Proceedings of the 3 rd International Seminar on Animal Industry, Bogor, Indonesia, 17-18 September 2015, pp 265-268 265 Effects of Selection on the Efficiency and Variability of Sow Reproduction and Maternal Abilities Parsaoran Silalahi 1,2,3, , Mohamad Agus Setiadi 3 , Dedy Duryadi 3 , Jean Gogué 4 , Yvon Billon 5 , Thierry Tribout 1,2 , Jean-Pierre Bidanel 1,2 1. INRA, UMR1313 GABI, 78350 Jouy-en-Josas, France Email : parsaoran.silalahi@jouy.inra.fr 2. AgroParisTech, UMR 1313, GABI, 75231 Paris Cedex 05, France 3. Bogor Agricultural University, 16680 Bogor, Indonesia 4. INRA, UE332 Domaine Expérimental de Bourges, 18520 Avord, France 5. INRA, UE1372 GENESI, 17700 Surgères, France Introduction Selection has been implemented in pig populations for several decades. Initially, the breeding goal was to increase growth, feed efficiency and carcass lean content. Since the early 1990’s, litter size has become a major component of the breeding goal in French Large White (LW) and Landrace (LR) breeds, leading to large genetic gains on litter size (Tribout et al., 2003; Guéry et al., 2009). This improvement may unfortunately be accompanied by detrimental effects on other economically important traits. For instance, an increase in piglet mortality was shown in French LW by Tribout et al. (2003) and Canario et al. (2007). Animal robustness, defined as the ability to maintain a good level of performance over a wide range of environments (Knap, 2005) could also unfavourably be affected by selection (Phocas et al., 2014). The adverse effects of selection are difficult to detect, as only a limited number of traits are routinely recorded in breeding programmes. The use of frozen semen is a simple and powerful way of measuring genetic trends for a large number of traits (Smith, 1977). This approach has been used in French LW population to estimate genetic trends from 1977 to 1998. Tribout et al. (2010) estimated genetic trends for growth and carcass traits. Preliminary results for reproduction traits were presented by Tribout et al. (2003). The objective of this study is to analyse the full set of data of the experiment described by Tribout et al. (2003; 2010) and investigate the effects of selection on the variability of sow and piglet performance as a potential indicator of their robustness. Materials and Methods Animals French large white sows were inseminated in the INRA GENESI experimental unit (17700 Surgères) with either stored frozen semen from 17 LW boars born in 1977 (G7) or with semen from 23 LW boars born in 1998 (G8). Thirty and 33 litters, respectively, were produced from G7 and G8 boars. After weaning, half of piglets from each litter and sex were transferred to the INRA experimental herd of Bourges (18520 Avord). Fifteen males from each group, as well as 74 G7 and 89 G8 females were chosen at random to produce a second generation by within group matings. Pigs from this second generation will be noted as D7 and D8. D7 and D8 sows were kept for 6 litters. They were managed under a batch farrowing system, with three weeks intervals between successive batches. They were inseminated twice when oestrus was detected. Farrowing was induced with prostaglandin on the 113 th day of gestation. The day of farrowing, total number born (TNB), number of stillbirths (SB) and number born alive (NBA) were recorded for each litter. All piglets born were individually weighed within 24 hours after farrowing. In order to disentangle sow and litter effects on piglet growth, cross fostering was used on a large scale in order to let sows from each group nurse 50% of their own piglets and 50% piglets from the other genetic group, and litter sizes were standardised (D7 and D8 sows nursed 7 and 12