In a case-control study of 2030 malformed [humans] infants, 6 selected birth defects were evaluated in relation to maternal ingestion during pregnancy of caffeine from tea, coffee and cola: 380 infants with inguinal hernia, 299 with cleft lip with or without cleft palate, 277 with cardiac defects, 194 with pyloric stenosis, 120 with isolated cleft palate and 101 with neural tube fusion defects were compared to 712 other malformed infants who served as controls. None of the point estimates of relative risk was significantly greater than unity. For ingestion of any caffeine relative to no ingestion, the upper 95% confidence bounds were of the order of .ltoreq. 3; for consumption of the equivalent of .gtoreq. 4 cups of coffee per day relative to < 2 cups, the upper bounds were .ltoreq. 2. Potential confounding factors did not explain the findings. Apparently, caffeine is not a major teratogen with regard to the 6 defects evaluated.