Resurrection and emendation of the Hypoxylaceae, recognised from a multigene phylogeny of the Xylariales

A multigene phylogeny was constructed, including a significant number of representative species of the main lineages in the Xylariaceae and four DNA loci the internal transcribed spacer region (ITS), the large subunit (LSU) of the nuclear rDNA, the second largest subunit of the RNA polymerase II (RPB2), and beta-tubulin (TUB2). Specimens were selected based on more than a decade of intensive morphological and chemotaxonomic work, and cautious taxon sampling was performed to cover the major lineages of the Xylariaceae; however, with emphasis on hypoxyloid species. The comprehensive phylogenetic analysis revealed a clear-cut segregation of the Xylariaceae into several major clades, which was well in accordance with previously established morphological and chemotaxonomic concepts. One of these clades contained Annulohypoxylon, Hypoxylon, Daldinia, and other related genera that have stromatal pigments and a nodulisporium-like anamorph. They are accommodated in the family Hypoxylaceae, which is resurrected and emended. Representatives of genera with a nodulisporium-like anamorph and bipartite stromata, lacking stromatal pigments (i.e. Biscogniauxia, Camillea, and Obolarina) appeared in a clade basal to the xylarioid taxa. As they clustered with Graphostroma platystomum, they are accommodated in the Graphostromataceae. The new genus Jackrogersella with J. multiformis as type species is segregated from Annulohypoxylon. The genus Pyrenopolyporus is resurrected for Hypoxylon polyporus and allied species. The genus Daldinia and its allies Entonaema, Rhopalostroma, Ruwenzoria, and Thamnomyces appeared in two separate subclades, which may warrant further splitting of Daldinia in the future, and even Hypoxylon was divided in several clades. However, more species of these genera need to be studied before a conclusive taxonomic rearrangement can be envisaged. Epitypes were designated for several important species in which living cultures and molecular data are available, in order to stabilise the taxonomy of the Xylariales.


Introduction
The Xylariaceae comprises one of the core families of the Sordariomycetes and is the largest family of the Xylariomycetideae that are known as cosmopolitan, ubiquitous wood degraders and predominant endophytes (Stadler 2011). The family is located within the Xylarialesthe only order of the Xylariomycetidae (Eriksson and Winka 1997; see Senanayake et al. 2015 andJaklitsch et al. 2016 for the current organisation of this order). It has always been recognised to comprise a homogenous evolutionary lineage based on its ascal and ascospore morphology (typically having eightspored asci with an amyloid ascal apical apparatus and brown to dark brown, ellipsoid ascospores; cf. Rogers 1979).
The prototype of a xylariaceous fungus is Xylaria hypoxylon (L.) Grev., the type species of the genus Xylaria Hill ex Schrank, which is probably the largest genus within the family, according to the current classification. The taxonomy of this fungus has only recently been stabilised by a lecto-and epitypification (Stadler et al. 2014b), following an extensive polyphasic study by Fournier et al. (2011).
Related families in the Xylariales/Xylariomycetideae such as the Diatrypaceae and Graphostromataceae have historically been distinguished from the Xylariaceae by having allantoid ascospores (Barr et al. 1993;Stadler et al. 2013). However, since about 50 years ago, the morphological characters of the asexual morph have been taken into account more seriously in the taxonomy of the Ascomycota. This has resulted in the restriction of the genus Hypoxylon Bull. for those taxa that have nodulisporium-like asexual stages. Genera such as Entoleuca Syd. and Nemania S. F. Gray were segregated from Hypoxylon as they produce geniculosporium-like asexual morphs Ju and Rogers 2002).
An attempt to classify the larger groups in this highly diverse group of Ascomycota by Dennis (1961) into "Xylarioideae", "Hypoxyloideae", and "Thamnomycetideae" proved invalid because these "subfamilies" have not been formally erected following the rules of botanical nomenclature (cf. Bitzer et al. 2008). Moreover, the "Thamnomycetideae", which Dennis (1961) had mainly based on the aberrant stromatal morphology of Thamnomyces Ehrenb., was recently proven obsolete. It was clearly shown that the phylogenetic and chemotaxonomic affinities of typical Thamnomyces are with Daldinia Ces. & De Not. and other "hypoxyloid" genera (Stadler et al. 2010b).
The first comprehensive classification of the Xylariaceae based on holomorphic concept was proposed by Ju and Rogers (1996), who divided the Xylariaceae into two major groups: the genera related to Xylaria with "geniculosporiumlike anamorphs" vs. the genera related to Hypoxylon, featuring "nodulisporium-like anamorphs". Ju and Rogers (1996) did not actually attempt formally to circumscribe these major groups by erecting subfamiliar taxa, but only listed the respective genera. Some authors have nevertheless used the taxonomically invalid terms "Xylarioideae" (for the former) and "Hypoxyloideae" (for the latter). Interestingly, these differences in anamorphic morphology also coincide with the presence of other features (e.g. inverse hat-shaped apical apparatus in many genera of the "Xylarioideae"; presence of stromatal pigments in Hypoxylon and allies). However, the immense morphological and phenotypic plasticity of the Xylariales has so far precluded a clear-cut segregation at the subfamiliar ranks, and even the definition of Hypoxylon by Ju and Rogers (1996) was still mainly based on gross stromatal morphology. Interfaces to other xylarialean families were also occasionally observed (cf. treatment of Creosphaeria, Lopadostoma, Jumillera and Whalleya below), but never taken seriously into consideration to remove any of the respective genera from the Xylariaceae, because the ascospore morphology was still regarded as a major discriminative character at this taxonomic level.
Starting from the work of Whalley and Edwards (1995), mycologists have begun to appreciate the value of chemotaxonomic data as complementary phenotype-based characters, which can be used to verify taxonomic concepts in the Xylariaceae. The latter authors mainly studied mycelial cultures of these fungi and identified their characteristic secondary metabolites after tedious chromatography work and extensive characterisation of the purified products, but did not have any high-throughput analytical methods at their disposal to screen numerous specimens for their occurrence to provide sound data on the distribution of the compounds in the family. While their work did not immediately result in taxonomic conclusions, this was later made possible by the advent of analytical techniques based on high performance liquid chromatography coupled with diode array detection and mass spectrometric detection (HPLC-DAD/MS). This technique, paired with extensive preparative chromatography work, has led to the discovery of over 150 secondary metabolites, most of which were new to science at the time of their discovery (Stadler and Hellwig 2005;Stadler 2011;Kuhnert et al. 2015). Recent studies have shown that in particular the metabolite profiles from cultures are strongly in agreement with molecular phylogenetic data (Bitzer et al. 2008;Stadler et al. 2010a, b, c), whereas the stromatal HPLC profiles proved highly valuable for delimitation of species in those genera that have so far been extensively studied (Stadler et al. 2008a, b, 2014a, Kuhnert et al. 2016. Comprehensive molecular phylogenetic investigations of the Xylariaceae using data from reliably identified specimens have been rather scarce until very recently. The studies by Sánchez-Ballesteros et al. (2000), Triebel et al. (2005) and Pelaez et al. (2008) exclusively used DNA loci such as the internal transcribed spacer region (ITS) of the nuclear ribosomal DNA (rDNA) that is widely accepted by the community as standard barcode of fungi (Schoch et al. 2012), but failed to resolve the taxa in the hypoxyloid clades, due to its limitations as phylogenetically informative locus. The studies by Hsieh et al. (2005, with focus on Hypoxylon and allies), and Hsieh et al. (2010, with focus on the xylarioid clades) provided a better resolution based on the protein-coding genes alpha-actin (ACT) and beta-tubulin (TUB2). Both genes exhibit intron-rich 5′ primed ends and thus met ideal requirements for phylogenetically informative loci, but areas non-universal regionsonly of limited use for phylogenetic analyes (Tang et al. 2007;Stielow et al. 2015). Interestingly, Hsieh et al. (2005) already chose to abandon ITS data from their great experience with comparative morphology as phylogenetically valid criterion at a time when many other mycologists promoted them, and some even proposed that this DNA locus can serve as "sole means of identification of a fungus". The development in fungal taxonomy and phylogeny that followed in the past decade certainly proved the latter authors right! Tang et al. (2009) were the first to attempt a multigene phylogeny derived from a combination of rRNA and protein-coding genes. Besides ITS, LSU, and TUB2, Tang et al. used the section "6-7" of the second largest subunit of the RNA-polymerase II gene (RPB2) that is commonly used for both barcoding and phylogenetic purposes within the fungal kingdom. Unfortunately, their phylogeny had many gaps and several data sets were derived from different specimens of the same taxon or came from non-verifiable sources (e.g. sequence data in GenBank retrieved from material that is not available in the public domain for verification of the taxonomy). Recent work by Daranagama et al. (2014Daranagama et al. ( , 2015 provided additional rDNA and RPB2 sequence data, in particular for important Anthostomella species and other related taxa. Because of these studies, and comprehensive treatments of the Xylariaceae in phylogenies obtained exclusively from housekeeping gene sequences by Hsieh et al. (2005Hsieh et al. ( , 2010, the data matrix for a multigene genealogy has steadily increased over the past decade. Notably, several of the genera that were regarded as xylariaceous by Ju and Rogers (1996) and earlier workers are characterised by libertella-like asexual stages (cf. Ju et al. 1993 for Creosphaeria; Rogers et al. 1997 for Jumillera and Whalleya), revealing affinities to the Diatrypaceae. Jaklitsch et al. (2014) provided a comprehensive polythetic study on the genus Lopadostoma based on rDNA and RPB2 data, revealing that these fungi appear rather distantly related to the core of Xylariaceae. Accordingly, Senanayake et al. (2015) placed Lopadostoma and Creosphaeria in the new family Lopadostomataceae, which was confirmed in a subsequent study by Jaklitsch et al. (2016). Concurrently and subsequently, several other studies have been initiated in order to classify the non-stromatic taxa of the Xylariales (Jaklitsch and Voglmayr 2012;Daranagama et al. 2015Daranagama et al. , 2016Maharachchikumbura et al. 2016), sometimes resulting in incongruent concepts according to the standpoint of the authors.
In particular the study of Jaklitsch et al. (2016) provided several examples that undermined the classical concepts of ascomycete taxonomy, because they proved conclusively that it is not feasible to rely on old herbarium specimens and morphological descriptions, unless the respective species have been recollected, cultured, studied in detail for their morphological characters, and sequenced by means of a multi-locus approach. The other above mentioned studies, however, heavily relied on collection of fresh material during field work in tropical areas and many non-stromatic Xylariales with affinities to the Xylariaceae in the classical sense were recovered and sequenced, but they could not be assigned to one of the existing teleomorph-based genera.
As a consequence, the number of genera and even families has been steadily increasing, while the core taxa of the Xylariaceae, i.e. arguably the largest and most important family, remain to be re-organised.
The present paper was dedicated to a multigene genealogy of the family, with emphasis on Hypoxylon and its allies and their affinities to the xylarioid clades. We selected numerous representative specimens and cultures from extensive taxonomic work in the past decade and hope to fill some gaps in our knowledge on the phylogenetic affinities of this family. We concentrated on the phylogenetic and chemotaxonomic aspects, because a concurrent comprehensive study giving an account on all xylariaceous genera is presently under preparation.

Materials and methods
Except for the typification details in the taxonomic part, all fungal names are given without authorities or publication details, in accordance with Index Fungorum (http://www. indexfungorum.org) and Mycobank (http://www.mycobank. org/).
With respect to the terminology of sexual vs. asexual structures, we have explicitly decided to treat the "old-fashioned" expressions (anamorph/teleomorph) as synonyms of the more modern terms asexual vs. sexual morph. The reason is that, particularly in the Xylariaceae and related taxa, the conidial states of most of the species are not yet known, or it has not been possible to obtain them in culture. Therefore, the characteristics of the teleomorph are thus by far more informative.

Taxon selection
All taxa were selected from previous work as indicated in Table 1, which also contains all references from which sequence data were derived. Representatives of Diatrypaceae (Diatrype disciformis), Lopadostomataceae (Creosphaeria sassafras), and the anthostomella-like Pyriformiascoma trilobatum were selected as outgroup. The type strain of Strain numbers of public culture collections, origin GenBank accession numbers and references are given. Type specimens are labeled with HT (holotype), ET (epitype), and PT (paratype). Unless references are given, the sequence data were generated in the present study. N/A: not available Calceomyces lacunosus was added, because it was reported to have a nodulisporium-like asexual morph and shoe-shaped ascospores and its phylogenetic affinities remained to be settled. The bulk of the taxa was selected from the large stromatic genera that comprise the Xylariaceae in the traditional sense, and for which numerous data derived from well-characterised specimens are available. The selection criteria included both morphological and chemotaxonomic characters. Type material or at least specimens representing the type species of the respective genera and some important other species were included whenever possible.

Molecular phylogenetic analyses
DNA was extracted from mycelial cultures using protocols and genomic DNA extraction kits described in Laessøe et al. (2010), Kuhnert et al. (2014a), and Otto et al. (2016). Partial sequences of four DNA loci were used as phylogenetic markers, namely a) internal transcribed spacer region (ITS); b) 28S large subunit (LSU) of ribosomal DNA; c) second largest subunit of the DNA-directed RNA polymerase II (RPB2); and d) beta-tubulin (TUB2) were generated using standard primers introduced by White et al. (1990;ITS4), Gardes and Bruns (1993;ITS1F), O'Donnell (1993;NL4), Vilgalys and Hester (1990;LR07 and LR0R), Liu et al. (1999;RPB2-5F, -7cF, and -7cR), and O'Donnell and Cigelnik (1997; T1, T2, T11, T21,  T22, T121, T222, and T224), Glass and Donaldson (1995;Bt2b), and protocols as described by Laessøe et al. (2010) and Otto et al. (2016). PCR products were also purified and sequenced as described in Laessøe et al. (2010) and Otto et al. (2016). Sequences were analyzed and processed in Geneious® 7.1.9 (http://www.geneious.com, Kearse et al. 2012). Raw sequence files were trimmed automatically to exclude flanking regions and bases with error probabilities above 1% and double-stranded sequences were compiled. The generated sequence data were complemented by available sequence data from GenBank and the data sets for each genetic marker were aligned using MAFFT 7.017 with G-INS-I as algorithm and default settings for gap open and gap extension penalties (Katoh and Kuma 2002). The constructed alignments were processed via the Castresana Lab Gblocks Server at low stringency settings (allowing smaller final blocks and gap positions within the final blocks; see Talavera and Castresana 2007;Castresana 2000) and the multigene alignment (MGA) was created by concatenation of all genetic markers in Geneious. The best fitting substitution model for each single gene alignment and the MGA was determined using jModeltest 2.1.7 (Darriba et al. 2012;Guindon and Gascuel 2003). Phylogenetic relationships were inferred using neighbor-Joining with Tamura-Nei as distance model (NJ, Saitou and Nei 1987) in Geneious Tree Builder, Maximum Parsimony (MP), and tree bisection and reconnection (TBR) as branch swapping algorithm in PAUP 4. 0a151 (Swofford 2002), maximum likelihood (ML) with GTR + G + I as substitution model in RAxML 7.2.8 (Stamatakis 2006(Stamatakis , 2014Stamatakis and Alachiotis 2010). Bootstrap support (BS) values in NJ, MP, and ML analysis were calculated from 500 replicates. Prior to MP analysis, the consistency index (CI), retention index (RI), and homoplasy index (HI) were calculated.

Molecular phylogeny
One-hundred and twelve fungal strains of 34 different genera were used in the analysis, of which 77 contained at least one sequence of the considered DNA loci. The five most species-rich genera were Hypoxylon with 29, Daldinia with 14, Xylaria with 10, Annulohypoxylon with seven, and Nemania with six included species. In total, 224 sequences were generated (36 ITS, 71 LSU, 75 RPB2, and 42 TUB2 sequences) and complemented with 206 sequences from GenBank. We used four different molecular markers and applied four tree reconstruction methods, which enabled us to make robust assumptions about the phylogenetic relationships of the studied taxa, but also led to substantial output data. Additional data files, e.g. the unprocessed MAFFT alignments, excluded data blocks, estimated model parameters or additional tree files have been provided in the Supplementary Information.
The calculated MAFFT alignments consisted of 1623 character positions in the ITS alignment, 2456 in the LSU alignment, 2335 in the TUB2 alignment, and 1325 in the RPB2 alignment. After curingwhich is required to exclude poorly aligned areas and positions of low informative characters within the alignmentsthe constructed multigene alignment (MGA) consisted of 2911 characters (314 of which were derived from the ITS alignment, 794 from LSU, 821 from RPB2, and 982 from TUB2). Of the MGA, 1082 characters were considered parsimony-informative.
The results of all phylogenetic tree constructions were superimposed on the Bayesian tree and support values that pose statistical support [≥0.95 for posterior probability (PP) and ≥50% bootstrap support (BS)] were indicated above the respective branches. However, significant support was assumed for BS values ≥70% and/ or PP values of ≥0.98.
R e p r es e n t a t i ve s o f t h e t y p e s p ec i es of t h e Diatrypaceae (Diatrype disciformis) and the most commonly reported species of the Lopadostomataceae (Creosphaeria sassafras), were initially designated as outgroup, because they are supposed to be close relatives of Xylariaceae in the classical sense. However, our analyses revealed that two further "xylariaceous" representatives, namely Calceomyces lacunosus and Pyriformiascoma trilobatum also clustered with this outgroup. Separation of the extended outgroup clade to the remaining taxa was fully supported in all phylogenetic trees.
Three major clades were observed: the first of those, clade (I) comprised Xylaria and all other genera studied with geniculosporium-like anamorphs (Fig. 1). Accordingly, clade I is from here on regarded to represent the Xylariaceae s. str. In the second major clade (II), Graphostroma platystomum clustered with Biscogniauxia, Camillea, and Obolarina as sister group to clade I, while Hypoxylon and all other genera with nodulisporium-like anamorphs appeared in a welldefined separate clade (III). This finding led us to abandon the conventional terminology concerning the classification of Xylariaceae. We will from here on use the newly introduced taxonomic terms that are explained in the taxonomic part further below.
Within clade I (Fig. 1), monophyletic clades of the coprophilous genera Podosordaria (PS) and Poronia (P) clustered with Sarcoxylon compunctum as sister to the remaining Xylariaceae s. str. clades. These included two Xylaria clades (X1 and X2), with X2 comprising Kretzschmaria deusta and Brunneiperidium gracilentum along with several other Xylaria spp., including the type species of the genus, Xylaria hypoxylon. A third clade (X3) was not supported and contained Xylaria polymorpha and further representatives of Xylaria and of the genera Amphirosellinia, Astrocystis, Collodiscula, and Stilbohypoxlon. The last clade of Xylariaceae s. str. split into clade N that comprised representatives of the genus Nemania and Euepixylon, and clade R with species of the genus Rosellinia and Entoleuca mammata. The latter clustered with non-pathogenic Rosellinia species that appeared in a sister clade to the two included plant-pathogens R. buxi and R. necatrix.
Within clade G, Graphostroma platystomum clustered in a well-supported subclade G1 with several species of Biscogniauxia. The second subclade G2 contained different species of Biscogniauxia (including the type species, B. nummularia), Camillea and Obolarina species.
Most representatives chosen for the current study were selected from the taxa with nodulisporium-like anamorphs and accordingly, clustered in Clade III. The topology of these clades was not consistently supported, but the backbone topology was characterised by distinct and strongly supported clades (Fig. 2). The genus Hypoxylon was divided into several subclades, sometimes only represented by one or two species, indicating that the taxon sampling may still be too low. Hypoxylon papillatum was the basalmost taxon of Clade III, followed by five well-supported clades comprising species of the genus Hypoxylon (H1-H5). H5 was a sister to a clade of the genera Jackrogersella (J) and Annulohypoxylon (A). And a sixth Hypoxylon clade (H6) formed the most derived monophyletic group of Hypoxylon as sister taxon to a clade comprising Pyrenopolyporus (Py) and Daldinia (D1-D2).
The separation of H. papillatum and the nodes of the basal sister clades H1&H2 were fully supported by the calculated support values of all phylogenetic tree algorithms. The clade H1 contained the frequent species H. fuscum, along with H. porphyreum and H. vogesiacum and was strongly supported, whereas clade H2 consisted of two well-supported subclades and an unsupported subclade. Hypoxylon hypomiltum and H. samuelsii formed one of the supported subclades and the second supported subclade consisted of H. rubiginosuma common species of the Northern temperate hemisphereand several other taxa that had previously been synonymised with this species, including H. cercidicola and H. petriniae. The unsupported H2 subclade comprised H. carneum, H. ochraceum, H. perforatum, H. musceum, and H. pilgerianum. The H3 clade was fully supported by NJ, Bayesian and RAxML analysis, but the node of H. crocopeplum and H. fendleri was not significantly supported, as only the NJ-BS value of 79% and a posterior probability value of 0.98 met the demanded threshold for statistical support. Nonetheless, the shown topology was predominantly observed in the phylogenetic analyses.
H4 comprised the generic type species, H. fragiforme, along with H. haematostroma, H. rickii, H. howeanum, and H. ticinense, and although the positions of H. haematostroma and H. rickii were not supported statistically, the clade was fully supported by NJ, MP, ML, and Bayesian analysis. The remaining clades split into two sister clades; one of those consisted of a multifurcated and poorly supported clade of H. griseobrunneum, H trugodes, and a subclade comprising the genera Hypoxylon (H5), Jackrogersella (J), and Annulohypoxylon (A); the other clade comprised a fully supp ort ed cl ad e o f t he " ba sal " H y p o x y l o n speci es H. monticulosum and H. submonticulosum (H6), a monophyletic clade comprising species of the resurrected (see taxonomic part) genus Pyrenopolyporus (Py) and a clade comprising daldinoid species and closely related genera (D1-D2).
Clade H5 consisted of the species H. lateripigmentum, H. investiens, and H. pulicicidum and was mostly well supported and was shown as a sister taxon to a clade comprising the genera Jackrogersella, Rostrohypoxylon and Annulohypoxylon. Although this topology was only supported by Bayesian and ML analysis, the subdivision of the latter genera into two clades (J&A)with Rostrohypoxylon terebratum in an intermediate position was strongly supported.
The fully supported genus Pyrenopolyporus (Py) was represented by the species P. laminosus, P. hunteri, and P. nicaraguensis. The sister taxon to clade Py split into three, mostly strong supported clades: clade D1 comprised the sister clades of D. andina, D. concentrica, D. dennisii, D. loculatoides, D. macaronesica, D. petriniae, D. pyrenaica, D. steglichii, and D. vernicosa (which are either temperate/ Fig. 1 Phylogeny of the Graphostromataceae and Xylariaceae sensu stricto. The phylogenetic relationships are depicted as Bayesian tree, inferred from a multigene alignment of ribosomal (ITS and LSU) and proteinogenic (TUB2 and RPB2) sequence information. In the Maximum Parsimony analysis a CI of 0.161, a RI of 0.524, and a HI of 0.839 was estimated, and yielded five most parsimonious trees with a length of 15,094 changes. The phylogenetic tree inferred by RAxML had a likelihood of −65,444.25 and the likelihood of the Bayesian tree was −65,951.87. Support values were calculated via neighbor-joining (NJ), maximum parsimony (MP), Bayesian (B), and maximum likelihood (MA) analysis and are indicated above (NJ/MP) and below (B/ML) the respective branches, if the bootstrap support (BS) values exceeded 50% or the posterior probability (PP) value was 0.95 or higher. Branches of significant support (BS ≥ 70% and PP ≥ 0.98) are thickened; accessions representing type material are highlighted in bold letters subtropical taxa or like D. andina, have been found in the tropics, but at very high altitudes). The second Daldinia clade D2 embodied two fully supported subclades of Rhopalostroma angolense and Thamnomyces dendroidea on one branch, and the predominantly tropical D. bambusicola, Daldinia caldariorum, D. eschscholtzii, D. theissenii, and D. placentiformis on the other. The third clade of Entonaema liquescens and Ruwenzoria pseudoannulata showed full statistical support, but the relative position of this clade to the Daldinia clades D1 and D2 was not supported sufficiently.
These data led us to propose significant changes in the taxonomy of the Xylariales. Below we propose a new scheme that takes the current molecular phylogenetic data into account and is congruent with the morphological concepts at the same time. As will become evident in the following taxonomic part, the reorganisation of the taxonomy was often found to be congruent with the morphology of the conidial states, which are compiled based on representative taxa in Figs. 3 and 4.  Epitypes of two further important species in Hypoxylon are here designated, based on the specimens that were used in the current phylogeny, in order to stabilise the taxonomy of the genus. Especially, H. rubiginosum has been misinterpreted very often in the course of the taxonomic history and was often regarded as a cosmopolitan species, while there is no record of this taxon for the tropics to the best of our knowledge. The specimen selected as epitype showed the same morphology as the type material, and it was even possible to detect their characteristic stromatal metabolites by comparison of the HPLC profiles of the type specimen. These data, as well as other results on the synonymy and geographic distribution of both taxa in the temperate Northern hemisphere, were compiled by Stadler et al. (2008b).

General description of the Hypoxylaceae
Saprobic on plant material, but many species are endophytes and some species are associated with insect vectors (Pažoutová et al. 2010a(Pažoutová et al. , b, 2013. Stromata (if present) erect, glomerate, pulvinate, discoid, effused-pulvinate, hemispherical, spherical or peltate; solitary or confluent, with broad or narrow attachment to the substrate; surface colored or black, pruinose or polished, planar or with perithecial mounds; waxy or carbonaceous tissue immediately beneath surface and between perithecia, with or without KOH-extractable pigments; the tissue below the perithecial layer inconspicuous, conspicuous, or massive, most often dark brown to black, persistent or loculate; interior sometimes conspicuously zonate (Daldinia) or hollow and filled with liquid (Entonaema).
Ostioles lower than the level of stromatal surface (umbilicate), at the same level of stromatal surface, or higher than the level (papillate) of stromatal surface, with or without discs formed by dehiscence of surrounding tissue.
Ascospores brown, unicellular in both mature and immature ascospores, ellipsoid or shorty fusoid, inequilateral, slightly inequilateral or nearly equilateral, with acute, narrowly rounded, or broadly rounded ends, in most species bearing a germ slit; perispore dehiscent or indehiscent in 10% KOH.
Conidiophores mononematous or infrequently synnematous, usually macronematous, hyaline or colored, smooth or roughened, simple or with a dominant main axis that is unbranched or bears one or more major branches.
Conidiogenous cells cylindrical, usually hyaline, one to several on each terminus of conidiophore, with conidiogenous regions at apex that are swollen to various degrees due to successive conidial production. Conidiogenous regions with poroid or, infrequently, denticulate conidial secession scars. Conidia produced sympodially in more or less basipetal succession, subglobose to ellipsoid, usually hyaline, with flattened base indicating former point of attachment to conidiogenous cells.
Notes: This ancient family name was coined originally by de Candolle (cf. de Lamarck and de Candolle 1805) in the sense of an order ("ordo") with the type genus Hypoxylon Bull. and its type species Hypoxylon coccineum (≡ Hypoxylon fragiforme) and was only occasionally used by other taxonomists. Under the previous taxonomic concepts, the Hypoxylaceae was, however, never generally accepted among mycologists as a separate family besides the morphologically similar Xylariaceae, but generally regarded as a synonym of the latter family.
Our above definition takes into account the few salient features by which the Hypoxylaceae differ from the Xylariaceae s. str., and the circumscription also includes the "special" features of the genera that are aberrant with respect to their ascal and stromatal morphology. Notably, hollow stromata filled with liquid also occur at least in one member of the Xylariaceae s. str., namely Xylaria mesenterica (Möller) M. Stadler, Laessøe & J. Fourn. 2008, originally reported by Möller (1901) as an Entonaema (cf. Stadler et al. 2008a). However, all genera formerly assigned to the Xylariaceae with cleistocarpous asci like Phylacia, Rhopalostroma and Thamnomyces belong to the Hypoxylaceae in the current sense. Chlorostroma is accommodated in the Hypoxylaceae due to its similarity to Hypoxylon aeruginosum and the highly similar secondary metabolite profiles (Laessøe et al. 2010).
The assignment of several genera formerly included in the broad concept of the Xylariaceae s. lat. to either of the families follows morphological descriptions, since no molecular phylogenetic data have so far been generated for their type species, and no living ex-type cultures seem to be available. The genera Jumillera and Whalleya (Rogers et al. 1997) are accommodated in the Lopadostomataceae at interim, based on the morphology of their conidial states (which is also supported for Whalleya, based on the comparison of molecular data from the literature). For some other genera of non-stromatic and anamorphic Xylariales, the assignment of the family was made based on DNA-based data, where those have been available. Several genera that are only known from morphological descriptions of the conidial states were accordingly placed in Xylariales incertae sedis. Substantial field work, in particular in tropical countries, remains to be accomplished to settle their phylogenetic affinities. The morphological concept that coincides with the organisation of families in the studied fungi is summarised in Table 2.
Other accepted genera of the Hypoxylaceae  ). e, f, g, h: geniculosporiumlike anamorph (characteristic of Xylariaceae s. str. and synanamorph of some Lopadostomataceae) e: conidiophore f: geniculate conidiogenous cells and conidia (from Nemania plumbea, EKKRF 1401) g: conidiogenous cell becoming geniculate after producing multiple conidia (from Rosellinia sp., Sir & Hladki 282) h: palisadic geniculosporium-like anamorph (from Stilbohypoxylon macrosporum, Sir & Hladki 972; i, k: xylocladium-like anamorph, characteristic of Graphostromataceae. i: conidiophore k: detail of ampullae with conidiogenous cells (from Camillea sp., Sir & Hladki 856) j, l: nodulisporium-like anamorph with periconiella-like branching patterns, characteristic of Graphostromataceae and some Hypoxylaceae j: conidiophore l: conidiogenous cells (from Biscogniauxia sp.; Sir & Hladki 187). m, n, o, p, q, r, s, t: "regular" nodulisporium anamorph with different branching patterns as defined by Ju and Rogers (1996) Note: This genus is transferred to the Hypoxylaceae because the scarce molecular data available (Miller et al. 2007) point to its affinities with Hypoxlon and the conidial state is said to be nodulisporium-like. The ITS and TUB2 sequences we obtained of the strain studied by Miller et al. (2007) that is deposited with CBS 119807 are 100% identical to those of Hypoxylon rubiginosum, which is why we have excluded it from our phylogeny and think we need to re-collect and sequence this taxon to get sure about its affinities. Type species: Xylaria hypoxylon (L.) Grev. 1824.
All remaining genera of the Xylariaceae according to the current concept are listed below. Unless a note is provided, the N o t e : T h i s g e n u s a p p e a r s v e r y c l o s e t o Kretzschmaria and certain species of Xylaria in the molecular phylogeny provided by Daranagama et al. (2015).
Note: The molecular data provided by Jaklitsch et al. (2016), where the taxonomic history of this genus is discussed in detail, point toward its being closely related to Anthostomella and the Xylariaceae in the current sense.  Note: A geniculosporium-like anamorph was reported in the protologue, but no molecular data on the genus are available as yet.
Type species: Hypocopra merdaria (Fr.) J. Kickx f. (1867). Notes: Recently, the first DNA sequences of a member of this genus have become available in the course of a study on the chemistry of Hypocopra rostrata (Jayanetti et al. 2014). The ITS sequence arising from this study (GenBank Acc. No. KM067909) was subjected to a homology search and the results proved beyond doubt that this species has close affinities to Xylaria and Podosordaria.
Hypocreodendron Henn., Hedwigia 36 (4): 223 (1897) Note: A geniculosporium-like anamorph was reported in the protologue, but no molecular data on the genus are available as yet. Note: A geniculosporium-like anamorph was reported in the protologue, but no molecular data on the genus are available as yet. Note: The phylogenetic affinities of this taxon are apparently with Nemania and other Xylariaceae in the current, restricted sense. It has brown ellipsoid ascospores but differs from the typical Xylariaceae in having a libertella-like conidial state (albeit with relatively short conidia) and an atypical crescent shaped apical apparatus (cf. Daranagama et al. 2015 Petrini (2013) suggests that it is still rather heterogeneous and may in the future be further subdivided. The type species, R. aquila, as well as various other taxa that have so far been studied for their anamorph morphology or by using molecular data, however, all appear to be members of the Xylariaceae s. str. as understood here.
Note: This genus is here proven to be xylarioid for the first time by using molecular phylogenetic data, confirming the suspicions by Rogers (1981) from his detailed morphological studies.
Note: Rogers (1981) described the specimen, originally collected from New Caledonia, which should be regarded as the isotype in K, and the holotype is housed in FH (D. Pfister and G. Tocci pers. comm.). There is evidence that the type species has recently been recovered from both China and Thailand. We are aware of an ongoing study on its pyhlogenetic affinities and will refrain from giving details of this unpublished work. However, from the morphological examination reported by Rogers (1981) there can be no doubt that the affinities of Squamotubera are with Xylaria.
Stilbohypoxylon Henn., Hedwigia 41: 16 (1902). Type species: Stilbohypoxylon moelleri Henn. 1902. Note: The anamorph of this genus is very similar to Xylaria, and Stilbohypoxylon is therefore retained in the Xylariaceae where it had been placed tentatively by Hennings, and which is also supported by molecular data. Type species: Wawelia regia Namysł. 1908. Note: The anamorph of this genus can be categorised as geniculosporium-like, and it is therefore retained in the Xylariaceae, even though no molecular data are available as now.
Genera excluded from the Xylariaceae in the current sense Emended generic description (modified from Barr et al. 1993).
Stromata on wood of living or dead angiosperm plants, effuse, erumpent from the bark of the host plant, bipartite, consisting of two layers with deciduous entostroma. Ascomata perithecial, immersed in the entostroma. Asci unitunicate, oblong to cylindrical, in spicate arrangement. Paraphyses sparse, elongate, tapering from wide base. Ascospores unicellular, allantoid and hyaline or brown and ellipsoid, at times with appendages, with or without germ slits, without dehiscent perispores. Stromatal pigments absent. Asexual morph of the nodulisporium-type, most often periconiella-like or xylocladium-like.
The concept of the Graphostromataceae, erected by Barr et al. (1993) for the monotypic genus and species G. platystomum, surprisingly held true in our multigene genealogy. We have added Biscogniauxia and Camillea, as well as Obolarina and Vivantia to this family, even though we realise that many species of the two former, relatively large genera remain to be studied in-depth and that virtually no reference sequence data are available for them in the public domain. However, previous studies on Biscogniauxia using molecular phylogenetic methods (Hsieh et al. 2005;Collado et al. 2001) have always shown the genus to be rather homogeneous and the same holds true for the numerous data on Biscogniauxia spp. in GenBank. Therefore, we do not expect many surprises in the future. Camillea primarily differs from Biscogniauxia by the more complex stromatal anatomy of several species and by its ascospore morphology (pale brown to uncolored, lacking a germ slit) and by the xylocladium-like conidiogeneous structures, which were mostly observed on the stromata. This genus was regarded by Laessøe et al. (1989) and Ju et al. (1998) as closely related to Biscogniauxia, but possibly more evolutionarily advanced. For a detailed treatment of these genera we refer to the papers mentioned above. Molecular data are widely amiss for many taxa of Biscogniauxia and most species of Camillea, and many of their species have never been cultured. Obolarina and Vivantia definitely appear related to Biscogniauxia and each genus only differs from the latter in a single salient morphological character (i.e., ascospore morphology and/or complexity of stromata). Further research using extensive studies on the anamorphic traits and molecular phylogenetic data may be helpful in their segregation. Interestingly, the most salient morphological feature of the species included in this family are the erumpent, bipartite stromata, which are often observed on the surface of still living woody plants. This is a rare case that a "macromorphological" character can be specific for a family in the Ascomycota, but it may be related to the endophytic lifestyle of these fungi. Notes: We selected the illustration of Bulliard since we could not find any data in the literature indicating that this species has ever been lectotypified. The location of the epitype is in an area where it is very commonly found in the beech forests. This area used to alternatively belong to France or Germany, during the time when German and French mycologists were proposing different alternative names for this area. Type species: Vivantia guadalupensis J.D. Rogers, Y.M. Ju & Cand. 1996.
Note: This genus is tentatively assigned to the Graphostromataceae because it has a nodulisporium-like anamorph (reminiscent of the periconiella-like type) and bipartite stromata and does, therefore, match the characteristics of the family better than those of the Hypoxylaceae or the Xylariaceae s. stricto. However, it remains to be studied by molecular phylogenetic methods.

Xylariales Incertae Sedis
The segregation of Lopadostomataceae (and as practiced here, Graphostromataceae and Hypoxylaceae) from the Xylariaceae s. lat., makes it difficult (if not impossible) to assign many genera to either of the new families, as neither anamorphic structures nor molecular data exist for this genera. A taxon that is only known from drawings or from old depauperate herbarium, specimens cannot be accommodated any longer in the Xylariaceae based solely on the ascospore morphology.
There had been two choices when interpreting the current phylogeny: a) reject the concepts brought forward by the recent studies that have revealed the Lopadostomataceae cited in the introduction, based on a polyphasic approach, or b) accept these concepts and apply it to the new data that are presented in the current study, although we still have various gaps.
Since we were not inclined to deny the tremendous progress on our understanding of the phylogenetic affinities in this highly complex and diverse group of Ascomycota, we have chosen option b).
Therefore, several genera were expelled from either of the families and placed at interim in Xylariales incertae sedis. We frankly hope that this procedure may point other mycologists toward these genera with unsettled phylogeny, promoting the recollection of fresh material to fill our gaps in the knowledge on the affinities of the Xylariomycetideae. All genera, whose affinities remain still unsettled and which we do not regard to belong to the Xylariaceae s. str., are listed below: Note: This genus was recently segregated from Anthostomella based on molecular phylogenetic data and morphological traits ). In their phylogenetic tree, it appeared related to Neoanthostomella and Biscogniauxia, rather than to Xylaria. It differs from Neoanthostomella in lacking pigmented ascospores. However, various xylariaceous and graphostromataceous taxa were not included in this phylogenetic study. We refrain from assigning it to one of the families in the current concept, until additional data have become available. Type species: Biporispora europaea J.D. Rogers, Y.M. Ju & Cand. 1999.
Notes: The type species of this monotypic genus is apparently a parasite of the stromata of Hypoxylon macrocarpum Pouz. The assignment to the Xylariaceae was tentative, based on the ascospore morphology, which was, however, regarded by the authors as atypical for the family. No data on anamorphic structures or molecular phylogeny are available. Recently generated, yet unpublished DNA sequence data actually point toward a placement of Biporispora in the Chaetosphaeriales (A. N. Miller, pers. comm.). Notes: This genus was recently erected based on a single cultured specimen and the placement in the Xylariaceae was based on DNA sequence data, where it appeared in the hypoxyloid clade ). However, the morphology does not match any known xylariaceous taxon. The teleomorph is unknown, while the anamorph is coelomyceteous and the conidia are not typical for Xylariaceae or Hypoxylaceae at all. The extype culture was only grown on PDA (rather than on the conventional media used for sporulation) where the coelomyceteous structures found in the type material were not observed again (as is usually the case, e.g. in Xylariales spp. featuring libertella-like anamorphic stages) and it was not examined at all for comparison with cultures of other taxa that have similar phylogenetic characteristics. We strongly suspect that it may constitute a contaminant and think that more collections should be made available for verification.
Cannonia Joanne E. Taylor  Note: The anamorph of this genus is unknown and no molecular data are available. The asci do not match any known xylariaceous taxon. They are claviform, short-pedicellate, evanescent and lack an apical apparatus. It is quite common to find only free ascospores as usually no intact asci can be observed (Trierveiler-Pereira et al. 2012). These authors already discussed that Carlos Spegazzini originally had placed it in Ceratostoma Fr., a genus generally possessing asci with an evanescent wall. Taylor & Hyde (1999) compared features of the Coniochaetaceae and Xylariaceae to place Cannonia in the latter family, but molecular phylogenetic data are definitely needed to clarify the phylogenetic affinities of this genus. Note: The current study, in which the ex-type strain of this monotypic genus is included, revealed a phylogenetic placement outside the major clades, close to Creosphaeria (Lopadostomataceae). A subsequent study including more representatives of the latter family and non-stromatic Xylariales should be carried out for comparison to shed more light on its affinities. Note: The anamorph of this genus (which has been referred to as a member of the Xylariaceae and was in fact, until recently, listed under this family in public databases) is unknown and no molecular data are available, except for a LSU sequence (GenBank Acc. No. AY346278) that proved to be only of little use for our current phylogenetic study. The authors have clearly stated that they prefer to place it in Xylariales incertae sedis and we agree. Note: For data on the phylogenetic position of this genus see Notes to Theissenia, with which Durotheca is obviously closely related. The anamorph of D. comedens was described by Ju et al. (2003) to be nodulisporium-like, but they observed rather long slender conidia that are somewhat reminiscent of scolecospores. Interestingly, the stromata of Durotheca comedens were found to contain lepraric acids, which is indicative of chemotaxonomic affinities t o  Type species: Fasciatispora nypae K.D. Hyde 1991.
Notes: The few sequence data of this genus show high homologies to Barrmaelia followed by Lopadostoma species. However, only a LSU sequence (GenBank Acc. No. KP744484) is available from material assigned to the type species. Further work must be carried out to demonstrate its affinities. Note: A nodulisporium-like anamorph was reported from the type species of this genus in the protologue, but it has ascospores reminiscent of the Apiosporaceae and neither a living culture nor molecular data are available.
Leptomassaria Petr., Annls mycol. 12 (5) Note: This genus is reminiscent of Anthostomella, but its anamorphic structures are not known and no molecular data are available.
Type species: Nipicola carbospora K.D. Hyde 1992. Note: The anamorph of this genus is unknown and no molecular data are available. Note: The anamorph of this genus is unknown and no molecular data are available. Note: The anamorph of this fungus is unknown, and if it were found to be geniculosporium-like, it would possibly have to be synonymised with Leprieuria.
Note: DNA sequences derived from the type species of this genus were reported to be basal in the phylogeny by Daranagama et al. (2015) with focus on Anthostomella. The data available on the simple conidiogeneous structures are reminiscent of the "sporothrix-like" conidiogeneous structures defined by Ju and Rogers (1996) and suggest that it is ancestral. It is not clear, whether it represents a separate lineage, or whether it can be categorised in one of the existing families. We think that more specimens need to be collected and studied before a final conclusion can be reached.
Note: This genus was recently segregated from Anthostomella based on molecular phylogeny and morphological traits and like Alloanthostomella, it showed affinities to Biscogniauxia, rather than to Xylaria. However, various xylariaceous and graphostromataceous taxa were not included in this phylogenetic study. It is characterised by an amyloid ascal apparatus and morphologically differs from the genus Anthostomelloides (which is clearly a member of the Xylariaceae based on molecular data) by the lack of a central periphysate ostiolar canal. In the phylogenetic tree, it appeared related to Neoanthostomella and Biscogniauxia, rather than to Xylaria. We refrain from assigning it to one of the families in the current concept, until additional data have become available.
Note: The anamorph of this genus is unknown and no molecular data are available. Note: According to the protologue, and in particular with respect to the morphology of its asci and ascospores (which are lacking germ slits; however, this also applies to a number of other xylarialean species), this genus may tentatively be assigned to the xylarioid Xylariales, but might represent yet another phylogenetic lineage within the Sordariomycetes. Studies on fresh material, possibly including a morphological characterisation of the anamorph and molecular phylogenetic data are needed to settle its affinities.

1914.
Notes: This genus appears heterogeneous even after the segregation of Durotheca by Laessøe et al. (2013) with respect to both teleomorph and anamorph morphology. The phylogeny of Ju et al. (2007) based on protein coding genes suggests affinities to both Biscogniauxia (Graphostromataceae) and Whalleya (Lopadostomataceae). Further studies will show whether Theissenia, which has been regarded as a basal or even doubtful member of the Xylariaceae s. lat., can be assigned to one of the aforementioned families or whether the genus and the related Durotheca will have to be elevated to a family of their own. From a phylogenetic point of view, it cannot be retained in the Xylariaceae and was also found to be quite distant to the taxa that are here regarded as Hypoxylaceae.
Note: Contrary to previous reports in the literature, the type is extant in B, where it was discovered in 2007 (M.S. personal observations, confirmed by J. Fournier). However, the stromata were soaked in ethanol for over a hundred years; no spores were found in the depauperate stroma, and even DNA extraction proved futile. The detailed description by Möller (1901) could point toward this genus being close to Xylaria or Sarcoxylon, but as no data on anamorphic structures were reported, we cannot safely assign it to any of the families treated herein. However, there is no doubt that X. pyriformis belongs to the Xylariales. Note: The anamorph of this genus is unknown and no molecular data are available.
Notes: This monotypic genus cannot be accommodated in either of the current families, since neither morphological data of the anamorph, nor molecular phylogenetic data are available, and their teleomorph and ascospore morphology is atypical for the Xylariaceae s. lat. even in the "traditional" definition. We think that its placement in the Xylariales is highly tentative and questionable. Notes: The type species of Areolospora is presently regarded as a synonym of Phaeosporis melasperma (Nyl.) Clem. 1909 (Sordariales), but no cultures and no molecular data are available on this taxon.
Type species: Basidiobotrys clautriavii (Pat.) Höhn. 1909. Notes: This genus had been originally proposed to replace Xylocladium, which is an anamorph stage of Biscogniauxia and Camillea. Hence, it was listed by Stadler et al. (2013) as a synonym of Xylocladium. However, no ex/type strains and illustrations are available, and there will be no way to relate it to any of the extent genera by modern polythetic methodology. It is actually not even possible to confirm its placement in the Xylariomycetideae and it should, therefore, be abandoned.
Notes: This genus is so far only defined on molecular data and the lack (!) of salient morphological features. We feel that it should be rejected, because the publication did not follow good taxonomic standards (cf. Stadler et al. 2013).
Note: The type material of this genus appears to be lost, and the description is vague. Hence, it is not possible to include it in any xylarialean taxon at this time.
Triplicaria P. Karst., Hedwigia 28: 195 (1889). Type species: Triplicaria hypoxyloides P. Karst. 1889. Note: The protologue suggests that the type species of this genus is a synnemata-forming hyphomycete that was assigned to Hypoxylon by some mycologists in the past, but might actually belong to many different genera in the current sense. Even an assignment to one of the Xylariales families in the current definition will never be possible and it is, therefore, suggested to abandon the genus name.
The genus Annulohypoxylon was erected by Hsieh et al. (2005) to accommodate the former sect. Annulata of Hypoxylon s. Ju and Rogers (1996) based on molecular phylogenetic data inferred from ACT and TUB2 DNA sequences. The segregation was also found in accordance with the morphological concept as inferred from the latter monograph. Concurrently, Quang et al. (2005a) had also provided first evidence that species of Hypoxylon sect. Annulata (i.e. Annulohypoxylon species) have different stromatal secondary metabolites. They reported cohaerins A and B from "Hypoxylon" (=Annulohypoxylon) cohaerens, which were then discovered as the first members of a novel class of azaphilone pigments. Moreover, they also studied several specimens of what is now regarded as Annulohypoxylon, using high performance liquid chromatography with diode array and mass spectrometric detection (HPLC-DAD-MS), showing that none of them contained any of the mitorubrin or daldinin type azaphilones that prevail in many species of Hypoxylon s. str. Later, various other pigments of the cohaerin/multiformin type were discovered from species of Annulohypoxylon (Quang et al. 2005b(Quang et al. , 2006Surup et al. 2013;Kuhnert et al. 2017). None of them has so far been encountered in another xylariaceous species, let alone any other fungus, even though some yet unidentified metabolites with cohaerin-like UV/visible spectra were detected in the stromata of certain species of Hypoxylon (cf. H. pulicicidum, Bills et al. 2012). A recent extensive study by Kuhnert et al. (2016) embarked on these data. Numerous type and authentic specimens of Annulohypoxylon including several new and recently erected species were studied in-depth using a more sophisticated HPLC-DAD-MS methodology. In addition, they combined published data from GenBank and newly generated DNA sequences and provided an updated phylogeny of the genus. Interestingly, the results confirmed that the species in which the cohaerin/multiformin type azaphilones had been detected showed a morphological peculiarity. With the exception of A. michelianum, they all have papillate ostioles. The species of the former group with papillate ostioles that have so far been cultured and sequenced also clustered together in the phylogeny based on TUB2 DNA sequences. In our mind, this segregation of the genus Annulohypoxylon s. Hsieh et al. (2005) which is confirmed by the current multigene genealogyinto distinctive clades warrants the establishment of a new genus for which we propose the name Jackrogersella.
Below we provide a description of the new genera, followed by an account of the species that remain in Annulohypoxylon according to the current taxonomy and their most important synoynyms. Epitypes have also been designated for two important species, which were concurrently evaluated by molecular phylogeny.
Jackrogersella L. Wendt, Kuhnert & M. Stadler, gen. nov. MB 819742 Etymology: In honor of Professor Jack D. Rogers, to acknowledge his tremendous accomplishments in Ascomycota taxonomy.
Differs from the genus Annulohypoxylon by containing cohaerin/multiformin type azaphilones as predominant stromatal pigments.   Rogers, Mycol. Mem. 20: 221 (1996). Notes: The genus Jackrogersella comprises the group of species formerly included in Hypoxylon sect. Annulata s. Ju and Rogers (1996) and Annulohypoxylon that have papillate ostioles and are lacking very conspicuous ostiolar disks. As can be seen in Fig. 8, some of these species such as J. gombakensis and J. ilanensis do have ostioles encircled by a disk, but they show similar secondary metabolite profiles as the remainder of Jackrogersella and also clustered with those in the phylogeny of Kuhnert et al. (2016). Notably, ostiolar disks are not exclusively found in the genus Annulohypoxylon, but even occur in other xylariaceous genera such as Hypoxylon (e.g. H. monticulosum Mont.) and Ruwenzoria (Stadler et al. 2010c).
The most salient feature to discriminate the new genus from Annulohypoxylon is, therefore, a chemotaxonomic trait: It is characterised by the specific occurrence of the unique cohaerin/multiformin type azaphilones while apparently lacking daldinone A, truncatone and other binaphthalenes that occur in many species of Annulohypoxylon s. str. as stromatal pigments. Interestingly, the binaphthalene derivative hinnulin may eventually turn out to be recognised as a chemotaxonomic bridging character, because it occurs in J. minutella, as well as A. purpureopigmentum (cf. Kuhnert et al. 2017). In our phylogeny, A. michelianum formed a separate, distinct clade that nevertheless clustered apart from Annulohypoxylon. Another new genus may have to be coined for this species, if the results of the molecular phylogeny will be confirmed during the course of further studies. The anamorph states of various species in the Jackrogersella clade have for instance never been evaluated, and also the information on their secondary metabolites produced in culture is incomplete. One species, J. nothofagi remains to be cultured and studied indepth on its phylogenetic affinities and its metabolite profiles, since the data on the specimens we have at hand are still inconclusive.
Sexual morph. Stromata effused-pulvinate, pulvinate, glomerate, discoid, hemispherical, or spherical, solitary or confluent, attached to substrate with a broad base; surface light-or dull-colored, usually blackened with age, pruinose or polished, planar or with inconspicuous to conspicuous perithecial mounds; waxy or carbonaceous tissue immediately beneath surface and between perithecia, with KOHextractable pigments in most cases; the tissue below the perithecial layer inconspicuous, conspicuous, or relatively large, dark brown to black, persistent. Perithecia spherical, obovoid, or less frequently tubular, monostichous, with carbonaceous stromatal layer surrounding individual perithecia. Ostioles higher than the level of stromatal surface, with the ostiolar openings papillate to conical papillate, with conspicuous to hardly noticeable discs formed by dehiscence of surrounding tissue. Asci eight-spored, cylindrical, stipitate, persistent, with apical ring discoid, amyloid or infrequently inamyloid, distinct. Ascospores light-to dark-colored, unicellular in both mature and immature ascospores, ellipsoid or shortly fusoid, inequilateral, slightly inequilateral or nearly equilateral, with acute, narrowly rounded, or broadly rounded ends, with a germ slit spore-length to much less than spore-length on the convex side or less frequently on the flattened (side; absent in A. macrosporum;) perispore dehiscent or indehiscent in 10% KOH, when dehiscing, with a thickened area visible at the position of ca. 1/3 ascospore length on the same side as the germ slit; epispore smooth. Asexual morph. Anamorph produced on young stromata, or in artificial culture, conidiophores mononematous, usually macronematous, hyaline or colored, smooth or roughened, with nodulisporium-like or rarely periconiella-like branching patterns (as defined in Ju and Rogers 1996), with holoblastic conidiogenesis.     ≡ Nummularia urceolata Rehm 1913. 1 The type material of this species seems to be lost (cf. Kuhnert et al. 2016). μm Dense amber, isabelline to olivaceous BNT, naphtoquinones, napththols *Ascospore sizes of D. placentiformis were taken from the type specimen and other herbarium material reported by Stadler et al. (2014a); the broader range of dimensions given in   Note: After re-examination of the type material, we believe that this species does not really fit into the concept of Annulohypoxylon by Hsieh et al. (2005) because the morphology of the ostiolar discs (Fig. 8) seems to be quite different. However, it still resembles Annulohypoxylon more than any other xylariaceous genus. As no information on the conidial state and no molecular data are available, we prefer to retain it ad interim in Annulohypoxylon. Notes: This species was included in the phylogeny of Kuhnert et al. (2016) based on material from French Guiana, whereas the holotype is from Brazil. The phylogeny revealed that its ITS sequence was located on a branch well outside the clade comprising what is here regarded as Jackrogersella, whereas the TUB2 sequence was nested inside the "Jackrogersella clade" (i.e. the clade comprising the species that are moved to the new genus), but on a rather long branch. Since the French Guianean specimen, as well as the holotype, were not found to contain cohaerins or multiformins as stromatal pigments, it is maintained in Annulohypoxylon for the time being, but its taxonomy needs further study. Notes: This species (for a recent treatment see Rubio & De la Pena 2016) occupies a separate clade in our phylogeny, chemotaxonomically belongs to Jackrogersella and its stromatal metabolites are the same as in the latter genus. A new genus would actually be justified from the outcome of our molecular phylogeny. However, we did so far not find any morphological trait t o segregate it from e ither Annulohypoxylon or Jackrogersella. We also were unable to observe any conidiogeneous structures in the two cultures we obtained, both of which did not dfferentiate much. For these reasons, we maintain it in Annulohypoxylon for the time being, until the phylogenetic position has been verified by additional studies.

Resurrection of Pyrenopolyporus
The current phylogeny revealed a sister clade Py of Daldinia and allies that was comprised of three representatives of a peculiar group of tropical xylariaceous fungi that are characterised by massive, peltate to discoid stromata and long tubular perithecia and have been assigned to Hypoxylon sclerophaeum by Miller (1961). Ju and Rogers (1996) segregated this species complex and resurrected various species that had been lumped by Miller (1961) under the aforementioned name. Various mycologists previously discussed that these species might represent intermediate forms to the genus Daldinia, whose species mostly differ in having stromata with conspicuous internal concentric zones (cf. Theissen 1909, Ju et al. 1997. Indeed, there are two species in the latter genus that deviate from typical Daldinia with respect to their stromatal anatomy: Daldinia placentiformis ) and the recently described D. korfii (Sir et al. 2016b).
These species show close similarities with respect to their stromatal morphology to Daldinia placentiformis, but deviate in their ascospore morphology and, where this is known, also in their anamorphic branching pattern and the production of secondary metabolites in their stromata and cultures (Bitzer et al. 2008). For instance, cultures of what is here regarded as Pyrenopolyporus have a characteristic virgariella-like conidial stage and produce cochliodinol and 8-methoxy-1-naphtol but no chromones, eutypinols and phytotoxic lactones of the "Ab-5046" type, which are characteristic for Daldinia and other taxa included in clade D of the current phylogenetic tree. The characteristic metabolites from their stromata and cultures are depicted in Fig. 10. We have also checked the stromata of several specimens and found that they contain BNT, hypoxylone (Bodo et al. 1983) and yet unidentified metabolites that are not present in D. placentiformis, which mainly contains daldinone A. One of the species included in this group, H. polyporus, has previously been validly described as Pyrenopolyporus hunteri by Lloyd (1917), meaning that the name is available. We think that it is better to resurrect this genus, rather than erect a new one; in particular as the name Pyrenopolyporus matches very well the macroscopic appearance of these fungi, whose stromata indeed resemble a polyporaceous basidiomycete. An account of morphological features of Pyrenopolyporus species and morphologically similar representatives of Hypoxylon and Daldinia is given in Table 3. Differs from the genera Daldinia and Hypoxylon by producing cochliodinol in its cultures. Differs from the genus Daldinia by having ascospores with indehiscent perispores that are not broadly ellipsoid with rounded ends and by forming stromata with a homogeneous context and long tubular perithecia at the same time. Differs from all other genera of the Hypoxylaceae by having massive stromata with long tubular perithecia that are lacking conspicuous ostiolar rings.
Sexual morph. Stromata discoid, peltate, hemispherical, convex on top, centrally attached to substrate, sessile or with a short and broadly attached central base, with crenate to entire margins, separate to coalescent; surface colored, planar or with inconspicuous to conspicuous perithecial mounds; granules black or colored immediately beneath surface, with or without KOH-extractable pigments, the tissue below the perithecia layer massive, soft-textured. Perithecia long tubular. Ostioles lower than the stromatal surface or slightly higher than the stromatal surface, umbilicate. Asci eight-spored, cylindrical, stipitate, persistent to evanescent, with apical ring discoid to cuneate, amyloid, distinct. Ascospores light-to dark-colored, unicellular in both mature and immature ascospores, ellipsoid-inequilateral, ellipsoid, slightly inequilateral, highly variable or irregularly shaped, narrowly rounded ends, with straight to rarely slightly sigmoid germ slit much less than spore-length  Rogers, Mycol. Mem. 20: 196 (1996).
Notes: The "new" old genus was resurrected and segregated from its allies based on a combination of molecular phylogenetic and chemotaxonomic traits that helped substantially to interpret its salient morphological features. However, this group of hypoxyloid pyrenomycetes actually has been regarded as an intermediate form between Hypoxylon and Daldinia for a very long time (cf. Bitzer et al. 2008). The peculiar secondary metabolite profiles with cochliodinol production in the cultures discriminate it from Daldinia, where the current generic concept includes species that have massive stromata with long tubular perithecia that are lacking the typical internal concentric zones on which the latter genus was originally based. Cochliodinol was also not detected in various species of Hypoxylon and allies studied previously (Bitzer et al. 2008). However, the ITS-based phylogeny of the aforementioned studies did not yet provide conclusive data that would justify the segregation of Pyrenopolyporus from Hypoxylon. This has now been accomplished as additional data from the studies of Kuhnert et al. (2014a) and Sir et al. (2016a, b) became available.
The genus Pyrenopolyporus may be amended in the future as there are various species in Hypoxylon that do not form the characteristic discoid or peltate stromata that are characteristic for the genus, but appear similar in their micromorphological features. These species (e.g. H. sclerophaeum) are only known from old stromatal collections. Their conidiogeneous structures remain to be studied, and no DNA sequence data are available for these species as yet. They will need to be collected and cultured to allow for an assessment of their anamorphic morphology and their phylogenetic position. On the other hand, there are other species in Hypoxylon with massive stromata and long tubular perithecia, which seem to belong to different lineages in the Hypoxylaceae. This concerns Hypoxylon kretzschmarioides, one of the species that is morphologically similar (cf. Ju and Rogers 1996). A re-examination of the holotype material has revealed that the perispore is dehiscent and it also did not show the characteristic HPLC profile that is common in all species of Pyrenopolyporus (J. Fournier & M.S., unpublished). In accordance with its purple stromatal pigments, BNT was detected as a major stromatal metabol i t e , b u t n o h y p o x y l o n e w a s f o u n d . P o s s i b l y, H. kretzschmarioides, which shows strong morphological similarities to Daldinia placentiformis, will eventually be transferred to the sister genus Daldinia. Another widely unknown species with massive stromata is Hypoxylon

Discussion
The present study attempted to segregate the large and heterogeneous family Xylariaceae based on polyphasic taxonomic methodology, and to provide a concept that will be feasible to include the numerous non-stromatic taxa that are now being recognised to have close phylogenetic affinities to the "Large Pyrenomycetes" that were historically the first members of the Xylariales known to science. Our molecular phylogeny agrees with the results of several concurrent studies focusing on different taxa. It ultimately proves that the traditional segregation of families in the Xylariomycetideae based on ascal and ascospore morphology as major discriminative criteria is artificial. This could actually have been predicted, since many xylariaceous taxa have an aberrant ascal and ascospore morphology, anyway. Examples are the lack of amyloid ascal plugs in many species of Hypoxylon that are otherwise morphologically very similar to the "typical" members of the genus, and the "cleistocarpous" genera Phylacia, Pyrenomyxa, Rhopalostroma, and Thamnomyces.
The current classification should be stable and allows more easily to place new taxa in one of the families that have been defined in a phylogenetic context. It will of course be necessary to (re-) collect, culture and sequence many taxa that are only known from old descriptions and herbarium material in order to verify their phylogenetic position. In some cases, (in particular for Hypoxylon and Daldinia and their respective relatives) a lot of work remains to be done. For instance, it will be necessary to sample further taxa and even generate sequence data of other DNA loci in order to attain a higher resolution. In any case, the use of a combination of protein coding genes and rDNA has been a great advantage in connecting the phylogeny with historical morphological concepts.