Two girls with Turner's phenotype and 45 X karyotypes in perpheral blood have good sexual development including menses. Buccal smears showed Barr bodies in 1/200 and 0/200 cells. Paraovarian fibroblasts from one girl gave 45 X on 40 spreads. Ovarian biopsies on both showed fibrous gonads with sparse follicles (est. total 8000/2 ovaries). Limited studies of estrogens and gonadotropins were normal. Traditional interpretation invokes mosaicism. A better interpretation is that these girls represent an expected extreme of a spectrum; that, because both X's are active in oogenesis before meiosis (Migeon and Jelalian, Nature 269:242, 1977), the XO complement causes an increased rate of atresia of follicles from the high fetal number, but does not prevent maturation. Somatic features of Turner's syndrome - almost uniform short stature and highly variable combinations of other features-can be explained by random inactivation during embryogenesis of the only X (Gartler and Sparkes, Lancet II:411, 1963) with ensuing cell death. This has never been seen because such cells are nonviable. This hypothesis would explain the high fetal wastage yet good viability of individuals who survive embryogenesis. The predicted reduction of cell number in early embryogenesis explains the uniform short stature in survivors. If such inactivation occurs, inactivation of the second X in normal cells must be prevented by interaction between the heterochromatic X(s) and the euchromatic X.