Abstract

Devices designed to detect motion (CowTrakker™, Bou-Matic, USA) were attached to Holstein-Friesian cows of either NZ or overseas origin. Animals of each strain either grazed pasture or were fed a total mixed ration. Milk samples were collected twice weekly and progesterone content analysed to identify ovulation events. Cows had tail paint applied from 2 weeks after calving, and were observed twice daily at milking for oestrous activity. Recording continued until 6 weeks after the planned start of mating (PSM). Cows were artificially inseminated following a progesterone-based synchrony programme at PSM. CowTrakker™ data were analysed using four algorithms (T3/1, T3/3, T10/1 and T10/3), which varied according to the period of the rolling average used (3 or 10 days) and the standard deviations from that average (1 or 3). These measurements were related to behavioural observations and the occurrence of ovulation. There was a significant number of false positive events detected by the CowTrakker with all algorithms (sensitivity: 12.3, 46.9, 62.3 and 81.5%; specificity: 67.1, 88.4, 95.4 and 99.0%, for T3/1, T3/3, T10/1 and T10/3, respectively) as compared to behavioural observations alone (sensitivity: 71.6%; specificity: 99.6%). The CowTrakker™ was a useful heat detection aid, but must be used in conjunction with careful behavioural observations in order to interpret the high proportion of false positive events that are recorded.

GA, Verkerk, RW Claycomb, VK Taufa, P Copeman, A Napper, and E Kolver

Proceedings of the New Zealand Society of Animal Production, Volume 61, Christchurch, 172-175, 2001