In an attempt to improve the success of a previous study of the prevention of neonatal hyperbilirubinemia employing phenobarbital in a single 30-mg nightly dose from 32 weeks' gestation until delivery, 246 patients were tested against 270 control subjects. A 39% reduction in significant hyperbilirubinemia was realized but this fell far short of previous findings with a larger dose. Premature infants received no apparent protection against jaundice. Mature infants had no severe jaundice (more than 16 mg/100 ml) and the resolution of hyperbilirubinemia was more rapid. No apparent difference in drowsiness or daytime sedation from previous experience was found. Female newborns responded much better (3:1) to prophylaxis. These disappointing results favor a return to the 1-g daily divided dose.