In order to evaluate the reproductive risks of low-frequency electromagnetic fields (EMF), it is important to include epidemiological and animal studies in the evaluation, as well as the appropriate basic science information in developmental biology and teratology. This review presents a critical review of in vivo animal studies and in vitro tests, as well as the biological plausibility of the allegations of reproductive risks. In vitro or in vivo studies in nonhuman species can be used to study mechanisms and the effects that have been suggested by human investigations. Only well designed whole-animal teratology studies are appropriate when the epidemiologists and clinical teratologists are uncertain about the environmental risks. Even the inference of teratogenesis cannot be drawn from culture experiments, because the investigator is not in a position to know whether any of his observations will be manifested in living organisms at term. Other aspects of reproductive failure such as abortion, infertility, stillbirth, and prematurity, cannot be addressed by in vitro or culture experiments. In fact, they are very difficult to design and interpret in nonprimate in vivo models. The biological plausibility some of the basic mechanisms involved in reproductive pathology were evaluated, concentrating primarily on the mechanisms involved in the production of birth defects. The studies dealing with mutagenesis, cell death and cell proliferation using in vitro systems do not indicate that EMFs have the potential for deleteriously affecting proliferating and differentiating embryonic cells at the exposures to which populations are usually exposed. Of course, there is no environmental agent that has no effect, deleterious or not, at very high exposures. The animal and in vitro studies dealing with the reproductive effects of EMF exposure are extensive. There are >70 EMF research projects that deal with some aspect of reproduction and growth. Unfortunately, a large proportion of the embryology studies used the chick embryo and evaluated the presence or absence of teratogenesis after 48-52 h of development. This is not a stage of development at which an investigator could determine whether teratogenesis occurred. The presence of clinically relevant teratogenesis can only be determined at the end of the gestational period. The chick embryo studies are also of little assistance to the epidemiologist or clinician in determining whether EMF represents a hazard to the human embryo, and the results are, in any event, inconsistent. On the other hand, the studies involving nonhuman mammalian organisms dealing with fetal growth, congenital malformations, embryonic loss, and neurobehavioral development were predominantly negative and are therefore not supportive of the hypothesis that low-frequency EMF exposures result in reproductive toxicity.