birds fed fewer meals per day had less TD, less hock burn, and better walking ability. These findings would be expected, bearing in mind the positive correlation between TD and GS and body weight. Based on previous studies (Sørensen et al., 1999), it might also be expected that the improvements would be lost when the data were adjusted for body weight. However, even when adjusted for body weight, the findings remained the same. Birds fed fewer meals per day still had less TD and better walking ability. The reason for this effect is unclear. It could reflect the more organized behavioral patterns induced by meal feeding, resulting in more overall activity, or it could reflect changes in the pattern of circulating hormones such as insulin and growth hormone, that may result from the pulsate feed intake similar to those reported by Buyse et al. (1997) for feedrestricted broilers. It is probable that the differences in FCR between the treatments was also reflected in TABLE 6. Simple correlations between some pairs of traits 1 Trial TD:BW28 TD:GS TD:GSA GS:BW35 GSA:BW35 GSA:Foot GSA:Hock ** 0.116** 0.108** 0.444** 0.489** 0.079* 0.275** ** ** 0.513** 1Abbreviations for traits: TD = tibial dyschondroplasia; BW28 = body weight at 28 d; GS = gait score; GSA = gait score excluding Score 4 and 5; Foot = foot pad burn; Hock = hock burn. *P < **P < 0.01.

6 954 SU ET AL. TABLE 7. Linear regressions of some leg disorders trait on body weight Trial Trait 1 Intercept Regression R 2 1 TD 0.264** 0.315** GS 1.321** 1.807** GSA 1.551** 1.932** TD 0.704** 0.914** GS 1.314** 1.718** GSA 1.577** 1.869** TD = tibial dyschondroplasia; GS = gait score; GSA = gait score excluding Score 4 and 5. *P < **P < differences in body composition, with birds fed fewer meals each day being leaner than those fed more meals or those that consumed feed ad libitum as reported by Jones and Farrell (1992b) and Cristofori et al. (1997). The improvements in FCR and leg weakness achieved in this trial were substantial. Taken together, they suggest that meal feeding beneficially alter the metabolism and behavior of the birds. If the improvements in FCR are maintained until slaughter, and there is no reason to believe they would not be, meal feeding may provide a viable method for improving production efficiency and limiting leg weakness. In the early feed restriction trial, TD and GS were reduced by feed restriction. However when observations were adjusted for body weight, many of the significant differences were lost. Only the age at start of restriction affected the prevalence of TD, birds that started feed restriction earlier had less TD. Similarly, only the age at which restriction started significantly affected the walking ability of the birds, but in this case, later restriction was associated with improved walking ability. These data imply that if improvements in walking ability are the objective, the restriction should be applied later in growth rather than earlier. It is interesting that there was no effect of severity of restriction or duration of restriction on walking ability. This result indicates that there is a limited improvement to be achieved by early feed restriction, and that this improvement can be achieved by the least severe and shortest period of restriction applied in this study. The growth and feed efficiency of the birds was affected by early feed restriction, with earlier start date, increasing severity, and longer duration being associated with lower body weights and better relative growth rate and FCR after restriction ended. This result is similar to the reports of Roth et al. (1993). Birds subjected to all restriction treatments (see Table 4), showed compensatory growth. However, contrary to the findings of Jones and Farrell (1992a), none of the birds on any of the restriction regimens adopted were able to fully compensate and achieve similar body weights at slaughter to the ad libitum birds. However, the birds in this study were slaughtered relatively young (42 d). Had they been killed at 49 d and the better relative growth rate had been maintained, full compensation might have been achieved. The restricted birds utilized feed more efficiently during the restriction period and the compensatory growth period, which is comparable to the findings from other studies (Cristofori et al., 1997). Because FCR is very important in the economics of broiler production and improvements in walking ability were achieved by relatively modest restriction programs, which also improved FCR, early feed restriction may provide a viable method for limiting the expression of leg disorders. Further work in this area is merited to optimize the manipulations. Considered together with the results of photoperiod manipulations reported in Sørensen et al. (1998), these trials suggest that TD is most limited by growth manipulations applied relatively early in life. Further, walking ability is most improved by manipulations applied later in life and both are most affected by manipulations such as meal feeding, which continue until assessment. Further studies are required to test these hypotheses. Based on the low overall prevalence of TD in these studies, and its weak correlation with gait score, it can also be tentatively concluded that TD is not one of the main factors determining walking ability in commercial broilers. At present, it is not known what the main factors contributing to poor walking ability are. Further work is required to identify them. In conclusion, these trials indicate that leg weakness can be reduced by meal feeding or early feed restriction. Because these manipulations can also result in improved efficiency, they may be a feasible method for limiting leg weakness. REFERENCES Bartels, J. E., G. R. McDaniel, and F. J. Hoerr, Radiographic diagnosis of tibial dyschondroplasia in broilers: A field selection technique. Avian Dis. 33: Buyse, J., E. Decuypere, and J. D. Veldhuis, Compensatory growth of broiler chickens is associated with an enhanced pulsatile growth hormone (GH) secretion: preferential amplification of GH secretory burst mass. Br. Poult. Sci. 38: Classen, H. L., and C. Riddell, Early growth rate and lighting effects on broiler skeletal disease. Poultry Sci. 69(Suppl. 1):35. (Abstr.) Cristofori C., A. Melluzzi, G. Giordani, and F. Sirri, Early and late quantitative feed restriction of broilers: effects on productive traits and carcase fatness. Arch. Geflügelkd. 61(4): Ducro, B. J., and P. Sørensen, Evaluation of a selection experiment on tibial dyschondroplasia in broiler chickens. Pages in: Proceedings of XIX World s Poultry Congress. Vol. 2. Amsterdam, The Netherlands. Duff, S.R.I., and B. H. Thorp, Patterns of physiological bone torsion in the pelvic appendicular skeletons of domestic fowl. Res. Vet. Sci. 39:307:312. Edwards, H. M., Jr., and J. R. Veltmann, Jr., The role of calcium and phosphorous in the etiology of tibial

7 MEAL FEEDING, FEED RESTRICTION, AND LEG WEAKNESS 955 dyschondroplasia in young chickens. J. Nutr. 113: Jones, G.P.D., and D. J. Farrell, 1992a. Early-life food restriction of broiler chickens: I. Methods of application, amino acid supplementation and the age at which restrictions should commence. Br. Poult. Sci. 33:579:587. Jones, G.P.D., and D. J. Farrell, 1992b. Early-life food restriction of broiler chickens: II. Effects of food restriction on the development of fat tissue. Br. Poult. Sci. 33: Kestin, S. C., T. G. Knowles, A. F. Tinch, and N. G. Gregory, The prevalence of leg weakness in broiler chickens and its relationship with genotype. Vet. Rec. 131: Plavnik, I., and S. Hurwitz, The performance of broiler chicks during and following a severe feed restriction at an early age. Poultry Sci. 64: Riddell, C., Pathology of the skeleton and tendons of broiler chickens reared to roster weights I. Crippled chickens. Avian Dis. 27: Robinson, R. E., H. L. Classen, J. A. Hanson, and D. K. Onderka, Growth performance, feed efficiency and the incidence of skeletal and metabolic disease in full-fed and feed restricted broiler and roaster chickens. J. Appl. Poult. Res. 1: Roth, F. X., M. Kirchgebner, and W. Windisch, Compensatory growth of broilers: Effects of feed restriction, varying in severity, duration and onset. Arch. Geflügelkd. 57(1):1 8. SAS Institute, SAS/STAT User s Guide: Statistical Version SAS Institute Inc., Cary, NC. Sørensen P., G. Su, and S. C. Kestin, The effect of photoperiod/scotoperiod on leg weakness in broiler chickens. Poultry Sci. 78: Weeks, C. A., and H. C. Davies Effect of leg weakness on feeding behavior of broilers. Br. Poult. Sci. 37(Suppl):