Ant–acacia mutualisms are conspicuous biotic associations in Savannah and neotropical ecosystems; however, the effects of tree size and forest structure on ant behaviour and tree traits are rarely examined. We tested two hypotheses related to these effects: (1) ant responses to disturbance are influenced by tree size and forest basal area; and (2) tree traits important to ants are predictable by tree size and forest basal area. We investigated these hypotheses in a dry tropical forest (Ometepe Island, Nicaragua) with the myrmecophytic Collins acacia (Vachellia collinsii Saff.) and the ant Pseudomyrmex spinicola (Emery 1890). We measured trees from three size classes and three basal area classes and quantified resources that are important for ants, including food resources (nectaries and Beltian bodies) and domiciles (thorns), as well as a measure of potential tree reproductive fitness (seedpods). We also evaluated ant responses to experimental disturbances. Three important findings emerged: (1) on average, 1140–1173% more ants responded to experimental disturbances of large trees than small- or intermediate-sized trees, respectively; (2) forest basal area did not affect ant responses to disturbance; and (3) neither tree size nor forest basal area was correlated with branch-level mean numbers of nectaries, food bodies or thorns. Our studies support the hypothesis that tree size is an important factor regarding ant behavioural responses to disturbance, but not forest basal area. Our work suggests that future studies of ant behaviour on myrmecophytes should consider tree size.