Aims: Desert dominant plants commonly facilitate plant communities within their canopies. Although substantial research has examined the direct consequences of this effect, a mechanistic understanding of indirect effects mediated via beneficiary plants is still relatively limited. We tested the hypothesis that the net positive outcome of dominant plants on beneficiaries extends to a series of interactions including indirect competition or facilitation. To test this hypothesis, we aggregated two years of field surveys with a manipulative experiment. Location: Atiquipa, Southern Peru. Methods: We surveyed the understorey plant community of the dominant tree Caesalpinia spinosa and compared it to that of open microsites. Field manipulations included removal of plant neighbourhoods of two target annual species in the understorey of the dominant plant Caesalpinia spinosa and adjacent open microsites. In the surveys, we measured plant density, and in the removal experiment, plant height, fruit set, and biomass of the targets. Results: In the surveys, density of Fuertisimalva peruviana, a target species, was negatively dependent on understorey neighbours’ density. Neighbourhood removal did not affect fruit set or biomass of Fuertisimalva peruviana in tree understories or open microsites. Neighbourhood removal around Plantago limensis increased biomass, plant height and fruit set, and changed the effect of the dominant tree from positive to negative for biomass. Conclusions: Understorey plant neighbours can mediate the direct effects of dominant plants in deserts, but responses to these indirect effects can be species-specific. The facilitative effects of dominant plants represent key coexistence mechanisms because they reduce stress and generate extended series of interactions not necessarily present in the open.