The history of cocaine use is reviewed. Cocaine teratogenesis has only recently been studied, and initial human studies had serious methodological flaws. These flaws included ascertainment bias, publication bias (studies finding cocaine effects have been more likely to be presented or published), and overemphasis on the perinatal period. Comparison with alcohol teratogenesis shows that alcohol is a more potent teratogen, which, however, produces major and specific effects (fetal alcohol syndrome) in less than 10% of offspring with heavy alcohol exposure during pregnancy. Nonspecific minor congenital anomalies or fetal alcohol effects are seen in a larger number. Personal experience with two groups of children exposed to cocaine in utero is reviewed. Insurance patients gained weight, took vitamins, and generally, their children did well in spite of cocaine use. Indigent patients were usually unmarried and often "street people," probably used more cocaine, generally used other drugs as well, often did not gain weight during pregnancy, and were much more likely to have children with problems. Surveys show that most cocaine users also use alcohol, often simultaneously. Those who use both agents are more likely to have troubled backgrounds and antisocial behavior and to drop out of treatment programs than those who use only alcohol. Cocaethylene or ethylbenzoylecgonine is formed in the liver when cocaine and alcohol are simultaneously ingested. It is a potent stimulant and dopamine uptake blocker that is more toxic to myocardial cells than is cocaine. Good nutrition is now known to be very important in preventing congenital anomalies and fetal death. A multihit model of neurologic handicap, which stresses the importance of a good postnatal environment, is briefly outlined.(ABSTRACT TRUNCATED AT 250 WORDS)