Towards a Natural Classification of Hyphodontia Sensu Lato and the Trait Evolution of Basidiocarps within Hymenochaetales (Basidiomycota)

Hyphodontia sensu lato, belonging to Hymenochaetales, accommodates corticioid wood-inhabiting basidiomycetous fungi with resupinate basidiocarps and diverse hymenophoral characters. Species diversity of Hyphodontia sensu lato has been extensively explored worldwide, but in previous studies the six accepted genera in Hyphodontia sensu lato, viz. Fasciodontia, Hastodontia, Hyphodontia, Kneiffiella, Lyomyces and Xylodon were not all strongly supported from a phylogenetic perspective. Moreover, the relationships among these six genera in Hyphodontia sensu lato and other lineages within Hymenochaetales are not clear. In this study, we performed comprehensive phylogenetic analyses on the basis of multiple loci. For the first time, the independence of each of the six genera receives strong phylogenetic support. The six genera are separated in four clades within Hymenochaetales: Fasciodontia, Lyomyces and Xylodon are accepted as members of a previously known family Schizoporaceae, Kneiffiella and Hyphodontia are, respectively, placed in two monotypic families, viz. a previous name Chaetoporellaceae and a newly introduced name Hyphodontiaceae, and Hastodontia is considered to be a genus with an uncertain taxonomic position at the family rank within Hymenochaetales. The three families emerged between 61.51 and 195.87 million years ago. Compared to other families in the Hymenochaetales, these ages are more or less similar to those of Coltriciaceae, Hymenochaetaceae and Oxyporaceae, but much older than those of the two families Neoantrodiellaceae and Nigrofomitaceae. In regard to species, two, one, three and 10 species are newly described from Hyphodontia, Kneiffiella, Lyomyces and Xylodon, respectively. The taxonomic status of additional 30 species names from these four genera is briefly discussed; an epitype is designated for X. australis. The resupinate habit and poroid hymenophoral configuration were evaluated as the ancestral state of basidiocarps within Hymenochaetales. The resupinate habit mainly remains, while the hymenophoral configuration mainly evolves to the grandinioid-odontioid state and also back to the poroid state at the family level. Generally, a taxonomic framework for Hymenochaetales with an emphasis on members belonging to Hyphodontia sensu lato is constructed, and trait evolution of basidiocarps within Hymenochaetales is revealed accordingly.


Introduction
Hyphodontia was erected with Gonatobotrys pallidulus as the generic type in 1958 [1] and was put in the family Chaetoporellaceae that is typified by Chaetoporellus [2]. Since its erection, the genus name Hyphodontia has been accepted worldwide and widely used in many papers. For this reason, Chaetoporellus, Grandinia, Kneiffiella, Lyomyces, Schizopora and Xylodon, once considered to be former synonyms of Hyphodontia, were all rejected against

Phylogenetic Analyses
Besides the newly generated sequences, additional reliable sequences downloaded from GenBank (Tables S1 and S2) were also incorporated in the datasets for phylogenetic analyses. Seven datasets were employed. (1) The ITS dataset was used to differentiate the species identities of studied specimens belonging to Hyphodontia sensu lato. All vouchers belonging to Hyphodontia sensu lato as listed in Table S1 were included and Auricularia cornea was included as an outgroup taxon. (2) The combined dataset of ITS, nLSU and mt-SSU regions was used to explore the phylogenetic relationships of members belonging to Hyphodontia sensu lato with each other and other main lineages within Hymenochaetales. Table S1, each with at least both ITS and nLSU sequences available, were included as ingroup taxa. Moreover, even nLSU sequences unavailable, the generic types Chaetoporellus latitans (a synonym of Kneiffiella abdita) and Hyphodontia pallidula were also included to confirm phylogenetic position of these genera. In addition, members of Polyporales and Thelephorales listed in Table S1 were included also as additional ingroup taxa, and Auricularia cornea was included as an outgroup taxon. (3)(4)(5) As certain recently speciated wood-inhabiting fungi may have nearly congruent ribosomal DNA sequences [44], three multilocus (six to seven genes) combined datasets for Hyphodontia, Lyomyces and Xylodon were separately used to supplementally differentiate species identities within each of these genera. All newly studied specimens with at least two genes available from these three genera, and the collections of Xylodon with tef1α and rpb2 sequences downloaded from GenBank were included (Table S1). (6) The combined dataset of ITS, nLSU, tef1α, rpb1 and rpb2 regions was selected for estimating divergence times of families within Hymenochaetales. The vouchers of (described or undescribed) species with at least two (mostly three to five) of these five genes available in families of Hymenochaetales were included (Table S1), besides those having been normally used for molecular clock analysis in previous analyses and representing all main lineages in Basidiomycota (Table S2). Neurospora crassa from Ascomycota was designated as an outgroup taxon. (7) Finally, the combined dataset of ITS, nLSU and mt-SSU regions was employed for ancestral state reconstruction. A single example of each (described and undescribed) species with sequences of all these three genes available and a confirmed taxonomic position at the family level within Hymenochaetales was selected from the dataset (2) with Auricularia cornea as an outgroup taxon (Table S1).

All vouchers of Hymenochaetales listed in
All datasets were aligned using MAFFT 7.110 [57] under the G-INS-i option [58]. Regarding combined multilocus datasets, each locus was aligned separately and then concatenated as a single alignment. All resulting alignments were deposited in TreeBASE (http://www.treebase.org; accessed on 18 May 2021). jModelTest [59,60] with calculation under Akaike information criterion was used to estimate the best-fit evolutionary model for each alignment subjected to phylogenetic analysis.
Maximum likelihood (ML) and Bayesian inference (BI) algorithms were utilized for phylogenetic analyses of the alignments of datasets (1-5). The ML algorithm was conducted using raxmlGUI 1.2 [61,62] with the calculation of bootstrap (BS) replicates under the auto FC option [63]. The BI algorithm was conducted using MrBayes 3.2 [64]. Two independent runs were employed, each including four chains and starting from random trees. The first 25% of the sampled trees every 1000th generation were removed, and the other 75% of trees were retained for constructing a 50% majority consensus tree and calculating Bayesian posterior probabilities (BPPs). Tracer 1.5 (http://tree.bio.ed.ac.uk/software/ tracer/, accessed on 18 May 2021) was used to judge whether chains converged.
Molecular clock analysis for the alignment of dataset (6) was performed using BEAST v2.6.0 [65]. The lognormal relaxed molecular clock model and the Yule speciation prior were set to evaluate the divergence times and their corresponding credibility intervals. Four time points were selected for calibration: (1) 90 million years ago (Mya) representing the minimum age of Agaricales by Archaeomarasmius leggetti, a fossil agaricoid species preserved in a Dominican amber [66,67]; (2) 125 Mya representing the minimum age of Hymenochaetaceae by Quatsinoporites cranhamii, a fossil poroid species collected from Apple Bay on Vancouver Island [68,69]; (3) 400 Mya representing the divergence time between Ascomycota and Basidiomycota by Paleopyrenomycites devonicus, a fossil fungi found in Great Britain [70,71]; and (4) 290 Mya representing the mean age of Agaricomycetes by the analyses of genome data [72]. According to these time points, the offset age with a gamma distribution prior (scale = 20, shape = 1) for Agaricales was set as 90 Mya, for Hymenochaetaceae as 125 Mya, and for Basidiomycota as 400 Mya, while the mean age with a normal distribution prior (SD = 1) for Agaricomycetes was set as 290 Mya. After 400 million generations, the first 10% of the sampled trees every 1000th generation were removed as burn-in. The resulting log file was checked for chain convergence using Tracer 1.5.
A consensus tree for the alignment of dataset (7) was generated by BI algorithm with 50 million generations and the first 10% of the sampled trees every 1000th generation as burn-in using BEAST v1.10.4 [73] and then used for ancestral state reconstruction. The resulting log file was checked for chain convergence using Tracer 1.5. The trait evolution of basidiocarps was evaluated using RASP 4.2 under the Bayesian Binary MCMC model [74,75]. Two kinds of morphological traits of basidiocarps were set for selected species: one is the basidiocarp shape, including pileate, pileate-resupinate and resupinate states; the other is hymenophoral configuration, including poroid, smooth, grandinioid and odontioid states.
The ITS dataset (1) including 582 collections resulted in an alignment of 925 characters with GTR+I+G as the best-fit evolutionary model. The ML search stopped after 350 BS replicates. In BI, after 50 million generations with an average standard deviation of split frequencies of 0.004289, all chains converged which was indicated by the effective sample sizes (ESSs) above 1400 and the potential scale reduction factors (PSRFs) close to 1.000. ML and BI algorithms generated similar topologies in main lineages with minor differences in statistical supports. Therefore, the tree generated by the ML algorithm was presented along with BS value above 50% and BPPs above 0.8 at the nodes ( Figure 1). In general, the ITS-based phylogeny delimited species well; at the generic rank, Hyphodontia and Fasciodontia were strongly supported (BS > 90%, BPP = 1), while Hastodontia, Kneiffiella, Lyomyces and Xylodon did not receive reliable support; the clade of Lyomyces and Xylodon (BS = 74%, BPP = 1) was moderately supported, while no affinity among genera was clarified with strong support (Figure 1). Among the genera Hyphodontia, Kneiffiella, Lyomyces and Xylodon, 28 undescribed independent lineages emerged from the newly studied specimens, of which 14 lineages are composed of at least two specimens and 14 are single-specimen lineages.
The combined dataset of ITS, nLSU and mt-SSU regions (2) including 380 collections resulted in a concatenated alignment of 3425 characters with GTR+I+G as the best-fit evolutionary model. The ML search stopped after 250 BS replicates. In BI, all chains converged after 50 million generations with an average standard deviation of split frequencies of 0.003928, which was indicated by the ESSs above 2690 and the PSRFs close to 1000. ML and BI algorithms generated similar topologies in main lineages, and thus only the topology generated by the ML algorithm is presented along with BS value and BPPs above 50% and 0.8, respectively, at the nodes ( Figure 2). The phylogeny generated by this dataset strongly supports Hymenochaetales as an independent order (BS = 98%, BPP = 1). Within Hymenochaetales, five other families, viz. Coltriciaceae, Hymenochaetaceae, Neoantrodiellaceae, Nigrofomitaceae and Oxyporaceae are strongly (BS = 100%, BPP = 1) or moderately (BS > 58%, BPP > 0.99) supported as five monophyletic lineages, while the six genera belonging to Hyphodontia sensu lato are all strongly supported as independent genera. Of the six genera, Fasciodontia, Lyomyces and Xylodon form a strongly supported clade (BS = 96%, BPP = 1), while Hyphodontia, Kneiffiella and Hastodontia represent independent lineages within Hymenochaetales. As Schizopora, the type genus of Schizoporaceae, is a later synonym of Xylodon, the clade including Fasciodontia, Lyomyces and Xylodon, recognized at the family level, should be called Schizoporaceae. Similarly, Chaetoporellus is a later synonym of Kneiffiella, so the family name Chaetoporellaceae typified by Chaetoporellus was reintroduced for the well supported clade of Kneiffiella (BS = 81%, BPP = 0.97). The clade of Hyphodontia (BS = 100%, BPP = 1) outside of previously arranged families Chaetoporellaceae and Schizoporaceae are described as one new family on the basis of this genus. Hastodontia is considered to be a genus with an uncertain taxonomic position at the family rank within The combined dataset of ITS, nLSU and mt-SSU regions (2) including 380 collections resulted in a concatenated alignment of 3425 characters with GTR+I+G as the best-fit evolutionary model. The ML search stopped after 250 BS replicates. In BI, all chains converged after 50 million generations with an average standard deviation of split frequencies of 0.003928, which was indicated by the ESSs above 2690 and the PSRFs close to 1000. ML and BI algorithms generated similar topologies in main lineages, and thus only the topology generated by the ML algorithm is presented along with BS value and BPPs above 50% and 0.8, respectively, at the nodes ( Figure 2). The phylogeny generated by this dataset strongly supports Hymenochaetales as an independent order (BS = 98%, BPP = 1). Within Hymenochaetales, five other families, viz. Coltriciaceae, Hymenochaetaceae, Neoantrodiellaceae, Nigrofomitaceae and Oxyporaceae are strongly (BS = 100%, BPP = 1) or moderately (BS > 58%, BPP > 0.99) supported as five monophyletic lineages, while the six genera belonging to Hyphodontia sensu lato are all strongly supported as independent genera. Of the six genera, Fasciodontia, Lyomyces and Xylodon form a strongly supported clade (BS = 96%, BPP = 1), while Hyphodontia, Kneiffiella and Hastodontia represent independent lineages within Hymenochaetales. As Schizopora, the type genus of   Phylogenetic relationships of species belonging to Hyphodontia sensu lato within Hymenochaetales inferred from the combined dataset of ITS, nLSU and mt-SSU regions. The topology generated by the maximum likelihood algorithm is presented along with the bootstrap values and the Bayesian posterior probabilities above 50% and 0.8, respectively, at the nodes.
With regard to the multilocus combined datasets, that of ITS, nLSU, mt-SSU, tef1α, rpb1 and rpb2 regions for Hyphodontia (3) including 15 collections resulted in a concatenated alignment of 5060 characters with GTR+G as the best-fit evolutionary model. The ML search stopped after 250 BS replicates. In BI, all chains converged after ten million generations with an average standard deviation of split frequencies of 0.002287, which was indicated by the ESSs above 5300 and the PSRFs equal to 1.000. The multilocus combined dataset of ITS, nLSU, mt-SSU, tef1α, rpb1, rpb2 and atp6 regions for Lyomyces (4) including 50 collections resulted in a concatenated alignment of 5342 characters with GTR+I+G as the best-fit evolutionary model. The ML search stopped after 300 BS replicates. In BI, all chains converged after ten million generations with an average standard deviation of split frequencies of 0.002946, which was indicated by the ESSs above 4400 and the PSRFs close to 1.000. The multilocus combined dataset of ITS, nLSU, mt-SSU, tef1α, rpb1, rpb2 and atp6 regions for Xylodon (5) including 194 collections resulted in a concatenated alignment of 5543 characters with GTR+I+G as the best-fit evolutionary model. The ML search stopped after 400 BS replicates. In BI, all chains converged after 15 million generations with an average standard deviation of split frequencies of 0.006113, which was indicated by the ESSs above 1000 and the PSRFs equal to 1.000. Regarding each of the multilocus combined datasets for Hyphodontia, Lyomyces and Xylodon, ML and BI algorithms generated similar topologies in main lineages, and thus only the topologies generated by ML algorithm are presented along with BS value and BPPs above 50% and 0.8, respectively, at the nodes. Based on the multilocus phylogenetic analyses, the midpoint-rooted phylogeny of Hyphodontia recovered three species (including two new) and two undescribed single-specimen lineages (Figure 3), that of Lyomyces recovered nine species (including three new) and four undescribed single-specimen lineages (Figure 4), while that of Xylodon recovered 24 species (including eight new) and seven undescribed single-specimen lineages ( Figure 5). With regard to the multilocus combined datasets, that of ITS, nLSU, mt-SSU, tef1α, rpb1 and rpb2 regions for Hyphodontia (3) including 15 collections resulted in a concatenated alignment of 5060 characters with GTR+G as the best-fit evolutionary model. The ML search stopped after 250 BS replicates. In BI, all chains converged after ten million generations with an average standard deviation of split frequencies of 0.002287, which was indicated by the ESSs above 5300 and the PSRFs equal to 1.000. The multilocus combined dataset of ITS, nLSU, mt-SSU, tef1α, rpb1, rpb2 and atp6 regions for Lyomyces (4) including 50 collections resulted in a concatenated alignment of 5342 characters with GTR+I+G as the best-fit evolutionary model. The ML search stopped after 300 BS replicates. In BI, all chains converged after ten million generations with an average standard deviation of split frequencies of 0.002946, which was indicated by the ESSs above 4400 and the PSRFs close to 1.000. The multilocus combined dataset of ITS, nLSU, mt-SSU, tef1α, rpb1, rpb2 and atp6 regions for Xylodon (5) including 194 collections resulted in a concatenated alignment of 5543 characters with GTR+I+G as the best-fit evolutionary model. The ML search stopped after 400 BS replicates. In BI, all chains converged after 15 million generations with an average standard deviation of split frequencies of 0.006113, which was indicated by the ESSs above 1000 and the PSRFs equal to 1.000. Regarding each of the multilocus combined datasets for Hyphodontia, Lyomyces and Xylodon, ML and BI algorithms generated similar topologies in main lineages, and thus only the topologies generated by ML algorithm are presented along with BS value and BPPs above 50% and 0.8, respectively, at the nodes. Based on the multilocus phylogenetic analyses, the midpoint-rooted phylogeny of Hyphodontia recovered three species (including two new) and two undescribed single-specimen lineages (Figure 3), that of Lyomyces recovered nine species (including three new) and four undescribed single-specimen lineages (Figure 4), while that of Xylodon recovered 24 species (including eight new) and seven undescribed single-specimen lineages ( Figure 5). Phylogenetic relationship among species of Hyphodontia inferred from the combined dataset of ITS, nLSU, mt-SSU, tef1α, rpb1 and rpb2 regions. The topology generated by the maximum likelihood algorithm is presented along with the bootstrap values and the Bayesian posterior probabilities above 50% and 0.8, respectively, at the nodes.   . Phylogenetic relationship among species of Lyomyces inferred from the combined dataset of ITS, nLSU, mt-SSU, tef1α, rpb1, rpb2 and atp6 regions. The topology generated by the maximum likelihood algorithm is presented along with the bootstrap values and the Bayesian posterior probabilities above 50% and 0.8, respectively, at the nodes.

Figure 5.
Phylogenetic relationship among species of Xylodon inferred from the combined dataset of ITS, nLSU, mt-SSU, tef1α, rpb1, rpb2 and atp6 regions. The topology generated by the maximum likelihood algorithm is presented along with the bootstrap values and the Bayesian posterior probabilities above 50% and 0.8, respectively, at the nodes. Phylogenetic relationship among species of Xylodon inferred from the combined dataset of ITS, nLSU, mt-SSU, tef1α, rpb1, rpb2 and atp6 regions. The topology generated by the maximum likelihood algorithm is presented along with the bootstrap values and the Bayesian posterior probabilities above 50% and 0.8, respectively, at the nodes.
Taking morphological characters and the phylogenies from the datasets 1-5 into consideration, one new family and 14 species are proposed, and the taxonomic status of an additional 30 species names is discussed with the inclusion of validating two ineffectively published, invalid names as two new species of Xylodon. Moreover, identification keys to each of the genera of Hyphodontia, Kneiffiella, Lyomyces and Xylodon are provided.
The combined dataset for molecular clock analysis (6) included 80 collections, of which 40 belonged to Hymenochaetales. This dataset resulted in a concatenated alignment of 8330 characters with GTR+I+G as the best-fit evolutionary model. Chain convergence was indicated by the ESSs above 2010. In Hymenochaetales, the youngest family is Neoantrodiellaceae occurring in a mean crown age of 7.29 Mya with a 95% highest posterior density (HPD) of 3.23-12. 49   . Maximum-clade-credibility chronogram and estimated divergence times of families within Hymenochaetales inferred from the combined dataset of ITS, nLSU, tef1α, rpb1 and rpb2 regions. The estimated divergence times of 95% highest posterior density for all clades were indicated as node bars and for families in Hymenochaetales were also provided in the upper-left of the tree as exact numbers, while the Bayesian posterior probabilities above 0.8 and the mean divergence times of clades were labeled above and below the branches, respectively, at the nodes.
The combined dataset for ancestral state reconstruction (7) including 59 collections resulted in a concatenated alignment of 2436 characters with GTR+I+G as the best-fit evolutionary model. Chain convergence was indicated by the ESSs above 290. Across the two basidiocarp traits, the resupinate habit and poroid hymenophoral configuration were evaluated as the ancestral state within Hymenochaetales (Figure 7). Below the order level, the resupinate habit remains in all families but also evolves to the pileate habit in Hymenochaetaceae, Neoantrodiellaceae and Oxyporaceae as pileate-resupinate basidiocarps. The ancestral poroid state in Hymenochaetales remains only in Oxyporaceae, and evolves to grandinioid state in Chaetoporellaceae and grandinioid-odontioid state in Hyphodontiaceae and Schizoporaceae. Noteworthily, the poroid state in Hymenochaetaceae and Neoantrodiellaceae seems to independently evolve back from the grandinioid state and be related to the emergence of a pileate habit in this lineage. Figure 7. Trait evolution of basidiocarps within Hymenochaetales. The mirrored consensus tree was generated by the Bayesian inference algorithms using BEAST, while the trait of a pie chart at each node was evaluated using RASP under the Bayesian Binary MCMC model. The trait represented by each color and letter in the pie chart is indicated in the upper-left for the left part and upper-right of the right part. The lineage for each family is indicated at the nodes along with the pentagram mark for assisting discussion.     bladder-like to clavate cystidia (35-55 µm in length) and basidiospores measuring 3.2-4 µm in length [78], whereas K. subglobosa based on a specimen from Island of Taiwan, has slightly thick-walled, tubular or cylindrical cystidia (40-180 µm in length) and longer basidiospores (4.2-5 µm in length) [30]. The phenomenon that two species have almost identical ITS regions but distinct morphological characters may represent an ongoing allopatric speciation event [79] and was also reported in Basidioradulum mayi and B. tasmanicum belonging to Hymenochaetales [44]. Therefore, we tentatively accept K. subefibulata and K. subglobosa as independent species. Notes: Like previous phylogenetic analysis [16], the current ITS-based phylogeny grouped sequences from two collections, each of Kneiffiella subefibulata and K. subglobosa together in a strongly supported clade ( Figure 1). However, the two collections of K. subefibulata do not form a separate subclade, but are in a grade at the base of the clade. Morphologically, both species have odontioid to hydnoid hymenophores with 3-5 aculei per mm, aculei up to 2 mm in length and microscopic elements without clamp connection [30,78]. K. subefibulata, typified by specimens from Mainland China, bears thin-walled, bladder-like to clavate cystidia (35-55 μm in length) and basidiospores measuring 3.2-4 μm in length [78], whereas K. subglobosa based on a specimen from Island of Taiwan, has slightly thick-walled, tubular or cylindrical cystidia (40-180 μm in length) and longer basidiospores (4.2-5 μm in length) [30]. The phenomenon that two species have almost identical ITS regions but distinct morphological characters may represent an ongoing allopatric speciation event [79] and was also reported in Basidioradulum mayi and B. tasmanicum belonging to Hymenochaetales [44]. Therefore, we tentatively accept K. subefibulata and K. subglobosa as independent species.
Hyphodontia alutaria (Burt) J. Erikss., Symb. bot. upsal. 16(no. 1): 104. 1958. Basionym: Peniophora alutaria Burt, Ann. Mo. bot. Gdn 12: 332. 1926Gdn 12: 332. (1925. Notes: In a previous study [16], one sequence from each of Hyphodontia alutaria and H. pallidula clustered together. These two sequences, and two further sequences of H. alutaria form a strongly supported clade in the current ITS-based phylogenetic tree, with the sequence of H. pallidula falling within the clade rather than at the base (Figure 1), which indicates that these two species may be conspecific. However, we did not have a chance to examine the voucher specimens for the ITS sequences. Moreover, the morphological characters separating these two species are quite distinct [19]. Therefore, for the moment we accept H. alutaria and H. pallidula as two independent species. Description: Basidiocarps annual, resupinate, adnate, cracked and brittle when dry. Hymenophore smooth to grandinioid, white to light-buff. Margin paler than or concolorous with subiculum, thinning. Hyphal system monomitic; generative hyphae with clamp connections, hyaline, thin-walled, dichotomous branching, tortuous, 2-3.5 µm in diam. Lagenocystidia abundant, thin-walled, with broad bases tapering abruptly towards the apices, apically encrusted. Basidia cylindrical, somewhat sinuous or utriform, occasionally encrusted with granular crystals, 25-30 × 3.5-5.5 µm, with four sterigmata and a clamp connection at the base; basidioles similar in shape to basidia, but smaller. Basidiospores broadly ellipsoid to ovoid, hyaline, smooth, thick-walled, acyanophilous, inamyloid, indextrinoid, Notes: The thick-walled basidiospores make Hyphodontia pachyspora distinct in Hyphodontia. Microscopically, H. pachyspora resembles H. arguta by the absence of septocystidia. It differs from H. arguta by the thick-walled and wider basidiospores [19,20]. Three sequenced specimens of H. arguta clustered into two lineages in the ITS-based phylogeny ( Figure 1): the first one including two specimens sequenced by Riebesehl et al. [34]; the second one including the specimen sequenced by Larsson et al. [7] and two specimens of H. pachyspora. The combined dataset of ITS, nLSU and mt-SSU regions also fully supported the second lineage ( Figure 2). We did not examine the specimen sequenced by Larsson et al. [7], but the specimens of H. pachyspora clearly show morphological distinctions from H. arguta as mentioned above. Whether the two sequenced specimens by Riebesehl et al. [34] really represent H. arguta is still ambiguous, as discussed by Kan et al. [36].    Description: Basidiocarps annual, resupinate, adnate, cracked and brittle when dry. Hymenophore odontioid, white to cream. Margin paler than or concolorous with subiculum, abrupt. Hyphal system monomitic; generative hyphae with clamp connections, hyaline, thin-walled, dichotomous branching, tortuous, 1.5-3.5 µm in diam. Lagenocystidia abundant, thin-walled, with broad bases tapering abruptly towards the apices, apically encrusted. Basidia subclavate somewhat sinuous to utriform, 20-25 × 3.5-4.5 µm, with four sterigmata and a clamp connection at the base; basidioles similar in shape to basidia, but smaller. Basidiospores broadly ellipsoid or ovoid, with a large oily drop, hyaline, smooth, thin-walled, acyanophilous, inamyloid, indextrinoid, (3.   Notes: Hyphodontia wongiae resembles H. wrightii by sharing odontioid hymenophore and absence of capitate cystidia; however, H. wongiae differs in its broadly ellipsoid to ovoid basidiospores, which are shorter and wider than those of H. wrightii (4.5-5.5 × 2.5-3 µm) [19]. Notes: Hyphodontia zhixiangii was originally described on the basis of Uzbek specimens [36]. Besides the holotype and two paratypes, the current phylogenetic analyses also recovered six Chinese specimens in the lineage of H. zhixiangii (Figures 1-3). Within this lineage, three Uzbek specimens clustered together with strong support in the phylogenies based on three and six genes (Figures 2 and 3). However, the monophyletic group of the six Chinese specimens was not supported. Therefore, these six specimens are confirmed to be H. zhixiangii from a phylogenetic perspective. This is the first record of H. zhixiangii from China.
Lyomyces      Description: Basidiocarps annual, resupinate, adnate, cracked and brittle when dry. Hymenophore white to cream, smooth or slightly tuberculate; margin paler, thinning, pruinose. Hyphal system monomitic; generative hyphae with clamp connections, hyaline, thinto slightly thick-walled, dichotomous branching, tortuous, frequently encrusted with crystals, 2.5-5.5 µm in diam. Leptocystidia numerous, usually encrusted with crystals, thin-or slightly thick-walled, 25-45 × 3.5-5.5 µm. Basidia utriform or clavate, 15-25 × 3.5-5.5 µm, with four sterigmata and a clamp connection at the base; basidioles similar in shape to basidia, but smaller. Basidiospores broadly ellipsoid, usually with a large oily drop, hyaline, smooth, thin-walled, acyanophilous, inamyloid, indextrinoid, (4.9-)5.0-6.1(-6.2) × Notes: Lyomyces sambuci is a worldwide species and has priority over the synonym L. serus, the generic type of Lyomyces [81]. However, its wide distribution is questioned, because L. sambuci has been considered to be a species complex in recent decades [20,30]. Recently, four new species were segregated from this complex [17]. Nevertheless, the reduced concept of L. sambuci is still a species complex, because two distinct lineages of L. sambuci with similar morphological characters were present in an ITS-based phylogeny [17]. For this reason, the lineage composed of the collections of 170SAMHYP and GEL 2414 was treated as an undescribed cryptic species, while the other lineage was accepted to be the true L. sambuci [17]. The current ITS-based phylogeny (Figure 1) confirmed these two lineages and also revealed an additional lineage with two collections labeled as L. cf. sambuci. This L. cf. sambuci lineage quite possibly represents a new species; however, we did not have the opportunity to examine specimens.
Greslebin et al. [84] and Hjortstam & Ryvarden [27] considered X. australis to be distributed also in South America but did mention essential morphological differences among specimens from different geographic regions. Given that two Chinese specimens (LWZ 20180920-12a and LWZ 20180922-47) with similar hymenophoral color as well as other morphological characters to X. australis represent an independent species (described below as X. yunnanensis), the South American specimens designated as X. australis are quite possibly an undescribed species. Further molecular evidence is needed to clarify this issue. Notes: Three collections previously treated as Xylodon brevisetus were separated into two quite separated lineages in the current ITS-based phylogeny. Firstly, two collections formed a strongly supported clade sister to X. crystalliger. Secondly, a single collection KHL 12386 was sister to X. victoriensis. In addition, two collections annotated as X. cf. brevisetus formed a strongly supported clade sister to the undescribed collection LWZ 20180922-26 from Yunnan, China (Figure 1). We apply the name X. brevisetus to the collection KHL 12386 from Sweden, which was also utilized in the phylogeny based on the combined dataset of ITS, nLSU and mt-SSU regions ( Figure 2) where it was also placed sister to X. Description: Basidiocarps annual, resupinate, adnate, cracked and brittle when dry. Hymenophore grandinioid to slightly odontioid, white to cream. Margin paler than or concolorous with subiculum, thinning. Hyphal system monomitic; generative hyphae with clamp connections, hyaline, thin-walled to slightly thick-walled, dichotomous branching, tortuous, 2.5-5.5 µm in diam. Clavate-sinuous to submoniliform cystidia, 35-40 × 6-7 µm. Basidia subclavate or subcapitate, 30-35 × 5.5-6.5 µm, with four sterigmata and a clamp connection at the base; basidioles similar in shape to basidia, but smaller. Basidiospores narrowly ellipsoid, with a large oily drop, hyaline, smooth, thin-walled, acyanophilous, inamyloid, indextrinoid, (5.1-)5.   Notes: Xylodon damansaraensis is characterized by the clavate-sinuous to submoniliform cystidia, which make it similar to X. brevisetus, X. crassisporus, X. spathulatus and X. subclavatus. Xylodon brevisetus and X. spathulatus differ from X. damansaraensis in the presence of capitate cystidia and slightly wider basidiospores (3-4 µm in width for both species) [19]. Xylodon crassisporus differs in the presence of capitate cystidia, thick-walled hyphae and thick-walled, wider basidiospores (4-4.5 µm in width) [25]. Xylodon subclavatus differs in the presence of hyphoid and capitate to subcapitate cystidia, and wider basidiospores (3.5-4 µm in width) [85].
Xylodon kunmingensis L.W. Zhou  Notes: Xylodon kunmingensis was described from Yunnan, China [42], while X. exilis was described based on specimens from Island of Taiwan [18]. Both species share a cream, odontioid hymenophore, thick-walled subicular hyphae, capitate cystidia, encrusted hyphal endings, and narrowly ellipsoid basidiospores similar in size [18,42]. Phylogenetically, the ITS-based tree grouped all eight type specimens of these two species and two newly sequenced specimens from Hubei and Liaoning Provinces, China, into a strongly supported clade (Figure 1). Similarly, in the tree inferred from the combined dataset of ITS, nLSU and mt-SSU regions all three samples (the holotype of X. exilis and the two newly sequenced specimens) formed a fully supported clade ( Figure 2). So, from both morphological and phylogenetic perspectives, X. exilis and X. kunmingensis are conspecific. Xylodon kunmingensis has priority as the correct name of this species. Description: Basidiocarps annual, resupinate, adnate, cracked and brittle when dry. Hymenophore smooth to grandinioid, white to cream. Margin paler than or concolorous with subiculum, abrupt. Hyphal system monomitic; generative hyphae with clamp connections, hyaline, thin-walled, dichotomous branching, tortuous, 2.5-4.5 µm in diam. Cystidia of two types: (a) leptocystidia thin-walled, with a wider base, gradually thinning, penetrating approximately half of their lengths through hymenium, 80-85 × 4-5 µm; (b) lagenocystidia thin-walled, with broad bases tapering abruptly towards the apices, apically encrusted, 20-25 × 3.5-4.5 µm. Basidia utriform or subclavate, 20-25 × 3.5-4.5 µm, with four sterigmata and a clamp connection at the base, encrusted with granular crystals; basidioles similar in shape to basidia, but smaller. Basidiospores ellipsoid, with a large oily drop, hyaline, smooth, thin-walled, acyanophilous, inamyloid, indextrinoid, (4.5-)4.6-5.      Notes: This species was published with a detailed description in a conference paper that is available online [82], but not in a form that is effectively published, as individual papers are available separately and there does not appear to be a title page with an ISSN or ISBN. Index Fungorum has considered the name to be invalid on the basis of the lack of effective publication, and therefore we validate it here by reproducing the type citation and original description. Notes: Xylodon niemelaei was first described as a poroid species of Hyphodontia [30], and later combined in Xylodon [10]. Xylodon rhizomorphus and X. reticulatus were both recently segregated from X. niemelaei as new species of Hyphodontia [15,86] and then combined in Xylodon [16,35]. Soon after that, Xylodon jacobaeus closely related to the above-mentioned three species was newly described [39]. However, morphological differences are not distinct and stable among these four species [15,30,39,86]. Moreover, the original ITS-based phylogenies did not clearly separate these four species. For example, in Zhao et al. [86], only a single collection of X. niemelaei was included to demonstrate X. rhizomorphus as an independent lineage from X. niemelaei; in Chen et al. [15], two or three collections for each species of X. niemelaei (including a collection misapplied as X. apacheriensis), X. reticulatus and X. rhizomorphus formed three subclades in the clade being composed of these three species, but the branch lengths were not long enough to further discriminate internal subclades as independent species; in Crous et al. [39] the branch lengths were also quite short.
In the current ITS-based phylogeny, a much greater sampling of 25 collections has been included. Not all collections were clustered in strongly supported subclades in the clade being composed of X. jacobaeus, X. niemelaei, X. reticulatus and X. rhizomorphus (Figure 1). Although three and four subclades were, respectively, revealed in the phylogenies based on three genes ( Figure 2) and seven genes ( Figure 5), the morphological characters for each subclade are not corresponding to the original concepts of these four species. Given the above, we consider the minor morphological differences among these collections to be variations of a single species X. niemelaei, and treat X. jacobaeus, X. reticulatus and X. rhizomorphus as later synonyms.    Description: Basidiocarps annual, resupinate, adnate, cracked and brittle when dry. Hymenophore odontioid, cream or light buff in young parts and buff-yellow in old parts. Margin white to cream, abrupt. Hyphal system monomitic; generative hyphae with clamp connections, hyaline, thin-walled, dichotomous branching, tortuous, usually encrusted with crystals, 2.5-4.5 µm in diam. Leptocystidia thin-walled, usually encrusted with crystals, 30-35 × 3-3.5 µm. Basidia subclavate, 25 × 3.5-5 µm, with four sterigmata and a clamp connection at the base, usually encrusted with crystals; basidioles similar in shape to basidia, but smaller. Basidiospores broadly ellipsoid, with a large oily drop, hyaline, smooth, thin-walled, cyanophilous, inamyloid, indextrinoid, (4.7-)4.8-6.5(-6.7) × Notes: Xylodon subflaviporus was named after its morphological similarity to X. flaviporus, and in an ITS-based phylogeny had a close relationship with X. flaviporus in a clade also including X. ovisporus [35]. However, X. subflaviporus was phylogenetically closer to X. ovisporus than to X. flaviporus [18]. In addition, the monophyly of X. subflaviporus was never well supported by previous phylogenetic analyses [18,35]. Similarly, the current ITSbased phylogeny (Figure 1) also supported that X. subflaviporus was closer to X. ovisporus, and failed to recover X. subflaviporus in a lineage with strong support. However, the phylogenies based on three genes ( Figure 2) and seven genes ( Figure 5) did strongly support X. subflaviporus as an independent lineage for the first time besides a closer relationship with X. ovisporus. Moreover, a lineage represented by a single Australian specimen LWZ 20180517-34 also appeared in the strongly supported clade with the two East Asian species X. subflaviporus and X. ovisporus (Figures 2 and 5 Description: Basidiocarp resupinate, effused, adnate, up to 300 µm thick in section; hymenial surface odontioid, aculei dense, conical, pale orange when fresh, orange gray to grayish orange on drying; margins thinning, fibrillose, paler concolorous, to indeterminate. Hyphal system monomitic. Generative hyphae, branched, septate, clamped; basal hyphae up to 4.7 µm wide, parallel to the substrate, thick-walled, loosely arranged, encrusted; subhymenial hyphae up to 3.5 µm wide, vertical, thin-walled, compactly arranged. Prominent patches of encrustation in the aculei. Cystidia like hyphal ends none. Basidia 20.0-26.0 × 4.7-5.3 µm, clavate, somewhat sinuous, 4-sterigmate, with basal clamp; sterigmata up to 4.0 µm. Basidiospores 4.2-5.2 × 3.0-5.0 µm, subglobose, smooth, thin-walled, inamyloid, acyanophilous. Notes: Although the name was originally published by Dhingra [82] with a detailed description, the publication was not effectively published (see comments above under Xylodon mussooriensis). Here, we validate the name by reproducing the type citation and original description.
Xylodon Notes: Xylodon subtropicus is a recently described poroid species based on two Asian specimens from Vietnam and China [15]. It is very similar to Hyphodontia radula, and the so-called morphological distinctions between these two species actually overlap [15]. Soon after the publication of X. subtropicus, the name X. raduloides was introduced as a replacement name in Xylodon for H. radula, to avoid the competing name Xylodon radula (Fr.) Tura, Zmitr., Wasser & Spirin, which blocked Poria radula, the basionym of H. radula, from being combined in Xylodon [16]. In addition, Fernández-López et al. [41] recognized three additional species, viz. X. laurentianus, X. novozelandicus and X. patagonicus, segregated from X. raduloides based on a combination of morphological, molecular and environmental evidence. However, the morphological differences are quite minor among these four species, and some species pairs lack morphological differentiation [41]. A species delimitation analysis found that the hypothesis of recognizing the independence of each of the four species was more probable than that of accepting them as a single species; however, regarding phylogenetic evidence, only these four species and an outgroup taxon Xylodon flaviporus were referred to and other close species, especially X. subtropicus, were not included [41]. Moreover, while only the ITS-based phylogeny strongly supported the independence of each of the four species, there was considerable missing data, and the lineages were not strongly supported on nLSU region alone and the combined ITS and nLSU regions [41].
The current ITS-based phylogeny covering the most comprehensive sampling of Xylodon till now recovered the four lineages of X. laurentianus, X. novozelandicus, X. patagonicus and X. raduloides, each not receiving strong support; the lineage of X. subtropicus being composed of two original Asian specimens described by Chen et al. [15] occupied a basal position of the four lineages (Figure 1). Like the phylogeny based on ITS region (Figure 1), that based on three genes also recovered the four lineages, and no one was strongly supported, neither was the clade being composed of these four lineages ( Figure 2). Alternatively, these four lineages together with the basal lineage of X. subtropicus formed a strongly supported clade (Figure 2). In the phylogeny based on seven genes ( Figure 5), the two original specimens of X. subtropicus with only ITS and nLSU regions available were not included; the lineage of X. raduloides was strongly supported, whereas those of X. laurentianus, X. novozelandicus and X. patagonicus were weakly to moderately supported; the clade consisting of these four lineages was strongly supported. In these three phylogenies (Figures 2 and 5), the newly sequenced Australian specimens merged in the lineage of X.
novozelandicus. Besides the topologies, the branch lengths among these lineages are also too short to clearly distinguish species and fall within the infraspecific distances observed in several other well accepted species in the genus (Figures 2 and 5). Taking into consideration the morphological similarity, the current phylogenies and the low level of divergence, we consider X. laurentianus, X. novozelandicus, X. patagonicus, X. raduloides and X. subtropicus to be conspecific with the last one as the correct name.
The more or less geographically structured phylogenetic signal recovered by Fernández-López et al. [41] indicates a species perhaps in the process of incipient speciation. It is also of interest that Fernández-López et al. [41] found some niche differentiation among the four lineages they recognized as species, and this differentiation merits further exploration on a wider sampling. Furthermore, Paulus et al. [87] found that New Zealand samples of X. subtropicus (as S. radula) had a high mating compatibility (73.7% crosses completely positive and 1.2% lacking clamps) but there was also compatibility between these and isolates from the Northern Hemisphere, albeit at a reduced level (47.8% crosses completely positive and 12.2% lacking clamps). Xylodon subtropicus presents an interesting test case for applying genome-wide markers such as restriction site-associated DNA markers (RADSeq) [88] to further investigate the boundary between population genetic and species level variation, especially if integrated with mating studies of this and closely related species.

Discussion
Even though species diversity of Hyphodontia sensu lato has been extensively explored worldwide, the phylogenetic relationships among genera of Hyphodontia sensu lato and the phylogenetic positions of these genera within Hymenochaetales were unclear. In this study, the most comprehensive phylogenetic analyses to date focusing on genera belonging to Hyphodontia sensu lato were performed. The analyses resulted in segregation of these genera at the family level, viz. two previous names Chaetoporellaceae and Schizoporaceae and a newly proposed name Hyphodontiaceae, and placed all these families within Hymenochaetales. Schizoporaceae includes Fasciodontia, Lyomyces and Xylodon, while the monotypic families Chaetoporellaceae and Hyphodontiaceae, respectively, accommodate Kneiffiella and Hyphodontia. Noteworthily, the genus Hastodontia was excluded from these three families and treated as a genus of uncertain placement at the family level within Hymenochaetales.
Hymenochaetales was erected as a monotypic order on the basis of Hymenochaetaceae [90]. An additional family Schizoporaceae was also accommodated in Hymenochaetales by Larsson [8] and recognized in the latest edition of the Dictionary of the Fungi [45]. Later, Oxyporaceae, Neoantrodiellaceae and Nigrofomitaceae were successively delimited within Hymenochaetales [91][92][93]. Besides these five families, four additional families have been accepted in Hymenochaetales by some authors: Coltriciaceae, Rickenellaceae, Repetobasidiaceae and Tubulicrinaceae.
Coltriciaceae was typified by Coltricia and also included Aurificaria, Coltriciella, Cyclomyces, Inonotus, Onnia and Phylloporia as exemplified genera [2]. Most of taxonomists treated these genera in Hymenochaetaceae, which makes Coltriciaceae as a synonym of Hymenochaetaceae [46,94]. However, a few studies indicated that Coltricia and Coltriciella were independent from Hymenochaetaceae [7]. The current study agrees with the topology of Larsson et al. [7]: the clade including Hymenochaetaceae and Coltriciaceae did not receive reliable statistical support ( Figure 2). Therefore, we accept Coltriciaceae for Coltricia and Coltriciella as an independent family from Hymenochaetaceae.
A recently published outline of Basidiomycota accepted Rickenellaceae in Hymenochaetales [46], which was followed by a subsequent outline of Fungi and funguslike taxa [95]. However, Rickenellaceae is a nomenclaturally superfluous name, because its original circumscription includes Repetobasidium, the type genus of a prior family Repetobasidiaceae. According to Art. 52.4 of the Shenzhen Code [96], Rickenellaceae could be legitimate only if it no longer includes Repetobasidium. The concept of Rickenellaceae was used under the name 'Rickenella family' by Larsson [8] and Ariyawansa et al. [92], but its circumscription was polyphyletic. Korotkin et al. [97] used the name 'Rickenella clade' for species in Rickenella and 13 additional genera falling within a clade that was poorly supported. In the current phylogenetic analysis (Figure 2), two species of Rickenella grouped together with six of the above-mentioned 13 genera in an also poorly supported clade, while the other included genera out of the above-mentioned 13 genera were separated from this clade. Regarding Repetobasidiaceae, since its erection in 1981, it does not appear to have been considered a family of Hymenochaetales, even though species of its type genus Repetobasidium were phylogenetically included in Hymenochaetales [7]. This may be due to the absence of the generic type Repetobasidium vile in any phylogenetic analysis. Therefore, the circumscription of Repetobasidiaceae is still ambiguous from a phylogenetic perspective. Tubulicrinaceae erected with Tubulicrinis as the type genus [2] was another family accepted at one time in Hymenochaetales [8]. The clade labeled as

Discussion
Even though species diversity of Hyphodontia sensu lato has been extensively explored worldwide, the phylogenetic relationships among genera of Hyphodontia sensu lato and the phylogenetic positions of these genera within Hymenochaetales were unclear. In this study, the most comprehensive phylogenetic analyses to date focusing on genera belonging to Hyphodontia sensu lato were performed. The analyses resulted in segregation of these genera at the family level, viz. two previous names Chaetoporellaceae and Schizoporaceae and a newly proposed name Hyphodontiaceae, and placed all these families within Hymenochaetales. Schizoporaceae includes Fasciodontia, Lyomyces and Xylodon, while the monotypic families Chaetoporellaceae and Hyphodontiaceae, respectively, accommodate Kneiffiella and Hyphodontia. Noteworthily, the genus Hastodontia was excluded from these three families and treated as a genus of uncertain placement at the family level within Hymenochaetales.
Hymenochaetales was erected as a monotypic order on the basis of Hymenochaetaceae [90]. An additional family Schizoporaceae was also accommodated in Hymenochaetales by Larsson [8] and recognized in the latest edition of the Dictionary of the Fungi [45]. Later, Oxyporaceae, Neoantrodiellaceae and Nigrofomitaceae were successively delimited within Hymenochaetales [91][92][93]. Besides these five families, four additional families have been accepted in Hymenochaetales by some authors: Coltriciaceae, Rickenellaceae, Repetobasidiaceae and Tubulicrinaceae.
Coltriciaceae was typified by Coltricia and also included Aurificaria, Coltriciella, Cyclomyces, Inonotus, Onnia and Phylloporia as exemplified genera [2]. Most of taxonomists treated these genera in Hymenochaetaceae, which makes Coltriciaceae as a synonym of Hymenochaetaceae [46,94]. However, a few studies indicated that Coltricia and Coltriciella were independent from Hymenochaetaceae [7]. The current study agrees with the topology of Larsson et al. [7]: the clade including Hymenochaetaceae and Coltriciaceae did not receive reliable statistical support ( Figure 2). Therefore, we accept Coltriciaceae for Coltricia and Coltriciella as an independent family from Hymenochaetaceae.
A recently published outline of Basidiomycota accepted Rickenellaceae in Hymenochaetales [46], which was followed by a subsequent outline of Fungi and fungus-like taxa [95]. However, Rickenellaceae is a nomenclaturally superfluous name, because its original circumscription includes Repetobasidium, the type genus of a prior family Repetobasidiaceae. According to Art. 52.4 of the Shenzhen Code [96], Rickenellaceae could be legitimate only if it no longer includes Repetobasidium. The concept of Rickenellaceae was used under the name 'Rickenella family' by Larsson [8] and Ariyawansa et al. [92], but its circumscription was polyphyletic. Korotkin et al. [97] used the name 'Rickenella clade' for species in Rickenella and 13 additional genera falling within a clade that was poorly supported. In the current phylogenetic analysis (Figure 2), two species of Rickenella grouped together with six of the above-mentioned 13 genera in an also poorly supported clade, while the other included genera out of the above-mentioned 13 genera were separated from this clade. Regarding Repetobasidiaceae, since its erection in 1981, it does not appear to have been considered a family of Hymenochaetales, even though species of its type genus Repetobasidium were phylogenetically included in Hymenochaetales [7]. This may be due to the absence of the generic type Repetobasidium vile in any phylogenetic analysis. Therefore, the circumscription of Repetobasidiaceae is still ambiguous from a phylogenetic perspective. Tubulicrinaceae erected with Tubulicrinis as the type genus [2] was another family accepted at one time in Hymenochaetales [8]. The clade labeled as Tubulicrinaceae with three species from two additional genera Hyphodontia and Sphaerobasidium, besides three species of Tubulicrinis, however, received no strong statistical support [8]. In addition, Tubulicrinis was shown to be polyphyletic with two species of Tubulicrinis distantly related to the so-called Tubulicrinaceae clade [7], and more importantly its type species T. glebulosus has not been included in any phylogenetic analysis. Another preliminary phylogenetic analysis on Hymenochaetales used Tubulicrinaceae for the clade including Athelopsis lunata (currently accepted as Sidera lunata) [98], Sphaerobasidium minutum and four species of Tubulicrinis, which received low statistical supports [92]. In the current phylogeny, the weakly supported clade including three species of Tubulicrinis and one of Sphaerobasidium was separated from Sidera lunata, and within this clade, the three species of Tubulicrinis did not cluster together ( Figure 2). Alternatively, Tubulicrinaceae was not accepted in Hymenochaetales, but treated as a later synonym of Hydnaceae by the latest edition of the Dictionary of the Fungi [45] and the recent outline of Basidiomycota [46]. It is not possible to apply the family names Rickenellaceae, Repetobasidiaceae and Tubulicrinaceae with confidence at the moment. Undoubtfully, a wider sampling including the generic types of the respective families is needed to judge whether the families should be taken up for monophyletic clades within Hymenochaetales. Nevertheless, the current phylogeny excludes the possibility of the six genera belonging to Hyphodontia sensu lato having overlapping circumscriptions with these three families at the family level ( Figure 2).
Zhao et al. [99] proposed that the taxonomic units at the same rank, especially higher ranks, should have roughly equivalent stem ages, and overlap between the ranges of different ranks should be minimized. This approach was also applied by He et al. [46]. Lücking [100] provides a critique, arguing that 'temporal banding should not be employed in an absolute manner, but rather as a tool complementing assessments of phenotypic disparity and differential diagnoses at given rank levels'. Except those for fossil fungi, almost all taxa in the current taxonomic system of fungi are introduced mainly for extant species (crown groups), but not unknown and extinct species (stem groups). For example, in the current case, Hyphodontiaceae is newly introduced to accommodate species of Hyphodontia. Moreover, excluding mass extinction events that are not known for fungi, the crown group emerges just before the extinction of the stem group [101], which indicates that the crown age is much closer to the genuine emergence time of the crown group than the stem age. Therefore, we use the crown age instead of the stem age, similar to Liu et al. [102], for complementary discussion of families within Hymenochaetales in the current study. The current estimations suggest that Chaetoporellaceae and Hyphodontiaceae share close crown age ranges, while those of Schizoporaceae, Hymenochaetaceae and Oxyporaceae are overlapping and somewhat older ( Figure 6). However, compared with these five families, Coltriciaceae, Neoantrodiellaceae and Nigrofomitaceae emerged considerably later ( Figure 6). Interestingly, Coltriciaceae, Neoantrodiellaceae and Nigrofomitaceae as well as Hymenochaetaceae formed a strongly supported clade together with the three genera Basidioradulum, Fibricium and Trichaptum with uncertain taxonomic positions at the family rank in the phylogeny based on a combined dataset of ITS, nLSU and mt-SSU regions (Figure 2). The clade including these four families emerged in the mean crown age of 191.36 Mya with a 95% HPD of 173.9-208.63 Mya, which, compared with those of Coltriciaceae, Neoantrodiellaceae and Nigrofomitaceae, was much closer to the crown ages of other four families in Hymenochaetales ( Figure 6). The phylogeny and the much younger crown ages for these three families both indicate that a larger concept of Hymenochaetaceae (inclusive of Coltriciaceae, Neoantrodiellaceae and Nigrofomitaceae) is worth considering. Further comprehensive phylogenetic analyses focusing on this clade with a wider taxon sampling will be helpful to generate a reliable phylogenetic resolution for the circumscription of Hymenochaetaceae.
Oxyporaceae was originally a monotypic family typified by Oxyporus [91]. Leucophellinus was subsequently added to this family [92]. Later, the type genus of Oxyporaceae was considered to be a later synonym of Rigidoporus, and species of Oxyporus were transferred to this genus as well as to Bridgeoporus in Hymenochaetales [103]. According to the current phylogeny based on three genes (Figure 2), Bridgeoporus, Leucophellinus and Rigidoporus are all accepted in Oxyporaceae. Notably, Oxyporaceae did not receive a strongly supported statistical value from the ML algorithm (BS = 72%), although it received a full statistical value from the BI algorithm ( Figure 2). Like Hymenochaetaceae, further research is required to confirm the circumscription of Oxyporaceae.
Besides the eight accepted families, about twenty genera have no available taxonomic position at the family rank in Hymenochaetales, partly due to a lack of systematic surveys on their species diversity. In developing a comprehensive classification of Hymenochaetales, more taxonomic work on species diversity and phylogeny of these poorly known genera needs to be performed.
As one of the most obvious traits of basidiomycetous fungi, basidiocarps play an essential role in the process of sexual reproduction via protecting reproductive organs and promoting basidiospore dispersal [104]. Previous studies at a higher taxonomic scale (at or above the order level) suggested that the ancestral shape of basidiocarps was resupinate and evolved several times to pileate-stipitate shape [105][106][107]. While this higher taxonomic scale provided a framework for the evolution of basidiocarp shape, the situation within certain orders has rarely been explored. Moreover, besides basidiocarp shape, the hymenophoral configuration is also crucial for sexual reproduction. It is common sense that, for example, poroid hymenia can provide stronger protection for reproductive organs but also an obstacle for basidiospore dispersal in comparison with smooth hymenia. Ancestral state reconstruction on such characters does not appear to have been performed. Species in Hymenochaetales have diverse traits of basidiocarps. This means that now that a phylogenetic framework for Hymenochaetales has been generated with an emphasis on a wide sampling of species belonging to Hyphodontia sensu lato, it is possible to explore for the first time the trait evolution of basidiocarps from the perspectives of basidiocarp shape and hymenophoral configuration simultaneously (Figure 7). The current result indicates that the evolutionary direction of basidiocarp shape below the order level is generally consistent with that at or above the order level (from resupinate to pileate). Meanwhile, neither smooth nor poroid hymenia, but the intermediate character (grandinioid and odontioid hymenia) perhaps representing a balance between protection and dispersal appears to have the adaptive advantage at least in certain lineages, where it has evolved independently. Interestingly, after evolving to the grandinioid state from the poroid state, the hymenophoral configuration in Hymenochaetaceae and Neoantrodiellaceae evolved back to the poroid state. This evolutionary event coincided with the evolution of basidiocarp shape from resupinate to pileate habit. However, in the lineage indicated by a pentagram mark (Figure 7), another evolutionary event occurred with a transition back to the poroid state from the grandinioid state, and this transition did not correspond to a transition from the resupinate to pileate habit. In Oxyporaceae, the ancestral state of basidiocarps was retained, resupinate habit and poroid hymenophoral configuration. So, a relationship between the pileate and poroid states of basidiocarps is indicated but is not strict. These findings provide a number of candidates for genomic analysis in a phylogenetic framework to investigate the evolutionary dynamics of transitions in basidiocarp shape and hymenophoral configuration, especially as far as reversals of hymenophoral configuration.