[The measuring frequency necessary for food intake and milk yield data in feeding studies in dairy cows]

An individual feeding experiment was carried out which comprised the daily milk analysis from Monday to Friday with proportional samples from the evening and morning milkings and the daily intake of the individual feed components being ascertained. For various measuring frequencies the following calculations were made: the mean values and their standard deviations, the absolute and relative control errors and their standard deviations, the standard deviations for the measuring error and the simple linear correlations and regressions between the various measuring frequencies for the individual parameters. From the results one can conclude that, in individual feeding experiments with dairy cows, five milk analyses per week do not bring about an essential improvement of accuracy in comparison with two analyses. If the composition of the ration remains the same and the same roughage is offered ad libitum, one can project the measuring of feed intake on five days per week for that of the whole week.