Comparative larval ultramorphology of some myrmecophilous Aleocharinae (Coleoptera, Staphylinidae), with a first description of the larvae of Amidobiatalpa (Heer O, 1841) and Oxypodahaemorrhoa (Mannerheim C.G., 1830), associated with the Formicarufa species group

Abstract The paper describes the external structures of the late larval stages of two Palearctic myrmecophilous staphylinids: Amidobiatalpa and Oxypodahaemorrhoa associated with the Formicarufa species group. This is the first-ever description of the larva of Amidobia, and the only complete, detailed account of the morphology of this developmental stage in the genus Oxypoda currently available. For the first time in these two genera, 13 and 10 larval diagnostic features, respectively, are proposed. Morphological differences have been established between known and the newly described larvae of five species (genera) of myrmecophilous and one non-myrmecophilous Aleocharinae, belonging to three tribes. Amidobiatalpa and O.haemorrhoa are probably typical, tiny predators, like most other Aleocharinae, including non-myrmecophilous ones. Being very small and highly mobile, they are ignored by worker ants. Not surprisingly, no particular larval morphological modifications were found to enable them to survive among ants. Such features have, however, evolved in the larvae of larger aleocharines, that is, those that are perceived by ants and are wholly integrated with their hosts in the ant nest (e.g. Lomechusa). This comparative analysis of the functional morphology of the larvae of known myrmecophilous Aleocharinae is a springboard to further such studies of these interesting insects.


Introduction
Red wood ants from the Formica rufa Linnaeus, 1761 species group are regarded as key insect species in European woodlands because of their vast numbers and the invaluable biocoenotic contribution they make to the ecosystems they inhabit. Being polyphagous predators, they have a major, multidimensional effect on the invertebrate fauna in that they limit the numbers of many harmful woodland phytophages. On the other hand, the presence of ants has a very positive effect on a range of tiny woodland creatures, such as aphids, which ants protect and defend, obtaining honeydew in return (Szujecki 1998, Laakso andSetälä 2000). Moreover, these ants build complex nests with an extensive inner space, which provides a distinctive microclimate with stable levels of temperature and humidity, as well as constant and diverse food resources. It is therefore an optimal but at the same time highly specific microhabitat for a great number of myrmecophilous invertebrates, predominantly insects (Hölldobler and Wilson 1990, Päivinen et al. 2002, Staniec and Zagaja 2008, Parmentier et al. 2014, Parker 2016.
Among the insects associated with ants, beetles (Coleoptera) are the richest and the most diverse in form. According to Hölldobler and Wilson (1990), there are many thousands of myrmecophilous coleopteran species in 35 families. By way of example, as many as 166 species of beetle have been found in the nests of Formica rufa (Päivinen 2002). By far the most numerous group of myrmecophilous beetles are from the family Staphylinidae (rove beetles), the majority of which, in turn, belong to the subfamily Aleocharinae (Wilson 1979, Kistner 1979, Kronauer and Pierce 2011, Parker 2016 represented, among others, by Amidobia talpa Heer O, 1841 and Oxypoda haemorrhoa Mannerheim C.G., 1830. The genus Amidobia Thomson C.G., 1858 (Athetini) contains eight Palearctic species, of which only A. talpa (Heer O., 1841) occurs in Europe (Löbl and Löbl 2015). This is a very small myrmecophile (body length ca 1.5 mm), inhabiting mainly the nests of Formica rufa and related species (Lohse 1974, Burakowski et al. 1981, Koch 1989. To date, this rove beetle has been reported from the nests of Formica rufa, F. pratensis Retzius, 1783, F. aquilonia Yarrow, 1955, F. polyctena Foerster, 1850, F. lugubris Zetterstedt, 1838, F. truncorum Fabricius, 1804, F. execta Nylander, 1846 and Lasius fuliginosus Latreille, 1798 (Päivinen et al. 2002, 2003, Staniec and Zagaja 2008, Parmentier et al. 2014. Its distribution range lies in central Europe and Fennoscandia, extending beyond the Arctic Circle in the north, and reaching the Caucasus, Siberia and the north of the Korean peninsula in the east. It is probably present throughout Poland, but there are still no records of it from many regions (Burakowski et al. 1981, Löbl andLöbl 2015).
The genus Oxypoda Mannerheim, 1830, one of the most species rich aleocharine genera, has a worldwide distribution (Newton et al. 2000, Löbl andLöbl 2015). Oxypoda haemorrhoa (Mannerheim C.G., 1830) (Oxypodini) belongs to the subgenus Bessopo-ra Thomson, 1859, which has 100 species in the Palearctic, 13 of which occur in Poland (Melke 2014, Löbl andLöbl 2015). This rove beetle is widely distributed from North Africa across continental Europe as far as Iceland and northern Fennoscandia; it has also been recorded in Siberia. In Poland, it probably inhabits all parts of the country except the higher mountain areas, although some regions currently lack records (Burakowski et al. 1981). This myrmecophile is one of the smaller European representatives of the genus (body length: 2.0-2.7 mm). It lives mainly in ant nests from the Formica rufa group: F. rufa, F. pratensis, F. polyctena, F. aquilonia and F. lugubris, but one also comes across it in the nests of F. truncorum, F. execta, F. sanguinea Latreille, 1798, F. suecica Adlerz, 1902, F. nigricans Emery, 1909 and Lasius fuliginosus (Päivinen et al. 2002, 2003, Staniec and Zagaja 2008, Lapeva-Gjonova 2013, Parmentier et al. 2014). It has sometimes been found in the neighbourhood of anthills, under plant debris (Burakowski et al. 1981).
Myrmecophilous Aleocharinae, like other beetles associated with ants, are fascinating organisms for research because of their highly interesting morphological, ecological and behavioural adaptations to the distinctive conditions found in ant nests (Kolbe 1971, Hölldobler andWilson 1990). In this context, however, very few data are available concerning the structure of their larval stages. Contemporary data on this subject relate to just three species: Pella laticollis Märkel F., 1844, Thiasophila angulata (Erichson W.F., 1837) and Lomechusa pubicollis Brisout de Barneville Ch.N.F., 1860. The firstmentioned is associated primarily with Lasius fuliginosus, the other two with ants from the Formica rufa group (Staniec et al. 2009, Zagaja et al. 2014.
Presumably, the larval structure of these myrmecophiles, among other characteristics, which actively live and forage in the anthill throughout their development, should well reflect the extent and nature of their integration with their hosts. Therefore, detailed morphological data of the larval forms should prove useful for discovering the distinctive adaptations of myrmecophilous species to life in ant nests and also the relations between them and their hosts.
The links of numerous Staphylinidae with such a characteristic habitat like ant nests are reflected in various degrees of specialization. In the context of host-guest interactions, Wasmann (1894) classified arthropods inhabiting anthills into ectoparasites, endoparasites, trophobionts, synechtry, synoics and symphiles. Myrmecophilous aleocharines can be placed in these last three categories. Lomechusa pubicollis is a representative of the symphilous Aleocharinae. Like all symphiles, this rove beetle is the most highly integrated with its host; evidence for this can be found inter alia in a number of adaptive morphological features in its larval stage, recently described by Staniec et al. (2017). In turn, both Amidobia talpa and Oxypoda haemorrhoa are probably synoics, that is, myrmecophiles feeding on debris or other organisms inhabiting the anthill; being small and highly mobile, they are probably overlooked by their hosts. On the other hand, both these beetles will readily feed on ant larvae, thus exhibiting typical features of synechtry (Parmentier et al. 2015, own observations). The myrmecophilous Thiasophila angulata (Erichson W.F., 1837) has a similar lifestyle ), but the morphology of its larva is no different from that of the larvae of other, related but non-myrmecophilous aleocharine beetles (Zagaja et al. 2014).
It is apposite, therefore, to pose the following questions: 1) Does the myrmecophily of A. talpa and O. haemorrhoa have any effect on the external structure of their larvae? 2) Are their ecological preferences of no great importance in this respect, as in the case of T. angulata? 3) How does the extent of guest-host integration affect the morphology of aleocharine larvae so far examined?
The chief aim of this paper is therefore to describe in detail the morphology, including the chaetotaxies and external ultrastructure, of the larval stages of Amidobia talpa and Oxypoda haemorrhoa and to compare them with the external larval structures of other, well-known myrmecophilous aleocharines.

Material examined
Larval stages were obtained by rearing 34 adults of Amidobia talpa and 9 adults of Oxypoda haemorrhoa. Specimens of both species were collected on May 5, 2017, at Lake Moszne (51°26'57.4"N, 23°07'34.0"E) and Lake Długie (51°27'04.0"N 23°09'39.9"E), situated in the Polesie National Park near Lublin (SE Poland). The insects were sifted from the nest material of Formica polyctena. Live beetles of A. talpa and O. haemorrhoa were placed in transparent plastic containers (diameter 10 cm, height 4 cm) filled with nest substrate and observed in the laboratory from May 9 to June 24, and from May 11 to June 21, respectively, at room temperature (22-25 °C). Adults and larvae of various species of ants, including F. rufa and small springtails, were supplied as a source of adult food.

Study techniques
Larvae of both species were killed in boiling water and preserved in ethanol (75%).
To prepare temporary microscope slides, some larvae were macerated in cold 10% KOH for two to three hours, immersed in lactic acid for subsequent preparation and mounting of antennae, mouthparts, sensory structures, chaetotaxy of the body, legs and urogomphi. They were then traced from photos taken with an Olympus DP72 or Olympus DP21 digital camera mounted on a binocular Olympus SZX16 or Olympus BX63 compound microscope (Figs 9,13,14,17,18,48,49,52,53,53a,54,54a,55,56). The final image adjustments were made using CorelDraw Graphics Suite 2018.

Measurements and their abbreviations
Measurements of the larvae of both species, made using an Olympus BX63 compound microscope in cellSens Dimension v1.9 software, are given in millimetres, as explained in detail in Pietrykowska-Tudruj and Staniec (2012). Measurements (Table 1) were made on freshly killed specimens. The terms of morphological structures, chaetotaxy (selected aspects only) and their abbreviations generally follow Ashe and Watrous (1984) and Staniec et al. (2018), with modifications in some of the figures. The material examined for the measurements is listed in Table 1. The material examined for morphological descriptions includes four or five specimens of the late-instar larva of each species. The voucher specimens are deposited in the collections of the Department of Zoology, Maria Curie Skłodowska University, Lublin.

Ecological preferences
Myrmecophilous species

Discussion and summary
This paper gives a detailed description of the external structure of the hitherto unknown larval stage of Amidobia talpa and Oxypoda haemorrhoa -Palearctic, myrmecophilous staphylinids belonging to the subfamily Aleocharinae -which are associated with the Formica rufa species group. It also gives the first description of the larva of Amidobia, and at present, the only complete, detailed account of the larval morphology of Oxypoda. The existing fragmentary descriptions of Oxypoda larvae, with only a few schematic drawings relating to just two species -O. spectabilis and O. longipes -were written 40-50 years ago (Pototskaya 1967, Topp 1978; in the context of contemporary comparative studies they are therefore practically useless. The diagnostic features presented above -13 for Amidobia and 10 for Oxypoda -are based mainly on this description of the late larval instars of A. talpa and O. haemorrhoa. They mainly involve the structural details of the mouthparts, and in the case of the former taxon, of the antennae and spiracles as well. These features were established with reference to well-known larvae of other aleocharine species of diverse ecological preferences (Ashe and Watrous 1984, Staniec et al. 2009, 2010, 2018, Zagaja et al. 2014), among which anthill symbionts, represented by the two titular staphylinids, are deserving of particular scrutiny.
Relationships between myrmecophiles and their hosts exhibit varying degrees of advancement (Wasmann 1894, Hölldobler 1971, Hölldobler and Wilson 1990, Stoeffler et al. 2011, which may be correlated with certain morphological adaptations in actively living developmental forms, including larvae. The data listed in Table 2 summarize the current state of our studies, which focus on the external structure of the larval stages of European myrmecophilous aleocharine species, especially in relation to the differing degrees of integration with their hosts (Stoeffler et al. 2011, Staniec et al. 2009, 2018, Zagaja et al. 2014. Five of the six species (genera) are known ant symbionts, associated with two species of these hymenopterans. Only Dinaraea aequata Erichson W.F. 1837 is a typical saproxylic (non-myrmecophilous) beetle, totally unconnected with these social insects. In this configuration, therefore, it functions as a control species.
By far the largest number of characteristic features of the external structure, compared with other myrmecophilous and non-myrmecophilous aleocharines, were found in the larva of the symphilous genus Lomechusa Brisout de Barneville Ch.N.F., 1860 (Table 2). These beetles are the most highly integrated with their hosts, both behaviourally and morphologically (Hölldobler 1967, 1970, Parmentier et al. 2014, Parker 2016, and this also applies to the larval stages . The morphological adaptations of their larvae are associated with the specific conditions prevailing within the anthill: absence of ocelli, a white body, and the close and continuous interaction in the host-guest system (e.g. absence of urogomphi; dense, asymmetrical chaetotaxy; membranous cuticle; short legs; some elements of mouthparts shortened). In addition, this distinctive structure of the larval stage of Lomechusa is accompanied by a passive lifestyle, possible trophallaxis and chemical mimicry.
The classification of the degree of integration of the other four myrmecophilous species of Aleocharinae is not so obvious. Nonetheless, there do seem to be certain differences between them in this respect. Stoeffler et al. (2011), referring to Hölldobler (1981, suggests that in Pella laticollis there is behavioural pre-adaptation towards a closer relationship with the host. Evidence for this could be the presence in adult beetles, in contrast to the other representatives of this genus, of glands modifying the behaviour of the ants, which enables the beetles to live unmolested in the near neighbourhood of the anthill. Again, on the basis of existing classifications (Wasmann 1894, Kistner 1979) and observations of different behavioural aspects of Thisophila angulata, Zagaja et al. (2017) place this species among host-integrated myrmecophiles. However, Stoeffler et al. (2011) demonstrated that the relations of T. angulata with ants resemble a pre-adaptation (initial phase) to a closer relationship with them, rather than complete integration, as in the case of P. laticollis. The relationships with hosts of the smallest of this group of beetles -Amidobia talpa and Oxypoda haemorrhoa -were not examined. It seems, however, that according to Wasmann's (1894) classification, they are probably synoics, that is, myrmecophiles feeding on detritus or other organisms inhabiting the anthill. Opportunistically, they may also consume eggs and small larvae of ants. Because they are small and highly mobile, these beetles are probably completely ignored by the host (Parmentier et al. 2015, own observations).
In view of the above it cannot be surprising that, with the exception of Lomechusa, discussed earlier, the other myrmecophilous larvae analysed here do not possess any outstanding features distinguishing them from non-myrmecophilous species (Table 2). Therefore, the morphological differences between the aleocharine larvae examined here are probably a reflection of the biotic and abiotic conditions specific to the particular microhabitat they occupy rather than of more general habitat preferences (myrmecophily or non-myrmecophily). The crucial aspect of this situation thus appears to be the trophic specialization of these tiny predators, a question as yet incompletely understood. That is why the greatest number of differences between them concerns the structural details of the mouthparts (e.g. shape of labrum and ligula, structure of epipharynx, mandibles, mala, length of articles of maxillary and labial palps), this differentiation being strictly linked with the food resources these larvae consume (Table 2). Other characteristic features of these larvae include the detailed structure of the antennae, the urogomphi, less often the structure of the spiracles (A. talpa) and head shape (O. haemorrhoa). The present morphological analysis has not revealed any features characteristic of the several tribes (Table 2). This might indicate, on the one hand, the need to reassess the systematics of the higher taxonomic units in Aleocharinae, but on the other, that the larval characteristics of these rove beetles are of minimal usefulness in phylogenetic studies.
Therefore, as studies to date have shown, the characteristic morphology of the aleocharine larvae examined to date is not due to their myrmecophily alone. Likewise, the larval stages of myrmecophiles, which exhibit behavioural pre-adaptations to integration with host ants (P. laticollis, T. angulata), do not possess any visible external structural features pointing to associations with ants (Staniec et al. 2009, Zagaja et al. 2014. By contrast, symphiles, i.e. aleocharine species wholly integrated with their hosts and obligatorily dependent on them (e.g. Lomechusa), do exhibit a far-reaching restructuring of the body, particularly that of the larva. The unique morphological features of larvae (Table 2) are the result of advanced adaptations to life in an anthill and to constant interactions with their inhabitants (Hölldobler 1967, Parker 2016. In this context, the structure of the newly-described larvae of Amidobia and Oxypoda is typical of tiny, predacious Aleocharinae, not associated with ants (Table 2). In all probability, because they are highly mobile and very small (max. lengths up to 2.6 and 3.3 mm respectively), they are, like the adult forms, entirely ignored by the worker ants. They can thus live unmolested among ants without the need to possess the morphological adaptations that have evolved in the larger and slower symphiles. Similar, co-existential strategies in other small insect species associated with ants were described by Parker (2016) and Parmentier et al. (2015).
This analysis of the comparative morphology of known myrmecophilous aleocharine larvae in the context of the type of interaction with hosts is merely a preamble to far more extensive research on this subject. Unfortunately, as knowledge of the larval stage, not only of myrmecophilous but of other members of this very numerous staphylinid subfamily, remains fragmentary, the formulation of more comprehensive generalizations is as yet not possible. Moreover, there is still no information whatsoever on the detailed external larval structure of a number of other interesting, symbiotic European aleocharines. This situation can be illustrated by the genus Dinarda Leach W.E., 1819. Its members exhibit behaviour testifying to quite an advanced degree of integration with hosts, including the possibility of their being fed by ants on the principle of regurgitation (Hölldobler and Wilson 1990). It may well be that the mouthparts of Dinarda larvae are adapted to this form of feeding in the same way as in Lomechusa ) and that they possess other features emerging from their close relationships with ants. Where the degree of advancement of integration with hosts is concerned, such features would help to place Dinarda right after members of Lomechusa, and certainly in front of the other myrmecophiles listed in Table 2. But further studies are needed in order to find a definitive answer to this question.