Different types of dietary fiber are fermented to various extents in vitro, but little is known about the effects of fiber on breath hydrogen and methane levels in vivo. Therefore, we studied the effects on breath hydrogen and methane of 15 g of guar, pectin, psyllium, soy polysaccharide, or cellulose in eight healthy subjects over a 12-h period. None of the fibers had a significant effect on breath hydrogen or methane concentrations, compared with the control (fasting). The four methane producers had lower breath hydrogen levels than the nonproducers 1 h after 15 g of lactulose (3 +/- 1 vs. 42 +/- 9, p less than 0.005) and 5-12 h after the different fibers (3.3 vs. 4.8 ppm; pooled SEM = 0.8; p less than 0.025). When the methane responses of the methane producers were expressed as increments relative to the control, there were small differences between treatments, with guar producing a larger response, 8.2 +/- 3.3 ppm, than cellulose, -2.9 +/- 2.3 ppm (p less than 0.05). The incremental methane responses of the different fibers in vivo were related to the previously reported production of propionic acid (r = 0.94, n = 5, p less than 0.02) and methane (r = 0.93, n = 4, NS) from in vitro fermentation of the same fibers. We conclude that methane producers have lower breath hydrogen levels than nonproducers. Purified fermentable and nonfermentable dietary fibers have no effect on breath hydrogen levels over 12 h in subjects previously consuming a normal diet. However, fermentable fibers may produce small increases in breath methane in methane-producing subjects.