The youngest pre-caldera volcanism of Deception Island is represented by a thick sequence of subaerial pyroclastic deposits which has been grouped as the Yellow Tuff Formation. Most of these deposits were related to the explosive activity of a central vent which was destroyed during the formation of the caldera. Two members can be distinguished in this formation. The lower member is mainly composed of 1 to 12 m thick massive pyroclastic flow deposits with interbedded air-fall and surge deposits. The upper member is in stratigraphical continuity with the lower member and consists of base surge deposits with minor air-fall and thin pyroclastic flow deposits. The pre-caldera deposits have undergone a palagonitic alteration which produced crystallization of smectites, Fe-oxides, zeolites and calcite.

Subantarctic Marion Island is one of the few localities where the congeneric albatrosses Phoebetria fusca and P. palpebrata breed sympatrically. Chicks of both species at Marion Island were induced to regurgitate their stomach contents after being fed. Liquid formed over half the diet by mass. Cephalopods occurred most frequently in both species' diets. In terms of mass, cephalopods formed the larger part of the diet of sooty albatrosses, whereas fish was more important to light-mantled sooty albatrosses. Crustaceans and birds were also recorded for both species. Squid of the families Onychoteuthidae, Histioteuthidae, Chiroteuthidae and Cranchiidae occurred most abundantly in both species. Most squid taken by both albatrosses were of species known to float after death, suggesting that scavenging plays an important role in the species' foraging behaviour. Light-mantled sooty albatrosses consumed more squid restricted to the south of the Antarctic Polar Front than did sooty albatrosses, supporting a trend to latitudinal segregation of the two species while foraging.

Early stages of toothfish are reported from cruises between 1978 and 2001. South Georgia was regularly sampled, Shag Rocks was briefly sampled in three summers, and Burdwood Bank was sampled in 1997. Most samples were using 8 m2 and 1 m2 nets in the upper 250 m and upper 3 m, respectively. Forty-three pelagic larval/early juvenile stages between 18.2 and 63 mm standard length (Ls) were taken between December and March. Four specimens were taken at Burdwood Bank, one at Shag Rocks, and 38 at South Georgia. Forty fish were from the upper 250 m, and 23 from the upper 3 m. They occurred onshelf and offshore over water depths between 10 and 3000 m. Estimated growth rate, at South Georgia from 11 cohorts (including published data), is c. 1% Ls, d−1 (n = 39). From this and a hatch size of 15 mm Ls, the predicted hatching season is November to mid-December. From seven trawl surveys during summer at South Georgia and Shag Rocks between 1986 and 2000, the mean depth of the smaller demersal juveniles 13–25 cm total length (LT) was < 140 m, indicating that early juveniles probably migrate inshore.

A regional chemical boundary termed the ‘salt line’, in the Vestfold Hills of East Antarctica, has been investigated using X-ray diffraction and electron probe analyses of surficial salts, and conductivity of surficial sediments. West of the salt line, halite and thenardite are abundant. These salts are derived from dispersal of marine aerosols, saturation of sediment by seawater during postglacial marine transgression, and glacial dispersal of salt-saturated fjord bottom sediments. East of the salt line, subglacial calcium carbonates and salts formed by chemical weathering of their substrates occur. Chemically and morphologically diverse minerals form the weathering products. They include two minerals not found previously in Antarctica, dypingite and hydromagnesite, and the first confirmed occurrence of brushite.

Three models were used to look at the Southern Ocean Ross Sea sector circulation and hydrography. Two were climate models of low (1°) to intermediate resolution (1/3°), and one was an operational high resolution (1/10°) ocean model. Despite model differences (including physics and forcing), mean and monthly variability aspects of off-shelf circulation are consistently represented, and could imply bathymetric constraints. Western and eastern cyclonic gyral systems separated by shallow bathymetry around 180°E redistributing water between the wider Southern Ocean and the Ross Sea are found. Some model seasonal gyral transports increase as the Antarctic Circumpolar Current transport decreases. Model flows at 900 m at the gyral eastern end compare favourably with float data. On-shelf model depth-averaged west–east flow is relatively consistent with that reconstructed from longline fishing records. These flows have components associated with isopycnal gradients in both light and dense waters. The climate models reproduce characteristic isopycnal layer inflections (‘V’s) associated with the observed Antarctic Slope Front and on-shelf deep water formation, and these models transport some 4 Sv of this bottom water northwards across the outer 1000 m shelf isobath. Overall flow complexity suggests care is needed to force regional Ross Sea models.

This study is the first published survey of diatom-environment relationships on sub-Antarctic Macquarie Island. Fifty-eight sites in 50 coastal and inland lakes were sampled for benthic diatoms and water chemistry. 208 diatom species from 34 genera were identified. Multivariate analyses indicated that the lakes were distributed along nutrient and conductivity gradients. Conductivity, pH, phosphate (SRP), silicate and temperature all explained independent portions of the variance in the diatom data. Transfer functions provide a quantitative basis for palaeolimnological studies of past climate change and human impacts, and can be used to establish baseline conditions for assessing the impacts of recent climate change and the introduction of non-native plants and animals. Statistically robust diatom transfer functions for conductivity, phosphate and silicate were developed, while pH and temperature transfer functions performed less well. The lower predictive abilities of the pH and temperature transfer functions probably reflect the broad pH tolerance range of diatoms on Macquarie Island and uneven distribution of lakes along the temperature gradient. This study contributes to understanding the current ecological distribution of Macquarie Island diatoms and provides transfer functions that will be applied in studies of diatoms in lake sediment cores to quantitatively reconstruct past environmental changes.

Although the lower size limit of microorganisms was previously believed to be c. 0.2 μm, there is evidence for the existence of microorganisms that can pass through 0.2 μm-pore-size filters called ultramicrobacteria or nanobacteria. However, information on the phylogeny and biogeography of these bacteria is limited. We obtained 53 isolates of 0.2 μm-passable bacteria from 31 samples collected at 26 locations worldwide, including the Arctic Svalbard Islands, deserts, and Maritime Antarctica. Phylogenetic analysis of near full-length 16S rRNA gene sequences revealed that 18 of the 53 isolates were < 97% homologous with previously cultured isolates, representing potentially novel species. Two isolates (order Rhizobiales) (100% identical) collected from Byers Peninsula, Livingston Island in Maritime Antarctica, were closely related (99.8% similarity) to an isolate collected from intertidal sediments in East Antarctica. In addition, the sequence of this Antarctic isolate showed ≥ 97% similarity to 901 sequences derived from known isolates and samples collected at geographically disparate locations under various environmental conditions. Interestingly, among 13 sequences showing ≥ 99% similarity, ten were isolated from cryospheric habitats such as Arctic, Antarctic, and alpine environments. This implies that such Rhizobiales strains occur in the cryospheric regions, however, their abundance and biomass may be scarce depending on the geographic location.

An inventory of Antarctic and Subantarctic mysid fauna is presented, together with a summary of the present state of knowledge of species and their taxonomic diversity, geographic and bathymetric distribution patterns. Fifty nine species of Mysidacea (Crustacea, Peracarida) are now known. Of these, 37 were reported for the Antarctic region and 31 for the Magellan region; six species occur further north in the Southern Ocean, but south of 40°S. 51% of the Antarctic Mysidacea are endemic, and the figure for the Magellan region is 48%. Most of the species live hyperbenthically, but some also occur bathy- or mesopelagically. Mysidetes has the most species in the Southern Ocean, and Eucopia australis is the species with the widest bathymetric distribution (600–6000 m depth). It is concluded that an emergence of species onto the Antarctic shelf in the Neogene was quite unlikely, because none of the mysid species is a true deepsea species, and most species occur on the shelf or at the shelf break. It is more probable that present day species colonized the Southern Ocean via shallower waters. The examples of the distribution of different genera suggest that the Mysidacea of the Southern Ocean probably had various geographical origins.

New polydolopid marsupial specimens have been recovered from the La Meseta Formation, a late early Eocene to probably early Oligocene unit cropping out in the northern third of Seymour (Marambio) Island, at some 100 km off the northern Antarctic Peninsula. Our review of the original materials, as well as the new specimens from the same levels, led us to: 1) revalidate the genus Antarctodolops Woodburne & Zinsmeister 1984, 2) regard Eurydolops seymouriensis Case, Woodburne & Chaney 1988 as a junior synonym of Antarctodolops dailyi Woodburne & Zinsmeister, and 3) recognize a new species of this same genus: A. mesetaense. As previously stated, the polydolopid radiation might be related to the expansion of the Nothofagus flora, as both have the same spatial distribution in southern South America and West Antarctica.

A review of the collembolan fauna of the South Shetland Islands is presented. Cryptopygus nanjiensis Yue & Tamura is synonymized with C. antarcticus Willem. A record of Tullbergia mediantarctica Wise from King George Island is considered a misidentification of Tullbergia mixta Wahlgren and Tillieria penai Weiner & Najt, described from the same island, is synonymized with T. mixta. The current fauna stands at eleven species, of which at least three are introduced. A checklist of Collembola currently considered to occur in the South Shetland Islands is supplied with distributional data.

The composition of suprabenthic Peracarida from a depth between 1 and 1.33 m above the seafloor was investigated. In order to study the abundance and diversity quantitatively, the samples were taken by means of the supranet of an epibenthic sledge off Kapp Norvegia and off the South Shetland Islands. The samples were taken during ANT XVII-3 with RV Polarstern in April and May 2000. These samples of EASIZ III received during the Antarctic autumn are compared with those of EASIZ II taken during the summer. While the autumn samples show far less abundance of all peracarid taxa, the diversity is about twice as high as during summer. In total 115 species of peracarid crustaceans were sampled at six stations yielding 1336 individuals standardized to 1000 m3 hauls. Amphipoda clearly dominated in abundance and diversity.

We surveyed the southern elephant seal population at Heard Island regularly from February 1992 to March 1993, and determined the haulout patterns of the major components of the population. While haulout patterns of moulting and immature seals may give broad indices of population trends, the breeding haulout of adult females was the only reliable haulout that could be used to determine annual pup production. During the breeding season 14 277 adult females were counted. Raw counts were corrected using two models, one purely mathematical and the other based on the haulout behaviour of adult female seals. The two models have slightly different assumptions, but both provided good fits to the observed haulout patterns and estimated total population with a coefficient of variation of less than 5%. Total pup production was estimated at between 17 000 and 18 000 for 1992. Previous counts of elephant seals from 1949–51, 1985 and 1987 were corrected using the same models. The two models gave estimates of the population that were within ± 2.5% for all but one year. The population declined by about 50% between 1949 and 1985 but there appears to have been little change from 1985–92. The previous decline may be related to changes in sea-ice.

A total of 135 bacterial strains were isolated from congelation (land fast) sea ice samples and ice algae biomass samples obtained from the coastal areas of the Vestfold Hills in East Antarctica (68°S, 78°E) during the summers of 1992–95. The sea ice isolates, along with reference strains. were analysed by numerical taxonomy and for DNA base composition in order to determine the biodiversity of sea ice bacteria. From these analyses 22 clusters of strains (phena) were obtained with most phena apparently representing novel bacterial taxa. The sea ice isolates could be categorized into three groups based on their ecophysiology: 1) slightly halophilic, psychrophilic bacteria often possessing fastidious growth requirements and which were predominantly isolated from sea ice algae biomass or from algae-rich ice samples; 2) halotolerant and psychrotolerant bacteria; and 3) non-halophilic bacteria isolated primarily from upper sections of congelation ice and other ice samples with low levcls of algal biomass.

The mitochondrial COI gene of the Antarctic springtail, Gressittacantha terranova, was sequenced across a polar coastal landscape at Terra Nova Bay, northern Victoria Land. Samples from two altitudinal transects in the foothills directly south of Campbell Glacier were compared with samples from Springtail Valley (northern foothills) as an external reference population. We found that mtDNA haplotypes clustered into two lineages (clades) with a mean sequence divergence of 10% (uncorrected distance). However, there was no phylogeographic structure found at this spatial (landscape) scale with haplotypes from both divergent clades found sympatric across most populations. At the landscape scale, the considerable genetic divergence revealed within G. terranova is around five times greater than any other continental Antarctic springtail examined to date. These data indicate a Pliocene divergence event in G. terranova around 4–5 million years ago. The unusual distributional profile of haplotypes - occurrence of multiple haplotypes at single sites and genetic contiguity between sites that are not physically contiguous - suggests a subsequent ‘reshuffling’ of haplotypes in the Holocene that has an ecological basis.

Several sub-Antarctic penguin populations have exhibited decreasing trends in the last 5–20 years. At Iles Kerguelen, the gentoo penguin (Pygoscelis papua) has decreased by approximately 30% over the last 15 years. This decrease is likely to be related to reduced food availability for this coastal species. We discuss this decrease with regard to the impact of overfishing and to the potential effects of climate changes on marine food webs.

By considering the basic stress equations for a unit volume of ice, a set of differential equations describing ice shelf flow is derived. In view of the lack of basal shear stresses at the bottom of ice shelf a model simulation which is restricted to the horizontal dimensions will not imply substantial errors. The model is applied to the Filchner-Ronne Ice Shelf, Antarctica, and model equations are solved in terms of finite differences on a 10 × 10 km grid. Present ice thickness data and boundary conditions, i.e. the balance velocities at the grounding line and strain rates at the ice front are entered as input. Using a non-linear Glen-type flow law (n=3) and a constant depth-averaged flow law parameter, representing an ice temperature of −17°C, a convincing velocity field is derived as a solution of the model equations. The model takes into account restrained flow across ice rumples where sufficient field data are available. A diagnostic run reproducing present velocity magnitudes is followed by two prognostic runs, each representing 2000 years of simulation. Transient ice thickness changes are obtained from solving the mass conservation equation. Two different assumptions concerning basal melting rates demonstrate its importance to ice shelf dynamics. Assumptions are: a) no basal melting, b) basal melting rates (−2m a−1 to +3m a−1) as derived from model results and geophysical field data.

Surface-based observations were taken of cloud and precipitation particles on the Avery Plateau (66°50.34′S 65°29.58′W), Antarctic Peninsula from 25 November to 13 December 1995. This paper considers cloud parameters on three days during this period when the cloud base reached ground level and snow was falling. It was found that on all three days more ice crystals were present in the cloud than would be expected from simple theoretical considerations. The rate of snowfall decreased as the number of ice crystals increased, the large number of ice crystals present effectively suppressing the formation of large precipitation-sized crystals. The source of the ice nuclei that allowed the formation of the large number of crystals is not known for certain but is thought to be the snow surface, possibly in the form of very fine ice crystals blown from the surface during blowing/drifting snow episodes.

The Bunger Hills were covered by the Antarctic Ice Sheet during the last glaciation. During deglaciation the ice sheet margin collapsed into the marine inlets and the sea entered the oasis before 7.7 ka BP. Raised beaches occur widely below 8.5 m and indicate uplift at 1.4 m ka−1 during the middle and late Holocene. After the coastal inlets were formed, the Edisto Ice Tongue and Apfels Glacier still impinged on land margins in the west of the oasis. Two sets of marginal moraines were formed; the Older Edisto Moraines after 6.2 ka BP and the Younger Edisto Moraines during the last few centuries. The margins of the Antarctic Ice Sheet and Apfels Glacier in the south have maintained their present positions since at least 5.6 ka BP and probably 10 ka BP.

Land-fast ice in the vicinity of Adélie Land was sampled during spring 1995. The ice was annual, thin, with no consistent snow cover, and exposed to oceanic conditions. Temporal and spatial variations of the vertical pigment distribution were studied in relation to environmental factors, during the break up of the ice. Different levels were sampled in the congelation ice and the platelet ice-like layer (PLI). Under-ice water and open water masses were also sampled. The algal biomass was greater in the PLI (24 ±14 μg chl a l−1 offshore and up to 9 mg chl a l−1 near-shore), than in the under-ice water, and fell to 0.9 ± 0.64 μg chl al−1 in open water masses. Homogenous low pigment concentrations were detected in the upper levels of congelation ice. A gradient was identified along a 7 km seaward transect, sampled in November, with the lowest biomass offshore. The integrated pigment concentrations in fast ice reached very high levels 500 mg chl a m−2 near the coast and 0.8 mg m−2 offshore), with apparently no relationship with either the ice thickness or snow cover. In the congelation ice nutrient concentrations were low and their distribution homogenous, whereas in the PLI high concentrations of nitrate (up to 100–300 μM NO3) and silicic acid [30–100 μM Si(OH)4] were detected, often related to high pigment concentrations and proximity to islands. The sea ice algae communities were diverse, but mostly composed of chain-forming and tube-dwelling pennate diatoms (Amphiprora, Berkeleya, Nitzschia and Navicula). Cell densities in PLI reached up to 1010 cells l−1. At very low biomass and cell densities 2 104 cells l−1) the phytoplankton also had a low diversity; some species were similar to those of the PLI, such as Navicula glaciei, but other were typically planktonic (Chaetoceros). At sea ice break-up it is estimated that a significant proportion of particulate matter (up to 0.5 g chl a m−2 near-shore) was transferred to the underlying water masses (on an average 15 t POC km−1 shoreline).

The notothenid species Trematomus bernacchii has previously been shown, by allozyme analysis, to be a complex of two cryptic species, one of which being more closely related to T. hansoni than to the other T. bernacchii cryptic species. Two T. bernacchii colour morphs, “white blotch” and “brown”, at McMurdo Sound, may correspond to these cryptic species. In this study, we present mitochondrial DNA sequences of the 12S and 16S ribosomal regions for six “white blotch” morphs, eight “brown” morphs collected in McMurdo Sound, one individual collected off the Antarctic Peninsula, and two T. hansoni individuals from McMurdo Sound. These sequences were compared with those of T. bernacchii and T. hansoni in the literature. Based on 14 phylogenetically informative sequences, no differences were found between “white blotch” and “brown” morphs. Furthermore, only one substitution separated these sequences from the previously published T. hansoni sequence, while 10 substitutions separated them from the previously published T. bernacchii sequence. Misidentified specimens, and sequence misreadings may be at the origin of these discrepancies. However, the presence of cryptic species within T. bernacchii and T. hansoni is not ruled out.