Between January 1, 1982 and December 31, 1992 we evaluated 200 children of mothers with pregestational or gestational diabetes mellitus through the University of South Florida Genetics/Dysmorphology Clinics. They were a portion of the 22,100 families seen during that period. Pregnant women with diabetes mellitus were not part of this study. One hundred and fifty-two of the 200 were offspring of mothers with gestational diabetes (classes A1 and A2). Class B1 was not encountered as a subclass of gestational diabetes in this series. Twenty-four of the 152 did not have anomalies. Forty-one of the 152 had another primary diagnosis to account for their malformations. Eighty-seven of the 152 had a constellation of anomalies or solitary structural defects as seen in diabetic embryopathy. Chromosomal, monogenic, and other teratogenic causes were excluded. The observed phenotypes matched those seen in offspring of mothers with diabetes mellitus classes B2 to T. They also corroborated the animal studies, indicating that the embryopathy of gestational diabetes has a pathogenesis similar to that in classes B2 to T, and recent epidemiological studies showing a statistically significant increase of anomalies as in diabetic embryopathy in the offspring of gestational diabetes mothers. As per established obstetric practice the testing for gestational diabetes was after gestation 16 weeks. Thus, it was impossible to prove that the anomalies of the 87 propositi were due to gestational diabetes. However, the anomalies occurred during organogenesis as in the other diabetic classes, and inferred that gestational diabetes is a human teratogen. If so, common "idiopathic" malformations may be actually caused by undiagnosed maternal gestational diabetes.