Biogeography, systematics and taxonomy are complementary scientific disciplines. To understand a species’ origin, migration routes, distribution and evolutionary history, it is first necessary to establish its taxonomic boundaries. Here, we use an integrative approach that takes advantage of complementary disciplines to resolve an intriguing scientific question. Populations of an unknown moss found in the Canary Islands (Tenerife Island) resembled two different Californian endemic species: Orthotrichum shevockii and O. kellmanii. To determine whether this moss belongs to either of these species and, if so, to explain its presence on this distant oceanic island, we combined the evaluation of morphological qualitative characters, statistical morphometric analyses of quantitative traits, and molecular phylogenetic inferences. Our results suggest that the two Californian mosses are conspecific, and that the Canarian populations belong to this putative species, with only one taxon thus involved. Orthotrichum shevockii (the priority name) is therefore recognized as a morphologically variable species that exhibits a transcontinental disjunction between western North America and the Canary Islands. Within its distribution range, the area of occupancy is limited, a notable feature among bryophytes at the intraspecific level. To explain this disjunction, divergence time and ancestral area estimation analyses are carried out and further support the hypothesis of a long-distance dispersal event from California to Tenerife Island.
Citation: Vigalondo B, Patiño J, Draper I, Mazimpaka V, Shevock JR, Losada-Lima A, et al. (2019) The long journey of Orthotrichum shevockii (Orthotrichaceae, Bryopsida): From California to Macaronesia. PLoS ONE 14(2): e0211017. https://doi.org/10.1371/journal.pone.0211017
Editor: Berthold Heinze, Austrian Federal Research Centre for Forests BFW, AUSTRIA
Received: December 6, 2017; Accepted: January 7, 2019; Published: February 13, 2019
Copyright: © 2019 Vigalondo et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information and Appendix files. All DNA sequences are available from GenBank database. The accession numbers appear in S2 Appendix.
Funding: This research was funded by the Spanish Ministry of Education and Science (CGL2007-61389/BOS) to F.L., the Spanish Ministry of Science and Innovation (CGL2011-28857/BOS) to V.M., the Spanish Ministry of Economy and Competitiveness (CGL2013-43246-P) to R.G., and the Spanish Ministry of Economy, Industry and Competitiveness (CGL2016-80772-P) to I.D. and the Spanish Ministry of Economy and Competitiveness (grants IJCI-2014-19691 and RYC‐2016‐20506) to J.P. J.P. also received the H2020 Marie Skłodowska-Curie Actions grant 747238. J.M.G-M and A.L-L. are grateful for funding from Parque Nacional del Teide. B.V. benefited from the support of the Ministry of Economy and Competitiveness through grant BES-2012-051976 of the Formación de Personal Investigador (FPI) programme. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Since Wegener’s plate tectonics theory was proposed, ancient fragmentation has long been considered the main process explaining common distribution patterns in plant biogeography , while dispersal is seen as a random and irrelevant process . However, more recent molecular tools and the development of dating and divergence time estimations have pointed to dispersal as a key process contributing to current species distributions [3–7]. In the case of oceanic islands, which originate without a former connection to a continental landmass, dispersal is considered to play a fundamental role in the generation of biodiversity and biogeographical patterns [2,8–12]. For the Macaronesian islands, a biogeographic region that encompasses the archipelagos of the Canary Islands, the Azores, Madeira and Cabo Verde (but see Vanderpoorten et al. 2007), it has been suggested that the endemic bryophyte component of the flora has a different biogeographical origin compared to angiosperms. This pattern has been explained, at least partially, by their different dispersal capabilities, since ancestors of a few endemic bryophytes seem to have colonized the islands from more distant continental pools [13,14]. Similarly, compared to tracheophytes, the larger distribution ranges of bryophytes have been attributed to their higher dispersal capabilities . In many cases, these ranges involve intercontinental disjunctions at the species level, while in vascular plants these mostly occur at a genus level [15,16].
Recent studies have provided evidence for the traditional hypothesis that vicariance through ancient fragmentation may explain the origin of widely disjunct distributions in a few bryophyte lineages [17–19]. There is growing evidence that long-distance dispersal (LDD), however, has shaped the bulk of transoceanic bryophyte distributions [20–23]. This phenomenon also applies to taxa present in Macaronesia [24–26]. Despite this, it is worth noting that processes like incomplete lineage sorting, slow evolutionary rates [27–29], and cryptic speciation (for review see [30,31]), may contribute to establishing apparently disjunct distributions in bryophytes. This may also be the result of incomplete taxonomical knowledge [32,33]. All of these caveats call for accurate species delimitation methods, which are a necessary first step in assessing distribution patterns, and performing biogeographic analyses in widely distributed bryophytes. To this end, a plethora of integrative approaches has proved useful [23,29,32–38].
During recent field surveys on Tenerife Island (Canary Islands), several saxicolous populations of an unknown Orthotrichum Hedw. species were found in the area of El Teide National Park, at altitudes around 2100 m a.s.l., growing in protected crevices of volcanic rocks and walls. A preliminary morphological examination of these specimens revealed that their main characteristics differed from any Orthotrichum species known in the Mediterranean and North Atlantic areas. Surprisingly, these populations resembled two Californian species: O. shevockii Lewinsky-Haapasaari & D.H. Norris, and O. kellmanii 1D.H.Norris, Shevock & Goffinet. Orthotrichum shevockii is a saxicolous moss described from two localities in dry mountain areas in the southern Sierra Nevada, California, between 1150 and 1600 m a.s.l.; it is restricted to granitic rock outcrops, ceilings and underhangs of large boulders where plants receive only indirect sunlight and moisture by capillarity supply from the rock surfaces. Orthotrichum kellmanii is another saxicolous species, known from just a few localities in the coastal mountains of central California, where the climatic regime is characterized by high winter rainfall and infrequent summer fog from the Pacific Ocean. It grows on sandstone rock outcrops in chaparral areas, at altitudes around 650 m a.s.l. These two similar species mainly differ in gametophytic traits. Orthotrichum shevockii is characterized by leaves with bi- to tristratose margins and highly papillose leaf cells , whereas the leaves of O. kellmanii have a completely bistratose lamina . However, the description of O. kellmanii based its diagnostic characters mainly on the presence of heterophyllous leaves (reproductive and vegetative stems leaves have different shapes), and a weakly cladocarpous growth across the substrate. The specimens from the Canary Islands have leaves with a highly variable degree of bistratosity among different individuals, from completely bistratose leaf laminae to bistratosity restricted to the leaf margins.
The bryophyte flora of Macaronesia is one of the best known among oceanic island regions worldwide . However, the knowledge of its diversity and levels of endemism is still incomplete, as suggested by the increasing number of recent descriptions and re-circumscriptions of species for this region (e.g. [29,35,38,41,42]). Herein, in order to evaluate the true nature of the populations discovered in the Canary Islands, we address the following questions: (1) What is the identity of the new moss found in the Canary Islands? (2) What are the taxonomical relationships between the Canarian Orthotrichum moss lineages and the Californian O. shevockii and O. kellmanii? (3) What is the evolutionary and biogeographical history of these lineages? To accurately address these questions, we used an integrative taxonomic approach combining morphological analyses, phylogenetic inferences, molecular dating and the estimation of ancestral ranges.
Material and methods
Field collecting permits were granted by both California state and federal ownership, where applicable, as well as from Parque Nacional del Teide.
The material for this study includes the collections sampled on Tenerife, herbarium specimens of O. shevockii and O. kellmanii from UC, CAS, CONN and NY herbaria (including type material), and specimens obtained during specific collecting campaigns to several Californian mountain ranges and neighboring regions of western Nevada. Specimens were selected to represent the whole distribution and ecological range of the species. For the target species, 30 samples were included in the morphological analyses: nine from Tenerife and 21 from California (Fig 1 and S1 Appendix). Based on the availability and quality of the specimens, a subset of 16 samples representative of the morphological diversity and geographic distribution of the species was selected for molecular analyses (Fig 1 and S2 Appendix). To provide a phylogenetic context for the assessment of the monophyly in these two putative species, along with the populations from the Canary Islands, in addition to our ingroup sequences, we included specimens of other Orthotrichum species from the western coast of North America and the Canary Islands, some endemic to these areas [25,33,39]. Three species of Lewinskya F.Lara, Garilleti & Goffinet, one of Macrocoma (Hornsch. ex Müll.Hal.) Grout, one of Nyholmiella Holmen & E.Warncke, and two of Zygodon Hook. & Taylor were selected as outgroups, resulting in a total of 66 samples (see S2 Appendix for voucher information and GenBank accession numbers).
(a) California and Nevada, (b) the Canary Islands, (c) Tenerife. Numbers indicate specimens included both in morphometric and phylogenetic analyses (see S1 and S2 Appendices). ** = original locality of O. shevockii, *** = original locality of O. kellmanii.
A morphological analysis was conducted on the 30 selected specimens to assess the differences between the Californian plants attributed to either O. shevockii or O. kellmanii, and those from Tenerife. A set of morphological characters, both qualitative and quantitative, was selected and studied according to our previous experience with Orthotrichaceae [32,33,43–45]. Qualitative traits of the gametophyte included plant habit and several leaf characters such as leaf shape, leaf margin, lamina bistratosity, and cell papillosity. Sporophyte characters are usually of great diagnostic value for the genus [46,47], and we therefore focused the study on capsule shape, exothecial band structure, stomata position, structure and ornamentation of the peristome calyptra, and vaginula hairiness.
For quantitative morphometric analyses, 16 characters were selected. Measurements and construction of the dataset protocol followed Vigalondo et al. . To detect a possible unknown underlying structure within the dataset, an exploratory multivariate analysis was performed (principal component analysis, PCA). A correlation matrix was used in the PCA to scale the morphological variables, and only principal components (PCs) accounting for more than 10% of the variance were considered in the results. Univariate analysis of variance (ANOVA) was conducted to assess the homogeneity of variances for each of the 16 quantitative variables for California and Tenerife specimens. Multivariate analyses were run twice: (i) discarding samples with missing values; and (ii) replacing missing values with the mean value of each character. Results from the two approaches were congruent, so to avoid reducing the sample size, we used the data set with missing values replaced with the mean for the final analyses (S1 Table). Descriptive statistics were ultimately computed for all quantitative variables, considering populations from Tenerife and California separately. The results were summarized in the form of beanplot graphs , representing the empirical density shape, mean, and all individual observations for each of the two evaluated geographical groups. All statistical analyses were conducted in R v.3.3.1 .
DNA extraction and sequencing
DNA was extracted from apices of stems and branches from dried herbarium specimens using the DNeasy Plant Mini Kit for DNA isolation (Qiagen). We selected four loci previously used for phylogenetic reconstructions of Orthotrichum [33,45]: three chloroplast loci, namely atpB-rbcL, rps4, and trnL-F, and the nuclear internal transcribed spacer II (ITS2). The primer pairs used for each locus were atb1/rbcL1 , rpsA/trnaS [51,52], trnC/trnF  and ITS2F/ITS2R .
Double-stranded DNA templates were prepared by PCR, which was performed using Ready-To-Go PCR Beads (Amersham Pharmacia Biotech Inc.) in a final reaction volume of 25 μL according to the manufacturer’s instructions. PCR amplification of atpB-rbcL, rps4, and trnL-F was performed using the protocol described in , while the ITS2 protocol followed . PCR products were purified using the Exo/SAP protocol (Thermo Fisher Scientific, Spain). Samples were incubated with 1 μL of Exo1 enzyme and 4 μL of FastAP following the manufacturer’s instructions. Cleaned PCR products were sequenced by Macrogen (www.macrogen.com). All new sequences were deposited in GenBank (see S2 Appendix).
Phylogenetic and dating analyses
Nucleotide sequence contigs were edited and assembled for each DNA region using Geneious 7.1.2 (http://www.geneious.com, ) and PhyDE v.0.9971 . Sequences were aligned manually and trimmed at the ends. Regions of ambiguous or incomplete data were identified with GBlocks  and excluded from subsequent analyses.
Phylogenetic analyses were performed using maximum likelihood (ML) and Bayesian inference (BI). The best-fitting substitution models for each locus were inferred under the Bayesian Information Criterion (BIC) in jModelTest v.2.1.3 . Maximum likelihood analyses were run with RAxML 8 , and the best ML tree was selected from 100 iterations and its support was assessed with 1000 replicates of bootstrap resampling under the ML criterion. Bayesian phylogenetic analyses were carried out using MrBayes v.3.2.1 . The Markov chain Monte Carlo (MCMC) was run for 2 to 5 million generations with two runs and four chains, sampling trees, and parameters every 1000 generations. After checking that stationarity had been reached (i.e. the average standard deviation of split frequencies remained below 0.01 for the last 10,000 generations), posterior probabilities (PP) were estimated from the 50% majority-rule consensus trees after a burn-in of 25% of the starting trees. The resulting trees for both ML and BI analyses were plotted using FigTree v.1.4.2 .
Insertions and deletions (indels) in non-coding regions are sometimes difficult to assess  and can lead to ambiguous alignments. To determine the effect of their inclusion, phylogenetic information from indels was coded as an adjacent block with the program SeqState , using the simple indel coding method . The analyses were performed with and without codified indels with the same parameters indicated above, using model F81 for the indel partition in MrBayes, as recommended by . The inclusion of the indels did not change the topology of the trees or result in increased statistical support as measured by PP, and further analyses were performed on the matrix treating the indels as missing data.
All independent gene data sets were combined in a single concatenated matrix, as no incongruences were identified in branches supported with posterior probability ≥ 0.95 and bootstrap support ≥ 85 when each gene was analyzed separately. The final analyses included only those sequences for which all loci were obtained, thus discarding 11 sequences from the final matrix. The resulting concatenated data set was analyzed in PartitionFinder  to select the best partitioning scheme and nucleotide substitution model, using the greedy algorithm with linked branch lengths under the BIC criterion. Three partitions were defined: ITS2 (HKY+G), rps4 (HKY+G) and the combined atpB-rbcL and trnL-F (GTR+G).
Divergence times were estimated using BEAST 1.8.0 . Because the inclusion of identical sequences in dating analysis results in many zero-length branches at the tip of the tree and can cause the model to over partition the dataset , we reduced the data set to haplotypes (30 sequences) using DnaSP 5.10.1 . This program considered all samples from the Canary Islands as a unique haplotype, although one of them appeared outside the main group of this area in the phylogenetic analyses. Thus, we compared BEAST analyses with the sample (BV050 ) either separate or included in the general haplotype from the Canary Islands. For all the analyses, clock and tree models were linked across partitions, and models of substitution were unlinked across the three partitions. Both strict and uncorrelated log-normal relaxed clocks were tested under two different speciation tree models: Yule and birth-death process. In the absence of fossil records of Orthotrichum, the uncertainty of dating estimates was modelled with a uniform distribution under two increasingly conservative nucleotide substitution rate assumptions incorporated into the ucld.mean parameter in BEAST. First (analysis I), we used an absolute substitution rate of mean = 4.453E-4 and stdev = 1.773E-6 substitutions/site/million years, inferred from relaxed-clock analyses across the Moss Tree of Life . Second (analysis II), to provide an alternative to time estimates that might be overestimated (given that Laenen et al.  performed the analyses at the generic level), we applied a distinct rate for the plastid (5.0E-4, stdev range 2–8E-4, subst./site/ma) and nuclear partitions (4.13E-3, stdev range 1.72–8.34E-3, subst./site/ma), proposed by Villarreal & Renner . All BEAST analyses were run for four independent chains of 40 million generations each, sampling every 10 thousand generations and their convergence was assessed by confirming that all parameters had reached stationarity and sufficient effective sample sizes (> 200) in all converged runs using Tracer v1.6 . The best model was selected through marginal likelihood estimates (MLEs) that were assessed using path-sampling (PS, ) and stepping-stone (SS, ) methods. The resulting MLEs were averaged across replicate runs to generate a single PS and SS value for each model. The obtained MLEs for all hypotheses were ranked, and Bayes factors were then calculated. In this study, the birth-death process model performed best (S2 Table). After discarding the burn-in steps, tree files from the four independent runs of the selected model were combined using LogCombiner 1.8  and the resulting maximum clade credibility (MCC) tree was summarized in TreeAnnotator 1.8  and viewed in FigTree v.1.4.2 .
Ancestral area estimation
To infer the historical biogeography of O. shevockii, we defined six geographical areas, also considering the whole distribution of the rest of the ingroup species: western North America (W), eastern North America (E), Caribbean, Central America and South America (N), Europe (U, including the Mediterranean and North Africa), Macaronesia (M), and Asia (A). We used the time-calibrated MCC tree obtained from BEAST, but removing the outgroups, to perform ancestral area estimations across the Orthotrichum ingroup with the R package BioGeoBEARS . BioGeoBEARS allows the use of the Lagrange DEC model (Dispersal-Extinction-Cladogenesis), which includes dispersal (d) and extinction (e) as free parameters, and a model (DEC+J) that includes an additional parameter J taking founder-event speciation into account ( and references therein). Since different ancestral area reconstructions are based on different assumptions, one can compare these two versions of the DEC model with a likelihood version of the Dispersal-Vicariance Analysis (DIVALIKE), and a likelihood version of the range evolution model of the Bayesian Binary Model (BAYAREA), with the option of also adding founder-event speciation to either of these two alternative models. However, in a recent study, Ree and Sanmartín  proposed that DEC+J might be a poor model of founder‐event speciation and statistical comparisons of its likelihood with a pure DEC model may be inappropriate. Consequently, we refrained from implementing the DEC+J in the present study and focused on the classical versions of the three biogeographical models implemented in BioGeoBEARS (DEC, DIVALIKE, BAYAREA). These three models were estimated under a maximum likelihood framework, and compared in terms of how well they fitted the data using the Akaike Information Criterion (AIC) [75,77]. Implementing the same best-fit model of ancestral area estimation (i.e. DEC; see results section), we determined the degree to which differences in tree topology and branch lengths yield different ancestral area estimations for nodes by using a customized script to run BioGeoBEARS on a subset of 100 BEAST trees randomly sampled from the posterior probability distribution obtained from BEAST analyses.
Only a limited number of specimens from California and Nevada could be clearly ascribed to either O. shevockii or O. kellmanii. This was not possible for the majority of specimens, which actually showed morphological traits of both taxa, traits that can also display a wide range of variation (Figs 2A–2D and 3H–3I). Moreover, all Californian specimens share some characteristics that are very uncommon in the genus Orthotrichum, such as: (1) stomata that appear restricted to the capsule neck, and only occasionally reach the base of the urn (Fig 4A and 4I); (2) exostome teeth that are usually lacunose, showing lacunae in both the external and internal layers around the median line, sometimes appearing within the teeth cell areas (Fig 4B, 4J and 4K); and (3) an endostome external layer that is frequently ornamented with oblique or vertical lines (Fig 4D and 4L).
(a-d) Orthotrichum shevockii from Western North America; (e-g) samples from Tenerife; (h-i) O. kellmanii from California. (a, c, f and i) capsule detail; (b, d, e, g and h) habit; (e) two different habits from the same voucher. Scale bars: a, c, e and i = 0.5 mm, b, d, f, g and h = 1 mm. Vouchers: a-b, Shevock 13404 (CAS 958716, paratype); c, Shevock 21948 (CAS 1040048); d, Shevock 21802 (UC 1754431); f, Losada-Lima, León & Díaz s.n. (TFC-Bry 15904); e-g, Losada-Lima s.n. (TFC-Bry 17428); h, Shevock 32935 (MAUAM 5097); i, Shevock 32935 (NY 1140598).
(a) Orthotrichum shevockii from western North America; (b) samples from Tenerife, (c) O. kellmanii from California. (a,b) from left to right: bistratose leaf (except base), leaf with a bistratose upper part and bistratose bands, leaf with dispersed bistratose bands in the upper part, leaf almost unistratose with bistratose patches around the apex. (c) top: bistratose leaf (except base), bottom: leaf with bistratose upper part and bistratose bands. Each leaf belongs to a separate individual. Cross-sections belong to a different leaf of the same individual. Arrowheads indicate bistratose strands or margins; in (c) they indicate the tristratose margins. Scale bars: leaves = 0.5 mm, cross sections = 100 μm. Vouchers: a, Shevock & Anderson 16754 (UC 1754230), Lara, et al. s.n. (MAUAM-Brio 3289), Shevock & York 13404 (CAS 958716, paratype of O. shevockii), Shevock 21802 (UC 1754431); b, J.M.B., J.G.M. & J.L.P s.n., (TFC-Bry 15957), Losada-Lima s.n. (TFCBry-17428), Losada-Lima, León & Díaz s.n. (TFC-Bry 15861), Losada-Lima, León & Díaz s.n. (TFC-Bry 15904); c, Shevock 32935 (NY 1140598).
(a-d) Orthotrichum shevockii from western North America; (e-h) samples from Tenerife; (i-k) O. kellmanii from California. (a, e and, i) capsules with stomata mostly restricted to the neck; (b, f and j): exostome structure, showing the teeth lacunosity; (c, g and k) endostome internal layer (IPL) papillose ornamentation; (c and k) exostome internal layer (PPL), black arrows indicate striae at base; (d, h and l) endostome external layer (PPL) ornamentation; white arrows indicate lines or striae, sometimes forming plaques. Scale bars: a, e and i = 200 μm; b-c, f-g, j-l = 20 μm; d-h = 10 μm. Vouchers: a, Lara, et al. s.n. (MAUAM-Brio 3289); b, d, Shevock 13404 (CAS 958716, paratype), c, Shevock 21802 (UC 1754431); e, Losada-Lima s.n. (TFC-Bry 15567); f, Losada-Lima, León & Díaz s.n. (TFC-Bry 15904); g, Losada-Lima s.n. (TFC-Bry 17406); h, Losada-Lima, León & Díaz s.n. (TFC-Bry 15952); i-j, Shevock 32935 (MAUAM 5097); k-l, Shevock 32935 (NY 1140598).
Other qualitative characters exhibit a wide range of variation among and within regions, and frequently even among stems from the same cushion. Plants commonly form short and dense cushions or tufts, but in some cases, they look longer and looser (Fig 2B and 2D). Concerning gametophyte traits, those related to leaves show great variation (Fig 3A). Leaf margins are typically bistratose for most of their length, occasionally 3–4 cells thick and rarely unistratose. Similarly, the leaf lamina is predominantly or completely bistratose in its upper 1/2-2/3 part, but can have sparse bistratose bands in its upper half or, rarely, only small bistratose strands or patches restricted to the apex. Papillosity of leaf lamina is also highly variable, with cells featuring 2(3–4) papillae on each side; papillae can be prominent, simple or bifurcate, or in other cases short or even negligible. Papillosity to a great extent appears to be related to leaf thickness: bistratose leaves have lamina cells with low papillae or almost smooth, while leaves only bistratose at the margins or with bistratose strands in the lamina show high and bifurcate papillae. Upper leaves, both vegetative and perichaetial, can be broadly to narrowly lanceolate, and their apices are usually acute, although sometimes perichaetial leaves are shortly acuminate (Fig 3A).
As for the sporophytic traits, capsule exothecial bands are formed by 4–8 isodiametric to rectangular cells, varying among samples, and usually extend along the whole urn length, but are occasionally restricted to the upper half only (Fig 4A). The ornamentation of the different components of the peristome is noticeably variable (Fig 4B–4D). The exostome outer layer consists of a reticulum at the basal part, where transverse lines are usually more apparent, and it is covered by a variable proportion of papillae. In contrast, the middle and upper parts of teeth have a denser ornamentation with a predominance of tall (occasionally low) papillae or, less frequently, vertical lines (Fig 4B). Ornamentation of the inner surface of the exostome can be papillose, reticulate, striate, or a mix, and predominantly shows very well-marked vertical striae at the base (Fig 4C). Regarding the ornamentation of the endostome, the internal layer is rugulose or papillose, with papillae sometimes densely disposed and variably prominent (Fig 4C), while the external layer can be smooth or variably ornamented with lines, sometimes densely grouped in plaques (Fig 4D).
The isotype material of O. kellmanii and two other samples originally ascribed to this species fit the variability encountered in the rest of the Californian samples (Figs 2H and 2I and 3C and 4I–4L). They also exhibit the above-mentioned characteristics. The only peculiarities noticed for these three samples concern leaf characteristics. As in other Californian samples, leaves are extensively bistratose, both in margins and lamina, but exceptionally show up to three layers of cells in areas adjacent to the nerve and near the apex (Fig 3C). Additionally, the perichaetial leaves are consistently shortly acuminate. Although most sporophytes show the typical structure described above, in some capsules the exothecial bands are unusually weak, made up of 3–6 cell rows and restricted to the upper part of the urn (Fig 4I). None of the material shows a cladocarpous growth pattern as described in Norris et al. .
The specimens from both Tenerife and California share all the qualitative characters mentioned above, and exhibit the same degree of variation for gametophytic (Figs 2 and 3) and sporophytic traits (Fig 4). The few peculiarities observed affect only the frequency of some leaf traits. In Tenerife, the leaf lamina of most samples is commonly partially bistratose, with bistratose margins and dispersed bistratose bands in the upper part of the lamina. Leaf cell papillae are commonly short, and perichaetial leaves are usually broadly lanceolate (Fig 3B).
Statistical analyses of morphological quantitative traits also showed no differences between California and Tenerife specimens. In PCA analyses, the three first principal components (PCs) accounted for 59.98% of the variance. The PCA biplot shows a dispersion of samples within the represented space, where specimens from California and Tenerife overlap without any geographical or taxonomical structure (Figs 5 and S1). With respect to the specimens originally identified as O. kellmanii, only one of them appears separated in the positive extreme of PC1. The most important variables in each of the three PCs are indicated in Table 1. When variables are considered independently comparing California and Tenerife, ANOVA analysis only shows significant differences for one variable: perichaetial leaf width (Table 1 and Fig 6).
The percentage of variance explained by each component is given between brackets. Arrows represent the variables included in the analyses. cos2 represents the squared loadings for variables. ! = samples originally identified as Orthotrichum kellmanii.
Yellow = Tenerife (Canary Islands), blue = California (western North America, including O. kellmanii isotype materials). Individual observations are represented by small horizontal lines (in the case of multiple observations with the same values, the corresponding number of lines were merged), mean per group is shown by a bold long line and the mean for all data by a dotted line. Estimated density of the data distribution is displayed by the density shape in grey (for details see ). Stars indicate ANOVA significance values: * 0.05.
Phylogeny, dating and ancestral area reconstruction
Information regarding sequence length and variability within each marker and the combined matrix is presented in Table 2, whereas pairwise differences among ingroup samples are shown in S3 Table. Phylogenetic analyses of ML and BI (Fig 7) group samples from California and Tenerife in the same monophyletic lineage with high support (BS = 75, PP = 1.0). This lineage is embedded within a clade with a PP of 0.93, composed of taxa restricted to California and Nevada along with Orthotrichum handiense F.Lara, Garilleti & Mazimpaka from Fuerteventura (Canary Islands). Samples of O. kellmanii are also placed within the clade of O. shevockii in BI and ML analyses.
Bayesian posterior probabilities (≥ 0.90) and maximum likelihood bootstrap values (≥ 70%) are shown above and below branches, respectively. Sequence labels of Orthotrichum shevockii are followed by identification number, geographical origin, and number identification between brackets as in Fig 1 and S1 and S2 Appendices. ** = paratype material of O. shevockii, *** = isotype material of O. kellmanii.
When considering a unique substitution rate (analysis I), dating analyses indicate that the split of O. shevockii between California and Tenerife populations dates back to the late Miocene–Pliocene (2.75 Ma; 95% highest posterior density interval (HPD): 0.44–6.69 Ma, Fig 8 node A). The results considering different rates for the nuclear and plastid partitions (analysis II) place the split 1.74 Ma (95% HPD: 0.17–3.73 Ma, S2 Fig node A). The common ancestor of the Californian clade dates back to 24.4 Ma (95% HPD: 10.9–31.45 Ma, Fig 8 node D) in analysis I and 18.44 Ma (95% HPD: 10.37–28.1 Ma, S2 Fig, node D) in analysis II. The best-fit model of ancestral area estimations, the DEC (Table 3), suggests that the present distribution of O. shevockii results from at least one long-distance dispersal event from western North America, which is supported by its inclusion in a highly supported western North American clade along with, as mentioned before, the Canary Island endemic O. handiense (Fig 7). When running the DEC analyses on 100 BEAST trees randomly sampled from the posterior probability distribution, the ancestral area estimates for the clades of interest (i.e. O. shevockii) were fully consistent with the former analysis (S3 Fig).
Maximum clade credibility tree from the relaxed molecular-clock analysis of the four loci in BEAST with an absolute nucleotide substitution rate (analysis I, mean = 4.453E-4, stdev = 1.773E-6subst./site/ma). Asterisks (*) at nodes refer to highly supported nodes (PP >0.95). Sequence labels are followed by identification number and geographical origin. In the case of Orthotrichum shevockii, sequence labels are also followed by an identification number between brackets as in Fig 1 and S2 Appendix (*** = isotype material of O. kellmanii). For the Orthotrichum group, letters in colored boxes in each node show the area or combination of areas with the highest probability of being ancestral, according to the reconstructions based on the DEC model implemented in BioGeoBEARS. Letters correspond to the following ancestral areas or combination of areas: W = western North America; E = eastern North America; N = Neotropics; M = Macaronesia; U = Europe; A = Asia. The complete probabilities of each biogeographical region are presented in S3 Fig.
Taxonomic relationships of Orthotrichum shevockii and O. kellmanii
Our results show a lack of morphological or molecular differences that would have supported the consideration of O. shevockii and O. kellmanii as separate species, based on the analysis of a significant number of samples and including type materials from both taxa. The morphological analyses revealed that a large number of qualitative traits exhibited a uniform variation range among these specimens. These traits include some basic qualitative traits regarding the gametophyte and sporophyte, but also others that are singular within the genus Orthotrichum [33,44,46,78], and thus can be considered as diagnostic characters, such as: stomata restricted to the neck, exostome lacunose, and endostome PPL ornamented with lines or striae (Fig 4). Moreover, quantitative traits did not reveal any evidence of taxonomic differentiation (Fig 5), and molecular analyses resulted in grouping the different samples ascribable to these two Californian species in a well-supported clade. Within this group, further segregation of samples seems to have no geographical, ecological or taxonomical meaning.
Our findings do not support the previous consideration of O. shevockii and O. kellmanii as two different species, and probably points to the fact that the description of both taxa was made based on very few samples, which exhibited extremes of the morphological variation with respect to some gametophytic traits. Orthotrichum shevockii was described  based on samples from two close inland xeric localities at the junction of the southern Sierra Nevada and western Mojave Desert, whose specimens have leaves with bistratosity mostly restricted to the margins. Orthotrichum kellmanii was described  based upon samples from two nearby coastal localities in which specimens showed completely bistratose leaves. Additionally, gametophores of these latter samples were larger than usual. In fact, these coastal localities of O. kellmanii are at considerably lower altitudes and receive more humidity due to the Pacific Ocean influence, with occasional summer fog, which could favor the greater development and shoot size of these populations. Under these circumstances, specimens of this moss species form long sympodial gametophytic axes with: (i) abundant basal short and ligulate leaves (identified as vegetative by ); (ii) and upper leaves (from female axes) that are progressively larger and lanceolate. This could explain why  interpreted the habit as weakly cladocarpic with extreme heterophylly, as the plants spread prostrate across the smooth sandstone rock surface. The development of shorter and somewhat different basal leaves, although rarely highlighted, is a common characteristic in Orthotrichum and related genera (see for example ), whereas true heterophylly related to male and female branches has only been reported for one European moss . Considering all of these arguments, it is clear that there is no morphological or molecular evidence that would further support the separation of O. kellmanii and O. shevockii as two distinct species. Therefore, we propose that O. kellmanii be synonymized under the priority name of O. shevockii.
Orthotrichum shevockii Lewinsky-Haapasaari & D.H. Norris, Bryologist 101(3): 435. 1998.
= Orthotrichum kellmanii D.H.Norris, Shevock & Goffinet, Bryologist 107(2): 210. 2004. syn. nov.
Taxonomic status of Tenerife populations
Once the circumscription of O. shevockii has been clarified, including the synonymization of O. kellmanii therein, we can now consider the identity of the Tenerife populations. Our analyses reveal that the moss populations newly found in the Canary Islands (Tenerife island) also correspond to O. shevockii. Specimens from these populations are rather uniform in gametophyte and sporophyte characteristics, and fall within the range of morphological variation encountered in western North America. As mentioned above, the fact that all samples from the Canary Islands and western North America share some diagnostic qualitative traits considered unusual within the genus (see Fig 4), along with their strong phylogenetic relatedness, is highly suggestive that these lineages represent the same species, from an evolutionary point of view.
Indeed, the molecular results (Fig 7), although based on only four molecular markers, with three of them belonging to the chloroplast compartment, agree with the morphological evidence, with the Canarian clade being nested within the Californian one. Additionally, ecological aspects point in the same direction. In Tenerife, for instance, O. shevockii occurs in arid areas at high altitudes, as a saxicolous moss associated with crevices, rock ceilings, and vertical faces on volcanic rocks. This is also the most frequent (micro-) ecological setting of O. shevockii in western North America, except for the occurrence of all the inland populations on acidic granitic (granodiorite) rocks, while the few coastal mountain localities are on a wide variety of rock types, from metavolcanics to marble.
The discovery of O. shevockii as new to the Canary Islands raises the number of species of Orthotrichaceous mosses to 14 (4 Lewinskya, 9 Orthotrichum, 1 Pulvigera) known for this archipelago [80,81]. It is evident that the increasing implementation of integrative taxonomic approaches substantially improves our knowledge of the real regional diversity of plant groups with reduced morphologies like mosses, and, in general, of the still incomplete cryptogamic floras of oceanic biogeographic regions like Macaronesia .
California–Macaronesia disjunction of Orthotrichum shevockii
Our results lead to the conclusion that the distribution of O. shevockii is disjunct and includes western North America (California and Nevada) and Macaronesia (Canary Islands, Tenerife). The connection of cryptogamic floras of Macaronesia and America has already been described, mainly referring to species present on the Azores and Madeira archipelagos, and involving disjunctions with tropical or Caribbean regions [14,82]. Some of these species have their main distribution in America, similar to O. shevockii. However, the disjunction reported here between the Californian region and the Canary Islands is quite uncommon. Other spore-producing organisms such as lichens, also have species with this type of distribution , but among bryophytes, species that are present in both regions tend to also expand their distribution into the Mediterranean basin [45,84,85].
The dating analyses place the colonization of Tenerife directly from western North America between 2.75 Ma (95% HPD: 0.44–6.69 Ma, analysis I) and 1.74 Ma (95% HPD: 0.17–3.73 Ma, analysis II), long after Tenerife island (11.9 Ma), and even the central and larger parts of the island (3 Ma) [86,87], where O. shevockii currently occurs, were formed. The phylogenetic inferences resolved O. shevockii within a clade composed of Californian endemic species, and the ancestral area estimation suggests a western North American origin for its ancestor (Figs 7 and 8). Our findings thus support that the present distribution of O. shevockii is the result of a long-distance dispersal event from California to the Canary Islands. This confirms the hypothesis that recurrent events of long-distance dispersal have occurred within the genus Orthotrichum from western North America (California) to the Macaronesian region, and in particular to the Canary Islands . These events have taken place at different times and reflect different dispersal windows, with the split of O. underwoodii and O. handiense being older  than the disjunction of O. shevockii (Fig 8).
The Californian origin of O. shevockii provides additional support for the hypothesis that the Macaronesian cryptogamic flora may be more related to the New World , at least for certain groups of bryophytes, whereas angiosperms are more related to Europe and North Africa . Moreover, it increases the evidence for bryophyte species with transoceanic distributions, which includes a number of Macaronesian taxa [25,82,88]. In the case of the Canary Islands, trade winds that cross the Atlantic Ocean run from east to west—opposite to the direction that has been identified for this long-distance dispersal—and cannot explain this disjunction. On the contrary, the high altitude subtropical jet stream that crosses over California and the Canary Islands [89,90] seems to be a suitable vector for wind-mediated dispersal events from west to east, as has been suggested in general for long-distance wind dispersal events in bryophytes  and, in particular for the North America-Europe disjunction .
Although O. shevockii’s range is constrained to few scattered locations in western North America (mountainous areas of California and nearby regions of westernmost Nevada), the presence of the species in the Canary Islands is restricted to a significantly smaller area. Restricted ranges in bryophytes are related to a recent origin, loss or lack of dispersal ability, extinction, preference for a specific habitat or a combination of some of these factors . Our dating analyses do not rule out a relatively recent origin for the disjunction, which could be placed between 0.17 and 6.69 Ma (see above). Concerning dispersal capabilities, all collected samples showed sporophytes with high numbers of spores that are small enough (ca. 12 μm in diameter), a particular size thought to be easily carried by wind over long distances [10,93]. Furthermore, it has been suggested that Macaronesian bryophyte species do not necessarily lose their dispersal ability, maintaining connections between islands, archipelagos and nearby continents [24,94–97]. Therefore, its restricted area is not a priori attributed to reproductive constraints, but likely to habitat limitations.
Most of the bryophyte species (endemic or otherwise) with restricted distributions in the Canary Islands grow in very rare habitats, which is especially observed among taxa restricted to the laurel forest or inhabiting high altitude scrublands . The latter match the distribution of O. shevockii in Tenerife, since it only occurs on rocks in open arid zones dominated by leguminous scrubs at altitudes around 2100 m a.s.l. On this archipelago, these altitudes are also found in La Palma Island, where the favorable habitat for this moss is more restricted than in Tenerife. This type of habitat is absent from the Azores and Cape Verde, the other Macaronesian archipelagos that exceed elevations of 2000 m a.s.l. . Therefore, the distribution of O. shevockii in Macaronesia, which is currently restricted to Tenerife, could be explained by the lack of a suitable habitat in the region, a relatively recent founding event or a combination of these two phenomena. Future research on the landscape population genetics of disjunct lineages like O. shevockii represent a unique opportunity to improve our mechanistic understanding of origin, evolution and distribution of insular bryophyte floras.
S1 Appendix. Selection of samples used for morphological analyses.
S2 Appendix. Specimens included in the molecular analyses and GenBank accession numbers.
New accessions from this study are in italics. Numbers between brackets after taxon ID correspond to the specimens included in molecular analyses as used in Figs 1 and 7. Samples originally identified as Orthotrichum kellmanii appear under this name in the table.
S1 Fig. Results of the principal component analysis (PCA) representing the first three components.
The percentage of variance explained by each component is given in brackets. Arrows represent the variables included in the analyses. cos2 represents the squared loadings for variables. ! = samples originally identified as Orthotrichum kellmanii.
S2 Fig. Molecular dating using a distinct nuclear and plastid nucleotide substitution rate.
Maximum clade credibility tree from the relaxed molecular-clock analysis of the four loci in BEAST from analysis II with a distinct rate for the plastid (5.0E-4 (2–8E-4) subst./site/ma) and nuclear partitions (4.13E-3 (1.72–8.34E-3) subst./site/ma). Black and grey circles at nodes refer to node support of PP >0.95 and PP >0.75 - <0.95, respectively. Identification number and geographical origin follow sequence labels. In the case of Orthotrichum shevockii, sequence labels are also followed by number identification between brackets as in Fig 1 and S2 Appendix (*** = isotype material of O. kellmanii).
S3 Fig. Biogeographical analyses.
Ancestral area reconstructions. Chronogram of the phylogenetic relationships among the four loci from analysis I, and ancestral area estimations for the Orthotrichum shevockii group and the evaluated ingroup estimated from analyses run using 100 BEAST trees randomly sampled from the posterior probability distribution. Pie charts show the relative probability of each area or combination of areas being ancestral, according to the ancestral area reconstructions under the DEC model implemented in BioGeoBEARS. Ancestral states: global optim, 6 areas max.; d = 0.0065; e = 0; LnL = −75.09. Letters in coloured boxes correspond to the following ancestral areas or combination of areas: W = western North America; E = Eastern North America; N = Neotropics; M = Macaronesia; U = Europe; A = Asia.
S2 Table. Marginal likelihood (MLE) and Bayes factor (BF) values for alternative clocks and models tested in BEAST.
The best model is marked in bold.
We thank Bernard Goffinet for providing working facilities to BV during her stay at the University of Connecticut, especially for the use of DNA sequences and vouchers of some Californian specimens used in the description of Orthotrichum kellmanii, and William R. Buck for his help to BV at the NY Herbarium. We also thank N. Matzke for assistance with BioGeoBEARS analyses, M. Flagmeier and S. Young for revising the English text, as well as to two anonymous reviewers and C. Matt Guilliams for their comments.
- 1. Raven PH, Axelrod DI. Angiosperm biogeography and past continental movements. Annals of the Missouri Botanical Garden. 1974;61: 539–673.
- 2. Cowie RH, Holland BS. Dispersal is fundamental to biogeography and the evolution of biodiversity on oceanic islands. Journal of Biogeography. 2006;33: 193–198.
- 3. Renner S. Plant dispersal across the Tropical Atlantic by wind and sea currents. International Journal of Plant Sciences. 2004;165: S23–S33.
- 4. Queiroz A de. The resurrection of oceanic dispersal in historical biogeography. Trends in Ecology & Evolution. 2005;20: 68–73.
- 5. Kadereit JW, Baldwin BG. Western Eurasian–western North American disjunct plant taxa: The dry-adapted ends of formerly widespread north temperate mesic lineages—and examples of long-distance dispersal. Taxon. 2012;61: 3–17.
- 6. Christenhusz MJM, Chase MW. Biogeographical patterns of plants in the Neotropics–dispersal rather than plate tectonics is most explanatory. Botanical Journal of the Linnean Society. 2013;171: 277–286.
- 7. Vargas P, Nogales M, Jaramillo P, Olesen JM, Traveset A, Heleno R. Plant colonization across the Galápagos Islands: success of the sea dispersal syndrome. Botanical Journal of the Linnean Society. 2014;174: 349–358.
- 8. Sanmartín I, Van Der Mark P, Ronquist F. Inferring dispersal: a Bayesian approach to phylogeny-based island biogeography, with special reference to the Canary Islands. Journal of Biogeography. 2008;35: 428–449.
- 9. Baldwin BG, Wagner WL. Hawaiian angiosperm radiations of North American origin. Annals of Botany. 2010;105: 849–879. pmid:20382966
- 10. Gillespie RG, Baldwin BG, Waters JM, Fraser CI, Nikula R, Roderick GK. Long-distance dispersal: a framework for hypothesis testing. Trends in Ecology & Evolution. 2012;27: 47–56. pmid:22014977
- 11. Alsos IG, Ehrich D, Eidesen PB, Solstad H, Westergaard KB, Schönswetter P, et al. Long-distance plant dispersal to North Atlantic islands: colonization routes and founder effect. AoB PLANTS. 2015;7: plv036. pmid:25876627
- 12. Vargas P, Arjona Y, Nogales M, Heleno RH. Long-distance dispersal to oceanic islands: success of plants with multiple diaspore specializations. AoB PLANTS. 2015;7: plv073. pmid:26174146
- 13. Carine MA, Russell SJ, Santos-Guerra A, Francisco-Ortega J. Relationships of the Macaronesian and Mediterranean floras: molecular evidence for multiple colonizations into Macaronesia and back-colonization of the continent in Convolvulus (Convolvulaceae). American Journal of Botany. 2004;91: 1070–1085. pmid:21653463
- 14. Vanderpoorten A, Laenen B, Rumsey FJ, Gonzalez-Mancebo JM, Gabriel R, Carine MA. Dispersal, diversity and evolution of the Macaronesian cryptogamic floras. In: Bramwell D, Caujapé-Castells J, editors. The biology of island floras. 2nd ed. Cambridge: Cambridge University Press; 2011. pp. 338–364.
- 15. Vanderpoorten A, Goffinet B. Introduction to bryophytes. Cambridge: Cambridge University Press; 2009.
- 16. Medina NG, Draper I, Lara F. Biogeography of mosses and allies: does size matter? In: Fontaneto D, editor. Biogeography of microscopic organisms, is everything small everywhere? Cambridge: Cambridge University Press; 2011. pp. 209–233.
- 17. McDaniel SF, Shaw AJ. Phylogeographic structure and cryptic speciation in the trans-antarctic moss Pyrrhobryum mnioides. Evolution. 2003;57: 205–215. pmid:12683518
- 18. Heinrichs J, Lindner M, Groth H, Hentschel J, Feldberg K, Renker C, et al. Goodbye or welcome Gondwana?–insights into the phylogenetic biogeography of the leafy liverwort Plagiochila with a description of Proskauera, gen. nov. (Plagiochilaceae, Jungermanniales). Plant Systematics and Evolution. 2006;258: 227–250.
- 19. Patiño J, Wang J, Renner MAM, Gradstein SR, Laenen B, Devos N, et al. Range size heritability and diversification patterns in the liverwort genus Radula. Molecular Phylogenetics and Evolution. 2017;106: 73–85. pmid:27664347
- 20. Muñoz J, Felicísimo ÁM, Cabezas F, Burgaz AR, Martínez I. Wind as a long-distance dispersal vehicle in the Southern Hemisphere. Science. 2004;304: 1144–1147. pmid:15155945
- 21. Lewis LR, Rozzi R, Goffinet B. Direct long-distance dispersal shapes a New World amphitropical disjunction in the dispersal-limited dung moss Tetraplodon (Bryopsida: Splachnaceae). Journal of Biogeography. 2014;41: 2385–2395.
- 22. Scheben A, Bechteler J, Lee GE, Pócs T, Schäfer-Verwimp A, Heinrichs J. Multiple transoceanic dispersals and geographical structure in the pantropical leafy liverwort Ceratolejeunea (Lejeuneaceae, Porellales). Journal of Biogeography. 2016;43: 1739–1749.
- 23. Carter BE, Larraín J, Manukjanová A, Shaw B, Shaw AJ, Heinrichs J, et al. Species delimitation and biogeography of a southern hemisphere liverwort clade, Frullania subgenus Microfrullania (Frullaniaceae, Marchantiophyta). Molecular Phylogenetics and Evolution. 2017;107: 16–26. pmid:27744015
- 24. Vanderpoorten A, Devos N, Goffinet B, Hardy OJ, Shaw AJ. The barriers to oceanic island radiation in bryophytes: insights from the phylogeography of the moss Grimmia montana. Journal of Biogeography. 2008;35: 654–663.
- 25. Patiño J, Medina R, Vanderpoorten A, González-Mancebo JM, Werner O, Devos N, et al. Origin and fate of the single-island endemic moss Orthotrichum handiense. Journal of Biogeography. 2013;40: 857–868.
- 26. Patiño J, Vanderpoorten A. Macaronesia is a departure gate of anagenetic speciation in the moss genus Rhynchostegiella. Journal of Biogeography. 2015;42: 2122–2130.
- 27. Szövényi P, Terracciano S, Ricca M, Giordano S, Shaw AJ. Recent divergence, intercontinental dispersal and shared polymorphism are shaping the genetic structure of amphi-Atlantic peatmoss populations. Molecular Ecology. 2008;17: 5364–5377. pmid:19121003
- 28. Stenøien HK, Shaw AJ, Shaw B, Hassel K, Gunnarsson U. North American origin and recent European establishments of the amphi-Atlantic peat moss Sphagnum angermanicum. Evolution. 2011;65: 1181–1194. pmid:21073451
- 29. Draper I, Hedenäs L, Stech M, Patiño J, Werner O, González-Mancebo JM, et al. How many species of Isothecium (Lembophyllaceae, Bryophyta) are there in Macaronesia? A survey using integrative taxonomy. Botanical Journal of the Linnean Society. 2015;177: 418–438.
- 30. Shaw J. Biogeographic patterns and cryptic speciation in bryophytes. Journal of Biogeography. 2001;28: 253–261.
- 31. Heinrichs J, Hentschel J, Feldberg K, Bombosch A, Schneider H. Phylogenetic biogeography and taxonomy of disjunctly distributed bryophytes. Journal of Systematics and Evolution. 2009;47: 497–508.
- 32. Medina R, Lara F, Goffinet B, Garilleti R, Mazimpaka V. Unnoticed diversity within the disjunct moss Orthotrichum tenellum s.l. validated by morphological and molecular approaches. Taxon. 2013;62: 1133–1152.
- 33. Medina R, Lara F, Goffinet B, Garilleti R, Mazimpaka V. Integrative taxonomy successfully resolves the pseudo-cryptic complex of the disjunct epiphytic moss Orthotrichum consimile s.l. (Orthotrichaceae). Taxon. 2012;61: 1180–1198.
- 34. Renner MAM, Devos N, Patino J, Brown EA, Orme A, Elgey M, et al. Integrative taxonomy resolves the cryptic and pseudo-cryptic Radula buccinifera complex (Porellales, Jungermanniopsida), including two reinstated and five new species. PhytoKeys. 2013;27: 1–113. pmid:24223490
- 35. Hedenäs L, Désamoré A, Laenen B, Papp B, Quandt D, González-Mancebo JM, et al. Three species for the price of one within the moss Homalothecium sericeum s.l. Taxon. 2014;63: 249–257.
- 36. Heinrichs J, Feldberg K, Bechteler J, Scheben A, Czumaj A, Pócs T, et al. Integrative taxonomy of Lepidolejeunea (Jungermanniopsida: Porellales): Ocelli allow the recognition of two neglected species. Taxon. 2015;64: 216–228.
- 37. Caparrós R, Lara F, Draper I, Mazimpaka V, Garilleti R. Integrative taxonomy sheds light on an old problem: the Ulota crispa complex (Orthotrichaceae, Musci). Botanical Journal of the Linnean Society. 2016;180: 427–451.
- 38. Sim-Sim M, Afonina OM, Almeida T, Désamoré A, Laenen B, Garcia Céa, et al. Integrative taxonomy reveals too extensive lumping and a new species in the moss genus Amphidium (Bryophyta). Systematics and Biodiversity. 2017;15: 1–13.
- 39. Lewinsky-Haapasaari J, Norris DH. Orthotrichum shevockii (Orthotrichaceae), a New Moss Species from the Southern Sierra, California. The Bryologist. 1998;101: 435–438.
- 40. Norris DH, Shevock JR, Goffinet B. Orthotrichum kellmanii (Bryopsida, Orthotrichaceae), a remarkable new species from the central coast of California. The Bryologist. 2004;107: 209–214.
- 41. Hutsemékers V, Vieira CC, Ros RM, Huttunen S, Vanderpoorten A. Morphology informed by phylogeny reveals unexpected patterns of species differentiation in the aquatic moss Rhynchostegium riparioides s.l. Molecular Phylogenetics and Evolution. 2012;62: 748–755. pmid:22138204
- 42. Vanderpoorten A, Patiño J, Dirkse G, Blockeel T, Hedenäs L. Early divergence of an Azorean endemic species in the moss genus Rhynchostegiella (Brachytheciaceae). Phytotaxa. 2015;210: 60–69.
- 43. Lara F, Garilleti R, Medina R, Mazimpaka V. A new key to the genus Orthotrichum Hedw. in Europe and the Mediterranean Region. Cryptogamie, Bryologie. 2009;30: 129–142.
- 44. Lara F, Garilleti R. Orthotrichum. In: Guerra J, Cano MJ, Brugués M, editors. Flora Briofitica Iberica Volumen V Orthotrichales: Orthotrichaceae; Hedwigiales: Hedwigiaceae; Leucodontales: Fontinalaceae, Climaciaceae, Anomodontaceae, Cryphaeaceae, Leptodontaceae, Leucodontaceae, Neckeraceae; Hookeriales: Hypopterygiaceae, Hookeriaceae, Leucomiaceae, Pilotrichaceae. Murcia: Universidad de Murcia—Sociedad Española de Briología; 2014. pp. 50–135.
- 45. Vigalondo B, Lara F, Draper I, Valcarcel V, Garilleti R, Mazimpaka V. Is it really you, Orthotrichum acuminatum? Ascertaining a new case of intercontinental disjunction in mosses. Botanical Journal of the Linnean Society. 2016;180: 30–49.
- 46. Lewinsky J. A synopsis of the genus Orthotrichum Hedw. (Musci, Orthotrichaceae). Bryobrothera. 1993;2: 1–59.
- 47. Vitt DH. A Revision of the genus Orthotrichum in North America, North of Mexico. J Cramer L, editor. Lehre; 1973.
- 48. Kampstra P. Beanplot: A boxplot alternative for visual comparison of distributions. Journal of Statistical Software, Code Snippets. 2008;28: 1–9.
- 49. R Core Team. R: A language and environment for statistical computing. [Internet]. Vienna, Austria: R Foundation for Statistical Computing; 2016. Available: http://www.R-project.org/
- 50. Chiang T-Y, Schaal BA, Peng C-I. Universal primers for amplification and sequencing a noncoding spacer between the atpB and rbcL genes of chloroplast DNA. Botanical Bulletin of Academia Sinica. 1998;39: 245–250.
- 51. Nadot S, Bajon R, Lejeune B. The chloroplast gene rps4 as a tool for the study of Poaceae phylogeny. Plant Systematics and Evolution. 1994;191: 27–38.
- 52. Souza-Chies TT, Bittar G, Nadot S, Carter L, Besin E, Lejeune B. Phylogenetic analysis of Iridaceae with parsimony and distance methods using the plastid gene rps4. Plant Systematics and Evolution. 1997;204: 109–123.
- 53. Taberlet P, Gielly L, Pautou G, Bouvet J. Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Molecular Biology. 1991;17: 1105–1109. pmid:1932684
- 54. Fiedorow P, Odrzykoski I, Szweykowska-Kulinska Z. Phylogenetic studies of liverworts using molecular biology techniques. Plant Cytogenetics. Katowice: Silesian University Press; 1998. pp. 244–249. Available: http://agro.icm.edu.pl/agro/element/bwmeta1.element.agro-article-6dc9232b-089d-4d20-ac57-e18af022d0d7
- 55. Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, et al. Geneious Basic: An integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics. 2012;28: 1647–1649. pmid:22543367
- 56. Müller J, Müller KF, Neinhuis C, Quandt D. PhyDE: Phylogenetic data editor. [Internet]. Distributed by the authors.; 2006. Available: http://www.phyde.de
- 57. Talavera G, Castresana J. Improvement of phylogenies after removing divergent and ambiguously aligned blocks from protein sequence alignments. Systematic Biology. 2007;56: 564–577. pmid:17654362
- 58. Darriba D, Taboada GL, Doallo R, Posada D. jModelTest 2: more models, new heuristics and parallel computing. Nature Methods. 2012;9: 772–772. pmid:22847109
- 59. Stamatakis A. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics. 2014;30: 1312–1313. pmid:24451623
- 60. Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, Hohna S, et al. MrBayes 3.2: efficient bayesian phylogenetic inference and model choice across a large model space. Systematic Biology. 2012;61: 539–542. pmid:22357727
- 61. Rambout A. FigTree v. 1.4. [Internet]. 2012. Available: http://tree.bio.ed.ac.uk/software/figtree/
- 62. Kelchner SA. The evolution of non-coding chloroplast DNA and its application in plant systematics. Annals of the Missouri Botanical Garden. 2000;87: 482–498.
- 63. Müller DK. SeqState. Applied Bioinformatics. 2012;4: 65–69.
- 64. Simmons MP, Ochoterena H. Gaps as characters in sequence-based phylogenetic analyses. Systematic biology. 2000;49: 369–381. pmid:12118412
- 65. Ronquist F, Huelsenbeck JP, Teslenko M. MrBayes version 3.2 manual: tutorials and model summaries. Available: http://mrbayes.sourceforge.net/mb3.2_manual.pdf. 2011.
- 66. Lanfear R, Calcott B, Ho SYW, Guindon S. PartitionFinder: combined selection of partitioning schemes and substitution models for phylogenetic analyses. Molecular Biology and Evolution. 2012;29: 1695–1701. pmid:22319168
- 67. Drummond AJ, Suchard MA, Xie D, Rambaut A. Bayesian Phylogenetics with BEAUti and the BEAST 1.7. Molecular Biology and Evolution. 2012;29: 1969–1973. pmid:22367748
- 68. Reid NM, Carstens BC. Phylogenetic estimation error can decrease the accuracy of species delimitation: a Bayesian implementation of the general mixed Yule-coalescent model. BMC Evolutionary Biology. 2012;12: 196. pmid:23031350
- 69. Librado P, Rozas J. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics. 2009;25: 1451–1452. pmid:19346325
- 70. Laenen B, Shaw B, Schneider H, Goffinet B, Paradis E, Désamoré A, et al. Extant diversity of bryophytes emerged from successive post-Mesozoic diversification bursts. Nature Communications. 2014;5: 5134. pmid:25346115
- 71. Villarreal JC, Renner SS. A review of molecular-clock calibrations and substitution rates in liverworts, mosses, and hornworts, and a timeframe for a taxonomically cleaned-up genus Nothoceros. Molecular Phylogenetics and Evolution. 2014;78: 25–35. pmid:24792087
- 72. Rambaut A, Suchard MA, Xie D, Drummond AJ. Tracer v1. 6. [Internet]. 2014. Available: http://beast.bio.ed.ac.uk/Tracer
- 73. Lartillot N, Philippe H, Lewis P. Computing Bayes Factors Using Thermodynamic Integration. Systematic Biology. 2006;55: 195–207. pmid:16522570
- 74. Xie W, Lewis PO, Fan Y, Kuo L, Chen M-H. Improving Marginal Likelihood Estimation for Bayesian Phylogenetic Model Selection. Systematic Biology. 2011;60: 150–160. pmid:21187451
- 75. Matzke NJ. Model selection in historical biogeography reveals that founder-event speciation is a crucial process in island clades. Systematic Biology. 2014;63: 951–970. pmid:25123369
- 76. Ree RH, Sanmartín I. Conceptual and statistical problems with the DEC+J model of founder-event speciation and its comparison with DEC via model selection. Journal of Biogeography. 2018;45: 741–749.
- 77. Matzke NJ. Probabilistic historical biogeography: new models for founder-event speciation, imperfect detection, and fossils allow improved accuracy and model-testing. Frontiers of Biogeography. 2013;5. Available: http://escholarship.org/uc/item/44j7n141
- 78. Lara F, Garilleti R, Goffinet B, Draper I, Medina R, Vigalondo B, et al. Lewinskya, a new genus to accommodate the phaneroporous and monoicous taxa of Orthotrichum (Bryophyta, Orthotrichaceae). Cryptogamie, Bryologie. 2016;37: 361–382.
- 79. Garilleti R, Lara F, Mazimpaka V. New differential characters for Orthotrichum rogeri Brid. (Orthotrichaceae, Bryopsida). Nova Hedwigia. 2002;75: 207–216.
- 80. González-Mancebo JM, Romaguera F, Ros RM, Patiño J, Werner O. Bryophyte flora of the Canary Islands: an updated compilation of the species list with an analysis of distribution patterns in the context of the Macaronesian Region. Cryptogamie, Bryologie. 2008;29: 315–357.
- 81. Ros RM, Mazimpaka V, Abou-Salama U, Aleffi M, Blockeel TL, Brugués M, et al. Mosses of the Mediterranean, an annotated checklist. Cryptogamie, Bryologie. 2013;34: 99–283.
- 82. Vanderpoorten A, Rumsey FJ, Carine MA. Does Macaronesia exist? Conflicting signal in the bryophyte and pteridophyte floras. American Journal of Botany. 2007;94: 625–639. pmid:21636431
- 83. Feuerer T, Hawksworth DL. Biodiversity of lichens, including a world-wide analysis of checklist data based on Takhtajan’s floristic regions. Biodiversity and Conservation. 2007;16: 85–98.
- 84. Shaw AJ, Werner O, Ros RM. Intercontinental Mediterranean disjunct mosses: morphological and molecular patterns. American Journal of Botany. 2003;90: 540–550. pmid:21659147
- 85. Werner O, Ros RM, Guerra J, Shaw AJ. Molecular data confirm the presence of Anacolia menziesii (Bartramiaceae, Musci) in southern Europe and its separation from Anacolia webbii. Systematic Botany. 2003;28: 483–489.
- 86. Fernández-Palacios JM, de Nascimento L, Otto R, Delgado JD, García-del-Rey E, Arévalo JR, et al. A reconstruction of Palaeo-Macaronesia, with particular reference to the long-term biogeography of the Atlantic island laurel forests. Journal of Biogeography. 2011;38: 226–246.
- 87. Otto R, Whittaker RJ, Gaisberg M von, Stierstorfer C, Naranjo‐Cigala A, Steinbauer MJ, et al. Transferring and implementing the general dynamic model of oceanic island biogeography at the scale of island fragments: the roles of geological age and topography in plant diversification in the Canaries. Journal of Biogeography. 2016;43: 911–922.
- 88. Devos N, Vanderpoorten A. Range Disjunctions, speciation, and morphological transformation rates in the liverwort genus Leptoscyphus. Evolution. 2009;63: 779–792. pmid:19154356
- 89. Krishnamurti TN. The subtropical jet stream of winter. Journal of Meteorology. 1961;18: 172–191.
- 90. Kuang X, Zhang Y, Huang Y, Huang D. Spatial differences in seasonal variation of the upper-tropospheric jet stream in the Northern Hemisphere and its thermal dynamic mechanism. Theoretical and Applied Climatology. 2014;117: 103–112.
- 91. Van Zanten BO, Gradstein SR. Experimental dispersal geography of neotropical liverworts. Beihefte zur Nova Hedwigia. 1988; 41–94.
- 92. Frahm J-P. Diversity, dispersal and biogeography of bryophytes (mosses). Biodiversity and Conservation. 2008;17: 277–284.
- 93. Wilkinson DM, Koumoutsaris S, Mitchell EAD, Bey I. Modelling the effect of size on the aerial dispersal of microorganisms. Journal of Biogeography. 2012;39: 89–97.
- 94. Hutsemékers V, Szövényi P, Shaw AJ, González-Mancebo J-M, Muñoz J, Vanderpoorten A. Oceanic islands are not sinks of biodiversity in spore-producing plants. PNAS. 2011;108: 18989–18994. pmid:22084108
- 95. Laenen B, Désamoré A, Devos N, Shaw AJ, González-Mancebo JM, Carine MA, et al. Macaronesia: a source of hidden genetic diversity for post-glacial recolonization of western Europe in the leafy liverwort Radula lindenbergiana. Journal of Biogeography. 2011;38: 631–639.
- 96. Patiño J, Bisang I, Hedenäs L, Dirkse G, Bjarnason ÁH, Ah-Peng C, et al. Baker’s law and the island syndromes in bryophytes. Journal of Ecology. 2013;101: 1245–1255.
- 97. Patiño J, Carine M, Mardulyn P, Devos N, Mateo RG, González-Mancebo JM, et al. Approximate bayesian computation reveals the crucial role of oceanic islands for the assembly of continental biodiversity. Systematic Biology. 2015;64: 579–589. pmid:25713307
- 98. Fernández-Palacios JM. The islands of Macaronesia. In: Serrano A.R.M., Borges P.A.V., Boieiro M. & Oromí P., editor. Terrestrial arthropods of Macaronesia Biodiversity, ecology and evolution. Sociedad Portuguesa de Entomología; 2011.