Hundreds of environmental chemicals affect prenatal development in experimental animals. However, only methylmercury and PCBs have been connected with such effects in humans during localized outbreaks of high exposure. In addition growth and development might also be affected by long-term intake of lead, fluorides or PCBs. Several factors may explain the discrepancy between human and animal data: the actual exposure of the population is below threshold levels, unspecific or delayed effects can be difficult to identify, etc. When experimental data are used to assess the hazards for the conceptus, due consideration should be given to actual ability of the study to detect effects. Thus, the limitations in statistical power, the relevance of the parameters considered and low-dose extrapolation should be taken into account. Finally, understanding toxicokinetics and biological mechanisms is needed to perform interspecies comparisons. Three examples of environmental chemicals showing different prenatal hazards are presented: thiabendazole, a benzimidazole compound with a moderate teratogenic potential, but which could represent a good model for biological extrapolation; nitrofen, a diphenyl ether herbicide which may pose a significant hazard, because of its high potential, toxicokinetics, and specific, hormone-like, teratogenic mechanisms; PCBs, well-known, global, cumulative pollutants which are not teratogenic in the laboratory animals, but may affect the human conceptus at high intake levels.