DNA extraction, amplification and sequencing

The proteinase K protocol (^{Aljanabi & Martínez, 1997}) was used. The DNA pellet was diluted in 100-500 μl of TE (Tris 10 mM pH 7.4, EDTA 0.2 mM), kept at room temperature overnight, and stored at −20 °C until the DNA amplification step.

The polymerase chain reaction (PCR) was performed to amplify the D1 and D2 domains of the nrLSU. A Taq DNA polymerase Recombinant Kit (Invitrogen) was used. Primer pairs LROR-LR3 (^{White, Bruns, Lee, & Taylor, 1990}) were used to amplify the first 600 bp. Because the DNA of 25 ascomata appeared to be deteriorated, mainly in the type specimens, we designed 4 additional Helvella -specific internal primers using the Lasergene Primer Select Ver. 7.1.0 (44) (DNAStar, Inc.) software. These primers amplify segments of approximately 300 bp and were designated as: LSUF1 (AGCGGAGGAAAGAAACCAACAG), LSUR1 (CTCTACTTGTGCGCTATCGGTCT), LSUF2 (AGACCGATAGCGCACAAGTAGAG), and LSUR2 (TCCCAACAGCTATGCTCCTACTC) (Fig. 2).

Figure 2 Location of the primers used in this work. Primers designed specifically for Helvella are shaded gray. 

DNA amplification was performed in an Eppendorf Mastercycler Personal 5332 thermocycler using the protocol of ^{White et al. (1990)} with some modifications. The amplification program included a denaturation period of 5 min at 95 °C, followed by 35 cycles of denaturation at 95 °C for 1 min, annealing at 52-54 °C for 90 s, and extension at 72 °C for 105 s. A final extension cycle of 72 °C for 10 min ended the process and the reaction was held at 4 °C. GFX PCR DNA columns and Gel Band Purification Kits (GE Healthcare) were used to purify PCR products, following the manufacturer's instructions.

The sequencing reaction was carried out with BigDye Terminator v3.1 (Applied Biosystems, Wellesley, MA). Sequencing products were purified using an AutoSeq G-50 Dye Terminator Removal Kit (GE Healthcare), following the instructions of the manufacturer. Sequences were visualized using an ABI-Prism 310 Genetic Analyzer (Applied Biosystems).

Phylogenetic analysis

Taxon sampling. Sequences of the nrLSU DNA were generated from 42 specimens (including 11 type specimens) of Helvella , representing 26 species, plus one sequence of Wynnella silvicola (Table 1). Furthermore, 42 additional sequences of Helvella were obtained from GenBank, plus another of W. silvicola , and 15 of the outgroup. The 92 sequences assembled for this study come from specimens of 12 countries.

DNA alignment. Sequences were edited using Chromas Pro Ver. 1.41 (Technelysium Pty, Ltd, Tewantin, Qld, Australia). Every sequence was then subjected to a BLAST search in GenBank and ambiguous sequences were removed. DNA alignments were checked by eye and manually corrected when necessary using MacClade 4.0 (^{Maddison & Maddison, 2000}).

Morphological data . A matrix of 15 morphological and chemical characters were used. Appendix I shows character coding and Appendix II the morphological and chemical matrix.

Phylogenetic inference . The first phylogenetic analysis was done to determine the position of the genus Wynnella and the second to establish the phylogeny within Helvella . In the first analysis, a matrix comprising species of Helvellaceae, Tuberaceae, Rhizinaceae, and Discinaceae with 101 LSU sequences and 487 bp was constructed; 459 bp remained when ambiguous regions were excluded. To determine the phylogeny of Helvella , 68 specimens producing a matrix of 499 bp and 15 morphological and chemical data was used. Phylogenetic analyses were conducted using maximum parsimony (MP), maximum likelihood (ML), and Bayesian inference. MP was executed in PAUP* 4.0b10 (Altivec) (^{Swofford, 2002}), with the following parameters: gaps as missing characters, heuristic searches with changes among character states having equal weights, 1,000 replicates, tree-bisection-reconnection (TBR) as the branch swapping algorithm, branches collapsed if maximum branch length was zero, and molecular data as the "DNA" option and morphological and chemical data as "mixed". MP bootstrap support (BS-MP) was obtained from 1,000 replicates (^{Felsenstein, 1985}). RAxML 7.2.6 (^{Stamatakis, 2006}), implemented in raxmlGUI 0.93 (^{Silvestro & Michalak, 2010}), was used for the ML search, consisting of 100 replicates to find the best ML tree using the GTRGAMMA model for molecular data and the GTR model for morphological and chemical data; bootstrap support (BS-ML) was calculated with 1,000 replicates using the same models. Bayesian analyses were conducted in MrBayes 3.1.2 (^{Ronquist & Huelsenbeck, 2003}). The best substitution model, GTR+G+I, was determined by Modeltest 3.7 (^{Posada & Crandall, 1998}). The parameters in the Bayesian search were set as GTR model, rates = invgamma, ngen = 100,000,000, nruns = 4, samplefreq = 100, and the first 50% of the samples were discarded. For morphological and chemical data, the program MrBayes was run with nst = 1 (basic model) and rates = gamma.

Updated generic description

Helvella(as Elvela) L., Sp. pl. 2: 1180 (1753)Type species: Helvella mitra (as Elvela) L., Sp. pl., Edn 2 2: 1180 (1753) Helvella crispa (Scop.) Fr., (Lundae) 2(1): 14 (1822)

Ascomata epigeous, stipitate, occasionally subsessile, solitary to gregarious. Apothecia cupulate, discoid, convex, saddle-shaped, bi-, tri-, or irregularly lobed, but never auricularioid nor completely fused with the stipe; hymenium white, cream, buff, brown, gray, red brown, dark brown, or black, occasionally mottled (shades of grayish brown); apothecial sterile surface smooth or ribbed, glabrous, subpubescent, or pubescent, same color as hymenium or lighter. Stipe terete, even, with clefts, costate, or lacunose, white, cream, buff, brown, gray, or black, glabrous, subpubescent, or pubescent, internally solid, hollow, or chambered; base with white mycelium. Asci cylindrical, hyaline, operculate, base pleurorhynchous or aporhynchous, non-amyloid (J-), 8-spored. Ascospores ellipsoid to broadly ellipsoid, subglobose, subfusoid to fusoid, frequently with a central guttule or rarely triguttulate, hyaline, smooth or verrucose, tetranucleate at maturity. Paraphyses straight, clavate, septate, simple or branched, normally thin-walled, rarely thick-walled or with a thick-walled cap or collar at the apex, hyaline, with light brown or dark brown pigments in the wall, cytoplasm, and/or encrusted on the wall, these pigments visible or not with cotton blue. Apothecial tissues divided in 2 layers: medullary and ectal excipulum; stipe tissues also separated in an outer and inner layer. Medullary excipulum and stipe inner layer of textura intricata , hyaline or light brown, some species with dextrinoid reaction (Dx+) with Melzer's reagent. Ectal excipulum and stipe outer layer of textura angularis , hyaline, with light or dark brown pigments in the wall, cytoplasm, and/or encrusted on the wall, pigments visible or not with cotton blue, some species Dx+. Hairs may be present on the apothecial sterile surface and on the stipe surface, with the same microstructure in all species that present them, grouping into hyphal clusters or fascicles, hyaline, light brown, or dark brown, pigments present in the wall, cytoplasm, and/or encrusted on the wall, pigments visible or not with cotton blue, some species Dx+. Basal mycelial hyphae septate, hyaline.

Habit, habitat, and distribution

Solitary or in groups, some scattered, in soil of conifer and angiosperm forests. Common in temperate and arctic habitats, almost exclusively distributed in the Northern Hemisphere; only Helvella elastica , H. fibrosa (Wallr.) Korf, H. lacunosa (sensu lato ), H. papuensis Dissing, and H. sp. (aff. H. pezizoides ) have been reported from a tropical region, in Papua New Guinea (^{Dissing, 1979}). Some species have been cited from Argentina and Australasia (^{Gamundi, 2010}; ^{Rifai, 1968}).

The misunderstood concept of Helvella

Since the last century, ^{Nannfeldt (1937)} argued that Acetabula (Fr.) Fuckel ENT#91;represented here by Helvella acetabulum (L.) Quél.ENT#93;, Cyathipodia Boud. (represented here by H. cupuliformis Dissing & Nannf.), Leptopodia Boud. (represented here by H. albella , H. alpestris , H. atra J. König, H. elastica , H. ephippium Lév., and H. stevensii ), Macropodia Fuckel or Macroscyphus Nees ex Gray ENT#91;represented here by H. macropus (Pers.) P. Karst., Peziza subclavipes W. Phillips & EllisENT#93;, and Paxina Kuntze (represented here by A. calyx , H. acetabulum , H. compressa , H. costifera Nannf., and H. leucomelaena in this paper) were untenable. They were described based on the shape of the apothecia and stipes, e.g., Macropodia with cup-shaped apothecia and Leptopodia with lobed apothecia. However, it is common to find ascomata with young cup-shaped apothecia that later become lobed (as in H. crispa and H. macropus ). None of these genera can be separated from Helvella because they are: (1) microscopically similar, and (2) do not form supported monophyletic clades.

Some authors did not follow the Nannfeldt's concept of Helvella . ^{Dennis (1981)} and ^{Breitenbach and Kränzlin (1984)} continued using the genera Cyathipodia , Leptopodia , Macropodia or Macroscyphus , and Paxina , as well as Helvella . In contrast, ^{Eckblad (1968)} placed Underwoodia columnaris Peck and U. fuegiana (Speg.) Gamundi under Helvella , and this placement was followed by ^{Harmaja (1974)}, who synonymized U. beatonii Rifai as H. beatonii (Rifai) Harmaja. However, ^{Abbott and Currah (1997)}, ^{Ainsworth, Sparrow, and Sussman (1973)}, ^{Dissing (1966}, ¹⁹⁷²⁾, ^{Gamundi (2010)}, ^{Korf (1972)}, and ^{Rifai (1968)}, considered Underwoodia as an independent genus. Phylogenetic studies have corroborated the last point of view (^{Hansen & Pfister, 2006}; ^{Læssøe & Hansen, 2007}; ^{O'Donnell, Cigelnik, Weber, & Trappe, 1997}, and this work). Although Underwoodia has a medullary excipulum of textura intricata and ectal excipulum of textura angularis as do all species of Helvella , and this character was used by ^{Eckblad (1968)} to consider Underwoodia as a synonym of Helvella , the apothecium of Underwoodia is fully adhered to the stipe, unlike any apothecium seen in Helvella . Also, ascospore ornamentation is very evident inside the asci (not seen in Helvella ascospores), and some species have hooked paraphyses (U. beatonii and U. columnaris ), while all the species of Helvella have straight paraphyses.

The taxonomic position of Wynnella has been controversial. On the basis of microscopic features, ^{Harmaja (1974)} transferred W. silvicola to H. silvicola . ^{Abbott and Currah (1988)}, following Harmaja, erected Helvella sect. Silvicolae to place this species. Subsequently, they raised its rank to subgenus (^{Abbott & Currah, 1997}). However, ^{Ainsworth et al. (1973)}, ^{Dissing (1966}, ¹⁹⁷²⁾, ^{Dissing, Eckblad, and Lange (2000)}, ^{Eckblad (1968)}, ^{Korf (1972)}, and ^{Parslow and Spooner (2009)} recognized Wynnella as an independent genus from Helvella . Phylogenetic studies of Helvellaceae have suggested a sister group relationship between Helvella and Wynnella (^{Hansen & Pfister, 2006}; ^{Harrington, Pfister, Potter, & Donoghue, 1999}; ^{Læssøe & Hansen, 2007}; ^{Landvik, Kristiansen, & Schumacher, 1999}; ^{O'Donnell et al., 1997}), which was also found in this study. Like Underwoodia and Helvella , Wynnella has a medullary excipulum of textura intricata and ectal excipulum of textura angularis , but its apothecia are auriculoid (similar to a rabbit's ear), different to any apothecia seen in Helvella . A complete circumscription of Helvella based mainly on ^{Nannfeldt (1937)}, ^{Dissing (1966}, ¹⁹⁷²⁾, and ^{Rifai (1968)}, and following the phylogenetic results previously obtained and the ones generated in this study, is presented above in Results.

Comments about diversity and distribution of the species of Helvella

Excluding synonyms and species transferred to other genera, 36 species have been recognized for Europe (^{Calonge & Arroyo, 1990}; ^{Dissing, 1966}; ^{Dissing et al., 2000}; ^{Häffner, 1987}; ^{Van Vooren, 2010}). In North America, 32 species have been registered (^{Abbott & Currah, 1997}; ^{Nguyen, Landeros, Garibay-Orijel, Hansen, & Vellinga, 2013}; ^{Weber, 1972}), 20 for Asia (^{Kaushal, 1991}; ^{Zhuang, 1995}), 3 for South America (^{Dissing, 1966}; ^{Gamundi, 2010}; ^{Wright & Albertó, 2006}), 3 for Africa (^{Dissing, 1966}), and 2 species for Australia (^{Rifai, 1968}). Helvella leucomelaena is the most widely cited, recorded from every continent. Similarly, H. lacunosa and H. solitaria P. Karst. are located on 4 continents. However, it has recently been revealed that H. lacunosa is a complex of species, where the American specimens do not correspond to its European counterpart (^{Nguyen et al., 2013}). This probably occurs with other species with a broad distribution such as H. leucomelaena and H. costifera as already seen from our data (Fig. 4), but we need more detailed studies to confirm it.

On the other hand, H. alpestris , H. dovrensis T. Schumach., H. fusca , H. latispora Boud., H. oblongispora , H. paraphysitorquata I. Arroyo & Calonge, and H. ulvinenii Harmaja have been reported only from Europe (^{Calonge & Arroyo, 1990}; ^{Dissing, 1966}; ^{Schumacher, 1992}). Helvella compressa , H. crassitunicata N.S. Weber, H. griseoalba N.S. Weber, H. maculata , and H. robusta S.P. Abbott are exclusively known from America (^{Abbott & Currah, 1997}; ^{Weber, 1972}, ¹⁹⁷⁵). Helvella papuensis is only recorded from a tropical region of Papua New Guinea (^{Dissing, 1979}) and H. aestivalis (R. Heim & L. Rémy) Dissing & Raitv., H. dovrensis , and H. verruculosa (Berk. & M.A. Curtis) Harmaja have an arctic and alpine distribution (^{Abbott & Currah, 1997}). It is necessary to examine Helvella species that have been described from Asia to determine their status.

Infrageneric classification of Helvella

The analysis combining DNA sequences, morphological, and chemical data generated trees with better-supported clades, which demonstrates the importance of "total evidence", where all relevant data are considered in phylogenetic analysis (^{Eernisse & Kluge, 1993}).

Our results do not completely support any previous infrageneric Helvella classification. However, there is enough current evidence to divide Helvella in 2 subgenera based on ascus development that is also supported by phylogenetic data (Fig. 4). Helvella subgen. Leucomelaenae is monotypic with sect. Leucomelaenae Dissing sensu ^{Weber (1972)} and includes species with aporhynchous asci. On the other hand, Helvella subgen. Helvella comprises species with pleurorhynchous asci. The present study failed to determine several species affinities at section level, as for H. acetabulum , H. costifera , H. griseoalba , H. robusta , and H. ulvinenii within Helvella sect. Acetabulum Dissing sensu ^{Weber (1972)}. Although H. atra , H. ephippium , and H. subglabra N.S. Weber, representing Helvella sect. Ephippium Dissing sensu ^{Weber (1972)}, formed a clade, this was unsupported statistically. Furthermore, the phylogenetic results did not support the association of the species of Helvella sect. Macropodes Dissing sensu ^{Weber (1972)}, which includes H. cupuliformis and H . macropus . It is necessary to use other DNA regions to resolve their affinity.

Most of the monophyletic clades are defined by more than one morphological feature. The exception is the clade representing Helvella sect. Leucomelaenae , which is characterized only by the presence of aporhynchous asci. Figure 4 shows that the main features used by ^{Abbott and Currah (1997)}, ^{Dissing (1966)}, and ^{Weber (1972)} to define infrageneric taxa (pubescence and apothecium shape) are present throughout the phylogram, and only in combination can they be used as diagnostic characters of the clades.

Helvella subgen. Helvella

≠ subgen. Helvella, in ^{Abbott and Currah (1997)}

Asci pleurorhynchous, i.e., ascus base forked, because a crozier is formed at the base.

Type species: Helvella crispa (Scop.) Fr.

^{Abbott and Currah (1997)} circumscribed subgen. Helvella for species with lobed apothecia and a costate or lacunose stipe; instead, we consider this subgenus to include all the species with pleurorhynchous asci. Phylogenetically supported sections in this subgenus are Elasticae , Helvella , and Lacunosae .

Helvella subgen . Leucomelaenae S.P. Abbott emend. Landeros & Guzm.-Dáv.

Asci aporhynchous, i.e., ascus base simple, because no crozier is formed at the base.

Type species : Helvella leucomelaena (Pers.) Nannf.

It is characterized by its aporhynchous asci. In contrast, ^{Abbott and Currah (1997)} delineated the subgenus for species with a cup-shaped apothecium and a costate stipe, without considered how the ascus was formed. We recognize sect. Leucomelaenae Dissing sensu N.S. ^{Weber (1972)}.

Taxonomically informative features

Comments on all the characters that have been used in Helvella are presented, mentioning if they have any taxonomical value.

Apothecia. ^{Weber (1972)} and ^{Abbott and Currah (1997)} considered the shape of the apothecia as one of the most important features for recognition of sections or subgenera within the genus. Helvella acetabulum , H. costifera , H. cupuliformis , H. ulvinenii , and H. verruculosa have cup-shaped apothecia throughout their entire development. In contrast, H. albella , H. compressa , H. elastica , and H. lacunosa exhibit lobed apothecia. However, some species like H. crispa , H. ephippium , and H. macropus have cupulate apothecia when young and lobed apothecia when mature, and sometimes even cupulate apothecia may have mature ascospores. Another relevant taxonomic character is the way in which the apothecium margin is attached to the stipe, a feature almost exclusively found in sect. Lacunosae and in H. robusta in sect. Acetabulum . Therefore, the classification of Helvella based on apothecia shape and attachment is problematic. In the light of the phylogenetic results (Fig. 4), it is desirable to re-evaluate these characters more carefully in each species by studying various collections and assessing the degree of variation.

Stipe. Its shape is a diagnostic character in the genus. ^{Weber (1972)} described the stipe as even, costate, lacunose, and sulcate. The first 3 are easy to distinguish, but sulcate is hard to differentiate from costate. For this reason, we consider sulcate and costate as synonyms, meaning that the stipe has only longitudinal ribs. Sometimes species with a smooth stipe can develop folds or clefts mainly in old ascomata ENT#91;e.g., Helvella atra , H. corium (O. Weberb.) Massee, H. pezizoides Afzel., H. stevensii ENT#93;, but they never look like true ribs as in H. acetabulum , H. costifera , and H. sulcata Afzel. On the other hand, H. costifera , H. palustris Peck, and H. sulcata have a costate stipe with one or 2 transverse ribs, but it never looks like a lacunose stipe, which has both longitudinal and tranversal ribs, giving the stipe the appearance of having deep holes, e.g., in H. lacunosa . Because we have seen specimens of H. acetabulum and H. crispa with either costate or lacunose stipes, we believe it is important to know the degree of variation among and within species. Therefore, we consider stipe shape (even, costate, and lacunose) as an important taxonomic character.

Hymenium and stipe color. ^{Weber (1972)} regarded hymenium and stipe color as valuable features at the infrageneric classification level, while ^{Abbott and Currah (1997)} did not use these as main characters for their sections. We agree with ^{Weber (1972)} that hymenium and stipe colors are useful, but only for species distinction and in combination with other characters. For instance, Helvella sect. Elasticae Dissing sensu ^{Weber (1972)} is defined by its tan to gray-brown hymenial surface and buff to nearly white stipe, besides the lobed apothecium and even stipe. Hymenium color is also helpful for identifying species. For example, fresh specimens of H. griseoalba have a hymenium with grayish tones, which separates them from fresh specimens of H. costifera , which have brownish tones.

Ribs. Rib structure is a useful character to distinguish some species. In Helvella acetabulum , sharp ribs contrast with the blunt ribs of H. costifera and H. griseoalba . Another feature to take into consideration is the distance from the ribs to the margin of the apothecial sterile surface. We recognize 2 categories in this character: (1) ribs missing or reaching a quarter or less in the lower part of sterile surface (e.g., H. leucomelaena , H. solitaria , and H. sulcata ), and (2) ribs reaching half or to the edge of the apothecium, i.e., in the same ascoma of H. acetabulum or H. costifera ribs may arrive halfway, some to three quarters, and others to the edge of the apothecial sterile surface. Although ribs can be simple, forked, or anastomosing, this feature is uninformative because some species have simple ribs when immature and forked when mature, or in the same stage ribs may be both simple and forked.

Asci. ^{Weber (1972)} considered ascus development (aporhynchous vs. pleurorhynchous) as a useful character. Indeed, she restricted Helvella sect. Leucomelaenae to include only species with aporhynchous asci. In this work, we found that Helvella sect. Leucomelaenae Dissing sensu ^{Weber (1972)} is monophyletic (Fig. 4), with aporhynchous asci its synapomorphy.

Ascospores. Ascospores have little taxonomic value, except in a few species of Helvella . Helvella crassitunicata and H. pocillum Harmaja have large ascospores (^{Landeros et al., 2012}). Likewise, H. macropus and H. terrestris (Velen.) Landvik develop fusoid to subfusoid ascospores (^{Abbott & Currah, 1997}; ^{Landvik et al., 1999}). ^{Abbott and Currah (1997)} established ascospore ornamentation as an important taxonomic feature. Unfortunately, ornamentation is variable among ascospores of the same species. The verrucose appearance is due to remnants of the secondary wall of the ascospores, which might adhere to the primary wall (Schumacher, pers. com. in ^{Landeros et al., 2012}). It is common in many ascospores of the same specimen that such remnants do not adhere, and therefore, ascospores remain smooth. For this reason, ^{Eckblad (1968)} described the ascospores of Helvella as having "false ornamentation".

Paraphyses. Their shape is useful at the generic level in Helvellaceae, but not for infrageneric classification or to separate species. In Helvella , all species have septate, simple or branched, and straight paraphyses, in at least one stage of their development. ^{Vite-Garín, Villaruel-Ordaz, and Cifuentes (2006)} described species with non-septate (e.g., H. acetabulum , H. corium , H. cupuliformis , and H. elastica ) and/or unbranched paraphyses (e.g., H. acetabulum , H. atra , H. corium , H. costifera , and H. crispa ). We studied some of the specimens ^{Vite-Garín et al. (2006)} mentioned and found that paraphyses are always septate and either branched or unbranched in the same specimen.

Hyphal fascicles. Microscopically, hairs on the apothecial surface are formed by hyphal fascicles. In the present work, following ^{Landeros et al. (2012)}, we use the terms: (1) glabrous, when such hyphal fascicles are absent; (2) subpubescent, when hyphal fascicles are shorter than 50 μm; and (3) pubescent, when hyphal fascicles are longer than 50 μm. ^{Dissing (1966)} considered hairs on the apothecial sterile surface as an important criterion in his infrageneric classification. In Figure 4, all species of Helvella sect. Lacunosae are glabrous-subpubescent, while all species of sections Helvella and Macropodes (H. cupuliformis and H . macropus ) are pubescent. However, in Helvella sect. Elasticae sensu ^{Dissing (1966)} there are species with glabrous, subpubescent, and pubescent ascomata. In this case, this feature is useful to recognize species, such as H. albella and H. elastica , which have a glabrous to subpubescent surface. Therefore, this is an important feature for recognizing species, but it must be used in conjunction with other characteristics to separate sections.

Pigments. For ^{Landeros et al. (2012)}, the presence of pigment in the ascoma has taxonomic value at species level. When present, pigments have to be evaluated from 3 points of view: (1) color, which may vary from dark to light brown; (2) structure location, pigments can be found in paraphyses, ectal excipulum cells, stipe outer layer cells, and apothecial and stipe hyphal fascicles, and (3) cellular location, in the wall, encrusted on the wall, and/or in the cytoplasm. We also recommend using cotton blue to evaluate whether the pigments are visible or not with this stain. Encrusted pigment on the wall (mainly in the paraphyses) is variable between ascomata of the same species; if present, it is always visible with cotton blue.

Dextrinoid reaction (Dx+ ). A red reaction in Melzer's reagent, assessed according to ^{Leonard (2006)}, is a meaningful character from taxonomic and phylogenetic perspectives. All species of Helvella sect. Elasticae present at least one Dx+ tissue, but this feature is not exclusive to this section; e.g., H. acetabulum , H. robusta , and H. subglabra also have at least one type of Dx+ tissue, and the reaction can be used to recognize these species. Because the reaction is especially evident in stipe tissues, it is important to look in the entire ascoma for it and to describe the tissue where it occurs.

Cell width in the ectal excipulum and stipe outer layer, and breadth of both layers. We noticed that cell width in the ectal excipulum and outer stipe layer of each ascomata is the same, independent of the tissue origin of the section. The thickness of the ectal excipulum may vary slightly; it is thinner toward the edge and wider in the center of the apothecium, while in the stipe the outer layer is uniform. Helvella oblongispora is unique in these layers being very wide, both composed of very large cells, almost the same width as in the medullary excipulum. Therefore, this feature is only important for recognizing H. oblongispora .

Ecology. ^{Hobbie, Weber, and Trappe (2001)} mentioned as uncertain the mycorrhizal status of Helvella ; however, ^{Tedersoo, Hansen, Perry, and Kjøller (2006)} considered all the species as forming ectomycorrhizae, hence studies are needed to confirm the status of the genus. Some taxa are associated with specific trees, e.g., H. fusca (^{Dissing, 1966}; ^{Landeros et al., 2012}) and H. leucopus var. populina I. Arroyo & Calonge (in ^{Calonge, 2000}) always grow near or under Populus L. Recently, on the basis of molecular data, 2 cryptic species of the H. lacunosa complex from western North America were found (^{Nguyen et al., 2013}); H. dryophila is associated with Quercus L. and H. vespertina with conifers. Furthermore, Helvella species can be restricted to a vegetation type or ecosystem, e.g., H. papuensis associates with tree species of Castanopsis (D. Don) Spach and Lithocarpus Blume in subtropical rain forests (^{Abbott & Currah, 1997}; ^{Dissing, 1979}) and H. aestivalis and H. verruculosa in the tundra of arctic and alpine regions (^{Abbott & Currah, 1997}).

Phenology is also cited as an important feature; H. ephippium fruits in July and August in USA (^{Weber, 1972}) and H. leucopus Pers. from March to May in Europe (^{Dissing, 1966}). The fruiting period can also be longer; H. elastica , H. lacunosa , and H. macropus fruit from June to November (^{Dissing, 1966}; ^{Weber, 1972}). Some species fruit at different times in different countries; H. leucomelaena fruits from May to July in Sweden, from January to May in France, from April to June in British Columbia and Washington, and from May to September in Alberta (^{Abbott & Currah, 1997}; ^{Dissing, 1966}). There are species that can be distinguished from other similar species on the basis of phenology; H. albella has an autumnal fruiting pattern, whereas H. compressa fruits in the spring. In addition to the information of vegetation and phenology, soil type was proposed by ^{Dissing (1966)} as an important feature to recognize some species. For example, he described H. corium , H. leucomelaena , H. queletii Bres., and H. solitaria growing in calcareous soil.

Species complex and misunderstood species

Helvella alpestris. This species is cited in ^{Index Fungorum (2015)} as a synonym of H. corium (O. Weberb.) Massee, but we treat it as an independent species, as ^{Mycobank (2015)}. ^{Boudier (1895)} described H. alpestris with ascospores 22-25 × 14-15 μm, stipe "sulcatum" (ribbed) and "... extus sub lente breviter velutinum, non squamoso -hirtum " (under lens surface slightly velutinous, non squamous-hirsute). ^{Dissing (1966)} considered H. alpestris as a synonym of H. corium , with ascospores 18-20-22 × 11-12.1-13 μm, stipe with grooves near the base, and surface pubescent to villose, which, according to Boudier's protologue corresponds to an incorrect interpretation of H. alpestris . Later, ^{Häffner (1987)} separated both species, describing H. corium with ascospores (15.3-) 17.3-21.3 (-22) × (8.5-) 9-12.7 μm, stipe even, and pubescent apothecial sterile surface. For Häffner, H. alpestris has ascospores (15-) 17-20.5 (-22.1) × 10.2-13.3 μm, stipe with well-defined ribs, and glabrous to subpubescent apothecial sterile surface. Except for the ascospore size, the interpretation of Häffner matches Boudier's description of H. alpestris . The specimens that we studied and sequenced have ascospores similar to those cited by Häffner; hence, in this paper we follow this interpretation. However, it is necessary to sequence material of H. corium to confirm if they are different.

Helvella atra . ^{Landeros et al. (2012)} mentioned several features to distinguish this species from H. subglabra . Furthermore, in the tree of Figure 4, the sequences of 2 specimens of H. atra from the USA are not grouped together, indicating that another species could be split. Therefore, it is necessary to include specimens from other parts of the world in the analyses, mainly from Europe, as H. atra was described from Iceland, to elucidate its status and typify it because there is no type material.

Helvella costifera . This species was described by Nannfeldt (in ^{Lundell & Nannfeldt, 1953}) from Sweden. Later, ^{Weber (1972)} described H. griseoalba and ^{Harmaja (1978)} H. hyperborea Harmaja. ^{Landeros et al. (2012)}, with a study of the type specimens and additional material, proposed that H. hyperborea is a synonym of H. costifera , but H. griseoalba is an independent species. We sequenced European specimens from Finland and Norway and found that they grouped in one supported clade (Fig. 4), except the holotype of H. griseoalba , which agrees with ^{Landeros et al. (2012)}. In fact, H. griseoalba is closer to the H. solitaria clade than to any member of H. costifera group. On the other hand, there is a species complex, formed by European specimens that correspond to H. costifera and Mexican specimens belonging to a different undescribed species (Fig. 4).

Helvella ephippium . This species has a small apothecium, less than 1.5 cm diam. Only H. pocillum , H. rivularis Dissing & Sivertsen, and H. terrestris can be as small, but all these species have a cup-shaped, well-defined apothecium, while H. ephippium has an involute, irregularly lobed to saddle-shaped apothecium. ^{Léveillé (1841)} did not mention the size of the ascospores, he only wrote "rondes et transparentes". ^{Massee (1895)} was the first to describe them, 16-18 × 10 μm. More recently, ^{Dissing (1966)} measured ascospores 19-20.4-21.5 × 11-11.6-12 μm, while ^{Weber (1972)} described them as 16.5-19 × 10.5-12 μm and ^{Abbott and Currah (1997)} as 17-21 × 10.1-13 μm. We included 2 sequences in our analyses, one as H. ephippium from Europe (unfortunately with no more information on locality and collector) (^{Alvarado et al., 2011}), and another as H. aff. ephippium from Mexico, which we proved to be unrelated (Fig. 4). On the basis of this phylogenetic result and the differences in ascospore size found by different authors, we think it is a complex. Thus, it is necessary to first lectotypify H. ephippium with the figure of ^{Léveillé (1841)} and then epitypify it with a specimen to clarify this group of species.

Helvella fibrosa ENT#91;=H. villosa (Hedw. ex Kuntze) Dissing & Nannf.ENT#93;. Not shown in the phylogenetic tree. ^{Korf (2008)} concluded that Peziza fibrosa is the correct name for this species because "H. villosa (Hedw. ex Kuntze) Dissing and Nannf." is a later homonym of a name proposed by ^{Schaeffer (1774)} for a different species. Also, ^{Korf (2008)} designated Hedwig's figure as the lectotype of H. fibrosa and specimen of Lundell, Nannfeldt, and Holm (Fungi exsiccate Suecici praesertim Upsaliensis # 3262 , 1985) as the epitype. On the other hand, ^{Weber (1972)} proposed H. pallidula N.S. Weber as a new taxon based on the absence of pigmentation in the ascoma tissues. We agree with ^{Häffner (1987)} and ^{Abbott and Currah (1997)}, in considering H. pallidula as a synonym of H. fibrosa because when we studied the holotype of the former we observed light brown pigments in the paraphyses, ectal excipulum, and outer layer of stipe, and yellowish pigments in the apothecial and hyphal fascicles of the stipe, as in H. fibrosa It is necessary to include sequences of specimens from several countries in different continents in the phylogenetic analyses to establish if all specimens cited as H. fibrosa correspond to the same species.

Helvella lacunosa and H. sulcata . These 2 species were described by ^{Afzelius (1783)}. Since the last century, their delimitation has been controversial, being considered as synonyms (^{Abbott & Currah, 1997}; ^{Dissing, 1966}; ^{Häffner, 1987}) and as separate taxa (^{Landeros & Guzmán-Dávalos, 2013}; ^{Weber, 1972}). Recently, 2 species of the H. lacunosa complex in the western USA were segregated and described (^{Nguyen et al., 2013}) because their DNA sequences do not correspond to the European material of this species. A careful study is needed to determine the number of species included in this complex and to elucidate the status of some species such as H. sulcata .

Helvella leucomelaena . We sequenced 4 specimens of this species (Table 1), 3 from Europe (one of them labeled as the isotype of Acetabula calyx ) and one from Australia, and all were grouped in a supported clade (Fig. 4). Acetabula calyx is cited in the ^{Index Fungorum (2015)} as a synonym of H. solitaria ; however, morphological data and LSU rDNA sequence of the isotype, indicate that it must be a synonym of H. leucomelaena . Of 4 sequences labeled as H. leucomelaena in GenBank (AJ972414, KC012682, KC122799, KC122798), only one (KC012682 from US) grouped with our sequences of this species (Fig. 4). We checked the source specimen ENT#91;Bellis Kullman, 21 May 2001 (TAA179637, TAAM)ENT#93; for sequence AJ972414 and found pleurorhynchous asci, so it corresponds to H. acetabulum ; in fact, AJ972414 groups with H. acetabulum from Mexico and Estonia. We did not study the US specimens from which sequences KC122798 and KC122799 were obtained; thus, we do not know whether they have aporhynchous or pleurorhynchous asci, but we suspect they correspond to H. acetabulum in their phylogenetic position (Fig. 4). ^{Harmaja (1977}, ¹⁹⁷⁸, ¹⁹⁷⁹⁾ proposed H. confusa Harmaja, H. oblongispora , and H. pedunculata Harmaja, which are very similar to H. leucomelaena . After morphologically studying the holotype of H. confusa (unpublished data) and specimens labeled as H. pedunculata by Harmaja (unpublished data), we have concluded that we do not have evidence to consider these 2 species as independent taxa. On the other hand, we sequenced 2 specimens labeled as H. oblongispora , one of them determined by Harmaja, and they were grouped in another clade, thus corroborating the observations of ^{Abbott and Currah (1997)}, who considered H. confusa and H. pedunculata as synomyms of H. leucomelaena , but H. oblongispora as an independent taxon.

Helvella macropus . It can be identified by its fusoid to subfusoid ascospores within the asci. Helvella terrestris also has fusoid to subfusoid ascospores, but larger, 50-65 × 12-15 μm (^{Landvik et al., 1999}) instead of 20-26 × 9-12 μm in H. macropus (^{Dissing, 1966}). ^{Peck (1902)} proposed H. macropus var. brevis Peck, based on specimens with smaller ascomata, with darker hymenia, and very short stipes. Later, ^{Harmaja (1974)} elevated it as H. brevis (Peck) Harmaja, considering the ascospore size and ecological features (^{Weber, 1972}), as did ^{Abbott and Currah later (1988)}. However, ^{Weber (1972)} argued that the variation between the 2 taxa merely represents different responses to environmental conditions. In fact, later ^{Abbott and Currah (1997)} reconsidered their position and placed H. brevis in synonymy with H. macropus . Figure 4 shows that specimens of H. macropus from USA and Mexico do not form a clade; thus, future studies need to include European sequences of H. macropus to determinate its relation with the American specimens.

Helvella stevensii . This species was considered a synonym of H. latispora by ^{Häffner (1987)} and ^{Abbott and Currah (1997)}, but we agree with ^{Dissing (1966)} in considering these species as independent taxa (see discussion by ^{Landeros & Guzmán-Dávalos, 2013}). On another hand, we studied authentic material of H. stevensii and the holotype of H. connivens , but we did not find any difference. In fact, their DNA sequences are identical; therefore, we agree with ^{Weber (1972)} and ^{Abbott and Currah (1997)} that H. connivens is the same as H. stevensii . DNA sequences of H. latispora specimens from Europe need to be analyzed to determine the relation of this species to H. stevensii in section Elasticae .