Fragmented Tasmania: the transition from Rodinia to Gondwana

The origin of the microcontinent VanDieland extends back to the late Paleoproterozoic, where it was positioned between East Antarctica and southwestern Laurentia, within the supercontinent Nuna and Rodinia. Paleo- to Mesoproterozoic events recorded in VanDieland have greater affinities with southwest Laurentia and East Antarctica, suggesting southern VanDieland was part of the Grenville Front, and the central Tasmanian part was adjacent to the Miller Range in the central Transantarctic Mountains. Late in the Neoproterozoic Rodinia break-up, VanDieland separated from East Antarctica and southwestern Laurentia, and moved north along the Terra Australis margin until its southern part was positioned next to the easternmost Robertson Bay Terrane of north Victoria Land. VanDieland comprises up to seven different crustal megaboudins or microcontinental ribbon terranes that likely had amalgamated by the end of the Cambrian; these ribbon terranes are bounded by major faults and suture zones. Some boundaries, such as the Arthur Metamorphic Complex, are well known. However, other boundaries, like the eastern edge of the Tyennan Zone, and the boundary between King Island and northwestern Tasmania, are more cryptic, as they are covered by younger geology or are under water. The boundaries are commonly defined by sedimentary and mafic volcanic infill that has been trapped between the crustal fragments. These rocks have previously been interpreted as allochthonous terranes but are more likely to represent inverted sections of attenuated transitional crust and back arc basin fill that formed along the eastern margin of the Gondwana plate during the Cambrian. This interpretation also provides an explanation for the previous tectonic analysis that suggests that Tasmania's mafic–ultramafic complexes were obducted westward onto older sequences and were subsequently transported southwards as other ribbons collided along the northeastern and western edges of the growing microcontinent, which existed in the overriding plate of a west-dipping subduction zone at the convergent margin between Gondwana and the proto-Pacific plate.


INTRODUCTION
The period from 600 Ma to 500 Ma includes some of the most fundamental changes in Earth history. Not only did animals develop hard skeletons and the 'snowball earth' time finish, but the Rodinian supercontinent had completely dispersed and the Gondwana supercontinent began to amalgamate. Facing the exterior ocean of Gondwana was the Terra Australis Orogen (Cawood 2005), which records the accretion of dismembered fragments and oceanic material at the edge of Gondwana. This paper examines the geological evolution and crustal architecture of western Tasmania. The region contains excellent evidence of some of the last breakup events of the remnants of Rodinia, at about 580 Ma Meffre et al. 2004). Thus, western Tasmania provides an opportunity to examine a fundamental transition in Earth history, from almost 300 Ma of continent destruction to 300 Ma of continent building. Furthermore, the Cambrian orogenesis in Tasmania apparently took place isolated from the rest of Gondwana, and terranes of western Tasmanian were not truly accreted onto the Gondwanan margin until the Middle Devonian (Cayley et al. 2002Moresi et al. 2014). Hence, the region provides opportunity to understand tectonic processes associated with Rodinia break-up and subsequent accretion on the Gondwanan margin during the Cambrian to Devonian stages of the Terra Australis Orogen.
Cayley (2011) coined the term 'VanDieland' for a Proterozoic micro-continent that included western Tasmania, the Selwyn Block (central Lachlan Fold Belt), the South Tasman Rise and the East Tasman Plateau. Van-Dieland extends over 1500 km from north to south ( Figure 1) and has been considered to represent a single, coherent entity (Berry et al. 2007;Cayley 2011;Berry & Bull 2012). Western Tasmania (Figures 1, 2) contains the most complete and accessible geological record of Van-Dieland. This region contains a wide range of pre-Ordovician rocks, including deep-water turbidites, nonmarine and shallow marine clastic deposits, carbonate rocks and cratonic fragments. The oldest rocks were deformed and metamorphosed in the Mesoproterozoic (Chmielowski 2009). Extensive areas of VanDieland contain Neoproterozoic to Cambrian rift assemblages of deep-water metasedimentary rocks, granitic plutons, within-plate basalts or ocean floor basalts, Cambrian boninite-associated maficÀultramafic complexes, eclogite, blueschist, greenschist and amphibolite facies metamorphic rocks (Berry & Crawford 1988;Crawford & Berry 1992;Black et al. 1997;Corbett et al. 2014). Low-tem-peratureÀhigh-pressure metamorphic rocks formed during the middle to late Cambrian Tyennan Orogeny (Meffre et al. 2000;Chmielowski & Berry 2012). Rocks with oceanic affinities have been interpreted as allochthonous sheets emplaced at one or more collision zones associated with the Tyennan Orogeny (e.g. Crawford & Berry 1992;Turner et al. 1998;Meffre et al. 2000;Stacey & Berry 2004;Berry et al. 2005;Berry & Bull 2012, Chmielowski & Berry 2012Seymour et al. 2013). The presence of these obducted maficÀultramafic rocks suggests that discrete continental crustal blocks may have been separated by oceanic or transitional crust (Moore et al. 2013). The Tyennan Orogeny is one example of the earliest phases of Gondwanan orogenesis (Crawford & Berry 1992;Everard et al. 2007;Berry & Bull 2012;Seymour et al. 2013) and appears to be part of a much larger orogenic system that includes the Ross Orogeny in Antarctica and the Delamerian Orogeny on mainland Australia (Boger & Miller 2004;Cawood 2005;Champion et al. 2009;Boger 2011). By the Ordovician, the extensive limestone sheet that stitches the many parts of western Tasmania indicates that the collage of cratonic and oceanic fragments had become a single continental ribbon (Seymour & Calver 1998;Moore et al. 2013), with final consolidation into Gondwana in the Middle Devonian .
In constructing a new tectonic model for VanDieland, we used detailed mapping by Mineral Resources Tasmania (including Jennings 1963;Gee & Legge 1979;Brown 1986;Brown et al. 1989;Corbett 2003;McClenaghan & Vicary 2005;Reed & Vicary 2005;Calver et al. 2006;Everard et al. 2007;Calver 2008;Vicary et al. 2008;and summarised in Data Management Group 2011;Brown et al. 2012;Corbett et al. 2014) and others (e.g. Powell & Baillie 1992;Exon et al. 1997b;Berry & Gray 2001;Hall 2001;Holm & Berry 2002;Noll & Hall 2005), and combined these data with airborne magnetic data and the ground gravity data collated by Geoscience Australia. Where appropriate, we supplemented the data with interpretations of the TASGO seismic survey that circumnavigated Tasmania (Hill & Webber 1995;Drummond et al. 2000). This approach allowed a reliable connection of the many geological observations along the coast with the sparse inland data. The geophysical interpretation was grounded in geological observations, while the geology could more confidently be extrapolated beyond the immediate outcrop areas.
We propose a tectonic model that defines each of the 'microplate' boundaries and provides a geological history for their amalgamation. We resolve major Cambrian structural and stratigraphic breaks and define their spatial extent, geometry and role in the evolution of an amalgamated VanDieland. We submit that, prior to the Tyennan Orogeny, western Tasmania was not a single cratonic fragment, but rather several microplates that had previously been extended to form megaboudins (ribbons) linked by transitional oceanic crust, which lay in an oceanic setting similar to the present southwest Pacific (Crawford et al. 2003b;Metcalfe 2011). In modern settings, these fragments can be full-or part-thickness cratonic fragments, as well as parts of fore-arc, back-arc, oceanic or turbidite packages (Cawood et al. 2009). This contrasts with the Ordovician to Devonian evolution of the Tasmanides, which was dominated by accretionary processes and resulted in turbidite-dominated successions that are thrust-repeated with trivial amounts of oceanic crust (Gray & Foster 1998), a setting that typifies parts of the modern northwest Pacific (Cawood et al. 2009). Our proposed model offers tectonic solutions for longstanding paradoxes such as the apparent change in Cambrian subduction polarity along the Gondwanan margin (cf. Crawford & Berry 1992;Miller et al. 2005;Foden et al. 2006), as well as an opportunity to assess and test different tectonic models, including those involving rapid changes of subduction polarity (Crawford & Berry 1992;Crawford et al. 2003b), oblique subduction on an outboard fragment (Cawood 2005), and complex tectonic scenarios where contemporaneous east-and west-dipping subduction zones were present in the early Cambrian and VanDieland was sucked obliquely into Gondwana by the west-dipping subduction zone in the Late Cambrian before rotating and moving away from Gondwana during the Early Ordovician (Cayley 2011). Finally, we place VanDieland in the context of Precambrian continental reconstructions.

REGIONAL GEOLOGY
Tasmania is typically separated into two regions. Western Tasmania comprises Mesoproterozoic and Neoproterozoic cratonic crust, whereas eastern Tasmania has oceanic affinities and was cratonised in the Middle Devonian Tabberabberan Orogeny (Figures 1, 2) (e.g. Reed et al. 2002;Black et al. 2004;Berry & Bull 2012;Seymour et al. 2013). The boundary between eastern and western Tasmania is believed to be present beneath the Tamar Graben ( Figure 1b) (Seymour et al. 2013), just west of which is seen a Lower Devonian sequence that has affinities with both eastern and western Tasmanian rocks (Rickards et al. 2002). Further south, the boundary is covered by Permian and younger rocks. Offshore seismic reflection data from eastern Tasmania indicated the presence of Proterozoic crust at depth along the southern half of the Tasmanian east coast (Drummond et al. 2000). In contrast, Seymour et al. (2013) suggested that eastern Tasmania is not floored by Proterozoic crust and placed the boundary further west. Eastern Tasmania has been widely considered as a single block or 'element' (Burrett & Martin 1989;Seymour & Calver 1998;Corbett et al. 2014).
The following section first outlines each of the zones that comprise western Tasmania and the South Tasman Rise. Figure 2 shows the pre-Silurian stratigraphic columns for each zone. Some zones replicate the "Elements" of Seymour & Calver (1995, 1998, but we have chosen to vary or subdivide others. In order to avoid confusion, we have called our divisions "Zones" and only used the term "Element" where it refers to areas defined by them.

King Island Zone
King Island largely comprises Mesoproterozoic metasedimentary rocks with detrital zircon populations older than 1350 § 90 Ma (Black et al. 2004). In the west, the rocks underwent regional metamorphism to greenschist or amphibolite facies at 1287 § 18 Ma (U/Th/Pb on monazite; Berry et al. 2005). On the west and north coasts, later metamorphism, deformation and granite intrusion at 760 § 12 Ma is coeval with ca 750À780 Ma granites (Black et al. 1997;Turner et al. 1998;Calver et al. 2013b). On the east coast, metasedimentary successions are unconformably overlain by Marinoan glaciomarine deposits with detrital zircon dates of ca 636 Ma (chemically abraded TIMS 238 U/ 206 Pb zircon; Calver et al. (2013a). This sequence is successively overlain by a cap carbonate, laminated black shale and the mafic volcanic (rift tholeiites) and intrusive rocks that form the eastern edge of the zone. This package of rocks is magnetic and relatively dense and is evident in regional geophysical data as a 35 km-wide zone (Figures 1, 3À5) (2004) suggested the sequence was fault repeated, and proposed a pre-repetition width of several tens to hundreds of kilometres of rift tholeiites. Based on stratigraphic and geochemical considerations, a volcanic passive margin sequence setting was interpreted. Sm/Nd analyses suggested the tholeiites were emplaced at ca 580 Ma (Meffre et al. 2004). A T DM model age of 1700 Ma was obtained from the most evolved samples (Ɛ Nd À3.1 at 579 Ma; Meffre et al. 2004). Mafic dykes in the sequence yielded a UÀPb SHRIMP date of 575 § 3 Ma .
The Braddon River Fault (Figures 1, 3À6) is prominent in regional magnetic images of offshore western Tasmania. It is exposed onshore at the Braddon River headwaters, where late movement has displaced early Paleozoic rocks (Baillie & Corbett 1985). Further south, it is defined by a 2 km-wide mylonite at the eastern edge of a Neoproterozoic mafic volcanic rock package (Brown 2011). The Braddon River Fault is imaged in seismic reflection data as steeply west-dipping and has a significant change in seismic character across it (Figure 7). South of King Island, it appears to have a displacement of approximately 70 km. However, further south much of this movement appears to be taken up in internal faulting within the Burnie Zone ( Figure 8).

Rocky Cape Zone
The Rocky Cape Zone occurs in the northwest corner of Tasmania and the northern Sorell Peninsula (Figures 1À4,6,8,9). The oldest rocks are the Mesoproterozoic Rocky Cape Group, which comprises alternating sequences of marginal marine quartzite and shelf siltstone ( Figure 9; Table 1; Everard et al. 2007;Halpin et al. 2014). The presence of the fossil Horodyskia williamsii suggests sedimentation at the basal part of the package occurred between 1400 and 1100 Ma (Calver et al. 2010). Authigenic monazite from quartzites yield dates that cluster in three populations, ca 1360À1290 Ma, ca 1280À1240 Ma and ca 1090 Ma (Halpin et al. 2014). Cambrian and Devonian granites that intrude the Rocky Cape Group contain slightly different zircon inheritance patterns, notably an excess population of 1650À1600 Ma ages that are poorly represented in the Rocky Cape Group (Black et al. 2010). These are likely to be sourced from the cratonic basement on which the Rocky Cape Group was deposited (Black et al. 2010).
Within the Smithton Basin, the Togari Group unconformably overlies the Rocky Cape Group (Everard et al. 2007). It includes conglomerate, dolomite, chert, diamictite, volcaniclastic rocks, siliceous metasedimentary rocks and tholeiitic basalt, which were deposited in an Sr/ 86 Sr ratios (Calver 1998) from the Togari Group and correlations with similar succession of the Adelaide Fold Belt suggest the lower Togari Group may be mid-Cryogenian (Everard et al. 2007). A rhyodacite in a rift tholeiite package in the upper Togari Group yielded an UÀPb zircon SHRIMP age of 582 § 4 Ma , while the uppermost siltstone unit contains early to middle Cambrian fossils (Everard et al. 2007). An upper Cambrian marginal marine to terrestrial clastic sequence overlies the Togari Group. This package contains serpentinite-rich detritus from maficÀultramafic rocks, similar to those that were obducted onto the adjacent Burnie Zone during the Cambrian Tyennan Orogeny (Everard et al. 2007).
The Rocky Cape and Togari groups are also present on the Sorell Peninsula ( Figure 8; Corbett 2003). We suggest that the Sorell Peninsula rocks have been displaced from the rest of the Rocky Cape Zone by approximately 50 km of sinistral movement on the Braddon River Fault ( Figure 11).

Burnie Zone
The oldest rocks of the Burnie Zone are the clastic metaturbidites of the Burnie (Oonah) Formation (Figures 10À13). Similar detrital zircon populations to the shallow water Rocky Cape Group can be interpreted as the two being coeval (Black et al. 2004). More likely, the Burnie Formation represents a younger package derived from the Rocky Cape Group or similarly sourced rocks. The minimum age for the Burnie Formation is ca 710 Ma based on a KÀAr biotite date in an alkaline dolerite that was intruded into wet sediments (Figure 13b;McDougall & Leggo 1965, recalculated in Black et al. 2004, suggesting the Burnie Formation correlates with the basal successions of the Togari Group in the Rocky Cape Zone (Calver 1998).
Further south, the Burnie Formation is overlain by a shallow water rift sequence of the Success Creek Group, and the deep-water turbidites and tholeiitic basalt, chert and mudstone of the Crimson Creek Formation (Brown 1986). The Success Creek Group has been correlated with ca 700 Ma succession in the Smithton Basin based on similarities in the stromatolites present (Brown 1986). The Crimson Creek Formation is considered to have been deposited on a rifting margin (Brown 1986), with the basalts interpreted as seaward-dipping reflectors (Direen & Crawford 2003).
Boninitic maficÀultramafic complexes within the Burnie Zone include duniteÀharzburgite, layered pyrox-eniteÀdunite or layered peridotiteÀpyroxeniteÀgabbro  In the upper Cambrian, the Owen Conglomerate and associated rocks were deposited in half graben (Noll & Hall 2005). By the Middle Ordovician, sedimentation had evolved from coarse clastics to fine clastics and then to the micritic dolomitic limestone of the Gordon Group (Seymour et al. 2013).

Pedder Zone
The Pedder Zone and the Tyennan Zone form the Tyennan Element of Seymour & Calver (1998) (Figures 1,  2, 14À16). The Pedder Zone is differentiated from the Tyennan Zone by the presence of the eclogite facies rocks of the Franklin Metamorphic Complex (Figure 15a). Metamorphic analysis of these eclogites yielded pressures between 1400 and 1960 MPa (depths of 40À60 km) and temperatures of~550À650 C (Chmielowski & Berry 2012), and so are likely to have formed at a boundary between cratonic blocks rather than within the interior of a single block. Calver et al. (2006) interpreted the oldest rocks in the Pedder Zone as clast-bearing proximal turbidites. A mylonite within this package yielded a typical 1900 to 1400 Ma detrital zircon population, but some had metamorphic overgrowths that yielded a date of 1220 § 36 Ma (Chmielowski 2009), suggesting the metaturbidites were either metamorphosed at 1220 Ma, or derived from rocks that had been metamorphosed then. The shallow marine to subaerial succession of the Clark Group overlies these metaturbidites, and the similarities in lithologies and detrital zircon populations suggest correlation with the Rocky Cape Group (Black et al. 2004;Calver et al. 2006). Several samples contained a metamorphic monazite population of 1367 § 7 Ma (Chmielowski 2009). Metaquartzite, dolomite and diamictite are in fault contact with, and are interpreted to overlie the Clark Group. These rocks are correlated with the Togari Group (Calver et al. 2006).
The siliceous Wings Sandstone, in the eastern part of the Pedder Zone, is characterised by an unusual detrital zircon population with dominant populations between ca 1400 and ca 900 Ma, with the youngest population 914 § 44 Ma (Black et al. 2004). The Ediacaran to Cambrian lithic volcanic metasediments, chert and minor basaltic tuff of the Ragged Basin, and slices of ultramafic rocks occur adjacent to the Wings Sandstone (Calver et al. 2006). There are no direct age controls on these east-dipping ultramafic slices; however, Crawford & Berry (1992) suggested they were obducted at the same time as other Tasmanian maficÀultramafic complexes at ca 520 to 515 Ma. Unconformably overlying all the older sequences are stratigraphic equivalents to the Owen Conglomerate, which are conformably overlain by Ordovician carbonate successions. Tilt filtered total magnetic intensity. (c) Gravity, isostatic onshore, free air offshore. The Rocky Cape-Burnie and Burnie-Pedder zone boundaries have been largely determined by the outcropping and inferred Precambrian boundaries. The former is interpreted as an east-dipping thrust fault, which we interpret to lie in front (west) of a much larger fault marked by a maficÀultramaficÀboninite package at the western edge of a Cambrian felsic volcanic sequence. We interpret this fault system as an east-dipping, leading imbricate fan (Boyer & Elliott 1982). In the magnetic and seismic images, these major boundary systems can be traced south from the Braddon River Fault until they are truncated by the southern extension of the eastern boundary of the King Island Zone (Figures 3, 10). Although the maficÀultramaficÀboninite package has been subsequently disrupted by the Devonian Tabberabberan deformation, we consider that this western boundary of the Burnie Zone on the Sorell Peninsula is more or less in place. In the north, the Braddon River Fault has a sinistral displacement of up to 70 km. However, the eastern boundary of the Burnie Zone is only displaced by approximately 5 km, and we suggest that most of the missing movement has been taken up along the sheared eastern boundary of a Neoproterozoic mafic volcanic package within the Burnie Zone. These volcanic rocks can be seen as both magnetic and gravity highs.
The oldest Tyennan ages of metamorphic monazites in the Franklin Metamorphic Complex are 529 § 10 Ma from a muscoviteÀquartzÀgarnetÀplagioclase schist, and the youngest 505 § 7 Ma from a quartzÀmuscoviteÀ-plagioclaseÀgarnetÀbiotite schist; most zircon and monazite ages cluster at approximately 510 Ma (Chmielowski & Berry 2012;Fergusson et al. 2013). On the southwest coast near Nye Bay (Figure 12b), the metamorphism is slightly younger, with monazite ages clustering at 505 § 2 Ma (Chmielowski & Berry 2012).

Tyennan Zone
The oldest mapped rocks in the Tyennan Zone are marginal marine metaquartzite and garnet schist, with a maximum deposition age of 1574 § 59 Ma based on the youngest detrital zircon population (Black et al. 2004). The basement to this package is unknown (Calver et al. 2006;Black et al. 2010). Metamorphic conditions for these rocks reached 1960 MPa and 545 C at 508 § 9 Ma (U/Th/ Pb monazite age; Chmielowski 2009) in a region near our interpreted boundary with the Pedder Zone. Drilling in central Tasmania intersected dolomite that was correlated with dolomite in the Togari Group of the Rocky Cape Zone ( (2001) suggested the entire sequence was inverted and allochthonous. We are not persuaded by this interpretation. Their mapping shows the metamorphic rocks in faulted contact with weakly metamorphosed packages, implying that the metamorphism took place before the faulting. If there were inversion, the sequence most likely came from the west. The Arthur Metamorphic Complex, at the western edge of the Burnie Zone, is characterised by temperatures and pressures of 350 C and 700 MPa (Turner & Bottrill 2001), which are significantly lower than those of the Forth Metamorphic Complex, suggesting they are not equivalents; other alternatives seem equally unlikely.
The Fossey Mountain Trough lies between the northern, coastal outcrops and the main Tyennan Zone in central Tasmania. This rift is filled with Mount Read Volcanics equivalents that are overlain by Owen Conglomerate and equivalent sequences. These are overlain by the Ordovician Gordon Group (Data Management Group 2011). North of the rift, there appears to be a broad facies change in the basement rocks, with the original sediments apparently less quartzite-rich. However, metaquartzite schist and dolomite are present in some of the north-easternmost exposures (McClenaghan & Vicary 2005), and these can be correlated with rocks to the south of the rift. The isostatic gravity data ( Figure 4) suggest that meridional pre-rift structures can be traced north under the rift. Elsewhere, drilling near Hobart intersected high-Al basalt that has similar geochemistry and petrology to basalt in the Mount Read Volcanics (Crawford & Berry 1992).

SorellÀBadger Head Zone
Metaturbidites at Badger Head, in the eastern part of this zone (Figures 14, 15b, Figure 13d). Over the 10 km east from the town of Penguin, the major faults generally dip to the west-northwest at about 70 . We infer that the many west-dipping faults to the east of are antithetic back-thrusts to the Penguin master fault. The deformation is older than the Lower to Middle Ordovician Moina Sandstone, which unconformably overlies many of the above lithologies. The eastern arm of the gravity low marks the Fossey Mountain Trough, a Cambrian feature largely comprising Mount Read Volcanics. Boxes i and ii show the locations of Figure 12 (a, b), where the boundary is detailed further.
implying syn-or post-Grenville Orogeny deposition. The rocks have been correlated with the Burnie Formation turbidites of the Burnie and Tyennan zones (Gee & Legge 1979). They are metamorphosed to lower greenschist facies, and locally contain retrogressed garnet (Reed et al. 2002).
To the west of these metaturbidites are mudstone, dolostone, volcaniclastic sandstone, conglomerate, chert, dolerite, rift tholeiite and rare rhyolite. These rocks are also associated with bedding-parallel broken formation (Calver & Reed 2001). Microfossil-bearing chert indicates deposition during the Cryogenian, which is supported by late Neoproterozoic d 13 C values from these rocks (Calver & Reed 2001). The sequence may correlate with the Crimson Creek Formation of the Burnie Zone (Reed et al. 2002).
The Andersons Creek Ultramafic Complex lies immediately east of the metaturbidites at Badger Head (Figure 15b). Magnetic and gravity modelling suggested the complex lay in the hangingwalls of east and west-vergent thrusts. The west-dipping thrust shallows beneath the Badger Head rocks, whereas the east-dipping thrust extends beneath Cambrian to Devonian platform sequences (Zengerer 1999). Highly deformed granite in the apex of the ultramafic complex yielded UÀPb SHRIMP dates of 658 § 5 Ma and 661 § 8 Ma (Black 2007,  simple facies change from shallow-water succession in the west to deep-water succession in the east. This model does not account for the presence of the 658 § 5 Ma granite in the footwalls of both thrusts; nor is it easily reconciled with the geology of the other tectonic zones throughout Tasmania and the regional west-dipping subduction system outboard of Tasmania (Cawood 2005;Squire & Wilson 2005;Stump et al. 2006). We accept the interpretation that the Badger Head metaturbidite package is allochthonous (Reed et al. 2002) and has been thrust eastwards above the ultramafic rocks. However, we prefer the interpretations by Powell & Baillie (1992) or Patison et al. (2001) that infer the presence of concealed Precambrian cratonic crust east of the Andersons Creek Ultramafic Complex as a basement to the Cambrian and Ordovician shelf sequences there, and consider that the eastern boundary of the cratonic crust lies under the Tamar (1997) and Norvick & Smith (2001) indicated that the East Tasman Plateau was rifted from the east South Tasman Rise in the Late Cretaceous. In order to avoid confusion with their terminology, we have grouped a slightly redefined east South Tasman Rise and the East Tasman Plateau into the Glomar Zone, after the ship that recovered the first pre-Mesozoic rocks in the region. Exon et al. (1995) briefly described the results from four cruises in the region, three of which had recovered pre-Mesozoic samples from 42 dredge sites or drill holes. Subsequent age dating on 14 samples has allowed initial conclusions to be drawn about the South Tasman Rise.
The east South Tasman Rise is topographically smoother, and the bathymetry is~1000 m. Seismic interpretations indicate the presence of up to 2 s TWT (~3 km) thickness of Cretaceous and Cenozoic sedimentary rocks (Exon et al. 1997b). Basement is interpreted to have been attenuated during Gondwana break-up; however, the present-day expression is consistent with separation during the Tasman rifting event. The free air gravity response has a strong linear pattern that trends at about 120 , although along its western edge, the gravity anomalies trend parallel to the boundary with the west South Tasman Rise.  (Figure 1c). This fault has generally been placed along mapped faults; it also lies east of all of the ultramafic rocks in the Pedder Zone, placing it east of the Wings Subzone that has been inferred to have been emplaced westwards (Crawford & Berry 1992). It is also interpreted to lie adjacent to the only eclogite facies metamorphic rocks of western Tasmania (Figures 14 box a, 15a).
Cores in zircons from deformed granites yielded dates of 1051 § 8 Ma and 1042 § 35 Ma, and metamorphic monazites gave dates between 1015 § 24 Ma and 920 § 7 Ma (Berry et al. 2008). An undeformed syenite in the southwestern Glomar Zone yields a UÀPb SHRIMP date of 1119 § 9 Ma and a T Nd DM age of 1270 Ma. Originally this location had been considered as part of the west South Tasman Rise, but the bathymetric, gravity and geological data (Figure 18) from the area have more in common with other parts of the east South Tasman Rise. All of these ages are unlike any from mainland Tasmania. As a result, we have chosen to regard the Glomar Zone as a separate geological entity.
Samples from the East Tasman Rise include granite gneiss, quartzite and marble with blue amphibole, all rocks unknown in eastern Tasmania, but the first two are also present in western Tasmania. No age date data are available.
The west South Tasman Rise is bathymetrically rough and typically lies in deeper water, mostly between 2000 and 4000 m. It also has a 200 mm À2 lower gravity response. In the southern part, linear gravity anomalies generally trend approximately 010 , but in the north,

Tasmanian issues
Despite the apparent structural, metamorphic and stratigraphic complexity outlined above, there are many geological features that are shared across many of the tectonic zones (Figure 2). For example, the oldest exposed rocks are generally quartzites or quartzite-rich sequences with a minor but persistent detrital zircon population at 1450 to 1380 Ma. These are generally marginal marine facies, suggesting they were deposited on an older cryptic basement. Neoproterozoic extension occurred between ca 780 Ma and 580 Ma and is characterised by rifting, sedimentation and intermittent granite production. During this period western Tasmania was disaggregated into smaller crustal blocks separated by MORB and maficÀultramafic complexes. The rifted blocks were deformed in the Tyennan Orogeny. Early west-or southdirected movement was accompanied by regional metamorphism, including blueschist facies in the Arthur Metamorphic Complex and eclogite facies along the extension to the Mt Hobhouse Fault (Figures 14, 15a, 16). The Mount Read Volcanics and the Owen Conglomerate were deposited during the waning stages of the Tyennan Orogeny across several tectonic zones, suggesting western Tasmania had reconfigured into a single crustal block. Subsequent deformation in the Tabberabberan Orogeny modified the boundaries and internal structures, but preexisting structures can still be recognised.
Despite these similarities, there remain other observations that need to be addressed in order to develop a comprehensive tectonic model of VanDieland. These are discussed below.

WHAT DO THE BOUNDARIES REFLECT?
Between the tectonic zones, most of the boundaries are defined by fault zones. However, several boundaries, such as the Tamar 'Lineament' and the Arthur Metamorphic Complex, are more distributed fault networks. Several of the boundaries may separate tectonic zones that are completely allochthonous with respect to their neighbouring zones (e.g. King Island Zone). Other boundaries may represent the inverted remnants of variably thinned crust reflecting that, at the end of the Proterozoic, western Tasmania was a series of crustal scale megaboudins (e.g. P eron-Pinvidic & Manatschal 2010; Direen et al. 2012).
The diversity in character and length of these boundaries suggests that different processes were active at the same time along the boundaries. For example, the high-pressure (to 1.9 GPa) metamorphism seen in the northern part of the PedderÀTyennan boundary is consistent with cratonÀcraton collision. By contrast, the southeastern segment of this boundary is characterised by lower metamorphic grade Proterozoic rocks (Calver et al. 2006) that do not indicate collisional tectonic processes. We suggest the Neoproterozoic to early Cambrian rifting of the Pedder and Tyennan zones resulted in ocean crust formation in the northern segment of the fault zone, while the southern zone was characterised by thin continental crust. We suggest a similar setting along the Rocky CapeÀBurnie zone boundary, where the Arthur Complex has attained blueschist facies metamorphic grades in the northern part (Turner & Bottrill 2001), but the equivalent boundary to the southern is defined by maficÀultramafic units that underwent lower grade metamorphism (Corbett 2003).  Approximately 7 km offshore from the west Tasmanian coast, the Arthur Metamorphic Complex is displaced almost 50 km south (sinistrally) by the Braddon River Fault (Figures 1, 8). On the Sorell Peninsula, the boundary between the Rocky Cape and Burnie zones is an east-dipping thrust fault (Figure 8).
The dominant Cambrian movement sense in the Arthur Metamorphic Complex is of southwest-directed (sinistral) movement, with a later west-verging thrusting event (Holm & Berry 2002). Both events took place after the obduction of the maficÀultramafic complexes at about 516 Ma. We interpret the Arthur Complex as crust that was extended at ca 770 Ma and then shortened in the Tyennan Orogeny and further modified in the Tabberabberan Orogeny. An implication of this is that the western part of the Burnie Zone may be underlain by equivalents to the older rocks in the Rocky Cape Block.

ORIGIN OF THE MAFICÀULTRAMAFIC COMPLEXES
The understanding that VanDieland was a completely separate crustal fragment in the Cambrian (Cayley et al. Crawford & Berry (1992) considered that they came from an unseen but speculated Cambrian arc system that lay to the east of VanDieland. In this interpretation, maficÀultramafic units were obducted at least 130 km westwards over the Tyennan and Pedder zones while leaving no evidence for their occurrence to the east, closer to their interpreted origin. We pose an alternative model, in which the mafic and ultramafic rocks with oceanic and transitional crust affinities were formed between micro-continental fragments that separated during 250 Ma of Neoproterozoic extension. In the Tyennan Orogeny, these basins were inverted and the individual ribbons re-amalgamated to produce the geometries and lithological distributions in western Tasmania. The model does not require an arc system as proposed by Crawford & Berry (1992). An east-dipping maficÀultramafic complex lies close to the western edge of Seymour & Calver's (1995) AdamsfieldÀJubilee Element (Figure 1a). Mapping implies that another significant boundary lies west of the maficÀultramafic complex (Calver et al. 2006). We interpret the maficÀultramafic rocks and the rocks either side to have been within a leading-edge thrust system, with the master fault being the Mt Hobhouse Fault (Moore et al. 2012b). While it is somewhat arbitrary as to whether one considers the AdamsfieldÀJubilee area part of the Tyennan Zone or the Pedder Zone, we have chosen the latter in order to emphasise the significance of the Mt Hobhouse Fault.

MOUNT READ VOLCANICS
The Mount Read Volcanics have been interpreted as a volcanic arc that formed above a west-dipping subduction zone during the late Cambrian (Crawford & Berry 1992). In this model, a transient, west-dipping subduction zone formed to the east of the 'Tamar Lineament.' However, there seems to be little independent evidence for this subduction zone. Recent geochronology results by McNeill et al. (2012) from the Mount Read Volcanics demonstrate that they were formed over 12 m.y., apparently longer than the "short-lived" event envisioned by Crawford & Berry (1992), and potentially questioning the validity of their interpretation. Meffre et al. (2004) described an east-dipping sequence at least 15 km thick of a rift-related tholeiitic volcanic passive margin sequence that crops out on King Island. This magnetic sequence can be traced southwards approximately 50 km to the west of the Mount Read Volcanics (Figures 1, 3). Inversion of this passive margin is likely to have resulted in the amalgamation of the King Island Zone with the rest of western Tasmania, and as this took place, the Mount Read Volcanics formed as the east-dipping subducting slab delaminated. We consider that this is a more satisfactory solution to the origin of the Mount Read Volcanics than that proposed by Crawford & Berry (1992).

WICKHAM OROGENY
The Tasmanian Cryogenian and Ediacaran geological record is dominated by events that can be interpreted as extensional (Figure 2 (2012) appears to be equally unlikely, since it places the Glomar Zone, with granites with ages of 1050 Ma and metamorphic rocks at 920 Ma, adjacent to King Island, with metamorphism at 1290 Ma and granites at approximately 760 Ma, and there are no known thermal events between 1050 and 920 Ma on King Island. We therefore consider the most likely location for the Glomar Zone before Gondwana breakup is close to its present position south of Tasmania, as originally suggested by Royer & Rollet (1997). This interpretation may be supported by the presence of Cretaceous northwest-trending basins on the east South Tasman Rise (Exon et al. 1997b), which are parallel with the dominant trends in the free-air gravity anomalies from with the Glomar Zone ( Figure 18). Furthermore, K/Ar ages determined on biotite in the Glomar Zone gave ages of 511 § 4 Ma and 451 § 4 Ma (Berry et al. 1997), consistent with Tyennan metamorphism. These data suggest there was not sufficient crustal extension or increase in the geothermal gradient to exceed the closure temperature of biotite, approximately 300 C, during Gondwana break-up.

Possible correlations with VanDieland
A fundamental part of deriving a coherent geological history for VanDieland is to place it within the wider context of Earth history. The Proterozoic events seen in VanDieland took place within the assembly and dispersal of Rodinia and the formation of Gondwana (Calver et al. 2014), and so events elsewhere should be reflected in the history of VanDieland. Because there is only one Proterozoic paleomagnetic data point of doubtful reliability from VanDieland (McWilliams & Schmidt 2003), other data assume greater importance in determining the most likely relationships with other cratons. We accept the reservation outlined by Andersen (2014) that identical detrital zircon populations can form at the same time on cratons distant from each other.
Nevertheless, even though these data are non-unique, they can exclude some possibilities. As well, where other data are available, we have attempted to integrate them as closely as possible into the discussion and Table 2.

ELSEWHERE IN AUSTRALIA
The provenance of the detrital zircons in Tasmania provides important constraints on where any Tasmanian crustal fragments may have come from. The dominant population, from approximately 1900 to 1650 Ma, is widespread in Australia and may have been derived from any of the Nuna assembly events (Evans 2013). In contrast, the 1460 to 1380 Ma zircons seen in VanDieland are less common elsewhere (Condie et al. 2009) and so provide a tighter constraint. In Australia, the only possibilities appear to be the Western Australian Madura Complex (near the South Australian border and covered by Cenozoic sedimentary rocks), which is known to include meta-igneous rocks containing zircons with the appropriate ages (Nelson 2006a, b), or the Musgrave Province in central Australia (Kirkland et al. 2013). However, if VanDieland has always been in generally the same position relative to the rest of Australia, as suggested in most Rodinian assemblies (e.g. Li et al. 2008b), the drainage required to transport the 1400 Ma zircons must have crossed the Mawson Craton, which includes the 1600 to 1500 Ma Hiltaba, Spilsby and St Peter suites, but there is little evidence of these older zircons in the VanDieland meta-sedimentary rocks. In contrast, 1600 to 1500 Ma zircons are abundant in the Neoproterozoic Adelaide Geosyncline metasedimentary rocks, which were also marginal to the Mawson Craton (Ireland et al. 1998;Preiss 2000).

EAST ANTARCTICA
There are few clear links between VanDieland and the pre-1300 Ma rocks of East Antarctica, including the Mawson Craton (Payne et al. 2009). However, the prominent 1900 to 1650 Ma detrital zircon peak in the Tasmanian quartzites overlaps with the 1730 to 1720 Ma Nimrod Orogeny in East Antarctica (Figure 19; Goodge et al. 2001). Clean metaquartzites in the area of the Princess Anne Glacier in the Queen Elizabeth Range also contain first cycle zircon populations with peaks between 1800 and 1700 Ma and lesser peaks from 1450 to 1400 Ma (Goodge et al. 2004), consistent with similar sources to the Tasmanian quartzites (Black et al. 2004). Although Goodge et al. (2004) did not record paleocurrent directions from the Antarctic quartzites, other units indicated sources from under the ice cap, and they considered that the quartzites had come from the same direction. Goodge et al. (2008) recorded the presence of a 1441 § 6 Ma (UÀPb zircon SHRIMP date) A-type granite clast in a moraine a few tens of kilometres north of the quartzites and indicated that it came from under the ice cap. This clast was correlated with the Granite-Rhyolite Province in Laurentia, suggesting southwestern Laurentia abut-  ages of dredged granite and orthogneiss from the Glomar Zone (Fioretti et al. 2005a;Berry et al. 2008). Borg & DePaolo (1991, 1994 used SmÀNd isotopic data to propose a strip of Nimrod Group basement that extended outboard of southern Victoria Land through to the Central Transantarctic Mountains and as far south as the Beardmore Glacier. This outboard terrane is characterised by T DM model ages of 1900 to 1600 Ma (Ɛ Nd À15.4 to À10.0 at 500 Ma), which are significantly younger than the >2000 Ma model ages for the inboard Antarctic terranes. Recent data support this division (Peucat et al. 2002;Goodge et al. 2012). Ca 1600 Ma granites are present in the Mawson Craton under the Antarctic ice (Peucat et al. 2002;Betts et al. 2008), but it is unlikely that the middle and lower VanDieland crust can be correlated with the deep Antarctic crust, because the Antarctic model ages are older than 2000 Ma where these intrusions are present.
Other correlations can also be made. Goodge et al. (2002) (Black 2007, unpublished data). A similarly aged (668 § 1 Ma, UÀPb zircon) gabbro was intruded into the middle of the Beardmore Group (Goodge et al. 2002), implying that rift-related sedimentation must have commenced earlier. The upper age of the Beardmore Group is poorly constrained, but an unconformably overlying package contained detrital zircons with ages as old as 589 § 6 Ma, which Goodge et al. (2012) interpreted as indicating that arc volcanism had commenced by then. However, the age is indistinguishable from that of 582 § 4 Ma zircons in a rift-related rhyolite within a basaltic sequence in the upper Togari Group in Tasmania , suggesting that similar rift-related rhyolites may have been the source to the 589 § 6 Ma zircons. Lastly, it may be no coincidence that both the early Beardmore Group and the possibly coeval Oonah Formation in Tasmania are quartz rich at the base, but become more carbonate rich near the top (Brown 1986;Goodge et al. 2002). et al. (2008b) and Li & Evans (2011) suggested that the Cathaysia Block lay between Laurentia and the Northern Australian Craton, which could suggest that VanDieland once lay adjacent to Cathaysia. This correlation is strengthened by the presence of similar zircon populations in the SorellÀBadger Head Zone to those from the Shihuiding Formation on Hainan Island at the southern end of the Cathaysia Block (Li et al. 2008a) ( Figure 19); both regions have detrital zircons with age peaks at ca 1000 Ma, and 820 to 720 Ma igneous zircons present in the southeastern-most Cathaysia Block cover the age of the 770 to 740 Ma Wickham Orogeny (Black et al. 1997(Black et al. , 2004Li et al. 2014).

Li
Zheng et al. (2011) considered that most, if not all, of the Cathaysia Block was underlain by Archean crust. However, Li et al. (2014) suggested that these Archean ages were derived from far-travelled zircons, and that a more appropriate basement age is approximately 1850 Ma or younger. Regardless of which is correct, neither age supports a correlation with VanDieland, where the data suggest a T DM model age of approximately 1700 Ma or younger. Furthermore, Cawood et al. (2013) considered that the presence of 1000 to 900 Ma arc and back-arc rocks along the western margin of the Cathaysia Block indicated it must have been located on the edge of Rodinia, perhaps off the Western Australian Craton, while Li et al. (2014) considered that these ages reflected plumerelated breakup. Both the nature of the southernmost Cathaysia Block on Hainan Island and whether these volcanic rocks are arc-related or break-up plume-related are beyond the scope of this paper. Thus, we consider that it is possible that VanDieland and southeasternmost Cathaysia might be able to be correlated, but that there are too many uncertainties to be more definite.

SYNTHESIS, CORRELATIONS AND CONSTRAINTS: A REGIONAL TECTONIC HISTORY
This section connects the previous interpretations and links them with others along the Australian and Antarctic Rodinian to Gondwanan margin.
The xenocrystic zircons in the Paleozoic granites in western Tasmania indicate the presence of an older, unexposed substrate with a strong 1650 to 1600 Ma igneous component. We suggest, first, that the source of these zircons is the Mazatzal Orogen of the Laurentian craton, which Goodge et al. (2010) indicated was adjacent to East Antarctica from at least 1400 Ma to ca 800 Ma, and second, that VanDieland lay between the two (Figures 20,  21). An implication is that we favour a broadly SWEAT- like reconstruction of Rodinia for Australia, Antarctica and Laurentia similar to that suggested by Goodge et al. (2010), Cawood et al. (2013) or .
The oldest rocks outcropping in the Tasmanian region are at the base of the Rocky Cape Group and on King Island, and were metamorphosed at ca 1300 Ma (Berry et al. 2005;Halpin et al. 2014). If a Rodinian fit close to those suggested by Goodge et al. (2010) or Li & Evans (2011) is valid, then it appears likely that the 1300 Ma metamorphism is connected to the Grenville events in southwestern USA or East Antarctica. We also suggest that the Grenville Front can be mapped at the northern edge of the Glomar Zone (Figure 18), since Grenville-age rocks have been recorded on the Glomar Zone but not in Tasmania.
As Rodinia began to break up, at approximately 830 Ma (Wingate et al. 1998), extension took place along the proto-Terra Australis margin. In VanDieland, it was marked by the 780 to 750 Ma Wickham Orogeny and sedimentation beginning in the Smithton Basin, coeval with the 777 § 7 Ma bimodal Boucaut Volcanics in the Adelaide Geosyncline (C.M. Fanning personal communication, 1994, in Preiss 2000. Continued extension is indicated by the intrusion of ca 710 Ma rift tholeiites in the Burnie Formation and also by the beginning of sedimentation in the Beardmore Group in the Transantarctic Mountains, which pre-dates 668 § 1 Ma pillow basalts and gabbro there (Goodge et al. 2002(Goodge et al. , 2004. Sedimentation also commenced at about 770 Ma in southwestern North America (Timmons et al. 2001;Yonkee et al. 2014). The 660 Ma intrusive age of the deformed granite in the SorellÀBadger Head Zone is also close to the age of the gabbro in the Beardmore Group, suggesting that they may have formed during the same extensional event. We interpret this rifting to have extended the Pedder, Tyennan and SorellÀBadger Head zones into crustalscale megaboudins (Figure 20).
After ca 650 Ma, strain became more distributed, allowing continued deposition in the Smithton Basin and equivalents, but perhaps also the Success Creek Group, which unconformably overlies the Oonah Formation (Brown 1986). Rifting also took place elsewhere along the Rodinian margin. In western Victoria, Neoproterozoic metasedimentary rocks were intruded by gabbro dykes that yielded primary zircons with an age of 643 § 4 Ma (Morand & Fanning 2006). In the Transantarctic Mountains, deposition continued in the Beardmore Group (Goodge et al. 2004). A seaway had developed along the entire western USA (Yonkee et al. 2014).
The final rifting event in VanDieland is marked by the 575 § 3 Ma MORB tholeiitic basalts of the Grassy Group on King Island Meffre et al. 2004) and the 582 § 4 Ma rift tholeiitic basalt of the Spinks Creek Volcanics in the Smithton Basin . The Crimson Creek Formation of the Burnie Zone and other mafic volcaniclastic packages in western Tasmania were probably deposited at this time or shortly before (Seymour & Calver 1998). These events left the King Island Zone isolated from both East Gondwana and VanDieland, while the Rocky Cape and Burnie zones remained contiguous. The Tyennan and Pedder zones were separated at their (now) northern ends, but connected by subcontinental lithosphere further south, similar to the present western North Atlantic Ocean, where the Hatton, Rockall and Porcupine banks are partly separated by V-shaped basins (P eron-Pinvidic & Manatschal 2010). The Tyennan and SorellÀBadger Head zones were also contiguous. This event slightly postdates the last rifting event in the western Lachlan Fold Belt (586 § 3 Ma; Greenfield et al. 2011), which is evidenced by ca 580 Ma detrital zircon population in the Delamerian Orogen (Ireland et al. 1998;Fanning & Morand 2002;Morand & Fanning 2009). In East Antarctica, 589 § 6 Ma zircon cores surrounded by 559 § 6 Ma rims in a clast of muscovi-teÀbiotite granite may record the same event (Goodge et al. 2012). The rims and the younger of the locally derived 580 to 540 Ma detrital zircon populations seen in the Cambrian Byrd Group in the Transantarctic Mountains (Goodge et al. 2002) may mark the transition to a convergent margin. Breakup was also completed in Laurentia (Yonkee et al. 2014), suggesting that VanDieland was left as an isolated microcontinental ribbon between the larger Gondwanan and Laurentian cratons.
After this, VanDieland drifted as at least two and perhaps as many as seven fragments on the proto-Pacific Ocean. In Tasmania, upper Neoproterozoic sedimentation was restricted to carbonate deposition. Lower Cambrian laminated siliceous siltstone and shale deposits suggest VanDieland was isolated from clastic sources (Burns 1964;Everard et al. 2007). The rifted fragments of VanDieland migrated northwards along the earliest Gondwanan margin and were not affected by the ca 560À530 Ma early phases of the Ross Orogeny (Allibone & Wysoczanski 2002;Stump et al. 2006;Goodge et al. 2012). Following a period of relative tectonic quiescence, the separated fragments of VanDieland began to converge at ca 530 Ma (Figures 20, 22), perhaps beginning to generate the high-pressure Franklin Metamorphic Complex at 529 § 10 Ma, before they underwent decompression at 512 § 4 Ma (Chmielowski 2009;Chmielowski & Berry 2012). Convergence between the Pedder and Tyennan zones occurred along the Mt Hobhouse Fault and may have coincided with the westward obduction of the maficÀultramafic complexes in the region at 516 § 1 Ma (Mortensen et al. in press). Final docking of these zones post-dated movement along the Mt Hobhouse Fault because the eastern edge of the Burnie Zone truncates the fault. During this event, the Wings Subzone in the Adamsfield area was trapped and obducted westwards onto the Pedder Zone. At a regional scale, these events are coeval with the 514 § 4 Ma U/Pb zircon age for the syn-tectonic Rathjen Gneiss in South Australia (Foden et al. 1999). Deformation associated with the Ross Orogeny also continued in Antarctica (Goodge et al. 1993) and was associated with felsic magmatism (Encarnaci on & Grunow 1996), associated with west-dipping subduction in the region (Boger & Miller 2004 Reed et al. 2002). If the two shortening events were coeval, they would be likely to have driven some maficÀultramafic complexes south, as described by Berry & Bull (2012). More regionally, Preiss (2001) interpreted an initial northwestÀsoutheast plate convergence at about 510 Ma. This was followed by northward propagating transpression. In the Koonenberry Block in western New South Wales, northeastÀsouthwest extension was accompanied by basaltic to rhyolitic volcanism (Greenfield et al. 2011). Deformation and magmatic activity also continued along the East Gondwanan Antarctic margin (Cawood & Buchan 2007, and references therein  ) and western New South Wales (Greenfield et al. 2011). This magmatism occurred in the overriding plate of a west-dipping subduction zone along the main East Gondwana margin, inboard of Tasmania Greenfield et al. 2011). The cessation of Tasmanian volcanism is coincident with movement on the Moyston Fault in western Victoria at ca 495 Ma ( 40 Ar/ 39 Ar on mica and hornblende; Miller et al. 2005). This fault has a major dip-slip component, with the hangingwall of amphibolite facies rocks and footwall of prehniteÀpumpellyite facies rocks (Cayley & Taylor 2001) and is interpreted to represent the relicts of a westdipping subduction zone (Miller et al. 2005). However, the fault also has a significant component of sinistral strikeslip movement, suggesting that oblique convergence in the region would have been likely (Cayley & Taylor 2001). The movement sense is consistent with the regional kinematics in the back-arc region in South Australia (Preiss 2001), suggesting northwest oblique convergence outboard of the Cambrian East Gondwana margin continued for at least 15 Ma. It is also coeval with subduction associated with the Narooma Complex on the south coast of New South Wales (Prendergast & Offler 2012). In the same period, our model suggests that the individual tectonic elements that comprise VanDieland were accreting to the west of the Narooma Complex (Moore et al. 2012a) but outboard of the Gondwanan margin.
After these accretion events, at the end of the Cambrian and into the Early Ordovician, western Tasmania underwent extension, perhaps caused by further rollback and/or by retreat of the eastern subduction zone. As a result, the locally derived Owen Conglomerate was deposited in half-graben on the west coast and in the north as far east as the eastern part of the SorellÀBadger Head Zone (Noll & Hall 2005;Reed & Vicary 2005). Similar aged, locally derived conglomerate is also present in western New South Wales (Greenfield et al. 2011) and probably also in north Victoria Land (Tessensohn & Henjes-Kunst 2005). Near VanDieland, subduction stalled, perhaps as a result of the collision of the Dimboola Complex into the early Gondwanan margin (Moore 2006). In Antarctica, granite intrusion and deformation continued into the Ordovician, until at least 480 Ma (Goodge et al. 2004;Rossetti et al. 2011). At least one granite, the 512 § 3 Ma Surgeon Island Granite, shows strong affinity in its inherited zircon population with granites in western Tasmania (Fioretti et al. 2005b). As well, the eastern Robertson Bay Terrane appears to be underlain by a different basement from the rest of north Victoria Land (Fioretti et al. 2005b). This suggests that VanDieland and the eastern Robertson Bay Terrane, separated in the Gondwanan breakup, were once part of the same microcontinental fragment. Thus, the terrestrial source of the Cambrian sedimentary rocks of the Robertson Bay Terrane (Henjes-Kunst & Sch€ ussler 2003;Tessensohn & Henjes-Kunst 2005) may have been the erosional products of the Tyennan Orogeny in VanDieland.
At this time, the only paleopole available from northwestern Tasmania places it close to the East Gondwanan coast, if not abutting it (Li et al. 1997). Evidence from Victoria suggests that the former is more likely. During the Late Ordovician to Silurian Benambran Orogeny, the Bendigo Zone was shortened by~200 km (Gray et al. 2006;Cayley et al. 2011), while the western Melbourne Zone was shortened by~40 km in the Middle Devonian Tabberabberan Orogeny (Foster & Gray 2007), which was accommodated by the Bendigo Zone over-riding the Selwyn Block . This shortening is well within the most likely error limits given by Li et al. (1997), a 95 D 10.4 (i.e.~700 km eastÀwest and~1200 km northÀsouth).
Our accretion model for VanDieland requires broadly similar plate kinematics to that defined on the late Cambrian eastern margin of Gondwana (e.g. Cayley & Taylor 2001;Cayley 2011). The horizontal accretion of micro-continental slivers or ribbons took place when a central core comprising the Tyennan and Pedder zones successively accreted or obducted the BurnieÀArthurÀRocky Cape, SorellÀBadger Head and King Island zones during northto northwest relative movement ( Figure 22). During this accretion, northÀsouth shortening took place in the Trial Harbour area on the west coast (McFarlane 2011), in central Tasmania (Calver et al. 2006) and the Arthur Lineament (Holm & Berry 2002). Our model suggests that the internal collisions were oblique, consistent with the kinematic observations. While the magnetic and gravity interpretation provides a clear sequence of accretion events, the polarities of the subduction zones between the accreted fragments are largely unconstrained. Exceptions are the collision between the combined Tyennan and Pedder zones and the Burnie Zone, where westward obducted maficÀultramafic complexes indicate east-dipping subduction (Crawford & Berry 1992), and the King IslandÀwestern Tasmanian subduction zone, which is also east dipping. Cayley (2011) proposed that VanDieland briefly accreted to the eastern margin of Gondwana at about 500 Ma. This is tentatively supported by trilobite fauna (Hally & Paterson 2014), which shows a convergence of shallow-water species in Gondwana and western Tasmania at that time.
At the global scale, these later events lie between the early Gondwana continent and the great circum-Pacific subduction system that has existed throughout the Phanerozoic (e.g. Coney 1992;Cawood 2005). We suggest that at least two Tasmanian fragments broke off in the dispersal of Rodinia at about 580 Ma (Yonkee et al. 2014) and these were left isolated in the newly formed ocean. Whether this was due to rifting or subduction roll back that caused back arc extension in the overriding plate is not clear. After VanDieland had departed, East Antarctica began to accrete adjacent terranes to form Gondwana (Boger 2011). However, in the region of the Transantarctic Mountains, the Neoproterozoic continental fragments remained stranded in a back-arc setting between a retreating west-dipping subduction zone and the Gondwana margin. VanDieland was left outboard of the subduction system that started in the Antarctic region at about 560 Ma (Goodge et al. 2012) and the larger subduction system that developed in the Cambrian along the early Gondwanan margin (Cawood 2005;Casquet et al. 2012). During inversion of this back arc setting the micro-continental fragments and megaboudins of Van-Dieland migrated northwards, and from ca 520 Ma to ca 495 Ma they successively accreted within this closing back arc system. Unlike the Beardmore and Bowers terranes along the Antarctic margin (Stump et al. 2006;Godard & Palmeri 2013), VanDieland did not accrete back onto the early Gondwanan margin at this time. Rather, it moved closer in the Early Silurian and, together with eastern Tasmania and the Lachlan Orogen, was finally integrated with the Gondwanan craton in the Middle Devonian.
The accretion of different fragments in different ways along the Gondwanan margin in the RossÀTyennanÀ Delamerian Orogeny would inevitably lead to different effects at different times along the same margin. The timings range from 554 § 10 Ma ( 87 Rb/ 86 Sr) in South Australia (Turner et al. 2009) and 559 § 6 Ma (SHRIMP) in Antarctica (Goodge et al. 2012), to 489 § 3 Ma ( 40 Ar/ 39 Ar) in western Victoria (Miller et al. 2005) and 484 § 8 Ma (SHRIMP) in Antarctica (Goodge et al. 2012), and perhaps even to ca 485 Ma (SHRIMP) in north Queensland (Paulick & McPhie 1999). Some microcontinental ribbons accreted directly onto the early Gondwanan margin, while others aggregated outboard. Our synthesis would suggest that VanDieland represents one of these microcontinental fragments, and was separated from the Western Lachlan Orogen by ocean crust that is now imbricated . It suggests that the accretion of VanDieland must have happened after the accretion of the Dimboola Complex in western Victoria. This may have occurred during the Early Ordovician but was finalised during early Silurian Benambran and Middle Devonian Tabberabberan

FUTURE DIRECTIONS
Many areas remain poorly controlled. At a more detailed scale, mapping of key areas will help to understand the more problematic boundary relationships. Because much of the region is difficult to access, mapping and age dating along the western Tasmanian coast are particularly important, and any results should be integrated with the potential field and seismic data. One important area is northwest from Nye Bay, where doubts remain as to the position of the eastern edge of the Burnie Zone and to the relationships of the mafic volcanic sequences there to the rest of the Burnie Zone. Another is along the Mt Hobhouse Fault, where mapping should more accurately locate its position and history. At the regional scale, there is a need for better paleomagnetic controls on VanDieland, but this can only come after tight age constraints can be placed on individual units. Age controls have been established in some of the glacial sequences of the Smithton Basin and associated rocks (Calver 2011;Calver et al. 2013a), but older rocks are generally less well constrained, as are supposedly equivalent sequences elsewhere. Tighter age controls have been established in some pre-570 Ma rocks on King Island and a comparison of poles with those in the Rocky Cape Zone would test the hypothesis that the two zones were from different parts of Rodinia. As well, too few SmÀNd model ages of the basement have been determined to make statistically valid comparisons with other regions. These should be obtained not only from the Paleozoic plutonic rocks, but also from plutonic rocks on the South Tasman Rise. Only then can the hypothesis proposed here be adequately tested.

ACKNOWLEDGEMENTS
The geophysical data used in this study have come from many years of acquisition and compilation by Geoscience Australia. The regional mapping was from Mineral Resources of Tasmania. Without these data, the study would not have been attempted. Research for this paper has been supported by Monash University as part of a PhD project. Their assistance has been invaluable. We also gratefully thank Clive Calver for his insights gained through many years of mapping much of Tasmania. Without these, the paper would have been much the poorer. Discussions with Nick Direen and Ross Cayley also helped clarify several issues. Reviewers Ross Cayley and Zheng-Xiang Li and an unnamed reviewer further improved this manuscript. This research was funded from ARC Discovery Grant DP11010253.