Introduction
The genus Micarea, described by Fries (Reference Fries1825) and conserved with M. prasina as the type species (Coppins Reference Coppins1989, ICBN appendix III), was resurrected from oblivion at the end of the 19th century (Hedlund Reference Hedlund1892). In his precocious revision of small crustose lichens, Hedlund (Reference Hedlund1892) emended Micarea to include 20 species. As circumscribed by him, the genus included species with unicellular or transversely septate ascospores, branched and anastomosing, apically unthickened paraphyses, an excipulum composed of paraphysis-like hyphae, a small-celled photobiont [4–8(–9) µm, later known as ‘micareoid’] and immarginate, often tuberculate apothecia.
Several decades later, Zahlbruckner (Reference Zahlbruckner1921–40) introduced an artificial but highly influential taxonomy in his Catalogum Lichenum Universalis, in which lecideoid lichens were sorted according to spore septation. Consequently, the genus Micarea was split and its species again transferred to other genera, mainly Lecidea, Catillaria, and Bacidia. When the Zahlbrucknerian ice sheet slowly melted in the 1960s and 70s, the pursuit for a more natural classification was revitalized and followed by the revival of previously described genera, as well as the description of numerous new ones.
One of the first to yet again re-establish Micarea was Anderson (Reference Anderson1974), who transferred Lecidea tuberculata Sommerf. to this genus. The inclusion of a species with a non-micareoid photobiont was in effect an emendation of Hedlund's generic delimitation, although Anderson did not discuss this. Subsequently, Vězda & Wirth (Reference Vězda and Wirth1976) published a new key to the genus, accepting Hedlund's work but also adding several species, such as Micarea bauschiana (Körb.) Vězda & Wirth, M. lutulata (Nyl.) Coppins (as M. umbrosa Vězda & Wirth) and M. sylvicola (Flot.) Vězda & Wirth. They noted that M. bauschiana and M. lutulata were probably closely related, but did not comment on the similarities between these species and M. sylvicola and M. tuberculata.
In his breakthrough revision of the European species of Micarea, Coppins (Reference Coppins1983) noted that the genus included several small groups of apparently closely related species. He recognized 11 infrageneric groups in Micarea, one of which (group ‘I’, hereafter known as the Micarea sylvicola group) consisted of M. bauschiana, M. lutulata, M. sylvicola, and M. tuberculata. Coppins concluded that this group was ‘almost worthy of subgeneric status’.
With the advent of molecular methods in taxonomy came the opportunity to test this hypothesis. Andersen & Ekman (Reference Andersen and Ekman2005), in a phylogeny based on mitochondrial ribosomal DNA, showed that Micarea was paraphyletic, and that several of the infrageneric groups identified by Coppins probably deserved generic recognition. In this phylogeny, the Micarea sylvicola group was represented by M. bauschiana and M. sylvicola, and formed a highly supported group together with Psora decipiens (Hedw.) Hoffm. in the Psoraceae. Subsequent studies based on three or more loci and better taxon sampling confirmed that the family Psoraceae consists of Protoblastenia, Psora and the Micarea sylvicola group (Ekman et al. Reference Ekman, Andersen and Wedin2008; Ekman & Blaalid Reference Ekman and Blaalid2011; Schmull et al. Reference Schmull, Miądikowska, Pelzer, Stocker-Wörgötter, Hofstetter, Fraker, Hodkinson, Reeb, Kukwa and Lumbsch2011). In these analyses, the M. sylvicola group has been represented by M. sylvicola only (Ekman et al. Reference Ekman, Andersen and Wedin2008; Schmull et al. Reference Schmull, Miądikowska, Pelzer, Stocker-Wörgötter, Hofstetter, Fraker, Hodkinson, Reeb, Kukwa and Lumbsch2011) or M. bauschiana and M. sylvicola (Ekman & Blaalid Reference Ekman and Blaalid2011).
As no name at genus level appears to be available for Micarea sylvicola and its close relatives, a new genus is described here to accommodate it. Furthermore, although there is ample evidence to show that the M. sylvicola group belongs in the Psoraceae, previous phylogenetic analyses did not include M. lutulata and M. tuberculata. In order to demonstrate that these species are unequivocally close relatives of M. sylvicola and M. bauschiana, we present a new phylogeny that includes all four species in the same analysis.
Materials and Methods
Taxon sampling
We selected 27 species of the Pilocarpaceae in the sense of Andersen & Ekman (Reference Andersen and Ekman2005) and Ekman et al. (Reference Ekman, Andersen and Wedin2008), and Psoraceae in the sense of Andersen & Ekman (Reference Andersen and Ekman2005) and Ekman & Blaalid (Reference Ekman and Blaalid2011) (i.e. including Psora, Protoblastenia, and the ‘Micarea sylvicola group’ described here as Brianaria, but excluding Eremastrella, Glyphopeltis, and Psorula). We did not include the genus Protomicarea in the Psoraceae, as this genus was shown by Schmull et al. (Reference Schmull, Miądikowska, Pelzer, Stocker-Wörgötter, Hofstetter, Fraker, Hodkinson, Reeb, Kukwa and Lumbsch2011) not to belong in this family. Protomicarea was tentatively referred to the Psoraceae by Hafellner & Türk (Reference Hafellner and Türk2001) and Lumbsch & Huhndorf (Reference Lumbsch and Huhndorf2010) but was not included in the phylogenetic analysis by Ekman & Blaalid (Reference Ekman and Blaalid2011). Two species (Brianaria sylvicola and B. tuberculata) were represented by two terminal units, resulting in an ingroup with 29 members. We used Sphaerophorus globosus as outgroup, the selection of which was based on the phylogeny by Miądlikowska et al. (Reference Miądlikowska, Kauff, Hofstetter, Fraker, Grube, Hafellner, Reeb, Hodkinson, Kukwa and Lücking2006), in which the Sphaerophoraceae is sister to the Psoraceae and Ramalinaceae.
Marker selection and sequence acquisition
For 26 of the 30 included terminals, we downloaded sequence data from GenBank (http://www.ncbi.nlm.nih.gov/genbank/) representing three different genes, viz. the largest subunit of the RNA polymerase II gene (RPB1), the internal transcribed spacer (ITS) region (including ITS1, 5.8S, and ITS2) of the nuclear ribosomal RNA gene, and the small subunit of the mitochondrial ribosomal RNA gene (referred to here as mrSSU). For the remaining four terminals, we produced new sequence data as described below. Most terminals were composed of sequence data from a single herbarium specimen. To avoid excessive amounts of missing data in the resulting matrix, we accepted two cases with terminals composed of sequence data from more than one herbarium specimen, viz. Psora decipiens and P. rubiformis. The sequence data used for this study is summarized in Table 1.
GenBank accession numbers for DNA sequences included in this study. Newly obtained sequences are in bold. Dashes represent missing data
PCR amplification and DNA sequencing
New ITS and mrSSU sequences were generated from four specimens representing three species of Brianaria, namely B. lutulata, B. sylvicola, and B. tuberculata. In addition, several unsuccessful attempts were made to obtain RPB1 sequences. Laboratory methods generally conformed to Ekman & Blaalid (Reference Ekman and Blaalid2011), except that we used the Hot StarTaq Mastermix Kit (Qiagen) and QIAquick PCR Purification Kit (Qiagen). In a few cases, we performed direct PCR from apothecial sections without preceding DNA extraction (Wolinski et al. Reference Wolinski, Grube and Blanz1999).
Sequence alignment
Sequences were aligned using MAFFT version 6.935 (Katoh & Toh Reference Katoh and Toh2008a ). The three ITS components, ITS1, 5.8S and ITS2, were aligned separately using the X-INS-i algorithm with MXSCARNA pairwise structural alignments and Contrafold base-pairing probabilities (Katoh & Toh Reference Katoh and Toh2008b ). A structural euascomycete mrSSU reference alignment was downloaded from the Comparative RNA Web Site (http://rna.ccbb.utexas.edu; Cannone et al. Reference Cannone, Subramanian, Schnare, Collett, D'Souza, Du, Feng, Lin, Madabusi and Muller2002). This alignment was used as a profile, to which our mrSSU sequences were added using the L-INS-i algorithm. This choice was motivated by the fact that the reference alignment included complete or near-complete mrSSU sequences that were much longer than our sequences. Subsequently, the reference alignment was removed and gap-only sites stripped. RPB1 sequences were aligned using the G-INS-i algorithm at the amino acid level and subsequently back-translated into DNA sequences. The choice of algorithm was motivated by the very few expected gaps in the RPB1 sequences once introns had been removed.
Ambiguous alignments were filtered out using Aliscore version 2.0 (Misof & Misof Reference Misof and Misof2009). All pairs of taxa were used to calculate the consensus profile. Gaps were treated as ambiguities and window size was set to 4. These are the most conservative options available in Aliscore.
Phylogenetic analysis
Maximum likelihood (ML) estimation of phylogeny was performed using GARLI version 2.0 (Zwickl Reference Zwickl2006) under a single GTR model with rate heterogeneity across sites, modelled as a discretized gamma distribution with six categories and a proportion of invariable sites. Phylogeny estimates were produced for 1) each of the three genes for the purpose of assessing potential gene tree conflicts, and 2) the complete concatenated but unpartitioned data, primarily for the purpose of generating an empirical branch-length prior for downstream Bayesian inferences. For the concatenated data, we performed 1000 optimizations from starting trees generated by stepwise random addition of taxa, every possible attachment point being evaluated. For all data sets, branch support was assessed using 1000 non-parametric bootstrap replicates. Majority-rule consensus trees from the single genes were subsequently input into Compat.py (Kauff & Lutzoni Reference Kauff and Lutzoni2002) for conflict identification above 70% bootstrap support.
The concatenated data were divided into seven potential character subsets, ITS1, 5.8S, ITS2, mrSSU, as well as RPB1 first, second and third codon positions. These subsets were subsequently input into PartitionFinder version 1.0.1 (Lanfear et al. Reference Lanfear, Calcott, Ho and Guindon2012) for a simultaneous exhaustive search for the best-fitting partitioning scheme and the best-fitting model of each partition under the constraint that only models with one, two, or six substitution rate categories could be selected. We used the Bayesian Information Criterion to select among models and partitioning schemes.
We performed Bayesian phylogenetic inference using Markov chain Monte Carlo (MCMC) as implemented in PHYCAS version 1.2.0 (Lewis et al. Reference Lewis, Holder and Swofford2010). Five different analyses were performed for the purpose of quantifying the support for the Psoraceae and the new genus Brianaria under a variety of model assumptions and branch-length priors.
The first analysis assumed independent best-fitting models for each of the partitions inferred by PartitionFinder. Rate heterogeneity was, when applicable, modelled as a discretized gamma distribution with six categories. We used flat Dirichlet priors on state frequencies, as well as the substitution rate matrix for six-rate models, a beta prime (1, 1) distribution on the transition and transversion rates for two-rate models, a uniform (0·001, 200) on the gamma distribution shape parameter, a uniform (0, 1) on the proportion of invariable sites, and a flat relative rate distribution, a transformed Dirichlet distribution described by Fan et al. (Reference Fan, Wu, Chen, Kuo and Lewis2011), on the subset rate multipliers. The prior on branch lengths was set to an exponential with rate parameter 25. This rate parameter was estimated by fitting an exponential distribution to the branch lengths obtained from the ML analysis. Curve-fitting was performed with EasyFit Professional version 5.5 (MathWave Technologies).
The second analysis was identical to the first, except that the prior on branch lengths was set to an exponential seeded by an exponential hyperprior with rate 10. This is a hierarchical model on branch lengths, in which the mean of the prior distribution is not fixed but treated as a parameter to be estimated. As a hierarchical model on branch lengths does not impose a specific exponential distribution, it should have less influence on posterior parameter distributions (Ekman & Blaalid Reference Ekman and Blaalid2011; Rannala et al. Reference Rannala, Zhu and Yang2012).
The third analysis was identical to the first except we allowed trees with polytomies to be sampled according to the model of Lewis et al. (Reference Lewis, Holder and Holsinger2005). The purpose of this analysis was to investigate whether the forced sampling of fully resolved trees in other analyses could cause excessive branch support (Lewis et al. Reference Lewis, Holder and Holsinger2005). We chose to set the polytomy prior (C) to 1. Thereby every tree topology was treated as a priori equally probable, regardless of the number of internal nodes. As there are c. 1000 times more unique topologies with at least one polytomy than there are fully resolved 30-taxon trees (Felsenstein Reference Felsenstein2004), our prior amounts to treating the class of polytomous trees as a priori far more likely than the fully resolved ones.
The fourth analysis was identical to the first, except that we used a fixed branch-length prior drawn from a gamma distribution with shape 0·647 and scale 0·062. The parameterization of the gamma distribution was taken from the above-mentioned EasyFit curve-fitting procedure. The rationale behind this analysis was that a branch-length prior violating the true distribution of branch lengths may bias posterior probabilities (Kolaczkowski & Thornton Reference Kolaczkowski and Thornton2007).
The fifth analysis was identical to the first, except that independent GTR+I+Γ models were used for the subsets. This analysis was carried out to safeguard against potential overestimates of branch support, in case of hidden inadequacies in the best-fitting models (Huelsenbeck & Rannala Reference Huelsenbeck and Rannala2004). It should be noted, however, that the potential adequacy of the GTR+I+Γ model is restricted to temporally reversible and homogeneous processes.
Each analysis included three runs, each with one cold and three heated chains, the hottest with power 0·5. We ran each analysis for 200 000 generations, sampling every 100th generation. The reason for the smaller number of generations compared to the commonly used MrBayes (Ronquist & Huelsenbeck Reference Ronquist and Huelsenbeck2003) is that a generation is defined differently in PHYCAS, one generation in this software corresponding to c. 100 generations in MrBayes. Average standard deviations of splits (with frequency ≥0·1) between runs, identical to the default measure used to diagnose MrBayes runs, were calculated from summaries provided by the ‘showsplits’ command in AWTY online (Wilgenbush et al. Reference Wilgenbush, Warren and Swofford2004) after having removed the first half of each tree sample as burn-in. Marginal likelihoods of the data were calculated with Tracer version 1.5 (Rambaut & Drummond Reference Rambaut and Drummond2009) using the importance sampling estimator originally suggested by Newton & Raftery (Reference Newton and Raftery1994) and modified by Suchard et al. (Reference Suchard, Kitchen, Sinsheimer and Weiss2003). Importance sampling, as well as the widely used harmonic mean, have, however, been shown to be unreliable when comparing models with high dimensionality (Lartillot & Philippe Reference Lartillot and Philippe2006). Therefore, we also calculated marginal likelihoods using the stepping-stone procedure described by Fan et al. (Reference Fan, Wu, Chen, Kuo and Lewis2011) and implemented in PHYCAS. This implementation requires a fixed tree topology, for the purpose of which we used the majority-rule consensus trees with all compatible groups. We took 1000 samples from each of 21 stepping stones, with the exception that 2000 samples were taken from the posterior. The fixed-topology requirement rendered stepping-stone estimation impossible under the polytomy model.
Results
The resulting filtered alignment consisted of 1731 sites, 352 of which belonged to the ITS, 807 to the mrSSU, and 642 to the RPB1. The number of variable alignment positions (assuming that gaps are treated as missing data) was 166, 300, and 289, respectively in the ITS, mrSSU, and RPB1. The total amount of missing data, including gaps, was 28%, RPB1 being clearly overrepresented due to technical difficulties amplifying this gene successfully.
We did not record any conflicts between the three genes. The best-fitting partitioning scheme, given the seven potential partitions, included five partitions: ITS1 + ITS2, 5.8S, mrSSU, RPB1 first and second codon positions, and RPB1 third codon positions. The best-fitting models for each of these partitions was found to be SYM+Γ, K80+I, GTR+I+Γ, K80+I+Γ, and K80+Γ, respectively. The total number of free parameters in this model, including the subset rate multipliers, was 26, compared to the 54 free parameters in the analysis with five independent GTR+I+Γ models. Kolmogorov-Smirnov tests of goodness-of-fit of preliminary maximum-likelihood branch lengths did not reject an exponential distribution (D=0·13, P=0·30), although a gamma distribution had better fit (D=0·09, P=0·75). Average standard deviations of split frequencies between MCMC runs ranged from 0·007 to 0·009 depending on the analysis, which is low enough to conclude that MCMC analyses had converged and that our tree samples represent valid samples from the posterior distributions.
The five Bayesian analyses are summarized in Table 2. Depending on model and branch-length prior, the posterior probability of the branch uniting the Psoraceae ranges from 0·98 to 1·00, whereas the posterior probability of the branch uniting the genus Brianaria ranges from 0·95 to 0·98. By comparison, ML bootstrap proportions were 0·73 for the Psoraceae and 0·89 for Brianaria. Stepping-stone estimation of marginal likelihoods seems to provide better resolution to discriminate between models and priors than importance sampling. A majority-rule consensus tree with all compatible groups from the first of the Bayesian inferences, the one with independent best-fitting model for each partition and an exponential (25) branch-length prior, is shown in Fig. 1.
Overview of Bayesian phylogenetic analyses performed with PHYCAS. The “best” model refers to the combination of partition models selected by PartitionFinder. Polytomies were modelled according to Lewis et al. (Reference Lewis, Holder and Holsinger2005). Stepping-stone estimation of the marginal likelihood was not possible because of a fixed-topology requirement in the implementation. Support for the Psoraceae refers to the posterior probability of the node uniting Brianaria, Protoblastenia, and Psora
Majority-rule consensus tree with all compatible groups, average branch lengths, and posterior probabilities of nodes resulting from Bayesian MCMC using PHYCAS under independent best-fitting models for five partitions and an exponential branch-length prior with rate parameter 25. Familial affiliations are indicated.
Discussion
The phylogenetic analysis indicates that Brianaria, including B. bauschiana, B. lutulata, B. sylvicola and B. tuberculata, forms a monophyletic group within the Psoraceae, which is in agreement with the findings of Andersen & Ekman (Reference Andersen and Ekman2005), Ekman & Blaalid (Reference Ekman and Blaalid2011) and Schmull et al. (Reference Schmull, Miądikowska, Pelzer, Stocker-Wörgötter, Hofstetter, Fraker, Hodkinson, Reeb, Kukwa and Lumbsch2011). Evidence also suggests that Brianaria is the sister group to the rest of the currently known members of the Psoraceae, viz. Psora and Protoblastenia. Conversely, there is no indication that Brianaria belongs in Micarea or the Pilocarpaceae, where its species have previously been included.
Differences in tree lengths and support for the Psoraceae and Brianaria between analyses based on best-fitting models for each partition are minuscule, whether or not allowing for polytomies and irrespective of the particular branch-length prior. Marginal likelihood differences are modest as estimated by importance sampling, whereas the more reliable stepping-stone estimation indicates reasonably strong support for a gamma distributed branch-length prior. This is not surprising, as the initial fitting of ML branch lengths provided better fit to a gamma distribution than to an exponential distribution. The analysis based on independent GTR+I+Γ models, on the other hand, resulted in a much worse marginal likelihood than other analyses, whether estimated by importance sampling or stepping stones, suggesting that it suffers from severe overfitting. However, whereas underfitting is known to cause severe topological bias, overfitting seems to be much less of a problem (Huelsenbeck & Rannala Reference Huelsenbeck and Rannala2004; Lemmon & Moriarty Reference Lemmon and Moriarty2004). Consequently, the overfitted analysis provides some indication that support for the Psoraceae and Brianaria is not overestimated. Ekman & Blaalid (Reference Ekman and Blaalid2011), based on more characters but fewer taxa in Brianaria, found support for a Psoraceae including Brianaria to be 0·96 and 0·95 for independent best-fitting and GTR+I+Γ partition models, respectively, when integrating over a wide interval of exponential branch-length priors. They did not, however, investigate the effect of allowing for polytomies or other than exponential branch-length priors. Nominal ML bootstrap support values for the Psoraceae and Brianaria nodes (0·73 and 0·89, respectively) seem to be in line with the posterior probabilities, given previous suggestions that bootstrap support at 0·70 corresponds to a 0·95 probability of a clade being real (Hillis & Bull Reference Hillis and Bull1993). This approximation assumes that rates of change are moderate and more or less equal across the tree, which may hold true in our case. Altogether, support for the monophyly of Psoraceae and Brianaria appears to be high and does not seem to be affected by any of the known causes of inflated posterior probabilities in Bayesian inference of phylogeny, viz. model underfitting, branch-length prior misspecification, or the forcing of fully resolved trees.
Taxonomy
Brianaria S. Ekman & M. Svensson gen. nov.
MycoBank No.: MB803358
Distinguished from Micarea s. str. by having a non-micareoid photobiont, dimorphic paraphyses, and a wider tube structure in the tholus, from Psora by the crustose thallus, lack of oxalate and anthraquinone crystals in the apothecia, and from Protoblastenia by the absence of anthraquinones in the apothecia and the lack of an excipulum.
Type species: Brianaria sylvicola (Flot. ex Körb.) S. Ekman & M. Svensson.
Thallus scurfy-granular-verruculose areolate, grey-greyish green. Photobiont of two types, 1) chlorococcoid, 5–12(–15) µm or 2) irregularly ellipsoid and up to 15×10 µm.
Ascomata immarginate, convex-hemispherical, often becoming tuberculate, 0·15–0·70 (–1·20) mm diam. Excipulum absent. Hypothecium 80–200 µm tall, composed of interwoven hyphae 1–3 µm thick. Hymenium 30–75 µm. Paraphyses dimorphic, either evenly distributed, sparingly branched, often anastomosing below, 0·8–1·5 µm wide or fewer in number, single; or in fascicles, simple or occasionally forked above, distinctly septate, 1·5–3·0 µm wide, up to 4 µm apically. Ascospores non-septate (sometimes 1-septate in B. tuberculata), 5·5–12·0×1·5–5·0 µm. Asci 8-spored, cylindrical-clavate, 25–45×7–12 µm. Tholus with a wide, dark tube structure that expands towards the top, without a pale axial body (‘Psora-type’ sensu Ekman et al. Reference Ekman, Andersen and Wedin2008).
Pycnidia immersed in thallus, 0·04–0·20 mm diam., black. Conidiogenous cells ± cylindrical, 5–10×1·0–1·5 µm. Conidia bacilliform to oblong to obovoid, 3–7×1–2 µm.
Chemistry
No lichen substances detected by TLC. Three different pigments occur in the apothecia of Brianaria, viz. blue-green, K−, N+ red (‘Pigment A’ sensu Coppins Reference Coppins1983, ‘Cinereorufa-green’ sensu Meyer & Printzen Reference Meyer and Printzen2000), brown, K−, N− or N+ orange-brown (‘Pigment F’ sensu Coppins Reference Coppins1983) and purple, K+ green, N+ red (‘Pigment B’ sensu Coppins Reference Coppins1983, ‘Melaena-red’ sensu Meyer & Printzen Reference Meyer and Printzen2000).
Etymology
The genus is named in honour of Brian Coppins, in recognition of his outstanding contribution to the taxonomy of crustose lichens in general, and to the genus Micarea in particular.
Ecology
All four species of Brianaria are essentially saxicolous and prefer shaded acid rock, often in rain-protected situations. Other substrata (e.g., wood, rusted iron) are occasionally inhabited, especially by B. sylvicola.
Notes
As delimited by Lumbsch & Huhndorf (Reference Lumbsch and Huhndorf2010), the Psoraceae consists of the genera Eremastrella, Glyphopeltis, Protoblastenia, Psora, Psorula and possibly also Protomicarea. Of these, Eremastrella, Glyphopeltis, Psorula and Protomicarea have subsequently been shown not to belong in this family (Ekman & Blaalid Reference Ekman and Blaalid2011; Schmull et al. Reference Schmull, Miądikowska, Pelzer, Stocker-Wörgötter, Hofstetter, Fraker, Hodkinson, Reeb, Kukwa and Lumbsch2011). Psora, the type genus of the family, differs from Brianaria in having a well-developed squamulose thallus, anthraquinones in the epithecium and calcium oxalate in the hypothecium (Timdal Reference Timdal1984, Reference Timdal, Nash, Ryan, Gries and Bungartz2002). Protoblastenia differs in having anthraquinones in the apothecial tissues, in having an exciple composed of parallel-radiate hyphae, and in having a preference for calcareous substrata (Kainz Reference Kainz, Nash, Ryan, Diederich, Gries and Bungartz2004; Kainz & Rambold Reference Kainz and Rambold2004). Brianaria, Psora and Protoblastenia are similar in having asci of the ‘Psora-type’ and immersed pycnidia containing bacilliform conidia.
Although not closely related, Brianaria is anatomically similar to Micarea, differing from Micarea s. str. (i.e. M. prasina Fr. and closely related species) primarily by having a non-micareoid photobiont, dimorphic paraphyses, and a slightly different tholus. Although similar, tube structures in members of the Pilocarpaceae tend to be thin without or with a slight tendency to expand near the apex. This appearance contrasts with the thick tube that expands near the apex found in the Psoraceae, including Brianaria (see Fig. 1 in Ekman et al. Reference Ekman, Andersen and Wedin2008).
In spite of the exclusion of Brianaria, the genus Micarea still includes species with a non-micareoid photobiont, such as M. lynceola (Th. Fr.) Palice and M. myriocarpa V. Wirth & Vězda ex Coppins, as well as species with dimorphic paraphyses, such as M. botryoides (Nyl.) Coppins and M. lithinella (Nyl.) Hedl. (Andersen & Ekman Reference Andersen and Ekman2005; Coppins Reference Coppins, Smith, Aptroot, Coppins, Fletcher, Gilbert, James and Wolseley2009).
The photobiont of Brianaria remains unidentified to genus. The photobiont in Psora decipiens and P. globifera has been identified as Myrmecia biatorellae (Geitler Reference Geitler1963; Galun et al. Reference Galun, Ben-Shaul and Paran1971; Tschermak-Woess Reference Tschermak-Woess and Galun1988), although Schaper & Ott (Reference Schaper and Ott2003) claimed to have found a species of Asterochloris (Schaper & Ott Reference Schaper and Ott2003) in Psora decipiens. Both Myrmecia and Asterochloris are members of the Trebouxiaceae in the Trebouxiales (Guiry & Guiry Reference Guiry and Guiry2012). The primary ‘micareoid’ photobionts in Micarea prasina, M. peliocarpa, and M. misella, on the other hand, appear to be Elliptochloris bilobata, E. reniformis, and E. subsphaerica (Voytsekhovich et al. Reference Voytsekhovich, Dymytrova and Nadyeina2011). The genus Elliptochloris has an unsettled position in the Prasiolales (Guiry & Guiry Reference Guiry and Guiry2012).
Descriptions of the species of Brianaria, as well as heterotypic synonyms, can be found in Coppins (Reference Coppins1983) and Czarnota (Reference Czarnota2007).
Brianaria bauschiana (Körb.) S. Ekman & M. Svensson comb. nov.
MycoBank No.: MB803360
Biatora bauschiana Körb., Parerga lich.: 157 (1860).—Lecidea bauschiana (Körb.) Lettau in Hedwigia 55: 28 (1914).—Micarea bauschiana (Körb.) Vězda & Wirth in Folia Geobot. Phytotax. 11: 95 (1976); type: Germany, Baden-Württemberg, “auf Porphyr bei Baden,” Bausch, distributed as Rabenhorst: Lich. Europ. 648 (M—lectotype, selected by Vězda & Wirth 1976, not seen; UPS—isotype, seen).
Brianaria lutulata (Nyl.) S. Ekman & M. Svensson comb. nov.
MycoBank No.: MB803361
Lecidea lutulata Nyl. in Flora Jena 56: 297 (1853).—Micarea lutulata (Nyl.) Coppins in D. Hawksw., P. James & B. Coppins, Lichenologist 12: 107 (Reference Hawksworth, James and Coppins1980); type: British Isles, “Jersey, Rozel meadow, bases of rocks,” 1873, Larbalestier (H-NYL 10696—lectotype, selected by Coppins Reference Coppins1983, seen).
Brianaria sylvicola (Flot. ex Körb.) S. Ekman & M. Svensson comb. nov.
MycoBank No.: MB803359
Lecidea sylvicola Flot., Lich. Schles.: 171 (1829), nom. inval. (Art. 32.1d, 34.1a).—Lecidea sylvicola Flot. ex Körb., Syst. Lich. German.: 254 (1855).—Micarea sylvicola (Körb.) Vězda & Wirth in Folia Geobot. Phytotax. 11: 99 (1976); type: Czech Republic/Germany/Poland, Lich. Schles. 171 (UPS—lectotype, selected by Hertel Reference Hertel1975, seen).
Nomenclatural note
Flotow (Reference Flotow1829) introduced the name Lecidea sylvicola, but did not himself consider this name valid (‘Lecidea sylvicola ad int.’) and did not provide a diagnosis or reference to a validly published diagnosis. The taxon was validated by Körber (Reference Körber1855), who provided a diagnosis and made explicit reference to nr. 171 in Flotow's exsiccate. The earliest known synonyms of L. sylvicola are L. aggerata Mudd and L. incincta Nyl., both of which were published in 1861 (Coppins Reference Coppins1983; Czarnota Reference Czarnota2007).
Brianaria tuberculata (Sommerf.) S. Ekman & M. Svensson comb.nov.
MycoBank No.: MB803362
Lecidea tuberculata Sommerf., Suppl. Fl. Lapp.: 160 (1826).—Micarea tuberculata (Sommerf.) R. A. Anderson in Bryologist 77: 46 (1974); type: Norway, Nordland, Saltdalen, Fiskevaagmöllen, March 1822, Sommerfelt (O—lectotype, selected by Coppins Reference Coppins1983, not seen; UPS—isotype, seen).
We thank Katja Fedrowitz, Mattias Lif and Veera Tuovinen for laboratory assistance.
