A classic example of extreme morphological adaptation to the environment is the neck of the giraffe (Giraffa camelopardalis), a trait that most biologists since Darwin have attributed to competition with other mammalian browsers. However, in searching for present-day evidence for the maintenance of the long neck, we find that during the dry season (when feeding competition should be most intense) giraffe generally feed from low shrubs, not tall trees; females spend over 50% of their time feeding with their necks horizontal; both sexes feed faster and most often with their necks bent; and other sympatric browsers show little foraging height partitioning. Each result suggests that long necks did not evolve specifically for feeding at higher levels. Isometric scaling of neck-to-leg ratios from the okapi Okapia johnstoni indicates that giraffe neck length has increased proportionately more than leg length-an unexpected and physiologically costly method of gaining height. We thus find little critical support for the Darwinian feeding competition idea. We suggest a novel alternative: increased neck length has a sexually selected origin. Males fight for dominance and access to females in a unique way: by clubbing opponents with well-armored heads on long necks. Injury and death during intrasexual combat is not uncommon, and larger-necked males are dominant and gain the greatest access to estrous females. Males' necks and skulls are not only larger and more armored than those of females' (which do not fight), but they also continue growing with age. Larger males also exhibit positive allometry, a prediction of sexually selected characters, investing relatively more in massive necks than smaller males. Despite being larger, males also incur higher predation costs than females. We conclude that sexual selection has been overlooked as a possible explanation for the giraffe's long neck, and on present evidence it provides a better explanation than one of natural selection via feeding competition.