Ant responses in a lycaenid-ant symbiosis are not facilitated by cuticular compounds alone

Abstract

Initiating partnerships in protective symbioses can be asymmetrical if there is a risk of attack from their symbionts. Myrmecophiles may encounter chemically mediated recognition systems that allow the host ants to distinguish nestmates from natural enemies, including non-nestmate conspecifics. The immature stages of the lycaenid butterfly Jalmenus evagoras form an obligate symbiosis with workers of Iridomyrmex mayri that protect them against natural enemies. However, the first instar larvae cannot anticipate this colony-specific chemical recognition system, since they are unlikely to encounter workers from the same colony that tended their mother. We show experimentally that workers of I. mayri can use chemical signals alone to distinguish between conspecifics and the larvae of J. evagoras; between nestmate and non-nestmate conspecifics and between larvae tended by nestmate and non-nestmate conspecifics. Nevertheless, we also show experimentally that while workers paid more attention to fourth than second instar larvae, they did not respond more aggressively to larvae that had been tended by non-nestmate versus nestmate workers. These data suggest that workers pay attention to other signals, perhaps via tactile, visual or vibratory sensory modalities, thereby allowing the butterfly myrmecophiles to mitigate the risks associated with the chemically mediated colony-specific recognition systems of their ant hosts.

Keywords: Lycaenidae; ants; colony recognition; cooperation; myrmecophily.

Figure 1.

Behavioural responses of workers of I. mayri towards filter paper strips containing cuticular extracts from conspecific workers from the same colony (familiar) or different colony (unfamiliar); larvae of J. evagoras from the same A. melanoxylon host plant (familiar) or different A. melanoxylon host plant (unfamiliar); and a hexane control (designated unfamiliar). The behaviours, reported as frequency per minute per eight workers, include (a) brief antennations; (b) long antennations; (c) dragging; (d) biting (for convenience, two extreme outliers for ‘familiar larvae’ were removed for the figure only) and (e) grooming. The length of the box represents the interquartile range, with the line within the box representing the median and X representing the mean. The results of statistical analyses are given in table 2, and we used Tukey’s post hoc test (p < 0.05) where there was a significant species by familiarity interaction term (treatments that do not have the same Greek symbol are significantly different). Note the different y-axis scales.

Figure 2.

Behavioural responses of workers of I. mayri towards second and fourth instar larvae of J. evagoras from the same (familiar) or different (unfamiliar) A. melanoxylon host plant. Long antennations (a) and biting (b) are calculated as frequency per minute per eight workers. The length of the box represents the interquartile range, with the line within the box representing the median and X representing the mean. The results of statistical analyses are given in table 2. Note the different y-axis scales.