Study design

The experimental design was planned as follows.

Three alternative patterns of feeding, as described below, were to be assigned at random to 30 goats that had recently kidded.

The goats were to be fed the control diet for 15 days after kidding prior to the experimental 60-day period starting.

It was planned to use average milk yield determined during this initial period as a covariate to adjust for the differences in the milk potentials of the different goats.