The E-Test (AB Biodisk, Sweden) is an antibiotic gradient strip that is applied to an inoculated agar plate and results in an elliptical zone of inhibition that intercepts the graded strip, producing a quantitative (microgram per milliliter) result. Pneumococci (100) were used in this study, 38 penicillin-susceptible [minimal inhibitory concentrations (MICs), less than or equal to 0.12 micrograms/ml], 42 intermediately resistant (MICs, 0.12-1.0 micrograms/ml), and 20 resistant (MICs, greater than 1 microgram/ml). E-Test strips were evaluated on Mueller-Hinton agar plates with 5% sheep blood. Agar dilution MICs were determined by the National Committee for Clinical Laboratory Standards (NCCLS) method. Penicillin MICs for the E-Test tended to be slightly lower (one log2 dilution) than reference MICs due to the continuous scale from which E-Test MICs were read. All but two penicillin-susceptible isolates were correctly categorized by the E-Test method. Of the 62 penicillin-resistant strains, 59 had E-Test MICs of greater than or equal to 0.12 micrograms/ml, with 88% of these strains having E-Test MICs within one doubling dilution of the reference MICs. However, using the current NCCLS breakpoint MICs, many of the penicillin-resistant strains with reference MICs of 2 micrograms/ml were categorized as intermediate by the E-Test, with MICs of 0.38-1 microgram/ml. For chloramphenicol, erythromycin, and tetracycline, correlation of the two methods was excellent. E-Test chloramphenicol MICs provided clearer separation of susceptible and resistant strains than did the reference method. We conclude that the E-Test is a reliable method for determination of MICs of the antibiotics evaluated for pneumococci.