The employment histories of 1,252 women experiencing a spontaneous abortion and 2,126 controls were compared to examine the relation between maternal employment before and during pregnancy and karyotype of the spontaneously aborted conceptus. Among private patients, there was no evidence of a positive association of work only before pregnancy, only during pregnancy, or both before and during pregnancy (v no work) with either chromosomally normal or abnormal abortions. In contrast, among public patients the odds of working only during pregnancy or both before and during pregnancy were raised for chromosomally abnormal abortions (adjusted odds ratios = 3.11 and 1.86, respectively) and slightly raised for chromosomally normal abortions (adjusted odds ratios = 1.35 and 1.31, respectively). Among workers, the frequencies of work before pregnancy in specific employment circumstances were compared among karyotyped cases and controls. For two work locations--factory and hospital/nursing facility--the data were of sufficient size to suggest that associations of twofold or greater with most types of abortion are unlikely.