Taxonomy, phylogeny, molecular dating and ancestral state reconstruction of Xylariomycetidae (Sordariomycetes)

Xylariomycetidae (Ascomycota) is a highly diversified group with variable stromatic characters. Our research focused on inconspicuous stromatic xylarialean taxa from China, Italy, Russia, Thailand and the United Kingdom. Detailed morphological descriptions, illustrations and combined ITS-LSU-rpb2-tub2-tef1 phylogenies revealed 39 taxa from our collections belonging to Amphisphaeriales and Xylariales. A new family (Appendicosporaceae), five new genera (Magnostiolata, Melanostictus, Neoamphisphaeria, Nigropunctata and Paravamsapriya), 27 new species (Acrocordiella photiniicola, Allocryptovalsa sichuanensis, Amphisphaeria parvispora, Anthostomella lamiacearum, Apiospora guiyangensis, A. sichuanensis, Biscogniauxia magna, Eutypa camelliae, Helicogermslita clypeata, Hypocopra zeae, Magnostiolata mucida, Melanostictus longiostiolatus, M. thailandicus, Nemania longipedicellata, N. delonicis, N. paraphysata, N. thailandensis, Neoamphisphaeria hyalinospora, Neoanthostomella bambusicola, Nigropunctata bambusicola, N. nigrocircularis, N. thailandica, Occultitheca rosae, Paravamsapriya ostiolata, Peroneutypa leucaenae, Seiridium italicum and Vamsapriya mucosa) and seven new host/geographical records are introduced and reported. Divergence time estimates indicate that Delonicicolales diverged from Amphisphaeriales + Xylariales at 161 (123–197) MYA. Amphisphaeriales and Xylariales diverged 154 (117–190) MYA with a crown age of 127 (92–165) MYA and 147 (111–184) MYA, respectively. Appendicosporaceae (Amphisphaeriales) has a stem age of 89 (65–117) MYA. Ancestral character state reconstruction indicates that astromatic, clypeate ascomata with aseptate, hyaline ascospores that lack germ slits may probably be ancestral Xylariomycetidae having plant-fungal endophytic associations. The Amphisphaeriales remained mostly astromatic with common septate, hyaline ascospores. Stromatic variations may have developed mostly during the Cretaceous period. Brown ascospores are common in Xylariales, but they first appeared in Amphisphaeriaceae, Melogrammataceae and Sporocadaceae during the early Cretaceous. The ascospore germ slits appeared only in Xylariales during the Cretaceous after the divergence of Lopadostomataceae. Hyaline, filiform and apiospores may have appeared as separate lineages, providing the basis for Xylariaceae, which may have diverged independently. The future classification of polyphyletic xylarialean taxa will not be based on stromatic variations, but the type of ring, the colour of the ascospores, and the presence or absence or the type of germ slit.

As a result, the taxonomic placements of many taxa lacking distinct stromata are uncertain (Daranagama et al. 2018;Wendt et al. 2018). Several recent studies have focused on the morphology and phylogeny of inconspicuous stromatic xylarialean taxa, including the re-examination of herbarium specimens (Daranagama et al. 2018). Those studies not only focused on providing morpho-molecular information but also placed them in higher ranks (e.g., Barrmaeliaceae, Fasciatisporaceae, Induratiaceae and Oxydothidaceae) Voglmayr et al. 2018;Hyde et al. 2020a;Samarakoon et al. 2020c). The genera that have been introduced into new families were previously accepted with uncertain morphologies and phylogenies. They are morphologically unique in having inconspicuous, immersed ascomata, that do not have key characters for delimiting higher ranks as compared to conspicuous stromatic xylarialean taxa. However, the morphology of asci and ascospore were cardinal characters coupled with molecular phylogenies towards establishing new higher ranks. There are many taxonomic uncertainties within xylarialean taxa that are not yet resolved.
We are researching xylarialean taxa towards resolving taxonomic uncertainties. Here we provide new collections with their morphology, analyse their DNA sequences, and investigate their phylogenetic relationships to better identify and classify them. We evaluated different stromatic characters of selected taxa in Xylariomycetidae to reconstruct the ancestral state. In addition, ascospore characters, i.e., colour, septation, and the presence or absence of a germ slit, were evaluated to understand the ancestral state of the xylarialean taxa.

Collection, isolation and morphological studies
Fresh specimens were collected and received from China, Italy, Russia, Thailand and the United Kingdom during the years 2016-2020. External examinations were made as described in Samarakoon et al. (2020b). Indian ink, Congo red and Melzer's reagent were used where necessary. The nature of the ascal ring is important for the release and ejection of ascospores and is often termed an apparatus or ring and we use the latter here. In the descriptions and notes, the ring bluing in Melzer's reagent is referred to as J+ (amyloid), and if not bluing in Melzer's reagent it is referred to as J− (inamyloid).

DNA extraction, PCR amplification and sequencing
Fresh mycelia were scraped from the margins of colonies on PDA plates, incubated at 25-30 °C for 4 weeks. When fungi failed to grow in culture, DNA was extracted directly from the fruiting bodies. Total DNA extraction kits were used according to the manufacturer's instructions [Sangon Biotech (Shanghai) Co. Ltd. China]. The primers and PCR protocols are summarised in Table 1. The total volume of 25 μL contained 12.5 μL of 2 × PCR Master Mix with dye [0.1 U Taq Polymerase/μL, 500 μM dNTP each, 20 mM Tris-HCl (pH 8.3), 100 mM KCl, 3 mM MgCl 2 ], 1 μL of each primer, 9.5 μL of double-distilled water and 1 μL (100-500 ng) of DNA template. All the PCR products were immediately subjected to 4 °C and were visualised on 1% agarose electrophoresis gels stained with GoldView I nuclear staining dye (1 µL/10 mL of agarose) with D2000 DNA ladder (Realtimes Biotech, Beijing, China). DNA sequencing was performed at Sangon Biotech (Shanghai) Co. Ltd., China.
Characters were assessed to be unordered and equally weighted. MrModeltest 2.3 was performed for each locus to estimate the best-fit evolutionary model under the Akaike Information Criterion (AIC) (Nylander 2004). Phylogenies were generated using maximum-likelihood (ML) and Bayesian Inference (BI) analyses using single and ITS-LSU-rpb2-tub2 and ITS-LSU-rpb2-tub2-tef1 combined alignments. For future studies, all the newly generated sequences were deposited in GenBank ) ( Table 2).
The ML analyses were performed with IQ-TREE Trifinopoulos et al. 2016) using the ML + rapid bootstrap setting with 1,000 replicates. The Bayesian tree was generated using MCMC sampling in MrBayes v3.1.2 (Huelsenbeck and Ronquist 2001;Zhaxybayeva and Gogarten 2002) for 10,000,000 MCMC generations using four chains and partition analysis with 100 sample frequencies.  Rehner and Samuels (1994)    The first 25,000 (25% of the total) trees were in the burn-in phase and were discarded. The remaining 75,000 trees were used to calculate the posterior probability (PP). The resulting trees were viewed with FigTree v.1.4.0 (Rambaut 2012), and the final layout was done with Adobe Illustrator® CS5 (Version 15.0.0, Adobe®, San Jose, CA). The final alignment and tree were registered on TreeBASE (http:// www. treeb ase. org/) under the submission ID: 28955.

Divergence time estimation
Divergence time estimation among the families in Xylariomycetidae was performed using the BEAST.v1.10.4 program. The analysis included 240 taxa represented by combined LSU, ITS, rpb2, tub2 and tef1 DNA loci (Supplementary Table 2). The XML file was obtained, including the partitioned alignment, using the BEAUti (BEAST package). The crown age of Xylariomycetidae was used as the secondary calibration node (mean = 168 MYA, SD = 16, Normal distribution) (Samarakoon et al. 2016;Hongsanan et al. 2017). The analysis was performed for 80,000,000 generations using BEAST.v1.10.4 (Suchard et al. 2018), obtaining logging parameters and trees for every 5000 generations. Effective sample sizes (ESS) of parameters were checked using Tracer v.1.6 (Rambaut et al. 2013) (ESS > 200

Ancestral character state analyses
The Bayesian Binary MCMC was performed in RASP 3.2.1 (Reconstruct Ancestral State in Phylogenies) to construct an ancestral character state (Yu et al. 2015). The time-calibrated maximum clade credibility tree reconstructed in BEAST was used for the analysis and exported to RASP 3.2.1. Each terminal in the tree was coded for seven stromatic characters Thailand Dendrocalamus giganteus MW240620 MW240549 --MW759501 -"-" Sequences were not obtained; "*" Regenerated sequences from previously published taxa Mostly single ascomata in an erumpent or superficial carbonaceous stroma (e.g., Astrocystis and Rosellinia) Character 6 Aggregated or single ascomata, mostly immersed or semi-immersed, inconspicuous, prominent or rudimentary carbonaceous clypeus (e.g., Amphisphaeria, Anthostomella, Arthrinium and Seiridium)) Character 7 Aggregated or single, superficial or semi immersed ascomata (e.g., Iodosphaeria and Leptosillia) (Table 3), and undetermined sexual morphs were treated as separate undetermined characters for the family level. In addition, the characters of ascospore septation (aseptate/ septate/apiosporous/undetermined), ascospore colour (hyaline/brown/undetermined) and ascospore germ slit (presence/absence/undetermined) were evaluated (Supplementary Table 2). Bayesian Binary MCMC trees were performed and visualised in RASP 3.2.1 using default settings as follows: 1,010,000 iterations for BayesTraits with a burnin of 10,000, sampling 1000 trees and with 10 ML trees; 50,000 generations for Bayesian Binary MCMC, with 10 chains, a sampling frequency of 100, a temperature of 0.1, state frequencies fixed (JC), and among-site rate variation equal.

Phylogenetic analyses
All gene regions resulted in the GTR + I + G model. Maximum likelihood tree topologies for each gene dataset and combined datasets were compared, and the overall tree topology was congruent to those obtained from the combined dataset. The RAxML analysis of the combined ITS-LSU-rpb2-tub2-tef1 dataset yielded the best-scoring tree ( Fig. 1). Bayesian posterior probabilities from MCMC were evaluated with a final average standard deviation of split frequencies less than 0.01. Delonicicolales clusters as basal to the Amphisphaeriales and Xylariales clades, with 100%/1.00 PP statistical support. Amphisphaeriales and Xylariales form distinct clades with 99%/0.92 PP statistical support, similar to a previous study in Hyde et al. (2020b). Amphisphaeriales comprised 27 clades (Clade Am) including 21 families, while Xylariales (Clade Xy) comprised 31 clades including 16 families. Uncertain clades with a single or few taxa are identified as six in Amphisphaeriales and 15 in Xylariales. Forty-nine of newly generated sequences from our study group with Xylariales and nine with Amphisphaeriales.
One of our collections (HKAS 107015) is similar to Appendicospora hongkongensis and is introduced here as a reference specimen. Two isolates (MFLU 19-2131, HKAS 106988), Neoamphisphaeria hyalinospora gen. et sp. nov. Fig. 1 Phylogram generated from maximum likelihood analysis based on combined ITS-LSU-rpb2-tub2-tef1 sequence data. Four hundred seventy-eight strains are included in the combined sequence analyses, which comprise 4830 characters with gaps. Single gene analyses were also performed and topology and clade stability compared from combined gene analyses. Nine taxa from Diaporthales, Hypocreales and Sordariales are used as out group taxa. Tree topology from ML analysis was similar to BI analysis. The best scoring RAxML tree with a final likelihood value of − 233,649.5773 is presented. The matrix had 2174 (45.01% of all sites) number of constant or ambiguous constant, 2178 number of parsimony informative sites and 3303 number of distinct site patterns. Estimated base frequencies were as follows; A = 0.2702, C = 0.2562, G = 0.2394, T = 0.2342; substitution rates AC = 0.9872, AG = 2.6813, AT = 1.0632, CG = 1.2043, CT = 5.5710, GT = 1.000; gamma distribution shape parameter α = 0.492847. Bootstrap support values for ML equal to or greater than 60%, PP equal to or greater than 0.9 (ML/PP) are given above or below the nodes.

Divergence time estimation
Three clades were obtained in Xylariomycetidae, including 39 families representing the orders Amphisphaeriales, Delonicicolales and Xylariales (Fig. 2)

Character analysis
Ancestral character state analyses resulting from Bayesian Binary MCMC (BBM) are shown in Figs. 2 and 3. Xylariomycetidae was reconstructed as derived from inconspicuous, immersed or semi-immersed ascomata with a prominent or rudimentary carbonaceous clypeus (Character 6) and shared a high percentage among Amphisphaeriales, Delonicicolales and Xylariales. Xylariaceae includes highly variable stromatic characters, and is a diversified group as compared to all the other families in Xylariomycetidae. Conspicuous, erumpent, bipartite, carbonaceous stromatic development (Character 2) and semi-immersed, erumpent or superficial, pseudostromatic development (Character 4) are mostly distributed in Graphostromataceae and Diatrypaceae, respectively. Even though the sexual morphs of Beltraniaceae and Castanediellaceae are undetermined, there is a high possibility that they will have inconspicuous, immersed or semi-immersed ascomata with a prominent or rudimentary carbonaceous clypeus stromata (Character 6), based on evidence from recent ancestors of the clade. It is therefore possible to predict the characters of the sexual morphs in some families that lack known sexual morphs through their ancestral characters. In addition, septate, hyaline ascospores and the absence of a germ slit are ancestral characters of Xylariomycetidae. Apiospores have evolved independently in several clades. Brown ascospores are often found in Xylariales, while Induratiaceae and Vamsapriyaceae have hyaline apiospores. Several xylarialean taxa have ascospores with germ slits, but these are not found in Amphisphaeriales and Delonicicolales. The Amphisphaeriales clade comprises a variety of characters and several groups with undetermined sexual morphs.

Taxonomy
In this paper, we follow the classifications in the studies of Hyde et al. (2020b) and Wijayawardene et al. (2020), and are updated according to recent relevant literature.  Hyde et al. (2020b) and Wijayawardene et al. (2020). Myelospermataceae is accepted in Xylariomycetidae family incertae sedis due to lack of molecular data.

Amphisphaeria Ces. & De Not
Notes: Amphisphaeria is the type genus of Amphisphaeriaceae, with A. umbrina as the type species (Cesati and de Notaris 1863). The genus is characterised by solitary or aggregated ascomata under a poorly-developed clypeus or clypeus lacking; unitunicate asci with J+ or J−, apical rings and light brown to dark brown, ellipsoid to fusiform, 1-3-septate ascospores and coelomycetous asexual morphs. Amphisphaeria accommodates 27 species, which are saprobes on woody branches and some monocotyledons, including grasses (Wang et al. 2004;Samarakoon et al. 2020b).

Fig. 2
Ancestral character state analysis focusing on stromatic characters in Xylariomycetidae, using Bayesian Binary MCMC method. Taxa with the Character 6 are shown in red, and unknown character are shown in blue. Dark red dots in nodes show the highest percentage of the node distribution with stromatic Character 6. The respective colour code represents each family in three orders, Amphisphaeriales (green), Delonicicolales (orange) and Xylariales (blue). Selected stromatic characters are given modified from previous publications (Deepna Latha and Manimohan 2012;Daranagama et al. 2018;Voglmayr et al. 2019a;Becker et al. 2020;Konta et al. 2020a, b) Amphisphaeria parvispora Samarak. & K.D. Hyde, sp. nov.

Apiospora Sacc.
Notes: Crous and Groenewald (2013) re-evaluated the morphology and phylogeny of Arthrinium (= Apiospora). Arthrinium species have densely arranged perithecial ascomata in a longitudinal stroma; clavate to broadly cylindrical asci and apiospores in the sexual and coelomycetous or hyphomycetous asexual morphs. The genus is widely distributed as endophytes, epiphytes, saprobes and plant pathogens on commercial crops and ornamentals (Crous and Groenewald 2013;Hyde et al. 2020b). Arthrinium was expanded with abundant sampling and isolation with morpho-phylo studies while accepting > 70 species in recent years (Hyde et al. 2020b). Pintos and Alvarado (2021) provided molecular data for the type species Ar. cariciola and accepted two genera as Apiospora and Arthrinium. Arthrinium species have variously shaped conidia and inhabit Cyperaceae and Juncaceae in temperate, cold or alpine habitats. Apiospora species have rounded/lenticular conidia and inhabit mainly Poaceae (and many other plant host families) in a wide range of habitats, including tropical and subtropical regions. Nearly 20 Apiospora/Arthrinium species have been recorded from China Farr and Rossman 2021;Feng et al. 2021). Etymology: The specific epithet refers to the location, Guiyang, from where the species was first collected.
Culture characteristics: Colonies on PDA reaching 55 mm diam. after two weeks at 25 °C, cottony, flat, spreading, with moderate aerial mycelium, circular, dense, entire margin, and light brown; reverse brown at the center and dirty white.
Notes: Apiospora sichuanensis clustered with A. pseudoparenchymatica in the combined gene phylogeny. Morphological comparison is not possible due to the lack of a similar morph for both species (Wang et al. 2018 Etymology: Named after the type genus, Appendicospora. Saprobic on dead rachis/fronds of palms and dicotyledonous twigs. Sexual morph: Ascomata immersed, under slightly raised areas, visible as brown or black dots, solitary or aggregated in clusters, in cross-section, conical to subglobose with a mostly flattened base. Ostioles centric, ostiolar canal periphysate or filled with white amorphous tissues. Peridium multi-layered, outer layer comprising brown, thick-walled, flattened cells of textura angularis, inner layer composed of hyaline, thin-walled cells of textura angularis. Paraphyses wider at the base, septate, embedded in a gelatinous matrix. Asci 8-spored, unitunicate, clavate to cylindrical, short pedicellate or sessile, lacking an apical ring, apically rounded. Ascospores uniseriate or 2-3-seriate, hyaline, clavate to broadly ellipsoidal, 1-septate, not constricted at septa, with or without appendages at one end. Asexual morph: Undetermined. Type genus: Appendicospora K.D. Hyde Notes: Appendicospora shares similar morphologies with Apiospora and Pseudomassaria with an uncertain taxonomic placement (Hyde 1995a, b;Bahl 2006). Several morphophylo studies showed that Appendicospora consistently grouped with Hyponectria and was best placed within the Hyponectriaceae, although it was suggested that further work was needed to confirm the taxonomic placement (Wang and Hyde 1999;Smith et al. 2003;Bahl 2006). The only available LSU sequence of Appendicospora sp. (HKUCC 1120) links the morphology and phylogeny of this group. The combined gene phylogeny in our study shows that Appendicospora forms a distinct clade from Apiosporaceae in the Amphisphaeriales. In addition, an inconspicuous taxon introduced as Neoamphisphaeria in this study clustered with Appendicospora with high statistical support (100%/1.00 PP). Appendicosporaceae forms a strong statistically supported clade with Anungitiomycetaceae, Iodosphaeriaceae, and Pseudomycetaceae. In addition, the divergence time estimates show that Appendicosporaceae has diverged at 89 (65-117) MYA (Amphisphaeriales), which is comparable with the common divergence trend at the family level (50-150 MYA) as described in Hyde et al. (2017). Based on distinct morphologies, phylogeny and divergence time estimates, we introduce Appendicosporaceae with the type genus Appendicospora and tentatively accommodate Neoamphisphaeria.
Notes: Neoamphisphaeria is similar to Amphisphaeria in having immersed ascomata with a brown peridium, long hyaline paraphyses, cylindrical asci and ellipsoid, 1-septate, mature ascospores. The distinct characters of Neoamphisphaeria are the ostiolar canal filled with amorphous hyaline cells, asci with a bilobed or dome-shaped apical ring and hyaline ascospores. In the phylogeny, Neoamphisphaeria clusters (100%/1.00 PP) with Appendicospora, which has a periphysate ostiolar canal, clavate asci and 2-3-seriate, hyaline, clavate ascospores with a bifurcated (moustache-shaped) appendage. With inconspicuous, immersed ascomata, asci with short or lacking pedicels with J−, apical ring and 2-celled hyaline ascospores and strong phylogenetic evidence, we accept Neoamphisphaeria as a new genus in Appendicosporaceae.
Notes: Neoamphisphaeria hyalinospora is similar to Amphisphaeria with its subglobose to conical ascomata, cylindrical asci with a short pedicel and 2-celled ascospores, but differs in the amorphous cells in the ostiolar canal and hyaline ascospores. The size and shape of N. hyalinospora asci and ascospores are similar to Keissleriella hyalinospora (formerly known as Amphisphaeria hyalinospora) (Müller and von Arx 1962). However, A. hyalinospora has dark brown, sparsely bristly hairs in the ostiolar canal, whereas N. hyalinospora has hyaline amorphous cells. Combined gene phylogenies showed that Neoamphisphaeria is not related to Amphisphaeria, but clustered with Appendicospora with high statistical support (100%/1.00 PP). Based on unique morphology and distinct phylogeny, we introduce N. hyalinospora as a new species.

Melogrammataceae G. Winter
Melogrammataceae was introduced by Winter (1887) to accommodate Melogramma. Species of the family are saprobes or hemibiotrophs on the bark of woody plants. Based on morphology and phylogenetic revisions, Jaklitsch and Voglmayr (2012) accepted Melogrammataceae in Xylariales. Senanayake et al. (2015) re-evaluated the phylogeny of Melogrammataceae and accepted it in Amphispheriales, and this has been confirmed in later studies by Hongsanan et al. (2017) and Hyde et al. (2020b).

Melogramma campylosporum
Culture characteristics: Colonies on PDA, reaching 9 mm diam. after three months at 25 °C, convex and a papillate surface, compact, lobate with zonate margin, hyphae embedded in the media, dark greenish brown; reverse dark brown in the center, yellowish brown marginal area, media becoming reddish brown.  (Farr and Rossman 2021). The LSU and ITS ses. j-m Asci (l, k in Melzer's reagent). n-s Ascospores. Scale bars: a-f = 1000 μm, g = 100 μm, j-m = 50 μm, h, n-s = 20 μm, i = 5 μm sequences of our two collections are identical to the ITS-LSU of acc. JF440978 (CBS 141086). The microscopic characters of our two collections are similar to M. campylosporum, described by Jaklitsch and Voglmayr (2012). This is the first record of M. campylosporum on Corylus avellane in Italy.

Sporocadaceae Corda
Sporocadaceae species typically possess appendage bearing conidia and are important as endophytes, saprobes or pathogens on leaves, twigs, branches, fruits of flowering plants and gymnosperms, and as parasites on humans and animals Hyde et al. 2020b). Following recent revision of the morphology and multigene phylogeny, there are 23 genera in Sporocadaceae .
Culture characteristics: Colonies on PDA reaching 23-25 mm diam. after four weeks at 25 °C, circular, flat, entire margin, white and yellowish brown as concentric zones; yellowish orange center, salmon pink marginal area.
Notes: Seiridium italicum is only known from its sexual morph and is similar to the generic description except it has J−, apical rings (Bonthond et al. 2018). Nonappendiculata and Nothoseiridium are close to S. italicum in phylogeny but lack sexual morphologies Crous et al. 2020 Xylariales Nannf.

Diatrypaceae Nitschke
Diatrypaceae was introduced by Nitschke (1869), which comprises saprobes, pathogens and endophytes on economic crops and forest trees worldwide, also occurring in aquatic habitats (Hyde et al. 2020b;Konta et al. 2020a (2017) with the type species A. polyspora on a dead twig of Hevea brasiliensis from Thailand. Allocryptovalsa has immersed ascomata, polysporous asci, and allantoid ascospores and forms a distinct phylogenetic affinity in Diatrypaceae with ITS-tub2 phylogeny. There are five Allocryptovalsa species. Etymology: The specific epithet refers to the location, Sichuan, from where the fungus was first collected.
Notes: Senanayake et al. (2015) designated a reference specimen for Diatrype disciformis (MFLU 15-0722) on a branch of Ostrya carpinifolia from Italy. In this study, we provide another specimen collected from the UK, which is morphologically similar with overlapping measurements of ascomata, asci and ascospores. However, our collection has larger stromata (3.3-5.6 mm vs 1.5-2 mm length) and a thinner peridium (17-20 μm vs 20-30 μm) as compared to MFLU 15-0722. Mehrabi et al. (2016) also described D. disciformis, which has up to 8 mm diam. of stromata. The known distribution of the fungus is on Alnus and Fagus from Europe and lead-contaminated soils from abandoned firing range in the USA (Acero et al. 2004;Vasilyeva and Stephenson 2009;2014;Sullivan et al. 2012;Vasilyeva and Ma 2014;Mehrabi et al. 2016).
Culture characteristics: Colonies on PDA reaching 30 mm diam. after one week at 25 °C, irregular, flat or effuse, fluffy to fairly fluffy, medium dense, margin fimbriate, and dirty white; reverse yellowish brown. Notes: Cultures were obtained by isolation from the internal tissue of the stromata. We did not see any spores in the stromata, but observed an asexual morph, which is similar to diatrypaceous asexual morphs on the stromata. Two areas in the stroma produced ascoma-like structures with a brown peridium filled with amorphous cells, probably immature ascomata. However, these observations did not provide enough morphology for the complete identification of the strain. Our strain forms a distinct single clade in Eutypa. The ITS and LSU sequences of our strain are 98% (4/559 gaps) and 99% (3/887 gaps) similar to those of E. lata, respectively. Combined ITS-tub2 (data not shown) and ITS-LSU-rpb2-tub2-tef1 ( Fig. 1) phylogenies showed that our new strain clustered in Eutypa sensu stricto as sister to E. armeniacae (ATCC 28120) with poor statistical support (58%/0.99 PP). Here we propose E. camelliae as a new species. The species was collected on a decayed stem which was a remaining stem part after pruning, and it would be interesting to carry out further investigations for pathogenicity on Camellia japonica. Etymology: The generic epithet refers to the Greek: melano-"black" + stictus "spot".
Halodiatrype and Pedumispora are the closest genera to our new genus. Halodiatrype has immersed ascomata, papillate ostioles with a brown outer amorphous layer and inner yellow cells of textura porrecta, 8-spored, unitunicate asci with oblong to allantoid or sub-inaequilateral, aseptate to septate, light brown ascospores , while Pedumispora has 1-4 immersed ascomata per stroma, papillate ostioles, 8-spored, fusiform asci with filiform, multi-septate, curved, longitudinally striate ascospores (Hyde and Jones 1992). Melanostictus clustered sister to Halodiatrype and Pedumispora as a distinct group in morphology and phylogeny. Hence, we introduce our two new collections into the new genus.
Culture characteristics: Colonies on PDA reaching 50-52 mm diam. after one week at 25 °C, irregular, flat or effuse, slightly raised, margin fimbriate, with aerial mycelium, and dirty white; reverse yellow to pale brown at the center, white at the margin.
Culture characteristics: Colonies on PDA reaching 50 mm diam. after one week at 25 °C, flat, circular, dense, margin entire, aerial mycelia, light brown; reverse brown at center with dirty white edges.
Notes: Zhang et al. (2017) introduced Biscogniauxia petrensis from a karst cave based on phylogeny and asexual morphology. Ma et al. (2020) reported B. petrensis as an endophyte in the root of Dendrobium harveyanum in Thailand, and Das et al. (2020) found it in adult mosquitoes from Korea. The conidia size of our isolate (4.3-6.5 × 2-3.5 μm) overlaps with the CGMCC 3.17912 (type) (4.5-7.5 × 2.5-4.5 μm) and MFLUCC 14-0151 (3.5-5 × 2.5-3 μm). The BLAST searches also confirmed that our isolate is similar to B. petrensis, and here we provide a new host record from Osmanthus fragrans in China.
Culture characteristics: Colonies on PDA reaching 55 mm diam. after two weeks at 25 ºC, flat, circular, dense, with a smooth surface, margin entire, and white to light brown.
Notes: Camillea tinctor exhibits a wide distribution in Thailand, and has been described on various timbers (Thienhirun 1997; Whalley et al. 1999;Suwannasai 2005). There is a slight difference in the size of stromata from different locations with overlapping size ranges of ascomata, asci and ascospores described from different studies (Whalley et al. 1999;Suwannasai 2005;Daranagama et al. 2018). Conspicuous orange staining of the host substrate is often reported, but we did not observe such staining in our specimen, as did San Martín González and Rogers (1993). Our collection is provided as a new host record from dead Bauhinia racemosa wood, and there is a need to determine if the fungus affects plants as a pathogen.

Lopadostoma quercicola
Culture characteristics: Colonies on PDA reaching 45-48 mm diam. after three weeks at 25 ºC, flat, dense, center raised, cottony, margin undulate, circular, whitish grey; reverse light brown and with dirty white zonations.  Notes: Jaklitsch et al. (2014) introduced Lopadostoma quercicola on Quercus pubescens from Austria, and the species is similar in morphology and phylogeny to our new collections. Size and shape of ascospores of our collections (8.3-11.4 × 3.4-4.4 μm) overlap in the range of WU 32079 (holotype) (9.5-12 × 4.3-5 μm). A BLAST search of ITS data of our strains is 100% similar with L. quercicola (WU 32079). The known host and geographical distribution is widespread in corticated branches of Quercus species from Europe, and from Austria, Croatia, France, Italy and Portugal on Q. cerris, Q. petraea, Q. pubescens, Q. robur and Q. suber. Morphology and phylogeny of our new collections are similar to Lopadostoma quercicola, and here we provide a new host record, Q. cerris from Italy.

Acrocordiella O.E. Erikss
Notes: Acrocordiella was introduced by Eriksson (1982) with the type species A. occulta. The genus is characterised by solitary or small groups of immersed ascomata without a clypeus, bitunicate-like asci with a thick-walled apex and wide ocular chamber, and ellipsoid to oblong ascospores with round or acute ends, with one or several transverse distosepta and large lumina (Jaklitsch et al. 2016). Three Acrocordiella species have been accepted with morphology and molecular data (Dissanayake et al. 2021b).
The LSU-ITS phylogenies in Jaklitsch et al. (2016) and Maharachchikumbura et al. (2018) showed that Acrocordiella and Requienella form distinct clades. With the addition of A. yunnanensis, LSU-ITS phylogeny showed A. occulta is not monophyletic and clusters in Requienella (Dissanayake et al. 2021b). Similarly, our ITS-LSU and ITS-LSU-rpb2-tub2-tef1 combined gene phylogenies confirmed that Acrocordiella is not monophyletic, even though they share similar morphologies. Boise (1986)

Vamsapriya Gawas & Bhat
Notes: Gawas and Bhat (2005) introduced Vamsapriya to accommodate V. indica described on dead and decaying bamboo twigs from India. The genus is characterised by catenate, phragmosporous conidia on synnematous conidiophores with non-cicatrised, monotretic conidiogenous cells in the asexual morph. Dai et al. (2017) linked the asexual and sexual morphs of Vamsapriya in a multigene phylogeny. Eight Vamsapriya species have been described, mostly on dead bamboo substrates, and the genus is accepted in Xylariaceae (Hyde et al. 2020b).
Apioclypea was introduced by Hyde (1994a, b), and is characterised by immersed, clypeate ascomata; cylindrical asci with J + /J−, apical ring and hyaline, apiospores covered by a mucilaginous sheath. Morphological comparisons between Apioclypea and Vamsapriya showed that both genera have similar morphologies in having immersed, clypeate ascomata, asci with short pedicels and hyaline apiospores with a mucilaginous sheath (Hyde 1994a, b;Hyde et al. 1998;Taylor and Hyde 2003). Two of our new collections clustered with Vamsapriya and are morphologically similar to the only known species, V. bambusicola. However, Vamsapriya has a thinner peridium and asci with J + , apical ring, as compared to Apioclypea with thicker peridia and some species with a J−, apical ring (A. nonapiospora, A. phoenicicola). Apioclypea has been placed in Clypeosphaeriaceae and Hyponectriaceae in different studies (Hyde 1994a, b;Hyde et al. 1998;Kang et al. 1999). Kang et al. (1999) re-examined the holotype of the generic type, A. livistonae and mentioned a J−, apical ring, which is J + in the generic description by Hyde (1994a, b). Currently, Apioclypea has seven species but lacks molecular data. Apioclypea sp. (HKUCC 6269) LSU sequence available in the GenBank (https:// www. ncbi. nlm. nih. gov/ nucco re/ AY083 836.1) is similar to unconfirmed taxa that belong to Lanceispora, Polyancora (Amphisphaeriales genera incertae sedis) and Leiosphaerella (Pseudomassariaceae). We treat Apioclypea and Vamsapriya as distinct genera until molecular data for the type of Apioclypea becomes available.
Notes: Paravamsapriya ostiolata is described from dead bamboo branches in Thailand, which has distinct ostioles, J− refractive apical rings, and hyaline, ellipsoidal to fusiform, aseptate, smooth-perispored ascospores with thin mucilaginous caps at the ends. The combined phylogeny shows that two isolates of P. ostiolata form a distinct basal clade in Vamsapriyaceae (Fig. 1, Clade Xy10).
Xylariaceous species are distributed worldwide as saprobes, pathogens and endophytes in wood, leaves and fruits or associated with insects. vectors. Maharachchikumbura et al. (2016) accepted 87 genera in the family, and Daranagama et al. (2018) accepted 37 genera on the re-examination of herbarium specimens. Following several recent studies, Hyde et al. (2020b) accepted 32 genera in the Xylariaceae.

Helicogermslita Lodha & D. Hawksw
Notes: Hawksworth and Lodha (1983) introduced Helicogermslita to accommodate a xylarialean species with ascospores having spiral germ slits and typified by H. celastri. Apart from this, Helicogermslita species share stromatic ascomata with white ectostroma and asci with J + , apical rings (Petrini 2003;Daranagama et al. 2018). Nine species have been described based only on morphology.  Fig. 26 Etymology: The specific epithet refers to the prominent clypeus.
Notes: Helicogermslita clypeata has immersed or semiimmersed or rarely superficial ascomata when mature, with a prominent clypeus and white ectostroma, J + , wedge-or inverted, hat-shaped, apical ring and broadly ellipsoidal ascospores with spiral germ slits, which are comparable to the generic type of Helicogermslita (Hawksworth and Lodha 1983). In addition, carbonaceous stromatic variations can be observed among the species (Laessøe and Spooner 1993). Ascospores of H. fleischhakii, H. gisbornia, H. johnstonii and H. mackenziei bear a cellular appendage, which is lacking in H. clypeata (Petrini 2003). Helicogermslita aucklandica and H. clypeata ascospores are similar in shape with one round spiral germ slit, while they differ in having larger ascospores (19-23.5 × 10-14 μm vs 13-18.5 × 5.7-7.5 μm) (Petrini 2003). Helicogermslita fleischhakii and H. gaudefroyi have a mucilaginous sheath around ascospores, which differs from H. clypeata (Laessøe and Spooner 1993;Petrini 2003). Helicogermslita clypeata differs from other Helicogermslita species in having large asomata with a wide clypeus. Yuea chusqueicola possesses spiral germ slits in their ascospores, but differs from H. clypeata in having small ascomata, large asci and ascospores covered with a mucilaginous sheath (Eriksson 2003). In addition, Anthostomella limitata and A. umbrinella have similar ascospores with spiral germ slits, but differ in having a mucilaginous sheath (Hawksworth and Lodha 1983;Lu and Hyde 2000;Petrini 2003). There is no molecular data for Helicogermslita, and it has been accepted in Xylariaceae based only on morphology.
In our phylogeny, Helicogermslita clypeata clusters in the Xylaria "PO" clade (U'Ren et al. 2016;Konta et al. 2020b), which is in Xylaria sensu lato. Xylaria "PO" clade comprises Amphirosellinia, Astrocystis, Collodiscula, Kretzmariella, Stilbohypoxylon and Xylaria species. Based on its distinct phylogeny and similar morphology, our collection is introduced as a new species of Helicogermslita in Xylariaceae. However, Leptomassaria simplex, the type species of Leptomassaria has ascospores with a spiral germ slit and a thick mucilaginous sheath similar to Helicogermslita (Hawksworth and Lodha 1983;Rappaz 1995). Lu and Hyde (2000) mentioned that Leptomassaria differs from Helicogermslita in having immersed ascomata and an irregular, wedge-shaped, ascal, apical ring. Daranagama et al. (2018) mentioned the possibility of linking Helicogermslita and Leptomassaria with observations based on herbarium specimens. In addition, we observed that Leptomassaria has poorly developed stromata, larger ascomata, Notes: Hypocopra is typified by H. merdaria and characterised by sessile, clypeoid, or reduced stromata; asci with complex apical rings and mostly aseptate, brown ascospores with a germ slit. Hypocopra species exclusively inhabit dung. Krug and Cain (1974) described 14 new species and one combination discovered on different dung types. There are 31 species, including three species with molecular data (Becker et al. 2020).
Notes: The macro morphology of our collection is similar to N. pouzarii in having mammiform stromata with a broad base . Nemania pouzarii has long asci with a long stipe, while N. paraphysata has short asci without a stipe (110-120 μm vs 65-105 μm). Nemania paraphysata has two types of paraphyses, and one of the types is trabeculae-like. The ITS sequence of N. paraphysata is similar to that of N. pouzarii ATCC 2612 (91%, 12/475 gaps) and N. macrocarpa CBS 109567 (88%, 22/362 gaps), while rpb2 is similar to that of N. maritima JF04055 (90%, 0/808 gaps) and N. macrocarpa WSP 265 (83%, 6/1046 gaps). The phylogenetic analyses show that N. paraphysata clusters basal to N. pouzarii and N. thailandica with 100% statistical support. With the presence of trabeculae-like paraphyses Ascospores (germ slit in white arrow; q in Indian ink). Scale bars: a = 1 cm, b-e = 1000 μm, g-j = 20 μm, f = 10 μm, k-q = 5 μm and ascospores with a knob at the base, N. paraphysata is introduced as a new species.
Notes: Our species is closely related to the Rosellinia emergens group as described in Petrini (2013), which consists of ascospores with L/W ≥ 4 and stromata generally < 1 mm × < 1 mm. A morphological comparison of R. capetribulensis, R. emergens, R. formosana, R. longispora, R. markhamiae, R. megalosperma and R. megalospora clearly placed our specimen in R. markhamiae (Sivanesan 1975;Bahl et al. 2005;Petrini 2013;Xie et al. 2019). All the above species share features of a germ slit along the entire ascospore length and surrounded at each end by slimy caps, except for R. capetribulensis, which has a straight germ slit and thin mucilaginous sheath and lacks slimy caps. Our specimen is similar to R. markhamiae, which was described on Markhamia hildebrandtii from Tanzania in having overlapping ascospore sizes (70-105 × 11.5-15 μm vs 72-135 × 9-14 μm) and J + apical ring (15-18.4 × 9.2-13 μm vs 13-15 × 7-11 μm). Here we introduce a new geographical record and the phylogenetic placement of R. markhamiae from Thailand, although further collections may show this to be a distinct species.

Anthostomella Sacc.
Notes: Anthostomella is a species-rich, polyphyletic genus, which is characterised by immersed ascomata beneath a dark clypeus, periphysate ostiolar canals, unitunicate, cylindrical asci with or without a J + , apical ring and mostly brown, aseptate ascospores with or without a dwarf cell or appendages at the ends and presence or absence of a germ slit (Lu and Hyde 2000;Daranagama et al. 2015). Saccardo (1875) introduced Anthostomella with three species; A. limitata, A. tomicoides and A. perfidiosa without designating a type. Different studies have accepted A. limitata (Eriksson 1966) and A. tomicoides (Francis 1975;Lu and Hyde 2000) as the generic type, emphasizing the generic description provided by Saccardo (1875). Eriksson (1966) interpreted Saccardo (1875)'s description and prioritised the non-appendiculate over appendiculate ascospores. Only A. limitata has non-appendiculate ascospores from the original three collections and is accepted as the generic type (Eriksson 1966). Francis (1975) argued that the original generic description is based on both appendiculate and non-appendiculate ascospore morphologies and the absence of clypeus in A. limitata is not well-suited to Anthostomella. However, A. limitata has a blackened clypeus (Lu and Hyde 2000). Following Lu and Hyde (2000) with the lectotype of A. tomicoides, Daranagama et al. (2015) accepted A. tomicoides as the generic type. Furthermore, Daranagama et al. (2015) provided a reference specimen for appendiculate ascospore bearing A. formosa and the clade accepted as Anthostomella sensu stricto.
Occultitheca J.D. Rogers & Y.M. Ju Notes: Occultitheca is a monospecific genus typified by O. costaricensis identified from decayed wood in Costa Rica (Rogers and Ju 2003). The genus is characterised by immersed ascomata, short pedicellate asci with J + , apical ring, and brown ascospores with hyaline dwarf cells and a straight germ slit. Rogers and Ju (2003) emphasised the large distance between the uppermost ascospore and the ascus apex as a key feature of the genus. Etymology: The specific epithet refers to the host genus Rosa.
Notes: There are several Anthostomella species that have asci with a J + , apical ring, ascospores comprising a large brown cell with a straight germ slit and apical or basal hyaline dwarf cell, such as A. foveolaris, A. hemileuca, A. tomicoides and A. unguiculata and differ with a short distance between the uppermost ascospore and the ascus apex as a key feature of the genus (Lu and Hyde 2000;Rogers and Ju 2003). Anthostomella tomicoides ascospores have very short germ slits, whereas Occultitheca rosae ascospores have a straight germ slit along the entire spore length (Lu and Hyde 2000). In addition, A. clypeata and A. clypeoides are similar to Occultitheca rosae in having immersed, clypeate ascomata, J + , apical ring and ascospores with a brown large and hyaline dwarf cell. The ascospores of both species lack germ slits and A. clypeoides has a mucilaginous sheath in their ascospores that differs from Occultitheca rosae (Lu and Hyde 2000;Lee and Crous 2003). Francis (1975) described A. sabiniana on Pinus sabiniana needles, which is similar to Occultitheca rosae in having clypeate, immersed ascomata, a J + apical ring, inequilateral, ellipsoidal, brown ascospores with a large and thickened cap at the polar end, straight germ slit and mucilaginous sheath, and short distance between the uppermost ascospore and the ascus apex. Occultitheca rosae mostly has 1-2 individual ascomata, which differs from O. costaricensis in having 2-12 ascomata in a cluster. In addition, O. rosae differs from O. costaricensis in having ascospores with a thin mucilaginous sheath. The  Notes: Daranagama et al. (2016a) introduced Pseudoanthostomella, with its type P. pini-nigrae, to accommodate anthostomella-like species. Pseudoanthostomella is similar to Anthostomella in having blackened, conical to dome-shaped, semi-immersed to immersed ascomata, which are mostly solitary or rarely aggregated in small groups, while it differs from Anthostomella in having eccentric, periphysate, ostiolar canals, distinctly clavate asci and ascospores with straight germ slits (Daranagama et al. 2016a). Furthermore, Pseudoanthostomella differs from Neoanthostomella in having asci with J + , apical ring and ascospores with germ slits and differs from Alloanthostomella in having aseptate, pigmented ascospores with germ slits. In this study, we re-examined the holotype of Pseudoanthostomella (Daranagama et al. 2016a). Saprobic on a dead aerial branch of Cytisus sp. Sexual morph: Ascomata 235-320 × 230-305 μm (x̄ = 275 × 255 μm, n = 10), immersed, visible as black, raised area, solitary, in cross-section globose to subglobose, with flattened top. Clypeus carbonaceous, black, thick-walled, short, comprising dark fungal hyphae and host epidermal cells. Ostioles centric or eccentric, ostiolar canal periphysate. Peridium 21-30 μm (x̄ = 25.5 μm, n = 15) wide, with two cell layers, outer layer comprising yellowish brown, thick-walled cells of textura angularis, inner layer composed of hyaline, thin-walled cells of textura angularis. Paraphyses 2.8-4.8 μm (x̄ = 3.7 μm, n = 20) wide, wider at the base, slightly shorter than the asci, numerous, guttulate, filamentous, septate, branched, constricted at septa, apically blunt. Asci 95-125 × 10-13.5 μm (x̄ = 110 × 12 μm, n = 20), 8-spored, unitunicate, cylindrical to clavate, pedicel short or absent, apically rounded, with a 1.6-2.2 × 4.5-5.6 μm (x̄ = 1.9 × 5 μm, n = 5), J + , discoid, inverted, hat-shaped, apical ring, bluing in Melzer's reagent. Ascospores 12.5-16 × 7.5-9 μm (x̄ = 14.5 × 8.3 μm, n = 25), L/W 1.75, overlapping uniseriate, dark brown to black, broadly ellipsoidal, aseptate, 1-2-guttulate, covered with a 2-3 μm (x̄ = 2.5 μm, n = 5) thick, mucilaginous sheath, germ slit on the ventral side of the ascospore, straight, along the entire spore length. Notes: We studied two collections of Pseudoanthostomella pini-nigrae collected on dead aerial branches of Cytisus sp. and a dead leaf of a cultivated plant from Italy and the UK. Both collections possess centric or eccentric and periphysate ostiolar canals, cylindrical to clavate asci 12-17 × 5.5-9 2.14 2.6-4.4 Grass and ascospores with straight germ slits. MFLU 15-3608 (Italy) has a noticeable black patch on the host surface and centric/eccentric ostioles as compared to MFLU 18-0877 (UK), which has only a centric ostiole. However, this could probably be due to the texture of the host tissue. These two collections are similar to the holotype of P. pini-nigrae in having globose-subglobose ascomata beneath a clypeus, a two-layered peridium, 8-spored, unitunicate, cylindricalclavate, short pedicellate asci with a discoid-inverted, hatshaped, J + , apical ring and broadly ellipsoidal ascospores with a thick mucilaginous sheath and a straight germ slit on the ventral side of the ascospore along the entire spore length (Daranagama et al. 2016a). We re-examined holotypes of P. delitescens, P. pini-nigrae, P. senecionicola and P. thailandica. Our specimen clusters with P. pini-nigrae and shares similar characters, such as globose to subglobose ascomata; cylindrical-clavate asci with J + , apical rings and equilateral ellipsoidal ascospores with a straight germ slit along the entire spore length. The overlapping measurements are given in Table 5. However, the presence of the clypeus varies from prominent to rudimentary, which may be due to the host texture.
Notes: Our specimen has superficial, mammiform, stromata, whereas the holotype of Xenoanthostomella chromolaenae has immersed clypeate stromata. We re-examined the holotype of X. chromolaenae (MFLU 20-0048) and confirmed the presence of a J + , discoid, apical ring in the asci and ascospores with a sigmoid germ slit (which were not described in the original description). Morphological comparison shows that both collections have an overlapping range in dimensions of paraphyses (2.5-4 μm vs 2.2-3.1 μm), J + , discoid, apical ring bearing asci (60-98 × 5-7.5 μm vs 64-80 × 6.4-8.5 μm) and ascospores (10.5-14 × 4-5.5 μm vs 9-12 × 3.4-5 μm) with sigmoid germ slits ). However, our specimen differs from the X. chromolaenae holotype in having superficial, stromatic, larger ascomata. The superficial and immersed stromatic character might be due to the hard surface of the rachis of Nephrolepis sp., while a dead stem of Chromolaena odorata has a soft surface. Anthostomella limitata, A. spiralis and A. xuanenesis have similar cylindrical asci with J + , discoid, apical rings, and ascospores with sigmoid/ diagonal germ slits, but differ in having immersed ascomata under a clypeus as compared to a superficial stroma in X. chromolaenae (Francis 1975;Lu and Hyde 2000;Taylor and Hyde 2003). The LSU and ITS sequences of our specimen are identical to those of X. chromolaenae (MFLUCC 17-1484). Based on similar morphology and phylogeny, we provide a new host record, Nephrolepis sp. for X. chromolaenae. If these are the same species, it would be most likely that X. chromolaenae has jumped to the introduced weed Chromolaena odorata from Nephrolepis sp. (a fern), which may be its original host.

Stromatic variation in xylarialean taxa
A stroma is an aggregation of vegetative mycelium with more or less differentiation, mainly based on the properties of the substrate or on the factors affecting growth, which cause variations in the type or amount of mycelium produced (Wehmeyer 1926;Laessøe and Spooner 1993). The formation of a stroma begins with the development and differentiation of immersed mycelium, appearing as a blackening of the substrate surface and followed by proliferation of mycelia (Wehmeyer 1926). The marginal zone comprises blackened tissues that protect the stromatic areas and provide nutrition for the developing ascomata (Wehmeyer 1926). A pseudostroma is an altered form of stromata consisting of host and fungus tissues (Laessøe and Spooner 1993) and is probably an intermediate evolutionary form between a clypeus and true stromata. Early mycologists used the position, structure, and arrangement of perithecia and stromata to distinguish the families of ascomycetous taxa (Barr 1987). For instance, Wehmeyer (1926) stated that it is more important to discuss the varying characters of the stromata than fixed types of stroma, i.e., eutypoid and valsoid.
Several genera that form conspicuous stromata are accommodated in the Xylariales, and the stromatal morphology often comprises firm, compact, external regions (ectostroma) and an inner region of stromata composed of loosely interwoven hyphae or pseudoparenchymatous cells, called entostroma (Barr 1990). The texture of stromata varies from fleshy, brittle and carbonaceous to tough and woody with different sterile and fertile regions. The diverse shapes and size of the stromata can range from sessile to stalked, subglobose or globose and erumpent from the substrate or superficial. The forms are crustose, applanate or pulvinate, and sometimes effuse. The pigmentation and shape of the final stromata structure can be recognizable as the ectostroma and entostroma, or in the incorporation of substrate cells known as a pseudostroma (Barr 1990).
In the family Diatrypaceae, extensive use is made of stromatic characters for generic segregation (Glawe and Jacobs 1987). Acero et al. (2004) grouped stromata of diatrypaceous taxa into diatrypoid, eutypoid and valsoid. However, it is emphasised that the evolutionary relationship among those stromatic characters among genera are not well-defined. The factors affecting the development of stromata, such as host species, substrate structure and environmental humidity, might be a reason for intra and interspecific stromatic variations (Rappaz 1987). However, in some classifications of Xylariales (e.g., Helicogermslita and Lopadostoma), the colour of stromatic tissue between the ostiolar necks or the presence or absence of a black stromatic line around individual ascomatal clusters are vital characters for species delimitation, rather than the diameter of stromata, the number of ascomata and their size (Barr 1987;Jaklitsch et al. 2014).
A clypeus is a form of modified stromatic features with stromatic tissue or melanised hyphae developing above the immersed or semi-immersed ascomata, which are shieldshaped and variable in development (Laessøe and Spooner 1993). Sometimes, several stromata are contained within a region of the substrate delimited from noninfected tissues by a blackened ventral and sometimes a dorsal zone or line (Barr 1990). Since stromatic tissues are formed from vegetative hyphae, they are considered a stable character (Barr 1990). In past studies that were published before molecular evidence, many taxa were lumped in Anthostomella or Xylaria in Xylariales, which were classified, based on stromatic characters, either being conspicuous (Xylaria) or inconspicuous (Anthostomella) stromata (Lu and Hyde 2000;Peršoh et al. 2009;Hsieh et al. 2010;Daranagama et al. 2018).
Even though the stromatic characters are easy to observe, they have sometimes been ignored, misinterpreted or overemphasised (Rogers 1979). The stromatic characters, such as size, shape, colour and external roughness, are sometimes highly variable and difficult to use for taxon identification (Rogers 1979). Thus, it is important to correlate other microscopic characters, such as conidial features, asci, and ascospores (Rogers 1979). Furthermore, Rogers (1979) suggested that the superficial stromatic nature of Xylariaceae and Sordariaceae probably originated from an astromatic common ancestor through similar environmental conditions. Even though there is a need to use stromatic variations for generic delimitation of some xylarialean taxa, the family level stromatic variations are more important for general identifications (Daranagama et al. 2018).

Significant aspects of micro characters
Besides the stromatic characteristics, amyloid reactions and the shape and size of the ascal apical ring are used for the identification of xylarialean taxa. The ascal apical ring evolved for the improved ejaculation of ascospores and is a character of most xylarialean taxa. Exceptions have formed in xylarialean taxa and have most likely been lost due to lifestyle adaptions. Suwannasai et al. (2012) proposed six types of apical ring, stacks of small rings (e.g., Hypocopra and Poronia), discoid or triangular (e.g., Daldinia and Hypoxylon), broad to discoid (e.g., Biscogniauxia), rhomboid to diamond-shaped (e.g., Camillea), inverted hat or urn-shaped (e.g., Kretzschmaria, Nemania, Rosellinia and Xylaria), and no visible apical ring under the light microscope (e.g., Ascotricha and Rhopalostroma). According to the reaction of the apical ring in Melzer's reagent, they are consistently iodine positive, consistently iodine-negative and vary in reaction (Suwannasai et al. 2012). The different ring types are likely to have evolved across xylarialean taxa and should reflect distinct genera when rings are distinct. We observed that the apical ring in some specimens stained blue with Melzer's reagent for only a short time. This may lead to misidentification in those taxa having a J−, apical ring (e.g., Nigropunctata bambusicola and Pseudoanthostomella senecionicola) when the species delimitation is based on the amyloid reaction. The amyloid reaction is constant within most taxa, but the interpretation can be problematic. However, positive reactions will continue to be a cardinal character for xylarialean taxa in delimiting species (Rogers 1979). In addition, the ascus stipe length is also important in xylarialean taxonomy (Hsieh et al. 2010). This study introduces Nemania longipedicellata with a characteristically long pedicel among closely related mammiform stromata forming Nemania species.
The number of ascospores per ascus, colour, septation and the germ slit play a vital role in the identification of xylarialean taxa. The number of ascospores is usually eight but may vary from four to polysporous (Barr 1990). Diatrypaceous generic delimitation is determined by the number of ascospores per ascus as a key character (Barr 1990;Carmarán et al. 2006;Senwanna et al. 2017;Konta et al. 2020a).
Another important character is the germ slit, through which the germ tube emerges. It is a typical structure in Xylariomycetidae, but is also found in other Sordariomycetes families (e.g., Auratiopycnidiellaceae, Ceratostomataceae, Coniochaetaceae and Lasiosphaeriaceae) (Hyde et al. 2020b) and even some Dothideomycetes (e.g., Botryosphaeriaceae and Venturiaceae) (Hongsanan et al. 2020). The germ slit or germ pore of the ascospores is a constant and diagnostic feature of certain xylarialean species (Beckett 1979). Generally, if the ascospore lacks a germ slit, it appears lighter (less melanised) than those that have it (Suwannasai 2005). This is due to the germ slit development process linked with the pigment deposition described in Beckett (1979). The germ slits are characterised by their shape, position on the spore, orientation along the long axis of the spore, and length. We observed straight to sigmoid, full ascospore length or less, concave or convex-sided germ slits from our collections during our study. Interestingly, we observed an equatorial and scattered distribution of germ pores in Paraxylaria as described above. The nature of the germ slit is distinct and likely to be a good character to differentiate between genera. The colour of the ascospores is generally a reflection of their habitat, as they need to be protected from UV light (Wong et al. 2019).
Generally, ascospores are brown to dark brown or hyaline, aseptate, septate or apiosporous and infrequently have a dwarf cell (Barr 1990). Ascospore appendages are important in spore dispersal (Jones and Moss 1978) and ascospores with distinct characters are likely to reflect distinct genera. More frequently, immature ascospores have cellular appendages, which can contribute to the characterization of species (Rogers 1979). Appendages in Xylariomycetidae, however, often occur and may not reflect generic status.

Paleomycological evidence of xylarialean taxa
Fossil remains of fungi and their taxonomic placement with extant taxa are useful for the reconstruction of past ecological conditions (van Geel and Aptroot 2006;Samarakoon et al. 2019a, b;Saxena et al. 2021). This is one of the challenges in the divergence time estimations and the ancestral reconstructions in xylarialean taxa, as there are very few taxonomically well-placed fossil specimens. Poinar (2014) described Xylaria antiqua from Tertiary Dominican amber (15-20 MYA), which has close affinity to extant X. allantoidea and X. grandis in having similar shaped ascospores and longitudinal germ slits. In addition, several other fossil ascospores have been identified, which are similar to extant xylarialean taxa. Saxena et al. (2021) reviewed fossil ascospores and accepted 55 Hypoxylonites species, which are similar to extant Hypoxylaceae and one Nigrospora (Apiosporaceae). van Geel and Aptroot (2006) described Amphisphaerella dispersella (Coniocessiaceae) fossil spores in the early Holocene deposits in the Netherlands with the characteristic equatorially arranged germ pores. Similarly, Palaeoamphisphaerella pirozynskii has been discovered from India aged to the Miocene, which is characterised by 8-10 equatorial germ pores similar to extant Paraxylaria. In addition, Saxena et al. (2021) accepted another two Palaeoamphisphaerella, namely P. keralensis and P. tankensis. However, it is important to research the integrated approach to identify taxonomically well-placed fossil xylarialean fungi for future studies (Samarakoon et al. 2019a, b).

Evolution and ancestral character reconstruction of Xylariomycetidae
Based on selected taxon sampling, Amphisphaeriales and Xylariales diverged around 150.5 (115-180) MYA within well-supported clades during the period of rapid divergence in the Early Mesozoic. The ancestral character analysis shows the earlier Xylariomycetidae to have evolved around 159 (124-193) MYA with ancestral ancestors with astromatic forms and hyaline, aseptate ascospores that lacked germ slits. Amphisphaeriales remained mostly as astromatic forms with aseptate and hyaline ascospores and are mostly known only as asexual morphs. Therefore, the unidentified morphologies are separately denoted in the analysis. In our ancestral character reconstruction based on seven stromatic characters, the immersed or semi-immersed inconspicuous ascomata with a prominent or rudimentary carbonaceous clypeus were an ancestral character of Xylariomycetidae. The divergence of ascomata types into different stromatic forms and ascospore modifications mainly occurred during the Cretaceous (66-145 MYA). The hypoxyloid (100-164 MYA), pseudostroma (95-156 MYA), bipartite and applanate (72-128 MYA), xylarioid (69-119 MYA) and rosellinoid (52.5-88.5 MYA) stromata types appear to have developed mostly during the Cretaceous. The pseudostromatic character found in some diatrypaceous evolved may have an intermediate development of the stromatic and astromatic development, which needs further characterizations to investigate unidentified sexual morphs. Brown ascospores are common in Xylariales, but first appeared in Amphisphaeriaceae, Melogrammataceae and Sporocadaceae during the early Cretaceous (94-155 MYA) (Fig. 2). The hyaline apiospores and filiform ascospores may have appeared as separate lineages basal to Xylariaceae. The ascospore germ slit appeared only in Xylariales during the Cretaceous (95-156 MYA) after the divergence of Lopadostomataceae.
The ancestral character state reconstruction plays an important role in a better understanding of morphological character evolution (Schmitt et al. 2009). The methodology has been used in various studies on fungi to show lifestyles and nutritional modes , morphological characterization of ascospores and appressoria (Chethana et al. 2021b), and geographical and host distributions (Píchová et al. 2018;Zhu et al. 2019). Even though the xylarialean taxa are highly diversified, no study of the ancestral character state reconstruction through molecular data has been performed. Rogers (2000) hypothesised that the aseptate ascospores with a germ slit, "truly xylariaceous" evolved from dark coloured one septate ascospores that lacked a germ slit, while passing several complex evolutionary stages. Considering the stromatic nature of Xylariaceae and Sordariaceae taxa, Rogers (1979) suggested that these two lineages may have evolved from astromatic ancestors, probably with bicellular ascospores. This is an important initiative to develop a hypothesis of the ancestor of Xylariomycetidae.
It has been theoreticised that fungi may have moved from aquatic to land environments in ancient times as symbionts of plants (Krings et al. 2012;Lutzoni et al. 2018). If this was the case, the early Ascomycota are likely to have been endophytes in symbiosis with plants and have subsequently evolved as saprobes and pathogens or maintained two or three of these lifestyles. The wide taxonomic distribution of viaphytes, i.e., fungi that undergo an interim stage as leaf endophytes and after leaf senescence, colonize other woody substrates via hyphal growth, suggests that viaphytic dispersal may be a deeply ancestral trait (U'Ren et al. 2016;Nelson et al. 2020). There is a broader host range for endophytic lifestyle compared to saprotrophs (Whalley 1996;U'Ren et al. 2016;Nelson et al. 2020). Xylarialean taxa can be found in living, asymptomatic leaves of angiosperms, gymnosperms, and bryophytes with close phylogenetic relationships (Davis et al. 2003;U'Ren et al. 2016). Interestingly, some of the taxa that appeared on freshly fallen branches or even on branches still attached to the parent tree (Whalley 1996), may have undergone rapid lifestyle changes. This shows that xylarialean taxa evolved as astromatic forms from endophytes and diversified into stromatic forms as adaptations to different conditions. The question then arises of how the endophytic species appear as saprobes. This might be a result of the endophytes emerging from spores from saprobes. Rodrigues et al. (1993) tested a high degree of genetic diversity of endophytic Xylaria cubensis, isolated from leaves of the Brazilian rainforest palm Euterpe oleracea. This is possibly due to the spore origin of endophytes. Ju et al. (2018) described the possibility of wind-borne xylariaceous fungal spores to establish endophytic associations. Furthermore, Ju et al. (2018) mentioned the efficiency of the modern molecular techniques in tracing propagules and infections rather than using tedious cultural procedures. However, studies on biochemical mechanisms and the transition of the saprobe to endophyte are poorly researched (Promputtha et al. 2007;Zhou et al. 2018). Rodriguez and Redman (2008) pointed out that both symbiosis and pathogens invade the host and remain dormant until favourable environmental factors. This is a vital strategy to exploit plant nutrients effectively (Davis et al. 2003). Some saprobic Ascomycota produce appressoria and that is evidence towards endophytic lifestyles and would also indicate that many saprobes may be host-specific (Chethana et al. 2021a, b). Oxydothis, the early colonizers of palms, and Linocarpon, which are very common on degrading palm fronds, have been shown to produce appressoria (Konta et al. , 2017. More endophytic taxa can be recovered from senescent leaves or decomposing leaves, wood, bark, fruits or flowers, and endophytism could be a stage to a complex lifecycle that can involve interactions with diverse host lineages (U'Ren et al. 2016). This provides evidence that endophytic fungi are the ancestral forms of saprobic nutritional modes. Xylarialean taxa however, show a wide range of host-substrate invasions of (1) living leaves and stems, often fruiting on living host material (e.g., Anthostomella), of (2) living stems as opportunists and rapidly and widely colonizing the host and fruiting (e.g., Biscogniauxia, Camillea, Daldinia and Hypoxylon), of (3) primarily as saprobes and then facultative parasites causing serious diseases (e.g., Dematophora necatrix, Kretzschmaria clavus and Xylaria), of (4) living as endophytes and often fruit on decayed material (e.g., Nemania and Xylaria), of (5) fruit on seeds and fruits with specific and discrete host-fungal relationships (e.g., Xylaria), of (6) invading dung (e.g., Hypocopra, Podosordaria and Poronia), of (7) associated with ant and termite nests, of (8) inhabiting litter and organic soils, and, of 9) damaging pathogens (e.g., Entoleuca and Rosellinia) (Rogers 2000). Voglmayr et al. (2019) emphasised the endophytic lifestyle as the primary stage of Leptosillia lined with ascomatal development as saprobes, while pathogenicity as the secondary evolved character.
Even though the exact ancestral state of xylarialean taxa is unclear, the stromatic nature of the taxa might have evolved for successful parasitism and saprotrophism in dry sites (Rogers 1979). Thus, the early xylarialeans were probably endophytes that formed simple anthostomella-like ascomata with small clypei on the host surface and later evolved into other stromatic forms. The rapid morphological diversifications resulted in the Cretaceous radiation of angiosperms (Rogers 2000;Phillips et al. 2019) and probably with multiple independent diversifications. Rogers (1979) suggested that Collodiscula was a primitive xylarialean taxon due to its two-celled ascospores, and in having Astrocystis as the nearest relative (Ju and Rogers 1990). Septate to aseptate, ascospores may have evolved, which is evident from the cellular appendages in some immature ascospores in xylarialean taxa. However, an independent evolution from aseptate to septate ascospores may have occurred due to rapid diversification. Another example is the long, filiform, aseptate ascospores of Linosporopsis. Voglmayr and Beenken (2020) mentioned that Linosporopsis are saprobes and endophytes, but not pathogens, and have unusual ascospore morphologies borne in asci in astromatic, clypeate ascomata. Furthermore, Voglmayr and Beenken (2020) proposed that it is not surprising that independently evolved leaf-inhabiting species of various ascomycete lineages occurred.
Much work needs to be carried out on Xylariomycetidae as the classification is only now being resolved Daranagama et al. 2018;Wendt et al. 2018;Konta et al. 2021). If, as theoreticised above, the stromatic Xylariomycetidae evolved from astromatic ancestors, it would be interesting to research the chemical profiles of a range of genera. The stromatic form is known to produce an array of chemical compounds that are probably useful in deterring insect predation (Becker and Stadler 2021), but what about the ancestral clypeate forms? Was there an evolution of chemicals produced in the astromatic forms as composed of those with stroma? The former would have been protected by the host plant and developed simple clypei for protection from UV light. Even though, as early predictions, the stromata development may have been related to moisture conservation, the stromatic forms would have needed insecticidal chemicals to prevent insect and other pest predation.
In summary, astromatic, clypeate ascomata with aseptate, hyaline ascospores lacking a germ slit may probably be the ancestral Xylariomycetidae, which evolved as the result of plant fungal endophytic associations. The Cretaceous period, with its rapid diversification of angiosperms, may have affected the diversification of xylarialean fungi with several independent lineages.

Conclusion
Before molecular data became available on a large scale, most of the xylarialean taxa were classified based on their stromatic nature, e.g., anthostomelloid (with indistinct stromata), hypoxyloid (with pulvinate atromata), rosellinioid (stroma with a single ascomata) and xylarioid (with welldeveloped stromata). Other forms had distinct characters, e.g., Astrocystis (star-shaped, split ectostroma) and diatrypoid taxa were placed in Diatrypaceae. With the molecular data, the criteria for taxonomic classification changed from previous schemes. It revealed that we could not use a stalk-like or well-developed stromata to place all taxa in Xylaria. Xylaria is polyphyletic and contains many distinct genera, and the well-developed stromata has evolved on multiple occasions. Therefore, we should not be conservative and should introduce higher ranks for xylarioid taxa (see Konta et al. 2020b;Maharachchikumbura et al. 2021). Similarly, taxa with poorly developed stromata with clypei or even lacking external structures should not be lumped into Anthotomella. Recent studies using molecular data have shown poorly developed stromata to be polyphyletic and found across Xylariales. The clypeus or immersed ascomata, lacking a well-developed stroma, is a primitive character and cannot be used to delimit genera without a combination of other characters, including molecular data. Furthermore, stromatic variations will not be a key character to introduce higher ranks, but the type of ring, the colour of the ascospores, and the presence or absence of the type of germ slit will be important in xylarialean taxa.