Epidemiological studies relating occupational exposures of working women to spontaneous abortions and malformation are reviewed and some methodological considerations are presented. The reproductive epidemiology is less developed than epidemiology in general and seems to involve some specific problems. The ascertainment of exposure and of outcome may be problematic. The exposures may be reported differently by the women depending on the outcome of the pregnancy; thus confirmation of exposure from an independent data source would be an asset. The recognition of spontaneous abortions by women is variable and this should be considered in the study design. An access to medical records could provide objective data on the outcome, but it may introduce other types of asymmetries. The types of occupational exposures of the women, suggested to carry a risk of spontaneous abortions, include anesthetic agents, laboratory work, copper smelting, soldering, and chemical sterilization using ethylene oxide and glutaraldehyde. Maternal employment in laboratories and exposure to solvents has been linked to a risk of congenital malformations in the offspring in five studies. Data on the teratogenic effects of anesthetic gases has been conflicting. In one study, employment in copper smelting was associated with malformations in the offspring.