1. There has been considerable debate on the importance of competition in ecological communities, but its importance in structuring ant assemblages has often been uncritically accepted. Here, we briefly review field experiments examining competition in ant assemblages and use a removal experiment to test the effect of the classical territorial dominant ant, Formica aquilonia. Ants of this species group are thought to structure communities through a dominance hierarchy. 2. First, we used pitfall traps to compare the abundance of other ants in replicated sites with low and high densities of F. aquilonia. We found differences in community composition, in particular, Camponotus herculeanus was more common in low-density sites, in accordance with predictions. Differences in ant assemblages were not owing to differences in measured habitat variables. 3. We removed F. aquilonia from a set of high-density sites, using physical and chemical methods, and repeated these procedures at procedural control sites. One year after removal, abundances of F. aquilonia at removal sites were similar to those at low-density sites. However, the composition of other species did not change in response to F. aquilonia removal. Replication rates were identical in the mensurative and experimental components of this study, so this is unlikely to be owing to the analysis being insufficiently powerful. 4. We suggest three possibilities for the lack of difference. First, the study may have been too short term or small scale to detect differences. However, previous studies have shown effects on smaller spatial- and temporal-scales. Second, priority effects may be important in the successful colonisation by F. aquilonia. Thirdly, boreal ant assemblages may be too depauperate for redundancy in ecological roles and for competition to play an important structuring role. 5. We thus recommend that long-term large-scale experiments be considered essential if we are to distinguish between competing hypotheses in community ecology.