Screening for bacteriuria by culture of voided midstream urine was done in 6803 puerperal women; significant growth was found in 8.1%. The urine was recollected by suprapubic aspiration and bacteriuria was confirmed in 52%, corresponding to an incidence of bladder bacteriuria of 3.7%. A history of past urinary tract infection, bacteriuria in pregnancy, operative delivery, epidural anesthesia, and bladder catheterization increased the risk of postpartum urinary tract infection. Only 21% of the women complained of dysuria; this symptom occurred significantly more often after operative delivery and in patients with previous urinary tract infection. Two hundred fifty-one women with bladder bacteriuria were subjected to different treatments by randomized allocation: 153 patients with amoxicillin-susceptible bacterias were selected for amoxicillin treatment of 1, 3, and 10 days' duration. The cure rates were 84%, 94%, and 98%, respectively; the single-dose therapy was significantly less effective than 10 days' treatment (p less than 0.05). Forty-six women with amoxicillin-resistant bacterial infections received cephalexin or nitrofurantoin therapy of 7 days' duration; the cure rate was 91%. Fifty-two women served as control subjects and received no treatment. Ten weeks later 27% still had persistent bacteriuria in their suprapubic aspiration control specimens. All therapeutic regimens except the single-dose method showed a cure rate that was significantly higher than the spontaneous cure rate (p less than 0.05). Multiparity seemed to be a predisposing factor for persistence of bacteriuria. The study indicates that puerperal patients with positive midstream urine specimens should not be automatically treated, but more thoroughly examined. In cases of confirmed bladder bacteriuria, treatment should be recommended; 3 days' therapy appears to be sufficient.