For this study, specimens from B, BAFC, BPI, ENCB, FH, IBUG, INBIO, MUCL, NY, S, and XAL herbaria (abbreviations follow Thiers, continuously updated), including the type specimens of Fomes weberianus, Ganoderma argillaceum Murrill, G. mexicanum Pat., G. microsporum R.S. Hseu, G. parvulum Murrill, G. perturbatum (Lloyd) Torrend, G. perzonatum Murrill, G. pulverulentum Murrill, G. rivulosum, G. sessiliforme Murrill, G. stipitatum (Murrill) Murrill, G. subamboinense var. subamboinense, G. subamboinense var. laevisporum, G. subincrustatum Murrill, and G. vivianimercedianum M. Torres were re-examined. Strains examined during this study were deposited at CBS, CIRM-CF, and BCCM/MUCL. The formation of chlamydospores was examined after growing the strains on malt extract agar medium at 25 °C over four weeks according to previous results of Bazzalo and Wright (1982).

The microscopic observations procedure followed Decock et al. (2007). Specimen sections were mounted in 5% KOH solution. Melzer’s reagent and cotton blue were used to test the amyloidity or dextrinoidity and cyanophyly of the microscopic structures, respectively. Microscopic characters were observed under a light microscope Axioscope 40 Carl Zeiss. Images were captured using Axio Vision 4 software on the same microscope. At least 30 structures of each mature specimen were measured. Basidiospores were measured without taking in account the apical umbo when not shrunk. Cuticular cells were measured from the middle part of the basidiome except in the case of some type materials, where only a fragment was received as loan. The 5% extremes of all microscopic measurements from each size range were given in parentheses and the arithmetic mean was provided in brackets. Color terms follow Kornerup and Wanscher (1963), and terms in descriptions are defined in Torres-Torres and Guzmán-Dávalos (2012).

Genomic DNA from herbarium specimens was extracted using three protocols: (I) CTab method with 1% PVP (Palomera et al. 2008), (II) salt-extraction method with 1% PVP (Aljanabi and Martinez 1997), and (III) Wizard Genomic DNA Purification Kit (Promega) with 1% PVP. Modifications of the protocols, following Doyle and Doyle (1987) or Palomera et al. (2008), were occasionally made. Genomic DNA was extracted from living cultures according to Amalfi et al. (2010, 2012).

The ITS region (ITS1, 5.8S, and ITS2) was amplified from dried specimens using the primer pairs G-ITS-F1/ITS4B (Cao et al. 2012) or ITS1F/ITS4B (Gardes and Bruns 1993), and from living cultures using the primer pairs ITS1F/ITS4 (White et al. 1990). The primers bRPB2–6F/bRPB2–7.1R (Matheny 2005) and CF2–EF983F/CR2–EF2218R (Rehner and Buckley 2005, Matheny et al. 2007) were used to amplify the rpb2 and tef1–α regions, respectively.

Polymerase chain reaction (PCR) to amplify the ITS from dried specimens followed Guzmán-Dávalos et al. (2003) with some modifications. Each 52 μl reaction solution contained 50 μl of PCR mix [35 μl of MilliQ water, 6 μl of 10 X Taq reaction buffer without MgCl₂, 3 μl of 50 mM MgCl₂, 3 μl of 5 mM dNTP, 3 μl of 2 μg/μl Bovine Serum Albumine (BSA)], 0.5 μl of each 10 μM primer, 0.15 μl of Taq DNA polymerase (5U/μl), and 1 μl of DNA template (1:5 dilution). PCR amplifications were performed in an ESCO Swift MaxPro thermocycler as described by Guzmán-Dávalos et al. (2003), except that the annealing temperature was following Cao et al. (2012). PCR products were purified with the GFXTM PCR DNA Purification Kit (GE Healthcare). Purified products (GFX) were sent to the Sequencing Department, University of Arizona and LaniVeg (CUCBA, University of Guadalajara). Polymerase chain reactions and amplification protocol of the ITS regions (including 5.8S), partial tef1–α gene, and the region of rpb2 from cultures were described in Amalfi et al. (2010, 2012) and Decock et al. (2013). Sequences were assembled and edited with SequencherTM 4.8 software (Gene Codes Corp., Ann Arbor, Michigan).

Two DNA sequence data sets were compiled: an ITS data set and a concatenated ITS, rpb2, and tef1–α data set. The combined data set comprised DNA sequences from 71 specimens/cultures from Africa, Europe, Meso– and South America, and Southeast Asia (Table 1). It included sequences from the types of G. microsporum, G. subamboinense, and G. subamboinense var. laevisporum. Sequences of 35 ITS, nine rpb2, and 13 tef1–α were downloaded from GenBank (www.ncbi.nlm.nih.gov/genbank/). It was subdivided into 11 partitions: ITS1, 5.8S, ITS2; rpb2 and tef1–α introns, and –1^st, –2^nd, and –3^rd codon positions of rpb2 and tef1–α. Ganoderma curtisii (Berk.) Murrill and G. lucidum (Curtis) P. Karst. were selected as the outgroup according to results shown by Moncalvo (2000).

The ITS data set was composed by 30 specimens/cultures, of which 29 originated from the Neotropics (Table 1). It was subdivided into three partitions: ITS1, 5.8S, ITS2. In this case, G. austroafricanum M.P.A. Coetzee, M.J. Wingf., Marinc. & Blanchette was selected as outgroup according to the results obtained by Coetzee et al. (2015).

All sequences were automatically aligned with MUSCLE (Robert 2004) and manually adjusted using PhyDe (Müller et al. 2010). PartitionFinder (Lanfear et al. 2012) was used to determine the best evolutionary model for each gene using the corrected Akaike information criterion (AICc). Maximum Likelihood (ML) analyses were conducted using RAxML 7.0.4 (Stamatakis 2006) and Bayesian Inference (BI) analyses with MrBayes v.3.2.2 (Ronquist and Huelsenbeck 2003). In the ML analysis, the default priors were used, including individual parameters for each partition, performing 1000 replicates under the GTRGAMMA model. BI analyses were run on CIPRES Science Gateway (Miller et al. 2010). Two independent runs, with 4,000,000 generations each, were carried out with a sampling frequency every 1000 generations and a burn-in of 25%. A 50% majority rule consensus tree with posterior probabilities (PP) was obtained. Convergence of the Markov chains to a stationary distribution was assessed by visual examination of the log likelihood values in the program Tracer v1.7.1 (Rambaut et al. 2018). Nodes were considered supported when bootstrap values (BS) were ≥ 75% and the PP was ≥ 0.85. The final alignments were deposited in TreeBASE (www.treebase.org), under accession ID: 24140 (http://purl.org/phylo/treebase/phylows/study/TB2:S24140).

The combined dataset contained 172 DNA sequences: 71 ITS, 50 rpb2, and 51 tef1–α. The final alignment comprised 526 bp in the ITS, 776 in the rpb2, and 1123 in the tef1–α. The concatenated data set (ITS + rpb2 + tef1–α) was 2425 bp long. From it, 23 ambiguous sites (12 from ITS1 and ITS2, 11 from tef1–α introns) were removed. The evolutionary models that best fit the individual dataset according to the AICc criterion were ITS1 = GTR+I+G, 5.8S = K80, ITS2 = GTR+I+G, rpb2 1^st = GTR+I, 2^nd = HKY+G, 3^rd codon positions = HKY+G, rpb2 intron= K80, tef1–α 1^st = GTR+I, 2^nd = HKY+G, 3^rd codon positions = GTR+G, and tef1–α intron = GTR+I. In BI analyses, the average standard deviation of split frequencies was 0.008100 in the concatenated data set and 0.008875 in the ITS data set. As far as our specimens from the Neotropics are concerned, the phylogenetic trees obtained from Bayesian (not shown) and Maximum likelihood inferences using the concatenated (Fig. 1) and the ITS (Fig. 2) data sets showed overall the same two clades, except for the unsupported branch of the specimen MUCL 43522 present in the concatenated ML and BI analyses, which collapsed in the ITS tree, and the placement of the specimen UMNFL100, G. subamboinense var. laevisporum, from Florida (Figs 1–2).

The Ganoderma weberianum-resinaceum lineage was resolved with strong support (PP 1, BS 100%). It was divided into two major clades, I and II (Fig. 1). Clade I (PP 1, BS 98%) corresponded to the G. resinaceum clade as defined by Moncalvo (2000), with G. austroafricanum, from South Africa, located in a basal position. This was further subdivided in an unsupported clade A (PP 0.66, BS 58) and moderately supported clade B (PP 0.89, BS 82). Clade A included specimens all originated from the temperate area of the Northern Hemisphere and a specimen from the highlands of central Kenya (MUCL 57035). Clade A was structured into two subclades, A1 (PP 1, BS 100), with the Kenyan specimen (MUCL 57035) in basal position, and A2 (PP 1, BS 89), with a specimen from China (MUCL 46912) in basal position. Clade B brought together several specimens originated from both North and South America, distributed into three well-supported subclades that corresponded to G. sessile Murrill and G. polychromum (Copel.) Murrill (PP 1, BS 89; PP 0.99, BS 100), as defined by Loyd et al. (2018). Two specimens from Argentina, tentatively identified as G. sessile and G. platense Speg., formed together a third distant well-supported subclade (PP 1, BS 99).

Clade II (PP 1/BS 99) was subdivided into two major clades: clade C (PP 1, BS 95) and clade D (PP 1, BS 100). Clade C corresponded to G. weberianum sensu Steyaert. It was further structured into two well-supported subclades, C1 (PP 1, BS 91) and C2 (PP 1, BS 74), with a geographic dichotomic pattern opposing the New World / Neotropics (C1) to the Old World / Paleotropics (C2).

New World / Neotropical clade (C1) had two low- to moderately supported terminal clades, C1.1 (PP 0.89, BS 86) and C1.2 (PP 0.80, BS 62). C1.1 included the ex-type strain of G. subamboinense var. laevisporum (BAFC 247, Bazzalo and Wright 1982) from Argentina and five specimens from Martinique, of which one was identified previously as G. subamboinense (Welti and Courtecuisse 2010) and three as G. weberianum (CIRM-CF, on-line catalog). Its sister clade, C1.2, comprised the type specimen of G. subamboinense (S F15183!) and ten specimens from Cuba and French Guiana, of which one was identified as G. weberianum (CIRM-CF, on-line catalog). It also comprised two specimens from Florida, which were both, alternatively identified as G. subamboinense var. laevisporum at GenBank, or as G. cf. weberianum in Loyd et al. (2018).

The Old World / Paleotropical clade (C2) contained specimens from both Africa and Asia. This clade was further subdivided into three well-supported subclades, viz. C2.1 (PP 0.99, BS 95), with two specimens from India, C2.2 (PP 1, BS 97) gathering specimens from Central Africa (Cameroon and Gabon), and C2.3 (PP 1, BS 91) with specimens from Australia and Southeast Asia (China, Philippines, and Taiwan). A specimen from Australia (DFP 8401) in clade C2.3 was identified as G. rivulosum by Steyaert (fide Smith and Sivasithamparam 2000) and reidentified as G. weberianum by Smith and Sivasithamparam (2000). The C2.3 also included the ex-type strain of G. microsporum (RSH0821, BPI) from Taiwan.

Finally, clade D (PP 1, BS 100) comprised specimens of Ganoderma sp. from Central Africa (DR Congo and Gabon) and from China, these latter tentatively identified as G. hoehnelianum Bres. at GenBank (Fig. 1).

The ITS data set comprised sequences from thirty specimens (Table 1) and the final alignment had 536 positions. Five positions, the alignments of which were judged to be ambiguous, were removed from the analyses. Clades C1.1 (PP 0.92, BS 100) and C1.2 (PP 0.78, BS 97) were confirmed by the ITS data set (Fig. 2). A specimen from Mexico (Guzmán-Dávalos 9569, IBUG), previously identified as G. weberianum by Torres-Torres et al. (2015), formed an additional branch, basal and sister to C1.1 and C1.2 (Fig. 2).

Taking in account both single and multilocus phylogenetic analyses, we considered that our Neotropical specimens formed two related but distinct, well-supported terminal clades, C1.1 and C1.2 (Figs 1–2) that could be equated each to a phylogenetic species. The additional branch formed by the single specimen from Mexico also might be equated to a distinct phylogenetic species.

From a morphological perspective, specimens in the phylogenetic species C1.1 and C1.2 were very similar, characterized by an overall reddish brown to violet brown pileal surface, light, cork-colored context, occasionally paler in the upper zone —described as not fully homogenous by Torres-Torres et al. (2012)—, with none to several (up to 4) dense, brown stripes or continuous lines of resinous deposits extending from the base of the context toward the margin. The cuticular cells were mainly cylindrical to clavate, apically rounded, regular, amyloid, and the basidiospores ovoid to broadly ovoid, with free to subfree pillars, and chlamydospores (“gasterospores” in Bazzalo and Wright 1982) in their context (Figs 3–4).

These chlamydospores were mostly subglobose in the context of the basidiomes and more variably shaped in pure culture on artificial media, thick-walled, hyaline to yellowish, and with dextrinoid content. In the specimens of C1.1, the chlamydospores were constantly, permanently smooth-walled (Fig. 3D–F, I) whereas in specimens of C1.2, they were smooth becoming roughened on aging, with free to partially anastomosed fine ridges with a meridian orientation (Fig. 4E–H, K). The Mexican specimen (Guzmán-Dávalos 9569, IBUG!) also presented chlamydospores in the context but they were punctuated, ornamented with thick pillars (Fig. 5).

Basidiospores were slightly wider in specimens from C1.1 compared to those of C1.2, mainly 8–9 × 6–7 µm (averaging 8.5 × 6.5 µm, Fig. 3C, H) vs. 8–9.5 × 5.5–6.5 µm (averaging 8.9 × 6.0 µm, Fig. 4C–D). Cuticular cells were moderately longer in specimens from C1.2 (up to 100 µm long, Fig. 4B, J) than specimens in C1.1 (up to 65 µm long, Fig. 3B, G).

In general, the morphology allowed the distinction of three morphotypes, which could be considered as three morphospecies. Each of these also corresponded to a phylogenetic species.

The present study, using single and multilocus phylogenetic inferences combined with morphological and in vitro culture studies, concordantly revealed two species of Ganoderma in the G. weberianum sensu Steyaert lineage, spanning over the Neotropics. Furthermore, a specimen from Mexico, represented only by the ITS sequence (Guzmán-Dávalos 9569, Figs 2 & 5), also could be equated to a morphological and phylogenetic species, pending confirmation when additional material is available. However, none of these three species could be equated to G. weberianum sensu Moncalvo (Moncalvo 2000), which is restricted to tropical Asia.

Clade C1.1 contained the ex-type strain of G. subamboinense var. laevisporum; hence, it could correspond to this taxon. The sister clade C1.2 contained the ex-type strain of G. subamboinense var. subamboinense; thus, it could correspond to the typical variety. Furthermore, both the molecular and morphological data would warrant recognition of both varieties at species level.

Ganoderma subamboinense was originally described by Hennings (1904) as Fomes subamboinense Henn. The type specimen originated from Brazil. Bazzalo and Wright (1982) accepted the species and distinguished a var. laevisporum from the typical variety. Both varieties were characterized by the presence of chlamydospores in their context, which were rough-walled with “veins anastomosing to form a sort of reticulum” in the typical variety, and smooth-walled in var. laevisporum, what was later confirmed by Gottlieb and Wright (1999). Ryvarden (2000), based on presumed morphological resemblance, reduced G. subamboinense s.l. (both varieties) as a synonym of G. multiplicatum (Mont.) Pat. Similarities between G. subamboinense var. laevisporum and G. multiplicatum var. vitalii Steyaert were previously reported (Bazzalo and Wright 1982). Inversely, Torres-Torres et al. (2012) recognized G. multiplicatum as an independent species that could be differentiated from G. subamboinense s.l., e.g., in having apically irregular cuticular cells, with many protuberances. Steyaert (1980) had already characterized G. multiplicatum var. vitalii with “mostly irregular” cuticular cells.

The revision of a fragment of the holotype of G. subamboinense var. subamboinense (S F15183!) (Fig. 4I–K) and of the holotype of G. subamboinense var. laevisporum (BAFC 25525!) (Fig. 3G–I) confirmed the previous observations (Bazzalo and Wright 1982, Gottlieb and Wright 1999). Both varieties are mainly characterized by a pale context with resinous incrustations or thin resinous brown bands, stretching from base towards the margin, cylindrical to clavate, apically regular, amyloid cuticular cells, ovoid to broadly ovoid basidiospores with free to subfree pillars, and chlamydospores in their context. The chlamydospores were striated in the typical variety and smooth-walled in var. laevisporum.

The study of the holotype of G. multiplicatum (K 123639!), originating from French Guiana, confirmed irregular cuticular cells with both lateral and apical protuberances, distinct from those of both varieties of G. subamboinense. Phylogenetic analyses inferred from an ITS data set (Bolaños et al. 2016) or a combined ITS–LSU data set (de Lima-Junior et al. 2014) also showed that G. subamboinense var. laevisporum (ITS sequence from the ex-type culture ATCC 52419) and G. multiplicatum (that should be considered as sensu auctores) formed two distinct clades, in two distant lineages. Hence, the synonymy of G. subamboinense and G. multiplicatum, as suggested by Ryvarden (2000), here also is rejected.

The macro- and microscopic features of both varieties of G. subamboinense, overall, corresponded to those of our specimens from C1.1 and C1.2 clades, as described above. Therefore, G. subamboinense var. subamboinense (C1.2) and G. subamboinense var. laevisporum (C1.1) could be applied to the taxa shown by these clades. However, for nomenclatural reasons, the varietal epithet laevisporum cannot be used, at any rank. The nomenclatural status of the epithet laevisporum was questioned; as previously highlighted, it was invalidly published (Moncalvo and Ryvarden 1997, Welti and Courtecuisse 2010). Moncalvo and Ryvarden (1997) noted that Bazzalo and Wright (1982) did not formally propose the combination G. subamboinense, making it invalid; consequently, the varietal epithet also was invalid. Moncalvo and Ryvarden (1997) validated the combination G. subamboinense but this did not automatically validate the varietal epithet, which, therefore, cannot be used. A name, therefore, should be found for the taxon shown by the C1.1.

Furthermore, several other names whose types originate from the Neotropics but currently of uncertain status or in the limbo of the G. weberianum-resinaceum complex (Moncalvo and Ryvarden 1997, Ryvarden 1985, 2000, 2004) also could be reconsidered for species represented by both clades. Taking into account the main features of our specimens as described above, such as a light-colored context, G. argillaceum, G. mexicanum, G. parvulum, G. perturbatum, G. perzonatum, G. praelongum Murrill, G. pulverulentum, G. sessiliforme, G. stipitatum, G. subincrustatum, and G. vivianimercedianum (Patouillard 1898, Murrill 1902, 1903, 1908, 1912, Steyaert 1962, 1980, Torres-Torres et al. 2008) were worth revisiting.

Ganoderma mexicanum (holotype: FH 458184!) is the earliest name potentially available. It should be treated together with G. sessiliforme (holotype: NY 98713!); both type specimens are originated from neighboring localities in Morelos State, Mexico, viz. Tepalcingo, D. de Jonacatepec (Patouillard 1898) and Cuernacava (Murrill 1912), respectively. Torres-Torres et al. (2012) emphasized the poorly conserved type of G. mexicanum (Fig. 6) and reported an additional specimen from Brazil. On the basis of these two specimens, Torres-Torres et al. (2012) described clavate to narrowly clavate, 35.2–72.4 × 6.8–10.5 μm cuticular cells with very thick wall, without apical granulations, and ellipsoid basidiospores, 9.3–10.6 × 6.2–7.4 μm, with subfree pillars. Our study of the type specimen of G. mexicanum showed smaller, clavate to narrowly clavate cuticular cells, 25–37 × 5–7.5 μm, averaging 28.5 µm long, with occasional apical granulations. Basidiospores were ovoid with free to subfree pillars, (6.5–) 7.8–9.4 × 5.2–6.5 (–7) μm, and smooth, dextrinoid chlamydospores, 10.3–15 μm, were observed in the context (Fig. 6).

The study of the type specimen of G. sessiliforme (Fig. 7A–F) and of a second specimen collected in an area also neighboring the type locality (Guzmán 2078, ENCB, Fig. 7G–K, cf. Torres-Torres et al. 2015) revealed a pale context with a few resinous incrustations, scattered, smooth-walled, dextrinoid chlamydospores, (8–) 10–12 (–13.5) x 7–11 μm, clavate to narrowly clavate, smooth to sometimes faintly apically granulated cuticular cells, 25–38 × 5–9 µm, averaging 32 µm long, and ovoid basidiospores, 8–9.3 (–10.7) × 6–7.7 (–8) µm, with free to subfree pillars.

Based on these observations and inversely to the previous conclusions of Torres-Torres et al. (2012, 2015), we did not found any consistent morphological difference between G. mexicanum and G. sessiliforme. Furthermore, the type specimen of both names originated from neighboring localities and, probably, related ecosystems. Therefore, G. sessiliforme and G. mexicanum are here considered as synonyms, the latter epithet (Patouillard 1898) having priority. However, the status and affinities of G. mexicanum are uncertain.

Patouillard (1898) suggested that G. mexicanum was a sessile form of G. lucidum but with smooth (“lisses”), ovoid basidiospores. There is also a typewritten, undated note from R. Singer in the type specimen folder emphasizing “This is merely Ganoderma sessile Murr.”, and then pencil corrected “same as [G. sessile]”. Previously, Murrill (1902) also pointed out similarities between G. sessiliforme and G. sessile. Loyd et al. (2018) also suggested that G. sessiliforme might represent a synonym of G. sessile.

Nevertheless, G. sessile has distinctly larger basidiospores, 11.2–14.4 (–16.4) × 7.2–8.8 µm (fide Torres-Torres et al. 2015) and a duplex and spongy context (Torres-Torres et al. 2012, 2015), both features which would justify distinguishing these species.

Gottlieb and Wright (1999) and Gottlieb et al. (2000) previously compared G. sessiliforme with G. subamboinense var. laevisporum. They argued that G. sessiliforme differed from G. subamboinense var. laevisporum by the size (up to 11 mm long) and ornamentation (“semirugose” under the SEM) of its basidiospores, and the lack of chlamydospores, both features that do not stand (cf. above). The characters of G. mexicanum, especially the light-colored context with resinous incrustations, the basidiospores size, and the presence of smooth chlamydospores, overall remind much those of G. subamboinense var. laevisporum and of our specimens from the clade C.1.1 but for the cuticular cells. The cuticular cells are more clavate and shorter, 25–38 μm in G. mexicanum compared to those of G. subamboinense var. laevisporum and specimens from C1.1, 30–50 μm (Figs 3G & 6C–D).

Ganoderma parvulum (type NY 985699!) is the second earliest name potentially available. It should be treated together with G. stipitatum; the types of both epithets were collected in Nicaragua by C.L. Smith (Murrill 1902, 1903), probably in neighboring localities. Steyaert (1980), based on type studies, although accepting these two taxa, considered they formed a species complex together with G. bibadiostriatum Steyaert, whose type was collected in Brazil (Steyaert 1962). Ryvarden, in a handwritten note dated from 1983, joint to the holotype of G. stipitatum (NY 985678!), concluded that G. parvulum and G. stipitatum, likely were synonyms (“the identity with G. parvulum Murr. is almost certain”), what he formalized later on, accepting a single species under G. stipitatum, with G. bibadiostriatum and G. parvulum as synonyms (Ryvarden 2000, 2004).

The examination of the type specimens of G. parvulum (NY 985699!) and G. stipitatum (NY 985678!) (Fig. 8A–E) revealed little developed, likely immature basidiomes. Murrill (1902) already suggested this, stating “It is possible that the specimens [of G. parvulum] I have are not quite mature”. The type of G. parvulum was characterized by a pale context (pale ochraceous, fide Murrill 1902; light “ochraceous buff”, fide Steyaert 1980), with resinaceous streaks, dark horny (fide Murrill 1902) or carob brown (fide Steyaert 1980). These resinaceous streaks were also present in the type of G. stipitatum (Murrill 1903, Steyaert 1980, Ryvarden 2004, pers. obs.). Cuticular cells were cylindrical to slightly clavate, average 50 µm long, and very few basidiospores were observed in holotype specimens. Ryvarden, in a handwritten note dated from 1983, emphasized the absence of basidiospores in the type of G. stipitatum (“spores are not present”, NY 985678!). Nonetheless, Steyaert (1980) reported basidiospores, although without commenting on their abundance, 7.5–8.5 × 5.5–6.5 µm, averaging 8.1 × 5.9 µm in G. parvulum and 7.0–10.5 × 4.5–6.5 µm, averaging 7.8 × 5.5 µm in G. stipitatum. The few basidiospores we observed were 7–9 × 5–6.5 µm in G. parvulum and 7–10 × 5–6.5 µm in G. stipitatum, with free to subfree and very thin pillars in both.

Chlamydospores were not reported in the literature for G. parvulum nor for G. stipitatum (Steyaert 1972, 1980, Gottlieb and Wright 1999, Ryvarden 2000, 2004, Welti and Courtecuisse 2010), with the sole exception of Torres-Torres et al. (2012). Torres-Torres et al. (2012) reported and illustrated double-walled chlamydospores with “inter-walled, very thick pillars” presumably from the type of G. parvulum (cf. Torres-Torres et al. 2012, fig. 22d). However, their fig. 8, which is captioned as type of G. parvulum, actually corresponds to the type of G. stipitatum (NY 985678!). The voucher specimen from which these punctuated chlamydospores were observed remained uncertain. Nonetheless, our study of the type of G. parvulum and G. stipitatum revealed scattered chlamydospores in the context of both. These chlamydospores were smooth-walled or also ornamented with anastomosed ridges (Fig. 8D–E).

The likely immaturity of the type of G. parvulum and G. stipitatum, to a certain extent, could prevent definitive taxonomic interpretations. Notwithstanding, we would follow Ryvarden (2000, 2004) in considering that these two epithets represent a single species. Furthermore, the main macro- and microscopic characteristics of G. stipitatum and to a lesser extent, of G. parvulum, as described above, overall, also correspond to those of G. subamboinense var. subamboinense. Therefore, G. parvulum, G. stipitatum, and G. subamboinense could be considered synonymous. In this case, the epithet parvulum (Murrill 1902) has priority over subamboinense (Hennings 1904) and stipitatum (basionym: Fomes stipitatus Murrill 1903), contrary to the conclusion of Ryvarden (2004).

Steyaert (1962, 1980), Bazzalo and Wright (1982), and Gottlieb and Wright (1999) recognized G. bibadiostriatum as a distinct species. Ganoderma bibadiostriatum was characterized by a distinctly brown (fide Bazzalo and Wright 1982, or cinnamon, fide Steyaert 1962, 1980) context and basidiospores 7.0–11.0 × 5.5–7.5 µm, averaging 9.3 × 6.5 µm (fide Steyaert 1980), or 9–11 × 6–8 µm (fide Gottlieb and Wright 1999). These features differed from those of G. parvulum. Through phylogenetic inferences based on ITS and LSU, de Lima-Junior et al. (2014) showed that Brazilian specimens identified as either G. parvulum or G. stipitatum (identifications that should be considered as sensu auctores) were gathered into a single clade, which nested within the G. tropicum clade sensu Moncalvo (2000), and were unrelated to G. subamboinense var. laevisporum, hence unrelated to the G. weberianum-resinaceum lineage. Therefore, contrary to Ryvarden (2000, 2004) but following Gottlieb and Wright (1999), we also rejected the synonymy of G. bibadiostriatum with G. parvulum and G. stipitatum. We suggest that the identity of the G. parvulum–G. stipitatum clade shown by de Lima-Junior et al. (2014) should be re-evaluated, and that it might well represent G. bibadiostriatum.

The status and affinities of G. perzonatum have been debated and still are uncertain. Ganoderma perzonatum was described from Cuba (Murrill 1908). Moncalvo and Ryvarden (1997) first related it to G. parvulum. Previously, Steyaert in 1962 and Wright in 1967, in two notes joint to the type specimen of G. stipitatum (NY 985678!) and to a second specimen annotated “probable type” [of G. stipitatum] (NY 985716!) also informally suggested that G. perzonatum and G. parvulum were synonymous. However, later on, Ryvarden (2004) retained the species that he associated to the G. resinaceum complex. The revision of the type specimen (NY 985702!) (Fig. 8F–I) confirmed the main features (Ryvarden 2004): a sessile, dimidiate habit with superposed pilei, a pale corky context with dark, resinous streaks, cuticular cells up to 100 µm long, and basidiospores 8.5–9.5 (–10) × 6–7 µm. Furthermore, chlamydospores with smooth or then ornamented with fine longitudinal ridges (Fig. 8H–I), also were observed, a feature previously unnoticed. These characteristics brought it back to G. parvulum, as first suggested by Moncalvo and Ryvarden (1997).

However, G. perzonatum would differ from G. parvulum in having larger, sessile, dimidiate basidiomes and markedly cylindrical, longer cuticular cells. A specimen originating from the type locality of G. perzonatum (MUCL 43522, La Havana, Cuba, Fig. 9) shared these characters. It also produced striated chlamydospores, both in the context of the basidiome and in pure culture, similar to those of G. parvulum. However, this specimen formed a short, isolated branch, basal to the C1.2 clade, in phylogenetic inferences of the combined data set (Fig. 1). Ganoderma perzonatum remains of uncertain interpretation. It could be included, for the time being, in the concept of G. parvulum (hence G. parvulum s.l.).

The taxonomic statuses of G. argillaceum, G. perturbatum, G. praelongum, G. pulverulentum, and G. subincrustatum also were widely debated. Ganoderma argillaceum and G. praelongum were considered as synonyms of G. resinaceum by Steyaert (1980). Bazzalo and Wright (1982) also included G. pulverulentum and G. subincrustatum to the list of the G. resinaceum presumed synonyms, to which Ryvarden (2000) added G. perturbatum and G. sessiliforme. Inversely, at the other extreme, Torres-Torres and Guzmán-Dávalos (2012) recognized under these epithets as many independent species. Gottlieb and Wright (1999) accepted G. praelongum with G. pulverulentum (Murrill 1908) as a synonym, whereas Welti and Courtecuisse (2010) recognized G. pulverulentum as an independent species.

Ganoderma argillaceum, G. praelongum, and G. pulverulentum differed from G. parvulum, G. mexicanum, and our Neotropical specimens in having larger basidiospores (respectively 9–11 × 6–8 µm, fide Gottlieb and Wright 1999; 9–11 × 6–8.5 µm, fide Gottlieb and Wright 1999; 9.6–12.8 × 6.2–8 µm, fide Torres-Torres et al. 2012). Ganoderma argillaceum (holotype NY, 01293316!) had basidiospores with abundant and thin pillars and lacked chlamydospores (Gottlieb and Wright 1999), whereas G. praelongum and G. pulverulentum had basidiospores with partially anastomosed pillars (Gottlieb and Wright 1999, Torres-Torres et al. 2012). Ganoderma vivianimercedianum also differed from our specimens in having larger basidiospores, 8.8–11.2 (–12) × 6.5–8 µm, and absence of chlamydospores (Torres-Torres et al. 2008). These names remained of uncertain status and affinities. They are most likely not synonyms of G. resinaceum s.s. from Europe (Bazzalo and Wright 1982) but affinities with North American species of the G. resinaceum clade, viz. G. sessile and G. polychromum (Loyd et al. 2018), should not be excluded.

Ganoderma perturbatum (BPI!) differs from our specimens from both clades (C1.1 and C1.2) in having larger basidiospores 10–12.8 × 8–9.4 µm with subacute apex and partially anastomosed pillars or short crest-like ornamentations. Ganoderma subincrustatum has cuticular cells generally with short and thick protuberances and basidiospores with partially anastomosed pillars (Torres-Torres and Guzmán-Dávalos 2012, Torres-Torres et al. 2015, pers. obs.). Their status and affinities also remained uncertain.

In conclusion, we are of the opinion that G. mexicanum could be selected as the earliest name available for the specimens of the clade C1.1. It is morphologically very similar if not identical to G. subamboinense var. laevisporum. We therefore suggest, for the time being, pending new material and DNA sequences data, to apply G. mexicanum to the clade C1.1. We also concluded that the name G. parvulum could be retained as the earliest name available for the taxa represented by the clade C1.2, previously reported in the literature as G. subamboinense var. subamboinense. Ganoderma perzonatum could represent another closely related taxon in the vicinity of G. mexicanum / G. parvulum.