Published online by Cambridge University Press: 02 September 2010
Individual observations made in different weeks of lactation (WL) of live weight (LW, lb), milk yield (MY and FCM, lb/day), age (years) live-weight change (LWC, lb/day) and digestibility of the organic matter consumed (D, %) were used to produce models pre-dicting organic-matter intake (OMI, lb/day), digestibleorganic-matter intake (DOMI, lb/day) and faecal output (lb OM/day) in 72 grazing cows (Analysis A). Estimates of herbage availability (HA, lb dry matter/cow per day) applicable to 44 of the cows, together with the above variables were used in a similar manner (Analysis B). The multiple regression methods of analysis were applied solely to between-cow observations by selection of a single set of data from the many available for each cow.
FCM and LW0·73 gave better results than MY and LW, and together with age, were all significant variables for the prediction of the dependent variable. HA was not a significant source of variation, D was only effective in models of DOMI and faecal output, and stage of lactation was not considered a consistently useful variable. The application of the multiple regression technique to the problem of partition of nutrients and the prediction of intake is considered. Precision of prediction is discussed in terms of tolerance limits for predicted values, and it is concluded that the lack of accuracy of measurement of intake in the individual grazing cow is the major obstacle to the accurate prediction of intake in grazing cows.