Accuracy of Sampling Procedures for Estimating Lactation Yields: A Review

Sixty research reports dealing with the estimation of lactation yields from samples taken at various intervals were reviewed. In general, these reports showed that the average error in lactation yields is primarily a function of the length of the interval between tests. The variation in errors in milk yield (s -----2.7%) from monthly sampling for a 24-hr period appeared lower than those for fat per cent (s--~ 3.5%) or milk fat yield (s ~-5.0%) from the same testing scheme. Actual distributions of errors showed that at least 90% of milk yields estimated from a single day's yield once a month are within ~ 5% of true production. Errors in milk yields computed from bimonthly samples are about 30% greater than those based on monthly tests. Yet, evidence clearly shows that accurate cow ranking and progeny testing can be based on samples taken as much as two months apart. Research on other proposed sampling plans was also reviewed and discussed.