Maternal age-specific rates of chromosome abnormalities in women undergoing chorionic villus sampling (CVS) were evaluated. These were rates derived from regression equations of the form rate = ln(b[age] + c) where age varied from 35 to 48 years; and b and c are parameters. For Down syndrome (47, + 21), b = 0.29 and c = -15.53; for other nonmosaic abnormalities, b = 0.25 and c = -14.11; and, for all abnormalities, b = 0.27 and c = -14.22. The predicted rates per 1,000 varied, respectively, from 4.2, 4.6, and 8.8 at age 35 years to 180.0, 120.0, and 300.0 at age 48 years. The predicted numbers of abnormalities were compared with those observed in an overlapping data set reported in the literature. For 47, + 21, the ratio of observed (O) to expected (E) numbers was 54/56.4 = 0.96, suggesting that these rates are at least not grossly erroneous. For non-47, + 21 chromosome abnormalities, the ratio of O to E was 66/51.1 = 1.29. The possibility of geographic variation in rates of non-47, + 21 chromosome abnormalities cannot be dismissed. Finally, if 1) the regression derived rates are correct, 2) there is no difference in selective loss of chromosome abnormalities between younger and older mothers, and 3) the ratio of rates at ages 35 and 40 years to those in the women of all ages are the same for all abnormalities at CVS as for 47, + 21 in live births, then the proportion of all women with viable conceptuses at time of CVS who have an embryo or fetus with cytogenetic abnormality is 3.8-4.4 per 1,000 conceptuses.