A randomized controlled investigation was undertaken to evaluate the role of maternal oral antibiotic therapy in decreasing the incidence of maternal and neonatal colonization at term with group B β-hemolytic streptococcus (GB-BHS). Data were collected to determine the optimum transfer media and the colonization rate in the study population. At delivery 1441 maternal-infant pairs were evaluated. One hundred sixty-eight women (11.6%) and 55 infants (3.8%) were colonized. Forty-four women colonized with GB-BHS at 38 weeks' gestation were randomly assigned to a treatment (500 mg potassium penicillin or erythromycin ethylsuccinate q.i.d.) or a control group to determine the value of antepartum oral antibiotic therapy in preventing infant colonization. There was a significant reduction in maternal (P= 0.008) and infant (P= 0.004) colonization rates in the treatment group. There were no observed complications of antibiotic therapy in mothers or infants. This study suggests that routine cultures for GB-BHS should be done at 38 weeks' gestation. Mothers colonized at this time may be considered candidates for prophylactic antibiotic treatment.