Biological standardization (BS) aims at equilibration of the activity of allergen preparations from different types of allergen source materials. The biological unit (BU), proposed in the Nordic guidelines for 20 patients, has been found reproducible among different countries in Europe, but to be relatively imprecise, with a 95% confidence interval of about one power of 10. A more precise estimate of the biological activity of allergens or difference in sensitivity between populations would be of value. We used Ch, i.e. the concentration of allergen eliciting a wheal of the same size as histamine in the individual patient, estimated by regression line analysis. The Ch of 36 patients included in a BS trial was used. One of the 36 Ch-values was drawn randomly, and then sent back to the sample. This procedure was repeated 10, 20, 30, 40 and 60 times to create "samples" of different sizes. Ten samples of each size were produced. With 60 "individuals", the 95% confidence interval of the sample and the confidence interval of the medians were reduced to less than a factor of 2, i.e. to 74 to 128% of the median of the medians.