What determines conditionality in ant–Hemiptera interactions? Hemiptera habitat preference and the role of local ant activity.

1. This paper describes spatial variation in density of a mutualist Hemiptera, and attempts to elucidate an understanding of the spatial variation in conditionality of its mutualism with ants. 2. Aggregations of the membracid treehopper, Campylenchia sp., occurred more frequently, and with higher number of individuals, in isolated trees as compared with their occurrence in the same tree species in wooded patches. 3. Using treehopper aggregations as the level of replication, there was a habitat × ant presence interaction associated with colony survival time. However, when the median survival time of all aggregations within each tree was used, only an additive effect of both habitat type and ant presence were apparent. This suggests that treehopper aggregations in some isolated trees experienced more benefit from ant presence than other trees. 4. The ant:treehopper ratio averaged over each tree was correlated with effect of ant presence on survivorship (survival time without ants/survival time with ants). Therefore, conditionality in the ant–membracid mutualism appears related to ant attendance level at the scale of individual trees. 5. Ant attendance at treehopper aggregations in isolated and matorral trees did not differ, and therefore there is no clear mechanism by which ants create the contrast in treehopper densities between the two habitat types. 6. The main force shaping initial treehopper densities in the two habitats is likely plant quality. Results suggest that growing situation (or more specifically plant health) is likely a good predictor of membracid density, while general ant activity within a tree is the best indicator of interaction strength. [ABSTRACT FROM AUTHOR]