A new genus, Complexispora, with two new species, C. multistratosa and C. mediterranea, and Epigeocarpum japonicum sp. nov.

Morphological analyses of three glomoid spore-producing fungi suggested that two of them were undescribed species of Glomeraceae (phylum Glomeromycota), and the third differed slightly from Dominikia glomerocarpica and Epigeocarpum crypticum, recently described in Glomeraceae. The first two fungi originated from the Mediterranean Sea sand dunes of the Peloponnese, Greece, and the third was originally found in a tree plantation in Yokohama City, Japan. Phylogenetic analyses of sequences of the 45S nuc rDNA region and the RPB1 gene showed that (i) the three fungi belonged to Glomeraceae; (ii) the first two represented a new genus, here described as Complexispora gen. nov. with C. multistratosa sp. nov. and C. mediterranea sp. nov. and (iii) the third enlarged the monospecific genus Epigeocarpum, as E. japonicum sp. nov.


Introduction
The family Glomeraceae with two genera, Glomus and Sclerocystis, was originally erected in the order Endogonales, class Zygomycetes (Pirozynski and Dalpé 1989). Except for Glomus and Sclerocystis species, which have been known to form arbuscular mycorrhiza and fungi suspected to live in such symbiosis, the order contained taxa that differed fundamentally in the mode of reproduction and nutrition, as, for example, Endogone species (Gerdemann and Trappe 1974). Therefore, Morton and Benny (1990) transferred Glomeraceae to a new order, Glomerales in Zygomycetes, which included solely arbuscular mycorrhizal fungi (AMF). Based on phylogenetic analyses of sequences of the 18S nuc rDNA gene, Redecker et al. (2000) found that Glomerales was polyphyletic, and Schwarzott et al. (2001) distinguished in Glomeraceae Glomus-groups A and B. Glomus-group A was represented, among others, by G. caledonium, G. mosseae and G. intraradices, and Glomus-group B by G. claroideum and G. lamellosum. Following phylogenetic analyses of sequences of the same gene, Schüßler et al. (2001) accommodated Glomerales in a newly erected phylum, Glomeromycota. Schüßler and Walker (2010) sequenced G. macrocarpum, designated it as the type species of Glomus and Glomeromycota, and, consequently, retained in Glomeraceae only species of Glomus-group A sensu Schwarzott et al. (2001); Glomus-group B was raised to the rank of a monogeneric family, Claroideoglomeraceae with Claroideoglomus, in Glomerales. Oehl et al. (2011) distinguished in Glomeraceae six subgroups depending on the shape of the spore subtending hypha, the thickness of its wall, the mode of pore closure and the formation of spores singly, in clusters, or in un-or organized glomerocarps (= sporocarps). Consequently, the researchers distributed glomoid spore-producing species of Glomeraceae in four genera. The feature linking these subgroups was the similar pigmentation of the subtending hyphal wall at and below the spore base to that of the spore wall in coloured spores. In coloured glomoid spores of Entrophospora species (Entrophosporaceae, Entrophosporales; Błaszkowski et al. 2022b), which are morphologically closest to members of Glomeraceae, the subtending hyphal wall, even at the spore base, is colourless or conspicuously lighter than the spore wall (Oehl et al. 2011;Błaszkowski et al. 2022b). However, these characters may be invisible in colourless spores due to lack of contrast. In addition, the presence of potential cryptic species, the simplicity of construction of spores, and the high phenotypic variability of the spore components may complicate the taxonomic diagnosis of members of Glomeraceae based on morphological analyses only.
Characterizing, identifying and classifying members of the Glomeraceae and other taxa of Glomeromycota based on morphological and phylogenetic analyses may also be difficult or impossible because a large number of species have not yet been sequenced. For instance, although the Glomeraceae is the richest family in Glomeromycota, only ca. 57% of species of this family, distributed in 18 genera, have known molecular phylogeny. Interestingly, of the 18 genera, 13, including the genus Epigeocarpum, were recognized only in the last 9 years (Błaszkowski et al. 2015;2018a, b, 2021aCorazon-Guivin et al. 2019a-c, Sieverding et al. 2014. So, it is expected that soon further new taxa at different ranks will be revealed among the still not sequenced members of Glomeraceae. Moreover, such members of Glomeraceae can also be among the ca. 23% of glomoid spore-producing species, currently classified in other families of Glomeromycota, with uncertain or unknown phylogenies. The aim of the studies described in this paper was to characterize the morphology and phylogeny of three fungi producing spores with features of species of Glomeraceae, including a species of almost identical morphology to the recently newly described Dominikia glomerocarpica and Epigeocarpum crypticum (Błaszkowski et al. 2021a).

Origin of study material, establishment and growth of trap and single-species cultures, extraction of spores and staining of mycorrhizal structures
The potentially new species were initially named Species 1-3. Spores of Species 1 and 2 were first grown in trap pot cultures. Then, the spores, as well as spores of Species 3, extracted from a field-collected glomerocarp (= sporocarp), were used to establish single-species pot cultures. Many attempts at establishing such cultures with Species 2 and 3 failed. Therefore, Species 1, 2 and 3 were characterized based on spores extracted from both types of cultures, only trap cultures and only a glomerocarp, respectively. The trap cultures, in which Species 1 and 2 were found, had been originally inoculated with field mixtures of rhizosphere soils and root fragments collected under two specimens of Ammophila arenaria (L.) Link ca. 100 m apart from each other. The plants colonized maritime sand dunes of the beach Voidokoilia (36°57′N 21°39′E) located on the Peloponnese Peninsula, Greece. The soil samples were collected by Dr. Dimitris Arrianas on July 18, 2012 and J. Błaszkowski on September 8, 2015. Data about the climate and soil chemical properties of the sampled site are in Błaszkowski et al. (2019). The glomerocarp of Species 3 was collected on September 22 of 2020 from beneath the litter in a plantation of Cryptomeria japonica (Thunb. Ex L.F.) D. Don in Yokohama City, Kanagawa Prefecture, Japan (35°30′N 139°30′E). The climate of the sampling site is warm and humid with an average annual temperature of 16.9°C and an annual rainfall of 1937 mm recorded in 2019 at the nearest meteorological station Yokohama by the Japan Meteorological Agency.
Methods used to establish trap and single-species cultures, growing conditions and methods of spore extraction and staining of mycorrhizal structures were as those described previously . Five to ten spores of uniform morphology of each AMF species were used to establish single-species cultures.

Microscopy and nomenclature
Morphological features of spores and phenotypic and histochemical characters of spore wall layers of the new species presented here were characterized based on 50-100 spores of each species mounted in water, lactic acid, polyvinyl alcohol/lactic acid/glycerol (PVLG, Omar et al. 1979) and a mixture of PVLG and Melzer's reagent (1:1, v/v). The preparation of spores for study and photography were as described previously (Błaszkowski 2012;Błaszkowski et al. 2012). The types of spore wall layers were defined by Błaszkowski (2012) and Walker (1983). Colour names were from Kornerup and Wanscher (1983). Nomenclature of fungi and the authors of fungal names are from the Index Fungorum website http:// www. index fungo rum. org/ Autho rsOfF ungal Names. htm. The terms "glomerospores" and "glomerocarps" were used for spores and sporocarps, respectively, produced by AMF as proposed by Goto and Maia (2006) and Jobim et al. (2019).
Voucher specimens of the proposed new species [spores permanently mounted in PVLG and a mixture of PVLG and Melzer's reagent (1:1, v/v) on slides] were deposited at ZT Myc (ETH Zurich, Switzerland; holotypes) and in the Laboratory of Plant Protection, Department of Environmental Management (LPPDEM), West Pomeranian University of Technology in Szczecin, Poland (isotypes).

DNA extraction, PCR, cloning and DNA sequencing
Genomic DNA of each of the three species was represented by eight samples. Each sample of Species 1 and 2 contained DNA extracted from single or 2-5 spores and Species 3 from fragments of a glomerocarp, each containing ca. 10-30 spores. The method of processing the spores prior to PCR, conditions and primers used for PCR, as well as cloning and sequencing of PCR products to obtain 45S sequences of the three species were as those described by Błaszkowski et al. (2021b). Primers and PCR conditions to obtain RPB1 amplicons are summarized in Table S2. Both 45S and RPB1 sequences were deposited in GenBank (45S: OQ437298-OQ437315; RPB1: OQ435056-OQ435061).

Phylogenetic analyses
Preliminary BLAST analysis on the 45S sequences obtained suggested that Species 1-3 are undescribed taxa of Glomeraceae. To confirm the status and find position of these species within Glomeraceae, three alignments were produced (45S, RPB1 and 45S + RPB1) separately using MAFFT 7 (Katoh et al. 2019) with the E-INS-i option (http:// mafft. cbrc. jp/ align ment/ server/). The 45S alignment contained 95 sequences of the 45S region or its part, which characterized Species 1-3 and 24 representative species of all recognized (17) genera of Glomeraceae, except for Simiglomus. For Sclerocystis sinuosa, to increase the phylogenetic signal, the 45S sequence was obtained by concatenation of the overlapping sequences AJ133706 (18S), AJ437106 (45S) and FJ461846 (28S) obtained from the same isolate MD126. The RPB1 alignment consisted of 61 RPB1 sequences of those species of the 45S alignment that have sequences of the two loci. This alignment represented 19 species in 16 genera of Glomeraceae and Species 1-3. The 45S + RPB1 alignment was produced by manual concatenation of the 45S and RPB1 alignments. In all alignments, the outgroup was represented by sequences of the same three Diversispora species (D. epigaea, D. spurca and D. trimurales).
The percentage intraspecific sequence variations of Species 1-3 and divergences of sequences of these species from those of their closest relatives ( Fig. 1; Figs. S1, S2) were calculated using BioEdit (Hall 1999). All comparisons were performed on sequences of the same length.
The phylogenetic position of Species 1-3 was reconstructed based on Bayesian inference (BI) and maximum likelihood (ML) phylogenetic analyses of the 45S and 45S + RPB1 alignments, performed via CIPRES Science Gateway 3.1 (Miller et al. 2010). The 45S alignment was divided into five partitions: 18S, ITS1, 5.8S, ITS2 and 28S. Five additional RPB1 partitions, representing 3 exons and two introns, were added in the alignments with the concatenated genes, similarly as in Błaszkowski et al. (2021a). In addition, a tree with only RPB1 sequences partitioned as in the analysis of the 45S + RPB1 alignment was generated to reveal potential conflict in the topologies of the trees obtained from analyses of sequences of single loci.
GTR + I + G was chosen in both BI and ML analyses as nucleotide substitution model for each nucleotide partition (Abadi et al. 2019). Substitution models selected by Model-Test-NG 0.1.5 (Darriba et al. 2020) were also tested in ML analyses but the trees obtained final log likelihood values lower than those where GTR + I + G was used.
Four Markov chains were run over 1 million generations in MrBayes 3.2 (Ronquist et al. 2012), sampling every 1000 generations, with a burn-in at 25% of sampled trees. The ML phylogenetic tree inference was performed with RAxML-NG 1.0.1 (Kozlov et al. 2019), using a maximum likelihood/1000 bootstrapping run and ML estimated proportion of invariable sites and base frequencies. The alignments and tree files were deposited as online resources.
We assumed that clades were supported when the Bayesian posterior probabilities and the ML bootstrap values were ≥ 0.95 and ≥ 70%, respectively. The phylogenetic trees were visualized and edited in TreeGraph 2 (Stöver and Müller 2010). To evaluate possible conflicts between the genes, the topologies of the ML trees (collapsed at bootstrap values < 70%) were compared. In addition, the trees were compared based on four measures referred to the ingroup: (i) the number of species clades supported with BI ≥ 0.95 and ML ≥ 70%, (ii) mean BI and ML values for species clades when supported, (iii) mean supports of nodes with BI ≥ 0.95 and ML ≥ 70% and (iv) the amount of resolution of each tree. The amount of resolution was calculated for the ingroup of the tree, dividing the number of significantly supported internal branches by the size of the ingroup (Thorley and Wilkinson 2000).
To detect possible other findings of the potential three new species, their 45S sequences were used as queries in BLASTn to retrieve nucleotide sequences from GenBank. The sequences were selected according to the percentage of identity > 96 with at least one of the queries. Sequences were clustered in Mothur 1.39 (Schloss et al. 2009) using a dissimilarity cutoff of 1%. The possible relatedness with any of the three new species was assessed repeating the BI and ML phylogenetic analyses of the 45S locus, including the representative environmental sequences in the 45S alignment.

General data and phylogeny
Among the analysed 45S and RPB1 sequences, 18 and 6, respectively, were newly obtained in this study. The ranges of intraspecific variability of 45S sequences of Species 1-3 were 0.2-1.2%, 0.3-3.3% and 0.2-0.5%, respectively, and the variability of RPB1 sequences of these species did not exceed 0.1%. Data about the numbers of variable and parsimony informative sites of the alignments are presented in Table 1.
The topologies of the trees generated in BI and ML analyses of the 45S and 45S + RPB1 alignments were identical ( Fig. 1;  Fig. S1). In all trees, Species 1 and 2 sequences clustered in two separate species clades belonging to a clade at the rank of genus. This new generic clade formed a highly supported clade together with the Glomus and Sclerocarpum clades, occupying a sister position to the Glomus clade. In all analyses, Species 1 and 2, as well as the generic clade containing the two species obtained full or very high supports (BI = 1.0; ML = 97-100%). Instead, the node linking the new generic clade with the Glomus clade was not supported in any analysis.
All analyses of both alignments placed Species 3 in a species clade sister to the clade with E. crypticum sequences. Both the species clade and the generic Epigeocarpum clade were fully supported ( Fig. 1; Fig. S1). All these analyses also showed that the sister of Epigeocarpum is the monospecific genus Kamienskia with K. bistrata.
The RPB1 tree showed similar statuses and relationships of the analysed species as the 45S and 45S + RPB1 trees, when the species had RPB1 sequences. Also, the supports of the new generic clade Complexispora and all species clades were significant.
In all trees, the considered measurements regarding the ingroup were identical or similar, except for the RPB1 tree where the BI and ML resolutions were slightly lower than in the 45S and 45S + RPB1 trees (Table S1).

Taxonomy
The results of the phylogenetic analyses and sequence comparisons described above proved that Species 1-3 are new taxa of Glomeraceae, of which Species 1 and 2 (below described as Complexispora multistratosa and C. mediterranea, respectively) represent a new genus (named Complexispora), and Species 3 extended the monospecific genus Epigeocarpum as E. japonicum. Diagnosis: Differs from glomoid spore-producing species of other genera of Glomeraceae in (i) having a swelling, colourless, laminate spore wall layer surrounding the Genus description: Producing hypogeous glomoid spores singly and in clusters. Spores yellow to brown, usually globose to subglobose, 62-230 µm diam., with a spore wall composed of four to six layers, of which two consist of tightly adherent sublayers (laminae). Only spore wall layer 1 stains in Melzer's reagent. Subtending hypha cylindrical to funnel-shaped, concolorous with the spore wall layer, with a wall composed of layers continuous with all spore wall layers, except for the innermost layer. Subtending hyphal pore closed by a septum continuous with the innermost spore wall layer, occasionally also by a septum connecting the inner surfaces of the main structural laminate spore wall layer. Forming mycorrhiza with vesicles and arbuscules staining dark in 0.1% Trypan blue. Etymology: multistratosa (Latin), referring to the multilayered spore wall of this species.

Description of new species
Diagnosis: Differs from other glomoid spore-producing species with a six-layered spore wall in (i) the phenotypic and histochemical properties of the two outer spore wall layers surrounding the other spore wall layers, (ii) morphology of the spore pore closure and (iii) nucleotide composition of sequences of the 45S nuc rDNA region and the RPB1 gene (see "Discussion" for details).

Complexispora mediterranea
Diagnosis: Differs from other glomoid spore-producing species with a four-layered spore wall in (i) the phenotypic properties of the outermost spore wall layer surrounding the other spore wall layers and (ii) nucleotide composition of sequences of the 45S nuc rDNA region and the RPB1 gene (see "Discussion" for details).
Ecology and distribution: In the field, associated with roots of A. arenaria in the Peloponnese Peninsula, Greece. The geographic position, climate and soil chemical properties of the sampled site are characterized in Błaszkowski et al. (2019). Not found in 239 other trap cultures inoculated with rhizosphere soils from other dune sites of the Peloponnese. Our sequence comparisons and phylogenetic analyses suggested that two 45S sequences (HE775333 and HE775339) ascribed to uncultured Glomeraceae might represent C. mediterranea. The identity of these sequences to the C. mediterranea 45S sequences is 97-98%. The sequences were obtained from roots of Briza media L. growing in Malesov, North Bohemia, Czech Republic. Attempts to grow C. mediterranea in singlespecies cultures with P. lanceolata as host plant failed. Etymology: Latin, japonicum, referring to Japan, origin of this species.
Diagnosis: Differs from E. crypticum, the sister relative, in (i) the reactivity of spore wall layers 2 and 3 in Melzer's reagent and (ii) nucleotide composition of sequences of the 45S nuc rDNA region and the RPB1 gene (see "Discussion: for details).
Ecology and distribution: The specimen of E. japonicum characterized in this study was found in a plantation of C. japonica in Yokochama City, Japan. The geographic position and climatic conditions of the sampling site are described in "Materials and methods"/ "Origin of study material". Sequence comparison and BI and ML analyses showed 24 environmental sequences, clustered in four OTUs (Fig. S3), whose phylogenetic affiliation suggested conspecificity to those of E. japonicum. All sequences originated from Japan. Most of them were obtained from five specimens (CE1408, CE1506, CE1509, CE1708 and CE1709) collected from different locations and originally identified as Glomus microcarpum. Attempts to grow E. japonicum in single-species cultures with P. lanceolata as host plant failed.

Discussion
The morphological and phylogenetic analyses of three glomoid spore-producing fungi presented above confirmed our hypotheses that they are undescribed species of the family Glomeraceae. Moreover, the phylogenetic analyses showed that two of these species clustered in a new generic clade of this family ( Fig. 1; Figs. S1, S2). Consequently, the clade was described under the name Complexispora gen. nov., and the two new species were characterized as C. multistratosa and C. mediterranea. The third fungus enlarged the monospecific genus Epigeocarpum (Błaszkowski et al. 2021a) and was described as E. japonicum sp. nov.
The only morphological structure that appears to be diagnostic for Complexispora is the swelling, colourless, laminate spore wall layer surrounding the main, much thicker, structural laminate spore wall layer 3a;4b,c,e). A spore wall layer with a similar phenotype and location Fig. 5 a-h Epigeocarpum japonicum. a. Glomerocarp with spores (sp). b. Glomerocarpic spores (sp) and glebal hypha (gh). c-f. Spore wall layers (swl) 1-3; septum (s) formed by invagination of swl3 in the slot of swl2 connecting the subtending hyphal lumen with the spore interior and glebal hyphae (gh) are visible in c and d-e, respectively. g, h.
Subtending hyphal wall layers (shwl) 1 and 2 continuous with spore wall layers 1 and 2; glebal hyphae (gh) and the insertion of swl3 forming a septum (s) in the slot between the subtending hyphal lumen and the spore interior are indicated in g and h, respectively. a. Dry specimen. b-d and g. Spores in PVLG. e, f and h. Spores in PVLG + Melzer's reagent. a. Light microscopy. b-h. Differential interference microscopy. Scale bars: a = 200 μm, b = 20 μm and c-h = 10 μm in the spore wall was previously revealed only in Funneliformis caesaris, originally described as Glomus caesaris (Fig. 4e-h;Oehl et al. 2002), which potentially belongs to this genus.
Morphologically, C. multistratosa most resembles F. caesaris. Both species produce single spores, which are similar in colour, size, the number of spore wall layers as well as in the position and phenotypic features of spore wall layers 1 and 5 (Figs. 2a-h and 3a-h) (Oehl et al. 2002). Spore wall layer 2 in both species swells in d). However, in F. caesaris, this layer is a permanent structure ( Fig. 3e-h), while in C. multistratosa, it deteriorates with age and may be completely sloughed off in older specimens (Fig. 3b, c). The spore wall of both species contains layer 3, which tightly covers the laminate layer 4 (Figs. 2b-h and 3a-h), but spore wall layer 3 of F. caesaris may be coloured (Fig. 3e, f vs. is always hyaline in C. multistratosa; Figs. 2b-h and 3a-d) and over twofold thicker. Moreover, the spore wall of F. caesaris is ca. 1.7-fold thinner than that of C. multistratosa and the spore wall components staining in Melzer's reagent in these species are layers 2 and 1, respectively (Figs. 2f-h and 3a, d). Importantly, in F. caesaris the spore subtending pore is closed only by spore wall layer 5, and in C. multistratosa the pore is closed by spore wall layer 6, which is lacking in the F. caesaris spore wall, and occasionally also by a septum separating the subtending hyphal lumen from the spore interior ( Fig. 3b-d). Finally, at the F. caesaris spore base, the subtending hypha and its wall may be 1.4-fold and 1.7-fold narrower and thinner, respectively. Phylogenetically, the two species cannot be compared because no sequence of F. caesaris has been deposited in public databases.
The only other species characterized to form glomoid spores with a six-layered spore wall is Funneliformis fragilistratus, originally described as G. fragilistratum (Skou and Jakobsen 1989). Spores of this species are also similar in size, colour and the phenotypes of their spore wall layers 1-4. Skou and Jakobsen (1989) described spore wall layers 5 and 6 as "hyaline membranous walls, each 1-2 µm thick", which correspond to flexible to semi-flexible layers in the terminology currently used in characterizing Glomeromycota spore morphology. The presence of two such layers would render F. fragilistratus a unique species among all glomoid spore-producing species of Glomeromycota. However, this finding appears to be an artifact of the condition of spores and the way they were crushed. Most importantly, phylogenetic analyses of 18S sequences placed this species in the Funneliformis clade (Krüger et al. 2012).
Among non-sequenced species of Glomeromycota forming glomoid spores, only Redeckera avelingiae, described as G. avelingiae, and G. bagyarajii were described to have a spore wall composed of four layers, of which none is ornamented, and the innermost layer is flexible to semi-flexible (Mehrotra 1997;Sinclair et al. 2000). Both species also produce spores in clusters and the spores are coloured similarly to those of C. mediterranea, but the largest globose R. avelingiae spores are up to 1.5-fold smaller and have a 2.4-5.6fold thinner spore wall. Moreover, the subtending hypha of R. avelingiae may be up to 2.1-fold narrower at the spore base, and its pore is closed by a septum continuous with spore wall layer 3 (vs. by a protrusion of spore wall layer 4; Fig. 4g). In R. bagyarajii, the main structural, thickest, laminate spore wall component is layer 2 (vs. layer 3 in C. mediterranea; Fig. 4b-g) and the spore wall is ca. 2.1-fold thinner. In addition, at the spore base the R. bagyarajii subtending hypha is 1.2-1.5-fold narrower, has a 1.9-3.5-fold thinner wall, and its pore may be up to 1.7-fold narrower. Finally, none of the spore components of R. avelingiae and G. bagyarajii stains in Melzer's reagent, as does spore wall layer 1 of C. mediterranea (Fig. 4e, f).
The only phenotypic character separating E. japonicum from E. crypticum is the reactivity of their spore wall components in Melzer's reagent (Błaszkowski et al. 2021a). In E. crypticum, only the thickest, laminate middle spore wall layer 2 stains in this reagent, while in E. japonicum the staining reaction occurs in the thinnest, flexible innermost layer 3 and often in some innermost laminae of the laminate layer 2 (Fig. 5e, f, h). However, this histochemical property probably is of little phylogenetic significance as Kaonongbua et al. (2010) concluded with respect to layer 2 of the innermost spore wall 3 of Acaulospora species. Instead, the 45S and RPB1 sequence divergences strongly confirmed the separateness of the two species (see "General data and phylogeny").
As we mentioned in "Ecology and distribution" regarding E. japonicum, this species has likely been previously identified in Japan as Glomus microcarpum because its 45S sequences were almost identical to the 45S sequences of our new species. However, we treat this species assignment as incorrect for the following reasons. First, the morphology of E. japonicum does not match the morphology of G. microcarpum originally described by the Tulasne brothers (Tulasne and Tulasne 1845) and that of the type material characterized by Berch and Fortin (1984) and Koske et al. (1986). Tulasne and Tulasne (1845) and Berch and Fortin (1984) described G. microcarpum with a one-layered spore wall. Koske et al. (1986) found that the G. microcarpum spore wall consists of two layers: a unit outer layer and a much thicker, laminate inner layer. Gerdemann and Trappe (1974), based on examination of spores originating from the field and cultures, found the presence of one laminate spore wall layer in specimens of this species coming from Idaho, Oregon and Washington. Błaszkowski (2012) and Oehl et al. (2011) distinguished two layers in the G. microcarpum spore wall and spores identified as G. microcarpum with such a spore wall structure were found in Northeast Brazil (unpublished data). Thus, there is no consensus in the literature on the morphology of G. microcarpum and, regardless of whether the spore wall structure of the true G. microcarpum is one-or two-layered, E. japonica clearly differs in this respect. Second, G. microcarpum has not yet been sequenced, although it has been described as one of the first two species of this group of fungi and until 2010 regarded as the type species of Glomus (e.g., Gerdemann and Trappe 1974). Therefore, in order to lift the doubts about the morphology of and to know the phylogenetic placement of G. microcarpum in Glomeromycota, its molecular phylogeny should be determined. This should be reconstructed from analyses of an epitype of G. microcarpum designated from a specimen (s) collected from Paris or its vicinity, where this species was originally discovered (Tulasne and Tulasne 1845).
The present study and literature data prove that epigeous and hypogeous glomerocarpic species of Glomeromycota differ strongly phylogenetically. Currently, they are distributed in eight genera belonging to two orders (Błaszkowski et al. 2021a(Błaszkowski et al. , b, 2022aJobim et al. 2019). However, the phylogenies of a number of glomerocarpic species remain unknown and certainly many species are undiscovered mainly due to much greater difficulty in finding them than other members of Glomeromycota . It is expected that these yet uncharacterized glomerocarpic fungi will represent new taxa at different ranks.
We excluded from phylogenetic analyses the monospecific Simiglomus with S. hoi. The genus was erected based on two sequences of the 18S gene (Oehl et al. 2011), which in our alignments is represented only by approximately 240 base pairs. In addition, the origination of these sequences is uncertain, as stated by Redecker et al. (2013). Further studies are needed to clarify the status of Simiglomus in Glomeraceae.