The in vitro activities of ciprofloxacin, levofloxacin, gatifloxacin, and moxifloxacin against a large collection of clinical isolates of Streptococcus pneumoniae (n = 4,650) obtained over a 5-year period, 1994-1995 through 1999-2000, were assessed as part of a longitudinal multicenter U.S. surveillance study of antimicrobial resistance. Three sampling periods were used during this investigation, the winter seasons of 1994-1995, 1997-1998, and 1999-2000; and 1,523, 1,596 and 1,531 isolates were collected during these three periods, respectively. The overall rank order of activity of the four fluoroquinolones examined in this study was moxifloxacin > gatifloxacin > levofloxacin = ciprofloxacin, in which moxifloxacin (MIC at which 90% of isolates are inhibited [MIC90], 0.25 μg/ml; modal MIC, 0.12 μg/ml) was twofold more active than gatifloxacin (MIC90, 0.5 μg/ml; modal MIC, 0.25 μg/ml), which in turn was fourfold more active than either levofloxacin (MIC90, 1 μg/ml; modal MIC, 1 μg/ml) or ciprofloxacin (MIC90, 2 μg/ml; modal MIC, 1 μg/ml). Changes in the in vitro activities of fluoroquinolones against S. pneumoniae strains in the United States over the 5-year period of the survey were assessed by comparing the MIC frequency distributions of the study drugs against the isolates obtained during the three sampling periods encompassing this investigation. These comparisons revealed no evidence of changes in the in vitro activities of the fluoroquinolones. In addition, the percentages of isolates in the three sampling periods for which MICs were above the resistance breakpoints were compared. Low percentages of resistant strains were detected, and there was no evidence of resistance rate changes over time. For example, by use of a ciprofloxacin MIC of ≥4 μg/ml to define resistance, the proportions of isolates from the three sampling periods for which MICs were at or above this breakpoint were 1.2, 1.6, and 1.4%, respectively. A total of 164 unique isolates (n = 58 from 1994-1995, 65 from 1997-1998, and 42 from 1999-2000) were examined for evidence of mutations in the quinolone resistance-determining regions (QRDRs) of the parC and the gyrA genes. Forty-nine isolates harbored at least one mutation in the QRDRs of one or both genes (1994-1995, n = 15; 1997-1998, n = 19; 1999-2000, n = 15). Among the 4,650 isolates of S. pneumoniae examined in the study, we estimated that 0.3% had mutations in both the parC and gyrA loci. The majority of mutations (67.3% of the mutations in 49 isolates with mutations) were amino acid substitutions in the parC locus only. Four isolates had a mutation in the gyrA locus only, and 12 isolates had mutations in both genes (8.2 and 24.5% of isolates with mutations, respectively). There was no significant difference in the number of isolates with parC and/or gyrA mutations detected during each study period. Finally, because of the magnitude of the study, we had reasonably large numbers of pneumococcal isolates with genotypically defined mechanisms of fluoroquinolone resistance and were thus able to determine the effects of specific resistance mutations on the activities of different fluoroquinolones. In general, isolates with mutations in parC only were resistant to ciprofloxacin but remained susceptible to levofloxacin, gatifloxacin, and moxifloxacin, whereas isolates with mutations in gyrA only and isolates with mutations in both parC and gyrA were resistant to all four fluoroquinolones tested.