Micromphale sect. Perforantia (Agaricales, Basidiomycetes); Expansion and phylogenetic placement

DNA sequences show that the traditional genus Micromphale appears to be polyphyletic. Nuclear ribosomal LSU and ITS DNA sequences place Micromphale sect. Perforantia Singer (typus sect. M. perforans) within Gymnopus, comprising a clade sister to a mixture of traditional Gymnopus taxa including G. fusipes (typus generis) plus traditional Marasmius sect. Androsacei. This study enlarges sect. Perforantia and shows that sect. Perforantia is a clade separate from those including Micromphale sect. Micromphale and sect. Rhizomorphigena. A new subsection Pinophili is proposed to include new species G. pinophilus and G. ponderosae. Eleven taxa are accepted at species rank, of which nine are proposed as new, mostly morpho-taxa.

Micromphale, as legislated in 1954, now comprises three sections. Singer (1948) had described two new heterotypic sections of Micromphale, adding to the typical section (Micromphale sect. Micromphale, nom. auton.; typus sect. et gen. Mi. venosum = Mi. foetidum). The first of these (Micromphale sect. Rhizomorphigena Singer; typus sect. Mi. westii) was accepted by Desjardin and Petersen (1989), and will be taken up elsewhere. The second, Micromphale sect. Perforantia Singer (typus sect. Mi. perforans), is more or less the object of this paper. Neither of Singer's (1948) sections was populated by more than its type species.
If Singer's (1948) diagnosis for sect. Perforantia was less than detailed, its type species provided a nucleus around which to gather similar basidiomata. Every diagnostic character (and several character combinations) applied equally to organisms placed elsewhere and short of studies of individual type specimens, clear-cut circumscription of the section awaited molecular analyses. Nevertheless, over time, additional taxa have been accreted.
Several attempts to phylogenetically place Micromphale have been made. For Moncalvo et al. (2002) (based on nrLSU sequences) a large clade labeled as /omphalotaceae comprised two subclades, /omphalioid and /lentinuloid, the latter including numerous taxa, including a polyphyletic Gymnopus. Within /lentinuloid, Mi. foetida occurred in a further subclade, /micromphale, but Mi. perforans formed a long branch of its own outside of /lentinuloid but within /omphalotaceae. Likewise, Mata et al. (2004) based on nrLSU sequences, found that M. foetidum was placed within a clade dominated by Gymnopus taxa. The first study including both Mi. foetidum and Mi. perforans was that by Wilson and Desjardin (2005) based on nrLSU sequences), who found that while both species were placed in a clade dominated by Gymnopus taxa, they did not form a monophyletic group. Mata et al. (2007) (based on nrITS sequences) more accurately placed Mi. foetidum and Mi. brassicolens within Gymnopus (but not monophyletic) but did not include Mi. perforans.
Based on shared taxonomic characters, both macro-and micromorphological, Micromphale sect. Perforantia may be difficult to separate from Marasmius sect. Androsacei. For example, usually overlooked, the pileipellis of basidiomata of Marasmius sect. Androsacei almost always occurs with a thin slime matrix, a feature traditionally used to define Micromphale and especially sect. Perforantia, further confusing the separation of Micromphale sect. Perforantia from Marasmius sect. Androsacei. Further, because of the peculiar pileipellis structure composed of diverticulate hyphae together with broom cell-like setulose hyphal termini, Marasmius sect. Androsacei was separated from mainstream Marasmius at genus rank as Setulipes (Antonín 1987), with type species S. androsaceus. The pileipellis structure was described as "a well-developed Ramealesstructure," a reference to a very similar (or identical) construction found in some taxa of Marasmiellus sect. Rameales. This construction and its hyphal structures are replaced in Micromphale sect. Perforantia by a repent pileipellis with hyphae usually encrusted to some extent but varying from diverticulate or setulose to without such differentiation. Joining a chorus of genera being dismembered as a result of molecular analyses indicating polyphyletic distribution of infrageneric complexes, the traditional concept of Marasmius (typus generis Ma. rotula) is being fractured. Accordingly, Marasmius sect. Androsacei is disassociated from mainstream Marasmius.
Superficial similarities among basidiomata of Micromphale sect. Perforantia and Marasmius sect. Androsacei prompted the current study. Morphological examination of many collections, including micromorphology, led to suspicion that undescribed taxa occurred in North America and that a comprehensive study including both complexes was unwieldy. It seemed more efficient, therefore, to focus on sect. Perforantia, with additional reports to follow.

Materials and methods
Abbreviation of genus names include Ma. for Marasmius and Mi. for Micromphale. Color names in quotation marks (i.e. "sayal brown") are from Ridgway (Ridgway 1912); see also (Petersen 2016), usually followed by an alphanumeric citation (i.e. 7E5) referring to plate, column and row in Kornerup and Wanscher (1967). PhC = Phase contrast microscopy; BF = bright field microscopy; TFB = Tennessee Field Book, fieldassigned number for notes and photographs; TENN = herbarium number assigned to a specimen for filing and future herbarium retrieval. RHP and KWH = abbreviations of authors' names; GSMNP = Great Smoky Mountains National Park. State and province abbreviations follow United States Postal Service designations.
All microscopic examination was undertaken using two microscopes: 1) Nikon dissecting microscope (model SMZ-27) with reflected light; and 2) Olympus compound microscope (model BX60) outfitted with bright field, phase-contrast and QC (model R5) camera.
Two types of cultures were employed to produce some DNA sequences: 1) over many years, mass-germinated fresh spore prints on malt extract agar (MEA) were dissected and polyspore isolates were allowed to overgrow. These cultures were stored in test-tube MEA slants in a cold room at ca. 8 °C; 2). Recently, it was discovered that surface-sterilized sections of stipes and/or rhizomorphs readily produced vigorous dikaryon growth on agar medium. Within 48-72 hrs, this mycelial "spray" could be excised into axenic status. Isolates were allowed to cover a 60-mm Petri dish, then transferred to MEA tube slants for storage. For sequencing, small MEA blocks were used as inocula in 15 ml potato-dextrose broth (PD, Difco) and allowed to grow at room temperature. When growth had progressed sufficiently, mycelium was teased away from agar blocks and the mycelium processed for DNA extraction.
For notes on examination of individual specimens, see under Results, Taxonomic characters.
Molecular procedures: Procedures for DNA extraction from cultures or dried herbarium specimens, PCR, sequencing and processing of sequence data and analyses of sequence data are described in Aldrovandi et al. (2015). Both forward and reverse sequences were obtained for each collection. Sequences were corrected and aligned using GCG (GCG 2000) either manually (Perforantia) or using the GCG Pileup program (Gymnopus LSU data set) followed by manual adjustment. In rare cases with older specimens, the ITS region was obtained in fragments using primers ITS1F-ITS2 and ITS3-ITS4 (White et al. 1990). In such cases, the nrLSU sequence was never obtained. Sequence data are not available for some taxa in this manuscript, a function of herbarium specimens that were too old or too scanty for DNA extraction.
Phylogenetic Analyses: PhyML analyses were carried out using GENEIOUS 8 (Geneious 2015) using an estimated transition/transversion ratio, proportion of variable sites and Gamma distribution parameter. The substitution model was GTR. The number of bootstrap replicates was 100. Percent identity was calculated for each pair of sequences using GENEIOUS 8 and manually averaged to obtain within and between average percent identities. These were converted to % divergence for this manuscript.
In order to determine the placement of sect. Perforantia within Gymnopus s.l., a representative file of 191 Omphalotaceae nrLSU sequences including representative taxa belonging to Omphalotus, Lentinula, Gymnopus, Marasmius, and Micromphale was constructed and analyzed using PhyML in GENEIOUS. Trees were visualized in FigTree 1.4.2 (Rambaut 2006). DNA sequences were deposited in GenBank (see Appendix 1 for GenBank numbers) and aligned sequences and trees were deposited in the Dryad Data Repository (doi: 10.5061/dryad.4081h).
Compared to most agarics, all macromorphological structures of Sect. Perforantia are quite small, and appear somewhat fragile. This is misleading, for basidiomata are quite pliant and resilient. Evidence suggests that basidiomata dry in situ and repeatedly re-expand with increased moisture, but whether viability also survives is unknown. Frequently, basidiomata are examined which exhibit myriads of spores attached to or embedded in the hymenium but with no mature basidia visible.
Pileus: In all cases, the juvenile pileus is strongly convex, maturing to Plano-convex, often with downturned margin and occasionally with evidence of small umbo. Pileus disc is usually somewhat darker than limb or especially margin, but disc colors range from medium tan to dark chocolate brown. Limb and margin are usually paler but can be unicolorous with the disc. Pileus margin is often sulcate-striate, sometimes deeply so, but this may be an effect of moisture and basidiome age.
Lamellae: In dried specimens, lamellar attachment to stipe is difficult to ascertain. Adnate attachment is typical but adnexed or decurrent attachment is also encountered. A small tell-tale character distinguishing sect. Perforantia from sect. Androsacei is lamellar color. In sect. Androsacei, lamellae are often almost as dark as the pileus, but at least in the same color series, while in sect. Perforantia, lamellae are usually off-white to dingy pale gray.
Lamellae in sect. Perforantia are often described as collariate or pseudocollariate when fresh, but in drying, lamellae secede from stipe apex and then often appear pseudocollariate. Lamellae are described below as "thin" versus "thickish," the former indicating a sharp lamellar edge, while "thickish" indicates lamellae with more or less parallel sides and blunt edge. There seems to be no correlation between this and presence of cheilocystidia. In most instances, figures are given for "total lamellae," the number of lamellae reaching the pileus margin, and "through lamellae," the number of lamellae reaching the stipe. The incidence of lamellar anastomosis or interveining is low (i.e. but see G. resinosae).
Rhizomorphs: Generally reported only as present or absent, an attempt has been made here to describe rhizomorphs in more detail, especially since rhizomorphs were used to establish cultures in some taxa. In G. quinaultii and G. sublaccatus, for instance, rhizomorphs are short, stout, forming small basal pads and colonizing adjacent Thuja scales, while in other taxa rhizomorphs may be long and gracile (i.e. in G. ponderosae unbranched, in G. glabrosipes branched). Occasionally, rhizomorphs are so plentiful as to form a loose thatch. Rhizomorphs branches are usually short and spur-like.
Taste and odor: The predominance of herbarium specimens in descriptions below, often with scant or absent notes, usually make taste and odor impossible to report. A more or less typical Micromphale odor and taste, variously described as fetid or of rotten cabbage, however, seems to occur in subsect. Perforantia (i.e. G. perforans, G. sequoiae), while the pine-dwelling taxa seem to exhibit only negligible taste and odor.
Pileipellis: Traditionally, the substance which absorbs water, creating a mucoid matrix for pileipellis, pileus and lamellar tramae, subhymenium and stipe medullary tissue has been called "gelatinous." Considering its consistency in KOH, I prefer to call it "slime." In my observations, in the balance between sol and gel, the substance is distinctly on the side of sol. As observed, this slime is always heterogeneous, often with effete basidiospores (and often of anamorphic fungi) and shards of encrusting material.
Where observed, pileal hairs are illustrated and are rather uniform in dimensions and shape. It is quite possible that pileal hairs occur more often toward pileus margin and have been overlooked.
Diverticulate pileipellis hyphal segments are common in G. pinophilus and G. ponderosae (subsect. Pinophili), with diverticula usually dichotomous. This feature is also found in cheilocystidia of the siccus-type.
Pleurocystidia: In the past, basidioles have often been described as fusiform. In fact, such fusiform hymenial structures are more common than and of comparable dimensions as basidia, which are universally clavate. Moreover, other hymenial elements, namely basidioles, are clavate, similar to basidia but without sterigmata. Recently, the fusiform elements have been recognized as pleurocystidia and we have adopted this usage here. Illustrations are furnished for all taxa, and these pleurocystidia extend to other groups within the clade /omphalotaceae.
A peculiar feature of pleurocystidia is the vague partitioning of cell contents, in which (under PhC + 1250×) distal cell contents appear paler than contents of the rest of the pleurocystidium, but without a well-defined separating membrane (i.e. see illustrations for G. quinaultii).
Basidia: Basidia are universally clavate and are dominated by 4-sterigmate forms. Two particular features have been noted in descriptions below: 1) in most taxa, basidia and pleurocystidia remain attached to subhymenial hyphae in squash mounts. Parenthetically, subbasidial hyphae usually remain turgid after basidia and pleurocystidia collapse, with clamp connection hook cells remaining turgid and therefore producing a beaded or catenulate appearance (i.e. see illustration for G. sublaccatus), superficially resembling cheilocystidia of the mainstream Gymnopus type. In several taxa, however, basidia and pleurocystidia are easily disarticulated from subbasidial hyphae. 2) In some taxa, basidia and pleurocystidia do not collapse when effete, remaining as empty walls. This phenomenon is reminiscent of a seed-bearing ear of corn enclosed in the leaf-like husk, and has been referred to in descriptions below as "husking" (i.e. see illustration for G. sublaccatus).
Basidiospores: Across sect. Perforantia, dimensions and shapes of spores differ little. These parameters cannot be used to identify individual specimens. Moreover, all spores are inamyloid.
Commentary. Antonín and Noordeloos (1997) recognized Gymnopus sect. Vestipedes subsect. Impudicae, including G. perforans, based, it would seem on the fetid odor of basidiomata of the included taxa. Later (Antonín and Noordeloos 2010), subsect. Impudicae was accepted at section rank, but basically included the same taxa as before. Wilson and Desjardin (2005) indicated that G. perforans formed its own clade sister to that of G. (Micromphale) foetidum and other mainstream odious Gymnopus taxa. Based on the present study, the original Micromphale sect. Perforantia is found to be more closely related to the Gymnopus (Marasmius) androsaceus complex, not inclusive of G. foetidum and other Gymnopus sect. Impudicae. Molecular phylogenetic information is available for only few taxa accepted (or proposed) here. For "morpho-taxa," some outline employed to justify their description in this paper is owed the reader. Unfortunately, almost every morphological character includes a caveat for collections which violate one or more characters. The following diagnoses are offered: Section Perforantia Subsect. Perforantia, subsect. auton.
Etymology. Bulliformis = bubble-like, referring to the appearance of cheilocystidia at lamellar edge.
The description which follows is compiled from examination of dried material. Colors when fresh may vary considerably from those below.
Commentary. Proposal of another species of Gymnopus (Micromphale) section Perforantia would seem problematic, especially from a region relatively rich in such taxa, but presence and uniqueness of diagnostic characters make such a proposal relatively safe. The diagnosis (above) lists characters which separate G. (Mi.) bulliformis from other members of sect. Perforantia with diminutive basidiomata.
The temperate rain-forests of the North American Pacific coastal region of United States and Canada apparently support several taxa of the androsaceoid/perforantioid complex. Loan of selected specimens (using Mycoportal as filter) under only Ma. androsaceus and Mi.perforans revealed enough new taxa to make intensive collecting worthwhile for such organisms. This might not be surprising, considering the variety of conifer trees, on the dead needles of which such diminutive basidiomata are to be found.
Separation of taxa in Micromphale sect. Perforantia is difficult. Micromphale perforans occurs in Europe and Scandinavia through northern North America, presumably continent-wide. It may be present in northeastern and western Siberia as well, but this is beyond the scope of this paper. From northern California comes M. sequoiae which closely resembles G. bulliformis but is reported as limited to dead foliage of Sequoia sempervirens. Both taxa are reported as lacking (or rare) cheilocystidia, neither taxon exhibits much differentiation in the pileipellis, and caulocystidia are quite similar, although those of M. perforans are reported as exhibiting some denser, more refringent (PhC) pigment at caulocystidial tips. Ordinarily, M. perforans exhibits a black, vestured stipe when mature, but Antonín and Noordeloos (2010) include dark brown shades as well, and M. sequoiae seems also to form a dark brown stipe. Desjardin (1985) used the following characters to separate M. sequoiae from M. perforans: 1) light brown to flesh-colored pileus versus pale pileus respectively; 2) mild odor versus strong, fetid odor respectively; 3) greyish orange to brown pubescent stipe versus black velutinous stipe respectively; 4) caulocystidia without concentrated apical pigment versus caulocystidia with concentrated apical pigment respectively; 5) habitat on Sequoia debris versus conifer needles, usually Picea and/or Abies respectively. Attempts were made to obtain ITS sequences from both collections of G. bulliformis, both unsuccessful. WTU-9305 yielded sufficient sequence to identify it as a member of the clade /perforantia and blasted weakly to G. perforans in GenBank. Whether G. bulliformis deserves phylogenetic distinction remains questionable until clean sequences can be obtained from "fresh" material.
Habitat and phenology. On dead broad-leafed leaves, most often Quercus leaves, fruiting especially on midribs and petioles of both red and white oak complexes; other adventitious substrates include Acer (at least A. rubrum), Cornus, Magnolia, Rhododendron, 2-needle Pinus; Appalachian Mountain chain from New England through northern South Carolina and northern Georgia, west to Arkansas and south to Gulf Coast; late Spring through early Autumn.
Commentary. Gymnopus foliiphilus is the most commonly collected North American taxon in sect. Perforantia. Preliminary field identification attempts to distinguish several taxa with similar basidiomata. Substrate segregates G. androsaceus and G. perforans (conifer needles, usually Picea and/or Abies), from G. foliiphilus. Geographic distribution is less secure. Both G. androsaceus and G. perforans are found in Europe and temperate North America, while G. foliiphilus seems limited to eastern North America. From all these taxa, a mimic, Marasmius pallidocephalus, is separated from G. androsaceus with difficulty in the field, based almost solely by lack of clamp connections of the former and phylogenetic placement.
Often in collections of G. foliiphilus, evidence of some bleaching of substrate can be detected. This is not dramatic -not to pale off-white -but distinct nonetheless. The phenomenon cannot be compared to M. perforans because fallen needles of Picea/Abies naturally bleach over time.
In G. folliiphilus, stipes are almost always bicolor, upward with some avellaneous to pinkish shade, downward to dark brown and finally black toward the base (note that the very junction of stipe and lamellae is always dark brown). The relative stipe length of these colors varies considerably, with the upward avellaneous shades from only the uppermost 10% to as much as the upper 50%. To some extent, the density and quality of the stipe vesture also varies, with upper surfaces producing shorter, less setoid, hyaline caulocystidia, often with long, slender, hyaline hyphae producing a sparsely silky or wispy appearance (40×) rather different from the hispid or barbed appearance of the lower stipe caused by setoid caulocystidia often gathered into synnematous sheaves. A seductive artifact in microscope mounts of hymenial structures are the subbasidial cells. As is typical, basidioles and basidia are produced in "bouquets" by subbasidial hyphae, which usually are tightly packed but which retain hyphal characteristics. In G. foliiphilus, subbasidial hyphae are catenulate or congestedly lobose. The result are structures which mimic the cheilocystidia of numerous Gymnopus taxa, especially in sect. Vestipedes. In G. foliiphilus (as in G. perforans), cheilocystidia are difficult to interpret, and when present, are consummately basidiiform.
Traditional generic characters are not consistent in sect. Perforantia and sect. Androsacei. For example, the pileipellis of G. androsaceus resembles a rameales structure of repent but diverticulate hyphae. Such a pileipellis is also present in Marasmiellus, taxa of which seem to belong to several relatively distantly related clades. In Micromphale, a gelatinous layer within the pileus trama can usually be demonstrated (i.e. M. foetidum, etc.), but in Micromphale sect. Perforantia the gelatinous layer is absent, but "replaced" by a thin slime matrix over and within the pileipellis.
A paper by Farnet et al. (1999) employed an agar medium reputed to promote production of rhizomorphs. For the present study, this medium (whole wheat flour 20 g/L; agar, Bacto 20 g/L; H 2 O 1 L) was used for numerous dikaryon isolates of various Marasmius and Micromphale collections. Ancillary to production of rhizomorphs, aerial mycelium of isolates of M. foliiphilus slowly changed from white to bright yellow ("empire yellow" 3A6), while aerial mycelium of M. perforans remained white. Desjardin (pers. comm.) indicated the possibility that Marasmius insititius Fr. (1838. Epicrisis: 386), fruiting on Quercus leaves in Sweden, might be similar to M. foliiphilus. Marasmius insititius has seen a checkered history. Recently, Antonín and Noordeloos (2010) excluded the epithet from Marasmius because: 1) no type specimen exists; 2) Fries's description is less than explicit; 3) in spite of Fries's physical location in central Sweden in 1838, a habitat on Quercus leaves might indicate his exposure to the organism in southern Sweden; and 4) Orton (1960): 303) had dismissed the epithet as a later heterotypic synonym of Marasmius calopus. Desjardin (1989) did not include M. insititius in his type specimen studies, presumably based on the above and its extralimital status for Marasmius of the southeastern United States. Svengunnar Ryman (UPS; pers. comm.) indicates that M. insititius is an unknown entity. It is not the purpose of this paper to attempt to exhume M. insititius, especially as this name was not taken up in the Scandinavian mycota by Noordeloos in Funga Nordica (Knudsen and Vesterholt 2012).
Once informed of our intention to propose a new species to represent the oakloving relative of Mi. perforans, Desjardin (pers. comm Diagnosis. Similar to Gymnopus foliiphilus, but differing as follows: 1) pileus darker at all ages, smooth (not pebbled), not convex when young nor everted in age; 2) lamellae more numerous and lamellulae in two ranks; 3) lamellar attachment adnexed to adnate (not pseudocollariate); 4) stipe black or bicolored; 5) stipe insertion subinsititious (not insititious); 6) outer pileipellis hyphae often significantly encrusted (not smooth or with flakes).
Habitat and phenology. Fruiting on sclerophyllous leaves (perhaps Quercus) at mid-to high-elevation; early summer.
Habitat and phenology. Known only from the holotype collection. "Scattered on outer bark of fallen tree, mostly fruiting from needles fallen on log" (teste Seidl) (perhaps Abies grandis). Seasonality unknown.
Habitat and phenology. Fruiting on needles of Tsuga (associated with Abies); mid-summer.
Commentary. Extraction of usable DNA from the two specimens examined was unsuccessful, so phylogenetic placement of G. glabrosipes remains unknown. Pileipellis involved in slime matrix, pileus tramal hyphae with gelatinizing walls, basidia and pleurocystidial bases strongly gelatinizing and absence of cheilocystidia all are diagnostic of the Mi. perforans complex. Conversely, the glabrous-shining stipe is not characteristic of that clade.
Two pilei were assessed for diverticulate hyphae with no success. Pileipellis was the same in both.
Habitat and phenology. In large troops on fallen needles of Picea, rarely also Pinus and/or Abies, in humus-rich, coniferous plantations; widespread over Europe and Scandinavia (perhaps also in Asia); July to November.
Commentary. Based on direct comparison between European and northeastern North American specimens, basidiomatal dimensions seem to differ. American basidiomata are, in general, shorter with smaller pilei than European basidiomata, but basidiomata of the American collections seem distributed in two size classes, the larger and more robust of which are commensurate with Europe basidiomata.
A limited phylogenetic tree based on ITS sequences (see Fig. 87 below), sequences from northeastern North American collections clearly form a clade separate from sequences from Europe (including Scandinavia). Base-pair separation between these two clades is 1.65%. Although heterogeneous, base-pair separation among northeastern North American sequences is 0.71%. Altogether, collections from Europe and North America are not considered to represent species-rank separation, but proposal of a subspecies seems warranted. Separation of all G. perforans sequences from a clade containing G. sublaccatus and G. sequoiae is more robust (2.30% see Fig. 87 below) but both G. sequoiae and G. sublaccatus fall within North American G. perforans. For more on this situation, see discussion under G. perforans.
In eastern North America, characters which separate G. foliiphilus fruiting on dead deciduous leaves, from G. perforans subsp. transatlanticus fruiting on conifer needles: 1) G. foliiphilus exhibits stipe vesture between villose and barbed, but G. perforans subsp. transatlanticus is distinctly spiked/barbed; 2) pileal hairs in G. foliiphilus are stouter than those of G. perforans subsp. transatlanticus, and secondarily septate (this may be an artifact); and 3) pileus surface hyphae of G. foliiphilus are often encrusted in small scabs, while those of G. perforans subsp. transatlanticus are generally smooth.
In examining numerous specimens eventually accepted as G. perforans subsp. tansatlanticus, it was unforeseen that a large number of specimens originally fruited on dead Tsuga needles, a host reported as rare by Antonín and Noordeloos (2010), although this is the case in Europe. Tolerance of this substrate affords the opportunity to subsist at lower elevations (and therefore at higher temperatures) in the southern Appalachian Mountains where spruce/fir forests exist only at the highest elevations (and therefore the coldest temperatures). In the future, it will be interesting to see if this condition extends across North America to the Pacific Northwest, where the spcies of spruce, fir, and hemlock are all different from eastern North American taxa.
In G. perforans subsp. transatlanticus, cheilocystidia may be confused with inflated basidia and/or pleurocystidia. Some shapes of clearly swollen elements assumed to be cheilocystidia are suggestive of sterigmata (but always two -four were not seen), and occasional objects of cheilocystidial size are shaped like fat pleurocystidia (broadly fusiform). Repeated preparations from a single pileus also showed numerous cheilocystidia.
Rhizomorphs are here described as of two types. The coarser of these are present in virtually every specimen observed, while the extremely fine type were seen in perhaps 25% of the specimens. It can be doubted that the second type actually belong to subsp. transatlanticus, but without some proof, they must be described as present. Etymology. pino-= referring to the genus Pinus; -phil = to love, referring to habitat on the needles of Pinus.
Habitat and phenology. Gregarious on dead needles of Pinus strobus in eastern North America; summer.
Commentary. Desjardin (1989) and Gordon (1994) treated G. pinophilus as an unnamed subset of Ma. androsaceus. The latter was perceived as fruiting on three sub-  strata, hardwood leaves, needles of Picea/Abies, and needles of Pinus. Gordon and Petersen (1997) understood that M. androsaceus s.l. in eastern North America could be divided into three sexually interINcompatible groups, one of which (mating group III) was represented by only a single collection (TFB 5627 TENN-F-53488 from Idaho). With subsequent collecting and DNA sequence production, this collection was determined as G. pinophilus. Mata et al. (2007) showed that Marasmius sect. Androsacei (represented by M. androsaceus) was actually embedded within Gymnopus, and Noordeloos and Antonín (2008) formally transferred the section as Gymnopus sect. Androsacei. Furthermore, Desjardin (1990) and Desjardin and Horak (1997) reported considerable morphological and phenological variation within the North Temperate Androsacei, in part by circumstantial evidence, based on the data above. Present phylogenetic analyses now show that the small alliance of G. pinophilus and G. ponderosae (Gordon's mating group III) is more closely related to the Mi. perforans complex than to the Ma. androsacei complex.
Observation on TFB 14097 (TENN-F-67846) revealed two sets of dimensions of spores, and TFB 10459 showed that these spores are formed by basidia in the same hymenium -thus spores must be judged as dimorphic. One cause for this might be 2-versus 1-nucleate condition of individual spores, but all seem to be produced by 4-sterigmate basidia.
Observations of pileipellis of TFB 10459 showed that broom cell-like hyphal termini are ephemeral -apparently they wash or gelatinize away, so when they were absent from some pilei of various collections, the strongly ornamented repent hyphae remained.
Recently, TFB 14097 was established in dikaryron and monokaryon cultures and a self-cross was performed (see Gordon and Petersen 1997 for methods). From an assay of 20 putative single-basidiospore isolates (SBIs), 12 clampless putative monokaryons were selected and paired in all combinations. As expected, a tetrapolar mating system was revealed. A 1 B 1 = 5*, 6; A 2 B 2 = 1*, 3, 4, 8; A 2 B 1 = 2*, 7, 9, 11, 12; A 1 B 2 not represented in the sample (* indicates a tester strain). Most growth of donors was submerged (except for structures noted below). Most pairings exhibited very subtle "barrage" or "flat" contact zones, most easily envisioned with the naked eye against back-lighting. Barrage = somewhat congested growth in "mustaches" (not within contact zone). Flat = slightly congested hyphae on both sides of a lightly overgrown crevasse.
Habitat and phenology. Fruiting on dead needles, twigs and debris of Pinus ponderosa (three-needle pine); distribution probably following distribution of Pinus ponderosa (at least California, Idaho, Washington); late autumn.
Commentary. Collections accepted as four separate species (G. "scoticus" nom. prov., G. "adventitius" nom. prov., G. pinophilus, G. ponderosae) fruit on dead needles of Pinus spp. Of these, the former two are found in Gymnopus sect. Androsacei, while the latter two belong in sect. Perforantia. Both G. ponderosae and G. adventitius fruit on needles of Pinus ponderosae and superficially resemble one another. Both exhibit black, glabrous-shining stipes and dark brown pilei, but the pileipellis of G. ponderosae comprises repent, encrusted hyphae, diverticulate hyphae and broom cell-like hyphal termini, while G. adventitius pileipellis does not show diverticulate hyphae and broom cell-like hyphal termini. Moreover, stipes of G. adventitius often produces adventitious rhizomorphic structures from wounds, not seen in G. ponderosae.
The small clade comprising G. pinophilus and G. ponderosae is found sister to that of core sect. Perforantia. Pileipellis construction, which includes diverticulate hyphae and pileipellis "broom cells" is characteristic of a "Rameales structure," indicative of sect. Androsacei, but hardly that of traditional Perforantia, although involved in minimal slime. Marasmioid cheilocystidia also separate this pine-loving alliance from sect.
Perforantia. Rhizomorphs, while not typically as long as basidiome stipe, are common, obvious and arise from the same needles as basidiomata.
TENN53488-TFB5627 (Idaho, G. ponderosae) was the only representative of mating group III of "Marasmius androsaceus" by Gordon (1994) and Gordon and Petersen (1997). The taxonomic concept of the latter, however, was broad ([see Desjardin and Petersen (1989); Gilliam (1976)], including two separate taxa now known to be members of sect. Androsacei (as well as G. ponderosae in sect. Perforantia), and perhaps other organisms for which haploid isolates were unavailable for pairing experiments.
The possibility of dimorphic spores is reported here also for G. pinophilus. Although reported for other mushroom groups (notably hygrophoroids) the cause of this phenomenon is not known in the group treated here. Basidia are uniformly 4-spored and basidia are conspicuously clamped, indicative of normal nuclear number and behavior. Etymology. Latinized; resembling Pyracantha, referring to spike-like setulae on pileipellis broom cell-like hyphal termini.
Habitat and phenology. Known from only the type specimen; fruiting on dead sclerophyllous leaves (probably Quercus) at high elevation; summer.
Commentary. Presence of slime sheaths surrounding pileus and lamellar tramae as well as the dimensions and shape of cheilocystidia are characteristic of taxa in sect. Perforantia. Conversely, pileipellis presents a very distinctive thatch of broom celllike hyphal termini rather than repent, encrusted hyphae in a slime matrix as is seen throughout the section. Such a differentiated pileipellis might also qualify for some infrageneric groups of Marasmiellus but would be expected to be without slime deposition. Unfortunately, DNA sequences could not be produced from the only known collection, so molecular placement remains unknown. Cheilocystidia in G. pyracanthoides resemble those of G. bulliformis (q.v.). They also resemble those described for G. trabzonensis. They may represent some stage of basidial development because basidioles also are clavate, but basidioles are present throughout the lamellae face, while the structures here described as cheilocystidia are found along the lamellar edge. No other differentiated cheilocystidial structures were observed, but some taxa in sect. Perforantia are known to lack differentiated cheilocystidia. Slime is produced as a thin sheath surrounding individual hyphae, with ornamentation appearing as flakes riding on the exterior of the gelatinized wall. It also appears to obliterate subhymenium. Likewise, stipe medullary hyphae exist in minimal (but present) slime. There is little evidence of a slimy (or gelatinized) matrix in the pileipellis.
Based on habitat on Thuja debris, an ancillary study was made of Collybia thujina Kauffman (MICH!), which immediately led to examination of material of Marasmius filopes Peck (NYS!) and Collybia piceina Kauffman (MICH!). Prior treatments had been published by Hesler (1959), Gilliam (1976), Redhead (1980) and Desjardin (1989). Basidiomata of all correctly identified collections were significantly smaller and more slender than those of G. quinaultii, and pileipellis organization was of a well-developed Rameales-structure, not similar to that in G. quinaultii. Desjardin (1989) agreed with Redhead that M. filopes, C. piceina and C. kauffmanii were taxonomic synonyms, all correctly placed in Marasmiellus by Redhead. I agree with this assessment, leaving G. quinaultii as a separate taxon.
The following description is a rearrangement of the protologue plus observations on dried material. Basidiomata (Figs 71, 72A) pliant, marcescent, reviving. Pileus 6-12 mm broad, when young convex to campanulate, often with a short, acute umbo, in age becoming broadly convex to plano-convex with or without a central papilla, occasionally plane with a shallow central depression; margin when young decurved or slightly incurved, even, entire, in age becoming straight, wavy, crenate, rugulose-striate to rugulose-sulcate ¼ of the distance to center; surface dry to moist, dull glabrous, hygrophanous, at first light brown 7D4-5 overall, rarely with disc reddish brown 8E5-7, in age disc remaining light brown or fading to brownish orange 6C3-4; margin in age fading to brownish orange, greyish orange 6B2-3 or orange white 5A2, in age rarely colored buff overall with a slightly darker disc; pileus trama light brown to brownish orange, soft, up to 1 mm thick. Lamellae adnate, free in age or rarely attached to a pseudocol- lar, close to subdistant, narrow to medium broad (up to 1 mm), rarely anastomosing or intervenose, total lamellae = 25-27, through lamellae = 14-16; at first pale greyish orange 6B2, fading in age to pale orange white 5-6A2, typically concolorous with the pileus margin at maturity; edge even, entire, concolorous; lamellulae in 1-2 series. Stipe 20-43 × 0.7-1.5 mm broad, terete or rarely apically compressed and cleft, equal or tapered downward, hollow, cartilaginous, insititious; context concolorous with stipe surface; apex pruinose, off-white, downward pubescent and often with furfuraceous base, when young, apical portion pale greyish orange 6B2, central portion light brown 7D4-6, base dark brown 7F5-7 to rusty brown, in age apex becoming pale brownish orange 7C3, central portion becoming brown 7E4-5, base becoming dark brown 7-8F4-8 or occasionally dark brown overall in age. Rhizomorphs (Fig. 69A) short, slender, black, poorly developed, scattered; sterile stipes rare. Taste strongly alliaceous after 1-2 minutes; odor mild or rarely slightly fetid when old and wet.
Habitat and phenology. Scattered to gregarious on branchlets and leaves of Sequoia sempervirens; presumably throughout the range of S. sempervirens (at least northern California); October-February.
Caulocystidia arise as side branches of stipe surface, incrusted hyphae. Early stages of caulocystidial development often bear a shagreened surface but soon becoming smooth. A unique character is the frequent internal secondary septation, as well as occurrence of a clamp connection near caulocystidial origin.
Commentary. With slime matrix covering the pileus surface and involving hymenial structures, with characteristic clavate cheilocystidia and with vestured stipe, UBC 25212 seems certain to belong in sect. Perforantia. There it joins G. perforans, G. foliiphilus and G. sequoiae with the same general characters. From Ma. androsaceus, G. sublaccatus differs in pigmented (not black), vestured stipe. If its substrate preference is limited to Pseudotsuga, this constitutes another difference. Finally, pileipellis of G. androsaceus is characterized by diverticulate hyphal segments and broom cell-like hyphal termini, unlike that of G. sublaccatus which lacks these structures.
Basidiomata of G. sequoiae resemble those of M. sublaccatus in stature and size, but seems limited to fruiting on needles of Sequoia sempervirens in northern California. Cheilocystidia are shaped like an incandescent bulb, sometimes slightly askew, but otherwise quite typical of cheilocystidia in this taxonomic complex (compare, for example, Figure 85. PhyML analysis of 191 nrLSU sequences within the Omphalotaceae using 100 bootstrap replicates. Clades were collapsed to better show broad relationships. Species names in blue represent /marasmiellus (Wilson and Desjardin 2005). Species names in red represent /gymnopus (Wilson and Desjardin 2005). The major clades reported in Mata et al. (2007) are overlaid in blue. Section Perforantia is indicated in green.  (Fig. 85, red) but with low bootstrap support (44%). Gymnopus fusipes, (Micromphale) foetidus, members of the former Marasmius sect. Androsacei [Clades A and B of Mata et al. (2007)] and Gymnopus brevipes (Micromphale sect. Rhizomorphigena) appear together in a well-supported clade (83%) confirming that Micromphale is polyphyletic with some species belonging to Gymnopus. Results of this phylogenetic analysis suggest that Gymnopus is also polyphyletic, consistent with findings by Wilson and Desjardin (2005) who informally segregated some Gymnopus taxa into /marasmiellus.
A phylogenetic analysis of only the Perforantia based on a concatenated nrITS and LSU data set is given in Fig. 86 together with average base-pair differences based Figure 87. PhyML analysis of Gymnopus perforans and G. sequoiae ITS sequences. Distance measurements percent base pair differences for the ITS region only. Bootstrap support greater than 70% is given to the left of the supported node. Haplotypes are indicated as h1 or h2; clones are indicated as c1, c2, etc. North American locations are indicated by postal codes. on nrITS alone. Clades for putative species G. pinophilus, G. foliiphilus, G. sequoiae and G. perforans are well-supported. Gymnopus pinophilus/ponderosae is distinct from the remaining taxa (approximately 19% ITS sequence divergence). Within-taxa divergence is low (Fig. 86). Gymnopus perforans (European and North American collections combined) had a within-species bp divergence of 0.98%. The remaining within-taxa divergence estimates are 0.24% bp divergence for G. foliiphilus, 0.32% bp divergence for G. sequoiae plus G. sublaccatus and 0.4% bp divergence for G. pinophilus.
An unrooted nrITS phylogeny of G. perforans and G. sequoiae plus G. sublaccatus only is given in Fig. 87. Examination of sequence motifs suggests that within North America there are likely three ancestral haplotypes with evidence for mating and meiotic recombination between the haplotypes. European haplotypes are more homogeneous but form a distinct European clade. Long branches represent environmental samples rather than fruitbodies (JX029948 Czech Rep. and AM901982 house dust Finland).

Discussion
The members of the clade proposed as subsection Pinophili were at first included in Gymnopus sect. Androsacei, based on pileipellis structure (well-developed Ramealesstructure) and cheilocystidial shape (siccus-type broom cells). It was surprising, therefore, that phylogenetic analyses showed that the clade was more closely related to sect. Perforantia than to sect. Androsacei. This may serve as cautionary for predictions concerning the larger and more complex sect. Androsacei.
Previously largely misunderstood, pleurocystidia are common and rather uniform across sect. Perforantia. At best, these structures were interpreted as basidioles, but this study repeatedly demonstrated basidioles as clavate, not fusiform. Pleurocystidia, conversely, are fusiform throughout development. More subtle is the apparent partition of pleurocystidial contents (see Fig. 42), seen in several taxa. This partition, when present, is always in the distal portion of the pleurocystidium, and may be observable only in PhC, not stained BF.
The three mating groups of Ma. androsaceus revealed by Gordon (1994) and Gordon and Petersen (1997), when exposed to molecular analysis, can be summarized as Ma. androsaceus s.s., Ma. androsaceus on deciduous leaves [recognized as two separate species (unpubl. data)], and G. ponderosae (this study). These three all exhibit a glabrous-shining, black stipe and similar pileipellis and cheilocystidial shapes. The former two belong in sect. Androsacei, but G. ponderosae is now found related to sect.

Perforantia.
With previously undisclosed taxa reported here from the temperate rainforest of North America (i.e. G. bulliformis, G. glabrosipes, G. sublaccatus, etc.), it must be expected that additional taxa will be discovered across eastern Russia, Japan and across Canada. Likewise, ranges will be more accurately estimated when additional collections are subjected to molecular analyses.
There has been considerable discussion about the accuracy of various species concepts and the best methodology for circumscribing new species [see Taylor et al. (2000)]. What is often not discussed is that speciation may not be strictly bifurcating. Rather, peripheral speciation (parapatric speciation) may occur when a small population becomes genetically isolated from other populations by establishment of Dobzhansky-Muller incompatibilities (Bank et al. 2012;Hörandl 2006;Mayr 1963). Such isolation may be geographic in nature or ecological but when examined phylogenetically with commonly used genes, such peripheral species may not be reciprocally monophyletic with the ancestral species. To fail to recognize such events results in reduced estimates of biodiversity at a time when accurate recognition of biodiversity is increasingly important.
Three species in this study may represent examples of situations where new species form at the geographical or ecological periphery of a species domain. Gymnopus ponderosae, is embedded within G. pinophilus and therefore does not satisfy the criteria of reciprocal monophyly for phylogenetic species (Kizirian and Donnelly 2004;Moritz 1994). Gymnopus pinophilus does, however, satisfy criteria for ecological speciation (Ecological Species Concept) and morphological differences are evident (Morphological Species Concept). In like manner, G. sequoiae and G. sublaccatus appear to be offshoots of one branch within North American G. perforans. In this latter case, genetic isolation may be in part geographical and in part ecological but it is clear that some genetic separation has occurred.

Appendix 1
Collections used for molecular analyses (Reviewers -GenBank numbers will be added after review)