The standard procedure for culturing blood is to inoculate 5 ml of blood into 50 ml of broth, resulting in a blood-broth ratio of 1:10. In infants, where only 0.5 to 1.0 ml of blood is available for culture, blood:broth ratios may be as dilute as 1:100. The purpose of this study was to determine whether a blood-broth ratio of 1:100 was as sensitive as a ratio of 1:10 for detecting low concentrations of bacteria in blood. Blood (0.5 ml) seeded with 2 to 30 organisms/ml of one of four common pediatric pathogens (Escherichia coli, Streptococcus pneumoniae, Group B Streptococcus or Haemophilus influenzae type b) was inoculated into culture broth so as to provide blood-broth ratios of 1:10, 1:30 or 1:100. There were no differences in the ability to recover organisms or in the time-to-positivity of cultures with different blood-broth ratios. Cultures were positive in 24 of 40 (60%) bottles at a ratio of 1:10, 28 of 40 (70%) at 1:30 and 30 of 40 (75%) at 1:100. Common pediatric pathogens can be recovered without delay from volumes of blood as small as 0.5 ml cultured at blood-broth ratios up to 1:100.