Welfare implications of non-nutritive visits to an electronic feeder by pregnant gilts

Electronic sow feeders (ESFs) are often used to feed pregnant sows. They collect detailed information on feeding behaviour which could potentially relate to sow welfare status. This study investigated associations between patterns of gilt feeding behaviour during pregnancy, specifically between non-nutritive visits to the ESF, and welfare indicators (performance of oral stereotypy behaviour; skin lesion counts; locomotory ability, and hair cortisol concentrations). Study gilts (n = 51; 8 replicates) were housed in a dynamic group (total group size range: 15–33 gilts; day 30–108 of pregnancy) and fed by an ESF which automatically recorded nutritive and non-nutritive visit start and end times. These data were used to calculate the average duration of non-nutritive visits (these took place after a gilt consumed her daily feed ration), the coefficient of variation (CV) of daily non-nutritive visits, the total number of non-nutritive visits throughout pregnancy, and the order of entry to the ESF four days post-mixing (a proxy metric of dominance status). Oral stereotypy behaviours (30 instantaneous scans/gilt/day; 3 hr/gilt/day) were observed directly 72 hr post-mixing, in mid and late pregnancy. Skin lesions were counted 24 hr and three weeks post-mixing, and in late pregnancy. Locomotory ability was scored using a visual analogue scale (0–150 mm) in mid and late pregnancy. Hair was shaved from gilts in late pregnancy, and cortisol concentration determined. Gilts that spent less time in the ESF and made fewer non-nutritive ESF visits performed more oral stereotypies 72 hr post-mixing (regression coefficient; REG: −1.6 ± 0.77, P = 0.041; REG: −0.02 ± 0.01, P = 0.043) and in mid-pregnancy (REG: −2.6 ± 0.83, P = 0.003; REG: −0.03 ± 0.01, P = 0.002, respectively). They also had lower skin lesion counts three weeks post-mixing (REG: 0.1 ± 0.02, P = 0.001; REG: 0.001 ± 0.0003, P = 0.001) and in late pregnancy (REG: 0.05 ± 0.023, P = 0.037; REG: 0.001 ± 0.0003, P = 0.041), and worse locomotory ability in mid-pregnancy (REG: −1.6 ± 0.58, P = 0.008; REG: −0.02 ± 0.01, P = 0.005). Gilts which showed more variability in the number of daily non-nutritive visits to the ESF performed more oral stereotypies 72 hr post-mixing (REG: 0.3 ± 0.15, P = 0.040) and in mid-pregnancy (REG: 0.44 ± 0.16, P = 0.010), and had lower skin lesion counts in late pregnancy (REG: −0.01 ± 0.004, P = 0.016). Gilts that entered the ESF earlier in the entry order four days post-mixing had higher locomotion scores in mid-pregnancy (REG: −0.3 ± 0.11, P = 0.010). There were no associations between feeding behaviour patterns and hair cortisol concentrations. However, given the numerous other associations, patterns of non-nutritive visits to the ESF may help inform on the welfare status of pregnant gilts. • Gilt feeding patterns associated with oral stereotypy, locomotion, and skin lesions. • Patterns of non-nutritive ESF use can relate to sow welfare status in pregnancy. • ESF data offer potential to automate sow welfare assessments.