Abstract

 Ant-attendance occurs in the larvae of many Lycaenidae butterfly species, this attendance is termed myrmecophily.  Myrmecophily occurs in the Lycaenidae caterpillars that possess three known types of organs, pore cupola organs (PCOs), dorsal nectar organs (DNOs), and tentacle organs (TOs).  Each of these organs plays a unique role in the caterpillars' relationship with the attending ant species.  Lycaenidae species lacking the DNO are defined as myrmecoxenous and are not as commonly attended by ants, although the ants will not aggressively attack them.  It is thus the secretions from the DNO that seem to play the largest role in myrmecophily.  The relationship between myrmecophilous species with ants is an evolves characteristic that has an ecological impact.  Caterpillars with nutritious secretions supply an alternate food source for ants to feed on, those caterpillars without a DNO may become the ants' next meal.  The relationship is also mutualistic.  Caterpillars are less likely to be parasitized if they have ant attendants and the ants get nutritive compensation from the caterpillar they help protect.  The nectar secretions from the DNO may vary depending on the larval age, the number of attending ants, and how long the attendance has been going on.  These factors will influence the energetic benefits that the ants get.  However, if energy gain is even minimal the ants will often keep attending the Lycaenidae larvae.  The host plant will also affect the energetics.  The plant is indirectly responsible for the DNO secretions as it is the caterpillars' food source and therefore will affect the quality of the secretion.  As ant attendance can not be constant, parasitism does occur and this too affects the ability of the caterpillar to supply the ants with nectar.  The various energetics and factors affecting myrmecophilous interactions will be further addressed.

Introduction

 Interspecific interactions are an interesting facet of the natural realm that have long intrigued scientists.  Some species have developed distinct and unique relationships with other species over time.  These relationships may vary in the degrees of positive and negative effects to each species, but the interaction between the two species is survival based.  Being able to survive close interaction with another species is the main evolutionary challenge of mutualism (Doebeli and Knowlton 1998).

 Ants often attend the larvae of many species in the family Lycaenidae (Order Lepidoptera) (Fiedler 1991, Fiedler et al. 1996).  This ant-attendance is known as myrmecophily.  Myrmecophilous larvae have three specific organs that elicit ant-attendance, pore cupola organs (PCOs), tentacle organs (TOs), and dorsal nectar organs (DNOs).  Ants will usually attack invaders or potential prey, but these organs appease the ants and supply them with nutritious secretions; the ants in turn provide a guard against predators and parasites of the caterpillars.

   Based on fossilized evidence, relationships have existed between butterfly larvae and ants since at least 15 million years ago (DeVries and Poinar, 1997).  Myrmecophilic associations and behavior are therefore evolutionarily and genetically ingrained into each species.  But not every interaction is of the same intensity.  Myrmecophilous relationships can range from truly mutualistic to commensalic to quasi parasitic, depending on the species of Lycaenidae (Fiedler 1991, Fiedler et al. 1996, Sanetra and Fiedler 1996).

 There are other factors, besides species, that have an effect on the interaction between the ants and lycaenid larvae.  Variations in actual ant-attendance, larval diet, habitat, the host plant, and the presence of parasitoids have ramifications on caterpillar behavior (Fiedler and Hummel 1995, Burghardt and Fiedler 1996a and b, Baumgarten and Fiedler 1997/98,  Thomas et al. 1998).
 

Myrmecophilic communication

 There are three known types of lycaenid larval organs that influence myrmecophily.  It is thought that these organs are the most important aspect in the maintenance of the relationship with ants (Fiedler 1991).  These organs have most likely evolved to be best adapted within each species.1

 The first of these organs are the pore cupola organs (PCOs).  The PCOs are small glandular structures that are derived from hairs, the hair shaft is perforated with many tiny pores with a diameter of 0.1-0.2 mm (Fiedler 1991).  The PCOs occur in both larvae and pupal forms of Lycaenidae and have been observed in all but one of investigated species so far.  PCOs vary in morphology and distribution between different subgroups of the Lycaenidae.  There are concentrations of pores around the spiracles and the DNO (if present).  Ants that antennally stimulate the organ are rewarded with a secretion of an as yet unknown chemical composition.  Pierce (1983) found that there were amino acids in the secretions, and the amount of amino acids were directly proportional to the attractiveness of the larva to the ants.  However not all lycaenids elicit attendance, even if they do have PCOs.  Therefore there must be some other source of communication between the ants and caterpillars.

 The second type of organs, tentacle organs (TOs), are eversible epidermal tubes located on the eighth abdominal segment of many lycaenids (Fiedler 1991, Fiedler et al. 1996).  The TOs have numerous setae branching from the top of each organ.  The TOs evert when the larvae are stimulated by ants, when the larvae are crawling from feeding to resting sites, or when the larvae are otherwise disturbed.  The tentacles are everted via localized pressure changes in the hemolymph.  If everted TOs are touched, they withdraw instantly; the TOs may evert and withdraw several times in succession.  The eversion of TOs has been observed to cause a change in behavior of nearby ants.  Ants within a few mm radius will react by running around and over the caterpillar at increased speeds.  It is thought that an ant alarm pheromone mimic is being released by the eversion of the TOs.  Proof for this hypothesis comes from the fact that TO eversions do not result in alarm behavior from all species of ants; this also shows that there is attendant specificity in some myrmecophilous relationships.

 The third organ type is the dorsal nectar organ (DNO).  The DNO is located on the seventh abdominal segment and is often surrounded by a concentration of PCOs and specialized setae (Fiedler 1991, Fiedler et al. 1996).  DNOs secrete clear fluid droplets when the organ is antennated by ants.  The DNO is often vigorously stimulated and the fluid is eagerly consumed by the ants.  The chemical composition of the droplets is thought to vary between species, but variable amounts of carbohydrates and amino acids have been described in several species (Maschwitz et al. 1975, Pierce 1983).  From the rapidity at which the ants imbibe the droplets from the DNO and the limited research on the composition of the droplets, the DNO secretions are evidently a valuable source of nutrition to the attending ants.

 There has also been some research into mechanical communication by the lycaenid larvae (DeVries 1990).  The caterpillars produce vibrations that are carried by the substrate the caterpillars are on or else it possibly for vibrations to travel through the air.  The source of these vibrations is unknown, and there is no proof yet of a connection to myrmecophily (Fiedler 1991)

Interactions

 Lycaenidae caterpillars do not actively seek ant-attendants, they wait for the ants to find them (Fiedler 1991).  Many ants are active foragers and can be aggressive.  If ants come across a potential prey item they may attack and kill it.  Many lycaenid larvae could be at risk, however the aforementioned organs work in a way to appease the ants.  Instead of killing the larva the ants will vigorously palpate the DNO and PCOs to evoke nutritious droplets.  The caterpillar's secretions appear to calm the ants as they drink the droplets.  Since the ants do not kill the caterpillar as prey, the benefits of nectar consumption must outweigh energy gain from consumption of the caterpillar itself (Fiedler et al. 1996).  The ants will remain on and around the caterpillar imbibing the droplets as they are secreted.  In exchange for DNO secretions the caterpillar gains ant-attendants which may serve as a guard to ward off predators and parasites.  Myrmecophily is a stable mutualistic interaction.

 Lycaenid caterpillars without a functional DNO have been designated myrmecoxenous species (Fiedler 1991, Fiedler et al. 1996).  Ants encountering these caterpillars will initially be calmed by the secretions from the PCOs.  However the PCO secretions are small and not as attractive to ants as the secretions from DNOs, relatively soon after contact the ant will ignore the caterpillar.  These caterpillars rarely have stable relationships with ants, but they are able to avoid being killed.  Their selective advantage is to be able to escape ant attacks.  Myrmecoxenous species do not have to invest energy in producing large droplets to feed the ants and are able to deter ants from attacking, but they do not receive the potential guardians.

 The main difference between myrmecophilous and myrmecoxenous Lycaenidae species is the respective presence or lack of a DNO (Fiedler 1991, Fiedler et al. 1996).  It is thus reasonable to assume that the DNO is the organ that perpetuates myrmecophily.

Degree of dependency - species specificity

 Although the majority of myrmecophilous relationships of the Lycaenidae are not ant specific, some are very specific (Fiedler 1991, Burghardt and Fiedler 1996a).  The degree of the caterpillar's dependence on the ants also varies between species.

 Caterpillars that can survive without ant-attendants are termed facultative myrmecophiles (Fiedler 1991, Fiedler et al. 1996).  There is a wide range of interactions between facultative myrmecophiles and ants, from occasional attendance to almost permanent.  The intensity of interactions may be a factor of the quality of the secretions (higher quality results in increased attendance) or the habitat of the caterpillars (accessibility by or presence of ants).  Facultative myrmecophiles are not ant specific or even very ant dependent, therefore their chances of survival under ecological stress would be greater than caterpillars which require specific ant-attendants (Thomas et al. 1998).

 Larvae which have relationships with specific ant species are termed obligate myrmecophiles (Fiedler et al. 1996, Thomas et al. 1998).  These larvae are almost absolutely dependent on their particular ant-attendants without which survival would be almost impossible.  When laying eggs the adult female butterfly searches for the correct host ant to ensure that her brood will be taken care of.  Obligate myrmecophiles also have host specific secretions which non-host ants may ignore (Fiedler et al. 1996).  Some of the obligate myrmecophiles rely on the attendants to drink the nectar droplets produced because they produce large drops regardless of being stimulated (Fiedler and Saam 1994, Fiedler et al. 1996).  If these drops are not removed fungal infections may occur in the region of the DNO, eventually leading to the death of the caterpillar.  Some ants that interact only with specific Lycaenidae species will kill other species, even if the foreign species is otherwise myrmecophilic (Fiedler et al. 1996).  Obligate myrmecophiles are also more susceptible to mortality due to environmental or ecological pressures because of their absolute dependence on their specific ant-attendants (Thomas et al. 1998).

Effects of attendance

 Various effects of ant-attendance have been observed in laboratory settings.  For most of the experiments the ant species used was Lasius flavus (e.g. Fiedler and Hummel 1995, Burghardt and Fiedler 1996).  Although L. flavus ants rarely attend caterpillars in nature (Fiedler 1991), they have a highly developed trophobiotic behavior and will readily attend ants in laboratory experiments (Fiedler 1991, Burghardt and Fiedler 1996a, Baumgarten and Fiedler 1997/98).

 DNO and TO activity were found to be higher within the first three minutes of interaction (Fiedler and Hagemann 1995, Burghardt and Fiedler 1996a).  Once the caterpillars have made contact with the ants and the ants are initially appeased, the caterpillars slow down the rate of droplet secretion and slowed down TO eversion rates.  This lower rate of activity is to minimize the amount of metabolic energy the caterpillar exerts once a stable association is formed, thereby prolonging the attendance time of the ants.  Although the rate of drop secretion was slowed down it became constant, indicating that the ants exerted pressure to supply secretions (Fiedler and Saam 1994).

 The has been some research on the effects of ant-attendance times and number of attending ants on some larval species, but it is fairly limited (Fiedler and Saam 1994, Fiedler and Hagemann 1995, Fiedler and Hummel 1995).  An increase in the number of attendant ants resulted in an increased rate of secretion, but only up to a point, after which attending time per individual ant decreased (Fiedler and Hagemann 1995).  This saturation level corresponds with the benefits each ant gets from attending, foraging costs for the ants increase with each added ant.  More ants result in better protection for the caterpillar, but less energetic return for each ant.  There seems to be an optimal level at which both species benefit the most, this level will most likely vary depending on factors such as ant species, population density of ants and caterpillars, and larval age.

 The attendance time of ants impacts development of the lycaenid larva.  Caterpillars of the species Aricia agestis were attended by ants for different periods of time (Fiedler and Hummel 1995).  Individuals that were permanently attended grew on average 10% heavier than larvae who were only attended for 30 minute periods/day.  However this increased size led to an increase in development time by one day.  There is an apparent trade off between vulnerability and adult size.  Non-attended larvae in the wild are at higher risk of predation and therefore will pupate earlier than attended ants.  Although the non-attended larvae will have a lower weight, they have a better chance of survival.  Predation of attended larvae is lower, they can take longer to develop because they are protected by an ant guard.  It has been suggested that overcompensation for the costs of ant-attendance by facultative myrmecophilous lycaenids occurs and results in these higher adult weights (Fiedler and Saam 1994, Fiedler and Hummel 1995).

 The investment into myrmecophily is beneficial, because appeasement of potentially dangerous ants will supply them with protectors against other predators.  Fiedler and Hagemann (1995) propose that once a stable association has been formed in the larval stage, the association is likely to last in the pupal stage, in which the PCOs are still available to the ants.  Protection throughout the pupal stage will lead to an increased survival rate by keeping predators away while the lycaenid is so vulnerable.

Effects of host plant and diet

 The ability to supply secretions from the various organs depends on what the caterpillar feeds on (Fiedler 1991, Fiedler 1995, Burghardt and Fiedler 1996b).  The caterpillar assimilates the resources from the food it ingests and then converts some of it into the secretions that are passed on to the ant-attendants.  The caterpillars need to be able to procure enough energy from the food in order to offset costs of development.  The caterpillars also need to be able to extract from their food enough carbohydrates and amino acids to be able to maintain myrmecophilic relationships.

 Pierce (1983) suggested that nitrogen-rich plants such as legumes would be an ideal food source for myrmecophilic species.  Although nitrogen content is important, and legumes are proposed as an ancestral host plant, there is a wide variety of non-nitrogen fixing plants that myrmecophiles use (Fiedler 1995).

 The quality of secretions by phytophagous caterpillars varies with the type and part of the plant that are ingested.  Polyommatus icarus caterpillars were raised on the inflorescences and the foliage and of several different plants (Burghardt and Fiedler 1996b).  The caterpillars that ate the inflorescences attained the highest weights and were able to deliver DNO secretions at higher rates.  The caterpillars that ate the foliage were of lighter weights and had low secretion rates.  Inflorescences typically have a higher nitrogen content and also contain more water than foliage, both of these factors will positively effect the caterpillar's ability to form secretions.  From these results one can conclude that the inflorescences are a better source of food than the foliage.  This experiment showed that the ability to maintain myrmecophily is related to food quality.  However caterpillars will still secrete droplets and develop even if under food stress (Burghardt and Fiedler 1996a), this is probably why some caterpillars will overcompensate (Fiedler and Saam 1994, Fiedler and Hummel 1995).

 Not all myrmecophilic lycaenids are herbivores, a few are aphytophagous and carnivorous (Fiedler 1991, Sanetra and Fiedler 1996).  These carnivorous caterpillars are mostly obligate myrmecophiles that are on the verge of parasites.  Caterpillars of the species Cigaritis acamas spend almost their entire larval life inside the host ant (Crematogaster) colony.  The caterpillars are fed via trophollaxis by the ants and also feed on the ant brood (Sanetra and Fiedler 1996).  However the ants do not protect the brood,. Even while the caterpillar consumed ant brood, ant workers attended the caterpillar, licking and grooming it.  A myrmecophilic interaction of this sort is extremely specific and not well understood.

Effects of parasitoids

 Ant-attendants may offer protection against parasitoids of Lycaenidae larvae, but sometimes the parasitoid will reach the larvae before the ants do, such is the case in many facultative myrmecophiles (Fiedler et al. 1992, Baumgarten and Fiedler 1997/98).  The majority of these parasites are from the orders Hymenoptera and Diptera.  There has been limited research into the effects of parasitism on myrmecophily.  It is difficult to lump all parasites together because there are so many differences between the parasites and the effects they have on caterpillars.  However there are some trends that can be seen from the research that has been done.

 This third-party, parasitoids, complicates an already complex relationship.  Parasitized larvae are still attractive to attending ants, because the parasitoid often leaves the myrmecophilous organs functional.  This is a highly evolved adaptation which has been refined over time, further implying that myrmecophily has been occuring for a very long time.  There are a couple of reasons why the parasitoids leave the organs functional:  1) to avoid hyperparisitism (Baumgarten and Fiedler 1997/98) and 2) to ensure emergence of the parasitoid (Fiedler et al. 1992).  The ants will protect and attend an already parasitized caterpillar if it is able to still produce nutritious secretions.  Even for several days after the coccoon formation, pupation or emergence of some parasitoids, the caterpillar carcass is still able to supply secretions from its myrmecophilous organs (Fiedler et al. 1992, Baumgarten and Fiedler 1997/98).  Much more time and energy needs to be put in to research about the effects of parasitoids on myrmecophily because it is such a broad field and the effects can be so varied.

Energetics

 The energetics of myrmecophilic relationships is another difficult field.  The ants must get some beneficial energy gain if they continue myrmecophily, instead of outright predation of Lycaenidae larvae.  The caterpillars overcompensate while under myrmecophilic conditions in response to the possible energy strain they may undergo.
 Fiedler and Hagemann (1995) estimated the lifetime droplet production for Aricia agestis and based on carbohydrate concentration came up with an average energy content of 5 J.  This was then correlated with previous researchers' data about different ant populations and the ants' energy use.  Fiedler and Hagemann found that even based on their lowest estimates the secretions provide a small, but valuable source of energy to the ants, more than the caterpillar would supply if it were eaten.

Conclusion

 Myrmecophily is wide spread throughout the butterfly family of Lycaenidae.  The evolution of these relationships between butterfly larvae and ants has led to many different intensities of interactions.  Although insects are considered to be simple, interspecific interactions make things more complicated.  There are many other factors that play a role in myrmecophily; plant host, diet, environmental conditions, species, and intensity of interactions to name a few.  Research has brought to light much information about how myrmecophily is controlled, but so much is still not understood.

Literature cited

Baumgarten, H-T. and Fiedler, K., Parasitoids of Lycaenid Butterfly Caterpillars: Different Patterns in Resource Use and Their Impact on the Hosts' Symbiosis with Ants.  Zoologischer Anzeiger 236, 1997/98: 167-180

Burghardt, F. and Fiedler, K., [1996a] Myrmecophilous behaviours in caterpillars of the butterfly, Polyommatus icarus (Rottemburg, 1775): temporal patterns and age dependency(Lepidoptera:Lycaenidae). Nota Lepidopterologica 19(3/4): 212-229

Burghardt, F. and Fiedler, K., [1996b] The influence of diet on growth and secretion behavior of myrmecophilous Polyommatus icarus caterpillars (Lepidoptera: Lycaenidae).  Ecological Entomology 21: 1-8

DeVries, P., Enhancement of symbioses between butterfly caterpillars and ants by vibrational communication. Science 248, 1990: 1104-1106 (cited in Fiedler 1991

DeVries, P. and Poinar, G., Ancient butterfly-ant symbiosis: direct evidence from Dominican amber.  Proc. R. Soc. Lond. B 264, 1997: 1137-1140

Doebli, M. and Knowlton, N., The evolution of interspecific mutualisms.  Proc. Natl. Acad. Sci. USA 95, 1998: 8676-8680

Fiedler, K., Systematic, evolutionary, and ecological implications of myrmecophiliy within the Lycaenidae (Insecta: Lepidoptera: Papilionoidea). Bonner Zoologische Monographien 31, 1991

Fiedler, K., Lycaenid butterflies and plants: is myrmecophily associated with particular hostplant preferences?.  Ethology, Ecology, and Evolution 7, 1995: 107-132

Fiedler, K. and Hagemann, D., The influence of larval age and ant number on myrmecophilous interactions of the African Grass Blue butterfly, Zizeeria knysna(Lepidoptera: Lycaenidae). Journal of Research on the Lepidoptera 31, 1995: 213-231

Fiedler, K. and Hummel, V., Myrmecophily in the brown argus butterfly, Aricia agestis (Lepidoptera: Lycaenidae): Effects of larval age, ant number and persistence of contact with ants. Zoology 91, 1995: 128-137

Fiedler, K., Holldobler, B.and Seufert, P., Butterflies and ants: the communicative domain.  Experientia 52, 1996: 14-24

Fiedler, K and Saam, C., Does ant-attendance influence development in 5 European Lycaenidae butterfly species? (Lepidoptera). Nota Lepidopterologica 17(1/2), 1994: 5-24

Fiedler, K and Saam, C., A "microbial cost" of butterfly-ant mutualism (Lycaenidae). Journal of the Lepidopterists' Society 49(1), 1995: 80-84

Fiedler, K., Seufert, P., Pierce, N., Pearson, J., and Baumgarten, H-T., Exploitation of lycaenid-ant mutualisms by broconid parasitoids.  Journal of Research on the Lepidoptera 31(3-4), 1992: 153-168

Maschwitz, U., W?st, M., and Schurian, K., Bl„ulingsraupen als Zuckerlieferanten f?r Ameisen. Oecologia 18, 1975: 17-21 (cited in Fiedler 1991)

Pierce, N., The ecology and evolution of symbioses between lycaenid butterflies and ants. Dissertation Harvard University, Cambridge, Mass., 1983 (cited in Fiedler 1991)

Sanetra, M. and Fiedler, K.,Behaviour and morphology of an aphytophgous lycaenid caterpillar: Cigaritis (Apharitis) acamas Klug, 1834 (Lepidoptera: Lycaenidae). Nota Lepidopterologica 18(1), 1996: 57-76

Thomas, J., Simcox, D., Wardlaw, J., Elmes, G., Hochberg, M., and Clarke, R., Effects of latitude, altitude and climate on the habitat and conservation of the endangered butterfly Maculinea arion and its Myrmica ant hosts. Journal of Insect Conservation 2, 1998: 39-46