Integrative taxonomy of five astome ciliates (Ciliophora, Astomatia) isolated from earthworms in Central Europe

. Four earthworm species, the endogeic Octolasion tyrtaeum (Savigny, 1826), the anecic Lumbricus terrestris Linnaeus, 1758 as well as the epigeic Eisenia fetida (Savigny, 1826) and Dendrobaena veneta (Rosa, 1886), were examined for the presence of astome ciliates. Based on the integrative taxonomic approach, five ciliate species were recognized in their gastrointestinal tracts: Metaradiophrya lumbrici (Dujardin, 1841), M. varians (de Puytorac, 1954), Anoplophrya lumbrici (Schrank, 1803), A. vulgaris de Puytorac, 1954 and A. nodulata (Dujardin, 1841). Their distinctness was assessed using the multivariate morphometric approach and molecular phylogenetic analyses. Although the two species of Metaradiophrya Jankowski, 2007 on the one hand and the two former species of Anoplophrya Stein, 1860 on the other, were not distinctly separated by the multivariate morphometric analyses, they were clearly delimited by the 18S rRNA gene sequences. Species within each genus also differed by their hosts, M. lumbrici and A. lumbrici occurred only in anecic earthworms while M. varians and A. vulgaris occured exclusively in epigeic earthworms. Only a single species, A. nodulata , was detected in endogeic earthworms. It was morphologically distinct from and did not cluster with the two other species of Anoplophrya but was nested within the paraphyletic assemblage containing other astomes from endogeic earthworms. This indicates that the evolution of endosymbiotic ciliates from earthworms has very likely proceeded through a specialization to various ecological groups of their host organisms.

with the aid of Pasteur micropipettes adjusted as described by Foissner (2014). Living ciliates were investigated at low (50-400´) and high (1000´, oil immersion) magnifications, using bright field and differential interference contrast optics (Foissner 2014). Images were captured by a Canon EOS 70D camera. Ciliates were measured from images, using the calibrated software ImageJ ver. 1.49 (Schneider et al. 2012).

Multivariate taxonomic methods
A multivariate approach was used to investigate the morphometric variation and distinctness of four astome species isolated from the gastrointestinal tract of earthworms. Altogether, 16 quantitative features, a single qualitative character and two derived ratios were scored on 33 specimens. The libraries NumPy (Oliphant 2015) and Pandas (McKinney 2010) were utilized to load and process the morphometric matrix in Python ver. 3.6.6. The similarity of specimens was measured by Gower's coefficient GOW (Gower 1971):  Table 1. where i and j are specimens, s ijk is the contribution provided by the k-th variable and w ijk is 1 or 0 depending upon whether or not the comparison is valid for the k-th variable. Values of s ijk for variables measured at interval and ratio scales are defined as follows: where x ik is the value of the k-th variable on the specimen i, x jk is the value of the k-th variable on the specimen j and R k is the range of values for the k-th variable. Values of s ijk for binary qualitative variables are 1, if i and j both have the attribute k present or 0 otherwise. The weight w ijk causes negative matches to be ignored. The function for computation of Gower's coefficient was obtained from https://github.com/scikit-learn/scikit-learn/pull/9555. OBERT T. & VĎAČNÝ P., Morphology and phylogeny of five astome ciliates 5 The similarity of specimens, as measured by Gower's index, was assessed by metric multi-dimensional scaling implemented in the scikit-learn package (Pedregosa et al. 2011). The SMACOF algorithm was run with 250 initializations, each run had 20 000 iterations and ε was set to 10 -8 to declare convergence. Plotting of the ordination diagram was done with the Matplotlib module (Hunter 2007).

Molecular methods
Single cells were picked and washed in five drops of Ringer's solution (0.6%). Thoroughly washed specimens were stored in 180 µl of the cell lysis buffer CLD (Promega, Fitchburg, Wisconsin, USA) at 6°C. Genomic DNA from single cells was extracted with the ReliaPrep™ Blood gDNA Miniprep System (Promega, Fitchburg, Wisconsin, USA). The 18S rRNA gene was amplified with the universal eukaryotic primers Euk A (5'-AAC CTG GTT GAT CCT GCC AGT-3') and Euk B (5'-TGA TCC TTC TGC AGG TTC AC-3') (Medlin et al. 1988). Individual polymerase chain reactions (PCR) included 5 µl of the extracted template DNA, 0.4 µl of the forward and reverse primers each (10 pmol/µl) and 10 µl of the GoTaq® Long PCR Master Mix (Promega, Fitchburg, Wisconsin, USA). The final volume was adjusted to 20 µl with deionized distilled water. PCR conditions were as follows: initial hot start denaturation at 95°C for 15 min, 30 identical amplification cycles (denaturing at 95°C for 45 s, annealing at 55°C for 1 min and extension at 72°C for 2.5 min) and final extension at 72°C for 10 min. The quality of the amplified DNA was checked by electrophoresing a 1% agarose gel. PCR products were purified using calf intestinal alkaline phosphatase and exonuclease I, E. coli (New England Biolabs ® Inc.) and sequenced on an ABI 3730 automatic sequencer (Macrogen, Amsterdam, The Netherlands).

Phylogenetic methods
Newly obtained sequences were examined in Chromas ver. 2.6.6 (Technelysium Pty Ltd.) and highquality sequence fragments were assembled into contigs in BioEdit ver. 7.2.5 (Hall 1999). Two 18S rRNA gene alignments were generated on the GUIDANCE2 server (http://guidance.tau.ac.il/ver2/), with the MAFFT algorithm and 100 bootstrap repeats (Sela et al. 2015). The first alignment (referred as 'general' henceforth) included 73 taxa and served to classify the newly obtained sequences into oligohymenophorean subclasses. Representatives from all subclasses were selected to cover their diversity and sampling basically followed our previous studies (Rataj & Vďačný 2018, 2019. Members from the subclass Peniculia were used to a posteriori root the trees (Fokam et al. 2011). The second alignment (referred as 'detailed' henceforth) contained 34 sequences from the oligohymenophorean subclass Astomatia and two outgroup sequences of Dexiotricha Stokes, 1885 from the subclass Scuticociliatia Small, 1967. Sampling in this dataset mostly followed Fokam et al. (2011) and Rataj & Vďačný (2018, 2019 and served to more closely analyze the phylogenetic position of the new astome isolates. To estimate the reliability of alignments, unmasked datasets and datasets masked with a cutoff value of 0.93 were constructed. The best evolutionary substitution models for all datasets were estimated and selected in jModelTest ver. 2.1.10 under the Akaike Information Criterion (Darriba et al. 2012) on the Cipres portal ver. 3.1 (http://www.phylo.org/) (Miller et al. 2010). Phylogenetic relationships among oligohymenophorean ciliates were reconstructed in the maximum likelihood, Bayesian and neighbor-joining framework under the GTR +Γ + I evolutionary model with parameters as estimated in jModelTest. Maximum likelihood analyses were performed in PhyML ver. 3.0 with the SPR swapping algorithm and 1000 non-parametric bootstrap replicates on the South of France bioinformatics platform (http://www.atgc-montpellier.fr/phyml/) (Guindon et al. 2010). Bayesian inferences were conducted on the Cipres portal ver. 3.1 in the program MrBayes on XSEDE ver. 3.2.6 (Ronquist et al. 2012). Prior parameters of the GTR +Γ + I evolutionary model were implemented with the 'lset' and 'prset' commands in the MrBayes command block. Two runs, each having four (one cold and three heated) simultaneous Markov chains, were five million generations long. The sampling frequency was set to one hundred and the burn-in fraction was specified as 25%. In addition, the second set of detailed alignments was also analyzed by the neighbor-joining algorithm in MEGA X (Kumar et al. 2018). All trees were computed as unrooted and were rooted a posteriori in FigTree ver.  (Shimodaira & Hasegawa 2001;Shimodaira 2002Shimodaira , 2008. Topologically unconstrained and constrained trees were constructed under the GTR +Γ + I evolutionary model, using the heuristic search, random sequence addition and the SPR swapping algorithm in the program PAUP* ver. 4.0b8 (Swofford 2003). Site-wise log likelihoods were calculated for unconstrained and constrained trees and served to calculate p-values for the tree topology tests in the program package CONSEL, using the commands makermt, consel and catpv (Shimodaira & Hasegawa 2001).
Average evolutionary distances among the five species of astome ciliates isolated from lumbricid earthworms, were calculated from their 18S rRNA gene alignment with the maximum composite likelihood model in MEGA X (Kumar et al. 2018). Standard errors of between group evolutionary distances were estimated with the bootstrap method and 1000 replicates. The rate variation among sites was modeled with a gamma distribution and the differences in the composition bias among sequences were considered in evolutionary comparisons (Tamura & Kumar 2002).

Description
The body size is about 105-230 × 55-130 µm, with an average of 170 × 100 µm. The shape is ovate to elliptical with an anterior body end rounded and posterior end broadly rounded to truncate. The cell is distinctly dorsoventrally flattened (Figs 2A, F, 3B, F-M, 4A-B, E-F). On the ventral side, about 10 µm away from the anterior body end, there is a conspicuous fibrillar hook composed of two unequally long arms. The longer arm is 25-35 µm long, flat and completely situated underneath the cell surface. The shorter arm is 8-13 µm long, usually appears slightly more robust at the base and projects from the cell      The nuclear apparatus is composed of a macronucleus and a micronucleus. The macronucleus is rodlike with both ends rounded. It begins about 27 µm away from the anterior body margin and ends about 12 µm above the posterior body margin. Its length spans a range from approximately 85 to 200 µm and its width ranges from 8 to 18 µm, averaging at 13 µm. The macronuclear surface is smooth and without any irregularities, however, small vesicules appear in its vicinity in dying cells. The micronucleus is typically situated close to the mid-portion of the macronucleus. The shape of micronucleus is circular to elliptical and its diameter is approximately 7 µm. The central region of micronucleus appears homogenous and brighter than its margin in the differential interference optics and might represent a central nucleolus There are invariably two staggered rows of contractile vacuoles, extending right and left of the macronucleus. Their number ranges from 3 to 12 with an average of 6 vacuoles in the right row and from 3 to 11 with an average of 6 vacuoles in the left row ( Somatic ciliature is holotrichous and composed of very densely ciliated and narrowly spaced kineties. The ventral ciliature is interrupted by a subapical suture that extends from the right body margin over the fibrillar hook towards the left body margin. Somatic kineties above the suture run towards the anterior body end where they curve onto the dorsal body side to meridionally extend over its surface towards the posterior body end. Somatic kineties below the suture run meridionally over the ventral side towards the posterior body end (

Occurrence
Metaradiophrya lumbrici was detected exclusively in a group of anecic earthworms, namely, in L. terrestris at three localities: in gardens in the Šúrska ulica street in Rendez and in the Jakubská ulica street in Rača as well as in floodplain soils in a riparian, willow-poplar forest near the Karlova Ves branch of the Danube River (Table 2). Ciliates were typically isolated from the middle part of the gastrointestinal tract, although very rarely one or two specimens were recorded also slightly above and below this gut region. (de Puytorac, 1954) Figs 6, 7

Description
The body size is about 110-180 × 85-95 µm. The body shape is ovate to elliptical with an anterior body end rounded and posterior end broadly rounded to truncate. The cell is distinctly dorsoventrally flattened (Figs 6A, F-I, 7A, D). The fibrillar hook is localized about 10 µm away from the anterior body end. Its longer arm measures on average 30 µm, while its shorter arm only 11 µm. The hook appears slightly flatter than in the previous species and the thickened base of the shorter arm has been never observed. There are 5 or 6 fibers attached to the upper right side of the longer arm, on average 26 (22-29) fibers to the ventral side of the longer arm and 11 or 12 fibers to the left side of the shorter arm (Figs 6A-C, 7B).  There are two staggered rows of contractile vacuoles arranged along the right and left side of the macronucleus: 5-8 vacuoles in the right row and 6 or 7 vacuoles in the left row (Figs 6A, E-I, 7A, D).
The cytoplasm is colorless and studded with granules about 0.5 µm across. The cortex is rigid and without granules (Fig. 7B). Swims moderately fast by rotation about the main body axis.
Somatic ciliature is holotrichous and composed of very densely ciliated and narrowly spaced kineties. The ventral ciliature is interrupted by a subapical suture extending from the right body margin over the fibrillar hook to the left body margin. Somatic kineties above the suture run towards the anterior body end where they curve onto the dorsal body side to meridionally extend over its surface towards the posterior body end. Somatic kineties below the suture run meridionally over the ventral side towards the posterior body end (Figs 6A-B, 7B-C). The number of ventral kineties ranges from 50 to 64 on the ventral side and from 53 to 64 on the dorsal side. There is a subterminal suture on the left and the right side of the body.

Occurrence
Metaradiophrya varians was solely recorded in the epigeic E. fetida at two comparatively distant localities, i.e., in compost heaps in the Jakubská ulica street in Rača and in the Botanical Garden of Comenius University ( Fig. 1; Table 2). Metaradiophrya varians inhabited only the middle part of the gastrointestinal tract and there were usually ten exemplars per earthworm. Our data on occurrence of M. varians and M. lumbrici indicate that both species are specialized on different ecological groups of host oligochaetes. The ecologically different L. terrestris and E. fetida never contained the same species of Metaradiophrya, even when they originated from closely situated localities or when they were cocultivated in the laboratory.

Description
The body size is about 50-120 × 35-100 µm, with an average of 90 × 60 µm. The shape is broadly elliptical to elliptical with both ends rounded. The cell is distinctly dorsoventrally flattened (

Occurrence
Anoplophrya lumbrici was recorded only in the anecic L. terrestris at three localities: gardens at the Šúrska ulica street in Rendez and at the Jakubská ulica street in Rača as well as in floodplain soils in a riparian, willow-poplar forest near the Karlova Ves branch of the Danube River (Table 2). Ciliates were typically isolated from the middle part of the gastrointestinal tract, although very rarely some specimens were recorded also below this gut region.

Description
The body size is about 100-145 × 60-80 µm. The body shape is broadly elliptical to elliptical with both ends rounded. The cell is distinctly dorsoventrally flattened (Figs 10A, D-F, 11A-B).
The nuclear apparatus is composed of a single macronucleus and a single micronucleus. The macronucleus begins about 6 µm away from the anterior body end extends through the cell's midline into the posterior body region. The length of macronucleus varies from 80 to 110 µm and its width ranges from 9 to 18 µm. The macronuclear surface is smooth. The macronucleus diminishes in size leaving behind a conspicuous hyaline envelope in dying cells. The micronucleus is situated conspicuously far away from the macronucleus, namely, near the middle of the left body margin and opposite to the row of contractile vacuoles. The micronucleus is globular and approximately 4 µm in diameter (Figs 10A, D-F, 11A-B).
There is a single row of contractile vacuoles arranged along the right cell margin. It is composed of three or four vacuoles being 5-9 µm across during diastole (Figs 10A, D-F, 11A-B). The cytoplasm is colorless and studded with granules measuring approximately 0.5-1.0 µm in diameter. The cortex is semi-rigid and without specific granules. Swims moderately fast by rotation about the main body axis.
Somatic ciliature is holotrichous and composed of kineties meridionally extending over both cell sides.
There are on average 34 (24-45) kineties on each body side. Individual kineties are very narrowly arranged and the interkinetal distance ranges from about 0.6 to 2.0 µm. Likewise, basal bodies are very narrowly spaced within kineties and the intrakinetal distance is ca 1 µm. There is an apical and a terminal suture at each cell pole where individual somatic kineties commence and terminate, respectively (Figs 10A-C, 11B-C).

Occurrence
Anoplophrya vulgaris was recorded in two species of epigeic earthworms, viz., in E. fetida and D. veneta. Ciliate 18S rRNA gene sequences originated from both hosts were identical. Host earthworms came from a compost heap in the Botanical Garden of Comenius University and at the Jakubská ulica street in Rača as well as from humous soil with a high content of decomposing plant material from a garden at the foothill of the Malé Karpaty Mts. (Table 2). Endosymbiotic ciliates typically occurred in the middle part of the gastrointestinal tract. There were usually 10 to 15 endosymbionts per host.

Description
Only five specimens were found, three were morphologically examined and two were used for molecular analyses. Therefore, the description is rather incomplete. The body size is about 100 × 50 µm. The The macronucleus begins about 10 µm away from the anterior body end and extends through the cell's midline. Its size is usually 75 × 13 µm. The macronuclear surface was slightly irregular. The macronucleus displays similar postmortem changes as in the two previous Anoplophrya species, i.e., it slightly diminishes in size leaving behind a hyaline envelope. The micronucleus was not observed (Figs 11D-E, 12).
There are two staggered rows of contractile vacuoles extending along the right and left side of the macronucleus: 3-6 vacuoles in the right row and 3-5 vacuoles in the left row. The average size of vacuoles ranged from 5-7 µm during diastole (Figs 11D-E, 12). The cytoplasm is colorless and filled with granules being approximately 1 µm in diameter. The cortex is semi-rigid and without specific granules. Swims moderately fast by rotation about the main body axis. Somatic ciliature is holotrichous and composed of densely ciliated meridional kineties. Due to the low number of ciliates, their number on the ventral and dorsal side could not be determined.

Occurrence
Anoplophrya nodulata was detected only in two out of five specimens of Octolasion tyrtaeum investigated. This endogeic earthworm originated from the upper 50 cm peat layer in the riparian zone of the Rašelinisko Pond in the vicinity of the Pusté Úľany Village in the Galanta District (Table 2).
Endosymbiotic ciliates were found only in the central part of the oligochaete gastrointestinal tract. No other ciliates were recorded in the digestive system of O. tyrtaeum.

Multivariate morphometric analyses
Altogether, 16 quantitative features, a single qualitative character and two derived ratios were scored on 33 individuals belonging to four astome species: M. lumbrici, M. varians, A. lumbrici and A. vulgaris. Morphometric data were compiled in Table 3 and served to calculate Gower's pairwise similarities among ciliate specimens. Multidimensional scaling was performed on the pairwise Gower's coefficients, using the scikit-learn package in Python. Two mutually isolated groups, species of Metaradiophrya and Anoplophrya, were distinctly separated along the first and the second ordination axis (Fig. 13). However, individual species within each genus were not distinctly segregated. Specifically, two specimens of M. varians were intermingled with some M. lumbrici individuals, indicating that the intraspecific variability in most morphometric features of the latter species is so wide that M. varians at least partially falls within its range. Nevertheless, both species can be unequivocally distinguished by a single morphometric character, the number of fibers associated with the ventral side of the longer arm of the fibrillar hook. Specifically, there are 30-37 (on average 33) fibers in M. lumbrici and 22-29 (on average 26) fibers in M. varians (Table 3). OBERT T. & VĎAČNÝ P., Morphology and phylogeny of five astome ciliates 23 A similar mixed pattern was revealed also for the two species of Anoplophrya. Anoplophrya lumbrici was depicted as a comparatively more variable species, somewhat overlapping with the cluster of A. vulgaris. The latter species, however, showed a trend of at least partial separation from A. lumbrici in the ordination diagram (Fig. 13). The morphometric data indicate that A. vulgaris is slightly larger and narrower than A. lumbrici (100-145 × 60-80 µm vs 50-120 × 35-100 µm), although their size ranges overlap distinctly.

Molecular characterization and phylogenetic position of ciliates isolated from lumbricid earthworms
In total, 19 new 18S rRNA gene sequences were obtained from M. lumbrici (eight sequences), M. varians (four sequences), A. lumbrici (two sequences), A. vulgaris (three sequences) and A. nodulata (two sequences). Their length, GC content and GenBank accession numbers are summarized in Table 4. Intraspecies sequence similarities were 100%, except for M. lumbrici where one out of the 1762 nucleotide positions was polymorphic. Namely, two KR-specimens (10/1 and 10/C) had guanine at the position 796 while the rest of KR-exemplars as well as all other M. lumbrici individuals displayed adenine there.
Genetic distances estimated under the maximum composite likelihood model showed that the evolutionary divergence between M. lumbrici and M. varians is 0.0138 ± 0.0032. A slightly larger distance was revealed between A. lumbrici and A. vulgaris (0.0168 ± 0.0037). However, evolutionary divergences between species of Metaradiophrya and Anoplophrya were more than two times greater ranging from 0.0304 to 0.0380. Anoplophrya nodulata was revealed as the most divergent taxon from both species of Metaradiophrya (0.0608-0.0627) as well as from the two other species of Anoplophrya (0.0717-0.0752) ( Table 5).
To determine the phylogenetic position of these five ciliate species, Bayesian and maximum likelihood (ML) as well as neighbor-joining (NJ) analyses were conducted (Figs 14-15). The newly obtained sequences were classified into the paraphyletic subclass Astomatia of the class Oligohymenophorea. Paraphyly of the Astomatia was caused in that the astome Haptophrya planariarum (von Siebold, 1839) isolated from flatworms clustered with species of the scuticociliate genus Dexiotricha with full statistical support both in the Bayesian and ML analyses. All astomes isolated from annelids formed a strongly statistically supported monophylum (posterior probability 1.00, 99% ML bootstrap). Astomes from polychaetes (Durchoniella spp.) branched off first and were followed by a paraphyletic assemblage of astomes from endogeic oligochaetes (Almophrya de Puytorac et Dragesco, 1969, Anoplophrya nodulata, Eudrilophrya de Puytorac, 1969, Metaracoelophrya de Puytorac et Dragesco, 1969, Njinella Ngassam, 1983and Paraclausilocola Fokam et al., 2011. Lineages of astome ciliates from anecic (M. lumbrici and A. lumbrici) and epigeic (M. varians and A. vulgaris) earthworms were nested within the crown radiation of this paraphyletic cluster (Figs 14-15). Conklin, 1930), A. vulgaris and A. nodulata were fully statistically supported in Bayesian, ML and NJ analyses of the general and detailed alignment (Figs 14-15). However, monophyly of M. lumbrici was fully statistically supported only in NJ analyses (100% bootstrap) of the detailed alignment (Fig. 15) and strongly in ML analyses (91% bootstrap) of the general alignment (Fig. 14), while it was left unsupported in ML analyses of the detailed alignment and in the Bayesian inferences of both alignments. Specifically, specimens of M. lumbrici clustered together in Bayesian analyses of the general alignment, but with a statistically insignificant posterior probability of 0.91 (Fig. 14). On the other hand, in the Bayesian tree inferred from the detailed alignment, specimens of M. lumbrici were placed in a basal polytomy of a cluster containing a fully statistically supported clade of A. lumbrici and A. vulgaris. This whole assemblage, however, obtained a statistically insignificant posterior probability of 0.82 (Fig. 14) and its relevance  Table 2 and a locality code as specified in Table 1, followed by the ordinal number of the specimen investigated. For specimen codes and further details, see Table 4. The scale bar denotes eight substitutions per one hundred nucleotide positions. European Journal of Taxonomy 559: 1-37 (2019) 28 should be therefore taken with great caution. Identical topology was revealed also in ML analyses of the detailed alignment (-ln L = 5056.68) (data not shown). Due to the discrepancy between the topology of the NJ tree on the one hand and the Bayesian and ML trees on the other, statistical tree topology tests were conducted. They revealed that the monophyly of M. lumbrici and its sister-group relationship to M. varians (-ln L = 5059.33) could not be rejected ( Table 6). The conflict between the distance NJ tree and the Bayesian and ML trees was possibly caused by a plesiomorphic/homoplastic trap. Meteradiophrya lumbrici differed from M. varians in 25 nucleotide positions, but a comparison of 18S rRNA gene sequences of M. lumbrici with those of three species of Anoplophrya revealed that 19 out of the 25 variable nucleotide positions are either plesiomorphies or possibly homoplasies. Only six positions (36,76,100,119,120,141) appear as molecular apomorphies of M. lumbrici (Fig. 16). This indicates that distance techniques might avoid the plesiomorphic/homoplastic trap when sequences of related species are highly similar (genetic distance between M. lumbrici and M. varians is only 0.0138 ± 0.0032; Table 5). On the other hand, Bayesian inference and maximum likelihood might not be resistant to the plesiomorphic/homoplastic trap when plesiomorphies/homoplasies significantly Species Specimen a Host species Locality code b Length (nt) GC (%) GenBank entry outnumber apomorphies. Therefore, taxonomic discrepancies in phylogenetic trees need to be analyzed also using distance methods and statistical tree topology tests.

Discussion
The gastrointestinal tract of earthworms from the oligochaete family Lumbricidae is inhabited by astomes belonging to only three genera, namely, Metaradiophrya, Anoplophrya and Maupasella (Heidenreich 1935;Lom 1961;de Puytorac 1972). There are several dozens of species in each genus and their identification is difficult because of comparatively few diagnostic morphologic features and lack of information about intraspecific variability and host range (e.g., Cépède 1910;Heidenreich 1935;Beers 1938;Williams 1942;Lom 1961;de Puytorac 1972). Because of these problems, most records of astome ciliates from lumbricid earthworms need to be taken with caution. Moreover, our integrative approach revealed that a combination of detailed microscopic observation, molecular data as well as identity and ecological group of host organisms is indispensable for reliable identification of astomes.   Heidenreich (1935) also considered A. alluri Cépède, 1910, A. lloydii Ghosh, 1918and A. striata (Dujardin, 1841 as junior synonyms of A. lumbrici. In addition, he also found A. striata as synonyms of A. alluri and A. nodulata. According to Lom (1961), A. alluri is also a junior synonym of A. nodulata. Whether the two latter species are conspecific and also identical with A. simplex Nana et al., 2018 is difficult to decide at the present state of knowledge. However, synonymization of A. alluri and A. nodulata with A. lumbrici is not justified in the light of both morphological (two rows of contractile vacuoles vs a single row of vacuoles) and molecular (Figs 14-15) data. Cépède (1910) (Nana et al. 2018). Therefore, conspecificity of A. alluri, A. lloydii and A. simplex is unlikely, although all three species display two rows of contractile vacuoles.
The present phylogenetic analyses did not support monophylies of the genera Metaracoelophrya, Metaradiophrya and Anoplophrya (Figs 14-15). Monophyly of Metaracoelophrya was also firmly refuted by all statistical tree topology tests (Table 6). To solve this taxonomic problem, details on morphology of the three species of Metaracoelophrya are needed. Likewise, monophyly of M. lumbrici, M. varians and Metaradiophrya sp. HQ446279 was consistently rejected, although monophyly of M. lumbrici and M. varians could not be excluded by any statistical test (Table 6). However, examination of the dissertation thesis of Fokam (2012) revealed that Metaradiophrya sp. HQ446279 (= "Metaradiophrya simplex" in his thesis) is obviously not a Metaradiophrya, since it does not possess a fibrillar hook composed of two unequally long arms. Finally, polyphyly of the genus Anoplophrya is caused in that A. nodulata clusters far away from the A. lumbrici + A. marylandensis + A. vulgaris clade. As mentioned above, the genetic divergence of A. nodulata and the two other species of Anoplophrya is comparatively big (0.0717-0.0752) and even greater than from the two species of Metaradiophrya (0.0608-0.0627).
Since A. nodulata is distinguished from the two other species of Anoplophrya by the contractile vacuole pattern (two rows vs one row) and by the ecological group of host earthworms (endogeic vs anecic or epigeic), it should be transferred to a distinct genus. However, due to the lack of detailed morphological data on A. nodulata, we prefer not to establish a new genus in the present study.
It is important to emphasize that all literature data on occurrence of astome ciliates in the digestive tract of oligochaetes from the family Lumbricidae are problematic due to taxonomic difficulties and lack of molecular data. Our integrative approach revealed that reliable identification of astome ciliates also requires molecular data. Moreover, our phylogenetic analyses also suggest that astome ciliates might be associated with certain ecological and systematic groups of their host organisms. Therefore, occurrence of individual astome species in ecologically different and phylogenetically distant earthworms seems to be unlikely and needs to be corroborated by molecular data.