Solving the taxonomic identity of Pseudotomentellatristis s.l. (Thelephorales, Basidiomycota) – a multi-gene phylogeny and taxonomic review, integrating ecological and geographical data

Abstract P.tristis is an ectomycorrhizal, corticioid fungus whose name is frequently assigned to collections of basidiomata as well as root tip and soil samples from a wide range of habitats and hosts across the northern hemisphere. Despite this, its identity is unclear; eight heterotypic taxa have in major reviews of the species been considered synonymous with or morphologically similar to P.tristis, but no sequence data from type specimens have been available. With the aim to clarify the taxonomy, systematics, morphology, ecology and geographical distribution of P.tristis and its morphologically similar species, we studied their type specimens as well as 147 basidiomata collections of mostly North European material. We used gene trees generated in BEAST 2 and PhyML and species trees estimated in STACEY and ASTRAL to delimit species based on the ITS, LSU, Tef1α and mtSSU regions. We enriched our sampling with environmental ITS sequences from the UNITE database. We found the P.tristis group to contain 13 molecularly and morphologically distinct species. Three of these, P.tristis, P.umbrina and P.atrofusca, are already known to science, while ten species are here described as new: P.sciastrasp. nov., P.tristoidessp. nov., P.umbrinascenssp. nov., P.pinophilasp. nov., P.alnophilasp. nov., P.alobatasp. nov., P.plurilobasp. nov., P.abundilobasp. nov., P.rotundisporasp. nov. and P.mediasp. nov. We discovered P.rhizopunctata and P.atrofusca to form a sister clade to all other species in P.tristis s.l. These two species, unlike all other species in the P.tristis complex, are dimitic. In this study, we designate epitypes for P.tristis, P.umbrina and Hypochnopsisfuscata and lectotypes for Auriculariaphylacteris and Thelephorabiennis. We show that the holotype of Hypochnussitnensis and the lectotype of Hypochnopsisfuscata are conspecific with P.tristis, but in the absence of molecular information we regard Pseudotomentellalongisterigmata and Hypochnusrhacodium as doubtful taxa due to their aberrant morphology. We confirm A.phylacteris, Tomentellabiennis and Septobasidiumarachnoideum as excluded taxa, since their morphology clearly show that they belong to other genera. A key to the species of the P.tristis group is provided. We found P.umbrina to be a common species with a wide, Holarctic distribution, forming ectomycorrhiza with a large number of host species in habitats ranging from tropical forests to the Arctic tundra. The other species in the P.tristis group were found to be less common and have narrower ecological niches.


Introduction
Species of the genus Pseudotomentella Svrček are recognised by their smooth, corticioid, membranaceous basidiomata, bi-or trifurcately echinulate basidiospores and their lack of cystidia (Larsen 1971a, Stalpers 1993, Kõljalg 1996. All species, for which a life strategy has been confirmed, are ectomycorrhizal (Agerer 1994, Kõljalg et al. 2000, Cline et al. 2005, Di Marino et al. 2007, Trocha et al. 2012, Branco et al. 2013, Malysheva et al. 2016) and the genus is widely distributed throughout the northern hemisphere (Larsen 1971a, Kõljalg 1996. Basidiomata are formed on the underside of dead wood, turf and stones, where their spores are probably dispersed by insects, as found by a study on a species in the closely related genus Tomentella Pers. ex Pat. (Lilleskov and Bruns 2005).
Pseudotomentella tristis (P.Karst.) M.J.Larsen is characterised by its brown to bluishgrey, sometimes green-tinged basidiomata, simple septate, monomitic hyphal system, wide subicular hyphae and large, yellow to brown basidiospores (Larsen 1971a, Stalpers 1993, Kõljalg 1996. In this current, morphological delimitation of the species, it is probably the most commonly collected Pseudotomentella species in the world: out of 1038 herbarium specimens registered in GBIF (13-08-2018), 497 are attributed to P. tristis -the second most common species being Pseudotomentella mucidula (P.Karst. ) Svrček with a total of 230 specimens. Even though there is no taxonomic study currently linking the type of P. tristis to molecular information, it is also a name frequently assigned to sequences recovered in molecular ecology studies of ectomycorrhizal com-munities in soil and on root tip samples (e.g. Kõljalg et al. 2000, Cline et al. 2005, Hrynkiewicz et al. 2009, Walbert et al. 2010, Izumi and Finlay 2011, Argüelles-Moyao et al. 2017, some of which even report it to constitute one of the most common species found (Kõljalg et al. 2000, Izumi andFinlay 2011). Entries in international sequence and specimen databases from 15 countries in the northern hemisphere indicate that it is a very widespread species (Kõljalg et al. 2005, Clark et al. 2015, GBIF 13-08-2018, Nilsson et al. 2019. Concordantly, it also has a very large ecological amplitude: sequences attributed to P. tristis have been encountered in habitats and with hosts ranging from the Swedish tundra with Salix polaris Wahlenb. (Hrynkiewicz et al. 2009) to the neotropics of Mexico with Abies religiosa (Kunth) Schltdl. & Cham. (Argüelles-Moyao et al. 2017).
Taxonomically and nomenclaturally, P. tristis is a species with a long history. Based on French material, Bulliard (1790) described Auricularia phylacteris Bull. -a fungus with a large corticioid basidiome, a plicated base and an initially pale, but with maturity darkening hymenium. Fries (1821) described Thelephora biennis Fr. with reference to A. phylacteris.
In 1828, Fries introduced the name Thelephora umbrina Fr. to describe a soft, brown, effused basidiome, which he stated that he had seen alive (Fries 1828). Karsten (1889) raised a subspecies, Hypochnus subfuscus ssp. tristis P.Karst., that he had previously described, based on Finnish material (Karsten 1883), to the level of species, giving it the name Hypochnus tristis (P.Karst.) P.Karst. In his protologue, he wrote of it as having a wool or felt-like basidiome and a thick, blackish hymenium, with hues of olive brown or green, coloured brown by the detaching spores. He noted the spores to be roundedly angular, aculeate, yellow or brown and 8-12 µm in diameter. In 1889, Karsten also described Hypochnopsis fuscata P.Karst. -a second species from Finland, whose description is similar to that of H. tristis, except that, according to the author, the colour of the hymenium is bluish-black and the spores are smooth, bluish with a dark wall and 3-4 µm in diameter (Karsten 1889). Bresadola (1897) described the species Hypochnus sitnensis Bres. from a Hungarian basidiome with a soft, chestnut brown colour, a smoke-coloured hymenium and spores similar in size to those of H. tristis. Berkeley and Broome (1873) created the new name Thelephora arachnoidea Berk. & Broome for their sparse description of a basidiome with a powdery, grey hymenium and a soft, black subiculum, based on a collection made in Sri Lanka. Burt (1926) described Hypochnus rhacodium Berk. & M.A. Curtis ex Burt from a US specimen. He wrote of it as having a crust-like and brittle texture, fuscous to dusky drab appearance and spores measuring 6-7 µm in diameter.
Following the original descriptions of A. phylacteris, T. biennis, T. umbrina, H. tristis, H. fuscata, H. sitnensis, T. arachnoidea and H. rhacodium, a large number of publications were made, proposing new combinations and synonymisations (Gmelin 1792, de Candolle 1815, Fries 1821, 1849, 1874, Quélet 1888, Saccardo 1888, Bresadola 1903, von Höhnel and Litschauer 1906, Burt 1916, Donk 1933, Litschauer 1933, Rogers 1948, Parker-Rhodes 1956, Svrček 1958. No new, morphologically similar species were published until 1967, when Larsen, based on a basidiome he collected in USA, described Pseudotomentella longisterigmata M.J.Larsen -a fungus with a greyish-green hymenium and unusually long sterigmata. In his description of P. longisterigmata, Larsen also combined Thelephora umbrina to Pseudotomentella. He then proceeded to synonymise all other species similar to P. umbrina described thus far (Larsen 1968). Hjortstam (1969), however, argued that Thelephora umbrina and Hypochnus tristis represented different species that could be separated mainly based on the colour and texture of basidiomata: T. umbrina could be recognised by its pale to dark chocolate brown hymenium and softer texture and H. tristis could be distinguished by its dark greenish-blue, sometimes brownish-tinted hymenium and firmer texture. Larsen (1971a) disagreed: writing that he had observed a continuum of variation between the distinct character states defined by Hjortstam as characteristic of the two taxa, he considered them as one. Larsen (1971a) further proposed that the taxon in question should have the species epithet tristis instead of umbrina, with reference to his interpretation of Fries's (1828) original description and to Rogers and Jackson (1943), who claimed Thelephora tristis to be a synonym of Coniophora olivacea (Fr.) P.Karst. He consequently made the combination Pseudotomentella tristis M.J. Larsen (1971a) and urged that the use of Thelephora umbrina and all its homotypic synonyms be discontinued. Larsen (1971b) described one additional species, Pseudotomentella atrofusca M.J.Larsen, based on a blackish-brown, soft basidiome with spores 5.5-6.6 µm in diameter, from the US. Kõljalg (1996) considered P. longisterigmata to be a synonym of P. atrofusca. Beside P. tristis, P. atrofusca is the only name left in common use today (GBIF 13-08-2018). It is employed for small-spored specimens in both North America and Europe, but is considerably less frequently collected (GBIF 13-08-2018).
Thus, in conclusion, ten names have so far been associated with the P. tristis group, as here defined. Of these, only two -P. tristis and P. atrofusca -remain in use today. Pseudotomentella tristis is under the currently employed, morphological delimitation regarded as a common species with a very wide geographic distribution and ecological amplitude. The purpose of the present study is to molecularly delimit species within the P. tristis group, describe their morphology and present knowledge on their ecology and geographical distribution -describing new species and designating types as needed.

Taxon sampling and information
We collected specimens of basidiomata extensively throughout Sweden, Norway and Estonia in the period 2010-2017. For the majority of the Swedish and Norwegian specimens, we recorded the vegetation type of each locality, following Fremstad (1997) and Pålsson (1998). We then sorted this information into the habitats "tundra", "co-niferous forest", "deciduous forest" and "mixed forest" and the soil pH types "low", "intermediate" and "high", following Pålsson (1998) and Halvorsen (2015). We also recorded potential hosts of each specimen, as indicated by nearby ectomycorrhizaforming plants. The Swedish specimens were photographed, weather permitting. We complemented the fresh material by examining all collections identified as P. tristis, P. atrofusca and Pseudotomentella sp. in GB and TU, along with collections identified as P. tristis in TUR and H, and relevant type specimens in S, H, BPI and ARIZ. Permission to extract DNA was granted. In addition, we studied Fries' collection of Thelephora umbrina in situ at UPS. Taxonomic author abbreviations follow IPNI (26-11-2018) and herbarium codes follow Index Herbariorum (Thiers 2018). Abbreviations of journal titles follow BPH Online (26-11-2018) and abbreviations of book titles follow Stafleu andCowan (1976-1988).

Morphological data
We studied all specimens macroscopically and at 20× magnification under a dissecting microscope. Photos of micromorphological characters and measurements were made using an Axioskop 2 microscope (Zeiss, Oberkochen, Germany), equipped with an AxioCam MRc camera (Zeiss) at 400× and 1000× magnifications and in the ZEN Blue software (http://www.zeiss.com/microscopy/int/home.html). Measurements were made on dried material, mounted in 3% (potassium hydroxide) KOH and in Melzer's reagent. We examined a minimum of three specimens per species, whenever the total number of specimens allowed it, and we measured 20-30 micromorphological structures of each type. Measurements were made to the nearest 0.1 µm, except basidial length, which was measured to the nearest µm. As a necessary aid in identification, we present the values recorded both as spans of the lowest to the highest value and as mean values. For the spans, the 5% smallest and largest measurements are denoted in brackets, in the cases where they differed from the remaining 90%. We calculated the mean values for each specimen omitting the 5% tails and values presented for each species are hence a span of such data. Spore measurements include lobes but exclude echinuli and the hilar appendage. We did not measure abnormally large spores, originating from two-sterigmate basidia. Measurements of basidial width were made at the widest part of the tip of the basidia; basidial length excludes sterigmata. We obtained the width of hyphae from unbroken, internodal sections.
The spore measurements follow Kõljalg (1996) in the recognition of the dorsal side of basidiospores observed in face view as the frontal face ( Fig. 1) and the sides of the spores seen in side-view as the lateral faces. Great care has been taken to correctly identify these faces while conducting measurements and not tilted versions of the same, as inclusion of such would undoubtedly lead to an increased margin of error.
For optimal usage of the species descriptions in this article, we recommend readers to utilise the online version, where high resolution pictures are available.

Molecular data
We generated sequences from four regions for the study: the complete ITS region, including the 5.8S gene, and about 1200 bases of the 5´end of the LSU nuclear ribosomal DNA; about 600 bases of translation elongation factor subunit 1 alpha (Tef1α); and approximately 500 bases of the mtSSU. DNA extractions, PCR reactions and sequencing were performed as described in Larsson et al. (2018). The primers used to amplify the complete ITS region and the 5´end of the LSU region were ITS1F (Gardes and Bruns 1993), LR21, LR0R and LR7 (Hopple and Vilgalys 1999); for Tef1α we used EF983F and EF1567R (Rehner and Buckley 2005); and for mtSSU we used MS1 and MS2 (White et al. 1990). Primers used for sequencing were ITS1, ITS4, MS1, MS2 (White et al. 1990), Ctb6 (https://nature.berkeley.edu/brunslab/tour/primers. html), Lr5 (Hopple and Vilgalys 1999), EF983F and EF1567R.
The species described in this article have been provided with links to the UNITE Species Hypotheses they are part of, in the cases where such exist. Upon publication of the article, the Species Hypotheses will be updated with their new names and the DNA sequences generated for the article will be made available in GenBank. At the next update of UNITE, the ITS sequences will then be copied from GenBank and clustered into the appropriate Species Hypotheses.

Molecular analyses
We used SplitsTree 4.14.4 (Huson and Bryant 2006) to explore the amount of intragenic conflict and possible presence of intragenic recombination, and RDP4 (Martin et al. 2015) to test for recombination. In RDP, all DNA regions were initially submitted to testing with the methods RDP, GENECONV, Chimaera and MaxChi, with Bonferroni correction and 0.01 as the significance level. We submitted sequences with significant signs of recombination to a second round of testing that made use of all recombination methods. Sequences with a positive result for more than two methods with p-values ≤ 10 -5 in the second round were regarded as probable recombinants.
To generate Bayesian phylogenetic trees from the alignments, we used BEAST 2.4.7 (Bouckaert et al. 2014), employing the standard version of the programme for gene tree estimation, and STACEY 1.2.4 (Jones 2017) for species tree inference under the multispecies coalescent model. We prepared the xml-files for the BEAST 2 runs in BEAUti 2.4.7 (Bouckaert et al. 2014). The following minimal partitions were assumed per unlinked genetic region (Table 2): ITS1, 5.8S, ITS2, LSU (nrDNA); Tef1α first positions, Tef1α second positions, Tef1α third positions (Tef1α); mtSSU (mtSSU). We used the automated best-fit tests implemented in PAUP 4.0a (Swofford 2002) to select optimal substitution models and substitution model partitions for each minimal partition. Using three substitution schemes, the following partitions and models had the highest ranking, according to BIC scores: ITS1+ITS2 (HKY+I+G), 5.8S+LSU (GTR+I+G), Tef1α first positions (JC+I), Tef1α second positions (K80+I), Tef1α third positions (HKY) and mtSSU (GTR+G). The BEAST 2 and STACEY analyses did not converge under the GTR model and invariant site fraction parameter (I), however, such that we used HKY+G for the two nrDNA substitution model partitions and the mtSSU region. Omitting the I parameter in a rerun of the partition test for Tef1α yielded the result JC+G for first+second positions and HKY for the third positions and these were hence the partitions and models we adopted in the BEAST 2 analyses. Table 1. DNA regions included per analysis and collection. Accession numbers of DNA sequences generated for this study start with "MK" and for sequences obtained from UNITE with "UDB"; all other sequences were acquired from GenBank. Type collections are shown in boldface. The respective gene trees included all available sequences, whereas collections and accession numbers whose sequences were included in the STACEY analysis are marked with * and those included in both the STACEY and ASTRAL analyses with **.  The substitution rate of each partition was estimated independently of the others in each BEAST 2 run. We set all individuals as separate species in the STACEY analysis. We set the trees of the minimal nrDNA partitions as linked, as did we for the Tef1α minimal partitions. We set the clock models of all minimal partitions as unlinked and a lognormal, relaxed clock model was assumed for each, as test runs had shown that all partitions had a coefficient of variation well above 0.1 (i.e. implying a relatively high rate variation amongst branches). The clock rate of each partition was estimated in the runs, using a lognormal prior, with a mean set to one in real space. We set the growth rate prior to lognormal, with a mean of 5 and a standard deviation of 2. The Collapse Height prior of the STACEY analysis was set to 10 -5 and a lognormal prior with a mean of -7 and a standard deviation of 2 was set to the PopPriorScale parameter.
We ran the Markov Chain Monte Carlo (MCMC) chains of the mtSSU and Tef1α regions for 10 million generations with tree and parameter files sampled every 1000 generations. For the nrDNA and STACEY analysis, we ran the MCMC chains for 100 million generations and sampled it every 5000 generations, and for 1 billion generations and sampled it every 25000 generations, respectively. All analyses converged well in advance of the 10% burn-in threshold and had effective sampling size values well above 200 for all parameters. Chain mixing was found to be satisfactory as assessed in Tracer 1.6.0 (Rambaut et al. 2014). After discarding the burn-in trees, maximum clade credibility trees were identified by TreeAnnotator 2.4.7 (Bouckaert et al. 2014). Posterior probabilities of the clusterings of the species trees output by STACEY were analysed in the associated software SpeciesDelimitationAnalyser version 1.8.0, with burn-in set to 10%, simcutoff to 1 and collapseheight to 10 -5 . The estimated similarity matrix was visualised by the R-script plot.simmatrix.R (Jones et al. 2015).
To generate Maximum Likelihood (ML) gene trees, we used PhyML 3.1 (Guindon et al. 2010). We set the substitution model to GTR+I+G for the Tef1α, nrDNA and mtSSU regions, since it was the best-fit model output by the automated model test in PAUP, using AICc and three substitution schemes. The tree topology search was conducted using NNI+SPR, with ten random starting trees. Non-parametric bootstrap analyses with 1000 replicates were performed on the resulting trees. We also inferred a species tree from the ML gene trees, using ASTRAL III (Zhang et al. 2018), with node support calculated as local posterior probabilities (Sayyari and Mirarab 2016).
The Bayesian and ML gene trees comprised the entire nrDNA, mtSSU and Tef1α datasets of this study, while the STACEY and ASTRAL species trees, including the ML trees underlying the latter, contained subsets thereof (Table 1) to avoid destabilising the analyses with large amounts of missing data.
We visually prepared the resulting trees from the Bayesian and ML analyses in FigTree 1.4.3 (Rambaut 2012) and Inkscape (Albert et al. 2018).

Molecular species delimitation
The STACEY analysis retrieved 13 well-supported clades, based on DNA from specimens morphologically belonging to the P. tristis group. We interpret these as species (Fig.  2). The corresponding clades in the ASTRAL analysis were also supported, when present as more than one leaf node (Fig. 4). We found three of the delimited species, P. tristis, P. umbrina and P. atrofusca, to constitute previously described taxa, while nine species are described as new to science: P. sciastra, P. umbrinascens, P. pinophila, P. alnophila, P. alobata, P. pluriloba, P. abundiloba, P. rotundispora and P. media. We chose not to describe P. sp. 1, since the only collection available is too small to be suitable as a type. The same applies to P. sp. 2, which in addition was retrieved as a singleton by both analyses.
The delimitation of the species recognised in the species tree equals the clusters output by SpeciesDelimitationAnalyser (Fig. 3). As displayed by the similarity matrix, each cluster has internal support and zero posterior probability of the included individuals belonging to any other cluster. In the case of specimens SS419 and SS420 and sequences HQ850125, HM370480, UDB012458 and KF041350, however, the similarity matrix shows weak support for them to belong to their respective clusters; but given the strong support for the corresponding species in the species tree, we interpret these as instances of intraspecific genetic structure. In addition to the species delimited based on clades in the species tree, we recognised two species, P. tristoides and P. sp. 3, based on their presence as highly supported nodes in the nrDNA gene tree (Fig. 5). Their sequences could not be included in the species tree analyses, due to lack of data for the other genetic regions. We did not describe P. sp. 3, since it has no physical material tied to it.

Phylogenetic relationships
The species tree analyses were congruent. The trees retrieved show that the species in the P. tristis group, with the addition of Pseudotomentella rhizopunctata E.C.Martini & Hen- tic, form a monophyletic clade with high support (Figs 2, 4). Its two daughter clades, one containing P. rhizopunctata and P. atrofusca and the other containing the remaining species of the P. tristis group (the "core P. tristis group"), were also well supported. The phylogenetic relationships within the core P. tristis group, however, were not; only the clades (P. sciastra, P. alobata), ((P. sciastra, P. alobata), P. tristis) and (P. pinophila, P. pluriloba) were supported by both analyses. In addition, the clades ((P. rotundispora, P. sp. 1), P. umbrina) and (P. rotundispora, P. sp. 1) were supported by the STACEY analysis.
No signal of intragenic recombination was detected in RDP4, but as indicated by the low phylogenetic resolution present also in the gene trees (Fig. 5, Suppl material 1: Figs S1-S4) and the network-like structure between splits observed in SplitsTree (Suppl material 1: Figs S5-S7) for the included genetic regions, there is considerable intragenic conflict between species. The exception to this pattern is the highly sup- ported, long branch of the clades (((P. sciastra, P. tristoides), P. alobata), P. tristis) and ((P. sciastra, P. alobata), P. tristis) in the nrDNA and Tef1α trees, respectively. In the neighbour nets of the nrDNA and Tef1α regions, this clade is reminiscent of the "dogbone" shape displayed by paralogy, a hypothesis that is reinforced by its placement at the very root of the Tef1α tree. The branch in question thus constitutes the only -but in itself a major -incongruence between the gene trees.
A methodological observation to future users of STACEY with limited amounts of data is that support for species-level nodes decreases dramatically unless at least one included leaf taxon has a complete coverage of all the genetic regions used.

Type studies
All of the 13 newly described and previously described species can be distinguished morphologically (Table 3), although the differences exhibited by some species pairs are small. We found the previously designated lectotype of P. tristis and neotype of P. umbrina to fall within the morphological variation of sequenced material, with which they could hence be epitypified. The lectotypes of H. fuscata and H. sitnensis, however, display the morphological characteristics of P. tristis, with which we thus considered them conspecific. Of the seven type specimens studied, we were only able to generate ITS sequence data for P. atrofusca. The European collections, studied of P. atrofusca, all belong to P. sciastra.
The type collections of P. longisterigmata and H. rhacodium differ morphologically from all other specimens studied. Their aberrant morphology, with extremely long sterigmata and a very hard and thick basidiome, respectively (see further under "Taxonomy"), however, suggest that they may be misshapen forms of other species.
Septobasidium arachnoideum (Berk. & Broome) Bres. was accepted in Septobasidium by the thorough study of Couch (1938) and hence, we consider it excluded from Thelephorales Corner ex Oberw. We designated a plate by Bulliard (1790) as lectotype of A. phylacteris and T. biennis. The morphology displayed by this plate and stated in the original descriptions of these species does not match any Pseudotomentella species known to date (Larsen 1971a, Stalpers 1993, Kõljalg 1996.

Morphology
We were able to discern a few morphological patterns amongst the clades of the nrD-NA trees. The most pronounced is perhaps the lack of hyphal cords and skeletal hyphae in the species of the core P. tristis group. These are characters that are present in P. atrofusca and P. rhizopunctata (Larsen 1971a, Martini andHentic 2003) and indeed in all other simple-septate Pseudotomentella species (Larsen 1971a, Kõljalg 1996. Generally, spore shape and dimensions, the width of subicular hyphae and the length of echinuli proved to be the most taxonomically informative characters. Subhymenial hyphal width, basidial dimensions and sterigmal length were moderately useful for Table 3. The most taxonomically informative micromorphological characters. Summary statistics for each species is marked in bold. "Frontal" and "lateral" refer to the corresponding spore faces and "L" and "W" denote mean length and width, respectively. All measurements are in µm, with the 5% smallest and largest values denoted in brackets, when differing from the remaining 90%.  distinguishing between species, whereas the Q value (spore length/width) was considerably less so. The presence of wide subicular hyphae, a blue green reaction in the hymenium and subhymenium and amyloid material present in and on the same, in the species of the clade containing P. tristis, P. alobata, P. sciastra and P. tristoides, is also worthy of notice. These are, however, characters also present in species of other clades, e.g. (P. pinophila, P. pluriloba) and (P. umbrina, P. rotundispora), where they are not shared amongst all species included and may thus represent plesiomorphic characters. The specimens examined of P. sciastra display considerable morphological variation; the spore and subicular hyphal measurements of TAA187322 deviate markedly from those of SS359 and O F110317. Interestingly, P. sciastra is also more genetically variable than the other species studied.
For many species, we recorded a blue-green reaction of subhymenial hyphae, basidia, encrusting material and sometimes also of spores, to occur in KOH. We only observed the reaction occurring close to air bubbles or adjacent to the edges of cover glasses and we did not record it close to the centre of preparations free from air bubbles, unless they had been made slowly and hence had allowed air to come into contact with the entire samples before the application of a cover glass. The same structures often, but not always, also had an amyloid reaction in Melzer's reagent. Both the blue green and the amyloid reaction could be used as species-separating characters (see further under "Taxonomy"). When present, the encrustation was most prevalent on the bases of basidia, but common also on the upper part of subhymenial hyphae. Occasionally, it was also appearing on subicular hyphae.

Ecology and geographical distribution
We found the majority of the collections and sequences included in this study to belong to P. umbrina (Table 4). As shown by the origin of these ( Fig. 6), P. umbrina is distributed across at least 12 countries in Europe and North America, where it has been found growing with 18 different hosts. It is present in Arctic/alpine vegetation above the treeline as well as in coniferous, deciduous and mixed forests and has been encountered on soils with pH ranging from low to high.
The other species in the P. tristis group have been encountered markedly fewer times, with a smaller number of hosts, in less diverse habitats and mostly within a smaller geographical range. They have all been collected on soil with intermediate to high pH or both. Similarly to P. umbrina, however, many species seem to form ectomycorrhiza with a range of hosts; they have been collected on the root tips of both broadleaved and coniferous trees, as well as orchid species. Three species seem to have a limited host range: P. pinophila has only been found inhabiting the roots of Pinus L. species, while P. alnophila and P. sp. 2 have been found exclusively on the roots of Alnus Mill. Two species, meanwhile, now have a different confirmed geographical distribution than previously documented: the only verified sequences and basidiomata of P. tristis here studied originate in Europe, while P. atrofusca now have no confirmed findings there -the only validated findings are currently the Arizona holotype and three Chinese root tip sequences.

Taxonomy
We provide descriptions of ten species new to science and of previously described accepted, dubious and excluded species in the P. tristis group. A worldwide key to all recognised and dubious species is also presented.
Key to the species in the P. tristis group Pseudotomentella species with brownish spores and subicular hyphae, lacking clamps and chlamydospores. Description. Basidiomata annual, resupinate, membranaceous, effused to several tens of centimetres in diameter. Mature parts continuous, with a rather firm, fibrous and compact, yet quite soft and elastic texture. Hymenium smooth, but sometimes strongly undulating; brown with a pinkish hue. Immature parts discontinuous, byssoid, with a cottony texture. Subhymenium and hymenium of immature parts blue grey to brown grey. Subiculum well developed, loose, fibrous, orange brown; often forms the outer edge of basidiomata, extending noticeably beyond the hymenium. All characters recorded in dried state.
Hyphal cords lacking, but loose bundles of subicular hyphae sometimes present. Hyphal system monomitic; clamp connections and reaction in Melzer's reagent absent from all hyphae. Subicular hyphae noticeably long and straight, thick-walled; forming a loose tissue. Individual hyphae (4.3-) 4.8-6.9 (-7.2) µm wide, with a mean width of 5.5-6.1 µm; orange brown to dark brown in both KOH and water.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae (4.0-) 4.1-6.8 (-7.2) µm wide, with a mean width of 5.5-5.7 µm; in the upper parts, hyaline to orange brown or orange green in KOH, with a blue green reaction in the presence of air; in the lower parts, pale orange brown to orange brown in KOH, unchanged in air; in water with strongly granular contents, orange green.
Encrustation granular, inamyloid; hyaline to orange brown or orange green in KOH, blue green in the presence of air; orange green in water; common to rare, usually scattered in occurrence on the upper parts of subhymenial hyphae and on the lower parts of basidia.
Habitat. The type collection was obtained in an old, mixed forest on soil with high pH. No additional sequences are available in UNITE.
Distribution. Basidiomata encountered in: Estonia and Norway.
Remarks. Within the P. tristis group, the basidiomata of P. abundiloba are recognised by their lack of hyphal cords and skeletal hyphae and their soft, yet rather firm and compact and ± elastic texture after drying, bluish to greenish colour of immature parts, wide subicular hyphae, long, abundantly lobed spores and inamyloid encrustation on subhymenial hyphae and basidia. Pseudotomentella abundiloba, P. pluriloba and P. media can appear similar, but none of them has abundantly lobed spores. Pseudotomentella media further differs by having smaller spores and narrower subicular hyphae, while P. pluriloba has narrower subicular hyphae, longer sterigmata and frontally wider spores and P. alobata has amyloid encrustation on its subhymenial hyphae and basidia.

Pseudotomentella alnophila
Description. Basidiomata annual, resupinate, membranaceous, effused. Mature parts continuous, with a soft cottony texture. Hymenium smooth; blue grey, sometimes with a slightly brown hue. Immature parts discontinuous, byssoid, with a soft cottony texture. Subhymenium and hymenium of immature parts pale blue grey to blue grey. Subiculum thin to well developed, loose, fibrous, orange brown; often forms the outer edge of basidiomata, extending noticeably beyond the hymenium. All characters recorded in dried state.
Hyphal cords lacking, but loose bundles of subicular hyphae sometimes present. Hyphal system monomitic, clamp connections and reaction in Melzer's reagent absent from all hyphae. Subicular hyphae noticeably long and straight, thick-walled; forming a loose tissue. Individual hyphae 4.0-5.0 (-5.1) µm wide, with a mean width of 4.5 µm; orange brown to dark brown in KOH and orange to orange brown in water.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae (3.2-) 3.4-5 (-5.6) µm wide, with a mean width of 4.1 µm; hyaline to pale orange brown in KOH, blue green in the presence of air; pale green in water, with strongly granular contents.
Habitat. The only specimens recorded to date of P. alnophila is the type collection and one other collection from the same locality, which is a mature and, at the collection site pure, stand of Alnus incana on clay soil with intermediate pH. In addition, UNITE sequence metadata show that the species forms ectomycorrhiza with at least Alnus mandschurica (Kõljalg et al. 2005, Nilsson et al. 2019.
Distribution. Basidiomata encountered in: Norway. Soil or root tip samples confirm presence also in: China.
Remarks. Within the P. tristis group, the basidiomata of P. alnophila can be recognised by their lack of hyphal cords and skeletal hyphae and their soft cottony texture after drying, bluish to greenish colour of immature parts, narrow hyphae, long spores, bluish-grey mature hymenium (sometimes with a slightly brown hue) and their association with Alnus. Pseudotomentella pluriloba, P. media and P. pinophila are similar, but they all have basidiomata that are compact and rather firm after drying and whose mature parts are some shade of brown, without any bluish hue. Pseudotomentella pluriloba also has slightly longer spores and echinuli and wider subicular hyphae, while P. media and P. pinophila have generally slightly smaller microcharacters. Pseudotomentella pinophila also has a different spore shape. In addition, neither of these species has been recorded as being associated with Alnus. Description. Basidiomata annual, resupinate, membranaceous, effused -often to several tens of centimetres in diameter. Mature parts continuous, with a cottony texture when fresh and a rather firm, fibrous and compact, yet quite soft and elastic texture when dried. Hymenium smooth, but sometimes strongly undulating; brown, purplish-brown or blue-greyish-brown when fresh, brown with a pinkish hue when dried. Immature parts discontinuous, byssoid, with a cottony texture both when fresh and when dried. Subhymenium and hymenium of immature parts blue to blue grey when fresh and blue grey to brown grey when dried. Subiculum well developed, loose, fibrous, orange brown; often forms the outer edge of basidiomata, extending noticeably beyond the hymenium.
Hyphal cords lacking, but loose bundles of subicular hyphae sometimes present.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae (3.1-) 3.4-6.9 µm wide, with a mean width of 4.0-4.5 µm; hyaline to pale green in KOH, blue green in the presence of air; yellow to pale orange yellow in water, with strongly granular contents.
Encrustation granular, amyloid; purple in KOH, dark blue green in the presence of air; dark brown in water; usually common and scattered in occurrence on the upper parts of subhymenial hyphae and on the lower parts of basidia.
Habitat. Data on habitat are scarce to date, but recent Scandinavian collections have been made in old growth coniferous or mixed forests on soil with high pH.
Distribution. Basidiomata encountered in: Norway, Slovenia and Sweden. No sequences originating from soil or root tip samples in UNITE.
Remarks. Within the P. tristis group, the basidiomata of P. alobata are recognised by their lack of hyphal cords and skeletal hyphae and their soft, yet rather firm and compact and ± elastic texture after drying, bluish to greenish colour of immature parts, wide subicular hyphae, long, unlobed spores and amyloid encrustation on subhymenial hyphae and basidia. Pseudotomentella abundiloba, P. pluriloba and P. media can appear similar, but none of them has spores which generally are unlobed. P. media further differs by having smaller spores and narrower subicular hyphae, while P. pluriloba has narrower subicular hyphae, longer sterigmata and frontally wider spores than P. alobata. Pseudotomentella abundiloba sometimes has encrusted subhymenial hyphae and basidia, but without amyloid reaction.  UNITE SH. SH005338.07FU Description. Basidiome annual, resupinate, membranaceous, effused. Mature parts continuous, with a cottony texture. Hymenium smooth, brown. Immature parts discon-tinuous, byssoid, with a cottony texture. Subhymenium and hymenium of immature parts initially whitish-grey to whitish-grey brown, when more mature blue grey to brown grey. Subiculum thin, loose, fibrous, pale brown; often forms the outer edge of the basidiome, extending noticeably beyond the hymenium. All characters recorded in dried state.
Hyphal cords connecting to the edge of the basidiome and thinning out underneath; whitish to pale brown. Individual cords dimitic; formed by a sheathing layer of skeletal hyphae and two layers of generative hyphae; the outer generative hyphae thinner and the inner ones wider, the latter often swollen interseptally. Skeletal hyphae 1.1-1.4 (-1.5) µm wide, with a mean width of 1.3 µm. Outer generative hyphae (2.2) 2.3-2.9 µm wide, with a mean width of 2.6 µm. Inner generative hyphae (3.8) 3.9-5.3 (5.5) µm wide, with a mean width of 4.5 µm. All hyphae pale yellowish-brown in both KOH and water.
Hyphal system dimitic, clamp connections and reaction in Melzer's reagent absent from all hyphae.
Subicular hyphae of two kinds: (1) generative hyphae noticeably long and straight, thick-walled; forming a loose tissue, in which (2) skeletal hyphae occur sparsely (most common in areas to where hyphal cords attach). Generative hyphae (1.7-) 1.8-2.8 µm wide, with a mean width of 2.3 µm; pale yellowish-brown to yellowish-brown in both KOH and water. Skeletal hyphae with the same width and colour as in the hyphal cords.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae (1.8-) 2.1-3.0 µm wide, with a mean width of 2.5 µm; hyaline to pale green in KOH, blue green in the presence of air; hyaline to pale blue green in water, with strongly granular contents.
Chlamydospores lacking. Habitat. The only specimen recorded to date of P. atrofusca is the type collection, which was collected on wood of Pinus ponderosa. Available UNITE sequences originate from root tips of Rhododendron decorum (Kõljalg et al. 2005, Nilsson et al. 2019.
Distribution. Basidiome encountered in: United States. Soil or root tip samples confirm presence also in: China.
Remarks. Within the P. tristis group, the basidiome of P. atrofusca can be recognised by its hyphal cords and skeletal hyphae. These features make it unique within the group and the risk for confusion with any other described species should be small. Outside the P. tristis group, P. rhizopunctata is somewhat similar, but differs from P. atrofusca by the presence of chlamydospores on its hyphal cords.

UNITE SH. SH005336.07FU
Etymology. The name refers to the middling morphological characters of the species, relative to other species in the P. tristis group.
Description. Basidiomata annual, resupinate, membranaceous, effused to approximately ten centimetres in diameter. Mature parts continuous, with a rather firm, fibrous and compact, yet quite soft and elastic texture when dried. Hymenium smooth, but sometimes strongly undulating; brown with a reddish hue, both when fresh and when dried. Immature parts discontinuous, byssoid, with a cottony texture when dried. Subhymenium and hymenium of immature parts blue to greenish-blue when fresh and pale blue grey or blue grey to grey brown when dried. Subiculum well developed, loose, fibrous, pale brown to pale orange brown; forms the outer edge of basidiomata, extending noticeably beyond the hymenium.
Hyphal cords lacking, but loose bundles of subicular hyphae sometimes present.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae (3.2-) 3.4-5.6 (-5.7) µm wide, with a mean width of 4.3-4.4 µm; pale brown, pale orange brown, pale greenish-brown or hyaline in KOH, blue green in the presence of air; orange green to brown in water, with strongly granular contents. Some subhymenial hyphae with a pink colour in water and an amyloid reaction.
Encrustation granular, probably amyloid (hard to observe due to the colour); blackish in KOH, dark blue green in the presence of air; blackish in water; scattered in occurrence on the upper parts of subhymenial hyphae and on the lower parts of basidia.
Habitat. P. media has been found to form ectomycorrhiza with at least Betula pendula, Larix decidua and Picea glauca (Kõljalg et al. 2005, Nilsson et al. 2019.
Distribution. Basidiomata encountered in: Estonia. Soil or root tip samples confirm presence also in: Canada, Italy and Russia.

Remarks.
Within the P. tristis group, the basidiomata of P. media can be recognised by their lack of hyphal cords and skeletal hyphae and their soft, yet rather firm and compact and ± elastic texture after drying, bluish to greenish colour of immature parts, narrow subicular hyphae, brown mature hymenium, long, laterally wide, angular-nodulose spores and subhymenial hyphae that are of ± the same width as the subicular hyphae. Pseudotomentella pinophila and P. pluriloba can appear similar, but the spores of P. pinophila are laterally narrower and generally star-shaped, while P. pluriloba has wider subicular hyphae, larger spores and subhymenial hyphae that are noticeably narrower than the subicular hyphae.

Pseudotomentella pinophila
Description. Basidiomata annual, resupinate, membranaceous, effused -often to several tens of centimetres in diameter. Mature parts continuous, with a cottony texture when fresh and a rather firm, fibrous and compact, yet quite soft and elastic texture when dried. Hymenium smooth, but sometimes strongly undulating; pale brown to pale greenish-brown when fresh, pale reddish-brown when dried. Immature parts discontinuous, byssoid with a cottony texture, both when fresh and when dried. Subhymenium and hymenium of immature parts blue to grey when fresh, pale blue or blue grey to dark blue grey or brown grey, sometimes with a green hue, when dried. Subiculum well-developed, loose, fibrous, pale orange brown; often forms the outer edge of basidiomata, extending noticeably beyond the hymenium.
Hyphal cords lacking, but loose bundles of subicular hyphae sometimes present. Hyphal system monomitic, clamp connections absent from all hyphae. Subicular hyphae noticeably long and straight, thick-walled; forming a loose tissue. Individual hyphae 3.0-4.9 µm wide, with a mean width of 3.6-4.1 µm; pale orange brown to pale pinkish-brown in both KOH and water.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae (3.1-) 3.2-4.7 (-5.2) µm wide, with a mean width of 3.9-4.0 µm; hyaline to pale green or occasionally pale brown in KOH, blue green in the presence of air; pale green to pale orange in water, with strongly granular contents; sometimes with an amyloid reaction in the cell walls. Encrustation granular, amyloid; bluish-black in both KOH and water; common to rare, usually scattered in occurrence on the upper parts of subhymenial hyphae and on the lower parts of basidia.
Habitat. Data on habitat are scarce to date, but recent Scandinavian collections have been made in old coniferous or mixed forests on soil with high pH. Pseudotomentella pinophila has been found to form ectomycorrhiza with at least Pinus densiflora, Pinus massoniana, Pinus sylvestris and Pinus thunbergii (Kõljalg et al. 2005, Nilsson et al. 2019). It should be noted however that, although the only hitherto documented hosts of P. pinophila belong to the genus Pinus and P. sylvestris has indeed been present at nearly all Nordic localities of collection, a few of these collections were made at localities where Pinus was not recorded as a possible host.
Distribution. Basidiomata encountered in: Norway and Sweden. Soil or root tip samples confirm presence also in: China and Republic of Korea.
Remarks. Within the P. tristis group, the basidiomata of P. pinophila can be recognised by their lack of hyphal cords and skeletal hyphae and their soft, yet rather firm and compact and ± elastic texture after drying, bluish to greenish colour of immature parts, narrow subicular hyphae, brown mature hymenium, long, laterally narrow and commonly star-shaped spores. Pseudotomentella sciastra, P. pluriloba and P. media can appear similar. Even though P. sciastra has star-shaped spores, it also has wider subicular hyphae than P. pinophila and, while P. pluriloba and P. media both share the characters of narrow hyphae, long spores and hymenia that are brown when mature with P. pinophila, they differ by having angular-nodulose spores, which are also laterally wider than the spores of P. pinophila.
Additional specimens studied. NORWAY. Akershus: Asker, Skaugumåsen, boreonemoral, mixed forest on on soil with high pH, 23 September 2010, S. Svantesson UNITE SH. SH030565.07FU Etymology. The name refers to the several lobes of the spores. Description. Basidiomata annual, resupinate, membranaceous, effused to approximately ten centimetres in diameter. Mature parts continuous, with a cottony texture when fresh and a rather firm, fibrous and compact, yet quite soft and elastic texture when dried. Hymenium smooth, but sometimes strongly undulating; brown to purplish-brown when fresh, reddish-brown when dried. Immature parts discontinuous, byssoid with a cottony texture, both when fresh and when dried. Subhymenium and hymenium of immature parts blue when fresh, blue grey to brown grey after drying. Subiculum well developed, loose, fibrous, brown with an orange hue; forms the outer edge of basidiomata, extending noticeably beyond the hymenium.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae (2.7-) 2.9-5.3 (-5.4) µm wide, with a mean width of 4.0-4.2 µm; pale orange green to hyaline in KOH, blue green in the presence of air; pale orange green to hyaline in water, with strongly granular contents.
Encrustation granular, amyloid, concolourous with the hyphae in both KOH and water; usually common and scattered in occurrence on the upper parts of subhymenial hyphae and on the lower parts of basidia.
Habitat. Data on habitat are scarce to date, but recent Scandinavian collections have been made in mature to old coniferous or mixed forests on soil with intermediate pH. Pseudotomentella pluriloba has been found to form ectomycorrhiza with at least Pseudotsuga menziesii (Kõljalg et al. 2005, Nilsson et al. 2019.
Distribution. Basidiomata encountered in: Finland and Sweden. Soil or root tip samples confirm presence also in: Canada and the United States.
Remarks. Within the P. tristis group, the basidiomata of P. pluriloba are recognised by their lack of hyphal cords and skeletal hyphae and their soft, yet rather firm and compact and ± elastic texture after drying, bluish to greenish colour of immature parts, wide subicular hyphae and noticeably narrower subhymenial hyphae, long, moderately lobed spores and amyloid encrustation on subhymenial hyphae and basidia. Pseudotomentella abundiloba, P. alobata and P. media can appear similar, but P. media differs by having smaller spores and narrower subicular hyphae which are ± the same width as its subicular hyphae, while P. abundiloba and P. alobata have frontally narrower spores with different lobation than P. pluriloba, as well as wider subicular hyphae and shorter sterigmata.  UNITE SH. SH030562.07FU Etymology. The name refers to the appearance of the spores, which often have rounded or weakly pronounced lobes and comparably short echinuli.
Description. Basidiomata annual, resupinate, membranaceous, effused -often to several tens of centimetres in diameter. Mature parts continuous, with a cottony texture when fresh and a rather firm, fibrous and compact, yet quite soft and elastic texture when dried. Hymenium smooth, but sometimes strongly undulating; brown to greenish-brown when fresh, concolourous when dried, but then sometimes with a red hue. Immature parts discontinuous, byssoid with a cottony texture, both when fresh and when dried. Subhymenium and hymenium of immature parts blue to blue green or grey when fresh, pale blue grey to grey blue when dried. Subiculum well developed, loose, fibrous, pale brown to pale orange brown; often forms the outer edge of basidiomata, extending noticeably beyond the hymenium.
Hyphal cords lacking, but loose bundles of subicular hyphae sometimes present.
Hyphal system monomitic, clamp connections absent from all hyphae. Subicular hyphae noticeably long and straight, thick-walled; forming a loose tissue. Individual hyphae 3.0-4.4 (-4.6) µm wide, with a mean width of 3.4-3.8 µm; brown to orange brown in both KOH and water.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae (2.7-) 3.0-4.6 (-6.3) µm wide, with a mean width of 3.6-3.7 µm; pale brown to brown in KOH, often with orange or green hues, blue green in the presence of air; brown to orange brown in water, with strongly granular contents; some subhymenial hyphae with a pink colour in water and an amyloid reaction in Melzer's reagent.
Encrustation granular, probably amyloid (hard to observe due to the colour); blackish in KOH, dark blue green in the presence of air; blackish in water; scattered in occurrence on the upper parts of subhymenial hyphae and on the lower parts of basidia.
Colours and reactions the same as for the upper parts of subhymenial hyphae, but in addition often with granular contents in KOH.
Habitat. Data on habitat are scarce to date, but recent Scandinavian collections have been made in old, coniferous, deciduous or mixed forests on soil with high pH. Pseudotomentella rotundispora has been found to form ectomycorrhiza with at least Castanea sp. and Populus tremula (Kõljalg et al. 2005, Nilsson et al. 2019.
Distribution. Basidiomata encountered in: Estonia, Norway and Sweden. Soil or root tip samples confirm presence also in: Austria, Italy and the United Kingdom.
Remarks. Within the P. tristis group, the basidiomata of P. rotundispora can be recognised by their lack of hyphal cords and skeletal hyphae and their soft, yet rather firm and compact and ± elastic texture after drying, bluish to greenish colour of immature parts, narrow subicular hyphae and short spores. The other species within the group can appear similar, but have either wider hyphae, longer spores or both.
Additional UNITE SH. SH030554.07FU Etymology. The name refers to the dark, star-like appearance of the spores. Description. Basidiomata annual, resupinate, membranaceous, effused -often to several tens of centimetres in diameter. Mature parts continuous, with a cottony texture when fresh and a rather firm, fibrous and compact, yet quite soft and elastic texture when dried. Hymenium smooth, but sometimes strongly undulating; blue grey when fresh and brown with a pinkish hue when dried. Immature parts discontinuous, byssoid with a cottony texture, both when fresh and when dried. Subhymenium and hymenium of immature parts blue grey when fresh or occasionally green or even yellow; blue grey to brown grey when dried. Subiculum well developed, loose, fibrous, pale orange brown to brown; often forms the outer edge of basidiomata, extending noticeably beyond the hymenium.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae 2.9-5.0 (-6.0) µm wide, with a mean width of 3.8-4.0 µm; hyaline to pale green in KOH, blue green in the presence of air; pale orange green to pale yellowish-green in water, with strongly granular contents.
Encrustation granular, probably amyloid (hard to observe due to the colour); blackish-brown in KOH, dark blue green in the presence of air; blackish-brown in water; scattered in occurrence on the upper parts of subhymenial hyphae and on the lower parts of basidia.
Habitat. Recent Scandinavian collections have been made in mature to old coniferous, deciduous or mixed forests on soil with intermediate to high pH. Pseudotomentella sciastra has been found to form ectomycorrhiza with at least Castanea sativa, Cedrus libani, Neottia ovata, Picea abies and Quercus sp. (Kõljalg et al. 2005, Nilsson et al. 2019. Distribution. Basidiomata encountered in: Estonia, Finland, Norway, Sweden, Turkey and the United Kingdom. Soil or root tip samples confirm presence also in: the Czech Republic, Mexico, Portugal (Madeira) and the United States.
Remarks. All studied European specimens previously identified as P. atrofusca belong to P. sciastra. The two species display considerable morphological differences (see key).
Within the P. tristis group, the basidiomata of P. sciastra are recognised by their lack of hyphal cords and skeletal hyphae and their soft, yet rather firm and compact and ± elastic texture after drying, bluish to greenish colour of immature parts, wide subicular hyphae and small, star-shaped spores. Pseudotomentella pinophila is similar, but has narrower subicular hyphae and larger spores. UNITE SH. SH030560.07FU Description. Basidiomata annual, resupinate, membranaceous, effused -often to several tens of centimetres in diameter. Mature parts continuous, with a cottony texture when fresh and a rather firm, fibrous and compact, yet quite soft and elastic texture when dried. Hymenium smooth, but sometimes strongly undulating; blue grey to purplishbrown when fresh, blue grey or blue-greenish grey to brown, with a reddish hue, when dried. Immature parts discontinuous, byssoid with a cottony texture, both when fresh and when dried. Subhymenium and hymenium of immature parts blue green, blue or blue grey when fresh and pale grey blue or pale blue grey to grey blue or blue grey when dried, sometimes with a green hue. Subiculum well developed, loose, fibrous, orange brown; often forms the outer edge of basidiomata, extending noticeably beyond the hymenium.
Hyphal cords lacking, but loose bundles of subicular hyphae sometimes present.
Hyphal system monomitic, clamp connections absent from all hyphae. Subicular hyphae noticeably long and straight, thick-walled; forming a loose tissue. Individual hyphae (4.5) 4.6-7.4 µm wide, with a mean width of 5.4-6.2 µm; orange brown to dark brown in KOH, orange brown to brown in water.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae 3.2-6.2 (7.2) µm wide, with a mean width of 3.9-4.5 µm; pale orange brown to pale green in KOH, blue green in the presence of air; pale green to pale greenish-orange in water, with strongly granular contents.
Encrustation granular, probably amyloid (hard to observe due to the colour); blackish in KOH, dark blue green in the presence of air; blackish in water; scattered in occurrence on the upper parts of subhymenial hyphae and on the lower parts of basidia.
Habitat. Data on habitat are scarce to date, but recent Scandinavian collections have been made in mature to old deciduous or mixed forests on soil with intermediate to high pH. Pseudotomentella tristis has been found to form ectomycorrhiza with at least Betula pendula and Fagus sylvatica (Kõljalg et al. 2005, Nilsson et al. 2019.
Distribution. Basidiomata encountered in: Estonia, Finland, Norway, Slovakia, Slovenia and Sweden. Soil or root tip samples confirm presence also in: Germany.
Remarks. We here select a Swedish specimen to serve as an epitype for both P. tristis and H. fuscata. This decision is based on four reasons: first, the present study has found P. tristis and H. fuscata to be conspecific; secondly, the lectotypes of these species were both collected in Finland (within the same county); thirdly we have found P. tristis to occur at several localities in both Finland and Sweden; and fourthly, the Swedish material chosen is both ampler and displays more variation with respect to maturity of the basidiome than the single available recent Finnish collection.
The type specimen of H. sitnensis was collected in Slovakia, i.e. far from the type locality of P. tristis. It displays the morphological characters of P. tristis, apart from the absence of an amyloid reaction in the encrusting material found on basidia and subhymenial hyphae. This might be an artefact of, for example, its drying conditions, time or intraspecific variation, but since the specimen is in too poor a condition to allow DNA sequencing with currently available methods, this cannot be ascertained. We therefore consider it a synonym of P. tristis and suggest it be epitypified in due course with locally sampled material that matches the type.
In the case of H. sitnensis, there is only one collection matching the locality and habitat description of the protologue as well as the collector stated. It predates the publication of the species. This collection must hence be regarded as a holotype. P. A. Karsten 770 is the lectotype of H. fuscata, as designated by Larsen in Nova Hedwigia (1971), but his note has been placed in P. A. Karsten 769, which has created confusion amongst mycologists studying these specimens. Mature spores that fall within the morphological span of P. tristis can easily be found in all collections that match Karsten's (1889) description of the species, with respect to locality and date. It would hence seem that the smooth, small, bluish spores he writes of in the protologue (see Introduction) probably were immature ones, studied in the presence of air.
Within the P. tristis group, basidiomata of P. tristis itself can be recognised by their lack of hyphal cords and skeletal hyphae and their soft, rather elastic texture after drying, bluish to greenish colour of immature parts, wide subicular hyphae, medium sized, commonly angular to nodulose spores and relatively long echinuli and sterigmata. Pseudotomentella tristoides is similar but has shorter echinuli and sterigmata, P. sciastra has smaller, star-shaped spores and H. rhacodium (only known from the type) has hard, brittle basidiomata after drying.
Additional The name refers to the overall similarity between this species and P. tristis. Description. Basidiome annual, resupinate, membranaceous, effused to approximately ten centimetres in diameter. Mature parts continuous, with a rather firm, fibrous and compact, yet quite soft and elastic texture. Hymenium smooth; brown with a reddish hue. Immature parts discontinuous, byssoid with a cottony texture. Subhymenium and hymenium of immature parts initially pale greyish-blue, when more mature dark blue grey. Subiculum well-developed, loose, fibrous, brown with an orange hue; forms the outer edge of the basidiome, extending noticeably beyond the hymenium. All characters recorded in dried state.
Hyphal cords lacking, but loose bundles of subicular hyphae sometimes present.
Hyphal system monomitic, clamp connections absent from all hyphae. Subicular hyphae noticeably long and straight, thick-walled; forming a loose tissue. Individual hyphae (4.7-) 4.9-7.1 (-7.6) µm wide, with a mean width of 6.0 µm; orange brown to dark brown in KOH, orange brown to brown in water.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae (3.1-) 3.2-5.3 (-5.4) µm wide, with a mean width of 4.6 µm; pale yellowish-brown in KOH, pale green to blue green in the presence of air; pale green to pale orange green in water, with strongly granular contents.
Encrustation granular, amyloid, concolourous with the hyphae in both KOH and water; scattered in occurrence on the upper parts of subhymenial hyphae and on the lower parts of basidia.
Habitat. The only specimen recorded to date of P. tristoides is the type collection, which was obtained in an old, mixed forest on soil with intermediate pH. UNITE sequence metadata show that the species forms ectomycorrhiza with at least Populus alba and Cephalanthera damasonium (Kõljalg et al. 2005, Nilsson et al. 2019.
Distribution. Basidiomata encountered in: Norway. Soil or root tip samples confirm presence also in: Estonia and the Czech Republic.
Remarks. Within the P. tristis group, the basidiome of P. tristoides can be recognised by its lack of hyphal cords and skeletal hyphae and its soft, yet rather firm and compact and ± elastic texture after drying, bluish to greenish colour of immature parts, wide subicular hyphae, medium sized, angular-nodulose spores and relatively short echinuli and sterigmata. Pseudotomentella tristis is similar but has longer echinuli and sterigmata, P. sciastra has smaller, star-shaped spores and H. rhacodium (only known from the type) has hard, brittle basidiomata after drying.  (1888) and Burt (1916)]. Corticium umbrinum (Fr.) Fr., Hymenomyc. eur.: 658 (1874  texture when fresh and a rather firm, fibrous and compact, yet quite soft and elastic texture when dried. Hymenium smooth, but sometimes strongly undulating; blue grey or purplish-grey to pale brown or brown when fresh, pale brown to brown when dried, sometimes with a reddish or greyish hue. Immature parts discontinuous, byssoid with a cottony texture, both when fresh and when dried. Subhymenium and hymenium of immature parts pale blue grey or pale purplish-grey to pale brown when fresh, pale brown when dried. Subiculum well developed, loose, fibrous, orange brown; often forms the outer edge of basidiomata, extending noticeably beyond the hymenium.
Hyphal cords lacking, but loose bundles of subicular hyphae sometimes present.
Hyphal system monomitic, clamp connections and reaction in Melzer's reagent absent from all hyphae. Subicular hyphae noticeably long and straight, thick-walled; forming a loose tissue. Individual hyphae 3.3-4.8 (-5.3) µm wide, with a mean width of 4.0-4.3 µm; orange brown to brown in KOH, orange brown in water.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae (2.8-) 2.9-5.0 (-5.9) µm wide, with a mean width of 3.7-4.2 µm; in the upper parts, pale green in KOH, sometimes with a faintly blue or brown hue; in the lower parts, orange brown to brown; in water, orange brown, with strongly granular contents.
The material in the Fries Herbarium cited by Burt (1916), as the type of Thelephora umbrina Fr., constitutes a collection made by Fries at locus classicus, Femsjö, but in Fries's own handwriting, it is identified as Corticium umbrinum (Fr.) Fr., a name he combined T. umbrina to in 1874 (Fries 1874). Therefore, Burt's typification cannot be considered a lectotype, but must be regarded as a neotype. We here designate an epitype from Femsjö, which matches the neotype morphologically.
Within the P. tristis group, basidiomata of P. umbrina can be recognised by their brown colour -blue or green colours are completely absent from immature parts and Pseudotomentella umbrinascens Svantesson, sp. nov. MycoBank No.: MB829031 Fig. 19 Type. SWEDEN. Bohuslän: Tanum (municipality), Tanum (parish), Greby Kleva, boreonemoral, deciduous forest on soil with high pH, RT90: E1236840, N6518916, 6 September 2016, S. Svantesson 335 (holotype: GB!, GenBank Acc. No. ITS: MK290697) UNITE SH. SH030563.07FU Etymology. The name refers to the overall morphological similarity to P. umbrina. Description. Basidiome annual, resupinate, membranaceous, effused to approximately ten centimetres in diameter. Mature parts continuous, with a cottony texture when fresh and a rather firm, fibrous and compact, yet quite soft and elastic texture when dried. Hymenium smooth; greenish-brown when fresh, pale brown when dried. Immature parts discontinuous, byssoid with a cottony texture, both when fresh and when dried. Subhymenium and hymenium of the immature parts initially yellowishwhite to pale brown, in the dried basidiome, when more mature pale brown. Subiculum well developed, loose, fibrous, pale yellowish-brown to pale orange brown; forms the outer edge of the basidiome, extending noticeably beyond the hymenium.
Hyphal cords lacking, but loose bundles of subicular hyphae sometimes present.
Hyphal system monomitic, clamp connections and reaction in Melzer's reagent absent from all hyphae. Subicular hyphae noticeably long and straight, thick-walled; forming a loose tissue. Individual hyphae 3.1-) 3.2-4.3 (-4.8) µm wide, with a mean width of 3.7 µm; pale orange brown to brown in KOH, orange brown in water.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae 3.6-5.7 (-7.1) µm wide, with a mean width of 4.5 µm; pale grey brown to grey brown or brown in KOH; orange brown to pale green in water (but not with the blue green reaction present in other species), with strongly granular contents.
Habitat. The only specimen recorded to date of P. umbrinascens is the type collection, which was obtained in an old, coastal, deciduous forest on soil with high pH. UNITE sequence metadata shows that the species forms ectomycorrhiza with at least Corylus avellana (Kõljalg et al. 2005, Nilsson et al. 2019.
Distribution. Basidiomata encountered in: Sweden. Soil or root tip samples confirm presence also in: Italy.
Remarks. Within the P. tristis group, basidiomata of P. umbrinascens can be recognised by their brown colour, their soft, rather elastic texture after drying and their microcharacters. Blue or green colours are completely absent from immature parts and from the subhymenium of mature parts. Pseudotomentella umbrina is very similar but has slightly different microcharacters (see key). Hypochnus rhacodium (only known from the type) is also similar but has basidiomata that are hard and brittle after drying. Description. Basidiome annual, resupinate, membranaceous, effused to approximately ten centimetres in diameter. Mature parts continuous, with a cottony to rather firm, fibrous and compact, yet quite soft and elastic texture. Hymenium smooth; bluish-grey to brownish-grey. Immature parts discontinuous, byssoid with a cottony texture. Subhymenium and hymenium of immature parts bluish-grey. Subiculum well-developed, loose, fibrous, orange brown; forms the outer edge of the basidiome, extending noticeably beyond the hymenium. All characters recorded in dried state.
Hyphal system monomitic, clamp connections absent from all hyphae. Subicular hyphae noticeably long and straight, thick-walled; forming a loose tissue. Individual hyphae 4.9-7.2 µm wide, with a mean width of 6.2 µm; orange brown to brown in KOH, orange to orange brown in water.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae (3.0-) 3.2-4.9 (-6.1) µm wide, with a mean width of 3.9 µm; pale brownish-green in KOH, blue green in the presence of air; brownish-green in water, with strongly granular contents.
Encrustation granular, probably amyloid (hard to observe due to the colour); dark brownish-green in KOH, dark blue green in the presence of air; blackish in water; scattered in occurrence on the upper parts of subhymenial hyphae and on the lower parts of basidia.
Habitat. The only specimen of P. longisterigmata recorded to date is the type collection, which was collected on coniferous wood in a coastal forest in the state of Washington, United States.
Distribution. Basidiomata encountered in: the United States.

Remarks.
The type collection is large and in seemingly good condition but repeated attempts at obtaining a useful DNA sequence from it proved unfruitful. The specimen exhibits a peculiar morphology, where the basidia carry sterigmata that are unusually long for the P. tristis group. They are often cylindrical rather than tapering and hence resemble generative hyphae -a growth form that basidia are sometimes seen reverting into in corticioid basidiomata formed under unfavourable conditions. It is therefore doubtful whether P. longisterigmata is a true species, but since this presently cannot be ascertained and, in order to stimulate its recollection, the name is here retained as a separate species.
There are relatively few spores in the hymenium of the holotype and many of them are immature. A more mature fruiting body of the species would hence probably have a browner colour.
Within the P. tristis group, the basidiome of P. longisterigmata can be recognised by its lack of hyphal cords and skeletal hyphae, its soft, yet rather firm and compact and ± elastic texture after drying, bluish to greenish colour of immature parts, wide subicular hyphae, large spores and long sterigmata. Pseudotomentella alobata, P. pluriloba and P. abundiloba are similar but all have smaller spores and shorter sterigmata.  Burt in Ann. Missouri Bot. Gard. 13: 322 (1926)).

Hypochnus rhacodium
Description. Basidiome annual, resupinate, membranaceous, effused. Mature parts continuous, with a hard and rather brittle texture. Hymenium smooth; pale brown to brown with a pink hue. Subiculum loose, fibrous and dark brown with an orange hue. All characters recorded in dried state.
Hyphal cords lacking, but loose bundles of subicular hyphae sometimes present.
Hyphal system monomitic, clamp connections and reaction in Melzer's reagent absent from all hyphae. Subicular hyphae noticeably long and straight, thick-walled; forming a loose tissue. Individual hyphae (5.6-) 5.7-7.3 (-8.0) µm wide, with a mean width of 6.5 µm; pale to dark orange brown in both KOH and water, sometimes with granular contents which turn blue green in the presence of air.
Subhymenial hyphae often somewhat sinuous, thin to thick-walled; forming a rather dense tissue. Individual hyphae (3.7-) 3.8-5.3 (-6.1) µm wide, with a mean width of 4.7 µm; brown to pale orange brown or pale green in KOH (but not with the blue green reaction present in other species); orange brown to brown in water, with strongly granular contents.
Encrustation granular, probably amyloid (hard to observe due to the colour); brownish-black in KOH, dark blue green in the presence of air; brownish-black in water; scattered in occurrence on the subicular hyphae.
Habitat. The only specimen of H. rhacodium recorded to date is the type collection, which was collected in Pennsylvania, United States. No further information on habitat or any further locality description is available.
Distribution. Basidiome encountered in: the United States.
Remarks. The hymenium of H. rhacodium is very thick and dense in comparison to all other Pseudotomentella species. It consists of tightly packed basidia, which are overlapping in length and have a total thickness equalling four-six basidial lengths. All other morphological characters fit well within the P. tristis group, thus suggesting an abnormal basidiome.
Within the P. tristis group, the basidiome of H. rhacodium can be recognised by its lack of hyphal cords and skeletal hyphae and its hard and brittle texture after drying. This feature makes it unique within the group and the risk for confusion with any other described species should hence be small. brown when older and which eventually turns black. It is further described as biennial and growing up from the ground and on to stones and branches, if they are present in its vicinity. It is hence doubtful whether the species in question belongs to the Thelephorales at all and, even though it has been synonymised with P. umbrina by Burt (1916), Litschauer (1933) and Svrček (1958), it does not match any Pseudotomentella described to date.

Discussion
In a world where unseen and undescribed new phyla hide in a grain of soil (Nilsson et al. 2016) and where visible, morphologically delimited taxa increasingly turn out to constitute cryptic species (e.g. Kroken andTaylor 2001, Shivas andCai 2012), it is reassuring to note that some visible, molecularly delimited species can still be separated morphologically. Similarly to many other fungal species, we, nevertheless, found P. tristis s.l. to constitute a complex of closely related and morphologically very similar species (e.g. Lücking et al. 2014, Jeppson et al. 2017). Thus Hjortstam's (1969) interpretation of P. tristis and P. umbrina as two different species was indeed correct, but so was Larsen's (1971a) argument that the variation he observed could not be separated into two species; under the name P. tristis are hiding no less than 13 species exhibiting morphological characteristics so close in range that they would indeed seem like a continuum to all mycologists without the aid of molecular analysis methods. In the light of our resurrection of P. umbrina as a separate taxon and the reviewed delimitation of P. tristis and P. atrofusca, this study not only proves the importance of combining molecular analysis methods with careful morphological studies, but also shows the power of these in conjunction with type studies. In the case of H. rhacodium and P. longisterigmata, the problems that can arise from species descriptions based solely on morphology are also clearly demonstrated.
From the perspective of functionality and usability of the international DNA sequence databases, it is satisfying to acknowledge that, while metadata from ecological studies have been very useful for understanding the molecularly delimited taxa presented here, future ecological studies querying such databases now have more reliable names to use. One species in the P. tristis group -P. umbrina -was indeed found to be widespread, have a wide ecological amplitude and, at least in northern Europe, to be commonly occurring. This is not to say that all the other species in the group are less widely distributed or have narrower ecological niches; some species, for example, P. pinophila, P. tristis itself and P. sciastra, have been collected in widely separated countries and habitats, but in comparison to P. umbrina they have rarely been encountered so far. More material is needed to establish the frequency of occurrence and ecological niches of all species in the species complex -information that might prove a helpful complement to morphology in the process of species identification, given the high degree of similarity between some species. With the current knowledge, it is quite paradoxical that the combination P. tristis was made by an American mycologist (Larsen 1971a), while the species in question now has no confirmed findings in North America. In contrast, P. atrofusca, a species believed to be widely distributed in Europe (GBIF 13-08-2018) and documented from the Russian far east (Kõljalg 1996), is now only known with certainty from the North American type collection and three sequences from China. South East Asia and Russia generally constitute large white spaces on the mycological map, even though findings so far indicate that species in the P. tristis group do occur in these areas. Even after taking the ecological knowledge gap into account, it is interesting to note that, unlike species in ectomycorrhizal genera such as Leccinum Gray and Hygrophorus Fr. that show strong host preferences and have more limited distribution ranges (den Bakker et al. 2004, Moreau et al. 2018, most species in the P. tristis group are able to form ectomycorrhiza with a broad range of hosts and are widely distributed. The fact that all species, except for P. umbrina, seem to be restricted to areas where soil pH is intermediate or high is possibly a factor that could help explain their difference in occurrence frequency.
The present study clarifies the application of the name P. tristis. In doing so, however, it renders hundreds of previous molecular ecology studies obscure with respect to this particular name. The name of P. tristis has served as something of a wastebasket for any and all Pseudotomentella species, owing both to the obscure nature of the underlying taxonomy and to the noisy state of taxonomic annotations in the public sequence databases. Thus, while the present study clarifies the use of the name P. tristis, it also raises doubts about previous molecular ecology results in the context of this name. To the extent that previous studies have relied on UNITE Species Hypotheses identifiers rather than Latin names when reporting molecular ecology results, this problem will be solved automatically. However, any study that tied species occurrences only to Scientific names may, from now on, convey incorrect information in the context of Pseudotomentella.
To the extent that it can be assessed given the moderate phylogenetic resolution, it is intriguing that the morphological characters that differ between species (e.g. spore size and shape, subicular hyphal width) do not seem to have a strong phylogenetic link. Whether these absences of patterns have the same cause, for example, an old rapid radiation, with extensive gene flow or are just artefacts of time and chance -causing both intragenic mutational conflict and genetic drift towards evolutionarily neutral shifts in morphology -is unclear, but could possibly be resolved by analysis of additional genetic regions. This may also shed some light on the presence of paralogous relationships between some of the taxa in the group and would possibly resolve some species into additional new species. The considerable genetic and morphological variation exhibited by Pseudotomentella sciastra, for example, may well indicate a species complex. Both ASTRAL and STACEY should be robust with the relatively small datasets used in the present study, in the sense that the employed datasets should not include less species than the analyses support. Additional specimen sampling may, however, aid in the distinction between populations and any possible, additional species and would thus, besides widened gene sampling, also be preferable in an extended study of the group.
Concerning morphology, the presence of amyloid material in and on basidia and subhymenial hyphae of Thelephorales species does not seem to have been reported. This is surprising, given its possible usefulness as a discriminatory character between species. Whether the cause of this is rarity or obscurity remains to be revealed by further studies in the field. Similarly worthy of notice is the blue green reaction observed in the same micromorphological structures of some species. Such a reaction has been mentioned by others studying Tomentella and Pseudotomentella (Larsen 1971a, Kõljalg 1996, but we would like to draw attention to the observation that the reaction in question here only seems to occur in the presence of air and also to its probable usefulness as a species-separating character. Finally, this study demonstrates clearly the necessity of applying a well-developed and consistent methodology when assessing the morphological characters of closely related species. It cannot be emphasised enough how important it is for those who endeavour to correctly identify Thelephorales species to carefully measure spores using the methodology originally described by Kõljalg (1996) and further explained in the Methods sections of this paper.