Using a technique previously described a study was done on the in vitro effect of 10 antimicrobials on the fermentative capacity of human fecal bacteria. The antimicrobials were: 5 antibiotics, oxytetracycline, chloramphenicol, neomycin, paramomycin and lincomycin; and 5 chemotherapeutic agents, bismuth subnitrate, cotrimoxazole, furazolidone, norfloxacin and metronidazole. The study was done with 162 stool samples, 27 to investigate the effect of bismuth subnitrate, and 135 to investigate the effect of the remaining 9 antimicrobials (15 for each one of them). The quantities of antimicrobials used were calculated to reproduce in vitro minimal inhibitory concentrations against fecal bacteria and/or concentrations reached by the respective antimicrobials or their active metabolites in the colonic contents or the stools after regular oral doses. For greater simplicity and clarity, whenever two concentrations of an antimicrobial were reproduced they were denominated low concentration and high concentration as it corresponded. Oxytetracycline, neomycin, paramomycin, cotrimoxazole and norfloxacin did not significantly affect fermentation by fecal bacteria. Chloramphenicol, lincomycin, bismuth subnitrate and furazolidone did not seem to affect significantly fermentation when the variations, increases and decreases produced in it, where evaluated as a whole. But, when increases and decreases were evaluated separately it was found that: chloramphenicol, at low concentration, increase fermentation significantly with 11 from 15 stool samples and diminished it although not significantly with the remaining 4 samples, and at high concentration, increased fermentation significantly with 8 from the 15 samples and diminished it although not significantly with the remaining 7 samples; lincomycin, at low concentrations, increased fermentation with 6 from 15 stool samples and diminished it with the remaining 9 samples although not significantly, and at high concentration, increased it significantly with 9 from the 15 samples and diminished it although not significantly with the remaining 6 samples; bismuth subnitrate, at low concentration, increased fermentation although not significantly with 5 from 27 stool samples and diminished it significantly with the remaining 22 samples, and at high concentration, increased it although not significantly with 2 from 10 stool samples and diminished it significantly with the remaining 8 samples; and furazolidone, at a unique concentration, increased fermentation significantly with 9 from 15 stool samples and diminished it although not significantly with the remaining 6 samples.(ABSTRACT TRUNCATED AT 400 WORDS)