New Taxonomic Arrangement of Dicranella s.l. and Aongstroemia s.l. (Dicranidae, Bryophyta)

The recent molecular phylogenetic study of the families Aongstroemiaceae and Dicranellaceae, which resolved the genera Aongstroemia and Dicranella as polyphyletic, indicated the need for changes in their circumscription and provided new morphological evidence to support the formal description of newly recognized lineages. Following up on these results, the present study adds another molecular marker, the highly informative trnK–psbA region, to a subset of previously analyzed taxa and presents molecular data from newly analyzed austral representatives of Dicranella and collections of Dicranella-like plants from North Asia. The molecular data are linked with morphological traits, particularly the leaf shape, tuber morphology, and capsule and peristome characters. Based on this multi-proxy evidence, we propose three new families (Dicranellopsidaceae, Rhizogemmaceae, and Ruficaulaceae) and six new genera (Bryopalisotia, Calcidicranella, Dicranellopsis, Protoaongstroemia, Rhizogemma, and Ruficaulis) to accommodate the described species according to the revealed phylogenetic affinities. Additionally, we amend the circumscriptions of the families Aongstroemiaceae and Dicranellaceae, as well as the genera Aongstroemia and Dicranella. In addition to the monotypic Protoaongstroemia that contains the newly described dicranelloid plant with a 2–3-layered distal leaf portion from Pacific Russia, P. sachalinensis, Dicranella thermalis is described for a D. heteromalla-like plant from the same region. Fourteen new combinations, including one new status change, are proposed.

It was thus perhaps not very surprising that a molecular-phylogenetic study of the northern temperate genera of Aongstroemia Bruch & Schimp. and Dicranella (Müll. Hal.) Schimp. [10] revealed a striking polyphyly in the existing delimitation of these genera. The traditional morphological circumscription of the genus Dicranella included plants of small size with a stem central strand, elongated linear-lanceolate to subulate leaves, costae with guide cells in the cross-section, predominantly dioicous sexual condition, and dicranoid peristome [16][17][18][19][20]. This allowed for a considerable variability of the other gametophytic and sporophytic traits, which found a reflection in the molecular differentiation of the thirteen analyzed species of the genus that were found in seven different phylogenetic lineages of haplolepidous mosses (the subclass of Dicranidae), three of which could be considered orphaned in the system of the representatives of dicranids analyzed to date. On the other hand, the genus Aongstroemia, originally introduced for a single species, A. longipes, was soon substantially expanded [16] to include most species of the modern Dicranella s.lat., and later again reduced [21] to harbor species sharing the julaceous habit originating from the ovate leaves, which are broadly rounded to acute or slightly attenuate at the tips and appressed to the stem. Despite this restriction in the generic concept of Aongstroemia, the rate of cryptic molecular diversity was similar to that revealed in Dicranella; Bonfim-Santos et al. [10] showed that the three analyzed species (out of the 11 accepted names in the genus) appear in three lineages belonging to three currently recognized families. Although the polyphyly of both genera was demonstrated quite convincingly, this study has not yet resulted in a taxonomic treatment. The major reason for this was the insufficient taxonomic sampling, particularly in Dicranella: of the 161 accepted species according to the Tropicos database [22], plus the 47 accepted species in Leptotrichella and five accepted names in Anisothecium, less than one tenth have been phylogenetically studied, which means that the generic and familial assignment of the bulk of the species remains pending after the splitting of Dicranella according to the obtained results. The recent description of a previously unknown dicranelloid moss from SW Portugal, which necessitated the erection of a new genus, Neodicranella Porley & Fedosov, following the assessment of molecular affinities [23], confirms that the diversity of dicranelloid mosses has not been fully captured, even in the relatively well-surveyed Europe. Although a thorough taxonomic revision of all included taxa and checking of the type material would be most appropriate, such a revision will hardly be possible in the near future for a complex and species-rich genus such as Dicranella, where many of the accepted species are based on a few historical collections from southern tropical countries, whose localities are difficult to access. Nomenclaturally, it would nevertheless be more relevant if the types of the generic names placed earlier in synonymy with Dicranella and Aongstroemia were designated, and their phylogenetic affinities known; a task that has been partly accomplished and the missing pieces of information do not threaten the stability of most proposed taxonomic solutions. Moreover, further accepting clearly polyphyletic taxa is in our opinion a less desirable alternative than establishing a baseline for further development of a phylogeneticallybased system of haplolepidous mosses, which can be further elaborated as soon as new information appears.
Although we generally followed the molecular sampling of Bonfim Santos et al. [10], who employed only organellar markers, plastid trnL-trnF and rps4 and mitochondrial nad5 intron 2, with respect to the absence of reasonably informative nuclear markers that would be generally used in subclass-spanning phylogenetic studies of mosses, we deepened the molecular sampling to include the highly informative trnK-psbA region, which has been used with success, e.g., in the treatments [24] or [25], and also sampled the two variable spacers flanking the gene for tRNA-Thr, which is located between the previously sampled trnL-trnF and rps4. The purpose of this was the testing of the weakly supported deeper nodes of Dicranidae, which was one of the unresolved questions in [10] that we aimed to address in our novel analyses. Secondly, we broadened the sampling in several critical groups, such as the South American representatives of Dicranella assigned to Anisothecium by Mitten [26], accessions of Neodicranella and several putatively new taxa of unclear affinity from North Asia, which were not included in [10]. We also deliver arguments for treating the two varieties of Dicranella schreberiana (Hedw.) Hilf. ex H.A. Crum & L.E. Anderson at the species level and resolve the molecular variation between D. varia (Hedw.) Schimp. and D. howei Renauld & Cardot. This study thus represents a state-of-the-art taxonomic treatment based on both previously published and newly obtained molecular and morphological data, which is expected to be updated, particularly for the southern and tropical taxa.

Results
The concatenated matrix consisted of 6381 aligned sites, of which 563 belonged to L partition, 975 to T partition, 728 to R, 2899 to K, and 1216 to N. Indels scored for the L, R, and N partitions yielded an additional 193 binary sites, and 251 indels were scored for the T and 218 for the K partition. The partitions corresponding to the dataset used by Bonfim-Santos et al. [10], i.e., L, R, and N, contained 907 variable and 603 parsimony-informative sites, the T partition had 526 variable and 346 parsimony-informative sites, and the K partition contained 1410 variable and 901 parsimony-informative sites.
The trees inferred from the combined L, R, and N data, and those with an added T region dataset, had essentially identical topology, with generally higher support values from the expanded dataset. The addition of indel data generally further improved the support values, without changes in topology at the supported nodes, however only when T indel data were not considered. At the same time, the trees estimated from the separate analysis of K data, which again had generally higher support node values in the version with SIC-coded indels included, yielded a topological incongruence compared to the trees derived from LTRN data, with respect to the estimated affinities of Chrysoblastella chilensis, Neodicranella hamulosa, and Archidium + Leucobryaceae and Grimmiales clades. Hence, we present here the results representing the total evidence of the fully concatenated dataset (LTRKN) in Figure 1, and the trees resulting from partial analyses of the LTRN and K data are presented as the supplementary Figures S1 and S2. In the following description of the results, we only comment on results differing from those obtained by Bonfim-Santos et al. [10], which was used as a reference.
In contrast to this study, after reconsidering the reading of several bases in the raw chromatograms, the position of Pseudoditrichum mirabile was newly assessed as unresolved among the basal protohaplolepidous clades (Flexitrichum, Scouleria + Drummondia + Hymenoloma clade) and the clade containing all other analyzed taxa, i.e., Bryoxiphium + rest of Dicranales incl. Grimmiales), rather than being found in the basal grade of Dicranales. This position was shared by the analyses resulting from the LTRN and K datasets ( Figures S1 and 2). Additional Dicranella staphylina accessions from northeastern Asia (Putorana, Yakutia, and Khabarovsk Territory) were found in the maximally supported clade with European accessions, although three accessions collected in the heart of the permafrost zone of northeast Asia (BF59, FDt107, and 116) differed in several substitutions, despite the absence of notable morphological differences, except for the slightly more robust habit. European accessions of the previously unanalyzed D. humilis from the Czech Republic and Russian Leningrad Province proved identical and distinct from the rest of the analyzed accessions of D. rufescens and Far Eastern accessions earlier referred to D. humilis based on their seemingly inclined capsules (yet all sporulating collections from that area were collected with immature sporophytes, which prevented the assessment of a basal membrane height). Additionally, the accession FDt119 from plants morphologically approaching D. humilis collected in Sakhalin Island was found to be molecularly distinct, in a sister position to the maximally supported D. rufescens clade. The topology between the maximally supported D. rufescens + humilis clade, D. crispa + subulata clade, and the crown clade of Dicranales has not been resolved, even with our deeper molecular sampling, yet the analysis of trnK data alone (Supplementary Figure S2) yielded an unsupported clade containing D. rufescens + humilis plus D. crispa + subulata lineages. Stronger support (BI PP 1 but without support from ML) was found for the clade containing these two lineages and the clade containing the core Dicranales. The ambiguous affinities of these two lineages might partly result from the ambiguous affinities of the Archidiaceae + Leucobryaceae clade, which was resolved in the sister position to the clade containing Grimmiales based on the LTRN data (now with BS 94/PP 1 support), as opposed to the unsupported (BS < 50/PP 0.79) sister position of the Archidiaceae + Leucobryaceae clade to the rest of the Dicranales, excluding D. rufescens + humilis and D. crispa + subulata lineages, in the analysis of K data alone. The signal from LTRN data was stronger than that of K data, weakening the support for Archidiaceae/Leucobryaceae plus Grimmiales clade to BS 67/PP 0.94 in the combined LTRKN analysis. Figure 1. Maximum likelihood tree, inferred from the concatenated data matrix from the chloroplast trnF-trnS and trnK-psbA and mitochondrial nad5 intron 2 sequence data (LTRKN dataset) of selected species of Dicranidae, focused on the genera Dicranella and Aongstroemia, rooted with Pseudoditrichum mirabile. Bootstrap support values higher than 60 inferred from ML analyses, without and with indel coding, and posterior probabilities higher than 0.7 inferred from BI, without and with indel coding, are shown above the branches; hyphens in place of support values denote lower support of the node, while a blank space indicates that the node is absent from the topology inferred from the particular analysis; maximally supported nodes are indicated by solid lines and asterisks. Newly studied terminals, as well as terminals for which at least one marker was obtained de novo, are printed in bold, and terminals for which the trnK-psbA sequence is available are underlined. For details, see Appendix A. Support for the core Dicranales clade has grown substantially in both ML (BS 85-100 according to the dataset) and BI (PP 1), and similar support was obtained for the sister relationship between Amphidium and the rest of the core Dicranales clade. The same is true for the three larger clades within the core Dicranales, the first one including Schistostega, Rhabdoweisiaceae, Ditrichaceae, Pottiaceae, Bruchiaceae, and a maximally supported clade containing accessions of Rhamphidium, Symblepharis krausei, and Dicranella vaginata; the second containing Dicranaceae, Fissidentaceae, Chrysoblastella, Bryowijkia, and Dicranellaceae; and the third containing Aongstroemiaceae. Support for the core Dicranales clade has grown substantially in both ML (BS 85-100 according to the dataset) and BI (PP 1), and similar support was obtained for the sister relationship between Amphidium and the rest of the core Dicranales clade. The same is true for the three larger clades within the core Dicranales, the first one including Schistostega, Rhabdoweisiaceae, Ditrichaceae, Pottiaceae, Bruchiaceae, and a maximally supported clade containing accessions of Rhamphidium, Symblepharis krausei, and Dicranella vaginata; the second containing Dicranaceae, Fissidentaceae, Chrysoblastella, Bryowijkia, and Dicranellaceae; and the third containing Aongstroemiaceae.
The nearly maximally supported (BS 98-100/PP 1) Dicranellaceae clade does not contain northern representatives of Dicranella s.l., except for the genus in its amended delimitation itself, and the newly analyzed accession of Dicranella polii from Madagascar appeared in a poorly supported clade with two accessions of the genus Garckea, which itself appeared nested within a maximally supported clade containing accessions of Microcampylopus, Campylopodium, and Leptotrichella flaccidula. This clade is resolved in sister position to the clade containing Aongstroemia filiformis s.lat. accessions. The Madagascan and Reunion accessions of that species appear molecularly distinct from the Neotropical accessions, which earlier were recognized as A. jamaicensis Müll. Hal. Within the Dicranella s.str. clade, accessions of D. cerviculata formed a clade sister to the remaining accessions of the genus. Within the latter, two molecularly identical accessions of D. heteromalla/Campylopus pyriformis-like plants from Southern Kuril Islands are separated in a maximally supported clade and are described below as Dicranella thermalis. The maximally supported sister clade to the D. thermalis clade consists of (1) the nearly maximally supported clade containing accessions of D. curvipes and (2) the unsupported clade containing the accessions of D. heteromalla. Within this clade, two smaller clades can be recognized, one with high support containing both European and non-European plants referable to this species, and the other unsupported clade containing only accessions from the Russian Far East, which might in the future receive formal status upon a detailed account of their molecular and morphological variability.
Within the now maximally (except for ML BS 87 in the LTRN-based tree) supported Aongstroemiaceae sensu [10], the sister group relationship between Dicranella varia s.lat. + D. howei + D. pacifica clade and the rest of the taxa was confirmed, yet the clade containing Aongstroemiaceae without the D. varia group now only has weak support (BS 63-78 only from K data and PP 0.67-0.97 according to dataset and indel scoring), with respect to the inclusion of the newly analyzed basalmost lineage containing the plants described below as a new genus, Protoangstroemia. Within the D. varia group, four maximally or nearly so supported lineages could be recognized. Apart from D. varia and D. howei, the newly analyzed D. pacifica appeared sister to D. varia + howei clade, and four accessions, containing the RF42, which earlier was assigned to D. varia but now is referred to Dicranella varia var. obtusifolia Berggren raised to the species rank below, form a lineage sister to the rest of the entire D. varia group. Within the Aongstroemiaceae s.str. clade, the basal grade consists of Protoangstroemia and the maximally supported lineages of Diobelonella, Dichodontium, and Neodicranella. However, the clade containing accessions of Neodicranella appears in a different position in the analysis of LTRN and K data (cf. Figures S1 and S2), essentially unresolved in the grade between Diobelonella and Dichodontium according to LTRN data but deeply nested within Aongstroemia s.lat., sister to the Dicranella grevilleana + Aongstroemia longipes clade according to K data. The crown clade (Aongstroemia s. lat.) contains the Dicranella grevilleana + Aongstroemia longipes, Dicranella schreberiana var. robusta, D. schreberiana var. schreberiana, Hygrodicranum bolivianum, H. herrerae, and Dicranella campylophylla + D. hookeri + Polymerodon andinus clades. Additional accessions of Aongstroemia longipes, Dicranella grevilleana, D. schreberiana var. schreberiana, D. campylophylla, and D. hookeri support the distinctness of Aongstroemia longipes from Dicranella grevilleana, Dicranella schreberiana var. robusta from var. schreberiana, and of D. campylophylla from D. hookeri, yet a more detailed study of taxa in this group needs to be performed in the future with respect to one isolated accession of Dicranella schreberiana s.lat. from Russia (RF40), the similarly isolated accession of D. hookeri RF65, and the nested position of D. campylophylla TJH13 within the clade, which otherwise contained specimens referable to D. hookeri.

Discussion
Our trees are largely congruent with those published by [10,23], yet bring more resolution to the relationships among the haplolepidous lineages, identify affinities of the seven previously unsampled species, and verify the previously assessed affinities using additional accessions of previously insufficiently sampled taxa.
Dicranella staphylina. The totally orphaned position of Dicranella staphylina within the system of haplolepidous mosses came as one the most surprising results of the phylogenetic reconstruction by [10]. This moss has to date been known from Europe essentially only from its gametophytic stage, which does not have any distinct autapomorphic traits. Several immature sporophytes have only been observed by [27]. They had yellow to orange seta (speculated to be red at maturity), erect, symmetrical, smooth capsules with irregular and incrassate exothecial cells and few stomata, longitudinally striate peristome teeth bifid to the middle, and unmatured spores 15-20 µm. The character of annulus was not mentioned. Neither of these characters is outstanding among northern Dicranella s.lat. species. Unexpectedly, sporulating plants not matching the description of any other known species of the genus were recently discovered in the north Siberian Putorana Plateau. They had yellow setae, asymmetric furrowed capsules, bright red peristome, and well-developed revoluble annulus ( Figure 2). Surprisingly, the molecular barcoding of these plants revealed their identity with the previously analyzed accessions of D. staphylina and, indeed, the gametophytic characters matched the European material except for plants from Khabarovsk Territory, discovered upon the subsequent herbarium revision, which lack the characteristic rhizoidal tubers. The difference in the capsule shape as compared to [27] might result from the ontogenetic stage, when young straight and smooth capsules may also become curved and furrowed with age. With respect to the isolated position of Dicranella staphylina, we propose a new genus and family to accommodate it.
The sister group relationship between D. staphylina and the clade containing both the rest of the order Dicranales (including Archidiaceae and Leucobryaceae) and the species currently recognized within the order Grimmiales opens the question of the ordinal placement of the lineage containing D. staphylina. While the resolved topology based on LTRN and K data differs in the assessment of affinities of the Archidium + Leucobryaceae clade, both topologies agree on the nested position of the currently recognized Grimmiales within Dicranales, should the D. staphylina-lineage remain in Dicranales. D. staphylina has a fairly typical dicranoid peristome with triangular, in basal and median part longitudinally striolate rather than filiform teeth, split to half of their length into unequal lobes, without a basal membrane (Figure 2), which clearly fits the description of the dicranoid peristome type by [28]. This favors the idea of including this lineage in the delimitation of Dicranales, rather than establishing an isolated new order to accommodate it, suggesting that the dicranoid peristome is the plesiomorphic character state for the whole large lineage, from which the more derived peristomes in Grimmiales and Pottiaceae evolved (see also [9] for a discussion of the secondarily modified peristomes in, e.g., Glyphomitrium Brid. or Pseudoblindia Fedosov, M. Stech & Ignatov of Rhabdoweisiaceae). The idea of a broad Dicranales, with the currently recognized Grimmiales being lowered to the rank of suborder, has further support from the absence of any derived morphological trait that is typical for the earlier diverging lineages, such as Catoscopium Brid., Distichium Bruch & Schimp., Bryoxiphium Mitt., or Pseudoditrichum Steere & Z. Iwats.
Dicranella rufescens and D. humilis. These two species share the red color of their stems [29], rather sparsely foliated stems with leaves hardly homomallous or secund, plane leaf margins (which however can be narrowly recurved on one side in D. humilis), and a weakly differentiated costa, especially in D. rufescens. A unique character of D. rufescens among other ex-Dicranella species is the high basal membrane (up to 10 rows), while the membrane of D. humilis does not extend four rows; basal membranes extending three rows are nevertheless rare in all other species except D. varia. Both species markedly differ in their capsule shape (characteristically straight and symmetric in D. rufescens, while inclined, slightly curved, and asymmetric in D. humilis). It was therefore important to confirm that D. humilis is indeed closely related to D. rufescens, which we accomplished. All Asian specimens referred to D. humilis on the available morphological grounds appeared in the D. rufescens clade. Thus, although our study confirmed the species status of D. humilis, further morphological study of additional Asian specimens is needed to clarify the differentiation of D. humilis and D. rufescens. The isolated position of the clade precludes any other taxonomic solution except for establishing a new genus for the two species of the lineage, with the familial placement being somewhat ambiguous. The very weak clustering with the Dicranella subulata + crispa based in trnK data ( Figure S2) might favor creating a family harboring both these lineages; however, the total evidence from all studied regions ( Figure 1) does not support this solution and favors the creation of a separate family for this monogeneric lineage. This is supported by the salient morphological differences between the lineages (absent versus well-developed revoluble annulus, basal membrane 3-10 versus 1-3 rows, sparse leaves, never clasping and shouldered versus leaves dense and contiguous, at least perichaetial leaves clasping) in the absence of other nonhomoplasic common characters.
Dicranella subulata and D. crispa. The morphological synapomorphies of this lineage were discussed at length by [10]. While the affinities with the preceding lineage have not been convincingly resolved, the same set of arguments can be used for segregating the two known representatives of this lineage to a separate genus and family (see Taxonomic treatment). The previous names adopted for Dicranella crispa and D. subulata include either names that are in use for other distinct genera, or the illegitimate genus names Dicranodon Béhéré and Leptotrichum Hampe ex Müll. Hal. Similarly, the possibility of raising Dicranella sect. Pseudodicranella Nyholm to the genus level is prevented by the name being invalid with respect to a missing Latin description and is illegitimate, as it includes the conserved type of Dicranella, D. heteromalla. Therefore, we propose to erect a new genus and family name for this group in the Taxonomy section.
Dicranellaceae. In agreement with [10], we concur with the proposal of reducing the delimitation of Dicranellaceae to only include members of the Dicranella heteromalla group, Microcampylopus/Leptotrichella/Garckea/Campylopodium polytomy, Aongstroemia filiformis s.lat., Eccremidium, and Cladophascum. With its remarkably distinct morphology [30], Bryowijkia, albeit robustly supported molecularly as a sister group to the above-specified assemblage, should remain separate at the family rank, as proposed by [31]. The question of the inclusion of Trichodontium falcatum (R. Br. bis) Fife remains open. The species was merged with Kiaeria pumila (Mitt.) Ochyra by [32], which was resolved as a member of Arctoa within Rhabdoweisiaceae [9]. However, the only available Trichodontium GenBank accessions AF435304/AF435353 from the specimen Streimann 51155 appeared in the clade with Leptotrichella flaccidula and Campylopodium medium in the analysis of [10], which suggests the possibility of an incorrect identification in one of the treatments, and the matter needs to be revisited in the future.
The maximally supported clade containing accessions of Aongstroemia filiformis s.lat. sister to the clade of other tropical Dicranellaceae also leaves us with no other option than to establish a new genus to accommodate it, as the species has never been included in genera other than Aongstroemia, Dicranella, Dicranum, and Thysanomitrion, which cannot be used for this purpose. This solution is put into effect later in the text. The geographically meaningful pattern of molecular variability, as supported by the analysis of an additional Aongstroemia filiformis specimen from Madagascar, seems to support the resurrection of A. jamaicensis from the synonymy of A. filiformis, but this task requires additional sampling and morphological study.
The well-supported tropical Dicranellaceae clade, consisting of analyzed accessions of Aongstroemia filiformis s.lat., Microcampylopus, Leptotrichella, Garckea, and Campylopodium also contains a single analyzed specimen of Dicranella polii. Its closest affinities were revealed to be with the previously analyzed Garckea species, with which it forms a clade moderately supported from ML and not supported from BI (BS 77-81, PP 0.86-0.88), nested within the well-supported clade containing the tropical Dicranellaceae, except for Aongstroemia filiformis s.lat. While the combination of D. polii under Garckea would make sense from a nomenclatural point of view, as the latter appears to be the oldest available generic name in this clade (Leptotrichella incl. the younger Microdus, Microcampylopus), the sporophytic characters currently used for delimitation between Garckea, Leptotrichella (generally considered synonymous to Dicranella [20,33,34], and Microcampylopus do not match the revealed phylogenetic affinities, and hence we prefer to postpone this taxonomic decision, pending a deeper sampling in this lineage. This brings, however, another piece of evidence that the tropical species referred previously to Dicranella s.lat., Leptotrichella, and Microdus might belong to this lineage, or to the lineage containing Rhamphidium species (see below for a discussion of Dicranella vaginata).
The affinities within Dicranella s.str. support the description of a new species, as realized below, and the continued recognition of D. curvipes from D. heteromalla at the specific rank, as suggested by [35], with the molecular support presented in a more limited dataset by [10]. Our additional data support the recognition of Dicranella curvipes as a separate entity, while further documenting the molecular variability within the lineage. Our review of gametophytic morphology revealed that, in some of Dicranella curvipes specimens, the leaves tend to be homomallous vs. mostly falcate secund in D. heteromalla; the costa is narrower (less than 1/5 of the leaf base width), well-delimited from leaf lamina (a unique trait in Dicranella s.str.), and unistratose throughout (vs. wider costa weakly delimited from leaf lamina, which is partly to nearly entirely bistratose distally in D. heteromalla); the cells in the basal leaf portion are narrowly rectangular and moderately thick-walled vs. shortrectangular to subquadrate, thin-walled in D. heteromalla). In addition, most specimens of D. curvipes have leaves with rather distinct shoulders. It needs to be acknowledged, though, that some specimens of D. curvipes (such as Kučera 21379, 21778) from the Russian Far East have other combinations of these characters and could not be identified without mature sporophytes. Moreover, there appears to be an internal differentiation of the clade consisting of plants currently assigned to D. heteromalla s.str., with the specimens from the Russian Far East (RF47, 49, FDt35, Kučera 21639) and one from the eastern United States (Goffinet 8162) showing several distinct molecular synapomorphies at the level of one-base substitutions. Although this clade is only weakly supported on the tree, with respect to the unequal sequenced regions in the studied accessions and ambiguous reads at several points, the lineage is probably molecularly distinct. The most diverged lineage in molecular terms is, however, the one harboring two accessions of plants collected on Iturup Island and originally identified as Campylopus pyriformis. Despite the few morphological traits differentiating these plants from D. heteromalla, the plants are described below as a morphologically semicryptic species; nevertheless, they are distinct with respect to their rate of molecular differentiation.
Aongstroemiaceae. The common characters of Aongstroemiaceae and features which differentiate the Dicranella varia group as the most alien element in Aongstroemiaceae were discussed in detail by [10]. Both morphology and molecular support for the clade containing D. varia, D. varia var. obtusifolia, D. howei, and D. pacifica require the generic rank to be used for this clade. In theory, the name Anisothecium could be applied to it, as Anisothecium varium is one of the six species cited in the protologue. However, we believe that this would be the least appropriate option for the typification of the genus, as [26] proposed this name in his "Musci Austro-Americani . . . " for a group of predominantly South American species, which mostly share vaginate or semivaginate leaves with distinctly widened leaf bases, while Dicranella varia, which was included based on a single specimen from Cuba, forms a distinctly discordant element in his circumscription of the genus, as was also emphasized in his key to the species. Therefore, we believe that Anisothecium is much more appropriately typified with one of the predominantly South American species with expanded leaf bases, as done below, and we propose a new genus name for the clade of Dicranella varia and closely related species.
Previous analysis [10] suggested that Dicranella varia was paraphyletic with respect to a specimen (RF42) from northern Siberia. Extended sampling of Asian material that was supposed to represent D. varia resulted in both plants being molecularly identical or closely related to specimens from Europe and plants identical to the previously studied RF42. The provenance of the latter specimens mostly included northern Siberia, while the lineage containing European plants included specimens collected throughout boreal Asia. Morphological examination of the north Siberian plants and comparison to D. varia s.str., as represented by both European specimens (the Central European lectotype from Leipzig, Germany, was reviewed by [36] and a specimen from the southern Siberia and the southern part of Russian Far East), confirmed the morphological differences between the two groups. It appeared that plants similar to the analyzed north Siberian ones had already been described. Lindberg and Arnell [37], who proceeded extensive bryophyte collections from the Russian Arctic, described Anisothecium rubrum var. obtusiusculum based on the plants from the lower course of the Yenisey River. Their description matches our plants well. They also mentioned that a similar taxon, Dicranella varia var. obtusifolia Berggren occurs in Svalbard and indeed provided a description that seems to match morphologically both the variety later described by [37] and the plants that we collected in northern Siberia. The type material held in MO (MO-407808, accession 2226886) shows a good match with the north Asian plants analyzed by us. Hence, we raise the variety earliest described by Berggren to the species rank in the newly established genus, as effected below in the Taxonomy section. The previously unsampled NW American endemic Dicranella pacifica W.R. Schofield, which shares with D. varia multiple characters including recurved leaf margins, inclined asymmetric capsules, and absent annulus [20,38], was confirmed as another member of this lineage. While the specific status of D. howei, which morphologically sometimes seems indistinct from D. varia [20], now appears unequivocal with respect to the resolved identity of D. varia var. obtusifolia, the elaboration of morphological differences remains the task for a future dedicated study with more numerous molecularly barcoded specimens.
The core Aongstroemiaceae clade contains two basal lineages, which were not sampled by [10]. The basalmost lineage is represented by a single collection of a plant from Sakhalin Island with a unique combination of the otherwise typical Aongstroemiaceae characters, including shouldered leaves; elongate laminal cells; a single costal stereid band and lack of guide cells; reddish setae; and short, dark, curved, smooth, or slightly furrowed exannulate capsules. On the contrary, the upper leaf lamina is bistratose to tristratose, a character that only occurs in some representatives of the family. This plant is therefore described below as a new monospecific genus. The second previously unsampled lineage is the likewise monotypic Neodicranella, whose affinities had not been well resolved in its protologue [23]. We were not able to convincingly assess its affinities, even now, due to the conflict in the resolved topologies between the LTRN and K datasets. Although the "dicranelloid" rather than "dichodontioid" habit would favor the affinities as assessed through analysis of the trnK data, we refrain, however, from merging Neodicranella with Aongstroemia at this point.
The crown clade of core Aongstroemiaceae contains the species recognized to date, aside from the above-mentioned conflicting position of Neodicranella, in at least four genera: Aongstroemia (type species A. longipes), Hygrodicranum (type species H. falklandicum Cardot, not analyzed), Polymerodon (monotypic), and Dicranella species with expanded, mostly vaginate leaf bases, which earlier were often assigned to Anisothecium (the analyzed Dicranella campylophylla is among the six Anisothecium species eligible for the type of the genus). The high molecular support for this clade, as well as the suite of morphological characters common in the analyzed taxa of this clade, strongly support the idea of recognizing this clade as one genus, for which the oldest available name is Aongstroemia (see below in the Taxonomy section). Within the genus, we were able to additionally analyze four specimens of Dicranella schreberiana s.str. from previously unsampled parts of Europe (Czech Republic, Greece, European Russia), which confirmed the distinctness from D. schreberiana var. robusta. Moreover, we morphologically revised the type specimen of Cynodontium canadense Mitt., which proved to be identical to D. schreberiana var. robusta. Having priority at species rank, Cynodontium canadense is combined under Aongstroemia later in the Taxonomy section, leaving however open the question of the specimen RF40, with the morphology rather suggesting D. schreberiana s.str., which was however found to be isolated from the clades representing both recognized varieties and did not form a monophylum with either of them. Additional analyzed specimens of D. campylophylla (including the specimens labelled as D. cardotii (R. Br. bis) Dixon, considered synonymous by [39] and subsequent authors) and D. hookeri brought more certainty to the taxonomic evaluation of these taxa. They were found to be closely related, yet probably distinct species, which however may be at times difficult to separate morphologically, as the accession D. campylophylla TJH13, which was unavailable for our study, was resolved in the polytomy formed by the accessions of D. hookeri, together with an accession labelled as Polymerodon andinus, which was downloaded from GenBank. The pattern is further complicated by the accession RF65, identified also as D. hookeri, which however is molecularly clearly distinct from all other accessions identified as this species, as well as from D. campylophylla. The type of Hygrodicranum, H. falklandicum Cardot remains unsampled, and hence the generic status of Hygrodicranum remains to be assessed. Based on the two analyzed accessions of H. bolivianum and one of H. herrerae, it appears rather safe to infer that both species are very closely related to D. campylophylla, yet possibly specifically distinct, although the protologue and illustrations of Hygrodicranum herrerae in [40] do not provide information that would distinguish this species from the descriptions and illustrations of D. campylophylla available in the literature [39]. The sequenced specimens IPG20 (H. herrerae), as well as TJH04 and TJH13 (D. campylophylla and D. campylophylla/hookeri) from Chile, are nearly identical, with bistratose leaf lamina and dorsally mamillose cells characteristic for both taxa. The specimen of Hygrodicranum bolivianum (Buck 39497), as well as the additionally studied Chilean specimen (Larraín 43529), matches the species description [41], which resembles some closely related Dicranella (Aongstroemia) species. Consequently, we propose to combine both H. bolivianum and H. herrerae in Aongstroemia and expect that H. falklandicum might be resolved in this clade as well, but until the species is analyzed, we prefer not to put this taxonomic change into effect. The same applies to the assessment of Polymerodon andinus (rps4 and nad5 sequences obtained from specimen M. Lewis 87608, 7/87 (DUKE) and Eucamptodontopsis pilifera (Mitt.) Broth. (nad5 sequence obtained from specimen S.R. Hill 27912, 2/97 (DUKE)). If the sequences indeed correspond to these taxa, then they should be considered conspecific with D. hookeri, but in the absence of type studies and a more representative selection of analyzed material, such a proposal is premature, as further corroborated by the affinities of GenBank sequences of Eucamptodontopsis brittoniae (E.B. Bartram) B.H. Allen (AF435285, AF435328), which appear to be closely related to Holomitrium species based on BLAST searches.
The caution with the taxonomic evaluation of this group of taxa can be illustrated by the example of Dicranella vaginata. This species was considered to be closely related to the group of South Hemispheric Dicranella species recognized as Anisothecium by [26] in the protologue of the genus. However, three Chilean accessions analyzed by us were found to be resolved in the maximally supported clade containing two Rhamphidium species and Symblepharis krausei (Lorentz) Ochyra & Matteri. Indeed, all species have the vaginate leaf base, elongate basal leaf cells with porose longitudinal cell walls, subquadrate upper leaf cells, mostly unistratose upper leaf lamina, and rather short erect or inclined, nearly symmetric capsule with markedly spiculose, deeply divided peristome teeth, different from the typical dicranoid, i.e., not spiculose, and less divided peristome shared by Aongstroemia species. Recent morphological studies found Dicranella vaginata very similar to Aongstroemia gayana [42,43], which further indicates the necessity of a modern polyphasic reassessment of the lineage containing the type of Rhamphidium.
Consequently, we propose to typify the genus Anisothecium with A. campylophyllum, which in our opinion best preserves Mitten's original idea to include in it mostly South American representatives of the then recognized broad genus Aongstroemia with broadened, mostly clasping leaf bases and dicranoid affinities. Neither A. varium nor A. vaginatum (see above) seem to qualify better for this purpose. The identity of A. jamesonii Mitt. is currently ambiguous, as it was considered synonymous either to A. vaginatum [44] or to Dicranella hookeri [45], which were not found to be closely related by us, and hence a new investigation of the type is necessary in light of this finding. Similarly, we have no molecular data for the remaining South American species of the original Anisothecium, A. convolutum (Hampe) Mitt. and A. planinervium (Taylor) Mitt.
The polyphyly of Aongstroemia, Dicranella, and Ditrichum demonstrated by the analyses in [8,10] and this study is yet another example of homoplasy of morphological characters that were considered taxonomically relevant in earlier classifications. The superficially similar small pioneer mosses that are adapted to opportunistic life strategies sometimes occupy a remarkably isolated phylogenetic position among the basal lineages of Dicranidae. They share with most other protohaplolepideous lineages the broad, typically circumholarctic ranges, usually associated with humid climates, suggesting that the early diversification of Dicranidae was associated with cool to mild conditions, and therefore might be underestimated in temperate and subarctic, and by analogy possibly also in subantarctic, areas. The early radiation might not have been followed by excessive diversification according to our current knowledge, but the pioneer strategy of their representatives might have allowed them to survive until the present.
In contrast, the later diverging lineages of (mostly) opportunistic pioneer mosses (Aongstroemiaceae s.str., Dicranellaceae s.str., Ditrichaceae s.str. and some groups of Pottiaceae) are remarkably more diverse, in terms of species numbers, morphologically, and ecologically, often occupying xeric environments (such as several groups of Pottiaceae) and displaying multiple transitions to annual life strategies. Although only a limited number of Aongstroemia and especially Dicranella species from outside the Holarctic have been studied, the preliminary rps4-based phylogenetic analysis of Brazilian Dicranella s.lat. species indicates that the studied Neotropical Dicranella species all belong to the lineage of Dicranellaceae [46].
Our results demonstrate the unexpectedly underestimated diversity of northern temperate and subarctic pioneer mosses with dicranelloid habit and the resulting limitations of the currently used floras, especially in North Asia. In addition, northern Asia is an area of higher molecular diversity of Dicranella s.l. species, while European accessions are typically uniform in sequences, which might indicate the role of northeastern Asia as a source of diversity in these lineages worldwide.
The stunning extent of convergence in the available morphological traits within the studied genera underlines the need for obtaining molecular data for the representatives of the as yet unevaluated taxa and also the revision of types for existing names. Given the number of poorly known taxa (>600 names in Dicranella and >260 in Aongstroemia, [22]), such a project would require the efforts of the whole bryological community.  (Figures 2 and 3).

Diagnosis:
The single species segregated into the newly established genus differs from other dicranelloid mosses in possessing non-vaginate leaf bases and rather shortly subulate leaf acumina, recurved leaf margins, costae with single central stereid band, leaf lamina unistratose or bistratose along upper margins, yellow setae, yellow-purplish to brownish, asymmetric, furrowed capsules, revoluble annulus, and rhizoidal gemmae irregular in shape, composed of bulging cells.
Etymology: The name (composed of the Greek ῥίζα, root, and Latin gemma, gem) refers to the characteristic rhizoidal tubers (commonly also referred to as gemmae) of the only currently known species of the genus.
Description: Plants bright green, lacking red pigmentation. Stems about 5 mm, forming rather dense tufts, with a central strand. Leaves up to 1 mm long, lanceolate, erectspreading to spreading, not secund; margins plane or recurved only at base or nearly throughout, smooth or denticulate distally, partly bistratose distally; costae percurrent to short excurrent, in transverse section with differentiated dorsal and ventral epidermis and single stereid band; leaf lamina unistratose, cells rectangular to elongate-rectangular, bulging in transverse sections, smooth. Rhizoidal tubers constantly present, in young

Diagnosis:
The single species segregated into the newly established genus differs from other dicranelloid mosses in possessing non-vaginate leaf bases and rather shortly subulate leaf acumina, recurved leaf margins, costae with single central stereid band, leaf lamina unistratose or bistratose along upper margins, yellow setae, yellow-purplish to brownish, asymmetric, furrowed capsules, revoluble annulus, and rhizoidal gemmae irregular in shape, composed of bulging cells.

Diagnosis:
The single species segregated into the newly established genus differs from other dicranelloid mosses in possessing non-vaginate leaf bases and rather shortly subulate leaf acumina, recurved leaf margins, costae with single central stereid band, leaf lamina unistratose or bistratose along upper margins, yellow setae, yellow-purplish to brownish, asymmetric, furrowed capsules, revoluble annulus, and rhizoidal gemmae irregular in shape, composed of bulging cells.
Etymology: The name (composed of the Greek ῥίζα, root, and Latin gemma, gem) refers to the characteristic rhizoidal tubers (commonly also referred to as gemmae) of the only currently known species of the genus.
Description: Plants bright green, lacking red pigmentation. Stems about 5 mm, forming rather dense tufts, with a central strand. Leaves up to 1 mm long, lanceolate, erectspreading to spreading, not secund; margins plane or recurved only at base or nearly throughout, smooth or denticulate distally, partly bistratose distally; costae percurrent to short excurrent, in transverse section with differentiated dorsal and ventral epidermis and single stereid band; leaf lamina unistratose, cells rectangular to elongate-rectangular, bulging in transverse sections, smooth. Rhizoidal tubers constantly present, in young

Diagnosis:
The single species segregated into the newly established genus differs from other dicranelloid mosses in possessing non-vaginate leaf bases and rather shortly subulate leaf acumina, recurved leaf margins, costae with single central stereid band, leaf lamina unistratose or bistratose along upper margins, yellow setae, yellow-purplish to brownish, asymmetric, furrowed capsules, revoluble annulus, and rhizoidal gemmae irregular in shape, composed of bulging cells.
Etymology: The name (composed of the Greek ῥίζα, root, and Latin gemma, gem) refers to the characteristic rhizoidal tubers (commonly also referred to as gemmae) of the only currently known species of the genus.
Description: Plants bright green, lacking red pigmentation. Stems about 5 mm, forming rather dense tufts, with a central strand. Leaves up to 1 mm long, lanceolate, erectspreading to spreading, not secund; margins plane or recurved only at base or nearly throughout, smooth or denticulate distally, partly bistratose distally; costae percurrent to short excurrent, in transverse section with differentiated dorsal and ventral epidermis and single stereid band; leaf lamina unistratose, cells rectangular to elongate-rectangular, bulging in transverse sections, smooth. Rhizoidal tubers constantly present, in young ζα, root, and Latin gemma, gem) refers to the characteristic rhizoidal tubers (commonly also referred to as gemmae) of the only currently known species of the genus.
Description: Plants bright green, lacking red pigmentation. Stems about 5 mm, forming rather dense tufts, with a central strand. Leaves up to 1 mm long, lanceolate, erect-spreading to spreading, not secund; margins plane or recurved only at base or nearly throughout, smooth or denticulate distally, partly bistratose distally; costae percurrent to short excurrent, in transverse section with differentiated dorsal and ventral epidermis and single stereid band; leaf lamina unistratose, cells rectangular to elongate-rectangular, bulging in transverse sections, smooth. Rhizoidal tubers constantly present, in young stage red, turning dark brown, irregularly shaped with protruding cells, 3-4 cells long and 1-3 cells wide. Perichaetial leaves differentiated, larger than lower leaves, from broadly sheathing base rather abruptly narrowed into squarrose or flexuose apex. Setae yellowish, straight; capsules incurved, longitudinally furrowed, without or weak strumae, yellow to purplish along ribs and around the mouth. Exothecial cells irregular in shape to rectangular, with evenly incrassate walls. Annulus well differentiated, composed of one row of large hyaline thick-walled cells, revoluble. Operculum long rostrate. Peristome bright red, teeth unequally split to half of their length, longitudinally striolate proximally, papillose distally. Calyptrae cucullate.
The genus is currently considered monotypic. Diagnosis: species combined in the newly established genus differ from other dicranelloid mosses in possessing reddish-brown mature stems; leaves from narrow-triangular base gradually narrowed to a subulate acumina; costae with single stereid band; leaf lamina unistratose or with bistratose margins, composed of elongate to linear cells; red setae; peristome with mostly high basal membrane and weakly developed annulus.
Description: Plants very small, in loose reddish-brownish tufts. Stems with central strand. Well-developed parts of stems reddish-brown and rhizoids with vinaceous pigmentation. Leaves up to 2 mm long, fuscous, weakly secund; margins plane throughout, denticulate at apex; costa rather weak, percurrent, sharply delimited from leaf lamina, with compact stereid band, ventral and dorsal epidermis, or with weakly developed dorsal band, composed of substereids and guide cells covered by ventral epidermis. Tubers consisting of one row of (1-)2-3(-6) much enlarged cells, pale reddish to wine-red. Perichaetial leaves similar to upper leaves. Setae reddish. Capsules erect to inclined, symmetric or curved, smooth or slightly furrowed. Exothecial cells short rectangular, in longitudinal rows. Annulus weakly differentiated, not revoluble. Peristome dicranoid, with high basal membrane.
Accepted rom Dicranella, the genus to which the spesly, and the suffix -opsis (from Greek ὄψις, the morphological similarity between the ly established genus differ from other diof widened to vaginate leaf bases and sutwo stereid bands and guide cells, bistrales and revoluble annulus. rk green, lacking red pigmentation. Stems abruptly tapering into a long, channeled, g, erect to squarrose-flexuose, patent or sere or very slightly denticulate at leaf tip, xcurrent, sharply delimited from leaf lamcells and dorsal and ventral or only dorsal ian leaf cells linear. Rhizoidal tubers, when out protruding cells, curved. Perichaetial rect to slightly inclined, symmetric or disinally ribbed, with more or less differentictangular, rather thin-walled cells between idened cells, revoluble. Peristome dicra- ψις, meaning aspect, appearance, sight), referring to the morphological similarity between the genera.
Diagnosis: Species combined in the newly established genus differ from other dicranelloid mosses in possessing a combination of widened to vaginate leaf bases and subulate leaf tips, plane leaf margins, costae with two stereid bands and guide cells, bistratose upper leaf lamina, red setae, ribbed capsules and revoluble annulus.
Description: Plants yellowish green to dark green, lacking red pigmentation. Stems with central strand. Leaves with oblong bases abruptly tapering into a long, channeled, subulate acumina, upper stem leaves sheathing, erect to squarrose-flexuose, patent or secund, crispate or not when dry; margins entire or very slightly denticulate at leaf tip, plane, unistratose; costae percurrent to short excurrent, sharply delimited from leaf lamina, with dorsal and ventral epidermis, guide cells and dorsal and ventral or only dorsal stereid band; distal leaf lamina bistratose, median leaf cells linear. Rhizoidal tubers, when present, dark brown, irregularly shaped without protruding cells, curved. Perichaetial leaves resemble upper stem leaves. Capsules erect to slightly inclined, symmetric or distinctly curved, not strumose, strongly longitudinally ribbed, with more or less differentiated exothecial bands and quadrate to short rectangular, rather thin-walled cells between them. Tentatively included genus (pending molecular confirmation): Bryotestua Thér. & P. de la Varde. As for Trichodontium (Dixon) Fife, see the Discussion.
The following synopsis only includes the genera where taxonomic novelties are proposed.
Dicranella Species with uncertain placement: all other accepted species (cf. [22]), pending morphomolecular studies, including Dicranella polii Renauld & Cardot and D. vaginata (Hook.) Cardot, for which our molecular phylogenetic data suggest placement outside Dicranella as recognized here, but additional sampling is needed to assign the generic affinities, as discussed above.

Diagnosis:
The new species resembles D. heteromalla in the rather robust plant size, non-shouldered leaves, wide costae occupying up to 1 2 of the leaf width and weakly delimited from leaf lamina, with thin-walled cells with large lumen forming ventral surface of costa in basal leaf portion, but differs from it in having homomalous rather than falcate secund leaves and weakly serrulate to nearly entire upper leaf margins.

Diagnosis:
The new species resembles D. heteromalla in the rather robust plant size, non-shouldered leaves, wide costae occupying up to ½ of the leaf width and weakly delimited from leaf lamina, with thin-walled cells with large lumen forming ventral surface of costa in basal leaf portion, but differs from it in having homomalous rather than falcate secund leaves and weakly serrulate to nearly entire upper leaf margins.
Type: Russia, Sakhalin Province, Iturup Island. South-West slope of Baranskogo Volcano, Goryachaya River. Etymology: The species name refers to the typical habitat of the species at the type locality.
Description: Plants medium-sized, stems up to 3 cm, single, with well-developed central strand, evenly foliate, tomentose in lower part. Leaves more or less appressed when dry, spreading when wet, gently falcate-secund, 2.5 -3.2 × 0.25 -0.35 mm, widest at base, from lanceolate base gradually tapering into canaliculate subulate acumen; margins plane, unistratose, weakly and bluntly toothed throughout or only in upper half, near apex with double teeth: costae strong, occupying 1/3-1/2 of leaf base, rather indistinctly Description: Plants medium-sized, stems up to 3 cm, single, with well-developed central strand, evenly foliate, tomentose in lower part. Leaves more or less appressed when dry, spreading when wet, gently falcate-secund, 2.5-3.2 × 0.25-0.35 mm, widest at base, from lanceolate base gradually tapering into canaliculate subulate acumen; margins plane, unistratose, weakly and bluntly toothed throughout or only in upper half, near apex with double teeth: costae strong, occupying 1/3-1/2 of leaf base, rather indistinctly delimited from the leaf lamina, with one row of guide cells, two stereid bands, and differentiated dorsal and ventral epidermis; sometimes ventral epidermis immediately covering guide cells or guide cells forming surface of costa ventrally; leaf lamina partly or completely bistratose distally, upper leaf cells 24-38 × 5-6 µm, elongate-rectangular, smooth, moderately thick-walled; basal leaf cells of the same length and 8-11 µm wide. Sexual condition and sporophytes unknown.
Differentiation: We did not find more characteristics to differentiate this molecularly distinct species from Dicranella heteromalla than those specified in the diagnosis. D. thermalis resembles Campylopus pyriformis (Schultz) Brid. in having wide costae, undifferentiated alar regions, and thin-walled cells with wide lumen on a ventral surface in the basal portion of leaf. In contrast to most Campylopus species, D. thermalis possesses two stereid bands.  [48], in which the type species of the genus was described as Dicranum filiforme P. Beauv.
Diagnosis: This genus differs from A. longipes and several other species of Aongstroemia in its traditional circumscription in its robust habit, leaves with sheathing leaf base, abruptly narrowed into a long, subulate leaf apex, and elongate to linear, extremely thick-walled basal leaf cells. From the genus Aongstroemia in its newly proposed circumscription, Bryopalisotia differs in having cylindric rather than ovoid or shortly ellipsoid capsules. Elongate to linear, extremely thick-walled basal leaf cells differ Bryopalisotia from A. guayana.
The genus is presently considered monospecific, although the below-stated synonymy should be revisited (see Discussion Plants minute to medium-sized, in loose to dense turfs. Stems julaceous or not, central strand present. Stem leaves with a broad sheathing base tapering into a blunt apex or abruptly narrowed to a short or long acumen. Margins entire, crenulate or weakly denticulate to dentate. Leaf lamina 1-2(-3) stratose; laminal cells variable in shape, usually smooth but mamillose or papillose in some species. Alar cells not differentiated. Costa subpercurrent to mostly short to long excurrent, weak to strong. Asexual reproduction via gemmae (on filamentous branches at the leaf axils) or rhizoidal tubers. Dioicous. Seta elongate, straight or flexuose. Capsule variable in shape, ovoid to curved and sometimes slightly strumose, smooth or furrowed when dry, operculate, with peristome teeth vertically pitted-striolate at base. Annulus not or poorly differentiated. Operculum conic or rostrate. Calyptra cucullate.
The following synopsis only includes genera where taxonomic novelties are proposed. Etymology: The generic name originates from the generic name Dicranella, where this species has been placed for a long time, and the prefix calcireferring to the ecological preference for calcareous substrates in the species included in the genus.
Diagnosis: Species combined in the newly established genus differ from other dicranelloid mosses in possessing non-vaginate leaf bases, partly to nearly entirely recurved leaf margins, smooth laminal cells, costae with well-differentiated stereids in one or two bands, red setae, dark reddish-brown, asymmetric, inclined capsules, and non-revoluble annulus, and by its ecological preference for base-rich mineral soil.
Description: Central strand present. Leaves lanceolate, gradually narrowed to blunt, acute or acuminate apex, without sheathing base, margins recurved on one or both sides; costa weakly or rather sharply delimited from leaf lamina, typically with guide cells and two or rarely only dorsal stereid band, differentiated dorsal and, in several species, also ventral epidermis; leaf lamina unistratose or with bistratose patches to entirely bistratose distally; leaf cells rectangular. Rhizoid tubers occasionally present, irregular in shape, with protruding cells, 100-140(-250) × 60-95 µm. Dioicous. Perichaetial leaves similar to lower leaves. Setae red. Capsules inclined, asymmetric, ovoid, gibbous, smooth or furrowed when dry, dark red when mature. Exothecial cells irregular in shape or rectangular, with thickened longitudinal walls. Annulus weakly differentiated, not revoluble. Peristome dicranoid.
Calcidicranella Description: Plants small, gregarious, light green or yellowish. Stems simple, ca. 0.1-0.2 cm, with strong round central strand and weak sclerodermis, evenly foliated. Leaves appressed, straight or slightly curved when dry, spreading when moist, 1.0-1.8(-2.2) mm, with wide, ovate bases and more or less distinct shoulders, above shoulders gradually narrowed towards blunt acumen, concave, lower leaves not widened, triangular; margins plane at base, narrowly recurved at shoulders and just above and below them or almost to the leaf tip, unistratose proximally and partly bistratose distally, uneven above, rarely throughout the margin; costae ending just below apices, rarely percurrent, rather strong, occupying ca. 1/7-1/5 of leaf base, distinctly delimited from leaf lamina, in transverse section with 2-4(-5) large ventral guide cells, differentiated dorsal epidermis and single weak stereid band; leaf lamina unistratose with occasional bistratose strands distally; upper leaf cells short rectangular to subquadrate, 12-20 × 7-12 µm, smooth, not bulging, proximally longer and wider, 44-90 × 10-17 µm, elongate-rectangular, 2-3 rows of cells along margins narrower, ca. 4-6 µm wide. Dioicous. Perichaetial leaves of the same length, but with wider and longer base, more abruptly narrowed to lanceolate or short subulate acumen. Setae 3-5 mm red to brownish. Capsules ca. 1 mm, asymmetric, curved, ovate, with short neck, strumose, brownish-red, distinctly furrowed, red rimmed distally, exothecial cells irregular in shape, thick-walled with equally thickened walls, longer and narrower along furrows, with few stomata proximally. Annulus not differentiated. Operculum conic. Peristome teeth red to brownish, 450-500 µm long, unequally split for nearly half of their length, longitudinally striolate proximally, papillose distally. Spores 14-17 µm, smooth, yellowish-brown, mature in summer. Rhizoidal tubers not seen.  Differentiation: C. obtusifolia resembles C. varia or C. howei in habit but differs in smaller plants with stems up to 5 mm, while stems of C. varia often extend to 1 cm. Leaf margins in C. obtusifolia are plane below shoulders, while C. varia has leaf margins recurved from the basal leaf portion and C. howei has leaf margins recurved mostly in the lower leaf part only, often only on one side. Leaf tips in C. obtusifolia are typically blunt, with costae ending a few cells below tips to being percurrent, while in C. varia/howei leaf tips are sharp and costae excurrent. Capsules of C. obtusifolia are strumose and distinctly longitudinally furrowed, while in C. varia/howei capsules are not strumose, smooth or rarely indistinctly furrowed. Exothecial cells in C. obtusifolia approach C. howei, they are irregular in shape, with equally thickened walls, while in C. varia longitudinal walls of exothecial cells typically are thicker than transverse ones. Although in many formal characters C. obtusifolia resembles North American C. pacifica, the latter species is much larger; moreover, with its contorted to crisped leaves and smooth capsules it is quite distinct from C. obtusifolia.
Distribution and ecology: A predominantly Arctic species, described from Svalbard and also known from a single locality in Nenets Autonomous District (European Russia), suite of localities along Yenisey River, in Taimyr Peninsula, Anabar Plateau and from a single locality in Yakutia. According to the protologue of Anisothecium rubrum var. obtusiusculum [37], it is also one of the most frequent mosses along the Yenisey River banks, although it rarely occurs in sufficient amounts, while in Svalbard it is either rare or not recognized from C. varia. It grows on bare loamy soil and silty sediments including saline ones on eroded slopes along rivers and in massives of baidzarakhs (thermokarst mounds), most often with Hennediella heimii var. arctica, Funaria spp., Tortula leucostoma, T. cf. cernua, Bryoerythrophyllum spp., Aloina brevirostris, Stegonia latifolia, Pohlia atropurpurea, Bryum spp., and many other pioneer mosses. At the same time, according to our field experience, it differs from other Dicranella s.l. species widespread in Siberian Arctic in occupied habitats, since these usually settle on acidic sandy sediments, typically with gemmiferous species of Pohlia, Pogonatum and Psilopilum species.  Etymology: The generic name originates from Aongstroemia (a genus of dicranoid mosses) and the prefix proto-(from Greek πρῶτος, first), which reflects the basalmost position of the genus within the core Aongstroemiaceae clade. The specific epithet reflects the provenance of the original collection, the Sakhalin Island.
Diagnosis: Differs from other Holarctic Dicranella s.l. species by the combination of distinctly shouldered leaves, distally regularly 2-3-stratose lamina, costa with a single stereid band and undifferentiated guide cells, elongate rectangular laminal cells and irregularly furrowed, curved capsules.
Description: Plants small, gregarious, light green or yellowish, mixed with other pioneer mosses. Stems simple, ca. 0.1-0.2 cm, with central strand and weak sclerodermis, evenly foliated. Leaves appressed, straight or slightly curved when dry, spreading when moist, gradually increasing in size distally, 1.5 -1.9 × 0.4 -0.53 mm, with wide, ovate base, τoς, first), which reflects the basalmost position of the genus within the core Aongstroemiaceae clade. The specific epithet reflects the provenance of the original collection, the Sakhalin Island.
Diagnosis: Differs from other Holarctic Dicranella s.l. species by the combination of distinctly shouldered leaves, distally regularly 2-3-stratose lamina, costa with a single stereid band and undifferentiated guide cells, elongate rectangular laminal cells and irregularly furrowed, curved capsules.
Description: Plants small, gregarious, light green or yellowish, mixed with other pioneer mosses. Stems simple, ca. 0.1-0.2 cm, with central strand and weak sclerodermis, evenly foliated. Leaves appressed, straight or slightly curved when dry, spreading when moist, gradually increasing in size distally, 1.5-1.9 × 0.4-0.53 mm, with wide, ovate base, widest at ca. 1/10-1/5 of leaf length with distinct shoulders, abruptly narrowed into gradually tapering blunt acumen, concave; margins plane, with few blunt distant teeth at shoulders and upper part of acumen to nearly entire, plane, partly bistratose proximally; costa weak, weakly delimited from leaf lamina, percurrent, in transverse section with ventral and dorsal epidermis and single band of substereids between them, without guide cells proximally, weakly differentiated distally; leaf lamina unistratose with bistratose strands proximally, 2-3 stratose distally; leaf cells elongate-rectangular, 37-62 × 6-13 µm, smooth, bulging on both sides, proximally somewhat longer, 45-75 µm long. Dioicous, male plants not seen. Perichaetial leaves with wider base, abruptly narrowed to short subulate acumen. Setae reddish, 5-7 mm, spirally twisted when dry and moving around after wetting. Capsules 1.2-1.5 mm long, asymmetric, curved, ovate, with short neck, weakly furrowed, not strumose, reddish-brown, red-rimmed distally; exothecial cells rectangular, moderately thick-walled with evenly incrassate transverse and longitudinal walls, longer and narrower along furrows, with few stomata in proximal part. Annulus not differentiated. Operculum conic or with short blunt oblique beak. Peristome teeth bright red, ca. 300 µm, unequally split for nearly half of their length, longitudinally striolate proximally, papillose distally. Spores 13-17 µm smooth, yellowish-brown, mature in autumn. Rhizoidal tubers not seen. Etymology: The generic name originates from Aongstroemia (a genus of dicranoid mosses) and the prefix proto-(from Greek πρῶτος, first), which reflects the basalmost position of the genus within the core Aongstroemiaceae clade. The specific epithet reflects the provenance of the original collection, the Sakhalin Island.
Diagnosis: Differs from other Holarctic Dicranella s.l. species by the combination of distinctly shouldered leaves, distally regularly 2-3-stratose lamina, costa with a single stereid band and undifferentiated guide cells, elongate rectangular laminal cells and irregularly furrowed, curved capsules.
Description: Plants small, gregarious, light green or yellowish, mixed with other pioneer mosses. Stems simple, ca. 0.1-0.2 cm, with central strand and weak sclerodermis, evenly foliated. Leaves appressed, straight or slightly curved when dry, spreading when moist, gradually increasing in size distally, 1.5 -1.9 × 0.4 -0.53 mm, with wide, ovate base, Differentiation: With its shouldered and then gradually narrowed leaves, elongate rectangular laminal cells and short curved capsules, P. sachalinensis habitually resembles a small Diobelonella, especially Asian populations with narrower leaves. However, it differs not only in its size but also in having bistratose leaf lamina. The same trait and plain margins differentiate P. sachalinensis from the somewhat similar Calcidicranella varia. Among species with partially bistratose lamina, P. sachalinensis differs from Dicranellopsis subulata in its non-subulate distal leaf portion, lack of guide cells and undifferentiated annulus; it differs from Calcidicranella pacifica in having shouldered leaves, narrower and longer leaf cells and lack of guide cells; and from C. howei in shouldered leaves and narrower costa.
Distribution and ecology: This newly described species is known from a single specimen, which was collected on silty alluvium sediments of Tym' River in the middle , the name Anisothecium should no longer be considered illegitimate. As argued above, the best candidate to typify the name is Anisothecium campylophyllum with respect to the good match with the general intent of the author and known phylogenetic affinities of this species.
Diagnostic characters: Stem leaves with a broad sheathing base tapering into a blunt apex (in less developed A. longipes plants) or abruptly narrowed to short or long pointed, spreading to squarrose leaf apex. Lamina cells rectangular, smooth or sometimes mamillose or papillose, sometimes (irregularly) bistratose. Tubers, if present, spherical without protruding cells. Capsules erect to inclined, symmetric to asymmetric, oval/obloid to curved and sometimes slightly strumose, on a straight, erect, red to brownish seta. Annulus not or poorly differentiated.
For a list of accepted species see below. Species with uncertain placement: all other accepted species (cf. [22]), pending morphomolecular studies, and also Aongstroemia orientalis Mitt., for which molecular phylogenetic data [10] suggest placement in Ditrichaceae, but additional sampling is required to assess its affinities within this family.

Taxon Sampling
The matrix of molecular data was largely based on that used for the backbone phylogeny of Dicranidae, with a focus on Dicranella and Aongstroemia [10]. With respect to the absence of dicranelloid taxa in some lineages of haplolepidous mosses, we reduced the matrix by leaving out or reducing the number of accessions in lineages where these taxa were absent, in order to decrease the complexity of the alignment. The outgroups were thus reduced to include only Pseudoditrichum, Flexitrichum, Scouleria, Drummondia, Hymenoloma, and Bryoxiphium, and we further substantially reduced the representation of Leucobryaceae, Rhabdoweisiaceae, Dicranaceae, and related families (leaving out completely Mittenia, Pleurophascum, Serpotortella, Hypodontiaceae, Octoblepharaceae, and Calymperaceae), and also Ditrichaceae including Aongstroemia orientalis and A. julacea, which will be treated in a dedicated future article. On the other hand, we added accessions of Dicranella staphylina, D. humilis, D. varia incl. its neglected var. obtusifolia, D. pacifica, D. grevilleana, D. schreberiana incl. its var. robusta, D. campylophylla, D. hookeri, D. heteromalla, D. curvipes, D. polii, D. vaginata, Aongstroemia longipes, A. filiformis, Neodicranella hamulosa, and unassigned dicranelloid plants from Pacific Russia, which were found to be related to Dicranella s.str. and to Diobelonella/Dichodontium/Neodicranella grade. Newly generated trnK-psbA data were added for at least one representative of each major lineage left. Laboratory protocols for isolation of DNA, amplification and sequencing followed the protocols described in [8,10,56,57]. Genbank accession numbers of the included specimens and vouchers of specimens studied de novo are compiled in Appendix A.

Phylogenetic Analyses
Sequences were aligned using MAFFT v. 7 [58] with the E-INS-i strategy and otherwise default settings, and the resulting alignment was improved manually at obviously misaligned sites. The concatenated dataset (available in http://purl.org/phylo/treebase/ phylows/study/TB2:S30163, accessed on 28 January 2023) was tentatively partitioned according to the sequenced regions (trnF-trnL, abbreviated L hereafter, trnL-rps4 (T), rps4-trnS (R), trnK-psbA (K), nad5 (N) with respect to their significantly differing coverage, rather than according to coding and non-coding regions. The best-fit partitioning scheme and models of nucleotide evolution were searched for in PartitionFinder2 [59]. The results of the greedy algorithm used suggested partitioning according to all of the initially suggested partitions, with the HKY + I+G model for the trnF-trnL partition and GTR + I+G for the remaining ones. Indel data were scored for individual partitions using the simple indel coding (SIC) approach [60] in SeqState 1.4.1 [61] and added to the dataset in three variants: (1) indels scored only for L, R, and N partitions; (2) indels scored for L, T, R, and N partitions; and (3) indels scored for all partitions. Based on the results of [10], we did not separately analyze the L, R, and N data with respect to the reported absence of conflicts in topology and relatively low resolution of trees obtained from single-gene analyses, but we explored the influence of previously unused regions, i.e., (a) data from spacers flanking trnT between trnL and rps4 (T) and (b) trnK-psbA (K) data, which were successively added to the working pilot analyses. Given the amount of phylogenetic signal, the K data were also analyzed separately from the 52 accessions for which these data were available.
Phylogenetic reconstructions were performed using Bayesian inference (BI) and maximum likelihood (ML). BI was run in MrBayes v.3.2.7 [62] in two parallel runs, each consisting of eight Markov chains run for 2,000,000 generations as default, and with further generations added if the convergence between runs did not reach 0.01, with the default number of swaps and a sampling frequency of one tree for each 100 generations. The chain temperature was initially set at 0.1 and lowered as necessary according to the acceptance rates. The models were sampled throughout the GTR model space and gamma-distributed rate variation across sites, and a proportion of invariable sites, as suggested by the Partition-Finder. PSRF values, were checked as being close to 1.000. ESS values were checked using Tracer v.1.7.2 [63] as being higher than 200. Consensus trees were calculated after omitting the burn-in of the first 25% of trees. The best-scoring maximum likelihood (ML) trees were searched using the new rapid hill-climbing algorithm in RAxML 8.2.12 [64] under the GTR model with gamma model of rate heterogeneity in 50 independent runs, each starting from a different random tree. The extended majority-rule consensus tree criterion was used to stop the bootstrapping used for the assessment of the node robustness. Analyses were performed using the grid computational services provided by the MetaCentrum Virtual Organization (see Acknowledgement). Trees were visualized using TreeGraph2 [65].

Morphological Studies
In addition to standard microscopic observations during the revision of herbarium specimens, preparation of taxon descriptions, and illustrations, images of peristomes were obtained by scanning electron microscopy (SEM) with a JSM-6380 (JEOL) at the User Facilities Center of M.V. Lomonosov Moscow State University. Peristomes mounted on stubs were coated with gold without any additional preparation, and light microscope illustrations were made under a stereomicroscope Olympus SZX-7 with a digital camera Infinity 8, with Z-stacking in Helicon Software [66].
Supplementary Materials: The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/plants12061360/s1. Figure S1: Maximum Likelihood tree, inferred from the concatenated data matrix from the chloroplast trnF-trnS and mitochondrial nad5 intron 2 sequence alignment (LTRN dataset) of selected species of Dicranidae. Bootstrap support values inferred from ML analyses without and with indel coding are shown above branches, posterior probabilities inferred from BI without and with indel coding are shown below branches. The same tree with saved branch lengths is shown in the lower left corner. For details, see File S1; Figure S2: Maximum Likelihood tree, inferred from and trnK-psbA of selected species of Dicranidae. Bootstrap support values inferred from ML analyses without and with indel coding are shown above branches, posterior probabilities inferred from BI without and with indel coding are shown below branches. The same tree with saved branch lengths is shown in the lower left corner. For details, see File S1. Acknowledgments: First of all, we are grateful to Marina Bonfim Santos and Michael Stech (Naturalis Biodiversity Center and Leiden University, Leiden, The Netherlands) as well as Henk Siebel (Natuurmonumenten, Hilversum, The Netherlands) for permission to use the results of their study. We acknowledge the molecular laboratory work performed by Alžběta Manukjanová (University of South Bohemia) and curators of herbaria, which provided specimens for DNA extraction and the morphological study (B, CONC, DUKE, LE, MHA, NSK, MO, NY). The work on SEM was performed under the financial support of the Ministry of Education and Science of the Russian Federation. We also are grateful to Juan Larraín for useful comments on the identity of several south-Hemispheric taxa and to John Brinda for checking the syntype specimen of Dicranella varia var. obtusifolia in MO.

Conflicts of Interest:
The authors declare no conflict of interest.

Appendix A
Studied specimens and GenBank accession numbers. Specimens are annotated with the isolate codes shown in the trees (indicated as "N/A" for sequences for which "isolate" annotation is omitted in GenBank), geographic origin and specimen ID (where available). Isolate codes of the originally studied specimens are in bold. Each isolate starts with a new line, so different isolates of the same species combined in the concatenated datasets are indicated with an "&" mark after the isolate code, which means that the present sequence(s) was combined with the one on the next line.