The potential problems in examining data for dose response relationships in occupational reproductive studies were discussed. Data from experimental teratologic studies suggested that with increasing dose levels, adverse response shifted from malformations and fetal loss to only fetal loss and ultimately to maternal death. Beyond a certain dose, gestational age of fetal loss could shift with increasing dose, possibly resulting in unrecognized fetal loss. If only birth defects were assessed, an inverse dose response relationship was observed with increasing dose beyond the dose which resulted in increasing fetal loss or beyond which malformed fetuses failed to survive to term. In order to obtain the most sensitive assessments of reproductive effects due to toxic exposures, several outcomes must be considered at several exposure levels. Many studies examined only recognized pregnancies, and failed to consider early fetal loss or menstrual irregularities resulting in subfertility. The ability to detect dose response relationships was often limited by the definition of exposure or dose. Analysis using the ordinal scale for the dose measure was considered the most prudent approach for reproductive studies. The importance of examining the relationship of timing of a particular exposure to reproductive outcome was emphasized. A cohort study which examines multiple outcomes and several exposure models is concluded to be the most useful approach for occupational reproductive studies.