Breath hydrogen and methane are specific end products of colonic fermentation, a process which may play a protective role against colon cancer. To assess the possibility of using these markers in epidemiological studies, we characterized the intra- and intersubject variability of breath hydrogen and methane excretion over 15 consecutive days among 32 men and women of various ethnic backgrounds (16 Asians, 8 Caucasians, 8 Hawaiians). Participants were asked to collect four end-expiratory samples each day, which we had shown previously would optimally characterize daily hydrogen excretion. There was substantial within-subject variation in breath hydrogen over the study, although breath methane levels were more constant over time. We found that about 4 days of measurement of breath hydrogen and 1 day of measurement for breath methane are required to correctly characterize individuals according to their long-term excretion of these gases. This was true for Asians and non-Asians. Although breath methane appears to be more practical to measure, it is a less sensitive marker of colonic fermentation than breath hydrogen. Whereas all subjects excreted hydrogen, only 28% of the subjects excreted methane, and methane excretor status of a few participants varied during the study. Because the breath test is noninvasive and reliable, we tested the multiple day collection protocol among colon cancer patients and controls and found it to be well accepted. We conclude that it is practical to measure breath hydrogen and methane in large epidemiological studies conducted at the individual level. The potential use for these markers is discussed.