Moss-inhabiting diatoms from two contrasting Maritime Antarctic islands

1Charles University in Prague, Faculty of Science, Department of Ecology, Viničná 7, CZ-128 44 Prague 2, Czech Republic 2Wladyslaw Szafer Institute of Botany, Polish Academy of Sciences, Lubicz 46, PL-31-512 Cracow, Poland 3Botanic Garden Meise, Department of Bryophyta & Thallophyta, Nieuwelaan 38, BE-1860 Meise, Belgium 4University of Antwerp, Department of Biology, Universiteitsplein 1, BE-2610 Wilrijk, Belgium *Author for correspondence: k.kopalova@hotmail.com


INTRODUCTION
The Maritime Antarctic vegetation is strongly affected by its geographical isolation as well as the climatic and environmental conditions prevailing in this region.A classical description made by Holdgate (1970) defined the Maritime Antarctic region as the zone between the southern limit of the extensive, closed phanerogamic vegetation to the southern limit of the extensive cryptogamic (mainly bryophyte) communities.This roughly comprises all landmasses between 70°S northwards to 55°S, including several islands and ar-chipelagos (South Sandwich, South Orkney, South Shetland Islands, Palmer Archipelago), as well as the west coast of the Antarctic Peninsula south to Marguerite Bay.The presence of vascular plants is limited to only two native species reaching their southern limits on north-west Alexander Island in the Maritime Antarctic region: Deschampsia antarctica Desv.and Colobanthus quitensis (Kunth) Bartl.(Ochyra et al. 2008).The Maritime Antarctic vegetation is therefore restricted to poorly developed tundra, of which lichens and mosses form the dominant component, mostly present on moist, low altitudinal, sheltered (north-facing) coastal habi-Background and aims -The Maritime Antarctic vegetation is a poorly developed tundra dominated by lichens and mosses, mostly present in moist environments, providing a favourable habitat for microorganisms.Although, diatoms represent one of the most common algal groups in the Antarctic region, moss-inhabiting diatoms are rarely studied.The moss vegetation on islands in the Maritime Antarctic region forms a favorable habitat for non-marine diatoms.These moss-inhabiting diatom communities are of particular interest as little is known about their species composition, ecological preferences and habitats.The present paper discusses the diversity of moss inhabiting diatoms from Byers Peninsula (Livingston Island) and Ulu Peninsula (James Ross Island), Maritime Antarctic region.Methods -The composition of the moss inhabiting diatom flora of 84 samples collected from Byers Peninsula, Livingston Island and Lagoons Mesa area, James Ross Island has been studied using light microscopy.Key results -A total of 130 taxa, belonging to 39 genera has been recorded.Detrended Correspondence Analysis using the entire sample set clearly separates the James Ross Island (JRI) communities from the Livingston Island (LI) communities showing mostly the dominance of taxa preferring terrestrial (JRI) instead of more moist and aquatic conditions (LI).A Principal Component Analysis of only the Livingston Island samples formed three groups clearly separated by their diatom species composition.Although some taxa seem to occur in high abundances in several assemblages, a number of taxa showed a distinct preference for a particular assemblage.Biotic stress caused by marine birds and mammals, expressed in higher nutrient and salinity levels, seems to play a key role in determining the species composition.Conclusions -Our results showed clearly the presence of a well developed moss-inhabiting diatom flora on both studied localities.The composition of the studied communities is determined by the type of habitat, moisture and biotic influences (salinity and nutrients).
Pl. Ecol. Evol. 147 (1), 2014 tats.(Ochyra et al. 2008).Habitat seems to play an important role in shaping the diversity of the moss communities that can be dominated by either one or rarely several species.Recently Ochyra et al. (2008) reported the presence of 111 moss species and two varieties belonging to 55 genera in the entire Antarctic region (excluding the sub-Antarctic islands).The most diverse moss flora has been reported from the South Shetland Islands, where 87 species and one variety are present (Li et al. 2009).
Although their occurrence is strongly influenced by their local environment, moss vegetations can sometimes cover up to several hectares in the Maritime Antarctic region providing unique microhabitats for a wide range of microbial epiphytes, such as cyanobacteria, chlorophytes and diatoms.Within Antarctica the more northerly-situated bryophyte communities in the maritime zone, such as the South Orkney Islands, contain a richer algal flora (Broady 1986) than the rest of the region.Habitats with relatively stable conditions tend to have a low moss species diversity compared to more complex habitats, where species diversity significantly increases (Ochyra et al. 2008).Species diversity decreases wherever conditions become colder and drier, for instance in the coastal zones of Continental Antarctica, whereas only a few species are able to survive the extreme conditions of continental inland sites (Jones 1996, Spaulding et al. 2010).
Temperature and extreme aridity are the most important features affecting the suitability of a microbial habitat.Since diatoms tolerate a wide range of environmental conditions, making them suitable bio-indicators (Rimet 2012), they represent one of the most common algal groups in terms of both species richness and number of individuals in the Antarctic region (Jones 1996, Van de Vijver & Beyens 1999, Sabbe et al. 2003).They are present in almost all aquatic and terrestrial habitats, either epiphytically on aquatic and terrestrial moss communities or as epilithon, epipsammon and epipelon biofilms in both lentic and lotic water-bodies (Round et al. 1990).Diatoms are also able to survive in non-submerged or even dry habitats such as terrestrial mosses (Van de Vijver & Beyens 1998).All of these moss-inhabiting diatom communities are of particular interest as little is known about their species composition and ecological and habitat preferences (Van de Vijver et al. 2004, Bertrand et al. 2004).
Over the past decades, there was a growing interest in the use of the Antarctic diatom flora to solve questions about biogeography, palaeoecology and processes related to environmental changes.Despite this increase on diatom research, only a few papers reporting the Antarctic moss-inhabiting diatom flora have been published.Most publications deal with moss-epiphytic diatom communities from the sub-Antarctic region (a.o.Hickman & Vitt 1974, Van de Vijver & Beyens 1998, 1999, Van de Vijver et al. 2001, 2004, 2008and Gremmen et al. 2007).In Van de Vijver & Beyens (1997a), one moss sample from King George Island (South Shetland Islands) was analysed together with 11 aquatic and one soil sample.Toro et al. (2007) reported on moss communities from Livingston Island, but they only discussed them in relation to the invertebrates living near them and did not mention any diatom communities associated with these mosses.Van de Vijver et al. (2011a) described Luticola adelae Van de Vijver & Zidarova from a moss sample taken near White Lake on James Ross Island but apart from the formal description, no further analyses were carried out on the sample.Actually, only one recent paper discusses the Maritime Antarctic moss-inhabiting diatom flora: Vinocur & Maidana (2010) provided the first analysis of the spatial and temporal variations in the diatoms associated with mosses on the South Shetland Islands.Unfortunately, their species list is apparently composed of a large number of cosmopolitan taxa that so far were never found in the Maritime Antarctic region but quite common on more temperate localities reducing the value of the entire analysis.
Recently, a thorough taxonomical and ecological revision of the Livingston and James Ross Island diatom flora started which not only resulted in the description of a large number of new taxa (Kopalová et al. 2011, 2012, Van de Vijver et al. 2010a, 2010b, 2013, Van de Vijver & Zidarova 2011, Zidarova et al. 2009, 2012) but also led to a better ecological characterisation of the aquatic diatom assemblages present on both islands (Kopalová & Van de Vijver 2013, Kopalová et al. 2013).
The present paper completes the ecological analysis of the Livingston and James Ross Island diatom assemblages discussing the terrestrial diatom communities associated with different moss species on the two islands.The main objectives of this study included a floristic analysis of the moss-inhabiting diatom flora of these two islands, a discussion of their biogeographical position within the Maritime Antarctic region and possible similarities and differences between them and with the other communities on the islands in relation to several habitat characteristics.

Field sampling
During the austral summer of 2009-2010, (Limnopolar Project POL 2006-06635), a total of 68 water-saturated and dry moss samples for diatom analysis were collected from Byers Peninsula (Livingston Island, South Shetland Islands).An additional set of 16 water-saturated and dry moss samples from the Lagoons Mesa from Ulu Peninsula (James Ross Island) was collected during the summer expedition LAGOS 2012 (Picto project 2010-0096).All moss samples were fixed with alcohol and stored in plastic vials.Sampling locations together with GPS co-ordinates are presented in table 1.
Due to the restricted logistic possibilities of working in these extreme conditions, only a limited number of environmental parameters were measured and/or determined.For all samples we noted: elevation (m a.s.l.), biotic influence (0 = none, 1 = heavy manuring and trampling by marine mammals or birds), habitat type (1 = lake, 2 = pond, 3 = stream, 4 = terrestrial) and dominant moss species present.Table 1 lists all samples with their characteristics.Moss species in the samples were identified using Ochyra et al. (2008).Sixteen different moss species, belonging to thirteen genera, were found in the entire sample set.On James Ross Island, only six species were identified, compared to Livingston Island where twelve different species were found.Only two of all moss species were in common between both islands.In order to determine the differences in diatom composition Habitat type: 1 = lake; 2 = pond; 3 = stream; 4 = terrestrial.ND = not determined.For more details, see text.due to moisture content, the F-value, referring to the F-classification of Jung (1936) was selected as a representative for moisture and used for each sample of Livingston Island.The F-value was not determined for the James Ross samples.It is a humidity scale based on water content as follows: FI = submerged mosses, FII = free floating mosses, FIII = very wet (water drips from the samples without pressure), FIV = wet (water drips with a slight pressure), FV = quasi-wet (water drips after moderate pressure), FVI = moist (little water produced after high pressure), FVII = quasi-dry (only a few drops of water can be squeezed out), FVIII = dry (contains no water).

Study area -Livingston Island
Livingston Island is the second largest island of the South Shetland Islands, with a total area of about 950 km 2 .Based on its ecological and climatological characteristics, this archipelago belongs to the Maritime Antarctic region (Chown & Convey 2007).The island (62°36'S 60°30'W), located 150 km north-west of the Antarctic Peninsula (fig.1), is almost entirely covered by permanent glaciers and icecaps leaving only 10% of the island ice-free.Byers Peninsula, the largest ice-free area (almost 61 km 2 ) forming the western tip of the island, is the most important biodiversity area on Livingston Island and is currently included within the list of the Antarctic Specially Protected Areas (ASPA No 126).More information on the climate, geology, hydrology and geomorphology of this area can be found in Chipev & Veltchev (1996) and Toro et al. (2007).Vegetation cover on Livingston Island, as typical for the Antarctic region, is scarce and has a mosaic structure (Toro et al. 2007).It is mainly formed by cryptogams, with lichens and mosses as dominant life forms with only the two above mentioned vascular plants forming small cushions (Toro et al. 2007).Several of the lakes have well developed monospecific stands of the benthic moss Drepano cladus longifolius (Wilson ex Mitt.)Broth.ex Paris, which might dominate overall lake productivity because of its large standing stocks (Li et al. 2009).

Study area -James Ross Island
James Ross Island is a large island with a total area of ~2600 km 2 , in the northern-western part of the Weddell Sea, close to the northern tip of the Antarctic Peninsula.It belongs to the transition zone between the Maritime Antarctic and Continental Antarctic region (Øvstedal & Lewis Smith 2001).More than 80% of the island is covered by an ice cap leaving only the northern part of the island, Ulu Peninsula, ice free (100 km 2 ).Olivero et al. (2008), Smellie et al. (2008) and Svojtka et al. (2009) discussed the geological history of the island.Ulu Peninsula is characterized by the presence of a large number of streams, seepages and lakes of glacial origin (Nedbalová et al. 2013).The human presence is limited to the Czech scientific base (Johann Gregor Mendel Station) that was constructed on Ulu Peninsula in 2006.The climate of James Ross Island is determined by cold, arid barrier winds from the south and by the location in the precipitation shadow of the Antarctic Peninsula (Engel et al. 2012).In comparison to the South Shetlands Islands, the climate is more arid with low precipitation, estimated to be less than 300 mm/yr.Owing to the dry air and often high wind speeds, evaporation rate is high.Further details on the climatic conditions can be found in Láska et al. (2010Láska et al. ( , 2011aLáska et al. ( , 2011b) ) and in Engel et al. (2012).Inland vegetation, lacking any vascular plants, is restricted to bryophytes and lichens.Their distribution is usually limited due to the deficiency of liquid water (Robinson et al. 2003).Although moss communities are not very frequent on James Ross Island, there are several patches of live or moribund moss (Láska et al. 2011b).On the other hand, the microflora, mostly composed of cyanobacteria, green algae and diatoms, is well developed in freshwater ecosystems such as seepages, lakes and streams (Komárek & Elster 2008, Kopalová et al. 2012, 2013).

Sample treatment and counting
Diatom samples were prepared using the method described in Van der Werff (1955).Subsamples were cleaned by adding 37% H 2 O 2 and heating to 80°C for about 1 h.Oxidation of organic material was completed by addition of KMnO 4 .Following digestion and centrifugation (10 min at 3700 x g), the resulting cleaned material was diluted with distilled water to avoid excessive concentrations of diatom valves on the slides, dried on microscope cover slips, and mounted in Naphrax®.Samples and slides are stored at the National Botanic Garden of Belgium (Meise, Belgium).In each sample, 400 diatom valves were identified and enumerated on random transects at x1000 magnification under oil immersion using an Olympus® BX51 microscope equipped with Differential Interference Contrast (Nomarski) optics.Identifications of Antarctic species are based on Van de Vijver et al. (2002a, b, 2004, 2010a, 2010b, 2011a, 2011b), Sabbe et al. (2003), Ohtsuka et al. (2006), Esposito et al. (2008), Van de Vijver & Mataloni (2008), Kopalová et al. (2009Kopalová et al. ( , 2011Kopalová et al. ( , 2012)), Zidarova et al. (2009Zidarova et al. ( , 2010Zidarova et al. ( , 2012)), Van de Vijver & Zidarova (2011) and references therein.For several species, identification up to species level was not possible due to their unclear taxonomic situation.All valves belonging to the genus Gomphonema were grouped as Gomphonema spp.The different taxa with affinity to Nitzschia perminuta (Grunow) Perag.were combined as N. perminuta-complex.Further morphological and taxonomic research (ongoing) will be necessary to establish their correct identity.

Data analysis
For a pairwise comparison of the moss-inhabiting diatom flora of Livingston Island and James Ross Island with similar bryophytic communities in the sub-Antarctic Region, the community coefficient of Sørensen (1948)  The geographic distribution of the taxa was based on literature data provided with illustrations or descriptions (table 2).When the identity of a taxon could not be determined this was shown using 'cf.' or 'spp.' and its distribution was listed usually as unknown (U).For Antarctic species, the geographic distribution was further subdivided in 'MA' when the species occurred only in the Maritime Antarctic region.Taxa present in the entire Antarctic region are listed as 'A'.Cosmopolitan taxa present as 'C'.
To determine the extent to which our sampling effort represented the total diatom flora of the two islands, the incidence-based species richness estimator (ICE, Chao et al. 2000) and the mean Chao2 richness estimator (Chao  Ordination was used to elucidate the principal patterns in species composition in the moss samples of Livingston Island.Squareroot-transformed abundance data with downweighting of rare taxa were used in the ordinations.All ordination analyses were performed using the computer program CANOCO version 4.5 (ter Braak & Šmilauer 1998).The statistical and numerical techniques used in this study are described in full detail in Jongman et al. (1995).

Species composition and diversity
The microscopic analysis of 84 samples revealed a total of 130 diatom taxa (including species, varieties and forms) belonging to 39 genera.Six samples (V3M3, M1 and M21 from James Ross Island and BYM-9, BYM-10 and BYM-40 from Livingston Island) contained (almost) no diatoms, even after counting an entire slide.Subsequently, these samples have been removed from further analysis.On Livingston Island (68 samples), 123 diatom taxa belonging to 39 genera were found, whereas from James Ross Island (16 samples), only 57 taxa from 23 genera were identified.Table 2 provides an alphabetical list of all observed species together with their biogeographical distribution.
Almost 53% of all observed species have a restricted Antarctic distribution with a majority of these (79%) confined to the Maritime Antarctic region whereas only 43 taxa (32%) have a typical cosmopolitan distribution, such as Fragilaria capucina Desm., Navicula gregaria Donkin and Nitzschia gracilis Hantzsch.
The similarity analysis indicates that the moss diatom flora of Livingston Island and James Ross Island shows a clear difference to the moss-inhabiting diatom flora from South Georgia, and the Prince Edward Islands with Sørensen index values for the complete dataset of both islands together ranging from 0.17-0.19.(table 3).We observed a similarity of only 0.57 between Livingston and James Ross Island.James Ross Island always presented a somewhat lower similarity with the other islands than Livingston Island (0.11-0.14 vs. 0.16-0.19).
Species richness per sample ranged from 9 to 46 for Livingston Island and 7 to 24 for James Ross Island.The distribution of species number per sample clearly differs between both islands.The average number (and standard deviation) of taxa per sample was 25 ± 8 for Livingston Island and 16 ± 6 for James Ross Island (fig.2).The highest species richness was recorded in several Livingston Island moss samples: BYM-11 (46 taxa), BYM-53 (45 taxa) and BYM-27 (40 taxa) whereas on James Ross Island, the maximum number of counted species was only 24 (sample M22) followed by samples M23 and V3M1 with 23 counted species.The species accumulation curve for Livingston Island (fig.3) indicates that this sample set contains a large part of the total diatom flora, although it is clear that theoretically not all species have been found.As for James Ross Island only 13 samples were analysed, it is clear that a considerable number of samples still will be needed to obtain a representative dataset for this island.Using species richness estimators, it is possible to evaluate how well the sampling effort reflected the true diatom richness.The expected total number of taxa in all samples is 138 (Chao2) or 142 (ICE) for Livingston, suggesting that our counting scored between 87 and 89% of the (theoretical) total number of taxa present in the samples overall.On the contrary, on James Ross Island, only a counting score of 62% (ICE) -69% (Chao2) of the (theoretical) total number of taxa was calculated.Based on these differences in species richness, the limited amount of samples from James Ross Island and the species accumulation curve, both datasets will also be treated separately in the following diversity and community analyses.
On Livingston Island, the 51 least abundant species (= 38% of all observed species) together made up only 1% of the total number of valves counted whereas the 7 most dominant species accounted for 50% of all counted valves.As can be seen in fig.2, a large number of species is restricted to only a few samples and only a few species occur in 50% or more of all samples.The genera Pinnularia (sixteen taxa), Luticola (thirteen taxa) and Psammothidium (nine taxa) were the most species rich genera.Other important genera include Diadesmis, Muelleria, Navicula and Nitzschia (seven taxa).The dominant species is Nitzschia perminuta with more than 15% of all counted valves followed by Fragilaria capucina (9.0%), Psammothidium incognitum (Krasske) Van de Vijver (6.6%) and Gomphonema spp.(6.3%).It should be noted however that both N. perminuta and Gomphonema spp.most likely represent complexes of several taxa that need to be split into several independent, most probably new taxa.
A considerable number of taxa appeared to be new for science (e.A very small proportion (< 0.1%) of all counted valves belonged to marine species (indicated as '*' in table 2) probably blown in by seaspray or wind or transported on the fur of marine mammals such as elephant seals (Mirounga leonina (Linnaeus, 1758) or the feathers of birds such as gentoo penguins (Pygoscelis papua Forster, 1781) or southern giant petrels (Macronectes giganteus Gmelin, 1789).

Community analysis
An initial detrended correspondence analysis (DCA) using the entire dataset was carried out to estimate gradient length (fig.4).The results showed that two samples were clear outliers.Sample BYM-02 contained exclusively a very large population of Eunotia paludosa Grunow whereas sample BYM-59 was entirely dominated by Psammothidium germainii (Manguin) Sabbe.A second DCA with the two outliers omitted showed gradient lengths for the first four axes of 3.322, 2.161, 2.495 and 2.139, suggesting that methods based on unimodal models (Correspondence Analysis) would be appropriate for the ordination of the entire sample set (ter Braak & Prentice 1988).Figure 4 shows clearly that the samples from James Ross Island (JRI) (▲) are entirely separated from the Livingston Island (LI) samples (•).All JRI samples are dominated by Pinnularia borealis, Hantzschia amphioxys and H. abundans, typical terrestrial species that only play a minor role on Livingston Island.
Since the LI sample set was almost five times as large and probably more diverse than the JRI sample set, a new ordination analysis was run, only using the LI samples.The initial DCA showed a maximum gradient length of only 2.0 making linear models (principal component analysis, PCA) more appropriate (ter Braak & Prentice 1988).The PCA analysis divides the LI samples intro three groups (fig.5).The distinction between these groups is clearly reflected in the species composition.The first two PCA axes (eigenvalues λ 1 = 0.168, λ 2 = 0.104) were highly significant (p = 0.001) and explained 27.3% of the variation in the diatom composition with an additional 15.7% explained on the next two axes.Table 4 shows the principal characteristics of the different groups including the dominant species.The samples on the right side of the diagram (×) form assemblage A. They can be subdivided into two sample groups: assemblage A1 groups coastal localities where the influence of marine mammals and/or birds is very high whereas assemblage A2 contains sites close to the Rotch Dome ice cap, formed after recent glacier retreats.On the left side, two assemblages can be formed within the samples.In the upper half of the diagram, assemblage B (▼) comprises mainly samples from more terrestrial sites at higher elevations, whereas the lower half (assemblage C, ■) is characterized by samples from aquatic habitats (pools and lakes) located at lower altitudes.There are clear differences in diatom species composition between the different assemblages (table 4).Only species with a cumulative fit of > 25% in the PCA diagram are shown (species acronyms are added to table 2).Although some taxa seem to occur in high abundances in several assemblages (such as Nitzschia perminuta, Chamaepinnularia krookiiformis (Krammer) Lange-Bert.& Krammer or Fra gilaria capucina sensu lato) it is clear that a number of taxa showed a distinct preference for a particular assemblage.Assemblage A1 is characterized by high frequencies of Cha maepinnularia krookiiformis, several Psammothidium species (P.

Species composition and general biogeography
This study focused on the moss-inhabiting diatoms of two islands located on both sides of the Antarctic Peninsula and therefore undergoing different climatological and ecological influences.Livingston Island is a typical example of the Maritime Antarctic region with relatively high precipitation rates reflected in a higher number of aquatic habitats with more luxuriant wet bryophyte vegetation compared to James Ross Island that has a much drier climate.On the latter island, the extent of aquatic and/or wet terrestrial moss vegetation is rather limited.This is clearly reflected in the observed diatom composition.Whereas the sampled mossinhabiting communities on James Ross Island are dominated by only typical terrestrial taxa such as Pinnularia borealis, Hantzschia amphioxys and Diadesmis arcuata, known for their preference of drier environments (Petersen 1935, Van de Vijver & Beyens 1997a), dry terrestrial moss vegetations were hardly present and therefore not sampled on Livingston Island resulting in a lower proportion of these terrestrial diatom species in the samples.Based on the differences in sampling effort and habitat types, comparing the species richness  between both islands (57 on JRI vs. 123 on LI) is rather difficult.
It is generally accepted that a decreasing diversity trend when going southwards exists in the Antarctic region (Jones 1996, Van de Vijver & Beyens 1999).This is confirmed by the similarity analysis.The moss-inhabiting diatom communities in the sub-Antarctic region show almost double the species number recorded from Livingston Island (Van de Vijver & Beyens 1997b, Van de Vijver et al. 2004, 2008).
The main reason for this decreasing diatom diversity is probably the lower environmental diversity.On the sub-Antarctic islands, the moss vegetation dominates all parts of these islands (Van de Vijver et al. 2002a), forming thick layers in the valleys, on slopes and even fell fields, whereas in the Maritime Antarctic region, mosses are restricted only to the wet areas around lakes, seepage areas and streams.Given the particular dry nature of the James Ross Island samples, it is not entirely sure whether the results in the present study provide a trustful overview of the moss-inhabiting diatom flora of this island.However, the absence of their typical wet moss habitat on James Ross Island is probably one of the main reasons for the lack of epiphytic diatoms on this island.On the nearby Antarctic Continent, the bryophyte flora is restricted to 30 moss species (Lewis Smith 1984), most of them being poikilohydric indicating that they are only wet during periods of snow melts in the summer months (Robinson et al. 2000).Based on the low annual precipitation (Láska et al. 2011a), mosses on James Ross Island apparently undergo a similar desiccation, providing a quite unstable environment for the epiphytic diatoms living on these mosses.Only typical dry-terrestrial diatom taxa are able to overcome these periods of temporal dryness and hence the moss diatom flora on James Ross Island has a lower species richness compared to Livingston Island were mosses remain continuously wet near the numerous lakes and pools on Byers Peninsula.
In the past, Antarctic diatom taxa were force-fitted into European and North-America relatives, which led to the incorrect conclusion that the Antarctic diatom flora is mostly cosmopolitan (Toro et al. 2007, Vinocur & Maidana 2010).However, the recent revisions of the non-marine diatom flora from the entire Antarctic region based on a narrower species concept and a more fine-grained taxonomy (e.g.Van de Vijver et al. 2010a, 2011b, Zidarova et al. 2012) resulted in the description of a large number of typical Antarctic taxa.Many of them were observed in the moss-inhabiting flora.This current revised taxonomy contradicts the Ubiquity Theory developed by Finlay & Clarke (1999) which stated that microorganisms, due to their large population size and small body size can be easily dispersed throughout the world reducing their overall diversity and possible local endemism.Antarctic diatoms clearly show a rather restricted distribution with many endemic taxa.This is confirmed by the results of the similarity analysis.Less than 25% of the taxa found on Livingston Island and James Ross Island are shared with the moss-inhabiting diatoms from the sub-Antarctic islands (on which a similar taxonomic revision was performed), even with localities situated in the southern Atlantic Ocean (South Georgia).Most of the species in common between these islands are mainly cosmopolitan taxa such as Navicula gregaria and Pinnularia borealis.It is however unclear whether these taxa are truly cosmopolitan since recent studies analysing the genetic similarities and differences between several Antarctic populations of P. borealis and Hantzschia amphioxys indicate a high degree of cryptic diversity (Souffreau et al. 2013).Together with the rather high number of still unidentified taxa (9-20%) for which further morphological and taxonomic analyses will be necessary to clarify their biogeographical preferences, the proportion of typical Maritime Antarctic taxa will likely increase, confirming the highly specific nature of the Antarctic diatom flora.

Moss-inhabiting diatom communities
The composition of freshwater diatom communities in the Maritime Antarctic region is determined by the amount of nutrients and the conductivity of their habitat (Jones 1993, Kopalová & Van de Vijver 2013, Kopalová et al. 2013) while on the Antarctic Continent lake communities seem to be more influenced by salinity (e.g.Verleyen et al. 2003, Gibson et al. 2006).The moisture availability of the moss habitats presents an additional stress factor for the diatom communities living on these habitats.Moss-inhabiting communities on the sub-Antarctic islands are controlled mainly by moisture of the moss habitat (Van de Vijver & Beyens 1997b, 1999, Van de Vijver et al. 2004, 2008, Vinocur & Maidana 2010).A similar result was found for the moss communities on James Ross and Livingston Island.All moss samples that were collected on James Ross Island were entirely terrestrial and, although not measured, had a rather low moisture content (Kopalová, pers. obs.).This had a clear influence on the moss-inhabiting community that was entirely composed of typical aerophytic species such as Pinnularia borealis and several Hantzschia taxa.On the sub-Antarctic island of South Georgia, these taxa were only found in the driest mosses (Van de Vijver & Beyens 1997b) and also on other, more distantly located islands, similar communities were observed (see for instance Van de Vijver et al. 2004, 2008).On Livingston on the other hand, dry mosses were almost not sampled which resulted in the observation of totally different communities.The first two axes of the PCA analysis of the Livingston Island samples most likely represent two determining factors: axis one seems to be a biotic axis related to nutrients and salinity whereas axis two represents a moisture axis.
Biotic stress resulting in higher nutrient and salinity input by marine birds and mammals, seems to play a first important role in determining the species composition in the Livingston moss diatom communities separating assemblage A2 from the other assemblages.Marine mammals are frequently found on the shores of this island and show a marked influence on the areas where they often stay for several weeks during their moulting period (Cruwys & Davis 1995).With their excrements, these animals considerably alter the diatom habitat by increasing both the salinity and the nutrient concentrations.Although these parameters were not measured for the moss samples, aquatic samples collected near these moss vegetations show the same trends.Kopalová & Van de Vijver (2013) discuss the environmental factors that determined the diatom composition in the waterbodies of Byers Peninsula and concluded that nutrients and salinity were the controlling factors.Assemblage B clearly represents wet terrestrial moss-inhabiting diatom communities.Psammothidium incognitum, originally described from wet mosses on southern Patagonia (Krasske 1939) was found to be one of the dominant terrestrial moss-inhabiting species on South Georgia (Van de Vijver & Beyens 1997b) and also on the sub-Antarctic islands in the southern Indian Ocean, the species was mostly found in wet terrestrial moss vegetations (Van de Vijver et al. 2002a) whereas it is less frequent in aquatic moss vegetations.The assemblage shows also a clear similarity with a South Georgian moss assemblage although some typical sub-Antarctic moss diatoms such as Psammothidium confusum (Krasske) Van de Vijver & Beyens are completely missing in the Maritime Antarctic region (Van de Vijver & Beyens 1997b).The aquatic moss assemblage on the other hand shows a large similarity with the epilithic and epipelic aquatic diatom community that was found in lakes on Byers Peninsula.Kopalová & Van de Vijver (2013) reported an almost identical species composition in the lakes.This might indicate that microhabitats in these permanent lakes apparently only play a minor role in the shaping of the diatom composition in these lakes.Most Nitzschia and Navicula species in the Antarctic region are exclusively aquatic species, hardly found outside permanent waterbodies (Van de Vijver et al. 2011b, Kopalová & Van de Vijver 2013).The results of this study confirm this observation since almost all species of these two genera were found only in the aquatic moss assemblage and not in the terrestrial moss group.Comparison with older data is hardly possible due to the taxonomic revision that started a couple of years ago (Van de Vijver et al. 2011b).

CONCLUSION
In general, a better knowledge of moss-inhabiting diatoms on the Maritime Antarctic islands is important not only from a taxonomical point of view, but also for further ecological and palaeoecological research as some of these moss vegetations present a typical habitat in which diatoms are able to survive, even during more stressful periods.The moss communities on the studied islands are influenced in the first place by the input of nutrients and salinity, most likely due to animal impact and secondly by the moisture content of the habitat.

Figure 1 -
Figure 1 -Geographic location of the studied islands: A, overview of the southern hemisphere with the location of several islands and archipelagos mentioned in the text; B, detailed map of Antarctic Peninsula region showing the position of James Ross Island and Livingston Island; C, the South Shetland Islands.Livingston Island is indicated; D, James Ross Island.The box indicates the locality of the studied area: Lagoons Mesa.
was used.This index has the following formula: 2c/(a+b+2c) where 'a' and 'b' are the numbers of species exclusively observed in each of the two sites and 'c' is the number of species shared by these sites.The comparison is based on the revised species lists of South Georgia (Van de Vijver & Beyens 1997b), Heard Island (Van de Vijver et al. 2004) and the Prince Edward Islands (Van de Vijver et al. 2008).For the Antarctic Continent, unfortunately, no recent data on moss-inhabiting diatoms are available.

Figure 2 -
Figure 2 -Distribution of samples for the two sample sets based on species richness.JRI: James Ross Island (black), LIV: Livingston Island (grey).

Figure 3 -
Figure 3 -Expected species accumulation curve (sample based rarefaction curves for the total sample set) for the Livingston Island moss samples.Each time, the 95% confidence interval is given.
g. Planothidium rostrolanceolatum Van de Vijver, Kopalová & Zidarova in Van de Vijver et al.).Several of them have been recently published (Van de Vijver et al. 2013) whereas others (such as Psammothidium sp., Halamphora sp. or Diadesmis sp1 and sp2) await a formal description.

Figure 4 -
Figure 4 -Detrended Correspondence Analysis (DCA) of the entire sample set.A clear division can be seen between the James Ross Island samples (triangles) and the Livingston Island samples (dots).

Figure 5 -
Figure 5 -Principal Components Analysis (PCA) of the Livingston Island sample set.Division into the three groups (group A: ×; group B: ▼; group C: ■) and subdivision of group A is indicated.The first two axes are shown.The acronyms of the species names are given in table 2.

Table 2 -List of all observed species with their acronyms in the investigated moss samples from James Ross Island and Livingston Island.
Distribution: C = Cosmopolitan; MA = Maritime Antarctic Region; A = Antarctic Region; U = Unknown.Marine species are marked with an *.