In recent spatial models describing interactions among a myrmecophilous butterfly Maculinea rebeli, a gentian Gentiana cruciata and two competing species of Myrmica ant, we predicted that apparent competition should exist between gentians (the food of young M. Rebeli caterpillars) and Myrmica schencki, which supports M. Rebeli in its final instar. Here we extend and quantify model predictions about the nature of this phenomenon, and relate them to ecological theory. We predict that: (i) Within sites supporting the butterfly, fewer M. Schencki colonies occur in sub-areas containing gentians than in identical habitat lacking this plant. (ii) Where G. Cruciata and M. Schencki do co-exist, the ant colonies will be less than half the size of those living >1.5 m from gentians; (iii) The turnover of M. Schencki colonies will be much greater than that of other Myrmica species in nest sites situated within 1.5 m of a gentian. All three predictions were supported in the field on 3-6 sites in two mountain ranges, although the exact strength of the apparent competition differed from some model predictions. Field data were also consistent with predictions about apparent mutualisms between gentians and other ants. We suggest that apparent competition is likely to arise in any system in which a specialist enemy feeds sequentially on two or more species during its life-cycle, as occurs in many true parasite-host interactions. We also predict that mor: complex patterns involving other Myrmica species and G. Cruciata occur in our system, with apparent competition existing between them in some sub-areas of a site being balanced by apparent mutualism between them in other sub-areas.