Quantitative predictions of rates of Down syndrome offspring as a function of maternal serum alpha-fetoprotein (MSAFP) values and of maternal age were reviewed. Comparisons were made of 35-year-equivalent-risk values by maternal age, that is, MSAFP values (at various maternal ages) that predict the risk (of a Down syndrome offspring) equivalent to that of an average 35-year-old woman irrespective of knowledge of MSAFP. This boundary was chosen because MSAFP values that predict a greater risk than this point for younger women or a lower risk for older women are likely in many jurisdictions to alter a decision about amniocentesis that would be reached without knowledge of MSAFP. There were marked differences among available studies in these 35-year-equivalent-risk values for older women. For instance, for age 41, the values among studies vary from 1.5 MoM to 2.1 MoM and the predicted proportion of 41-year-old women likely to be affected clinically by MSAFP determinations by the above criterion is 4.4%-18%, depending on which rate schedule is used. At age 39, the variation is 1.1 to 1.9 MoM or 7.1% to 38%. Possible explanations for the variation include sampling fluctuation. Relatively few data on Down syndrome have been used to calculate the parameters of the gaussian distributions used by the studies reviewed. It is suggested that great caution be used before employing published rate schedules for genetic counseling of older women, at least until sufficient data are collected and the nature of the discrepancies among studies is clarified.