Mission impossible completed: unlocking the nomenclature of the largest and most complicated subgenus of Cortinarius, Telamonia

So far approximately 144,000 species of fungi have been named but sequences of the majority of them do not exist in the public databases. Therefore, the quality and coverage of public barcode databases is a bottleneck that hinders the study of fungi. Cortinarius is the largest genus of Agaricales with thousands of species world-wide. The most diverse subgenus in Cortinarius is Telamonia and its species have been considered one of the most taxonomically challenging in the Agaricales. Its high diversity combined with convergent, similar appearing taxa have earned it a reputation of being an impossible group to study. In this study a total of 746 specimens, including 482 type specimens representing 184 species were sequenced. Also, a significant number of old types were successfully sequenced, 105 type specimens were over 50 years old and 18 type specimens over 100 years old. Altogether, 20 epi- or neotypes are proposed for recently commonly used older names. Our study doubles the number of reliable DNA-barcodes of species of C. subgenus Telamonia in the public sequence databases. This is also the first extensive phylogenetic study of the subgenus. A majority of the sections and species are shown in a phylogenetic context for the first time. Our study shows that nomenclatural problems, even in difficult groups like C. subgenus Telamonia, can be solved and consequently identification of species based on ITS barcodes becomes an easy task even for non-experts of the genus.


Introduction
So far approximately 144,000 species of fungi have been named (Willis 2018) but sequences of the majority of them do not exist in the GenBank or UNITE. Moreover, only a small percentage of the names in the GenBank, about 4800 species, are based on sequences from type materials or other reliable sources (Schoch et al. 2014).
Currently species identification of fungi in academic studies is almost solely based on nrDNA ITS barcodes (Lindahl et al. 2013). Thus, those collections with taxonomically correct names that are not in any public sequence repositories are basically omitted in academic research. Therefore, the quality and coverage of public barcode databases is a bottleneck that hinders the study of fungi (Schoch et al. 2014). Depositing the ITS sequences in public repositories like GenBank does not automatically make them useful for identification. Two excellent platforms for delivering sequencebased identification information for the end-users include RefSeq under GenBank (Schoch et al. 2014) and UNITE . However, in both cases an extra step by an expert, in addition to the normal sequence submission, is required, but unfortunately often is left undone, making part of the already existing information unusable.
Cortinarius (Pers.) Gray is the largest genus of Agaricales with thousands of species world-wide (Kirk et al. 2008). They are important ectomycorrhizal fungi and often discovered in ecological studies. Only three large studies of type specimens based on ITS sequence data in Cortinarius have been made so far. Two of them are from C. subgen. Phlegmacium (Fr.) Trog: Liimatainen et al. (2014) includes over 230 sequences of type specimens representing over 150 species and Frøslev et al. (2007) has over 50 sequences of 79 species. The third one is from C. subgen. Telamonia (Fr.) Trog and includes over 60 sequences of 33 species .
The nuclear ribosomal internal transcribed spacer (ITS), which has been proposed as the universal barcode marker for fungi (Schoch et al. 2012), is also the main locus used in the species level taxonomy of Cortinarius. The treshold value for barcoding Cortinarius species has been proposed to be 99% (Garnica et al. 2016). However, there already is evidence that a few morphologically distinct Telamonia species only have 1 base difference (99.8% similarity) in the ITS region, e.g. C. laniger Fr./C. solis-occaus Melot (Niskanen et al. 2012) and C. paragaudis Fr./C. pinigaudis Niskanen, Kytöv. & Liimat. (Niskanen et al. 2011) and in the case of C. confirmatus Rob. Henry the intraspecific variation is > 1%, although the species has a wide morphological and ecological range and based on ITS sequences there are 3 supported subclades which might be separate taxa .
In this study our aim was to provide a revision of Cortinarius, subgen. Telamonia as well as an extensive ITS database for the identification of the species. Almost all type specimens of the species described in the subgenus were studied and an epi-or neotype is proposed for all recent frequently used older names when possible. In addition, a phylogenetic tree is produced as a framework for the infrasubgeneric classification of the species; including many that are included in a phylogenetic analysis for the first time.

Taxon sampling
The type specimens of Telamonia species published over many years by Ammirati, D. Antonini, M. Antonini, Bergeron, Bidaud, Bohus, Bouteville, Bresadola, Carteret, Chevassut, Consiglio, Daniele, Eyssartier, A. Favre, J. Favre, Fellner, Ferville, Fillion, Henry, Hesler, Høiland, Hongo, Karsten, Kauffman, Kühner, Landa, Lindström, Matheny, McKnight, Moser, Moënne-Loccoz, Murril, Nespiak, Orton, Pearson, Peck, Ramm, Reumaux, Sasia, Seidl, Smith, Schwöbel, Soop, Svrček, Velenovský, and Vialard were sampled as well as all the Telamonia collections published and illustrated in Brandrud et al. (1989Brandrud et al. ( , 1992Brandrud et al. ( , 1994Brandrud et al. ( , 1998. A total of 482 types are included here. An additional 1 3 183 previously published sequences of Telamonia types were added to our dataset for the best overview of current available data. We aimed to have at least two sequences per species in our study. Therefore, some additional sequences, either our own unpublished ones or from databases GenBank and UNITE, were included. Information on the sequences of type specimens is available in Supplementary Table 1 and information on other sequences included in the phylogenetic analysis is available in Supplementary Table 2. Fungarium acronyms follow Index Herbariorum (Thiers 2013).

Species concept
Based on criteria mentioned in the introduction we have used 1 % (5 differences) as a cut-off value for species. When type sequences differ in at least 5 sites from one another we have treated them as different species. We are not claiming that all the variation below 1 % is automatically intraspecific. Separating species below the 1% cut-off value, however, does require careful study. Therefore, we have added 'aff.' prefix to the Latin name in cases where there are 3 to 4 differences to another type sequences. With 2 differences we have used the s. lato notation in the Fig. 1 and Supplementary Tables 1 and 2. Using this approach indicates places where determining taxonomic synonyms might be problematic and require further study. Also, when macroscopic, microscopic and/or ecology data differ considerably although the ITS sequences are the same, we have not placed the taxa in synonomy. Furthermore, in cases where a species complex has previously been shown to include several species supported by morphology and small, but constant barcode gaps, we have avoided making synonymys.
One cannot emphasize enough that using a small cut-off value requires good quality sequences. In this study all the specimens have been sequenced from both directions and the chromatograms of the sequences were checked and edited manually before any preliminary analyses. When small, less than ten base or indel changes and/or odd differences are found between sister species or within species those differences have been confirmed by combining the relevant chromatograms and checking manually the base sites that differ. Also, base or length polymorphisms sites are not counted as a difference and an indel is counted as one difference despite its length.

Molecular analyses
DNA was extracted from a few milligrams of dried material (a piece of lamella) with the NucleoSpin Plant kit (Macherey-Nagel, Düren, Germany). The same protocol was used for all materials. Primers ITS 1F and ITS 4 (Gardes and Bruns 1993;White et al. 1990) were used to amplify ITS regions. The same primer pairs were used in direct sequencing. For problematic material the primer combinations ITS 1F/ITS 2 and ITS 3/ITS 4 were also used. PCR amplifications were performed in a 25 µl reaction mix with about 70 ng extracted DNA, 1 U Phusion High-Fidelity DNA polymerase and 1× HF buffer (ThermoFisher), 200 mM of each dNTP and 0.5 µM of each primer. The PCR were run on a MBS 0.2 G Thermal Cycler (Thermo Hybaid) with the following settings: denaturation for 30 s at 98 °C, followed by 35 cycles of denaturation for 10 s at 98 °C, annealing for 30 s at 50 °C, and extension for 30 s at 72 °C. The PCR products were purified using an ExoSAP-IT purification kit (Amersham Biosciences). Sequencing was performed on both strands using a BigDye Terminator v1.1 Sequencing kit (Applied Biosystems). Reactions were performed in 10 µl with 1 µl of PCR product, 1.3 mM of primer (ITS 1F or ITS 4), 1 µl 5X sequencing buffer, and 1 µl of Terminator Ready Reaction Mix. Reactions were run for 1 min at 96 °C, followed by 30 cycles of 30 s at 96 °C, 15 s at 50 °C, and 4 min at 60 °C. Unincorporated dye terminators and primers were removed by Sephadex G-50 DNA Grade Fine (Amersham Biosciences) purification system, and the reactions were analysed by ABI 3730 DNA Analyzer (Applied Biosystems) automatic sequencer. Sequences were assembled and edited with Sequencher 4.1 (Gene Codes, Ann Arbor, Michigan, USA). A total of 755 new ITS sequences were produced for this study. Collections and GenBank sequences used for the phylogenetic analysis are given in Supplementary Tables 1  and 2.
The short ITS sequences of type specimens were excluded from the phylogenetic analysis. To improve the resolution of phylogenetic analyses we included 146 published LSU sequences from GenBank to our dataset. The chosen LSU sequences are from different parts of Telamonia and they were mostly obtained from Garnica et al. (2005), Harrower et al. (2011), and Stensrud et al. (2014. Sequences from section Dermocybe Pers. were selected as an outgroup based on Stensrud et al. (2014). A total of 919 ITS and 146 LSU sequences were aligned separately for both regions using MAFFT 7 (Katoh and Standley 2013) with the G-ING-i algorithm (Katoh et al. 2005). The alignments were then manually improved in SeaView (Galtier et al. 1996). The phylogenetically informative indels in the ITS region were coded as characters following the simple indel coding algorithm (Simmons and Ochoterena 2000) with FastGap 1.2 (Borchsenius 2009). The binary and aligned nucleotide data were concatenated in Mesquite 3.2 (Maddison and Maddison 2017). The alignment is 2008 nucleotides long (including gaps) and is available at TreeBASE under S26824 (http:// www.treeb ase.org/treeb ase-web/home.html). A phylogenetic tree was generated from the concatenated dataset using maximum likelihood (ML) analyses with 1000 bootstrap replicates under the GTRGAMMA model for nucleotide partitions (ITS + LSU) and the default setting for binary (indel) data in RAxML 8 (Stamatakis 2014).

Molecular results
The phylogenetic tree resulting from the analysis, ITS and LSU regions including binary data from gap coding of the ITS region, is shown in FIG. 1 and a schematic drawing of the relationships of the sections based on the phylogenetic analysis in Fig. 2. Altogether, we recognize 80 sections which all form monophyletic groups in our analysis and examples of the sections are shown in Supplementary Figs. 1-11. We almost entirely used section names with clear identity, i.e., the concept of the type species of a section was well known. A total of 482 types representing 184 species were successfully sequenced. Of these, about half of the species had one or more synonyms. A significant number of old types were successfully sequenced, 105 types over 50 years old and 18 types over 100 years. All the major Cortinarius taxonomists have described new species that already had an older name, but the portion of younger taxonomic synonyms in terms of the total number of described species (synonym rate) varies among the different authors. Here are synonym rates for authors who have described most of the Telamonia species, based only on morphological characters: Peck 17%, Kauffman 7%, Smith 40%, Moser 47%, Henry 55%, and French Atlas team 72%. Current names of Cortinarius species used in this study with their synonyms are listed in Table 1. All the names of the types are listed in alphabetical order in Supplementary Table 1, followed by the current name.
Sometimes it was only possible to amplify part of the ITS region, in most cases it was then the ITS1 region that was succesful. Often in Cortinarius the ITS1 region alone is enough for a proper identification, but especially in the case of small Telamonia species several sister species can have an identical or almost identical ITS1 region. Therefore, all of the unclear cases are marked in Supplementary Tables 1  and 2 with a prefix 'cf.' in the Latin name under the current name.

Neo-and epitypifications
All older names without a type specimen that are included in the Cortinarius subgen. Telamonia key in Funga Nordica (Niskanen et al. 2012) and not yet typified are typified here with the exceptions of C. paleaceus Fr. and C. miniatopus J.E. Lange (not included in Niskanen et al. 2012) and C. psammocephalus (Bull.) Fr. (nomen dubium, no type proposed). In addition, for C. colus Fr. an epitype that differs from the current use of the name is proposed, and in the case of C. alboviolaceus (Pers.) Fr., C. flabellus (Fr.) Fr. and C. hinnuleus Fr. the best fitting candidate from two or more avalaible ones in a species group was selected. For C. anthracinus Fr. and C. cinnabarinus Fr. neotypes have been chosen by Høiland (1983) but were not sequenced in this study. Altogether, neotypes for 11 species originally described by Fries, Liljeblad and Persoon are proposed as well as epitypes for 9 species described by             Notes-The name C. hinnuleus has been collectively used for several deciduous forest species that have a yellowish brown to reddish brown pileus, distant lamellae with an earthy odour, white universal veil and strongly verrucose, subglobose to obovoidly subglobose spores. They collectively more or less fit to the Fries's protologue (Fries 1838) that describes a species with fulvous cinnamon pileus, distant lamellae and a white veil ring on the stipe that grows early in the season in deciduous forests. The species in the photograph of Brandrud et al. (1989), plate A19, fits Fries's protologue as well as Sowerby's colour plate and therefore we propose it as an epitype of the species. Notes-Fries' protologue does not perfectly fit to any currently known Cortinarius species. Since there is no clear solution, we decide to follow the Nordic concept of this name (Brandrud et al. 1992, Niskanen et al. 2012. For more nomenclatural discussion of this name and the reasoning for the current interpretation see the booklet of Brandrud et al. (1992). Notes-This species has recently been called C. colus (see also C. colus above) in the Nordic literature and listed as a synonym of C. miniatopus in Brandrud et al. (1989). However, the concept included two species, one with large spores currently named C. subminiatopus Kytöv., Niskanen & Liimat., (photograph Brandrud et al. (1989;A55)) and a sister species with smaller spores (7.0-9.0 x 4.5-5.5 μm, av.= 7.5-8.2 x 5.0-5.2 μm, Q=1.45-1.70, Qav.= 1.52-1.62). The macroscopic description of C. miniatopus by Lange (1940) fits both species well but the spore size given is 6.5-7 x 4.3-4.5 μm. Although the spore size in the protologue is even smaller than that of the small-spored species we conclude that the small-spored species fits best to the original description and here propose collection H6041343 as the epitype of the species. Illustrations: Brandrud et al. (1989: pl. A31), Fries (1867-1884 Descriptions of the species: Brandrud et al. (1989: pl. A31) as C. hemitrichus, Niskanen et al. (2012) as C. hemitrichus.

Cortinarius paleaceus
Notes-The name C. paleaceus has often been applied to C. flexipes (Pers.) Fr. coll. However, no odour, which is very typical of species of C. sect. Flexipedes Kytöv., Niskanen & Liimat., is mentioned in Fries's protologue (Fries 1838) and the lamellae are described as whitish when young. In addition, a plate from Fries (1867-1884) illustrates a species with pale lamellae and context of the stipe, a species that looks like C. hemitrichus, and not like C. flexipes and relatives that have darker lamellae and stipe context. Based on this we conclude that our current interpretation of C. hemitrichus best represents also this species and a neotype making these two names synonyms is suggested. Both names, C. paleaceus and C. hemitrichus, were described by Fries (1838) in the Epicrisis. Here we choose to continue the use of the name C. hemitrichus as the current name of the species to avoid a name change and confusion.  fig. 2 (1793).
Notes -This species was described by Bulliard (1793) and the only original material is the painted figure that has been chosen as a lectotype of the species in Brandrud et al. (1998). The plate illustrates a rather slender, brown species with a wide, convex to low convex, sometimes low umbonate, scaly pileus, and a scaly stipe, the lamellae are brown. However, it is not obvious that Agaricus psammocephalus would be a Cortinarius. The illustrated basidiomata are also reminiscent of species in the genus Inocybe and the clustered growing habit reminds one of a saprotrophic fungus. The epithet psammocephalus was combined in the genus Cortinarius by Fries (1838), who intepreted it as a species growing in coniferous forests. Because Bulliard worked in the Paris region, already Brandrud et al. (1998) concluded, that Fries's species most likely is different from Bulliard's species that supposedly was growing in a deciduous forest. Currently, the name is applied to a species pair C. castaneopallidus Carteret/C. quercoconicus Liimat., Kytöv. & Niskanen that usually have a much narrower, acutely umbonate pileus (Bidaud et al. 2004, plate 481;Brandrud et al. 1998, plate D57).
Taking into consideration that i) the basidiomata illustrated in the Bulliard's plate do not fit the species for which the name has currently been used, ii) the plate may represent a species from another genus, and iii) we have not found another candidate for the name from the genus Cortinarius, we refrain to use the name for a species in genus Cortinarius and treat is as a nomen dubium. Illustration. Brandrud et al. (1994: pl. C04).
Notes-The protologue by Fries (1818) is very short but mentions the main characteristics of the species currently considered as C. traganus (Brandrud et al. 1994;Niskanen et al. 2012): Basidiomata with a smell. Pileus pale lilac, stipe whitish purplish and bulbous, context yellow. Fries (1818) also refers to an illustration of Schaeffer (1774) that then becomes the type of the species. A majority of the figures in the illustration represent our interpretation of C. traganus (Fig. I-V), but Fig. VII clearly shows a typical characteristic of C. cyanites Fr.: the context of the stipe and pileipellis have become vinaceous red on exposure. In Fig. IX the spores are round which does not fit either of the above species, a potential species could be found from C. sect. Anomali where species with round spores and bluish colours occur. It seems that the type of C. traganus is a mixed illustration, but since the majority of the figures and the protologue fit the current concept of C. traganus, we here choose an epitype to support this interpretation.

Studies of type specimens
There are two ways for naming a barcode in a sequence database: either sequence a named voucher specimen based on a morphological identification or sequence a type specimen. Paradoxically, the first approach is currently the most widely used although the core reason for using the DNAbased identification is the unreliability of the morphological identification. The gold standard should be sequencing the type specimens to achieve an unambiguous, good quality identification database, but this unfortunately has thus far been generally neglected.
To improve the sequence-based identification of the important ectomycorrhizal genus Cortinarius and create a solid base for future taxonomic work 482 type specimens were sequenced. This is more than twice as many as the largest type study of Cortinarius so far . We were able to successfully sequence many old type specimens; 105 types which were over 50 years old and 18 over 100 years old. This shows that most available Cortinarius type specimens can likely be sequenced regardless of the age of the specimen. The dataset, including the already published type sequences in this group, contains a total of 363 species. About half of these species' names, altogether 184, are published now for the first time in GenBank, thus doubling the reliably of barcoded species of Cortinarius, subgen. Telamonia in the public sequence databases. Also 33% of the species represented here have been described over the last decade using DNA sequences alongside morphology and ecological data. Adding DNA tools for fungal taxonomy has accelerated the process of discovering and describing fungus diversity.

Synonyms
Our dataset shows that many species have been described several times. Of the 363 species recognized in this study 31% have a synonym, the synonym rate is even higher with species described using only morphological characters (46%). The two main reasons for synonyms are that the interpretation of the existing names has been challenging and there have been problems based on the morphological species concept. The high number of species, convergent evolution and the small number of useful morphological characters for classification have not made the task any easier. Also, the lack of uniform and stable infrageneric classification has made it more difficult to find potential, already existing descriptions of the species and thus many species have been subsequently named as new again. In the future, the problem of synonyms will be much reduced when sequences from type specimens are available and the description of new species without ITS barcodes are strongly discouraged.
One example of the difficulty of interpret existing names is C. impolitus Kauffman. It was the species described the most times by multiple authors over decades, e.g. by Kauffman (1918) and Smith (1944) from North America and by Velenovský (1939), Pearson (1946), Favre (1955) and Lindström (Brandrud et al. 1998) from Europe (Table 1 and Figure 1). The species is small and brown which partly explains the problem but it also has two good characters, odour of Pelargonium in the lamellae and narrow basidiospores, but despite these characteristics it has been very challenging to recognize it from the works of different mycologists based on morphology only.
Examples of the second problem, the challenges of using the morphological species concept, are C. macropodius Rob. Henry and C. luridus Rob. Henry that overall had the highest number of synonyms, 13 and 9 respectively. In this case, all synonyms come from the French authors and are due to a too narrow species concept. Some of the synonyms are also placed in different infrageneric groups in their classification system. This error rate and unnatural classification make it very difficult to use the earlier parts of the Atlas des Cortinaires series for identification of Cortinarius. However, the individual descriptions of the species are usually of good quality and 61 species names that have been described by the team are the oldest names for the species: representing about 15 % of all the currently known species of C. subgenus Telamonia. In recent years they have also included molecular data into their work which has greatly improved the outcome (e.g. Bidaud et al. 2017).
When looking at the rate at which the different authors described synonyms it is self-evident that it was easier to describe new species earlier when more species were undescribed. For example, the error rate of Kauffman is only 7% whereas Smith's error is double that, most likely because he was partly describing the species from the same area where Peck and Kauffman had previously worked. Half of the Smith's synonyms are Kauffman's species. The error rate of Moser and Henry are rather similar, which is a bit surprising since they mainly worked in different habitats and with a different species concept.

Interpretation and typification of the early names without type materials
Many early names without type specimens have been redescribed by later authors. From all the old names used in this study only 10 of them are without synonyms: C. armillatus (Fr.) Fr., C. bibulus, C. bovinus Fr., C. cinnabarinus Fr., C. colymbadinus Fr., C. dolabratus Fr., C. evernius (Fr.) Fr., C. gentilis, C. glandicolor (Fr.) Fr., and C. helvelloides. About half of them are rather characteristic and easy to interpret so taxonomists after Fries understood his concept and therefore did not describe those species again, i.e. C. armillatus and C. evernius. On the other hand, some of these species are really difficult to interpret and might not have been described again just because of the restricted distribution, infrequent occurrence or just a matter of chance, i.e. C. bovinus and C. dolabratus.
Interpreting the early names, like those of Fries and Persoon, when often no physical specimen is left to study and the descriptions themselves are short, vague and without microscopical characters, is extremely difficult. In many cases their species concept most likely included several species and was too generalized. They surely did their best but the state of knowledge in those times was far from what we know now. For example, C. paragaudis and C. praestigiosus, two species which based on current, widely accepted concept are far from each other both phylogenetically and morphologically, were included as varieties of one species in Fries' concept (Fries 1874). In Cortinarius sect. Bovini only one species was described by Fries although the section includes at least seven species in Sweden . Of course, Fries might not have found all those species in the areas he collected or did not have time to work with them, but it is still rather certain that many Fries' names included several species. Thus, due to the broad species concept there often is not any correct one candidate for epi-or neotypification. And even if there has been a clear concept behind the early species descriptions, it is often very difficult to interpret based on short and vague descriptions.
The interpretation of a name based only on morphology is a demanding, often impossible, task. In this study ca. 80% of the species described by Fries have been described again. The poor record can not be explained by a few poor studies or unprofessional authors-all major Telamonia authors have misinterpreted Fries' names or simply overlooked them. Studying the type specimens of Karsten's species gave a similar result. Karsten's descriptions are somewhat better than Fries' since they also include microscopical characteristics, but the critical difference is that Karsten's specimens are available and can be sequenced, thus we really can confirm the true identity of his species. The result was that all the seven Telamonia species described by Karsten, which we studied, have been redescribed later by other authors confirming the conclusion from Fries' materials. No current data supports the claim that the early names could be interperated correctly and consistently by anyone.
Because of the problems mentioned above the interpretation of early names in general is not a very meaningful thing to do and often the outcome is highly questionable. The majority of early names should probably be treated as nomen dubium. Therefore, we only typified those early names that have been widely used, e.g. appear on many national check lists or are commonly used in books like Funga Nordica (Niskanen et al. 2012). In these cases, the typification is a quicker and a more efficient way to stabilize nomenclature than trying to convince users to stop using the name. Also, it is important to point out that when typifying early names, we do not claim that the outcome would be correct, i.e. would represent the original concept of the author. We simply try to find the species that would best fit to the original description and in the case of several equally suitable candidates choose the most practical solution, i.e. the one that causes fewest changes in the current use of the name, the species itself would be the most common and wide spread of the candidate species and/or the easiest to recognize.
Another problem with the old names is the references to the illustrations. At those early times authors did not know that the references would later turn out to be the most important part of the descriptions-based on the current International Code of Nomenclature for algae, fungi, and plants (https ://www.iapt-taxon .org/nomen /main.php), they are considered as 'original material' of the species. At the time there was not a huge amount of published illustrations to choose from. It seems that in some cases Fries referred to an illustration that did not fit perfectly to his concept of the species but was the closest one with some similarity. This is e.g. obvious with C. colus and C. turgidus.

Nomenclatural coverage of the dataset and conclusions
In this study we tried to sequence all species level type specimens belonging to Cortinarius subgen. Telamonia that have not been previously studied. Our aim also was to stabilize all commonly used early names for which a type specimen does not exist. Obviously, all names in Telamonia are not in this dataset. Some type specimens could not be sequenced, especially Hongo's and Murrill's types failed almost without exception. Also, Henry's material was difficult to sequence and in addition, many of his type specimens were not found, the names are nomenclaturally invalid, or had other problems. Most of the Peck's material could not be acquired from NYS during the time of the molecular study of this paper. Some of Favre's type specimens were too small to sample or have already been sequenced but not published by other authors. Melot's type specimens are in his personal collection and despite several attempts to aquire them on loan, they were not avalaible for molecular study. Unless this situation changes the identity of the names remains unclear and it would be better not to use them to avoid confusion rising from the different interpretations of the names. There are a few authors whose materials we have not studied, e.g. Bon and Lamoure, but the number of Telamonia species they described is relatively small, only some tens of species.
After this study there will only be a few dozen valid names that have not yet been studied with molecular methods and where the type specimens are good quality for sequencing and available for study. Most likely many of them have an earlier name which already have been studied.
There are a few exceptions, however, for example if one is working with the sub-alpine Telamonia species the names described by Favre (e.g. 1955) and Lamoure (1977Lamoure ( , 1978 are relevant, for Eastern North America species described by C. H. Peck's should be checked (Burnham 1919;Gilbertson 1962), and for European Mediterranean areas the works of local authors would be appropriate to study (e.g. Mahiques and Ortega 2002). Otherwise, if a new sequence does not have ≥ 99% similarity to any published type sequence it can be rather certain that it derived from an undescribed species, given that the quality of the sequence is good.
Overall, our data set contains about 300 species from Europe and 150 species from North America and many of which they have in common. There may only be a few hundred more Telamonia species to be found from Europe, but certainly in North America the quest has just begun. The situation in Africa, Asia and Central and South America is praticially unknown, but it would not be an exaggeration to predict that the world-wide diversity of Cortinarius subgen. Telamonia would be a four-digit number. Thus far, members of the subgenus have not been found in the Nothofagus forests of New Zealand (Soop et al. 2019) and from Nothofagus forests of South America only one species is confirmed (Garnica et al. 2005).
As species are discovered and named the easier the identification based on ITS will become. Unfortunately, the same does not apply to morphological identification. All the current keys we use would require extensive rewriting and even though there often are morphological and/or ecological differences between the species, identifying many of the species of Telamonia using keys without deeper experience and knowledge of the group will be challenging if not impossible. Having local keys (i.e. Scandinavian boreal Telamonias or Telamonias of the Pacific North West) and in certain cases only trying to identify sections or species complexes rather than species would be the most realistic approach when using morphological identification.
Many times, the biggest obstacle for efficient identification and naming of alfa diversity are the nomenclatoral problems, i.e. what is the correct name for the species or is it an undesribed one? The species of C. subgenus Telamonia have been considered one of the most challenging cases in the Agaricales at the species level. Its high diversity combined with convergent, similar appearing taxa have earned it a reputation of being an impossible group to study, one better left in the forest. Our study shows that nomenclatoral problems, even in difficult groups like Telamonia, can be solved and identification based on ITS barcodes becomes an easy task even for non-experts.

Relationships of the sections within Cortinarius subgen. Telamonia
The relationships of the sections within Telamonia remain unclear in our phylogenetic analysis. The grouping of the sections in the tree, however, does not seem random and makes sense when compared to the morphological characteristics. Therefore, the main findings that we feel would be of importance are summarized below and could be used as starting hypotheses for future studies.
Based on our phylogenetic analysis C. subgen. Telamonia is roughly divided into two main entities (Fig. 2): (i) The basal groups of the tree ("Basal Telamonias") that only contain species with medium-to large-sized basidiomata (the apex of the stipe is > 4 mm wide) with the exception of a few species in C. sect. Brunnei. (ii) The monophyletic upper part of the tree ("Crown Telamonias") that mainly contains species with small basidiomata (the apex of the stipe is < 4 mm wide), and the following sections including species with mainly small-to medium-, less commonly large-sized basidiomata: Hinnulei Melot, Rubricosi Moënne-Locc. & Reumaux Leiocastanei Niskanen, Kytöv. & Liimat., and the monotypic sections Pseudoduracini Liimat., Niskanen & Kytöv., Friesiorum Liimat., Kytöv. & Niskanen, and Vinaceobrunnei Ammirati, Niskanen & Liimat.. The most basal part of this clade also includes sections Anthracini Melot, Crassispori Kytöv., Niskanen & Liimat. and Squalidi Liimat., Ammirati & Niskanen. Within the "Crown Telamonias" some further grouping can be observed. Brandrud et al. (1989) initially classified the species with small basidiomata into two sections, Incrustati Melot and Hydrocybe (Fr. ex Rabenh.) P. Karst. Although not forming well supported clades, these two earlier groups seem to correlate with the phylogeny to some extent. In the Fig. 2 the groups are named as /Squamicybe (Incrustati s. Brandrud et al.) and /Erubescentes (Hydrocybe s. Brandrud et al.). The new names are introduced because the type species of sect. Hydrocybe, Cortinarius duracinus, does not belong to "Crown Telamonias" but to "Basal Telamonias" and the type species of sect. Incrustati, C. luxnymphae, was not available for study and thus the identity of the species remains unclear.
The previous members of the C. sect. Incrustati are all placed in/Squamicybe (Fig. 2) in two monophyletic groups/ Eusquamicybe and/Paludosi but the group also includes sections of species with medium-to large-sized basidiomata. Many species of this group have a ±scaly pileus, a universal veil that forms distinct girdles on the stipe and a stipe/ context of the stipe that becomes darker towards the base, especially with age. No part of the basidiomata turns reddish (except in C. sect. Rubrocincti that resembles more the species in /Erubescentes  Brandrud et al. 1989, C. sect. Alnicolarum) and C. helvelloides (C. sect. Helvelloides), also belong to this larger group. Together with C. sect. Saniosi they form a monophyletic group /Paludosi, although without support, indicating that within C. subgen. Telamonia the ability to form mycorrhizae with Alnus may only have evolved once.
A majority of the species classified earlier in C. sect. Hydrocybe are placed in another, monophyletic, group / Erubescentes (Fig 2). They all have small basidiomata and a smooth pileus and in most species the stipe/context of the stipe does not become darker towards the base. In addition, in quite a few species either the base of the stipe, universal veil and/or basal mycelium turns ± reddish with time.

Sections
The aim of this study was not to solve the infrasubgeneric classification of C. subgen. Telamonia but to show the preliminary placement of the studied species and existing sections ( Figs. 1 and 2). Examples of the species belonging to the sections are shown in Supplementary Fig. 1-11. We included representative photographs for all but the following three sections: C. sect. Cacaodisci Kytöv., Niskanen & Liimat., C. sect. Pseudoduracini, and C. sect. Squalidi. In this study, 80 previously described sections and 9 subsections are used, and additional 11 section names are considered synonyms. A small number of species are not currently placed in any of the sections. The "Basal Telamonias" with mediumto large-sized basidiomata have been easier to study and are thus better known and only four species, C. hepaticus Kytöv., Niskanen & Liimat., C. hillieri Rob. Henry, C. uraceisporus Niskanen, Kytöv. & Liimat. and one C. sp., remain outside the currently accepted sections. In the "Crown Telamonias", that have been more overlooked mainly due to their small size, 18 species included in our phylogenetic analysis remain unclassified. Some of them, like C. denigratus Ammirati, Beug, Niskanen, Liimat. & O. Ceska and the related C. spp from North America that form a monophyletic group and differ > 4% (> 20 indels and substitutions) from other species of Cortinarius subgen. Telamonia, might be considered as a new section in the future. Some may be grouped with existing sections with futher analysis using additional DNA regions, i.e. C. ferrugineovelatus Kytöv. Liimat. & Niskanen and C. umbrinobellus Liimat., Niskanen & Kytöv. that share morphological characteristics with the species in C. sect. Praestigiosi but were currently placed in a basal position of the branch containing that section.
We wanted the section names to be as unambiguous as the species names as far as possible and therefore we only accepted section names that can be interpreted without a doubt, i.e. the type specimen of the type species of a section is sequenced. Exceptions were made with four names: sect. Anthracini Melot, sect. Brunneotincti M.M. Moser, sect. Cinnabarini Melot, and sect. Parvuli Melot. For these sections we have not been able to study the type specimen of the type species for several reasons or the sequencing failed but we believe that the concept of the type species is rather uniform and clear (e.g. Niskanen et al. 2012). Therefore, it seems acceptable to use these common section names. We are aware that this kind of approach is risky as the case of sect. Testaceofolii Liimat., Niskanen & Kytöv. shows. At the time, it was clear that C. biformis Fr. sensu Funga Nordica (e.g. Niskanen et al. 2012) was a different species than C. testaceofolius H. Lindstr. & Soop. However, later it turned out that the neotype Moser had selected for C. biformis was in fact an older synonym for C. testaceofolius, an outcome that no one had previously thought possible. Therefore, sect. Biformes Moënne-Locc. & Reumaux and sect. Testaceofolii are now synonyms.
If the species concept is often difficult to apply, then classification above species becomes even more subjective. In general, we should try to avoid having too many monotypic entities since they are less meaningful in classification.
However, the risk with bigger entities is having units which would have very little, if any, exclusive morphological characters that would define those groups, since one of the main reasons of having a higher-level classification is to recognize groups with unique character states. For example, C. armillatus, C. sect. Armillati, C. subgen. Telamonia, genus Cortinarius represent four levels of classification in which the species C. armillatus belongs to all four groups that have their own, unique defining characters that other groups in higher or lower levels do not have.
In this study, one example of the difficulties of delimiting a section is C. sect. Uracei. With a wider concept it includes several previously recognized sections, C. sect. Cinnabarini, C. sect. Colymbadini, and C. sect. Miniatopodes Moënne-Locc. & Reumaux, that all form a monophyletic clade with good support value and morphological differences from C. sect. Uracei s. str. Therefore, keeping all above-mentioned sections separate would be an arguable choice, but then there would be at least more than four monotypic sections inside the clade Uracei that would need a new name. In this case we have currently delimited C. sect. Uracei in a broad sense because the group is also supported by morphological characters. The other existing sections are treated at the subsection level.
We have tried to delimit the sections to be the widest monophyletic group with a reasonable support value and with at least some shared morphological character states. This approach and level of grouping mainly corresponds to the concepts previously used to delimit the sections in the genus Cortinarius in the era of molecular data (e.g. Ammirati et al. 2013Ammirati et al. , 2017Dima et al. 2014;Liimatainen et al. 2015Liimatainen et al. , 2017Liimatainen et al. , 2020Niskanen et al. 2009Niskanen et al. , 2011Niskanen et al. , 2013San Fabian et al. 2018;Soop et al. 2019).
The sections identified here vary from monotypic entities i.e. C. sect. Brunneocalcarii Niskanen, Liimat. & Kytöv. to middle-sized groups i.e. C. sect. Armillati and C. sect. Disjungendi to very diverse groups i.e. C. sect. Bovini and C. sect. Uracei. The imbalance is unlikely to be solved due to the speciation history of different groups, likely some of them have diversified more than the others which has led to the current species-poor and species-rich groups. Also, this dataset only contains a fraction of the true diversity of C. subgen. Telamonia worldwide and therefore the number and the species diversity of sections will change when more data are available. Most of the monotypic sections will most likely turn out to be multi-species sections as shown e.g. by Soop et al. (2019). This is the first extensive phylogenetic study of C. subgen. Telamonia. The great majority of sections and species are shown in a phylogenetic context for the first time. Also, many sections previously included in phylogenetic studies now contain more species and therefore seem to have better support values. For example, Harrower et al. (2011) used the same two DNA regions, ITS and LSU, in their study and got less than BS 50% support for C. sect. Firmiores (Fr.) Hennings when including three species in their analysis. In our study that contains 20 species the support value for the same section was BS 88%. For C. sect. Armillati the corresponding values were BS 54% (2 species) and 90% (7 species). Fig. 1 shows our current view of the number of the sections in Cortinarius subgen. Telamonia and which species we include in them. The earlier delimitations based on morphology have been partly incorrect and included only a part, often a small fraction, of the species Brandrud et al. 2012;Niskanen et al. 2012). The classification presented here is a major step forward and can be used as a basis for a more thorough revision of morphological characterstics of the groups in the future. Now that the nomenclatoral history of the last 100 years has been sorted out for many taxa, everyone can benefit from the outcome and continue to improve the understanding of this diverse group of species. Fortunately, all current Cortinarius taxonomists produce an ITS barcode of the type specimen of new species and upload and annotate the new sequence in GenBank. We hope that mycologists working on Cortinarius and other genera will build on the findings reported here.

Conclusions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.