Species list

To obtain an overview of the Austrian terrestrial isopod fauna, we updated the lists provided by Strouhal [12,14] and Schmölzer [16], which are the most detailed lists for Austria. As these include species that have since been synonymized or subspecies that have been raised to species, the validity of all species was checked against the list by Schmalfuss [9] and WoRMS (accessed 06/2025). The list was further complemented with other reports (e.g., [23]) and our own data.

Sample preparation of fresh material

Terrestrial isopods were collected between 2021 and 2025 from various habitats across Austria and preserved in 80% methanol-denatured ethanol (Fig 2; S1 Table). In addition, five individuals collected in neighboring Slovenia were included as well. Species identification was based mainly on Schmölzer [24] and the interactive key in https://bodentierhochvier.de as well as various original descriptions of species. As in several genera only males can be unambiguously identified to species, we largely employed a reverse taxonomic approach (following [46]), in which specimens are first clustered based on their DNA sequence data into putative species (molecular Operational Taxonomic Units = mOTUs) and then these clusters are assigned a species name based on the identification of the associated individuals. Specimens that themselves cannot be identified to species level (i.e., females, juveniles, damaged) are thus assigned the species name by genetic similarity. A map of all recent sampling localities (historic localities are not shown as their exact position are often vague) was made using QGIS version 3.28.14. The Copernicus GLO-30 Digital Elevation Model was downloaded from OpenTopography (https://portal.opentopography.org/raster?opentopoID=OTSDEM.032021.4326.3), boundary source(s): geoBoundaries, Federal Office for Metrology and Survey, Austria (https://www.geoboundaries.org/index.html) [47].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Map.

Sampling localities of all recently collected and successfully DNA barcoded Austrian specimens are shown. Blue = sampling points. Thick black line = country border of Austria. Narrow black lines = federal state borders. Historical localities without available coordinates are not depicted on the map. Map based on Copernicus GLO-30 Digital Elevation Model downloaded from OpenTopography (https://portal.opentopography.org/raster?opentopoID=OTSDEM.032021.4326.3), boundary source: geoBoundaries, Federal Office for Metrology and Survey, Austria [47].

https://doi.org/10.1371/journal.pone.0352446.g002

DNA was extracted from a single leg each. For very small-bodied individuals (mainly Trichoniscidae) a section of the mid-body was removed with sterile needles. The tissues were incubated in 10 μl Proteinase K (GeneOn) and 90 μl ATL lysis buffer (DNeasy Blood and Tissue kit, Qiagen) at 56°C for 2–4 hours. DNA was purified using 100 μl Ampliclean (Nimagen) magnetic beads, following the manufacturer’s recommendations. DNA was eluted in 40 μl nuclease-free H₂O. The extracted DNA samples were stored at −20°C.

A fragment of the cytochrome c oxidase subunit I (COI) gene, commonly used as a DNA barcode region for animal species, was PCR amplified using the LCO1490/HCO2198 primer pair (5’-GGTCAACAAATCATAAAGATATTG-3’/5’-TAAACTTCAGGGTGACCAAAAAATCA-3’; [48]). If these did not yield a PCR product, a shorter 316 bp fragment (so called mini-barcode) was amplified using the primer pairs BF1 (5’-ACWGGWTGRACWGTNTAYCC-3’) and BR2 (5’-TCDGGRTGNCCRAARAAYCA-3’; [49]) or BF1 and HCO2198. 15µl PCR reactions included 0.05µl DreamTaq DNA polymerase (Thermo Scientific), 1.5µl 10x DreamTaq Green Buffer, 0.3µl dNTPs (10µM each), 1.5µl each primer (10mM), 8.65µl H₂O and 1.5µl DNA extract. If amplification failed repeatedly, another taq-polymerase was employed: 7.5µl 2x QIAGEN Multiplex PCR Master Mix (Qiagen), 1.5µl of each primer (10mM each), 2 µl DNA extract and 2.5µl H₂O were combined and run with following PCR cycling program: 95°C for 15 min, 35 cycles of 30 s at 95°C, 45 s at 46°C and 1 min at 72°C and a final elongation of 5 min at 72°C. PCR products were visualized on 1.5% agarose-TBE gels. Prior to sequencing, primers and dNTPs were inactivated with 0.6 µl shrimp alkaline phosphatase (rSAP, NEB) and 0.3 µl exonuclease I (NEB) (incubated for 15 min at 22°C and 20 min at 85°C). Cleaned PCR products were bidirectionally sequenced using the respective PCR primers with the Sanger sequencing service of Macrogen. The resulting electropherograms were assembled, manually curated and aligned with Geneious PrimeⓇ 2025.0.3.

Ethics statement

Specimens were collected under the following collection permits: RU5-BE-64/023–2022 (Lower Austria), 2024-009.459-1/6 (Burgenland), A4/NR.AB-10007-12-2022 (Burgenland), MA22–302672/2024 (Vienna), ABT13–523696/2023–9 (Styria), VK3-NS −2790/2024 (005/2024) (Carinthia), FE3-NS-3710/2024 (005/2024) (Carinthia). This work did not require ethical approval from a human subject or animal welfare committee.

Sample preparation of historic material

From the historic collection of the Natural History Museum of Vienna, Austrian specimens identified by former Oniscidea expert Hans Strouhal were selected. Strouhal was the curator of the crustacean collection at the Natural History Museum Vienna from 1946 to 1962 with a strong focus on terrestrial isopods. Animals collected before 1900 were prioritized to avoid fixation in formaldehyde (the original fixative is rarely documented for the historic material), which further reduces the quality of historic samples due to issues like deamination of cytosine to uracil, crosslinks or single strand breaks [34]. 80 historic individuals were selected for DNA extraction.

DNA extraction of all historical samples was performed in the clean room of the Natural History Museum Vienna. Through UV radiation and filter systems, this facility enables contamination-free DNA extraction, which is particularly important for museum material. The DNA extraction protocol for historical samples was similar to that used for fresh samples (see above). To account for the low DNA concentration in museum samples, a larger amount of tissue was utilized compared to the fresh material. Small specimens were placed whole in the extraction buffer and the remaining exoskeleton preserved after digestion. To compensate for the increased proportion of tissue, the amount of proteinase K and ATL lysis buffer was doubled and also the relative amount of magnetic beads increased (up to 2x) to 300μl to retain small DNA fragments as well. DNA was eluted in 20–30 µl H₂O.

At first, PCR amplification via the standard primers for full length COI as well as mini-barcode primers BF1 and BR2 was tested as described above. As this failed on most instances, a DNA repair step with the NEBNext® FFPE DNA Repair v2 (E7360) module following the manufacturer’s instruction (except that all volumes were reduced by 50% to reduce costs) was tested for 21 DNA extracts. This kit is supposed to fix several typical problems of historic DNA (e.g., deamination, single-strand breaks, etc). Then the PCR with mini-barcodes was repeated and PCR products were Sanger sequenced as described above.

We further generated NGS Illumina libraries for genome skimming to obtain COI barcode sequences. Only dsDNA can be ligated to Illumina adaptors to construct genomic libraries. As a large fraction of hDNA is single-stranded, we tested a method that converts ssDNA to dsDNA prior to library construction for some samples. After DNA extraction, the dsDNA was quantified with the ThermoFischer Scientific® Qubit dsDNA HS Assay kit (Q32854) and ssDNA with the Qubit ssDNA Assay Kit (Q10212). Only for samples with sufficient DNA amounts library preparation commenced. First, potential DNA damages were repaired with the NEBNext® FFPE DNA Repair v2 module (see above) for each DNA extract. Library preparation was performed either directly using the repaired DNA or ssDNA was first converted into dsDNA using the ‘G-tailing’ protocol [41]. The G-tailing method adds a series of guanines to the ssDNA which then serves as a priming site for a C-primer to synthesize the second DNA strand [41]. 38.5 µl repaired DNA was denatured for 15 min at 96°C, then 5 µl 10x TdT buffer, 5 µl CoCl₂ (2.5 nM), 0.5 µl Terminal Transferase (20 U/µl) and 1 µl dGTP (5mM) (all from New England Biolabs) were added and incubated for 30 min at 37°C and 10 min at 70°C to add approximately 6–8 guanine bases. After clean-up using Ampliclean magnetic beads, DNA was eluted in 18 µl H₂O of which 16.5 µl were used in the following reaction. Second strand synthesis utilized 4.5 µl dNTPs (10mM each), 3 µl 10x NEBbuffer, 3 µl Klenow Fragment (5 U/µl), 3 µl primer (comprising 6 cytosine bases; 15 mM) and 16.5 µl of the G-tailed DNA and was incubated at 23 °C for 180 min and 75 °C for 20 min. The product was then cleaned-up with magnetic beads, with the elution volume being adapted to the respective library protocol that followed.

Library preparation was performed either with the NEBNext® Ultra™ II DNA Library Prep Kit for Illumina or the NEBNext® Ultra™ II FS DNA Library Prep Kit for Illumina (which has an additional fragmentation step at the beginning of the library preparation). We used half of the volumes of all reagents (including input DNA volumes) and otherwise followed the recommendation by the manufacturer. Adapters were stubby y-yoke adapters following Glenn et al. [50], to which a 5 bp barcode region had been added next to the ligation site of the adapter. The adapters gained full length during PCR amplification, which also added an index. This allowed a flexible indexing strategy, on the one hand by the barcodes (which always add the same unique barcode on both sides of the library and which represents the first 5 bp of each read) and on the other hand the two PCR primers, which adds the standard Illumina indices. This allowed multiplexing of several samples with a large number of unique barcode and index combinations. PCR amplification utilized the NEBNext Ultra II Q5 Master Mix included in the library preparation kits, setting up two PCR reactions of 10µl with 15 PCR cycles for each library to reduce amplification bias; these replicates were each pooled before beads clean-up and quantification. All clean-up steps were performed with Ampliclean magnetic beads. Prior to PCR amplification, 1.5-2x beads to library (or DNA) ratios were used to retain short DNA fragments, after PCR 1x beads to library ratios were used to remove primer dimers. Prior to sequencing and final pooling, all libraries were quantified with the Qubit 4 Fluorometer by Thermo Fisher Scientific and 4200 Tapestation D1000 tapes to assess fragment size distribution and to allow pooling at equal molarity. As some libraries had fragments >1000 bp, the final sequencing pool was run on Sage Science’s BluePippin DNA Size Selection System with a 1.5% agarose cassette retaining only DNA fragments between 250–600 bp prior to sequencing. Sequencing commenced with Genohub sequencing service with an Illumina HiSeq 4000. Illumina reads were trimmed with TrimGalore v. 0.6.2 (https://github.com/FelixKrueger/TrimGalore) with a minimum per base quality of 30. To retrieve COI sequences, reads were assembled with MitoFinder [51], providing all Isopoda whole mitochondrial genomes available on NCBI as references for gene identification.

Species delimitation

All COI sequences were combined in a single alignment. The complete dataset has been submitted to BOLD (S1 Table; dx.doi.org/10.5883/DS-ISOBA). Austrian representatives of Asellus aquaticus (a non-oniscidean freshwater isopod) were included as an outgroup. To obtain estimates of putative species (=mOTUs), species delimitation was performed with ASAP (Assemble Species by Automatic partitioning; [52]) using the ASAP web service (https://bioinfo.mnhn.fr/abi/public/asap/asapold.html) employing p-distances and standard settings. To further test if the mOTUs or putative species delimited by ASAP are reciprocally monophyletic, a maximum likelihood tree was inferred using RAxML [53], employing the GTR + I + G substitution model and 1000 rapid bootstrap replicates. RAxML was run on http://usegalaxy.eu [54]. The tree was visualized with FigTree ver. 1.4.3 The goal of the phylogenetic analysis is to test the support of the mOTUS and species, not to assess phylogenetic relationships among taxa. With only a single and highly variable gene, the relationships of these evolutionary old taxa cannot be adequately reconstructed. Pairwise genetic distances were calculated in MEGA-X [55] using p-distances and pairwise deletion.

To compare our data to that of Raupach et al. [25], who provided the most comprehensive DNA barcode dataset for terrestrial Isopoda from central Europe, and to assess if the mOTUs recovered for the species were the same, we downloaded the corresponding COI dataset from NCBI. The data was aligned and a joint phylogenetic and molecular species delimitation analysis was performed with RAxML and ASAP as described above.

Species diversity and DNA barcoding

Of the 64 nominal (sub)species known from Austria, we were able to provide DNA reference barcodes for 49 (plus one of an unidentified Trichoniscus species). This corresponds to 76% of the known Austrian Oniscidea fauna.

In his latest review of the Austrian Oniscidea, Strouhal [14] listed 64 species and subspecies. However, he omitted Haplophthalmus montivagus (see also [16,22] and included T. illyricus twice: on the one hand T. illyricus (see [12]) and on the other T. illyricus carinthiacus (see [14]). According to Schmölzer [16], the nominate form (which would be T. illyricus illyricus) does not occur in Austria. Comparison to Schmalfuss [9] and WoRMS (accessed 06/2025) showed that three of the previously listed species and subspecies are now considered junior synonyms of other species occurring in Austria: Androniscus stygius tschameri and Androniscus stygius kesselyáki (both now A. stygius), Tracheoniscus pseudoratzeburgi apenninorum (synonym of T. arcuatus), and Protracheoniscus amoenus amoenus (synonym of P. politus). In addition, some of the subspecies have now been raised to species level.

The following three nominal species have been newly reported for Austria since Strouhal [14] and Schmölzer [16]: Armadillidium arcangelii (see also [23]), and Philoscia affinis affinis and P. muscorum muscorum in this study. In total, the number of nominal Austrian Oniscidea species and subspecies remains unchanged at 64 (Table 1), although the species composition has changed slightly.

Of the newly reported species, A. arcangelii is not native to Austria and has been introduced by human activities; so far only a single localized population in Vienna is known [23]. The presence of the two Philoscia species is surprising as no Philoscia species had been included in any of the previous species lists for Austria [12,14,16]. Philoscia muscorum muscorum was found only in western and northwestern Vienna and all individuals exhibited the identical COI haplotype, which is also the most common haplotype in Germany (see S2 Fig; [25]). This may indicate that also Philoscia muscorum muscorum is an introduced species, though range expansion or simply being missed in previous studies is a possibility. Conversely, P. affinis affinis had a wider geographic distribution in Austria with records from several localities in Lower Austria and Styria and it includes a larger number of COI haplotypes. This rather suggests it is native in Austria. The historic collection of the Natural History Museum Vienna holds a very detailed and extensive Oniscidea collection from Austria. In total, seven lots of Austrian Philoscia specimens from Vienna, Carinthia, Tyrol and Lower Austria dating back to as early as 1890 are present. As all were females, an unambiguous identification was not possible, though some had been historically identified as P. muscorum. Although this underpins the long-term presence of Philoscia in Austria, the number of historic records is relatively small compared to many other species; whereas Philoscia was quite commonly encountered in our study. Maybe it has become more common in recent years. Why these historic records were not considered by previous authors (e.g., [12,14,16]) and why these species were never encountered by them remains an open question.

Some of the 49 (sub)species, for which DNA barcodes were obtained, are very rare and difficult to collect, including Mesoniscus alpicola, which is found exclusively in caves, and Oroniscus mandlii, which is known from a single mountain and of which only two specimens had been previously collected [13]. These newly generated DNA barcodes are an important addition to the Austrian Barcode of Life (ABOL) initiative (www.abol.ac.at; [8]). Until now, however, there had been little data on isopods. The remaining ~ ¼ of species, for which we were not able to produce DNA barcodes, belong mainly to the small-bodied Trichonsicidae including several in-soil or subterranean species. Also, species and populations from western and/or northern Austria were underrepresented, as our study was biased towards eastern and southern Austria (especially Vienna, Lower Austria, Styria and Carinthia).

The presence of multiple, genetically very clearly differentiated mOTUs in several nominal species suggests the presence of additional species. Reported intraspecific COI distances for Oniscidea are usually below 4% (e.g., [56,57]) or below 6.5% [58] with interspecific distances usually exceeding 9% [56–60]. However, minimum interspecific distances may be underestimated, as several of these studies did not necessarily focus on putative sister species. Data from marine isopods suggests intraspecific distances of <6% and interspecific distances exceeding 10% [61–65].

With the data available, it is not possible to unambiguously decide which of the mOTUs represent actual species and which are rather genetically divergent intraspecific lineages. The comparison to the previous studies and the divergence patterns in the studied species, suggests that our most conservative estimate (ASAP9.8) probably underestimates actual species numbers. However, even under this conservative approach several species are split into multiple mOTUs, which most likely represent cryptic species. This includes the ant-associated species Platyarthrus hoffmannseggii, the cave-living M. alpicola and three Trichoniscus species (T. noricus, T. pusillus and T. steinboecki) as well as L. germanicum. For the latter, our preliminary morphological data on the male pleopod II endopod further supports the distinction between the two L. germanicum mOTUS (here termed L. germanicum and L. cf. germanicum). The respective morphological differences may be minor, but consistent with differences observed among other Ligidium species [24]. Together with the 18.5–20.0% p-distance in COI, this strongly suggests that these represent reproductively isolated species. The mOTU identified as L. germanicum probably represents the “true” L. germanicum as its distribution spans southern Germany, where the species was first described from. We hesitate to formally describe Ligidium cf. germanicum as a formal species herein as this would require a larger and more comprehensive sample of the species and a more in-depth morphological characterization.

Conversely, the two subspecies P. fiumanum fiumanum and P. fiumanum salisburgense were lumped together at a 9.8% threshold as were A. opacum and A. carniolense, but they were separated in the partitions with thresholds of 4.8% and 6.2%. Together with their morphological differences, this suggests that these (sub)species are indeed separate species. It further strengthens the assumption that several of the other mOTUs separated by lower threshold values represent distinct species as well. Our least conservative estimates (ASAP4.8 and even more BINs) probably overestimate species diversity and we assume that the actual species diversity is closer to ASAP6.2. Of course, no universal barcoding threshold exists and each species (or mOTU) pair needs to be assessed independently in an integrative framework including morphological and nuclear data. This requires extensive revisions including material from other countries as some of these mOTUs may correspond to previously synonymized species or subspecies. For this reason, we do not propose any taxonomic changes herein. The example of P. collicola highlights that for some presumably well-known species the extent of morphological variability is not fully understood, which in turn may lead to erroneous species identifications. Integrative approaches that document the morphological variability across the species’ geographic range are needed. This may be particularly important for the small-bodied Trichoniscidae where we document several examples of putative cryptic species, showing the need for a more thorough revision. Similarly, the non-monophyly of the genera Porcellio and Haplophthalmus may indicate the need for further systematic revisions. However, our phylogenetic analysis was based on a single, fast evolving gene and more detailed analyses are required before any systematic and taxonomic changes are proposed. In summary, our results indicate the presence of multiple hitherto undescribed species, suggesting that the diversity of terrestrial isopods in Austria may be even higher than the currently accepted 64 species.

Museomics

Including historic museum specimens in DNA barcoding studies has many advantages. They can fill gaps in DNA barcoding reference databases – which is particularly important for rare, difficult to access and collect, threatened, protected or even extinct species – and allow the integration of expert knowledge of previous experts, by integrating specimens identified by the respective historic experts. The gold standard would, of course, be the integration of historic type specimens. Due to their high scientific value, the latter can be difficult for many small-bodied taxa, as DNA extraction is often at least partly destructive (see also [45]). For this reason we did not attempt sequencing historic types herein. Our results underline the overall value of such museomic (sensu [44]) approaches. DNA barcodes of three species were only obtained from historic museum specimens and these species would not have been represented otherwise.

A variety of different museomic approaches exist. We tested and compared Sanger-based as well as Illumina-based approaches. For Sanger-based approaches, targeting mini-barcodes (instead of longer fragments) and the treatment of the DNA with a DNA repair kit prior to PCR greatly increased PCR amplification and sequencing success. This is probably the most effective approach in terms of costs and lab work to generate DNA barcodes of historic specimens and could probably also be integrated in large-scale sequencing approaches [66,67]. Higher success rates of mini-barcodes compared to full-length barcodes had been demonstrated before, sometimes with success rates close to 100% (e.g., [68–70]). The higher overall success rates in those studies compared to ours, may be due to differences in fixation, preservation and storage of the specimens. DNA quality appears to degrade faster in ethanol-fixed specimens than in dried specimens (e.g., [42]) and some of the specimens studied by us were probably fixed in formalin. Unfortunately, this was not well documented in the past. Of course, mini-barcodes contain less information compared to full length-barcodes due to their shortness, but their overall effectiveness in species delimitation has been demonstrated [69,71] and we would argue that a short barcode is preferable to no barcode, though we would always try to obtain full barcodes if possible. We do not know which of the repair steps included in the DNA repair kit is responsible for the increased PCR amplification success. Probably it is either the repair of the deaminated bases (which would have prematurely terminated the PCR amplification) or the repair of single-stranded nicks, as these nicks would result in much shorter fragments when the DNA is denatured during PCR.

Illumina-based target capture was also successful in generating high quality COI sequences, with the advantage that these COI sequences were usually (but not always) longer than the mini-barcodes and, of course, also other genetic markers may have been retrieved if wanted. However, the overall efforts in costs and lab work are higher compared to Sanger-based mini-barcodes, especially when G-tailing or commercial kits are included in the library preparation. We observed only minor differences in success rates between Illumina libraries with G-tailing compared to those without. Maybe other methods developed for converting ssDNA into dsDNA are more effective (e.g., [72–74]. Despite the higher effort compared to Sanger-based approaches, it is worthwhile to employ Illumina-based approaches for scientifically valuable specimens for which Sanger-based approaches failed or if genetic data beyond DNA barcodes is desired.

In summary, our study provides important reference COI barcodes for the Austrian terrestrial isopod fauna. These will be relevant for future metabarcoding-based biodiversity and monitoring studies of soil organisms. Especially, for many of the small-bodied taxa such reference data has been missing. Our results also highlight the potential of cryptic diversity in this group and the utility of museomic approaches in DNA barcoding approaches.