Parasites of parasites of bats: Laboulbeniales (Fungi: Ascomycota) on bat flies (Diptera: Nycteribiidae) in central Europe

Background Bat flies (Streblidae and Nycteribiidae) are among the most specialized families of the order Diptera. Members of these two related families have an obligate ectoparasitic lifestyle on bats, and they are known disease vectors for their hosts. However, bat flies have their own ectoparasites: fungi of the order Laboulbeniales. In Europe, members of the Nycteribiidae are parasitized by four species belonging to the genus Arthrorhynchus. We carried out a systematic survey of the distribution and fungus-bat fly associations of the genus in central Europe (Hungary, Romania). Results We encountered the bat fly Nycteribia pedicularia and the fungus Arthrorhynchus eucampsipodae as new country records for Hungary. The following bat-bat fly associations are for the first time reported: Nycteribia kolenatii on Miniopterus schreibersii, Myotis blythii, Myotis capaccinii and Rhinolophus ferrumequinum; Penicillidia conspicua on Myotis daubentonii; and Phthiridium biarticulatum on Myotis capaccinii. Laboulbeniales infections were found on 45 of 1,494 screened bat flies (3.0%). We report two fungal species: Arthrorhynchus eucampsipodae on Nycteribia schmidlii, and A. nycteribiae on N. schmidlii, Penicillidia conspicua, and P. dufourii. Penicillidia conspicua was infected with Laboulbeniales most frequently (25%, n = 152), followed by N. schmidlii (3.1%, n = 159) and P. dufourii (2.0%, n = 102). Laboulbeniales seem to prefer female bat fly hosts to males. We think this might be due to a combination of factors: female bat flies have a longer life span, while during pregnancy female bat flies are significantly larger than males and accumulate an excess of fat reserves. Finally, ribosomal DNA sequences for A. nycteribiae are presented. Conclusions We screened ectoparasitic bat flies from Hungary and Romania for the presence of ectoparasitic Laboulbeniales fungi. Arthrorhynchus eucampsipodae and A. nycteribiae were found on three species of bat flies. This study extends geographical and host ranges of both bat flies and Laboulbeniales fungi. The sequence data generated in this work contribute to molecular phylogenetic studies of the order Laboulbeniales. Our survey shows a complex network of bats, bat flies and Laboulbeniales fungi, of which the hyperparasitic fungi are rare and species-poor. Their host insects, on the other hand, are relatively abundant and diverse. Electronic supplementary material The online version of this article (doi:10.1186/s13071-017-2022-y) contains supplementary material, which is available to authorized users.


Bats and bat flies
Bats (Mammalia: Chiroptera) contain more than 1,300 described species and are the second-most diverse group of living mammals, after rodents [1]. The lineage has evolved numerous adaptations that uniquely and effectively expand their ecological amplitude. These include flight, echolocation and a generally nocturnal lifestyle [2]. Moreover, bats are capable of utilizing a wide variety of food sources including insects, other vertebrates, blood, fruit and nectar. Bats vary greatly in their sociality ranging from solitary to the largest congregations of mammals numbering tens of millions of individuals. Finally, they vary in the roosts they use during day and night, from large and permanent structures such as caves and mines, to intermediate structures such as buildings and hollow trees, to ephemeral structures such as leaf tents and plant foliage [3]. Owing in part to these unique adaptations, bats are also parasitized by a plethora of arthropod lineages, including mites, ticks, bugs, fleas and flies. Among these parasites, the flies (Diptera: Hippoboscoidea: Streblidae and Nycteribiidae) are among the most conspicuous. Commonly known as "bat flies", these specialized insects are known only from bats where they live in the fur and on the flight membranes and feed on blood [4]. Bat flies currently are divided into two families. The family Streblidae contains about 230 described species, which are cosmopolitan but generally richer in the Western Hemisphere. The family Nycteribiidae contains about 275 species, also occurring worldwide but generally more speciose in the Eastern Hemisphere [5]. Both families are considered tropical or subtropical in distribution, though certain species occur at northern or southern latitudes.
The flies are "semi-permanent" parasites with adaptations that help keep them associated with their bat hosts. Adult females rear three larval stages internally, nourished by milk glands, and adults of both sexes generally do not often leave the host bats. One exception are gravid females, who temporarily leave their host to deposit the terminal third-instar larvae on the roost substrate. Here, the larva soon forms a puparium and at approximately three weeks development the new adult emerges to seek a host bat in order to feed and find mates [4]. However, some studies [6] suggest that the duration of European bat fly pupal development is dependent in part upon the presence of bats in the roost. Thus, although larval stages are telescoped and tied to the host via the female fly, there is necessarily a period of off-host development that presents opportunities for flies to interact with other fly and host species inside the roost, and with the roosting environment itself [7]. Despite opportunity to colonize multiple host species, the emerging consensus is that bat flies are quite specific to individual host species, at least in ecological time. Although some experimental work has evaluated host choice and specificity [8,9], much of our understanding of host specificity comes from large surveys of bats and parasites. When care is taken in the handling of bats and bat flies in the field, a remarkably high degree of host specificity is noted [10]. In particular, bat social structure, the roost environment, and species fidelity to roosting structures play key roles in parasite dynamics [11,12]. Bat species that live in large groups and roost in large and long-lived structures that they return to with high fidelity increases parasitism generally, which in turn should raise transmission frequency and opportunities for selection toward parasitism in organisms associated with bats and with bat parasites.

Laboulbeniales
Laboulbeniales (Fungi: Ascomycota) are microscopic parasites of myriad arthropod hosts. They grow onto the integument of three subphyla within Arthropoda: Chelicerata (infecting only mites), Myriapoda (millipedes), and Hexapoda (insects) [13]. Unlike other multicellular fungi, Laboulbeniales do not grow mycelia or hyphae. Instead they form a thallus (plural: thalli), which is derived from the enlargement and subsequent divisions of a single twocelled ascospore. A thallus typically consists of three major parts: a multicellular receptacle, which attaches to the host through the (often blackened-melanized) foot or rootshaped haustorium that may penetrate the body of the host; a single or multiple perithecia, the spore-forming structure(s); and appendages bearing antheridia, which produce spermatia.
While about 80% of the currently described species of Laboulbeniales are found on Coleoptera, only 10% parasitize Diptera [14]. Laboulbeniales parasitizing flies belong to eight genera: Arthrorhynchus Kolenati, Dimeromyces Thaxt., Gloeandromyces Thaxt., Ilytheomyces Thaxt., Laboulbenia Mont. & C.P. Robin, Nycteromyces Thaxt., Rhizomyces Thaxt. and Stigmatomyces H. Karst. The genus Laboulbenia is by far the largest genus with close to 600 recognized species with a wide array of hosts, and only 24 of those are known from flies [15]. Stigmatomyces is the second-largest genus in the order, with 144 described species, all of which occur on flies [16]. The genera Arthrorhynchus, Gloeandromyces and Nycteromyces are specific to bat flies, while all other genera have never been recorded from these two host families.
Arthrorhynchus is apparently restricted to Eastern Hemisphere species of the Nycteribiidae, which are also most diverse in the Eastern Hemisphere. Thaxter [17] recognized three species: A. cyclopodiae Thaxt., A. eucampsipodae Thaxt. and A. nycteribiae (Peyr.) Thaxt. Two additional species, A. diesingii and A. westrumbii, were described in the 1850s [18], but were soon synonymized with Arthrorhynchus nycteribiae (as Helminthophana) [19]. Merola [20] described a fourth species, A. acrandros. This species, however, is very similar to A. nycteribidae and likely represents the same species. New collections are needed before making a final decision about synonymy because Merola's original material is likely lost (W. Rossi personal communication). Both species A. acrandros (Italy) and A. cyclopodiae (Papua New Guinea) are only known from the type collection.
A comprehensive study was conducted by Blackwell [21] to screen nycteribiid bat flies for presence of Arthrorhynchus spp. She screened 2,517 individuals, of which 56 were infected by A. eucampsipodae or A. nycteribiae, denoting an infection prevalence of 2.2% [21].
The diversity of the other two bat fly infecting genera is restricted, as is knowledge of their distribution and biology. Two species of Gloeandromyces are known: G. nycteribiidarum (Thaxt.) Thaxt. and G. streblae Thaxt. Both species have been reported only once [22], although several aspects of the biology of their hosts have recently been the focus of different studies. Gloeandromyces nycteribiidarum was described on Megistopoda aranea (Coquillett, 1899) (as Pterellipsis aranea) (Streblidae) from Grenada, G. streblae on Strebla wiedemanni Kolenati, 1856 (as S. vespertilionis) (Streblidae) from Venezuela. Nycteromyces is monotypic. Its single species, N. streblidinus Thaxt., was described on Strebla wiedemanni (as S. vespertilionis) from Venezuela [23] and has not been reported since. Interestingly, G. streblae and N. streblidinus were described from the same bat fly specimen ("No. 2073/a," deposited at the Harvard Museum of Comparative Zoology). Both Gloeandromyces and Nycteromyces seem restricted in distribution to the Western Hemisphere, where they are associated with host species of the Western Hemisphere clade of the paraphyletic Streblidae [24]. The currently known distributional records, fungus-bat fly and fungus-bat associations are summarized in Fig. 1 and Additional file 1: Table S1. From the two countries considered in this paper, only Arthrorhynchus nycteribiae (Peyr.) Thaxt. is  Table S1 reported, from the nycteribiid Penicillidia conspicua Speiser, 1901. Recently, studies on the distribution, population dynamics, host-choice and host-parasite interactions of different Laboulbeniales enabled assessment of the biology of this peculiar fungal group [25][26][27]. Such studies have successfully focused on locally abundant host insects forming colonies or aggregates (ants and ladybirds). Since bat flies are easily collected during bat surveys and are restricted to microhabitats where they may reach high population densities, they are interesting research targets for similar Laboulbeniales-oriented studies. With this collaborative project we aim to contribute to the research of parasitism of bat flies by Laboulbeniales fungi. The current scantiness of material precludes hypotheses about parasitism, host specificity, etc. As for five species of Laboulbeniales from bat flies only the type collection exists -A. acrandros, A. cyclopodiae, G. nycteribiidarum, G. streblae and N. streblidinusour first effort has been to collect bat flies and screen these and existing bat fly collections for the presence of Laboulbeniales. Despite having many features that contribute to the difficulties in studying the Laboulbeniales [28], they do remain intact on dead host individuals. This way, historical collections of hosts can be used to record parasite prevalence, host specificity and population dynamics through time [13].
Here we extend geographical and host ranges, and discuss host associations. Our results also illustrate how even decade-old insect collections (with detailed collection data) can be used to uncover new host-parasite networks and the underlying factors.

Sampling sites
Bat flies were collected during bat surveys in the Romanian Carpathians and the Dobrogean Plateau and in various, mainly mountainous parts of Hungary (Transdanubian Mountains, Mecsek Mountains, North Hungarian Mountains) (Fig. 2). Study areas included roosting sites localized in caves and mine galleries. Dates for capturing bats ranged from 1998 to 2015.

Capture of bats and collection of bat flies
Bats were captured mostly close to roosting sites using mist nets or harp traps located at drinking, foraging and swarming sites. All bats were identified [29] and sex and age was determined for each specimen (whenever possible). Ectoparasitic bat flies were removed using forceps. The only exception to this method is the bat flies from Hungarian bats captured in 1998-1999, which were collected with the help of a Fair Isle Apparatus. The Fair Isle Apparatus method was originally developed for collecting ectoparasites of birds [30]. Bats are placed in a small plastic box but keeping their head outside while ethyl acetate is used as fumigant inside the box. After parasite removal, bats were immediately released at the place of capture.
Preservation and long-term storage of bat flies was in 70 or 87% ethanol in separate vials (one vial per bat host). A small number of the Romanian flies were collected from corpses of bats found in caves or early emerging ones (Myotis daubentonii, n =12, Cheile Turzii, Romania). Identification of bat flies was based on Theodor [31] and Theodor & Moscona [32]. Voucher specimens are deposited at Eszterházy Károly University (Hungary) under accession nos. 12EP01 through 12EP95, 14EP01 through 14EP21, 15EP01 through 15EP07 and P001 through P178. In this study, we collected all ectoparasites from each individual bat to allow prevalence data to be calculated.

Identification of Laboulbeniales
Bat flies were screened for ectoparasitic fungi using a binocular microscope. Position and density of thalli were detailed and thalli were removed with an entomological pin and slide-mounted for identification [13]. Identification was based on the descriptions and drawings in Thaxter [33]. Voucher slides are deposited at BP (Botanical Department, Hungarian Natural History Museum; nos. 107915 & 107916) and FH (Farlow Herbarium, Harvard University; no. D. Haelew. 1015c) (details in Additional file 2: Table S2).

DNA extraction from Laboulbeniales, PCR and sequencing
Isolation of Laboulbeniales DNA was carried out using the Extract-N-Amp Plant PCR Kit (Sigma-Aldrich, St Louis, MO, USA) and the heat extraction protocol, as outlined in a previous study [28]. PCR amplification of small subunit (SSU) and large subunit (LSU) ribosomal DNA (rDNA) was performed using the following primer combinations: SL122/NSL2 [28,34] and NS1/NS4 [35] for SSU, LIC24R/LR3 [36,37] and LR0R/LR5 [37] for LSU. PCR conditions were: denaturation at 94°C for 3 min, followed by 35 cycles at 94°C for 1 min, 50°C for 45 s, and 72°C for 1:30 min, and a final extension step of 72°C for 10 min. In the case of unsuccessful PCR, a modified touchdown protocol was applied: initial denaturation at 95°C for 10 min, followed by 30 cycles at 95°C for 1 min, 62°C for 1 min (decreasing 1°C every 3 cycles), and 72°C for 1:30 min; then 30 cycles with denaturation at 95°C for 30 s, 55°C for 30 s, and 72°C for 1 min, and a final extension step of 72°C for 7 min. PCR amplification of the internal transcribed spacer (ITS) rDNA was attempted, but without success. PCR products were purified using QIAquick PCR purification kit (Qiagen, Stanford, CA, USA), then sequenced as described in Haelewaters et al. [28]. Generated DNA sequences were assembled and edited in Sequencher 4.10.1 (Gene Codes Corporation, Ann Arbor, MI) and blasted in NCBI GenBank (http://ncbi.nlm.nih.gov/blast/Blast.cgi).

Statistical analysis
Laboulbeniales infection fractions for Penicillidia conspicua, P. dufourii and Nycteribia schmidlii were statistically analyzed using generalized linear mixed models (GLMM), using the R package lme4 [40]. We took the binomial distribution with logit link function for the binary infection score per bat fly (aggregating the binary scores per bat gave no indication of overdispersion). The GLMM contained fixed effects for the three bat fly species, the sex of the bat fly, and their interaction, and random effects for the location and year of data collection. First, the sex effect on infection fraction was studied per bat fly species, and, if allowed, the sex effects were aggregated over the three species.
Hypothesis testing was done using likelihood ratio tests, with P-values calculated based on χ 2 distributions, declaring an effect significant when P < 0.05. Outcomes of the test statistics are reported as X df 2 with df = the number of parameters tested simultaneously.

Visualization of the host-parasite-parasite network
The bat-bat fly-Laboulbeniales associations were visualized with the help of the R package bipartite [41]. We used weighted data and the function plotweb to build a network showing host-dependencies and prevalence.

Bats and bat flies
We sampled 1,594 bats of 28 species. Of these, 997 bats were captured in Hungary (24 species), of which 361 carried bat flies (parasite prevalence of 36%). In Romania, 597 bats were captured (10 species), of which 186 had bat flies (31%). Taken together, 547 bats were infected with bat flies, accounting for an overall prevalence of 34%. Details are presented in Table 1. Infected bats were found with different numbers of bat flies. The highest number of bat flies harvested from a single bat was 21. This was Myotis daubentonii with only Nycteribia kolenatii bat flies from Hungary (sample P161). Overall, M. daubentonii (220 infected bats) carried the highest number of bat flies, with 22 specimens having ≥ 10 bat flies each. In comparison, also many Miniopterus schreibersii bats were infected by bat flies (179), but only one carried ≥ 10 bat flies. Altogether 270 different bat specimens had a single bat fly. Numbers of infected bats are given in Table 2, along with minimum and maximum number of bat flies collected per bat. Of the 28 sampled bat species, 13 were not found to host bat flies (46%). Collection data of all surveyed 1,494 bat fly specimens are listed in Additional file 2: Table S2.
Among the bat fly specimens collected in this study, a new country record for Hungary was found, Nycteribia pedicularia Latreille, 1805, collected from Myotis daubentonii (20 specimens) and Myotis myotis (2 specimens). This fly species was recovered from the sampling localities in Western Hungary (Isztimér, Komló, Mánfa, Olaszfalu, Őcsény and Pogány).
Although in all three bat fly species the female infection fractions were higher, in none of them a significant difference between infection fractions of female and male bat flies was found (GLMM: P. conspicua: X 1 2 = 3.19, P = 0.074; P. dufourii: X 1 2 = 2.52, P = 0.113; N. schmidlii:   The interaction between bat fly species and sex was not significant (GLMM: X 2 2 = 1.35, P = 0.510). We therefore removed this interaction from the model. In the resulting GLMM we found a significant difference in infection rate between female and male bat flies (averaged over species; X 1 2 = 4.56, P = 0.0327, higher in females), and highly significant differences in infection rates among bat fly species (averaged over sexes; X 2 2 = 52.83, P < 0.0001). No extra variation due to year of data collection was found, but variation due to location of data collection was considerable X 1 2 = 12.12 (P = 0.00025, obtained by halving the P-value from the χ 1 2 distribution). Infected P. conspicua bat flies were sampled from three different bat hosts: Myotis daubentoni (4 infected bat flies, n = 7); Miniopterus schreibersii (33, n = 142); and Rhinolophus euryale (1, n = 1). Infected P. dufourii bat flies were found only on the bat Myotis myotis and infected N. schmidlii only on the host Miniopterus schreibersii. A detailed summary, including collecting data, is presented in Additional file 3: Table S3.
Host-parasite-parasite network Figure 6 shows the association of bat flies with their bat host as well as the association of Laboulbeniales and their arthropod hosts. The principal bat host for bat flies infected with Laboulbeniales species was M. schreibersii for P. conspicua and N. schmidlii. Penicillidia dufourii was found mainly on M. myotis but also occurred commonly on M. blythii and M. schreibersii.

Distribution and associations of bat flies and their bat hosts
Bat flies generally have one or two preferred bat hosts upon which they are collected with higher probability compared to other hosts [31,42]. Our dataset supports these main hosts for the bat flies with a few notable deviations. In the case of Basilia nana we found most flies on Myotis bechsteinii (n = 41), one of the two major hosts, and fewer individuals from the other host, M. nattereri (n = 4). Phthiridium biarticulatum has two major hosts, and we mostly found them on Rhinolophus ferrumequinum (n = 26), while the other host species, R. hipposideros, is practically missing from our dataset. The new country record for Hungary, Nycteribia pedicularia, has Myotis capaccinii as its main bat host species, but we found it on M. daubentonii (n = 20) and M. myotis (n = 2).
The new host associations in case of the bat flies Nycteribia kolenatii, Penicillidia conspicua and Phthiridium biarticulatum only extend the list of known hosts, since our collecting records support the existing major host species.

Morphological identification of Arthrorhynchus
In our circumscription of the genus, three species are accepted: A. cyclopodiae, A. eucampsipodae and A. nycteribiae. The three species can be distinguished based on two morphological characteristics: cell III and the perithecial tip (Fig. 4). Arthrorhynchus nycteribiae is easily separated from the other two species by the elongated cell III, which carries the appendage ( [33]: Plate XLVIII, 7-9). In addition, its perithecial tip is four-lobed, with each of the four lobes conspicuously three-lobed ( [33]: Plate XLVIII, 10). The perithecial tip of A. eucampsipodae is very similar to that of A. nycteribiae but the four lobes are not individually three-lobed. The perithecial tip in A. cyclopodiae also ends in four lobes but these are not short and broad as in A. eucampsipodae and A. nycteribiae, but erect. Arthrorhynchus cyclopodiae and A. eucampsipodae have similar appendage structures, however there are some differences. In A. eucampsipodae, cell III is constricted in the middle, the upper part of it being conspicuously smaller compared to the lower half; it is also narrower than the basal cell of the appendage ( [33]: Plate XLVIII, 14). Cell III of A. cyclopodiae is also constricted in the middle but the upper half is inflated and mostly broader than the base of the basal cell of the appendage ( [33]: Plate XLVIII, 3). Variation in the thallus morphology of A. eucampsipodae was noted by Thaxter [33], who mentioned that its perithecia may be straight or slightly curved distally.
Thalli from the locality Szársomlyó (Fig. 4b) showed straight perithecia, although the other half, those collected in Felsőtárkány (Fig. 4a) showed markedly curved (and somewhat longer) perithecia. Both populations were identical in the cell III and the appendage structure.

DNA sequence data of bat fly-associated Laboulbeniales
Morphological identification may be insufficient to assess the diversity of Laboulbeniales on (temperate) bat flies. There is the phenomenon of position specificity in Laboulbeniales, in which multiple morphotypes of the same species occupy limited portions of the host's integument [43]. More generally, many fungal species are cryptic, with lineages "hidden" within morphological species complexes. This cryptic diversity is rapidly being uncovered by applying a molecular phylogenetic approach to taxonomic studies [44][45][46]. In Laboulbeniales, cryptic diversity is a subject of current research, although advances in this field have been hindered by the difficulties in working with and extracting DNA from these fungi [28]. Preservation of material is of utmost importance in successfully extracting DNA from Laboulbeniales: freshly collected host specimens preferably stored in ≥ 95% ethanol give the best results [28]. To fully understand and discuss the diversity of Laboulbeniales on specific host groups, we need to generate DNA sequences of as many species as we encounter, to be able to add a phylogenetic component to morphological data. For this study, we generated the first DNA sequences of A. nycteribiae. We generated sequences for two isolates. These isolates were made from bat flies from different bat hosts collected in different locations. Additionally, we applied two distinct DNA isolation methods in two of the collaborating laboratories. The SSU sequences match for 99% (over 522 bp) and the LSU sequences are 100% alike (over 450 bp). This confirms molecular identity of our isolates. Blasting the SSU rDNA sequence against several species in the Stigmatomycetinae shows that Hesperomyces and Prolixandromyces are more closely related to Arthrorhynchus, compared to Corethromyces, Rhadinomyces and Stigmatomyces. No significant similarity was found with C. bicolor.
Traditionally, the Laboulbeniales (as well as the class Laboulbeniomycetes) are excluded from major phylogenies of the Ascomycota. This is due to a lack of sequence data and the difficulty of working with these fungi . Currently, no broadly inclusive phylogeny of the order is available because of the lack of adequate sampling. However, preliminary data suggest that the Stigmatomycetinae are not a monophyletic group (D. Haelewaters, unpublished data). Generating more DNA sequences and increasing taxon sampling will be important to the resolution of relationships within the Laboulbeniales and to better understand how species have evolved and developed some of their unique traits (such as the formation of a thallus). As to bat fly-associated Laboulbeniales, determining the position of the genera Arthrorhynchus, Gloeandromyces and Nycteromyces within the phylogeny of the order may reveal that they form a single clade or, on the contrary, that multiple colonizations on bat flies from other host groups have happened with subsequent diversification sensu De Kesel & Haelewaters [47].
The number of thalli of A. eucampsipodae was insufficient for molecular work. We will continue to sample fresh bat flies and screen for Arthrorhynchus spp. from different populations of different host species. We hope this will lead to more sequences. It will be a difficult endeavor to recollect A. cyclopodiae, since it is currently only known from the type locality. However, comparing isolates of A. nycteribiae taken from different populations of the same host species and from different host species, will give a good estimate of the diversity of Laboulbeniales in this temperate system.

Diversity of bat fly-associated Laboulbeniales in central Europe
Laboulbeniales exhibit three types of specialization: high specificity to host species (host specificity), growth restricted to certain areas of the host body (position specificity), and speciation resulting from co-habiting hosts (ecological specificity) [43,[47][48][49].
Arthrorhynchus eucampsipodae is known from species of Basilia, Cyclopodia, Eucampsipoda, Nycteribia and Penicillidia bat flies (Additional file 1: Table S1). Arthrorhynchus nycteribiae has been found on species of Nycteribia, Penicillidia and Phthiridium. In our study, we found only Nycteribia schmidlii, Penicillidia conspicua and P. dufourii infected by Arthrorhynchus eucampsipodae and A. nycteribiae. These three species, together with Phthiridium biarticulatum, are most commonly encountered with Laboulbeniales [21]. Except for being seemingly restricted to Eastern Hemisphere species of Nycteribiidae, there is no strict host specificity, as previously reported [21].
Nycteribia schmidlii was host for both A. eucampsipodae and A. nycteribiae in our study. No double infection was found and both fungi were sampled from bat flies from different populations of M. schreibersii in Hungary (Additional file 2: Table S2 Do bat flies sharing the same bat host carry the same species of Laboulbeniales? If they do share hosts, this would represent an example of ecological specificity. In this situation, the bat skin/fur itself acts as a microhabitat. Also caves, ant nests, and fragmented habitats in salt marshes can be cohabited by multiple, often unrelated hosts [47,49,51,52]. In our dataset, we found three bat fly hosts for A. nycteribiae. These are Penicillidia conspicua, P. dufourii and Nycteribia schmidlii. The fungus was present on two specimens of P. dufourii and on a single specimen of N. schmidlii, while it was associated with 37 specimens of P. conspicua. It seems that P. dufourii and N. schmidlii are "accidental hosts" and P. conspicua the "main host". Penicillidia dufourii uses Myotis myotis and M. blythii but it is also found on Miniopterus schreibersii (20% in our material). This is probably due to the roost sharing habits of these three bat species; they frequently form mixed colonies in caves [53]. Miniopterus schreibersii is the main bat host for N. schmidlii and P. conspicua. Furthermore, P. conspicua and P. dufourii sometimes choose non-primary hosts if the opportunity arises [54]. Four P. conspicua bat flies and one specimen of N. schmidlii from the same Miniopterus schreibersii bat were found infected with A. nycteribiae. Given the record of A. nycteribiae on N. schmidlii is the only one known in the literature (also not mentioned in [21]), we can safely assume that this infection was the result of an accidental transmission. Indeed, in this occasion the bat host served as the microhabitat allowing transmission of ascospores between hosts. In some cases, shifting between co-occurring hosts can lead to adaptation and eventually to speciation [47]. However, "successful colonization of a new host is probably a rare event" [51] as seems the case with Eastern Hemisphere bat flies.
Our dataset of~1,500 bat flies from 15 different bat host species, different habitat types, and geographic locations from Western Hungary to Eastern Romania allows us to speculate on the diversity of Laboulbeniales associated with bat flies in central Europe. We think that the potential for undiscovered species of Laboulbeniales on central European bat flies is very low, and expected only on rarely collected (in this study) bat fly species. Our study revealed no fungi other than A. eucampsipodae and A. nycteribiae after screening 914 Nycteribia kolenatii, 159 N. schmidlii, and 152 P. conspicua specimens. Also Blackwell [21], after screening 2,937 bat fly specimens, did not make any note of undescribed diversity of Arthrorhynchus.

Unbalanced sex ratios of infection
Generally, female and juvenile bats are more heavily infected by ectoparasites compared to males [3,[55][56][57][58][59][60]. Furthermore, pregnant individuals are more parasitized by bat flies than non-pregnant females [57]. Dick and Patterson [61] found significantly more male than female bat flies on Venezuelan bats (>36,500 bat flies included in the survey). This phenomenon was potentially explained by selective host grooming, which removes and/or kills the larger females.
For bat fly-associated Laboulbeniales other factors seem at play. First, female bat flies live longer than males and have an average life span of about 5-6 weeks [67]. Laboratory studies of Basilia hispida Theodor, 1967 bat flies (Nycteribiidae) revealed that males lived for at least 97 days, while females lived for at least 156 days [67]. It could be that the fungal parasites need this time for successful development, maturation, and build-up of inoculum. Second, during pregnancy, female bat flies are significantly larger than males-at extrusion, prepupae may comprise about one third of the body size of the female fly. Moreover, a descriptive modeling study on tsetse flies (Glossinidae: Glossina spp.), which belong to the same superfamily as bat flies (Hippoboscoidea), demonstrated that during each pregnancy females accumulate an excess of fat reserves [68], which are manifested as lobes in the haemolymph for maximal exposure [69,70]. Such reserves represent higher nutritional resources in female flies for parasites such as Laboulbeniales relative to males. A combination of these factors may lead to a greater infection prevalence on female bat flies, as we have observed.

Conclusions
Our knowledge about Laboulbeniales fungal ectoparasites of bat flies is poor. Seven species in three genera are recognized based on morphological descriptions, although one species is doubtful. Of those seven, four species are only known from the type collections, which are between 65 and 116 years old. For this paper, we collected bat flies from captured bats in central Europe (Hungary, Romania) and screened them for presence of Laboulbeniales. Our survey shows a complex network of bats, bat flies and Laboulbeniales. New bat-bat fly associations are reported: Nycteribia kolenatii on Miniopterus schreibersii, Myotis blythii, Myotis capaccinii and Rhinolophus ferrumequinum; Penicillidia conspicua on Myotis daubentonii; and Phthiridium biarticulatum on Myotis capaccinii. While the bat flies were relatively abundant and diverse (studied material), the Laboulbeniales associated with them were rare and species-poor. Laboulbeniales were found on 45 of 1,494 screened bat flies: Arthrorhynchus eucampsipodae was reported on Nycteribia schmidlii, and A. nycteribiae on N. schmidlii, Penicillidia conspicua and P. dufourii. Penicillidia conspicua was infected most often, followed by N. schmidlii and P. dufourii. The bat fly Nycteribia pedicularia and the fungus Arthrorhynchus eucampsipodae represent new country records for Hungary. We generated SSU and LSU ribosomal DNA sequences for A. nycteribiae. These are the first sequences for any species of bat fly-associated Laboulbeniales.

Additional files
Additional file 1: Table S1. Summary of all records of Laboulbeniales associated with bat flies thus far. (XLS 53 kb) Additional file 2: Table S2.