Familial resemblance in intellectual skills is well documented, but its interpretation is a source of continuing controversy. The critical problem is that a family's shared genes are confounded with its shared experiences, and controls possible in animal research (selective mating, cross-fostering, and uniform or randomised environments) do not directly apply to human subjects. Conventional twin and family methods reveal substantial genetic variance in intelligence quotient (IQ) test scores, but the same methods also document significant environmental influences. Research designs which can identify the nature of these environmental factors may effect progress in the 'IQ debate' (ref. 1). The families of monozygotic (MZ) twins provide a new research design which permits a unique assessment of maternal influences in quantitative traits. We describe here initial applications of the design to verbal IQ, with results suggesting that maternal effects significantly contribute to familial similarity in verbal intelligence.