Emmanuelia, a new genus of lobarioid lichen-forming fungi (Ascomycota: Peltigerales): phylogeny and synopsis of accepted species

The former family Lobariaceae, now included in Peltigeraceae as subfamily Lobarioideae, has undergone substantial changes in its generic classification in recent years, based on phylogenetic inferences highlighting the polyphyly of the speciose genera Lobaria, Pseudocyphellaria and Sticta. Here we introduce the new genus Emmanuelia, named in honor of Prof. Emmanuël Sérusiaux for his extensive work on the Peltigerales. Emmanuelia currently comprises twelve species. It is superficially similar to the lobarioid genus Ricasolia, but differs by its apothecia, rimmed by overarching and often crenulate to lobulate margins, with the parathecium (proper excipulum) and the amphithecium (thalline excipulum formed by the thallus cortex) apically separated and of a different structure. Also, ascospore dimensions and shape differ between the two genera, with the ascospores of Emmanuelia being longer and narrower. Molecular phylogenetic analyses using DNA nucleotide sequences of the internal transcribed spacer region (ITS) and the small subunit of mitochondrial ribosomal DNA (mtSSU) confirm that Emmanuelia belongs to the Lobaria s.lat. clade and forms a monophyletic group sister to the lineage consisting of Dendriscosticta, Lobariella and Yoshimuriella. None of the available generic names of lobarioid lichens can be applied to this group, and consequently a new name is proposed for this new genus, which is typified with E. ravenelii comb. nov. Eleven other species are transferred to Emmanuelia: E. americana comb. nov., E. conformis comb. nov., E. cuprea comb. nov., E. elaeodes comb. nov., E. erosa comb. nov., E. excisa comb. nov., E. lobulifera comb. nov., E. ornata comb. nov., E. patinifera comb. nov., E. pseudolivacea comb. nov. and E. tenuis comb. nov. The genus is represented in North America by three species, including E. lobulifera, which is resurrected from synonymy with E. (Lobaria) tenuis, a South American species, and E. ornata, whose populations were previously treated under E. (Lobaria) ravenelii.


Introduction
Recent years have witnessed major systematic rearrangements of lobarioid lichens. In less than a decade, the number of genera circumscribed in this lineage of conspicuous macrolichens was multiplied by four. Essentially, the three long-established genera (Lobaria, Pseudocyphellaria, Sticta), which were diagnosed by single morphological features (the presence/absence of cyphellae or pseudocyphellae), did not survive the advent of molecular phylogenetics and were partitioned into twelve genera (Galloway & Elix 2013;Moncada et al. 2013;Galloway 2015;McCune et al. 2014). As an example, moon lichens, which were characterized by the presence of crater-like pores on the lower cortex, are no longer considered as a monophyletic group under the genus name Sticta, as this trait evolved in two unrelated lineages. Thus, the genus Dendriscosticta, more closely related to Lobaria, was introduced to accommodate the additional lineage (Moncada et al. 2013).
In addition to the profound changes in generic concepts, lobarioid lichens were not spared from a recent systematic revision at the family level: in a recent study by Kraichak et al. (2018), under a temporal-banding proposal (Avise & Johns 1999;Kraichak et al. 2017), the authors proposed treating the families Lobariaceae and Nephromataceae as synonyms of Peltigeraceae. While the mechanistic approach of temporal banding classifications has been criticized, the broad agreement in morphological, anatomical and chemical features and the absence of a clear diagnostic character for each of the three previously separated families justifies this revised classification (Lücking 2019). As a consequence, lobarioid lichens, long treated as Lobariaceae, are now recognized as members of the subfamily Lobarioideae within Peltigeraceae (Lumbsch & Leavitt 2019).
In the present study, yet another new genus, Emmanuelia, is erected to accommodate a group of lobarioid lichens that cannot be placed in any of the existing genera. The species of interest belong to a lineage mostly restricted to the Neotropics and the southeastern United States, and were previously treated as members of the genus Lobaria and subsequently considered part of Ricasolia (Moncada et al. 2013;Käffer et al. 2016;Lehnen et al. 2017;Etayo et al. 2018). Yoshimura (1998) treated the South American taxa as L. quercizans group, also implying a close relationship to Ricasolia. This group of lichens includes, among others, shade-loving species of the Atlantic Forest biome in South America such as L. tenuis (Käffer et al. 2009), as well as L. ravenelii, a well-known taxon characteristic of the Atlantic-Gulf Coastal Plain in North America (Jordan 1973). To address their phylogenetic affinity, we reconstructed the phylogeny of the Lobaria s.lat. clade by using sequence data of two loci obtained from seven related genera of Lobarioideae. Our molecular analysis confirmed that L. ravenelii and other related species should be accommodated in a new segregated genus of Lobaria s.lat.

Taxon sampling and phenotypic characterization
The present study is based on detailed examination of lichen specimens provided by NY (William and Lynda Steere Herbarium, New York, USA) and numerous freshly collected specimens from fieldtrips to Brazil, the Caribbean Islands and Galapagos Islands. Thirty-four representatives of the taxonomic group of interest were selected based on preliminary phylogenetic analysis of the internal transcribed spacer region (ITS), along with ten specimens from related genera. Morphological features were observed at various laboratories, using various dissecting microscopes (Olympus SZ60, Leica Zoom 2000) and compound microscopes (Olympus BX53, Nikon Eclipse 80i, Zeiss Axioskop). Secondary chemistry was investigated through thin-layer chromatography (TLC) using solvent G and following the protocol by Orange et al. (2001). Detailed descriptions were provided for the generic type of the newly introduced genus (E. ravenelii) and for the resurrected species (E. lobulifera). For all other species we added short diagnostic descriptions.

DNA extraction, amplification, and sequencing
Genomic DNA was isolated using Nucleospin Plant II Midi kits (Macherey-Nagel, Bethlehem, Pennsylvania, USA), following the manufacturer's guidelines or following the protocol by Cubero et al. 1999. We amplified and sequenced the internal transcribed spacer region (ITS) using primers ITS1F (Gardes & Bruns 1993) and ITS4 (White et al. 1990), and the small subunit of mitochondrial ribosomal DNA (mtSSU) using primers SSU1 and SSU3R (Zoller et al. 1999). Standard PCR protocols were carried out using GoTaq Green Master Mix (Promega, Madison, Wisconsin, USA), following the manufacturer's guidelines. The thermal cycling parameters were set as follows: 94°C for 3 min, followed by 35-40 cycles of 94°C for 1 min, 52°C for 1 min and 70°C for 1 min, with a final extension of 70°C for 10 min. The quality and size of the amplicons were visually checked on a 1% w/v agarose gel stained by SYBR Safe DNA Gel Stain (Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA). Amplicons were cleaned using the ExoSAP-IT protocol (USB Corporation, Cleveland, Ohio, USA) and sequenced on a ABI Prism 3100 Genetic Analyzer (Applied Biosystems, Foster City, California, USA), or by Macrogen, Inc. (Seoul, South Korea). The forward and reverse sequences obtained were assembled and edited using Geneious 10.0.7 (Biomatters Ltd.,Auckand,New Zealand) or Sequencher version 4.9 (Gene Codes Corporation, Ann Arbor, Michigan, USA). The resulting sequences were submitted to GenBank (Table 1).

Alignment and sequences analyses
The sequences generated for each gene were aligned with selected sequences from species of the Lobaria s. lat (Katoh 2002;Katoh et al. 2009), and the ends of each alignment were trimmed. The final matrices were obtained using default MAFFT settings, with two alignment iterations, as implemented in the Guidance Web Server (Penn et al. 2010a, b;Sela et al. 2015). Sites with low-quality scores (i.e., with confidence scores below 0.93) reported by GUIDANCE 2 were excluded from the datasets, which resulted in alignments of 589 bp (initially 779 bp) and 700 bp (initially 941 bp) for ITS and mtSSU, respectively. As strongly supported topological conflicts were not observed when the loci were analyzed separately, the two markers were combined into a concatenated matrix of 1289 bp. The concatenated dataset included 55 terminals, all represented by ITS, and 47 of which were also represented by the mtSSU marker. Parti-tionFinder 2 (Lanfear et al. 2016) was used to determine the best partitioning schemes and nucleotide substitution models for the subsequent maximum likelihood (ML) analysis on the concatenated dataset. Two initial subsets were considered (ITS, mtSSU) and the default configuration settings were used (branchlengths = linked, models = GTR+G, model_selection = AICc) with the greedy algorithm (Lanfear et al. 2012) and PhyML (Guindon et al. 2010). An ML analysis was performed on the 2-gene dataset using RAxML-HPC2 8.2.12 (Stamatakis 2014) on the CIPRES portal (Miller et al. 2010; https://www. phylo.org), using the rapid hill-climbing algorithm and bootstrapping with 1000 pseudoreplicates under a GTR+G model of evolution for each subset provided by Partition-Finder 2 (subset 1: ITS; subset 2: mtSSU).
Relationships among the Emmanuelia elaeodes species aggregate (E. americana, E. elaeodes, E. ravenelii, E. tenuis) remained largely unresolved in our ML phylogenetic tree (Fig. 1A). For this reason, we constructed a haplotype network for these closely related species, using the TCS v1.21 program (Clement et al. 2000) as implemented in PopART software (Leigh & Bryant 2015). The ITS sequences of these species were re-aligned using the general MAFFT settings as implemented in the Guidance Web Server. Since the resulting alignment contained relatively few ambiguous portions, the dataset was loaded with all sites included. Sites with undefined states were then masked, and sequences containing significantly more undefined states than others were removed from the analysis.
Based on the 2-gene dataset, a strict molecular clock model was employed to date the evolutionary origin of the genus Emmanuelia, using the Bayesian program BEAST 1. 10.4 (Drummond & Rambaut 2007). We initially conducted a run using a relaxed, log-normal, uncorrelated clock: this preliminary run supported a clock-like rate of evolution, as the standard deviation estimate of the clock (i.e., the 'ucld.stdev' parameter estimate) was close to zero. Consequently, a strict clock prior was applied. The dataset was analyzed with unlinked substitutions models across the two loci, and the most appropriate nucleotide substitution model for each locus was determined based on the AICc model selection criterion as implemented in jModelTest2 (Darriba et al. 2012) and using five substitution schemes (ITS: TrN+I+G; mtSSU: HKY+I+G). A Yule prior was assigned to the speciation process (Yule 1924;Gernhard 2008). The 'ucld.mean' prior (mean substitution rate) was set to a diffuse gamma distribution (shape 0.001, scale 1000). The time to the most recent ancestor ('tmrca') for the ingroup node (Lobaria s.lat. clade) was calibrated at 57.6 myr, using a normal prior distribution with the standard deviation set to 13 myr; this calibration followed the results of the time-calibrated Lobariaceae phylogeny by Widhelm et al. (2019). All other priors were held to default values. The BEAST analysis was run for 50 million generations, sampling parameters every 5000 steps, and performed on the CIPRES Science Gateway (Miller et al. 2010). Convergence, mixing, and effective sample sizes (ESS) of parameters were checked in Tracer 1.6 ( Rambaut et al. 2014). All ESS values were above 200. A burn-in of 10% was discarded from the run. A maximum credibility tree with a cut-off of 0.5 of posterior probabilities was generated with the remaining 9,000 trees in TreeAnnotator version 1. 10.4 (BEAST package). The results of the ML and Bayesian analyses were visualized with the R package ggtree (Yu et al. 2017).

Results
Emmanuelia emerged as a strongly supported monophyletic group (ML bootstrap support [BS]=100, posterior probabilities [PP] = 1) on a fairly long branch within the Lobaria s.lat. clade (Fig. 1A). In contrast, backbone support values within Emmanuelia were low, suggesting a rather recent radiation. Emmanuelia was recovered as sister to a clade consisting of Dendriscosticta, Lobariella and Yoshimuriella by both ML and Bayesian inferences, and not directly related to Ricasolia, justifying the introduction of a new genus to accommodate this group of lichenized fungi. Within Emmanuelia, strong support was obtained for a clade of four species: E. americana, E. elaeodes, E. ravenelii and E. tenuis. Relationships within this clade remained mostly unresolved under ML. The TCS haplotype network (Fig. 1B) for 13 specimens within the E. elaeodes species aggregate further highlighted the lack of signal in the ITS marker to segregate these species (with the exception of E. ravenelii, which appeared well-differentiated in both TCS and ML analyses).
Although Ricasolia is not directly related to Emmanuelia, it is morphologically most similar to the latter, whereas the related genera Dendriscosticta, Lobariella and Yoshimuriella are easily distinguished. Yet, analyses of morphological and anatomical characters revealed a number of subtle differences between Emmanuelia and Ricasolia. In Emmanuelia the apothecial margins are typically overarching and lobulate, with a rough outer surface, whereas in Ricasolia they are slightly prominent, more or less entire, and with a smooth outer surface. In Emmanuelia the parathecium (i.e., proper excipulum) appears to be apically separated from the amphithecium (i.e., thalline excipulum formed by the thallus cortex) by the photobiont layer which reaches up to the apex, and also different in structure (prosoplectenchymatous vs. paraplectenchymatous; Fig. 3A-B). In Ricasolia the parathecium and amphithecium are apically connected, due to the photobiont layer stopping short below the apex, and at least the upper part of the parathecium is paraplectenchymatous ( Fig. 3C-D). A further difference is found in the ascospores, with those of Emmanuelia generally acicular and arranged in a bundle and those of Ricasolia fusiform and irregularly arranged to uniseriate. In the three common species of Ricasolia the ascospores are as follows: 45-60 × 4-5 µm, ~9-12 times as long as broad (R. quercizans), 40-60 × 5-7 µm, ~7-9 times as long as broad (R. amplissima), and 25-45 × 6-10 µm, ~4-5 times as long as broad (R. virens). By contrast, for A -Best-scoring phylogenetic tree of the genus Emmanuelia within the Lobariaceae family, obtained from inferences of two-loci molecular data (ITS, mtSSU). Values above branches represent ML bootstrap values (indicated when greater than or equal to 50). Colors in taxa names of Emmanuelia spp. represent sample origin (Brazil: dark blue; Mainland United States: orange; Caribbean: purple; Galapagos Islands: light blue). B -Relationships among Emmanuelia intermedia complex ITS haplotypes, using statistical parsimony networks as implemented in TCS. Each colored circle represents one observed haplotype and one individual. Individuals whose sequences contained significantly more undefined states than others were removed. Circles are color-coded by origin. Hatch marks and black dots represent hypothetical haplotypes needed to connect the network but not observed among the samples. The low backbone support for phylogenetic relationships within Emmanuelia may be due to a recent rapid radiation, an evolutionary scenario recently highlighted within the Lobarioideae (Lücking et al. 2017b;Simon et al. 2018  Ricasolia and Yoshimuriella have similar crown ages, whereas Dendriscosticta and particularly Lobaria s.str. appear to be older. The fairly restricted geographic distribution of Emmanuelia, as compared to other more widespread related genera, tends to support the scenario of a recent diversification. Notably, the crown divergence time for Emmanuelia coincides with the onset of the uplift of the northern Andes (Hoorn et al. 2010). The latter may explain the rather recent diversification of genera such as Lobariella (Moncada et al. 2013), but most of the Emmanuelia species occur in the Atlantic Forest of southeastern Brazil, in a geologically much older formation (Peucker-Ehrenbrink & Miller 2007;Brotzu et al. 2007;Colombo & Joly 2010). Possibly the recent diversification of Emmanuelia largely in the Atlantic Forest, with subsequent expansion northwards, is related to paleoclimatic events, as apparent from other groups of organisms in this biome with similarly recent diversification times (e. g., Fouquet et al. 2012;Batalha-Filho et al. 2013;Machado et al. 2018).
Low taxonomic resolution possibly resulting from recent diversification is particularly evident in the E. elaeodes aggregate, where morphological identifications do not necessarily correspond to phylogenetic topology (with the exception of E. ravenelii). Two possible reasons for this may be offered: (1) the ITS and mtSSU markers may not be sufficiently informative to disentangle species within this complex, or (2) this complex may represent a single species with morphological variation. In particular, E. tenuis may represent the phyllidiate counterpart of E. elaeodes. Further studies are needed to resolve the taxonomy of this complex.

Taxonomy
Emmanuelia Ant. Simon,Lücking & Goffinet,gen. nov. MycoBank MB 834643 Diagnosis: A lobarioid genus lacking cyphellae, pseudocyphellae and maculae, with a short, ± uniform lower tomentum (no veins), primarily associated with a green alga, sometimes with a dendriscocauloid cyanomorph, with gyrophoric acid (major) and 4-O-methylgyrophoric acid (congyrophoric acid; minor or absent) as secondary compounds. Morphologically and chemically similar to the genus Ricasolia, but differing in the apothecia with overarching margin and separation of the parathecium and amphithecium, consistently narrower and longer, acicular ascospores, and a subtropical to tropical distribution. Molecularly, the new genus is characterized by a short, unique rDNA sequence motif within the highly conserved 5.8S region in the ITS. The section defining Emmanuelia is as follows (deviations underlined): 5-CGAATCATCGAATCTTTGAACGCACATTGCGC-CCCYYGGYAC-3. Comments. Morphologically, Emmanuelia can be easily differentiated from Lobaria s.str. by the tomentum on the lower surface and the shape of the ascospores (Yoshimura 1998;Moncada et al. 2013). However, the newly introduced genus is quite similar to Ricasolia. In particular, like some species of Ricasolia, at least one species of Emmanuelia, E. ornata (previously often identified with the name Lobaria patinifera), produces dendriscocauloid cyanomorphs emerging from the green-algal thallus (e. g., Jordan 1972;Tønsberg et al. 2016; Fig. 6C-D). The two genera differ in their geographical distribution and ecology: Emmanuelia is a subtropical to tropical taxon found from southeastern North America to southern South America, whereas Ricasolia appears to be a strictly temperate, Northern Hemisphere taxon (Cornejo et al. 2017); the two genera are somewhat sympatric in the southeastern United States, but Emmanuelia replaces Ricasolia in coastal areas (Jordan 1973).
Morphologically, Emmanuelia differs from Ricasolia by its apothecia, rimmed by overarching and often crenulate to lobulate margins with a rough surface, whereas in Ricasolia the margins are only slightly prominent, more or less entire, and with a smooth surface, an observation also noted by Yoshimura (1998) when comparing tropical species of his 'Lobaria quercizans group' to Ricasolia quercizans s.str. Anatomically, in Emmanuelia the parathecium (proper excipulum) appears to be apically separated from the amphithecium (formed by the thallus cortex), by the photobiont layer reaching up to the apex, and also different in structure (prosoplectenchymatous vs. paraplectenchymatous), a characteristic referred to as 'apothecium type II' (Yoshimura 1971;Yoshimura & Osorio 1975). In Ricasolia the parathecium and amphithecium are apically connected, due to the photobiont layer stopping short distinctly below the apex, and at least the upper part of the parathecium is paraplectenchymatous. A further difference is found in the ascospores, with those of Emmanuelia generally acicular and arranged in a bundle and those of Ricasolia fusiform and irregularly arranged to uniseriate.
Below we provide brief diagnostic descriptions and comments for ten of the 12 species included in the new genus, and more detailed accounts for the type species, E. ravenelii, and the reinstated eastern North American E. lobulifera. Diagnostic description. Primary photobiont a trebouxioid alga. Thallus medium-sized to large, rather loosely adnate, composed of irregularly arranged lobes. Lobes leathery, rather broad, up to 12 mm wide, sinuose with ± rounded apices. Upper surface ± even but often with abundant pycnidial warts, glabrous. Lower surface thinly tomentose except for bare marginal zone, tomentum brown. Apothecia scattered, laminal, to 7 mm in diam., margin strongly prominent, lobulate. Ascospores acicular, straight to slightly curved, 1(-3)-septate, 60-80 × 2.5-3 µm, ~20-30 times as long as broad.

Secondary chemistry. Gyrophoric acid (major).
Comments. This material was first considered conspecific with Emmanuelia elaeodes (see comments below) but differs in the overall much smaller size and in the 3-5-septate ascospores, as well as in its distribution. Yoshimura (1998) considered Lobaria conformis a synonym of L. patinifera, with the latter name in his treatment being misapplied to what is here recognized as E. ornata. However, E. conformis differs from both E. ornata and E. patinifera (see key below). Notably, several studies had reported L. conformis as a separate species prior to Yoshimura's (1998) treatment (e.g., Osorio & Fleig 1987Sipman 1993
Comments. This taxon is characterized by the purplish red color of the underside in rewetted herbarium specimens (Yoshimura & Osorio 1975;Osorio & Fleig 1987); this corresponds to a specific pigment detectable by TLC (Yoshimura & Osorio 1975). A similar effect has been reported for, e.g., Cora rubrosanguinea from Ecuador (Lücking et al. 2017a).

Secondary chemistry. Gyrophoric acid (major), 4-O-methylgyrophoric (congyrophoric) acid (minor).
Comments. We first named the Brazilian material Ricasolia intermedia, a species described from Mexico. The type material of Lobaria elaeodes, Ricasolia intermedia, and L. conformis, also described from Mexico, all agree in thallus morphology and lobe disposition. However, after discovering the differences in ascospore size and shape between Emmanuelia and Ricasolia, we noticed that the ascospores of R. intermedia, not given in the protologue, were described as rather short and broad by Stizenberger (1895), namely ~35-45 × 10 µm, ~3.5-4.5 times as long as broad. On the type material, also rather broad ascospores are depicted, without measurements but ~7 times as long as broad. In contrast, the ascospores of the types of L. elaeodes and L. conformis fit those of Emmanuelia. Therefore we consider R. intermedia a genuine member of that genus, whereas the Brazilian material represents the genus Emmanuelia and must bear the epithet elaeodes. The Mexican L. conformis is also considered a species of Emmanuelia but it differs from E. elaeodes by the smaller thallus with narrower lobes and the multiseptate ascospores.
Specimens examined. BRAZIL

Secondary chemistry. Gyrophoric acid (major).
Comments. This is a rather characteristic species, which in the past had been synonymized with Emmanuelia (as Lobaria) ravenelii and even Ricasolia quercizans (e.g., Stizenberger 1895), but its rather narrow, scrobiculate lobes are distinctive and its separation is supported by molecular data.

Secondary chemistry. Gyrophoric acid (major).
Comments. Müller (1891) described Sticta excisa from Colombia and Jamaica. There are three syntypes in G (Colombia, Jamaica) and M (Colombia). The material from Jamaica does not bear apothecia, so its identity cannot be ascertained. Therefore the well-developed specimen from Colombia in M was selected as lectotype. Emmanuelia excisa is somewhat intermediate between E. patinifera and E. elaeodes but it differs from the first in the narrower, radiating, adnate lobes and smaller, horizontal apothecia, and from the second in the more robust thallus and distinctly lobulate apothecial margins. The ascospores of E. excisa are the among the broadest thus far known in the genus, but still narrower than those of Ricasolia quercizans (Fig. 4). Description. Primary photobiont a trebouxioid alga. Thallus irregular in outline, small to medium-sized, to 8 cm diam., composed of radiating, stiff , repeatedly branched and spreading lobes. Lobes adjacent to imbricate and overlapping, from 1 mm wide at branching point and to 5 mm wide above, with truncate to spatulate apices, adnate, with free, plane margins; margins entire, occasionally with small simple lobules inward. Upper surface smooth, glabrous, light greenish or rarely brownish grey and light brownish towards margin when dry, greenish when wet, matte; margin lacking pruina. Phyllidia mostly laminal, somewhat obliquely oriented, squamiform, almost orbicular to palmate, mostly unbranched to sparsely branched, either at base or along margin bearing one or more lobules, or dichotomously branched, typically ~0.5 mm in diam., to 1 mm long. Medulla rather compact, white, KC+ pink. Cephalodia (with Nostoc) internal, globose, to 0.2 mm in diameter. Lower surface smooth at margin, rugose, verrucose inward, thinly tomentose, with hairs in short fascicules, light brown to cream-colored. Rhizines simple, abundant to scattered, to 0.8 mm long, whitish, darkening inward. Apothecia not observed. Upper cortex paraplectenchymatous, 20-30 µm thick, homogeneous, consisting of 4-6 cell layers. Photobiont layer 15-30 µm thick, its cells ~6 µm diam. Medulla 60-100(120) µm thick. Lower cortex paraplectenchymatous, 15-20 µm thick, with 3 cell layers; surface papillose to microtomentose in section in between short fascicles. Secondary chemistry. Gyrophoric acid (major), 4-O-methylgyrophoric acid (minor).

Ecology and distribution.
On hardwood trees on the coastal plain of the southeastern United States.

Comments.
Our study revealed that Lobaria lobulifera (Moore 1969), previously synonymized under L. tenuis, is a distinct species, here resurrected as E. lobulifera. The shared presence of phyllidia led Yoshimura (1971) and Jordan (1973) to synonymize the two taxa, but both the position of the phyllidia (largely laminal vs. marginal) and their shape (squamiform vs. elongate) differ between E. lobulifera and E. tenuis, and their distinction is supported by phylogenetic data. The different distribution of the two taxa provides another argument, with E. tenuis mostly known from Brazil (e.g., Lücking 37544b, 40067), including the type, and E. lobulifera from the southeastern United States (e. g., Kaminski LK450, Lendemer 21578, 41467, and Rosentreter 19739), with its type collected in Florida (Moore 1969 Comments. This species is here reported for the first time from the United States and the Galapagos Islands. North American material was previously considered a form of Lobaria ravenelii with erumpent cephalodia (Jordan 1973). In tropical America this taxon was mostly named L. patinifera, but the type of the latter is entirely different from the material with narrow, adnate lobes and cephalodia to which this name has been applied (see below). The dendriscocauloid cephalodia are reminiscent of those of Ricasolia species, and since apothecia are apparently absent in this species, it cannot be readily assigned to Emmanuelia without molecular data. ≡ Ricasolia patinifera (Taylor) Müll. Arg., Flora 71: 24. 1888.

Comments.
The name Lobaria patinifera has been commonly misapplied to specimens with an adnate thallus, with narrow lobes and forming abundant, dendriscocauloid cephalodia. Revision of type material revealed that Emmanuelia patinifera is characterized by rather broad, loosely attached lobes, large, distinctly lobulate apothecia, and lack of external cephalodia. The type material of Sticta casarettiana and the other taxonomic synonyms listed here correspond well to this morphology. The ascospores of E. patinifera were described as 3-7-septate by some authors, but this was likely based on either wrongly identified specimens or misinterpretation of the septa of the very narrow ascospores. Typically the ascospores are 1-septate, occasionally 3-septate. Emmanuelia patinifera is a peculiar species, emerging from the earliest split in the genus in our ML tree: it has the largest thallus and most robust lobes, as well as the largest apothecia and ascospores in the genus; the apothecia are obliquely oriented, and the parathecium (proper exciple) is formed by peculiar, thick, palisadic hyphae. Magain et al. (2012) reported this species from Reunion (Mascarene archipelago), but the ITS sequence generated does not fall into Emmanuelia and their specimen may belong to Lobariella. Specimens examined. BRAZIL Comments. This species is characterized by the rather narrow lobes combined with large apothecia, and the lack of secondary substances, an unusual feature in the genus, which is otherwise chemically uniform.

Secondary chemistry. Gyrophoric acid and 4-O-methylgyrophoric acid.
Ecology and distribution. On hardwood trees in the Atlantic-Gulf Coastal Plain in North America, also known from the Antilles and reported from Panama (Büdel et al. 2000). Comments. While E. ravenelii has been considered a synonym of E. erosa [Sticta erosa] by Tuckerman (1882) and other later authors, the two species can be clearly distinguished by some morphological features, such as the absence of pruina on the upper, more distinctly scrobiculate surface of the latter, as well as a different geographical distribution (North America and Greater Antilles versus South America, respectively). Emmanuelia (as Lobaria) ravenelii has been repeatedly reported from Brazil (Kalb 1983;Brako et al. 1985). The specimens distributed in Kalb's Lichenes Neotropici (No. 237)

Secondary chemistry. Gyrophoric acid (major).
Comments. For a detailed discussion, see under E. lobulifera (above). The correct application of this name remains somewhat uncertain. In our phylogeny, specimens with marginal phyllidia are found in two different clades: one small separate clade (37504, 39705) and one larger clade intermingled with non-phyllidiate specimens identified as E. elaeodes (37502, 37544b, 40067). The two differ in the size and disposition of the phyllidia, which in the small separate clade are larger and oriented in the same direction as the lobes, and in the larger mixed clade are smaller and somewhat obliquely arranged. Both taxa were found sympatric at one locality. Vainio's original material appears to correspond to the form with smaller, obliquely arranged phyllidia, which would mean that the separate clade, currently labeled aff. tenuis, requires a name.