Myrmecophily

One of the most recent attempts to investigate the structure underlying interspecific mutualism was performed by Leimar and Axén (1993). They studied the interaction of lycaenid butterfly larvae with ants. The caterpillars of many lycaenid species have a 'dorsal nectar organ', an exocrine gland that secretes a liquid containing carbohydrates and free amino acids. The ants harvest the liquid and in turn provide the larva with protection (see Pierce 1987 for more detailed information on the characteristics of this relationship). Since the sweet substance is costly to produce (Pierce 1987, Leimar and Axtén 1993), the larva would profit from obtaining ant attendance without releasing any liquid. On the other hand, ants may profit from harvesting the food rewards instead of defending the caterpillar. So both would do better defecting and thus T > R > P > S. The results from Leimar and Axén (1993) suggest that the caterpillars do respond to ant attendance, i.e. retaliate. Furthermore, w might be sufficiently high, since the butterfly larvae never leave their host plants (Pierce 1987). However, there are several differences to the classical IPD:

1. The rewards to the players are asymmetrical.
2. The moves are not necessarily simultaneous.
3. The decision rules are influenced by a third party: predators and parasitoids attacking the larva (Leimar and Axtén 1993).
4. The investments in each interaction can be varied gradually (Leimar and Axén 1993).
5. Individual recognition is not likely between lycaenids and ants (Pierce 1987).

Hence the structure of lycaenid-ant interaction seems to resemble an IPD, but it is presumably more complex.