Variability and repeatability in gestation length related to litter performance in female pigs on commercial farms.

Forecasting gestation length (GL) in sows enables producers to be prepared to provide assistance at farrowing, or for timely treatment to induce parturition. The objectives of the present study were to determine GL across parities, the repeatability and correlation of the GL, and associations of GL with the three litter size variables (total pigs born, pigs born alive, and dead piglets) and longevity. This study was conducted on 94 farms and encompassed 66,254 farrowing records of 13,715 sows born during 1999. Variance components and correlation analyses were used to determine the repeatability and the correlations of GL. Mixed-effects models were used to analyze associations of GL with litter size variables and longevity. The mean of GL across parities was from 115.2 to 115.4 d. The proportions of GL 114, 115, and 116 d for all farrowing events were 19.2, 30.8, and 22.2%, respectively. The GL between parities were correlated (0.40</=r</=0.56; P<0.01). The repeatability of GL was 0.50 on commercial farms. Shorter GL was associated with more numbers of total pigs born and dead piglets (P<0.01). Sows with GL 113 to 116 d had more pigs born alive than those with GL</=112 and >/=117 d (P<0.01). Sows with GL</=112 d were 1.57 and 2.21 times as likely to be culled as those with GL 115 d (P<0.01). It is recommended that swine producers maintain records of GL at each parity in individual sows.