DNA barcode identification of lichen-forming fungal species in the Rhizoplaca melanophthalma species-complex (Lecanorales, Lecanoraceae), including five new species

Recent studies using sequence data from multiple loci and coalescent-based species delimitation have revealed several species-level lineages within the phenotypically circumscribed taxon Rhizoplaca melanophthalma sensu lato. Here, we formally describe five new species within this group, R. occulta, R. parilis, R. polymorpha, R. porterii, and R. shushanii, using support from the coalescent-based species delimitation method implemented in the program Bayesian Phylogenetics and Phylogeography (BPP) as the diagnostic feature distinguishing new species. We provide a reference DNA sequence database using the ITS marker as a DNA barcode for identifying species within this complex. We also assessed intraspecific genetic distances within the six R. melanophthalma sensu lato species. While intraspecific genetic distances within the five new species were less than or equal to the lowest interspecific pairwise comparison values, an overlap in genetic distances within the R. melanophthalma sensu stricto clade suggests the potential for additional Copyright Steven D. Leavitt et al. This is an open access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC-BY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. MycoKeys 7: 1–22 (2013) doi: 10.3897/mycokeys.7.4508 www.pensoft.net/journals/mycokeys A peer-reviewed open-access journal

Rhizoplaca Zopf, as currently circumscribed, is a morphologically diverse, polyphyletic genus (Arup and Grube 2000) represented by ca. 19 lichen-forming fungal species (Index Fungorum: http://www.indexfungorum.org/). Within the past decade a number of studies have indicated that traditional phenotype-based species circumscriptions fail to recognize multiple species-level lineages within this genus (Zhou et al. 2006;Leavitt et al. 2011a). The Rhizoplaca melanophthalma species-complex (sensu Leavitt et al. 2011a) includes a morphologically diverse assemblage of species, including individuals ranging from placodioid crustose and umbilicate forms, to completely vagrant, or obligatory unattached forms. Substantial chemical variation has been identified within this group, including at least four chemotypes within R. melanophthalma sensu lato (s.l.) (McCune 1987;Ryan 2001); and eight previously unrecognized extrolites recently identified from specimens within this complex (Leavitt et al. 2011a).
Within R. melanophthalma s. l. (Fig. 1), Leavitt et al. (2011a) circumscribed six 'candidate' species that were supported using multiple lines of evidence from molecular sequence data, including: fixed nucleotide characters, genealogical exclusivity, Bayesian population clustering, and the coalescent-based species delimitation program Bayesian Phylogenetics and Phylogeography (BPP; Yang and Rannala 2010). The lat- ter, has recently been shown to outperform other species delimitation methods under a variety of scenarios (Camargo et al. 2012). In spite of a high degree of morphological and chemical variation within this species complex, most of the candidate species identified in this study were morphologically and/or chemically polymorphic, and diagnostic morphological/chemical characters were not identified for the majority of the independent species-level lineages (Leavitt et al. 2011a). Recently, additional data, including broader geographic sampling and three additional genetic markers, provide additional support that the candidate species identified in Leavitt et al. (2011a) represent species-level lineages (Leavitt et al. 2013).
While morphological and chemical character differences have traditionally served as proxies for identifying reproductively isolated groups, multilocus coalescent-based species delimitation methods can provide a more direct assessment of gene flow and independent lineage status through genetic analysis. Coalescent-based methods can provide a more direct and replicable approach for assessing hypotheses of evolutionary independence, regardless of whether putative lineages differ in potentially subjective phenotypic character systems (Fujita et al. 2012). Character evolution in lichens is still poorly understood, leading to potentially confounding morphological/chemical taxonomic features (e.g. Leavitt et al. 2011b;Leavitt et al. 2011c;Pino-Bodas et al. 2011;Pérez-Ortega et al. 2012;Pino-Bodas et al. 2012). While we are strong advocates for the application of independent data types (i.e. ecology, geography, morphology, genetics, and chemistry) in developing an integrative taxonomy, there is an increasing need to formally recognize the existence of phenotypically cryptic species-level lineages in lichen-forming fungi (Hibbett et al. 2011).
In this paper we use support from the coalescent-based species delimitation method implemented in the program BPP (Yang and Rannala 2010) as the diagnostic feature distinguishing new species from other taxa. While in practice, most modern species descriptions include a character-based diagnosis, it has been argued that coalescent-based diagnosis serve the same purpose as a standard diagnosis when the species in question is not diagnosable on the basis of morphology alone (Leache and Fujita 2010;Fujita and Leache 2011;Fujita et al. 2012).
Rhizoplaca melanophthalma s.l. is frequently used in air quality bio-monitoring studies (Dillman 1996;Ugur et al. 2004), and differences in pollution accumulation patterns among closely related species have not been tested. In addition, R. melanophthalma s.l. has been shown to have pharmaceutical potential for treating drug genotoxicity in human blood (Geyikoglu et al. 2007). Therefore, accurate specimen identification may have important implications for bio-monitoring and pharmaceutical research using R. melanophthalma s.l. Furthermore, several lineages within the R. melanophthalma species-complex are broadly distributed (Leavitt et al. 2013), and these may potentially serve as valuable groups for assessing dispersal capacity and landscape-level genetics in response to changing climatic conditions, assuming accurate specimen identification. Given the overall importance of accurate specimen identification, including phenotypically cryptic lineages, the objectives of this study are to (1) formally describe five new species within this group and (2) provide a reference DNA sequence database using the ITS marker as a DNA barcode for identifying species within this complex.

Candidate species and taxon sampling
Using multilocus sequence data generated from Rhizoplaca specimens collected throughout the Intermountain region of western North America, Leavitt et al. (2011a) circumscribed six candidate species, 'C2', 'C3', 'C4a', 'C4b', 'C4c', and 'C4d', within R. melanophthalma s.l. Based on initial sampling, these six species-level lineages were strongly supported by a variety of operational criteria for species delimitation (Leavitt et al. 2011a). Furthermore, many of the candidate species within R. melanophthalma s.l. were shown to occur sympatrically with strong evidence of reproductive isolation among lineages (Leavitt et al. 2011a), and thus de facto species status. However, these candidate species were not formally described due to the limited geographical sampling. Increased geographic sampling, including collections from Antarctica, Central Asia, Europe, and North and South America, along with additional genetic markers corroborate the previously recognized candidate species (Leavitt et al. 2013).

Data analysis
In this study, we used the ITS alignment and phylogeny reported in Leavitt et al. (2013) (Fig. 2A; supplementary file 1; TreeBase ID 13903). This data consisted of 228 sequences and 524 aligned nucleotide position characters. A full description of multiple sequence alignment and phylogenetic reconstruction methods is given in Leavitt et al. (2013). In short, the multiple sequence alignment was performed using the program MAFFT v6 (Katoh et al. 2005;Katoh and Toh 2008), and phylogenetic relationships were estimated using maximum likelihood using the program RAxML v7.2.8 (Stamatakis 2006;Stamatakis et al. 2008).
In the present study, we calculated pairwise distances to characterize both intraand interspecific variation within and among candidate species-level lineages. Pairwise distances can be viewed as a rough measure for the overall sequence divergence (Del-Prado et al. 2010). Average genetic distances were computed using PAUP* (Swofford 2002) based on pairwise comparisons of all sequences within each candidate species individually, overall intraspecific distances from all species, and pairwise interspecific distances. Pairwise distances between different haplotypes were reported as the number of nucleotide substitutions per site (s/s).
The Barcode of Life Data Systems (BOLD; Ratnasingham and Hebert 2007) provides an informatics workbench aiding the acquisition, storage, analysis, and application of DNA barcode data, including a BLAST-based identification tool for fungi using the ITS region. Data from the candidate species circumscribed in Leavitt et al. (2011a), including ITS sequences, electropherograms, and collection information, were submitted to the BOLD database, project name 'Rhizoplaca melanophthalma-species complex'. This custom database includes collections form the western USA representing the six previously recognized candidate species.
We assessed the utility of the ITS region for BLAST-based specimen identification in the R. melanophthalma group by conducting searches of the newly reported ITS sequences in Leavitt et al. (2013) against the custom BOLD database. The queried sequences were generated from R. melanophthalma s.l. specimens from Antarctica, Austria, Chile, China, the Czech Republic, Iran, Kazakhstan, Kyrgyzstan, Russia, Spain, Switzerland, and additional specimens from the United States of America (supplementary file 1). Top species matches obtained from the BLAST searches against the BOLD database for each specimen were recorded and compared with results from a standard, tree-based method  Leavitt et al. (in review). Values at each node indicate nonparametric-bootstrap support; only support values > 50% are indicated B Box plots of ITS genetic distances within each new species, all intraspecific distances, and all interspecific distances. In each box plot, the box shows the interquartile range (IQR) of the data. The IQR is defined as the difference between the 75th percentile and the 25th percentile. The solid and dotted line through the box represent the median and the average length, respectively; and C The coalescent-based species-tree for the Rhizoplaca melanophthalma species-complex estimated from five genetic markers (ITS,IGS,group I intron,and MCM7 loci) in Leavitt et al. 2011a. to address accuracy of identification in a thoroughly sampled phylogeny. Species were scored as successfully discriminated if the samples were recovered within monophyletic clades in the ML analysis corresponding to candidate species and if top species matches for each new sequence obtained from the BLAST searches against the BOLD database corresponded to the monophyletic clade where it was recovered.
The marginal posterior probability of speciation (speciation probability) was estimated in Leavitt et al. (2013) from multi-locus sequence data using the program BPP v2.1 (Yang and Rannala, 2010). This method accommodates the species phylogeny as well as lineage sorting due to ancestral polymorphism. BPP has recently been shown to outperform other coalescent-based species delimitation methods, with robust performance using a modest number of genetic markers (Camargo et al., 2012). Full details are reported in Leavitt et al. (2013) and summarized only briefly here. The multi-locus species tree, representing the six candidate species, was used as the fully resolved guide tree. Because the prior distribution of θ and τ 0 can result in strong support for models containing more species (Leaché & Fujita, 2010), Leavitt et al. (2013) explored three combinations of priors, including a moderate and conservative combination. The most conservative combinations of priors favors fewer species by assuming large ancestral population sizes, and relatively shallow divergences among species. The moderately conservative set of priors assumed intermediate ancestral population sizes, and relatively shallow divergences among species. High speciation probabilities (SP ≥ 0.95) were estimated at all nodes using both the default prior gamma distributions for θ and τ 0 and a more moderate combination of these priors, with the exception of the split between the two vagrant species (R. haydenii and R. idahoensis). Under the most conservative combination of priors for θ and τ 0 , speciation probabilities match those supported using default priors, with the exception of lower probabilities for a split between 'C4d' and 'C4c' (SP < 0.50).

Results
For each candidate species, the pairwise distances among the different ITS haplotypes were estimated and the distribution of distances plotted (Fig. 2B). Intraspecific genetic distances in candidate species-level lineages 'C3', 'C4a', 'C4b', 'C4c' and 'C4d' were less than or equal to the lowest interspecific pairwise comparison values (0.006 substitutions/site), with the exception of outlier values in 'C4b' and 'C4d' (Fig. 2B). The outlier values within clade 'C4b' were based on comparisons from two sequences retrieved from GenBank (EF095278 and EF095287). We were unable to verify the quality of these sequences. Within candidate lineage 'C2', intraspecific values largely fell below the lower quartile of interspecific pairwise comparison values (Fig. 2B), although there was substantial overlap between intra-and interspecific genetic distance values. Mean distance values, standard deviations and the range of intraspecific distances within the six candidate species are reported in Table 1.
Ecology and distribution. In its narrower circumscription, this taxon is known from Antarctica, Asia (including Central Asia and China), Europe, North and South America. The species has also been recorded from alpine areas in the tropics. However, additional studies are required to verify the identity of these populations. It typically occurs on exposed calcium-poor rock (e.g. basalt, granite, schist), but sometimes on calcium rich sandstone and limestone. It ranges in distribution from arid lowland woodlands into upper montane coniferous forests and the lower portions of the alpine tundra. Specimens examined. See supplementary file 1.
Phylogenetic notes: Strongly supported as monophyletic lineage in both concatenated multilocus gene tree (ML bootstrap = 100%: posterior probability = 1.0) and the ITS gene topology (ML bootstrap = 99%, this study); and strong speciation probability inferred from multiple loci (BPP speciation probability ≥ 0.98). R. occulta belongs to a closely related, and well-supported, monophyletic lineage including R. paralis, R. polymorpha, R. porterii, and the obligatory vagrant species R. haydenii and R. idahoensis.
Ecology and distribution. Growing usually on exposed calcium-poor rock (e.g. basalt, granite, schist) in pinyon-juniper woodlands but also occurs free on soil. So far known only from collections in western North America. R. occulta included a total of five individuals from Idaho (3 individuals), Nevada (1), and Utah (1), USA, and included GenBank accessions identified as R. cerebriformis ined. (AF159942) and R. subidahoensis ined. (AF159944).
Etymology. The name is derived from the Latin "occultus," meaning hidden, and refers to the fact that this species was hidden within the phenotypically circumscribed taxon Rhizoplaca melanophthalma sensu lato.
Specimens examined.  Leavitt et al. (2011a), which is supported as a lineage distinct from all other populations according to coalescentbased genetic analysis of multiple genetic loci. Within the R. melanophthalma speciescomplex, the occurrence of orsellinic, lecanoric, and gyrophoric acids appear to be restricted to R. parilis. However, the occurrence of these compounds varies widely within this species, with the proportional occurrence of each compound ranging between 0.43 -0.64 (Leavitt et al. 2011a). The mean genetic distance among ITS haplotypes was estimated to be 0.003 ± 0.004.
Ecology and distribution. This species usually occurs on exposed calcium-poor rock (e.g. basalt, granite, schist), but sometimes on calcium rich sandstone and limestone. Its habitat ranges from pinyon-juniper woodlands to montane coniferous forests and the lower portions of alpine tundra. This taxon is currently known from Asia (including Central Asia and China), Europe, and North and South America Etymology. The specific epithet is chosen from the Latin parilis, meaning equivalent, like, or similar, in reference to the morphological similarity between the new species and the other species within the R. melanophthalma species-complex.
Specimens examined.  Leavitt et al. (2011a), which is supported as a lineage distinct from all other populations according to coalescent-based genetic analysis of multiple genetic loci. This species is morphologically quite variable. While some individuals are morphologically similar to R. melanophthalma sensu stricto, vagrant forms partly embedded in badland soils in western Idaho also belong within this species. The mean genetic distance among ITS haplotypes was estimated to be 0.0 ± 0.
Phylogenetic notes: Strongly supported as monophyletic lineage in both concatenated multilocus gene tree (ML bootstrap = 82%: posterior probability = 1.0), and weak statistical support in the ITS gene topology (ML bootstrap = 66%, this study); and strong speciation probability inferred from multiple loci (BPP speciation probability ≥ 0.97). R. polymorpha belongs to a closely related, and well-supported, monophyletic lineage including R. occulta, R. parilis, R. porterii, and the obligatory vagrant species R. haydenii and R. idahoensis.
Ecology and distribution. Currently only known from collections in western North America. Its habitat includes pinyon-juniper woodlands and montane coniferous forests, but unattached forms are also known from the McBride Creek Badlands in Western Idaho.
Etymology. The specific epithet was selected based on the morphologically polymorphic forms within this species, including both umbilicate and vagrant forms.
Specimens examined. Description. Morphologically similar to R. melanophthalma sensu stricto, but consists of specimens recovered within 'clade IVd' in Leavitt et al. (2011a), which is supported as a lineage distinct from all other populations according to coalescent-based genetic analysis of multiple genetic loci. This species is also characterized by the absence of a group I intron in the nuclear SSU rDNA at the 1516 position (Gutiérrez et al. 2007), which is present in all other species within the R. melanophthalma species-complex. The mean genetic distances among ITS haplotypes was estimated to be 0.002 ± 0.002.
Phylogenetic notes: A monophyletic lineage in both concatenated multilocus gene tree with weak statistical support (ML bootstrap < 50%; posterior probability < 0.5), and with strong statistical support in the ITS gene topology (ML bootstrap = 94%, this study); and strong speciation probability inferred from multiple loci (BPP speciation probability ≥ 0.97). R. porterii belongs to a closely related, and well-supported, monophyletic lineage including R. occulta, R. parilis, R. porterii, and the obligatory vagrant species R. haydenii and R. idahoensis.
Ecology and distribution. This species usually occurs on exposed calcium-poor rock (e.g. basalt, granite, schist), but sometimes on calcium rich sandstone and limestone. Its habitat ranges from pinyon-juniper woodland into montane coniferous forests and lower alpine tundra. This taxon is currently known only from the western USA (Idaho and Utah).
Etymology. The new taxon is named in honor of Dr. Lyndon D. Porter, whose research on Rhizoplaca melanophthalma proved invaluable to the present work.
Specimens examined. See supplementary file 1.  Leavitt et al. (2011a), which is supported as a lineage distinct from all other populations according to coalescentbased genetic analysis of multiple genetic loci. The mean genetic distances among ITS haplotypes was estimated to be 0.001 ± 0.002.
Phylogenetic notes: A monophyletic lineage in both concatenated multilocus gene tree with strong statistical support (ML bootstrap = 100%; posterior probability 1.0), and with strong statistical support in the ITS gene topology (ML bootstrap = 100%, this study); and high speciation probability inferred from multiple loci (BPP speciation probability = 1.0).
Ecology and distribution. Found growing only on sun-exposed basalt boulders in subalpine meadows in southwestern USA. Currently known only from subalpine habitats on the Aquarius Plateau in southern Utah, USA.
Etymology. The new taxon is named in honor of the late Dr. Sam Shushan, a pioneer in western North American lichenology.

Discussion
In this study we described five new species within Rhizoplaca melanophthalma s.l. Our results indicate that a molecular-based approach for specimen identification in the common lichen-forming R. melanophthalma species-complex can effectively assign individuals from cosmopolitan populations to previously circumscribed 'candidate' species (Leavitt et al. 2011a;Leavitt et al. 2013). Molecular data indicate that the genus Lecanora, as currently circumscribed, is not monophyletic (Arup and Grube 1998;Lumbsch 2002;Grube et al. 2004). The Rhizoplaca melanophthalma speciescomplex clearly falls outside of the core group of Lecanora sensu stricto (including the type species L. allophana; see Brodo and Vitikainen 1984), and we choose to describe the new species within the heterogeneous, and also non-monophyletic, genus Rhizoplaca, pending future taxonomic revisions.
In spite of the limitations in delimiting taxa using molecular data, the effective use of genetic data appears to be essential to appropriately and practically identify natural groups in some phenotypically cryptic lichen-forming fungal lineages (Divakar et al. 2010b;Leavitt et al. 2011b;Leavitt et al. 2011c;Molina et al. 2011b;Pino-Bodas et al. 2011;Pino-Bodas et al. 2012), including R. melanophthalma s.l. This does not preclude the fact that additional studies investigating morphological and chemical characters may potentially identify independent characters, or combinations of characters, supporting species circumscribed using molecular data. In fact, under the general lineage species concept (GLC; de Queiroz 1998de Queiroz , 1999de Queiroz , 2007, more independent properties associated with putative species boundaries are associated with a higher degree of corroboration, resulting in a truly integrative approach to species discovery. However, robust species delimitations using molecular data in phenotypically cryptic species can provide working hypotheses about what constitutes separately evolving metapopulation lineages (de Queiroz 1998;1999;Mayden 1999;de Queiroz 2007;Fujita et al. 2012). DNA barcoding provides an objective approach for specimen identification within these taxonomically difficult groups.
Within the Rhizoplaca melanophthalma species-complex, DNA barcoding can be performed in a variety of ecological, bio-monitoring and population genetic studies in order to quickly sort specimens into genetically divergent groups. In R. melanophthalma s.l., this barcode application for specimen identification may provide a valuable framework for assessing biogeographic patterns, bio-monitoring research, and prove to be an important tool in making critical conservation-related decisions. The application of molecular-based identification could also be used as a way for both specialists and non-specialists alike to discriminate species that are otherwise difficult to identify, making specimen identification more accessible and more accurate at the same time.
In spite of the progress in recognizing independent species-level lineages within R. melanophthalma s.l., high intraspecific distances within R. melanophthalma sensu stricto, suggest additional species-level lineages may potentially be hidden within this lineage. In order to address this question, we are currently developing novel genetic markers (i.e. microsatellites) specific for this group in order to assess population structure and gene flow within this broadly distributed species.
Using ITS sequence data, specimens within the Rhizoplaca melanophthalma species group can be identified by means of DNA barcoding using the publicly available database in BOLD (http://www.boldsystems.org/). Based on our broad intercontinental sampling, it appears that Rhizoplaca melanophthalma s.l. specimens can be accurately identified to species using the BLAST-based identification tool for fungi in BOLD. This provides an objective approach for a broad array of researchers to accurately identify species within this group using ITS sequence data from their collections, regardless of their level of taxonomic expertise.
Although in some cases the ITS region has been shown to be effective for molecular identification using DNA barcoding (Kelly et al. 2011;Schoch et al. 2012), including species within the Rhizoplaca melanophthalma species-complex, the reliance on a single locus for inferring relationships and circumscribing species is problematic because the history of a single gene might not be representative of the organismal history. Within the R. melanophthalma species-complex relationships among species-level lineages are largely unsupported in the ITS topology ( Fig. 2A) and differ greatly from the coalescence-based species tree estimated from multilocus sequence data (Fig. 2C;Leavitt et al. 2011a;Leavitt et al. 2013). Additionally, genealogical concordance among independent genetic markers can provide strong evidence that distinct clades represent reproductively isolated lineages among well-separated groups (Dettman et al. 2003;Pringle et al. 2005). Although different datasets and operational criteria may give conflicting or ambiguous results due to the multiple evolutionary processes associated with speciation, the use of multilocus sequence data and multiple empirical methods are known to establish robust species boundaries in many lichen-forming fungal lineages (reviewed in Lumbsch and Leavitt 2011). Within the R. melanophthalma species-complex, the majority of candidate species showed high levels of genealogical concordance among three ribosomal and two protein-coding markers (Leavitt et al. 2011a). The coalescentbased species delimitation method BPP (Yang and Rannala 2010) also supported the distinctness of all new Rhizoplaca species described here (Leavitt et al. 2013) and, under a variety of scenarios, has been shown to be among the most accurate coalescent-based species delimitation methods (Camargo et al. 2012).
While in some cases data have supported the taxonomic use of secondary metabolic characters for delimiting lichen taxa (Tehler and Kallersjo 2001;Blanco et al. 2004;Schmitt and Lumbsch 2004;Molina et al. 2011a), other studies found no correlation between chemotypes and lineages identified using molecular phylogenetic reconstructions (Articus et al. 2002;Buschbom and Mueller 2006;Divakar et al. 2006;Nelsen and Gargas 2009;Velmala et al. 2009;Myllys et al. 2011). In spite of some general patterns in the distribution of secondary metabolites among species within the Rhizoplaca melanophthalma species-complex, it appears that chemical characters cannot consistently be used to diagnose independent species-level lineages. For example, within this complex the occurrence of orsellinic, lecanoric, and gyrophoric acids appear to be restricted to R. parilis. However, the occurrence of these compounds varies widely within this species, with the proportional occurrence of each compound ranging between 0.43 -0.64 (Leavitt et al. 2011a). Additionally, all sampled specimens of vagrant Rhizoplaca species only produced usnic acid, although additional morphological differences are required to accurately identify distinct vagrant species. Previous studies have used TLC to characterize lichen secondary metabolic products within Rhizoplaca (McCune 1987;Zhou et al. 2006). However, it appears that within R. melanophthalma s.l. some extrolites would be masked by other compounds, or likely found at levels undetectable by TLC; and HPLC has been shown to provide a more sensitive approach for determining secondary metabolite diversity within the R. melanophthalma complex (Leavitt et al. 2011a). In spite of the increased sensitivity of HPLC, unambiguous secondary metabolic characters corroborating most of the species within R. melanophthalma s.l., including the most genetically divergent clades, were not identified (Leavitt et al. 2011a).
As molecular sequence data become more readily available, they will allow us to better understand the diversity of lichenized fungi. Their use in identifying species will become increasingly important and routinely applied. Other disciplines such as ecology, conservation and physiology will benefit from a more objectively based species circumscription, enabling us to interpret distribution and ecological patterns better and more accurately monitor environmental disturbance and climate change. Bovini (Bovini s. str.). The species are characterized by brown to dark brown basidiomes without bluish colors and exsiccatae with a dark brown to blackish brown pileus. The universal veil is white, brownish white or grayish white, in some species becoming grayish brown with age, and the odor is indistinct or slightly raphanoid. To date, species are only known from Europe, except C. oulankaënsis which also occurs in Canada in British Columbia. By studying more material from western North America, we wanted to determine if C. bovinus found from Alaska, U.S.A. and Alberta, Canada is conspecific with European samples or does it represent an autonomous species.

Methods
Material gathered by the authors from North America was studied morphologically, ecologically and sequenced to infer phylogentic relationships with other species in Bovini. DNA was extracted from dried material (a piece of lamella) with the NucleoSpin Plant kit (Macherey-Nagel, Düren, Germany). Primers ITS 1F and ITS 4 (White et al. 1990, Gardes and were used to amplify ITS regions, and specific primers cort6F and b7.1R (Frøslev et al. 2005) for the rpb2 region. The same primer pairs were used in direct sequencing. PCR amplification and sequencing followed Niskanen et al. (2009). Sequences were assembled and edited with Sequencher 4.1 (Gene Codes, Ann Arbor, Mich., USA). Using a BLAST query of the public databases (GenBank: http://www. ncbi.nlm.nih.gov/ and UNITE: http://unite.ut.ee/), we checked if identical or similar sequences for our species exist in public databases. For the phylogenetic analysis ITS and rpb2 sequences of the species belonging to the well-supported ingroup of section Bovini, C. bovinus, C. bovinaster, C. bovinatus, and C. oulankaënsis, were included. Cortinarius anisatus, C. neofurvolaesus, and C. sordidemaculatus were chosen as outgroup species.
The combined ITS and rpb2 alignment of 11 specimens was produced with the program MUSCLE (Edgar 2004) under default settings. The alignment comprised 1286 nucleotides (including gaps). The alignment is available at TreeBASE under S14159 (http://www.treebase.org/treebase-web/home.html).
Bayesian inference (BI) was performed with MrBayes 3.1.2 (Ronquist and Huelsenbeck 2003). The best substitution model for the alignment was estimated by both the Akaike information criterion and the Bayesian information criterion with jModelTest version 0.1.1 (Posada 2008). A GTR model, including a gamma shape parameter, was chosen for both DNA regions. Two independent runs with four chains in each were performed for 1 000 000 generations sampling every 100th generation. All trees sampled before stationarity were discarded with a 25% safety margin (burn-in of 2 500 trees [250 000 generations]). Sampled trees from both runs were combined in a 50% majority rule consensus phylogram and posterior probabilities (PP) were calculated. The analysis was run with computer clusters of the CSC, IT Center for Science, Espoo, Finland.
Morphological descriptions are based on material collected by the authors including specimens in all stages of development. Color notations in the description follow Munsell (2009) soil color charts. Microscopic characteristics were observed from dried material mounted in Melzer's reagent (MLZ). Measurements were made in MLZ with an ocular micrometer using 100× oil-immersion lens. Basidiospores were measured from the veil or top of the stipe, 20 spores from one basidiocarp. The length and width were measured for each spore, and their length/width ratios (Q value) were calculated. The lamellar trama and basidia also were examined, and the pileipellis structure was studied from scalp sections taken from the pileus center.

three new species of foetid Gymnopus in New Zealand
Their /micromphale clade was nested within a broader /lentinuloid clade including Rhodocollybia, Marasmiellus ramealis (Bull.) Singer, Marasmius scorodonius (Fr.) Fr. and Lentinula. Mata et al. (2004), based on an LSU analysis, also identified a clade containing sequences of M. foetidum and material named as Setulipes androsaceus (L.) Antonín and Gymnopus fusipes (Bull.) Gray, the type species of those respective genera. They adopted a broad concept of Gymnopus incorporating these genera together with Marasmiellus. Wilson and Desjardin (2005) used LSU to examine the group and identified a /gymnopus clade containing G. fusipes, M. foetidum, S. androsaceus at its core with Micromphale perforans (Hoffm.) Gray lying on its boundary. These results were broadly supported by Mata et al. (2006) in their analysis using ITS1-5.8-ITS2 but they demonstrate clustering of G. fusipes, S. androsaceus and Micromphale on the periphery of a concentration of Gymnopus-labelled samples. On the basis of these results the currently generally accepted concept of Gymnopus is broad (e.g. Noordeloos 2012), and incorporates a number of previously recognised genera. Hughes et al. (2010) erected the genus Connopus to accommodate the Gymnopus acervatus group within the gymnopoid clade and presented LSU and ITS data indicating its placement close to Rhodocollybia. Their LSU analysis supports a core gymnopoid clade containing G. fusipes, S. androsaceus, which once again places Micromphale foetidum and M. perforans on a boundary with a sister group containing Rhodocollybia, Marasmiellus juniperinus Murrill and various Gymnopus species. The /gymnopus, clade as interpreted by Hughes et al., contains significant substructure. A multi-gene analysis including more representatives may indicate the recognition of further segregates at genus-level. In this paper we accept our newly described species within the current broad concept of Gymnopus whilst recognising their close alliance to the historical concept of the genus Micromphale.
For this study we analysed ITS1-5.8-ITS2 data for related New Zealand collections together with representative sequences from Genbank, many from the studies cited above. The structure of our ITS tree is consistent with these previous analyses, and once again identifies a /micromphale clade closely linked to core Gymnopus species. ITS data generated for a number of representative collections of our newly described taxa support species concepts based on morphology.

Morphological protocols
Spore dimensions are stated as the mean ± 1.5 SD of 20 measurements, thus covering 86% of measurements under an assumed normal distribution model. Fresh or dried material was examined mounted in 10% KOH or Melzer's reagent. Material was handsectioned. Some micrographs were obtained under DIC conditions. Measurements were always taken without DIC optics and an extended objective iris in order to maximise boundary contrast.

Phylogenetic protocols
DNA extraction and sequencing followed the protocols outlined in Cooper and Leonard (2012). We downloaded from Genbank selected sequences used in cited publications, together with close BLAST matches, Table 1. General sequence management was carried out using Geneious (Drummond et al. 2011). Data exchange between applications was facilitated using Alter (Glez-Peña et al. 2010). Sequence alignment was carried out using MAFFT within Geneious . A maximum likelihood analysis was executed using RAxML (Stamatakis 2006), with 100 bootstrap runs, launched from Topali 2.5 (Milne et al. 2004). The substitution model of GTR+G was recommended by Topali 2.5. We selected a sequence of Anthracophyllum archeri (Berk.) Pegler as the outgroup.

Results
Our analysis places the New Zealand taxa in a monophyletic clade close to G. foetidum and G. brassicolens historically recognised in the genus Micromphale (Fig. 1). The combination of sequence data and morphological analysis of many collections indicate two major groups which we equate with the newly described species G. imbricatus and G. ceraceicola. In addition we recognise a further species, G. hakaroa, which is poorly distinguished from G. imbricatus on the basis of ITS sequences but which is morphologically consistently different. Minor sequence variation in the G. ceraceicola group does not correlate with morphology and we choose to recognise these specimens as a single species. More information and images of collections may be found on the Landcare Research website (Systematics Collections Data). Micromorphology. Pileipellis a partially gelatinised radially arranged clamped cutis of smooth hyphae to 5 μm diameter, with brown extra-cellular encrustation. Epidermal layer to 140 μm. Subepidermis of thick glassy-walled non-gelatinised smooth hyaline hyphae, weakly dextrinoid. Basidia clavate to 40 × 8 μm. Sterigmata to 7 μm, 4-spored. Basidioles cylindrical, tapering towards apex, 40 × 4 μm. Spores hyaline, lacrymoid, 7.9 ± 1 × 4.5 ± 0.6 μm, Q = 1.8 ± 0.1 including apiculus. Cheilocystidia and pleurocystidia not observed. Stipitipellis a cutis of brown parallel hyphae, to 5 μm wide. Caulocystidia smooth, hyaline, agglutinated into fascicles. Habitat. Colonies of a few to hundreds of fruitbodies on bark of fallen, dead branches and twigs, especially Nothofagus.

Gymnopus ceraceicola
Distribution. Broadly distributed and common in both North and South Islands of New Zealand.
Etymology. Ceraceicola, indicating association with a basal waxy layer, although this feature is common to the three species described here.
Habitat. Forming imbricate colonies of dozens to hundreds of fruitbodies on bark and decorticate wood of dead branches and twigs, especially Kunzea and Leptospermum but occurs with other trees. Also occurs at the stem base of live trees.
Distribution. Broadly distributed and common in both North and South Islands of New Zealand.
Etymology. Imbricatus, pertaining to the often tiered and overlapping eccentrically stemmed caps. Diagnosis. G. hakaroa is distinguished from G. ceraceicola by smaller stature and a pruinose stipe lacking fascicles of agglutinate caulocystidia. It is distinguished from G. imbricatus by non-imbricate growth, a consistently central stipe, and smaller basidiospores.
Macromorphology. Pileus 3-10 mm diam. convex, rusty tawny to umber, minutely felty, weakly radially furrowed and striate towards the margin. Lamella cream to yellow, waxy. Lamellae present, in series of three: intercalated short/long/short. Stipe central, cartilaginous, to 5 × 0.6 mm, equal, umber to black, paler towards base, smooth to minutely pruinose. Stipe base insititious and always associated with an obvious waxy to chalky cream layer of partially gelatinised hyphae covering the substrate, usually green with algal cells. Fruitbodies with garlic/rotten cabbage smell, especially when crushed.
Habitat. Forming imbricate colonies of dozens to hundreds of fruitbodies on decorticate dead wood.
Distribution. Currently G. hakaroa is only known from a single location on the Canterbury Port Hills in the South Island of New Zealand.
Etymology. Hakaroa, a Maori name for the Bank's Peninsula region of New Zealand. Notes. Sequence data (Fig 1) indicates a close phylogenetic relationship to G. imbricatus but there are consistent and substantial morphological differences.

Dicussion
Gymnopus imbricatus, as its name suggests forms dense populations of small imbricate fruitbodies. It is most commonly associated with tea-tree (Kunzea ericoides and Leptospermum scoparium) and often found on the bark at the base of living trees. Gymnopus hakaroa is larger, with a dark minutely pruinose cap and again forms dense populations  on the bark of dead logs. These two species have smooth stems. Gymnopus ceraceicola is distinguished by larger fruitbodies, a pruinose stipe, and is more commonly associated with southern-beech forests on dead fallen logs. The species of Gymnopus described here belong in the /micromphale clade of Moncalvo et al. (2002) and share the diagnostic feature of this clade of a foetid odour likely due to the presence of mercaptan-like compounds. In New Zealand this feature is shared with Mycetinis curraniae (G. Stev.) J.A. Cooper & P. Leonard, a marasmioid fungus distinguished by its ornamented hymeniderm pileipellis. Another very distinctive character common to all three Gymnopus species, and visible in the accompanying photographs (Figs 2 and 6), is the presence of a waxy layer of partially gelatinised hyphae on the substrate from which the fruitbodies emerge. This layer is usually green from the presence of embedded algal cells. Interestingly, some published images of G. foetidus in the northern hemisphere also show a similar layer, e.g. Antonín and Noordeloos (2010). Detailed examination of our material does show algal cells deeply embedded within the context of the waxy layer and the basal portion of the stipe (Figs 9 and 10), but it would seem unlikely that algal cells are present in sufficient numbers to confer any significant nutritional benefit to the fungus. The morphologically similar Marasmiellus affixus (Berk.) Singer, described from Australia and commonly known as the 'little stinker', is also associated with a waxy algae-infected layer. The association of M. affixus with alga was noted by Singer (1973) and has been speculated to be a basidio-lichen, although this has not proven (Lepp 2011). A partial, poor quality ITS1 sequence for M. affixus obtained during this work (not deposited) suggests it has affinity with Marasmiellus ramealis (Bull.) Singer rather than the taxa treated here.

DNA barcoding and morphological studies reveal two new species of waxcap mushrooms (hygrophoraceae) in Britain
A

introduction
In Europe, waxcap mushrooms (Hygrophoraceae, Hygrocybe s.l.) are conspicuous and often attractive features of nutrient-poor short turf. Hotspots of waxcap taxonomic diversity are usually biocide-free, unfertilised or semi-improved, grazed grasslands or mown lawns with low levels of soil disturbance and long periods of ecological continuity. Such sites are often, but not always, of low botanical interest and this has undoubtedly delayed their recognition as sites of international conservation importance. Historically, therefore, even the best "waxcap grasslands" (using Arnolds ' (1980) terminology) have rarely received adequate long-term protection. More recently, waxcap assemblages are increasingly being recognized as useful bioindicators for identifying sites of conservation priority (e.g., Boertmann 1995, Nitare 2000, McHugh et al. 2001, Griffith et al. 2002, Newton et al. 2003, Evans 2004, Genney et al. 2009). Some species, such as the pink waxcap H. calyptriformis, have emerged as flagships for the still nascent practice of targeted conservation of fungi, and waxcaps as a group are becoming mascots for fungal conservation in general.
Waxcaps are regarded as nitrogen-sensitive organisms because fruiting is inhibited by applications of nitrogenous fertilizers (Arnolds 1989). However, their belowground ecology, in particular their nutritional mode(s), remains unclear despite recent attention by several researchers. Indirect evidence from carbon and nitrogen stable isotope ratios suggests that at least some taxa are biotrophic (Seitzman et al. 2011), and further evidence of a mycorrhiza-like association with plants has been demonstrated recently (Halbwachs et al. unpublished data).
Taxonomic treatments of European waxcaps have recognized from one to seven genera (e.g. Kummer 1871, Orton 1960, Orton and Watling 1969, Kühner 1980, Moser 1983, Kovalenko 1989, Arnolds 1990, Bon 1990, Boertmann 1995, Candusso 1997, Krieglsteiner 2001, Bresinsky 2008, Vizzini and Ercole 2012. Molecular phylogenetic analysis indicates that these fungi are not monophyletic and that at least two major phylogenetic clades can be recognized (Babos et al. 2011, Lodge et al. in press). Basidiomata of one group are characterized by vivid yellow, orange and red colours whereas those of the second group lack muscaflavin pigments and are pallid to brown, sometimes showing olive, pink or purple tints (Babos et al. 2011). At least three major groups can be recognized based on hyphal arrangement and compartment lengths within the hymenophoral trama (Boertmann 1995(Boertmann , 2010. These categories are partly supported by phylogenetic evidence (Babos et al. 2011, Lodge et al. in press).
Waxcap identification in Britain and Ireland currently adheres to Boertmann's (1995Boertmann's ( , 2010 taxonomic concepts. In turn, these concepts are based on basidiomatal macroscopic and microscopic morphology, although it is accepted that some taxa can show overlapping variation. As a result, 51 species (plus eight infraspecific taxa) of Hygrocybe s.l. are currently accepted in the online Checklist of the British and Irish Basidiomycota (CBIB; http://www.basidiochecklist.info/). However, only a handful of these are individually recognised as species of conservation concern. Five waxcaps were assessed as Vulnerable or Near Threatened in the Great Britain & Isle of Man unofficial Red Data List (Evans et al. 2006) and only the date waxcap, H. spadicea, is currently recognised as a priority species in the UK Biodiversity Action Plan. Not only does morphological identification of waxcaps underpin their current RDL assessment (Evans et al. 2006), it also contributes to the designation of UK sites as Important Fungus Areas (Evans et al. 2001) and Sites of Special Scientific Interest (SSSI). Indeed, waxcaps are currently one of the few groups of fungi for which SSSIs can be designated; any site with at least 18 recorded Hygrocybe s.l. species "should be considered for SSSI status" (Genney et al. 2009). Waxcap taxonomy and identification are, therefore, fundamental to their effective conservation.
Recent developments in DNA-based methods of identification ("DNA barcoding") are revolutionizing rapid diagnosis of diversity in mushrooms and other Fungi (Dentinger et al. 2011, Schoch et al. 2012). This study is part of a UK-wide initiative that is applying a DNA barcoding approach to waxcaps and revealing surprising levels of unknown diversity. We currently believe that at least 96 species are present in the UK as defined by DNA sequence-based methods (Defra science and research project WC0787). This has involved morphological and molecular analysis, or reanalysis, of 83 fungarium specimens in K whose sequences were published by Brock et al. (2009), 124 newly-sequenced specimens from K, E, and MICH, and more than 600 new field collections mostly from 2011 and 2012.
This paper focuses on our treatment of two unusual waxcaps that, because of their viscid pilei and subregular hymenophoral tramal hyphae, are assigned to the segregate genus Gliophorus. They share some morphological characters with the parrot waxcap G. psittacinus, which encompasses a wide range of basidiomatal pigmentation based on current concepts. Numerous colour forms can be recognised (Boertmann 2001) but, partly because this character is known to be influenced by ageing and weather conditions, formal taxonomic resolution into recognisable segregate species has proved more challenging. Four varieties are listed in Index Fungorum (http://www.indexfungorum.org). Our unusual collections lacked green pigments and one group matched the type description of Hygrophorus perplexus A.H. Sm. & Hesler, a North American species. This is recorded in Europe where, as one of the few accepted parrot waxcap segregates, it is currently recognised as Hygrocybe psittacina var. perplexa (Boertmann 2012). Molecular analysis, including sequences derived from type specimens of Hygrophorus perplexus, collections filed as H. psittacina var. perplexa in K and downloaded from GenBank labelled as H. psittacina, confirmed the presence of two new species lacking green pigments, which we describe here.

Taxon and specimen sampling
A total of 20 collections corresponding to the G. psittacinus complex were sequenced and morphologically examined in the current study. These comprised 12 recent UK field collections now in K, four existing K collections from UK and Jersey and four US type collec-tions in MICH. Table 1 shows the relevant collection details. Further details for specimens of G. europerplexus and G. reginae are provided in the taxonomic treatment below. Geographical coordinates of collections were converted from Ordnance Survey National Grid References, based on the OSGB36 datum, to latitude and longitude (WGS84 datum).

Morphological analysis
Spore measurements are rounded to the nearest half micron and preceded by associated data in square brackets. For example, [60, K(M)181128*, K(M)181129] would indicate that 60 spores in total were measured either in water from prints (G. reginae) or in Melzer's reagent from lamellar squashes (G. europerplexus) from the collections K(M)181128 and K(M)181129. Collections sequenced during this study, such as K(M)181128 in this example, are denoted throughout by *. Measurements of basidia and other hyphal elements are rounded to the nearest micron. Colours given in parentheses refer to those shown in a standard mycological identification chart (Anon 1969).

DNA extraction and sequencing
DNA was extracted using either an enzymatic digestion-glass fiber filtration protocol in 96-well plate format with a vacuum-manifold or the Whatman FTA® card method described in Dentinger et al. (2010). Full and partial nuclear ribosomal internal transcribed spacer regions (ITS) were amplified and sequenced with primers ITS1F/ITS3 and ITS2/ ITS4 (White et al. 1990, Gardes and or with primers ITS8F and ITS6R (Dentinger et al. 2010) following the cycling conditions in Dentinger et al. (2010). PCR products were visualized by UV fluorescence after running out 2 μL PCR products in a 1% agarose gel containing 0.005% ethidium bromide. Prior to sequencing, positive PCRs were cleaned by adding 0.5 μL ExoSAP-IT to every 2.5 μL PCR reaction mix and incubating this mix for 15 min at 37 C followed by 15 min at 80 C. Unidirectional dye-terminator sequencing used the ABI BigDye kit (Foster City, CA) following the manufacturer's instructions except reducing the total reaction volume to 5 μL. Sequencing reactions were cleaned using ethanol precipitation and resuspended in distilled water before loading into an ABI PRISM 3730 DNA Analyzer in the Jodrell Laboratory, Royal Botanic Gardens, Kew. Complementary unidirectional reads were aligned and edited using Sequencher4.2 (GeneCodes, Ann Arbor, MI). All new sequences have been deposited in the International Nucleotide Sequence Database (Accession numbers: KF218257-KF218275).

Phylogenetic analysis
Six additional sequences labelled as H. psittacina (Brock et al. 2009, Babos et al. 2011 were downloaded from GenBank and combined with our dataset (Table 1). The se- quences were trimmed to minimize uneven ends across the dataset and aligned using the RNA structure-based algorithm Q-INS-i implemented in MAFFT v7.023b , Katoh and Toh 2008, Katoh and Standley 2013. Phylogenetic analysis under the maximum likelihood criterion was performed using the Pthreadsparallelized version of RAxML v7.0.3 (Stamatakis 2006, Ott et al. 2007) with a GTR-GAMMA model. Branch support was assessed using nonparametric bootstrapping with the "thorough" option and 1000 replicates. The final alignment and phylogenetic tree are available from TreeBase (#14384; http://purl.org/phylo/treebase/phylows/ study/TB2:S14384 ).

Gliophorus reginae
Distribution. Known from a cemetery in West Wales (Pembrokeshire) and fields in central England (Worcestershire, Staffordshire and Derbyshire). The earliest known collection was made by C. Lovatt in Staffordshire in 1996 who noted that she had recorded similar specimens in 1994. It has fruited on private land at Willow Bank (Worcestershire) almost every year from 2000 onwards and recorded there in five discrete fruiting patches in a single field of ca. 0.8 ha.
Ecology. In unimproved short (grazed or mown) acid-neutral rough pasture or other grassland. This species is often a relatively late fruiter and can continue producing basidiomata in January, long after other waxcaps have finished.
Etymology. Latin reginae meaning "of a queen", named for the royal purple colour of the basidiomata and to celebrate the diamond jubilee of Her Majesty Queen Elizabeth II in 2012 and the 60 th anniversary of her coronation in 2013.
Conservation status. Collectors noted that although basidiomata of this species resembled G. psittacinus, some characters such as pileal colour and radial splitting, were more characteristic of H. calyptriformis. Furthermore, dried collections of the latter and G. reginae often attained a similar reddish-coral tint in the fungarium that was distinct from the pale saffron of G. psittacinus. This similarity facilitated a rapid search of the British G. psittacinus collections at Kew, but no additional G. reginae specimens were  Etymology. Named to distinguish this European taxon from the morphologically similar Hygrophorus perplexus A.H.Smith & Hesler, a species with North American type material.

Conservation status.
Initially, it seemed likely that historic British collections assigned to H. psittacina var. perplexa would be redetermined as G. europerplexus following DNA sequencing. However, two specimens filed in K under the former name yielded distinct ITS sequences (Fig.1). Therefore, the distribution of G. europerplexus is currently unknown and it should be assessed as Data Deficient.

Discussion
The traditional Gliophorus (=Hygrocybe) psittacinus species concept is relatively broad (Boertmann 2001), mainly due to difficulties in defining unambiguous morphological discontinuities in basidiomatal characters. Our ITS sequence analysis revealed a clade comprising two terminal clusters, G. reginae and G. europerplexus, together with a singleton, a paratype of H. perplexus, that is clearly divergent from that comprising the currently-accepted European taxon G. psittacinus and the N. American G. perplexus (Fig.  1). It is clear that the name G. psittacinus currently represents a species complex and further work is required to characterise and describe the component taxa. We have assigned our two new species to the segregate waxcap genus Gliophorus Herink based on recent supporting molecular phylogenetic evidence (Babos et al. 2011, Lodge et al. in press).
One of the novel taxa, G. reginae, is recognisable in the field having a relatively stout stipe, sometimes yellowing at the base, and distinctive deep purple or reddishbrown pileus. Our field observations suggest that a colour form of this might be shown in Boertmann's photograph of Danish specimen DB 2000/33 taken at Lysnet, E. Jylland, on 17 Oct. 2000 (Boertmann 2001 fig. 4, 2010 p.91). The remaining currentlyaccepted European taxon in the parrot waxcap group is H. psittacina var. sciophanoides (Boertmann 2010(Boertmann , 2012. This species was originally described as Hygrophorus sciophanoides by Rea (1922) based on a painting of an English specimen found in 1909 in Derbyshire designated as Rea 937, but no type specimen is preserved at Kew. Rea's "uncommon" fungus had a rosy pink striate pileus 1-3 cm diam. with pale pink lamellae, a concolorous stipe 2-5 cm × 2-3 mm and flesh described as pale yellow becoming white. Although Rea did not use the word "viscid" in his description, nevertheless he synonymised Cooke's (1889) concept of Hygrophorus sciophanus, a slightly viscid-pileate species with decurrent gills, with H. sciophanoides. Cooke (1889) quoted a description of two Scottish specimens, one pale and sterile and the other darker and yielding "very pale clay-coloured" spores from Notices of British Fungi No. 1560 (Berkeley and Broome 1876). The latter authors also recorded some small Welsh specimens in Notices of British Fungi No. 1885, again noting the existence of light and dark forms (Berkeley and Broome 1881). Cooke's (1886-1888) six illustrations (No. 905 Plate 937A) of English material from Kendal resemble the specimen depicted by Rea. G. reginae can be similarly pink and striate (Fig. 2F), but Rea and Cooke both described and illustrated basidiomata that were strikingly more slender than those of G. reginae. This together with the lack of type material of H. sciophanoides and the existence of other colour forms of the H. psittacina complex that can develop pink tints with age, leads us to conclude that H. sciophanoides should be regarded as a nomen dubium. Our attempts to sequence Welsh material collected in 1950 (K(M) 69657) and determined by Pearson as H. sciophanoides were unsuccessful.
In Europe, Hygrophorus sciophanus (Fr.) Fr. is currently regarded as a synonym of H. psittacina var. perplexa with Hygrophorus perplexus A.H. Sm. & Hesler as basionym. By contrast, Hesler and Smith (1963) argued that their taxon had a very similar lamellar attachment, "never decurrent", to that of Hygrophorus psittacinus, a character that distinguished it from H. sciophanus. Indeed Fries distinguished the lamellar attachment of Agaricus psittacinus, described as "adnatis", from that of A. sciophanus, "decurrentibus" (Fries 1821) and, later, of H. sciophanus, "subdecurrentibus" (Fries 1836(Fries -1838. Rea (1922), on the other hand, described the attachment in H. sciophanus as "attenuato-adnate", as shown in Rea 936, a painting of a French collection, and he cited an illustration approved by Fries. The latter painting (Fries 1877-1884.1) appears to bear out Rea's description, but in the same volume (p. 66), Fries used the word "decurrentibus" in the diagnosis and commented that the illustration showed "lamellarum insertio minus typica". The original concept of H. sciophanus thus is unclear and various interpretations exist in the literature. Three collections originally filed as Hygrophorus sciophanus preserved in K were sequenced and determined to be highly divergent from the Gliophorus sequences in our dataset, belonging instead to Hygrocybe sensu stricto (data not shown). In our view, H. sciophanus should be regarded as a nomen dubium.
Our analysis showed that the ITS sequence derived from the holotype specimen of H. perplexus is certainly distinct from the second of our new species, G. europerplexus. Two specimens identified as H. psittacina var. perplexa (Table 1) collected in 2003 and 2008 were also sequenced, but they are phylogenetically distinct, forming a clade near to G. psittacinus (Fig. 1) and may represent a further novel taxon. The single anomalous sequence from a paratype of H. perplexus, which comes near G. europerplexus in our analysis, reveals additional cryptic diversity within this species complex in North America and highlights the difficulty in correctly naming waxcap species using morphology alone. Attempts should be made, therefore, to sequence additional European and North American specimens currently filed as G. perplexus, Hygrophorus perplexus, Hygrocybe perplexa and H. psittacina var. perplexa to gain a better understanding of the distribution of G. europerplexus and other emerging segregate taxa.