Introduction
Aims of research
The aim of this research was to investigate the little known sedimentary rocks of the Sperm Bluff Formation, lower Taylor Group (Beacon Supergroup) that record deposition during the onset of Devonian subsidence in the northern McMurdo Sedimentary Basin (Reference BradshawBradshaw in press). The Sperm Bluff succession is unusual within the Taylor Group in that it contains abundant intraformational conglomerates as distinct from basal breccia-conglomerates immediately above the Kukri Erosion Surface. These pebble–boulder conglomerates are known only in the Clare Range, Saint Johns Range and the eastern Olympus Range. An initial aim of this research was to test the hypothesis that the sediments were derived from a southward continuation of the northern Victoria Land terranes lying beneath the Ross Sea (Allibone et al. Reference Allibone, Cox and Smillie1993). Our principal aims were to investigate the lithology and depositional setting of these conglomeratic rocks, the likely provenance of the conglomerates and sandstones, and to redefine the stratigraphic position of the Sperm Bluff Formation within the Taylor Group. We also see our results as helping to resolve the controversy over the depositional environment of the Taylor Group. Previous interpretations have ranged from marine/marginal marine (Bradshaw Reference Bradshaw1981) to non-marine (Woolfe Reference Woolfe1990).
Field area
In the north, the localities are separated by large glaciers, but in the south on Mount Cerberus the land is largely ice-free. Camps were moved by helicopter and travel to outcrops was on foot. The topography of the field area is steep. Outcrops of Sperm Bluff Formation are difficult to access, commonly cliffs, making measurement challenging. The three-dimensional form of cross-beds was, in many instances, difficult to assess.
A total of ten sections were examined and six of these were measured in detail. The sites chosen were based on the abundance of conglomerate as reported by Turnbull et al. (Reference Turnbull, Allibone, Forsyth and Heron1994) based on fieldwork for the Bull Pass-Saint Johns Range geological map.
The Taylor Group
The Taylor Group forms the lower half of the Beacon Supergroup (Devonian–Triassic) and in southern Victoria Land rests unconformably on a Precambrian to early Palaeozoic basement complex (Ross Orogen). The Kukri Erosion Surface at the base of the Taylor Group is widely developed throughout the Transantarctic Mountains. Although relatively level in the Ohio Range, where the surface appears to represent a shore platform trimmed across a weathered landscape prior to burial by Lower Devonian marine sediments (Bradshaw Reference Bradshaw2010, Fig. 4), the surface in southern Victoria Land shows significant relief (up to 30 m) and is buried by a variety of coarse grained lithologies ranging from pebble to large boulders.
The Taylor Group is dominated by quartzose sandstones with subsidiary finer beds and local coarse units. The group contains seven formations (see McKelvey et al. Reference McKelvey, Webb and Kohn1977 for summary) and in the north is divided by the Heimdall Erosion Surface at the base of the Altar Mountain Formation (Fig. 1). Apart from poorly preserved palynomorphs from the Terra Cotta Siltstone Formation that suggest an Early Devonian age (Kyle Reference Kyle1977), and a microflora from near the top of the youngest formation, the Aztec Siltstone, for which Helby & McElroy (Reference Helby and McElroy1969) suggested an early Late Devonian age, the only reliably datable body fossils in the Taylor Group are diverse fossil fish in the Aztec Siltstone. These were initially dated as Late Devonian (Woodward Reference Woodward1921), but subsequent studies indicate a Middle to early Late Devonian age (Young Reference Young1988, Turner & Young Reference Turner and Young1992, Young & Long Reference Young and Long2005). The Late Devonian age of Helby & McElroy (Reference Helby and McElroy1969) is now less certain as the range of the critical taxon (Geminospora lemurata Balme) is Middle to Late Devonian in Australia (Young Reference Young1993). Consequently, the Taylor Group is regarded as entirely Devonian in age (Reference BradshawBradshaw in press). Both basement and Taylor Group rocks were intruded by thick Jurassic-age dolerite sills of the Ferrar Group, with a 40Ar/ 39Ar age of 176 ± 1.8 Ma (Fleming et al. Reference Fleming, Heimann, Foland and Elliot1997).
Subdivision of the Taylor Group in the McMurdo Dry Valleys to Mackay Glacier region with maximum thicknesses (left), and proposed correlation of the Sperm Bluff Formation (right). Significant erosion surfaces also indicated. Early Devonian microfossils have been extracted from the Terra Cotta Siltstone and Mid to Late Devonian fish remains and microfossils from the Aztec Siltstone.
The environment of the Taylor Group has been a contentious issue for some time, with strong support for both non-marine and marine or marginal marine conditions. The lack of any marine body fossils in the lower six formations makes the issue difficult to resolve, although dissolution of calcareous shells followed by compaction that obliterates moulds in such a thick sandstone succession would not be surprising. Non-marine proponents (Plume Reference Plume1982, Woolfe Reference Woolfe1990) use the colour of sediments combined with sedimentary structures, such as desiccation and syneresis cracks and palaeosols, to support their case, although these are also compatible with emergence in marginal marine conditions. Woolfe (Reference Woolfe1990, p. 301) stated that the Taylor Group has many features typical of alluvial plain deposition, including abundant red beds. Several others have referred to the Taylor Group as a red-bed sequence (Plume Reference Plume1982, Sherwood et al. Reference Sherwood, Kirk and Woolfe1988, p. 122, Woolfe & Barrett Reference Woolfe and Barrett1995, p. 13). Convincing red-beds are, however, found only in the highest formation (Aztec Siltstone). The bulk of the Taylor group is pale-yellow coloured sandstone, with subsidiary green, grey or black siltstone and only occasional maroon siltstones. The latter probably owe their colour to the hydrothermal effects of major dolerite intrusions. Large dune sands that Woolfe (Reference Woolfe1990) and Wizevich (Reference Wizevich1997) implied could only have been made under non-marine desert conditions, are equally consistent with migrating dune fields along a sandy coastline, especially during low stands of sea level. Plant remains (Woolfe Reference Woolfe1990) are not restricted to non-marine sediments. Rafts of plant debris seen around the coast of New Zealand after storms can become incorporated into shallow water coastal sediments. They are also the most common fossils in submarine fan deposits of the Southern Alps.
Marine proponents (Bradshaw Reference Bradshaw1981, Gevers & Twomey Reference Gevers and Twomey1982, Savage Reference Savage2005, Gilmer Reference Gilmer2008, O'Toole Reference O'Toole2010) use trace fossils in conjunction with lithology and sedimentary features, together with widespread and significant erosion surfaces that suggest sea level changes controlled sedimentation. In places we would expect coastal emergence and the development of coastal dunes and interdigitation of marine with fluvial sediments.
Previous work on the Sperm Bluff Formation
Gunn & Warren (Reference Gunn and Warren1962) first drew attention to the Sperm Bluff Formation conglomerates on Mount Suess (Fig. 2) where they recorded 20 m of mixed conglomerates and cross-bedded sandstones overlying c. 30 m of pale-yellow, cross-bedded medium-grained sandstones that rested on weathered granite. Gunn & Warren (Reference Gunn and Warren1962) included both the sandstones and overlying conglomerates in their “Basal Beds”, and the overlying feldspathic and quartzose sandstones were named “Lower Arenite”.
Map of the northern McMurdo Dry Valleys to Mackay Glacier region where outcrops of Sperm Bluff Formation are found showing locations referred to in the text. Rectangles 1 and 2 indicate areas that are enlarged in Fig. 3. Coloured areas show exposed rock. GT = Gargoyle Turrets, MAT = Mount Allan Thomson, MC = Mount Cerberus, MJ = Mount Jason, MS = Mount Suess, RG = Ringer Glacier, SB = Sperm Bluff, SP = Sponsors Peak, WV = Wheeler Valley.
A section on Mount Suess was measured by Kohn & McPherson (Reference Kohn and McPherson1973), who described the lower quartzose sediments above the Kukri Erosion Surface as New Mountain Sandstone Formation. They also apparently recognized the Heimdall Erosion Surface below the mixed coarse sandstones and conglomerates which they identified as equivalent to the Odin Arkose Member of the Altar Mountain Formation (Kohn & McPherson Reference Kohn and McPherson1973, McKelvey et al. Reference McKelvey, Webb and Kohn1977). Barrett & Kohn (Reference Barrett and Kohn1975) believed that the New Mountain Sandstone Formation had been completely stripped off the Kukri Erosion Surface at many northern sites, but on Mount Suess had become preserved in a depression below the Heimdall Erosion Surface. They placed this surface at the base of the conglomerates following Kohn & McPherson (Reference Kohn and McPherson1973).
A short distance (15 km) to the south-west of Mount Suess, on what is now called Gargoyle Turrets (Fig. 3a), Gunn & Warren (Reference Gunn and Warren1962, fig. 36) reported different, finer sediments resting on granite. These sediments were later equated with the Terra Cotta Siltstone Formation by McKelvey et al. (Reference McKelvey, Webb and Kohn1977, fig. 5, section MG), and recorded as overlain by 295 m of New Mountain Sandstone Formation.
Maps showing location of measured sections and sampling sites in the Sperm Bluff Formation (green dots). a. Mount Suess, Sperm Bluff and the Saint Johns Range (see rectangle 1 in Fig. 2). The grey background is ice and snow. Letters refer to measured stratigraphic sections. b. Sampling sites on Mount Cerberus (Repeater Ridge) in the eastern Olympus Range (see rectangle 2 in Fig. 2).
A more detailed study of the conglomerates on Mount Suess was undertaken by Turnbull et al. (Reference Turnbull, Allibone, Forsyth and Heron1994), during research for the geological map of the Bull Pass to Saint Johns Range area. Within the mapped area Turnbull et al. (Reference Turnbull, Allibone, Forsyth and Heron1994, p. 21) noted a relief of 77 m on the Kukri Erosion Surface, much more than previously recorded, but no locality or description was given. Turnbull et al. (Reference Turnbull, Allibone, Forsyth and Heron1994) placed the conglomerates within a new unit, the Sperm Bluff Conglomerate, and designated Mount Suess as the type section. Further outcrops were found in the Clare Range, Saint Johns Range, and the eastern Olympus Range. The Sperm Bluff Conglomerate was described as comprised dominantly of well-rounded, clast-supported pebble to cobble, polymict conglomerate, interbedded with and/or overlain by fine to coarse quartzose sandstones. Turnbull et al. (Reference Turnbull, Allibone, Forsyth and Heron1994) regarded the conglomerate as the equivalent of the Beacon Heights Orthoquartzite (Fig. 1). They also defined an informal “Queer member” on Gargoyle Turrets for the finer beds first described by Gunn & Warren (Reference Gunn and Warren1962).
Turnbull et al. (1994) reported that the conglomerates contained large proportions of volcanic, metasedimentary and quartz clasts, with a lesser number of plutonic basement clasts (Turnbull et al. Reference Turnbull, Allibone, Forsyth and Heron1994, fig. 20).
Analytical techniques
Stratigraphic sections were measured where possible using standard techniques. These placed the conglomerate in sedimentological context, built up a picture of the complexity of the depositional systems, and allowed correlation between sites. Six stratigraphic sections were measured: one on Mount Suess (type section), three on Sperm Bluff (B, D & F) and two on Gargoyle Turrets (A & B). Where a stratigraphic section could not be measured, detailed descriptions were made tied to GPS locations
Clast counts were made on conglomerates and pebbly sandstones at several stratigraphic levels within the Sperm Bluff Formation, and at different sites, to determine percentages of clast lithotypes. At least 100 clasts were counted at each site using 10 cm intervals on a 1 m x 1 m grid.
Point counts of thin sections from 22 samples were made using an automated Prior James Swift point counter. The samples varied from sandstone to pebbly sandstone in grain size and included some matrix samples from boulder to cobble conglomerate units. Rhodizonic acid was used to stain for plagioclase and sodium cobalt nitrite to stain for alkali feldspar following standard techniques. The dyes bond to calcium and potassium respectively, staining alkali feldspar yellow and plagioclase red. Often the alteration products (dominantly sericite) of feldspar take up the stain obscuring composition, although relics are commonly present. Three hundred counts were made per sample at a stage interval of 3 mm. Categories included in counts are feldspar (alkali varieties and intergrowths), quartz (straight extinction, undulose extinction), polycrystalline (greater than three grains), lithics (volcanic), matrix (sericite, felsitic), cement (quartz overgrowths) and other subordinate minerals.
Palaeocurrent directions were measured wherever possible. The majority of these were from foresets of abundant cross-beds. Most measurements came from sandstone lithofacies and very few came from sandstones interbedded within the conglomerates themselves. Due to the proximity of the South Magnetic Pole, correction for declination was made using data from the http://www.ngdc.noaa.gov/geomag/geomag.shtml website (accessed March 2004). Each site required a unique correction, calculated from latitude, longitude, elevation and date of record. Rotation of palaeocurrent measurements for subsequent deformation was not necessary because beds are generally flat-lying (median dip 4°W).
Nomenclature
The name Sperm Bluff Formation is here proposed to replace Sperm Bluff Conglomerate as defined by Turnbull et al. (Reference Turnbull, Allibone, Forsyth and Heron1994). Mount Suess remains the type section. We consider the generic name more appropriate as conglomerate never exceeds 30% of the outcrop at any given location and many of the beds mapped as Sperm Bluff Conglomerate by Turnbull et al. (Reference Turnbull, Allibone, Forsyth and Heron1994) contain no actual conglomerate. Sandy lithofacies, including pebbly sandstone, cross-bedded sandstone and siltstone make up the bulk of the formation. Thus, the name Sperm Bluff Conglomerate does not accurately represent the lithologies present. Because of its potential usefulness in regional correlation, the informal “Queer member” (Turnbull et al. Reference Turnbull, Allibone, Forsyth and Heron1994) has been retained and is formalized by designating Gargoyle Turrets as the type section.
We retain the name Sperm Bluff Formation, rather than try to incorporate it into known Taylor Group formations for two reasons. Firstly, the formation is unusual in the number and thickness of conglomerate units and deserves to be recognized. Secondly, correlation with specific formations is tentative and may remain so given the lack of age control and the distance from more typical Taylor Group outcrops.
Basal sediments and Kukri Erosion Surface
The basal contact of the Sperm Bluff Formation with the Kukri Erosion Surface is rarely exposed. On Mount Suess, Sperm Bluff and Mount Cerberus the Kukri Erosion Surface has a rolling rather lumpy aspect, but the actual contact is often buried by scree. On Mount Suess the relief on the surface is up to 20 m and the basement granite below is weathered to a depth of 2 m. On the western face of the mountain the north-westward sloping unconformity is well below the level of the conglomerates, and is obscured by several metres of scree (11 m) below outcrops of quartzose sandstone. Further south-east along this face, where there is a local basement high of a dark dyke-like rock, the Kukri Erosion Surface slopes steeply to the south-east and is mantled by very coarse breccia-conglomerate (Fig. 4). Unfortunately, this locality forms the head of a steep, narrow canyon in the cliff line that was too dangerous a site to collect samples from the breccia-conglomerate. Adjacent to the head of the canyon, beneath the snow, we exposed a section comprising weathered granite overlain by 4 m of sediment (Fig. 5). The lowest unit is a red-brown mudstone layer (70 cm), which is succeeded by 2.6 m of upward-coarsening green and brown, bedded, siltstones and fine-grained sandstones. A 50 cm thick, medium-grained greenish sandstone that contained small blocks of a dark grey muddy siltstone, forms the highest layer, below cross-bedded pale quartzose sandstones that underlie the remainder of the slope. It is not clear whether the large blocks seen from a distance in the breccia-conglomerate are of dark basement rock, or dark muddy Taylor Group sediment, or a combination of both. It is clear, however, that the overlying, pale quartzose sandstone progressively onlaps westwards onto a basement high.
South-east end of the Beacon section on Mount Suess showing a. intraformational conglomerate at the base of the Sperm Bluff Formation (arrows). Lack of safe access to this outcrop prevented us providing an appropriate scale, but in this photograph the breccia is c. 10 m away. The cliffs above and to the left are of Sperm Bluff Formation, with a Jurassic dolerite sill at the top of the section. b. Telephoto of basal conglomerate at this site containing large, angular grey boulders that may be ripped-up finer sediments or local basement. The Kukri Erosion Surface slopes steeply to the right. The boulders are buried by pale, cross-bedded sandstones similar to the New Mountain Sandstone, and also contain Heimdallia chatwini.
Measured type section of the Sperm Bluff Formation on Mount Suess (base at left), with lithofacies and suggested depositional environment shown to right of each column. Bold arrows immediately to right of column indicate current directions, with north at top of page. Site of trace fossils also marked. Coloured circles indicate position of clast counts. Conglomerates and coarse sandstones are found in the middle of the section.
In a limited outcrop on the east side of Repeater Ridge near Mount Cerberus (Fig. 3b) the succession starts with large (up to 2 m x 1 m), rounded and weathered boulders of lineated schist lying on the Kukri Erosion Surface. These are overlain by bedded, dark-brown and green muddy often pyrite-rich sandstones that contain scattered angular quartz clasts (up to 6 cm long), which are followed by more typical white sandstones. Elsewhere, thinly-bedded, dark sandstones and mudstones lay directly on the Kukri Erosion Surface that truncates complexly folded schists.
The basal contact at Gargoyle Turrets is less dramatic. Here the underlying granite is weathered for up to 1.70 m and the basal contact is planar. In section A (Figs 3a & 6) the earliest deposits are green sandstones and ripple cross-laminated dark siltstones. Higher sandstones are pink and contain layers of quartz and feldspar granules. In section B, a green, moderately-sorted sandstone infills a shallow hollow in the erosion surface. This is overlain by a thin, poorly-sorted, stratified layer containing large rounded clasts of the underlying green sandstone (up to 30 cm diameter) as well as large angular pebbles of vein quartz and feldspar that differ from the underlying non-porphyritic granite (Fig. 7).
Measured section (A) of Sperm Bluff Formation on Gargoyle Turrets (base at left) with lithofacies and suggested depositional environment shown to right of each column. Bold arrows immediately to right of column indicate current directions, with north at top of page. Position of trace fossils also marked. The section begins with the interbedded siltstone/sandstone lithofacies resting on granite. Sediments coarsen upwards to pebbly sandstone.
Coarse basal sediments on Gargoyle Turrets, section B. Weathered granite is visible below the overhang at the bottom of the photograph. Scale bar = 10 cm (bottom left).
On the north-western end of Sperm Bluff (section F) cobble to boulder conglomerates of the Sperm Bluff Formation rest directly on granite (Figs 8 & 9).
Coarse boulder conglomerate resting directly on jointed granite (lower arrow), overlain by sandstone and further conglomerate layers. West side of Sperm Bluff (section F). Mount Suess in middle distance with Beacon outcrop below summit sill (upper arrow).
Measured section (F) on Sperm Bluff (base at left) with lithofacies and suggested depositional environment shown to right of each column. Bold arrows immediately to right of column indicate current directions, with north at top of page. Coloured circle indicates position of clast count. The basal beds are boulder-conglomerate resting directly on granite basement. Sandstones below an upper conglomerate contain Heimdallia chatwini and Skolithos linearis. The upper conglomerate is recognized in section D (Fig. 11).
Sedimentary lithofacies
Six lithofacies can be recognized in the Sperm Bluff Formation: conglomerate, pebbly sandstone, cross-bedded sandstone, low-angle cross-stratified sandstone, bioturbated sandstone and interbedded siltstone/sandstone lithofacies. These are described below.
Conglomerate lithofacies
The conglomerate lithofacies comprises massive to bedded, moderate to well indurated, sub to well-rounded, granule to cobble and occasional boulder, clast-supported, polymict conglomerate (Fig. 10). Thin shale and sandstone interbeds are locally present. Clast types include volcanic rocks (predominantly rhyolitic), quartzite and vein quartz clasts (Table I). Minor chert, mudstone, siltstone, sandstone, gneiss, schist, tourmaline metasomatised rock and feldspar crystals also occur. The most abundant clast type, rhyolite, is unknown in outcrop in southern Victoria Land. The conglomerate lithofacies represents 17% of the type section on the south-west face of Mount Suess where the thickest conglomerates occur (8 m bed within 10.50 m thick coarse unit, Fig. 5). The conglomerates fill broad channels c. 63 m above the Kukri Erosion Surface. The main conglomerate horizon thins south-eastwards and direct correlation of individual beds is not possible, even over short distances. On the western side of Sperm Bluff the conglomerate rests directly on granite (section F) and lenses out southwards over 80 m (Figs 3a & 8). Conglomerate is repeated twice higher in the Sperm Bluff F section above thick sandstone intervals (Figs 9 & 11). The conglomerate lithofacies is present in lesser amounts at all other localities visited, except for Gargoyle Turrets. Two thin conglomerates are present at Wheeler Valley, Ringer Glacier and Mount Cerberus (see Fig. 3a & b), but these sections are thin and poorly exposed.
Clast supported conglomerate on Sperm Bluff (section D). Clasts are predominantly of rhyolite varieties and grey quartzite. The chisel indicates a well rounded boulder that was initially thought to be granite, but under microscopic analysis proved to be hydrothermally altered rhyolite.
Measured section D on Sperm Bluff (base at left), with lithofacies and suggested depositional environment shown to right of each column. Bold arrows immediately right of column indicate current directions, with north at top of page. Position of trace fossils also indicated. The coloured circle marks the site of clast count SBD.C1. The contact with basement is obscured by scree and the section starts 65 m higher at the first outcrop up the slope. The section ends against a dolerite sill.
The main clast lithotypes found in the Sperm Bluff Formation. Relative proportions of these lithotypes found at different localities are shown in Fig. 17.
Pebbly sandstone lithofacies
The pebbly sandstone lithofacies consists of pebbles, granules and very coarse sand grains. Pebbles occasionally line foresets or are scattered throughout medium to coarse sandstone beds (Fig. 12). The sandstone beds are generally 50 cm to 8 m in thickness and display decimetre to metre-scale trough cross-bedding. Feldspathic sandstone interbeds without pebbles may be up to 3 m thick. Rare mudstone lenses may be present in higher beds. Pebble composition is comparable to lithotypes in the associated conglomerate lithofacies, primarily rhyolite, quartzite, and vein quartz. The pebbly sandstone lithofacies represents 8% of the Mount Suess type section (Fig. 5) but up to 30% at Gargoyle Turrets (Fig. 6) where it exists in place of conglomerate. It is subordinate at all other sites.
Pebbly sandstone lithofacies showing a. rhyolite and quartzite pebbles lining foresets on Mount Suess. The large grey clast bottom right is rhyolite. b. Pebble stringers of rhyolite and quartzite in sandstone on Gargoyle Turrets. Scale bars are 10 cm long.
Cross-bedded sandstone lithofacies
The cross-bedded sandstone lithofacies is buff, tan or pinkish in colour and comprises medium to very coarse-grained, subfeldspathic to feldspathic sandstones (Fig. 13). Trough (decimetre to metre scale) and tabular (decimetre scale) cross-beds occur, although the latter are rare. Troughs often occur stacked on top of each other, and palaeocurrent measurements show a dominant palaeoflow direction, although rare bimodal palaeoflow directions also occur (Fig. 13c). Sparse trace fossils include Diplichnites gouldi (Gevers), Skolithos ichnosp., Helminthopsis ichnosp., Tigillites ichnosp. and the possible undertracks of indeterminate nestling depressions. The cross-bedded sandstone lithofacies is the dominant lithofacies of the Sperm Bluff Formation representing 40–60% of outcrop at all sites visited.
Cross-bedded sandstone lithofacies. a. Sandstone on Mount Cerberus. Scale bar is 10 cm long. b. Same image with cross-sets highlighted. c. Herringbone cross-lamination in arkosic sandstone on Sperm Bluff section F. Scale bar is 10 cm long.
Low-angle cross-stratified sandstone lithofacies
The low-angle cross-stratified sandstone lithofacies contains predominantly low-angle cross-stratified sandstone beds with lesser massive or parallel-bedded sandstones (Fig. 14a). The sandstones are subfeldspathic, well-sorted, fine to medium-grained with occasional coarse to very coarse-grained, quartz-rich beds. Low-angle stratification is on a decimetre to metre scale with the largest reaching 3 m in thickness. Foresets may show parting lineations. This lithofacies comprises between 20 and 30% of any outcrop.
Bioturbated sandstone lithofacies on Gargoyle Turrets. a. Outcrop of medium-grained sandstones on the most southerly turret. The lower beds show abundant Heimdallia chatwini. Height of face c. 3 m. b. Bedding plane surface at Gargoyle Turrets showing abundant Heimdallia chatwini. Similar densities of this trace fossil are seen elsewhere only in the Windy Gully and New Mountain Sandstones in the McMurdo Dry Valleys to the south. Scale on rock is in cm.
Bioturbated sandstone lithofacies
The bioturbated sandstone lithofacies contains burrowed, massive to thinly parallel-bedded to laminated, fine to medium-grained sandstones. These are moderately sorted, mottled in colour and texture, and reach a bioturbation index ranging from moderate to completely homogenized massive sandstone beds.
This lithofacies contains a variety of trace fossils, although the steepness of the topography and restricted outcrop often limits detailed identification (see Figs 5, 9 & 11 for stratigraphic positions). These fossils were found in situ except where indicated. Trace fossils include Arenicolites ichnosp., Didymaulyponomos rowei Bradshaw, Didymaulyponomos ichnosp., Diplichnites gouldi, Diplocraterion parallelum Torell (in fallen block), Diplocraterion ichnosp., Heimdallia chatwini Bradshaw, Helminthopsis ichnosp., ?Planolites ichnosp., Skolithos linearis Haldeman (in fallen block), Skolithos ichnosp, Tigillites ichnosp. and Zoophycus ichnosp.
The bioturbated sandstone lithofacies is minor and best observed at the type section on Mount Suess where it makes up < 10% of the outcrop (Fig. 5). It is also found on Sperm Bluff in section F (Fig. 9), section D (Fig. 11) and at the east Sperm Bluff sampling site (Fig. 3a), also above the top of the measured sequence at Gargoyle Turrets (on the south turret) (Figs 6 & 14).
Interbedded siltstone/sandstone lithofacies (Queer Member)
The interbedded siltstone/sandstone lithofacies comprises siltstone and sandstone interbedded on a range of scales and constitutes the informal “Queer member” of Turnbull et al. (Reference Turnbull, Allibone, Forsyth and Heron1994) here formalised. At Gargoyle Turrets this lithofacies rests directly on granite (Fig. 15a) and the lowest <1 m of beds are coarse and poorly sorted (see Fig. 7). Near the base of the stratigraphic section (Fig. 6), siltstone dominates with beds up to 70 cm thick interbedded with arkosic sandstone beds 20–30 cm thick. Higher in the section, arkosic sandstone dominates with beds up to 3 m although siltstone beds still occur up to 70 cm in thickness.
Interbedded siltstone/sandstone lithofacies exposed at the base of section B, Gargoyle Turrets. a. Interbedded siltstones and sandstones of the Queer Member rest on massive granite (see Fig. 7 for detail). Figure for scale (arrow) is standing on the Kukri Erosion Surface and is 1.7 m tall. b. Lenticular-bedded fine sand in siltstone layer. Scale in cm. c. Paired mud-drapes on lamina and mud-drapes on foresets in cross-bedded sandstone layers. Scale in cm. d. Irregularly-bedded pink sandstone and grey siltstone beds. Scale bar is 1 metre long.
Sandstone beds are pink, whitish or green (Fig. 15d), medium to coarse-grained, subfeldspathic to feldspathic. The sandstone contains well-rounded vein quartz, angular alkali feldspar and rhyolite granules. Most beds are cross-bedded on a scale of 4–20 cm, and rare examples of both normally and reversely graded beds occur. Siltstone beds range in colour from grey, green-grey, to brown, purplish black and dark purple-brown, and in texture from coarse siltstone to mudstone. The siltstone beds are fissile and appear to form broad troughs between dune crests in the sandstones.
Finely laminated siltstone and sandstone show flaser and lenticular bedding (Fig. 15b), mud drapes (Fig. 15c), pinch and swell morphologies (Fig. 15d), siltstone rip up clasts and syneresis and desiccation cracks, all suggesting shallow water. The syneresis cracks are similar to mudcracks in cross-section, but their random orientation and lack of a continuous pattern in plan view supports a syneresis origin. Small (10 cm x 1 m) channels of interbedded sandstone and siltstone constitute some units.
Attempts were made to extract a microfauna/flora from the mudstones to help date the deposit and better support an interpretation. This was not successful, probably due to the close proximity of a Ferrar Dolerite sill.
At Gargoyle Turrets, the interbedded siltstone/sandstone lithofacies represents nearly 30% of the measured section, of which c. 30% consists of siltstone and c. 70% sandstone. A minor development of this lithofacies is found at the bottom of the section on Mount Suess, and locally in the basal sediments on Mount Cerberus.
Palaeocurrent analysis
Palaeocurrents were measured at all localities, primarily from cross-beds in sandstones and pebbly sandstones. In this paper palaeocurrents are reported as directions towards.
On Mount Suess palaeocurrents were measured in the cross-bedded sandstone, pebbly sandstone and low-angle cross-stratified sandstone lithofacies (Fig. 16, see also Fig. 5). These all indicate a strong unimodal flow broadly to the west and south-west.
Rose diagrams for palaeocurrent data from the Sperm Bluff Formation arranged under location and lithofacies. Bin size is 20°. North is indicated at the top of each circle.
Palaeocurrents at Sperm Bluff are more complex since they were measured from a wider variety of lithofacies including the cross-bedded sandstone, the pebbly sandstone, the low-angle cross-stratified sandstone, and cross-beds within the conglomerate lithofacies (Fig. 16, see also Figs 9 & 11). The cross-bedded sandstone lithofacies has two dominant palaeoflow directions, one toward the west (SW–NW) and the second toward the north-east with a large variation. The low-angle cross-stratified sandstone lithofacies shows a strong foreset slope to the north-east. Cross-beds from pebbly sandstones within the conglomerate lithofacies show a variable palaeoflow direction to the east, south-east, and south-west (Fig. 11). The direction of palaeoflow also appears to be related to the size of cross-beds. Cross-beds < 50 cm indicate flow towards both the north-west and south-east, with minor currents towards the south-west. Cross-beds > 50 cm show a dominant palaeoflow towards the east.
Gargoyle Turrets has a bimodal distribution of current directions with dominant palaeoflow directions to the east and south (Fig. 16, see also Fig. 6). Flow directions relate to the size of the structure measured. Small-scale cross laminations (< 6 cm) in siltstones to medium sandstones show palaeoflow towards the south, east and west. Cross-beds (6–20 cm) in medium to coarse-grained sandstone show a strong current direction to both the north-east and south. Cross-beds (< 20 cm) in coarse sandstone, granule conglomerate and pebbly sandstones, show a dominant current to the east. When broken down by lithofacies, the east directed palaeoflow is primarily found in the pebbly and cross-bedded sandstone lithofacies, whereas the interbedded siltstone/sandstone lithofacies was more variable. The cross-bedded sandstones at this locality show palaeocurrent directions c. 180 degrees opposed forming herringbone cross-beds (Fig. 6).
Wheeler Valley shows dominant palaeoflow to the south although palaeocurrent measurements were limited by outcrop. Similarly, only limited data is available for Mount Cerberus, although the six measurements suggest that in the low-angle cross-stratified sandstone lithofacies palaeoflow was to the south-west and in the cross-bedded sandstone lithofacies palaeoflow was to the north-east (Fig. 16).
All sites have polymodal palaeoflow directions suggesting that a combination of different currents was affecting these sites, some of which were diametrically opposed.
Composition
Clast counts
The Sperm Bluff Formation comprises sandstone and conglomerate with a range of well-rounded pebble to cobble sized clasts, most commonly of rhyolitic volcanic rocks and quartzite. Subsidiary clast lithotypes (listed under “other” in Table I) include gneiss, coarsely crystalline quartz, sandstone (possibly cannibalized Taylor Group), schist, siltstone and feldspar crystals. It is worthy of note that despite the abundance of plutonic clasts reported by Turnbull et al. (Reference Turnbull, Allibone, Forsyth and Heron1994) our detailed investigation yielded few convincing plutonic clasts. The main clast types present are volcanic (predominantly rhyolitic), quartzite and vein quartz (Table I). Subsidiary clasts are listed under “other”.
Rhyolite conglomerate clasts tend to be porphyritic although aphyric clasts do occur. The porphyritic clasts have a range of different groundmass textures, which place them into different categories including high-level intrusives, shallow intrusive/extrusive, tuff and ignimbrite. Aphyric clasts are generally pebble sized or smaller and exceed the proportion of porphyritic clasts in only two counts (both from the pebbly sandstone lithofacies).
The quartzite clasts are white, well-sorted, fine- to medium-grained sandstones and make up a large proportion of the clasts in the Sperm Bluff Formation. In thin section, the majority are fine-grained meta-quartzite, although coarser examples exist. They exhibit undulose extinction and have undergone limited to extensive grain boundary migration suggesting the grain size in some cases is not primary. Several of the quartzite clasts examined contain dark lamina where zircon crystals are concentrated.
Commonly, the largest clasts at any locality (up to 250 mm) are vein quartz clasts with a distinctive coarse rod-like texture. Thin section examination proved all such clasts to be > 90% quartz. These clasts commonly show interlocking crystals that support the view they were formed in a vein setting.
To aid comparison between sites, the proportions of the three clast types are shown according to lithofacies and locality (Fig. 17). The conglomerate lithofacies is subdivided into lower conglomerates (Sperm Bluff F) and upper conglomerates (Mount Suess and Sperm Bluff F & D). Lower conglomerates show a clear decrease in volcanic component southward as quartzite increases. The “other” component also increases slightly to the south. Clasts in the upper conglomerates have consistent proportions of lithotypes throughout the region. While a slight decrease in the volcanic component is apparent towards the south, compensated for by an increase in the “other” lithologies, quartzite remains constant.
Individual clast counts (numbered) from coarse lithologies in the Sperm Bluff Formation arranged by location and lithofacies. Clast counts are shown by given categories. Red = volcanic, yellow = quartzite, green = other.
In the pebbly sandstone lithofacies the relative proportions of clast types is highly variable but is most similar to conglomerates in the more northern localities. This probably reflects the fact that the Gargoyle Turrets locality is closer to the northern sites than to the southern ones and that at this locality the conglomerate lithofacies is replaced with pebbly sandstone lithofacies.
Point counts
Point counts on 22 samples show that sandstones range from quartzose to arkosic in composition (Fig. 18). Cross-bedded and pebbly sandstone lithofacies are dominantly sub-arkose to arkosic in composition whereas low-angle cross-stratified sandstone lithofacies are subarkosic to quartzose sandstones.
Quartz-Feldspar-Lithic fragments (QFL) compositional discrimination diagram of sandstones from the Sperm Bluff Formation based on relative position in the stratigraphic succession.
Quartz grains exhibit both normal and undulose extinction although the latter tends to dominate. Minor amounts of polycrystalline (> 3 sub grains) quartz grains are present in most samples but dominate the coarser grained sediments suggesting that these grains broke up into individual monocrystalline grains in the finer grained rocks. Monocrystalline quartz is present in comparable amounts in every count and are sub-angular to sub-rounded, although it appears that some grains were euhedral in original form suggesting derivation from the same source as the porphyritic rhyolite conglomerate clasts. Quartz never comprises less than 42% of any sandstone examined.
Feldspar grains are sub-angular to sub-rounded. They show all degrees of alteration from partial to complete replacement (mostly by sericite). Only alkali feldspar is preserved in these rocks. The lack of plagioclase is notable although if it had been replaced by sericite in the source area during weathering or metasomatism, this could explain the abundance of sericite matrix. Alkali feldspar is dominantly perthitic orthoclase, although subsolvus microcline and hypersolvus granophyric intergrowths occur in subordinate amounts (up to 10%). Rarely, simple Carlsbad twins are preserved. The alkali feldspars are readily picked out by the sodium carbide-nitrate stain. Altered feldspar comprises between 1 and 45% of the sections with a mean and median about 17%.
Lithics, where present, appear to be felsitic and likely to be partially devitrified rhyolitic volcanic fragments. Rare large volcanic lithics with ragged edges are clearly made up of small rhyolite fragments and like the rhyolite conglomerate clasts, they contain phenocrysts and devitrified patches. Smaller aphyric rhyolite lithics are more pervasive, occurring in almost all thin sections, with quartz and feldspar within the grains in a felsitic texture of a devitrified rhyolite.
Interpretation of depositional setting
The variation in thickness and dominance of different lithofacies in the Sperm Bluff Formation suggests lateral facies changes reflecting a variety of contemporaneous depositional environments along a sandy, low-gradient coastline. Superimposed changes in relative sea level are suggested by the stratigraphic columns (Figs 5, 6, 9 & 11). We interpret the conglomerate lithofacies as forming in fluvial channels on a coastal delta plain. Marginal marine depositional environments of various character are represented by the low-angle cross-stratified sandstone lithofacies and the interbedded siltstone/sandstone lithofacies (Queer Member). Deeper water to shoreface is suggested by the change in sedimentary structures within the cross-bedded sandstone lithofacies and the pebbly sandstone lithofacies. The bioturbated sandstone lithofacies is interpreted as proximal inner shelf indicating deposition in a deeper water setting. Supporting details justifying interpretations are given below.
Because the outcrop of the Taylor Group is a single narrow north–south strip between the uplifted edge of the Transantarctic Mountains and the ice-cap, it is difficult to define the orientation of the successive sedimentary environments.
Wave-dominated delta environment
In the conglomerate lithofacies on Mount Suess and Sperm Bluff, broad channels up to 50–60 m wide are filled with clast-supported conglomerate and are associated with smaller channels (a few metres across) filled with sandstone, and less commonly, siltstone. The conglomerates in the channels are moderately sorted, with sub-angular to well-rounded clasts and a sand-siltstone matrix, all indicating bedload transport in channels. The channels have a strong basal scour with fining and upward-thinning beds and are typical of fluvial channels on a delta plain.
Conglomerates also occur in packages with a lower part of coarsening and upward thickening beds, which then fine and thin upwards. This is typical of bar sequences at the mouth of fluvial channels on a delta front (Figs 5, 9 & 11). These features indicate that there was adequate accommodation to preserve the gradational upwards coarsening succession and suggests an environment that was close to base level. We interpret these conglomerates as deposits of fluvial channels close to shoreline on a delta plain (e.g. Bhattacharya Reference Bhattacharya2010). On Sperm Bluff, two rare palaeocurrent measurements indicate a palaeoflow in these channels towards the east to south-east.
The low-angle cross-stratified sandstone lithofacies comprises well-sorted sandstones with abundant low-angle cross-stratified flaggy beds that contain parting lineations and sparse bioturbation. We interpret this lithofacies as having been deposited in a sandy foreshore setting (Clifton Reference Clifton2006, Plint Reference Plint2010). Low-angle foresets commonly dip toward the north-east, but also to the south-west (Fig. 16).
The associated pebbly sandstone lithofacies is interpreted to have been deposited on the subaqueous delta slope in a shoreface setting. It contains abundant pebbly cross-beds that are often quite large and steep. Where this lithofacies underlies the conglomerate lithofacies, it takes the form of cross-bedded, thickening and coarsening up beds that are typical of mouth-bar deposits on a prograding delta (Figs 5 & 9) (Bhattacharya Reference Bhattacharya2010). Where the pebbly sandstone lithofacies overlies the conglomerate (at Sperm Bluff and Mount Suess) or replaces it (at Gargoyle Turrets) the lithofacies represents deepening water in a shoreface environment. The consistent, mainly westward palaeoflow directions indicate the dominant direction of longshore currents at these localities. There are slight changes in this westward palaeoflow direction both upsection and at laterally correlative localities.
The trough cross-bedded sandstone lithofacies, in which only a single Diplichnites gouldi trackway has been found, is probably upper shoreface at most localities. Trough cross-beds are large and laterally continuous and stack on top of each other to great thickness. This lithofacies is also interpreted as deposited in a shoreface setting with a strong westward longshore current (Bhattacharya Reference Bhattacharya2010).
These four lithofacies are associated with each other across the field area. The coarsening upward succession at Mount Seuss and Sperm Bluff from low-angle cross-stratified sandstones into coarse, trough cross-bedded pebbly sandstone lithofacies, and then into clast-supported conglomerate lithofacies channels is consistent with a wave-dominated delta on a low gradient prograding sandy shoreface (Bhattacharya & Giosan Reference Bhattacharya and Giosan2003, Bhattacharya Reference Bhattacharya2010). Rare transport directions at Sperm Bluff indicate a palaeoflow in the subaerial channels towards the east to south-east, with longshore palaeoflow directions dominantly toward the west (Fig. 16). The action of longshore drift on a wave-dominated delta leads to the development of spits and barrier bars that cause the down-drift deflection of river mouths, especially when fluvial discharge decreases. This creates an asymmetrical delta along a probably curved coastline explaining the variation in palaeocurrent directions (Bhattacharya & Giosan Reference Bhattacharya and Giosan2003).
The restricted occurrence of the conglomerate, pebbly sandstone and cross-bedded sandstone lithofacies at Wheeler Valley, Ringer Glacier and Mount Cerberus (Repeater Ridge) suggests that similar processes were operating to the south, associated with multiple small sandy deltas along a shallow marine shoreline. The decrease in volcanic clasts southwards and the continued presence of quartzite clasts, suggests that southern localities such as Mount Cerberus were supplied by a different river system to that in the north.
Estuary or lagoonal environment
The interbedded siltstone/sandstone lithofacies (Queer Member), found primarily on Gargoyle Turrets, but possibly also in minor form on Mount Suess and Repeater Ridge near Mount Cerberus, comprises cross-bedded sandstone dunes with flaser bedding alternating with siltstone background sedimentation with lenticular bedding (Fig. 15). These structures occur in environments where periods of activity alternate with periods of quiescence, and where both sand and mud are readily available. This is consistent with an estuarine or lagoonal setting (Boyd Reference Boyd2010, Dalrymple Reference Dalrymple2010).
Tidal influence is indicated by lenticular, wavy or flaser bedding in the mudstones, as well as flaser bedding in the interbedded cross-bedded sandstones where foresets have mud-drapes (Dalrymple Reference Dalrymple2010). Flaser bedding commonly forms where wave action and sand deposition alternate with slack-water conditions. Flaser bedding is very rarely preserved in fluvial environments (Martin Reference Martin2000). The measured palaeoflow directions in cross-beds from the interfingering cross-bedded sandstone lithofacies show bi-directional flow forming herringbone cross-bedding strongly suggestive of a tidal influence with deposition occurring where tidal circulation is restricted such as at the tidal inlet channel (Boyd Reference Boyd2010). Syneresis cracks also occur in some of the muddy beds, suggesting substantial salinity changes during sedimentation and an estuarine environment is favoured.
A tidal estuarine or lagoonal setting for the interbedded siltstone/sandstone lithofacies conflicts with the interpretation of Turnbull et al. (Reference Turnbull, Allibone, Forsyth and Heron1994) who favoured a fluvial setting based on what they described as a series of fining upward cycles at Gargoyle Turrets. Our detailed stratigraphic sections record both normally and reversely graded beds, but no fining upward cycles (Fig. 6). Fining upward cycles should constitute a succession of beds that fine and thin upwards, and not one or two isolated beds.
Inner shelf environment
The bioturbated sandstone lithofacies comprises moderately sorted, moderately to heavily bioturbated fine to medium sandstones. Beds are massive and mottled to thinly-bedded. This lithofacies is likely to represent the lower shoreface to proximal inner shelf as it has the richest distribution of trace fossils (Suter Reference Suter2006, MacEachern et al. Reference MacEachern, Pemberton, Gingras and Bann2010), including individual beds dominated by Skolithos linearis, Heimdallia chatwini and Diplocraterion parallelum respectively. The Heimdallia-rich beds on Sperm Bluff (eastern side) also contain rare Zoophycus ichnosp. (the first recorded from the Taylor Group). This ichnotaxon is known to occur in offshore beds of the Devonian Catskill Delta in New York (Miller Reference Miller1991). The interbedding of the bioturbated sandstone lithofacies with the associated deltaic and foreshore lithofacies suggests fluctuation of relative sea level, with the coastline moving landwards for short intervals before the next lowering of base level and influx of coarse material. Interbedded sandstones containing dense Skolithos linearis and others containing deep Diplocraterion parallelum or Heimdallia chatwini would be consistent with marine incursions.
In Antarctica Heimdallia is confined to the lower part of the Taylor Group. Elsewhere it is found only in the Tumblagooda Sandstone (Kalbarri Group) of Western Australia (Trewin & McNamara Reference Trewin and McNamara1995). Although the age of the Tumblagooda Sandstone is poorly constrained, it is thought to be Devonian in age and is over 1000 m thick (Trewin & McNamara Reference Trewin and McNamara1995). Heimdallia, together with Diplichnites is found largely in the middle of the Tumblagooda Sandstone in Facies Association 2. Trewin & McNamara (Reference Trewin and McNamara1995) interpreted the underlying medium to coarse-grained trough cross-bedded sediments (Facies Association 1), which are devoid of trace fossils, as fluvial, forming a broad outwash plane that sloped W–NW (unimodal W–NW currents). Facies Association 2 sediments are interpreted as a mixed development of fluvial sandsheets, over which aeolian dunes migrated across the outwash plane, with flooded hollows between them in which sediments became reworked by Heimdallia-producing animals. Trewin & McNamara (Reference Trewin and McNamara1995) considered it probable that this outwash area was coastal rather than part of an inland drainage system, and that dune transport direction was consistently up the palaeoslope away from the hypothetical coastline. The upper part of the Tumblagooda Formation (Facies Association 4) contains trace fossils such as Skolithos linearis, Diplocraterion and Daedalus, with more varied current directions, and a marine environment with fluvial influence is likely. This interesting Australian succession is in some ways more closely similar to the lower part of the Taylor Group in the McMurdo Dry Valleys area (Windy Gully Sandstone to lower Altar Mountain formations) than to the Sperm Bluff Formation, as it contains no significant conglomerates or pebbly sandstones.
Depositional setting - conclusions
We interpret the depositional setting of the Sperm Bluff Formation as the interaction of a wave dominated delta with a sandy, low-gradient coastline. Sedimentary environments ranged from proximal inner shelf, shoreface, estuary or lagoonal, to fluvial channel on the delta plain. Changes upsection represent responses to sea level change and can be used to correlate isolated outcrops across the Sperm Bluff Formation. The interpretation of estuarine deposits directly on the irregular Kukri Erosion Surface is consistent with the initial drowning of a sediment-starved landscape during sea level rise.
A useful modern analogue is seen in the Canterbury Plains, South Island, New Zealand (Fig. 19). Rivers, such as the Ashley River, drain the Southern Alps depositing gravels in braided river channels on the Canterbury Plains. However, beaches on the wave-dominated coast are uniformly composed of fine sand forming a low-gradient foreshore to shoreface and the closest gravels are c. 1 km upstream. River mouths become deflected by the strong longshore currents, and a spit is created from both aeolian sand-dunes and low gradient foreshore deposits. During major floods, the river cuts a fresh channel through the spit and a new estuary is formed. The coastline is relatively straight, but the numerous rivers discharging onto it create multiple wave-dominated deltas. Local uplift produced rocky shore-platforms that would be similar to the Heimdall and Kukri Erosion Surfaces. Pegasus Bay has a broad, low-gradient distal shoreface to proximal inner shelf environment.
Suggested depositional setting of the Sperm Bluff Formation using the coastline of Pegasus Bay north of Christchurch, New Zealand, as a modern analogue.
Interpretation of provenance
Plotting sandstone composition on a tectonic discrimination diagram (Fig. 20) shows a range from continental interior through transitional continental to basement uplift. The transitional continental block provenance implies that no active volcanism was occurring at the time of deposition. Alkali feldspar types present include perthitic and granophyric intergrowths and are consistent with a granitic source such as the Cambrian Granite Harbour Intrusives. Detrital zircons from the Taylor Group in the Darwin Glacier area to the south contain abundant detrital zircons of Cambrian age (Wysoczanski et al. Reference Wysoczanski, Forsyth and Woolfe2003). Our clast counts (Table I) show that the conglomerates are dominated by volcanics, quartzite, and “others”, with the latter including a high proportion of vein quartz. The presence of granite clasts (Turnbull et al. Reference Turnbull, Allibone, Forsyth and Heron1994) could not be confirmed. U-Pb zircon SHRIMP geochronology of the volcanic clasts (Wysoczanski et al. Reference Wysoczanski, Forsyth and Woolfe2003) gives best constrained ages of 486 ± 7 Ma and 484 ± 8 Ma and shows that the volcanic source is contemporaneous with the younger Dry Valley 2 magmatic suite of the Granite Harbour Intrusives (DV2). Wysoczanski et al. (Reference Wysoczanski, Forsyth and Woolfe2003) were not, however, able to establish a geochemical correlation with DV2. Volcanic rocks of Cambro-Ordovician age are unknown in southern Victoria Land and discussion of the petrology, geochemistry and geochronology of the volcanic rocks and the quartzite clasts is beyond the scope of the present paper and will be presented later.
Tectonic discrimination diagram showing counts for samples collected from the Sperm Bluff Formation at the localities indicated. Key to sample localities: MS = Mount Suess, SB = Sperm Bluff, QM = Gargoyle Turrets, RG = Ringer Glacier, MC = Mount Cerberus.
Correlation with other formations in the Taylor Group
On a regional scale, the Taylor Group represents a prolonged period of relatively stable siliclastic sedimentation during the Devonian. In the southern Victoria Land part of the McMurdo sedimentary basin it appears that subsidence was greatest in the south, where the lowest units are thickest, and that sedimentation spread northwards over a highly irregular land surface. The Windy Gully Sandstone is not found north of Mount Jason but the New Mountain Sandstone certainly extends north to the Barwick Valley and Sponsors Peak (Fig. 2). At these localities the formation contains Heimdallia chatwini and has been truncated by the Heimdall Erosion Surface and overlain by Altar Mountain Formation (Gilmer Reference Gilmer2008, Reference BradshawBradshaw in press). Turnbull et al. (Reference Turnbull, Allibone, Forsyth and Heron1994) considered that the most northerly occurrence of the New Mountain Sandstone Formation was also on Mount Jason. However, Sykes & Pocknall (Reference Sykes and Pocknall1991) recognized an ichnofauna of dense Heimdallia chatwini in 20 m of quartzose sandstone overlying the Kukri Erosion Surface on Mount Allan Thomson, north of the Mackay Glacier (Fig. 2). These sediments were originally identified as Altar Mountain Formation (Sykes & Pocknall Reference Sykes and Pocknall1991), but were later emended to New Mountain Sandstone by Pocknall et al. (Reference Pocknall, Chinn, Sykes and Skinner1994), making this the formation's most northerly known exposure.
Deposition of the Taylor Group was interrupted by the Heimdall Erosion Surface (Fig. 1), that is remarkably extensive and shows very little relief. The erosion surface is overlain by the feldspathic sandstones of the Odin Arkose Member at the base of the Altar Mountain Formation, with a thin basal conglomerate and abundant Skolithos linearis. The Heimdall Erosion Surface cannot be recognized south of the Ferrar Glacier, but its position can be inferred by a marked influx of feldspathic detritus in sections to the south, and the appearance of Skolithos linearis.
North of the Olympus Range in the Balham Valley, the Odin Arkose Member rests directly on Skelton Group metasediments (McKelvey et al. Reference McKelvey, Webb and Kohn1977, fig. 5), and this region became known as the ‘Balham Valley High’ (Bradshaw Reference Bradshaw1981). In the Barwick Valley the Odin Arkose Member oversteps the New Mountain Sandstone Formation to rest on granite gneiss (Gilmer Reference Gilmer2008).
Initially, the abundance of feldspar in parts of the Sperm Bluff Formation led Savage (Reference Savage2005) to correlate this formation with the Odin Arkose Member, because at that time the remainder of the Taylor Group was regarded as feldspar-poor. Later studies to the south of the Sperm Bluff Formation outcrops have demonstrated that at many localities (e.g. Mount Jason) the lower part of the New Mountain Sandstone is both coarse and exceptionally rich in feldspar (Gilmer Reference Gilmer2008, O'Toole Reference O'Toole2010) in some places also containing clasts of basement granite, rhyolite and crystalline rod-like vein quartz.
The Gargoyle Turrets section is the easiest to correlate with the main McMurdo Dry Valleys succession. The section begins with interbedded siltstone and sandstones (Queer Member) that contain structures suggestive of tidal conditions. This lithofacies is very similar to the Terra Cotta Siltstone in the south, and would make Gargoyle Turrets the only locality where this formation rests directly on granite. The nearest known outcrop of Terra Cotta Siltstone is on the north side of Mount Jason, 35 km to the south (see Fig. 2), where the formation is 5 m thick and overlies 30 m of Windy Gully Sandstone. On Mount Suess, at the south-eastern end of the outcrop, four metres of dark silts and fine sandstones below cross-bedded sandstones may also be a thinner record of the Queer Member.
On Mount Suess, the lower part of the section is composed of well-sorted, quartzose sandstones that show low-angle, flaggy cross-bedding comparable in appearance to the New Mountain Sandstone. The similarity is strengthened by the presence of rare Heimdallia, which in Antarctica is known only from the Windy Gully Sandstone and New Mountain Sandstone formations.
At Sperm Bluff section F (Fig. 9) the succession begins with a coarse boulder-conglomerate (8 m) that rests directly on granite. This is followed by pebbly sandstone (4 m) and a second conglomerate (2 m), overlain by a thick succession (29 m) of cross-bedded quartzose sediment, the higher levels of which contain abundant Heimdallia chatwini and Skolithos linearis below two thin higher conglomerates. The presence of unequivocal Heimdallia chatwini suggests that both the sandstones and basal conglomerate should be correlated with the New Mountain Sandstone Formation.
At Sperm Bluff east (Fig. 3a) only the highest exposures of the Sperm Bluff Formation were accessible, but the presence of well-sorted sandstones densely burrowed by Heimdallia chatwini and also containing Zoophycus ichnosp. suggests the whole of this section could be correlated with the New Mountain Sandstone Formation.
On Gargoyle Turrets the interbedded siltstones and sandstones (Queer Member), which are here correlated with the Terra Cotta Siltstone Formation, are succeeded by thick pebbly sandstones, and then, on the southernmost turret, by well-sorted sandstones containing abundant Heimdallia chatwini (Fig. 14). The position of these sandstones stratigraphically above a Terra Cotta Siltstone-like succession, would suggest strongly that they correlate with the New Mountain Sandstone Formation.
On Mount Suess, the quartzose lower sediments pass upward into first pebbly feldspathic sandstones and then into clast-supported conglomerate. The conglomerates in turn grade upwards into cross-bedded feldspathic sandstones that closely resemble the Odin Arkose Member at the base of the Altar Mountain Formation, and then up into well-sorted, medium to fine-grained quartzose sandstones, containing the burrow Beaconites barretti Bradshaw. The presence of Beaconites in these sandstones supports a correlation with the Altar Mountain Formation, as in the McMurdo Dry Valleys region Beaconites first appears just above the top of the Odin Arkose Member. The burrow is unknown in older formations (Heimdallia chatwini was mistakenly identified as Beaconites in the New Mountain Sandstone Formation by Wizevich Reference Wizevich1997). We were unable to find a distinct Heimdall Erosion Surface on Mount Suess although one was reported by Kohn & McPherson (Reference Kohn and McPherson1973). The influx of coarse detritus, however, may equate with the base of the Odin Arkose Formation.
Conclusions
-
• The Sperm Bluff Formation represents initial Taylor Group deposits in the northern part of the intra-cratonic southern Victoria Land sub-basin during the Early Devonian.
-
• The Formation comprises a mixture of conglomerate, sandstone and rare mudstone.
-
• Six lithofacies suggest the development of a wave-dominated sandy delta building southwards into a shallow sea, and the deposition of foreshore, shoreface and inner shelf sandstones with trace fossils.
-
• The Sperm Bluff Formation records rises and falls in relative sea level during its deposition.
-
• Westward flowing longshore currents caused asymmetry in the delta, and near Gargoyle Turrets the growth of a bar or spit across an interdistributary bay, to create a tidal lagoon.
-
• The conglomerates indicate derivation principally from a highland that lay to the north between the basin and the palaeo-Pacific margin.
-
• The northern highland supplied Ross-age rhyolitic material that is at present unknown in outcrop.
-
• We found no evidence to support the hypothesis that the source area was a continuation of the northern Victoria Land terranes beneath the Ross Sea.
-
• The erosion of rhyolite volcanics concurrent with granites of similar age suggests the volcanics may have been preserved by down-faulting during the late stages of the Ross Orogeny.
-
• The highland also supplied cobbles of well-sorted fine to medium grained quartzites, that predate the Ross Orogeny and which may have been derived from the Skelton Group.
-
• The Sperm Bluff Formation is considered equivalent to the Terra Cotta Siltstone, New Mountain Sandstone and Altar Mountain formations of the Taylor Group succession in the McMurdo Dry Valleys succession.
-
• The trace fossils Heimdallia chatwini and Beaconites barretti have assisted this correlation.
Acknowledgements
The authors wish to thank Antarctica New Zealand and Scott Base staff for excellent logistic support in the field. We thank the Ministry of Foreign Affairs & Trade and the University of Canterbury for scholarship funding for the principal author, and a research grant for the expedition. Thanks to mountaineer Duncan Ritchie for his help in the field, to Anekant Wandres for assistance with geochemical and clast analysis, and to John Bradshaw for proof-reading the manuscript and for many Beacon discussions. We also thank Nigel Trewin (Aberdeen) and Morag Hunter (Cambridge) for comments that have significantly improved this paper.
