The behaviour of 80 gilts and 14 boars was observed over a 54-day period in
a dynamic service system with pens of 20 gilts and 4-5 boars. The female p
opulation of this system was changed on a regular basis (four animals in fo
ur animals out per week) while boars were resident. Altogether 933 mating a
ttempts (MAs) were recorded and their quality was classified using quantita
tive behavioural criteria. Only 9% and 18% of the observed MAs were classif
ied as 'very good' and 'good', while 14, 24 and 35% were 'fair', 'poor' and
'very poor', respectively. Differences between and within boar teams in th
e mating quality parameters were significant tall p < 0.05). A significant
correlation was found between time to first mount (TTFM) and MA duration (r
= +0.58, p < 0.01). MAs followed a diurnal pattern and had a peak at about
0900 h (after feeding time). The number of MAs a gilt received per hour of
oestrus decreased (r = -0.55, P < 0.001), while mating quality score impro
ved (r = +0.38, p < 0.005). The overall pregnancy rate was 84%. Gilts with
a positive PD result had received more total MAs and total 'good' + 'very g
ood' MAs (T = 2.33, p < 0.05 and T = 2.25, p < 0.05, respectively), and wer
e significantly heavier at the time of selection (T = 2.80, p < 0.01). Mati
ng quality grade differed significantly with boar social status (chi(2) = 3
5.8, df = 16, p < 0.01) but this had no influence on mating frequency, time
to first mount or mating duration tall p > 0.05), indicating that subordin
ate boars were not deprived of sexual partners. Gilt lesion scores reached
a maximum 3 days after entry then subsequently declined (r = -0.75, p < 0.0
5). It was concluded that with a good knowledge of the factors influencing
the activity of the sexual partners kept in such a system, the dynamic serv
ice system could be a valuable alternative in future pig production. (C) 20
00 Elsevier Science B.V. All rights reserved.