bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1

2 Character evolution of modern fly-speck fungi and implications for interpreting

3 thyriothecial fossils

4

5 Ludovic Le Renard*1, André L. Firmino2, Olinto L. Pereira3, Ruth A. Stockey4,

6 Mary. L. Berbee1

7

8 1Department of Botany, University of British Columbia, Vancouver BC, V6T 1Z4, Canada

9 2Instituto de Ciências Agrárias, Universidade Federal de Uberlândia, Monte Carmelo, Minas

10 Gerais, 38500-000, Brazil

11 3Departamento de Fitopatologia, Universidade Federal de Viçosa, Viçosa, Minas Gerais, 36570-

12 000, Brazil

13 4Department of Botany and Plant Pathology, Oregon State University, Corvallis, Oregon 97331,

14 USA

15

16 Email: [email protected]

17

18 Manuscript received ______; revision accepted ______.

19

20 Running head: Fly-speck fungi character evolution

21

22

23

p. 1 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

24

25 Abstract

26 PREMISE OF THE STUDY: Fossils show that fly-speck fungi have been reproducing with

27 small, black thyriothecia on leaf surfaces for ~250 million years. We analyze morphological

28 characters of extant thyriothecial fungi to develop a phylogenetic framework for interpreting

29 fossil taxa.

30 METHODS: We placed 59 extant fly-speck fungi in a phylogeny of 320 Ascomycota using

31 nuclear ribosomal large and small subunit sequences, including newly determined sequences

32 from nine taxa. We reconstructed ancestral character states using BayesTraits and maximum

33 likelihood after coding 11 morphological characters based on original observations and literature.

34 We analyzed the relationships of three previously published Mesozoic fossils using parsimony

35 and our morphological character matrix, constrained by the molecular phylogeny.

36 KEY RESULTS: Thyriothecia evolved convergently in multiple lineages of superficial, leaf-

37 inhabiting ascomycetes. The radiate and ostiolate scutellum organization is restricted to

38 Dothideomycetes. Scutellum initiation by intercalary septation of a single hypha characterizes

39 Asterinales and Asterotexiales, and initiation by coordinated growth of two or more adjacent

40 hyphae characterizes Aulographaceae (order incertae sedis). Scutella in Microthyriales are

41 initiated apically on a lateral hyphal branch. Patterns of hyphal branching in scutella contribute

42 to distinguishing among orders. Parsimony resolves three fossil taxa as Dothideomycetes; one is

43 further resolved as a member of a Microthyriales-Zeloasperisporiales clade within

44 Dothideomycetes.

45 CONCLUSIONS: This is the most comprehensive systematic study of thyriothecial fungi and

46 their relatives to date. Parsimony analysis of the matrix of character states of modern taxa

p. 2 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

47 provides an objective basis for interpreting fossils, leading to insights into morphological

48 evolution and geological ages of Dothideomycetes clades.

49 Key words: ancestral state reconstruction; Ascomycota; Dothideomycetes; fly-speck fungi; fossil

50 fungi, leaf cuticle; morphological characters; phylogeny; scutellum; thyriothecium.

p. 3 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

51 Fly-speck fungi are ascomycetes that appear as minute (50 μm to 2 mm in diameter) black

52 reproductive structures called thyriothecia. Thyriothecia are distinguished by having scutella,

53 which are small, thin, darkly pigmented, shield-like plates that cover spore-producing cells. The

54 margins of scutella are appressed to the surfaces of living or dead leaves or stems of vascular

55 plants (Arnaud, 1917, 1918, 1925; Stevens and Ryan, 1939; Wu et al., 2011b; Wu et al., 2014).

56 Fly-speck fungi are common throughout the world (Arnaud, 1918; Doidge, 1942; Batista, 1959;

57 Batista et al., 1969; Holm and Holm, 1991), especially in tropical rainforests, where they grow

58 on leaf surfaces (Batista, 1959; Hughes, 1976; Hofmann and Piepenbring, 2011; Firmino, 2016).

59 Scutella of fly-speck fungi are also common as fossils (Elsik, 1978; Kalgutkar and Jansonius,

60 2000). Our interest in these fungi is rooted in a desire to interpret the phylogenetic significance

61 of the hundreds of fossil forms found in palynological preparations and on preserved cuticles of

62 plant fossils (Kalgutkar and Jansonius, 2000; Taylor et al., 2015). Fossil fly-speck scutella have

63 long been classified into form genera and form species that carry minimal phylogenetic

64 information. Phylogenetic classification of the fossils can potentially be improved by analysis of

65 the systematic distribution of scutellum characters among extant fly-speck fungal lineages.

66 Scutella of thyriothecia are recognizable in part because they develop above the plant

67 cuticle, in contrast to many other kinds of minute ascomycete fruiting bodies that begin within

68 host tissue and break through to the leaf surface only when spores are ready for dispersal. How

69 fly-speck fungi interact with their host plants is unclear, but some appear to be asymptomatic

70 biotrophic parasites (on living hosts) while others are found in leaf litter (Hofmann, 2009; Wu et

71 al., 2011b). Hyphae of some fly-speck fungi enter the host directly through a stoma (Hofmann,

72 2009; Firmino, 2016). Many other leaf inhabiting fungi, including fly-speck fungi, enter the host

73 using appressoria. Appressoria are specialized hyphal cells that accumulate turgor pressure and

p. 4 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

74 direct growth into the host through a penetration peg, a small melanized ring (Parbery and

75 Emmett, 1977). Appressoria then link the superficial fungal mycelium and various hyphal

76 structures inside host tissue such as haustoria, hypostromata made of densely packed or stromatic

77 hyphae, (Hofmann, 2009; Firmino, 2016), or palmettes of flattened, intercellular, dichotomous

78 hyphae (Ducomet, 1907; Arnaud, 1918).

79 Several aspects of fly-speck fungus biology make it challenging to relate their

80 morphology to phylogeny. First, thyriothecia are small, and so it can be difficult to find enough

81 individuals of the same species for morphological or molecular systematic analysis. Second, fly-

82 speck fungi appear in mixtures with other epiphyllous fungi, which can complicate

83 distinguishing among species (Dilcher, 1965; Phipps and Rember, 2004). Hofmann (2009) and

84 Firmino (2016) noted mixed species of Asterina growing together, and Ellis (1976) described

85 closely related, co-occurring species in the collection of Microthyrium microscopicum that is the

86 type for its genus. For sequence analysis, species should ideally be cultured. While some fly-

87 speck fungi grow in pure culture (Firmino, 2016), other species do not (Hofmann, 2009); overall,

88 they are poorly represented in culture collections.

89 A long-held view that thyriothecial fungi were monophyletic (Arnaud, 1918; Doidge,

90 1919; Clements and Shear, 1931; Gäumann, 1952; Kirk et al., 2008) has been corrected by a

91 series of molecular phylogenetic analyses. Wu et al. (2011b) found a broad sampling of

92 thyriothecial fungi to be scattered across four orders of Dothideomycetes, although with low

93 support from bootstrap values or posterior probabilities. Mapook et al. (2016) and Hernández-

94 Restrepo et al. (2019) showed that species of thyriothecial and thyriothecium-like fungi in

95 Muyocopronales form a strongly supported sister group to Dyfrolomycetales, still in

96 Dothideomycetes but distant from any other thyriothecial fungi. Even more distant is a

p. 5 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

97 monophyletic group of two thyriothecial Micropeltis species nested in class Lecanoromycetes

98 rather than Dothideomycetes (Hongsanan and Hyde, 2017).

99 In spite of this recent progress, the challenges involved in collecting sufficient cells from

100 fly-speck fungi remain, and their sequences are still poorly represented in GenBank and in

101 phylogenetic studies. About 150 sequences were available, representing 18 genera, as of January

102 2019, contrasting with the more than 50 genera documented in recent publications (Hofmann,

103 2009; Wu et al., 2011a; Wu et al., 2011b; Hongsanan et al., 2014; Wu et al., 2014; Mapook et al.,

104 2016), and close to ~150 generic names available in families of thyriothecial fungi

105 (Wijayawardene et al., 2014).

106 To interpret the fossil record of fly-speck fungi, our goals include increasing the sampling

107 of extant representatives and related taxa in the class Dothideomycetes. Our project provides new

108 cultures, herbarium specimens, and sequences into the public domain. Using new and previously

109 published data, our first goal is to infer a molecular phylogeny that samples as widely as possible

110 from the lineages of fly-speck fungi and their possible close relatives. Secondly, we score

111 morphological characters for selected species in each lineage, and reconstruct ancestral

112 morphological character states throughout the phylogeny. Thirdly, we use the molecular tree,

113 comparative morphology, and inferred ancestral character states to discuss the phylogenetic

114 affinities of three fly-speck fossils.

115

116 MATERIALS AND METHODS

117 Taxon sampling—We assembled a dataset of 320 ascomycete taxa including species of

118 59 fly-speck fungi. We obtained sequence data from new collections, from pure cultures where

119 possible, found in Canada, Costa Rica (Permit nºR002-2014-OT-CONAGEBIO) and France.

p. 6 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

120 Pure cultures were established by crushing individual thyriothecia under an inverted compound

121 microscope using insect pins (Austerlitz Insect pins, size 0; Slavkov u Brna, Czech Republic) and

122 then transferring individual spores or groups of spores onto water agar with 1 mM ampicillin and

123 1 mM kanamycin for germination. Germinating spores were transferred to antibiotic-free

124 nutritive media under a dissecting microscope. The resulting cultures have been deposited in the

125 Westerdijk Institute, Netherlands (Appendix S1, strains labeled with an asterisk).

126 Sequence data for nuclear ribosomal large subunit (LSU) and small subunit (SSU) DNA

127 from thyriothecial taxa and their closest relatives were recovered from GenBank using a series of

128 BLAST searches (Appendix S1). The use of these two loci was pragmatic; for most of the

129 relatives of thyriothecial fungi including other leaf-inhabiting fungi and lichenized fungi, only

130 ribosomal DNA data were available. We also included published sequences from lichenized

131 fungi or lichen associated fungi that shared the epiphyllous habitat with thyriothecial fungi. We

132 selected outgroups to represent classes related to Dothideomycetes. In addition, near full-length

133 to full-length LSU and SSU data were retrieved from 13 whole genome projects with the help of

134 BLAST searches and using the JGI MycoCosm Portal (Grigoriev et al., 2013) (Appendix S2).

135 DNA extraction, amplification, sequencing—Strains successfully isolated in pure

136 culture include Lembosina sp. CBS 144007, Lembosina aulographoides CBS 143809,

137 Microthyrium macrosporum CBS 143810, and cf. Stomiopeltis sp. CBS 143811. They were

138 harvested from culture media for genomic DNA extraction using the Qiagen DNeasy Plant

139 Minikit (QIAGEN Inc., Mississauga, Ontario). A pure culture corresponding to the voucher of

140 cf. Stomiopeltis sp. UBC-F33041 was isolated and its DNA extracted, but it could not be

141 maintained over repeated transfers to new cultures. The polymerase chain reaction (PCR)

142 reactions were performed using PureTaq Ready-To-Go PCR beads (Amersham Biosciences,

p. 7 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

143 Piscataway, New Jersey), following manufacturer instructions. Amplifications were performed

144 using fungal specific rDNA primers reported in James et al. (2006): BMB-BR, NS1, NS3, NS6,

145 NS8, NS20, NS23, NS51, ITS2, LIC1460R, LIC2197, LR0R, LR3R, LR5, LR5R, LR7R, LR8,

146 LR10, LR11, LR10R, LR12, LR13. Uncultured dried specimens Micropeltis sp. UBC-F33034,

147 Scolecopeltidium spp. UBC-F33033 and UBC-F33035, and Asterotexiaceae sp. UBC-F33036

148 were used as templates for direct DNA amplification (Lee and Taylor, 1990) of ~1 kb of the 5'

149 end of the LSU, a region known to be informative for thyriothecial fungi (Hofmann et al., 2010;

150 Wu et al., 2011b; Guatimosim et al., 2015). A cocktail of 12.5 µL of water and primers at a

151 concentration of 1.0 µM each was pipetted into a PCR reaction tube containing a PureTaq

152 Ready-To-Go PCR bead, and kept on ice. Thyriothecia were: (i) mounted in a drop of sterile

153 water, observed under the compound microscope to ascertain identity, presence of ascospores,

154 and absence of contaminant material; (ii) crushed firmly between a slide and a coverslip, and (iii)

155 ~10-13 µL of the crushed ascospore material in water was pipetted from the coverslip and the

156 slide into a waiting PCR reaction tube, which was then vortexed briefly. The final volume of

157 liquid in each tube was ~25 µL, and the final concentration of each primer was 0.5 µM. The PCR

158 reactions ran for 40 cycles with an annealing temperature of 50ºC and an elongation temperature

159 of 72ºC. Direct amplifications of the 5' end of the LSU using primers LR0R-LR8 with an

160 annealing temperature of 55ºC initially yielded weak PCR product. Weak PCR products were

161 then diluted 100 to 1000 times and re-amplified with internal primers to increase product

162 concentration for sequencing (see Mitchel, 2015, for similar methods).

163 Phylogenetic inference—The LSU and SSU sequences were first aligned using the

164 iterative refinement method G-INS-I in MAFFT v7 (Katoh and Standley, 2013). Introns were

165 trimmed out manually and the resulting sequences re-aligned with MAFFT. Ambiguously

p. 8 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

166 aligned positions were excluded manually. The resulting concatenated dataset spanned 4552 bp.

167 We ran JModelTest 2 (Darriba et al., 2012) on the CIPRES portal (Miller et al., 2010) and

168 selected the GTR+G model for individual genes. We ran RAxML v.8 (Stamatakis, 2014) on two

169 partitions (one for each gene) and 1000 independent maximum likelihood searches, again

170 running on CIPRES. Branch support was assessed using 1000 bootstrap replicates (BS)

171 (Felsenstein, 1985). Bayesian analysis was performed using MrBayes 3.2 Metropolis Coupled

172 Monte Carlo algorithm (Ronquist et al., 2012). For the Bayesian analysis, we ran four

173 independent searches, each with eight MCMC chains for 160 million generations sampling every

174 5000 trees, using 128 processors of 2 Gb each on a Compute Canada cluster

175 (cedar.computecanada.ca). Following test runs, the heating parameter was adjusted to 0.03,

176 optimizing the search space so that swap frequencies were between 0.3 and 0.7, as the MrBayes

177 3.2 manual advises. We discarded 50% of the samples as burn-in. We used Tracer 1.5 to

178 determine that effective sample size exceeded 200 for all parameters, and the R package RWTY

179 (Warren et al., 2017) to verify the convergence of independent runs. We considered clades with

180 >70% likelihood bootstrap support and >0.95 Bayesian posterior probability to be moderately

181 well supported.

182 Preliminary results suggested that some taxa at various ranks were not monophyletic. To

183 explore the fit of the data to alternative hypotheses, we set monophyly constraints as single

184 individual branches and found the most likely tree given each constraint using 100 independent

185 “thorough” searches in RAxML. We then generated per-site log likelihoods of each likelihood

186 tree given the constraint, and performed an approximately unbiased (AU) test (Shimodaira,

187 2002) using CONSEL (Shimodaira and Hasegawa, 2001). We rejected hypotheses receiving a

188 probability below 0.05 in the AU test (Appendix S3). We specifically tested in turn the

p. 9 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

189 monophyly of (i) a putative clade comprising Asterinales and Asterotexiales, (ii) a putative clade

190 comprising Microthyriales and Zeloasperisporiales, (iii) Radiate thyriothecia, (iv) Asterina and

191 (v) Lembosia.

192 Observation and scoring of morphological characters—Original observations of

193 specimens were particularly important for coding thyriothecial characters that also appear in

194 fossils. Table 1 lists the 26 specimens examined directly for morphological analysis. Scoring for

195 the remaining 294 taxa was based on our interpretations of published illustrations and the

196 literature (Appendix S4). When possible, we used morphological and sequence information from

197 the same specimen (Appendix S4).

198 Table 2 shows our working definitions for scoring 11 morphological or habitat

199 characters, considering sporocarp features of both asexual, conidium-producing structures and

200 sexual ascomata. For direct observation of the hyphal structure of the dorsal surface of scutella,

201 and to look for appressoria, we peeled thyriothecia together with their surface hyphae off of leaf

202 surfaces using nail polish, as in Hosagoudar and Kapoor (1985). We examined thyriothecial

203 features using a Leica DMRB (Leitz, Wetzlar, Germany) differential interference contrast (DIC)

204 microscope, and photographed them at 400x or 1000x with a Leica DFC420 digital color camera.

205 Many scutella were more cone-shaped than strictly flat, and investigating their hyphae required

206 views from multiple focal planes. We attempted to use confocal microscopy to visualize hyphal

207 organization, staining a specimen of Microthyrium with 0.1 mg/mL calcofluor white (Sigma, St.

208 Louis, Missouri). The calcofluor revealed internal asci and some septa, but hyphae of the

209 scutellum were obscured by autofluorescence and pigmentation, and we pursued this approach

210 no further. Subsequently, we assembled stacks of bright field or DIC images at different focal

p. 10 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

211 planes to capture details of scutella and mycelia, following Harper (2015, p. 64, §4.3.1) and

212 suggestions from Bercovici et al. (2009).

213 Published written descriptions provide some character state information but are rarely

214 detailed enough to allow coding of the more subtle characters. Of the 11 scored characters, 10–

215 100% could be scored per specimen, across the 320 specimens studied (Appendix S4).

216 Appressoria are an example of a character that is rarely reported or illustrated, visible in surface

217 mycelium of extant and fossil thyriothecial taxa, but rarely noted in publications on other fungi.

218 Six characters specific to sporocarps could be scored for 67–79% of the taxa included (Appendix

219 S4). Thyriothecial development could be coded in extant and fossil taxa because different stages

220 are exposed on the surfaces of leaves. Development in non-thyriothecial Dothideomycetes is

221 more difficult to observe, again leading to much missing data (Appendix S4).

222 Analysis of morphological characters—We performed maximum likelihood ancestral

223 character state reconstruction on the most likely tree in Mesquite for the scored characters.

224 Maximum likelihood reconstructions were computed using the Markov k-state 1-parameter

225 (MK1) model (Lewis, 2001). Likelihood ratio tests showed that the MK1 model, which assumes

226 equal rates of forward and reverse transitions between states for each character was more likely

227 than an asymmetrical, 2-parameter model for the five binary characters in our dataset. We

228 reconstructed character states on the single, most likely tree.

229 To take phylogenetic uncertainty into account, we also performed reconstruction using

230 BayesTraits V3 (Pagel and Meade, 2007). In BayesTraits, we used the Multistate ML search

231 algorithm, which estimates different rates for every state transition, with the option MLtries set

232 to 100, a setting which led to consistent results in preliminary runs. Each character was

233 reconstructed for 208 nodes out of the 319 nodes of our tree. The proportion of each state for

p. 11 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

234 each node was calculated from its mean value over the 5000 trees chosen randomly from the

235 among the 32,000 trees in the posterior distribution of trees in chain one of our MrBayes

236 analysis. We considered any character state reconstructions with > 0.7 of both posterior

237 probability (from BayesTraits) and proportional likelihood (maximum likelihood) to be

238 supported.

239 Analysis of fossils—We selected three fossils reported as thyriothecial from the literature

240 that were well enough illustrated to be coded for morphology. These include the oldest dispersed

241 thyriothecium-like structure (Mishra et al., 2018), a dispersed Lichenopeltella-like form (Monga

242 et al., 2015) for which cell patterns could be analyzed, and Asterina eocenica found on a leaf

243 associated with mycelium (Dilcher, 1965). As for extant taxa, we coded character states from the

244 illustrations (Appendix S4).

245 To infer the phylogenetic relationships of each of the three fossils, we used the most

246 likely, 320-taxon tree from the rDNA analysis as a topological constraint. We analyzed the

247 relationships of the fossils together and then one at a time, using the 11-character morphological

248 data set (Appendix S4). With branch-and-bound searches in PAUP 4.0a166 (Swofford 2003), we

249 found the most parsimonious positions of the fossils, given the rooted constraint tree. We did not

250 further pursue the simultaneous analysis of all three fossils because the result was a poorly

251 resolved, uninformative phylogeny (not shown). Instead, we focus on the results from adding one

252 fossil at a time. Alignments, phylograms, and cladograms are available from TreeBASE

253 (Submission ID: 25326).

254 RESULTS

255 Distribution of thyriothecium forming fungi—Thyriothecium forming fungi appear

256 polyphyletic, as expected (Fig. 1). All but Micropeltidaceae are included in class

p. 12 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

257 Dothideomycetes (Fig. 1). Within Dothideomycetes, thyriothecial fungi appear widely

258 distributed, arising from a poorly resolved backbone (Fig. 1; Appendices S5 and S6).

259 Asterotexiales form a well-supported clade rich in thyriothecial forms, including the type

260 species Asterotexis cucurbitacearum (Fig. 1; Appendices S5 and S6). Of the 35 Asterotexiales

261 taxa sampled, 20 have thyriothecial sporocarps and 19 have epiphyllous habitats (Appendix S4).

262 Clade A, one of three clades diverging near the base of the Asterotexiales, includes 19 taxa with

263 thyriothecia, and 18 of the species that are epiphyllous (Appendices S4 and S6). However, Clade

264 A also includes Buelliella poetschii, an apothecioid lichen parasite, and Mycosphaerella

265 pneumatophorae, which produces a perithecioid ascoma in (rather than superficially on) plant

266 tissue. Note that the "ioid" suffix refers to sporocarp shape, not to development of the ascomata

267 (see character state definitions in Table 2).

268 Clade B species were recently classified in Asterinales (see Ertz and Diederich, 2015) but

269 here, they group instead with Asterotexiales (Appendices S5 and S6). Clade B includes 14

270 species, most of which are apothecioid parasites of lichens, although two are corticolous, and one

271 species, Morenoina calamicola is thyriothecial and epiphyllous (Appendices S4, S5 and S6). The

272 third clade is represented by Geastrumia polystigmatis, an epiphyllous species known only by its

273 asexual state, and it has yet to be formally classified to order. Aulographaceae, represented by six

274 thyriothecial taxa, appears as the sister group of Asterotexiales, but with low branch support.

275 A clade of five orders including Microthyriales and Venturiales receives 88% bootstrap

276 support and 100% posterior probability (Fig. 1; Appendices S5 and S6). Of the orders,

277 Microthyriales and Zeloasperisporiales are thyriothecial and epiphyllous, but their close relatives

278 in Natipusillales are freshwater cleistothecioid fungi that fruit on decorticated wood (Fig. 1;

279 Appendices S4 and S6). Although Venturiales include a few thyriothecial species, most of its

p. 13 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

280 members are perithecioid leaf and stem parasites causing, for example, apple scab and black knot

281 of plum (Fig. 1; Appendices S4 and S6).

282 Most Capnodiales are perithecioid, but the order also includes thyriothecial species (Fig.

283 1; Appendix S4). The order Asterinales is nested in Capnodiales with 99% Bayesian posterior

284 probability but with negligible likelihood bootstrap support (Appendix S6). Asterinales is

285 represented here by the type species Asterina melastomatis, and by sequences from three

286 collections of Parmularia styracis, the type species of Parmulariaceae (Appendices S5 and S6).

287 Asterinales also includes anamorphic foliicolous fungi and Thyrinula eucalypti, a plant pathogen

288 that causes leaf spots around its thyriothecia. Two thyriothecial taxa in Capnodiales are not

289 closely related to one another or to Asterinales: Peltaster fructicola and Schizothyrium pomi. The

290 genus Stomiopeltis is polyphyletic and Stomiopeltis betulae is resolved in Microthyriales while

291 two species tentatively identified as Stomiopeltis form a clade with Tothia fuscella in Venturiales

292 (Fig. 1; Appendices S5 and S6).

293 Muyocopronales represents a strongly supported example of convergent origin of

294 thyriothecia (Fig. 1; Appendices S5 and S6). Muyocopronales thyriothecia have several cell

295 layers of lightly pigmented, pseudoparenchymatous cells below their outer, darkly pigmented

296 scutellum (see illustrations of Mapook et al., 2016b). The thickness of the scutellum makes

297 detailed analysis of the hyphal branching pattern challenging.

298 Micropeltidaceae provide another strongly supported example of convergent thyriothecial

299 morphology. The family is nested in Ostropomycetidae in Lecanoromycetes rather than

300 Dothideomycetes (Fig. 1; Appendices S5 and S6). The scutellum of Micropeltidaceae is

301 distinctive in that it is ostiolate and formed as a flat, compact reticulum-like network of

p. 14 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

302 overlapping hyphae. As illustrated by Hofmann and Piepenbring (2006), the hyphae do not

303 radiate from the center to the margins of the scutellum.

304 Character states of most recent common ancestors of thyriothecial clades—Bayesian vs

305 likelihood results—Different methods of character state reconstruction are generally congruent,

306 but BayesTraits more frequently gives unresolved ancestral states compared with likelihood

307 (Table 3; Appendices S6 and S7). BayesTraits reconstructs states over a posterior distribution of

308 trees, so topological conflict could have explained low support for states. Surprising, in our

309 analysis, unresolved ancestral states in BayesTraits are often associated with clades that have

310 strong topological support, indicating that conflicting branching order is not the only cause of

311 lack of resolution. Rather, the unresolved nodes are associated with missing morphological data

312 and high estimated rates of character state transitions. BayesTraits estimates the forward and

313 reverse rates separately for each character, and where data are missing, the information content is

314 insufficient to parameterize the separate rates. An example of this is evident in the reconstruction

315 of the scutellum branching pattern in the ancestor of Zeloasperisporiales and Natipusillales

316 (Table 3). For this character, BayesTraits estimated an average forward transition rate from

317 pseudoparenchymatous (0) to anisotomous branching (2) as 4.3 but it estimated the reverse rate

318 as 56. Near the reconstruction, the topology is well supported and most of the surrounding

319 branches receive ≥92% posterior probabilities but BayesTraits shows the reconstructed state as

320 completely equivocal (0.2 for each of five states) (Appendices S6 and S7). In contrast, the same

321 node is reconstructed as anisotomous with 0.85 proportional likelihood (Appendices S6 and S7)

322 because the MK1 model can take advantage of all available data to estimate the single transition

323 rate that it applies to each character. We consider results to be most reliable when supported by

324 both likelihood and Bayesian reconstructions. However, when the topology is well supported, we

p. 15 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

325 consider a reconstruction supported by likelihood to be more useful than an equivocal Bayesian

326 result.

327 Reconstruction of evolutionary origin of character states—Sporulation substrate,

328 lichenization, sporocarp type—The sporulation substrate of ancestral Dothideomycetes is

329 reconstructed with likelihood (Fig. 1; Appendix S8) as immersed in plant tissue, or with

330 BayesTraits as being saxicolous (Appendix S8). The most recent common ancestors of

331 thyriothecial lineages appear to have adapted to superficial growth on leaf or twig surfaces (Fig.

332 1; Appendix S8). In the case of Muyocropronales, adaptations to superficial growth (Appendix

333 S8) preceded origin of thyriothecia. The ancestral Dothideomycetes and the most recent common

334 ancestors of thyriothecial fungi are reconstructed as non-lichenized (Appendix S9), although

335 Micropeltidaceae is nested in the Ostropomycetidae among lichenized fungi. Reconstructions

336 suggest that Dothideomycetes initially produced apothecioid sporocarps (Appendix S10).

337 Thyriothecial Micropeltidaceae appear to have arisen from apothecioid ancestors while

338 thyriothecial Asterinales, Capnodiales and Muyocopronales appear to have originated from

339 perithecioid ancestors. In three clades, Asterotexiales Clade A, Aulographaceae, and the group of

340 Microthyriales plus Venturiales clade, ancestors are reconstructed as thyriothecial or unresolved

341 (Appendix S10).

342 Lower walls—Unlike most ascomata, thyriothecia usually lack a differentiated lower

343 wall, a layer of pigmented fungal tissue that would separate sporogenous tissue from the

344 substrate (Fig. 1; Appendix S11). Lower walls are hypothesized to be present in the ancestor of

345 Dothideomycetes based on 100% proportional likelihood and 58% support from BayesTraits

346 (Node 1 in Fig. 1; Appendix S11). Out of 59 thyriothecial species included here, only four have a

p. 16 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

347 pigmented, differentiated lower wall (Appendices S4 and S11) and ancestral species

348 reconstructed as thyriothecial are also reconstructed as lacking a differentiated lower wall.

349 Dehiscence—The mechanism of spore release from sporocarps varies among the closest

350 relatives of Dothideomycetes, and among thyriothecial taxa (Appendix S12). Likelihood

351 reconstructions suggest that ancestral Dothideomycetes opened with an ostiole (Fig. 1);

352 BayesTraits reconstructions favor opening by a regular slit and show ancestral forms with slits

353 subsequently giving rise to ostiolate clades that diversified further (Table 3; Appendices S6, S7

354 and S12). A transition to irregular slits is reconstructed somewhere before the most recent

355 common ancestor of Asterotexiales and Aulographaceae (Figs. 1 and 2; Appendix S12). A

356 convergent transition from ostioles to irregular slits appeared along a branch nested in

357 Capnodiales that leads to Asterinales. Developmentally, the irregular slits result from cracking

358 between the adjacent hyphae that make up the scutellum (Fig. 2L, 2M). The precise location of

359 the opening remains unclear until dehiscence. Slit-forming genera sometimes produce ostioles in

360 their asexual or spermatial states (Fig. 2O). In Aulographaceae, and possibly also in species of

361 Lembosia, the slit appears early in development as a line of lightly pigmented cells (Fig. 3), but

362 the crack follows longitudinal walls and propagates beyond the less pigmented area. Regular slits

363 that follow a pre-determined pattern without further propagation are uncommon among

364 Dothideomycetes. Regular slits appear to have originated twice among thyriothecial taxa, being

365 present in Parmularia in Asterinales and in Inocyclus angularis in Asterotexiales, but not in their

366 inferred common ancestor (Appendix S12). Regular slits are also found in taxa producing

367 hysterothecia and lirella (such as Farlowiella charmichaeliana, Hysterographium fraxinii,

368 Melaspileopsis cf. diplasiospora (Nyl.) Ertz & Diederich Stictographa lentiginosa (Lyell ex

p. 17 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

369 Leight.) Mudd, which at maturity may look more like apothecia than thyriothecia (Appendix

370 S12).

371 Ancestral ostioles appear to have been retained in Microthyriales and Muyocopronales

372 (Figs. 1 and 4; Appendix S12), but the mode of dehiscence and ancestral states are unknown in

373 Zeloasperisporiales. The common ancestor of Micropeltidaceae probably released spores through

374 an ostiolate thyriothecium which is hypothesized to be derived from the apothecium that is

375 shared by other members of Ostropomycetidae in Lecanoromycetes (Appendix S12).

376 Locule circumference shape—In most Ascomycota, spores are formed in a cavity called a

377 locule (Kirk et al., 2008). In Dothideomycetes, the circumference of the locule within the ascoma

378 or conidioma is commonly round but may be elongate. Elongate locules are scattered across the

379 phylogeny. Thyriothecial species with elongate locules are inferred to be closely related to

380 species with round locules (Appendix S13), contrary to predictions from the current

381 classification (Wu et al., 2011b). For example, ascoma shape, which was used to separate the

382 elongate Lembosia from circular Asterina, fails to predict relationships. Instead, species of both

383 genera appear in both Asterotexiales and in Asterinales (Fig. 1; Appendix S6). In tree topology

384 tests, constraining either Asterina or Lembosia to be monophyletic is rejected with strong support

385 (Appendix S3).

386 Radiate development—Scutella of thyriothecia are either made up of radiating hyphae

387 that can be traced most of the distance from their distal tips to their central point of origin (Figs.

388 1–5), or of hyphae that are not radiate and cannot be traced back to their origin (Figs. 1 and 5).

389 Radiating patterns of thyriothecia are restricted to Dothideomycetes (Fig. 1; Appendices S4 and

390 S14). They are reconstructed to have originated convergently from ancestors with non-radiating

391 patterns. Radiate scutella appear to have arisen independently in Asterinales and in

p. 18 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

392 Muyocopronales (Appendix S14). One or multiple additional transitions to radiate thyriothecia

393 may have occurred among Asterotexiales, Aulographaceae, and in the clade including

394 Microthyriales and Zeloasperisporiales (Table 3; Appendix S14). Non-radiating thyriothecial

395 patterns also appear to have originated convergently, being found in distantly related species in

396 Capnodiales in Dothideomycetes, and in Micropeltidaceae in Ostropomycetidae (Fig. 1;

397 Appendices S4 and S14).

398 Phylogenetic distribution of hyphal branching patterns among thyriothecial scutella—

399 Likelihood consistently reconstructs pseudoparenchymatous sporocarp walls of polyhedral cells

400 as the ancestral state at the base of Dothideomycetes and Lecanoromycetes, (Figs. 1 and 5, Table

401 3; Appendix S15). However, corresponding BayesTraits reconstructions are equivocal

402 (Appendices S6, S7 and S15). Among thyriothecial fungi, branching patterns of the hyphae

403 forming the scutellum wall are usually distinctive compared with the untraceable hyphae in the

404 more widespread pseudoparenchymatous ascomatal walls of most Dothideomycetes.

405 Scutella with isotomous overlapping branching are reconstructed as possibly ancestral at

406 the base of Clade A in Asterotexiales (proportional likelihood 68%, BayesTraits 52%) and for

407 Asterinales after the divergence of Parmularia styracis (proportional likelihood 100%,

408 BayesTraits 56%), with increasing support for reconstructions with this state near the tips of the

409 tree (Table 3; Appendices S6, S7 and S15). Isotomous branching patterns in thyriothecia expand

410 the scutellum circumference mainly by duplicating files of cells through apical, nearly equal,

411 isotomous dichotomies, resulting first in a “Y” shaped cell with a basal septum, and then in three

412 cells, after additional septa delimit the two arms of the “Y” as separate cells (Figs. 2A, 2B, 2H

413 and 5B). In some cases, one of the two branches overlaps the other, growing downward and

414 disappearing below its neighbors (Figs. 2D, 2I and 5C). Septa are sometimes aligned around part

p. 19 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

415 of the circumference of the scutellum, resulting in partial concentric rings of septa (cf. Fig. 3.8

416 "d" in Hofmann 2009). Pigmentation in older parts of the scutellum contrasts with more

417 translucent tips at the margin of scutellum hyphae (Figs. 2H, 2J and 2K).

418 Scutella with pseudomonopodial branching are restricted to Aulographaceae. The

419 common ancestor of Aulographaceae is reconstructed with pseudomonopodial branching with

420 99% proportional likelihood and 28% posterior probability (Table 3; Appendices S6, S7 and

421 S15). In this growth form, the tips of dominant hyphae grow at the circumference of the

422 expanding scutellum, sequentially producing short lateral branches (Figs. 3B inset, 3D and 5E)

423 that overlap with neighboring hyphae. Hyphae curve as they grow, and septa only form in the

424 oldest hyphal segments. This results in a scutellum of irregular, multi-lobed cells (Fig. 5E).

425 Distal tips at the appressed margin are rounded, and uniform in diameter (Figs. 3B, 3D and 3H).

426 Within Aulographaceae, the construction of scutella is similar in Lembosina (Fig. 3B) and

427 Aulographum (Fig. 3D; Appendix S4).

428 Scutellum hyphae in some thyriothecial species in Venturiales and Capnodiales in

429 Dothideomycetes, and in all Micropeltidaceae in Lecanoromycetes are “untraceable” (Figs.1, 3A,

430 and 3C). Scutella may show some evidence of radiate construction, but details of hyphal

431 branching are obscured by dark pigmentation, or curved or meandering scutellum hyphae that

432 overlap and obscure one another, resulting in a “textura intricata" or “textura epidermoidea”

433 (Kirk et al., 2008; Figs. 3A, 3C and 5F).

434 In Microthyriales, scutellum hyphae usually branch dichotomously without overlap.

435 While their branching is usually isotomous, anisotomous branching also occurs, producing

436 hyphal tips of unequal width (Figs. 4B inset, 4D, 5B and 5D). A septum sometimes forms at the

437 base of only one of the two new hyphal tips, resulting in a roughly 'L' shaped proximal cell (Figs.

p. 20 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

438 4B and 5D; cf. Fig. 15E Wu et al. 2011). The anisotomous expansion is reconstructed in the

439 common ancestor of Zeloasperisporiales with 81% proportional likelihood (Fig. 1) but only 21%

440 posterior probability from BayesTraits (Appendices S6 and S7). Septa are angular (Figs. 4B, 4G,

441 4H and 5D) and constrictions at septa are infrequent, resulting in trapezoidal to nearly

442 rectangular cells with sharp angles. Partially aligned septa in adjacent hyphae sometimes result in

443 a pattern of incomplete concentric rings. Early stages of thyriothecial development may be

444 isotomous in Microthyriaceae (Figs. 4C, 4I, 4J and 5D) making the distinction between

445 anisotomy and isotomy most visible in fully grown structures (Fig. 5D).

446 Scutellum margins—Thyriothecia are unusual in that the margins of each scutellum are

447 pressed to the surface of a host--a leaf or a lichen--and the characters of the hyphae at the

448 scutellum margin differ across clades. Other kinds of sporocarps lack a direct homolog to the

449 scutellum margin, because their outer walls curve up from the substrate. In Asterotexiales,

450 Asterinales, Aulographaceae, in species of Stomiopeltis, and in Muyocopronales, the scutellum

451 margins often look crenulate or finely scalloped, because the dome-shaped hyphal tips bulge out

452 around the circumference (Figs. 2H, 2J, 3C and 3D; Appendix S16). In some Stomiopeltis-like

453 species (Fig. 3A), tips of hyphae are free; we interpret this as a variant of a crenulate margin

454 (Table 2). In other species, for example in Prillieuxina baccharidincola (Fig. 2K), the scutellum

455 hyphae anastomose with other superficial hyphae on the leaf surface, so that the margin of the

456 scutellum is poorly defined and continuous with the surface mycelium. In Microthyriales and

457 Zeloasperisporiales, scutellum margins are often entire, the scutellum hyphae ending in truncate

458 tips aligned around the circumference (Figs. 4D, 4F and 4L). In Microthyrium, the hyphal tips of

459 scutella that have finished radial expansion will narrow and continue to grow, forming a short

p. 21 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

460 fringe at the scutellum periphery (Figs. 4B and 4L). Fringe hyphae appear to be more tightly

461 appressed to the host surface than central portions of scutella.

462 Lateral appressoria—Appressoria are widely distributed across thyriothecium-forming

463 Dothideomycetes and vary in their morphology and position. Conspicuous, pigmented, lateral

464 appressoria are only produced by Asterinales and Asterotexiales (Figs. 2P–2S; Appendices S4

465 and S17). The appressoria are usually unicellular (Figs. 2P and 2R) but in some species they may

466 be two-celled (Figs. 2Q and 2S) (Hofmann, 2009, Figs. 3.8; 3.14; 3.21; 3.24; 3.31; 3.56). Lateral

467 appressoria may be swollen, lobed or minimally differentiated, and they branch from surface

468 hyphae. We coded thyriothecial taxa of Asterinales and Asterotexiales that lack appressoria or

469 produce them below the scutellum rather than on surface hyphae as absent for this character

470 (Table 1).

471 Other kinds of appressoria appear in other thyriothecial clades. All Aulographaceae

472 examined have intercalary appressoria on swollen cells with a conspicuous melanized ring (Fig.

473 3G), and strains isolated in pure culture form an appressorium regularly in each segment of

474 somatic hyphae. Species of Stomiopeltis (Venturiales) form infrequent, intercalary appressoria

475 (Figs. 3E and 3F). The hyphae bearing the appressoria may be swollen (Figs. 3E and 3G).

476 Inconspicuous, hyaline, intercalary appressoria are produced by some Microthyriales (Figs. 4E

477 and 4K). BayesTraits reconstructs lateral appressoria as present in the most recent shared

478 ancestor of Asterotexiales, and in the most recent shared ancestor of Asterinales, while

479 likelihood reconstructs them as ancestral only for lineages that have diverged more recently from

480 one another (Table 3; Appendix S17).

481 Initiation of thyriothecia—Development of thyriothecia begins with distinctive, lineage-

482 specific patterns of coordinated hyphal growth and septation in the species where observation

p. 22 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

483 was possible. We observed different developmental stages of 13 taxa (Table 1) and analyzed

484 development for 14 additional taxa from published illustrations (Appendix S4). Otherwise, data

485 on development is scarce and ancestral state reconstructions are equivocal where data are

486 missing (Appendices S4, S6, S7 and S18).

487 In Asterinales and Asterotexiales a single “generator hypha” (Hofmann, 2009) initiates

488 the thyriothecia by forming multiple, closely spaced intercalary septa (Figs. 2A–2C, 2H–2J and

489 5C). Each of the delimited cells then gives rise to transverse, adjacent, dichotomizing hyphal tips

490 that together form the scutellum. The generator hypha remains above the dorsal surface of the

491 mature scutella and is visible even in mature thyriothecia (Figs. 2I–2J and 5C). Ascomata in one

492 species of Asterotexiales, Rhagadolobiopsis thelypteridis, are initiated directly over host

493 stomata, as shown by Guatimosim et al. (2014) and in Fig. 2G here; in Asterotexis

494 cucurbitacearum, they are initiated directly from ascospores (Figs. 2E and 2F) (shown

495 previously by Guerrero et al., 2011; Guatimosim et al., 2015).

496 Of all the Aulographaceae observed, three species of Lembosina and one species of

497 Aulographum initiate their ascomata with coordinated development from multiple generator

498 hyphae. Adjacent intercalary portions of the multiple generator hyphae develop abutting,

499 transverse branches that contribute to forming a single scutellum (Figs. 3H and 5E). At a later

500 stage, radially arranged hyphae form pseudomonopodial branching units that cover the upper

501 surface of the hyphal aggregate (Fig. 3D).

502 In contrast to the widespread intercalary origin among Asterinales and Asterotexiales, all

503 Microthyriales examined initiate their ascomata at the tip of a generator hypha (terminal), located

504 on a short lateral branch of a surface hypha (Figs. 4C, 4I and 5D). The generator hypha does not

505 become highly septate, but instead its tip gives rise to a succession of dichotomously branched,

p. 23 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

506 closely appressed hyphal tips that develop into a scutellum. The scutellum in Microthyriales

507 overgrows the generator hypha; the generator hypha can no longer be seen in older scutella (Figs.

508 4B, 4G and 4H).

509 Case studies using thyriothecial fossils—Morphological characters of thyriothecial

510 scutella and mycelia offer a range of variation, allowing phylogenetically informed

511 interpretations of published fossils. We looked for the oldest geological occurrence of

512 thyriothecia, and also selected two other thyriothecial fossils that have a reasonably complete set

513 of morphological characters (Fig. 6; Appendix S4). Each phylogenetic analysis that included a

514 fossil yielded more than one equally parsimonious tree, given the morphological dataset and the

515 rDNA likelihood constraint tree. All trees from the morphological datasets were the same length,

516 232 steps, whether or not one of the fossils was included.

517 Triassic dispersed scutellum–The oldest evidence of a radiate fungal scutellum is a fossil

518 from the Early Triassic of India (Induan, ~251 Ma) referred to as a "fungal thallus" but not

519 identified to species or genus. The fossil was from a palynological sample that was rich in land

520 plant cuticles and spores (Mishra et al. 2018, Fig. 8j). It shares five characters with extant

521 thyriothecial taxa (Appendix S4). Like extant radial thyriothecia, it is made up of hypha-like

522 filaments that are closely appressed along their sides. The filaments are regularly septate into

523 trapezoidal cells that form partial, concentric rings. Some of the filaments dichotomize. A central

524 circular opening appears to be an ostiole surrounded by a ring of very dark cells. Where the

525 scutellum margin is in focus, it appears to be crenulate (Fig. 6A; Appendix S4).

526 We found 28 equally parsimonious positions for this fossil relative to the constraint tree.

527 Based on its five coded characters, the fossil could cluster with thyriothecial taxa in Asterinales,

528 Asterotexiales, or Microthyriales. Missing character state data from some extant taxa increased

p. 24 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

529 the number of possible, equally parsimonious positions for the fossil to include positions among

530 the asexual taxa of Zeloasperisporiales. A strict consensus of the 28 trees resulted in a polytomy

531 of 45 lineages, all stemming from the most recent ancestor of all Dothideomycetes (Fig. 1;

532 Appendix S19).

533 Microthyriaceous fossil–Trichothyrites setifer (Cookson) Saxena & Misra, a fossil from

534 the early Eocene (Ypresian, 56–47.8 Ma) sediments of India is illustrated by Monga et al. (2015)

535 (Fig. 6B). While this specimen was from dispersed material, rather than in situ on a leaf surface,

536 it shares seven characters with extant Lichenopeltella pinophylla and six characters with other

537 thyriothecial Microthyriales (Appendix S4). It has a circular, multicellular thallus. The radiate

538 scutellum has an entire margin and shows the presence of a lower wall (Fig. 6B, arrowhead). The

539 scutellum branching pattern is mostly isotomous, but also shows anisotomous dichotomies. Its

540 central ostiole is surrounded by papillate cells. Papillate ostioles are uncommon but occur in

541 Chaetothyriothecium elegans and L. pinophylla. However, C. elegans lacks a differentiated

542 lower wall.

543 We found 16 equally parsimonious trees resolving T. setifer in our phylogeny of living

544 fungi. The strict consensus position of this fossil suggests it shares affinities with the clade

545 containing Microthyriales and Zeloasperisporiales (Fig. 1; Appendix S20). Among the 16

546 reconstructions, T. setifer is alternatively placed in Microthyriales; as sister to Microthyriales

547 (10/16); as sister to asexual taxa in Zeloasperisporiales (that are missing data for characters of

548 the sporocarp) (3/16); as the sister to the perithecioid Natipusillales (2/16); or as the sister to all

549 these lineages (Fig. 1; Appendix S20).

550 Asterina fossil–The early Eocene (56-48 Ma) Asterina eocenica was found on leaves of

551 Chrysobalanus L. (Fagaceae Dumort.) and consists of circular, multicellular thalli and mycelia

p. 25 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

552 that are preserved in different states of development (Appendix S4), allowing comparison with

553 extant Asterinales and Asterotexiales (Dilcher, 1965). The fossil thyriothecium lacks a lower

554 wall, and dehiscence was by means of irregular slits. The pattern of branching of the scutellum is

555 isotomous, with overlapping hyphae, and the margin is crenulate, as in living Asterina. The

556 sporocarp was initiated with an intercalary generator hypha (Fig. 6C). The mycelium bears

557 lateral appressoria.

558 A strict consensus of the 23 equally parsimonious trees places A. eocenica in a polytomy

559 at the base of Dothideomycetes (Appendix S21). Asterina eocenica shares all 11 morphological

560 characters with three extant taxa in Asterinales and with nine in Asterotexiales (Fig. 6C;

561 Appendix S4). Consistent with the shared characters, 18 of the individual phylogenies show A.

562 eocenica within or as sister to Asterotexiales (Appendix S22), and five show it in, or as sister to

563 Asterinales (Appendix S23).

564

565 DISCUSSION

566 Results from our molecular phylogeny of 320 species including 59 species of

567 thyriothecial taxa combined with critical analysis of 11 morphological characters of extant

568 Dothideomycetes offer grounds for cautious optimism for incorporating thyriothecial fossil data

569 into the broader picture of fungal evolution. Ancestral state reconstructions show that convergent

570 evolution of thyriothecia in Dothideomycetes likely began among fungi that produced

571 pseudoparenchymatous-walled sporocarps on surfaces of plant cuticles. Our reconstructions

572 suggest that radiate scutella arose at least three times within epiphyllous Dothideomycetes. As

573 radiate thyriothecial forms evolved, patterns of hyphal branching and septation diversified,

574 giving rise to distinctive scutella. As the following fossil case studies show, a phylogenetic

p. 26 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

575 approach takes into account morphological variation within orders and convergence among

576 orders. We apply the approach to fossils but the same methods could be used to evaluate possible

577 relationships among extant taxa that are difficult to sample using molecular techniques.

578 Case studies: interpreting characters of fossils—A dispersed scutellum as evidence of

579 Triassic Dothideomycetes—Among our sampling of extant taxa, radiate scutella are restricted to

580 Dothideomycetes. Further, ancestral state reconstructions show radiate thyriothecia as a derived

581 character state that arose convergently, but only among epiphyllous Dothideomycetes. Consistent

582 with results from the parsimony analysis, the combination of an ostiole and a scutellum with

583 dichotomous branching that is found in the fossil is ancestral for Asterinales, Microthyriales, and

584 possibly for Asterotexiales. The fossil could represent a member of any of these groups.

585 Although it cannot be classified more precisely, parsimony analysis shows that the scutellum-

586 like "fungal thallus" of Mishra et al. provides evidence of Dothideomycetes. The fossil from the

587 Early Triassic (~251 Ma) (2018, Fig. 8j) likely represents the oldest evidence of thyriothecia.

588 Fossil evidence of Microthyriales—Monga (2015, Pl.2 Fig.18) illustrates the fossil

589 Trichothyrites setifer. Like the "fungal thallus" discussed above, T. setifer is from a sample rich

590 in land plant material, but its substrate is unknown. Trichothyrites specimens (synonym of

591 Notothyrites, see Kalgutkar and Jansonius, 2000), interpreted as members of Trichothyriaceae

592 have frequently been reported in the fossil record throughout the Tertiary (Kalgutkar and

593 Jansonius, 2000). Although classified in Microthyriaceae, extant Lichenopeltella shares multiple

594 characters with Trichothyriaceae Theiss., including a scutellum with isotomous and anisotomous

595 dichotomies; a scutellum margin that is entire, a papillate ostiole, and a differentiated lower wall

596 that can be seen through the translucent scutellum (Spooner and Kirk, 1990). Thyriothecia with a

597 papillate ostiole and a lower wall are characteristic of Trichothyriaceae. Anisotomous

p. 27 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

598 dichotomies were not present in Lichenopeltella pinophylla, the only extant taxon we sampled

599 with this morphology, but they do occur in the genus and are illustrated for Lichenopeltella

600 arctomiae Pérez-Ort. & T. Sprib. (Fig. 1 in Pérez-Ortega and Spribille, 2009).

601 Although T. setifer is more similar to Lichenopeltella than any other taxa in our dataset,

602 some of its shared characters including ostiolate thyriothecia are reconstructed as ancestral in the

603 larger clade encompassing Zeloasperisporiales, Natipusillales and Microthyriales. Three of six

604 taxa in Zeloasperisporiales are only known from asexual states and are missing data for the

605 sporocarp. A differentiated lower wall also occurs in non-thyriothecial ascomata of

606 Natipusillales. As a result of plesiomorphic characters and missing data, equally most

607 parsimonious placements of T. setifer were: in Microthyriales; as the sister lineage to

608 Microthyriales, Natipusillales or Zeloasperisporiales; and as sister to the clade including all of

609 these taxa.

610 At present, T. setifer (Monga et al., 2015), from the early Eocene (56–47.8 Ma),

611 represents the oldest convincing evidence of the clade including Microthyriales and Venturiales.

612 Investigating Paleocene deposits may very well yield further and older evidence of the

613 divergence of Trichothyriaceae. For more formal testing, the molecular phylogenetic positions of

614 extant taxa of Actinopeltis Höhn., Lichenopeltella, Trichothyrium Speg., and Trichothyrina

615 (Petr.) Petr. should be established using molecular data, and their morphology analyzed for

616 comparison with the fossils.

617 Asterina-like fossils could represent either Asterinales or Asterotexiales—Contrary to our

618 expectations, combining close zmorphological observations and the molecular phylogeny

619 revealed no clade-specific morphological differences between Asterinales and Asterotexiales.

620 Guatimosim et al. (2015) showed that Asterinales and Asterotexiales both include taxa with

p. 28 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

621 scutellum dehiscence by irregular cracks, and superficial hyphae bearing lateral appressoria. The

622 orders share crenulate scutellum margins. Lembosia species (characterized by elongate

623 thyriothecial locules) occur nested among both Asterinales and Asterotexiales species with round

624 locules.

625 Developmental and anatomical characters for Asterinales and Asterotexiales were

626 sometimes illustrated in great detail in early works (Arnaud, 1918; Doidge, 1919; Viégas, 1944).

627 More recently, Hofmann's (2009) thesis contains elegant drawings from diverse species,

628 documenting the common intercalary origin of thyriothecia, starting with septation in a generator

629 hypha. At the time of her thesis, the distinction between Asterinales and Asterotexiales had yet to

630 be discovered and Hofmann lacked sequence data to link her observations to clades. Our analysis

631 of species of Asterinales and Asterotexiales shows that development and scutellum branching as

632 Hofmann described are common to both orders.

633 Liu et al. (2017) interpreted Asterotexiales as "Asterinales sensu stricto" and suggested

634 that the Asterinales, represented by a clade of six species and their LSU rDNA sequences, all

635 from Brazil, should be considered "Asterinales sensu lato". However, the six Brazilian

636 Asterinales species that were sequenced by Guatimosim et al. (2015) include Asterina

637 melastomatis (from the epitype specimen of the Asterinales), which means that the ordinal name

638 applies to their clade. If a single sequence had been involved, or if the disputed Brazilian

639 Asterinales were from a clade of fungi commonly amplified from leaf surfaces in other studies,

640 contamination by non-target DNAs might seem a possible explanation for the presence of

641 morphologically similar species in phylogenetically divergent groups. However, the Brazilian

642 sequences give congruent results for species in five genera of thyriothecium forming taxa:

643 Batistinula, Prillieuxina, Parmularia, Asterina, and Lembosia. In a survey of Eucalyptus L'Hér.

p. 29 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

644 spp. Crous et al. (2019) found additional thyriothecial species of Thyrinula Petr. & Syd. to be

645 closely related to asexual species of Blastacervulus. Our analysis shows these two genera to be in

646 Asterinales, and future detailed analysis of their scutellum morphology will provide data useful

647 in reconstructing ancestral character states of Asterinales. Other thyriothecial species from leaves

648 of Eucalyptus spp. lack sequence data but are diverse in their scutellum morphology (Swart,

649 1986) and some of these may also represent Asterinales. In the absence of any contradictory

650 evidence, the sequences from Guatimosim et al. (2015) and from Crous et al. (2019) must be

651 assumed to come from their target fungi.

652 It is possible that more molecular data may resolve Asterinales and Asterotexiales as

653 sister groups, consistent with their morphological similarities. Tree topology tests did not rule

654 out a sister relationship between the two orders. Although the orders consistently appear

655 unrelated in published phylogenies, their separation never receives strong bootstrap support or

656 high posterior probabilities (Guatimosim et al., 2015; Firmino, 2016; Liu et al., 2017 and Fig. 1).

657 With cultures now available representing Asterotexiales (Asterotexiaceae sp.2 CBS 143813) and

658 Asterinales (Blastacervulus robbenensis CBS 12478) it should soon be possible to expand

659 available sequences beyond the ~1 kb of 28S ribosomal DNA per taxon that currently limits

660 phylogenetic resolution. If Asterinales and Asterotexiales are not sister groups, their thyriothecial

661 characters represent surprising convergence. If Asterinales and Asterotexiales are sister groups, a

662 phylogenetic analysis of their morphological characters, including dehiscence type, scutellum

663 branching and pattern of initiation of sporocarp, should likely allow unambiguous assignment of

664 fossils to their clade.

665 Asterina eocenica from the Middle Eocene (48.6-37.2 Ma) of western Tennessee

666 (Dilcher, 1965; Elsik and Dilcher, 1974) is the oldest fossil that combines a convincing

p. 30 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

667 Asterinales/Asterotexiales suite of characters. The exquisitely preserved scutella of A. eocenica

668 show clear isotomous dichotomies with overlap (Figs. 64 and 65 in Dilcher, 1965). This places

669 the fossil with clades of thyriothecia in which hyphae branch dichotomously at their apices to

670 produce radial scutella (Farr, 1969; Reynolds and Gilbert, 2005; Hofmann and Piepenbring,

671 2011; Guatimosim et al., 2015). Dichotomous branching is infrequent among fungi, where

672 branching is usually initiated some micrometers behind the hyphal apex. Asterina eocenica has

673 nine out of 11 characters shared by Asterinales and Asterotexiales. The development of

674 thyriothecia of A. eocenica is typical of Asterinales and Asterotexiales, with initiation by closely

675 spaced septa forming in a generator hypha. The generator hypha persists on the dorsal side of the

676 scutellum (Figs. 59, 60, 63, 64 and 65 in Dilcher, 1965). Like Asterinales and Asterotexiales, A.

677 eocenica has slit-like dehiscence (Figs. 63 and 65 in Dilcher 1965), an undifferentiated lower

678 wall (Figs. 61, 63, 65 and 66 in Dilcher, 1965), and some elongate, one-celled tapering lateral

679 hyphal projections described as appressoria (Figs. 57 and 59 in Dilcher, 1965). In their otherwise

680 detailed and useful review of fossils, Samarakoon et al. (2019) assign A. eocenica to crown

681 group Asterinales (=Asterotexiales) and with a minimum age as Paleocene, 54 Ma. The

682 Paleocene age appears to be an error because it is inconsistent with an analysis by Elsik and

683 Dilcher (1974), which assigned a more recent Middle Eocene age to the fossil's source material

684 from the Lawrence Clay Pit. None of the characters in A. eocenica justify positioning it in the

685 crown group of either order as opposed to a stem relationship of either order, and as the

686 parsimony analysis shows, it could equally well represent Asterinales and Asterotexiales. A

687 strong implication of the close similarity of Asterinales and Asterotexiales is that even a

688 beautiful, complete fossil like A. eocenica cannot be assigned to one order versus the other.

689 CONCLUSION

p. 31 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

690 Much cryptic diversity in Dothideomycetes consists of minute leaf dwelling fungi (Arx

691 and Müller, 1954; Müller and von Arx, 1962; von Arx and Müller, 1975). As our analysis shows,

692 epiphyllous fungi together with lichenized, lichenicolous and aquatic fungi are all relevant to the

693 evolution of thyriothecia, but they have been studied by different mycologists who sequenced

694 different loci for phylogenetic analysis (Dhanasekaran et al., 2006; Miadlikowska et al., 2006;

695 Etayo, 2010). The sequencing of the 5' end of the LSU was common across most studies, but

696 choices of other loci varied. We anticipate that sequencing of more Dothideomycetes genomes,

697 including some from the new isolates from this study from cultures available through the

698 Westerdijk Institute, will contribute to improved resolution of relationships among extant taxa.

699 Our molecular phylogeny of thyriothecial taxa and their closest relatives leads to testable

700 predictions about the sequence of early events in their evolutionary history. A remaining

701 limitation of this study is that we could include only a small fraction of living fly-speck species.

702 The fidelity of association of characters with extant lineages should be tested across a broader

703 range of taxa using morphological and molecular phylogenetic analysis. Undoubtedly, further

704 study will help reveal convergence and variation, and perhaps additional distinctive

705 synapomorphies. Our analyses suggest that analyzing and coding scutellum characters can link

706 fossils on leaf cuticle to phylogenetic lineages, offering rewarding insights into fungal evolution

707 through geological time. However limited fossilized characters may be, documenting equivalent

708 anatomical features in extant taxa helps communicate the significance of fossil morphologies,

709 increasing their value for paleomycology.

710 ACKNOWLEDGEMENTS

711 This research was enabled in part by support provided by WestGrid (www.westgrid.ca),

712 Compute Canada Calcul Canada (www.computecanada.ca) and CIPRES Science Gateway

p. 32 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

713 (www.phylo.org). The authors also wish to express their gratitude to UBC Bioimaging facilities

714 for use of their fluorescence microscope, UBC InterLibrary Loan service for providing crucial

715 literature, Robert Lücking for gift to UBC Herbarium of lichenized taxa in genera

716 Microtheliopsis and Asterothyrium, OTS and Conagebio, and more specifically Francisco

717 Campos Rivera and Bernal Mattarita Carranza for allowing collection of tropical thyriothecium

718 forming fungi in the research station of La Selva in Costa Rica. Funding for this research was

719 provided by a Natural Sciences and Engineering Research Council of Canada Discovery Grant

720 RGPIN-2016–03746 to MLB.

721

722

723

724

725

726

p. 33 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

727 FIGURE LEGENDS

728 Figure 1. Phylogeny and character state transitions leading to convergent evolution of fly-speck

729 fungal morphology. Numbers on boxes to the left match labeled transitions between

730 reconstructed ancestral states that received > 0.7 proportional likelihood in the maximum

731 likelihood phylogeny to the right. Thickened branches in the tree indicate bootstrap support >

732 70% and posterior probabilities > 0.95. Arrowheads indicate the most parsimonious consensus

733 positions (for A. eocenica, alternative consensus positions) of fossils. Green boxes surround

734 clades that include taxa with radiate thryiothecia. Solid black circles after taxon names indicate

735 thyriothecial species; question marks indicate possible thryiothecia still in need of anatomical

736 analysis; slashed circles indicate taxa known only from their asexual state; and tick marks

737 indicate thyriothecia with differentiated lower walls.

738 Figure 2. Morphological characters of Asterinales and Asterotexiales in face view. (A–J)

739 Development. Initiation of ascomata in Asterinales and Asterotexiales usually involves

740 intercalary septation of a superficial generator hypha but is sometimes (D) uninterpretable, (E, F)

741 from spores, or (G) from stomata. (H–J) The highly septate generator hypha usually persists

742 above the scutellum throughout development (white arrowhead). (A–D, H–J) The generator

743 hypha gives rise to the scutellum through (i) the formation of new, lateral hyphal tips that branch

744 with isotomous dichotomies, occasionally (D, I, inset) with one tip that overlaps the other (o).

745 (H, J) The appressed scutellum margin shows lightly pigmented tips and is usually crenulate at

746 maturity (black arrowhead). (K) Hyphae of the appressed marginal of scutella Prillieuxina

747 baccharidincola instead anastomose with surrounding mycelium (white arrowhead). (I, K, L–O)

748 Dehiscence. (I, K, L–M) Asterinales and Asterotexiales ascomata usually dehisce with irregular

749 slits. (K) Mature scutellum with irregular radial slits, black arrowheads. (L-M) Irregular slits

p. 34 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

750 follow the longitudinal walls of scutellum cells (white arrowheads). (N-O) Ostiolate openings are

751 less common. (N) Ostiole in young scutellum. (O) Asexual or spermatial stage, with evenly

752 pigmented cells lining ostiole. (P–S) Lateral appressoria showing characteristic melanized ring at

753 lower focal plane (insets, representing area in the dashed grey boxes). Scales (A–D, H–I) 10 µm;

754 (E–G) 20 µm; (J–K) 50 µm; (L–S) 5 µm. (A, I, J, P, Q) Asterina melastomatis VIC 52822, 1, 4,

755 6, 2 and 2 focal planes (f.p.) respectively; (B) Asterina chrysophylli VIC 42823, 4 f.p.; (C, O)

756 Asterotexiaceae sp. CBS 143813, 2 f.p. each; (D) Hemigrapha atlantica BR 14014, 1 f.p.; (E, F)

757 Asterotexis cucurbitacearum VIC 42814, 1 f.p. each; (G, L) Rhagadolobiopsis thelypteridis

758 EG156; (H, R) Batistinula gallesiae VIC 42514, 10 and 2 focal planes (f.p.) respectively; (K N),

759 Prillieuxina baccharidincola VIC 42817; (M) Asterotexiaceae sp. UBC F33036, 11 f.p.; (S)

760 Asterina sp. (VUL. 341b), 2 f.p.

761 Figure 3. Morphological characters of Aulographaceae and Stomiopeltis spp. s.l. (Venturiales) in

762 face view. (A–D) Scutella. (A) Mature Stomiopeltis-like scutellum. Hyphae are radiate but

763 untraceable due to frequent overlap of neighboring hyphae. The ostiole is wide (white

764 arrowhead) and at the distal appressed margin, hyphal tips are free (black arrowhead). (B)

765 Mature Aulographaceae scutellum with radiate, pseudomonopodial hyphal branching. Inset (m),

766 a close up of hyphal branching in the area in the dashed grey box. An irregular slit for dehiscence

767 would form in the elongate, lighter area (white arrowhead). The appressed distal margin has free

768 hyphal tips. (C) Early development of a Stomiopeltis-like scutellum. The appressed distal margin

769 is irregularly crenulate (arrowhead). (D) Young Aulographaceae scutellum showing multiple,

770 intercalary generator hyphae, pseudomonopodial hyphal branching, and an irregularly crenulate

771 appressed margin (arrowhead). (E–G) Intercalary appressoria with a melanized ring

772 (arrowheads). (H) Initiation of Aulographaceae ascomata; coordinated intercalary septation of

p. 35 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

773 multiple generator hyphae (arrowheads) gives rise to scutella. Scales, (A, B) 20 µm; (C, D) 10

774 µm; (E–H) 5 µm. (A, C, F) Stomiopeltis sp. UBC F33041, 5, 6 and 1 stacked focal planes (f.p.)

775 respectively. (E) Stomiopeltis sp. UBC F33040, 1 f.p. (B) Lembosina aulographoides CBS

776 143809, 4 f.p.; (D, G, H) Aulographum sp. CBS 143545, 1 f.p.

777 Figure 4. Morphological characters of Microthyriales in face view. (A) Calcofluor white staining

778 of Microthyrium sp. reveals radially arranged asci (arrowhead) in a hymenium below the

779 scutellum. The inset shows uneven fluorescence of scutellum hyphae due to dark pigments. (B,

780 G, H) Mature scutella are radiate and insets show details of anisotomous (a) and isotomous (i)

781 hyphal branching. Dehiscence is with an ostiole that is lined by small, darkly pigmented cells.

782 (C, D, I, J) Early development of scutella. The appressed margins of young scutella are entire. (I)

783 The scutellum is initiated at the tip of a lateral generator hypha. (C, D, J) New hyphal tips arise

784 from the generator hypha with closely spaced, isotomous, dichotomous branches. (D, J) The

785 young scutella overgrow and conceal the generator hyphae. (E, K) Intercalary appressoria in a

786 regular (E) or swollen (K) hyphal segment, showing melanized rings (arrowheads). (F, L)

787 Margins of mature scutella are considered to be entire if hyphal tips are blunt, at least in part

788 (arrowheads) even if narrow portions of the tip hyphae extend further into an irregular,

789 meandering fringe as in (B, L). Scales, (A–C) 20 µm; (D) 10 µm; (E–I) 5 µm. (A) Microthyrium

790 sp., 44 focal planes (f.p.); (B–F), Microthyrium macrosporum CBS 143810, with 4, 3, 1, 1 and 1

791 focal planes (f.p.), respectively; (G), Lichenopeltella pinophylla, CBS 143816, 14 f.p.; (H–L), M.

792 ilicinum CBS 143808, with 4, 1, 3 and 2 f.p. respectively.

793 Figure 5. Scutellum initiation and interpretation of branching. (A) Ancestral Dothideomycetes

794 had pseudoparenchymatous walls (a–b). (B) Radiate scutella in thyriothecia grow by successive

795 dichotomies followed by cross-wall formation. Isotomous dichotomies produce daughter hyphae

p. 36 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

796 of similar width. Hyphae occasionally overlap so that one file of cells seems to vanish (top);

797 anisotomous dichotomies produce daughter hyphae that differ in width and rarely overlap

798 (bottom). (C) Asterotexiales/Asterinales pattern. Scutellum hyphae show isotomous dichotomies,

799 and some hyphae overlap in some specimens (c, e) but not in all (d1-2) specimens. Overlap can be

800 missing in young specimens (e1). Initiation of the scutellum often begins with a single intercalary

801 generator hypha. The generator hypha persists above the scutellum at maturity (c, d1-2, e1-2). (D)

802 Microthyriales pattern. Young scutella are initiated at the tip of a short lateral generator hypha

803 (g1, h1). The generator hypha gives rise to scutellum hyphae that eventually overgrow it as they

804 branch with isotomous dichotomies (g1-2, h1-2). The generator hypha is not visible in the mature

805 scutellum (g2-3, h2-3). In mature scutella, hyphae show anisotomous and isotomous dichotomies

806 and hyphae do not overlap (g3, h3). (E) Aulographaceae pattern. Scutella with pseudomonopodial

807 branching (i2). The scutellum is usually initiated by coordinated growth of multiple adjacent,

808 intercalary generator hyphae (i1). (F) Stomiopeltis (UBC F33041) pattern. Scutellum hyphae are

809 generally radiate but they overlap and are irregular (j–l) and they cannot be traced to generator

810 hypha or from the center of the scutellum to the margin. Sources of drawings: (a) Fumiglobus

811 pieridicola UBC-F33205; (b) Pleospora herbarum UBC-F4461; (c) Batistinula gallesiae VIC

812 42514; (d) Asterotexiaceae sp. UBC F33036; (e) Asterina chrysophylli VIC 42823; (f)

813 Asterotexiaceae sp. CBS 143813; (g) Microthyrium ilicinum CBS 143808; (h) from M.

814 macrosporum CBS 143810, (i) Aulographum sp. CBS 143545, (j) Stomiopeltis sp. UBC F33041,

815 (k) Stomiopeltis sp. CBS 143811 and (l) Peltaster cerophilus redrawn from Fig. 2M in

816 (Medjedović et al., 2014).

817 Figure 6. Reproduction of fossil illustrations used in our case studies. (A) Triassic dispersed

818 “fungal thalli,” courtesy of Mishra et al. (2018), reproduced with the permission of Springer

p. 37 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

819 Nature; B.S.I.P. Slide No.15524, J36-4. (B) Early Eocene dispersed Trichothyrites setifer with

820 lower wall (arrowhead) seen through the translucent scutellum, reproduced with permission from

821 Monga et al. (2015); B.S.I.P. slide no. 15297, U41/3. (C) Middle Eocene Asterina eocenica on

822 cuticle of Chrysobalanus sp. described in Dilcher (1965), illustration courtesy Steven

823 Manchester, reproduced from Taylor et al. (2015), with permission from Elsevier; Florida

824 Museum of Natural history, l.f. 87. Scales, 20µm.

825 SUPPORTING INFORMATION

826 Additional Supporting Information may be found online in the supporting information section at

827 the end of the article

828 Appendix S1. Isolates included in this study; names in bold indicate newly contributed

829 collections and voucher/strain numbers beginning with 'CBS' and followed by '*' represent

830 cultures newly deposited in the Westerdijk Institute. When LSU and SSU were derived from two

831 different collections, bold indicates the voucher/strain name that is used in phylogenies.

832 Accessions labeled 'JGI' refer to sequences extracted from JGI genomes project, for which no

833 GenBank accession exists.

834 Appendix S2. JGI sequence data retrieved from MycoCosm portal.

835 Appendix S3. Results from CONSEL for six topologies tested against most likely tree.

836 Appendix S4. Matrix of characters used in ancestral character state reconstructions and in

837 parsimony analyses of fossils.

838 Appendix S5. Consensus tree from Mr. Bayes analysis with four runs of eight chains each,

839 running 160 million generations and sampling every 5000 trees, from which we discarded 50%

840 of the samples as burn-in. Posterior probabilities mapped on branches, clades are labeled and

841 ordered as in Fig. 1.

p. 38 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

842 Appendix S6. Maximum likelihood phylogeny labeled with codes for nodes reconstructed in

843 ancestral character state analyses. Taxa forming thyriothecia are indicated with pink diamonds.

844 Most likely tree obtained out of 4000 independent ML searches of a 4552 bp alignment of LSU

845 and SSU rDNA data for 320 taxa, with bootstrap support > 70% in red and a posterior

846 probability > 0.95.

847 Appendix S7. Detailed output of ancestral state reconstruction for each of 208 nodes as

848 proportional likelihoods (white rows) or posterior probabilities (grey rows). Node labels in

849 column 1 are designated in Fig. 1. Nodes labeled in Column 2 are designated in the phylogeny in

850 Appendix S6.

851 Appendix S8–S18. Ancestral character state reconstructions for 11 morphological characters,

852 given the most likely DNA sequence topology. Reconstructions from MK1 analysis are shown

853 above branches as proportional likelihoods in pie charts with red outlines. Reconstructions from

854 BayesTraits are shown below branches as posterior probabilities in pie charts with black outlines.

855 In each tree, names in grey indicate taxa that could not be coded for the character, with ancestral

856 states that could not be reconstructed. Bootstrap support >0.7 is reported in red above branches

857 and posterior probabilities >0.95 are in black below branches.

858 Appendix S8. Ancestral character state reconstructions for substrate (Sub).

859 Appendix S9. Ancestral character states reconstruction (Lic), as lichenized or non-lichenized.

860 Appendix S10. Ancestral character state reconstruction for sporocarp type (Spo).

861 Appendix S11. Ancestral character state reconstruction for differentiated or undifferentiated

862 lower wall of sporocarp (Low).

863 Appendix S12. Ancestral character state reconstruction for sporocarp dehiscence (Deh).

864 Appendix S13. Ancestral character state reconstruction for locule circumference (Loc).

p. 39 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

865 Appendix S14. Ancestral character state reconstruction for presence or absence of a radiate

866 sporocarp (Rad).

867 Appendix S15. Ancestral character state reconstruction for scutellum branching (Bra).

868 Appendix S16. Ancestral character state reconstruction for margin of sporocarp (Mar).

869 Appendix S17. Ancestral character state reconstruction for the presence or absence of lateral

870 appressoria on superficial mycelium (App).

871 Appendix S18. Ancestral character state reconstruction for Sporocarp initiation (Ini): terminal

872 on generator hypha, intercalary on generator hypha, from spore, from host stomata or from

873 multiple generator hyphae.

874 Appendix S19. Consensus of 28 equally parsimonious phylogenetic positions of the Triassic

875 “fungal thallus” from India described in Mishra et al. (2018).

876 Appendix S20. Consensus of 16 equally parsimonious phylogenetic positions of Trichothyrites

877 setifer from the Eocene of India described in Monga et al. (2015).

878 Appendix S21. Consensus of 23 equally parsimonious phylogenetic positions of Asterina

879 eocenica from the Eocene of USA, described in Dilcher (1965).

880 Appendix S22. Consensus of the 18 equally parsimonious phylogenetic trees placing Asterina

881 eocenica with taxa in Asterotexiales.

882 Appendix S23. Consensus of the 5 equally parsimonious phylogenetic trees placing Asterina

883 eocenica with taxa in Asterinales.

884

885

p. 40 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

886 LITERATURE CITED

887

888 Arnaud, G. 1917. Sur la famille des microthyriacées. Compte Rendu de l’Académie des Sciences

889 Paris 164: 574–577.

890 Arnaud, G. 1918. Les Asterinées. Annales de l'École Nationale d'Agriculture de Montpellier 16:

891 1–288.

892 Arnaud, G. 1925. Les Asterinées. IVe partie. Annales de l’École National d'Agriculture de

893 Montpellier, série 10 5: 643–722.

894 Arx, J. A. v., and E. Müller. 1954. Die Gattungen der amerosporen Pyrenomyceten. Beiträge zur

895 Kryptogamenflora der Schweiz 11: 1–434.

896 Batista, A. C. 1959. Monografia dos fungos Micropeltaceae. Publicações do Instituto de

897 Micologia da Universidade do Recife 56: 1–519.

898 Batista, A. C., J. L. Bezerra, T. T. Barros, and F. B. Leal. 1969. Sobre um novo gênero de

899 Microthyriaceae da Nova Caledônia. Instituto de Micologia, Universidade do Recife 637: 1–11.

900 Bercovici, A., A. Hadley, and U. Villanueva-Amadoz. 2009. Improving depth of field resolution

901 for palynological photomicrography. Palaeontologia Electronica 12: 1–12.

902 Clements, F. E., and C. L. Shear. 1931. The genera of fungi, second ed. New York.

903 Crous, P. W., M. J. Wingfield, R. Cheewangkoon, A. J. Carnegie, T. I. Burgess, B. A.

904 Summerell, J. Edwards, et al. 2019. Foliar pathogens of eucalypts. Studies in Mycology 94: 125–

905 298.

906 Darriba, D., G. L. Taboada, R. Doallo, and D. Posada. 2012. jModelTest 2: more models, new

907 heuristics and parallel computing. Nature Methods 9: 772–772.

p. 41 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

908 Dhanasekaran, V., R. Jeewon, and K. D. Hyde. 2006. Molecular taxonomy, origins and evolution

909 of freshwater ascomycetes. Fungal Diversity 23: 351–390.

910 Dilcher, D. L. 1965. Epiphyllous fungi from Eocene deposits in western Tennessee, U.S.A.

911 Palaeontographica Abteilung B 116: 1–54.

912 Doidge, E. M. 1919. South African Microthyriaceae. Transactions of the Royal Society of South

913 Africa 8: 235–282.

914 Doidge, E. M. 1942. A revision of the South African Microthyriaceae. Bothalia 4: 273–420.

915 Ducomet, V. 1907. Recherches sur le développement de quelques champignons parasites à thalle

916 subcuticulaire. Thèse doctorale, Faculté des sciences de Paris, Rennes, France.

917 Ellis, J. P. 1976. British Microthyrium species and similar fungi. Transactions of the British

918 Mycological Society 67: 381–394.

919 Elsik, W. C. 1978. Classification and geologic history of the microthyriaceous fungi. IVth

920 International Palynological Conference (1976–77), Lucknow, India.

921 Elsik, W. C., and D. L. Dilcher. 1974. Palynology and age of clays exposed in Lawrence clay

922 pits, Henry County, Tennessee. Palaeontographica Abteilung B: 65–87.

923 Ertz, D., and P. Diederich. 2015. Dismantling Melaspileaceae: a first phylogenetic study of

924 Buelliella, Hemigrapha, Karschia, Labrocarpon and Melaspilea. Fungal Diversity 71: 141–164.

925 Etayo, J. 2010. Lichenicolous fungi from the western Pyrenees. V. Three new ascomycetes.

926 Opuscula Philolichenum 8: 131–139.

927 Farr, M. L. 1969. Some "black mildew", "sooty mold", and "fly-speck" fungi and their

928 hyperparasites from Dominica. Canadian Journal of Botany 47: 369–381.

929 Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap.

930 Evolution 39: 783–791.

p. 42 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

931 Firmino, A. L. 2016. Taxonomy, phylogeny and biogeography of Asterinales on native forest

932 species from Atlantic Forest and Cerrado. Tese doutourado, Universidade Federal de Viçosa,

933 Viçosa, Minas Gerais, Brazil.

934 Gäumann, E. A. 1952. The fungi. A description of their morphological features and evolutionary

935 development. Hafner Publishing Company, New York.

936 Grigoriev, I. V., R. Nikitin, S. Haridas, A. Kuo, R. Ohm, R. Otillar, R. Riley, et al. 2013.

937 MycoCosm portal: gearing up for 1000 fungal genomes. Nucleic acids research 42: D699-D704.

938 Guatimosim, E., H. J. Pinto, R. W. Barreto, and J. Prado. 2014. Rhagadolobiopsis, a new genus

939 of Parmulariaceae from Brazil with a description of the ontogeny of its ascomata. Mycologia

940 106: 276–281.

941 Guatimosim, E., A. Firmino, J. Bezerra, O. Pereira, R. Barreto, and P. Crous. 2015. Towards a

942 phylogenetic reappraisal of Parmulariaceae and Asterinaceae (Dothideomycetes). Persoonia 35:

943 230–241.

944 Guerrero, Y., T. A. Hofmann, C. Williams, M. Thines, and M. Piepenbring. 2011. Asterotexis

945 cucurbitacearum, a poorly known pathogen of Cucurbitaceae new to Costa Rica, Grenada and

946 Panama. Mycology 2: 87–90.

947 Hernández-Restrepo, M., J. D. P. Bezerra, Y. P. Tan, N. P. Wiederhold, P. W. Crous, J. Guarro,

948 and J. Gené. 2019. Re-evaluation of Mycoleptodiscus species and morphologically similar fungi.

949 Persoonia 42: 205–227.

950 Hofmann, T., R. Kirschner, and M. Piepenbring. 2010. Phylogenetic relationships and new

951 records of Asterinaceae (Dothideomycetes) from Panama. Fungal Diversity 43: 39–53.

952 Hofmann, T. A. 2009. Plant parasitic Asterinaceae and Microthyriaceae from the Neotropics

953 (Panama). Ph.D., Johann Wolfgang Goethe-University, Frankfurt, Germany.

p. 43 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

954 Hofmann, T. A., and M. Piepenbring. 2006. New records and host plants of fly-speck fungi from

955 Panama. Fungal Diversity 22: 55–70.

956 Hofmann, T. A., and M. Piepenbring. 2011. Biodiversity of Asterina species on neotropical host

957 plants: New species and records from Panama. Mycologia 103: 1284–1301.

958 Holm, K., and L. Holm. 1991. Ascomycetes on Myrica gale in Sweden. Nordic Journal of

959 Botany 11: 675–687.

960 Hongsanan, S., and D. K. Hyde. 2017. Phylogenetic placement of Micropeltidaceae. Mycosphere

961 8: 1930–1942.

962 Hongsanan, S., Y.-M. Li, J.-K. Liu, T. Hofmann, M. Piepenbring, J. D. Bhat, S. Boonmee, et al.

963 2014. Revision of genera in Asterinales. Fungal Diversity 68: 1–68.

964 Hosagoudar, V., and J. Kapoor. 1985. New technique of mounting meliolaceous fungi. Indian

965 Phytopathology 38: 548–549.

966 Hughes, S. J. 1976. Sooty moulds. Mycologia 68: 693–820.

967 James, T. Y., F. Kauff, C. L. Schoch, P. B. Matheny, V. Hofstetter, C. J. Cox, G. Celio, et al.

968 2006. Reconstructing the early evolution of Fungi using a six-gene phylogeny. Nature 443: 818–

969 822.

970 Kalgutkar, R. M., and J. Jansonius. 2000. Synopsis of fossil fungal spores, mycelia and

971 fructifications. AASP Contributions Series, 351. American Association of Stratigraphic

972 Palynologists Foundation.

973 Katoh, K., and D. M. Standley. 2013. MAFFT Multiple sequence alignment software Version 7:

974 Improvements in performance and usability. Molecular Biology and Evolution 30: 772–780.

975 Kirk, P. M., P. F. Cannon, D. W. Minter, and J. A. Stalpers. 2008. Ainsworth & Bisby's

976 Dictionary of the Fungi, 10th ed. CABI Europe, UK.

p. 44 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

977 Kumar, M. 1996. Palynostratigraphy and paleoecology of Early Eocene palynoflora of Rajpardi

978 lignite, Bharuch District, Gujarat. Palaeobotanist 43: 110–121.

979 Lee, S. B., and J. W. Taylor. 1990. Isolation of DNA from fungal mycelia and single spores. In

980 M. A. Innis, D. H. Gelfand, J. J. Sninsky, and T. J. White [eds.], PCR Protocols: A Guide to

981 Methods and Applications, 282–287. Academic Press, San Diego.

982 Lewis, P. O. 2001. A likelihood approach to estimating phylogeny from discrete morphological

983 character data. Systematic Biology 50: 913–925.

984 Mapook, A., K. Hyde, S. Hongsanan, C. Phukhamsakda, J. Li, and S. Boonmee. 2016.

985 Palawaniaceae fam. nov., a new family (Dothideomycetes, Ascomycota) to accommodate

986 Palawania species and their evolutionary time estimates. Mycosphere 7: 1732–1745.

987 Medjedović, A., J. Frank, H.-J. Schroers, B. Oertel, and J. C. Batzer. 2014. Peltaster cerophilus

988 is a new species of the apple sooty blotch complex from Europe. Mycologia 106: 525–536.

989 Miadlikowska, J., F. Kauff, V. Hofstetter, E. Fraker, M. Grube, J. Hafellner, V. Reeb, et al. 2006.

990 New insights into classification and evolution of the Lecanoromycetes (Pezizomycotina,

991 Ascomycota) from phylogenetic analyses of three ribosomal RNA- and two protein-coding

992 genes. Mycologia 98: 1088–1103.

993 Miller, M. A., W. Pfeiffer, and T. Schwartz. 2010. Creating the CIPRES Science Gateway for

994 inference of large phylogenetic trees. Gateway Computing Environments Workshop (GCE), New

995 Orleans, LA: 1–8.

996 Mishra, S., N. Aggarwal, and N. Jha. 2018. Palaeoenvironmental change across the Permian-

997 Triassic boundary inferred from palynomorph assemblages (Godavari Graben, south India).

998 Palaeobiodiversity and Palaeoenvironments 98: 177–204.

p. 45 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

999 Mitchell, A. 2015. Collecting in collections: a PCR strategy and primer set for DNA barcoding

1000 of decades-old dried museum specimens. Molecular Ecology Resources 15: 1102–1111.

1001 Monga, P., M. Kumar, V. Prasad, and Y. Joshi. 2015. Palynostratigraphy, palynofacies and

1002 depositional environment of a lignite-bearing succession at Surkha Mine, Cambay Basin, north-

1003 western India. Acta Palaeobotanica 55: 183–207.

1004 Müller, E., and J. A. von Arx. 1962. Die Gattungen der didymosporen Pyrenomyceten. Beiträge

1005 zur Kryptogamenflora der Schweiz 11: 1–922.

1006 Pagel, M., and A. Meade. 2007. BayesTraits v.2 Computer program and documentation

1007 available at http://www.evolution.rdg.ac.uk/BayesTraits.html.

1008 Parbery, D., and R. Emmett. 1977. Hypotheses regarding appressoria, spores, survival and

1009 phylogeny in parasitic fungi. Revue de Mycologie 41: 429–447.

1010 Phipps, C. J., and W. C. Rember. 2004. Epiphyllous fungi from the Miocene of Clarkia, Idaho:

1011 reproductive structures. Review of Palaeobotany and Palynology 129: 67–79.

1012 Reynolds, D. R., and G. S. Gilbert. 2005. Epifoliar fungi from Queensland, Australia. Australian

1013 Systematic Botany 18: 265–289.

1014 Ronquist, F., M. Teslenko, P. Van Der Mark, D. L. Ayres, A. Darling, S. Höhna, B. Larget, et al.

1015 2012. MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large

1016 model space. Systematic Biology 61: 539–542.

1017 Shimodaira, H. 2002. An approximately unbiased test of phylogenetic tree selection. Systematic

1018 Biology 51: 492–508.

1019 Shimodaira, H., and M. Hasegawa. 2001. CONSEL: for assessing the confidence of phylogenetic

1020 tree selection. Bioinformatics 17: 1246–1247.

p. 46 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1021 Spooner, B. M., and P. M. Kirk. 1990. Observations on some genera of Trichothyriaceae.

1022 Mycological Research 94: 223–230.

1023 Stamatakis, A. 2014. RAxML version 8: a tool for phylogenetic analysis and post-analysis of

1024 large phylogenies. Bioinformatics 30: 1312–1313.

1025 Stevens, F. L., and M. H. Ryan. 1939. The Microthyriaceae. The University of Illinois press,

1026 Urbana.

1027 Swart, H. J. 1986. Australian leaf-inhabiting fungi XXII. Microthyrium-like fungi on Eucalyptus.

1028 Transactions of the British Mycological Society 87: 81–91.

1029 Taylor, T. N., M. Krings, and E. L. Taylor. 2015. Fossil Fungi, 1st ed. Elsevier/Academic Press

1030 Inc., San Diego.

1031 Viégas, A. P. 1944. Algun fungos do Brasil II Ascomycetos. Bragantia 4: 1–392.

1032 von Arx, J. A., and E. Müller. 1975. A re-evalutation of bitunicate ascomycetes with keys to

1033 families and genera. Studies in Mycology 9: 1–159.

1034 Warren, D. L., A. J. Geneva, and R. Lanfear. 2017. RWTY (R We There Yet): An R package for

1035 examining convergence of bayesian phylogenetic analyses. Molecular Biology and Evolution 34:

1036 1016–1020.

1037 Wijayawardene, N., P. Crous, P. Kirk, D. Hawksworth, S. Boonmee, U. Braun, D.-Q. Dai, et al.

1038 2014. Naming and outline of Dothideomycetes–2014 including proposals for the protection or

1039 suppression of generic names. Fungal Diversity 69: 1–55.

1040 Wu, H., K. D. Hyde, and H. Chen. 2011a. Studies on Microthyriaceae: placement of Actinomyxa,

1041 Asteritea, Cirsosina, Polystomellina and Stegothyrium. Cryptogamie, Mycologie 32: 3–12.

1042 Wu, H., Q. Tian, W.-J. Li, and K. Hyde. 2014. A reappraisal of Microthyriaceae. Phytotaxa 176:

1043 201–212.

p. 47 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1044 Wu, H.-X., C. L. Schoch, S. Boonmee, A. H. Bahkali, P. Chomnunti, and K. D. Hyde. 2011b. A

1045 reappraisal of Microthyriaceae. Fungal Diversity 51: 189–248.

1046

1047

p. 48 bioRxiv preprint

1048 Table 1. Matrix of specimens and their observed character states. Specimens examined Sub Lic Spo Low Deh Loc Rad Bra Mar App Ini (which wasnotcertifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. doi: Asterina melastomatis (Asterinales) VIC, https://doi.org/10.1101/2020.03.13.989582 sup no thyr und irr circ yes iso cren yes int 42822, BR, AL Firmino

Asterotexiaceae sp. (Asterotexiales) UBC, sup no thyr und irr circ yes iso cren no ? F33042, CA, L Le Renard

Asterotexiaceae sp. (Asterotexiales) UBC, sup no thyr und ost circ yes ? cren no int F33036, CR, L Le Renard ; Asterotexis cucurbitacearum (Asterotexiales) this versionpostedMarch14,2020. sup no thyr und irr circ yes ? cren no spo VIC, 42814, BR, OL Pereira & AL Firmino

Aulographum sp. (Aulographaceae) UBC, sup no thyr und irr elng yes ? cren no mult F33038, CA, L Le Renard

Batistinula gallesiae (Asterinales) VIC, 42514, sup no thyr und irr circ yes iso cren yes int The copyrightholderforthispreprint BR, AL Firmino, DB Pinho & OL Pereira

Buelliella poetschii (Asterotexiales) BR, sup no apo dif exp circ no pseud not no ? 5030020189138; 5030020188124, BE, D Ertz

p. 49 bioRxiv preprint

Fumiglobus pieridicola (Capnodiales) UBC,

sup no per dif ost circ no pseud not no ? (which wasnotcertifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. F23788, CA, T Bose doi:

Hemigrapha atlantica (Asterotexiales) BR https://doi.org/10.1101/2020.03.13.989582 lich no thyr und irr elng yes iso ent no ? 5030012458426, ES, D Ertz

Inocyclus angularis (Asterotexiales) VIC, sup no thyr dif reg elng yes ? cren no ? 39748, BR, RW Barreto

Lembosina aulographoides (Aulographaceae) yes sup no thyr und irr elng mono cren no mult UBC, F33037, CA, L Le Renard ; this versionpostedMarch14,2020. Lembosina sp. (Aulographaceae) UBC, sup no thyr und irr elng yes mono cren no mult F33044, CA, L Le Renard

Lembosina sp. (Aulographaceae) UBC, sup no thyr und irr elng yes mono cren no mult F33045, CA, L Le Renard

Lichenopeltella pinophylla (Microthyriales) The copyrightholderforthispreprint sup no thyr dif irr circ yes aniso ent no ter UBC, F33032, CA, L Le Renard

Micropeltis sp. (Micropeltidaceae) UBC, sup no thyr und ost circ no untra ana no ? F33034, CR, L Le Renard

p. 50 bioRxiv preprint

Microthyrium ilicinum (Microthyriales) UBC,

sup no thyr und ost circ yes aniso ent no ter (which wasnotcertifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. F33031, FR, L Le Renard doi:

Microthyrium macrosporum (Microthyriales) https://doi.org/10.1101/2020.03.13.989582 sup no thyr und ost circ yes aniso ent no ter UBC, F33039, CA, L Le Renard

Parmularia styracis (Asterinales) VIC, 42450, sup no thyr dif irr elng yes ? cren no ? BR, M Silva & OL Pereira

Prillieuxina baccharidincola (Asterinales) sup no thyr und irr circ yes ? anast no int VIC, 42818, BR, AL Firmino ; this versionpostedMarch14,2020. Rhagadolobiopsis thelypteridis

(Asterotexiales) VIC, 31939, BR, E sup no thyr und irr elng yes iso cren no sto

Guatimosim

Schizothyrium gaultheriae (S. pomi

(Capnodiales) was sequenced) UBC, F3143, sup no thyr und irr circ no untra anast no ? The copyrightholderforthispreprint

CA, M Barr

Scolecopeltidium sp. (Micropeltidaceae) UBC, sup no thyr und ost circ no untra anast no ? F33033, CR, L Le Renard

p. 51 bioRxiv preprint

Scolecopeltidium sp. (Micropeltidaceae) UBC,

sup no thyr und ost circ no untra anast no ? (which wasnotcertifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. F33035, CR, L Le Renard doi:

Seuratia millardetii (Lichenostigmatales) https://doi.org/10.1101/2020.03.13.989582 sup ? apo dif exp circ no pseud not no ? UBC, F33043, CA, L Le Renard

cf. Stomiopeltis sp. (Venturiales) UBC, sup no thyr und ost circ yes untra cren no ? F33040, CA, L Le Renard

cf. Stomiopeltis sp. (Venturiales) UBC, sup no thyr und ost circ yes untra cren no ? F33041, CA, L Le Renard ; this versionpostedMarch14,2020. 1049

1050

1051 Notes: Specimen data includes systematic affinities followed by herbarium code (BR = Meise Botanic Garden; UBC = University of

1052 British Columbia; VIC = Universidade Federal de Viçosa), specimen accession number, ISO country code and collector(s).

1053 Sub = sporulation substrate: sup = superficial on plants; lich = lichenicolous; The copyrightholderforthispreprint

1054 Lic = lichenized; no = non-lichenized

1055 Spo = sporocarp: apo, apothecioid; per, perithecioid; thyr, thyriothecioid

1056 Low = sporocarp lower wall: und, undifferentiated; dif, differentiated

1057 Deh = dehiscence: ost, ostiole; reg, regular slit; irr, irregular slit; exp, exposed at maturity

p. 52 bioRxiv preprint

1058 Loc = locule circumference: circ, circular; elng, ellipsoid to elongate (which wasnotcertifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. 1059 Rad = radiate sporocarp: no, absent radiate; yes, present doi:

1060 Bra = scutellum branching: pseud= pseudoparenchymatous; iso, isodichotomous; aniso, anisodichotomous; mono, pseudomonopodial https://doi.org/10.1101/2020.03.13.989582

1061 dichotomies; untra, untraceable

1062 Mar = margin: not, margin not appressed; cren, crenulate, aligned; ent, entire, aligned; anast, anastomosing with surrounding

1063 mycelium

1064 App = lateral appressoria; no, absent; yes, present

1065 Ini = initiation of sporocarp, position and number of generator hyphae: int, single, intercalary; ter, single, terminal; mult, multiple, ; this versionpostedMarch14,2020. 1066 intercalary; spo, from spores; sto, from stomata The copyrightholderforthispreprint

p. 53 bioRxiv preprint

1067 (which wasnotcertifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. doi: https://doi.org/10.1101/2020.03.13.989582 ; this versionpostedMarch14,2020. The copyrightholderforthispreprint

p. 54 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1068 Table 2. Definitions of characters and their states with reference to examples from illustrations

1069 in figures.

Character, character states, and Definition

examples in figures in this paper

Sub- Sporulation substrate

0–On rock or soil (= saxicolous) On rock or other inorganic substrates.

Rock inhabiting (Saxomyces sp.),

ground dwelling lichen (Simonyella

variegata)

1–Superficial on plants (= foliicolous) On leaves, stems, fruit; superficial, closely associated

On living leaves, Asterina with cuticle.

melastomatis, Fig. 2J, or on dead

leaves, Microthyrium macrosporum

Fig. 4B

2–On lichens only (= lichenicolous), On lichen thallus.

Hemigrapha atlantica

3–Immersed in plants (= lignicolous In bark or in woody parts of plants, in or under a plant's

or corticolous) epidermis.

Taxa bursting through woody parts

(p.Dothidea 55 sambuci), on decorticated bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

material (Natipusilla bellaspora), or

as part of corticolous lichens thalli

(Trypethelium elutheriae)

4–Other Various habitats that are not the focus of this study and

Coenococcum geophillum, that are poorly sampled here. Includes taxa that inhabit

Phaeotrichum atkisonii dung, are endoparasites, or are mycorrhizal.

Lich- Lichenized

0–Non-lichenized

Asterina melastomatis Figs. 2P and

2Q

1–Lichenized

Graphis scripta

Spo- Sporocarp type

0–Apothecioid Sporogenous tissue exposed at maturity.

Abrothallus usneae

Flask-shaped, with an ostiole

1–Perithecioid

Acrospermum graminearum

2–Thyriothecioid Flat, appressed, sporocarp, sporogenous tissue covered

Microthyrium macrosporum by a pigmented scutellum (comprise catathecia, which

p. 56 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

unlike most thyriothecia have a pigmented vental wall).

3–Cleistothecioid No preformed opening (include chasmothecia).

Lepidopterella palustris

Low- Lower wall

0–Undifferentiated Lower wall of sporocarp either absent or made up of

Microthyrium spp. one or a few layers of unpigmented cells, contrasting

with the darker upper sporocarp wall.

1–Differentiated Lower wall of sporocarp present and concolorous to the

Lichenopeltella pinophylla Fig. 4G rest of the sporocarp. It does not include host tissue and

is thus not a hypostroma.

Deh- Dehiscence

0–Ostiole Any dorsal pore by which spores are freed from an

Opening of most Pleosporalean fungi ascigerous or pycnidial sporocarp.

1–Irregular slit Elongated opening formed by cracking or splitting of

Prilleuxinia baccharidincola Fig. 2K the scutellum. The splitting is irregular in that upon

opening, it usually propagates radially between lateral

walls of neighboring hyphae but it can also propagate

transversely across hyphae. The split does not appear to

follow pre-formed lines of dehiscence.

p. 57 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

2–Regular slit Elongated opening formed in a regular fashion that

Opening found in hysterothecia and does not propagate irregularly.

lirella, Hysteropatella elliptica,

Graphis scripta

3–No dehiscence Sporocarp closed when mature, no mechanism of

Piedraia hortae dehiscence as part of sporocarp growth.

4–Exposed at maturity Sporocarp mostly open, sporogenous tissue is exposed

Abrothallus spp., Seuratia millardetii throughout its development

Loc- Locule circumference

0–circular Sporocarp enlarges from a center in all direction at

Asterinaceae Figs. 2H and 2J about the same pace so that it appears round in dorsal

Microthyriaceae Figs. 4A–4D and view.

4G–4J

1–elongate Sporocarp elongates as it enlarges from a central area

Lembosina aulographoides Fig. 3B into one or more 'arms.'

Rad- Radiate sporocarp wall

0–Absent No signs of a common central origin; distal hyphae

Micropeltis spp., Schizothyrium pomi cannot be traced back to central area of a sporocarp.

p. 58 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1–Present Most or all of the distal hyphae of the scutellum can be

Asterina spp., Microthyrium spp. traced back to their parental hyphae at or near the

Figs. 2, 3B, 3D and 4 center of the sporocarp.

Bra- Branching pattern (in hyphae of sporocarp wall)

0–Pseudoparenchyma Any wall of hyphal origin where linear hyphal

Most pseudothecia Figs. 5Aa and filaments can no longer be discerned. Instead, walls are

5Ab composed of polyhedral cells, and the original planes

of transverse cell divisions cannot be recognized.

1–Isotomous Scutellum hyphae radiate, isotomous dichotomies

Asterina spp. Fig. 5Ce predominate where in each pair of dichotomous

branches, one hyphal branch is wider than its sister.

Hyphal segments occasionally show evidence of

overlapping their neighbors.

2–Anisotomous Scutellum hyphae radiate, isotomous dichotomies and

Microthyrium macrosporum anisotomous dichotomies coexist but hyphal segments

Fig. 5Dh do not show evidence of overlapping their neighbors.

3–Pseudomonopodial Scutellum hyphae radiate with few isotomous

Lembosia spp., Aulographum spp. dichotomies; abundant evidence of overlap;

Fig. 5Ei pseudomonopodial dichotomies predominate and are

most evident at the periphery of the scutellum

p. 59 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

4–Untraceable Scutellum hyphae not radiate or if radiate, they cannot

Schizothyrium sp., Micropeltis spp., be traced from the sporocarp center to its margin.

Stomiopeltis sp. Isotomous dichotomies are few and neighboring

Figs. 3A, 3C and 5Fj–5Fk hyphae frequently overlap.

Mar- Appressed margin

0–No appressed margin Sporocarp without defined, appressed margin. Instead,

Pleospora herbarum, appressed the sporocarp wall usually curves upwards, forming a

margin not applicable; sporocarp is flask or disc shape.

flask-shaped

1–Crenulate to free Margin of sporocarp made of well-defined hyphal tips,

Batistinula gallesiae Fig. 2H each tip bulging out or elongating beyond the

longitudinal wall shared by adjacent cells.

2–Entire Hyphal tips appear truncate, blunt, at least in young

Microthyrium macrosporum specimens. In mature specimens, the blunt tips may

Figs. 4C–4D give rise to a fringe of tapering hyphae that are 2-5

times narrower than their mother cells.

3–Anastomosing Marginal hyphae of sporocarp continue outwards,

Schizothyrium pomi, Prillieuxina connecting or merging with surrounding superficial

baccharidincola Fig. 2K mycelium; hyphal tips not discernable.

App- Lateral appressoria

0–Absent Appressoria absent or intercalary if present

Aulographum sp. Fig. 3G

p. 60 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1–Present Appressoria lateral, branching directly from superficial

Asterina melastomatis Figa. 2P and hyphae.

2Q

Ini- Initiation of sporocarp

0–Terminal; at tip of generator hypha Generator hypha is a short lateral branch of a

Microthyrium spp. superficial hypha on a plant surface. The scutellum is

Figs. 5Dg and 5Dh derived from a single cell of the generator hypha, and it

overgrows generator hyphae at maturity.

1–Intercalary; below generator hypha Generator hypha is an intercalary (more rarely

Asterinales, Asterotexiales terminal) portion of a hypha that is superficial on a

Figs. 5Cc, 5Cd and 5Ce plant surface. The generator hypha remains above the

scutellum at maturity.

2–From spores Generator hypha absent; a scutellum grows directly

Asterotexis cucurbitacearum from each part-spore and becomes confluent with

Figs. 2E and 2F surrounding scutella as it enlarges.

3–From host stomata Generator hypha indiscernible. Scutellum arises

Rhagadolobiopsis thelypteridis directly from hyphae growing up through the stoma

Fig. 2G (likely from a hypostroma) and covers guard cells and

stomatal complex at maturity.

p. 61 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

4–Multiple generator hyphae More than one intercalary generator hypha participates

Aulographum sp. Fig. 5Ei1 in forming a single thyriothecium. A generator hypha

gives rise to the first dichotomous scutellum cells and

this differentiation propagates to adjacent portions of

the same hypha as well as to nearby hyphae, all of

which become generator hyphae and contribute to the

same scutellum.

1070

p. 62 bioRxiv preprint

1071 Table 3. Comparison of ancestral character states at selected nodes as reconstructed from proportional likelihoods in Mesquite (which wasnotcertifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. 1072 (unshaded rows) and from posterior probabilities in BayesTraits (shaded rows). doi:

1073 https://doi.org/10.1101/2020.03.13.989582

Ancestor reconstructed Sub Lic Spo Low Deh Loc Rad Bra Mar App Ini

Asterotexiales, clade of sup no thyr und irr circ yes iso cren yes int

Asterina spp (NAstx04) sup no thyr und irr circ yes iso cren yes -

Asterotexiales Clade A sup no thyr und irr circ yes iso cren no int ;

(NAstx03) sup no thyr und irr - yes - cren yes - this versionpostedMarch14,2020.

Asterotexiales (NAstx01) sup no thyr und irr - no - not no -

- no - und irr - - - - yes -

Aulographaceae (NAulo1) sup no thyr und irr elng yes mono cren no mult

- no thyr und irr - yes - cren no - The copyrightholderforthispreprint Zeloasperisporiales (NZelo2) sup no thyr - - - yes - - no -

sup no thyr - - - yes - - no -

Zeloasperisporiales + sup no thyr und - circ no aniso ent no -

Natipusillales (NZelo1) - no - - - - yes - - no -

p. 63 bioRxiv preprint

(which wasnotcertifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. Microthyrium (NMThyr4) sup no thyr und ost circ yes aniso ent no ter doi:

- no thyr und - - yes - - no - https://doi.org/10.1101/2020.03.13.989582

Microthyriales (NMThyr2) sup no thyr und ost circ yes aniso ent no -

- no thyr - - - yes - - no -

Microthyriales + Venturiales sup no thyr - ost circ no pseud not no -

(NMVP) - no - - - - yes - - no -

Muyocopron ;

+ sup no thyr und ost circ yes - cren no - this versionpostedMarch14,2020.

Mycoleptodiscus (NADM08) - no thyr und ost circ yes - cren no -

Dyfrolomycetales + sup no apo dif ost circ no - not no -

Muyocopronales (NADM04) - no - - - - yes - - no -

Asterinales, no Parmularia sup no thyr und irr circ yes iso cren no int The copyrightholderforthispreprint styracis (NAste2) sup no thyr und irr - yes - - yes -

Asterinales (NAste1) sup no thyr - irr circ yes - cren no -

sup no thyr und irr - yes - - yes -

Capnodiales ss + Asterinales sup no per dif ost circ no pseud not no - p. 64 bioRxiv preprint

(NCapnod5) - no - - - - no - - no - (which wasnotcertifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. Dothideomycetes (NDothid) imm no apo dif ost circ no pseud not no - doi:

sax no - - reg elng no - - no - https://doi.org/10.1101/2020.03.13.989582

Dothideomycetes + imm no apo dif exp circ no pseud not no -

Arthonimycetes + sax no - - - elng no - - no -

Trypetheliales (NDothTryp)

Micropeltidaceae (NMpelt1) sup no thyr und ost circ no untra ana no -

sup no thyr und ost - no - - no - ; this versionpostedMarch14,2020. 1074

1075 Notes: For 'Ancestor reconstructed', the node name given in parentheses is mapped onto the phylogeny in Appendix S6.

1076 Sub = sporulation substrate: sax, saxicolous; sup, superficial on plants; lich, lichenicolous; imm, immersed on plants

1077 Lic = lichenized

1078 Spo = sporocarp: apo, apothecioid; per, perithecioid; thyr, thyriothecioid The copyrightholderforthispreprint

1079 Low = sporocarp lower wall: und, undifferentiated; dif, differentiated

1080 Deh = dehiscence: ost, ostiole; reg, regular slit; irr, irregular slit; exp, exposed at maturity

1081 Loc = locule circumference: circ, circular; elng, ellipsoid to elongate

1082 Rad = radiate

p. 65 bioRxiv preprint

1083 Bra = scutellum branching: pseud, pseudoparenchymatous; iso, isodichotomous; aniso, anisodichotomous; mono, pseudomonopodial (which wasnotcertifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. 1084 dichotomies; untra, untraceable doi:

1085 Mar = margin: not, margin not appressed; cren, crenulate, aligned; ent, entire, aligned; anast, anastomosing with surrounding https://doi.org/10.1101/2020.03.13.989582

1086 mycelium

1087 App = lateral appressoria; no, absent; yes, present

1088 Ini = initiation of sporocarp, position and number of generator hyphae: int, single, intercalary; ter, single, terminal; mult, multiple,

1089 intercalary

1090 ; this versionpostedMarch14,2020. 1091 The copyrightholderforthispreprint

p. 66 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Ancestral character states transitions Buelliella_poetschii Ertz_18116(BR) Hemigrapha_atlantica Ertz_14014(BR) Asterotexiaceae_sp.2 CBS_143813 Rhagadolobiopsis_thelypteiridis EG_156 Asterotexiaceae_sp.1 UBC − F33036 Ps 10 Mycosphaerella_pneumatophorae AFTOL − ID_762 2 Lembosia_albersii MFLU13 − 377 Lembosia_xyliae MFLU14 − 4 Inocyclus_angularis VIC_39747 Asterotexis_cucurbitacearum VIC_42814 Asterotexiales A Asterotexis_cucurbitacearum PMA_M − 141224 9 Asterina_sp. MFLU13 − 619 Ps 0 9 10 Asterina spp.

8 Morenoina_calamicola MFLUCC_14 − 1162 Ps 4 Geastrumia_polystigmatis FJ147177.1 Aulographum spp. 8 Lembosina spp. Aulographaceae

Z_eucalyptorum CBS_124809 Z_wrightiae MFLUCC_15 − 215 Z_cliviae CPC_25145 Ps 7 Z_siamense IFRDCC_2194 Zeloasperisporiales 3 Z_ficusicola MFLUCC_15 − 222 D Z_searsiae CPC_25880 7 Natipusillales O Heliocephala_zimbabweensis MUCL_40019 6 Heliocephala_gracilis MUCL_41200 T Ps 0 Tumidispora_shoreae MFLUCC_12 − 409 ? T Chaetothyriothecium_elegans CPC_21375 Stomiopeltis_betulae CBS_114420 ? Microthyriales H Microthyrium spp. I 6 Lichenopeltella_pinophylla UBC − F33032 D

Ps ? 5 Venturiales E cf.Stomiopeltis_sp._1 CBS_143811 Tothia_fuscella CBS_130266 cf.Stomiopeltis_sp._2 UBC − F33041 O 4 Pleosporomycetidae M Ps 0 Muyocopron_lithocarpi MFLUCC_10 − 41 Y 5 Muyocopron_lithocarpi MFLUCC_14 − 1106 Muyocopron_castanopsis MFLUCC_14 − 1108 4 Muyocopron_dipterocaerpi MFLUCC_14 − 110 C Mycoleptodiscus_indicus UAMH_8520 Muyocopronales Muyocopron garethjonesi MFLU_16 − 2664 Paramycoleptodiscus_albizziae CPC_27552 E Arxiella_dolichandrae CBS_138853 Dyfrolomyces_rhyzophorae JK_5349A Dyfrolomyces_tiomanensis NTOU3636 Dyfrolomycetales Ps 3 T 2 E Stomiopeltis_versicolor GA3_23C2b #? Houjia_yanglingensis YHJN13 S 1 Uwebraunia_commune CBS_110747 Ps 2 Schizothyrium_pomi CBS_228.57 "Schizothyriaceae" 0 Johansonia_chapadiensis CBS_H − 20484 F

? Blastacervulus_eucalypti CBS_124759 2 3 Blastacervulus_eucalyptorum CPC_29450 iP 1 Asterina_melastomatis VIC_42822 0 A Lembosia_abaxialis VIC_42825 Asterinales s.s. Asterina_chrysophylli VIC_42823 Prillieuxina_baccharidincola VIC_42817 Batistinula_gallesiae VIC_42514 Parmularia_styracis VIC_42587

Peltaster_fructicola JN573665.1 "Capnodiales" Ps 1 M "Dothideomycetidae" Abrothallales + Tubeufiales

Arthoniomycetes + Trypetheliales

Potential fossil calibrations Scolecopeltidium_sp.1 UBC − F33033 Scolecopeltidium_sp.2 UBC − F33035 Micropeltis_zingiberacicola IFRDCC_2264 Micropeltidaceae s.s. M Micropeltis_sp. UBC − F33034 F Triassic fossil 'fungal thalli' Cryptodiscus_gloeocapsa TSB_30770 Absconditella_sphagnorum EU940095.1 Bryodiscus_arctialpinus Baloch_SW057(S) Ostropomycetidae Acarosporina_microspora AFTOL − ID_78 Early Eocene Stictis_radiata AF356663.1 T Lecanoromycetes Trichothyrites setifer Eurotiomycetes A Middle Eocene Asterina eocenica Leotiomycetes

Substrate Sporocarp lower wall Dehiscence Scutellum hyphal pattern Isotomous, overlapping 2 Immersed in Pseudoparenchyma iP Undifferentiated Ostiole Exposed Isotomous, plant tissue 0 unspecified 3 non-overlapping On plants and Ps plant surfaces Differentiated Irregular slit(s) Untraceable Pseudomonopodial 1 4 bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Young sporocarp Mature sporocarps

1 Pseudoparenchyma 3 Isotomous 4 Anisotomous

(d1) (g ) (g ) (g ) (a) (b) (c) (d1) (d2) 1 2 3

(e2) (f) 1 (e1) (h1) (h2) (h3)

2 Tracing radiate branching Isotomous dichotomies overlapping 5 Pseudomonopodial 6 Untraceable rgence angle of dive

betwee n bra nching t (k) ips (i2) (j)

es no overlap Anisotomous dichotomi (i1)

(l)

Legend

Example traced Isotomous dichotomies Branching involving overlap Generator hyphae ontological relationship from centre to margin Anisotomous dichotomies Pseudomonopodial dichotomies Scutellum opening bioRxiv preprint doi: https://doi.org/10.1101/2020.03.13.989582; this version posted March 14, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.