Materials and methods Directory UMM :Data Elmu:jurnal:A:Applied Animal Behaviour Science:Vol67.Issue4.2000:

existence of a greater number of pigs feeding at any moment, and thereby encouraging social facilitation, and the possibility of an excessive number of animals preferring the Ž . same feeding point in large groups have been suggested by Spoolder et al. 1999 as potential sources of increased feeding competition in large group systems. Without evidence to the contrary, it is possible that such mechanisms may operate in large groups to elevate the demand for drinker access also. Although the volume of drinking water offered to growing pigs is generally not restricted, the ability of the pigs to gain access to the drinking points themselves has received little attention. The current Welfare Codes have adopted caution in advising the provision of one nipple drinker per 10 growing pigs fed a dry diet. Producers have frequently extended this ratio to one drinker per 20 animals without encountering problems. In view of this discrepancy and the increased popularity of large group housing systems, a reassessment of the current drinker provision recommendation is required. The consequences for drinking behaviour, overt aggression and performance of the pen as a whole resulting from the restriction of drinker allocation from one drinker per 10 pigs to one drinker per 20 pigs for animals housed in groups of 20 and 60 have Ž . recently been reported by Turner et al. 1999 . These findings indicate that, under the environmental conditions and feeding regime used in the experiment, the water intake, behaviour and performance of the pen as a whole were not compromised by the more restricted drinker allocation, large group size or the interaction of these two factors. However, the consequences for the detailed drinking behaviour of specific individuals within the pen have not yet been reported. Thus, the purpose of this investigation was, firstly, to describe in detail the drinking behaviour of growing pigs. The second objective was to examine how the drinking behaviour, aggression and performance of pigs of different weight categories, selected as examples of individuals likely to be different in their competitive ability, was influenced by changes to drinker allocation, group size and the interaction of these two factors.

2. Materials and methods

2.1. Animals and housing Ž Over the period January to July, 640 Large White = Landrace growing pigs start . weight 36 5.0 kg were allocated to one of four treatments for 5 weeks. Four replicates over time were allocated according to a 2 = 2 factorial design of two nipple Ž . Ž . drinker to pig ratios 1:10 vs. 1:20 and two group sizes 20 vs. 60 giving four treatment combinations. Each group comprised of 55 males and 45 females, and floor space allowance per pig remained constant. The groups of 20 were selected from three smaller commercial pens, and the groups of 60 from nine. The animals were housed in a commercial grower house, on a fully slatted concrete floor, and an Ž . automatically controlled natural ventilation ACNV system was in operation. Dry pelleted feed was offered ad libitum via one 5-space hopper for the groups of 20, and three such hoppers for the groups of 60. Hoppers were filled daily at 0900 h. Artificial light was provided for 8 h daily. Daily minimum and maximum ambient temperature was recorded for each pen 1 m above floor height. 2.2. Drinkers Ž The nipple drinkers were of standard commercial type Arato, ‘‘80 Pig Drinker’’, . Clacton-on-Sea, UK , suitable for the weight of pigs. A flow meter was attached to the pipe above the point at which it diverged to supply the individual nipples within each pen. The flow meters were calibrated regularly and daily water use per pen was recorded. The flow rate of the individual drinkers was measured at the beginning of each replicate, and showed a mean value of 670 S.E.M. 30.8 mlrmin. Drinkers were placed a minimum of 1 m apart and located adjacent to the feeding area, and more distant from the lying area. 2.3. Focal animals Two males and one female of each of three weight categories chosen to represent the Ž heaviest, mean and lightest weight animals in the pen heavy; 41.9 S.E.M. 0.57 kg, . medium; 35.7 S.E.M. 0.51 kg and light; 30.9 S.E.M. 0.63 kg , were selected on the basis of start weight alone. Detailed observations were concentrated on these nine focal animals per pen. 2.4. BehaÕiour 2.4.1. Drinking behaÕiour The time of onset and duration of every drinking bout of the focal pigs were recorded during a complete 24-h period using time lapse videos collected during their fourth week on trial. A drinking bout was defined as starting when the mouth of a pig made contact with a nipple drinker and terminating when this contact ended. Bouts below 5-s duration were not recorded as water may not have been drawn in that time. The interaction of treatment and weight category was compared for its effect on the drinking behaviour and diurnal pattern of drinking time of the focal animals. 2.4.2. Skin lesion score The body of the focal animals was divided into face, ears, neck, shoulders, flanks, back, rump and combined tail and ano-genital areas. The number of fresh cuts, scratches and abrasions in each area were counted, by a single observer, at 3 days post-mixing, and at the end of the 1st and subsequent weeks, up to week 4. The dimensions of the lesions were not quantified. 2.5. Performance Ž . Average daily gain ADG was calculated from liveweight measurements made on each individual pig in the pen on the day of mixing and at the end of weeks 2 and 4. 2.6. Statistical analysis Ž Interactions between the four treatment combinations hereafter abbreviated to treat- . ment and focal pig weight category on drinking behaviour, lesion score and ADG were investigated using analyses of variance with weight class, treatment and replicate as factors. Interactions between drinker ratio and group size on the drinking behaviour of the pooled focal pigs were examined using two-way analyses of variance. Changes in lesion score value over time were investigated using a one-way analysis of variance. The existence of significant correlations between the individual parameters of drinking behaviour and between mean lesion score and parameters of drinking behaviour were examined using Pearson Product Moment Correlations. A Wilcoxon Signed Rank Test was used to examine differences in drinking bout length during the period of the day with the maximum compared to minimum duration of drinking behaviour.

3. Results