The study population consists of 424 three-generation families originally ascertained through nonsyndromic cleft lip with or without cleft palate (CL +/- P) surgical probands by Carter et al [J Med Genet 19:246-261, 1982] in London, England. Carter et al proposed that the multifactorial threshold model (MF/T) could explain the data. The goal of our study was to test that hypothesis, plus alternatives, rigorously. Two approaches were used: 1) Carter et al had proposed that these data were consistent with the predictions of the MF/T as presented by Carter [Br Med Bull 25:52-57, 1969]. However, we tested those predictions using standard chi 2 tests and found statistically significant departures from the predictions in these families. 2) Complex segregation analysis under the mixed model was performed. Again, the MF/T model could be rejected, as could a model of a major locus alone. The best-fitting model included both major locus and multifactorial components. When the data were analyzed in two parts based on the proband's phenotype (CL vs CL + P) there was some evidence of heterogeneity in that there was a significant proportion of sporadic cases in the families of CL probands but not in the families of CL + P probands. Our results provide no support for the MF/T model. The results from segregation analyses of CL +/- P in these families were most consistent with autosomal major gene inheritance plus multifactorial contributions.