In the ideal situation, the evaluation for sepsis in the young infant should include collection of multiple blood cultures before the institution of antibiotics. Unfortunately, in some infants, it may not be possible to obtain more than a single blood culture at the time of initial evaluation. If this single culture ultimately grows coagulase-negative staphylococci and the infant has been treated with antimicrobial therapy in the interim, it is often difficult to determine whether the positive culture represents true infection or contamination. Our data suggest that peripheral blood cultures yielding high colony counts most likely represent infection. Furthermore, in this high-risk patient population, low colony-count growth should not be ignored as contamination, particularly if there are significant clinical findings or if the infant has a central catheter or hematologic abnormality. Future studies should examine these important issues.