Rewriting the evolutionary history of the lichen genus Sticta (Ascomycota: Peltigeraceae subfam. Lobarioideae) in the Hawaiian islands

Hawaiian lichen species have been thought to be widespread, with low endemism. Nine species of the genus Sticta (Peltigeraceae subfamily Lobarioideae) have previously been reported for Hawaii, all supposedly cosmopolitan or Pantropical or widespread in the Paleotropics except for the putative endemic S. plumbicolor. This study is the first one employing a molecular phylogenetic approach to Hawaiian Sticta, elucidating the relationships of these conspicuous and ecologically important macrolichens. We sequenced the ITS fungal barcoding locus and used a maximum likelihood approach to reconstruct phylogenetic relationships of Hawaiian Sticta from a large dataset of more than 200 species. Thirteen species were identified among Hawaiian Sticta, four more than previously recorded. Of these, seven are new to science and putatively endemic to Hawaii. Only four previously reported species were confirmed: S. fuliginosa, S. limbata, S. plumbicolor and S. tomentosa. Together with S. plumbicolor and S. scabrosa subsp. hawaiiensis (described elsewhere), putative endemism in Hawaiian Sticta is estimated at 69%. The 13 species correspond to nine or ten colonization events, predominantly from the Australasian realm. Thus, the evolutionary history of Sticta lichens in the Hawaiian archipelago is very different from what has been assumed, and matches that of other organisms in many aspects. The seven new species, all with cyanobacterial photobionts, are Sticta acyphellata, a small, stipitate Sticta with isidia and lacking cyphellae; S. antoniana, a mid-sized Sticta with abundant marginal lobules, apothecia, and a thick, grey-brown lower tomentum ending abruptly to leave a bare marginal zone; S. emmanueliana, a small, shortly stipitate Sticta forming small lobes with marginal isidia and black cilia; S. flynnii, a small, shortly stipitate Sticta with largely unbranched thallus with marginal isidia and a veined underside producing large, irregular cyphellae; S. hawaiiensis, a small Sticta with a suborbicular thallus with laminal isidia, conspicuous white cilia, and papillae on the membrane of the cyphellae; S. smithii, a small, stipitate Sticta with marginal, flattened isidia and small cyphellae; and S. waikamoi, a small to mid-sized Sticta with a much-branched thallus with slightly canaliculate lobes and marginal, dark isidia, and a thick, dark brown lower tomentum with strongly contrasting whitish cyphellae.

Sticta is one of the best-known lichen taxa, due to its usually large, conspicuous thalli. It is easily recognized by its distinctive pores on the lower side, the cyphellae, which facilitate gas exchange (Green et al. 1981;Galloway 2007), although a small group of species centered around S. wrightii was shown to form a separate clade and was segregated as Dendriscosticta (Moncada et al. 2013a). Species of Sticta are characteristic of humid, coolto warm-temperate environments with high precipitation or humidity. At tropical latitudes they are most diverse and abundant in wet montane forests and alpine grasslands (páramos in the Neotropics), but are also found occasionally in lowland rain forests. Given that under favorable conditions they can accumulate large biomass and that most species are associated with nitrogen-fixing cyanobacteria, species of Sticta are significant contributors to local nitrogen input as biological fertilizers (Kelly & Becker 1975;Becker 1980;Green et al. 1980;Green & Lange 1991;Antoine 2004;Benner et al. 2007;Benner & Vitousek 2012). Equally important is their role in the water cycle, since their thalli are capable of storing up to 800% of their dry weight in water and influence the microclimate of their immediate environment (Green et al. 1985;Guzmán et al. 1990;Green & Lange 1991;Beckett 1995;Zotz et al. 1998). Species of Sticta are highly sensitive to human-induced environmental changes and pollution, and have become extinct in many areas of North America and Europe (Wirth 1995;Brodo et al. 2001;Pišút 2005;Hodkinson et al. 2014;Magain & Sérusiaux 2015;Lendemer & Goffinet 2016;Simon et al. 2018a;Ekman et al. 2019). This sensitivity makes them excellent indicators of environmental health, including the effects of habitat disturbance and global climate change (Scheidegger et al. 1995;Zoller et al. 1999;Radies et al. 2009).
Checklists of Hawaiian lichens (Elix & McCarthy 1998, 2008Smith 2013) list 11 names under Sticta, with one, S. crocatoides [sic] f. sandwicensis, an orthographic error for S. crocata f. sandwicensis, actually representing Pseudocyphellaria sandwicensis (Moncada et al. 2014b;Lücking et al. 2017b). The remaining ten names correspond to seven species, viz. Sticta ambavillaria, S. cyphellulata, S. filix, S. fuliginosa, S. plumbicolor, S. tomentosa and S. weigelii, plus three varieties of the latter: S. weigelii var. beauvoisii, S. weigelii var. lutescens and S. weigelii var. peruviana. Sticta weigelii var. beauvoisii was recently accepted as a separate species (McDonald et al. 2003;Galloway 2006). The application of the name S. lutescens is unclear; Zahlbruckner (1925) gives S. lutescens as synonym of S. xanthosticta, which in turn is a synonym of Pseudocyphellaria crocata (Galloway 2001), but the lectotype of S. lutescens, described from Indonesia (Java), is a Sticta species. Sticta weigelii var. peruviana was described as S. sylvatica var. peruviana from Peru (Delise 1825) and subsequently considered a separate species (Nylander 1859). This taxon has been misunderstood, since the type material lacks isidia and bears abundant apothecia instead (Moncada 2012). One additional species, S. limbata, was identified in an ecological study (Benner & Vitousek 2012) but has not been included in the checklist (Smith 2013). Since identifications of Hawaiian material as S. weigelii var. lutescens and var. peruviana are likely incorrect and may refer to specimens of S. weigelii s.lat., we assume that the eleven names of Sticta reported for the archipelago (excluding S. crocatoides var. sandwicensis) represent nine species, with eight of them widespread and only S. plumbicolor (Zahlbruckner 1903(Zahlbruckner , 1925 putatively endemic to Hawaii, resulting in an inferred endemism of 11%. Of the nine species reported for Hawaii, one is a green algal taxon (S. filix). The others are cyanolichens, two exclusively apotheciate (S. ambavillaria, S. tomentosa), one sorediate (S. limbata), and five isidiate or phyllidiate (S. beauvoisii, S. cyphellulata, S. fuliginosa, S. plumbicolor, S. weigelii). Based on published records, the eight non-endemic species presumably represent six different distribution patterns: S. filix is a chiefly Australasian species (described from New Zealand; Galloway 2007;Ranft et al. 2018); S. cyphellulata is Australasian (described from Australia) but also occurs in India and the Mascarenes (Galloway 1998(Galloway , 2008Magain & Sérusiaux 2015;Simon et al. 2018b); S. ambavillaria (originally from Africa) has a Gondwanan distribution (Neotropics and African Paleotropics; Marcano et al. 1996;Bock et al. 2007; van den Boom et al. 2011;Eliasaro et al. 2012;Aptroot 2016;Simon et al. 2018b); S. beauvoisii (established from material from Cuba) is predominantly North American (Galloway 1995(Galloway , 2006McDonald et al. 2003); S. tomentosa (originally from Jamaica) has a Pantropical range (van den Boom et al. 2011;Moncada et al. 2014a); and S. fuliginosa, S. limbata (both described from Great Britain) and S. weigelii (described from Martinique) are subcosmopolitan (Galloway 2006;Moncada et al. 2014a;Magain & Sérusiaux 2015). This would paint an undifferentiated, diffuse picture of the assembly of Hawaiian Sticta, apparently with stochastic colonization events from both the Eastern (Australasia) and the Western Hemisphere (North America) and almost no separate evolutionary history on the archipelago.
A recent molecular phylogenetic revision of Sticta focusing on the northern Andes revealed that the morphological concept traditionally applied to delimit species in Sticta is ill-defined and that names such as S. fuliginosa and S. weigelii, assumed to represent cosmopolitan taxa, correspond to several distantly related lineages (Moncada et al. 2014a). Hence, we suspected that the names listed for Hawaii do not accurately reflect the diversity and taxonomic affinities of the species occurring there. This assumption is supported by two previous studies of Hawaiian lobarioid Peltigeraceae. Specifically, the seven presumably widespread species of Pseudocyphellaria s.lat. reported for Hawaii (Smith 2013) turned out to represent 12 species in three genera (Crocodia, Podostictina, Pseudocyphellaria), nine of them putatively endemic to the archipelago (Moncada et al. 2014b); for the genus Lobariella, Lücking et al. (2017c) found that instead of a single widespread species reported for Hawaii (Smith 2013), at least four taxa are present, three of them putatively endemic. In both cases, putative endemism had to be corrected from the previous 0% to 75%, close to the level characterizing vascular plants.
Since Sticta was the only genus in this subfamily with at least one presumably endemic species in Hawaii (S. plumbicolor), we expected that a molecular phylogenetic revision would reveal a degree of endemism even higher than for the other two groups. To test this hypothesis we collected fresh samples of Sticta from 14 sites on the four largest islands (Kauai, Oahu, Maui, the Big Island) and extracted DNA for molecular phylogenetic analysis. Based on the results, it is evident that the evolutionary history of Sticta in the Hawaiian archipelago must indeed be rewritten.

Morphological study
Herbarium material of Sticta originating from Hawaii was revised at the US National Herbarium and at the herbarium of the University of Hawaii at Manoa (HAW). We performed thin-layer chromatography to determine the chemical constituents of selected samples, following the methods outlined in Orange et al. (2010). New fresh material was collected during field work in June 2013 on the islands of Oahu, Maui and Kauai (Fig. 1), in collaboration with Clifford Smith and Philip Thomas (University of Hawaii at Manoa), Pat Bily (The Nature Conservancy, Maui) and Daniel Kaniaupio-Crozier (West Maui Soil & Water Conservation District). In addition, we received some specimens collected by C. Smith on the Big Island (Hawaii). Specimens were initially identified using keys provided by Swinscow & Krog (1988), Joshi & Awasthi (1982) and Galloway (1994bGalloway ( , 2001Galloway ( , 2007 for southern South America, West Africa, India and Nepal, Australia and New Zealand. For alternative taxonomic scenarios, we then compared these identifications with the older concept provided by Magnusson and Zahlbruckner (1943) and the refined concept from Moncada et al. (2014a).

Molecular phylogeny
We obtained new sequences of the nuclear ITS barcoding locus for 149 samples, including 111 from Hawaii. Molecular work was performed at the Field Museum's Pritzker Laboratory for Molecular Systematics and Evolution. DNA was extracted using the QIAGEN DNeasy Plant Mini Kit. Dilutions of 10:1 up to 10:2 were used for PCR amplifications, with the primer pairs ITS1F and ITS4 (Gardes & Bruns 1993;White et al. 1990). The 25 µl PCR reactions contained 2.5 µl buffer, 2.5 µl dNTP mix, 1 µl of each primer (10 µM), 5 µl BSA, 2 µl Taq, 2 µl genomic DNA extract and 9 µl distilled water. The thermal cycling parameters were set as follows: initial denaturation for 3 min at 95°C, followed by 30 cycles of 1 min at 95°C, 1 min at 52°C, 1 min at 73 °C, and final elongation for 7 min at 73°C. Amplification products were mounted on 1% agarose gels stained with ethidium bromide and, after cutting of the target bands, purified using the QIAGEN QIAquick PCR Purification Kit or Nucleo Spin DNA Purification Kit (Macherey-Nagel). Fragments were sequenced using the Big Dye Terminator reaction kit (ABI PRISM, Applied Biosystems). Sequencing and PCR amplifications were performed using the same sets of primers. Cycle sequencing was executed with the following setting: 25 cycles of 95°C for 30 sec, 48°C for 15 sec, 60°C for 4 min. Sequenced products were precipitated with 10 µl of sterile dH2O, 2 µl of 3 M Napa, and 50 µl of 95% EtOH, and subsequently loaded on an ABI 3100 (Applied Biosystems) automatic sequencer. Sequence fragments obtained were assembled with DNASTAR SeqMan 4.03, manually inspected and adjusted and, after quality control, submitted to GenBank (Table 1).
The obtained sequences were aligned with previously published sequences of the genus Sticta (Moncada et al. 2014a;Magain & Sérusiaux 2015;Simon et al. 2018b;Widhelm et al. 2018;Mercado-Díaz et al. 2020; Table S1). Sequences were assembled in BIOEDIT 7.0.9 (Hall 2011) and aligned with MAFFT 7.164 using the '--add' option (Katoh & Frith 2012;Katoh & Standley 2013), using the '--auto' option and with subsequent manual inspection. The alignment was subjected to analysis of ambiguously aligned regions using the GUIDANCE webserver (Penn et al. 2010a(Penn et al. , 2010b, but based on the results, no columns were removed, resulting in an alignment length of 684 bases (File S1). Phylogenetic analysis was performed using maximum likelihood in RAxML 8.2.0 (Stamatakis 2014) on the CIPRES Science Gateway (Miller et al. 2010), with non-parametric bootstrapping using 402 pseudoreplicates (based on an automated saturation criterion) under the GTRGAMMA model. Trees were visualized in FIGTREE 1.4.2 (Drummond & Rambaut 2007).

Time-calibrated tree
Using a subset of 190 ingroup sequences representing different species each, a relaxed, uncorrelated lognormal molecular clock model was employed to date the evolutionary origin of the Hawaiian colonizations by Sticta lineages. BEAST 1.7.5 and 1.8.4 (Drummond & Rambaut 2007;Drummond et al. 2012) were used for this purpose, locally and on the CIPRES Science Gateway (Miller et al. 2010), with the GTR substitution model with base frequencies estimated and Gamma and invariant sites with four Gamma categories. Speciation was estimated through a Yule process with the 'yule.birthRate' prior set to an exponential distribution with 0.7 as mean. Based on Simon et al. (2018b) and Widhelm et al. (2018), we set treeModel.rootHeight (Lobaria-Sticta split) to 45 Mya and the Sticta crown node to 25 Mya, both with a normal distribution and 5 and 3 my standard deviation, respectively. The final analysis was run for 10 million generations and every 1,000th tree was sampled, for a total of 10,001 trees and a burn-in of 2,500 trees. The resulting log file was analysed in TRACER 1.5 and the maximum clade credibility tree was compiled using TREE ANNOTATOR 1.7.5 (Drummond & Rambaut 2007).

Phylogenetic relationships
The ITS-based phylogeny binned the Hawaiian Sticta samples into ten clades, of which two were closely related and eight were not closely related and were dispersed across the tree ( Fig. 2; Fig. S1). The ten clades exhibited three different topological patterns: (1) six clades corresponded to distinct lineages representing exclusively Hawaiian specimens; (2) two clades were more widely distributed in the tropics but the Hawaiian material formed (near-)exclusive shallow subclades; and (3) two clades were more widely distributed or even subcosmopolitan and the Hawaiian specimens were dispersed within these clades, with no phylogenetic distinction from non-Hawaiian samples.

Clade-based taxonomic assessment
Morphological analysis revealed that the ten clades corresponded to 13 distinct phenotypes. The six distinct, exclusively Hawaiian lineages were each morphologically homogeneous; none of these correspond to known species and they are therefore considered undescribed taxa new to science. All are marginally isidiate and most are caulescent, but they differ in thallus size and particularly in the development and color of the lower tomentum. The early-diverging lineages 1 to 4, below described as S. acyphellata, S. emmanueliana, S. flynnii and S. smithii, represent diminutive species which in most cases do not correspond to previously reported names (e.g., Magnusson & Zahlbruckner 1943;Magnusson 1955), except for lineage 4, which has been reported as S. cyphellulata (Elix & McCarthy 1998, 2008. These taxa have apparently been overlooked in previous surveys or were mistaken for young forms of other species. For instance, of 175 specimens of Sticta in HAW, only two corresponded to taxa representing these lineages; previously these had been identified as S. plumbicolor and S. weigelii, respectively. The other two lineages would correspond morphologically to either S. fuliginosa s.lat. (lineage 5) or S. weigelii s.lat. (lineage 6); in HAW, two specimens of lineage 5 had been identified as S. fuliginosa and three specimens of lineage 6 as S. weigelii. However, lineage 5 belongs in a complex centered around the recently reinstated S. ciliata and within that complex is most closely related to the recently described S. parvilobata from Puerto Rico.
The two clades containing shallow, nested subclades correspond to two more widely distributed taxa, S. scabrosa and S. tomentosa. Sticta scabrosa is one of the few species of the genus typically found at lower elevations and in more exposed situations, tolerant to disturbances, and sometimes even with a weedy aspect. The Hawaiian specimens differ consistently in two distant positions of the ITS and exhibit a tendency towards scrobiculate to foveolate (pitted) lobe tips; these are considered a subspecies (subsp. hawaiiensis; Moncada et al., unpublished). In their ecology they are similar to the nominal subspecies, which is widely distributed in the Neotropics. In Hawaii, this is by far the most frequent and widespread Sticta: 59 of the 127 sequenced specimens (46%) and one third of the collections in HAW correspond to this taxon. By contrast, the Hawaiian material of S. tomentosa differs consistently from Neotropical and Paleotropical specimens only in one ITS position, with one specimen from Ecuador even corresponding to the Hawaiian haplotype; therefore, given the absence of any morphological differences, we consider this material to represent S. tomentosa s.str. However, the Hawaiian material clustering phylogenetically with S. tomentosa contains two additional, very distinct morphotypes. One agrees with S. tomentosa in overall morphology, including the characteristic bluish hue, the whitish underside with thin tomentum, and the apothecia soon becoming biatorine, but features abundant laminal and marginal phyllidia and also differs in the branching pattern of the lobes, which are narrower than in typical S. tomentosa and often show shallow constrictions. This material matches the type of S. plumbicolor, which had been considered the only Hawaiian endemic species in this genus. The second morphotype strongly deviates from S. tomentosa in the strongly ascending, distinctly brownish (no bluish hue) lobes with rather thick and darker lower tomentum, with only the lobe tips abruptly bare of tomentum, the strongly dissected, lacinulate-lobulate lobe tips, and the apothecia remaining zeorine. This morphology corresponds perfectly to what Zahlbruckner (in Magnusson & Zahlbruckner 1943) had identified with the name S. ambavillaria and is here described as S. antoniana (see below).
Finally, the two widespread clades with Hawaiian material nested or dispersed within correspond to three distinct morphotypes. One clade is currently being described as S. andina, with an S. weigelii morphology but unrelated to the latter (Moncada et al., unpublished). Sticta andina differs from S. weigelii in the much thicker lower tomentum, It is more common in high-altitude forest and shrubland, and is the most frequently encountered species in the Andean páramos. The Hawaiian material represents a single haplotype that is identical to specimens from the Northern Andes and from the Azores. The second clade represents the widespread taxa S. fuliginosa and S. limbata. Both are morphologically quite distinct, producing laminal isidia vs. marginal soredia, but based on ITS data cannot be resolved, a phenomenon that has not yet been clarified (Moncada et al. 2014a;Magain & Sérusiaux 2015). Both taxa are present in Hawaii and correspond morphologically to specimens from other regions. No green algal Sticta species was collected despite intensive searching. The report of S. filix by Tuckerman (1866) is likely erroneous, as already suggested by Magnusson & Zahlbruckner (1943). As against the nine species previously reported from Hawaii, with one presumably endemic (11%), we found 13 taxa in the studied material, corresponding to ten phylogenetic lineages and three additional, distinct morphotypes that are not resolved with ITS. While the case of S. fuliginosa vs. S. limbata is a well-known phenomenon of two subcosmopolitan species (Moncada et al. 2014a;Magain & Sérusiaux 2015), the other two cases are putative endemics in the S. tomentosa clade and are similar to the case reported for the putative Hawaiian endemics Pseudocyphellaria philipiana and P. pomaikaiana, each exhibiting distinct phenotypes but sharing identical ITS sequences (Moncada et al. 2014b). Since there is no phylogenetic support in the ITS fungal barcoding for these phenotypes to be recognized as species, two alternative scenarios are possible: (1) the phenotypes represent habitat-induced morphological plasticity of identical genotypes, as known for instance from photosymbiodemes; (2) the phenotypes represent distinct lineages but the ITS is not resolved to reflect this. The first scenario is not likely, since the taxa in question co-occur in the same habitats and often even side-by-side and the variation is discrete, not gradual. Support for the second scenario comes from the well-documented case of S. fuliginosa vs. S. limbata, which cannot be separated based on ITS but cannot be considered morphodemes of the same species, as they represent highly distinctive forms of vegetative reproduction (laminal isidia vs. marginal soredia). High  morphological divergence going along with very similar genotypes has also been found in Hawaiian vascular plants, such as the Silversword alliance in the Asteraceae (Baldwin et al. 1991;Baldwin & Sandersson 1998;Carlquist et al. 2003).

Usefulness of the ITS barcoding marker for species delimitation
The available data from Hawaiian Sticta allow further assessment of the usefulness of the fungal ITS barcoding marker for species delimitation in this genus. Multi-marker studies have previously shown that ITS phylogenies are highly congruent with those of other markers in this genus (Magain & Sérusiaux 2015;Simon et al. 2018b;Widhelm et al. 2018), with no evidence of artifacts. This is also supported by the strong correlation between ITS-delimited species-level clades and their phenotypes (Moncada et al. 2013b(Moncada et al. , c, 2015Magain & Sérusiaux 2015). Thus, given the large amount of data available, there is little evidence for potential type I errors when using ITS for species delimitation in this genus. On the other hand, a few well-documented cases suggest the possibility of type II errors, that is, species not resolved by ITS alone, for example S. fuliginosa vs. S. limbata, S. filix vs. S. lacera, or in the present case also the three discrete morphodemes found in Hawaiian S. tomentosa s.lat. (Magain & Sérusiaux 2015;Ranft et al. 2018; this paper). This underlines the point that the use of ITS in revising species taxonomy in the genus Sticta can considerably improve previous treatments based on morphology alone. It should also be noted that due to the broad coverage of the current ITS phylogeny of the genus, branches appeared relatively compressed even if corresponding to substantial differences (Fig. S1). For the species recognized below, with the exception of the S. ciliata and S. tomentosa complexes, similarity (based on ITS substitutions and indels combined) to the most closely related taxa oscillated between 94.0% and 96.1% (Table 2), distinctly below the lowest fixed threshold level of 98.5% applied as default for species hypotheses (Nilsson et al. 2019). Given the critical situation of conservation of native biodiversity in island biota such as Hawaii, it also seems prudent to be more discriminant in recognizing lineages as taxa, following the concept of recognizing evolutionarily significant units (Casacci et al. 2014;Cornejo et al. 2017).

Biogeography and evolutionary history
Among the 13 taxa of Sticta now known from Hawaii, four are demonstrably widespread: S. andina, S. fuliginosa, S. limbata and S. tomentosa. The other nine are putative endemics, representing eight species and one subspecies. Notably, the only previously considered endemic species, S. plumbicolor, is phylogenetically unresolved from S. tomentosa based on ITS, although it displays a distinctive morphology not known from collections of S. tomentosa found elsewhere. Overall, this results in a degree of 69% putative endemism, comparable to the 75% found in Lobariella and in Pseudocyphellaria s.lat. but overall slightly lower (Moncada et al. 2014b;Lücking et al. 2017c). Of the nine names previously recorded for Hawaii, only four could be confirmed: S. fuliginosa, S. limbata, S. plumbicolor and S. tomentosa. Thus, between the prior taxonomy and the taxonomy to be implemented based on the results, the Sørensen overlap is only 36%, which is higher than for Pseudocyphellaria s.lat. (11%; Moncada et al. 2014b) but slightly lower than for Lobariella (40%; Lücking et al. 2017c). Our findings thus challenge the assumption that Hawaiian lichens are widespread. In addition, lichen diversity in Hawaii appears to be substantially underestimated; if the results from the studies of lobarioid Peltigeraceae are any indication, the current number of 880 listed species represents only around 50% of the real diversity, according to which Hawaiian lichen diversity would surpass that of vascular plants (Imada et al. 2012). Overall, given the significantly increased degree of putative endemism and the substantial changes in the taxonomy of Hawaiian Sticta, the evolutionary history of this genus in the archipelago Table 2. Quantitative analysis of ITS differences and similarity between newly recognized Hawaiian species of Sticta and their closest relatives. Length = alignment length (alignment reduced to corresponding target clade); % = similarity based on number of substitutions or indels (and total) relative to alignment length. Total similarity values below 98.5% are bolded and underlined.

Species
Species is very different from what had previously been assumed (Magnusson & Zahlbruckner 1943;Magnusson 1955;Smith 2013). Thus far, the lichen biota on islands has been investigated using molecular approaches in only a few cases. For the genus Sticta, the most important studies are those of Simon et al. (2018b) for Madagascar and the Mascarenes and Mercado-Díaz et al. (2020) for Puerto Rico. Particularly the first offers some basis for comparison, given that the evolutionary history of the genus Sticta in the Hawaiian archipelagos is strikingly different. The diversity of Hawaiian Sticta is the result of multiple independent colonization events, with almost no evidence of radiation or post-colonization speciation except perhaps for the closely related S. flynnii and S. smithii and the three phylogenetically unresolved species in the S. tomentosa complex. In contrast, in Madagascar and the Mascarenes a subclade of Sticta underwent a radiation giving rise to at least 31 species following a single colonization event (Simon et al. 2018b). Regarding other lobarioid taxa in Hawaii, the patterns in Pseudocyphellaria (Moncada et al. 2014b) are comparable to that of Hawaiian Sticta, whereas that of Lobariella reflects a potential microradiation, which, however, yielded only four species (Lücking et al. 2017c).
Under the assumption of a molecular clock, our inferences for Hawaiian Sticta suggest recent, multiple independent colonization events, mostly between 1 Mya and 2.5 Mya, and in one case (S. acyphellata) 6 Mya (Fig. 3). For comparison, the Lobariella microradiation in Hawaii was estimated, also based on ITS, at (1-)8 Mya (Lücking et al. (2017c). In contrast, the substantial Madagascar-Masquarenes radiation was dated in our analysis to between 13 Mya (crown node) and 17 Mya (stem node), between three and 15 times older than any of the inferred Hawaiian colonizations. Given that lichen fungi clearly do have a potential for insular radiations comparable to that of vascular plants (Simon et al. 2018b), the absence of such radiation in Hawaiian Sticta is therefore attributable to recent colonization of the archipelago by this genus.
While molecular clock approaches have to be considered with caution, particularly when using the length-variable ITS marker, a comparison with other time trees of the genus Sticta supports our results showing a substantial relative difference between the colonization of Hawaii and that of Madagascar and the Masquarenes. With three markers (mtSSU, nuLSU, RPB1) but with reduced ingroup sampling, Simon et al. (2018b) computed the Sticta stem node at 42 Mya, the crown node at 16 Mya, and the stem node of the Madagascar-Masquarenes radiation at 12 Mya, about 30% younger than our ITS-based analysis. Hawaiian lineages were not included. With the Lobaria-Sticta split recovered at 46 Mya and the Sticta crown node at 27 Mya, Widhelm et al. (2018), adding two further markers (MCM7, ITS), estimated the stem and crown nodes of the Madagascar-Masquarenes radiation at 27 Mya and 15 Mya, respectively. Their crown age estimate was close to ours, whereas their stem age suggested a much older divergence of this clade. The latter is explained by a deviating underlying topology, different from other published phylogenies, with the Madagascar-Masquarenes radiation sister to the remainder of Sticta. In the same analysis, the divergences of the included Hawaiian lineages were estimated at 5-10 Mya, about 1.5-2 times (crown node) and 2.7-5.5 times (stem node) younger than the Madagascar-Masquarenes radiation. Thus, independent of the markers employed and the underlying topology, the finding that Hawaiian colonization by the genus Sticta was much more recent than in the case of Madagascar and the Masquarenes remains.
Besides an obvious correlation between time of colonization and the potential for subsequent radiation, even in lineages that did not radiate, the phylogenetic distinctiveness of putative endemics among Hawaiian Sticta appears to be a function of time. The nine putative endemic taxa can be divided into two groups: S. acyphellata, S. emmanueliana, S. flynnii and S. smithii are taxonomically well-distinguished, whereas the other five taxa are either morphologically cryptic (S. hawaiiensis, S. scabrosa subsp. hawaiiensis, S. waikamoi) or cannot be resolved based on the ITS (S. antoniana, S. plumbicolor). The first four species exhibit inferred mean divergence times of 2-6 Mya, the other five only 1-2 Mya (Fig. 3). Thus, the degree of phylogenetic and phenotypic distinctiveness of the putative endemics clearly correlates with time of isolation. In addition, the first four species all emerged from basally diverging clades in the genus Sticta, with Australasian affinities, whereas the other five taxa are placed in a large clade corresponding to a later-diverging subclade and exhibit affinities with Neotropical lineages (Fig. 2). This indicates two distinct periods of Hawaiian colonization by the genus Sticta: according to our timetree, one chiefly from Australasia in the Pliocene and another chiefly from South America in the Pleistocene. By contrast, the Madagascan-Masquarene radiation goes back to the Miocene (Simon et al. 2018b;this paper).
The notion that the absence of radiation in Hawaiian Sticta may be due to colonization events more recent than for Madagascar and the Masquarenes is consistent with findings for vascular plants. For instance, Hawaiian lobeliads in the family Campanulaceae radiated into six genera comprising 126 species, following a single colonization event about 13 Mya in the Miocene (Givnish et al. 2009), comparable to the Madagascan-Masquarene radiation in Sticta. While Madagascar is a continental island and Mauritius and Réunion are comparatively close (less than 900 km apart), the Hawaiian archipelago is nearly 4,000 km distant from North America and over 6,000 km from eastern Asia and Australasia (Wagner & Funk 1995;Fleischer et al. 1998). Why were vascular plants able to colonize Hawaii much earlier than lichens in the genus Sticta? Besides the stochastic nature of dispersal history, one reason could be that epiphytic macrolichens in Peltigeraceae subfam. Lobarioidaeae thrive in more or less undisturbed forest, so their successful establishment depends on the prior formation of such ecosystems after initial colonization by vascular plants. By contrast, forest ecosystems on a continental island such as Madagascar were present long before the geological events that led to its isolation, so initial colonization by macrolichen lineages could have led to rapid radiation. The multiple colonization of the genus Sticta in Hawaii offers some statistical basis to assess the biogeographic affinities of Hawaiian lineages, as briefly outlined above. In the present case, five of the ten lineages and 13 species have Australasian affinities, including Hawaiian specimens of S. fuliginosa, which appear to be closest to those studied from New Zealand (Fig. S1). On the other hand, Hawaiian Sticta limbata matches specimens from Western Europe rather than from New Zealand. Three lineages and taxa appear to have New World affinities (S. andina, S. scabrosa, S. waikamoi), whereas two additional lineages corresponding to four species (S. hawaiiensis and the S. tomentosa complex) remain unresolved based on the geography of their closest relatives, but in our taxon sampling are also closely related to Neotropical taxa. Thus, the Hawaiian Sticta biota exhibits a slight prevalence of Australasian elements. This is comparable to the genus Pseudocyphellaria (Moncada et al. 2014b) and also agrees with the origin of the majority of vascular plant lineages. For instance, dominant forest trees of the genera Acacia (Fabaceae), Cheirodendron (Araliaceae) and Metrosideros (Myrtaceae) have Indopacific-Australasian relationships (Mueller-Dombois 1987;Wright et al. 2001;Percy et al. 2008;Brown et al. 2012;Mitchell et al. 2012), presumably due to the northern subtropical jet stream as predominant dispersal agent (Geiger et al. 2007). Biogeographic relationships with North, Central and South America have been detected in some plant groups, such as the Hawaiian Silverswords, whose closest relatives are the North American Tarweeds (Baldwin et al. 1991;Barrier et al. 2001;Carlquist et al. 2003). The only available biogeographical study of Hawaiian lichens found that most Cladoniaceae have Australasian relationships (Stenroos & Smith 1993), whereas Neotropical affinities are apparent in the genus Lobariella (Yoshimura 1984(Yoshimura , 1998Yoshimura & Arvidsson 1994;Lücking et al. 2017c). Thus, our results in the genus Sticta fit findings for other organisms and also support the notion of a complex evolutionary history of the Hawaiian biota.
Sticta offers another example of how an island biota can evolve unique phenotypes, with the first species known in the genus to entirely lack cyphellae, S. acyphellata. A comparable situation was found in the Hawaiian Phaeophyscia laciniata (Esslinger 1978), and in the presumably endemic genus Ramalinopsis (Follmann 1974) which represents a foliose species nested within the fruticose genus Ramalina (Kistenich et al. 2018). Hawaiian Lobariella revealed a novel, unique, near-fruticose phenotype in L. flynniana (Lücking et al. 2017c). The phenomenon of developing novel morphotypes on islands has been well documented for vascular plants (Fosberg 1936;Pax et al. 1997;Kidd 2005). Thus, in general it appears that Sticta and other lichen-forming fungi do not behave differently from vascular plants or other organisms in terms of their evolutionary history but match them in aspects such as endemism, diversification (related to age of origin), and the evolution of peculiar phenotypes. This has also been found for other groups of organisms such as bryophytes, which traditionally have been thought to exhibit evolutionary stasis (Medina et al. 2018).
Etymology. The epithet refers to the lack of cyphellae, a unique feature within the genus. Distribution and ecology. Notably, this unique species was only found in a partially disturbed secondary forest on the densely populated and strongly altered island of Oahu, where it grew epiphytic in the shaded understory near the base and on the roots of tree trunks, between bryophytes.
Notes. Sticta acyphellata is a unique species, given its diminutive size and the complete lack of cyphellae. It is the first species in the genus known with this feature. In a strictly morphological sense it would have to be classified as Lobaria s.lat., but the molecular data place it clearly within Sticta s.str. It falls near the base of the tree (Fig. 2; Fig. S1) near other cyanobacterial, stipitate species such as S. caulescens, S. gaudichaudii, S. hypochra, S. gracilis and S. cyphellulata, all from the Southern Hemisphere and/or Australasia (Galloway 1994b(Galloway , 1998(Galloway , 2001(Galloway , 2007. Two other putative Hawaiian endemics, S. flynnii and S. smithii (see below), also belong in this clade, suggesting a previously unrecognized level of diversification in this group of species and also indicating the Southern Hemisphere and Australasia as major sources of colonization of the Hawaiian archipelago.
Etymology. The species name honors the legacy of Anton Zahlbruckner, for his invaluable contributions to lichenology, including Hawaiian lichens, a work he did not see completed before his death (Magnusson & Zahlbruckner 1943), and who first noted the distinctiveness of this species.
Distribution and ecology. Sticta antoniana was found on the islands of Maui and Kauai, in both cases in more or less undisturbed montane forest at mid elevations (between 1000 and 1500 m), on shaded tree bark.
Etymology. We are delighted to dedicate this new species to our colleague and friend, Emmanuël Sérusiaux, on the occasion of his official retirement from formal duties but certainly not from lichenology. Emmanuël has made numerous invaluable contributions to lichenology in almost all taxonomic groups and geographic areas, but especially in tropical lichenology. We congratulate him on his great achievements! Distribution and ecology. Sticta emmanueliana appears to be similar in ecology to S. antoniana, having been found at exactly the same localities on the islands of Maui and Kauai, in both cases in more or less undisturbed montane forest at mid elevations (between 1000 and 1500 m), on shaded tree bark. An older collection is also from Maui, at higher elevation (1800 m).
Notes. Based on the marginal isidia and black cilia, this new species would perhaps be identified with the name Sticta cometiella, originally described from Mexico. However, the latter, apparently a strictly Neotropical taxon, is entirely unrelated, clustering in a different clade in another portion of the tree, and differs also in the frequently laminal isidia (Moncada 2012;Moncada et al. 2014a). Another similar stipitate species with marginal isidia and cilia is S. duplolimbata, which has an Australasian distribution (Galloway 1998). However, the latter is also unrelated to S. emmanueliana but in turn sister to S. cometiella (Fig. S1). This is another example of superficially similar phenotypes that evolved independently in different lineages of the genus, as already shown for S. fuliginosa and S. limbata (Moncada et al. 2014a;Magain & Sérusiaux 2015).
Phylogenetically close to S. emmanueliana are the Paleotropical S. marginifera and S. caliginosa from New Zealand; both are stipitate and marginally isidiate but lack cilia (Galloway 1998(Galloway , 2007. Sticta emmanueliana is strongly supported sister to a clade formed by the latter two species (Fig. 2; Fig. S1), forming a basal, paraphyletic grade. It differs from S. caliginosa (New Zealand) substantially in ten substitutions and 13 indels and from S. marginifera (Paleotropics) in seven substitutions and 18 indels in the ITS (  Moncada & Lücking, sp. nov. (Fig. 5A-E)
Etymology. This new species is dedicated to Timothy Flynn, Curator at the herbarium (PTBG) of the National Tropical Botanical Garden and co-collector of the type material.
Distribution and ecology. Sticta flynnii has only been found twice at the type locality, growing in undisturbed montane forest at mid elevations (between 1000 and 1500 m) on shaded tree bark. An older collection comes from Haleakalā National Park on Maui, indicating a distribution and ecology similar to those of S. antoniana, S. emmanueliana and S. hawaiiensis. Like several other new species described here, due to its small size this taxon likely has been overlooked and might be more common.
Notes. Sticta flynnii is one of three new Hawaiian species, and putative endemics, clustering in an early diverging clade close to S. gracilis and S. cyphellulata ( Fig. 2;  Fig. S1). Quite a number of cyanobacterial species in the genus share the caulescent morphology with marginal isidia, beside S. cyphellulata also S. brevipes, S. hypochra, S. longipes and S. marginifera (Galloway 1994b(Galloway , 1998. All differ in morphological details such as size, the robustness and degree of branching of the thallus and the nature of the isidia. Sticta flynnii is a smaller species compared to the others mentioned here but it produces among the largest cyphellae, reaching up to 3 mm in diam. (usually up to 1 mm in the other species). This feature also distinguishes it from the closely related S. smithii.
Etymology. The epithet refers to the archipelago of Hawaii, as this new species is a putative endemic in a highly derived, apparently subcosmopolitan species complex. Distribution and ecology. Besides Sticta antoniana and S. emmanueliana, this is a third species with a similar ecology, having been found at the same two localities on the islands of Maui and Kauai, in both cases in more or less undisturbed montane forest at mid elevations (between 1000 and 1500 m), on shaded tree bark. Another, older collection stems from Oahu, in a more disturbed habitat.
Notes. The Sticta ciliata complex is one of more than a dozen mostly unrelated lineages that exhibit S. fuliginosa morphology, with broadly rounded lobes featuring laminal isidia and generally a pale underside. This distinctive lineage was first recognized in a broad phylogenetic analysis by Moncada et al. (2014a), based on material from Colombia, but Magain & Sérusiaux (2015) eventually established that S. ciliata, described from Ireland, is a representative of this clade. The data now available characterize this clade as a species complex, with uniform morphology and anatomy (small size; broad, laminally isidiate lobes; marginal whitish cilia; cells of the cyphella membrane with numerous tiny papillae).
The complex has been shown to be present in Western Europe, the Neotropics, and now Hawaii. It forms several distinct lineages ( Fig. 2; Fig. S1): two chiefly in South America, two in Western Europe (one including the type), two in the Caribbean (Puerto Rico; recently described as S. parvilobata; Mercado-Díaz et al. 2020) and one in Hawaii. The Hawaiian taxon S. hawaiiensis appears to be most closely related to S. aff. parvilobata from Puerto Rico, from which it differs in only one substitution, whereas S. parvilobata s.str. (also Puerto Rico) differs in ten substitutions. Sticta ciliata s.str. deviates in six substitutions, whereas the other, unnamed clades exhibit differences between three and six substitutions and between zero and two indels in the ITS (Table 2; Fig. 6). It therefore seems prudent to formally recognize the Hawaiian material as a distinct taxon, as the two named species in this complex show substantial differences, whereas the more similar lineages have not yet been named. Given the difference of only one substitution between S. hawaiiensis and the Puerto Rican S. aff. parvilobata, the latter could be considered to represent the same species, which would be a remarkable disjunction in this clade. However, the substitution present in S. aff. parvilobata is unique among all lineages in the entire clade (Fig. 6) and therefore we consider it unlikely that this lineage is conspecific with S. hawaiiensis. Overall, S. ciliata s.lat. appears to represent a relatively recent radiation in active, phenotypically cryptic speciation but with a distinct geographic signal (Magain & Sérusiaux 2015;Mercado-Díaz et al. 2020). Diagnosis: A diminutive, epiphytic, stipitate Sticta with a cyanobacterial photobiont, marginal, arbuscular, typically flattened isidia, and a pale underside with small cyphellae.
Etymology. The epithet honors Clifford Smith for his invaluable contributions to our knowledge of Hawaiian Figure 6. Comparison of variable sites in the fungal ITS barcoding marker in the Sticta ciliata complex (screenshot from BIOEDIT). All constant columns and the few autapomorphic, parsimony-uninformative singleton sites were deleted. Formally named taxa are highlighted. Note that the variation is structured in blocks that largely correlate with geography. The Hawaiian lineage is most similar to two unnamed lineages, namely S. aff. ciliata 4 (four substitutions) and S. aff. parvilobata (one substitution). While the difference towards S. parvilobata appears minor, it corresponds to two unique substitutions in the latter not present in any of the other lineages.
lichens and for his tireless conservation efforts to preserve native Hawaiian ecosystems, including its lichens (Smith 1991;Tunison et al. 1992;Ellshoff et al. 1995;Smith 2002;Rohrer et al. 2006). Distribution and ecology. Like Sticta acyphellata, this new species was only found in a partially disturbed secondary forest on the densely populated and strongly altered island of Oahu, where it grew epiphytic in shaded conditions on tree trunks between bryophytes.
Notes. Sticta smithii is overall most similar to S. flynnii, sharing the small, stipitate, cyanobacterial thallus with marginal isidia and the cyphellate underside. A major difference is the size of the cyphellae, becoming large and irregular in S. flynnii but remaining diminutive in S. smithii. Also, the distribution and ecology of the two species appear to differ, as judged from the limited data.
Etymology. The epithet is a noun in apposition referring to the Waikamoi Preserve (The Nature Conservancy), the largest private nature reserve in the state of Hawaii.
Distribution and ecology. This new species is so far only known from a single collection, found at high altitude in mixed conifer forest. In spite of the well-preserved appearance of this forest, no conifer is native to Hawaii and all have been introduced. It is therefore difficult to speculate about the ecology of this species. Its closest relatives, S. aff. cordillerana, S. rhizinata and S. aff. rhizinata, are from North and South America (McDonald et al. 2003;Moncada & Lücking 2012;Moncada et al. 2014a).
Notes. Sticta waikamoi is one of two species in Hawaii (the other being S. andina) best corresponding to the morphology of what has been called S. weigelii. The latter species in a strict sense appears to be a Neotropical taxon and is not directly related (Fig. S1); it differs in the thinner lower tomentum and the often yellow cyphellae. In Hawaii, S. waikamoi can be confused with S. andina, which is found in the same habitat, but differs in the generally narrower lobes and is also not directly related ( Fig. 2; Fig. S1). Although S. waikamoi is closely related to S. rhizinata, it appears to be a much smaller species, and rhizines, a characteristic feature of the latter (Moncada & Lücking 2012), are sparse and not conspicuous.
Phylogenetically, the new species forms part of a complex of four lineages, including Sticta rhizinata from Colombia and two as yet undescribed taxa from Colombia and North America (Fig. S1). Overall, S. waikamoi differs from S. rhizinata in four substitutions, and from the other two taxa in six to seven substitutions and up to one indel. Given these limited differences, one may also consider the alternative of applying subspecies level to the Hawaiian material. However, compared to the case of S. scabrosa subsp. hawaiiensis (Moncada et al., unpublished), there are twice as many differences in the ITS between S. waikamoi and S. rhizinata (Table 2) and these correlate with the deviating morphology. We therefore consider species level to be more appropriate, and we apply it also to raise awareness of this apparently very rare taxon until it can be studied from more material. Figure S1. Best-scoring ML circle tree of Sticta based on the ITS barcoding marker. Hawaiian specimens are marked in blue (exclusive Hawaiian clades) and orange (Hawaiian specimens nested within more widely distributed taxa). Bootstrap values are indicated above branches. Download file Table S1. Genbank accession numbers of ITS sequences of non-Hawaiian (or previously accessioned) representatives of Sticta (and outgroup) used in this study. Download file File S1. Alignment of the fungal ITS barcoding marker for 859 OTUs of Sticta used in this study (Fasta format). Download file