Journal of Biogeography (J. Biogeogr.) (2010) 37, 1648–1656 ORIGINAL ARTICLE Bryozoans of the Weddell Sea continental shelf, slope and abyss: did marine life colonize the Antarctic shelf from deep water, outlying islands or in situ refugia following glaciations? David K. A. Barnes1* and Piotr Kuklinski2,3 1 British Antarctic Survey, Natural Environment Research Council, High Cross, Madingley Road, Cambridge, CB3 0ET, UK, 2 Institute of Oceanology, Polish Academy of Sciences, ul. Powstancow Warszawy 55, Sopot 81-712, Poland, 3Department of Zoology, Natural History Museum, Cromwell Road, London SW7 5BD, UK ABSTRACT Aim At the height of glaciations such as the Last Glacial Maximum (LGM), benthic life on polar continental shelves was bulldozed off nearly all of the Antarctic shelf by grounded ice sheets. The origins of the current shelf benthos have become a subject of considerable debate. There are several possible sources for the current Antarctic shelf fauna, the first of which is the continental slope and deep sea of the Southern Ocean. The high levels of reported eurybathy for many Antarctic species are taken as evidence supporting this. A second possible source for colonists is the southern margins of other continents. Finally, shelves could have been recolonized from refugia on the continental shelves or slopes around Antarctica. The current study investigates whether the patchily rich and abundant biota that now occurs on the Antarctic continental shelf recolonized from refugia in situ or elsewhere. Location Weddell Sea, Antarctica. Methods We examined bryozoan samples of the BENDEX, ANDEEP III and SYSTCO expeditions, as well as the literature. Using similarity matrices (Sørensen coefficient), we assessed similarities of benthos sampled from around Antarctica. By assessing numbers of species shared between differing depths and adjacent shelf areas, we evaluated the origins of cheilostome bryozoan communities. Results Bryozoans decreased from 28, 6.5 and 0.3 colonies per trawl, and 0.16, 0.046 and 0.0026 colonies per cm2 of hard surface from shelf to slope to abyssal depths. We found little and no support for recolonization of the Weddell Sea shelf by bryozoans from the adjacent slope and abyss, in the scenario of LGM faunal wipe-out. The Weddell Sea shelf bryozoan fauna was considerably more similar to those on other Antarctic shelves than to that of the adjacent (Weddell Sea) continental slope. The known bryozoan fauna of the Weddell Sea shelf is not a subset of the Weddell Sea slope or abyssal faunas. *Correspondence: David K.A. Barnes, British Antarctic Survey, Natural Environment Research Council, High Cross, Madingley Road, Cambridge CB3 0ET, UK. E-mail: dkab@bas.ac.uk 1648 Main conclusions We consider that the composition of the current Weddell Sea bryozoan fauna is most easily explained by in situ survival. Thus we consider that at least some of the Weddell Sea fauna persisted throughout the LGM, although not necessarily at the same locations throughout, to recolonize the large area currently occupied. Keywords ANDEEP, Antarctica, benthos, ice sheet, Last Glacial Maximum, Pleistocene refugia, Southern Ocean. www.blackwellpublishing.com/jbi doi:10.1111/j.1365-2699.2010.02320.x ª 2010 Blackwell Publishing Ltd Recolonization of Antarctic shelves by marine life INTRODUCTION Biodiversity in the polar regions is dominated by the speciesrich, continental shelf sea-bed (e.g. in Antarctica, see Clarke & Johnston, 2003; Barnes et al., 2009). These authors, among many, show that one of the taxa particularly well represented on the Antarctic shelf is the cheilostome bryozoans, a group of clonal, colonial, sessile suspension feeders. However, few bryozoan species have been reported from below the shelfbreak (typically between 500 and 1000 m) on the continental slope (but see López-Fé, 2005; Barnes, 2008). As with the apparent comparative rarity (of known species) of other taxa below shelf depths, this is in part because the shelf around Antarctica is so much better sampled than the continental slope (c. 1000–3000 m) and abyss (c. 3000–5000 m). Nevertheless, the ANDEEP expeditions alone have made nearly 100 collections from such depths, and have shown a variety of taxa to be richly represented (Brandt et al., 2007). They are not rich or common, but bryozoans have been found at abyssal depths in all the world’s oceans except the Southern Ocean (Hayward, 1981). Hayward (1981) and other sources have listed the location of a few species as ‘Southern Ocean’, but the positions of all the samples in question are all north of the Polar Front, which is generally considered as the boundary of the Southern Ocean (see Moore et al., 1999 for mean and extreme positions of the Polar Front). Thus, to date, no bryozoans have ever been reported from abyssal depths of the Southern Ocean sensu stricto. Whether, and which, bryozoans (or other taxa) occur in the Southern Ocean deep sea, such as at abyssal depths, may be important for understanding Antarctica’s recent biological and physical history. Variation in Earth’s orbit around the Sun generates cyclicity in surface temperatures such that every 41,000 years (and most recently every 100,000 years) in the past few million years have caused glaciations (‘ice ages’). Current evidence suggests that, during each glaciation, ice sheets have extended to cover the continental shelves of the Polar Regions, and sonar pictures of the Antarctic continental shelf are covered with grounding scours from the Last Glacial Maximum (LGM) (Anderson et al., 2002). If the rich and abundant life that now occurs on polar continental shelves was completely bulldozed off, species would have to recolonize from somewhere else. Around Antarctica, there are a number of possibilities for where this fauna could recolonize from. 1. The continental slope and deep sea of the Southern Ocean. Thatje et al. (2005) considered that living on the slope would be difficult during glacial maxima due to frequent catastrophic cascades of rock and sediment generated from the grounded ice at the shelf-break. However, as few bryozoans have yet been reported from deeper than shelf depths, they at least can provide little support for recolonization from the deep. 2. The southern margins of other continents, sub-Antarctic islands and some Antarctic archipelagos not reached by the ice advance. Both adults and larvae of marine species can, and do, travel into the Southern Ocean from north of the Polar Front (Barnes et al., 2006). However, the high levels of endemism of Journal of Biogeography 37, 1648–1656 ª 2010 Blackwell Publishing Ltd most Antarctic taxa on the shelf (Arntz et al., 1997) can provide little support for these areas as sources for colonization. Furthermore, new genetic evidence suggests that ocean current directions transport adults and larvae from Antarctic shelf areas to the outlying islands, rather than vice versa (Linse et al., 2007; Mahon et al., 2008). 3. Small refugia on the continental shelf or slope around Antarctica. Various authors have suggested this, and even mechanisms for the existence of refugia (e.g. Dayton & Oliver, 1977; Brandt, 1991; Thatje et al., 2005). However, the only evidence to date is that pockets of land may have remained icefree throughout successive glaciations (Convey & Stevens, 2007). Bryozoans have been well studied at mid- to high southern latitudes, have high levels of Antarctic endemism, and appear to have a strong distribution pattern with depth (e.g. Schopf, 1969). Thus this taxon should be a powerful model group to investigate the possibility of recolonization of the Antarctic shelf from deep water. In the current study, we examine continental shelf, slope and abyssal samples from the Weddell Sea from the BENDEX, ANDEEP III and SYSTCO expeditions, as well as distributions of bryozoans reported in the literature. Our null hypothesis (H0) is that endemic (c. 60%) Antarctic bryozoans on the shelf are a subset of those that occur deeper in the abyss (and that the slope is a transition zone) because they have recolonized the shelf from the deep sea. Our first alternative hypothesis (H1) is that these bryozoans are a subset of those that occur on outlying islands within the Polar Front, because they have recolonized the Antarctic shelf from such areas (or they have recolonized from both deeper water and outlying islands). Our final hypothesis (H2) is that bryozoans endemic to the Antarctic shelf are distinct because they have survived glaciations in situ in shelf refugia. MATERIALS AND METHODS We examined continental shelf, slope and abyssal samples from the eastern Weddell Sea collected by the BENDEX, ANDEEP III and SYSTCO expeditions (details of sample location, depth and collection apparatus are shown in Table 1). The samples we compare originate from three different towed apparatus (Agassiz trawl, Rauschert dredge and Epibenthos sledge), and tow lengths differ even within apparatus types, representing sources of error for work on macro- and mega-benthos in the deep sea and across depths. However, they all collect boulders of a similar size range, thus most bias is limited to the number of such substrata collected, which we have negated by also comparing colonies per unit area of rock. Less biased apparatus, such as cores or camera systems, is inappropriate for the collection of bryozoans, which occur mainly on hard surfaces and are too small for identification by camera. Bryozoans on rocks were preserved dry, whereas unattached bryozoans were fixed and preserved in 96% ethanol. Most of the (BENDEX) shelf samples contained bryozoans (typically on rocks), so we selected five samples at random and identified all the cheilostome bryozoan colonies within each of these 1649 D. K. A. Barnes and P. Kuklinski Table 1 Weddell Sea benthic sample details. Sample Latitude Longitude Depth Apparatus PS65/121-1 PS65/069-1 PS65/279-0 PS65/324-1 PS65/326-1 PS67/078-11 PS67/074-7 PS71/17-10 PS67/078-9 PS67/074-6 PS67/102-11 PS67/080-9 PS67/110-8 70°50.08¢ S 70°25.87¢ S 71°7.43¢ S 72°54.55¢ S 72°51.70¢ S 71°9.39¢ S 71°18.48¢ S 70°04.78¢ S 71°09.39¢ S 71°18.35¢ S 65°35.40¢ S 70°39.07¢ S 65°0.52¢ S 010°35.54¢ W 008°37.43¢ W 011°29.83¢ W 019°47.30¢ W 019°39.22¢ W 13°59.33¢ W 13°58.55¢ W 03°19.66¢ W 13°59.30¢ W 13°57.71¢ W 36°29.00¢ W 14°43.36¢ W 43°2.09¢ W 268.0 413.6 119.2 647.2 605.2 2157.0 1055.0 2189.7 2156.0 1030.0 4794.0 3103.0 4698.0 Agassiz trawl Rauschert dredge Agassiz trawl Rauschert dredge Rauschert dredge Agassiz trawl Agassiz trawl Agassiz trawl Epibenthos sledge Epibenthos sledge Agassiz trawl Epibenthos sledge Epibenthos sledge Samples from continental shelf (unshaded); slope (light grey); or abyss (dark grey). Trawling distances were 537 (± 125) m on the shelf; 1530 (± 201) m on the slope; 3228 (± 177) m on the abyssal plain. Bias in sampling across apparatus and trawl distance is discussed in the Materials and Methods. Slope samples cover about three times as much area as shelf samples (see trawl distances in Table 1), so patches of bryozoans were present over an order of magnitude less area at slope depths compared with on the shelf. Likewise, patches of bryozoans are present on nearly an order of magnitude less area in the deep sea than at slope depths, and nearly two orders of magnitude less than on the shelf. The apparatus used was not quantitative, but averages of 28, 6.5 and 0.3 bryozoan colonies per trawl on the shelf, slope and abyss, respectively, suggest bryozoans may be about four orders of magnitude less abundant in the deep sea than on the shelf. Virtually all (> 99%) the bryozoans found across depths occurred on hard substrata (such as boulders), so semi-quantitative abundance comparisons can be made in terms of colonies per area of boulder. Shelf boulders had an average of 0.16 bryozoan colonies per cm2, compared with 0.046 and 0.0026 at slope and abyssal depths, respectively. However, these values include boulder area only for sites that had bryozoans (or other encrusting fauna) covering them. Most slope and abyssal samples had no encrusting fauna on hard substrata and were not kept (due to logistical constraints, e.g. storage space on samples to morphospecies (and in most cases to named species). We investigated all continental slope and abyssal samples that contained cheilostome bryozoans. Where necessary, large rocks with colonies on were sawn into fragments to aid placement of colonies under light and scanning electron microscopes. Identifications were made using Hayward (1981, 1995, and references therein), López Gappa (1986), López de la Cuadra & Garcı́a Gómez (2000) and Gontar (2002, 2008). We evaluated the number of bryozoan colonies per unit area of rocks and per sample. The area of each rock was measured by draping a flexible but inelastic net (demarked in a permanent grid of square centimetres) over its entire surface area, and counting the number of centimetre grid squares (following Barnes, 2008). Following identification, we compared the number of species between depths and, using the literature, with what was previously recorded for the region. We constructed species-accumulation curves using both boulder and colony as a sample unit to compare patterns between depths and between different locations around the Weddell Sea region. We collated literature to show known bathymetric ranges for all species with deepest recorded depths below the shelf break. Similarity matrices were generated, based on binary (species presence/absence) data, using the Sørensen coefficient to construct dendrograms using EstimateS software (v. 8.2, Colwell, 2009). We calculated numbers of species shared between differing depths and adjacent shelf areas, and we attempted to evaluate whether or not cheilostome bryozoan distribution supported our various hypotheses. RESULTS All the shelf samples examined contained cheilostome bryozoans. Five of the 17 slope samples (29%) and three of the 22 abyssal samples (13.6%) also had at least one bryozoan colony. 1650 Figure 1 Number of (a) bryozoans colonies and (b) species with sample depth in the Southern Ocean. Values are from the current study (with trend as line of best fit) except for literature data (unfilled diamonds or circles) from Gontar & Zabala (2000), López de la Cuadra & Garcı́a Gómez (2000), Barnes (2008), and unpublished BIOPEARL II cruise data. Journal of Biogeography 37, 1648–1656 ª 2010 Blackwell Publishing Ltd Recolonization of Antarctic shelves by marine life Table 2 Bryozoan species occurrence in the Weddell Sea by depth of sample. Species Acanthophragma polaris Adelascopora jeqolqa* Aimulosia antarctica Amastigia gaussi* Amphiblestrum rossi Amphiblestrum henryi Andreella uncifera ,à Antarctichaetos bubeccata* Arachnopusia aviculifera Arachnopusia ferox*, Arachnopusia gigantea Aspidostoma coranatum Buffonellaria frigida* Bugulella klugei*, Caberea darwini* Camptoplites areolatus* Camptoplites latus*,§ Cellaria aurorae Cellaria coronatum* Cellarinella edita Chaperiopsis quadrispinosus Cornucopina pectogemma* Cornucopina lata*, ,§ Cornucopina nupera ,§ Crassimarginatella inconstantia* Dakariella concinna*, Dakariella dabrowni Dakariella sp? ,§ Dendroperistomata projecta* Ellisina antarctica Escharella mamillata Escharella watersi* Exallozoon simplicissimum* Exochella hymanae Exochella rogickae Fenestrulina exigua Fenestrulina fritilla Fenestrulina proxima Galeopsis bullatus*, Himantozoum taurinum ,à,§ Icelozoon lepralioides* Kymella polaris* Lacerna watersi Melicerita obliqua* Micropora brevissima Microporella stenoporta Notoplites drygalskii* Orthoporidra brachyrhyncha* Orthoporidra stenorhyncha Osthimosia curtioscula* Osthimosia notialis Pyriporoides uniserialis Ralepria conforma* Reteporella erugata* Reteporella frigida Reteporella gelida* 119 268 414 605 647 1030 1055 2156 2157 2189 3102 4698 4802 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X Journal of Biogeography 37, 1648–1656 ª 2010 Blackwell Publishing Ltd X X X X 1651 D. K. A. Barnes and P. Kuklinski Table 2 Continued Species Reteporella hippocrepis Smittina alticollarita Smittina antarctica Smittina glebula Smittina incernicula Smittina rogickae Smittinella rubringulata*, Smittoidea ornatipectoralis Talivittaticella frigida* Thrypticocirrus phylactelloides Thrypticocirrus rogickae Thrypticocirrus contortuplicata Toretocheilum absidatum Toretocheilum turbinatum* Trilaminopora trinervis Turritigera cribrata* Undescribed species Total 119 268 414 605 647 1030 1055 2156 2157 2189 3102 4698 4802 5 3 6 2 2 3 X X X X X X X X X X X X X X X X X 14 X 14 X 10 X X 12 21 16 12 Data are presence/absence and shown as shelf (unshaded); slope (light grey); abyss (dark grey). *New depth record. First record for Weddell Sea. àFirst record for Southern Ocean. §First record for Southern Ocean abyss. research ships). Thus true values for slope and abyssal density of colonies were one to two orders of magnitude lower than 0.046 and 0.0026 bryozoan colonies per cm2. As well as abundance, bryozoan richness clearly decreased rapidly with depth in the Weddell Sea (Fig. 1). The 43 bryozoan species in the shelf samples include four that have not been previously reported from the Weddell Sea (Table 2). A further five of the 32 species we found in the slope samples are new reports for the Weddell Sea, as are four of the six abyssal species (and two of these species are recorded for the first time from the Southern Ocean). Key to our hypotheses, we established that at least six bryozoan species occur in the Southern Ocean deep sea (> 3000 m) – these were the first abyssal bryozoans to be reported from the abyss south of the Polar Front. Such data show that although 201 bryozoan species are described from the Weddell Sea, the fauna is still not well known across depths. Seventeen bryozoan species were known previously from the continental slope of the Weddell Sea (Hayward, 1981; Zabala et al., 1997; Gontar & Zabala, 2000), and 38 for this depth around Antarctica (see López-Fé, 2005 and Barnes, 2008 for further species). We have considerably increased these values, to 44 and 59. On average, at shelf and slope depths, each new sample is adding 0.39 new species records to the known Weddell Sea fauna (Table 3). Treating boulders as a sample unit, few if any speciesaccumulation curves approached asymptote (Fig. 2). Thirteen species of bryozoans occurred in both the Weddell Sea shelf and the slope samples (22.6% similarity; Table 3). Strikingly, no species were common to slope and abyssal 1652 samples, but one species (Melicerita obliqua; Fig. 3) occurred in shelf and abyssal samples. By referring to literature species– depth data, we found that Camptoplites latus and Cornucopina lata, which we found in abyssal samples, had been found previously in Antarctic Sea shelf samples (see Gontar & Zabala, 2000 and Hastings, 1943, respectively). Of the species found in the current study, 31 are known only from the shelf, five only from the slope, and two only from the deep sea. Eight of the species have never been reported from the Antarctic shelf, to our knowledge. The known depth ranges of bryozoans, which were found in the current study and occur at slope depths and below, are shown in Fig. 4. The currently known bryozoan fauna of the Weddell Sea abyss is more similar to the deep sea fauna (of the South Atlantic and Indian oceans) than to that of the shallower Southern Ocean (Fig. 5). Non-metric multi-dimensional scaling of samples showed an essentially similar pattern of separation between abyssal and shallower samples (associated stress 0.1, plot not shown). The Weddell Sea shelf bryozoan fauna was considerably more similar to those on other Antarctic shelves than to that of the adjacent (Weddell Sea) continental slope. The bryozoan fauna of the Weddell Sea shelf, known to date, is thus not a subset of the Weddell Sea slope or abyssal faunas. It is also not a subset of shelf faunas from the outlying islands (26 species are present that have not been reported from the Scotia arc or Bouvet Island at any depth, e.g. Arachnopusia gigantea). Finally, the Weddell Sea shelf fauna is also not a subset of the outlying islands and the Weddell Sea slope and deep sea. Furthermore, at least 12 Journal of Biogeography 37, 1648–1656 ª 2010 Blackwell Publishing Ltd Recolonization of Antarctic shelves by marine life Table 3 Rate of bryozoan species being recorded for the first time at a locality, bryozoan species richness and shelf–slope similarity in bryozoan species composition for the Weddell Sea and adjacent islands. New records per sample Total species Shelf–slope % similarity South Sandwich Islands South Georgia Islands South Orkney Islands Weddell Sea Bouvet Island 0.5 65 ND 0.6 132 10.1/31.2* 0.7 110 22.6 0.39 201 22.6 3.5 34 ND We define a sample as a single successful deployment of a benthic apparatus (e.g. Agassiz trawl). Data are from current study (bold), López de la Cuadra & Garcı́a Gómez (2000), Barnes (2008), Kaiser et al. (2008), and BIOPEARL I unpublished cruise data. *Values for Shag Rocks. ND, no data could be found. species (e.g. Cellarinella weddelli) have only ever been reported from the Weddell Sea shelf. Other species are known only from the shelves of the Weddell Sea and other locations (e.g. South Georgia, Camptoplites assymetricus; South Shetland Islands, Arachnopusia ferox). Antarctic shelf faunas broadly cluster together, with the exception of the very isolated, young and steep-sided Bouvet Island. (a) DISCUSSION (b) Figure 2 Cheilostome bryozoan species accumulation in the Weddell Sea region with sample number. (a) Accumulation of species with boulder as sample unit and (b) bryozoan colony as the sample unit. Additional data for Scotia arc sites from Barnes (2008) and unpublished BIOPEARL II cruise data. (a) Similarity of Antarctic abyssal and wider geographical deep sea biological groupings, as we found here with bryozoans, is not surprising. Many conditions are similar, and abyssal water in the southern Atlantic, Indian and Pacific oceans is Antarctic Bottom Water. What is much more surprising is the apparent lack of connectivity in bryozoan assemblages between the shelf, the slope and the deep sea, perhaps especially in an area of bottom water formation such as the Weddell Sea. It has long been suggested that species have made emergent and submergent migrations between the Antarctic shelf and the deep sea (Hessler & Wilson, 1983; Brandt, 1991), in some groups now supported by molecular evidence (see Held & Wägele, 2005). In our results, we could find no support for our null hypothesis that Weddell Sea shelf bryozoans were a subset of those in the abyss or the continental slope. Bryozoans decrease with depth across oceans and seas (see Zabala et al., 1997 for the Weddell Sea, and Schopf, 1969 for a global review), and are clearly very rare at abyssal depths in the Southern Ocean, hence (b) Figure 3 Melicerita obliqua Thornely occurs from shelf to abyssal depths in the Weddell Sea. The details of this specimen are NHM 2009.11.18.1, 65°35.40¢ S; 36°29.00¢ W, depth 4794 m, Agassiz trawl, bleached. (a) Group of zooids, some ovicellate, (b) ovicellate zooid. Scale bars: (a) 200 lm; (b) 100 lm. Journal of Biogeography 37, 1648–1656 ª 2010 Blackwell Publishing Ltd 1653 D. K. A. Barnes and P. Kuklinski Figure 4 Known depth ranges of Southern Ocean bryozoans found below 1000 m in the present study (from data of the present study and the literature). Additional data are from Hayward (1981, 1995), Zabala et al. (1997), Gontar & Zabala (2000), López de la Cuadra & Garcı́a Gómez (2000), Moyano & Ristedt (2000), Rosso & Sanfilippo (2000), López-Fé (2005), Barnes (2008), and unpublished BIOPEARL II cruise data. Thirteen other species (not found in the current study) have also been recorded in the Southern Ocean deeper than 1000 m (1000–2320 m; see literature cited). Figure 5 Similarity (Sørensen coefficient) of cheilostome bryozoan faunas from shelf, slope and abyssal depths for the Weddell Sea and surrounding areas. Data are taken from Barnes & Griffiths (2008), with additional data from the current study and literature cited in Fig. 4. they are reported here for the first time. Schopf (1969) suggested that availability of (and distance between) substrata ‘oases’ might underlie such rarity. Food availability is probably also very important, as the deep sea fauna tend to deposit-feed or scavenge, neither of which is possible for bryozoans. Of the six abyssal species found in the current study, three have never been found at shelf depths (see Hayward, 1981 for original description). One of these is an undescribed species of Dakariella. From the sampling carried out to date, the deep Weddell Sea does not appear to have either the species composition (or arguably the density of colonies) to be a major source for shelf recolonization in the scenario of large-scale glacial faunal wipe-out. We found more (but still little) evidence to support shelf recolonization from the adjacent Weddell continental slope. As suggested for other Antarctic benthos (Brey et al., 1996), bryozoans appear to have wide bathymetric ranges – enough for some to occur from shelf to slope, or even shelf to abyssal depths (Fig. 4). It was notable that the similarity between shelf and slope fauna in the Weddell Sea was at a similar level to that of the South Orkney Islands (c. 22%; Table 3), where ice probably covered little of the shelf (Herron & Anderson, 1990). If the shelf of the 1654 Weddell Sea has been recolonized from its adjacent slope (since the LGM), it should show greater shelf–slope similarity than at a location where the shelf fauna had remained intact in situ, such as the South Orkney Islands. Species-accumulation curves from continental slope samples (constructed using colony as the sample unit) did not approach asymptote (Fig. 2b). Thus there is still a possibility that the Weddell abyss or slope has a limited role as a species source for shelf recolonization. Any such argument, though, is weakened by shelf species-accumulation curves also not reaching asymptote. Furthermore, new sampling appears to be increasing the number of known endemics found only above 1000 m (Gontar, 2002, 2008). Even with an absence of evidence, there are significant theoretical problems for recolonization of the shallows from the deep, as it requires migration against a downwelling current, and very few cheilostome bryozoans are known to have larvae that spend much time in the water column. Exactly where, and to what depth, ice sheets were grounded during the LGM and previous glacial maxima is hard to know (e.g. due to modern ice scour obscuring grounding lines), but ice sheets probably grounded only to 130–250 m in the South Journal of Biogeography 37, 1648–1656 ª 2010 Blackwell Publishing Ltd Recolonization of Antarctic shelves by marine life Orkney Islands (Herron & Anderson, 1990). Our second hypothesis was that parts of the shelves of outlying islands within the Polar Front, such as the Scotia arc or Bouvet (Bouvetøya), which were not covered in grounded ice during glacial maxima, acted as supply sources to the Weddell Sea shelf. In this case, rafting adults of species (on kelp or pumice) might at least travel with prevailing currents of the Weddell Sea gyre (it is unlikely that larvae would remain viable for the many weeks such a journey would probably take). However, we also found little evidence to support this as a possibility. This was mainly because of the low levels of faunal similarity; the closest archipelago (South Sandwich Islands) and island (Bouvet) were the least similar of all shelf areas within the Polar Front. If species had rafted on kelp (e.g. from South Orkney Islands) or pumice (e.g. from South Sandwich Islands), the Weddell Sea fauna would be expected to show higher similarity with bryozoans from the shallows (than across all shelf depths) of either location, but on testing this was not the case. It remains a possibility that the Weddell Sea shelf was recolonized partly from the Weddell Sea slope and partly from the Scotia arc shelf. For this scenario to be true, the many species that occur on the Weddell Sea shelf must be present, but so far unfound, at either the Weddell Sea slope or the Scotia arc shelf (or have rapidly evolved in situ since the last glaciation). The third hypothesis, of survival of some Weddell Sea fauna throughout the last glaciation in situ (in shelf refugia), is the simplest explanation that is consistent with the evidence we have found so far. This hypothesis grew in acceptance in the scientific community from very low in 2006, to dominant in terrestrial polar biology by 2008, but is still in its infancy in marine biology (Convey et al., 2009). New molecular work and other techniques may establish which species survived in situ and, perhaps more importantly, which areas remained ice-free. For this hypothesis to be correct, there must always have been some area of the Weddell Sea shelf that was not covered by grounded ice at any one time. This may have been through diacrony, whereby some ice edges advanced while others retreated (Thatje et al., 2005), or permanently ice-free areas (as suggested on land by Convey & Stevens, 2007), or both. We do not doubt that other species will be found in the Weddell abyss, slope and shelf with increased sampling, but the number of samples and rock area that will need to be taken and examined is likely to be considerable. Likely candidates to be found in the Southern Ocean abyss include Camptoplites bicornis, which has been reported from the Pacific abyss and Antarctic shelves, and some of the species found around the margins of the Polar Frontal Zone at Îles de Kerguelen (Hayward, 1981). However, it has taken more than a decade, many considerable grants and major scientific collaboration to achieve the series of ANDEEP cruises that yielded the current material, so progress in discovering the Southern Ocean abyssal fauna will be slow. We suggest that, as with new discoveries in Antarctic terrestrial biodiversity (Convey & Stevens, 2007), the implications of marine biological finds can have important ramifications for constructing Antarctica’s glaciological past. Journal of Biogeography 37, 1648–1656 ª 2010 Blackwell Publishing Ltd ACKNOWLEDGEMENTS We thank Angelika Brandt, Wolf Arntz and Katrin Linse for provision of samples, Huw Griffiths for aiding analysis, and Claus-Dieter Hillenbrand for discussion leading to an improved manuscript. We are also very grateful to Mike Tarbecki for rock-saw work to cut colony-bearing areas of rock off large boulders. Finally, we would like to thank Julian Gutt and an anonymous referee for very constructive comments leading to a much improved manuscript. This study was completed thanks to the financial support to one of us (P.K.) from the Polish Ministry of Science and Higher Education (539/N-CAML/2009/0). REFERENCES Anderson, J.B., Shipp, S.S., Lowe, A.L., Wellner, J.S. & Mosola, A.B. (2002) The Antarctic Ice Sheet during the Last Glacial Maximum and its subsequent retreat: a review. Quaternary Science Reviews, 21, 49–70. Arntz, W.E., Gutt, J. & Klages, M. (1997) Antarctic marine biodiversity: an overview. Antarctic communities: species, structure and survival (ed. by B. Battaglia), pp. 3–14. Cambridge University Press, Cambridge. Barnes, D.K.A. (2008) A benthic richness hotspot in the Southern Ocean: slope and shelf cryptic benthos of Shag Rocks. Antarctic Science, 20, 263–270. Barnes, D.K.A. & Griffiths, H.J. (2008) Biodiversity and biogeography of southern temperate and polar bryozoans. Global Ecology and Biogeography, 17, 84–99. Barnes, D.K.A., Hodgson, D.A., Convey, P., Allen, C.S. & Clarke, A. (2006) Incursion and excursion of Antarctic biota: past, present and future. Global Ecology and Biogeography, 15, 121–142. Barnes, D.K.A., Kaiser, S., Griffiths, H.J. & Linse, K. (2009) Marine, intertidal, freshwater and terrestrial biodiversity of an isolated polar archipelago. Journal of Biogeography, 36, 756–769. Brandt, A. (1991) Zur Besiedlungsgeschichte des antarktischen Schelfes am Beispiel der Isopoda (Crustacea, Malacostraca). Berichte zur Polarforshung, 98, 1–240. Brandt, A., Gooday, A.J., Brandão, S.N. et al. (2007) First insights into the biodiversity and biogeography of the Southern Ocean deep sea. Nature, 447, 307–311. Brey, T., Dahm, C., Gorny, M., Klages, M., Stiller, M. & Arntz, W.E. (1996) Do Antarctic benthic invertebrates show an extended level of eurybathy? Antarctic Science, 8, 3–6. Clarke, A. & Johnston, N.M. (2003) Antarctic marine benthic diversity. Oceanography and Marine Biology – An Annual Review, 41, 47–114. Colwell, R.K. (2009) EstimateS: Statistical estimation of species richness and shared species from samples. Version 8.2. Available at: http://viceroy.eeb.uconn.edu/estimates. Convey, P. & Stevens, M.I. (2007) Ecology: Antarctic biodiversity. Science, 28, 1877–1878. 1655 D. K. A. Barnes and P. Kuklinski Convey, P., Stevens, M.I., Hodgson, D.A., Smellie, J.L., Hillenbrand, C.-D., Barnes, D.K.A., Clarke, A., Pugh, P.J.A., Linse, K. & Craig Cary, S. (2009) Exploring biological constraints on the glacial history of Antarctica. Quaternary Science Reviews, 28, 3035–3048. Dayton, K. & Oliver, J.S. (1977) Antarctic soft-bottom benthos in oligotrophic and eutrophic environments. Science, 197, 55–58. Gontar, V.I. (2002) New species and new genera of Cheilostomata from the Weddell Sea, Antarctica (Bryozoa). Zoosystematica Rossica, 10, 285–292. Gontar, V.I. (2008) Three new species of the genus Smittina from the Weddell Sea, Antarctic (Bryozoa: Cheilostomata: Smittinidae). Zoosystematica Rossica, 17, 7–9. Gontar, V.I. & Zabala, M. (2000) Bryozoa. Berichte zur Polarforschung, 372, 24–43. Hastings, A.B. (1943) Polyzoa (Bryozoa) I. Scrupocellariidae, Epistomiidae, Farciminariidae, Bicellariellidae, Aeteidae, Scrupariidae. Discovery Reports, 22, 301–351. Hayward, P.J. (1981) The Cheilostomata (Bryozoa) of the deep sea. Scientific Results of the Danish Deep-Sea Expedition around the world (1950–52) (ed. by T. Wolff), pp. 21–68. Galathea Report, Vol. 15. University of Copenhagen, Copenhagen. Hayward, P.J. (1995) The Antarctic cheilostomatous Bryozoa. Oxford University Press, Oxford. Held, C. & Wägele, J.-W. (2005) Cryptic speciation in the giant Antarctic isopod Glyptonotus antarcticus (Isopoda: Valvifera: Chaetiliidae). Scientia Marina, 69, 175–181. Herron, M.J. & Anderson, J.B. (1990) Late Quaternary glacial history of the South Orkney Plateau, Antarctica. Quaternary Research, 33, 265–275. Hessler, R.R. & Wilson, G. (1983) The origin and biogeography of malacostracan crustaceans in the deep sea. Evolution, time, and space: the emergence of the biosphere (ed. by R.W. Sims, J.H. Price and P.E.S. Whalley), pp. 227–254. Systematics Association, Special Publication 23, Academic Press, London. Kaiser, S., Barnes, D.K.A., Linse, K. & Brandt, A. (2008) Epibenthic macrofauna associated with the shelf and slope of a young and isolated Southern Ocean island. Antarctic Science, 20, 281–290. Linse, K., Cope, T., Lorz, A.-N. & Sands, C. (2007) Is the Scotia Sea a centre of Antarctic marine diversification? Some evidence of cryptic speciation in the circum-Antarctic bivalve Lissarca notorcadensis (Arcoidea: Philobryidae). Polar Biology, 30, 1059–1068. López de la Cuadra, C.M. & Garcı́a Gómez, J.C. (2000) The cheilostomate Bryozoa (Bryozoa: Cheilostomatida) collected by the Spanish ‘Antártida 8611’ Expedition to the Scotia Arc and South Shetland Islands. Journal of Natural History, 34, 755–772. 1656 López Gappa, J.J. (1986) A new bryozoan genus from the Weddell Sea, Antarctica. Polar Biology, 6, 103–105. López-Fé, C.M. (2005) Cheilostomate Bryozoa of the Bellingshausen Sea (Western Antarctica): a preliminary report of the results of the ‘Bentart 2003’ Spanish expedition. Bryozoan studies 2004 (ed. by H.I. Moyano G., J.M. Cancino and P.N. Wyse-Jackson), pp. 173–179. A.A. Balkema, London. Mahon, A.R., Arango, C.P. & Halanych, K.M. (2008) Genetic diversity of Nymphon (Arthropoda: Pycnogonida: Nymphonidae) along the Antarctic Peninsula with a focus on Nymphon australe Hodgson 1902. Marine Biology, 155, 315– 323. Moore, J.K., Abbott, M.R. & Richman, J.G. (1999) Location and dynamics of the Antarctic Polar Front from satellite sea surface temperature data. Journal of Geophysical Research, 104, 3059–3073. Moyano, G.H.I. & Ristedt, H. (2000) The Sclerodomidae (Bryozoa) of the eastern Weddell Sea collected during the Antarctic I, II, V Expeditions of the R/V Polarstern. Gayana, 64, 83–107. Rosso, A. & Sanfilippo, R. (2000) Shallow-water bryozoans and serpuloideans from the Ross Sea (Terra Nova Bay, Antarctica). Ross Sea ecology (ed. by F. Faranda, I. Guglielmo and A. Ionora), pp. 515–525. Springer Verlag, Berlin and Heidelberg. Schopf, T.J.M. (1969) Geographic and depth distribution of the phylum Ectoprocta from 200 to 6,000 meters. Proceedings of the American Philosophical Society, 113, 464–474. Thatje, S., Hillenbrand, C.-D. & Larter, R. (2005) On the origin of Antarctic marine benthic community structure. Trends in Ecology and Evolution, 20, 534–540. Zabala, M., Orejas, C. & Alvá, V. (1997) Bryozoans of the Weddell Sea. Berichte zur Polarforschung, 249, 55–61. BIOSKETCHES David Barnes has a PhD from the Open University, UK. He is currently a researcher at the British Antarctic Survey and teaches at the University of Cambridge. His main interests include benthic ecology, biodiversity and biogeography of polar seas. Piotr Kuklinski has a PhD from the Institute of Oceanology in Poland. He is currently a researcher at the Institute of Oceanology and a scientific associate at the Natural History Museum, London. His main interest is in benthic ecology and bryozoan taxonomy of polar regions. Editor: John Lambshead Journal of Biogeography 37, 1648–1656 ª 2010 Blackwell Publishing Ltd