Hidden Diversity within Tetralophozia filiformis (Marchantiophyta, Anastrophyllaceae) in East Asia

Tetralophozia filiformis s.l. is known from a number of localities mostly in amphi-oceanic areas in Northern Hemisphere, including Atlantic Europe, amphi-Pacific Asia, South Siberia, and western North America. The newly obtained collections of this ‘species’ show strong variation in morphology of the taxon across amphi-Pacific Asia although connected by some ‘intergrading’ modifications. This implies the genetic diversity within this unit earlier recognized as a single taxon. Authors used molecular-genetic, morphological, and chorological methods to understand if the geographically correlated morphological variation also correlates with genetic differences and if it is possible to distinguish some additional taxa within the series of specimens originating from the various areas in amphi-Pacific Asia. It was found that Tetralophozia filiformis is a complex of at least three morphologically similar species, including one long forgotten name (Chandonanthus pusillus) that should be reinstated as separate species and one taxon (Tetralophozia sibirica) that should be described as new. Tetralophozia filiformis and Chandonanthus pusillus are lectotypified, and the new combination is provided for the latter. The three accepted taxa distinctly differ one from another in distribution patterns, preferable climate characteristics, and genetic distances, besides minor differentiations in morphology. The main morphological distinguishing features are the leaf cell size, height of undivided part in leaf lamina, and leaf dentation characteristics. Taking into account the robust correlation between the climate-based and molecular-genetic-based clusters, one more (fourth) taxon could be probably segregated from Tetralophozia filiformis.


Introduction
The best known species of the genus Tetralophozia (R.M.Schust.) Schljakov is the broadly distributed Arctic-Montane circumpolar T. setiformis (Ehrh.) Schljakov. Other taxa of the genus are locally distributed or rare. One of the lesser known is the predominantly East Asian Tetralophozia filiformis (Steph.) Urmi, originally described from Yunnan Province of China as Chandonanthus filiformis Steph. [1]. The distribution of Tetralophozia filiformis was reviewed in detail by Urmi [2], who reported this species in Europe (Spain), East Asia (Japan, Taiwan, China, North-East India, Bhutan, Malaysia), Canada, and transferred the taxon from Chandonanthus Mitt. to Tetralophozia. Urmi's paper [2] inspired further interest to this taxon, including in Russia, where at the beginning of the new millennium, T. filiformis was identified for the first time in the relatively harsh oroboreal conditions of southern Siberia. It was a new report for Russian liverwort flora [3]. Following the latter, several collections of T. filiformis were made in south Russian Asia (Konstantinova

Molecular Genetic
Seven accessions of ITS1-2, nine of trnL-F and a single for trnG-intron were produced and deposited into GenBank. The specimens from Amur Province and Khabarovsk Territory were excluded from phylogenetic estimation due to the presence of trnL-F data only but were used in p-distance calculation. The combined alignment ITS1-2 + trnL-F for 21 specimens consists of 1375 sites, among which 867 belong to ITS1-2 and 508 belong to trnL-F. The number of conserved positions in ITS1-2/trnL-F was 611 (70.47%)/392 (77.17%), the number of variable positions was 241 (27.80%)/106 (20.87%), and the number of parsimony-informative positions was 105 (12.11%)/37 (7.28%).
The single most parsimonious tree with a length of 773 steps was obtained in MP analysis, consistency index-0.732394, and retention index-0.509677. The ML analysis yielded a single most likely tree (-ln L = 5058.463250; Figure 1). The obtained topologies are congruent among each other in the sense of relationship in the genus Tetralophozia but provide unsupported affinity among majority of genera in Anastrophyllaceae, as shown previously [8,9]. All specimens of Tetralophozia were placed in one unsupported clade. The subclade with three specimens of T. setiformis (bootstrap support 99% in MP, 100% in ML, or 99/100) is in a sister relationship to the subclade with the T. filiformis complex. Two subsequently diverged specimens of T. filiformis from China are sister-related (72/90) to specimens of T. pusilla from Japan and Korea (93/98). Both Siberian accessions compose a subclade with 99/99 support in relation (81/90) to T. filiformis + T. pusilla; these specimens we ascribe to a new species described here-T. sibirica.

Climate Variables and the Genetic Diversification
The bioclimate variables, as mentioned in the Section 5, were obtained for 22 geographic localities. These 22 localities corresponded to 22 specimens (both studied and unstudied by the authors). In accordance with the data obtained by molecular-genetic methods (arguing the recognition of three taxa in Asia within the T. filiformis complex), the specimen localities fell into four categories: (1) T. filiformis s. str., marked as squares, (2) T. pusilla, marked as triangles, (3) newly described T. sibirica, marked as circles and (4) specimens that we did not study but whose coordinates are precisely known, marked as snowflakes. The obtained bioclimate variables are shown in Table 2. Then, DCA was performed for a three-dimensional grid diagram. The position of each locality within a three-dimensional grid is described by formal values placed in Table 3. Then, the correlation between values of each bioclimate and values obtained in the axis of the DCA (Table 3) is placed in Table 4. Graphically, the distribution of collecting localities is presented in Figure 2. The taxonomic units revealed in the present account are encircled. Two observations are noticeable:

1.
Tetralophozia pusilla and T. filiformis are closely related in the molecular-genetic respect and grow in similar climatic environments too. Tetralophozia sibirica grows in harsh northern environments and is genetically well different from other two taxa that also correspond distance in climate diagram.

2.
Strongly geographically distanced Alaskan and Spanish populations were found to be similar in measured climate variables.
Locality 21 presumably should not be far climatically from Alaskan localities, but shows a strong difference from the climates of all other involved localities. Whether this is local aberration in the climate or regularity correlating with morphology and/or genetics was not tested here.

Climate Variables and the Genetic Diversification
The bioclimate variables, as mentioned in the Section 5, were obtained for 22 geographic localities. These 22 localities corresponded to 22 specimens (both studied and unstudied by the authors). In accordance with the data obtained by molecular-genetic methods (arguing the recognition of three taxa in Asia within the T. filiformis complex), the specimen localities fell into four categories: (1) T. filiformis s. str., marked as squares, (2) T. pusilla, marked as triangles, (3) newly described T. sibirica, marked as circles and (4) specimens that we did not study but whose coordinates are precisely known, marked as snowflakes. The obtained bioclimate variables are shown in Table 2. Then, DCA was performed for a three-dimensional grid diagram. The position of each locality within a three-dimensional grid is described by formal values placed in Table 3. Then, the correlation between values of each bioclimate and values obtained in the axis of the DCA (Table 3) is placed in Table 4. Graphically, the distribution of collecting localities is presented in Figure 2. The taxonomic units revealed in the present account are encircled. Two observations are noticeable: 1. Tetralophozia pusilla and T. filiformis are closely related in the molecular-genetic respect and grow in similar climatic environments too. Tetralophozia sibirica grows in harsh northern environments and is genetically well different from other two taxa that also correspond distance in climate diagram. 2. Strongly geographically distanced Alaskan and Spanish populations were found to be similar in measured climate variables.
Locality 21 presumably should not be far climatically from Alaskan localities, but shows a strong difference from the climates of all other involved localities. Whether this is local aberration in the climate or regularity correlating with morphology and/or genetics was not tested here.  Table 3.  Table 3.

Taxonomy
As found in the molecular genetic analysis, the Korean-Japanese populations are the most closely related to the populations from Southwest China (Yunnan and Sichuan Provinces). These two groups of populations may be treated as infraspecific units, i.e., as two subspecies of the same species, especially considering the differences in the geographic patterns of the two units. However, we prefer to keep them as separate species, thus following the concept of Stephani [1], and not to create additional infraspecific taxa. Concerning the group of populations from South Siberia (and possibly the specimen from the oroboreal environments in the Russian Far East) and considering the robust genetic difference between it and Korean-Japanese plus Yunnan-Sichuan populations, the group of populations from South Siberia (and possibly the specimen from the oroboreal environments in the Russian Far East) should be treated as the distinct species.

Morphology
At first glance, there is the morphological continuum connecting the largest 'modification' with the smallest 'modification' within single species. However, the comparison of the morphology with the molecular genetic data showed morphological hiatuses between revealed entities along with strictly defined regularities in the distribution. Tetralophozia sibirica, a species of the smallest size in the complex, has the largest cells in the leaf lobe base. The 'intermediate' cell size is shown by T. filiformis, which is characterized by the largest plants within the complex. Finally, T. pusilla, which is intermediate in plant size, possesses the smallest leaf cells. A similar regularity is observed in the stem cross section: the largest cells in the outer layer are in the habitually smallest T. sibirica (10-12 µm in diameter), while the smallest cells are in T. pusilla (7-10 µm in diameter) and then intermediate in size in T. filiformis (7-13 µm in diameter). This 'cell size'; feature certainly possesses a quantitative nature and cannot be used alone. The further feature discriminating T. sibirica from other entities is the strongly thickened outer cell walls in the stem cross section (versus slightly thickened to virtually thin), while inner cells are commonly thick-walled (versus invariably thin in the remaining taxa). The trigones in the stem cross section are always convex in T. sibirica (versus mostly concave in two other). The same regularity is observed in the inner cells of the stem cross section.
Moreover, T. filiformis, the largest taxon of the group, has the thinnest cell walls in the inner part of the stem cross section and smaller (moderate in size, not large) trigones. One feature is a very speculative, although should also be mentioned. The leaf lobes in T. pusilla and T. filiformis are somewhat tuned not only to the stem apex, but also slightly so to the dorsal side to the stem, which provides to the plants the appearance somewhat similar to depauperate Herbertus.
The papillae are prominently coarse in Tetralophozia filiformis (well observable in relatively fresh specimens, less than 10 years old), while the papillae in two other taxa of the complex are slightly developed to virtually absent.
Two other features that distinguish these three taxa transform gradually in the row T. sibirica-T. pusilla-T. filiformis. The teeth occurring on the leaf lobes gradually become larger from T. sibirica, where they by 1-3 in the sinus area and developed only near the base, plus basal tooth on each lateral side of the leaf, up to 8 cells long). Then, T. pusilla possesses more numerous (2-4 per side in the sinus), plus lateral side basal teeth of the leaf also become larger and sometimes even branched. Tetralophozia filiformis shows more prominent teeth, developed in the lower 1/3-1/2 of the lobe, and the lowest of them is more than 10-13 cells long, besides each lateral side base have additional commonly curved and branched, to 20 cells long tooth. The latter taxon shows the same feature in the underleaves, whose lobes are toothed below their middles.
The undivided part in the leaf lamina also varies among the three taxa. The highest undivided part, 3-5 cells, is in the smallest species (T. sibirica); other taxa have larger leaves but are characterized by a lower (2-3 cells) or the same (to 6 cells) height of the undivided part. The leaves are larger in T. filiformis s. str. than in T. sibirica, which gives the impression of the much smaller undivided part in the leaf lamina in the former although in absolute value it is similar.

The Distribution of 'Narrow' Taxa of Tetralophozia filiformis Complex
The map provided by GBIF (https://www.gbif.org/ru/species/2689448, accessed on 18 May 2021) for the worldwide distribution of Tetralophozia filiformis s.l. is in general Plants 2022, 11, 3121 9 of 26 traits the same with the map provided almost 40 years ago by Urmi [2], with the exception of newly added reports from Russian Siberia and the Far East [3,4,10], the Republic of Korea [7] and Vietnam (the data cited in Table 5 are the new record of the species for the country). The general distribution of this species complex could be characterized as amphioceanic oro-boreotemperate, although the specimens in the southern extremes, including North Vietnam, were collected in orosubtropical forests. Scrutinizing the distribution of the T. filiformis complex in Asia, there are three taxonomical entities discussed in this account whose distribution coincides well with certain climate characteristics. Tetralophozia filiformis s. str. has a pronounced Sino-Himalayan distribution, extending from Indian Sikkim, West Bengal, Nepal, Bhutan to southwest China (Yunnan, Sichuan Provinces) and then to the Khoang Lien Range in northernmost Indochina (Lao Cai Province of Vietnam). The taxon is distributed in a warm monsoon climate. We do not know if the Taiwanese material belongs to the same taxon, although considering the wide penetration of many Sino-Himalayan species to Taiwan Island, T. filiformis s. str. can also occur there. The same may be suggested on the distribution of T. filiformis in Sabah (Malaysia).  Tetralophozia pusilla, a long-forgotten taxon, is distributed in a humid oceanic climate, and its distribution can be characterized as Japanese-Korean temperate. The northernmost collections belong to T. sibirica. These collections occur in a pronounced continental climate, and the type of distribution can be characterized as oroboreal Asian.
The question remains then regarding how the populations growing outside Asia should be named. Considering that the occurrences of Tetralophozia filiformis s.l. in western North America are confined to cool and moderate temperate vegetation zones in the areas under oceanic climate conditions, we may assume that these American populations belong to T. pusilla. In contrast, the Spanish populations are unlikely to belong to the same species as the Japanese-Korean T. pusilla. Rather, they should belong to T. sibirica or another taxon that is not described yet. However, this assumption needs further study.
The DCA performed based on climate variables showed that three East Asian taxa of the complex are well differentiated by the climate conditions that confirmed the ecological requirements of all recognized taxa. Moreover, the climates in the localities of T. filiformis s. str. and T. pusilla are much closer to one another than to T. sibirica, which is prominently a continental climate taxon and is characterized by more robust molecular differences with the pair T. pusilla-T. filiformis s. str. Unexpectedly, the climatic localities from Spain were found in the same cluster as Alaska in the USA, which should induce further work on this species complex beyond Asia. If the climatic conditions are again found to correlate with the molecular-genetic differences, it would be quite an unusual distribution type when Pacific American specimens are related not to East Asian populations but to the temperate Atlantic.
Key to Tetralophozia in Amphi-Pacific Asia 1.
Tetralophozia filiformis (Steph.) Urmi Stephani (with very limited exceptions) did not designate holotypes in the materials he studied. Moreover, the vast majority of the collections he studied were sent to him on loan and were returned to the sender after Stephani separated a small part of the specimen for his own herbarium in Leipzig (now is in G). Therefore, at least two type specimens for each taxon are presumed to exist in the majority of cases, and the formal lectotype should be designated. A discussion on this issue, including also the questionable lectotypification of Stephani's taxa by Bonner, is provided by Engel and Merrill [13] with the corresponding references on this issue. This is why the vast majority of taxa described by Stephani should be lectotypified.   Stephani (with very limited exceptions) did not designate holotypes in the materials he studied. Moreover, the vast majority of the collections he studied were sent to him on loan and were returned to the sender after Stephani separated a small part of the specimen for his own herbarium in Leipzig (now is in G). Therefore, at least two type specimens for each taxon are presumed to exist in the majority of cases, and the formal lectotype should be designated. A discussion on this issue, including also the questionable lectotypification of Stephani's taxa by Bonner, is provided by Engel and Merrill [13] with the corresponding references on this issue. This is why the vast majority of taxa described by Stephani should be lectotypified.
Description (based on specimens examined placed in the Table 5). Plants brown to yellowish brown in herbarium, when fresh commonly with green-brownish shoot apices, ascending, never distinctly creeping or erect, forming very loose mats, commonly with admixture of some pleurocarpous mosses, 15-50 mm long and 0.6-0.75 mm wide (dry plants 0.4-0.6 mm), shoots look loosely vermiform, leaves commonly obscurely turned to the dorsal side. Rhizoids virtually absent to sparse, 400-600 µm long, grayish to nearly colorless, soft, undulate, obliquely spreading, separated or in unclear fascicles, originating from the stem near underleaf bases. Branching sparse, terminal, Frullania-type Description (based on specimens examined placed in the Table 5). Plants brown to yellowish brown in herbarium, when fresh commonly with green-brownish shoot apices, ascending, never distinctly creeping or erect, forming very loose mats, commonly with admixture of some pleurocarpous mosses, 15-50 mm long and 0.6-0.75 mm wide (dry plants 0.4-0.6 mm), shoots look loosely vermiform, leaves commonly obscurely turned to the dorsal side. Rhizoids virtually absent to sparse, 400-600 µm long, grayish to nearly colorless, soft, undulate, obliquely spreading, separated or in unclear fascicles, originating from the stem near underleaf bases. Branching sparse, terminal, Frullania-type (superficially looking as dichotomous), rarely ventral intercalary; stem cross section slightly transversely elliptic with smooth external wall, 130-150 × 160-180 µm, with loosely defined cortex, outer cells with thick walls, irregular in shape, ca. 7-13 µm in diameter, inner cells to 25 µm in diameter, with thin walls and large to moderate in size, concave trigones. Leaves transversely inserted, mainly subimbricate or (rarely, in weaker shoots only) obliquely spreading, somewhat turned dorsally; subimbricate leaves distinctly concave, with lower 1/4-1/3 erect spreading, then suddenly curved and going subparallel to the stem; (3-)4lobed, lobes subequal (although weak plants have predominantly 3-lobed leaves, with dorsal lobe larger), 600-800 µm long and (at the level of leaf lamina) 250-350 µm wide, undivided part 30-50 µm (2-5 cells) high, lobe apices prominently acuminate, with straight axis and 2-4(-5)-celled uniseriate ends, sinus strongly recurved in the base, each lobe with 2-4 acute 3-12-celled teeth in lower 1/3-1/2, besides each lateral side of the leaf have one more additional, commonly curved and branched, to 20 cells long tooth. Underleaves bilobed, 400-500 µm long, 150-250 µm wide (at the level of lamina), undivided part 20-40 µm (2-3 cells) high, lobes prominently acuminate with straight or distinctly recurved axes and 3-6-celled uniseriate ends, if uniseriate end shorter then the lobe ends by biseriate end to 8-10 cell-pairs long, sinus margin recurved in the base, lobe lower half in the sinus side with 2-4 teeth, leaf lateral bases with 3-5 teeth, the basal tooth largest. Cells in the lobe base 12-25 × (7-)10-12(-15) µm, oblong to nearly isodiametric, walls strongly vermiculately thickened, with prominent, large, and convex trigones; cuticle distinctly papillose-verrucose, sometimes coarsely so (Generative structures in molecularly studied specimens are absent) (Figures 6-10).       Comment. The type specimen of the taxon is somewhat smaller than other specimens from Yunnan and Sichuan. It has shorter but the same in width leaves and somewhat wider underleaves (although of the same length as other materials). Moreover, the plants in the lectotype are slightly crumbled, which gives the impression of smaller and shorter plants than commonly in the species. The old, dried herbarium specimens commonly have a less pronounced leaf surface armature, including papillae. The latter may be the consequence of the partial collapsing of the cells and indenting of surface elements inside. The latter is the reason the cuticle elements are not as coarse as in relatively fresh material.    [1] wrote the collector name as Yoshinaga, although the collection was only sent to Stephani by Yoshinaga (one of his Japanese correspondents), while the label distinctly indicates the collector name as K. Tamura.
of leaf lamina) 250-450 µm wide, undivided part 50-90 µm (3-6 cells) high, lobe apices prominently acuminate, with straight axis and 1-3-celled uniseriate ends, sinus strongly recurved in the base, each lobe with 1-3 acute 1-6-celled teeth in lower third or basal teeth completely absent (as in the type specimen), each lateral side of the leaf have one more additional, commonly curved and sometimes branched, long (to 12 cells long) tooth. Underleaves bilobed, 300-400 µm long, 120-150 µm wide (at the level of lamina), undivided part 30-50 µm (2-3 cells) high, lobes prominently acuminate with straight or distinctly curved axes and 3-4-celled uniseriate ends, sinus recurved in the base, lobe bases in the sinus with 2-4 teeth, lateral underleaf bases with 3-5 teeth, the basal tooth the largest, sometimes branched and strongly curved, 1-2 basal teeth sometimes terminating with slime papilla. Cells in the lobe base 10-20 × 10-12 µm, oblong to nearly isodiameteric, strongly vermiculately thickened, with prominent, large, and convex trigones; cuticle papillose, sometimes scarcely so. Generative structures unknown. (Figures 11-14) Comment. The same written in the comment about papillae on the leaf and stem surface under T. filiformis should be applied to the type of C. pusillus. The papillae are difficult to observe and the surface sometimes looks virtually smooth. This may be explained by the indenting of surface elements inward and also because papillae in fresh material even are not prominently high in this species.

The Tetralophozia Overview
The genus Tetralophozia started to be widely accepted in Europe from Schljakov [14], who first made the new combinations to raise Chandonanthus subg. Tetralophozia R.M. Schust. into the genus and transferred Chandonanthus setiformis into Tetralophozia. Chandonanthus subg. Tetralophozia R.M. Schust. was described 16 years prior to Schljakov's new combination [15]. Seven years after Schljakov [14], Urmi [2] transferred Chandonanthus filiformis to Tetralophozia, Váňa [16] did the same with Blepharostomum cavallii Gola and Schuster [17]-with Chandonanthus piliferus Steph. The genus Tetralophozia, as accepted by Söderström et al. [18], includes 4 species worldwide. Tetralophozia setiformis generally has an arctic-alpine circumpolar distribution. The species is widely spreading southward in the Holarctic by the mountain ranges in Europe (Alps, Carpathians) but far less so in North America and Asia, where the southernmost localities lie at 43°N in the Russian Far East mainland. The species is also illustrated here ( Figure  15), based on the somewhat depauperate plants from one of the southernmost localities (43ºN) in Asian mainland in Primorsky Territory. The species is almost exclusively epilithic or grows on humus in cliff crevices, as exclusion in some sites in Northeast Asia may occur on lying branches of Pinus pumila, a dwarf shrubby pine. The Tetralophozia filiformis complex, discussed in the present paper in detail, has (as one, although complex Comment. The same written in the comment about papillae on the leaf and stem surface under T. filiformis should be applied to the type of C. pusillus. The papillae are difficult to observe and the surface sometimes looks virtually smooth. This may be explained by the indenting of surface elements inward and also because papillae in fresh material even are not prominently high in this species.

The Tetralophozia Overview
The genus Tetralophozia started to be widely accepted in Europe from Schljakov [14], who first made the new combinations to raise Chandonanthus subg. Tetralophozia R.M. Schust. into the genus and transferred Chandonanthus setiformis into Tetralophozia. Chandonanthus subg. Tetralophozia R.M. Schust. was described 16 years prior to Schljakov's new combination [15]. Seven years after Schljakov [14], Urmi [2] transferred Chandonanthus filiformis to Tetralophozia, Váňa [16] did the same with Blepharostomum cavallii Gola and Schuster [17]-with Chandonanthus piliferus Steph. The genus Tetralophozia, as accepted by Söderström et al. [18], includes 4 species worldwide. Tetralophozia setiformis generally has an arctic-alpine circumpolar distribution. The species is widely spreading southward in the Holarctic by the mountain ranges in Europe (Alps, Carpathians) but far less so in North America and Asia, where the southernmost localities lie at 43 • N in the Russian Far East mainland. The species is also illustrated here (Figure 15), based on the somewhat depauperate plants from one of the southernmost localities (43 • N) in Asian mainland in Primorsky Territory. The species is almost exclusively epilithic or grows on humus in cliff crevices, as exclusion in some sites in Northeast Asia may occur on lying branches of Pinus pumila, a dwarf shrubby pine. The Tetralophozia filiformis complex, discussed in the present paper in detail, has (as one, although complex unit) an oro-temperate-oro-boreal amphioceanic range, and all records are from rocky substrates. Tetralophozia cavalli (Gola) Váňa is confined to the Central African high mountains (Ruwenzori Mt., Virunga Mts., Kilimanjaro Mt.). The species occurs on the bark of trees from middle to high elevations [16]. Tetralophozia pilifera (Steph.) R.M. Schust. is a New Guinean endemic species growing mostly in epiphytic habitats and occurring much more rarely in epixylous habitats [19]. All recognized taxa of the genus are quite variable morphologically, although there have been no attempts to confirm whether this is indeed an environmentally induced variation that does not correlate with molecular-genetic features. unit) an oro-temperate-oro-boreal amphioceanic range, and all records are from rocky substrates. Tetralophozia cavalli (Gola) Váňa is confined to the Central African high mountains (Ruwenzori Mt., Virunga Mts., Kilimanjaro Mt.). The species occurs on the bark of trees from middle to high elevations [16]. Tetralophozia pilifera (Steph.) R.M. Schust. is a New Guinean endemic species growing mostly in epiphytic habitats and occurring much more rarely in epixylous habitats [19]. All recognized taxa of the genus are quite variable morphologically, although there have been no attempts to confirm whether this is indeed an environmentally induced variation that does not correlate with moleculargenetic features. As it was admitted starting from Urmi [2], Chandonanthus pusillus Steph. was treated as the synonym of Tetralophozia filiformis, although Schuster [20] called this synonymy 'presumable' and noted "T. filiformis is more strongly armed with spinescent teeth and cilia, which may occur far up the lobe margins; in C. pusillus, lobes are entire or bear 1-2 small teeth just above their bases". The type localities of Chandonanthus filiformis and C. pusillus are, although in East Asia, situated at a strong distance from one another. The type locality of Chandonanthus filiformis is in Ma'ershan ('Ma Cul Chan' by Delavay), which is in Yunnan Province of China ([1]; G00112121!). The locality is situated within the small As it was admitted starting from Urmi [2], Chandonanthus pusillus Steph. was treated as the synonym of Tetralophozia filiformis, although Schuster [20] called this synonymy 'presumable' and noted "T. filiformis is more strongly armed with spinescent teeth and cilia, which may occur far up the lobe margins; in C. pusillus, lobes are entire or bear 1-2 small teeth just above their bases". The type localities of Chandonanthus filiformis and C. pusillus are, although in East Asia, situated at a strong distance from one another. The type locality of Chandonanthus filiformis is in Ma'ershan ('Ma Cul Chan' by Delavay), which is in Yunnan Province of China ( [1]; G00112121!). The locality is situated within the small Ma'ershan floristic province belonging to the Jinsha River floristic Subregion. This small province houses 19 stenochoric endemic vascular plants, which is quite a lot among 84 floristic provinces within the administrative Yunnan Province of China [21]. The type locality of Chandonanthus pusillus is in Komagadake Mt. in Yamanashi Prefecture ( [1]; G00283305!). The latter is situated within the Southern Japanese Alps [=Minami Alps]-a quite distinctive area due to high level of taxonomic diversity as it was recognized by United Nations Educational, Scientific and Cultural Organization (UNESCO) (http://www.unesco.org/new/en/natural-sciences/environment/ecologicalsciences/biosphere-reserves/asia-and-the-pacific/japan/minami-alps, accessed on 18 May 2021). The bryofloristic richness of this region is also obvious [22,23].

Specimens
The studied specimens of the Tetralophozia filiformis complex that are kept in VBGI, JNU, and KPABG were collected by the authors or other collectors. Types were loaned from G (two types: Chandonanthus pusillus and C. filiformis). Where possible, the specimens were processed via molecular-genetic analysis. The basic principle that was accepted in this work was not only to study quite old-type materials (that are also almost always impossible to sequence) but also to try to collect specimens in type localities or nearby-to reveal their real morphological variability in classic localities and to obtain material suitable for molecular-genetic analysis. Therefore, we visited Komagadake Mt. in Japan and the Ma'ershan area (the position of the type locality is quite indefinite) in Yunnan Province of China. In addition, other areas in southern China (Sichuan Province), the Korean Peninsula and Russian Siberia were explored. In total, 17 specimens (both collected by us, other collectors and requested from other herbaria) were studied: 4 from Russian Siberia, 1 from the Russian Far East, 6 from the Korean Peninsula, 1 from Japan, 1 from North Vietnam and 3 from China. Additionally, three specimens of T. setiformis were included for moleculargenetic comparison. All specimens are listed in Table 5, along with the GenBank accession numbers.

Molecular-Genetic Study
The monophyly of the genus Tetralophozia remains questionable since two species from the genus Plicanthus R.M. Schust. were subsequently found within it [8,24]. Due to the absence of suitable sampling of Plicanthus, we were not able to clarify the generic concept of both genera and concentrated here only on affinity among known Tetralophozia species. For molecular estimation, we selected 11 specimens of Tetralophozia and 11 species from the family Anastrophyllaceae, and Lophozia ascendens (Warnst.) R.M. Schust. from Lophoziaceae was chosen as an outgroup. In total, ITS1-2 nrDNA and trnL-F cpDNA sequence data for 14 samples were taken from our previous studies, GenBank accession numbers for them are provided in Figure 1.
DNA was extracted with a DNeasy Plant Mini Kit (Qiagen, Hilden, Germany). The primers suggested by White et al. [25] for ITS1-2, Taberlet et al. [26] for trnL-F and Shaw et al. [27] for trnG-intron were used for amplification and sequencing. Polymerase chain reaction (PCR) was carried out in 20 µL volumes with the following protocol: 3 min at 94 • C, 30 cycles (30 s 94 • C, 40 s 56 • C for ITS1-2 and trnL-F or 64 • C for trnG-intron, 60 s 72 • C), 2 min of final extension at 72 • C. Amplified fragments were visualized on 1% agarose TAE gels by EthBr staining, purified using the QIAquick Gel Extraction Kit (Qiagen, Germany), and used as a template in sequencing reactions with the ABI Prism BigDye Terminator v. Newly generated sequences were assembled and aligned with previously obtained sequences in BioEdit 7.0.1 [28]. Alignments for ITS1-2 and trnL-F were produced manually; all positions were considered. Due to the absence of appropriate data, the trnG-intron dataset was not produced. Preliminary phylogenetic estimation revealed congruent results from both datasets; thus, they were combined in a single dataset ITS1-2+trnL-F for subsequent analyses by the maximum parsimony (MP) method with TNT v.1.5 [29] and the maximum likelihood (ML) method with PhyML v.3.0 [30]. The MP analysis involved a New Technology Search for the minimal length tree by five iterations and 1000 bootstrap replicates, and default settings were used for other parameters. The software ModelGenerator [31] selected the best-fit evolutionary model of nucleotide substitutions, namely, TN+I+G. The stopping frequency criterion for bootstrapping suggested 450 replicates as enough to reach BS convergence with Pearson average ρ100 = 0.997020 realized in RAxML v7.2.6 [32]. Thus, ML analysis was performed with the TN+I+G model, 500 bootstrap replicates and gamma distribution of the rate heterogeneity among sites with four rate categories.
The average pairwise p-distances for the genus Tetralophozia were calculated in Mega 11 [33] based on each DNA locus using the pairwise deletion option for counting gaps.

Climate Analysis
Since the climate is obviously changing in the distribution range of the Tetralophozia filiformis complex, we obtained the bioclimate variables for collecting localities of the specimens: 1) studied by molecular and/or morphological methods, and 2) randomly selected from the GBIF database (https://www.gbif.org/ru/occurrence/search?taxon_ key=2689448, accessed on 18 May 2021) and for two specimens from Spain on which the first report of the taxon for Europe was based [2]. In total, 22 localities were selected, and 19 bioclimate variables were identified based on information provided in The obtained data were then tested using multivariate analysis (using Past ver. 4.03c [34]). The hierarchical clustering was based on Ward's method [35], and Euclidean distance was used to check the results shown by detrended correspondence analysis (DCA). DCA was visualized in a three-dimensional grid graph, with the third dimension given by the color gradient.

Conclusions
The study illustrates that when only a few materials and from limited regions are available, the distinctive traits of the regional 'populations' may be overlooked, and a potentially new taxa could be neglected. In our case, when we had an experience restricted to the specimens collected in South Siberia, we certainly could not recognize that a taxon different from Tetralophozia filiformis is in hand because all our attention was attracted to the robust difference of collected material from T. setiformis (another locally known taxon). The involvement of additional material from type localities (including types) derived from other regions in East Asia revealed noticeable variation within widely treated T. filiformis in four main groups of species features: molecular genetics, morphology, ecology (including climate characteristics), and geography. The situation is somewhat similar to the recently published Ptilidium himalayanum, the species molecularly more different from the pair P. ciliare-P. pulcherrimum than the taxa constructing the pair and occupying a distinctly defined area in the Sino-Himalaya [36].
Treating Tetralophozia sibirica as the terminal link in the adaptation process to the cold and dry climate within T. filiformis s. lat. complex, then it could be stressed the main morphological pathways of this 'adaptation' (shortening leaf lobes, not so deeply divided leaves, sparser leaf dentation) seem imaginable in regard to the northerly distributed Arctic-alpine T. setiformis (showing the same features in much more pronounced manner). This regularity to possess shorter divided leaves northward may be compared with trends observed in other genera, e.g., (1) Scapania with deeply divided leaves are not present in Arctic-distributed taxa (in spite of generally high taxonomic diversity of this genus in 'high' latitudes), (2) the wide distribution of Gymnomitrion taxa with shallowly-lobed leaves in the Arctic and the absence taxa with deeply-divided leaves there.