The effects of variations in laboratory technique on the speed and sensitivity of isolation of Pseudomonas pseudomallei from blood were evaluated prospectively. Pseudomonas pseudomallei was isolated from 154 of 546 cultures from 325 patients with suspected or confirmed melioidosis. Subcultures after 12 to 24 and 36 to 48 hours of incubation were positive in 52.3% and 80.8% respectively. The yields from 20 ml (blood to broth ratio 1:4) and 50 ml (blood to broth ratio 1:10) brain heart infusion broth bottles were equivalent in patients not receiving treatment for melioidosis. During therapy, the 50 ml bottles grew Pseudomonas pseudomallei significantly faster than the 20 ml bottles (p less than 0.01), and gave a higher overall yield for cultures processed in antimicrobial removal devices (p less than 0.05). These devices themselves increased the speed of isolation of the organism from treated patients (p less than 0.01). In most cases, all bottles collected from a patient before treatment were positive, and a single 20 ml bottle had an estimated relative sensitivity of 85.7% (95% confidence interval 77.1-94.3%). Early subculture should be employed routinely for the laboratory diagnosis of septicaemic melioidosis. However, blood culture techniques do not need to be sophisticated. Culture of 5 ml blood in 20 ml broth is a simple and sensitive procedure suitable for regions where melioidosis is currently under-diagnosed.