Local distribution of the lycaenid butterfly, Jalmenus evagoras, in response to host ants and plants

Abstract

The caterpillars of Jalmenus evagoras are tended by ants as they feed upon Acacia trees. In the area of Brisbane, Australia, J. evagoras require ants of the Iridomyrmex anceps species group; predation and parasitism are so intense that larvae and pupae deprived of attendant ants cannot survive (Pierce 1983). We investigated the efficiency with which J. evagoras locate and exploit the "host ant" resource by sampling 737 quadrats in 30 sampling grids and six study sites containing appropriate host plants; ants were collected at baits located in the center of each quadrat. J. evagoras was found in all habitats where I. anceps cooccurred with host Acacia. Nine of the ten sampling grids which had three or more I. anceps/Acacia "host" quadrats also had colonies of J. evagoras present (or immediately adjacent), including sites as far as 35 km apart. Of 19 sampling grids on which "host" quadrats were rare (i.e., less than three quadrats), none had J. evagoras (P<0.001). Within sample grids, I. anceps was distributed indepedently from Acacia trees, suggesting that they are not dependent for their survival on either Acacia or on J. evagoras. Within montane pasture habitats, I. anceps and at least one other ground-dwelling Iridomyrmex species were distributed in mutually exclusive "ant mosaic" territories which were stable during a one month period. I. anceps did not colonize or tend pupae of J. evagoras experimentally placed in adjacent territories of a different, nontending species of Iridomyrmex, demonstrating the integrity of territory boundaries. Sampling of ants in Acacia trees revealed that, in the absence of J. evagoras, Iridomyrmex workers are not common above ground level, and that their numbers decline in larger trees (P=0.02). In I. anceps territories, eight of nine J. evagoras pupae placed in trees over 3.0 m tall were not found after 24 h whereas all ten controls placed in low trees were found and tended (P=0.00012). This may explain why J. evagoras tends to oviposit in trees less than 2.0 m tall. An alternative hypothesis, that smaller trees have higher content of total nitrogen, and are threfore more nutritious, was not supported. We conclude that the local distribution and host tree selection by J. evagoras is dependent upon the distribution, patchiness, and foraging behavior of the host ant, I. anceps, and its spatial overlap with a number of species of host Acacia.