A TLC-direct bioautography (DB) assay using Bacillus subtilis as test bacteria was developed. Various factors affecting the microorganism's viability on the TLC plates were studied and verified for the flumequine standards. The Dhenasar's method called "direct sample determination" was used for TLC; the antibiotic samples were spotted on the TLC plates and subjected to bioautography without developing with a mobile phase. The best preincubation and incubation times of bacterial broth were found to be 1 h at 37 degrees C and 6 h at 37 degrees C. The optimal viscosity of broth was obtained by the addition of agarose to obtain a 0.05% solution in the Mueller-Hinton broth. The best incubation time of seeded TLC plates was 17 h at 37 degrees C. The plates were visualized by spraying with 0.2% aqueous 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazolium bromide solution and incubated again for 0.5 h at 37 degrees C. The method was validated by determination of linearity, interday and intraday precision, LOD, and LOQ. The calibration curves showed good linearity in the range 0.005-0.5 microg (0.5-50.0 microg/mL). The regression coefficients were 0.9970 and 0.9955 for intraday and interday plots, respectively. The LOD of flumequine equalled 0.5 microg/mL, i.e., 5 ng of the antibiotic in the spot. The sensitivity of the developed TLC-DB test was compared with that of the two most commonly used standard antimicrobial susceptibility assays: agar disc diffusion and agar cylinder diffusion. The obtained minimum inhibitory concentration values clearly indicate much higher sensitivity of the TLC-DB method compared to the standard antimicrobial susceptibility assays.