It is hard to divorce most human activities, including science, completely from politics. Politics is about perceived certainty whilst science is about doubt – they make strange bedfellows. Politicians detest probabilities whilst scientists abhor the absolute. Nowhere is the relationship between politics and science more publicly developed than in the Antarctic Treaty System. In the only continent devoted to peace and science it might be supposed that, after more than thirty years, the role of science would be both more robust and more pivotal in decision-making than elsewhere in the world. So it appears at present but will it remain so?

Middle Eocene land mammals from La Meseta Formation, Seymour (Marambio) Island are reviewed. A taxonomically diverse fossil land-vertebrate assemblage with small and medium-size mammals has been recovered from four localities. The depositional setting is shallow marine and most of the mammal-bearing beds are in reworked, moderate to high energy subtidal facies. The characteristics of these mammals not only confirm but also strengthen the biogeographical relationships between southern South America (Patagonian Province) and the Antarctic Peninsula during the Paleogene and rule out the possibility of a major barrier between these areas. The Antarctic ungulates (Astrapotheria and ?Litopterna) are plesiomorphics in retaining low crowned cheek teeth and are more similar to those from the Pancasamayoran local faunas of southern South America (Patagonia).

Melt waters cover c. 20% of the McMurdo Ice Shelf during the austral summer. The streams, ponds, and lakes up to 104 m2 in area occur in two types of terrain systems with differing morphological, chemical, and biological characteristics: pinnacled ice (PI) areas with sparse sediment cover, low relief, and little biomass; and ice-cored moraine (ICM) areas with 10–20 cm sediment cover, hummocky topography with up to 20 m relief, occasional mirabilite deposits, and dense benthic cyanobacterial mats. Pond water composition in the two areas is markedly different. PI area melt waters have low salinities, <2270 mg 1−1 total dissolved salts (TDS), and near neutral pH, mean = 7.8. The chemical composition of PI waters closely follows that of diluted sea water, suggesting that the release of ions from the sea ice matrix of the ice shelf is the major solute source. In contrast, ICM area melt waters have a wide range of salinities, up to 60 400 mg 1−1 TDS and alkaline pH, mean = 9.3. The chemical composition in c. 40% of the ICM ponds investigated did not resemble that of sea water, but had higher relative abundances of SO2−4, Na+, K+ and Ca2+. Leaching of local salt deposits, particularly mirabilite, weathering of surficial sediments, and morphological features promoting closed-basin brine evolution are possible contributing factors to the enrichments.

Only one living specimen of Adamussium colbecki was found amongst the shells of this species collected in the Lazarev and Weddell Seas, and those of the limpet Nacella cf. concinna from three locations in the southern Weddell Sea. In order to explain this peculiarity four hypotheses are discussed. The possibility that the shells belong to living deep-water populations is in contradiction to the high number of trawl samples taken in the area and the abundance of empty shell material. 14C dates show the shells are recent, thus excluding the possibility that the shells belong to autochthonous fossil populations exposed on the seafloor. Lateral transport of living specimens or shells over long distances by marine currents or by ice is unlikely due to their state of preservation and the water circulation pattern. The most likely explanation is that the shells are allochthonous, transported only over short distances by marine currents from adjacent areas underneath floating ice shelves where both species have living populations.

The macroinfauna at depths 30–115 m was sampled in March–April 1989, c. two months after an oil spill that resulted from the grounding of the Bahia Paraiso. Stations consisted of the oil-spill site and a comparable control location, and two historical sites previously sampled in 1971. The historical sites were located at two distinct points along a known continuum of increasing physical stability with depth, attributed to disturbances from glacial calving. Macroinfaunal assemblages at most stations were characterized by very high densities and numbers of taxa. There were no significant differences (P<0.05) between the oil-spill and control sites in numbers of individuals, species, or families; nor were there any major differences in dominant fauna or overall community composition. The absence of a detectable impact on the fauna is consistent with results of hydrocarbon analyses, which showed that subtidal sediments were nearly devoid of contamination emanating from the Bahia Paraiso. The assemblage at the shallower of the two historical sites, however, showed a substantial change over the 18-yr period between studies. This change consisted of a shift toward a more species-rich and abundant macroinfauna characteristic of the more physically stable parts of the harbour. This change may be related to the fact that the glacier face near this site has retreated c. 250 m over the last 20 yrs, resulting in less physical disturbance of the adjacent seafloor.

The occurrence and distribution of the South Georgia fungal flora, particularly Ascomycotina and Basidiomycotina, is assessed in terms of habitat and substrate preference. The 113 taxa reported comprise 37 basidiomycetes, 49 ascomycetes, six myxomycetes and at least 21 lower fungi. Peat and litter substrata associated with tall tussock grassland have a rich macro-fungal flora, and numerous species occur in bog and mire communities, some in abundance from mid to late summer. Many micro-fungi and ascomycetes colonize dying leaves and inflorescences of specific vascular plants, and a few colonize bryophytes and lichens. At least a dozen species, probably non-indigenous, are associated with rotting timber and other imported materials at former whaling stations. An intensive survey of the South Georgia mycoflora is necessary to gain better understanding of their role in decomposition and nutrient cycling processes in the principal plant communities.

In the cold Antarctic environment labile organic compounds may accumulate in soil due to relatively low utilization rates by heterotrophic microorganisms. Microbial fermentation of these compounds might contribute to the development of strongly acid soils. To test this and assess concentrations, extracts of a range of soils in the Windmill Islands, Budd Coast were analysed by GLC and HPLC for the presence of low molecular weight sugars, polyols and organic acids. Concentrations of sugars and polyols up to 3300 mg g−1 were detected in cryptogam dominated soils. Some, such as trehalose, may have principally originated in the soil microflora. Soils from occupied penguin rookeries were found to possess oxalic, acetic, propionic and succinic acids at levels up to 1000 mg g−1 soil. Most other soils, however, lacked these acids at detectable levels (1–5 mg g−1 soil). No correlation was established between organic acid accumulation and soil pH although those dominated by moss and lichen had been acidified significantly when compared with barren soils. Thus while substantial pools of these readily utilized carbohydrates were probably present in cryptogam dominated soils, there was little accumulation of organic acids which could account for the acidity of mineral soils typical of the Windmill Islands.

An investigation into the feeding habits of two demersal nototheniids, Trematomus bernacchii and T. centronotus, showed that the most important prey were polychaetes, molluscs and euphausiids for T. bernacchii and polychaetes and amphipods for T. centronotus. Epifaunal (e.g. Barrukia cristata) and tube-dwelling polychaetes (Amphicteis cfr. midas and Amythas membranifera) were common in the diet of both species. Bivalvia including Adamussium colbecki were found in the diet of T. bernacchii. Epifaunal gastropods (Trochidae) were an occasional prey for T. centronotus. Amphipods (mainly Acanthonotozomatidae) and the euphausiid Euphausia frigida were the main crustacean food of T. centronotus and T. bernacchii respectively. Our data suggest a difference in the feeding behaviour of the two nototheniid species, although both appear capable of feeding on common epibenthic invertebrates.

Plate counts were made of bacteria surviving in materials from Shackleton's and Scott's camps from the first decade of this century. Several millions of bacteria per g of material were detected in samples of pony dung and lesser numbers in dried peas, pearl barley, chaff and straw. No coliforms had survived in the dung: apparent positives in the presumptive coliform counts proved to be sporing Bacillus spp. when tested in a confirmatory coliform test. Subsamples of the colonies growing on agar plates all proved to be either Bacillus spp. producing endospores or actinomycetes (Micromonospora spp.) with single spores along the hyphae.

Apatite fission-track data from samples of Precambrian basement, Late Permian Triassic sedimentary rocks and inferred Cretaceous intrusive bodies are used to constrain the low-temperature (i.e. sub ~110°C) thermal history of the northern Prince Charles Mountains, East Antarctica. Two discrete phases of cooling have been identified, both of which are attributed to regional exhumation associated with rifting episodes. A phase of late Palaeozoic cooling, that began during the Carboniferous, is inferred to have been associated with the initial formation of the Lambert Graben. A more recent phase of cooling was initiated during the Early Cretaceous and is estimated to have locally involved the removal of at least 2 km of material using an assumed palaeotemperature gradient of ~25°C km−1 at the time of cooling. This latter phase of exhumation was closely accompanied by the emplacement of a variety of mafic alkaline rocks at ambient palaeotemperatures less than ~60°C and was probably related to renewed extension of the Lambert Graben during the break-up of eastern Gondwana. The results of this study suggest that final exhumation of high-grade Precambrian basement of the northern Price Charles Mountains was largely controlled by Phanerozoic rifting events.

Within the Jones Mountains, which form part of the Thurston Island crustal block, up to 700 m of Miocene (c. 10 Ma) pillow basalt and palagonitized volcaniclastic rocks unconformably overlie Jurassic granitic basement and Cretaceous volcanic rocks and dykes. New geochemical analyses demonstrate the alkalic nature of the basalts, which range in composition from alkali basalt to basanite. Unradiogenic Sr-isotope ratios (0.7031–0.7034), coupled with low LILE/HFSE ratios (e.g. Th/Ta c. 1.4, Rb/Nb 0.3–0.9) indicate a predominantly asthenospheric source for the basalts. The Jones Mountains basalts are geochemically similar to the alkalic basalts of Marie Byrd Land, but have consistently lower K/Ba and higher Ba/Nb ratios than Late Cenozoic alkalic basalts along the Antarctic Peninsula. These regional variations in geochemical composition apparently reflect differences in tectonic setting and are not the result of lithospheric interaction or partial melting/crystallization effects. The generation of alkalic magmas along the Antarctic Peninsula was causally related to the formation of slab windows following ridge crest-trench collision and the cessation of subduction, whereas the Jones Mountains alkalic basalts may represent the expression of the northward propagation of the head of the Marie Byrd Land plume.

At two localities adjacent to Hatherton Glacier, Darwin Mountains, a thick glacio-lacustrine sequence underlies the main diamictite facies of the Darwin Tillite (Metschel Tillite), and at one of these locations a glacio-lacustrine sequence also conformably overlies the diamictite. These deposits record the initial advance and final retreat of the Gondwana ice cap. Other, pro- and periglacial phases within these sequences record several minor advances and retreats of the ice sheet. Carbonaceous sediments above and below the main diamictite phase show that vegetation was established near the ice margin, suggesting that present-day ice margin temperatures are colder than those experienced during the Permo-Carboniferous (Gondwana) glaciation. It appears that there was little or no erosion of the Devonian Taylor Group which underlies the glacial sequence. The apparent erosional nature of the contact between the Permo-Carboniferous glacial sequence and the overlying Permian coal measures is attributed to surface winnowing of unconsolidated tills and locally the units are conformable.

Fossiliferous horizons of the Ferrar Group at Carapace Nunatak of southern Victoria Land have yielded the richest and most diverse freshwater Jurassic biota hitherto recorded from Antarctica. Fossil conchostracans are the most important in terms of number of individuals and distributional area. Scanning electron microscopy is used to establish a new genus and species (Carapacestheria balli), and Cyzicus (Lioestheria) disgregaris Tasch is attributed to Carapacestheria disgregaris (Tasch) emend. The conchostracan fauna of the Ferrar Group, characterized by Carapacestheria, is probably of early Middle Jurassic age.

We examine some aspects of the performance of the United Kingdom Meteorological Office's new climate model over Antarctica. Pressure and temperature fields are presented as a basic check on the model climate. The gradient of pressure between mid-latitudes and high southern latitudes is too great, resulting in an Antarctic trough that is too deep by 4–6 hPa. Temperature is well modelled though the interior is slightly too cold in winter. Precipitation is interesting because of its relevance to mass balance and therefore changes in sea level. The simulation of the pattern of accumulation is good despite somewhat high values at places in the coastal areas, with an areally-averaged value of 182 mm y−1. We also look at the phenomena of the coreless winter and the katabatic winds which are a consequence of the intense radiative cooling. These two effects may provide a useful diagnostic of the model performance.

The near Earth space environment (known as Geospace) is dominated by the interaction between the solar wind and the geomagnetic field, which creates the magnetosphere. Considerable energy flows from the solar wind into the magnetosphere and ends up in the Earth's upper atmosphere (the thermosphere and ionosphere). The coupling of the geomagnetic field with that of the solar wind (known as the interplanetary magnetic field, or IMF) produces a variety of electro-dynamic responses with signatures such as electric fields and currents in the polar ionospheres. These produce, inter alia, motion of the ionospheric plasma (at altitudes between 100 and 1000kms) which can be monitored from the ground using radar techniques. Analysis of such plasma motion provides a very powerful means of investigating the nature of the interactions taking place at the boundaries between the magnetosphere and the solar wind. To do this effectively requires simultaneous measurements over as large an area (in latitude and longitude) as possible.