In an attempt to resolve conflicting evidence from the literature concerning the existence of a paternal age effect in 46,X,i(Xq) Turner's syndrome, we have analysed data on all known cases ascertained in the main population centres of Scotland and on others ascertained in England, using population controls matched for year of birth. There was a significant (P = 0.02) increase of 2.3 years in the mean paternal age of the Scottish cases, and a smaller and non-significant increase in their mean maternal age. Logistic regression analysis confirmed that the primary association was with paternal, rather than maternal, age. For the English cases, however, there were small and non-significant decreases in their mean maternal and paternal ages. The differences between the two groups were also significant, but cannot be explained by any likely source of ascertainment bias. We therefore conclude that there is no evidence for a universal paternal age effect in this condition, but that at least one mechanism of origin, occurring with variable frequency, may be associated with increased paternal age. Using data from this and earlier published studies, we estimate the incidence of individuals with a 46,X,i(Xq) cell line to be between 3.3 and 13 per 10(5) female livebirths.