Sareomycetes: More Diversity than Meets the Eye


 Since its resurrection, the resinicolous discomycete genus Sarea has been accepted as containing two species, one with black apothecia and pycnidia, and one with orange. We investigate this hypothesis using three ribosomal (nuITS, nuLSU, mtSSU) regions from and morphological examination of 70 specimens collected primarily in Europe and North America. The results of our analyses support separation of the traditional Sarea difformis s.l. and Sarea resinae s.l. into two distinct genera, Sarea and Zythia. Sarea as circumscribed is shown to comprise three phylospecies, with one corresponding to Sarea difformis s.s. and two, morphologically indistinguishable, corresponding to the newly combined Sarea coeloplata. Zythia is maintained as monotypic, containing only a genetically and morphologically variable Z. resinae. The new genus Atrozythia is erected for the new species A. klamathica. Arthrographis lignicola is placed in this genus on molecular grounds, expanding the concept of Sareomycetes by inclusion of a previously unknown type of anamorph. Dating analyses using additional marker regions indicate the emergence of the Sareomycetes was roughly concurrent with the diversification of the genus Pinus, suggesting that this group of fungi emerged to exploit the newly-available resinous ecological niche supplied by Pinus or another, extinct group of conifers. Our phylogeographic studies also permitted us to study the introductions of these fungi to areas where they are not native, including Antarctica, Cape Verde, and New Zealand and are consistent with historical hypotheses of introduction.


Introduction
Conifers, particularly in the families Araucariaceae, Pinaceae, and Cupressaceae, produce resins in their tissues (Langenheim 2003) as part of a complex defense system to protect against herbivores (Smith 1961;Rudinsky 1966; van Buijtenen and Santamour 1972), pathogenic fungi (Whitney and Denyer 1969;Gibbs 1972;Hart et al. 1975;Yamada 2001), protists (Krupa and Nylund 1972;Bunny and Tippett 1988), and bacteria (Hemingway and Greaves 1973;Hartmann et al. 1981). To protect against fungi, resins have the potential to act in several different manners. First, they present a physical barrier to penetration by fungal hyphae (Verrall 1938;Shain 1971;Rishbeth 1972;Prior 1976). When soft, resin can ow, trapping fungal hyphae and spores; when hard, the resin is di cult to penetrate. Furthermore, the components of the resin can inhibit the growth of fungi, acting as a chemical barrier (Cobb et al. 1968;Hintikka 1970;De Groot 1972;Fries 1973;Väisälä 1974;Chou and Zabkiewicz 1976;Bridges 1987;Yamamoto et al. 1997). Despite this apparently inhospitable environment, there are a number of so called "resinicolous" fungi that have evolved to exploit this niche (Cappelletti 1924; Selva and Tuovila 2016).
The study of fungi growing on conifer resins has a long history, dating back to the fathers of mycology (Persoon 1801;Fries 1815Fries , 1822). The rst species described was Helotium aureum, described in 1801 by Christiaan Persoon, though he made no mention of the resinicolous habit (Seifert and Carpenter 1987).
Thus, the rst author to describe fungi dwelling on resin was Elias Fries, who described three such fungi in 1815. Sphaeria resinae and Lecidea resinae were described as sharing the same habitat and easily confused; these were later determined to represent the anamorphic and teleomorphic states of the same fungus, currently known as Sarea resinae (Fr.) Kuntze (Ayers 1941;Hawksworth and Sherwood 1981). The third species, Racodium resinae, described from Picea resin, is a synnematous hyphomycete now called Sorocybe resinae (Fr.) Fr. (Seifert et al. 2007). These three Friesian species were followed by Cytospora resinae, described by Christian Ehrenberg in 1818 (Ehrenberg 1818); this was later determined to be a synonym of Fries' Sphaeria resinae (Fries 1823; von Thümen 1880). The last of these early species was described in 1822, again by Fries, as Peziza difformis [now Sarea difformis (Fr.) Fr.]. No new resinicolous taxa were noted until Arnold (1858).
The two species assigned to the genus Sarea, S. resinae and S. difformis, are the most commonly collected and reported of these resinicolous fungi. A search of the Global Biodiversity Information Facility (GBIF) database for S. resinae yielded 1261 records, and a search for "Sarea resinae" on Google Scholar yields 249 results, with S. difformis giving 519 records and 196 results, respectively. In contrast, Sorocybe resinae gives only 24 records and 56 results (accessed 13 July 2020). In addition to frequent reports, the two Sarea species have also been a subject of some interest regarding their systematic placement, which has been unclear (Reeb et al. 2004;Miadlikowska et al. 2014). A recent study resolved the uncertainty and has supported the erection of a new class in Pezizomycotina, Sareomycetes (Beimforde et al. 2020). This study, as well as a recent study that yielded 31 endolichenic isolates of Sarea species (Masumoto and Degawa 2019), have illustrated that both Sarea species are genetically diverse. This pattern is present in published sequences of both Sarea species deposited in public repositories. Sequence similarity and phylogenetic analyses also suggest that Arthrographis lignicola Sigler, though morphologically unlike Sarea species, is a close relative (Giraldo et al. 2014). This, combined with the wide distributions of these species, suggest a higher than known diversity, both obvious and cryptic, in Sareomycetes. The aim of this study is to assess this diversity.
To assess this diversity within Sareomycetes, an integrative taxonomic approach was employed. Fresh and fungarium specimens of orange (Sarea resinae) and black (S. difformis) species from around the world were borrowed or collected and examined morphologically. Where possible, DNA was extracted, and several regions ampli ed and sequenced. Two multi-locus datasets were assembled to investigate species boundaries and their phylogenetic relationships, as well as to provide further insights on the evolutionary history of Sareomycetes on a temporal and spatial scale.

Specimens Examined and Microscopic Examination
During the course of this study, a number of specimens of Sarea were collected and examined by the authors. The host range and distribution of these specimens were broad, with collections from the United States (California, Georgia, Maine, Massachusetts, Minnesota, New Hampshire, Rhode Island and Vermont) made by the rst author and collections from Austria, Cape Verde, Spain and Switzerland made by the second author. Further specimens were collected by and lent by Tomás J. Curtis (Ohio), Alden C. Dirks Microscopic examination of hymenial elements was conducted using free-hand sections cut under a dissecting microscope (Wild M5; Leica Geosystems, Heerbrugg, Switzerland) and of the excipulum using sections made on a freezing microtome. Microtome sections were prepared by stabilizing water-hydrated apothecia on a freezing stage (Physitemp BFS-MP; Physitemp Instruments LLC, Clifton, NJ) with a diluted gum arabic solution and sectioning with a sliding microtome (Bausch & Lomb Optical Co., Rochester, NY) set at approximately 25 µm. The resulting sections were applied serially to a clean glass slide and allowed to adhere by drying in the remaining gum arabic. Slides were prepared under a dissecting microscope (Olympus SZX9; Olympus Corporation, Tokyo, Japan) and studied with a compound microscope (Olympus BX40; Olympus Corporation, Tokyo, Japan). Digital images were captured with an Olympus XC50 USB camera (Olympus Corporation, Tokyo, Japan). Hand sections were studied with a compound microscope (Motic B1; Motic, Hong Kong, China). Except for two fresh collections studied alive in tap water ( Fig. 1, b1-d2, Fig. 2, b1-d3) and a culture studied on PDA (Fig. 1, n), all the other specimens ( Fig. 1, g1-m2, o1-o4, Fig. 2, e2-e9, f2-f9, g2-g9, h2-h9, i2-i9, j2-j9, k2-k9, l2-l9, m2-m9, Fig. 3, b1-d4), were pre-treated in 5% KOH prior to morphological studies. Melzer's reagent (MLZ) was used to test amyloidicity and Congo red (CR) to contrast cells walls. Images were captured with a Moticam 2500 USB camera and processed with the software Motic images Plus 2.0 (Motic, Hong Kong, China). The 95% con dence intervals of the median were calculated with SPSS 15.0 (SPSS Inc., Chicago, IL) for each morphological feature. Measurements are given as follows: (the smallest single measurement) smallest value for percentile of 95% -Largest value for percentile of 95% (largest single measurement). Whenever possible, biometric values are based on ≥ 10 measurements for each character on an individual specimen. Culturing Some specimens were grown in axenic culture. Cultures were generated from ascospore shoots. A living apothecium was placed oriented upward on a dab of petroleum jelly on a lter paper. This assemblage was then placed in the lid of an upside-down, sterile petri dish containing either potato dextrose agar (PDA) or cornmeal agar (CMA) prepared according to the manufacturer's instructions (HiMedia Laboratories Pvt. Ltd., Mumbai, India). The lter paper was saturated with water, and the chamber sealed with Para lm (Bemis Company, Inc., Neenah, WI). After incubation at room temperature for one or two days, the lid was removed and replaced with another sterile lid. The culture was then allowed to grow at 25 °C for up to one month before sampling. DNA Extraction, PCR, and Sequencing DNA extractions were performed from axenic culture when available and from fresh or preserved apothecia or pycnidia otherwise. Fresh or plentiful dried material was extracted by grinding 1-2 apothecia, 3-4 pycnidia, or a rice grain-sized slice of a culture and employing the DNeasy Plant Mini Kit (QIAGEN, Venlo, Netherlands) following the manufacturer's recommendations. Preserved or scanty material was extracted by (Castresana 2000) was used to automatically remove ambiguously aligned regions in the nuITS and mtSSU MSAs using the least stringent parameters but allowing gaps in 50% of the sequences.

Phylogenetic Tree Inference
The online version of RAxML-HPC2 hosted at the CIPRES Science Gateway (Stamatakis 2006; Stamatakis et al. 2008;Miller et al. 2010) was used to estimate a three-locus phylogeny under a Maximum Likelihood (ML) framework based on a dataset comprising specimens with at least two available sequenced markers. Several specimens of Pycnora were included as outgroup to root phylogenetic trees. Prior to concatenation, and to test for topological incongruence among sequence datasets, we inferred ML trees independently for each locus with RAxML-HPC2, using 1000 bootstrap pseudoreplicates, and assumed bootstrap values ≥ 70% as signi cant for con icting relationships among the same set of taxa (Mason-Gamer and Kellogg 1996). Because no con icts were detected, the RAxML analysis was conducted using the GTRGAMMA substitution model for the four delimited partitions (nuITS1 + 2, 5.8S, nuLSU, mtSSU) and 1000 rapid bootstrap pseudoreplicates were implemented to evaluate nodal support. Evolutionary relationships were additionally inferred in a Bayesian context using MrBayes v. 3.2.6 (Ronquist et al. 2012). Optimal substitution models and partition schemes for these four sequence data partitions were estimated with PartitionFinder v. 1.1.1 (Lanfear et al. 2012) considering a model with linked branch lengths and the Bayesian Information Criterion (BIC). This analysis favoured the SYM + Γ model for the nuITS1 + 2 partition, the K80 + I + Γ for the 5.8S + nuLSU, and the HKY + I + Γ for the mtSSU. The analysis was then conducted with two parallel, simultaneous four-chain runs executed over 5 × 10 7 generations starting with a random tree, and sampling after every 500th step. The rst 25% of data were discarded as burn-in, and the 50% majority-rule consensus tree and corresponding posterior probabilities were calculated from the remaining trees. Average standard deviation of split frequencies (ASDSF) values below 0.01 and potential scale reduction factor (PSRF) values approaching 1.00 were considered as indicators of chain convergence. Tree nodes showing bootstrap support (BP) values equal or higher than 70% (RAxML analysis) and Bayesian posterior probabilities (PP) equal or higher than 0.95 (MrBayes analysis) were regarded as signi cantly supported. Phylogenetic trees were visualized in FigTree v. 1.4 (available at http://tree.bio.ed.ac.uk/software/tracer/) and Adobe Illustrator CS5 was used for artwork.

Species Discovery-Validation Approach
Based on the existence of well-delimited and highly supported clades in the three-locus phylogenetic tree inferred above, we assessed species boundaries independently for the orange and black Sarea. To this end, we used the distance-based Automatic Barcode Gap Discovery method (ABGD) (Puillandre et al. 2012), restricting the analyses to specimens with available data for the fungal barcode nuITS. The analyses used the Kimura two-parameters (K2P) model to estimate genetic distances, a transition/transversion value of 3.95 (orange Sarea) and 3.07 (black Sarea) calculated with MEGA v.5.2 (Tamura et al. 2011), a Pmax of 0.01, and different values for the relative gap width (X). Subsequently, the Bayes Factor Delimitation (BFD) method, which allows for topological uncertainty in gene trees and incongruences among gene trees, was chosen to compare the two species boundary hypotheses generated for the black Sarea on the basis of our morphological study of the specimens, and the ABGD and phylogenetic results (Table 1). *BEAST (Heled and Drummond 2010; Drummond et al. 2012) was used to build the two competing models. These comprised a three-locus dataset in which specimens with identical sequences were removed to avoid sequence redundancies; the number of specimens left was 85, including outgroup specimens. The same optimal substitution models and partition schemes selected in the MrBayes analysis were used for the *BEAST analyses except for the substitution model TrNef + I + Γ, which was preferred for the 5.8S + nuLSU partition. An uncorrelated relaxed lognormal molecular clock was chosen for the three markers based on a preliminary assessment of the adequacy of strict clocks in MEGA 5.0 (Tamura et al. 2011) (see Table S2). The mean clock rate was xed to 1.0 for nuITS whereas rates were co-estimated for nuLSU and mtSSU under a uniform prior (1 × 10 − 5 , 5). A birth-death process tree prior was imposed after conducting preliminary Bayes factors comparisons of Maximum Likelihood Estimates (MLE) calculated with Path Sampling and Stepping-Stone (Lartillot and Philippe 2006; Xie et al. 2011) for models implementing alternative tree priors (see Table S2). By using this tree prior we accommodated incomplete sampling and speciation of nodes in the topology. The *BEAST analyses used a piecewise linear and constant root model for population size (Grummer et al. 2014). Hyperpriors for the birth-death process tree prior and species population mean were given an inverse gamma distribution with an initial value of 1 or 0.1, shape parameter of 1 or 2 and scale of 1 or 2, respectively. Default (but informative) priors were given for the remaining parameters across all analyses. Finally, *BEAST runs of 1.5 × 10 8 generations, saving every 15000th tree, were performed using the CIPRES Science Gateway (Miller et al. 2010 (Table S3).
Alignments of the nuSSU, nuLSU, mtSSU, RPB1, RPB2 and tef1-α were carried out in MAFFT v. 7.308 as implemented in Geneious v. 9.0.2 using the same algorithm parameters as above. Manual optimization of the resulting MSAs consisted in removing clearly ambiguously aligned and intronic regions in rDNA marker datasets (nuSSU, nuLSU and mtSSU), as well as non-coding regions (introns) in the protein-coding markers (RPB1, RPB2 and tef1-α). Sequences of the latter three datasets were also translated into amino acids to spot misaligned regions generating stop codons. Finally, "N"s were used to ll gaps at the ends of shorter Additional settings included selection of an uncorrelated lognormal relaxed clock for each marker and a birth-death prior, and the use of a rooted, strictly-bifurcating ML topology obtained in RAxML as a starting tree. This ML tree was previously transformed into ultrametric using the function chronos in the R package ape (Paradis et al. 2004). In the prior settings step, we forced to co-estimate the average rate of evolution of each locus by setting the priors for the ucld.mean parameter to uniform (10 − 5 , 0.01). The taxa and prior distributions used to set the fossil calibrations are detailed in

Molecular Sequence Data
Molecular data were obtained from 70 collections. From these, we produced 212 sequences: 70 nuITS, 63 nuLSU, 61 mtSSU, 9 RPB2, and 9 nuSSU (Tables S1 & S3 (nuLSU) and − 2375.8252 (mtSSU). The nuITS and mtSSU phylogenies showed strong nodal support for (a) a clade including all orange Sarea s.l. (hereafter referred to as Zythia resinae; see section Taxonomy below), and (b) a clade assigned to the new genus Atrozythia (see section Taxonomy below) including two species composed of a few specimens each (Figs. S1-S3). The two taxa referenced below as Sarea coeloplata 2 and S. difformis s.s. also formed well delimited and highly supported clades in these two phylogenies; however, S. coeloplata 1 was monophyletic with high support only in the mtSSU topology. A supported sister relationship was found for Zythia and Atrozythia, whereas a clade comprising the three Sarea species was only supported in the mtSSU topology, in which S. coeloplata 1 and S. difformis appeared as sister species. The nuLSU phylogeny only delimited the S. coeloplata 2 clade with support, and a specimen assigned to the new species A. klamathica was found interspersed in a non-supported clade including Z. resinae specimens (Fig. S2

Species Delimitation
The ABGD analyses conducted on the Zythia (orange specimens) and Sarea (epruinose black specimens) nuITS datasets did not reveal clear barcode gaps. In Zythia, ABGD rendered 6 to 52 different partitions (i.e., putative species) when the relative gap width (X) was set to 0.5, but initial and recursive partitions only converged in the 52-partitions solution (Fig. S4). With X = 1, convergence was found for 1 and 52-partition solutions. In agreement with our morphological data, we hereafter considered the existence of only one Zythia species. In Sarea, although a barcode gap was not strictly found, ABGD analyses using varying levels of X (0.5, 1 and 1.5) suggested the combination of specimens assigned to S. difformis and S. coeloplata 1 into one single partition (Fig. S5). As this solution contradicted our morphological observations of specimens suggesting the existence of three species in Sarea, a hypothesis in agreement with the multilocus phylogenetic results, we compared the two alternative species delimitation models with the BFD method. Marginal likelihood values for the considered models calculated through Path Sampling and Stepping-Stone are shown in Table 1. Bayes factor comparisons favoured the three species model over the two species model.

Genetic Polymorphism, Neutrality Tests and Phylogeographic Structure
Genetic diversity indices, such as the numbers of segregating sites and haplotypes, were greater for Zythia resinae than for any Sarea species across different markers ( Table 2). The nucleotide diversity index behaved in a similar way except for the mtSSU marker: though four times as many specimens of Z. resinae as S. difformis were included in their respective analyses, S. difformis showed slightly higher values than Z. resinae. Haplotype diversity values were comparable among species and markers, although S. coeloplata 2 consistently showed lower values. However, these results must be interpreted with caution due to the uneven number of studied specimens for each species, e.g. Z. resinae incorporated three to eight times more individuals in the analyses than the remaining species. Neutrality tests gave signi cant negative values of Fu's Fs in S. coeloplata 1 and Z. resinae based on nuITS data (Table 2), indicating a population expansion. Negative values of Tajima's D and Fu's Fs were also obtained for the same species as well as S. difformis using the nuLSU dataset; however, these were not statistically signi cant. Tajima's D tests of mtSSU data generated positive values for all species, Polymorphism statistics and neutrality tests results for each marker (nuITS, nuLSU and mtSSU), and Sarea spp. and Zythia resinae. Columns contain the number of sequences (n), their length (in bp), the number of positions in the alignment with gaps and missing data, the number of segregating sites (s), the number of haplotypes (h; value after vertical bar was calculated considering gaps in the alignment), haplotype diversity (Hd), nucleotide diversity (π) using the Jukes and Cantor (1969) correction, and results of neutrality tests. but these were not signi cant as well.
Tokogenic relationships among the 48 nuITS haplotypes of Zythia resinae revealed no geographic structure as haplotypes from North America, Northern/Central Europe and Eastern Asia were widespread across the network (Fig. 5A). Identical haplotypes were shared among widely distant regions: (a) North America and Eastern Asia, and (b) North America, the whole of Europe and the Macaronesian islands. The two studied New Zealand haplotypes were not closely related: whereas one was relatively close to a haplotype shared between North America and Eastern Asia, the other was linked to a haplotype shared between Northern/Central Europe and the Macaronesia. The Caribbean haplotype was close to a North American one. As for Sarea s.l., the network delimited the three considered species well (Fig. 5B). These showed differing levels of intraspeci c diversity. For instance, haplotypes of S. difformis were separated from each other by a higher number of mutations than haplotypes of S. coeloplata 1 & 2. At the geographical scale, whereas haplotypes from any of the considered Northern Hemisphere regions were widespread across the network, we found no haplotypes shared between widely distant localities, except for an Antarctic haplotype shared with Northern/Central Europe and the Iberian Peninsula. These observations may also be due to the limited number of specimens studied compared to the scenario revealed for Z. resinae. Finally, in S. coeloplata 1 & 2, some Iberian Peninsula and Macaronesian haplotypes showed an increased number of separating mutations; further, S. coeloplata 1 haplotypes from these two regions were closely related.

Age Estimates for the Crown Nodes of Sareomycetes and Main Lineages Within
The maximum clade credibility (MCC) tree with 169 fungal taxa and divergence estimates obtained with BEAST showed posterior probabilities (PP) of 1.0 for all inner nodes except for the sister relationship between the clades allocating Coniocybomycetes + Lichinomycetes and Lecanoromycetes + Xylobotryomycetes + Eurotiomycetes that received a support of PP = 0.96 (Fig. S6) The ve chronograms inferred for estimating a time frame for the diversi cation of Sareomycetes showed high posterior probabilities supporting relationships among the main lineages except for the sister relationship between Sarea difformis and S. coeloplata 1 (PP = 0.93-0.94). Similar to previous results, divergence ages obtained with Erysiphales and Melanohalea nuITS substitution rates generated much more recent time estimates (Table S5). All in all, the origin and diversi cation of Zythia, Atrozythia and Sarea occurred during the Tertiary (Table S5). Thus, the crown nodes of Zythia and Sarea were estimated in the Eocene and Miocene, whereas that of Atrozythia in the Oligocene-Miocene (Fig. 6). The split between the two Atrozythia species (A. klamathica and A. lignicola) probably occurred during the Miocene. The crown nodes of the three Sarea species were placed in the Oligocene-Miocene. Finally, the different dating strategies estimated that intraspeci c diversi cation in the three studied genera occurred < 10 Ma, in the Neogene and Pleistocene (Figs. S8-S12).

Taxonomy
Although the terms "holotype" and "lectotype" as de ned in Article 9 of the International Code of Nomenclature for Algae, Fungi, and Plants (ICN) (Turland et al. 2018) do not apply to names at ranks higher than species, they will be used by analogy here to indicate type species of monotypic genera or type species selected by their authors and type species selected by later authors, respectively (Art. 10, Note 1).
Exclamation points after specimen identi ers indicate that they were examined by the authors. Diagnosis: Apothecia of Atrozythia differ from Zythia in their color (black vs. orange) and from Sarea because of their white to light blue gray pruina. Paraphyses in Atrozythia are unbranched whereas those in Sarea are always branched or anastomose, at least in the basal cells. Zythia can have unbranched paraphyses but differs from Atrozythia in the amount and color of lipid guttules, orange and abundant vs. yellowish and sparse, respectively. Atrozythia has a hyaline ectal and medullary excipulum that are sharply delimited by a narrow dark brown pigmented layer; in Zythia there is no brown pigmented layer between these layers. In Sarea the medullary excipulum is always differentiated by its dark brown color.
Etymology: from the Latin for black (ater) and the genus name "Zythia," referring to the macroscopic resemblance to Zythia species, but with a dark coloration. Notes: This genus currently encompasses two species, both apparently uncommon or under-collected, with one known only in an apothecial teleomorphic state and the other only in a hyphomycetous anamorphic state. Both are found on dead or living conifers; there are some indications of a resinicolous habit in the type species, A. klamathica sp. nov., but additional information is needed to elucidate the ecology of these fungi. In our phylogenetic analyses, the a nities of this group apparently lie closer to the genus Zythia than to Sarea, but Atrozythia species are located on a relatively long branch compared to these two genera. There are apparently no closely matching, unnamed environmental sequences on GenBank assignable to this genus, possibly suggesting rarity rather than merely being overlooked. Individual cells at middle layer of ectal excipulum (5)6.5-9(10) × 2-3.5 µm at margin, (6.5)8.5-12(15.5) × 2-3 µm at lower ank and base, cell walls 0.5-1.5(3.5) µm thick. Medullary excipulum of slightly gelatinized textura intricata, tightly packed, cells neither with intercellular spaces nor particular orientation, Notes: This species is known from two specimens (of which the holotype was sequenced twice) and is illustrated in Fig. 3. It was probably observed once in Alaska (https://www.inaturalist.org/observations/41563051), but no specimen was collected. Little is known about its ecology or possible anamorphic states. Sequence and morphological data are su cient to separate it from Sarea and Zythia, and it shows a closer a nity to the latter. Although apparently collected only twice, it is possible (given the rarity with which Sarea difformis is found on cupressaceous hosts) that A. klamathica is the fungus which was isolated as an endophyte of cupressaceous plants in central Oregon and reported as S. difformis (Petrini and Carroll 1981). Due to the lack of detailed data in the report, this supposition can neither be con rmed nor refuted. Culture work with fresh material should be done. Emended Description: Apothecia discoid, roundish to ellipsoid, scattered or gregarious, erumpent from the resin, consistency coriaceous and apothecia slightly to moderately contracted when dry, expanding and eshy when moist, 0.2-1.3 mm diam., up to 0.5 mm high, sessile, entirely black (267.Black). Disc and receptacle rough; margin distinct, slightly raised when immature or dry but not protruding from the hymenium after rehydration, 0.5-1 mm thick, rough or radially cracked, concolorous with hymenium and receptacle. Hymenium and tissues in section light purple (222.l.P) to deep purple (219.deepP), pigments turning brilliant blue (177.brill.B) to deep blue (179.deepB) in KOH. Asci (34)46.5-53.5(78) × (9.5)12.5-14.5(18.5) µm, clavate, multispored, mature asci 10-30 µm below the hymenial surface prior to spore discharge, ascus dehiscence rostrate, inner wall material expanding, protruding c. 9-15 µm, reaching the hymenial surface at spore discharge; apex hemispherical, thick-walled, strongly staining in CR, apex with an apical chamber, apical wall 2-3.5 µm thick, chamber later disappearing and apical tip thickening, becoming 7-11 µm thick, projecting into the ascus, becoming dome-like, inner wall not or faintly amyloid, outer wall intensely amyloid; lateral walls 0.5-1.5 µm thick, asci covered with an amyloid gel layer; base short-stipitate and arising from a crozier. Ascospores (1.7)2.1-2.3(3) µm diam, globose to subglobose, hyaline, inamyloid, aseptate, wall slightly thick and with one eccentric medium gray (265.med.Gy) lipid guttule. Paraphyses embedded in gel, cylindrical, unin ated to slightly clavate, straight or slightly curved at the apex, terminal cell (4)6-7.5(11.5) × 1.5-2.5(3) µm, covered with a deep brown (59.d.Br) to brown black (65.brBlack) amorphous exudate, lower cells (4.5)7.5-8.5(11.5) × 1.5-2.5 µm, basal cells (6.5)9-10(12) × 1.5-2.5 µm, bifurcate in lower cells, hyaline, septate, septa strongly staining in CR, basal cells ± equidistantly septate, but lower and terminal cells shorter, walls smooth, sparse tiny yellow gray (93.yGray) lipid guttules in all cells. Excipulum at margin and upper (-lower) ank composed of two well differentiated layers, lower ank to base not always differentiated into two types of tissues. Ectal excipulum strongly gelatinized, (41)57-67(92) µm thick at lower ank and base, (28)49-60(86) µm thick at margin and upper ank, cells loosely packed and surrounded by a light grayish brown (60.l.gy.Br) to medium brown (58.m.Br) gel, running parallel each other (sometimes interwoven), frequently bifurcated and oriented perpendicular to the outer surface, cortical layer with shorter, parallel and very tightly packed cells without intercellular spaces, walls strongly pigmented and surrounded by a dark brown (59.d.Br) to brown black (65.br.Black) amorphous exudate. Ectal cells (6.5)10-12.5(18.5) × 1.5-3 µm at upper ank and margin, (7) Notes: The concept of Sarea difformis is here restricted to those specimens presenting a purple pigment in the hymenium which turns blue when a strong base is applied, a character clearly visible in one isoneotype (FH 00995483!) and illustrated in Fig. 1 Notes: A specimen collected by Norman at the type locality and stored under the name Biatorella coeloplata in TROM is here designated the lectotype. Norman (1884) described a form, Biatorella coeloplata f. carbonata, for older apothecia; we use a single specimen to lectotypify this form as well as the species. Since it is clear that even Norman considered the two forms merely different developmental stages of the same fungus, we see no reason to consider this form a separate taxon.
The type of Tympanis abietis was not available for examination from CO. Its true a nities are unclear, but Le Gal's (1953) statement "L'hyménium est plongé dans une matière brunâtre qui en agglutine les éléments" in her description of the holotype likely place it in one of the two clades we assign to S. coeloplata s. lat.; morphological re-examination of the type should be conducted to verify its placement.
The description above applies to both Sarea coeloplata 1 and Sarea coeloplata 2 as presented in our phylogenetic analyses. We have been unable to separate the two morphologically, and thus we cannot assign the examined type to one clade or the other. We have observed morphological variations among collections (illustrated in Fig. 1) and are con dent that the di culty of characterizing the members of these two clades may be overcome by careful analyses involving DNA analysis and morphological examination of single apothecia. This will avoid the problem of mixed collections. For more information, see our discussion of mixed collections below.  Emended Description: Teleomorph apothecial. Apothecia brilliant orange-yellow (67.brill.OY) to deep orange (51.deepO), erumpent from the resin, discoid to cupulate, roundish or slightly ellipsoid, coriaceous and darker when dry, eshy and lighter after rehydration, hymenium and receptacle concolorous, margin usually differentiated and protruding slightly beyond the hymenium; sessile with broad attachment, sub-stipitate to prominently stipitate. Hymenium and tissue colors not changing in KOH. Asci and ascospores exhibiting morphology and reactions as in Sarea. Paraphyses cylindrical, unin ated to slightly or moderately clavate, straight or bent at the apex, completely surrounded by gel that contains hyaline or grey yellow (90.gy.Y) amorphous lumps, all cells with a high amount of brilliant orange-yellow (67.brill.OY) to vivid orange-yellow (66.v.OY) lipid guttules; terminal cell and 1-2 cells below covered by medium yellow (87.m.Y) rough amorphous exudate; usually branched at apical cells or cells below, rarely unbranched, frequently with anastomoses, septa frequently constricted and equidistantly septate with terminal and lower cells shorter (moniliform). Excipulum and medulla not well differentiated in section, although two layers can be noted mostly from the margin to the anks because of the arrangement of cells and amount of pigments. Ectal excipulum in lower ank to margin strongly gelatinized, pigmented due to a high amount of brilliant orangeyellow (67.brill.OY) to vivid orange-yellow (66.v.OY) lipid guttules or not pigmented, cells moderately packed and running parallel each other and surrounded by hyaline gel sometimes including hyaline or grey yellow (90.gy.Y) amorphous lumps, cortical layer with shorter, parallel or unoriented, tightly packed cells without intercellular spaces, amorphous rough exudate covering the cortical cells, hyaline or colored between deep orange-yellow (72.d.OY) to brown orange (54.brO), usually more abundant at the margin, sometimes even appearing as glassy processes. Amyloid reaction present mostly in the ectal excipulum at the margin and anks, or absent. Medullary excipulum composed of textura intricata, cells changing from ectal excipulum to medulla progressively, hyaline, less spaced and gelatinized; subhymenium somewhat similar or differentiated from medulla because of the presence of pigmented lipid guttules, cells without intercellular spaces and without gel. Anamorph pycnidial; see descriptions of Pycnidiella Höhn. and Pycnidiella resinae (Ehrenb.) Höhn. in Sutton (1980: 544) and Sarea resinae (Fr.) Kuntze in Hawksworth & Sherwood (1981: 365).
Notes: The history of typi cation in the genus Zythia is somewhat complicated. This is due both to the sparse protologue and apparent confusion among some authors as to whether or not Fries' Sphaeria resinae had been a combination of Ehrenberg's Cytospora resinae. This has been discussed at length in a recent publication on the matter (  Emended Description: See emended description above for Zythia and notes below. Notes: The status of the basionym of Zythia resinae is somewhat confused, with authors treating Cytospora resinae either as a new name or as a new combination of Fries' Sphaeria resinae. Examination of the protologue (Ehrenberg 1818) shows no references, direct or indirect, to Fries' earlier name, and Ehrenberg explicitly includes his species in the index of new species and attributes it to himself ("mihi"); we thus accept this as having been a species novum. It is desirable to conserve Cytospora resinae with the same type as Sphaeria resinae (UPS F-541747) because these names are 1) almost always treated as synonyms, 2) share the same epithet (and thus will demand a replacement name for one if they are taken out of synonymy and included in the same genus), and 3) are likely indistinguishable based on morphological features. This has been proposed in another publication (Mitchell and Quijada In Press).
We do not provide an additional description for Z. resinae since at present it is the only accepted species in this genus, and our emended description of the genus serves as a description of this species. It has been noted, however, that collections in our phylogenetic analyses do exhibit morphological variation, some visible in Fig. 2. Examples of this variation were found in the excipular tissues, i.e.: slightly amyloid reaction in the excipulum of specimens in clade 8 (Fig. 2, j2), specimens with sessile apothecia in clades 3, 6 and 9 (Fig. 2, e1, i1, m1) vs. stipitate apothecia in clades 5 and 12 (Fig. 2, h1, k1), specimens with a strongly pigmented cortical layer in clades 2 and 3 (Fig. 2, f2, e2), an almost hyaline ectal excipulum in clades 1, 6 and 12 (Fig. 2, g2, i2, k2), ectal excipulum with high content of pigments in clades 9 and 13 (Fig. 2, m2, l2) and margin with glassy processes in clade 12 (Fig. 2, k2) (clade names are from Fig. S1). We also found examples of variation in the hymenium, i.e.: the presence of an additional amyloid thick gel layer in specimens in clade 3 (Fig. 2, e5), and paraphyses simple and not branched in the apical or lower cells in clades 6, 8 and 9 (Fig. 2, i9, j9, m9) vs. bifurcate or branched at apical cell in clades 2, 3, 6 and 12 (Fig. 2, f9,  e9, l9, k9). We have not separated species within what is almost certainly a species complex because of questions of the prevalence of mixed collections and our inability to examine type material of Lecidea resinae. For additional information, see our discussion of mixed collections below. The specimen issued as "Lecidea resinae Fr." under number 277 of Leighton's Lichenes Britannici Exsiccati (FH 00964658!) is Biatoridium monasteriense J. Lahm ex Körb., which had not been described at the time of issue (Leighton 1858). Mudd (1861), citing this and other specimens, described Z. resinae as having a green thallus, brown apothecia, a thin margin, ellipsoid spores, and having been collected on elms (Ulmus sp.). None of these traits characterize any species in Sareomycetes. That his conception of Z. resinae was incorrect and at least partly based on B. monasteriense is con rmed by Magnusson's examination and reidenti cation of one of Mudd's specimens in the Rehm herbarium (Magnusson 1935). Mudd (1861) also described the new variety Biatorella resinae var. rubicundula, which has been accepted as being an synonym of a Strangospora species (Fries 1874; Rehm 1889a); unfortunately, type material could not be located at K or BM for examination (Angela Bond & Gothamie Weerakoon, pers. comm.). Many subsequent authors cite specimens cited or issued by Mudd  A similar case to the preceding arose in Southern California around the turn of the twentieth century. Hasse reported Z. resinae from the area three times, rst in a publication by McClatchie (1897), then in two of his own (Hasse 1898(Hasse , 1908. He describes the substrate of the specimens as bark, and in the last publication describes the species with black apothecia turning brown when moist, and without margins. These features are all uncharacteristic of species in Sareomycetes. Examination of a specimen labelled "Lecidea (Biatora) resinae Fr." (= Zythia resinae) sent by Hasse to George Knox Merrill (FH 00964657!) revealed that it was a specimen of Strangospora moriformis (Ach.) Stein. Additionally, the collecting information matches that given in his 1898 publication, suggesting that this is the specimen he based that report on. An additional Farlow Herbarium specimen (FH 00480746!) matches the collecting information and description of the 1908 publication and was originally determined by Hasse as "Biatorella resinae (Fr.)" (= Zythia resinae) but later changed by him to "Biatorella moriformis (Ach.) Th. Fr." (≡ Strangospora moriformis) with the later identi cation con rmed by an annotation by Magnusson. These specimens, along with his description, suggest that his concept of Z. resinae was at the time partly or completely based on S. moriformis, but that he later realized his error. By 1913, Hasse removed Zythia resinae from his list of Southern California lichens entirely (Hasse 1913). We list the following misapplications:

Species Diversity
The number of species in Sarea s.l. (= Sareomycetes) has long been a matter of discussion. Hawksworth and Sherwood (1981) traced the idea of there being only a single species for both black and orange fungi to Johann Hepp's (illegitimate) publication of Peziza myriospora in his Die Flechten Europas (1857), noting that he designated two forms ("a" being orange and "b" being black). Hepp's designation of these two forms could not be veri ed by examining either the specimen in the complete, unbound exsiccata (FH 00964656!) or the specimen from the Patouillard Herbarium (FH 00964655!). Each contains a single specimen, and the labels make no mention of color or forms. The labels of the specimens at Kew were also examined (Lee Davies, pers. comm.), as well as the schedae with the same results (Hepp 1857a). It is possible that these "forms" were annotations in the bound, boxed specimens prepared by Hepp (Sayre 1969), but this has yet to be veri ed. A Sarea species dominates in both specimens in FH, and Hepp cited Peziza resinae (Fr.) Fr. (= Zythia resinae) as a synonym of his proposed new name; it is likely that he considered both orange and black fungi to be a single species (Hepp 1857b). Consideration of the orange and black apothecia as representing a single species carried into the 20th century (Nylander 1857b, 1866; Koerber 1865; Leighton 1872; Fink 1935). As stated by Hawksworth and Sherwood (1981), the orange and black fungi, each treated as a single species, rested in separate genera (for S. resinae, Biatorella; for S. difformis, Retinocyclus) for much of the 20th century. Based on morphological similarities, they were then united in a single genus, Sarea, where they stood as two separate species, easily differentiated by color (Hawksworth and Sherwood 1981).
The current study employed integrative taxonomy (Goulding and Dayrat 2016; Haelewaters et al. 2018; Lücking et al. 2020) to assess the number of species in Sareomycetes. In addition to two species in the new genus Atrozythia, one previously undescribed and one not previously recognized as a relative of this group, it was determined that the black and orange fungi deserve to each be treated in separate genera with at least three and one species, respectively. The black fungi are recognized as the core genus Sarea and are recovered here as three phylospecies and two morphospecies. Sarea difformis, the type species of the genus, is quite distinctive and specimens are easily identi able based on the purple pigment in the hymenium and (sometimes) stipe that turns blue in application of strong base (e.g. Figure 1, g1-5). The remaining morphospecies and two phylospecies represent Biatorella coeloplata, here combined as Sarea coeloplata; the type could not be assigned to a single phylospecies due to issues addressed in our discussion of Mixed Collections. The orange fungi are recognized in the genus Zythia and are provisionally retained as a single species, Zythia resinae. The results of morphological and molecular analyses indicate that there are likely many species, but due to inability to examine the type specimen of Lecidea resinae and the issues caused by mixed collections, we refrain from naming any new species.
This previously unrecognized diversity was suggested by available data on the Sareomycetes. Both Sarea and Zythia are widely distributed in Europe, North America, Asia, and Africa, with Z. resinae also present in Australasia (Hawksworth and Sherwood 1981 . In addition to the broad geographic range, Sareomycetes species are found on the resin of a wide variety of host species. Sarea species are found on the resin of seven genera in Pinaceae and Z. resinae is found on twelve or thirteen genera in Cupressaceae and Pinaceae (see Table S6). This broad host range is again not necessarily indicative of cryptic diversity (

Biogeography and Host Speci city
Little to no phylogeographic pattern in the studied Sareomycetes species is recovered in our analyses. This may be due to the fact that conifers in Pinaceae and Cupressaceae have been widely introduced around the world for ornamental and commercial purposes (Farjon 2017). We hypothesize that a number of Sareomycetes strains have been distributed worldwide, travelling on the resin of hosts, or as endophytes. The most obvious example is the introduction of S. coeloplata 1 to Antarctica reported in a study of the wood decay fungi on huts dating from the early 20th century (Held et al. 2003;Arenz et al. 2006). This fungus presumably was inhabiting the pinaceous timber brought to build the Discovery Hut on Ross Island (77° S), during the Discovery Expedition (1901)(1902)(1903)(1904). Our haplotype network suggests that the origin of the fungal strain was in Northern or Central Europe, where the countries supplying materials for these expeditions are located. The persistence of this species over the course of a century is perhaps an indication of how easy it would be to accidentally introduce these fungi to a new area. Another clear and relatively recent introduction is that of both Zythia resinae and S. coeloplata 1 to Cape Verde (reported in this study). Since no conifers are native to Cape Verde, we can again be sure that this is a case of human introduction (Hansen and Sunding 1993;Arechavaleta Hernández et al. 2005;Farjon 2017); Pinus spp. and Cupressus spp. have been widely introduced to Cape Verde (Frahm et al. 1996). At least two haplotypes of Zythia resinae and Sarea coeloplata 1 from Macaronesia (Cape Verde and the Canary Islands) are identical, or closely related, to haplotypes from the Iberian Peninsula. This makes sense since these archipelagos have close historical relationships with Spain and Portugal. The reports of Zythia resinae from New Zealand almost certainly represent a third instance of anthropogenic introduction. Pinaceae and Cupressaceae are, the only families known to host fungi in Sareomycetes; of these families, only two species in Cupressaceae are native to New Zealand (De Lange and Rolfe 2010), but all reports of Zythia are from Abies, Pinus, and Pseudotsuga, in Pinaceae (Gadgil and Dick 1999;Beimforde et al. 2020). A nal apparent indicator of ease of transmission through wood projects are a series of seven nuITS sequences uploaded to Genbank and misidenti ed as "Hormococcus conorum" and "Zythia pinastri" (NCBI, NLM, Bethesda (MD) 2020a, b, c, d, e, f, g). Since these are part of a project titled "Imported wood products to United States as vectors for potential invasive fungal species," it may be surmised that these were generated from imported wood products. On the other hand, the almost complete lack of genetic structure in the geographic distributions of species and the extensive geographic distribution in the Northern Hemisphere of some genetic lineages may be also due to long-distance dispersal of minute spores by wind, or even migratory birds, which use coniferous trees as perches in their migration routes (Hallenberg and Kúffer 2001;Muñoz et al. 2004;Wilkinson et al. 2012;Viana et al. 2016). Based on age estimates for the divergence among closely related haplotypes in all Sareomycetes species, intercontinental dispersal of lineages could have occurred during the Quaternary (< 2.59 Ma), and this could have been concomitant with events of population expansion, as suggested by neutrality test results in the nuITS and nuLSU markers. Larger datasets assembled with a populationgenetics scope are needed to evaluate these hypotheses. Nevertheless, there are exceptions to this general pattern, since seven clades in total contain only specimens from a relatively restricted, and sometimes sympatric, ranges: one from the eastern US (Zythia resinae clade 13 in Fig. S1), one from New England (Zythia resinae clade 5 in Fig. S1), one from the Paci c Slope (Atrozythia klamathica), and three from Japan (Zythia resinae clades 2, 4, & 7 in Fig. S1). Without broader sampling, particularly in Asia and Africa, and considering all available environmental sequences, it is di cult to determine if these are truly lineages of limited range, or a sampling artifact.
Likewise, there is little overall pattern of host speci city, except perhaps at the host family level. This might be expected, since resin composition is broadly similar within each conifer family (Langenheim 2003;Lambert et al. 2005) but still varies among species (Lambert et al. 2007) and even varies within a single species (Tappert et al. 2011). If there is a pattern of speci city even at family level, it appears not to hold for all species. For example, Sarea coeloplata 1 was found growing on Thuja occidentalis (ACD0147.1) in addition to a number of species in Pinaceae. Similarly, Zythia resinae clade 8 (in Fig. S1) encompasses primarily specimens on Pinaceae, but also a specimen found growing on Cupressus forbesii (JM0077), and the two known specimens of Atrozythia klamathica are from hosts in different families. Although this could be explained by a complete lack of host speci city, an alternative explanation is that different strains/species in Sareomycetes in some way selectively grow on resin containing or lacking certain components. Production of speci c resin components need not mirror evolutionary relationships (Tappert et al. 2011), so what currently appears random may still contain a hidden pattern. Nonetheless, there are clades suggestive of host speci city at the host generic or speci c level, even if most clades are found on mixed hosts. Zythia resinae clade 4 (in Fig. S1) contains only samples found associated with Chamaecyparis obtusa. Zythia resinae clade 5 (in Fig. S1) and an unnumbered clade appearing only in our three-gene and mtSSU analyses appear to be found only on Chamaecyparis spp. and Cupressus spp., respectively. Perhaps signi cantly, each of these clades also shows a fairly restricted geographic pattern, noted above, and each of these clades are among the least well-sampled, supported groups in our phylogeny. Wider, more robust sampling could change the pattern seen. Ultimately, a more detailed understanding of the speci c ecology of species in Sareomycetes is needed to generate and test hypotheses regarding host speci city.
Our dating analyses provide additional insight into host speci city in Sareomycetes at the temporal scale (Fig. 6). The results of our dating analyses match well with estimates of the diversi cation of the tree host genera of these fungi. Our estimate of 120. This suggests that the Sareomycetes evolved to exploit the new niche of resin provided by Pinus or another, now extinct taxon (Smith et al. 2017;Leslie et al. 2018). The origins of the genera Atrozythia, Sarea, and Zythia and subsequent diversi cation in Sarea (speci c estimates given in Table S5) also correspond well with a later period of diversi cation of host genera in Cupressaceae and Pinaceae of these fungi in the Cenozoic Era (Leslie et al. 2018). This occurred during and following a period of global cooling (Scotese 2016) together with some of the last important geological events, including Cenozoic orogenies, which in uenced the worldwide distribution of conifers. Close evolutionary histories among fungi and their hosts are well known in several parasitic and ectomycorrhizal fungal clades (Takamatsu 2013

Mixed Collections
An unexpected complicating problem was uncovered during these investigations. Prior to this study, authors have noted that both Sarea and Zythia species can be found growing on the same piece of resin (Hawksworth and Sherwood 1981;Spier and Aptroot 2000;Yatsyna 2017). This was noted in our study of specimens: Atrozythia klamathica was found growing alongside Zythia resinae (JM0068), and Zythia resinae was found growing with Sarea difformis (e.g. PV-D863), Sarea coeloplata 1 (e.g. ACD0147), and Sarea coeloplata 2 (e.g. IGB316). Less obviously, it was discovered that multiple clades of Zythia resinae or species of Sarea can be found mixed in a single collection. This was rst seen when sequencing multiple loci for specimen BHI-F779. An initial DNA extraction, PCR, and sequencing yielded sequences matching Zythia resinae clade 1; a subsequent round of sequencing from the same collection yielded sequences matching Zythia resinae clade 13. Later, Sarea difformis was detected living alongside Sarea coeloplata 1 (JM0072) and Sarea coeloplata 2 (JM0011). This ability to share substrate with closely related species, while ecologically interesting, poses serious challenges to the identi cation of morphological synapomorphies and matching them with the corresponding phylogenetic clade. Given the frequency with which we have found mixed collections, it cannot be excluded that some of the specimens we sequenced and examined morphologically contain mixes of Sarea coeloplata 1 and Sarea coeloplata 2, or mixes of multiple Zythia resinae clades. This could account for the lack of consistent morphology observed during our investigations of these species and informs our decision to not name these clades.
Based on our experience, future investigation of this family should be conducted by extracting DNA, examining micromorphology, and performing culture work from single apothecia. While this can be a challenge, given that apothecia are typically < 1 mm in diameter, we feel that this is the only reliable way of accurately characterizing this group of fungi.

Color Changes in Sections
We observed that microtome cut sections of Zythia resinae stored out of light in dried gum arabic solution on glass slides for a period of several months showed a marked degradation of pigment. Only the high concentration of pigments in the ectal excipulum and in the epithecium remained evident. A similar pattern was observed in sections permanently mounted in glycerin. In addition to color loss, the encrusting layer over the ectal excipulum and the epithecium was found to dissolve, further altering morphological characters of the fungus.
Such changes posed a challenge to morphological examination, since they create arti cial morphological patterns that differ from those seen in recent or fresh material, or even in fungarium material. For these reasons, to accurately assess pigment-related and other morphological characters, we recommend that any morphological examination of Zythia species be done on newly sectioned material rather than material sectioned by previous investigators and stored on glass slides or mounted.

Ascus Dehiscence
Previous authors have reported the asci of Sareomycetes as "not functionally bitunicate" (Hawksworth and Sherwood 1981;Nash et al. 2008) or of the broadly de ned "archaeascé" type (Letrouit-Galinou 1973). Our observations indicate that all three genera, and Atrozythia in particular, have ascus dehiscence characterized by a rupture of the outer layer at the tip of the ascus and protrusion of an inner wall. The inner wall extends some distance beyond the outer wall, varying among species. This agrees with the electron microscopic examination of a Sarea species performed by Bellemère (1994). It is not clear in our observations whether there is any zone of full wall separation between the inner and outer layers; we thus view this as the "rostrate" type of ascus dehiscence (Eriksson 1981;Bellemère 1994a).
The controversy regarding the ecology of species in Sarea and Zythia is long-standing; they have often been thought of as lichens. This is re ected in the taxonomy of the synonymous names. This idea goes back to Fries' original publication, in which he placed Zythia resinae in the lichen genus Lecidea and included the phrase "crusta tenuissima membranacea contigua cinerascenti" apparently describing a lichen thallus (Fries 1815). Hawksworth and Sherwood (1981) Hepp (1857b) included an unnumbered, mixed specimen of Zythia resinae and a Sarea sp. in his exsiccata, Die Flechten Europas. His opinion of whether it was a lichen or fungus, however, is obscured by the fact that the specimen was provided as an example of something easily confused with the black-apothecial lichen he included as number 332 ("Calicium inquinans γ. sessile"). Other authors referred to species in Sareomycetes as intermediate between lichens and fungi, sometimes placing them in named groups (e.g. "Lichenes ambigua," "Lichenes parasitici," "Pseudolichenes," "Hybridolichenes," and "Fungilli lichenoides") (Anzi 1860; Fries 1860; Koerber 1865; Ohlert 1870; Lettau 1912). One of the more unusual cases is that of Carlo Cappelletti (1924), who stated that S. difformis could be found both lichenized and non-lichenized in different samples. This situation is known in some fungi (Wedin et al. 2004(Wedin et al. , 2006, but the fact that Cappelletti reported this relationship in several resinicolous fungi casts doubt on his observations. Additionally, the concepts of some authors accepting species in Sareomycetes as lichens have been based on incorrectly identi ed material; these cases are treated in the "Excluded Names and Misapplications" section of the Taxonomy section above. Other mycologists and lichenologists, including the majority of modern authors, treat species in Sareomycetes as non-lichenized ( Are Sareomycetes Parasitic? The occurrence of these fungi on resinous wounds, has inevitably raised the question of whether they are parasitic (Kujala 1950;Groves and Wells 1956;Malençon 1979;Hawksworth 1980;Suto and Kanamori 1990). This question has been investigated by attempting to satisfy Koch's postulates, with varying results. The rst of these was conducted by Ayers (1941), who used one of his cultures of Z. resinae to attempt to infect Pinus strobus; he saw no effect. Researchers in the northwestern USSR used inoculation studies to investigate a disease of pines. They called the fungus they identi ed as the causal agent "Biatoridina pinastri," which they proved was the anamorphic state of a Sarea species (Shchedrova 1964(Shchedrova , 1965. In a broad study of conifer associated discomycetes, Smerlis (1973) concluded that Z. resinae was mildly pathogenic, producing cankers on every pinaceous host tested. An inoculation study conducted in the 1980s to determine the cause of a disease of Pinus koraiensis in northeastern China also found no evidence of infection by Z. resinae and identi ed the true causal agent, Tympanis confusa (Sūn et al. 1983;Cuī et al. 1984;Xiang et al. 1985; Xiang and Song 1988;Kobayashi et al. 1990). A similar study in Japan on a disease of Pinus thunbergii gave the same results; inoculations with Z. resinae produced no symptoms, but inoculations with a species of Ascocalyx did (Kobayashi and Kusunoki 1985;Kobayashi and Zhao 1989). Additional studies to determine the causal agent of the resinous stem canker of Chamaecyparis obtusa determined that Z. resinae did not cause symptoms on hosts in Pinaceae or Cupressaceae, and identi ed the causal agent as Cistella japonica (Hayashi and Kobayashi 1985;Yokozawa et al. 1986Yokozawa et al. , 1989Suto 1987Suto , 1992Suto , 1997Suto , 1998Kobayashi et al. 1990). The varying results and generality of these tests leave unresolved the question of pathogenicity of species in Sareomycetes; some authors assume pathogenicity and others accept a saprobic lifestyle, as summarized by Beimforde et al. (2020).

Sareomycetes as Endosymbionts of Photosynthetic Organisms
Other aspects of the ecology of species in Sareomycetes have been established with more certainty. These fungi have frequently been isolated as endophytes of conifers in Pinaceae (Petrini and Fisher 1988 (Peršoh 2013), and possibly deciduous woody plants (Novas and Carmarán 2008). Apart from the Pinus-dwelling mistletoe, which presumably allows the fungus close access to the resin seeping from any wounds created by the mistletoe, the occurrence of these fungi in these various hosts is di cult to explain. A closer look at the Cupressaceous endophytisms reveals a similarly di cult-to-explain pattern: a Sarea species was isolated (Petrini and Carroll 1981), but the current work represents the rst report of a Sarea species fruiting on a cupressaceous host. Several studies have found Sarea and Zythia species living within thalli of foliose and fruticose lichens in Europe and Asia (

The Ecology of Atrozythia Species
This uncertainty extends to our new genus, Atrozythia. Some cellulolytic capacity has been reported for Atrozythia lignicola (Sigler and Carmichael 1983), and the fungus has been recovered from both diseased (Sigler and Carmichael 1983)  . Additional study is needed to determine if A. lignicola is resinicolous, since all other members of Sareomycetes seem to be, or if it has some other lifestyle. Atrozythia klamathica, known thus far from only two specimens, was found fruiting directly on the resin of Chamaecyparis lawsoniana and Tsuga heterophylla; it presumably shares a similar ecology with other members of Sareomycetes.

Taxonomic Placement
The placement of species in Sareomycetes in the fungal tree of life has had a long and confused history, which we attempt to elucidate here with more details than Beimforde et al. (2020). In the late nineteenth century authors grouped species generally among the eshy discomycetes (Crouan and Crouan 1867; Cooke 1871; Saccardo 1889) and more speci cally with the Dermateaceae (Karsten 1885; Saccardo 1889) or Patellariaceae (Fuckel 1871), or among the lichens in Lecideaceae, allied with Biatora (Tuckerman 1872; Stein 1879). A number of mycologists between 1889 and 1934 (starting with Rehm) placed species among the Patellariaceae (see Table S7). Researchers later placed species variously in Lecanorales and Helotiales, or declined to place them; for instance, the rst (and several subsequent) edition(s) of the Dictionary of the Fungi list Retinocyclus as belonging with the lichens or in Helotiales, and Sarea as being of uncertain placement Bisby 1943, 1950). Placement was stabilized in 1981, when Hawksworth and Sherwood, based on morphological similarities with Agyrium rufum, placed Sarea and Zythia in Agyriaceae (Lecanorales) (Hawksworth and Sherwood 1981). Molecular evidence indicating that A. rufum was unrelated to the remainder of Agyriaceae (Lumbsch et al. 2007; Lumbsch and Huhndorf 2010) resulted in the move of Sarea and Zythia to Trapeliaceae (Hodkinson and Lendemer 2011). Not all authors followed these placements. In the course of an electron microscopical study of asci, Bellemère stated that the placement of both Sarea difformis and Zythia resinae based on ascus ultrastructure was uncertain, and noted that the two species differed in their method of ascus dehiscence (Bellemère 1994b). This study must be considered with some caution, since the substrate of the Z. resinae specimen used was said to be stone, indicating that the specimen was likely misidenti ed. Schultheis, Tholl, Baral and Marson placed Sarea difformis under the heading "Ascomycetes Incertae Sedis" (Schultheis et al. 2001). The application of molecular techniques was needed to properly place these taxa.
The history of the multiple publications attempting to elucidate the taxonomic position of these fungi using molecular data is outlined by Beimforde, et al. (2020). Reliance on these publications is likely the reason for uncertain placements or placements in Leotiomycetes by several subsequent authors (Lumbsch et  Recently, use of information from six genes and sampling taxa throughout Pezizomycotina resulted in the erection of a new class, Sareomycetes, to accomodate Sarea and Zythia (Beimforde et al. 2020). This placement explains over two centuries of confusion and uncertainty.

Conclusion
Our studies of species in Sareomycetes have revealed the existence of three genera, one described as new. Sarea is restricted to the group of species traditionally identi ed as Sarea difformis, but shown to be at least three phylospecies, Sarea difformis s. str., with a purple hymenial pigment, and two cryptic species lacking such a pigment and identi able morphologically with the type of Biatorella coeloplata, combined here as Sarea coeloplata. Zythia is resurrected for Zythia resinae (= Sarea resinae), which is retained provisionally as a single, highly diverse species. Atrozythia and the new species Atrozythia klamathica are described, and a combination is made for Arthrographis lignicola. The family name Zythiaceae is resurrected as an earlier name for Sareaceae. This family displays few biogeographic patterns and little evidence of host speci city. It is shown to have arisen in the late Jurassic or Cretaceous; subsequent diversi cation occurred roughly concurrently with the diversi cation of Cupressaceae and Pinaceae. Further work on this family is recommended, including: type studies on Lecidea resinae and Tympanis abietis, use of precise methodologies to study the two phylospecies assignable to S. coeloplata and to split the Zythia resinae complex, and collection of the data required to do population genetic analyses at least for Zythia.

Declarations
Ethics approval and consent to participate Not applicable.
Adherence to national and international regulations Not applicable.

Consent for publication
LoBuglio, Jason Karakehian, and Danny Haelewaters are thanked for their advice and instruction in molecular methods. Kanchi Gandhi is also thanked for his advice on nomenclatural issues.
The collections staff of FH are also thanked for their aid in searching the collections, permission for sampling, and processing loans from other institutions; the collections staff of these institutions are also thanked for sending material and allowing sampling, where appropriate. Additionally, all collectors are thanked for their contributions of specimens. The Nature Conservancies of Massachusetts, Rhode Island, and Vermont, Clearwater Creek State Park, the Arnold Arboretum, all National Forests in California, Redwoods National Park, and Jedediah Smith Redwoods State Park are thanked for issuing collecting permits. The Mycological Society of America is thanked for arranging permitting for the forays at their annual meetings. The Friends of the Farlow and The New England Botanical Club are thanked for their nancial support, and the Boston Harbor Islands National Recreation Area is thanked for both permission to collect and funding for sequencing some specimens associated with the All-Taxa Biodiversity Inventory being conducted there. Without them, the geographic sampling we obtained would have been impossible.