Abstract #T15

# T15
Assessing human-conditioned sorting behavior in dairy cows in farm research trials.
D. Manriquez*1, L. Chen2, G. Albornoz2, J. Velez2, P. J. Pinedo1, 1Department of Animal Science, Colorado State University, Fort Collins, CO, 2Aurora Organic Dairy, Platteville, CO.

Cow-human interactions can influence and modulate group and individual behaviors of dairy cows. Our objective was to test the effectiveness of human sorting to separate subgroups of lactating dairy cows in on-farm studies and to assess the level of conditioning to this activity. As part of a previous nutritional trial cows were sorted daily for 150 d into 2 contiguous sub-pens. Subsequently, we compared 3 sorting methods applied to the same cows: (1) human active sorting (AS) at the pen gate; (2) human presence as passive sorting (PS); and (3) non-human gate sorting (GS). We hypothesized that after a training period cows become conditioned to sorting. Holstein cows (n = 176; parity = 2.5 ± 1.3), were randomly assigned into 2 subgroups (A = 91 cows; B = 85 cows) to be sequentially separated by 3 sequentially applied sorting methods (AS; PS; and GS). Each sorting method was evaluated once per day after the morning milking during 5 d. The counts of correctly allocated cows and the error rate (misplaced-cows/d) of each subgroup were compared by sorting method. Additionally, an individual error index rate was calculated and analyzed considering treatment group, parity, and general activity ratio (daily rumination / daily activity). When AS was applied, the total proportion of animals correctly sorted was of 99.8%, whereas PS had 94.8% of sorting accuracy (P < 0.001). Non-human GS could not be accurately assessed because the cows lost the self-sorting behavior overcrowding one side of the pen making impossible the data collection. The average of animals correctly placed was greater in AS when compared with PS (175 ± 1.7 cows vs. 166.6 ± 3.5 cows; P = 0.005). Cows that had longer walking distance to their research sub pen had greater individual error rate, especially when PS was applied. No association was found between parity and general activity ratio with the individual error index. Researchers observed a clear self-sorting behavior in response to human sorting, for both AS and PS, with average values for correct placement above 90%. Therefore, after a period of training, lactating dairy cows became operant conditioned to human sorting, which represent an opportunity for animal separation without intense human labor or stressing practices for the animals.

Key Words: sorting, operant conditioning