Marijuana is among the most widely used psychoactive substances in the Western world. In the United States, about 255 of Americans 18 to 25 years of age use it to some degree (U.S. Department of Health, Education, and Welfare 1980). Considerable marijuana use also appears to be occurring among pregnant women (Sokol et al. 1980; Hingson et al. 1982; Linn et al. 1983; Fried et al. 1984; Gibson et al. 1983). It is only in the last few years, however, that critical attention has been focused on the possibility that these substances can cause birth defects and postnatal behavioral aberrations, although delta-9-tetrahydrocannabinol (6 9-THC), the principal psychoactive ingredient in marijuana, is known to cross the placenta (Abel 1983). Before examining the data relevant to this issue, the first part of this review will examine some of the general methodological considerations which should be kept in mind in evaluating research in this area. CLINICAL AND EPIDEMIOLOGICAL STUDIES Since experimental administration of drugs to pregnant women is unethical, evaluation of potential teratogens is limited to clinical observations or epidemiological investigations. Although clinical reports can be of considerable importance in alerting physicians and health care providers to possible agents causing abnormal development, they are often difficult to evaluate. For example, two early clinical reports of malformations in children born to marijuana users (Hecht et al. 1968; Carakushansky et al. 1969) were inconclusive since the mothers of these children were users of other drugs as well. When clinical reports are followed by epidemiological studies involving larger numbers of patients, a better appreciation of incidence and causation is possible. Such epidemiological studies can be divided into two types, retrospective and prospective, each of which has its own strengths and shortcomings.