Effect of group composition and feeding system on behaviour, production and health of dairy heifers in deep bedding systems

An experiment was conducted in eight Danish commercial dairy herds, using 1 68 Danish Friesian heifers for 5 months, from November 1995. Three groups w ere established in each herd: two small, homogeneous groups (SHom) with fiv e light heifers (130-250 kg) or five heavy heifers (250-380 kg) and one lar ge, heterogeneous group (LHet) with 10 light and heavy heifers (130-380 kg) . The space allowance per heifer was 4.5 m(2) including a 2.7-m(2) resting area (deep litter). Four herds were fed restrictive amounts of concentrate per pen and ammonium-treated straw ad libitum (RCS) and four herds were fed one ration of either silage or total mixed ration (TMR ad libitum (OR). Th e group composition had no general affect on the live weight gain (LWG). Wh en using the RCS feeding system, the daily LWG in the LHet groups compared with the SHom groups decreased by 95 g for the light heifers, whereas it in creased by 31 g for the heavy heifers. The average feed intake was 7% lower in the LHet groups than in the SHom groups (P = 0.02). When heifers were f ed RCS, light heifers spent less time eating concentrates (P = 0.01) and fo r heavy heifers, there was a tendency to spend more time eating concentrate s (P = 0.10) in the LHet groups compared with SHom groups. The effect of gr oup composition on synchronization of resting behaviour (when all heifers l ie down simultaneously) revealed a tendency towards less synchronization in the LHet groups when fed OR (P = 0.09), but not with RCS feeding (P = 0.16 ) compared with the SHom groups. For heavy heifers, the aggression was decr eased (P = 0.02) when fed RCS, whereas it increased when heifers were fed O R (P = 0.001) in the LHet groups compared with the SHom groups. Grouping cr iteria regardless of the feeding system did not influence the heifers? heal th. Whereas productivity was not decreased by increasing group size and gro up heterogeneity, the welfare of light heifers seems to some extent to be a ffected negatively by housing in a large heterogeneous group. Contrary to e xpectations, this negative effect could not be diminished by offering one r ation (TMR or silage) nd libitum to the heifers.