Leaf Damage and Associated Cues Induce Aggressive Ant Recruitment in a Neotropical Ant-Plant

Abstract

Induced chemical responses following herbivory are common in plants. Plant responses that change the level of physical or biotic defense are less well documented and poorly understood. Many Azteca spp. ants are obligate inhabitants of Cecropia spp. trees. In such ant–plant associations the ants are thought to be analogous to chemical defenses; previous experiments have demonstrated that ant occupation of C. obtusifolia reduced herbivory and plant competition and increased growth. Experiments, conducted over two years, on the dynamics of ant defense demonstrate that leaf damage caused a fivefold increase in the number of Azteca spp. ants on damaged leaves of C. obtusifolia compared to that on disturbed but undamaged control leaves. Ant activity peaked 8–12 min after damage, and differences between damaged and control leaves remained evident for 24 h. Such rapid induction of ant recruitment is likely to be particularly effective against unpredictable and mobile herbivores. The magnitude of the induced ant response to damage was strongly correlated with the number of ants patrolling the leaves before damage occurred. Ant responses to disturbance were not influenced by the presence of damage that had been applied 24 h previously. However, ant responses to subsequent damage, 24 h after initial damage, resulted in greater recruitment than to previously undamaged leaves. Ant recruitment to several other cues associated with herbivory was also tested. Presence of pyralid caterpillars that naturally feed on C. obtusifolia induced a low level of ant recruitment, and most larvae were removed from leaves by the ants within 10 min. Exposure to plant sap collected from damaged conspecifics and a commercially available green leaf volatile (hexanal) commonly released by plants after damage, both resulted in a doubling of ant numbers relative to controls. However, the levels of recruitment in response to these stimuli were insufficient to account for the high numbers of ants and persistence of recruitment observed on experimentally damaged leaves. Experimental wounding of leaves with minimal leaf tissue removal (using pin pricks) revealed that leaf wounds per se can only partially explain the induced ant recruitment following leaf damage. The type of herbivory and size of leaf wounds may be important cues for ant recruitment. Severed C. obtusifolia leaves that were freshly damaged failed to elicit an induced ant response when held adjacent to conspecific leaves with ants. However, induction of ant recruitment on damaged plants did significantly induce a low level of ant recruitment on neighboring conspecifics, providing evidence for interplant communication. Induced ant responses in the Cecropia–Azteca system are the result of multiple physical and chemical cues associated with herbivory. Ant responses to herbivory, although not previously studied in detail, are likely to be common among myrmecophytic plants and are likely to be an important component of antiherbivore defense in such systems.