A preliminary examination of sham chewing behaviour in group-housed, nulliparous sows

Stereotypies are relatively invariant motor acts, repeated frequently, that have no apparent function (Mason and Rushen 2008). Whilst the origins and mechanisms of stereotypic behaviour remain unclear, it has been suggested that sows may develop stereotypies, such as sham chewing, as a means to cope with their environment (Mason and Rushen 2008). Despite the move from stall to group-housing during gestation, stereotypies are still anecdotally observed in Australian sow herds. Sham chewing is the most common and frequently observed stereotypy in group-housed sows (Vieuille-Thomas et al. 1995). The characteristics and welfare implications of stereotypic behaviour in group-housed sows have received little examination and their causation remains unknown. This preliminary study is part of a larger project investigating the relationships between sham chewing and the welfare and productivity of group-housed gestating sows.

This study aimed to characterise sham chewing behaviour in group-housed gestating sows with regard to average bout duration, bout frequency and the persistence of sham chewing, and develop a valid method of sampling this stereotypy. Archive video footage of 20 group-housed, nulliparous sows (two groups of 10 sows) in their first gestation was utilised. Gilts were twice artificially inseminated and within 7 days of insemination randomly mixed into groups of 10 (1.8 m²/gilt). A standard commercial gestation pelleted diet (13.1 MJ/kg DM and 12.8% protein; 31.3 kg per feeder per drop and 2.5 kg/sow/d) was delivered onto the floor in four feeding bouts drops (~0730, 0930, 1100 and 1500 h). Water was supplied ad libitum. One video camera with built-in infrared lights was positioned above each pen during gestation and video recordings were conducted from 0700 to 1700 h on d 3 (D3) and 8 (D8) post-mixing. An ethogram was developed; a sow was deemed visible if the observer could clearly view the snout and jaw, and sham chewing was defined as jaw movement without contact with any substrate. The performance of sham chewing was assessed using continuous sampling, instantaneous point sampling at 2 min intervals (2 min IPS), and instantaneous point-sampling at 5 min intervals (5 min IPS). Wilcoxon signed-rank tests (Z) were used to compare continuous sampling with the two instantaneous point sampling methods. Wilcoxon signed-rank tests were used to compare sham chewing characteristics on D3 and D8.

The average duration of sham chewing bouts was 54 s on D3 (s.d. 47) and 78 s on D8 (s.d. 173), and the difference in average bout duration between the 2 days was not significant (Z = –1.11, P = 0.27). The average frequency of sham chewing bouts was 8 bouts/sow/d on D3 and 5 bouts/sow/d on D8, and the difference in average frequency of bouts between the 2 days was significant (Z = –2.16, P = 0.03). 85% of sows were identified as performing sham chewing behaviour on D3 and 70% of sows on D8; however, the difference between the days was not significant (Z = –1.34, P = 0.18). Sham chewing was observed across the day, both pre- and post-feed drops. The proportion of sows identified as performing sham chewing behaviour with continuous sampling was significantly different to 5 min IPS on both D3 (Z = –2.45, P = 0.01) and D8 (Z = –2.65, P = 0.01), and 2 min IPS on D3 (Z = –2.00, P = 0.05) but not D8 (Z = –1.73, P = 0.08). When comparing the two instantaneous sampling methods, the proportion of sows identified as performing sham chewing behaviour was significantly different on D8 (Z = –2.00, P = 0.04) but not D3 (Z = –0.82, P = 0.41). Using 5 min IPS, 65% of sows were identified as performing sham chewing behaviour on D3 and 35% of sows on D8. While for 2 min IPS, 55% of sows were identified as performing sham chewing behaviour on D3 and 55% of sows on D8. Thus, while 2 min IPS was more accurate in identifying sows sham chewing than 5 min IPS, both instantaneous point sampling methods failed to identify some sows that continuous sampling identified.

These findings suggest that a point sampling frequency of 30 s is likely to be an effective method of sampling sham chewing. This is currently being investigated, with the subsequent aim of examining sham chewing behaviour in 200 group-housed gestating sows and its implications for their welfare. Whilst this preliminary study involves a small number of animals, it provides valuable data on the characteristics of sham chewing behaviour in group-housed sows and the development of an appropriate sampling method for use in larger scale studies.