Many studies have reported digit ratio (2D:4D) to be sexually dimorphic, (males lower 2D:4D than females). However, Kratochvíl and Flegr ([ 2009]: Biol Lett 5:643-646) have suggested that 2D regressed on 4D has an allometric regression line with nonzero Y-intercept that is shared by males and females. Thus, 2D is shorter than expected when 4D is long, and males have lower 2D:4D than females because they have longer fingers. In this study, it is shown that this suggestion may be incorrect because sex differences in slope were not considered. Participants were recruited in an Internet study and had an age range of 12–30 years. The expected sex difference in 2D:4D was found, and the regression of 2D on 4D showed a significant sex difference in slope (males lower than females). A comparison of 10 age groups (12 years, 13 years…, 21–30 years) showed that sexual dimorphism for fingers was age dependent, varying from monomorphic to very dimorphic. Changes in sexual dimorphism of 2D:4D were much less marked, but there was a significant reduction in mean 2D:4D with age. The tendency for slopes of 2D regressed on 4D to be lower in males compared with females was significant in eight age groups. Sex difference in 2D:4D varied across the age groups and was positively related to the magnitude of the difference in female and male slopes. In contrast to the report of Kratochvíl and Flegr, it was found that the regression of 2D on 4D showed sex differences in slope, and such differences gave rise to the sexual dimorphism in 2D:4D. Am. J. Hum. Biol. 22:604–608, 2010. © 2010 Wiley-Liss, Inc.