Studies of Laboulbeniales (Fungi, Ascomycota) on Myrmica ants: Rickia wasmannii in the Netherlands

An important group of fungal insect parasites is the Laboulbeniales (Ascomycota). Th ese are microscopic in size and live attached to the cuticle of their arthropod hosts. Rickia wasmannii is a common European species limited to the ant genus Myrmica (Hymenoptera, Formicidae). We present new records of R. was-mannii in the Netherlands on three host species: Myrmica ruginodis , M. sabuleti , and M. scabrinodis . Our data show a mass infection of M. sabuleti by R. wasmannii . Th e average parasite prevalence is 38% (n = 3,876). Th e prevalence was much lower on the other Myrmica species. So far, R. wasmannii infections have been found only on Myrmica species in the rubra -group and the scabrinodis -group. We provide possible explanations for this observation. To date, Rickia wasmannii is known on nine Myrmica species in sixteen European countries; an overview is included in tabulated form.


Introduction
The order Laboulbeniales (Ascomycota: Laboulbeniomycetes) consists of microscopic ectoparasites of Arthropoda, mostly true insects.Within the subphylum Hexapoda representatives of nine orders are known as hosts (Weir and Hammond 1997): Blattodea, Coleoptera, Dermaptera, Diptera, Hemiptera, Hymenoptera, Mallophaga, Orthoptera, and Thysanoptera.[Note that the termites, previously ranked in the order Isoptera, were recently included in the order Blattodea, based on phylogenetic data (Beccaloni and Eggleton 2013).]In addition, Acari (subphylum Cheliceriformes) and Diplopoda (subphylum Myriapoda) are known to host Laboulbeniales (Haelewaters et al. 2012, Weir andHammond 1997).Laboulbeniales are unusual in their determinate growth pattern and lack of hyphae.Diversity in the group is largely underexplored and many questions related to the taxonomy and biology of these fungi remain unresolved.
In the Netherlands, a single worker of Myrmica sabuleti Meinert, 1861 (as M. scabrinodis) infected with Rickia wasmannii is known (Haelewaters 2012).Its discovery has initiated the search for Laboulbeniales on ants in this country.For this research project, ants were collected with many pitfall traps and subsequently screened for the presence of R. wasmannii.

Methods
The study site spans somewhat less than 1 km 2 , situated east of Maastricht (Limburg, the Netherlands) near the border with Belgium.This area has a rolling landscape and calcareous soil.Many different habitats were sampled: calcareous grassland, thicket, moist forest, forest edge, felling area, agricultural field edge, and hollow road.
Pitfall traps were filled with a formol solution (3%) or with a saturated salt solution.A lid was placed above each trap to exclude rainfall.Pitfall trapping was performed during three inventories for which in total 424 pitfall traps were placed in 29 series: (i) a year-round inventory of the insect fauna with nine trap series (each with five pitfall traps) at a railroad verge with semi-natural grassland and thicket, (ii) a twomonth-long inventory with eleven trap series (each with five pitfall traps) of the arthropod fauna of a nature restoration parcel with a species-rich hay-meadow (dominated by Arrhenatheretum elatioris Braun-Blanq.),and (iii) a four-day-long inventory of nine different habitats (nine traps series, one series per habitat, each with 36 pitfall traps) during a student course.For an overview of the study area, see Figure 1.
Screening for thalli of Laboulbeniales was done at 45× magnification with a Euromex Z-1740H stereomicroscope (Arnhem, the Netherlands).Infected ants were sent to Harvard University for study of their associated fungi.Thalli were carefully removed from the host integument using a Minuten Pin (BioQuip #1208SA) and embedded in Amann solution (Benjamin 1971) or PVA Mounting Medium (BioQuip #6371A).Cover slips were ringed with transparent nail varnish.Morphological analyses, measurements, and identifications were done using an Olympus BX40 light microscope with Olympus XC50 digital camera and MicroSuite Special Edition software 3.1 (Soft Imaging Solutions GmbH).Voucher specimens are deposited at the Farlow Herbarium, Harvard University (FH).
Chi-square tests were performed to analyze whether infection rate of the infected ant species was different from a hypothesized even distribution of Rickia wasmannii presence over the infected species.In social insects, it sometimes seems better to analyze the number of infected nests instead of the number of infected workers.However, the infected Myrmica species are very abundant in our study area, are polygynous and occur in clusters of mini-populations.They are hardly territorial and contact between workers of different nests is likely (Garnas et al. 2007, Wilson 1971).Each pitfall trap (and trap series) thus likely samples ants from many nests.Our sampling effort with 424 pitfall traps is so widespread over the study area and combines workers from so many nests, that we performed our test on worker individuals.

Results
Twenty-seven ant species were recorded during this study (Table 1).Only three species in the genus Myrmica bore thalli of Rickia wasmannii.
Only workers were found infected; neither gynes nor males were found infected, not even in the highly infected Myrmica sabuleti.Myrmica ruginodis and M. scabrinodis are reported as hosts of Rickia wasmannii in the Netherlands for the first time.

Discussion
Rickia wasmannii is mentioned in the literature in the literature to occur on nine host species (Table 2), most of which are widespread across Europe (Radchenko and Elmes 2010).It is reported from sixteen European countries, and expected to occur in Belgium, Denmark, Ireland, and Portugal (Espadaler and Santamaría 2012, Haelewaters 2012, Witek et al. 2014).

Ant species with Rickia infection
Santamaría and Espadaler (2014) mention the "high host phylogenetic specificity" of Laboulbeniales, but caution is needed when interpreting support for this assertion.Although specific to the genus Myrmica, host species of Rickia wasmannii belong to two clades or so-called species groups that are not phylogenetically closely related: rubragroup and scabrinodis-group (Jansen et al. 2010).
The Palearctic species in the genus Myrmica are classified into 17 taxonomic species groups based on their morphology (Radchenko and Elmes 2010), three of which are represented in our study: the rubra-group, encompassing M. rubra and M. ruginodis; the scabrinodis-group, with M. rugulosa, M. sabuleti, and M. scabrinodis; and the schencki-group, to which (in most parts of Europe) only M. schencki belongs.The monophyly of these morphological species groups was confirmed using molecular data (Jansen et al. 2010).So far, only species of the rubra-group and scabrinodis-group have been found with Rickia wasmannii infection.
Our research confirms this finding: although six Myrmica species were screened (n = 6,314), Rickia wasmannii was only present on M. ruginodis, M. sabuleti, and M. scabrinodis (Table 1).Several researchers screened multiple species of Myrmica in Hungary/Romania (Tartally et al. 2007), Slovakia (Bezdĕčka and Bezdĕčkova 2011), and the Czech Republic (Bezdĕčkova and Bezdĕčka 2011), but found only species in the scabrinodis-group infected by R. wasmannii.In addition, our study in the Netherlands revealed a single lightly infested worker in the rubra-group (M.ruginodis).In all four studies, R. wasmannii was found only in the rubra-and scabrinodis-groups.
Inadequate sampling of potential hosts may (partly) explain this pattern, although recently in Europe several studies have been conducted on Laboulbeniales on Myrmica, including various studies in which different species groups were sampled and screened.To avoid taxon-sampling errors in these kinds of observation, we suggest that systematic collections in natural history museums be screened for Rickia wasmannii on Myrmica.Screening museum collections previously has yielded important contributions on patterns of host utilization (Weir and Hammond 1997: 80,000 insects screened) and the distribution of Hesperomyces virescens Thaxt.(Haelewaters et al. 2014: 4,000 ladybirds screened).
Host shift could be another explanation for the restricted presence of R. wasmannii on two species groups that are not sister clades.Host shifts have been suggested for morphologically similar Laboulbenia species between Cicindelinae and other Car- abidae living in the same habitat (Arndt et al. 2003, Rossi 2011), and De Kesel and Haelewaters (2014) provide morphological and ecological data to support the hypothesis that a species of Laboulbenia shifted between Carabidae and Staphylinidae.To explain the extremely small size of Rickia lenoirii on Messor spp.ants (Hymenoptera, Formicidae), Santamaría and Espadaler (2014) suggest that Laboulbeniales can shift between myrmecophilous Acari and their ant hosts.

Geographical variation
Parasite prevalence (= infection frequency) is often used to quantify differences in populations of Laboulbeniales in a given host community (De Kesel 2011, Haelewaters et al. 2012).It is beyond doubt that in our study area Myrmica sabuleti is the main host of Rickia wasmannii, considering the actual prevalence of R. wasmannii on M. sabuleti workers infected, compared to the low parasite prevalence on M. ruginodis and M. scabrinodis (Table 1).
In our study site, parasite prevalence was high in the scabrinodis-group (Table 1).Myrmica sabuleti and M. scabrinodis occur in other habitats; the former species gener-  (Bezdĕčka and Bezdĕčkova 2011;Bezdĕčkova and Bezdĕčka 2011;Csata et al. 2013;Espadaler and Santamaría 2012;Haelewaters 2012;Tartally et al. 2007;Witek et al. 2014), completed with the current findings from the Netherlands.*In Hungary and Romania, M. scabrinodis is the most common host of R. wasmannii (Csata et al. 2014;Tartally et al. 2007).In our study area in the Netherlands, M. sabuleti is the most commonly infected host species.
ally lives in drier areas than the latter.While in our study area in the province Limburg both species live in sympatry, we observed noticeable differences in occurrence and parasite prevalence.In the rubra-group, only one specimen of M. ruginodis was found with Rickia wasmannii.The fact that we did not find any infected M. rubra out of 974 workers raises questions, since this species is the second-most commonly found infected species in Europe after M. scabrinodis, with examples in Austria, Germany, Luxembourg, Romania, and Switzerland (Espadaler and Santamaría 2012, Table 2).Also in Romania, so far, infection by Rickia wasmannii only occurs in the rubra-group and scabrinodis-group (Csata et al. 2013, Tartally et al. 2007).In Romania, Myrmica scabrinodis often is the only or main host; in a single population R. wasmannii was found only on M. rubra, although thalli were also observed in populations of M. scabrinodis and M. slovaca; M. gallienii and M. rubra; and M. ruginodis (Csata et al. 2013).
Myrmica scabrinodis is the main host in Romania (studied areas, Csata et al. 2013), as is M. sabuleti in our studied site in the Netherlands.These observations suggest (1) that R. wasmannii has a considerable host choice plasticity within the genus Myrmica and (2) that infection rates of Rickia wasmannii show geographical variation, potentially with different dominant host species across regions.It is not known what factors cause this geographical variation.Future research in the Netherlands is needed to confirm that M. sabuleti is the main host in a wider region.

Figure 1 .
Figure 1.The study site with all the sampling localities indicated.Trajectories (lines) and areas (circles) were sampled with pitfall traps during three inventories.The separate inventories that are mentioned in the text are given: (i) in black (45 pitfall traps in nine series), (ii) in white (55 pitfall traps in eleven series), and (iii) in grey (324 pitfall traps in nine series).Inset: map of the Netherlands with location of the study site in Limburg.

Table 1 .
Total number of ant species studied for infection with R. wasmannii, over all three series of pitfall traps.

Table 2 .
All published records of R. wasmannii on different Myrmica hosts