The aim of this study was to examine hydrogen (H2) production with the hydrogen breath test (HBT) after ingesting primarily digestible carbohydrate (CHO) during 3 h of 75% maximal oxygen consumption exercise. This was done to indicate CHO overflow in the colon which may occur when gastric emptying, intestinal transit and CHO absorption are not matched and CHO accumulates in the colon where it is subject to bacterial degradation. Further, this study was designed to assess breath H2 production as a function of the type of CHO ingested and the type of exercise. A group of 32 male triathletes performed three exercise trials at 1-week intervals with either a semisolid (S) intake, an equal energy fluid intake (F) or a fluid placebo (P). Each trial consisted of cycling (sessions 1 and 3) and running (sessions 2 and 4). The mixed-expired H2 concentrations in the resting and "recovery" periods (5 min after each session) did not change significantly in time and did not differ among intakes. There were also no significant differences in H2 concentrations between resting and "recovery" conditions. During exercise, H2 concentrations decreased three to six-fold in comparison to resting and recovery levels and differed among intakes (ANOVA; P < 0.05). The H2 concentrations were almost continuously lower with P than with F and S. The H2 concentrations were significantly higher during running than during cycling. During exercise, we found that CHO overflow could be compared among intakes and between exercise types by using the HBT, provided the influence of other factors on H2 excretion--ventilation and intestinal blood flow--was similar for each condition.