R. C. DE VEKEY and R. LOVEGROVE Building Research Establishment, Garston.


Most UK standards for masonry units measure compliance with strength specifications using the mean result of destructive tests on a sample of 10, randomly selected from a batch or lot. Other countries in Europe use different sample sizes for example, 6 in Germany and 20 in Holland for some products. On the basis that the distribution of strengths for a batch of such products can be assumed to be normal (Gaussian) the size of the sample required to give a producer’s risk of 1 in 40 has been calculated as a function of the variability of the population and the difference between the specified value and the mean of the batch (fi—X). This indicates that a sample size in the range 5-10 is more logical than arbitrary since the response curve for samples in excess of 10 is very sensitive to changes in variability and i-t -X and could give control problems if exploited.