This nonrandomised retrospective case-comparison survey was based on the hypothesis that craniofacial morphometric features can be used to identify individuals at greater risk for having children with a cleft. The theory of aetiological heterogeneity suggests there may be differences according to gender and cleft type. From a completely ascertained sample of 286 children with cleft lip and/or palate born in the West of Scotland between 1 January 1980 and 31 December 1984, a sample of 83 parents of the children with non-syndromic clefts volunteered for lateral cephalometric examination. A comparison group was derived from the archives of Glasgow Dental Hospital. Multivariate statistical analyses were applied to identify which parental craniofacial parameters, if any, determine predisposition to orofacial clefting. Compared to the male comparison group, the fathers of children with CL(P) were shown to have reduced mandibular and symphyseal areas (P < 0.001), reduced maxillary area (P < 0.01) and a shorter palatal length (P < 0.01). The cranial base angle was more acute (P < 0.01) and the cross-sectional area of the cranium on lateral skull radiographs was significantly smaller (P < 0.001). However, the occipital subtenuce was larger in the fathers (P < 0.05). The craniofacial morphology in the mothers of children with CL(P) was characterized by a longer mandible (P = 0.011), an increase in the anterior facial height (P < 0.05) and greater facial length (P < 0.01). Anterior cranial base and the clivus length were also larger in the mothers (P < 0.05). The cranial parameters showed a similar trend to the paternal group with a reduced cranial area (P < 0.01) and an increase in the occipital subtenuce length (P < 0.001). Different cephalometric parameters distinguish fathers from a male comparison group and mothers from their female counterparts. An awareness of these parameters might be of value in the prediction of liability to clefting and may prove to be important in the quest for clues to the pathogenesis of both CP and CL(P).