Eighteen highly-trained runners ran two half marathons in mild environmental conditions, 3 wk apart, consuming either 426 +/- 227 mL of a flavored placebo drink (PLACEBO) or an equivalent volume of water (386 +/- 185 mL) and a commercial gel (GEL) supplying 1.1 +/- 0.2 g/kg body mass (BM) carbohydrate (CHO). Voluntary consumption of this fluid was associated with a mean BM change of approximately 2.4%. Runners performed better in their second race by 0.9% or 40 s (P = 0.03). Three runners complained of gastrointestinal discomfort in GEL trial, which produced a clear impairment of half-marathon performance by 2.4% or 105 s (P=0.03). The effect of GEL on performance was trivial: time was improvedby 0.3% or 14 s compared with PLACEBO (P = 0.52). Consuming the gel was associated with a 2.4% slower time through the 2 x 200 m feed zone; adding a trivial approximately 2 s to race time. Although benefits to half marathon performance were not detected, the theoretical improvement during 1-h exercise with CHO intake merits further investigation.