The Isolator 1.5 microbial system (ISO 1.5) (Wampole Laboratories, Cranbury, N.J.) was compared with the BACTEC NR660 aerobic NR6A bottle (NR6A) (Becton Dickinson Diagnostic Instrument Systems, Sparks, Md.) for the detection of fungemia in hospitalized pediatric patients. For 4,825 paired blood cultures evaluated retrospectively from April 1992 to December 1994, at least one blood culture system was positive for 89 clinically important fungal isolates involved in 36 episodes of fungemia in 34 patients. Sixty isolates (44 Candida albicans, 12 Candida parapsilosis, and 4 Candida tropicalis isolates) were recovered from both systems, 13 were recovered from NR6A bottles only (10 C. albicans, 1 C. parapsilosis, and 2 Cryptococcus neoformans isolates), and 16 were recovered from ISO 1.5 tubes only (8 C. albicans and 5 C. parapsilosis isolates and 1 C. tropicalis, 1 Candida lusitaniae, and 1 Rhodotorula glutinis isolate) (P > 0.05). For the 60 Candida isolates from both systems, the mean time to detection was the same in each system. Thirty-seven isolates were detected by both systems on the same day, 9 isolates were detected earlier by NR6A, and 14 isolates were detected earlier by ISO 1.5 (P > 0.05). Of the 36 clinically important episodes of fungemia, 22 were detected by both systems (13 C. albicans isolates and 9 other Candida isolates), 4 were detected by NR6A only (3 C. albicans isolates and 1 C. neoformans isolate), and 10 were detected by ISO 1.5 only (3 C. albicans isolates, 6 other Candida isolates, and 1 R. glutinis isolate) (P > 0.05). Of the 22 episodes in which cultures from both systems were positive at some point during the episode, 12 were detected on the same day by both systems, 8 were detected earlier by NR6A, and 2 were detected earlier by ISO 1.5. Thus, for our pediatric population, NR6A is comparable to ISO 1.5 in both yield and time to detection of yeasts in fungemic patients.