In this work we studied the effect of size-dependent predation and cannibalism on the spatial distribution of the wolf spider Lycosa subfusca, which shares its habitat with a bigger spider: the tarantula Brachypelma vagans, in the Yucatan Peninsula. We examined (i) in the field, the predation interaction between medium tarantulas and large, medium and small wolf spiders; (ii) in the laboratory, ... [Show full abstract] size-dependent cannibalism in the wolf spider; and (iii) the simultaneous occurrence of small, medium and large tarantulas, and small, medium and large wolf spiders in three sites in the field (secondary forest, backyard and grazed lawn). In predation interactions, tarantulas attacked large wolf spiders more frequently than medium ones (55% vs 25%), and never attacked small ones. Cannibalism experiments in the wolf spider showed that attack was always initiated by the larger individual. Attack frequencies increased with spider size, but capture success was similar regardless of spider size. Medium and large wolf spiders were much more abundant in the secondary forest, where tarantulas were absent, than in the other two sites. In contrast, small wolf spiders were much more abundant than large ones in both the backyard and grazed lawn, where tarantulas were common. Spatial segregation of large wolf spiders in the secondary forest appears to be related to the absence of tarantulas. On the other hand, small wolf spiders may prefer sites with tarantulas to avoid predation by larger wolf spiders. In conclusion, the spatial distribution of Lycosa subfusca seems to be the result of the predation by tarantulas and conspecifics.