Whole blood prothrombin times were determined serially in five groups of randomly selected well newborn "term" infants, none of whom had clinical evidence of bleeding. Prothrombin values in infants given menadione sodium bisulfite (vitamin K) parenterally shortly after birth were stable. The remainder of the infants who were not given vitamin K were divided into subgroups. Twenty-four hours after the first feeding, prothrombin activity of infants fed cow's milk did not differ significantly from that of infants given vitamin K at birth. In contrast, prothrombin times 24 hours after the first feeding of breast fed infants, of infants given bottled human milk, or of infants receiving only sterile water, were all significantly longer than that of infants given vitamin K parenterally. The responses of these newly born infants to the various feeding regimens appear to represent an in vivo assay for dietary vitamin K. The data document hypoprothrombinemia in the majority of infants who do not receive vitamin K parenterally and emphasize the importance of the administration of a small prophylactic dose of the vitamin to all newly born infants.