Effects of Age and Stocking Density on Leg Weakness in Broiler Chickens 1

Effects of Age and Stocking Density on Leg Weakness in Broiler Chickens 1 P. Sørensen,*,2 G. Su,* and S. C. Kestin *Department of Animal Breeding and Genetics, Danish Institute of Agricultural Sciences, Foulum, DK-8830 Tjele, Denmark and Department of Clinical Veterinary Science, University of Bristol, Langford, Bristol BS40 5DU, United Kingdom ABSTRACT The effects of stocking density (STD) on ability. Further deterioration had occurred by 7 wk of leg weakness in broiler chickens was assessed in two trials. The interaction between age and STD on leg weakness was further evaluated in one trial. In Trial 1, walking ability was assessed at 28, 42, and 49 d of age. Birds were stocked at 833, 625, or 435 cm 2 per bird. In Trial 2, birds were stocked at 625 or 455 cm 2 per bird and assessed for tibial dyschondroplasia (TD) by radiographic examination at 28 d and walking ability at 35 d. Foot pad burn, hock burn, and angulation of the hock joint were also assessed at slaughter on Day 42. Body weight was measured during both trials. At 4 wk of age, leg weakness was a relatively minor problem; few severely lame birds had a gait score (GS) of 4 or 5 at any density. However, 2 wk later, the birds had substantially poorer walking age. At all ages, males exhibited greater leg weakness than did females, and the proportion of severely lame birds increased with age of assessment. The effect of STD was consistent across both trials; higher STD were associated with poorer walking ability and reduced live weights. In Trial 2, higher STD resulted in more foot and hock burns. Females were more sensitive to STD than were males However, there was no effect of STD on the prevalence of TD or angulation of the hock joint. The effect of high STD on walking ability was apparent even at 4 wk of age. Adjusting the observations for differences in BW did not alter the findings. It was concluded that the lower STD substantially reduced the prevalence of leg weakness. (Key words: leg problems, tibial dyschondroplasia, stocking density, age, body weight) 2000 Poultry Science 79:864 870 INTRODUCTION Many studies have examined the effect of bird density on growth and slaughter quality of broiler chickens (Proudfoot et al., 1979; Proudfoot and Hulan, 1985; Grashorn and Kutritz, 1991; Cravener et al., 1992). Summarizing the findings, in comparison with birds housed at lower densities, individual birds that were provided 500 cm 2 of walking area had reduced BW and an increasing frequency of breast blisters and ammonia burns of foot pad and hocks (Proudfoot et al., 1979; Cravener et al., 1992; Thomsen, 1992; Martrenchar et al., 1997), scabby hip syndrome (Proudfoot and Hulan, 1985), and downgrading of the carcass (Grashorn and Kutritz, 1991). It is now recognized that high stocking densities (STD) (500 cm 2 per bird or less) deleteriously affect bird growth and quality. In addition to compromising production parameters, high densities may also deleteriously affect bird welfare. Broilers stocked at commercial densities appear con- Received for publication September 27, 1999. Accepted for publication February 23, 2000. 1 Parts of this program were supported by the Danish Ministry of Agriculture, The Danish Poultry Council, and the United Kingdom Ministry of Agriculture Fisheries and Food. 2 To whom correspondence should be addressed: Poul.Sorensen@ agrsci.dk. strained from moving readily by their companions. Lewis and Hurnik (1990) found that the average distance traveled by individual birds was reduced by 20% when density was increased (1320 to 660 cm 2 per bird), and Blokhuis and van der Haar (1990) found a reduced percentage of chickens performing pecking or scratching and walking behaviors at 500 cm 2 per bird at 7 wk of age compared with birds housed at a lower density. Recently, it has been reported (Reiter and Bessei, 1995) that broilers given more exercise have less leg weakness. In these studies, broilers were exercised on a treadmill, and their walking ability was measured. Birds given more exercise had better walking ability. Thus, it might be expected that a relationship would exist among STD, activity, and leg weakness; birds at lower densities would move around more and thus limit the expression of leg disorders. Su et al. (1999) presented results showing that the prevalence of leg weakness was reduced by meal feeding or early feed restriction. In both cases, reductions in leg weakness were relatively modest and were coupled with reduced growth rate. It was speculated that husbandry manipulations that lasted throughout the life of the birds were more effective at reducing leg weakness than manip- Abbreviation Key: GS = gait score; LS = least squares STD = stocking density; TD = tibial dyschondroplasia. 864

AGE, STOCKING DENSITY, AND LEG WEAKNESS 865 ulations applied for a short time early in life. Housing broilers at lower densities might be a particularly effective and commercially acceptable method for reducing leg weakness because reduced STD would be effective throughout life. The purpose of this paper was to examine the effects of different STD on incidences of leg weakness. In Trial 1, the interaction of age and STD on the development of leg weakness was also investigated. MATERIALS AND METHODS Assessment of Traits The main index of leg weakness in Trial 1 was gait score (GS), and the indices in Trial 2 were GS and tibial dyschondroplasia (TD). Gait score and TD were assessed according to the methods of Kestin et al. (1992) and Ducro and Sørensen (1992), respectively, as outlined in Sørensen et al. (1999) and Kestin et al. (1999). Briefly, birds were scored individually for gait in both trials by the same experienced assessor, who assessed the walking ability of the birds while they were moving spontaneously in the rearing environment. Birds were assigned a score from a 6-point scale, where 0 = a perfectly normal bird to 5 = a bird that could not walk at all. In Trial 2, the left leg of each bird was assessed for the prevalence of TD using an X-ray fluorescence device 3 as described by Bartels et al. (1989). The size of the cartilage plug was scored on a scale from 0 to 3, where 0 = no occurrence of TD to 3 = cartilage almost filled the proximal head of the tibio-tarsus, as illustrated in Kestin et al. (1999). Simultaneously with TD assessment, curvature of the tibia was also subjectively estimated by X-ray according to the method of Kestin et al. (1999) using a scale of 0 to 3, where 0 = no curvature and 3 = severe curvature. At gait scoring and TD assessment, birds were individually weighed. Postmortem, carcasses were assessed for prevalence of foot pad burn, hock burn valgus or varus angulation, and breast blisters according to the method of Sørensen et al. (1999). The foot pads and hocks of each bird were evaluated as the carcasses passed on the evisceration line, and combined scores for both legs were assigned from 0 to 3, where 0 = no sign of damage to 3 = extended burn and inflammation. The valgus or varus angulation was subjectively evaluated and scored at the same time on a scale of 0 to 3, where 0 = no angulation of the hock (less than 5 ) and 3 = severe angulation (>40 ). Bird Husbandry In both trials, Ross 208 chicks were obtained from a commercial hatchery, sexed, and tagged at 1 d old. In Trial 1, a photoperiod of 23 h was provided with a light intensity of approximately 10 lx at bird height with less than 0.3 lx during the scotoperiod. Birds were provided a broiler starter diet (ME = 3,125 kcal/kg, 22% CP) from 3 Lixiscope, Lixi Inc., Downer s Grove, IL 60515. 1 to 14 d of age, a broiler grower diet (ME = 3,150 kcal/ kg, 20% CP) was fed from 15 to 42 d of age, and a broiler finisher diet (ME = 3,125 kcal/kg, 20% CP) was provided thereafter; all for ad libitum consumption. The starter diet contained nicarbazin and the grower monensin as coccidiostats. Water was available for ad libitum consumption from bell drinkers. Feeder and drinker spaces were identical in each pen. Approximately 2 kg/m 2 of wood shavings litter were supplied to each pen at the start of the trial. In Trial 2, bird husbandry was as described by Su et al. (1999) unless otherwise stated. A photoperiod of 21 h was provided with a light intensity of approximately 15 lx at bird height during the photoperiod and less than 0.3 lx during the scotoperiod. Thirty minutes of dusk (approximately 4 lx) was provided toward the end of the photoperiod. Birds were fed a broiler starter diet (ME = 3,100 kcal/kg, 21.8% CP) from 1 to 14 d of age and a broiler grower diet (ME = 3,150 kcal/kg, 20.2% CP) thereafter. Feed and water were available for ad libitum intake, and feeder and nipple drinker spaces were identical in each pen. Two kg/m 2 of chopped straw litter wer supplied to each pen at the start of the trial. During the progress of the trial, the moisture of the litter was estimated each week by taking four samples of litter from four pens per density and determining the DM content according to the standard EEC method (EEC, 1971). Experimental Methods In Trial 1, 120 chickens were allocated at random to each pen but were balanced for sex. Pen area was adjusted by moving one wall to achieve a STD of 833, 625, or 435 cm 2 per bird. In total, 1,032 birds were assessed in Trial 1. In Trial 2, 275 chickens were allocated at random to each pen, of which 150, picked at random, were tagged and sexed at 1doldandsubjected to detailed study. The size of the pen area was adjusted similarly to achieve a STD of 625 or 455 cm 2 per bird. In total, 2,331 birds were assessed in Trial 2. The choice of densities was governed by commercial practice; densities are approximately 625 cm 2 per bird in the UK and 400 cm 2 per bird in Denmark. Summary details of the conditions used in these trials are shown in Table 1. Statistical Analysis In Trial 1, a linear model comprising a factorial arrangement of density and sex with pen nested within density was used to analyze the effects of STD at each age. A model containing a factorial arrangement of density, age, and sex with pen nested within the combination of density and age was used to analyze the effects of STD and age simultaneously. The model used in Trial 2 comprised a factorial arrangement of density, litter type, and sex with pen nested within the combination of density and litter type. In the second set of models, BW at the time of measurement was introduced as a covariable to account for BW effects on TD and walking ability. Because

866 SØRENSEN ET AL. TABLE 1. Summary of experimental treatment details Replicate per Birds per Birds assessed Feeder space Density Trial treatment replicate per replicate (cm per bird) (cm 2 per bird) 1 3 120 120 1.6 833, 625, or 435 2 8 275 150 1.4 625 or 455 the data were unbalanced over the various factors included in the models, the comparisons of age and STD effects were based on least squares (LS) means, which accounted for the influence of other factors such as sex, litter type, etc. To examine in detail the effect of BW on TD and walking ability and the effect of age on walking ability, a regression analysis was conducted using a simple regression model in which either BW or the logarithm of age was taken as the independent variable. The AN- OVA and the regression analyses as well as the LS means were calculated by using the GLM procedure of SAS (SAS Institute, 1994). Previous studies have indicated that most of the birds with a GS of 4 or 5 are atypical of the bulk of lame birds (Sørensen et al., 1999) and may have infections in their joints or bones (Kestin et al., 1994). Therefore, in most analyses, GS data analyses were also performed with birds with a score of 4 or 5 excluded from the data set. These data are presented in the tables as GS excluding scores 4 and 5 (GSA). Trial 1 RESULTS The LS means of the traits measured at the three STD at each of the three ages are shown in Table 2. Further, Table 3 presents the LS means of these traits for the three STD pooled over the three ages and for the three ages pooled over the three STD. The proportion of birds scored with a 4 or 5 are shown in Table 4. It should be noted that those with a score of 4 or 5 increased (P < 0.05) with age (Table 4). At 4 wk of age, leg weakness was relatively minor; GS were low, and very few severely lame birds had GS of 4 or 5. However, 2 wk later, the birds had substantially higher GS, having deteriorated at the rate of 0.45 GS units per week. Further deterioration had occurred 1 wk later; the birds had deteriorated by 0.2 GS units. Stocking density affected walking ability at all ages measured. At 4 wk of age, birds at 435 cm 2 per bird had higher scores than birds at 833 cm 2 per bird. At 6 wk of age, birds at 435 cm 2 per bird had higher scores than birds at 625 cm 2 per bird, which, in turn, were greater than those of birds at 833 cm 2 per bird. At 7 wk of age, birds at 435 and 625 cm 2 per bird had higher GS than did birds at 833 cm 2 per bird, and the proportion of birds given a score of 4 or 5 was higher for birds in pens at 435 cm 2 per bird and 625 cm 2 per bird than those for birds in pens at 833 cm 2 per bird. Body weight was also decreased by high STD at 7 wk of age; birds at the highest density were lighter than birds at the lower two densities. When the GS observations were adjusted for BW, the findings remained the same. When pooled over the three ages, walking ability at 435 cm 2 per bird was poorer (P < 0.05) than that at 625 cm 2 per bird, and, in turn, the latter was poorer (P < 0.05) than that at 833 cm 2 per bird. Based on the LS means for GS estimated from Model 2 at 4 wk of age, the differences in walking ability between birds in pens at 833 cm 2 per bird and 625 cm 2 per bird and between birds in pens at 625 cm 2 per bird and 435 cm 2 per bird were 0.046 and 0.128 GS units, respectively. At 6 wk of age, the differences were 0.181 and 0.312 GS units, respectively, and, at 7 wk of age, the differences were 0.282 and 0.113 GS units, respectively. Figure 1 shows the curves of averaged GS related to age for males and females in each of the three densities. The relationship between walking ability and age was nonlinear but fitted the function of GS = a + b log age. Gait score differences caused by STD were larger for fe- TABLE 2. Least squares means for gait score (GS) and BW at 28, 42, and 49 d of age for stocking densities of 833, 625, and 435 cm 2 per bird in Trial 1 Age Model 1 1 Model 2 (d) Trait 2 833 cm 2 per bird 625 cm 2 per bird 435 cm 2 per bird 833 cm 2 per bird 625 cm 2 per bird 435 cm 2 per bird 28 GS 0.854 b 0.891 ab 0.999 a 0.838 b 0.884 b 1.012 a GSA 0.854 b 0.872 ab 0.987 a 0.839 b 0.884 ab 0.998 a 42 GS 1.617 c 1.795 b 2.098 a 1.601 c 1.782 b 2.094 a GSA 1.563 c 1.724 b 2.030 a 1.530 c 1.725 b 2.066 a 49 GS 1.814 b 2.103 a 2.212 a 1.815 b 2.097 a 2.210 a GSA 1.730 b 1.904 a 2.027 a 1.702 b 1.885 a 2.066 a BW, g 2,841 a 2,811 a 2,679 b a c Estimates in a row within model with no common superscripts differ (P < 0.05). 1 Model 1 excludes and Model 2 includes BW at 49 d as covariable in the analysis of GS. 2 GSA = GS excluding score 4 and 5; BW = BW at 49 d of age.

AGE, STOCKING DENSITY, AND LEG WEAKNESS 867 TABLE 3. Simultaneous estimation of least squares means for gait score (GS) and BW for stocking densities (833, 625, and 435 cm 2 per bird) across ages (28, 42, and 49 d) and for age across densities in Trial 1 Density Age Model 1 Trait 2 833 cm 2 per bird 625 cm 2 per bird 435 cm 2 per bird 28 d 42 d 49 d 1 GS 1.431 c 1.587 b 1.755 a 0.911 c 1.825 b 2.038 a GSA 1.388 c 1.495 b 1.666 a 0.901 c 1.759 b 1.889 a BW 2,841 a 2,813 b 2,692 c 2 GS 1.419 c 1.573 b 1.756 a 0.906 c 1.810 b 2.033 a GSA 1.360 c 1.491 b 1.692 a 0.903 c 1.758 b 1.882 a a c Estimates in a row within variable with no common superscripts differ (P < 0.05). 1 Model 1 excludes and Model 2 includes BW at 49 d of age as a covariable in the analysis of GS. 2 GSA = GS excluding score 4 and 5; BW = BW at 49 d of age. males than for males at all ages (0.65 GS units; range: 0.53 to 0.91). Trial 2 The LS means for the traits measured are shown in Table 5. There was no difference in the prevalence of TD, tibial curvature, BW at 28 d, or angulation of the hock joint at either STD. Walking ability at 35 d was poorer at the higher density, and this finding did not change when the data set was adjusted to remove birds with a score of 4 or 5. The proportion of birds that scored 4 or 5 was low and was not affected by density. Birds at a STD of 455 cm 2 per bird were lighter at 35 d than were those at an STD of 625 cm 2 per bird. Birds at a density of 455 cm 2 per bird had higher scores for foot pad, hock burn, and breast blisters at slaughter at 42 d of age than birds at the lower density. When the observations were adjusted for differences in BW, the findings did not change. The difference in LS means of GS (from Model 2) between the two densities was 0.195 GS units at 35 d. In Trial 2, the DM content of the litter decreased through 42 d in birds stocked at both densities (Figure 2). The differences in DM between densities were significant (P < 0.05) at 2, 4, 5, and 6 wk of age. Correlation coefficients for the main traits measured in Trials 1 and 2 are shown in Tables 6 and 7, respectively. Walking ability at the different ages was strongly correlated. Correlation coefficients ranged from 0.41 to 0.56. Gait score and 49-d BW correlations, however, had a lower range than those among GS at the different ages. In Trial 2, weak correlations were found between TD and BW and between TD and walking ability (0.054 and 0.101, respectively). A strong correlation was found between walking ability and BW (0.511), and moderate correlations were found between hock burn and walking ability as well as BW (0.269 and 0.273, respectively). When correlation coefficients were calculated for each density group in each trial, the correlations mentioned previously were consistent among different groups. DISCUSSION In both trials, birds appeared overtly healthy and, at the lower densities, grew well. The distribution in GS and prevalence of severely lame birds (having a GS of 4 or 5) were similar to commercial flocks (Kestin et al., 1992) and those in previous studies (Sørensen et al., 1999; Su et al., 1999; and Kestin et al., 1999). The findings from these trials were reasonably consistent. Walking ability deteriorated with age. At 4 wk of age, the broilers had good walking ability; <1% of birds had a GS of 4 or 5. By 6 wk of age, the walking ability of the birds had become substantially poorer, and, by 7 wk, the walking ability had deteriorated further. The rate of deterioration in walking ability was faster between 4 and 6 wk than between 6 and 7 wk. Whether birds would have continued to experience poorer walking ability as they grew further is not known. Both sexes were affected by leg weakness. However, when adjusted for differences TABLE 4. Number and percentage of the birds with a gait score of 4 or 5 for different stocking densities and at different ages for both trials Density Trial Age 833 cm 2 per bird 625 cm 2 per bird 335/355 cm 2 per bird Total (d) (no.) (%) (no.) (%) (no.) (%) (no.) (%) 1 28 0 a 0 2 a 0.6 1 a 0.3 3 c 0.3 42 7 a 2.1 10 a 3.1 10 a 3.0 27 b 2.7 49 10 b 3.1 23 a 7.4 24 a 7.3 57 a 5.9 Total 17 b 1.7 35 a 3.6 35 a 3.5 87 2.9 2 35 4 a 0.4 1 a 0.1 5 0.2 a c Estimates in a row (or a column for the last column) with no common superscripts differ (P < 0.05).

868 SØRENSEN ET AL. TABLE 5. Least squares means for the two stocking densities for the traits measured in Trial 2 Model 1 2 Model 2 Trait 1 625 cm 2 per bird 455 cm 2 per bird 625 cm 2 per bird 455 cm 2 per bird TD 0.083 a 0.085 a 0.083 a 0.085 a TC 0.101 a 0.093 a 0.101 a 0.092 a GS 1.264 b 1.423 a 1.242 b 1.437 a GSA 1.253 b 1.420 a 1.231 b 1.434 a Foot 0.687 b 0.918 a 0.687 b 0.920 a Hock 0.272 b 0.576 a 0.267 b 0.574 a Angul 0.011 a 0.012 a 0.011 a 0.009 a Blist 0.141 b 0.186 a 0.143 b 0.185 a BW28, g 983 a 984 a BW35, g 1,507 a 1,489 b a c Estimates in a row within a model with no common superscripts differ (P < 0.05). 1 TD = tibial dyschondroplasia, TC = tibial curvature, GS = gait score, GSA = GS excluding score 4 and 5, Foot = foot pad burn, Hock = hock burn, Angul = valgus or varus angulation, Blist = breast blisters, BW28 = BW at 28 d of age, and BW35 = BW at 35 d of age. 2 Model 1 excludes and Model 2 includes BW as a covariable in the analyses of leg weakness traits. in live weight, males were approximately one-half GS unit higher than females at all ages. Because birds with a score of 4 or 5 were humanely killed when observed or died naturally, the proportion of birds that actually had a GS of 4 or 5 was greater at 6 and 7 wk than that presented in the tables. Because the number of birds and feeder space per bird in each pen were the same in all STD groups, some of the effects of high STD on live BW might have been a function of the difficulty birds had in accessing the feeders. Because of the strong correlation between GS and BW, it would normally be expected that birds at the lower densities, which were heavier, would have poorer walking ability; however, this was not the case. Birds at lower densities had better walking ability, despite higher live weight. This result meant that, when the observations were adjusted for differences in BW, the differences in walking ability between densities became larger. Birds with a score of 4 or 5 were light because they could not access feed easily (Kestin et al, 1992), which resulted in a nonlinear relationship between BW and GS. Because there were few birds with a score 4 or 5 at 28 d, the correlation coefficient between BW at 49 d and GS at 28 d was higher than that between BW at 49 d and GS at 49d. When these very lame birds were excluded from the analysis, the correlation coefficient between BW at 49 d and GS at 49 d increased considerably (Table 6). At 4 wk of age, from casual observation, there appeared to be adequate space for all birds to move around, even at the higher densities. At 6 and 7 wk of age, it was clear that, at higher densities, bird movement was more constrained. The improved walking ability in birds kept at lower densities might have been due to their greater level of overall activity. There is preliminary evidence that greater exercise is associated with improved walking ability (Reiter and Bessei, 1995). Whether deterioration in walking ability could be limited by encouraging the birds to exercise by, for example, increased lighting levels, which is believed to increase activity (FAWC, 1992), or by placing the feeders and drinkers some distance apart to encourage activity remains to be investigated. Based on the findings of this study, a thorough investigation of the role of bird activity and exercise at different ages on the development of leg weakness is merited. Females were more sensitive to density than were males with regard to walking ability, but the reasons are not clear. The low overall prevalence of TD found in Trial 2 is similar to that reported elsewhere (Sørensen et al., 1999; Su et al., 1999; Kestin et al., 1999) and probably reflects the progress broiler breeders have made reducing this problem. The lack of effect of STD on the prevalence of TD indicates that STD is not important in regulating the expression of this condition. This lack of importance is probably because TD lesions first develop very early in life before STD becomes important. It is probable that TD made only a modest contribution to the overall prevalence of leg weakness in this study. Not surprisingly, higher STD were associated with more foot and hock burn, and both were correlated with walking ability. In these studies, this result was probably TABLE 6. Correlation coefficients between gait scores at different ages and between gait score and BW in Trial 1 Traits GS42 1 GS49 BW49 Traits GSA42 GSA49 BW49 GS28 0.517** 0.415** 0.283** GSA28 0.498** 0.432** 0.281** GS42 0.560** 0.247** GSA42 0.547** 0.311** GS49 0.186** GSA49 0.352** 1 GS42 = gait score at 42 d, GSA42 = GS excluding score 4 and 5 at 42 d of age; and BW49 = BW at 49 d of age. **Correlation coefficient is different from zero (P < 0.01).

AGE, STOCKING DENSITY, AND LEG WEAKNESS 869 FIGURE 1. Gait scores (GS) in relation to age at different stocking densities (455, 625, and 833 cm 2 per bird) for the two sexes. Fem. = female. a reflection of poorer litter quality, indicated by the increased moisture content of the litter from pens housed at higher densities. Moreover, it was the heavier birds that were more affected with hock burn, as indicated by the positive correlation with BW. The correlation between walking ability and hock burn was of a similar order to the correlation between BW and hock burn. These findings are consistent with other studies (Kestin et al., 1999; Sørensen et al., 1999; Su et al., 1999). These studies taken together suggest that hock and foot burn make significant contributions to the overall prevalence of poor walking ability. The pathology of lameness in broiler chickens is complex, and the roles of foot and hock burn have not been investigated in detail. Recent studies have shown that birds with a GS of 4 or 5 are lame because they have infections in their bones and joints (mainly Staphylococcus aureus or Escherichia coli) or because they suffer from extreme angular bone deformity (Kestin et al., 1994; McNamee et al, 1998; Butterworth, 1999). Less disabled birds are lame principally because they suffer from angu- TABLE 7. Correlation coefficients of some particular pairs of traits in Trial 2 TD:BW28 1 0.054** TD:GS 0.101*** TD:GSA 0.099*** GS:BW35 0.511**** GSA:BW35 0.527**** GS:Hock 0.268*** GSA:Hock 0.270**** GSA:Foot 0.063** Hock:BW35 0.273**** 1 TD = tibial dyschondroplasia, BW28 = BW at 28 d of age, GS = gait score, GSA = GS excluding score 4 and 5, BW35 = BW at 35 d of age, Hock = hock burn, and Foot = foot pad burn. **P < 0.01. ****P < 0.0001. FIGURE 2. Dry matter of litter in relation to age of the chicken in Trial 2. The dry matter content at a high stocking density (455 cm 2 per bird) was significantly lower than that in litter from chickens reared at a stocking density of 625 cm 2 from 28 d of age and later. lar bone deformity, which bestows abnormal loads onto joints (Kestin et al., 1994). Lame birds are less mobile than sound birds (Weeks and Kestin, 1997). But whether hock and foot burns are promoted by other causes of lameness, because lame birds spend more time in contact with the litter, or whether hock and foot burns are the primary cause of lameness is not known. Whatever the cause of hock and foot burns, their control should be part of a leg weakness control program, which might be accomplished through good litter management. In Trial 2, GS at 35 d of age (adjusted for BW effects) for the lower STD was 0.195 GS units lower than that for the higher STD. In similar studies, we have found a difference between extremes of meal feeding treatments of 0.240 GS units (Su et al., 1999), 0.105 GS units for feed restriction treatments (Su et al., 1999), and 0.523 GS units between different broiler genotypes (Kestin et al., 1999). It would appear that the improvement in walking ability to be gained by manipulations of STD is of a similar size. There is increasing evidence that birds with a GS of 3 or more suffer from pain when they walk. Summarized, these include large behavioral differences between lame and normal birds; lame birds spend one-half the time that sound birds do in standing activities (Weeks and Kestin, 1997). Walking ability and agility in lame birds improves after the administration of an analgesic (Mc Geown et al., 1999), and lame birds self-select analgesic drugs when offered the opportunity (Danbury et al., 1999). Thus, There is good reason to believe that the welfare of these birds is compromised by leg weakness. Commercially, the use of the high STD is driven by the need to maximize, as far as possible, the economics of floor space utilization. Stocking density is limited when deleterious effects of high STD such as reduced bird growth and quality (Proudfoot et al., 1979; Proudfoot and

870 SØRENSEN ET AL. Hulan, 1985; Grashorn and Kutritz, 1991; Cravener et al., 1992) limit returns. The present studies indicate that high STD can also affect the welfare of the birds by exacerbating painful conditions such as leg weakness. These welfare effects should be considered when commercial STD is derived. Because leg disorders are painful, it is important that leg weakness control strategies are adopted. Meal feeding (Su et al., 1999), early feed restriction (Su et al., 1999), and the use of genotypes with a reduced propensity to develop leg disorders (Kestin et al., 1999) have already been identified to limit the expression of leg disorders. Reduced STD may be a valuable alternative method to limit leg weakness. REFERENCES Bartels, J. E., G. R. McDaniel, and F. J. Hoerr, 1989. Radiographic diagnosis of tibial dyschondroplasia in broilers: a field selection technique. Avian Dis. 33:254 257. Blokhuis, H. J., and J. W. van der Haar, 1990. The effect of the stocking density on the behaviour of broilers. Arch. Geflügelk. 54:74 77. Butterworth, A. J., 1999. Infectious components of broiler lameness. World s Poult. Sci. J. 55:327 352. Cravener, T. L., W. B. Roush, and M. M. Mashaly, 1992. Broiler production under varying population dentities. Poultry Sci. 71:427 433. Danbury, T. C., J. P. Chambers, C. A. Weeks, A. F. Waterman, and S. C. Kestin, 1999. Self-selection of the analgesic drug carprofen by lame broiler chickens. Vet. Rec. 146:307 311. Ducro, B. J., and P. Sørensen,1992. Evaluation of a selection experiment on tibial dyschondroplasia in broiler chickens. Pages 386 389 in: Proceedings of XIX World s Poultry Congress, Vol. 2. WPSA, Amsterdam, The Netherlands. EEC, 1971. Analytical methods in official control of feedstuff. Off. J. Eur. Commun. L 279/7:860. FAWC, 1992. Report on the welfare of broiler chicken. Farm Animal Welfare Council, MAFF Government Buildings, Hook Rise South, Tolworth, Surbiton, Surrey UK. Grashorn, M., and B. Kutritz, 1991. Der Influss der Besatzdichte auf die Leistung moderner Broilerherkünfte. Arch. Geflügelk. 55:84 90. Kestin, S. C., S.J.M. Adams, and N. G. Gregory, 1994. Leg weakness in broiler chickens, a review of studies using gait scoring. Pages 203 206 in: Proceedings of the 9th European Poultry Conference. Vol 1. United Kingdom Branch of the WPSA, Glasgow, Scotland. Kestin, S. C., T. G. Knowles, A. F. Tinch, and N. G. Gregory, 1992. The prevalence of leg weakness in broiler chickens and its relationship with genotype. Vet. Rec. 131:190 194. Kestin S. C., G. Su, and P. Sørensen, 1999. Different commercial broiler crosses have different susceptibilities to leg weakness. Poultry Sci. 78:1085 1090. Lewis, N. J., and J. F. Hurnik, 1990. Locomotion of broiler chickens in floor pens. Poultry Sci. 69:1087 1093. Martrenchar, A., J. P. Morisse, D. Huonnic, and J. P. Cotte, 1997. Influence of stocking density on some behavioural, physiological and productivity traits of broilers. Vet. Res. 28:473 480. Mc Geown D. A., T. C. Danbury, A. E. Waterman-Pearson, and S. C. Kestin, 1999. Effect of carprofen on lameness in broiler chicken. Vet. Rec. 144:668 671. McNamee, P. T., J. J. McCullagh, B. J. Thorp, H. J. Ball, D. Graham, S. J. McCullough, D. McConaghy, and J. A. Smyth, 1998. A study of leg weakness in two commercial broiler flocks. Vet. Rec. 143:131 135. Proudfoot, F. G., and H. W. Hulan, 1985. Effects of stocking density on the incidence of scabby hip syndrome among broiler chickens. Poultry Sci. 64:2001 2003. Proudfoot, F. G., H. W. Hulan, and D. R. Ramey, 1979. The effect of four stocking densities on broiler carcass grade, the incidence of breast blisters, and other performance traits. Poultry Sci. 58:791 793. Reiter, K., and W. Bessei, 1995. Influence of training on the locomotor ability of fast and slow growing broilers. Pages 206 217 in: Aktuelle Arbeiten zur artgemässen Tierhaltung. KTBL-Schrift, Darmstadt, Germany. SAS Institute, 1994. SAS/STAT User s Guide: Statistical Version 6.08. SAS Institute, Inc., Cary, NC. Sørensen P., G. Su, and S. C. Kestin, 1999. The effect of photoperiod/scotoperiod on leg weakness in broiler chickens. Poultry Sci. 78:336 342. Su, G., P. Sørensen, and S. C. Kestin, 1999. Meal feeding is more effective than feed restriction at reducing leg weakness in broiler chickens. Poultry Sci. 78:949 955. Thomsen, M. G., 1992. Influence of increasing stocking rates on performance and carcass quality of broilers. Pages 285 287 in: Fourth European Symposium on Poultry Welfare. C. J. Savory and B. O. Hughes, ed. Universities Federation for Animal Welfare, Herts, Great Britain. Weeks, C. A., and S. C. Kestin, 1997. The effect of leg weakness on the behaviour of broilers. Pages 117 118 in: Proceedings of the 5th Poultry Welfare Symposium. P. Koene and H. J. Blokhuis, ed. Agricultural University, Wageningen, The Netherlands.