Because of concern about the potential adverse effects of consuming caffeinated beverages during pregnancy, pregnant women may choose to replace caffeinated with decaffeinated coffee. In a population-based study of 7,855 livebirths in California's San Joaquin Valley, we examined the relation of maternal decaffeinated and caffeinated coffee consumption during pregnancy to measures of fetal growth and gestational duration. Mothers answered a questionnaire in the hospital at the time of completing the birth certificate. Compared with women who drank neither decaffeinated nor caffeinated coffee, those who consumed only decaffeinated coffee showed no increased odds of small-for-gestational age birth, low birth weight, or preterm delivery, nor lowered mean birth weight or shortened mean gestational age. Women who consumed caffeinated coffee alone had an adjusted odds ratio of 1.3 [95% confidence limits (CL) = 1.0, 1.7] for preterm delivery, whereas those who consumed both caffeinated and decaffeinated coffee had an adjusted odds of 2.3 (95% CL = 1.3, 4.0). When caffeinated and decaffeinated coffee were considered as continuous variables, we found a reduction in adjusted mean birth weight of −3.0 gm per cup per week (95% CL = −5.9, −0.6) for caffeinated coffee and an increase of +0.4 gm per cup per week (95% CL = −3.7, 4–5) for decaffeinated coffee. (Epidemiology 1999;10:242–249)