Behaviour of pre-pubertal gilts and its relationship to farrowing behaviour in conventional farrowing crates and loose-housed pens

• This study investigated gilt behaviour in pre-pubertal tests and at farrowing in loose-housed pensor conventional crates.
• Gilt behaviour in pre-pubertal tests was influenced by first test type experienced.
• Piglet-directed aggression at farrowing was lower in free-farrowing pens compared with conventional crates.
• Gilts that crushed one or more piglets were slower to contact during the first pre-pubertal test.
• Gilts showing severe piglet-directed aggression at farrowing tended to make more contactduring the first test.

Individual variation in the reproductive performance of sows has the potential for greater negative impacts in loose-farrowing systems. Therefore, the ability to select gilts that will perform well would be a major advantage. This study investigated the behaviour of gilts during pre-pubertal tests and farrowing behaviour in conventional crates and PigSAFE (Piglet and Sow Alternative Farrowing Environment) pens. Gilts underwent two phases of behavioural testing. First, gilts were subjected to three individual human interaction and three startle object tests randomly allocated to test sessions over 3 days (i.e. gilts had either a human or startle test first). Three weeks later, gilts underwent three human interaction and three novel object tests, in their stable group of six. Gilts farrowed in individual PigSAFE pens or conventional crates and behaviour was observed for 8 h from the first piglet birth. Data were analysed using linear mixed models and Spearman's rank correlations. A novel finding was the effect of individual test order: gilts that had the human interaction or startle object test first behaved differently. The first test was different whichever test type, with a higher latency to interact with the object or human, and gilts experiencing the startle test first interacted more with the human in all three subsequent tests. Gilts farrowing in crates and pens showed differences in behaviour, most notably, a lower frequency of piglet-directed aggression was seen in pens (P < 0.05). Piglet-directed aggression was studied further by comparing gilts that exhibited no aggression, to those showing aggressive behaviour, but no injurious biting, to those causing injury or death. This latter severely aggressive group spent more time alert, piglet focused and standing (P < 0.05) compared with the other two groups and tended to show greater (P < 0.1) contact duration in the first individual pre-pubertal test. Gilts that crushed one or more piglets were slower (P = 0.038) to contact either the human or startle object in the first individual test, than those that did not crush. The impact of first individual test on behaviour in subsequent tests indicates that previous test experience could be influencing subsequent behaviour. Differences in gilts showing severe piglet-directed aggression and between ‘crushers’ and ‘non-crushers’ suggests that it could be possible to use pre-pubertal behaviour to predict maternal ability.