Poor statistical correlation was obtained between tube dilution test results and disk test results by using standard procedures. Significant regressions were obtained although they were not linear. Different regressions were obtained with different bacterial species. It is suggested that for the aminoglycosides studied (gentamicin, kanamycin, tobramycin, and sisomicin) both disk and dilution tests are useful in separating resistant from sensitive organisms. However, poor relationships were obtained between the two test types among sensitive organisms.