Thirty-two human isolates of enterococci were tested for antibiotic synergy by using penicillin and one of six aminoglycosides. Three methods were used: synergy screen, microdilution checkerboard, and time-kill curves. The synergy screen accurately predicted synergy for gentamicin-penicillin combinations, and this synergy was later confirmed by time-kill curves. The microdilution checkerboard method suffered from inherent variation, and agreement with time-kill curves ranged from 92% (twofold reduction in minimum inhibitory concentration) to 4.2% (fourfold reduction in minimum inhibitory concentration). We suggest that enterococci be screened for synergy (i.e., presence or absence of high-level resistance) by using the criterion of growth or no growth in the presence of 2,000 microgram of an aminoglycoside per ml. The microdilution checkerboard test for synergy is not recommended.