The aim of this study was to test the hypothesis that Grevy's zebra (Equus grevyi Oustalet) compete for critical resources with pastoral people and their domestic livestock. The study was conducted in three areas to approximate a manipulative experiment, with pastoralists and their livestock being excluded from one of the areas. Forage was seasonally limited in the areas used by pastoralists. Although no single factor determined the space use of Grevy's zebra, the distribution of food - particularly in the pastoralist areas - and predators were important. Use of space by Grevy's zebra was constrained by their dependence on water and by the presence of the people in the pastoralist areas. Zebra dispersed from areas when food became limiting. Different sex classes exhibited different times of dispersal that may have been related to different thresholds for dispersal. In the area where there were no pastoral people, temporal patterns of water use by Grevy's zebra were driven by avoidance of predation. In contrast, Grevy's zebra in the pastoralist areas drank at night to avoid the monopolisation of water by people and their domestic livestock by day. By drinking at night, the zebra, and particularly their foals, were probably more susceptible to predation. Breeding patterns in Grevy's zebra were explained by condition dependent oestrus: oestrus and conception were highly dependent on stochastic patterns of climatic variation. Juvenile survival was related to the amount mothers move, and was lowest in areas used by pastoralists. Grevy's zebra were found to compete for critical resources with pastoralists and their livestock. This highlighted the importance of protected areas to sustain recruitment into the population. Release from competition is, however, necessary to further improve the likelihood of persistence of Grevy's zebra in their natural range. To achieve this, pastoralist stocking rates must be reduced.