Devir, S., Institute of Agricultural Engineering, ARO, PO Box 6, Bet Dagan, 50250, Israel, IMAG-DLO, Institute of Agricultural Engineering, PO Box 43, 6700 AA, Wageningen, Netherlands, The Technion-Israel Institute of Technology, Haifa, 32000, Israel, Department of Statistics and Operational Research, ARO, PO Box 6, Bet Dagan, 50250, Israel
Maltz, E., Institute of Agricultural Engineering, ARO, PO Box 6, Bet Dagan, 50250, Israel, The Technion-Israel Institute of Technology, Haifa, 32000, Israel

A multivariate linear model was developed to evaluate dairy cow body weight (BW) during lactation on a commercial dairy farm. The model was calibrated separately for first, second and greater than second parities, and also for winter and summer calvers, fed a total mixed ration of 65% concentrates. Average weekly BW was predicted using the following variables on a weekly basis: 4% fat-corrected milk (FCM) yield, time of conception (PREG), time from calving (w), initial body weight (1BW) measured within 2 days after calving, and weeks remaining until the end of the calving season (SWR). The ability of the model to predict mean BW along lactations was tested in the same herd following calibration and during the same period. The prediction capacity of the model for average BW of groups of cows was examined on two timescales: (i) time from calving (physiological timescale) and (ii) any time during the year (calendar timescale). Results indicated that predicted mean BW deviated from measured values either by 1-2% with a maximal deviation of c. 5% on the physiological timescale or by up to 2% on the calendar timescale. © 1995, Cambridge University Press. All rights reserved.