Abstract

The immature stages of hoverflies of the subfamily Microdontinae (Diptera: Syrphidae) develop in ant nests, as predators of the ant brood. The present paper reviews published and unpublished records of associations of Microdontinae with ants, in order to discuss the following questions. (1) Are all Microdontinae associated with ants? (2) Are Microdontinae associated with all ants? (3) Are particular clades of Microdontinae associated with particular clades of ants? (4) Are Microdontinae associated with other insects? A total number of 109 associations between the groups are evaluated, relating to 43 species of Microdontinae belonging to 14 genera, and to at least 69 species of ants belonging to 24 genera and five subfamilies. The taxa of Microdontinae found in association with ants occur scattered throughout their phylogenetic tree. One of the supposedly most basal taxa (Mixogaster) is associated with ants, suggesting that associations with ants evolved early in the history of the subfamily and have remained a predominant feature of their lifestyle. Among ants, associations with Microdontinae are known from subfamilies Ponerinae, Dolichoderinae, Formicinae, Myrmicinae, and Pseudomyrmecinae. These subfamilies comprise more than 95% of all ant species. Interestingly, no associations are known with “dorylomorph” ants (army ants and relatives).

1. Introduction

Ants “run much of the terrestrial world,” is the claim of Hölldobler and Wilson [1] in the opening lines of their landmark book The ants. This may be true, but the colonies of ants—on their turn—are to some extent affected by many species of myrmecophilous organisms which live in their nests, especially insects and other arthropods. Some of these are not detrimental to the ants or can even be considered beneficial, for example, because they clean up the nests or provide the ants with certain nutrients. Other species of myrmecophilous insects, however, are predators of the ant brood or the adult ants. The larvae of hoverflies of the subfamily Microdontinae (Diptera: Syrphidae) exemplify the latter category.

The nature of the feeding habits of the slug-like larvae of Microdontinae has long remained uncertain. Several authors have suggested that they live as scavengers or feed on pellets of food ejected by the worker ants [25]. More recently, however, accumulated evidence showed that larvae of at least a number of species of Microdon Meigen and Omegasyrphus Giglio-Tos are predators, feeding on eggs, larvae, and pupae of ants [610]. There are a few reports of Microdontinae larvae feeding on aphids and coccids attended by ants [1113], but these could so far not be confirmed. Little is known about the degree of taxonomic specialization exhibited by Microdontinae with respect to their host ants, but available evidence suggests that Microdon species are highly specialized, although this may differ between species [1417]. It seems probable that a certain degree of host specialization is required for predators living in ants nests, because the predators need to make sure that they are not recognized by the ants as hostile intruders. For some Microdon species it has been established that their larvae use “chemical mimicry” to prevent them from being attacked by the ants: the fly larvae possess cuticular hydrocarbons similar to those of the ants [14, 15].

The impact of larvae of Microdontinae on ant colonies is potentially large. Duffield [7] reported that third-instar Microdon larvae could consume 8–10 ant larvae in 30 minutes, and Barr [6] stated that a Microdon larva may consume up to 125 ant larvae during its life. With an average number of five or six Microdon larvae per nest [6], over 700 ant larvae would be consumed per nest. A more indirect way in which Microdon larvae possibly affect the fitness of ant colonies was revealed by Gardner et al. [18]. They found that workers of a Microdon-infested polygynous ant colony are less closely related to each other than workers of uninfested colonies. They explained this by arguing that it is harder for a Microdon larva to intrude in a genetically homogeneous colony, because in such a colony the worker ants smell more alike and will therefore more easily recognize an intruder. So, a decreased genetic diversity will reduce the chance of becoming infested with Microdon larvae.

Worldwide, 454 valid species of Microdontinae are known [19], which may be only half or less of the actual species number (estimation by the author based on unpublished data). Approximately 12,500 species of ants are known [20]. Little is known about associations between species of Microdontinae and species of ants. Because of the potential impact of these flies on ant colonies, and hence on ecosystems, it is interesting to learn more about these associations. Besides, this information may be useful for research on subjects like the evolution of host association, chemical mimicry, and (triggers of) cryptic speciation. The present paper aims to summarize available knowledge of associations of Microdontinae with ants, in order to answer the following questions.(1)Are all Microdontinae associated with ants?(2)Are Microdontinae associated with all ants?(3)Are particular clades of Microdontinae associated with particular clades of ants?(4)Are Microdontinae also associated with other insects besides ants?

2. Material and Methods

2.1. Host Associations

The literature has been reviewed and records on associations of Microdontinae with ants and other insects were assembled. Omitted from the dataset were references to host associations for which considerable doubt exists as to whether the identifications are correct. This is especially the case with several older references to European species, since it became clear that certain taxa actually comprise cryptic species complexes, as in Microdon analis (Macquart)/M. major Andries and M. mutabilis (Linnaeus)/M. myrmica (Schönrogge et al.) [16, 21]. The following records were excluded because of this reason (names as in cited publication): Microdon mutabilis in nests of Lasius niger (Linneaus), Myrmica ruginodis Nylander, and Formica fusca Linnaeus [2]; Microdon eggeri Mik in nests of Lasius niger [2]; Microdon eggeri in nests of Formica sanguinea Latreille [22]; Microdon devius (Linnaeus) in nests of Formica sanguinea and Lasius fuliginosus (Latreille) [2325]; Microdon devius in nests of Formica fusca, and Formica rufa Linnaeus [25]; Microdon mutabilis in nests of Formica fusca, F. rufa, F. rufibarbis Fabricius, Lasius niger, L. brunneus (Latreille), and L. flavus (Fabricius) [25]. These records were, however, included in a more generalized way, that is, as associations of species of Microdon s.s. with the ant genera Formica Linnaeus, Lasius Fabricius, and Myrmica Latreille. The records reported in the literature on European Microdon (the only genus of Microdontinae occurring in Europe) have not been fully surveyed, as this would not add information to the generic level at which this study was conducted.

Weber [26] reported larvae “of the Microdon type” from nests of the ant Ectatomma ruidum (Roger) (subfamily Ectatomminae). However, his figure does not show a Microdon larva but a larva belonging to another family of Diptera Cyclorrhapha (possibly Phoridae). Hence, this record was excluded from the dataset analyzed in this paper.

In addition to the survey of the literature, associations found in entomological collections were recorded. Such records were noted when an empty puparium was mounted together with an adult specimen, and the label mentioned a genus or species of host ant. Records were taken from the following collections: Natural History Museum, London (BMNH); National Museums of Scotland, Edinburgh (RSME); United States National Museum, Washington D.C. (USNM); Zoölogisch Museum Amsterdam (ZMAN, recently included in the collection of Naturalis Biodiversity Center (RMNH), Leiden).

2.2. Taxonomy and Phylogeny

Classification of Microdontinae follows Reemer and Ståhls [19]. Classification of ants is updated to modern standards according to Bolton [27]. A recent phylogenetic hypothesis for intrageneric relationships of Microdontinae is obtained from Reemer and Ståhls [28], who presented a tree based on parsimony analysis of combined molecular and morphological characters. All specific taxa were pruned from this tree in order to obtain a tree of generic relationships only. For ants, several recent phylogenetic hypotheses are available (e.g., [29, 30]), which are incongruent at some points. Therefore, in the present study, the tree of extant subfamilies as compiled by Ward [31] is used, because this summarizes relationships which are well supported by all recent studies.

3. Results

Table 1 lists 109 recorded associations of Microdontinae with ants, 105 of which are based on the literature and four are based on collection surveys. These records concern 43 species of Microdontinae belonging to 14 genera, and at least 69 species of ants belonging to 24 genera and five subfamilies (Ponerinae, Dolichoderinae, Pseudomyrmecinae, Formicinae, and Myrmicinae). The distribution of recorded association over the major biogeographic regions is as follows: Nearctic 62, Palaearctic 18, Neotropical 18, Australia/Oceania 6, Afrotropical 4, and Oriental 1.

Figure 1 presents a phylogenetic hypothesis for 28 (out of 43) genera of Microdontinae, with indications of known associations with subfamilies of ants. Figure 2 presents a phylogenetic hypothesis for all extant subfamilies of ants, with indications of known associations with Microdontinae.

4. Discussion

4.1. Are All Microdontinae Associated with Ants?

The larval habits remain unknown for the majority of microdontine taxa: 14 out of 43 genera are now known to be associated with ants. The present results, however, indicate that associations with ants are found well distributed over the tree representing the most recent phylogenetic hypothesis of Microdontinae (Figure 1). Spheginobaccha de Meijere (tribe Spheginobacchini) is the sister group to all other Microdontinae (tribe Microdontini), but the larvae of this taxon are presently unknown. Within the tribe Microdontini (the remaining part of the tree), Mixogaster Macquart is the first genus to branch off (a strongly supported clade; see Reemer and Ståhls [28]), and larvae of a species belonging to this genus have been found in an ant nest [35]. These results do not give a definite answer to the question, but they indicate that associations with ants are a dominant feature of larval biology for all Microdontinae, which has evolved early in the evolution of the group. Obviously, as already exclaimed by Cheng and Thompson [66], “one wants to know what the larvae of Spheginobaccha do!”

4.2. Are Microdontinae Associated with All Ants?

The ant genera which have been recorded in association with Microdontinae belong to five subfamilies: Ponerinae, Dolichoderinae, Pseudomyrmecinae, Myrmicinae, and Formicinae. The four latter subfamilies all belong to the “formicoid clade” (Figure 2), as defined by Ward [31].

So far, no species of Microdontinae are known to be associated with the dorylomorph ant subfamilies (Figure 2), which also belong to the formicoid clade. This group includes the army ants: four subfamilies which are characterized by a nomadic lifestyle and mass foraging. The lack of records of associations of Microdontinae with army ants is remarkable, as these ants are relatively well studied and are known to host extremely rich communities of myrmecophiles [1]. It is tempting to hypothesize that the nomadic behaviour of these ants somehow prevents Microdontinae from getting adapted to them. However, when species numbers of the ant subfamilies are taken into account (Figure 2), it is clear that making such a statement would be jumping to conclusions. Together, the five subfamilies known to be associated with Microdontinae contain more than 12,000 species of ants, which is more than 95% of the world’s ant diversity. With so few records available, chances that microdontine larvae are found in assocation with other groups of ants are small. These chances are even smaller when the geographical bias of the records is taken into consideration: a large majority of the records originate from the Palaearctic and Nearctic regions, whereas the subfamilies outside of the formicoid clade are predominantly tropical.

4.3. Are Certain Clades of Microdontinae Associated with Certain Clades of Ants?

So far, only one record of a poneroid ant associated with Microdontinae (Hypselosyrphus Hull) is known. Whether this is an exception or the tip of an iceberg remains uncertain until more data on associations of tropical taxa become available.

Figure 1 indicates that associations with the ant subfamilies Formicinae and Myrmicinae occur on several parts of the microdontine tree, without any obvious pattern. Associations with both subfamilies are even found within the same genus. For instance, Microdon (s.s.) mutabilis is associated with ants of the genus Formica (Formicinae), whereas the closely related Microdon myrmicae, which until recently was not separated from M. mutabilis, is associated with Myrmica ants [16]. Larvae of different species of Paramixogaster Brunetti were also recorded in association with ants of Formicinae and Myrmicinae (Table 1). These records suggest that shifts in host association between Formicinae and Myrmicinae occur relatively frequently. Whether this is also true for other ant subfamilies, or for other genera of Microdontinae, cannot be deduced from the presently available data. For most other genera of Microdontinae only one association is known (Table 1). An exception is Stipomorpha Hull, of which the larvae of two species were found in Crematogaster Lund nests. Another exception is Oligeriops Hull, of which two species were found in nests of Iridomyrmex Mayr. Whether these records indicate some degree of parallel evolution remains an open question, at least until a larger number of associations is be known.

4.4. Associations with Other Insects?

Wasmann [23, 25] reported having found Microdon larvae in the nests of wasps and termites. This record was repeated by other authors [2, 4] but has never since been confirmed. Wheeler [32] reported a finding of Microdon larvae in the chambers of termite nests, but those were abandoned by the termites and occupied by ants of the genus Camponotus Mayr. He wrote “These ants regularly take possession of the chambers adjacent to the tree trunk supporting the termitarium and permit the termites to inhabit the remainder of the structure.” A similar explanation may be true for Wasmann’s reports of Microdon larvae in wasps and termites nests.

Another, apparently independent, record of an association of Microdon with termites was mentioned by Séguy [67], who stated that the larvae of a Microdon species were attracted to exuding saps on certain fruit trees that were attacked by termites. However, the source of this record is unclear and no figures of the larvae are provided, so whether this report really concerns Microdon larvae remains doubtful.

Pendlebury [68] described Paramixogaster icariiformis Pendlebury and hypothesized that its larva lives in the nest of the wasp species that it mimics, without presenting any other evidence than their similarity in appearance.

So, there are no convincing records of Microdontinae living in the nests of other insects than ants. All published records suggesting such associations can be considered doubtful.

5. Concluding Remarks

With so few associations known among the total of 12,500 described ant species and 454 described species of Microdontinae, any conclusion about evolutionary trends claiming general validity would be premature. Despite this, the present paper is the first to demonstrate in a phylogenetic context that it seems likely that all Microdontinae are associated with ants. Vice versa, associations with Microdontinae are found among a large diversity of ant subfamilies, suggesting that all ants may be prone to “infestation” by Microdontinae. Exceptions may occur, such as the army ants, with which no associations are known so far.

At least as interesting as the questions discussed in this paper is the question as to the exact nature of the associations between Microdontinae and ants. Available evidence for a few Palaearctic and Nearctic species shows that these species are predators of immature stages of ants (see Introduction). The species for which this feeding mode is known all belong to Microdon s.s. (in the sense of Reemer and Ståhls [19]) and Omegasyrphus. Whether the larvae of other genera of Microdontinae also feed this way remains to be discovered.

Acknowledgments

The author would like to thank the following persons for sharing information or helping with studying the collections they are curating: Ben Brugge (ZMAN), Jean-Paul Lachaud, Mirian Nunes Morales (DZUP), Gabriela Pérez-Lachaud, Graham Rotheray (RSME), and Manuel Zumbado (INBio). André van Loon and Gunilla Ståhls are thanked for commenting on an earlier version of the paper. Jean-Paul Lachaud was very helpful with updating the taxonomy of the ant hosts.