The problem to be considered arose in connexion with applications for plant breeder's rights. An applicant has to submit material for scrutiny by the authority, one aspect considered being uniformity. The applicant may wish to conduct a preliminary test, of the same type as the authority's, to allow a calculation of the probability of passing the regulation test. In both tests only a sample is examined and the population frequency of off-types remains unknown. It is assumed here that the authority sets a critical frequency (e. g. 0–01) and specifies a size of sample (e.g. 300). The result is regarded as a failure if the observed frequency exceeds the critical frequency (e.g. if more than three off-types are observed in a total of 300). Given that this is the test procedure it is possible to calculate the conditional probability of passing the regulation test, knowing the result of a preliminary test.