At least two spontaneous abortions were karyotyped in 273 women during cytogenetic surveys in New York City and Honolulu. These pairs were analyzed using maximum-likelihood logistic-regression analysis to adjust for maternal age and location. There was a significantly increased risk for a chromosomally normal spontaneous abortion after a previous abortion with a normal karyotype. There was no increased risk for trisomy in a second spontaneous abortion following either a previous trisomic abortion or an abortion with another abnormal karyotype. This is unexpected, given the increased risk for trisomy found among live births and at prenatal diagnosis in young women with a previous trisomic birth. The most likely explanation is that the increased recurrence risk for trisomy is restricted to trisomy for only one or a few chromosomes, for reasons such as parental trisomy mosaicism. These data predict no increased risk of chromosome abnormality in future pregnancies after either (1) spontaneous abortions with trisomies of a kind that are always lethal in utero or (2) multiple early abortions in the presence of normal parental karyotypes.