One hundred thirty-four indigent patients at term who had premature rupture of membranes and a cervix unfavorable for induction of labor (80% effacement or less, 2 cm dilation or less) were randomized to compare expectant with intervention management. Women with any medical or obstetric condition warranting immediate intervention were excluded from the study. Patients treated expectantly were placed at bed rest and observed for labor or infection. Patients managed by intervention were given oxytocin if labor did not ensue within 12 hours of rupture of the membranes. Patients in the intervention protocol had longer labor (P less than .02) and a higher incidence of both cesarean delivery (P less than .05) and intraamniotic infection (P less than .05). There was only one case of proven neonatal sepsis, and this occurred in a patient managed by induction of labor. There was no statistically significant difference between groups in mean length of maternal hospitalization.