The Taxonomically Richest Liverwort Hemiboreal Flora in Eurasia Is in the South Kurils

The long coexistence of various floral elements, landscape diversity, and island isolation led to the formation of the richest Eurasian hemiboreal liverwort flora in the southern Kurils. This land that covers less than 5000 square kilometres and houses 242 species and two varieties of liverworts and hornworts. The flora ‘core’ is represented by hemiboreal East Asian and boreal circumpolar taxa. Other elements that have noticeable input in the flora formation are cool-temperate East Asian hypoarctomontane circumpolar and arctomontane. The distribution of some species is restricted to the thermal pools near active or dormant volcanoes or volcanic ash deposits; such species generally provide specificity to the flora. Despite the territorial proximity, the climate of each considered island is characterized by features that, in the vast majority of cases, distinguish it from the climate of the neighbouring island. The last circumstance may inspire the difference in the liverwort taxonomic composition of each of the islands. The comparison of the taxonomic composition of district floras in the Amphi-Pacific hemiarctic, boreal, and cool-temperate Asia revealed four main focal centres: East Kamchatka Peninsula and Sakhalin Island, the southern Sikhote-Alin and the East Manchurian Mountains, the mountains of the southern part of the Korean Peninsula, and the South Kurils plus northern Hokkaido. The remaining floras involved in the comparison occupy an intermediate position between these four centres.


Introduction
The South Kurils include Iturup, Kunashir, Shikotan, and a group of small Habomai islands. All are located in the southern part of the Kuril Island Chain to the north of Hokkaido and cover a small area of approximately 5000 square kilometres. Despite a small size, the South Kurils have unique vascular plant flora in which arctomontane and cool-temperate elements are present [1]. The liverworts in this promising flora were only poorly studied until the beginning of the 21st century. By 2005, only 53 liverwort species were known to the South Kurils [2], and no hornwort species were reported. The related group, mosses, was studied considerably more due to studies conducted in the late 1970s and early 1980s by moss specialists Bardunov and Cherdantseva [3]. Great advances in the knowledge of bryophytes in the South Kurils were made in the middle of the first decade of the 21st century, when approximately 4000 specimens of bryophytes were collected due to the joint efforts of a group of bryologists from 2005 to 2007. These collections have formed the background for the first reliable reports on the taxonomic diversity of these groups. For the South Kurils, as they are understood in this work (see Section 3), slightly over 200 species were reported [4], which is the great number for the district's hemiboreal flora of relatively small-sized area.
Naturally, even at the time of publication for the checklist mentioned above [4], it could not be considered completely exhaustive since there were no data on taxonomic diversity in a number of areas on the islands. The incompleteness of the data presented in that paper (l.c.) has become more evident 13 years after publication. During the elapsed time, (1) new taxa have been described that did not appear in the previous checklist, (2) the taxonomic concept of a number of species was revised and sometimes drastically altered, (3) incorrect or doubtful species identifications were corrected, and (4) new materials were collected in a number of previously unstudied or insufficiently studied areas. It is worth mentioning that, in contrast to the materials on which the previous work was based [4], all newly collected materials were delivered alive to the cryptogamic biota laboratory (herbarium acronym VBGI), where oil bodies were studied and photographed for the vast majority of taxa. Several of the obtained photographs were published by Bakalin et al. [5]. The identity of several specimens was also tested using molecular genetic analysis. These facts suggest that the study of newly collected materials was conducted using a broader scale approach. As a result of the undertaken research, it is necessary to compile a new list of liverworts for the South Kurils. The publication of the new checklist along with a phytogeographic discussion is the main goal of the present study.

Results and Discussion
The main result of the field research and subsequent identification is an updated list of liverworts of the three southern Kuril Islands that includes 242 species and two varieties. These species are listed below. A total of 183 taxa are known for Iturup Island, 179 for Kunashir, and 148 for Shikotan. Thirty-two taxa are newly recorded for South Kurils, including four taxa (Calypogeia japonica Steph., Calypogeia tosana (Steph.) Steph., Nipponolejeunea pilifera (Steph.) S. Hatt., Solenostoma hokkaidense (Váňa) Váňa, Hentschel and Heinrichs). Although these were newly found in Iturup from field work in 2015 (Iturup Island), no localities were provided in the publication devoted to that island [6]. Below, we provide the list of the species recorded in the South Kurils with the data on the distribution within I (Iturup Island), K (Kunashir), and S (Shikotan). Doubtful records are listed together with confirmed records, although with corresponding references. Reidentified or, for other reasons, excluded species are listed in a separate section; this list includes 33 taxa. The field numbers are provided for the newly recorded taxa (the taxon known prior for the South Kurils in general is supplemented with field numbers only for the islands for which it is newly recorded), and the field number citation immediately follows the island name abbreviation (I, K, or S). No field numbers are cited for the species already known for the corresponding islands. The cited specimens from Iturup were collected by Bakalin (if otherwise not mentioned), whereas those from Kunashir and Shikotan were collected by Bakalin and Klimova. Some other changes (simply nomenclatural or resulting from the revisions of certain groups) are explained in the comments. If the species was reported for the South Kuril flora after 2009 but before the present account, the corresponding note and reference are provided. New records for islands are marked by asterisks following the island name abbreviation (in total 41 taxon were newly recorded on specific island), whereas new records for the South Kurils are indicated by an asterisk proceeding to the species name in the list. Simple name changes, e.g., all Calycularia crispula Mitt. Were transferred to C. laxa Lindb. and Arnell or, similarly, all Bazzania ovifolia (Steph.) S. Hatt. are now treated as B. denudata (Lindenb. et Gottsche) Trevis.) are not marked as newly recorded species, but are referenced with corresponding notices that the name used here is not an original name under which the species were previously reported in the flora. The homotypic synonyms used before are given in the checklist in the brackets. The simple transfer of species reports to the report under the heterotypic synonym of the same species and are not marked as new reports, although they are discussed in the species annotation.
The taxa are arranged alphabetically, and nomenclature follows Söderström et al. [7] with some updates from the recent literature (the most valuable is the 'narrow' genus concept in Solenostomataceae and acceptance of Pseudolophozia as distinct from Barbilophozia). The abbreviated island names where the taxon occurred are followed by the species name. The taxonomic or other notices (if any) are placed at the end of the comment sentence. view of current advances in the systematic of Frullania.
*  [4] was collected in Kunashir, but mistakenly indicated as Shikotan in the cited paper. It contains plants with characteristically large cells in the leaf, but the insertion and decurrency of the leaf is more similar to that in F. pleniceps. We are not sure plants in the specimen belong to 'true' F. pleniceps or F. connivens.   [4]) was argued by Zhu and So [24]. Our specimens belong to Microlejeunea punctiformis.
Mylia anomala (Hook.) S.Gray-I, K, S.  [4] based on study of previously dried material without oil bodies available for study and prior to the description of Nardia pacifica (described by Bakalin and Klimova [25]). Some of these reports surely belong to N. pacifica. Nardia japonica was confirmed based on fresh material comparison for Iturup and Shikotan islands only (the studied specimens are provided here).
Nardia pacifica Bakalin-I; comment: reported for Iturup in [25] based on study of fresh material with oil bodies available. Some of earlier reports of N. japonica may belong to this species.
Nardia subclavata (Steph.) Amakawa-I, K, S. Nardia unispiralis Amakawa-I. ; comment: this species is newly recorded for Russia and this record was briefly discussed by Bakalin et al. [6], but no specimen citation were provided in l.c.
Bazzania ovifolia (Steph.) S. Hatt.-the synonymy with B. denudata is confirmed by Bakalin [9] and all reports are transferred to the latter.
Calypogeia azurea Stotler et Crotz-excluded from the flora of the Russian Far East, all records transferred to C. orientalis [13].
Lejeunea ulicina (Taylor) Gottsche, Lindenb. et Nees -excluded from the flora of the South Kurils due to transfer of all specimens to Microlejeunea punctiformis-morphologically similar, but nevertheless different species, as it was argued by Zhu and So [24].
Cryptocoleopsis imbricata, a species of young volcanogenic substrate (scoria and volcanic ash mainly that in the crevices of tufa and other pyroclastic-derived cliffs). Iturup and Shikotan common species mainly possess boreal and Arctic-boreal circumpolar distributions, with the exception of two bright East Asian taxa: Hattorianthus erimonus and Wiesnerella denudata. All these species may also be found on Kunashir Island. Kunashir and Shikotan common taxa mostly have East Asian (temperate to arctic-alpine) or widely temperate distributions as determined from their southern location in comparison with Iturup Island: Blepharostoma minus, Cololejeunea macounii, Frullania koponenii, Frullania takayuensis, Lejeunea japonica, Metzgeria temperata, Microlejeunea punctiformis, Phaeoceros carolinianus, Plectocolea rigidula, Scapania rigidula, Solenostoma bilobum, and Trichocolea tomentella. Other floral elements within the common species fraction (boreal, arctomontane, hemiboreal circumpolar, etc.) are in the minority and include only five species.
The specificity of the certain island is quite large: The distribution of 37 species is restricted to Iturup, 25 to Kunashir, and 17 to Shikotan.
As mentioned above, the total known taxonomic liverwort diversity in the South Kurils includes 242 species, which is the highest number for hemiboreal flora of similar and even much larger sizes. If to directly compare the number of known species in the South Kurils with the species numbers for European countries [36], then very few countries would be found ahead of the South Kurils in terms of the number of species, although such a comparison is certainly not appropriate on the basis of logic. Moreover, the highest numbers exceeded South Kurils flora by no more than 30%. Among the countries with similar taxonomic diversity to the South Kurils are the following (despite the significant differences in the area, the knowledge history, and landscape diversity): Austria-260 species, Switzerland-263, Germany-252, Finland-222, Ireland-235, Poland-233, Sweden-262, and Slovakia-223 species. The South Kurils are far ahead of floristically rich and well-studied countries such as Belgium (192 species) and Denmark (144 species), and somewhat lag behind the most taxonomically rich countries: Great Britain (293 species), Spain (276), France (308), and Norway (277).
Of course, such a comparison is surely inappropriate from the size of the compared units but clearly illustrates the taxonomic richness of the liverwort flora of the South Kurils. There are few data to compare the diversity in the South Kurils with adjacent areas. The large area (approximately 200,000 square km) lying mostly in the temperate zone of the Korean Peninsula houses 326 liverwort and hornwort species [37]. However, if we compare the number of species with that known for Hokkaido Island, which is much larger in area (83,424 square km), more diverse in terms of landscape and climate, and lies mostly in the temperate zone, then the comparison will be in favour of the South Kurils: 182 species are known in Hokkaido [38] versus 242 in the South Kurils. The Chichibu-Okutama Mountains in central Japan (belonging to the warm temperate zone at lower elevations) house 230 species [39].
An illustrative comparison can be done with the adjacent areas that are relatively well-studied: the Paramushir Island liverwort flora (with a clearly hemiarctic vegetation) includes 85 species [40]. Attu Island, situated in a similar zone, counts 112 species [41]. However, these islands are located in a hemiarctic vegetation zone. Hokkaido Island, mentioned above, is much larger. Among smaller islands situated near (in similar climatic conditions with the South Kurils) is Rishiri Island, counting 92 species in a small-sized area of only 183 square km. The large hemiboreal flora of the southern part of Primorsky Territory (approximately 80,000 km 2 ) includes approximately 150 species ( [42], with our additions). When comparing some local floras in temperate Europe, the South Kurils are also ahead of the majority of districts. The list of liverworts of Auvergne [43] includes 189 species for land covering 13,796 square km. The Alpi Apuane counts 128 species [44] known on the land approximately 1000 square km in size. The Azores house 153 species on 2351 square km of land.
Therefore, it was found that the flora of the South Kurils is richer than many larger places; the species richness of the South Kurils with 242 species in an area of <5000 square kilometers is comparable to the biodiversity of countries that are many times larger. We provide a graphical species-area diagram with species number plotted as a function of area ( Figure 1). The Arrhenius SAR is fitted to the areas in Europe. The diagram is based on Table 1. As it easy to see, almost all hemiboreal and temperate floras in East Asia (blue colored dots) are above of the Arrhenius SAR (species-area relationships) fitted for Europe, whereas all hemiarctic sites in Northeast Asia (green colored dots) are below the Arrhenius SAR. This clearly highlights the diversity of the South Kurils versus other places in Eurasia. Meanwhile, the low position of Hokkaido Island (dot number 11 in Figure 1) evidently shows that its flora is strongly undercollected.
Plants 2022, 11, x FOR PEER REVIEW 14 The comparison provided above is formalized in the provided diagram (Figu and based entirely on Table 1.   Table 1. The data on taxonomic diversity and area size in selected floras used for graphic diagram presented in Figure 1. The data on diversity are from [36][37][38][41][42][43][44] and the present paper. The data on the area are from https://en.wikipedia.org/wiki/Main_Page, (accessed on 19 August 2022). Figure 1 Land Name Size, Square km (Sometimes Approximate)

Number of Species
Color in Figure 1 1 The comparison provided above is formalized in the provided diagram ( Figure 1) and based entirely on Table 1.
The question arises: what causes such a high diversity of the liverwort flora in the South Kurils? Possible explanations may be (1) landscape diversity, including differentiation between islands, (2) position on the migration path from/to East Asia (cf. [45]), and (3) the island's isolation effect.
The landscapes in the islands vary in lower elevations from gentle sloped plains to hilly plains (in fact, most of Shikotan Island is a hilly plain). At higher altitudes, the relief occasionally becomes similar to the alpine landscapes with steep slopes and narrow peaks: the most distinct examples are the near-top part of the Bogdan Khmelnitsky Volcano (Iturup) and in the immediate vicinity of Rurui Mt. (Kunashir). However, more often, the relief is intermediate between alpine-type and hilly, somewhat smoothed. The intensity of volcanic activity and age passed from the eruption events nearby (from modern to Miocene in the past) also affects the age of the substrata and the degree of their degradation, drainage abilities, and occupancy by vegetation. Towards the tops of the mountains, the amount of precipitation increases quite significantly (by 30%, exceeding that observed at lower altitude levels), whereas the temperatures decrease. In addition, the temperatures change along the latitudinal gradient (gently rising to the south). Landscape and climatic diversity suggest a diversity of community types ranging from broadleaved (including cool-temperate Magnolia hypoleuca Siebold and Zucc., Fraxinus L. spp., Phellodendron sachalinense (F. Schmidt) Sarg., Kalopanax septemlobus (Thunb.) Koidz.), dark coniferous (including cool-temperate Taxus cuspidata Siebold and Zucc.) forests, crooked forests, shrubby areas, Sasa impassable 'meadows', mountain tundra, and tundralike communities on heavily wind-stressed slopes.
In addition to the diversity of landscapes and vegetation types, which increase the potential number of occurring plant species, the South Kurils are one of the inevitable links on the route of the most important floral exchange between Japan and Northeast Asia and even farther, with Pacific North America. In general, this phenomenon was formulated by Engler [46] and scrutinized by Takeda [47], Tatewaki [48], and others. Despite the active volcanism on the two largest islands of the South Kurils, which partially destroyed the flora (and, in this sense, impoverished its composition due to the 'falling out' of species that were not able to actively generate propagules), the South Kurils are a final destination for several species on the way to movement to the north and to the south. The most striking examples of migrants from the south are Alobiellopsis parvifolia, Aneura maxima, Calypogeia japonica, Cololejeunea spp., Frullania spp., Hattorianthus erimonus, Nipponilejeunea pilifera, and others, which do not pass the South Kurils, and the South Kurils are the northernmost distribution point in Pacific Asia. There are fewer migrants from the north that were 'stopped' in the South Kurils; among them are Cephaloziella uncinata, C. elashista, Endogemma caesipticium, and Gymnomitrion concinnatum; all those listed do not occur in Japan [38]. The macro-climate in the islands situated northward becomes much colder, and the North Kurils have typical hemiarctic vegetation, thus the conditions are probably not suitable for temperate taxa. Another reason for the existence of the area edge of southern taxa here are the wide oceanic straits between islands northward: the distance between Kunashir and Iturup Islands is ony 21 km, whereas the gap between Urup and Simushir Islands is 107 km, and from Simushr to Onekotan Island is 291 km, with only small islates between them that potentially make the probability of dispersal lower.
Finally, the third (particularly difficult to measure) factor promoting the high taxonomic diversity of liverworts in the South Kurils is the isolation effect in the islands. In this series, Shikotan Island stands out. This island is located away from the Great Kuril Chain, having no active volcanism evidence starting from the Miocene and is the final destination point of migration from the south in the Lesser Kuril Chain because there are no more islands immediately northwards of Shikotan. The presence of Dasiphora fruticosa, which is absent in Kunashir, and the absence of dwarf pine (Pinus pumila (Pall.) Regel), common on other large islands of the South (and North) Kurils, are widely known examples of Shikotan specificity [1]. The isolation from other islands of the South Kurils and Hokkaido is due to the high distance of Shikotan from the nearest island of the South Kurils (Kunashir), along with separation from Hokkaido Island through the long but flat Nemuro Peninsula (reducing the potential for migration of oro-boreal and arctomontane species), small-sized and flat Habomai Islets, and oceanic straits. This isolation might contribute to the conservation of a number of species that are currently not known on other islands of the South Kurils. There are 17 species known on Shikotan and not recorded in other South Kurils islands. Interestingly, six of them (Cololejeunea subkodamae, Lejeunea flava, Lejeunea otiana, Nardia insecta, Pseudomoerckia blyttii, Riccia huebeneriana, Scapania mucronata) are not known even in Hokkaido (the main potential 'source' of liverworts for Shikotan Island), and two more (Moerckia flotoviana, Nardia insecta) are not seen in Japan at all. However, considering the incomplete data on the diversity of liverworts in Hokkaido, at least some of the listed taxa may be expected there.
A comparison of bioclimates following Table S1 by the DCA method is presented in Figure 2.  (Table S1) comparison by the DCA method.
Prior to statistical comparison, it was assumed that the landscape and orogra position would be decisive for the relationship between climate data from differen calities (conditionally speaking, all localities in the lower altitudinal belt could be un whereas the localities from higher elevations should form a separate cluster). Indeed data on the average annual temperature (Table S1) seemed to confirm this assump However, when all available data were analysed together, it was found that the local (with three exceptions) formed fairly clear clusters corresponding to the islands. Mo ver, in Figure 2, the sequence of clusters corresponds to the island sequence in the d tion of Iturup-Shikotan-Kunashir. The loosest cluster is formed by the climate data localities on Iturup Island. In addition, two localities are separated from the group o other Iturup localities. The climate in one of the highest elevations of the island (the cr of the Stockap Mountain (actually dormant volcano), point 8 in Figure 2) with a ra cold climate is left out. Point 10 is nearest to it and belongs to the atrio of Tyatya Volc in Kunashir. Both localities are characterized by a humid cool climate with negative m annual temperatures. It is interesting that the peak of Kamuy Mt. (1200 m a.s.l. in the north of the South Kurils) is clearly localized with other Iturup local climates (comple merging with the climate of the Vetrovoy Isthmus, point 2 in Figure 2). Another p distanced from the rest of Iturup climate localities is point 7. This locality character the climate of the isthmus to the Atsonopuri Volcano-one of the few places on Itu Island where good coniferous forests of Picea jezoensis (Siebold & Zucc.) Carrière are veloped. In this sense, the embedding of this locality into the Shikotan cluster is q understandable (Picea jezoensis is the most common tree on Shikotan Island). These ferences in the climatic conditions confirm that the differences observed in the f composition of each island are not stochastic effects arising from undercollecting.
An analysis of the phytogeographic position of the small-sized floras (similar in to the islands considered in the present account involved in the comparison) showed results to be somewhat similar, but more statistically sound, in comparison with w was revealed in the previous study by Bakalin et al. [6]. The South Kurils cluster (IV indeed distanced from true East Asian (III) and circumboreal floras (I) (Figure 3).  (Table S1) comparison by the DCA method.
Prior to statistical comparison, it was assumed that the landscape and orographic position would be decisive for the relationship between climate data from different localities (conditionally speaking, all localities in the lower altitudinal belt could be united, whereas the localities from higher elevations should form a separate cluster). Indeed, the data on the average annual temperature (Table S1) seemed to confirm this assumption. However, when all available data were analysed together, it was found that the localities (with three exceptions) formed fairly clear clusters corresponding to the islands. Moreover, in Figure 2, the sequence of clusters corresponds to the island sequence in the direction of Iturup-Shikotan-Kunashir. The loosest cluster is formed by the climate data for localities on Iturup Island. In addition, two localities are separated from the group of all other Iturup localities. The climate in one of the highest elevations of the island (the crater of the Stockap Mountain (actually dormant volcano), point 8 in Figure 2) with a rather cold climate is left out. Point 10 is nearest to it and belongs to the atrio of Tyatya Volcano in Kunashir. Both localities are characterized by a humid cool climate with negative mean annual temperatures. It is interesting that the peak of Kamuy Mt. (1200 m a.s.l. in the very north of the South Kurils) is clearly localized with other Iturup local climates (completely merging with the climate of the Vetrovoy Isthmus, point 2 in Figure 2). Another point distanced from the rest of Iturup climate localities is point 7. This locality characterizes the climate of the isthmus to the Atsonopuri Volcano-one of the few places on Iturup Island where good coniferous forests of Picea jezoensis (Siebold & Zucc.) Carrière are developed. In this sense, the embedding of this locality into the Shikotan cluster is quite understandable (Picea jezoensis is the most common tree on Shikotan Island). These differences in the climatic conditions confirm that the differences observed in the floral composition of each island are not stochastic effects arising from undercollecting.
An analysis of the phytogeographic position of the small-sized floras (similar in size to the islands considered in the present account involved in the comparison) showed the results to be somewhat similar, but more statistically sound, in comparison with what was revealed in the previous study by Bakalin et al. [6]. The South Kurils cluster (IV) is indeed distanced from true East Asian (III) and circumboreal floras (I) (Figure 3). This cluster (IV) also clearly includes the flora of Rishiri Island. Mountain flora with a full elevation gradient of vegetation (from temperate broadleaved forests to crooked forests and dark coniferous forests near the peaks) of the middle part of South Korea (III) are also combined into one clade. The southern Sikhote-Alin floras and Changbaishan also form one clade (II) (although it should be noted that the apparent proximity of Changbaishan Mt. flora to Olkhovaya Mt. flora in the diagram is just apparent, since these floras are quite far away along the third axis, shown by the colour gradient). The oceanic floras of the Kamchatka Peninsula and Sakhalin Island form another distinct cluster (I). Eight local flora are not included in any of the clusters. The reasons for this exclusion are insufficient knowledge (a small number of species, which makes the analysis unreliable), regional (including climatic) features, and incompleteness of the spectrum of communities along the altitudinal profile. Thus, the flora of Bystrinsky Nature Park is quite rich but developed in a subcontinental climate, which leads to differences in its structure from suboceanic floras. The flora of Ayan surroundings belongs to the hemiarctic floras of the Sea of Okhotsk coast, with a subcontinental climate and dry winters. Tardoki  This cluster (IV) also clearly includes the flora of Rishiri Island. Mountain flora with a full elevation gradient of vegetation (from temperate broadleaved forests to crooked forests and dark coniferous forests near the peaks) of the middle part of South Korea (III) are also combined into one clade. The southern Sikhote-Alin floras and Changbaishan also form one clade (II) (although it should be noted that the apparent proximity of Changbaishan Mt. flora to Olkhovaya Mt. flora in the diagram is just apparent, since these floras are quite far away along the third axis, shown by the colour gradient). The oceanic floras of the Kamchatka Peninsula and Sakhalin Island form another distinct cluster (I). Eight local flora are not included in any of the clusters. The reasons for this exclusion are insufficient knowledge (a small number of species, which makes the analysis unreliable), regional (including climatic) features, and incompleteness of the spectrum of communities along the altitudinal profile. Thus, the flora of Bystrinsky Nature Park is quite rich but developed in a subcontinental climate, which leads to differences in its structure from suboceanic floras. All of them, with the exception of Wiesnerella denudata (having a broader distribution, although with an area core in East and Southeast Asia), are chiefly East Asian in distribution.

Materials and Methods
To analyze the specifity and the position of the liverwort flora of the South Kurils among floras of Northeast Asia, we worked in three directions: (1) to compile the checklist of the South Kuries using available herbarium collections and literature data, (2) to compile the database on liverwort distribution within selected local floras in Northerast and cool-temperate East Asia, and (3) to collect available data for the comparison of taxonomic richness with some European countries and natural units (such as Alpi Apuane and Auvergne). More details are discussed below.
The available herbarium specimens of taxa whose distribution in South Kurils might be questioned in light of the currently accepted distribution concepts were revised. A suite of the new materials was collected in 2015 (Iturup Island), 2018 (Kunashir Island), and 2020 (Shikotan Island) and subsequently identified. The field expedition of 2015 to Iturup Island was remarkable since a group of bryologists (both moss and liverwort specialists) from various Russian institutes have visited Iturup Island, where they carried out a study mainly in the northern part of the island, which previously remained completely unexplored. Based on these collections, some newly found liverwort species on the island were mentioned by Bakalin et al. [6], but their locations were not provided in the paper. Several new distribution data were published in some taxonomic papers, including those on Ricciaceae [29], Jubulaceae [49], Lejeuneaceae [33], Radulaceae [28], Frullaniaceae [18] and several others cited in the checklist.
Data on the area and latitudinal characteristics of the islands are shown in Table 2, and the places of liverwort collection in the last ten years are shown on the map (Figure 4) and Table 3. The Habomai Islands were not studied, but data on them are provided for comparison in Table 2. The total area of the South Kurils is 5018 km 2 (4966 km 2 without Habomai). It should be noted that, in contrast to our previous work [6], we did not classify the islands northwards of Iturup as the South Kurils. This was done based on (1) the lack of at least minimally appropriate data on liverworts of these islands and (2) the lack of new information on these islands, which would enrich our work and would allow us to discuss the position of these islands from a phytogeographical point of view.   Table 3. Map with marked locality numbers is provided in Figure S1.   Table 3. Map with marked locality numbers is provided in Figure S1.    General data on the climate of the South Kurils were provided by Bakalin et al. [4]. All climate data provided before (l.c.) are based on data from weather stations located in large settlements, at low altitudes, and in open areas. Although the general trends are universal across the South Kurils, the climate is oceanic, cool, with the coldest month being February, with high snow cover (up to 2.5 m) and cool, humid summers. However, they hardly reflect the existing climate under which liverworts survive. Since it is obvious that the meso-relief elements and the elevation above sea level should influence the local climate features, we compiled a list of localities (some of those are the same as the collecting localities) for which we obtained data on 19 bioclimates (https://www.worldclim.org/, accessed on 24 February 2020). The bioclimate data are provided in Table S1 using  Since the highest precision with which the bioclimates may be measured is 30", this parameter clearly illustrates mesoclimatic conditions, not microclimatic conditions. Therefore, bioclimate data can be used only with certain reservations in regard to analysing bryophytes commonly growing in microniches. However, as a general pattern descriptor, the biocli-mates work and correlate quite well with, for instance, the fact that the vast majority of arctomontane species grow in the tundra belt and do not occur in coniferous-broadleaved forests despite the great variation in microclimates. Table S1 shows that the average annual temperature varies from slightly negative values in the apical parts of the mountains to +5 • C at lower altitudes. The total amount of precipitation varies from 1000 to 1400 mm, and it clearly increases with elevation above sea level and (to a lesser extent) in the direction from west to east. When comparing pairs of bioclimates, BIO16, 17 and BIO18, 19, it is obvious that the coldest quarter is also the driest (for Kuril bryophytes, this is not as important since the snow cover still covers most of the habitats). At the same time, the wettest quarter is not the warmest, although they distinctly overlap one another (the first shifts closer to the beginning of summer).
To describe in terms of statistics whether the climate characteristics are different from one island to another and to identify the value changes of the climate with elevation, we used the DCA method, the matrix for which is shown in Table S1. This method was previously used in a similar situation in the work with North Vietnam liverwort distribution [50].
The South Kurils flora was variably treated in various phytogeographic classifications, although almost all classifications referred to it as the East Asian floristic region. Various views on this issue were discussed by Bakalin [51]. Contrary to floristic regionalization, the vegetation of the lower altitudinal belts was recognized as hemiboreal starting from Ahti et al. [52] and is now widely accepted. Although it is worth mentioning, in the large-scale altitudinal variation (with mountain tundra vegetation in upper reaches) across the South Kurils, it is possible to postulate the vegetation as hemiboreal only in a very conditional sense. The main features of vegetation and landscapes were described in our previous work [4] and do not need to be repeated here. However, we provide the most characteristic types of habitats in Figures 5-7.
As noted above, the collegial fieldwork in the northern part of Iturup Island carried out in 2015 did not result in a taxonomic checklist (although several species were mentioned in the paper text). However, the attempt to understand the phytogeographic position of Iturup and adjacent islands was made based on data on both moss and liverwort distributions (as far as it was available at that moment) in North Pacific Asia. It was suggested that the bryophyte flora of South Kurils belongs to its own phytogeographic province, which occupies an intermediate position between the East Asian and circumboreal floristic regions but does not definitely belong to either region. The intermediate nature of the liverwort flora of the South Kurils (between the Circumboreal and East Asian floristic regions) was shown even earlier [45].
Since 2009, the data on the taxonomic diversity of the South Kuril Islands have been significantly supplemented in the course of conducted work. The taxonomic lists of each island were used as a single unit in the analysis. Fortunately, for comparison purposes, progress was also achieved in the Korean Peninsula, where three more national park liverwort floras have been published since 2009. Each national park in Korea is somewhat comparable by the area with other floras involved in the analysis. The latter made it possible to enrich the matrix with new data, making the comparison more reliable. The compared floras are listed in Table 4. The floras located far to the north from South Kurils (areas in Northeast Asia with continental to ultracontinental climates) were removed from the matrix used before, which made it possible to avoid the visual effect of "sticking together" of taxonomically distant floras in hemiboreal and cool temperate amphi-Pacific Asia. The compiled matrix was analysed by the DCA method, as was also done in the previous work on liverworts from the vicinity of the Ayan Settlement [53] and in the abovementioned work on Northern Iturup [6] and North Vietnam [50]. In total, 475 taxa were involved in the analysis.     As noted above, the collegial fieldwork in the northern part of Iturup Island carried out in 2015 did not result in a taxonomic checklist (although several species were mentioned in the paper text). However, the attempt to understand the phytogeographic position of Iturup and adjacent islands was made based on data on both moss and liverwort

Conclusions
The isolation effect, combined with landscape diversity and position on the migration route from East to Northeast Asia, led to the formation and preservation of the richest district liverwort hemiboreal flora in Eurasia. There are no floras equal in diversity and of similar sizes in Europe. In contrast, in Pacific Asia, the highest taxonomic diversity per small-sized area is reached south of Kyushu, where on the small island of Yakushima, only 505 km 2 in size, 310 species of liverworts and hornworts are known [75]. This is, however, an absolutely incomparable case because the flora of Yakushima Island in lower elevation belts is clearly subtropical.
It is difficult to imagine how much more the number of known taxa in the South Kurils will increase in future studies; however, given that the southern tip of Iturup Island has not yet been studied, there are no or only fragmentary data on a number of areas of Kunashir Island (especially in its middle part), and the liverwort systematic progress means the description of a coupes of new taxa, several new findings are highly probable. Considering the high taxonomic diversity of the liverworts, it is likely that conservation measures are needed to preserve the taxonomic diversity in this pristine corner of hemiboreal insular amphi-Pacific Asia.