Molecular Phylogenetics and Evolution 44 (2007) 217–227www.elsevier.com/locate/ympev

Molecular phylogenetics and delimitation of species in Cortinarius section Calochroi (Basidiomycota, Agaricales) in Europe

Tobias Guldberg Frøslev a,¤, Thomas Stjernegaard Jeppesen b, Thomas Læssøe a, Rasmus Kjøller a

a Section of Microbiology, Department of Biology, University of Copenhagen, Øster Farimagsgade 2D, DK-1353 Copenhagen K, Denmarkb Botanical Museum, University of Copenhagen, Gothersgade 130, DK-1123 Copenhagen K, Denmark

Received 21 July 2006; revised 24 October 2006; accepted 9 November 2006Available online 26 November 2006

Abstract

Cortinarius is the most species rich genus of mushroom forming fungi with an estimated 2000 spp. worldwide. However, species delim-itation within the genus is often controversial. This is particularly true in the section Calochroi (incl. section Fulvi), where the number ofaccepted taxa in Europe ranges between c.60 and c.170 according to diVerent taxonomic schools. Here, we evaluated species delimitationwithin this taxonomically diYcult group of species and estimated their phylogenetic relationships. Species were delimited by phylogeneticinference and by comparison of ITS sequence data in combination with morphological characters. A total of 421 ITS sequences were ana-lyzed, including data from 53 type specimens. The phylogenetic relationships of the identiWed species were estimated by analyzing ITSdata in combination with sequence data from the two largest subunits of RNA polymerase II (RPB1 and RPB2). Seventy-nine specieswere identiWed, which are believed to constitute the bulk of the diversity of this group in Europe. The delimitation of species based on ITSsequences is more consistent with a conservative morphological species concept for most groups. ITS sequence data from 30 of the 53types were identical to other taxa, and most of these can be readily treated as synonyms. This emphasizes the importance of critical anal-ysis of collections before describing new taxa. The phylogenetic separation of species was, in general, unambiguous and there is consider-able potential for using ITS sequence data as a barcode for the group. A high level of homoplasy and phenotypic plasticity was observedfor morphological and ecological characters. Whereas most species and several minor lineages can be recognized by morphological andecological character states, these same states are poor indicators at higher levels.© 2006 Elsevier Inc. All rights reserved.

Keywords: Calochroi; Cortinariaceae; Fulvi; Molecular taxonomy; Phlegmacium

1. Introduction

Cortinarius is currently considered the largest mushroomgenus with the number of species estimated at c. 2000 (Kirket al., 2001). However, much confusion exists concerning thenumber and status of taxa and the application of names.Cortinarius is a good example of how limited our knowledgeis about many mushroom groups. Hitherto, recognition anddelimitation of species and higher taxa has relied almostexclusively on basidiome morphology and ecology. Within

* Corresponding author. Fax: +45 35 32 23 21.E-mail address: tobiasgf@bi.ku.dk (T.G. Frøslev).

1055-7903/$ - see front matter © 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.ympev.2006.11.013

Cortinarius, the use of anatomical and chemical charactershas proven valuable for establishing morphologically coher-ent taxonomies for several groups (e.g. Brandrud, 1996, 1998;Melot, 1990). But widely accepted consensus taxonomies arelacking for most groups, and the phylogenetic and taxo-nomic value of traditionally emphasized characters has notbeen thoroughly tested.

1.1. Cortinarius section Calochroi—the /Calochroi clade

Phylogenetic studies have indicated that the sectionsCalochroi and Fulvi as delimited by Melot (1990) andBrandrud et al. (1990, 1992, 1995, 1998), with the exclusion

218 T.G. Frøslev et al. / Molecular Phylogenetics and Evolution 44 (2007) 217–227

of the Cortinarius percomis-group, constitute a well delim-ited lineage (Frøslev et al., 2005; Garnica et al., 2003b, 2005;Peintner et al., 2001, 2004)—the /Calochroi clade. The cladecontains taxa with bright yellow/greenish anthraquinonoidpigments treated in section Fulvi and taxa without thesepigments for the major part treated in section Calochroi.Both sections have been placed in the polyphyletic subge-nus Phlegmacium deWned by a glutinous pileus and a drystipe. Considerable confusion exists over the taxonomy andnomenclature of the species in the clade. In this study theconcept of section Calochroi is extended to cover all speciesin the /Calochroi clade.

The European species belonging to section Calochroihave been treated by several workers but variously circum-scribed (Bidaud et al., 1994, 2001, 2003, 2004; Brandrud,1998; Brandrud et al., 1990, 1992, 1995, 1998; Consiglioet al., 2004b; Henry, 1943, 1951; Moser, 1983, 1960). Theseworks are taxonomically and nomenclaturally discordant:Brandrud (1998) recognized 30 anthraquinonoid taxa insection Fulvi, and Brandrud et al. (1990, 1992, 1995, 1998)accepted c. 16 non-anthraquinonoid taxa; Bidaud et al.(1994, 2001, 2003, 2004) accepted 63 non-anthraquinonoidand 110 anthraquinonoid taxa and placed them in sectionsCalochroi, Fulvi, Laeticolores and Caerulescentes; andMoser (1960, 1983) accepted 32 non-anthraquinonoid and35 anthraquinonoid taxa and placed them in sectionsPhlegmacium, Multiformes, Calochroi, Coerulescentes,Scauri and Fulvi.

Most species are rare and have narrow ecological prefer-ences, and many are included in national red lists in Europe(Arnolds and Ommering, 1996; Bendiksen et al., 1998;Benkert et al., 1992; Courtecuisse, 1997; Gärdenfors, 2000;Stoltze and Pihl, 1998). However, as species recognition isproblematic, appropriate conservation measures are diY-cult to establish.

1.2. Species delimitation

Additional data are required to test the divergent taxo-nomies within section Calochroi, and to assess the pheno-typic plasticity of the morphological characters emphasizedby diVerent taxonomic schools, and to evaluate the mor-phological characters traditionally used for delimitation ofhigher taxa.

Many studies of Cortinarius have employed phyloge-netic analyses of nuclear ribosomal gene sequence data(Frøslev et al., 2005; Garnica et al., 2003b, 2005; Høilandand Holst-Jensen, 2000; Kytövuori et al., 2005; Liu et al.,1997; Moser and Peintner, 2002; Peintner et al., 2001, 2002,2003, 2004; Seidl, 2000). However, most of these includedfew collections within species.

Sequence data from three genetic markers (ITS, RPB2and RPB1) was analyzed in a multi-gene phylogeneticstudy of Cortinarius p.p. including several taxa of sectionCalochroi (Frøslev et al., 2005). Several (morphological)species were represented by more than one collection andall three markers showed little variation within these, con-

Wrming conspeciWcity, and showing that inference from ITSalone is indicative of the species level phylogenetic delimita-tions of multi-gene analyses. Despite the inherent impossi-bility of applying a strict phylogenetic species concept(sensu Taylor et al., 2000) using only one genetic marker,and the low power of phylogenetic resolution at higher lev-els, the ITS seems appropriate for delimiting phylogeneticspecies that comply with a morphological concept in Corti-narius. An example of this is the extensive study by Kytövu-ori et al. (2005) where three closely related species ofCortinarius subgenus Telamonia were delimited by phyloge-netic inference of ITS data and morphological examination.This study also showed the importance of includingsequence data from type specimens for settling nomencla-tural questions. Other recent studies have also successfullyapplied ITS as a species level marker in combination withevaluation of morphological data for diVerent groups (e.g.Cortinarius: (Garnica et al., 2003a,b; Peintner et al., 2003;Seidl, 2000); Hebeloma: (Aanen and Kuyper, 2004; Aanenet al., 2000); Hygrophorus: (Larsson and Jacobsson, 2004);Lepiota sensu lato: (Vellinga, 2003; Vellinga et al., 2003);Xerocomus: (Taylor et al., 2006)). Despite the lack of infor-mation on gene Xow between individuals of Cortinarius, itseems that the evaluation of morphological characters inlight of phylogenetic information from ITS (and othergenetic markers) generally provides species delimitationsthat are more coherent than delimitations based on the sub-jective evaluation of phenetic data alone.

Here we extend the sampling of taxa in section Calochroito represent most of the known European morphological,geographical and ecological diversity. We delimit species byinference and comparison of ITS sequence data in combi-nation with morphological analyses, and estimate the phy-logeny of species by inference of ITS, RPB1 and RPB2data. We examine the distribution and plasticity of mor-phological characters by mapping them on the resultingtree. We aimed to stabilize the nomenclature and to testvarious species concepts by sequencing holotypes and refer-ence collections.

2. Materials and methods

2.1. Taxon sampling

As many of the morphological characters of the treatedgroup are ephemeral and/or only present on very youngbasidiocarps, taxon sampling focused on well-documentedmaterial. One hundred and sixty-three collections made bythe two Wrst authors were selected to cover the largest pos-sible morphological and geographical span (withinEurope). These were selected to represent the morphologi-cal diversity of more than 700 collections. All relevant typespublished by Bidaud et al. (1994, 2001, 2003, 2004), Brand-rud et al. (1990, 1992, 1995, 1998), Consiglio et al. (2003,2004a,b), Delaporte and Eyssartier (2002), Eyssartier(2004), and Frøslev et al. (2006) were sampled, as well asmany types of the species published over the years by the

T.G. Frøslev et al. / Molecular Phylogenetics and Evolution 44 (2007) 217–227 219

workers, R. Henry, R. Kühner, M. M. Moser and P. D.Orton. Relevant collections published and illustrated inBidaud et al. (1994, 2001, 2003, 2004), Brandrud et al. (1990,1992, 1995, 1998), Consiglio et al. (2004b, 2003, 2004a) andEyssartier (2004) were included, and several collectionsfrom the herbaria of B. Dima, A. Bidaud, G. Eyssartier, R.Henry, P. Moënne-Loccoz and J. Vesterholt were included.Additional sequences belonging to the ingroup wereacquired from the sequence databases GenBank andUNITE (Kõljalg et al., 2005), and unpublished sequenceswere supplied by A. Taylor (Sveriges Lantbruksuniversitet,Uppsala, Sweden) and K. Liimatainen (University of Hel-sinki, Finland). Information on the sequences and collec-tions from which sequences were obtained is available assupplemental information (1).

2.2. Molecular analyses

Genomic DNA was extracted either with various CTABprotocols (e.g. Frøslev et al., 2005), with Plant Minikit(Qiagen) or MagAttract (Qiagen). The internal transcribedspacer (ITS) region was ampliWed and sequenced with theprimer combination ITS1F (Gardes and Bruns, 1993) andITS4 (White et al., 1990). The primer combinations ITS1F/ITS2 and ITS3/ITS4 were used on old and/or problematicmaterial. AmpliWcation and sequencing of the largest andsecond largest subunits of RNA polymerase II was carriedout according to Frøslev et al. (2005). Sequences wereassembled in Sequencher 3.11 (Gene Codes Corporation)or Vector NTI (Invitrogen Life Science Software). Thesequences have been deposited in GenBank (see supple-mental information (1) for accession numbers). Alignmentswere produced with MAFFT v5.6 (Katoh et al., 2005)under the settings E–INS–i. Due to the size of the ITS data-set, an initial alignment was produced combining severalseparate alignments of subsets, and a rough maximum par-simony analysis was carried out. Independent lineages ofidentical or near-identical sequences were identiWed, andrepresentative sequences from each lineage were thenselected. An alignment of the representative sequences wasproduced with MAFFT and the remaining (identical andnear-identical) sequences were added by eye. Cortinariuscaesiocortinatus was chosen as outgroup based on theresults of Frøslev et al. (2005).

Equally weighted maximum parsimony analyses wereperformed with PAUP* (SwoVord, 2003). A heuristicsearch strategy with TBR branch swapping, 100 randomaddition replicates (RAS), maxtrees of 5000, and steepestdescent and multiple trees in eVect, was employed treatinggaps as missing data. Bootstrap proportions (BS) wereassessed with 1000 replicates with 10 RAS and maxtrees of100, and otherwise as above.

A restricted ITS dataset with a representative sequencefor each species was combined with RPB1 and RPB2sequence data. A concatenated alignment was produced,and analyzed with MrBayes 3.1 (Huelsenbeck and Ron-quist, 2003). The combined dataset was assumed to have

three partitions (ITS, RPB1 and RPB2). Each partition wasallowed to evolve under a separate model including agamma shape parameter and a proportion of invariablesites. Two independent runs were carried out with 3,200,000generations and sampling every 100th generation. The last5000 trees sampled from each run were combined in a 50%majority rule consensus cladogram.

2.3. Morphological studies

All collections made by TF and/or TSJ were studied asfresh material, and detailed macroscopical descriptions weremade, and most collections photographed. Macro-chemicalalkaline reactions have been carried out with 2%, 10% and30% KOH on pileal and stipital surfaces as well as on thecontext. Microscopical features were studied in fresh or driedspecimens with material mounted in water or 2% KOH. Pri-mary speciWc determinations were carried out using pub-lished and unpublished determination keys, Xoras, researchpapers and internet resources (Bidaud et al., 1994, 2001, 2003,2004; Brandrud et al., 1990, 1992, 1995, 1998; Frøslev andJeppesen, 1999–2006; Frøslev et al., 2006; Moser, 1983, 1960;Moser and Peintner, 2002; Orton, 1955). Morphological andecological characters were mapped on the cladogram. Char-acters were mapped as present (Wlled circle) or absent (emptycircle). Character states observed in at least one basidiocarpfrom one collection were mapped as present. The followingcharacters were mapped: presence of: anthraquinonoid pig-ments (A), a pink alkaline reaction due to the presence ofsodagnitin (Sontag et al., 1999) on pileus (B), bulb (C), con-text (D), the presence of blue (E) and yellow (F) colors on thepileus, a pale/colorless pileus (G), presence of blue (H), andyellow (I) colors on the lamellae, pale/colorless lamellae (J),and association with frondose (K) and coniferous (L) trees.Characters present in all or nearly all species were notmapped (e.g. absence of hypoderm, glabrous and non-Wbrouspileus, abruptly bulbous stipe, amygdaliform to citriformspores, coarse ornamentation of spores). The type of anthra-quinonoid pigment were mapped according to Brandrud(1998): ATR, atrovirin; FLA, Xavomannin; FDM, Xavoman-nin 6,6�-dimethylether; FTM, Xavomannin 6,6�,8-trimethyle-ther; 4OH-FDM, 4OH- Xavomannin 6,6�-dimethylether;PHL, phlegmacin; PME, phlegmacin-8�-methylether.

2.4. Species delimitation and genetic variation

Independently evolving lineages with low internalgenetic variation were identiWed from the phylogeneticanalyses of the all-inclusive ITS alignment. Thereafter, cor-relations with other characters (e.g. morphology, ecology,chemical reactions) were studied to reach a consensus spe-cies delimitation based on currently available knowledge—i.e. a species concept that assumes a species to be the leastmonophyletic unit correlated with other characters.

The amount of genetic variation within and among spe-cies was assessed in terms of absolute nucleotide diVerenceswithin species and between the most closely related species.

220 T.G. Frøslev et al. / Molecular Phylogenetics and Evolution 44 (2007) 217–227

Pair wise genetic distances between all (complete)sequences in terms of absolute nucleotide diVerences werecalculated in PAUP*, treating gaps as missing data.

As the hyphae of Cortinarius and most other species ofagarics are dikaryotic and as ITS is a multi copy gene thepossibility of intragenomic polymorphism exists. Intrage-nomic polymorphisms have been reported from severalgroups of fungi (Aanen et al., 2001; Leonardi et al., 2005;Taylor et al., 2006). Intragenomic variation was assessed asmixed peaks in chromatographic data from an individual.

3. Results

3.1. Sequence data

Sequence data from the ITS region were obtained for295 samples including 52 types. Old and/or ill preservedmaterial often proved diYcult to amplify and sequence, andthe types of the following names did not yield usablesequence data: Cortinarius cuprescens, C. caesiolatens,C. caesiolatens var. subelegantissimus, C. elegantior var. bas-icroceus, C. fulgidus, C. nymphaecolor, C. pantherinus,C. parelegantior, C. parelegantior var. volvaceus, C. porri-color, C. prasinoides, C. rioussetorum, C. simillimus,C. splendens var. papillatosporus and C. subglobispermusfrom Bidaud et al. (deposited in G and PC), C. cupreorufusfrom Brandrud et al. (1995) (isotype deposited in S),C. aereus, C. eufulmineus, C. Xavovirens, C. olearioides,C. olivascentium, C. olivellus, C. orichalceolens, C. osful-gentis, C. parafulmineus, C. pelitnocephalus, C. pseudofulgens,C. pseudofulmineus, C. subelegantior and C. xanthochlorusfrom the herbarium of R. Henry (deposited in PC), C. citri-nus, C. caroviolaceus, C. subturbinatus and C. xanthochrousfrom the herbarium of P. D. Orton (deposited in K) andC. homomorphus from the herbarium of R. Kühner (depositedin G). The problematic collections were almost entirely anthra-quinonoid containing species or material older than 40yr.

More than 30% of the samples had intragenomic poly-morphisms in terms of base polymorphisms or of spacerlength polymorphisms consisting of single nucleotide or fewnucleotide indels. For a few species, multiple intragenomicspacer length polymorphisms were often simultaneouslypresent (e.g. Cortinarius xanthophyllus, C. elegantior) makingwhole regions of the ITS impossible to access by directsequencing. The sequences were combined with 126 earlierpublished and unpublished sequences including one typesequence for a total sequence dataset of 421 ingroupsequences. After trimming the ends, an alignment of 780positions was achieved. Six RPB1 and 28 RPB2 sequenceswere obtained (GenBank accessions: RPB1: DQ680081–DQ680086 and RPB2: DQ680087–DQ680113) and com-bined with sequences published in Frøslev et al. (2005).

3.2. ITS analyses and delimitation of species

The maximum parsimony analysis of the 421 ITSsequence data set yielded 21 equally parsimonious trees of

988 steps (CID0.388, RID0.925). One of the resulting phy-lograms is shown in Fig. 1. The majority of the sequenceswere distributed in well-separated phylogenetic units of lowinternal genetic variation, and normally low morphologicalvariation. In total 80 (–81) such phylogenetic species wereidentiWed. Sixty-seven phylogenetic species were repre-sented by several (two to 17) sequences. Nine phylogeneticspecies receive bootstrap support below 90%. Disregardingthe divergent North American sequences assigned to Corti-narius aureopulverulentus, the bootstrap support of this spe-cies is 82%. The three species that receive support below BS70% all belong to trios of closely related species (i.e. theC. quercus-ilicis group and the C. saporatus group). Corre-lation with morphological characters was generally highand allowed for morphological characterization and delim-itation of all but a few phylogenetic species based oncurrent morphological knowledge. The morphospeciesC. ionochlorus and C. atrovirens were ITS identical, as wereC. xanthophyllus and C. claroXavus. For 68 species we wereable to Wnd a name, which we Wnd applicable under the cur-rent botanical code. For the remaining 13 spp., we haveprovided the provisional names: Cortinarius osloensis inedand C. sp.1–C. sp.12. These species are indicated and namedon the tree (Fig. 1) and in the supplemental information (1).

3.3. Genetic variation and diVerentiation of the ITS region

The 65,703 pair wise genetic distances for completesequences are illustrated in Fig. 2. Genetic distance shows anormal distribution with an average of 38 nucleotide diVer-ences (corresponding to a genetic distance of 0.059 using aKimura 2-parameter model of DNA evolution). The localmaximum at zero nucleotide diVerences relates to the highnumber of identical sequences.

Most species represented by two or more sequences weregenetically polymorphic. In most species, polymorphismswere (also or only) observed intragenomically. Several sin-gle sequences acquired from sequence databases had one ora few unique nucleotide diVerences. These diVerences areincluded in this analysis but were not considered for closeevaluation of genetic diVerentiation. For samples where wehave been able to check chromatographic data, the largestinfraspeciWc variation was seen within C. platypus, whichhad six polymorphic sites, three of which were also seen asintragenomic polymorphisms. Three polymorphic positionsconsistently separated the species into two groups. Cortina-rius lilacinovelatus had Wve polymorphic positions, three ofwhich were also observed intragenomically, and two poly-morphic positions that consistently separated the speciesinto two groups. Several other species had one to threepolymorphic sites of which several were often intrage-nomic—e.g. Cortinarius humolens, which had three poly-morphic sites, all of which were also observedintragenomically.

Most phylogenetic species were separated from eachother by at least 12 (average 38) nucleotide diVerences(Fig. 2), except for the two species pairs Cortinarius

T.G. Frøslev et al. / Molecular Phylogenetics and Evolution 44 (2007) 217–227 221

luteolus/Cortinarius odorifer separated by Wve nucleotides,and Cortinarius sulphurinus/Cortinarius verrucisporus (sixnucleotides) and the two species trios: Cortinarius quercus-ilicis/C. murellensis/C. sp.11 (three to four nucleotides) andC. saporatus/Cortinarius caroviolaceus/C. sp.12 (six to ninenucleotides). Thus, with the exception of a few instances,the (maximum) infraspeciWc distances were well below the(minimum) interspeciWc distances.

3.4. Multi-gene phylogenetic analyses

The two independent Bayesian runs reached a standarddeviation of split frequencies of 0.02 after about 2,000,000generations. Trees sampled during the last 500,000 genera-tions were almost identical between the two runs, andshowed only minor diVerences (see below) and were com-bined for the consensus cladogram. A 50% majority rule

Fig. 1. Phylogenetic species in the /Calochroi clade (Cortinarius section Calochroi). One of 21 equally parsimonious trees based on maximum parsimonyanalysis of 421 ITS sequences. Bootstrap proportions above 50 are indicated above branches. Branches collapsing in the strict consensus tree are markedwith an asterisk. After each taxon label the number of sequences belonging to that lineage is indicated. Two branches have been truncated.

C. aureofulvus (3)

C. osmophorus (5)

C. flavovirens (8)

C. humolens (8)C. sp.2 (1)

C. pseudoglaucopus (3)C. suaveolens (2)

C. xanthophyllus (9)

C. cedretorum (4)

C. xanthochlorus (8)

C. sp.4 (1)C. prasinus (8)

C. rufoolivaceus (4)C. cupreorufus (3)

C. langeorum (4)C. osloensis (2)

C. citrinus (9)

C. saporatus (8)

C. caroviolaceus (3)C. sp.12 (3)

C. luteolus (4)C. odorifer (4)

C. elegantissimus (9)

C. rapaceotomentosus (3)C. sp.3 (1)

C. sulphurinus (10)

C. sp.1 (1) USA

C. alcalinophilus (16)

C. splendificus (4)

C. olearioides (9)

C. quercus-ilicis (13)

C. arcuatorum (7)C. sp.10 (1) USA

C. dibaphus (6)

C. atrovirens (15)

C. aureocalceolatus (6)

C. odoratus (4)

C. splendens (5)

C. meinhardii (4)C. moseri (1)

C. aureopulverulentus (6)

C. napus (2)

C. natalis (7)C. caesiocortinatus

1 change

C. elegantior (10)

Continued

*

*

*

**

100

100

100

100

96

99

100

100

94

94100

94

100100

65

8971

7790

100

100

90

97

100

100

85

100

100

100

100

99

100

98

82

99

100

100

61

61

71

9999

6461

63

93

93

5964

100

73

100

60

(USA)

**

*

*

*

*

*

*

*

52

75

*

*

*

53

100C. sp.11 (1)

C. murellensis (2)

C. albertii (7)

C. sp.6 (1)

C. corrosus (4)

C. insignibulbus (4)

C. parasuaveolens (11)

C. sancti-felicis (2)C. dalecarlicus (1)

C. magnivelatus (2) USA

C. caesiocinctus (4)

C. lilacinovelatus (12)

C. selandicus (2)

C. calochrous (13)

C. sp.5 (1)C. sp.7 (3)

C. piceae (5)

C. vesterholtii (6)

C. chailluzii (2)C. subgracilis (2)

C. haasii (9)

C. splendidior (3)C. sublilacinopes (4)

C. catharinae (16)

C. cisticola (8)

C. platypus (12)b(3)

C. sp.9 (2)

C. fulvocitrinus (4)

C. nymphicolor (9)

C. sp.8 (3)

C. sodagnitus (8)

C. barbaricus (1)

C. barbarorum (5)C. verrucisporus (1) USA

C. ochraceopallescens (8)

*

*

*

**

*

*

100

97

99

95

97

91

87

94

8599

99

100

100

100

100

95

91

100100

99

100

97

98

85

97

89

97

100

63

87

66

83

63

96

a(8)

*

*

*

59

53

222 T.G. Frøslev et al. / Molecular Phylogenetics and Evolution 44 (2007) 217–227

cladogram of the three-gene analyses is presented in Fig. 3.Morphological and ecological characters are indicated foreach species, as well as the subclades and groupings dis-cussed below. Some of the groups that correspond more orless to classical taxonomic entities are indicated.

Several groupings were relatively well supported andcorresponded to those in the less extensive analyses inFrøslev et al. (2005), but several relationships receiving lesssupport diVer between the two studies. Several of the spe-cies, that appear unresolved in the base of the cladogramlack information from at least one genetic region.

The /calochroid subclade is supported by .58 BPP. Mostnon-anthraquinonoid species are found in this derived cladecorresponding to section Calochroi in the old sense as cover-ing only non-anthraquinonoid species. All 13 newly samplednon-anthraquinonoid species were placed in this subclade.The small /sulphurinus subclade is composed of two mor-phologically similar European species and a sequestrateNorth American species of unknown pigment chemistry. The/rufoolivacei subclade received little support (.74 BPP) and isa morphologically heterogeneous group. All species treatedin section Fulvi subsection Rufoolivacei (Brandrud, 1998)were placed in the /rufoolivacei subclade except Cortinariusaureofulvus. The /fulvi subclade is well delimited (1.00 BPP)and correspond to subsection Elegantiores (Brandrud, 1998),and the /arcuatorum subclade (1.00 BPP) is composed of asmall morphologically well deWned group of species aroundCortinarius arcuatorum. These two are furthermore well sup-ported (.98 BPP) as sister groups.

4. Discussion

4.1. Species delimitation

The species as distinguished with ITS data are welldelimited in the sense that they all represent independent

phylogenetic lineages supported by bootstrap values of 63or more, and in most cases 100 or just below. All phyloge-netic species can also be delimited by other characters (e.g.morphology, ecology, chemistry). The majority show sig-niWcantly lower (maximum) infraspeciWc genetic distancesthan the observed (minimum) interspeciWc distances. Forspecies where seven or more collections were sampled, mostinfraspeciWc polymorphisms were also observed as intrage-nomic polymorphisms. These polymorphic individuals mayrepresent heterozygous individuals in a large population.

For the anthraquinonoid species, the broad species con-cept applied in Brandrud (1998) and in Brandrud et al.(1990, 1992, 1995, 1998) based on pigment chemistry andmorphology, is well supported by the phylogenetic Wndingsof the present study. In contrast, a narrower species conceptbased on emphasis of extreme character states (Bidaudet al., 2003, 2004) is not supported.

The non-anthraquinonoid group is primarily composedof species with a calochroid appearance (a more or less yel-lowish-to pale or bluish-pileal coloration, varying degreesof violaceous coloration on the lamellae and stipe, palecontext, a broad marginate bulb, a simplex pileipellis andcoarsely ornamented spores). The unambiguous phyloge-netic separation of most taxa allows for a re-evaluation ofdiagnostic characters, and it seems evident that many char-acters have been over- and underemphasized. The pheno-typic plasticity especially of basidiocarp coloration and sizeoften exceeds diVerences between phylogenetically separatespecies, and thus poses challenges for morphological identi-Wcation of species and higher taxa.

The present delimitation of the non-anthraquinonoid spe-cies deviates signiWcantly from those in Brandrud et al. (1990,1992, 1995, 1998) and in Bidaud et al. (1994, 2001), but liessomewhat closer to that in Moser (1983). Several of the mor-phospecies accepted and often treated in diVerent sections,subsections and series in Bidaud et al. (1994, 2001) are

Fig. 2. Pairwise genetic distances (n D 65703) between ITS sequence samples in terms absolute nucleotide diVerences for 363 samples.

0

500

1000

1500

2000

2500

3000

3500

0 10 15 20 25 30 35 40 45 50 55 60 65 70

Absolute pairwise distance (nucleotide differences)

Num

ber o

f pai

rs

5

T.G. Frøslev et al. / Molecular Phylogenetics and Evolution 44 (2007) 217–227 223

less lamellae, (K) association with frondose trees, (L) association with conifer

ous trees.

Fig. 3. Phylogenetic relationships of species in the /Calochroi clade (Cortinarius section Calochroi) based on Bayesian analyses of ITS, RPB1 and RPB2data. Fifty percent majority rule consensus cladogram based on 10,000 trees sampled in two Bayesian runs. Bayesian posterior probabilities are indicatedabove branches. Clades and groups discussed in the text are indicated, as well as major non-anthraquinonoid lineages as “N.A.”. Morphological and eco-logical characters ware mapped as present (Wlled circle) or absent (empty cirlce): (A) anthraquinonoid pigments—pigment type were mapped according toBrandrud (1998): ATR, atrovirin; FLA, Xavomannin; FDM, Xavomannin 6,6�-dimethylether; FTM, Xavomannin 6,6�,8-trimethylether; 4OH-FDM, 4OH-Xavomannin 6,6�-dimethylether; PHL, phlegmacin; PME, phlegmacin-8�-methylether. (B) a pink alkaline reaction with sodagnitin on pileus, (C) d.o. onthe bulb, (D) d.o. on the context, (E) pileus with blue, (F) pileus with yellow, (G) pale/colorless pileus, (H) blue lamellae, (I) yellow lamellae, (J) pale/color-

C. albertiiC. barbaricusC. barbarorumC. caesiocinctusC. selandicusC. lilacinovelatusC. catharinaeC. sp.6C. nymphicolorC. sp.8C. platypusC. sp.9C. calochrousC. sp.5C. sublilacinopesC. sp.7C. cisticolaC. haasiiC. splendidiorC. chailluziiC. subgracilisC. piceaeC. dalecarlicusC. fulvocitrinusC. sodagnitusC. insignibulbusC. parasuaveolensC. sancti-felicisC. magnivelatus USAC. corrosusC. ochraceopallescensC. verrucisporus USAC. vesterholtiiC. osmophorusC. aureofulvusC. citrinusC. flavovirensC. sp.3C. sulphurinusC. sp.1 USAC. caroviolaceusC. saporatusC. sp.12C. osloensisC. humolensC. sp.2C. pseudoglaucopusC. sp.4C. cedretorumC. cupreorufusC. elegantissimus

C. langeorum

C. luteolusC. odoriferC. prasinusC. rufoolivaceusC. rapaceotomentosusC. suaveolens

C. xanthophyllus

C. xanthochlorusC. natalisC. alcalinophilusC. splendificusC. murellensisC. sp.11C. quercus-ilicisC. elegantiorC. olearioidesC. arcuatorumC. sp.10 USAC. dibaphusC. napusC. aureocalceolatusC. aureopulverolentusC. moseriC. atrovirensC. odoratusC. splendensC. meinhardiiC. caesiocortinatus

.74

.79

1.00

.81

.87

.58

.90

.83

.57

.59

.89

.90

.84.68

100

.87

.58

.83.92

.99

.83 1.00

.97

.95

.87

1.001.00

.88

1.00.99

.74

.98

.961.00

.84

1.00.68

1.00

1.00

.98

1.00

1.00

.92.99

1.00

1.001.00

FT

M &

FD

MP

HL

/calochroid

/rufoolivacei

/fulvi

/arcuatorum

/sulphurinus

FT

M

FTM?

AT

R&

FLA

4OH

-F

DM

PHLFDM

FDM

FDM?

?

PMEPME?PME

=sect. Fulvi subsect. Elegantiores

s. Brandrud 1998

sect. Calochrois. auct

sect. Fulvisubsect. Rufoolivacei

s. Brandrud 1998

N.A.

N.A.

N.A.

hostlamellacolor

pileuscolor

sodagnitinreaction

B C D E F G H I J K LA

pile

usbu

lbco

ntex

t

blue

yello

wpa

le

blue

yello

wpa

le

frond

.co

nif.

anth

raqu

in.

sect. Laetocoloress. Bidaud et al. 2004

=sect. Fulvi s. Bidaud et al. 1998

((

))

I

II

III

224 T.G. Frøslev et al. / Molecular Phylogenetics and Evolution 44 (2007) 217–227

phylogenetically indistinguishable. Most of these relate to thelevel of plasticity within color states especially of the pileusand the stipe.

The infrageneric groupings of Bidaud et al. (1994, 2001,2003, 2004) are almost entirely based on very sharp andnarrow (often overlapping) cut-oV values/intervals forbasidiocarp and basidiospore size, and pileal and stipitalcoloration. One of the weaknesses of this approach is thatspecies with a variable size and coloration can be perceivedas several species belonging to diVerent infra-generic taxa.For example: Cortinarius albolutescens treated in subsec-tion Calochroi serie Platypus is ITS identical to collectionsof C. catharinae, C. pseudoparvus and C. pallens placed insubsection Calochroi serie Parvus, and C. calochrous in sub-section Calochroi serie calochrous, and C. xanthochrous insubsection Arquati serie arquatus. Conversely, the over-looked subtle diVerences between similar species haveresulted in collections of separate species being assigned tothe same epithet, C. insignibulbus, C. ochraceopallescensand C. catharinae being assigned to C. calochrous forexample.

Generally, only a few morphospecies, which are widelyaccepted as separate species turned out to be phylogeneti-cally indistinguishable. Cortinarius atrovirens and C. ionoch-lorus are usually distinguished by their ecology anddiVerent coloration of the lamellae. Similarly, what hasbeen determined as either C. xanthophyllus or C. claroXavus(non sensu Moser), based on the presence of a blue (sub-)cuticular pigmentation, possess identical ITS sequences. Inboth cases, the morphological characters used to separatethe taxa are few and/or of a continuous nature. Similarly,Cortinarius natalis exhibits a remarkable morphologicalvariability, but our current morphological knowledge doesnot allow for unambiguous separation of this geneticallyhomogeneous group into distinct morphological groups.These few instances, where morphology or other charactersindicate the potential existence of ITS-identical but mor-phologically separate species, present obvious cases forevaluating the general applicability of ITS as a speciesmarker in the group with more variable genetic markers,denser sampling, and detailed morphological studies.

Conversely, some lineages showed a relatively high levelof phylogenetic variation, which was not correlated unam-biguously with other characters. Three of six polymorphicsites consistently separated C. platypus into two groups(marked a and b in Fig. 1). Although the species show ahigh level of morphological variation, this did not seem tobe correlated with phylogenetic separation.

ITS sequence data were obtained from 52 holotype spec-imens. Thirty-seven of the phylogenetic species found con-tained at least one type sequence. Thirty holotypesequences were either identical to sequences of older holo-types (with nomenclatural priority) or were unambiguouslyassigned to species with older (un-sequenced) types con-nected to diagnoses that do not contradict conspeciWcity.Most of these types can therefore be readily treated assynonyms.

4.2. Phylogenetic relationships and morphological characters

The phenotypic plasticity of morphological characterswas generally high (which also counts for many non-mapped characters which are not easily scored as present/absent—e.g. presence of scales on cap, basidiocarp size,variations in color intensity and hue etc.). Nonetheless,combinations of morphological, chemical and ecologicalcharacters can delimit most species. But, while most speciesand several monophyletic groups can be recognized bymorphological and ecological character states, these samestates are poorer indicators of higher-level taxa.

Section Calochroi as a whole is phylogenetically welldelimited, and species share several morphological and eco-logical characters with only a few exceptions (these are notmapped on Fig. 3). Stipes are bulbous with a more or lessabruptly marginate bulb. Cap cuticles are rarely innatelyWbrillose (except for e.g. older specimens of C. xanthochlo-rus). Basidiospores are amygdaliform to citriform with acoarse ornamentation. Pileipellis is of the simplex type(sensu Brandrud et al., 1990—i.e. lacks a well-diVerentiatedhypoderm of inXated cells) except for C. aureocalceolatus.Many species are thermophilous and grow in more or lesscalcareous soils, and the main mycorrhizal host tree speciesare to be found in the Fagales or Pinaceae.

The presence/absence and the type of anthraquinonoidpigments seem to be apomorphic characters for several lin-eages and constant for species. The largest non-anthraqui-nonoid lineage (i.e. the /calochroid subclade containingonly one anthraquinonoid species) corresponds more orless to section Calochroi in the old sense. Several non-anthraquinonoid species, sometimes treated in sectionCalochroi s. auct. (e.g. C. arcuatorum, C. dibaphus, C. suaveo-lens and C. saporatus) (e.g. Brandrud et al., 1990, 1992,1995, 1998), do not belong in the /calochroid subclade, andthe majority of these also deviate from the general mor-phology exhibited by the species in that subclade.

In the Bayesian consensus cladogram the species treatedin section Fulvi subsection Rufoolivacei (Brandrud, 1998) orsection Laeticolores (Bidaud et al., 2004) do not form amonophyletic group, although results from the two sepa-rate Bayesian runs (data not shown) were indicative of this.All species with the pigment rufoolivacin (except C. aureo-fulvus) were thus placed in a derived lineage along withC. rapaceotomentosus except for C. sp.4. The subclade alsocontains several non-anthraquinonoid species and most ofthese are found in the derived group III of along with somerelatively pigment poor anthraquinonoid species (i.e.C. humolens, C. osloensis and C. sp.2) containing the pig-ment Xavomannin-trimethylether (FTM). At least the twoEuropean taxa in the /sulphurinus subclade most probablyhave plegmacin-8�-methylether (PME) as main pigment,but share this with the seemingly unrelated C. Xavovirens.The /fulvi subclade exclusively contains species with bothXavomannin-dimethylether (FDM) and -trimethylether(FTM). Species with either only FDM (C. fulvocitrinus,C. citrinus and C. xanthochlorus) or FTM (C. osloensis and

T.G. Frøslev et al. / Molecular Phylogenetics and Evolution 44 (2007) 217–227 225

C. humolens), species with 4OH-FDM (C. splendens and C.meinhardii) and species with both atrovirin and Xavoman-nin (C. odoratus and C. atrovirens) were not supported asmonophyletic entities. The alkaline reactions of the anthra-quinonoid species are related to the type of pigment presentand thus carry much the same information, but diVerencesin topology and intensity of reactions can distinguishbetween similar species in several instances (data notshown). The morphologically similar and anthraquinonoidpigment rich species pairs (C. atrovirens/C. odoratus and C.meinhardii/C. splendens) appear more or less unresolved atthe base of the cladogram. This may indicate that the pos-session of pigments in large quantities is a primitive statefor species in section Calochroi—a hypothesis that also canbe established from the potential evolution of pigmentcharacters (Brandrud, pers. comm.).

The presence (and topology) of sodagnitin as indicatedby a pink alkaline reaction likewise seems to be phylogenet-ically informative and constant for species. The presence onsurfaces (cap and/or bulb) is (almost) restricted to species inthe /calochroid subclade. One major (little supported) line-age, group I, is characterized by the presence of a pink alka-line reaction, whereas it is absent from species in group II.A few smaller groups within the /calochroid subclade alsoseem to be characterized by either the presence or absenceof this pigment. The well supported /arcuatorum subcladecontains a few species sharing the apomorphic presence ofa pink alkaline reaction in the context. The two Europeanspecies in the /sulphurinus subclade are characterized by apink reaction on the mycelial strands most probably causedby sodagnitin. Otherwise sodagnitin seems to be absentfrom the anthraquinonoid species.

The homoplasy and phenotypic plasticity of pileus andlamellae coloration is high for species in the /calochroidsubclade. Group II and I are, however, mainly character-ized by species with a yellow pileus (and absence of bluecolor) and species without yellow pileus (but presence ofblue color), respectively. A relatively high level of variationin intensity and occasional absence of these colors was,however, observed for most species. In the part of the clad-ogram containing mainly anthraquinonoid species the col-oration of pileus and lamellae is predominantly in hues ofyellow. The coloration relates to the type of anthraquino-noid pigment and thus carries similar information as thepigment type.

4.3. Host preference and distribution

Although most species are primarily found associatedwith only one or a few host tree species (and genera), hostpreference for either coniferous vs. frondose trees generallyseems to be speciWc, with the exceptions C. dibaphus, C. ele-gantior and (possibly) C. haasii. The majority of the speciesgrow with tree species of either Pinaceae or Fagales, butmost species also form mycorrhizas with other hosts (e.g.Corylus, Carpinus, Tilia, Dryas, Helianthemum, Cistus).Cortinarius moseri is the only species in section Calochroi

that is assumed to grow exclusively with Alnus. Most spe-cies grow with frondose trees, and the choice of host seemsto be of little value for delimiting higher taxa.

Most species identiWed here have widespread distribu-tions in Europe. For widely sampled species, sequencesfrom diVerent places in Europe are identical or near identi-cal in ITS (and morphology), and variations do not seem tobe correlated. Less densely sampled species may artiWciallyappear to have restricted distributions, but a few species arein fact only known from restricted areas. Cortinarius splen-didior, C. murellensis and C. natalis are only known fromthe Mediterranean area, and their distribution seem to becorrelated with the distribution of sclerophyllous Quercusspecies.

4.4. Species diversity

It is evident from these results that section Calochroi inEurope alone is more speciose than most agaricalean gen-era, although it is only one of many lineages in the genusCortinarius. The majority of the species in section Calochroiare moderately to extremely rare, and although we havetried to cover the morphological diversity in Europe, thereare unsampled species. However, by including numerouscollections/types of other workers and all available ingroupsequences from public databases, only very few lineagesunsampled by us were encountered. Therefore, we feel con-Wdent that this sample describes the bulk of European vari-ation belonging to section Calochroi. The existence of atleast 75 species in Europe is indicated by our study. Thiscontrasts to the c.170 proposed by others (Bidaud et al.,1994, 2001, 2003, 2004), notwithstanding that several of thespecies accepted here are not included in these treatments.

The extra-European diversity is largely unsampled inthis study as the focus was on the taxonomy of Europeantaxa. Five of the six extra-European sequences belong tofour separate unique non-European lineages, indicatingthat the number of species on a global scale is possiblymuch greater.

4.5. Barcoding Cortinarius

The use of single DNA sequences as species identiWershas lately received much attention, and has been termed“barcoding” (Hebert et al., 2003). Considerable concernhas, however, dealt with whether infraspeciWc variationoverlaps with interspeciWc variation and thereby introduc-ing the possibility of erroneously assigning a sequence to awrong taxon (Mayer and Paulay, 2005). The two distribu-tions in the pair wise genetic distances (Fig. 2) hardly showany overlap, and it is tempting to infer this as the gapbetween infraspeciWc and interspeciWc diVerences (some-times termed the barcoding gap). However, four groups ofphylogenetically distinct species show interspeciWc dis-tances below this threshold, C. luteolus/C. odorifer, C. sapo-ratus/C. sp.12 and C. sulphurinus/C. verrucisporus and C.quercus-ilicis/C. murellensis/C. sp.11. Although this narrows

226 T.G. Frøslev et al. / Molecular Phylogenetics and Evolution 44 (2007) 217–227

the barcoding gap considerably, there is still no signiWcantoverlap the observed inter- and infraspeciWc distances. Thetwo cases of ITS identical morphospecies, C. atrovirens(incl. C. ionochlorus) and C. xanthophyllus (incl. C. claroXa-vus), also need to be examined in this context (see above).All of the most polymorphic species do not have veryclosely related sister-taxa, and therefore do not present aproblem for barcoding. BLAST-searches, clustering algo-rithms or phylogenetic analysis will thus assign any ingroupITS sequence to the correct (phylogenetic) species.

5. Conclusions

This study presents the Wrst thorough sampling of taxain the species rich /Calochroi clade (D section Calochroi),and the Wrst molecular phylogenetic evaluation of thewidely divergent proposed taxonomies. Although smallerthan some estimates, the number of species (more than 70)in section Calochroi in Europe is greater than many basid-iomycetous genera, with indications of a much larger globaldiversity. Most species are rare, but seem to be widespreadand genetically and morphologically homogeneousthroughout their European distribution. Despite thishomogeneity, the phenotypic plasticity of many morpho-logical characters has been underestimated or misinter-preted resulting in the publication of many synonymousnames and misidentiWcations, and the high level of mor-phological homoplasy has resulted in the acceptance ofmany unnatural higher taxa. ITS data are presented for 22species not previously sequenced, and the known geneticvariation is extended for most included taxa. Thirty of theincluded 53 type sequences are identical to other taxa, andas morphological characters used to delimit these speciesare rather vague in most cases, many can readily be treatedas synonyms. Our Wndings emphasize the importance ofcritical analysis of collections before describing new taxa.No signiWcant overlap was found between intra- and inter-speciWc genetic distances indicating that ITS can functionas a sequence “barcode” for the treated group.

Acknowledgments

We thank Andre Bidaud, Tor Erik Brandrud, GiovanniConsiglio, Balint Dima, Guillaume Eyssartier, Jacob Heil-mann-Clausen, Ilkka Kytövuori, Kare Liimatainen, PierreMoënne-Loccoz, Tuula Niskanen, Ursula Peintner, JuhaniRuotsalainen, Andy Taylor and Jan Vesterholt for provid-ing collections and/or sequences for this study. We thankthe curators of the herbaria in Copenhagen (C), Geneva(G), Innsbruck (IB), Kew (K) Stockholm (S) and Paris (PC)for arranging loans and Bart Buyck for his hospitality. Wethank the Jacob E. Lange fund for supporting the Weldworkof TF and TSJ. We thank Andy Taylor for linguisticimprovements and comments on the manuscript. Further-more we thank Tor Erik Brandrud for valuable discussions,David S. Hibbett, Duur K. Aanen, Ole Seberg and twoanonymous reviewers for comments on early versions of

the manuscript. This study was partly funded by a facultyscholarship to TF from The Faculty of Science, Universityof Copenhagen, a stipend from the International School ofBiodiversity Sciences (ISOBIS) to TF and a SYNTHESYSgrant for TF to visit the herbarium in Paris.

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.ympev.2006.11.013.

References

Aanen, D.K., Kuyper, T.W., 2004. A comparison of the application of abiological and phenetic species concept in the Hebeloma crustulini-forme complex within a phylogenetic framework. Persoonia 18,285–316.

Aanen, D.K., Kuyper, T.W., Boekhout, T., Hoekstra, R.F., 2000. Phyloge-netic relationships in the genus Hebeloma based on ITS1 and 2sequences, with special emphasis on the Hebeloma crustuliniforme com-plex. Mycologia 92, 269–281.

Aanen, D.K., Kuyper, T.W., Hoekstra, R.F., 2001. A widely distributedITS polymorphism within a biological species of the ectomycorrhizalfungus Hebeloma velutipes. Mycological Research 105, 284–290.

Arnolds, E., Ommering, G., 1996. Bedreigde en kwestbare paddestoelen inNederland. Rapport IKC Natuurbeheer nr. 24, Wageningen.

Bendiksen, E., Højland, K., Brandrud, T.E., Jordal, J.B., 1998. Truede ogsårbare sopparter i Norge—en kommentert rødliste. FungiXora, Oslo.

Benkert, D., Dörfelt, H., Hardtke, H.-J., Hirsch, G., Kreisel, H., Krieglste-iner, G.J., Lüderitz, M., Runge, A., Schmid, H., Schmitt, J.A., Winter-hoV, W., Wöldecke, K., Zehfuß, H.-D., Einhellinger, A., Groß, G.,Große-Brauckmann, H., Nuss, I., Wölfel, G., 1992. Rote Liste dergefährdeten Großpilze in Deutschland. - Deutsche Gesellschaft fürMykologie e.V. & Naturschutzbund Deutschland e.V. (NABU) (Hrsg.).IHW-Verlag, Eching.

Bidaud, A., Carteret, X., Eyssartier, G., Moënne-Loccoz, P., Reumaux, P.,2003. Atlas des Cortinaires. Pars XIII (1). Éditions Fédération myco-logique Dauphiné-Savoie, Lyon.

Bidaud, A., Carteret, X., Eyssartier, G., Moënne-loccoz, P., Reumaux, P.,2004. Atlas des Cortinaires. Pars XIV (2). Éditions Fédération myco-logique Dauphiné-Savoie, Lyon.

Bidaud, A., Moënne-Loccoz, P., Reumaux, P., Carteret, X., Eyssartier, G.,2001. Atlas des Cortinaires XI (2). Éditions Fédération mycologiqueDauphiné-Savoie, Lyon.

Bidaud, A., Moënne-Loccoz, P., Reumaux, P., Henry, R., 1994. Atlas desCortinaires. Pars VI. Éditions Fédération mycologique Dauphiné-Savoie, Annecy.

Brandrud, T.E., 1996. Cortinarius subgenus Phlegmacium section Phlegma-cium in Europe. Descriptive part. Edinburgh Journal of Botany 53,331–400.

Brandrud, T.E., 1998. Cortinarius subgen. Phlegmacium sect. Fulvi—chemo-taxonomy versus morphological taxonomy. Journal des J.E.C. 0, 4–9.

Brandrud, T.E., Lindström, H., Marklund, H., Melot, J., Muskos, S., 1990.Cortinarius Flora Photographica I (English version). Cortinarius HB,Matfors.

Brandrud, T.E., Lindström, H., Marklund, H., Melot, J., Muskos, S., 1992.Cortinarius Flora Photographica II (English version). Cortinarius HB,Matfors.

Brandrud, T.E., Lindström, H., Marklund, H., Melot, J., Muskos, S., 1995.Cortinarius Flora Photographica III (English version). Cortinarius HB,Matfors.

Brandrud, T.E., Lindström, H., Marklund, H., Melot, J., Muskos, S., 1998.Cortinarius Flora Photographica IV (English version). Cortinarius HB,Matfors.

T.G. Frøslev et al. / Molecular Phylogenetics and Evolution 44 (2007) 217–227 227

Consiglio, G., Antonini, D., Antonini, M., 2003. Il Genere Cortinarius inItalia I. Associazione Micologica Bresadola, Brescia.

Consiglio, G., Antonini, D., Antonini, M., 2004a. Il Genere Cortinarius inItalia II. Associazione Micologica Bresadola, Brescia.

Consiglio, G., Bidaud, A., Antonini, D., Antonini, M., La Rocca, S., 2004b.Il Genere Cortinarius in Italia. Parte terza. Alcune specie interessantidella sezione Calochroi. Bollettino dell’Associazione Micologica edEcologica Romana 61, 3–43.

Courtecuisse, R., 1997. Liste rouge des champignons menaces de la regionNord - Pas-de-Calais. Cryptogamie Mycologie 18, 183–219.

Delaporte, A., Eyssartier, G., 2002. Cortinarius rapaceotomentosus sp. nov.,un nuveau cortinaire proche de Cortinarius europaeus. Bulletin de laSociété Mycologique de France 118, 11–18.

Eyssartier, G., 2004. Notes sur cortinaires et psathyrelles rares et nouvelles.Fungi non Delineati 28, 1–55.

Frøslev, T.G., Jeppesen, T.S., 1999-2006. http://www.cortinarius.com.Frøslev, T.G., Jeppesen, T.S., Læssøe, T., 2006. Seven new calochroid and

fulvoid species of Cortinarius. Mycological Research 110, 1148–1160.Frøslev, T.G., Matheny, P.B., Hibbett, D., 2005. Lower level relationships

in the mushroom genus Cortinarius (Basidiomycota, Agaricales): acomparison of RPB1, RPB2, and ITS phylogenies. Molecular Phyloge-netics and Evolution 37, 602–618.

Gardes, M., Bruns, T.D., 1993. ITS primers with enhanced speciWcity forbasidiomycetes—application to the identiWcation of mycorrhizae andrusts. Molecular Ecology 2, 113–118.

Garnica, S., Weiss, M., Oberwinkler, F., 2003a. Morphological and molec-ular phylogenetic studies in South American Cortinarius species.Mycological Research 107, 1143–1156.

Garnica, S., Weiss, M., Oertel, B., Oberwinkler, F., 2003b. Phylogeneticrelationships of European Phlegmacium species (Cortinarius, Agarica-les). Mycologia 95, 1155–1170.

Garnica, S., Weiss, M., Oertel, B., Oberwinkler, F., 2005. A framework fora phylogenetic classiWcation on the genus Cortinarius (Basidiomycota,Agaricales) derived from morphological and molecular data. CanadianJournal of Botany 83, 1457–1477.

Gärdenfors, U., 2000. Rödlistade arter i Sverige—The 2000 Red List ofSwedish Species. ArtDatabanken, SLU, Uppsala.

Hebert, P.D.N., Cywinska, A., Ball, S.L., DeWaard, J.R., 2003. BiologicalidentiWcations through DNA barcodes. Proceedings of the Royal Soci-ety B: Biological Sciences 270, 313–321.

Henry, R., 1943. Essai d’une clé dichotomique analytique provisoire desti-nee aI faciliter l’étude des Cortinaires du groupe des Scauri. Revue deMycologie 8 (suppl.), 1–56.

Henry, R., 1951. Les Scauri. Bulletin de la société mycologique de France67, 225–322.

Huelsenbeck, J.P., Ronquist, F., 2003. MRBAYES 3: Bayesian phyloge-netic inference under mixed models. Bioinformatics 19, 1572–1574.

Høiland, K., Holst-Jensen, A., 2000. Cortinarius phylogeny and possible tax-onomic implications of ITS rDNA sequences. Mycologia 92, 694–710.

Katoh, K., Kuma, K., Toh, H., Miyata, T., 2005. MAFFT version 5:improvement in accuracy of multiple sequence alignment. NucleicAcids Research 33, 511–518.

Kirk, P.M., Cannon, P.F., David, J.C., Stalpers, J., 2001. Ainsworth andBisby’s Dictionary of the Fungi, ninth ed. CAB International, Walling-ford, UK.

Kytövuori, I., Niskanen, T., Liimatainen, K., Lindström, H., 2005. Cortina-rius sordidemaculatus and two new related species, C. anisatus and C.neofurvolaesus, in Fennoscandia (Basidiomycota, Agaricales). Karste-nia 45, 33–49.

Kõljalg, U., Larsson, K.H., Abarenkov, K., Nilsson, R.H., Alexander, I.J.,Eberhardt, U., Erland, S., Høiland, K., Kjøller, R., Larsson, E., Pennanen,T., Sen, R., Taylor, A.F.S., Tedersoo, L., Vrålstad, T., Ursing, B.M., 2005.UNITE: a database providing web-based methods for the molecular iden-tiWcation of ectomycorrhizal fungi. New Phytologist 166, 1063–1068.

Larsson, E., Jacobsson, S., 2004. Controversy over Hygrophorus cossus set-tled using ITS sequence data from 200 year-old type material. Myco-logical Research 108, 781–786.

Leonardi, M., Paolocci, F., Rubini, A., Simonini, G., Pacioni, G., 2005.Assessment of inter- and intra-speciWc variability in the main species ofBoletus edulis complex by ITS analysis. FEMS Microbiology Letters243, 411–416.

Liu, Y.J., Rogers, S.O., Ammirati, J.F., 1997. Phylogenetic relationships inDermocybe and related Cortinarius taxa based on nuclear ribosomalDNA internal transcribed spacers. Canadian Journal of Botany 75,519–532.

Mayer, C.P., Paulay, G., 2005. DNA barcoding: error rates based on com-prehensive sampling. PLoS Biology 3, e422.

Melot, J., 1990. Une classiWcation du genre Cortinarius (Pers.) S.F.Gray.Documents Mycologiques 20, 43–59.

Moser, M., 1983. Die Röhrlinge und Blätterpilze. Gustav Fischer Verlag,Stuttgart.

Moser, M.M., 1960. Die Gattung Phlegmacium (Sleimköpfe) Die PilzeMitteleuropas, Band IV. Julius Klinkhardt, Bad Heilbrunn.

Moser, M.M., Peintner, U., 2002. Die phylogenetischen Beziehungen derCortinarius aureopulverolentus-Gruppe. Journal des J.E.C. 5, 28–38.

Orton, P.D., 1955. The genus Cortinarius. I. Myxacium and Phlegmacium.The Naturalist (supplement), 1–80.

Peintner, U., Bougher, N.L., Castellano, M.A., Moncalvo, J.M., Moser,M.M., Trappe, J.M., Vilgalys, R., 2001. Multiple origins of sequestratefungi related to Cortinarius (Cortinariaceae). American Journal of Bot-any 88, 2168–2179.

Peintner, U., Horak, E., Moser, M.M., Vilgalys, R., 2002. Phylogeny ofRozites, Cuphocybe and Rapacea inferred from ITS and LSU rDNAsequences. Mycologia 94, 620–629.

Peintner, U., Moncalvo, J.M., Vilgalys, R., 2004. Toward a better under-standing of the infrageneric relationships in Cortinarius (Agaricales,Basidiomycota). Mycologia 96, 1042–1058.

Peintner, U., Moser, M.M., Thomas, K.A., Manimohan, P., 2003. Firstrecords of ectomycorrhizal Cortinarius species (Agaricales, Basidiomy-cetes) from tropical India and their phylogenetic position based onrDNA ITS sequences. Mycological Research 107, 485–494.

Seidl, M.T., 2000. Phylogenetic relationships within Cortinarius subge-nus Myxacium, sections DeWbulati and Myxacium. Mycologia 92,1091–1102.

Sontag, B., Fröde, R., Bross, M., Steglich, W., 1999. Chromogenic Triterpe-noids from Cortinarius fulvoincarnatus, C. sodagnitus and RelatedToadstools (Agaricales). European Journal of Organic Chemistry1999, 255–260.

Stoltze, M., Pihl, S., 1998. Rødliste 1997 over planter og dyr i Danmark.Miljø- og Energiministeriet, Danmarks Miljø-undersøgelser og Skov-og Naturstyrelsen.

SwoVord, D.L., 2003. PAUP*. Phylogenetic Analysis Using Parsimony(*and Other Methods) Version 4. Sinauer Associates, Sunderland,Massachusetts.

Taylor, A.F.S., Hills, A.E., Simonini, G., Both, E.E., Eberhardt, U., 2006.Detection of species within the Xerocomus subtomentosus complexin Europe using rDNA-ITS sequences. Mycological Research 110,276–287.

Taylor, J.W., Jacobson, D.J., Kroken, S., Kasuga, T., Geiser, D.M., Hibb-ett, D.S., Fisher, M.C., 2000. Phylogenetic species recognition and spe-cies concepts in fungi. Fungal Genetics and Biology 31, 21–32.

Vellinga, E.C., 2003. Phylogeny of Lepiota (Agaricaceae)—evidence fromnrITS and nrLSU sequences. Mycological Progress 2, 305–322.

Vellinga, E.C., de Kok, R.P.J., Bruns, T.D., 2003. Phylogeny and taxonomyof Macrolepiota (Agaricaceae). Mycologia 95, 442–456.

White, T.J., Bruns, T.D., Lee, S., Taylor, J.W., 1990. In: Innis, M.A., Gelf-and, D.H., Sninsky, J.J., White, T.J. (Eds.), PCR protocols: A Guide toMethods and Applications. Academic Press, New York, pp. 315–322.