In this paper we examine the age pattern of sterility in a natural fertility population of 16 English parishes. We examine estimators of sterility proposed by historical demographers. We demonstrated through Monte Carolo simulation of reproductive histories that the estimators proposed earlier work well only if the ages to which the estimates pertain are substantially modified from the original formulation. The new estimates show a much larger positive effect of childbearing on sterility than would the earlier ones. We also present estimates of the age pattern of sterility due solely to the process of ageing by eliminating secondary sterility induced by childbearing. This curve rises slowly until age 40, after which the proportion sterile increases rapidly with age. We find no evidence of a sharp rise in the risk of sterility in the 30s. We find strong evidence of a decline in fecundity by examining age-specific fertility rates only for those women who are known to be fecund because they later bear children. The evidence suggests only a moderate decline until ages 35-39 and a much steeper decline thereafter. Finally, we illustrate the danger of the use of a clinical test of infertility commonly employed. We show that if women are judged to be infertile because they have not become pregnant within one year of unprotected exposure, then a large fraction of those so judged will be falsely diagnosed. We conclude that the one-year period is too short.