Phylogenetic and morphological studies in Xylodon (Hymenochaetales, Basidiomycota) with the addition of four new species

Abstract Xylodon (Hymenochaetales, Basidiomycota) is the largest segregate genus of Hyphodontia s.l. Based on molecular and morphological data, 77 species are accepted in Xylodon to date. Phylogenetic analyses of ITS and 28S sequences, including 38 new ITS and 20 28S sequences of Xylodon species, revealed four species new to science. The new taxa X.exilis, X.filicinus, X.follis and X.pseudolanatus from Taiwan, Nepal, Réunion, Belize, and USA are described and illustrated. In addition, species concepts for Odontiavesiculosa from New Zealand and Xylodonlanatus from U.S.A. are revised and the new name X.vesiculosus is proposed. Phylogenetic analyses of the ITS region placed X.spathulatus, X.bubalinus and X.chinensis in a strongly supported clade and demonstrated that they are conspecific. Palifer and Odontiopsis are synonymised under Xylodon based on morphological and sequence data. The following new combinations are proposed: X.erikssonii, X.gamundiae, X.hjortstamii, X.hyphodontinus, X.septocystidiatus and X.verecundus. Line drawings of X.cystidiatus, X.hyphodontinus, X.lanatus and X.vesiculosus, as well as photographs of X.raduloides basidiomata, are provided. A key to X.lanatus and similar species is presented.

The most recent generic description of Xylodon was published by Riebesehl and Langer (2017). With few exceptions, the hymenophore in Xylodon is odontioid or poroid with many different cystidia types and basidiospore shapes.
Palifer Stalpers & P.K.Buchanan (1991), based on Peniophora verecunda G.Cunn. from New Zealand, is another segregate genus of Hyphodontia s.l. recognised by Hjortstam and Ryvarden (2009). It is characterised by encrusted cystidia and remained monotypic until 2007 when three species were transferred to the genus (Hjortstam and Ryvarden 2007a). After a thorough morphological study of Palifer species and related taxa, Gorjón (2012) concluded that Palifer was probably a synonym of Xylodon but did not propose any new combinations. Palifer is represented by only one nuclear ribosomal internal transcribed spacer (ITS) sequence in the public record and phylogenetic analyses showed it to be embedded in Xylodon (Larsson et al. 2006). Riebesehl and Langer (2017), however, declined to synonymise Palifer with Xylodon based on one DNA sequence alone and chose to emphasise its morphological features.
In this study, we conducted an in-depth phylogenetic study of 36 Xylodon species represented by 96 strains or collections, including 58 new ITS and large subunit (28S) ribosomal DNA sequences. Phylogenetic analyses of the ITS and 28S sequence data uncovered four new taxa, X. exilis, X. filicinus, X. follis and X. pseudolanatus, that are described and illustrated. In addition, the species complex of X. spathulatus was identified and resulted in the synonymisation of two taxa. The genera Palifer and Odontiopsis are re-evaluated and placed in synonymy with Xylodon, resulting in a number of new combinations. Morphological studies in Xylodon lanatus and Odontia vesiculosa were conducted and a key to morphologically similar species is provided. Line drawings of X. cystidiatus, X. hyphodontinus and X. vesiculosus are presented and X. vesiculosus is described.

Molecular study
Pieces of dried basidiomata served as material for DNA extractions with the E.Z.N.A.® Fungal DNA Mini Kit (Omega Bio-Tek, VWR, USA). Two nuclear ribosomal DNA markers were used in this study: the ITS region and the D1-D2 domains of 28S. The ITS region includes the internal transcribed spacers 1 and 2 as well as the intercalary 5.8S rRNA gene. For amplification of ITS, different combinations of the following primers were used: ITS1-F (Gardes and Bruns 1993), ITS1, ITS2, ITS3, ITS4, ITS5 (White et al. 1990) and ALR0 (Collopy et al. 2001). The last one was modified in one position . NL1, NL4 (O'Donnell 1993), LR0R (Bunyard et al. 1996) and LR5 (Vilgalys and Hester 1990) were used, also in different combinations, for the amplifications of the D1-D2 domains of 28S. PCR products were purified with innuPREP PCRpure Kit (Analytik Jena, Berlin, Germany) and the DNA sequencing was implemented by Eurofins Genomics (Ebersberg, Germany).
Newly generated sequences were edited with MEGA7 (Kumar et al. 2016). Their quality was checked following the five guidelines by Nilsson et al. (2012) and they were deposited in NCBI GenBank (Benson et al. 2018;Tab. 1). Other sequences used in this study were downloaded from the same database. Phellinus gabonensis Decock & Yombiyeni (Hymenochaetales) was chosen as the outgroup for rooting the phylograms. The two different alignments were calculated with MAFFT v.7 (Katoh and Standley 2013), using the L-INS-i algorithm for ITS and G-INS-i for 28S. Minimum Evolution (ME) and Bayesian inference (BI) trees were calculated for both datasets. The ME phylograms were computed with MEGA7, using the Tamura-Nei model (Tamura and Nei 1993) including 1000 bootstrap (BS) replications, partial deletion of gapped positions with 95% site coverage cut-off and other default settings. The BI phylograms were constructed with MrBayes 3.2.6 (Ronquist and Huelsenbeck 2003), using DNA substitution models estimated by MrModeltest 2.4 (Nylander 2004) with 10 million generations and one tree saved for every 1000 generations; other parameters used default settings. A partitioned analysis was done for the ITS alignment with independent DNA substitution models and parameter values for ITS1, 5.8S and ITS2. MEGA7 and FigTree 1.4.2 (Rambaut 2012) were used for processing the phylograms. Table 1. List of accepted species in Xylodon with some closely related species from other genera, including specimens used in the phylogenetic study. Newly generated sequences are shown in bold. Xylodon species without available ITS or 28S sequences are marked with 'not available' (n.a.); these have to date not been studied using ribosomal sequence data.

Morphological study
The studied specimens are deposited in herbaria CFMR, FR, KAS, LIP, MSK, and TUB (acronyms follow Index Herbariorum, http://sweetgum.nybg.org/science/ih). Morphological descriptions and figures employed dried basidiomata. Preparations in 3% potassium hydroxide (KOH) aqueous solution were used for microscopic measurements and most drawings. Crystalline deposits on hyphae were additionally examined in Melzer's reagent (Mz) and tap water. Amyloid and dextrinoid reactions of basidiospores were tested with Mz. Spore wall cyanophily was determined in Cotton Blue-Lactophenol solution (CBL). The following abbreviations are used to describe arithmetic averages for 30 basidiospores, randomly selected in squash preparations of one specimen: Lspore length, Wspore width, Q = length/width ratio.

Phylogeny
The aligned ITS data matrix consisted of 92 taxa and 847 positions. The partial deletion of gapped positions resulted in 463 positions that were used in the ME phylogenetic analysis. The data matrix was partitioned as follows: ITS1 = positions 1-373, 5.8S rRNA gene = 374-541 and ITS2 = 542-847. The GTR + G model was used as DNA substitution model for ITS1 and ITS2 and SYM + I for 5.8S in the BI analysis.
The aligned data matrix of the D1-D2 domains of 28S rRNA gene consisted of 47 taxa and 634 positions; 532 positions were used in the ME analysis. The GTR + I + G model was chosen as the DNA substitution model for the BI analysis. A high degree of agreement was observed between the ME and BI trees; therefore, the ME phylogram with BS and integrated posterior probability (PP) values from the BI phylogram are presented in Figures 1, 2. The sum of branch lengths in the resulting ME phylograms was 3017 for ITS ( Fig. 1) and 1009 for 28S (Fig. 2). Multiple sequence alignments and trees are deposited in TreeBASE (http://purl.org/phylo/treebase/phylows/study/ TB2:S23512). The ITS phylogram (Fig. 1) shows a number of clades with low BS support which is well-documented for Xylodon (see Chen et al. 2018, Riebesehl and. Figure 1 includes 83 sequences of Xylodon specimens or strains of which 38 were generated in this study. No significant distances were observed amongst sequences of X. niemelaei, X. rhizomorphus, and X. reticulatus nor amongst sequences of X. spathulatus, X. chinensis, and X. bubalinus. The strong BS (99 and 97) and PP (1) values of these two clades indicate that the taxa within each clade may be conspecific. Seven collections of X. raduloides form two distinct subclades (96 and 98 BS). Similarly, four collections of X. flaviporus formed two subclades (99 and 93 BS). The newly generated ITS sequences show that X. cystidiatus, X. hyphodontinus, X. serpentiformis and X. subclavatus form distinct lineages in Xylodon. The four new species introduced herein form   Table 1. distinct lineages as well. Xylodon pseudolanatus and X. exilis are sister groups with 5.9% differences between their ITS sequences (X. pseudolanatus: FP-150922 and X. exilis: TUB-FO 42565). They cluster in a well-supported clade (99 BS), within a weakly supported lineage that includes X. follis. The closest relative of X. filicinus is X. hastifer; they form a clade (93 BS) that is sister to X. hyphodontinus s.l.
The 28S phylogram (Fig. 2) includes 39 sequences of Xylodon of which 20 were generated here. Notable new 28S sequences include X. australis, X. hyphodontinus, X. serpentiformis, the four new species described herein, and furthermore Hyphodontia borbonica. As the 28S phylogram features several lineages with low BS support, the clades between the 28S and ITS trees are not identical throughout. Although clearly resolved with ITS sequences, the 28S gene analyses were not able to resolve the closely related X. raduloides and X. subtropicus. Some clades that were well supported with ITS sequences were also well supported in the 28S phylogram, for example, the X. niemelaei and X. reticulatus (100 BS) and the X. chinensis and X. spathulatus (99 BS) clades.

Xylodon filicinus
Distribution and ecology. From the lower mountainous belt in Taiwan, on dead fern rachises.
Etymology. from Latin filix -fern, refers to the occurrence on dead fern rachises. Additional specimen examined. TAIWAN, the same locality and the same substrate as holotype, leg. E. Yurchenko, 31 Jul 2011 (field No. 18;MSK-F 12870;dup. in KAS).

Xylodon follis
Distribution and ecology. The species is so far known from Réunion (Mascarene Archipelago) and inhabits dead wood.
Etymology. from Latin follis -bag or bubble, referring to shape of the spores, basidioles and capitate cystidia found in this species.

Xylodon pseudolanatus
Distribution and ecology. South-eastern USA and Central America, on dead wood of angiosperms.
Etymology. From Greek pseudo--false, refers to its similarity to X. lanatus. Remarks. The diagnostic features of this species are the minutely odontioid hymenophore, bundles of sparsely to moderately encrusted hyphae, projecting from aculeal apices, embedded capitate cystidia, cylindrical basidia and narrowly ellipsoid basidiospores. Some hymenial elements in this species are intermediate in morphology between basidioles, capitate cystidia and hyphal ends. Xylodon pseudolanatus can be distinguished from similar species in the key below (see Discussion). Remarks. This new combination is based on the phylogenetic analyses of the ITS and 28S sequences as well as morphological study of specimens, including the holotype of O. hyphodontina. Originally, the collections from Martinique and French Guyana were identified as O. ambigua, but the molecular data clearly show that these collections are embedded in Xylodon (Figs 1, 2). Although H. ambiguum is the oldest name for this taxon, it cannot be transferred to Xylodon because the name is preoccupied by X. ambiguus (Peck) Kuntze ( = Veluticeps ambigua (Peck) Hjortstam & Telleria). Odontiopsis ambigua, P. sphaericosporus and O. hyphodontina were recognised as conspecific by Hjortstam (1987Hjortstam ( , 1991. Odontiopsis hyphodontina is the next oldest name and is chosen to represent this taxon. As O. hyphodontina is also the type of Odontiopsis Hjortstam & Ryvarden, Odontiopsis concomitantly becomes a synonym of Xylodon.

Xylodon hyphodontinus (Hjortstam & Ryvarden) Riebesehl
The newly generated ITS and 28S sequences of X. hyphodontinus hold comparable positions in a clade that includes three distinct lineages in both phylogenetic trees (Figs 1, 2). Specimens KAS-GEL9222 from Kenya and LIP GG-GUY13-044 from French Guyana each represent distinct lineages from the third lineage of LIP GG-MAR15-127 and LIP GG-MAR12-238 from Martinique. As species in Hyphodontia s.l. can be readily distinguished with ITS or 28S sequences, these three lineages should result in the recognition of three different species. However, we were not able to identify any definite morphological differences amongst the lineages in comparison with the holotype material from Tanzania. Cultures are not available for these specimens, thus intercompatibility tests are not possible. As a result, we decided to treat all three lineages as X. hyphodontinus at this time.  Feb 1954 (PDD-18112 -holotype). Cunningham (1959) described this new taxon as Odontia vesiculosa G. Cunn. Earlier, Odontia vesiculosa Burt was used for another species (Povah 1929). Consequently, Odontia vesiculosa G. Cunn. is an illegitimate name and a new name is required for this taxon (see Art. 6.11, 7.4 and 58.1 in Turland et al. 2018).
Remarks. This species was considered conspecific with Xylodon lanatus from North America (Burdsall and Nakasone 1981, Wu 1990, Gorjón and Greslebin 2012), but we observed significant morphological differences. For example, in X. vesiculosus, the basidiomata have a denser, tough-membranaceous texture compared to the soft woolly basidiomata of X. lanatus. In addition, the aculei in X. vesiculosus are larger and the basidia are thin-walled in contrast to the smaller aculei and basally thick-walled basidia found in X. lanatus (compare Figs 9, 10). See Discussion for a key to X. lanatus, X. vesiculosus and allied taxa. Remarks. Molecular and morphological analyses demonstrate that the three taxa listed above are very similar. The 11 samples of X. niemelaei, 3 of X. rhizomorphus and 3 of X. reticulatus formed a strongly supported clade (99 BS, 1 PP) in the ITS phylogram (Fig. 1). In addition, three samples representing two of the species are found in a strongly supported clade (100 BS, 1 PP) in the 28S tree (Fig. 2), differing in only one position in the associated alignment.  Xylodon niemelaei was described and illustrated in detail by Wu (1990) and Langer (1994). It is characterised by a poroid hymenophore, embedded and hymenial capitate cystidia, small, subulate or fusoid hymenial cystidia and encrusted hyphal ends mainly developed at the pore edges but sometimes also in other areas. At the morphological level, the bladder-like embedded cystidia and hyphal encrustations appear identical in X. niemelaei (Langer 1994), X. rhizomorphus (Zhao et al. 2014) and X. reticulatus . Spore size and spore quotient overlap in these three species. Xylodon rhizomorphus occurs in south-western China, whereas X. reticulatus occurs in Taiwan and Japan. Xylodon niemelaei is reported also from these three countries and furthermore from Réunion, Africa and South America (Langer 1994), but the last two reports require morphological and molecular confirmation. Remarks. Based on both molecular data and morphology, we place the taxa X. bubalinus and X. chinensis in synonymy under X. spathulatus. In our phylogenetic analysis of ITS sequence data, the recently described X. bubalinus (4 collections) and X. chinensis (2 collections) from China form a well-supported clade with X. spathulatus (4 collections) from Europe (97 BS, 1 PP) that is sister to X. apacheriensis (Fig. 1). Within this clade are several subclades, with very low bootstrap support (<55), thus subspecies or varieties cannot be identified. The 28S rRNA gene analysis also supports conspecificity between X. chinensis and X. spathulatus (99 BS, 0.65 PP) (Fig. 2). Xylodon spathulatus has three main diagnostic features: prominent (1-2 mm tall) aculei of varied shape, numerous apically acute cystidia with 1 -4 slight constrictions and capitate cystidia with a resinous cap. It is described and illustrated by Eriksson and Ryvarden (1976) and Langer (1994). Minor morphological variation amongst the three taxa was observed. For example, X. chinensis has ventricose cystidia, similar to those in X. spathulatus, but they are sometimes septate at the constrictions. Distinctly ventricose cystidia were not observed in X. bubalinus, which instead had hyphoid or subulate cystidioles (Wang and Chen 2017, Fig. 2f ). Encrusted hyphal ends at apices of the aculei in X. bubalinus and X. chinensis are typical of those in X. spathulatus. Resinous caps enclosing capitate elements are often absent as in the case of X. bubalinus, X. chinensis, X. spathulatus KAS-GEL2690 (from Germany) and X. spathulatus MSK-F 12931 (from Russia). Spore shape and size are similar amongst the three taxa and the spore quotient 1.3 -1.4(-1.5) overlaps (Eriksson and Ryvarden 1976, Wang and Chen 2017. A few spores in X. chinensis were up to 6 × 5 μm and may be due to better climatic conditions. The description of X. spathulatus is modified to include variable aculei from conical and subulate to distinctly spathuliform and the variable presence of cystidia with resinous caps, mucronate apices and a submoniliform type that are aseptate with more or less blunt apices. Thus X. spathulatus is a highly variable but distinctive species that is widely distributed from northern Europe (Eriksson and Ryvarden 1976) to southern China  and has a preference for old-growth forests (Dvořák et al. 2017). Reports of X. spathulatus from North and South America (Ginns andLefebvre 1993, Hjortstam andRyvarden 2007b) should be confirmed by molecular sequence data. ≡ Schizopora cystidiata A. David & Rajchenb., Mycotaxon 45:140 (1992).

Discussion
We recognise 77 species of Xylodon based on studies by Riebesehl and Langer (2017), Chen et al. (2017Chen et al. ( , 2018, Kan et al. (2017), Wang and Chen (2017) and results herein (Table 1). Our phylogenetic analyses included 122 ITS and 28S sequences representing 37 Xylodon species. The other 40 accepted species in Xylodon are based on morphological studies.
In the following discussion, we highlight some of the significant results.

Odontiopsis is a synonym of Xylodon
The monotypic Odontiopsis Hjortstam & Ryvarden was described in 1980 based on O. hyphodontina from Tanzania. Hjortstam (1987) transferred Hydnum ambiguum to Odontiopsis and placed O. hyphodontina in synonymy. Later, Pteridomyces sphaericosporus was placed in synonymy with O. ambigua by Hjortstam (1991). Analyses of ITS and 28S sequences placed specimens originally identified as O. ambigua in the Xylodon lineage. Due to nomenclature rules to choose the earliest possible epithet to represent a taxon (see Art. 11.4 in Turland et al. 2018), the name for this taxon is Xylodon hyphodontinus and Odontiopsis is reduced to a synonym of Xylodon.

Palifer is a synonym of Xylodon
Species in Palifer have apically encrusted cystidia that are characteristic of the genus and distinctly different from the lagenocystidia of Hyphodontia s.s. Ryvarden 2009, Riebesehl and. Palifer is defined primarily by morphology because there is only a single ITS sequence available. Phylogenetic studies place P. verecundus amongst the Xylodon species (Larsson et al. 2006, Fig. 1). The recently described X. mollissimus has cystidia that are similar to those of Palifer species and ITS sequence analyses place it in a clade with Xylodon sp. 1 (Kan et al. 2017, Fig. 1). Although not closely related, X. mollissimus and P. verecundus are embedded within Xylodon and demonstrate that the distinctive cystidia developed in Palifer is not a phylogenetically significant character. Thus, we reduce Palifer to a synonym of Xylodon and propose the following transfers:

Xylodon lanatus and allied species
Xylodon lanatus was originally described by Burdsall and Nakasone (1981) based on collections from North America and New Zealand. A comparative morphological study of specimens, annotated as X. lanatus from Taiwan and North America, revealed that X. lanatus is a complex of morphologically similar species. The New Zealand specimen, X. vesiculosus, was discussed above and is considered to be a distinct species. The specimen X. lanatus  from Taiwan, depicted in Langer (1994), is X. exilis. The specimen of X. lanatus cited by Hjortstam and Ryvarden (1984) from Nepal is also X. exilis. In the protologue of X. lanatus (Burdsall and Nakasone 1981), the authors illustrated the paratype (HHB-6925 from Florida, U.S.A.) which is correctly identified as X. pseudolanatus. Hjortstam and Bononi (1987) reported X. lanatus from Brazil while noting the controversial taxonomic position of this species.
We accept X. lanatus, based on the type (CFMR HHB-8925; Figs 3f, 10) and paratype (CFMR HHB-4305), as a distinct species but with a restricted concept. We retain the same diagnostic features, noted in the protologue (basidiomata with welldeveloped woolly subiculum, terminal vesicular structures on subicular hyphae, poorly differentiated subhymenium, encrusted thick-walled hyphae in tooth apices and capitate cystidia) and add that walls of basidia and subhymenial hyphae directly under hymenial elements are slightly but distinctly thickened. The illustration of Xylodon lanatus from Taiwan, provided by Wu (1990), also shows basidia with walls thickened below, but hyphal pegs appear different from X. lanatus s.s. We have also determined that X. echinatus (Yurchenko et al. 2013) is the most morphologically similar species to X. lanatus s.s. A key to the taxa in the X. lanatus group is presented here.  (Figs 1, 2). ITS sequences from holotypes of X. reticulatus and X. rhizomorphus were included in the analyses (Fig. 1). The ITS sequences were 98.2-99% similar amongst the taxa, differing at 6-11 sites. Minor morphological differences were noted amongst the taxa. We keep the taxa X. niemelaei, X. reticulatus and X. rhizomorphus as separate species following the results of phylogenetic studies of Chen et al. (2017) and Fernández-López et al. (2018a). The last work was published shortly before the completion of our study and therefore could not be considered further. However, in our reconstruction, the phylogenetic distances between these taxa are very short, and comparable to those between the OTUs of X. spathulatus. Taking into account that these three taxa remain as monophyletic branches, we suppose that they can be subspecies or varieties of one species.
Xylodon bubalinus, X. chinensis and X. spathulatus are conspecific Phylogenetic analyses of ITS sequences of 10 samples, including sequences from holotypes of Xylodon bubalinus and X. chinensis and 28S sequences of X. spathulatus and holotype of X. chinensis show that the three taxa are conspecific (Figs 1, 2). Amongst the taxa, ITS sequences were 98.7-99.8% similar, differing at up to 8 sites. The hymenophore is quite variable in this group and the presence of the different types of cystidia is also variable. The correct name for this group is X. spathulatus with X. bubalinus and X. chinensis reduced to synonyms.

The classification of Xylodon australis in Xylodon is confirmed
Xylodon australis is sequenced for the first time and shown in the 28S phylogenetic tree (Fig. 2) and its placement in Xylodon is confirmed. The sequenced specimen is from Argentina and was studied by Greslebin et al. (2000). They reported differences in spore morphology in specimens from Argentina, Australia and New Zealand. A molecular study may be able to resolve this species complex.
The paratype material of Xylodon dimiticus may be an independent species Chen et al. (2017) showed that the holotype material of Xylodon dimiticus (Jia J.Chen & L.W.Zhou) Riebesehl & Langer Dai 11686 is conspecific with X. nongravis. In addition, they also proposed that the paratype material Dai 15321 may be an independent species as shown in the phylograms. We support this view and included Dai 15321 in our phylo-grams as Xylodon sp. 1. The NCBI BLAST search of the ITS sequence of Dai 15321 shows an identity of 93% with the X. dimiticus holotype material Dai 11686 as well as with sequences of X. nongravis. Although this low similarity value indicates that Dai 15321 is a different species, a further study is needed to identify morphological differences.

Additions to the distribution and morphology of Xylodon serpentiformis
A BLAST search of the newly generated Xylodon serpentiformis ITS sequences revealed that they are 99% identical to a sequence from South Korea identified as Hyphodontia sp. (KUC20121019-31, Jang et al. 2016). Xylodon serpentiformis is known from Taiwan and the Canary Islands (Langer 1994). Based on the small distance between Taiwan and South Korea and the similarities of the sequences, the distribution of X. serpentiformis is expanded to include South Korea. Langer (1994) cited a specimen of X. serpentiformis from the Canary Islands, but this material needs molecular confirmation. A distinctive feature, described for this species, was the presence of flexuous, thick-walled tubular tramacystidia in the aculei (Langer 1994). After our morphological analysis of the holotype of Hyphodontia serpentiformis Langer (TUB-FO 40677) and three more specimens (TUB-FO 40675, TUB-FO 40985, TUB-FO 42688), we emend the diagnosis of X. serpentiformis as follows: aculei consisting mostly of flexuous, agglutinated projecting hyphae, hyphae slightly thickened or moderately thickwalled at base, then thinning toward apex and partly collapsing at maturity; spores broadly ellipsoid, ellipsoid, subovoid, sometimes narrowly ellipsoid.

Sequences of Xylodon raduloides form two subclades in phylogenetic trees
The two subclades of Xylodon raduloides in the ITS phylogeny ( Fig. 1) appear also in the 28S phylogram (Fig. 2), although with the inclusion of the sister species X. subtropicus. Both subclades include specimens from Germany and Australia. Micromorphological distinctions between the clades were not observed, but we noted that, in one of the clades (KAS-JR 02, 03, 09, 26), the hymenophore is pale cream to cream whereas, in the second clade (KAS-JR10, LR 18813), it is yellow-brownish (Fig. 13). Nevertheless, the morphological as well as sequence-based differences are not sufficient to recognise two separate species.

Sequences of Xylodon flaviporus form two subclades in phylogenetic trees
In the ITS phylogram, two subclades are also present in the Xylodon flaviporus lineage (Fig. 1). Subclade 1 (FR-0249797, KAS-GEL 5047) comprises specimens from Réunion whereas subclade 2 (FCUG 1053, KAS-GEL 3462) comprises specimens from Romania and Taiwan. All other ITS sequences of X. flaviporus from the NCBI GenBank are from the northern hemisphere (Romania, South Korea, Taiwan, Turkey and USA) and clustered together in subclade 2 (data in Fig. 1 are reduced to specimens from Romania and Taiwan). We did not find micromorphological differences between specimens of the two subclades and the variation in the ITS sequences is too small to merit recognition of two different species.

Xylodon ramicida and X. quercinus
Xylodon ramicida and X. quercinus are morphologically similar (Ariyawansa et al. 2015), exhibiting slight differences in spore width and shape and different substrate preferences. Their ITS sequences are 98.8-99% similar, differing at just 6 -7 positions (X. quercinus: Miettinen 15050 and X. ramicida: Spirin 7664). Taking into account the similarity values of other Xylodon species (X. spathulatus 98.7-99%, X niemelaei 98.2-99%) and small morphological differences between X. ramicida and X. quercinus, we believe that X. ramicida is a well-defined subspecies within X. quercinus. More sequences from both taxa are required, however, before the taxonomic status of X. ramicida can be clarified.

The taxonomic status of Xylodon detriticus
In this study, we accept Hyphodontia detritica (Bourdot) J. Erikss. in Xylodon as X. detriticus. This combination was introduced by Ţura et al. (2011), recognised as invalid in MycoBank (Art. 36.1a and b, Melbourne Code) and supported in the work by Rosenthal et al. (2017). The first sequenced specimen of this species was GB Nilsson 990902 (Larsson 2007). We have studied this specimen and it is identical to the concept of Hypochnicium detriticum (Bourdot) J. Erikss. & Ryvarden (Eriksson and Ryvarden 1976). The alignment between GB Nilsson 990902 and X. detriticus UC2023108 (Rosenthal et al. 2017) in ITS2 (ITS1 is unavailable for the previous specimen) showed nearly 100% similarity. We discovered that the ITS2 sequences between Lagarobasidium detriticum MA-Fungi 5758 (Dueñas et al. 2009) and GB Nilsson 990902 were 62% identical. Consequently, the taxonomic identity of Lagarobasidium detriticum MA-Fungi 5758 needs to be investigated. Viner et al. (2018) came to the same conclusion, but we could not integrate their further results, because this study was already finished when the work of Viner et al. was published.

Conclusion
The usefulness of ITS sequences alone in defining and identifying species in Xylodon is approaching its limits. Further studies in Xylodon will require sequences from additional genetic markers with more variation. Fernández-López et al. (2018b) published the first phylogenetic tree for Xylodon with rpb2 sequences, but it contains only sequences of six different species. Nevertheless, the topology is very similar to our ITS and 28S trees.
Morphological features for defining species in Xylodon is also limited. Species, such as X. spathulatus with its variability in aculei morphology and in cystidia occurrence and shape, present challenges for identification. In other cases such as X. hyphodontinus, ITS sequence differences are significant whereas morphological differences are elusive.