A substantial increase in blood cultures positive for a Serratia marcescens strain unusually sensitive to antibiotics was noted in two large hospitals within six months. Because the patients' illnesses seemed incompatible with Serratia bacteremia, contamination of blood cultures was suspected. Investigation suggested that pediatric-sized vacuum tubes containing ethylenediamine tetraacetic acid (EDTA) were the source of the organisms, and the epidemic strain of Serratia was recovered from 41 (35%) of the 116 tubes cultured. Mock trials showed that reflux from tube to syringe can occur while vacuum tubes are being filled. Because contaminated EDTA tubes were sometimes inoculated before blood culture bottles in these hospitals, cross-contamination occurred. Most evacuated specimen tubes are not guaranteed sterile by the manufacturer. False-positive blood cultures stemming from the use of nonsterile tubes can be eliminated by inoculating blood culture bottles before other specimen tubes. Because false-positive blood cultures may lead to unnecessary antibiotic therapy, health-care workers should guard against the potential hazard associated with use of these tubes.