Résumé
Water is an essential and often limiting resource that pervades all aspects of animal ecology. Yet, water economics are grossly understudied relative to foraging and predation, leaving ecologists ill‐equipped to predict how the intensifying disruption of hydrological regimes worldwide will impact communities. For savanna herbivores, reliance on surface water can increase exposure to predators and competitors, and thus strategies that reduce the need to drink are advantageous. Yet, the extent to which increasing dietary water intake while decreasing water loss enables animals to forego drinking remains unknown. We studied water budgets of sympatric African savanna antelopes that differ in size, bushbuck ( Tragelaphus sylvaticus , ~35 kg) and kudu ( T. strepsiceros , ~140 kg). We hypothesized that both species compensate for seasonally declining water availability by increasing consumption of moisture‐rich plants and reducing faecal water loss, and that these adjustments are sufficient for small‐bodied—but not large‐bodied—herbivores to avoid spending more time near permanent water sources as the dry season advances. We tested our predictions using temporally explicit data on antelope movements, diets, plant traits and drinking behaviour in Gorongosa National Park, Mozambique. Water content declined between the early and late dry seasons in roughly half of plant taxa consumed by antelope. Although both species reduced faecal water loss and shifted their diets towards relatively moisture‐rich plants as the dry season progressed, dietary water intake still declined. Contrary to expectation, kudu reduced selection for surface water in the late dry season without adjusting total time spent drinking, whereas bushbuck increased selection for surface water. We developed a generalizable approach for parsing the importance of dietary and surface water for large herbivores. Our results underscore that variation in surface‐water dependence is a key organizing force in herbivore communities, that simple allometric predictions about the behavioural and ecological consequences of this variation are unreliable. Understanding wildlife water economics is a research frontier that will be essential for predicting changes in species distribution and community composition as temperatures rise and droughts intensify.