Sexing of human hip bones.

Eight measurements of the hipbone were taken from 62 human skeletons of unknown sex. The subjective judgements of eight experienced anatomists were used to obtain subsets of male, female and unknown sex specimens. Principal components analysis was employed as a group-finding procedure; this confirmed the subjective determinations. Discriminant function analysis, using the subjectively determined 'known' specimens, was conducted. This yielded scores which sexed the unknown subset in agreement with the groups found by principal component analysis. Further discriminant functions were calculated to exclude two hip bone measures which were shown to be poor indicators of sex. The specimens with discriminant function scores closest to the cut-off point were seen, in the graph of the first two principal components, to lie in a cluster intermediate between the main gender groups and, upon re-examination, these were found to be rather ambiguous specimens with small overall measures and a narrow sciatic notch and pubis. The results of this study indicate that mid-pubic width is a more useful sexing character than pubic length.