s /Applied Animal Behaviour Science 44 (1995) 257-281 259 showed the least agonistic interactions (s, 0.49 f 0.21 h‘; sg, 2.17 f 0.24 h‘; esd, 1.29 + 0.11 h‘; P < 0.001). However, sows in stalls had fewer clearly resolved conflicts (s, 16.83 & 6.89%; sg, 51.07f4.48%; esf, 80.13 + 1.59%; P<O.OOl) and more high levels of aggression (s, 86.9+6.21%; sg, 20.12 f 4.4%; esf, 42.53 + 4.01; P < 0.001). By the fourth pregnancy, stall-housed sows spent 50.14+9.62% of the period O&30-17:30 h showing stereotypies or activities, such as drinking and rooting or chewing at pen fittings, which were sometimes stereotyped. Comparable figures for grouphoused sows were 8.19 f 1.79 for sg and 3.69 f 0.7% for esd (P < 0.001). The body weights of stallhoused sows were lower by the fourth parity (s, 220.9 f 6.86 kg; sg, 239.1 f 4.2 kg; esf, 237.2 f 2.87 kg; P < 0.02). Sows in stalls took longer to lie down (s, 20.4 f 3.5 s; sg, 11.1 f 0.9 s; P < 0.0001) and their lying time was not correlated with muscle proportions so they had less control over their lying movements than sows in groups. After slaughter, they had lower muscle weights (P < 0.05 for eight limb muscles) and considerably lower leg bone breaking strengths (e.g. femur 808.9 k 36.4 kg vs. 1120.4f 57.4 kg; P <O.OOl ) than group-housed sows. Both muscle and bone strength effects were presumably due to lack of exercise in stalls. Confinement in stalls resulted in substantially more indications of poor welfare than in either of the group-housing systems. The influence of man’s assistance at parturition on the neonatal behaviour of Altamurana breedpopulation lambs P. Cagnetta, G. Vonghia, L. Melodia Department of Animal Production, University of Bari, Bari, Italy