Prolonged second stage of labor is associated with increased perinatal mortality. Factors influencing second-stage duration are poorly understood. This study was undertaken to characterize those factors. A population of 473 nulliparous women and 491 multiparous women with spontaneous vaginal deliveries were analyzed extensively with history, physical examination (including clinical pelvimetry), labor and delivery data, and neonatal measurements. On the basis of stepwise multiple linear regression, epidural analgesia (p less than 0.0001), active-phase duration (p less than 0.0001), parity (p less than 0.0001), height (p less than 0.0004), birth weight (p less than 0.0003), and station at complete dilatation (p less than 0.027) predicted second-stage duration. The sum of their effect, however, accounted for less than 25% of the variability in second-stage length (total R2 = 0.233), leaving 75% of the variance unexplained.