The optimal atmosphere of incubation for the isolation of group A beta-hemolytic streptococci from throat cultures has been the subject of considerable debate. To resolve this issue, we examined 5992 consecutive throat cultures performed at a private pediatric office in children with acute upper respiratory tract infections. All throat cultures were streaked onto duplicate blood agar plates, one of which was then incubated anaerobically and the other aerobically. beta-Hemolytic streptococci were isolated in cultures from 1885 (31.5%) of the patients; 1479 (24.7%) were identified as group A, and 406 (6.8%) were identified as non-group A. Group A streptococci were recovered significantly more often from the plates incubated anaerobically than from those incubated aerobically (1467 vs. 1054; anaerobic only, 425; aerobic only, 12; P less than 0.01). Non-group A streptococci were also recovered significantly more often from the plates incubated anaerobically than from the plates incubated aerobically (397 vs. 170; anaerobic only, 236; aerobic only, 9; P less than 0.01). Anaerobic incubation maximizes the yield of the throat culture. The additional cost and effort of anaerobic incubation are small, and would appear to be justified by the increased isolation rate of group A beta-hemolytic streptococci. The significance of the increased isolation rate of non-group A beta-hemolytic streptococci with anaerobic incubation needs to be investigated further.