On the identity of Chamaedrilus glandulosus (Michaelsen, 1888) (Clitellata, Enchytraeidae), with the description of a new species

Abstract The taxonomy of Chamaedrilus glandulosus (Michaelsen, 1888) s. l., most commonly known previously as Cognettia glandulosa, is revised. A recent molecular systematic study has shown that this taxon harbours two cryptic, but genetically well separated lineages, each warranting species status. In this study these two lineages are scrutinized morphologically, on the basis of Michaelsen’s type material as well as newly collected specimens from Central and Northern Europe. Chamaedrilus glandulosus s. s. is redescribed and Chamaedrilus varisetosus sp. n. is recognized as new to science. The two species are morphologically very similar, differing mainly in size, but seem to prefer different habitats, with Chamaedrilus glandulosus being a larger aquatic species, and Chamaedrilus varisetosus being smaller and mainly found in moist to wet soil.


Introduction
In 1888 Michaelsen described an enchytraeid worm, Pachydrilus sphagnetorum var. glandulosus Michaelsen, 1888, as a variant of P. sphagnetorum Vejdovský, 1878. The description was based on material from the banks of the Bille and Elbe rivers in Hamburg, northern Germany. These two taxa were then transferred to Marionina Michaelsen, 1890(in Pfeffer 1890, and P. sphagnetorum var. glandulosus was considered a good species, Marionina glandulosa, separate from M. sphagnetorum (Michaelsen 1900). Later Friend (1919) assigned both species to Chamaedrilus Friend, 1913, an action seldom noticed by subsequent authors. For instance, when Nielsen and Christensen (1959) established Cognettia, they transferred Marionina glandulosa to their new genus without considering its previous placement in Chamaedrilus. Nielsen and Christensen's (1959) concept of Cognettia came to embrace a number of terrestrial and freshwater enchytraeids and until recently it has been widely accepted. However, as noted by Schmelz and Collado (2010) and now more closely investigated by ourselves , Cognettia is indeed a junior synonym to Chamaedrilus. For details about the complex taxonomical history and a formal revision of Chamaedrilus, see .
Several cryptic forms have been found within well-known morphology-based taxa of former Cognettia . The morphospecies Chamaedrilus sphagnetorum s. l. was found to be a non-monophyletic assemblage of at least four species; these have been revised and described by . The taxon Ch. glandulosus, on the other hand, traditionally distinguished from sphagnetorum by the possession of secondary septal glands and longer spermathecal ectal ducts (Nielsen and Christensen 1959), was shown by both nuclear and mitochondrial DNA evidence to consist of two separately evolving lineages in Northern Europe. These two lineages appeared as sister species, i.e., representing a monophyletic group . According to Christensen (1959) Ch. glandulosus s. l. reproduces both by fragmentation and parthenogenetically, but the eggs must be activated by spermatozoa for normal development (Christensen 1961). However it is still possible that at least one of the two cryptic species occasionally reproduces biparentally. Uniparental reproduction makes species delimitation problematic, in particular when referring to the biological species concept (Mayr 1942). However, as discussed by , asexual organisms form distinct clusters and can be delimited using the unified species concept by de Queiroz (2007). According to this concept, the sole requirement of a species is that it is a separately evolving metapopulation lineage, and criteria (e.g. morphological differences, reproductive isolation, or gene tree monophyly) from any of the more traditional species concepts can be used to delimit the lineages. The greater the number of criteria supporting a divergence, the stronger the case is for speciation, but, even a single piece of evidence, if properly substantiated, may be enough to establish lineage separation.
The aim of this study is to revise the taxonomy of Chamaedrilus glandulosus s. l. by delimiting Ch. glandulosus s. s., with the designation of a lectotype, and describing Ch. varisetosus sp. n.

Material and methods
This study is based on two syntypes of Pachydrilus sphagnetorum var. glandulosus Michaelsen, 1888, from the original syntype series of ten, borrowed from the Zoological Museum of Hamburg University (ZMUH), Germany, of which one is here designated as lectotype, plus material analysed by , and new specimens collected in northern and central Europe. A list of all examined specimens, with locality data and GenBank accession numbers for DNA-barcodes is given in Table 1.
Newly collected specimens were DNA-barcoded using the cytochrome c oxidase subunit I (COI) marker, as described by ; DNA was extracted from a few posterior-most segments of each worm, using Epicentre Quick-Extract DNA Extraction Solution 1.0, following the manufacturer's instructions, while the rest of the specimen was used for morphological studies, i.e., as a voucher. All new barcodes were matched with COI sequences of Cognettia glandulosa 'A' and 'B' from . For tissue samples of the over 100 years old syntypes, newly designed primers were tested to amplify a short part of COI, as well as a fragment of the ribosomal 16S mtRNA gene, respectively, but these attempts were unsuccessful.
Unless otherwise mentioned in the descriptions, all information refers to the studied material only, in that the two taxa treated in this paper have previously been classified as one and the same species. Michaelsen's syntypes were first studied as temporary mounts in glycerol. The newly designated lectotype was then stained with paracarmine and permanently mounted in Canada balsam on a slide as outlined by Erséus (1994), and so were all other voucher specimens (including the types of Ch. varisetosus sp. n.). All measurements and observations were made on preserved and somewhat compressed animals under a compound microscope (Leitz Laborlux K). As the posterior parts of the specimens were used for DNA extraction, the body size is arbitrarily given as the length of the 20 anteriormost segments and the width in segment XII (latter representing not clitellum but general body width). This size estimate was used also in . In the descriptions, body measurements are given as the range followed by the mean ± 1 standard deviation. Differences in size between the two species were visualised with boxplots ( Fig. 1, where asterisks denote the outliers), and tested by using two-sided t-tests performed in SPSS v. 22 (SPSS Inc., Chicago). Sketches were drawn using a camera lucida and used as templates for producing digital illustrations with Adobe PhotoShop.
The geographical distributions consider the origin of our material as well as that of COI barcode matches in BOLD (Barcoding of Life Data Systems, Ratnasingham and Hebert 2007). The Barcode Index Numbers (BIN) (Ratnasingham and Hebert 2013) are given under Remarks, for respective species. The BIN system clusters the sequences to produce operational taxonomic units that are assumed to closely correspond to species (http://www.boldsystems.org).
All specimens studied, including new types, are deposited in the Swedish Museum of Natural History (SMNH), Stockholm, the University Museum Bergen (UMB),  Norway, and the Zoological Museum Hamburg (ZMUH), Germany; all COI barcodes are deposited in GenBank (see Table 1). Paralectotype. ZMUH V 429b, immature specimen, in alcohol; same collection data as for lectotype.
Additional type material (not studied). Paralectotypes ZMUH V 429b, 8 specimens in alcohol, same collection data as for lectotype.
Other material. See Table 1. In total 15 specimens, of which one from Finland, one from Norway and 13 from Sweden (whereof one mature and three submature). All specimens except one are DNA barcoded (Table 1).
Diagnosis. Can be separated from all other European species of Chamaedrilus except Ch. varisetosus by its unique combination of 2-4 pairs of well-developed secondary pharyngeal glands, two chaetae per lateral bundle in preclitellar segments, and three chaetae in all other bundles, spermathecae with comparatively long ectal ducts, and genitalia shifted forward 3-4 segments (in relation to normal placement in Enchytraeidae). No characters completely separate this species from Ch. varisetosus sp. n., but specimens of Ch. glandulosus are usually larger and have only two chaetae in the lateral bundles of preclitellar segments, whereas Ch. varisetosus usually has three chaetae in lateral bundles of III-V. Furthermore, Ch. glandulosus is found in aquatic habitats only (i.e. submerged under water for most of the time), whereas Ch. varisetosus is found in both aquatic and terrestrial habitats; so far we have not found them occurring together. Description. EXTERNAL CHARACTERS: Size: length of 20 anteriormost segments 3.49-6.68 mm, mean 4.55±0.87 (n=11); body width in XII 0.24-0.56 mm, mean 0.42±0.10 (n = 14). Chaetae sigmoid without nodulus, 60-100 µm long, chaetal formula 2,(3)-3:3-3, with 3 lateral chaetae per bundle from VII-IX; in sexually mature specimens, ventral chaetae, or both ventral and lateral chaetae, missing in the segment bearing male pores (VIII or IX). In the sexually mature and submature specimens examined, clitellum poorly developed.
Habitat and distribution. Occurs in freshwater habitats, in sand and gravel bottoms in lakes and small streams, and climbing on vegetation and dead wood in water. Barcoded specimens document occurrence in Finland, Germany, Norway and Sweden, but the species is probably more widely distributed, not only in Europe. For instance, Ch. glandulosus s. l. has also been reported from North America: the records by Nurminen (1973) and Healy (1996) are insufficiently described and cannot even tentatively be assigned to any of the two species, and the records by Schlaghamerský (2013) and Schlaghamerský et al. (2014) are likely to be Ch. varisetosus, see under Habitat and distribution for that species.
Biology. Seems to reproduce mainly parthenogenetically; specimens with developing genitalia are found from June to July (Sweden).
Remarks. Michaelsen (1888;1900) described this species as sturdier than Ch. sphagnetorum, with 2 chaetae per preclitellar lateral bundle and three chaetae in all other bundles. This together with the fact that Michaelsen's type material was collected at an aquatic site makes us confident that our new material is conspecific with Michaelsen's species. Michaelsen (1888) described the spermathecae in vivo as very long ("they often project, in spite of much meandering, up to the segment VII") and the ampullae to consist each of an ectal enlargement followed by a long connecting tube and an expanded ental chamber (Fig 2F). In our new material the spermathecae seem to be either not fully developed or much con-tracted after fixation: they show simple oval ampullae, not differentiated into ectal and ental compartments. In the mature lectotype we can only follow the spermathecae to what we interpret as the ampullar ectal enlargement. Chamaedrilus glandulosus is larger than Ch. varisetosus described below. Both the length of the 20 anteriormost segments (P = 1.5E-5) and the width in segment XII (P = 5.5E-5) differ significantly between the two species (Fig. 1).
This species is represented in BOLD by BIN: AAT8923. Other material. See Table 1. Twenty-seven immature specimens, of which 2 from the Czech Republic, 12 from Norway, and 13 from Sweden, all DNA-barcoded.
Etymology. The species is named after the variation in numbers of chaetae in the lateral preclitellar bundles.
Diagnosis. The new species can be separated from all other European species of Chamaedrilus except Ch. glandulosus s. s. by its unique combination of 3-4 pairs of well-developed secondary pharyngeal glands, two chaetae in most lateral bundles in preclitellar segments, and three chaetae in all other bundles, spermathecae with comparatively long ectal ducts, and genitalia shifted forward 3-4 segments (in relation to normal placement in Enchytraeidae). No characters completely separate this species from Ch. glandulosus, but specimens of Ch. varisetosus are generally smaller, have shorter chaetae and smaller internal organs, and usually have a few preclitellar lateral bundles with three chaetae (Ch. glandulosus constantly has two chaetae per lateral bundle in preclitellar segments). Furthermore, Ch. varisetosus is mainly found in moist to wet soils, whereas Ch. glandulosus is only found in aquatic habitats.
Habitat and distribution. Found both in aquatic and terrestrial habitats. In freshwater found on stony bottoms in rivers, on land found in both deciduous and coniferous forest as well as in grassland soils. Known from Canada (BOLD record), the Czech Republic, Finland (BOLD record), Norway and Sweden, but may be more widely distributed in Europe and North America. Schlaghamerský's (2013) description of C. glandulosa from Michigan fits our description of Ch. varisetosus. This and Schlaghamerský's et al. (2014) records from Minnesota and Wisconsin are likely to refer to the same species.
Biology. Parthenogenetic reproduction more limited in time (maturing specimens found in August in Norway) than fragmentation (observed in May-September in Sweden and Norway). Worms with regenerating tails and/or heads rather frequent. This species may correspond to the population studied by Christensen (1959), in which the number of mature worms was high for a short period during the autumn. The variation in number of the lateral chaetae corresponds to that given in the diagnosis by Nielsen and Christensen (1959).
Remarks. This species is represented in BOLD by BIN: AAT9501.

Discussion
The two species treated in this paper, Chamaedrilus glandulosus sensu stricto and Ch. varisetosus sp. n., are easily separated morphologically from other species of Chamaedrilus by a unique combination of characters: the secondary pharyngeal glands are well developed in several segments, there are two chaetae in most preclitellar lateral bundles, but no enlarged chaetae, the genital organs are shifted forwards, and the spermathecae have comparatively long ectal ducts. The two species are morphologically similar and they have therefore been regarded as a single taxon by previous authors (e.g., Nielsen and Christensen 1959;Schmelz and Collado 2010). As demonstrated in the present paper, they can only be separated by their body size, chaetal size (and prevailing number) and, when fully grown, by the proportions of most internal organs. Genetically, however, they are well separated from each other , and they are also ecologically separated, with Ch. glandulosus found in aquatic habitats, whereas Ch. varisetosus is predominantly found in moist to wet soil. Ecological and physiological differences have been found between cryptic lineages in morphospecies of various organisms (e.g. Beauchamp et al. 2002;Feckler et al. 2014;Sattler et al. 2007), and if such lineages are not formally recognized and named, the differences may continue to be overlooked or neglected.  found Chamaedrilus glandulosus and Ch. varisetosus sp. n. to be sister species, nested within a part of the sphagnetorum-complex, making the latter non-monophyletic. The sphagnetorum-complex also turned out to be morphologically more heterogeneous than Ch. glandulosus s. l. , which could probably be, at least partly, explained by its non-monophyly. However, not even the two morphologically indistinguishable species, Ch. sphagnetorum s. s. and Ch. pseudosphagnetorum  came out as sister species in the phylogenetic study .
Without the genetic data, the delimitation of Ch. glandulosus and Ch. varisetosus would have been much more challenging, all the more so because these worms, like those in the sphagnetorum complex, are mostly found sexually immature. It should also be considered that these species, even when mature, actually reproduce uniparentally, as mentioned in the introduction and discussed earlier by . Uniparental reproduction makes species delimitation harder; however, we still believe this is possible using the unifying species concept (see Introduction). In the present case, we have a combination of genetic, ecological and morphological differences, supporting the split of Ch. glandulosus s. l. into two species. It should further be noted that it is not known with certainty if Christensen (1959;1961) studied both species, or only one of them. As mentioned in the description, Ch. varisetosus seems to correspond well with the taxon studied in his 1959 paper and also fits the description given by Nielsen and Christensen (1959). Until the mode(s) of reproduction is (are) studied again for the two species, we cannot exclude the possibility that one or both species may reproduce biparentally, at least occasionally.
Genetic studies discovering cryptic and unnoticed diversity need to be followed by formal taxonomic revision, including careful morphological scrutiny, updated descriptions and species names, if possible based on barcoded types. We believe that an integrative approach, combining genetic and morphological data with as much as possible of ecological and physiological information, will strengthen studies of enchytraeid systematics.