Methods are specified enabling a sample to be used as the basis for determining a statistical tolerance interval, which may be two-sided or one-sided. They are applicable only where the sample unity may be assumed to have been selected at random and are independent, and if the distribution is normal (if appropriate, transforming the variate may make it normal). Tables 1 and 2 are applicable to the case where the standard deviation for the population is known, tables 3 and 4 where the mean and the standard deviation are unknown.