THERE is considerable evidence that, as a result of some early embryonic event, the genes of one X chromosome in each somatic cell of the human female are not expressed1. Studies of somatic tissues and cultured cells (including fibroblast clones) from human embryos, heterozygous for the common electrophoretic variant of glucose-6-phosphate dehydrogenase (G6PD), indicate that a single active X is present in cells from various tissues, at least by 5 weeks from conception2. On the other hand, the presence of the heteropolymorphic form of the enzyme in oocytes of the heterozygous adult female3 and a 16-week-old foetus4 is compelling evidence that, in meiotic stage germ cells, there are two active X chromosomes; however, the basis for the two active chromosomes in oocytes is not clear. It is conceivable that germ cell progenitors escape inactivation because inactivation occurs when cells destined to become germ cells have already been imprinted. Alternatively, only a single X chromosome may be active in all cells of the early zygote, with the second X activated only in germ cells at some time after their differentiation. The third possibility, that germ cells are subject to inactivation, but that reactivation occurs when meiosis commences, has been suggested by Gartler et al. because they did not observe the heteropolymer in ovaries from a 12-week-old human foetus heterozygous for the G6PD variant5,6. The ovary at that developmental stage contains approximately 23% germ cells, with a preponderance of premeiotic and leptotene germ cells7. But the implication that two functional X chromosomes are required for the onset of mammalian meiosis is not supported by evidence that XO mice produce normal gametes8. Using human foetal material, we have now obtained evidence against the third possibility.