We measured the serum bilirubin concentrations in 2,416 consecutive infants admitted to our well baby nursery. The maximal serum bilirubin concentration exceeded 12.9 mg/dL (221 mumol/L) in 147 infants (6.1%), and these infants were compared with 147 randomly selected control infants with maximal serum bilirubin levels less than or equal to 12.9 mg/dL. A serum bilirubin concentration greater than 12.9 mg/dL was associated strongly with breast-feeding (P = .0000) and percentage of weight loss after birth (P = .0001), as well as with maternal diabetes, oriental race, decreased gestational age, male sex, bruising, and induction of labor with oxytocin. Risk ratios and the risk of jaundice were calculated for hypothetical infants in the presence and absence of these variables. These calculations show that, in certain infants, "nonphysiologic" jaundice is likely to develop and its presence in such infants might not require laboratory investigations. In others, a modest degree of hyperbilirubinemia could be cause for concern. An awareness of these factors and their potential contribution to serum bilirubin levels permits a more rational approach to the action levels used for the investigation of jaundice in the newborn. We need a new definition of physiologic jaundice.