In rural Bangladesh, the continuation rate of oral contraceptives (OCs) is the lowest among all methods except for condoms, and the pregnancy rate for OC users is comparable to that for users of traditional methods, which lies halfway between that for foam and condom users (high) and Depo-Provera and IUD users (low). To determine why postpartum amenorrheic OC acceptors have shorter pregnancy intervals than postamenorrheic OC acceptors who were matched on time of acceptance as well as on age and parity, subintervals of the pregnancy interval are examined. Results reveal significantly shorter duration of use by amenorrheic acceptors. This short duration of use overrides their significantly longer duration to pregnancy since discontinuation, to give overall shorter duration to pregnancy since last birth.