Competitive interactions often play an important role in local community structure, and particularly so in ant communities. We test predictions derived from a competition model proposed for Scandinavian ant communities by comparing ant nest densities inside and outside fifty randomly-selected territories of competitively dominant wood ants within 30 km of Umeå, northern Sweden. As predicted by the model, competitively-intermediate encounter species, as well as other territorial species, showed complementary occurrences with both studied wood ants, Formica aquilonia Yarrow and F. lugubris Zett.. In contrast, complementary abundances, i.e. lower density of competitively-inferior submissive species in the presence of wood ants, as predicted by the model, was not supported for all submissive species. Of the two studied wood ant species, the nest density of submissive species was negatively correlated only with F. aquilonia. Submissives as a group, as well as Myrmica-species, showed complementary abundances with F. aquilonia, but one Leptothorax and two Serviformica species had higher nest densities in the presence of this competitively-dominant species. We propose that, for Leptothorax, these deviations from the model predictions may be because of limited niche overlap with dominant wood ants, a small worker force and a timid behaviour which does not elicit aggression in wood ants. For the two Serviformica species, we propose a combination of protection against social parasites, inter-specific social control performed by F. aquilonia, and dominance relationships between competitively-inferior submissive species as reasons for the higher nest density inside F. aquilonia territories. Monogyny, and thereby smaller nests, lower worker force, smaller territory (as shown by this study) in F. lugubris, as compared to the polygyny in F. aquilonia, may also help explain the differences in their effects on subordinate species. Our analyses indicate that the linear competition hierarchy model proposed for Scandinavian ants accurately predicts the outcome of interaction between and community composition for dominant territorial and encounter species, but that it needs refinement with respect to the relationship between territorial and submissive species and the resulting community composition. Further studies are needed, especially addressing the complex relationships between these latter groups, and the effects of different competitively-dominant wood ants, to determine the mechanisms determining the outcome of these relationships and to more accurately predict community composition.