Parallel artificial membrane permeability assay (PAMPA) was used to measure the effective permeability, P(e), as a function of pH from 4 to 10, of 17 fluoroquinolones, including three congeneric series with systematically varied alkyl chain length at the 4'N-position of the piperazine residue. The permeability values spanned over three orders of magnitude. The intrinsic permeability, P(o), and the membrane permeability, P(m), were determined from the pH dependence of the effective permeability. The pK(a) values were determined potentiometrically. The PAMPA method employed stirring, adjusted such that the unstirred water layer (UWL) thickness matched the 30-100 microm range estimated to be in the human small intestine. The intrinsic permeability coefficients (10(-6)cm/s), representing the permeability of the uncharged form of the drug, are for 4'N-R-norfloxacin: 0.7 (R=H), 49 (Me), 132 (n-Pr), 365 (n-Bu); 4'N-R-ciprofloxacin: 2.7 (H), 37 (Me), 137 (n-Pr), 302 (n-Bu); 4'N-R-3'-methylciprofloxacin: 3.8 (H), 20 (Me), 51 (Et), 160 (n-Pr), 418 (n-Bu). Increasing the alkyl chain length in the congeneric series resulted in increased permeability, averaging about 0.34 log units per methylene group, except that of the first (H-to-Me), which was about 1.2 log units. These results were compared to Caco-2 and rat in situ permeability measurements. The in situ closed loop technique used for obtaining permeability values in rat showed a water layer thickness effect quite consistent with in vivo expectations. The rat-PAMPA correlation (r2=0.87) was better than that of rat-Caco-2 (r2=0.63). Caco-2-PAMPA correlation indicated r2=0.66. The latter correlation improved significantly (r2=0.82) when the Caco-2 data were corrected for the UWL effect.