Zircon inheritance, sources of Devonian granitic magmas and crustal structure in central Victoria

Abstract In central Victoria, inherited zircon in Devonian igneous rocks and detrital zircon in metasedimentary country rocks and an amphibolite-facies xenolith show that Mesoproterozoic parts of the underlying Selwyn Block cannot be the source for all the silicic magmas. Zircon inheritance in S-type samples reveals significant thermal events at 525–425 Ma and 1200–1100 Ma. Both S- and I-type samples have prominent zircon age peaks at 420–410 Ma, which record high-grade metamorphism of the deep crust during the terminal phases of the Benambran and Bindian orogenies. All I-type rocks have 650–500 Ma peaks, suggesting derivation from an arc-related metavolcanic source in the upper Selwyn Block. Protoliths of the greenschist-facies Ordovician metasediments and the amphibolite-facies Cambrian metasedimentary xenolith were deposited in distal backarc settings. Most inherited zircon cores are metamorphic, and the strongest zircon inheritance occurs in hornblende-bearing I-type rocks, highlighting their largely crustal origin. Zircon populations at ca 1400 Ma, thought to signal sediment derivation from East Antarctica and Rodinia-Nuna, are mostly absent in I-type samples and some S-types. The ca 1400 Ma signal probably applies to the upper, metasedimentary Selwyn Block, so Devonian S-type magmas were sourced mainly in the deeper sections. Zircon inheritance in the Devonian igneous rocks was not influenced by the exposed metasedimentary country rocks. Two samples from one of the smaller plutons have contrasting patterns of zircon inheritance, suggesting relatively small-scale source heterogeneity. Many rounded and corroded cores in zircon crystals yield the same ages as the crystallisation dates for the rocks, and thus are antecrysts. Higher whole-rock Zr contents generally correlate with higher proportions of inherited zircon, and differentiation does not affect this relationship. The degree of partial melting of a magma source and the efficiency of crystal entrainment are critical in governing zircon inheritance. KEY POINTS Mesoproterozoic sections of the Selwyn Block cannot be the sources for all the Devonian silicic magmas in central Victoria. I-type rocks have 650–500 Ma zircon age peaks, suggesting derivation from arc-related metavolcanic rocks in the upper Selwyn Block. Hornblende-bearing I-type rocks have the strongest zircon inheritance patterns, indicating the largely crustal origins of I-type magmas. Exposed metasedimentary country rocks were not involved in magma genesis.


Introduction and regional framework
In some previous studies, inherited zircon crystals in granites and silicic volcanic rocks have been shown to be almost completely derived from the source rocks of the magmas, with very little contamination by zircon derived from the country rocks that the magmas passed through or intruded into (e.g. Bea et al., 2021). As a result, if inherited zircon crystals within individual granite samples are analysed in sufficient numbers to form a statistically relevant sample of the inheritance, U-Pb age/frequency plots can provide insights into the magma sources. Thus, inherited zircon populations in granites can be useful for understanding the nature of unexposed basement geology in regions that contain abundant granitic rocks of similar age. Central Victoria is such a fortunate region. Moore et al., 2016). Its northern tip, the Selwyn Block, forms the true crustal basement to the younger Paleozoic rocks in central Victoria. VanDieland probably consisted of strips of continental and arc crust that eventually amalgamated with Gondwanaland (e.g. Moore et al., 2015). Figure  1 illustrates the inferred position of VanDieland during the Neoproterozoic, within the context of the supercontinent of Rodinia and Nuna. At this time, VanDieland is inferred to have lain between East Antarctica and southwestern Laurentia. Nuna assembly involved events that occurred between 1900 and 1650 Ma and, based on zircon populations in metasedimentary and granitic rocks, an unexposed basement of 1650-1600 Ma age is inferred to exist beneath Tasmania and, by extrapolation, all of VanDieland. These metasedimentary and granitic rocks also contain a more distinctive 1460-1380 Ma detrital zircon population, which Moore et al. (2015) suggested was derived from East Antarctica and southwestern Laurentia ( Figure 1). Some of the Proterozoic formations in the Adelaide Superbasin also contain zircon populations with prominent ca 1400 Ma peaks (Lloyd et al., 2020).
Given the foregoing, it has been proposed that zircon populations derived from the Selwyn Block should contain a small but distinct 1400 ± 50 Ma population. The reasoning behind this is that such a detrital population is present throughout the exposed VanDieland microcontinent (in Tasmania) as well as in the southwestern USA, and that this population was ultimately derived from large, 1.50-1.34 Ga, granite-rhyolite provinces ( Figure 1) that were major sediment sources (Black et al., 2004;Moore et al., 2015;Mulder et al., 2015Mulder et al., , 2018. Apart from within these specific provinces, 1400 Ma zircons appear to be scarce in the geological record, making the presence of inherited zircon of this age quite diagnostic of Selwyn Block origins. Circa 1650 Ma zircon may also be present but this is deemed far less distinctive since zircon of this age is more common and widespread. An additional characteristic of Selwyn Block zircon would be the presence of very few Archean grains among the detrital populationsinsufficient to have been derived from a primary source such as the Wyoming Craton (D. Moore, written comm., 2022). The present work serves partly as a test of these predictions.
Recent work by Habib et al. (2022) adds a further perspective that is useful in interpreting our results, as well as in moderating expectations regarding the contributions of the Selwyn Block to the Devonian igneous rocks in central Victoria. These authors studied the zircon provenance in lower Paleozoic metasedimentary rocks of Tasmania and southernmost Victoria. They found that transport of zircon from distal regions into the Paleozoic sedimentary basins varied in both time and space, and that this pattern of change reflects the complex nature of the paleotopography and paleogeography. Thus, we may expect the same sorts of heterogeneity among the rocks that formed the crustal sources of inherited zircon in the igneous rocks studied here. Habib et al. (2022) also found that a shift in the detrital zircon signature in the Paleozoic metasediments of western Tasmania implies that VanDieland (and therefore the Selwyn Block) began docking against southern Australia during the Cambrian Tyennan Orogeny. This event correlates with the Delamerian Orogeny in South Australia and the Ross Orogeny in North Victoria Land, Antarctica, and spanned the middle Cambrian to the Early Ordovician (e.g. Turner et al., 1998). This suggests that zircon crystals of Early to Middle Ordovician age (ca 480-460 Ma) could be represented in the Selwyn Block and therefore could also be present among the inherited zircon populations in the Devonian silicic igneous rocks. Since docking of VanDieland with mainland Victoria appears to have already begun in the Cambrian, inherited Selwyn Block zircon could also record metamorphism during late extensional phases of the Benambran Orogeny (440-430 Ma).

Local geology
To understand the geological evolution of this part of southeastern Australia it is important to appreciate the nature of the basement in the central part of the state of Victoria. However, this subject is a matter of debate. The state has been divided into a series of tectonometamorphic zones (VandenBerg et al., 2000), separated by major faults, some of which extend to the Moho (Figure 2). The present study addresses the nature of the basement in the Stawell, Bendigo and Melbourne zones (Vandenberg et al., 2000) as well as the Bassian Basement Terrane (BBT) of Chappell et al. (1988), collectively identified here as 'central Victoria', and all forming part of the western Lachlan Orogen. Note that the BBT and the Melbourne Zone are closely related, and both are underlain by the Selwyn Block.
The prevailing view is that the Cambrian to Ordovician, chlorite-grade metaturbidites of the Bendigo and Stawell zones, in the western sector, were deposited on a thick basement of Cambrian seafloor metabasalts, with minor, intercalated, deep-water sediments ( Figure 3a). This Cambrian crust is seen as the remnant of an extensive sea floor, the western parts of which had been subducted westward beneath the Stavely Arc-modelled as having evolved from a passive margin, through a Japan-style arc to a Cambrian Andean-style margin (Cayley et al., 2018;Schofield, 2018). Further east, the Melbourne Zone and BBT contain mainly Silurian to Early Devonian lower greenschist-facies quartz-rich metaflysch (Phillips & Wall, 1980), with some minor pelagic Middle to Late Ordovician units, all originally deposited on a basement of the Proterozoic to Cambrian Selwyn Block (Cayley et al., 2002;Rossiter & Gray, 2008). Here also, mainly Cambrian metabasaltic rocks, with minor andesitic and ultramafic rocks, cherts and mudrocks, were emplaced in slivers along the bounding structures (Heathcote and Governor faults).
In the Devonian, the Paleozoic metasedimentary cover rocks of the region were intruded by numerous stocks and batholiths of high-K, calcalkaline, I-and S-type granitic rocks. In the Melbourne Zone, coeval silicic volcanic complexes are also numerous, and include I-, S-and A-type rocks. The magmas that formed these post-orogenic, extension-related, granitic and silicic volcanic rocks were generated mainly by partial melting of sources in the Paleoproterozoic to Mesoproterozoic parts of the Selwyn Block (e.g. Clemens & Buick, 2019) during the late, extensional phase of the Tabberabberan Orogeny. The same is assumed to have been the case for the magmas of the Wilsons Promontory batholith, in the BBT, although these granites are Early Devonian and belong to the Benambran/ Bindian rather than the Tabberabberan tectonic cycle, as do the Early Devonian post-tectonic granitic plutons in the Bendigo and Stawell zones ( Figure 2).
The early to middle Silurian Benambran Orogeny involved mainly the Stawell and Bendigo zones, the BBT and possibly the Selwyn Block. The Early to Middle Devonian Tabberabberan Orogeny affected these, as well as the Melbourne Zone. In the models of Moresi et al. (2014), the Selwyn Block is regarded as a relatively rigid body, against which the Ordovician turbiditic rocks were deformed as the Selwyn Block docked. This would explain the presence of Benambran deformation on both boundaries of the block but not within it. However, Benambran deformation could have affected the crust of the Selwyn  Moore et al., 2015). Note the belt of ca 1.4 Ga granitic and silicic volcanic rocks in East Antarctica and Laurentia (black crosses). This is inferred to be the source of detrital zircon of this age in VanDieland and which is therefore suggested to be present in the inherited zircon populations of Devonian silicic magmatic rocks in Tasmania and central Victoria. Note that the Mawson Craton is now commonly known as the Mawson Continent (e.g. Lloyd et al., 2020).
Block if it were not completely rigid. Following this compressional phase, there may have been insufficient mantle heat advected into this part of the orogen to permit widespread partial melting and formation of Early Devonian, post-orogenic, granitic magmas. However, there are two potentially Middle Devonian silicic magmatic units in the Melbourne Zone. The S-type, pre-caldera Hollands Creek Ignimbrite in the Tolmie Igneous Complex, on the eastern margin of the Melbourne Zone, is late Givetian to early Fransian in age (Long & Werdelin, 1986). Also, the very small S-type Bulla Granodiorite (actually a monzogranite) just north of Melbourne has been dated to 392 ± 5 Ma (Bierlein et al., 2001). After the Tabberabberan compression, there was evidently a considerable heating event in the deep crust of the whole region, leading to the widespread and voluminous Late Devonian, mainly silicic magmatism.
Using chemical and isotopic constraints, Clemens (2020) showed that the granitic rocks of the Stawell and Bendigo zones could not have been formed from the same source rocks as their counterparts in the Melbourne Zone. The inference is that, in agreement with its previously inferred position, the Selwyn Block does not extend far west of the Heathcote Faultthe boundary between the Bendigo and Melbourne zones ( Figure 2). However, Clemens (2020) also demonstrated that the Devonian granitic rocks of the Bendigo and Stawell zones could not have been produced by any process that involved partial melting of the Cambrian to Ordovician country rocks, or their deeper equivalents, and nor could they have been formed by mixing of such crustal melts with any plausible mantle-derived basaltic magmas. This conclusion is based on chemical and isotopic characteristics of the rocks as well as modelling of the Sr and Nd isotope variations in the granitic rocks and the Cambrian and Ordovician metasedimentary and meta-igneous rocks. Thus, the logical inference is that, beneath the Cambrian metabasaltic rocks, there lies a terrane, of probable Proterozoic age, that forms the true crustal basement of the Bendigo Zone ( Figure 3b). This inferred crustal block must contain a substantial component of extended ancient arc crust, as this kind of material is the only viable source of the various granitic rocks of the Stawell and Bendigo zones. However, this terrane is unlikely to be composed of rocks related to the Proterozoic formations that are exposed in far western Victoria and in South Australia (Clemens, 2020); it must be a distinct and potentially exotic block, or dismembered remnants of such a block.  Wilson et al. (2017) and Willman et al. (2010). The numbered locations indicate the approximate locations of the new samples in the present study: 1, Victoria Hill, Bendigo (sample VH2); 2, Mount Bute Granite (sample MB1); 3, Ercildoun Granite (sample AD1); 4 and 5, Mount Wombat pluton, Strathbogie batholith (samples S1 and S12, respectively); 6, Mount Disappointment laccolith (samples MD27 and MD 50); 7 and 8, Mount Alexander and Baringhup plutons, Harcourt batholith (samples HAR17 (H17) and HAR19 (H19)); 9, Lysterfield Granodiorite (sample L1); 10, Tynong pluton, Tynong batholith (sample TY2); 11, Oberon pluton, Wilsons Promontory batholith (sample WPB7 (PS)). Lower-case letters indicate the locations of silicic igneous rock samples from previous work: a, Oberon pluton, Wilsons Promontory batholith, samples 2340 and W96H2 from Elburg (1996); b, rhyolite from the Toombullup Ignimbrite in the Tolmie Igneous Complex ; c, an ignimbrite from the Lake Mountain Rhyodacite in the Marysville Igneous Complex (sample E27-2) from unpublished work by JDC, F. Finger and D. Frei; and, from the work of Bierlein et al. (2001); d, the Bulla Granodiorite (sample CV108); e, the Pyalong pluton of the Cobaw batholith (sample WV99); and f, the Tullaroop Granodiorite (sample WV8).
The present work explores the characteristics of the sources for the Devonian granitic magmas of this region, the relationships between these sources and the implications for the geological and tectonic evolution of the western Lachlan Orogen. These subjects are approached through a study of inherited cores of zircon crystals in local S-and I-type granitic rocks, as well as detrital zircon grains in the Paleozoic metasedimentary country rocks and an amphibolite-facies metasedimentary xenolith that is inferred to have been derived from part of the Selwyn Block (Clemens & Buick, 2019).
In addition to the conclusion of Bea et al. (2021), that inherited zircon crystals in granites are derived from their source rocks but not their wall rocks, these authors also found that the rates of magma segregation and extraction should not affect the degree of inheritance observed, except in cases where partial melting takes place through heating by underplated or intraplated mafic magmas. The reasoning here is that melting reactions are slower than zircon dissolution reactions in slowly heated terranes and much faster in underplating or intraplating situations, in which temperatures can also reach higher levels. In central Victoria, heat for partial melting is inferred to have been due to post-orogenic, extensional tectonics and intraplating of mantle-derived magmas, with rapid mobilisation of crustally derived silicic magmas. These conditions allow for considerable disequilibrium and zircon preservation at temperatures that would have destroyed most inheritance in other types of settings (e.g. Bea et al., 2007;Harris et al., 2000;Watt & Harley, 1993). From Bea et al. (2021) we would expect two things to be the case in central Victoria: (1) there should be a considerable degree of zircon inheritance in the granitic rocks; and (2) there should be minimal influence of wall rocks on the zircon populations. We test both these model predictions. Section (a) is a representation of the present interpretation from the surface geology and geophysical data, after Cayley et al. (2011, figure 17f), as modified by Clemens (2018). Section (b) is an alternative interpretation from Clemens (2018), consistent with the surface geology and the chemical and isotopic constraints provided by the Devonian granitic rocks, but at odds with the current interpretations of geophysical data. See the text for further details.

Rock types in the Selwyn Blockpreliminary deductions
The inference from Clemens (2018) and Clemens and Buick (2019) is that the Devonian silicic magmas were formed through high-temperature metamorphism and partial melting of backarc volcaniclastic metagreywackes, with minor metadacites, meta-andesites and metapelites, all in the deeper parts of the Selwyn Block. Clemens and Buick (2019) constructed an Rb-Sr errorchron for the metamorphic age of all analysed Selwyn Block xenoliths from the granitic rocks and confirmed the existence of the inferred high-grade thermal event, which peaked at ca 364 Ma. However, the cross-Bass Strait correlations of Moore et al. (2016) suggest that the Mesoproterozoic parts of the Selwyn Block do not generally contain suitable source rocks for the full variety of Devonian silicic magmas. Certainly, on King Island the Mesoproterozoic Surprise Bay Formation consists of >1 km thickness of quartzo-feldspathic schists with minor quartzites, pelitic schists and rare calcareous lenses, intruded by volumetrically minor metabasic amphibolites (Calver, 2007). If metasedimentary rocks similar to these were abundant in the Selwyn Block beneath central Victoria, they could have provided sources for some of the Devonian S-type granitic magmas. However, there do not appear to be significant volumes of intermediate metavolcanic rocks that could have been sources for the abundant, contemporaneous I-types, many of which have strong crustal signatures and, as we shall demonstrate, contain abundant inherited zircon. In this context, note that most of the central Victorian I-type rocks cannot have been derived through mixing between mantle-derived and crustal magmas (e.g. Clemens, 2020). This conclusion contrasts with multicomponent mixing models such as those of Collins (1996Collins ( , 1998 and Keay et al. (1997), although it should be recalled that these mixing models relate to a different geographic region, far to the east in the Lachlan Orogen. For the central Victorian case, Clemens and Buick (2019) suggested that the main source of the Devonian granitic magmas could lie in the deeper, Paleoproterozoic section of the Selwyn Block. The lithological characteristics of this part of the Selwyn Block remain essentially unconstrained by geological data and poorly constrained by seismic data. However, based on the compositions and phase relations of the granitic magmas, together with their Sr, Nd and O isotopic characteristics, it is reasoned that the most abundant source rocks for the Stype magmas are likely to be metagreywackes originally deposited in backarc settings.

Zircon survival in granitic magmas
Zircon has low solubility in granitic melts, and its saturation in granitic magmas depends on both temperature and melt composition, including H 2 O content (Harrison & Watson, 1983;Watson, 1996). This behaviour is considered to be sufficiently well parameterised that Zr contents of magmatic rocks have been proposed to accurately constrain the minimum temperatures of magma genesis (e.g. Hanchar & Watson, 2003 and references therein). Si egel et al. (2018) provided a thorough analysis of the fundamental conceptual flaws inherent in the use of bulk-rock Zrbased thermometry, including the commonly incorrect but necessary assumption that all zircon is autocrystic and the fact that a number of different processes can change the effective bulk composition of a magma or melt. Some recent studies have addressed this complexity by integrating phase-equilibrium modelling with estimates of zircon saturation in melts as a function of temperature (e.g. Kirkland et al., 2021;Laurent et al., 2022). However, even with this more sophisticated approach, the problem of the presence of xenocrystic, antecrystic and inherited zircon remains. This is particularly relevant for the present study because it has been demonstrated that some granitic magmas form, ascend and cool quickly enough for their inherited zircon cargo to be incompletely dissolved, even though the melts were very probably zircon-undersaturated when the magmas segregated from their partially molten sources (e.g. Bea, 1996;Bea et al., 2007;Clemens, 2003;Harris et al., 2000;Watt & Harley, 1993). That is, in silicic melt-zircon systems, kinetic factors favour disequilibrium, which is actually a boon for those trying to trace magma origins. The presence of small populations of very old zircon grains in many of the present samples suggests that the statistics of zircon survival, even for small age populations, are more favourable than some equilibrium experimental results would suggest.
Existing zircon U-Pb zircon data for central Victoria Squire et al. (2006) is a significant data source for zircon populations in upper Cambrian to Lower Devonian metasedimentary rocks from central Victoria. These authors present data from the upper Cambrian Saint Arnaud Group metaturbidites, including the Leviathan and Albion formations (samples ST18, ST21, ST23 and ST43), all in the Stawell Zone. To facilitate comparisons, we combined the data for these samples into a composite Stawell Zone Cambrian dataset. From the same work, rocks of the Melbourne Zone include the uppermost Ordovician Riddell Sandstone (metagreywacke) of the Sunbury Group (sample ST36), the upper Silurian quartz-rich metagreywacke of the Humevale Siltstone (sample ST39), and the Lower Devonian Norton Gully Sandstone (sample ST47). As for the Cambrian, we combined the Silurian and Devonian data for the Melbourne Zone to create a composite, Melbourne Zone Silurian-Devonian dataset. Data for the Lower Devonian Glen Creek Lithic Sandstone were excluded because the detrital components of this unit appear to be locally derived, from sources close in age to the depositional ages, and it contains essentially no older zircon populations. Readers can view plots for the individual Cambrian, Silurian and Devonian metasedimentary units in Squire et al. (2006, figure 4), and we provide descriptions of the age spectra below.
Some additional, igneous-rock zircon samples are available. Elburg (1996) published data for zircon crystals in two rocks (samples 2340 and W96H2) from the Oberon pluton (Clemens & Wallis, 2022;Wallis & Clemens, 2018) in the Early Devonian Wilsons Promontory batholith, which lies just inside the BBT, at the southeastern edge of the Melbourne Zone ( Figure 2). These data were combined with our new analyses from the Wilsons Promontory to create a composite dataset for the Oberon pluton; see below.  published data for a peraluminous Atype rhyolite sample of the Upper Devonian Toombullup Ignimbrite in the Tolmie Igneous Complex (sample TOOM). This complex straddles the Governor Fault, a suggested terrane boundary that separates the Melbourne Zone from the Tabberabbera Zone, to the east. Unpublished data for the zircon population in an Upper Devonian, S-type ignimbrite (sample E27-2) from the Lake Mountain Rhyodacite (Marysville Igneous Complex) are also available (courtesy of Prof. Dirk Frei, who analysed the sample, for JDC, in 2014). Lastly, Bierlein et al. (2001) dated a set of granitic rocks from western and central Victoria using the SHRIMP U-Pb technique on zircon. In the process, these authors amassed data on inherited grains, particularly for sample CV108 of the Middle Devonian S-type Bulla Granodiorite (monzogranite) in the Melbourne Zone, sample WV99 of the Late Devonian S-type Pyalong pluton in the Cobaw batholith, mainly in the Bendigo Zone, and sample WV8 from the I-type Tullaroop Granodiorite in the Bendigo Zone.
The approximate locations of all these samples from the literature are shown in Figure 2.

New samples
Most 'new' rock samples in this study (MD27, MD50, L1, S1, S12, TY2, HAR17 [aka H17] and HAR19 [aka H19]) are materials collected for work previously published by Clemens and Benn (2010), Clemens and Bezuidenhout (2014), Clemens and Phillips (2014), Clemens et al. (2016), Clemens (2018, Clemens and Buick (2019) and Clemens et al. (2020). Additional samples (VH2, MB1 and AD1) were collected expressly for the present study, and a mafic schliere (WPB7 [aka PS]) from the Oberon pluton of the Wilsons Promontory batholith was collected as part of petrogenetic work on that batholith (e.g. Clemens & Wallis, 2022;Clemens et al., 2020). Table 1 shows the structural zone from which each sample was taken, the stratigraphic unit or plutonic body, the rock type or, in the case of the granitic rocks, whether it is S-or I-type, its age and the method by which the age was determined. Table 2 gives the majorand trace-element compositions of the samples. Granitic sample (MB1) was collected from the Late Devonian Mount Bute Granite in the Stawell Zone. In the Bendigo Zone, samples include a Lower Ordovician lower greenschistfacies metagreywacke (VH2), from Victoria Hill near Bendigo, and a Late Devonian I-type granitic rock (AD1) from the Ercildoun Granite. From the Melbourne Zone, samples include the amphibolite-facies metapelite xenolith (S12) from the Strathbogie batholith, three Late Devonian S-type granitic rocks (S1, HAR17, MD27 and MD50), from the Strathbogie and Harcourt batholiths and from the Mt

Analytical techniques
Supplemental data (Paper 1) provides information on the techniques used to analyse the new samples for their whole-rock major-element contents and Sr and Nd isotope ratios. It also describes the methods by which the zircon crystals were harvested and gives details of how these were prepared and analysed (by sector-field LA-ICP-MS) for their U and Pb isotope ratios, data processing and quality . Selection of CL images of zircon crystals separated from sample AD1, a weakly peraluminous, I-type, porphyritic biotite micromonzogranite from the Ercildoun Granite (Mount Bolton pluton) near the town of Addington. The crystals are relatively large and prismatic, with an average size of $150 Â $50 lm, with long prisms and acute but commonly multiple pyramids and few pinacoids. The apparently inherited cores are large and abundant. They mostly have well-rounded shapes and oscillatory zoning, although some show mottled zonation. Some are equant, but most are somewhat elongate. Some show evidence of brittle fracture after having undergone previous rounding, probably in a sedimentary environment. The rims are narrow and show typical igneous oscillatory zoning. Some cores are CL-dark, others are CL-bright, and a few have embayed margins against the late overgrowths. For analysed crystals, spot numbers are shown, along with the corresponding ages in Ma (yellow for those with concordance between 90 and 110%, and white for more discordant analyses). Grains not analysed are given the designation 'AD1-n.a'. Note that some of the discordant cores are CL-dark, indicating high U contents, and are surrounded by radial fractures (e.g. spot AD1-71 toward the centre of the figure). Such cores are probably metamict, but a few were analysed to check. On the other hand, many discordant cores are CL-bright and show no signs of such cracking (e.g. AD1-74, -77 and -89 toward the LHS of the figure). Figure A3.3 of Supplemental data (Paper 3) shows essentially the same image.  control. The complete dataset is presented in the Supplemental data (Paper 2).

Zircon petrography
Detailed descriptions of the zircon population in each sample can be found in the Supplemental data (Paper 3). Much of each description makes use of SEM cathodoluminescence (CL) images. Examples of zircon crystals with 'old' cores, from an I-type granitic rock, are shown in Figure 4. On textural grounds, such 'old' cores are commonly assumed to be inherited grains. However,  found that, in the samples that we describe here, many of these cores are actually magmatic and were probably precipitated from the magma systems that formed the host igneous rocks, i.e. they are antecrysts, in the sense of Charlier et al. (2005). We also used zircon textures to infer the metamorphic or igneous origins of the analysed genuinely inherited cores (see below).

Zircon age spectra
All zircon populations, whether from the literature or the present study, were processed or reprocessed to produce probability density distributions (PDDs), with spot ages with between 90 and 110% concordance accepted for inclusion in the models. For relatively small samples of data points, as in some of the present cases, a bandwidth (1r uncertainty) that is rather greater than the measured instrumental uncertainty is preferred (Vermeesch, 2018a). Thus, a bandwidth of 30 Ma was applied to all data points (see e.g. Andersen et al., 2018). A bin width of 5 Ma was used for the plots. In most igneous samples, the crystallisation ages are ca 370 Ma, and we found that a few rather ancient zircons are present in some samples. To filter out non-inherited zircons, we therefore plotted all age spectra between 400 and 3600 Ma, 400 Ma being 370 plus the adopted 1r of 30 Ma. In the PDDs, the peaks calculated using the 30 Ma bandwidth are shown overlaid on the age spectra calculated using the 1r instrumental uncertainties, so that relationships between the two forms of presentation are apparent. In data collected using beam techniques (i.e. SHRIMP or LA-ICP-MS), there is a potential problem with analytical spots that span or drill through boundaries between different age zones in crystals with multiple age components. The result of this would be the production of mixed ages. Since we used a relatively fine laser beam of 20 lm, and quite carefully located our analytical spots to avoid such boundaries, such a problem is minimised for the present datasets; see the Supplemental data (Paper 3) for examples. As a further illustration of how the data appear, we present Wetherill concordia plots for a selection of our samples in the Supplemental data (Paper 4). For each sample, we show the data points in some critical regions along the concordia and also provide a list of age peaks, as identified in the PDD models.
An additional question here is whether the relatively small numbers of spots analysed in some samples mean that a proportion of the smaller zircon populations may have been overlooked. According to Vermeesch (2004), the optimum number of datapoints in a zircon provenance study is 117. This number should be viewed as a threshold beyond which inclusion of additional analyses brings minimal statistical improvement. Among our new data, samples AD1, S12 and L1 fall below this level, as do the previously unpublished data for E27-2, the published data for sample TOOM, from the study of  and all samples (CV108, WV8 and WV99) from the study of Bierlein et al. (2001). Table 3 shows the numbers of datapoints (90-110% concordant spot ages) for each of these samples, and the estimated maximum age fraction that may have been missed, at the 95% confidence level. The fractions are quite low for most samples ( 1.3%). However, the smaller numbers of datapoints gleaned from the study of Bierlein et al. (2001) mean that approximately 1.5% may have been missed in samples CV108, WV8 and WV99. Nevertheless, these are relatively small proportions, and the fact that even these three samples contain numerous very ancient grains, with spot ages up to 2629 Ma, suggests that any abundant and characteristic-aged zircon would have been detected, even in the samples with just 29-35 datapoints, and this should certainly be the case for most of the samples here, with 74-361 datapoints (Supplemental data, Paper 2).
For Victorian Ordovician metasediments it is possible to calculate an Rb-Sr errorchron using the 10 data points from Turner et al. (1993), O'Halloran (1996 and Anderson (1997). We used IsoplotR (Vermeesch, 2018b) to construct a model-3 regression (maximum likelihood with overdispersion), yielding an Early Ordovician apparent date of 474.8 ± 1.8 Ma. U-Pb zircon age peaks at ca 430 Ma have been recorded from mid-crustal metasedimentary xenoliths in the Amboyne Granodiorite, which crops out around Deddick Valley, in the east of the state (Maas et al., 2001). Maas et al. (2001) correlated this with a deep-crustal, amphibolite-facies, metamorphic event that occurred during the terminal phase of the Benambran Orogeny (440-430 Ma). They noted that these xenoliths also yield zircon age peaks at around 500 and 1200-1100 Ma, regarded as typical of Paleozoic sedimentary sequences of the Gondwana margin in eastern Australia, New Zealand and East Antarctica. Likewise, the inheritance patterns of our S-type granitic samples (S1, MD27, HAR17 and the Oberon pluton composite), as well as the Pyalong Granite and the Bulla Granodiorite samples WV99 and CV108 from Bierlein et al. (2001), show evidence of thermal events between 525 and 425 Ma and in the range of ca 1200-1100 Ma. This latter age range also appears as a minor component of the age spectrum of our Selwyn Block metasedimentary enclave from the Strathbogie batholith (sample S12). However, as will be seen, the inheritance patterns in our dataset are complex, with major age peaks commonly falling outside these ranges and suggesting a variety of different sediment provenances.
The age spectra that we obtained are shown in Figures 5-10. Figures 5 and 6 also include the age spectrum for the Selwyn Block schist xenolith, for comparison with the spectra for the Cambrian to Silurian and Middle Devonian metasediments that form the country rocks into which the granitic plutons were emplaced. In this connection, it is of interest to consider what the zircon data imply about the tectonic settings in which the protoliths of the metasedimentary country rocks may have formed. Recognising that detrital zircon populations reflect the tectonic settings of depositional basins, Cawood et al. (2012) developed a method of distinguishing between convergent, collisional and extensional settings. This involves plotting the distribution of the difference between the measured crystallisation ages of individual zircon grains in a sediment or metasediment and the depositional age (i.e. CA À DA) as a cumulative histogram. Such histograms can also be used to distinguish between various environments within the three broad categories.
In Figure 11 we show cumulative histograms for VH2, from the Ordovician of the Bendigo Zone (DA % 480 Ma, from the youngest detrital zircon with concordance between 95 and 100%), and S12, the Selwyn Block Cambrian metasedimentary xenolith from the Strathbogie batholith (DA % 493 Ma, also from the youngest concordant detrital zircon). Although there are very thin metamorphic zircon rims in S12 (Supplemental data, Paper 3, Figure A3.6), these are far too narrow to be analysed. Cawood et al. (2012) suggested that extensional settings have CA À DA >150 Ma in the youngest 5% of the zircons, and that all convergent settings have CA À DA <100 Ma in the youngest 30% of zircons. From these broad relationships, it seems clear that both rocks in Figure 11 were deposited in convergent settings. Comparing the detailed shapes of the curves for VH2 and S12 with those for different convergent settings (transferred to Figure 11 from Cawood et al., 2012, figure 2), the best overall fit is to the backarc region and the poorest to the forearc and intra-arc. The result for sample VH2 is in keeping the inference that the granitic plutons in the Lachlan Orogen were all emplaced in a distal continental backarc setting . It is also in accord with the inference that the parts of the Selwyn Block that formed the sources for the Devonian granitic magmas of central Victoria were most probably formed in the backarc basin of an Andean-type margin (e.g. Clemens & Buick, 2019).
Sample S9 is closely similar in mineralogy and texture to S12 and comes from the same location in the Strathbogie batholith (location 5 in Figure 2). Clemens and Buick (2019) used pseudosections to determine the depth and temperature at which S9 crystallised. They found that this xenolith (and presumably S12) originated at a depth of $17 km and experienced a peak metamorphic temperature of $775 C. These upper amphibolite-facies conditions are in keeping with the presence of thin leucosomes in these xenoliths, denoting the occurrence of arrested partial melting. From our observations, some other granitic plutons in central Victoria contain similar high-grade metapelitic xenoliths (e.g. the I-type You Yangs and Cobaw batholiths). Assuming that these xenoliths are all upper amphibolitefacies pelitic to psammopelitic rocks, as they appear to be in hand specimen, this implies the presence of a widespread horizon of distal backarc aluminous metasediments of late Cambrian age at mid-crustal depths in the Selwyn Block. This lithological characteristic does not correspond closely with the cross-Bass Strait correlations made by Moore et al. (2016) but is in accord with the inferred tectonic setting of the broader region through the Paleozoic, at least . This suggests that there may be considerable variation between VanDieland in Tasmania and the Selwyn Block in central Victoria, at least in the unexposed middle crust.

Metasedimentary rocks of the Stawell Zone
These are shown as a combined dataset in (Figure 5b) but are described individually below.

The late Cambrian St Arnaud Group-samples ST18 þ ST21
A major peak centres on 594 Ma in the late Neoproterozoic. A second major cluster centres on 1039 Ma. There is a small blip at 1409 Ma and two other small clusters centred on about 2100, 2564 and 2744 Ma. The earliest dates are at 3024 and 3329 Ma.

The late Cambrian Leviathan Formation-ST23
Major peak clusters are at 599, 924 and 1074 Ma. There is no ca 1400 Ma peak, but there is a wide range of small peaks between 3169 and 2524 Ma, with more prominent clusters at 1899, 2599 and 2789 Ma.  Squire et al., 2006); and (c) Selwyn Block schist xenolith from the Strathbogie batholith (sample S12). Spectra shown in dark colours are the data processed using a bandwidth equal to 1r measurement uncertainties on 206 Pb/ 238 U spot ages. These are overlaid with lighter-coloured spectra calculated using a constant bandwidth of 30 Ma, as recommended for detrital zircon studies (Andersen et al., 2018). The pink band at 1400 ± 50 Ma indicates the age that has been suggested as distinctive of zircon in the Selwyn Block. See text for further details.  Squire et al., 2006); and (c) Selwyn Block schist xenolith from the Strathbogie batholith (sample S12). Spectra shown in dark colours are the data processed using a bandwidth equal to 1r measurement uncertainties on 206 Pb/ 238 U spot ages. These are overlaid with lighter-coloured spectra calculated using a constant bandwidth of 30 Ma, as recommended for detrital zircon studies (Andersen et al., 2018). The pink band at 1400 ± 50 Ma indicates the age that has been suggested as distinctive of zircon in the Selwyn Block. See text for further details.

The late Cambrian Albion Formation-ST43
This sample, analysed by Squire et al. (2006), contains a small peak at 414 Ma (earliest Devonian) and a large grouping of zircons centring on 584 Ma. Other strong peaks are at 969 and around 1150 Ma. There is no Selwyn Block peak at ca 1400 Ma. There are small prominent peaks at 1344 and 1844 Ma and a cluster between 2600 and 2400 Ma, with the oldest zircons dating to 2809 Ma.
Combining the data for all the Cambrian units, and applying a bandwidth of 30 Ma, the resulting PDD shows prominent peaks at 594, 1054 (cluster at ca 1074-950), 2594 and 2804 and 3019 Ma (Figure 5b). Note the absence of any distinct age peaks that would be characteristic of sediment sources that include the ca 1400 Ma belt of granitic and rhyolitic rocks in East Antarctica and Laurentia, i.e. within the pink band in Figure 5b.

Metasedimentary rocks of the Bendigo Zone
Lower Ordovician greywacke of the Castlemaine Group-VH2 (Figure 5a) In this sample, there are two prominent groups of age peaks, one centring on 524 Ma and another at 1049 Ma. There is a small peak at 1364 Ma and a broad range of low peaks between 2999 and 1624 Ma. The ca 1400 and 1650 populations may indicate Selwyn Block sources, which is perhaps unexpected. Applying a 30 Ma bandwidth results in a PDD with just two significant age peaks: one at 524 and the other at 1049 Ma ( Figure 5a). As for the Cambrian rocks of this zone, at this resolution, the age spectrum contains no distinct signal related to the ca 1400 Ma granitic and rhyolitic rocks of East Antarctica and Laurentia. If granitic rocks in this zone were to contain an influence from these Ordovician metasediments, the 1049 and 524 Ma peaks would be the only ones likely to be in evidence. The 524 Ma population could be derived from the Ross Orogen, and the 1049 Ma group from the Transantarctic Mountains, where Grenville-aged zircon is common (e.g. Boger, 2011;Goodge et al., 2004).

Metasedimentary rocks of the Melbourne Zone
The Silurian and Devonian rocks have been combined into a composite dataset, illustrated in Figure 6a Zone rocks lack a ca 1400 Ma signal that might correlate with the silicic magmatic province of that age in East Antarctica and Laurentia.

Upper Silurian Humevale Siltstone-ST39 (Figure 6a)
An unusual feature of this rock is the presence of a population of late Silurian (Homerian) zircons with a spot age peak at 429 Ma, suggesting a magmatic source only very slightly older than the earliest possible Gorstian depositional age of 427.4 Ma. This is not visible in the spectrum, except as a shoulder on the strong cluster centred on 554 Ma. There is another major cluster between 1300 and 924 Ma, centred on 1054 Ma. Combining ST39 and ST47 to form a single Silurian-Devonian dataset and applying a 30 Ma bandwidth filter to the data, results in quite a complex pattern with major peaks at 1054 and 549 Ma (similar to the Ordovician rocks), as well as large peaks at 2654,2454,2009,1789,1644,1429 and 1219 Ma (Figure 6a). The population at 1429 Ma could signal detrital zircon derived from either a then-exposed portion of the Selwyn Block or from a continental source in East Antarctica or Laurentia. If a granitic rock in the Melbourne Zone were to have a population of zircon derived from the local Silurian-Devonian strata, a prominent peak at ca 1789 Ma ought to be present.

Granitic rocks of the Stawell and Bendigo zones
Late Devonian I-type Mount Bute Granite (monzogranite)-MB1 (Figure 7d) This pluton intrudes Cambrian metasedimentary rocks of the Stawell Zone and was dated at 370 ± 4 Ma (Bierlein et al., 2001) and 371.3 ± 2.8 ; see Table 1. Inherited zircons cluster in several age groups centred on 1024, 894, 609 and 424 Ma, with smaller peaks at 2409 and 1814 Ma. This sample lacks ca 1400 and 1650 Ma zircons and crops out well to the west of the projected subsurface limit of the Selwyn Block (Figure 2). The PDD constructed with a 30 Ma bandwidth shows major peaks at 2409, 1814, 1024, 894, 609 and 424 Ma. No Selwyn Block signature is present, and the spectrum bears almost no elements in common with the Cambrian metasediments or the Selwyn Block xenolith. Thus, there is no basis for a suggestion of influence from the Selwyn Block or the country rocks. The inheritance pattern, however, suggests that a separate, unknown crustal component was involved.
Late Devonian I-type Ercildoun Granite (porphyritic monzogranite)-AD1 (Figure 7c) Where sampled, this pluton intrudes Ordovician metaturbidites of the Bendigo Zone, and has previously been dated using K-Ar and Rb-Sr techniques, providing dates between 364 ± 10 and 359 ± 4 Ma (VandenBerg et al., 2000).  dated the pluton to 369.8 ± 5.6 Ma. The zircon inheritance pattern shows several age peaks between 669 and 414 Ma (late Neoproterozoic to earliest Devonian), with a large, late Mesoproterozoic grouping at 1169-1079 Ma. There is a prominent group at 1404 Ma, which is similar to the signature of the Selwyn Block xenolith, although this pluton lies to the west of the projected subsurface limit of this block. There are then prominent age peaks at 2709, 2579 and 1724 Ma (Neoarchean and late Paleoproterozoic). In the spectrum generated using a 30 Ma bandwidth, this rock has prominent zircon age peaks at 2709, 2579, 1724, 1404, 1169, 1079, 669, 579, 509 and 414 Ma. This pattern bears little resemblance to that of MB1, the Ordovician country rock (VH2) or the metasedimentary Selwyn Block xenolith (S12). However, the presence of the 1404 Ma peak does suggest the possibility of a Selwyn Block influence, perhaps a meta-igneous component in the magma since this is an I-type rock.
Apart from minor overlap in the Cambrian parts of the spectra, there is little similarity between the two populations of zircon crystals from these granitic plutons in the Stawell and Bendigo zones. The presence of earliest Devonian zircon cores in AD1 is one major point of difference between the two. Given the above observations, the  (Bierlein et al., 2001). Spectra shown in dark colours are the data processed using a bandwidth equal to 1r measurement uncertainties on 206 Pb/ 238 U spot ages. These are overlaid with lighter-coloured spectra calculated using a constant bandwidth of 30 Ma, as recommended for detrital zircon studies (Andersen et al., 2018). In (a) and (b), the peaks are shaded blue to red and denote transitions between S-and I-type character. The pink band at 1400 ± 50 Ma indicates the age that has been suggested as distinctive of zircon in the Selwyn Block. Thin red lines indicate calculated two-stage Nd model ages for the rocks, with data for HAR17 and HAR19 from Clemens (2018). See text for further details.
idea of a significant wall-rock component in these zircon populations receives little support, and the implication is that the granite source rocks, at depth, may have differed significantly. Since AD1 is in the Bendigo Zone, and MB1 is in the Stawell Zone, the deep crustal components in these two zones appear to differ significantly. It may be that the Selwyn Block projects a little further west than indicated in Figure 2.
Late Devonian Mount Alexander pluton-HAR17 (Figure 7a) Liu et al. (2017) published a crystallisation age of 373 ± 2 Ma for the Mount Alexander pluton and  determined the age of HAR17 to be 376.5 ± 2.0 Mathe same, within uncertainty. Our Ordovician metasedimentary sample VH2 could be regarded as typical of the local wall rocks. The sample location lies just inside the area inferred to be underlain, partly, by the western fringe of the Selwyn Block ( Figure 2). Despite this batholith being classified as an I-type, Clemens (2018) compiled evidence that favours its origin in partial melting of a mainly metasedimentary (volcaniclastic?) source. This sample is somewhat exceptional within the Mount Alexander pluton in that it is the most peraluminous collected from the batholith. It has chemical features indicating S-type affinity and contains modal red-brown biotite and garnet, and yet has relatively unevolved Sr isotopic characteristics. This unit is inferred to have formed from a more clay-rich sedimentary component in the heterogeneous magma source. The zircon age spectrum of HAR17 is relatively simple. It contains no strong Ordovician signature, but a moderate spike at 414 Ma. The rest of the pattern is dominated by two groups of peaks, with a prominent mid-Cambrian signal at 519 Ma. The largest peak then follows at 999 Ma (earliest Neoproterozoic) and this is accompanied by a lesser, late Mesoproterozoic peak at 1164 Ma. The rest of the spectrum consists of three small populations of Paleoproterozoic zircon at 2239, 2019 and 1704 Ma. There is no sign of a Selwyn Block signature, which seems a little curious because I-type sample AD1, from much further west in the Bendigo Zone, and outside the inferred subsurface limit of the Block, does have what might be interpreted as a strong Selwyn Block signal, with a prominent peak at 1404 Ma. Indeed, HAR19, the more Itype sample from the Baringhup pluton of the Harcourt batholith (see below), does have an age peak at 1449 Ma, compatible with a Selwyn Block source. Thus, the western edge of the Selwyn Block may have a complex, fragmented structure, as for example among the crustal blocks in the Grampians-Stavely and Glenelg zones, as modelled in Schofield (2018, figure 2.23).
Late Devonian Baringhup pluton-HAR19 (Figure 7b) This weakly peraluminous sample was collected 6.6 km west of the location of HAR17 but in a distinct pluton. It is less potassic than HAR17 and contains brown biotite and no garnet, and yet has one of the highest initial Sr isotope ratios in the batholith. This sample is considered to be more of an I-type than HAR17, a judgement borne out by the character of its zircon population, with fewer and smaller distinct cores present (see Supplemental data, Paper 3).  determined the age of this sample as 378.2 ± 2.0 Ma, which is at least 1 Myr older than the best constrained date for the adjacent Mount Alexander pluton (373 ± 2 Ma). Note, however, that the two dates are the same, within uncertainty, and we probably cannot resolve an age difference of less than about 7 Ma, in any case .
The zircon age spectrum (Figure 7b) is somewhat different to that of the more S-type HAR17 (Figure 7a). There is a very prominent age peak at 454 Ma, very much larger than in HAR17, which has only small peaks at 474 and 414 Ma. Also, there are no peaks at ages greater than 1634 Ma in HAR19, whereas HAR17 has several other Paleoproterozoic age peaks, ranging up to 2239 Ma. This could be regarded as the signature of a distinct difference in magma source, such that there is less metasedimentary heritage, which would also be consistent with HAR19 having more I-type character than HAR17. The rest of the pattern is even simpler than in HAR17, with prominent peaks at 1634, 1449, ca 1069, 799, 679 and 544 Ma. Note that, unlike HAR17, HAR19 does have a zircon signature compatible with Selwyn Block derivation of the magma (i.e. the presence of ca 1650 and 1400 Ma zircon). On the assumption that all the magmas of the Harcourt batholith were derived from the Selwyn Block, these relationships suggest that the particular sources of these two magma batches lay in different horizons (levels?) within the block, consistent with the substantial differences in their Sr isotope ratios calculated at a reference date of 370 Ma (HAR17 ¼ 0.70807 ± 17 and HAR19 ¼ 0.71366 ± 25; Clemens, 2018).
Late Devonian I-type Tullaroop Granodiorite-WV8 (Figure 7e) The zircon population in sample WV8 was analysed by Bierlein et al. (2001), who dated its crystallisation to 378 ± 2 Ma. We have taken the data presented in Bierlein et al. (2001) and used these to construct an age spectrum for its inherited zircon population (Figure 7e). Unlike all the other I-type rocks in the Stawell and Bendigo zones, WV8 appears to contain no inherited zircon population of 1100-1000 Ma age or in the range 1800-1700 Ma. Instead, this rock contains populations at ca 2200-2100 and ca 1270 Ma, which are absent from all the other igneous rocks of this zone. The implication of this is that the source rocks were distinctive for the Tullaroop magmas. There is also no sign of a shallow (Neoproterozoic to Cambrian) Selwyn Block signature and essentially no similarities with zircon populations in the Cambrian to Ordovician metasedimentary country rocks.

Devonian granitic and silicic volcanic rocks of the Melbourne Zone and the boundary with the Tabberabbera Zone
A-type rhyolite from the Toombullup Ignimbrite-TOOM ( Figure 8a) The published date for this unit is 377 ± 2 Ma . Many of the zircon crystals in the sample have distinct rounded cores with oscillatory-zoned rims (e.g. Supplemental data, Paper 3, Figure A3.13). However, this sample is unique among the igneous rocks in this study, in that it contains a single magmatic zircon population with no inheritance at all (Table 3). What appear to be rounded and potentially inherited cores returned the same ages as the clearly magmatic rims on the crystals. This is interpreted to mean that the cores are antecrysts-zircon crystals formed from the host magma system and later partially resorbed as the magma was processed at shallower levels, prior to eruption. In the case of the Toombullup Ignimbrite, thermobarometry has demonstrated that the magma was formed at a high temperature (>856 C), with early crystallisation at a pressure <500 MPa, followed by later magma evolution at <150 MPa and $800 C (Clemens, Birch & Dudley, 2011). It is a particular characteristic of high-temperature, ferroan, A-type magmas that they commonly carry no restitic materials from their source rocks. In the present case, this explains the apparent complete absence of inherited zircon. This is in keeping with the inferred high-temperature origins of many A-type granitic magmas, as well as the geothermometry of this particular rock (Clemens, Birch & Dudley, 2011;Clemens et al., 1986;Collins et al., 1982). The high temperatures of A-type magmas typically result in dissolution of any entrained zircon, with reprecipitation as purely magmatic crystals.
S-type ignimbrite from the Lake Mountain Rhyodacite-E27-2 (Figure 8b) This sample has a rather simple pattern of age peaks. The crystallisation age is 374 ± 2 Ma (unpublished data). In the age spectrum, there is a single strong peak at 424 Ma with a lower-amplitude hump that represents the presence of some Neoproterozoic and Cambrian to Late Ordovician zircon. There is a late Mesoproterozoic group at 1149-1084 Ma and a final late Paleoproterozoic peak at 1779 Ma. There is no Selwyn Block signature. The number of points analysed was 114, which is slightly suboptimal but Figure 8. Zircon age spectra (probability density distributions) for Devonian granitic and silicic volcanic rocks of the Melbourne Zone and the boundary with the Tabberabbera Zone: (a) sample TOOM from the strongly peraluminous, A-type Toombullup Ignimbrite; and (b) from the S-type Lake Mountain Rhyodacite (sample E27-2). Sample TOOM contains no inherited zircon, so the single peak illustrated represents the magmatic crystallisation age. The blue band at 1400 ± 50 Ma in (b) indicates the age that has been suggested as distinctive of zircon in the Selwyn Block, and the blue vertical line represents the Nd model age for a different sample (9399) from the same unit, calculated from data in Clemens and Birch (2012). See text for further details.
should not have caused these peaks to be missed. Although the sample was analysed for dating of the crystallisation event, many crystal cores were targeted for analysis, although many of these returned spot ages equivalent to the crystallisation age, again supporting an antecrystal origin (Supplemental data, Paper 3, Figure A3.14).
S-type monzogranite from the Mount Wombat pluton of the Strathbogie batholith-S1 (Figure 10a) The crystallisation age for S1 is 381.5 ± 2.6 . Even though this pluton hosted the metapelitic Selwyn Block xenolith (S12), the age peaks of these two samples are a poor match. Specifically, S1 has a major peak at 634 Ma, which is weak to absent in S12, the ca 900 Ma peak is strong in S1 and weak in S12, and the 1724 and 1274 Ma peaks in S1 are absent from S12, although S12 does have small peak at 1784 Ma. The S12 xenolith spectrum is then barren until the Paleoarchean peaks at 3534 and 3399 Ma appear, whereas the granitic rock has good Paleoproterozoic and Mesoarchean peaks at 3014, 2834 and 2444 Ma, but nothing older.
In agreement with Clemens and Phillips (2014) and Clemens and Buick (2019), this granitic magma was unlikely to have been derived from rocks similar in chemistry or mineralogy to the xenolith S12; compare also Figures 5c and 10a. Rather, it is most likely to have been derived from a deeper and older Selwyn Block source-most probably a less aluminous metagreywacke, with less evolved Sr and Nd isotope characteristics (e.g. Figure 12). Likewise, zircon derivation from the Silurian-Devonian country rocks intruded by the granitic magma seems highly unlikely. Although both rocks have peaks at around 1430 Ma (a possible Selwyn Block signature), overall there is a very poor correspondence between the spectra. In particular, the Silurian-Devonian metasediments have major zircon populations at 1789, 1644 and 544 Ma, none of which are present in the granitic rock's zircon population. Likewise, the granitic rock has a prominent zircon age peak at 979 Ma, which is absent from the country-rock metasediments.

S-type granodiorites from the Mount Disappointment nested laccolith-MD27 and MD50
Two related samples of rocks from this pluton were chosen for study because they have somewhat different mineralogical characteristics and chemistry. The crystallisation age was assumed to be ca 370 Ma (Rossiter, 2001)

MD27 (Figure 10b)
This spectrum has two very prominent peaks at 529 and 434 Ma. There is another prominent but low peak at 789 Ma and then a cluster centred on ca 1000 Ma. The pattern plotted using a 30 Ma bandwidth has peaks at 1049 and 959 Ma, in this cluster. In this filtered spectrum, further peaks occur at 1789, 1584 and 1449, and a small one at 3149 Ma. Apart from this last particularly ancient peak, the spectrum is generally similar to that of the Lake Mountain Rhyodacite, although the granite preserves much stronger inheritance and also displays the ca 1400 Ma Selwyn Block signature, which is absent in the ignimbrite sample. There are general similarities with the Melbourne Zone metasedimentary samples, and most particularly with the Silurian-Devonian, which indeed forms the pluton's wall rocks. The 1789 Ma peak, in sample MD27, is also prominent in the Silurian-Devonian rocks. However, there is no genuine match. For example, the prominent 1644 Ma peak in the Silurian-Devonian metasediments has no counterpart in the granitic rock. On the other hand, unlike the Strathbogie monzogranite (sample S1, described above), there is quite a good correspondence with the Selwyn Block xenolith (S12).  (Bierlein et al., 2001); (e) sample WV99 from the Pyalong pluton of the Cobaw batholith (Bierlein et al., 2001); and (f) a composite of three samples from the Oberon pluton of the Wilsons Promontory batholith in the Bassian Basement Terrane, with data for monzogranite 2340 and leucogranite W96H2 from Elburg (1996). Note that WPB7 is from a schliere hosted in a monzogranite from the same unit as 2340. Spectra shown in dark colours are the data processed using a bandwidth equal to 1r measurement uncertainties on 206 Pb/ 238 U spot ages. These are overlaid with lighter-coloured spectra calculated using a constant bandwidth of 30 Ma, as recommended for detrital zircon studies (Andersen et al., 2018). The pale blue band at 1400 ± 50 Ma indicates the age that has been suggested as distinctive of zircon in the Selwyn Block. The thin blue lines indicate calculated two-stage Nd model ages for the rocks, with data for 2340 and W96H2 from Elburg (1996). See text for further details.

MD50 (Figure 10d)
The age spectrum for this sample is presented directly below that for MD27, for ready comparison. The peaks around 530 and 430 Ma are present in both samples from Mount Disappointment. However, the peak at 789 Ma, although prominent in MD27 is entirely absent in MD50. Both rocks have clusters at around 1150-950 Ma but the 1314 Ma peak in MD50 is not present in MD27. Both samples have peaks at ca 1450 Ma, indicating probable Selwyn Block magma derivation. As can be seen in the plots, the older zircon populations of the two Mount Disappointment rocks differ from each other. Together with the other contrasts, this suggests magma derivation from a heterogeneous source within the Selwyn Block. Like MD27, MD50 has some age peaks similar to those of the Melbourne Zone Silurian-Devonian metasediments. However, the high-grade metapelitic sample from the Selwyn Block (sample S12, e.g. Figure 6c) also shares many zircon age peaks with MD27 and MD50. Indeed the similarity between S12 and the Mount Disappointment age spectra could be used to infer that the Mount Disappointment magmas originated through partial melting of Selwyn Block metasediments of similar type to this xenolith. On the other hand, the Sr and Nd isotope signatures of the rocks (Figure 12) rule out this possibility. Instead, in agreement with Clemens and Buick (2019), we suggest granitic magma derivation through partial melting of the deeper Paleoproterozoic or Mesoproterozoic sections of the Selwyn Block.
Late Devonian S-type Pyalong pluton of the Cobaw batholith-WV99 (Figure 10e) This pluton (marked 'e' in Figure 2) occurs as an outer sheath on the Cobaw batholith-a 'stitching pluton' that straddles the Heathcote and Mount William faults, which separate the Melbourne and Bendigo zones. The Pyalong pluton was dated to 376 ± 2 Ma by Bierlein et al. (2001). Using these data, we calculated the PDD plot of the inherited zircon population (Figure 10e). This age spectrum is rather similar to all the other zircon populations from S-type granitic rocks in the Melbourne Zone, except that it lacks any peaks around 1700 Ma and has a prominent peak at ca 2440 Ma. The only other S-type rock with a similar Paleoproterozoic population is S1 from the nearby Strathbogie batholith, although S1 has only a relatively small peak here. An additional signature of WV99 is the complete absence of a peak near 1500-1400 Ma, and therefore no zirconbased evidence for a Selwyn Block connection. In sample S12, from the Cambrian part of the Selwyn Block, the absence of peaks in the range of 2500-2400 Ma further underlines this lack of a connection to the shallow sections of the block.
Early or Middle Devonian S-type Bulla Granodiorite-CV108 (Figure 10c) This very small pluton (in bright green and labelled 'd' in Figure 2) is unique in that it has been dated to 392 ± 5 Ma (Bierlein et al., 2001), which would make it the only Middle Devonian pluton in the region. However, within the dating uncertainty, the Bulla Granodiorite could also be grouped, with the Wilsons Promontory batholith, as Early Devonian. Whatever the epoch to which we assign the Bulla Granodiorite, the U-Pb zircon data of Bierlein et al. (2001) can be used to construct the inherited zircon PDD age spectrum in Figure 10c. There are strong similarities with the age spectra for many of the other S-type granitic rocks. However, like WV99, from the Pyalong pluton, CV108 has no ca 1400 age peak, as would be expected for magmas derived through partial melting of the upper Selwyn Block. Equally unlikely is a significant contribution from the Melbourne Zone Silurian-Devonian metasedimentary rocks, with a major zircon age peak at ca 1790 Ma, which has no counterpart in the age spectrum for the Bulla Granodiorite.
Late Devonian I-type Lysterfield Granodiorite-L1 (Figure 9a) This hornblende-bearing I-type pluton intruded the volcanic rocks of the Mount Dandenong Igneous Complex and has Melbourne Zone Silurian-Devonian metasedimentary rocks in its contact aureole. The published biotite/ hornblende Ar-Ar date is 364 ± 7 Ma (Richards & Singleton, 1981). As Figure 9a, shows, the whole-rock Nd model age of 1.46 Ga lies just within the band of ages that are Figure 11. Cumulative histogram of the parameter CA À DA (¼zircon crystallisation ageÀsediment depositional age) for Ordovician greenschist-facies metagreywacke VH2 from the Bendigo Zone and amphibolite-facies metapelitic schist xenolith S12 from the Strathbogie batholith. Curves for various arcrelated settings are shown, as well as two horizontal red lines-stage 1 at 5% and stage 2 at 30% of the detrital zircon population. The positions and shapes of these curves suggest formation of both protoliths in backarc convergent settings; after Cawood et al. (2012). See text for further discussion.
thought to be most characteristic of a Selwyn Block contribution to an inherited zircon population. This is the igneous rock that contains the highest proportion of inherited zircon in the present dataset (59%; Table 4; Figure 13), and the age spectrum shows prominent peaks at 1744, 1219, 1144, 704, 629, 494 and 424 Ma. Neither of the peaks suggested to be diagnostic of Selwyn Block influence (1650-1600 and 1460-1380 Ma) is present, despite the very high degree of inheritance recorded in the sample. In relation to the local Silurian-Devonian metasedimentary rocks, only the rather small 1219 Ma peak may be shared, whereas the major peaks in the metasedimentary rocks (1789( , 1644( , 1054 are absent in the pattern for L1. The 504 Ma peak in the Melbourne Zone Ordovician metasediment may correlate with the 494 Ma peak in the granodiorite, but the prominent cluster (in the Ordovician metasediment) that crests at 1069 Ma does not appear to be present in L1. The Selwyn Block xenolith has essentially no peaks in common with L1, with the prominent peaks at 989 and 534 Ma unequivocally absent in the igneous rock. Despite the hefty population of inherited zircon crystals that was evidently carried by the L1 magma, these features suggest minimal to zero contribution from known metasedimentary rocks, of whatever age, in the region.
Late Devonian Monzogranite, Tynong pluton, Tynong batholith-TY2 (Figure 9b) This sample comes from the low-Al series of hornblende-biotite monzogranites in the Tynong pluton (Clemens et al., 2016), which intrudes lower Silurian to Lower Devonian quartz-rich metaturbidites of the Melbourne Zone (i.e. the kinds of Silurian-Devonian rocks whose zircon inheritance pattern is portrayed in Figure 6a). Regmi et al. (2016) published a U-Pb zircon date of 372.0 ± 3.4 Ma for a Tynong granitic rock and  dated the present TY2 sample to 365.8 ± 2.7 Ma, using a similar technique. This pluton has the least crustally evolved initial Sr-isotope ratio of any analysed from the geographical area covered by Figure  2 ($0.706), and the eNd value ($0) also points to an enriched mantle source, a crustal source with low crustal residence time prior to magma formation or a mixed mantle and crustal source. From Table 4 and Figure 13, it is readily apparent that the degree of inheritance in the zircon population of TY2 is relatively low, at 25%. However, this suggests that a pure mantle origin for the magma is unlikely; a substantial crustal component is more probable. Turning to the inheritance pattern (Figure 9b), this rock has zircon age peaks at 924, 848, 614, 534 and 429 Ma, none of which correspond with dates expected for Selwyn Block provenance and consistent with the relatively young whole-rock Nd model age of 1.13 Ga (Clemens et al., 2016). Also, in comparing Figure  9b with Figure 6a (for the Silurian-Devonian metasedimentary country rocks), it is apparent that there cannot have been any significant contamination of the magmatic population with zircon crystals scoured from the wall rocks. The same is true for the Melbourne Zone Ordovician (Figure 6b) and the Selwyn Block xenolith (Figure 6c). The crustal component of the magma that formed sample TY2 evidently had source rocks that are not represented by any of the Melbourne Zone Silurian-Devonian metasedimentary rocks or the known underlying upper (Cambrian) parts of the Selwyn Block.  Clemens (2020), and data sources from appendix 1 of that publication.
Early Devonian granitic rocks of the BBT Early Devonian S-type Oberon pluton-composite sample of leucogranite W96H2, monzogranite 2340 and mafic schliere WPB7 (PS) (Figure 10f) The crystallisation age of the Wilsons Promontory batholith was determined as 395 ± 4 Ma by Elburg (1996), who used the SHRIMP technique with zircon crystals separated from samples 2340 and W96H2. Using LA-SF-ICP-MS,  were also able to date the Oberon pluton of this batholith with zircon crystals from schliere WPB7 (also known as PS). The result was 397.2 ± 2.7, confirming the Early Devonian crystallisation age of the pluton. To examine the inheritance pattern, we combined the data from Elburg (1996) and our new analyses of zircon crystals in WPB7 (PS) because there are so few concordant inherited zircon spots in the Elburg (1996) dataset that it would be difficult to confirm any statistically significant differences between the datasets for these two rocks (e.g. Andersen et al., 2018). Note here that the leucogranite (W96H2), although part of the same pluton as the other two samples, is actually an independent and slightly older magma batch (Clemens & Wallis, 2022). In the composite dataset, the youngest inherited zircons form prominent peaks at 514 and 439 Ma, with a lower-amplitude shoulder at about 600 Ma (Figure 10). The rest of the inherited zircon crystals are mainly Mesoproterozoic and Neoproterozoic, with peaks at 1034, 809 and 719 Ma, a shoulder peak at about 1100 Ma, and relatively small peaks at 1664, 1484 and 1219 Ma. The overall impression is of weak inheritance of any zircons older than earliest Mesoproterozoic. The presence of the population at 1664 Ma may represent a Selwyn Block signature but there is no 1400 ± 50 Ma peak to accompany it, and this latter age is considered most characteristic of zircons derived from Selwyn Block sources. The match between this composite and the Selwyn Block xenolith (S12) is also very poor, again suggesting that, if the Wilsons Promontory magmas were derived from the Selwyn Block, the actual sources would have to be rather less crustally evolved (less pelitic), a conclusion borne out by the relatively low whole-rock eNd 370Ma values (-3.9 to À3.2; Table 2) compared with other S-type rocks in the region, and even many of the I-types.

Discussion
Do the inherited zircon age spectra of the igneous rocks share similarities?
From the above descriptions and Figures 7-10, it is clear that there are some shared age peaks among the igneous rocks. Both S-and I-type samples have strong zircon age peaks in the region of 420-410 Ma, which most likely record latest Silurian to earliest Devonian thermal events associated with the late, extensional phase of the Silurian orogenic cycles in the western Lachlan Orogen; see below. Also, among the I-type rocks, all have peaks in the range of 650-500 Ma. Although some pairs of I-type rocks have other shared peaks, there is little commonality in the rest of the spectra, and some samples have major peaks that are not reflected in any other sample (e.g. the ca 1140 and 1270 Ma peaks in the Lysterfield Granodiorite and the Tullaroop Granodiorite, respectively). Among the S-type rocks studied here, all have peaks in the range of 1060-950 Ma, similar to some zircon data from the Tasman Rise in Bass Strait (Moore et al., 2015), and five of the six have peaks between ca 1790 and 1660 Ma. However, like the I-type rocks, no other peaks are shared by more than two of the samples, and there are some rocks that have peaks not present in any others (e.g. ca 1100 Ma in the Mount Wombat pluton, ca 1580 Ma in Mount Disappointment pluton sample MD27 and ca 1220 Ma in the Oberon pluton). Indeed, the Pyalong pluton contains two quite unique peaks in its spectrum-ca 2579 and 880. In this context, it is also worth noting that the specific, minor, Neoarchean to Paleoproterozoic peaks that are present in several of the samples are not shared by any other sample. Overall, the zircon age spectra reveal few close similarities between different S-or I-type plutons. This suggests that each pluton was constructed from magmas that tapped different combinations of source rocks-a characteristic interpreted as a clear reflection of widespread source heterogeneity and original sediment source variability.

Degrees of inheritance and their variations
With regard to degree of inheritance, Table 4 shows that, as predicted, both I-and S-type rocks in the region can contain substantial proportions of inherited crystals. For each sample, the percentage inheritance was calculated from the number of 90-110% concordant 206 Pb/ 238 U spots with ages >30 Ma older than the crystallisation age, as a proportion of all similarly concordant spots. The results are illustrated graphically in Figure 13. The highest proportion of inherited zircon (59%) occurs in the strongly I-type, hornblende-biotite Lysterfield Granodiorite (sample L1). Of the plutonic rocks, the lowest proportions of inherited zircon crystals occur in the I-type Tynong pluton sample TY2 (25%) and the strongly S-type monzogranite from the Oberon pluton of the Wilsons Promontory batholith (13% for the ordinary rock and 19% for the mafic schliere). The Mount Oberon leucogranite in this pluton is a separate magmatic pulse (Wallis & Clemens, 2018). Although far more silicic (80.26 wt% SiO 2 ; Table 2), this sample has a much higher degree of inheritance (40%) than the monzogranite, which has 70.30 wt% SiO 2 .
These relationships run counter to the common perceptions that I-type rocks should contain less inherited zircon than S-types, and that zircon inheritance should be negatively correlated with SiO 2 content (e.g. Collins et al., 2020;Williams, 1995). This latter idea may relate to adherence to the restite-unmixing model for the origin of chemical variations in granitic rocks (see e.g. Chappell et al., 1987;Clemens, 1989;Wall et al., 1987;White et al., 1999). This model posits that most mafic minerals and 'mafic' enclaves in granitic rocks have restitic origins, and the logical inference from this would be that restitic zircons should be more abundant in the more mafic rocks. In some cases, this relationship may indeed hold but, as demonstrated by Villaros et al. (2009), the negative correlation between whole-rock Zr and SiO 2 contents in granites is most commonly due to co-entrainment of zircon and the peritectic mineral assemblage from the magma source. The present work also demonstrates that it is not always the case that I-type rocks have systematically lower degrees of inheritance than S-type rocks. As Kemp et al. (2005) pointed out, peraluminous I-type granitic rocks commonly had substantial supracrustal components in their source terranes, and Clemens (2018) presented an example of I-type rocks derived from a dominantly metasedimentary source. Nevertheless, our Lysterfield Granodiorite sample L1 is a borderline metaluminous, hornblende-bearing granodiorite, and yet has the highest degree of zircon inheritance in our dataset. It seems possible that the volcanic rocks have weaker inheritance than plutonic rocks in the same area. If this relationship were the case, it could potentially relate to higher temperatures in the volcanic magmas, although this would need a separate study to confirm. This is because we have very few data from volcanic rocks in the region, the analyses on the zircons from the volcanic rocks did not specifically target inherited cores, and one of the two volcanic samples (TOOM) is a high-temperature peraluminous A-type rock that completely lacks inheritance.

Archean inheritance
In addition to the percentages of inherited zircon analyses, Table 4 shows the proportions of Archean dates as percentages of total inherited zircons in each sample. The Archean dates mostly make up 0% of the age spectra and do not exceed 3.9% (or $1.8% of the total population). The only rocks with Archean inheritance are the I-type Ercildoun pluton (sample AD1) and the S-type Mount Wombat (S1), Mount Disappointment (MD27), Bulla (CV108) and Pyalong (WV99) plutons. As already mentioned, among the I-type samples, only AD1 (the Ercildoun pluton) contains Archean zircons, although this rock contains the highest proportion recorded in the present study-3.9% of the inherited population. Many of the S-type and transitional Stype samples (i.e. HAR17 and the Wilsons Promontory rocks, as well as the Lake Mountain Ignimbrite E27-2) contain no Archean zircons. However, other S-type rocks do have small Archean populations-Mount Wombat pluton S1 ¼ 2.1%, Mount Disappointment pluton MD27 ¼ 1.0% and MD50 ¼ 3.4%, the Bulla Granodiorite ¼ 3.4% and the Pyalong pluton ¼ 2.7%. These low levels of Archean inheritance bear out the inference of Moore et al. (2015) that few very ancient zircon grains should be present in any granitic rock crystallised from a magma that formed through partial melting of rocks in the Selwyn Block.

Igneous and metamorphic inheritance and ages of events
One of the interesting features of the inheritance in the granitic rocks surrounds the determination of the ages of the different igneous and metamorphic events recorded in the inherited zircon cores. We examined each of the inherited cores (as recognised above) to determine whether the laser-analytical spot sampled an igneous or a metamorphic grain. The commonly used chemical criterion for such a determination uses zircon Th/U ratios (e.g. Kohn & Kelly, 2018). However, for our dataset, this parameter displays no correlation with the textural criteria. We examined and enlarged the SEM CL images (e.g. Figure 4 and Supplemental data, Paper 3) together with reflected-light optical images of the crystals, taken after analysis, to identify the location of each analysed spot. If this analysis was within a core showing concentric oscillatory or sector zoning, we classified this as an igneous core. For cores that lack zoning or show weak, cloudy, patchy, mottled, wavy or planar zoning, we deemed the ages to be of metamorphic grains; see, e.g. Wu and Zheng (2004). We used the 90-110% concordant 206 Pb/ 238 U spot ages to create PDDs separately for the igneous and metamorphic grains in each igneous rock in our new dataset. The full procedure is given, the resulting PDDs presented, and descriptions provided in the Supplemental data (Paper 5).
Having produced these data, we can consider the proportions of igneous vs metamorphic inheritance and how these vary with total inheritance. The first important observation is that, with only one exception (Mount Disappointment sample MD50), zircon inheritance is dominated by metamorphic grains. Figure 14 shows the proportion of metamorphic cores (in %) plotted against total inheritance. As for total inheritance (Figure 13), the proportion of metamorphic inheritance is higher in the I-type samples, again underscoring their strong crustal connections.
One of the major implications of the mismatches between the age spectra for the pair of rocks from the Mount Disappointment pluton (i.e. MD27 and MD50, ca 2 km apart) and the composite to transitional S-and I-type characters of several of the batholiths in the region (e.g. You Yangs, Cobaw, Harcourt and Tynong) is that the crustal source rocks for the granitic magmas must vary considerably, in some cases on relatively small spatial scales (kilometres or less in 3D), and that that variation is preserved through pulsed delivery of relatively small volumes of magma to the growing plutons. This is already apparent from the plots in Figures 7-10 but becomes even more obvious in the PDD plots in the Supplemental data (Paper 5), in which the igneous and metamorphic spectra are distinguished. Note that strong, metre-scale isotopic heterogeneity has been documented in some granitic rocks (e.g. in the Dartmoor pluton in southwest Britain; Clemens et al., 2021).
Combining all the age peaks in the PDDs of the Supplemental data (Paper 5), and applying a nominal Figure 13. Bar chart showing the proportions (%) of inherited zircon present in each of the analysed samples, with I-type rocks in blue, S-type plutonic rocks in red and a single S-type volcanic sample (Lake Mountain Rhyodacite E27-2) in pink. The rocks are arranged such that inheritance proportion increases left to right. For spot 206 Pb/ 238 U ages that are 90-110% concordant, the proportions were calculated as % inheritance ¼ 100 Â (number of spot ages ! crystallisation age þ 30 Ma)/total number of spot ages. See text for data sources but note that the three I-type rocks with the least degrees of inheritance are the Tarnagulla pluton (two samples) and the Buangor Granodiorite, with data drawn from Bierlein et al. (2001). The sample with the highest proportion of inherited zircon is the I-type Lysterfield Granodiorite (marked L1) and the rock with the least inheritance, from our own sample set, is the Tynong Granite (marked TY2).
bandwidth of 30 Ma, we can identify the main igneous and metamorphic epochs recorded in the inherited zircon crystals from the central Victorian region. Table 5 shows the resulting list, with some attempts made to correlate these with known orogenic or magmatic events. Among the igneous events, the ca 1600 and 1400 Ma dates that are considered to represent a Selwyn Block signature do indeed appear in these processed and filtered data.

Origin of the ca 420 Ma inherited zircon population
The prominence of an inherited zircon population that dates to ca 420 Ma is one of the most striking features of the individual zircon age spectra of the granitic rocks (Figures 7-10), as well as the results of the combined data, as listed in Table 5. This latest Silurian to earliest Devonian peak is commonly present in age spectra of both the inherited igneous and metamorphic zircon cores in the granitic rocks (Supplemental data, Paper 5). If it were not for the absence of such a population in the Silurian-Devonian metasedimentary country rocks and the Cambrian, highgrade xenolith from the Strathbogie batholith ( Figures 5  and 6), this ca 420 Ma age peak might be construed as evidence for upper-crustal assimilation or the involvement of these rocks in the partial melting events that formed the Devonian host granitic magmas.
Given the evidence for the formation of the Devonian granitic magmas from Proterozoic source rocks in the Selwyn Block (e.g. Clemens & Buick, 2019 and references therein), the presence of latest Silurian to earliest Devonian zircon of partly igneous origin would seem paradoxical. However, the age of these zircon crystals corresponds to the time after the terminal phase of the Benambran Orogeny. If we assume that a late, extensional phase of the Benambran cycle was somewhat similar to that in the subsequent Tabberabberan, it would seem logical that crustal heating and partial melting would have taken place around this time. Recall that, from evidence in xenoliths from a granitic rock, Maas et al. (2001) identified a deep-crustal, amphibolite-facies, metamorphic episode associated with the terminal phase of the Benambran Orogeny further east in Victoria. Indeed, as mentioned above, we have outcropping rock evidence for silicic magmatism in the Early to Middle Devonian-the Wilsons Promontory batholith (11 in Figure 2), the Bulla Granodiorite (d in Figure 2) and the Hollands Creek Ignimbrite (near b in Figure 2). We suggest that earliest Devonian silicic melts were formed and a proportion of these failed to segregate from their source rocks (perhaps owing to insufficient thermal input from the mantle) to become migmatites. Such quartzo-feldspathic to pelitic migmatites may have had textural similarities with samples S12 and S9 (described by Clemens & Buick, 2019), Figure 14. Proportion of inherited zircon cores with metamorphic zoning textures plotted against the proportion of total inheritance for the 10 new samples. Note the preponderance of metamorphic heritage as well as the fact that the I-type rocks (in blue) show more inheritance and more metamorphic inheritance than the S-type rocks (in red). See the text for further analysis of these relationships. Events in italics are a possible cross-match. All assignments of dates to known orogenic events are tentative. a The A-type granite-rhyolite belt refers to the large silicic igneous province that developed in East Antarctica and Laurentia at 1.5-1.34 Ga (Figure 1).
which contain quartzo-feldspathic layers with igneous textures, as well as biotite-enriched melanosomes with more metamorphic textures. Later, in the extensional phase of the Tabberabberan Orogeny, there was evidently widespread and intense crustal melting that produced the numerous and voluminous Late Devonian silicic magmas (e.g. Clemens & Buick, 2019). Indeed, between relatively brief compressional episodes, the entire Lachlan Orogen was probably in a state of extension (e.g. Collins, 2002). We suggest that the thermal event responsible for the crustal melting also involved Early Devonian metamorphic and migmatitic rocks in the middle to lower crust. Inherited zircon from such sources would make up the prominent ca 420 Ma population that is present in nearly all the igneous rocks studied here, and this would also explain the presence of ca 420 Ma inherited zircon cores with both igneous and metamorphic textures. It could even be that the middle and deep crust in the region remained at high, near-solidus temperatures throughout the period from the terminal phase of the Benambran to the terminal phase of the Tabberabberan Orogeny. If that were the case, the Selwyn Block may not have behaved as a rigid entity but rather could have readily extruded at a high angle to the direction of compression during the Tabberabberan Orogeny. If this speculation were correct, it would also suggest that, owing to the rheological contrast, a decollement would have developed beneath the thickened Paleozoic cover rocks of central Victoria. There does indeed appear to be a basal detachment structure, known as the Thomas Fault, between the younger Paleozoic cover and the Selwyn Block. The present interpretation of the origin of this structure (e.g. Cayley et al., 2002Cayley et al., , 2011 is that it formed during tectonic reactivation that detached a thick and undeformed foreland basin sedimentary pile from an underlying and rigid Selwyn Block, causing folding and uplift as this Silurian-Devonian pile was thrust eastward against the backstop of the Mt Useful Fault Zone, on the western edge of the Melbourne Zone. Although it is conceivable that both reasons for the presence of this detachment are correct, in different stages of the Tabberabberan tectonic cycle, the rigid vs the plastic state of the Selwyn Block represents a fundamental divergence in the two models.

Comparisons with inherited zircon populations in Tasmanian granitic rocks
The study by Black et al. (2010) represents a valuable source of data on the age populations of inherited zircon in the Paleozoic granitic rocks of Tasmania. For inherited zircon in I-type rocks of eastern Tasmania the major age peaks are at about 560 and between 1100 and 1000 Ma. In western Tasmania, the I-type rocks have zircon with more complicated age patterns and major peaks at about 1850, 1800, 1690, 1590, 1160, 1070 and 400 Ma. The inheritance patterns in central Victorian I-type granitic rocks show some similarities with the western Tasmanian I-types, with major peaks at approximately 1160, 1070 and 400. However, in the Tasmanian rocks, the prominent peaks at about 1850, 1800, 1690 and 1590 appear to have no counterparts in the central Victorian I-type patterns.
Among the S-type rocks, the eastern Tasmanian examples have zircon inheritance patterns with major peaks at about 910, 580, 500, 430 and 400 Ma. These patterns share many similarities with the central Victorian S-types, particularly in the prominent ca 1060-940, 580-500 and 430-400 peaks. In western Tasmania the S-type rocks contain major zircon peaks at about 1720, 1620 and 1470 Ma, and the patterns share almost nothing in common with the central Victorian S-types. Thus, as far as inherited zircon patterns are concerned, the granites of central Victoria appear most akin to the eastern Tasmanian granites. It should also be noted that 1400 ± 50 Ma inherited zircon (the proposed Selwyn Block and VanDieland signature) appears to be present in around 27% of Tasmanian granitic rocks. This is quite similar to the proportion in central Victoria (35%), except that, in Tasmanian granites, most of the occurrences are in I-type rocks.
What determined the degree of zircon inheritance in the igneous rocks?
As we have described above, and illustrated in Figure 13, the degree of inheritance in the granitic and silicic volcanic rocks varies considerably. Compared with I-type rocks, one might expect that S-type granitic rocks should exhibit greater inheritance, but we have shown that this is not the case in the Devonian rocks of central Victoria. Indeed, as stressed above, two of the I-type samples in our study have higher degrees of inheritance than any of the S-types, and some S-type rocks have lower degrees than most of the I-types. We have discussed how magma (liquid) chemistry and temperature are important factors in zircon solubility, and how the rates of magma generation, transport and emplacement can govern the survival of entrained zircon grains.
As is apparent in Figure 15a, in general, the higher the Zr content of a granitic rock (former magma), the greater the proportion of inherited zircon will be present. The obvious exceptions are the Wilsons Promontory monzogranite and its Zr-rich schliere, which both have relatively low proportions of inherited zircon. Note that the leucogranite sample (W96H2) was not formed as part of the same magma that gave rise to the monzogranite and its garnetbiotite schlieren (Clemens & Wallis, 2022). Despite having a far higher Zr content than the host monzogranite, the schliere still has about the same degree of inheritance as its host rock. This suggests two things: first, as seems logical, that mechanical concentration of zircon crystals (e.g. during differentiation by crystal segregation) has little effect on the proportion of these crystals that are inherited grains, and second that fairly large increases in the numbers of analysed grains do not result in large increases in the inherited proportion. This second point is emphasised in Figure 15b, where it is apparent that there is little or no correlation between the number of zircon spots analysed and the degree of inheritance detected. Indeed, if one were to include a linear fit to the I-type data, a very weak negative correlation would result. From the previous discussion, this is unlikely to represent solely a preservation effect. Rather, it probably correlates with zircon abundance in the crustal source rocks that contributed to the magma volume and the efficiency with which accessory zircon grains are entrained into magmas escaping from their source rocks. This inference is supported by relationships in Figure 15c, which show that inheritance, most particularly in I-type rocks, correlates with FM (FeO T þ MnO þ MgO), which is taken as a measure of crystal entrainment efficiency during silicic magma formation (e.g. Clemens & Stevens, 2012;Clemens, Stevens & Farina, 2011). For completeness, it can also be mentioned that there are no correlations between degree of inheritance and Nd model age, initial Sr isotope ratio or ASI value (mol. Al 2 O 3 /(CaO À 3.33 P 2 O 5 þ Na 2 O þ K 2 O). Thus, the main influences are the abundance of zircon in the host rocks and the mechanical efficiency of entrainment of accessory (mainly restitic) zircon grains into the magmas. Finally, the lack of correlation evident in Figure 15b also suggests that it may be necessary to analyse only a few tens of grains to obtain a good idea of the zircon inheritance pattern in a granitic rock. This is a somewhat surprising implication that merits further investigation, but it is in line with the low probabilities of missing any particular inherited population with 'suboptimal' numbers of analysed grains (e.g. Table 3).
Where is the subsurface western edge of the Selwyn Block, and is ca 1400 Ma detrital zircon characteristic of the block and magmas derived from it? Cayley et al. (2011) interpreted a 2006 seismic survey to indicate that the western margin of the Selwyn Block dips at about 45 west, placing its western extremity at about 45 km depth, well to the east of the Mount Bute and Ercildoun plutons (numbered 2 and 3 in Figure 2). However, since the dip is only well constrained along a single seismic line, it is possible that the Selwyn Block boundary does not have a 45 dip along its entire western margin. The magnetic properties of the granitic rocks can provide an additional constraint on the subsurface extent of the block. There is a change from magnetic to nonmagnetic granites eastward in central Victoria. All the granites thought to be derived from the Selwyn Block are 'reduced' and have ilmenite as their sole Fe-Ti oxide mineral; the Ercildoun Granite is one of these 'nonmagnetic' plutons. However, granodiorites that crop out 10-20 km west of the Ercildoun Granite (e.g. the Mount Bute Granite-number 2 in Figure 2) are magnetite-bearing, and many produce conspicuous responses in magnetic surveys; see, e.g. Simons and . This suggests that, if the Selwyn Block does extend further westward than indicated in Figure 2, it is unlikely to be present much beyond the trace of the Avoca Fault (AF in Figure 2).
Given the possible geometric relationships just discussed, the Ercildoun Granite could have been derived from the westernmost part of the Selwyn Block, and this would explain the presence of ca 1400 Ma zircons in this pluton. As we have shown above, for the region as a whole, a 1400 Ma zircon population is a feature that occurs in some Devonian silicic magmatic rocks. Nevertheless, a majority of the Devonian silicic rocks analysed in the present work contain no inherited zircon of this age. The presence of ca 1400 Ma detrital zircon in the Selwyn Block xenolith from the Strathbogie batholith (sample S12) does imply that part of the Selwyn Block has this signature, and that some of the Late Devonian granitic magmas were probably derived from this part-specifically the Strathbogie batholith and the Mount Disappointment laccolith. At this point, it is worth recalling that the presence of 1400 Ma zircon in xenolith S12 and in some of the granitic rocks does not imply that these granitic magmas were formed by partial melting of the specific kind of rock represented by sample S12. This has already been ruled out on the grounds of chemistry and isotopic studies (e.g. Figure  12) and can be further excluded by the present study of zircon inheritance patterns. In sample S12, the youngest detrital zircon with 95-105% concordance has a 206 Pb/ 238 U spot age of 493.4 ± 7.6 Ma, so this xenolith seems to have been sourced in the latest Cambrian (uppermost) part of the Selwyn Block. In contrast, Clemens and Buick (2019) suggested that the Late Devonian S-type granitic magmas here were formed from the deeper Paleoproterozoic to Mesoproterozoic parts of the block. Source rocks formed in the Siderian to Calymmian periods should not contain 1400 Ma igneous zircon.
Given the above considerations, the overall conclusion is that significant parts of the Selwyn Block beneath central Victoria do not closely resemble the exposed parts of VanDieland in Tasmania and the Bass Strait islands, and do not contain rocks with a ca 1400 Ma zircon population derived from the 1.4 Ga, A-type granite-rhyolite province in East Antarctica and Laurentia. This does not mean that sediment sources for the Proterozoic rocks of VanDieland were not from these continental blocks but only that the 1.4 Ga granitic rocks were not always represented in the detrital load carried into the particular basins or sub-basins. This is all the more sensible if we consider that many of the Devonian granitic magmas are inferred to have formed by partial melting of the deeper sections of the Selwyn Block, which are older than 1400 Ma. In this connection, it is worth noting that the Nd model ages of the granitic rocks in central Victoria are mostly >1.4 Ga, with only the I-type Tynong sample (TY2) having a younger model age of 1.13 Ga. Several samples have borderline 1.46-1.40 Ga model ages-the I-type Lysterfield Granodiorite (L1) and the three S-type samples from the Wilsons Promontory batholith (2340, W96H2 and WPB7). However, even so, none of the samples with model ages of 1.40-1.13 Ga contain ca 1.4 Ga inherited zircons (Figures 9 and 10). Given that the inheritance patterns vary considerably among the granitic rocks, one further implication is that the sediment sources for the basin or basins in which the Selwyn Block protoliths were deposited most likely varied in both space and time, as suggested by Habib et al. (2022) for Proterozoic VanDieland in general.
Did Paleozoic country rocks in Central Victoria donate zircon crystals to the Devonian silicic magmas?
For the more easterly section of the Lachlan Orogen, Keay et al. (1999) concluded that, although there are some general similarities between the age spectra of inherited zircon in the Silurian granitic rocks and detrital zircon in the local Ordovician metasediments, the mismatches imply that the metasedimentary country rocks were not involved in magma genesis. However, these authors interpreted the presence of a ca 495 Ma inherited zircon population in the granitic rocks to mean that the granites were derived from sources younger than the late Cambrian. This could indeed be the case, if those 495 Ma (and younger) inherited zircon cores are igneous in origin, but the zircon textures do not seem to have been studied in detail, or at least any such data were not published. If these cores were instead metamorphic, such as the planar-zoned example shown in figure 3a of Keay et al. (1999), it would simply mean that a metamorphic event of that age had affected the source rocks, which could have protoliths of any older age. Highgrade regional metamorphism in the region, during the Wenlockian epoch (433-427 Ma), was responsible for the formation of the Omeo Metamorphic Complex. Thus, inherited metamorphic zircon of this younger vintage might also be present in Silurian granitic rocks. The possible involvement of relatively young Paleozoic crust in the genesis of the granitic rocks led various workers to suggest two-or three-component source-or magma-mixing models for the origins of the Silurian granitic magmas (e.g. Gray, 1984;Gray & Kemp, 2009;Keay et al., 1997).
From our descriptions and Figures 5-10, it would seem that the variety of solid wall rocks around and beneath the various igneous bodies contributed very little or nothing to the zircon populations in the Devonian silicic igneous rocks of central Victoria. Thus, we conclude that the inherited zircon populations in our study rocks reflect the zircon populations in the magma source rocks, in the deeper crust, as modified by dissolution or recrystallisation in the magmas, during ascent and emplacement. This means that, with minor caveats, we can use the data to determine some aspects of the magma source and its potential relationships to neighbouring terranes. In that regard, we suggest that two-or three-component source-or magma-mixing models that involve the local Paleozoic metamorphic rocks, such as have been suggested for Silurian granites further east, would not be appropriate for Devonian central Victoria.

What lies beneath the Bendigo and Stawell zones?
It is widely accepted that a wedge-shaped Proterozoic to Cambrian Selwyn Block underlies the whole of the Melbourne Zone and also extends some distance to the west, beneath the eastern edge of the Bendigo Zone, eventually thinning to zero thickness at depth, along the westdipping thrust plane that separates the two zones. The standard interpretation of the basement geology beneath most of the Bendigo Zone and all of the Stawell Zone, to the west, is that they consist of Cambrian deep-water turbiditic metasediments underlain by a thick sequence of Cambrian metabasaltic and metaboninitic rocks. However, using the chemical and isotopic characteristics of the Devonian granitic rocks of the Stawell and Bendigo zones, Clemens (2020) inferred the presence of a significant layer or at least slivers of heterogeneous, backarc, metasedimentary and metavolcanic rocks of Proterozoic to Cambrian age. Modelling showed that the granitic rocks of these zones could not have been produced by any combination of known mantle and crustal source rocks, be it through source mixing or magma mixing. It was also deduced that these basement units would not represent an extension of the Selwyn Block, since the granitic rocks of the Stawell and Bendigo zones have chemical and isotopic features that distinguish them clearly from their temporal equivalents in the Melbourne Zone. Instead, the basement in the Stawell and Bendigo zones was suggested to contain thinned crust, perhaps in isolated slices, that could be derived from the craton to the west.

Conclusions and implications for continent assembly and evolution
From findings above, we derive the following conclusions. 3. Based on the compositions and phase relations of the granitic magmas, together with their isotopic characteristics, the most abundant source rocks for the S-type magmas are likely to be metagreywackes deposited in a backarc setting, with the crustal components of the I-type magmas coming from associated dacitic and andesitic igneous or volcaniclastic rocks. 4. The zircon inheritance patterns of the S-type granitic samples provide evidence of thermal events between 525 and 425 Ma and in the range of ca 1200-1100 Ma. This second age range also forms a minor component of the age spectrum of the late Cambrian, Selwyn Block, metasedimentary enclave from the Strathbogie batholith. 5. Both S-and I-type samples have prominent zircon age peaks in the region of 420-410 Ma, which most likely record latest Silurian to earliest Devonian thermal events associated with and postdating the Bindian Orogeny. All I-type rocks have peaks in the range of 650-500 Ma reflecting probable derivation from metavolcanic sources of this age, most likely formed in an arc-related environment. However, since such volcanic rocks do not appear to be present among the Neoproterozoic rocks of the Selwyn Block, the only prospective source must be the uppermost, Cambrian section of the block. The ca 504 Ma Mount Read Volcanics in western Tasmania do not appear to contain suitable rock types, as a considerable volume of andesitic to dacitic rocks would be required to produce the Devonian I-type magmas. Thus, it seems that some extant inferences regarding the rock units in the Selwyn Block may be erroneous, even for its Cambrian section. 6. From the geological relations and their zircon inheritance patterns, it is likely that the protoliths of the greenschist-facies Ordovician metasediments and the amphibolite-facies Cambrian metasedimentary enclave were deposited in the distal backarc of an Andean-type margin. Together with the observations above, this suggests that the Cambrian section of the Selwyn Block must contain substantial volumes of intermediate metavolcanic rocks of arc affinity. 7. The degrees of inheritance among the zircon populations of the Devonian silicic igneous rocks vary widely and do not correlate with the S-or I-type character of their hosts. Indeed, both the lowest and the highest degrees of inheritance are found in strongly Itype granitic rocks. This emphasises the largely crustal origin of most I-type magmas in the region. 8. Zircon populations of ca 1400 Ma, thought to signal sediment derivation from East Antarctica and Laurentia, are mostly absent in the I-type samples and also some of the S-types. However, this zircon inheritance signal probably only applies to the upper parts of the Selwyn Block. This is interpreted to mean that the Devonian S-type granitic magmas in central Victoria were mainly produced through partial melting of the deeper sections of the Selwyn Block, which have no known surface expression in southern Victoria, the Bass Strait Islands or in mainland Tasmania. 9. Neither the S-type nor the I-type rocks contain zircon age populations that bear close resemblances to any of the Cambrian to Lower Devonian metasedimentary rocks that they intrude. Thus, these metasediments were not involved in the genesis of the granitic magmas and nor did they contaminate the magmas with measurable amounts of wall-rock-derived zircon xenocrysts. 10. Samples from chemically and isotopically distinct parts of the Harcourt batholith and the Mount Disappointment laccolith have different patterns of zircon inheritance. This suggests derivation of the magmas from source horizons that are heterogeneous on a spatial scale of <22 km in the horizontal direction in the case of the Harcourt batholith and <2.4 km in the case of the source crust beneath the Mount Disappointment laccolith. 11. What appear to be inherited cores in many of the zircon crystals have the same ages as the previously published crystallisation dates for the host rocks, and so represent antecrysts. This conclusion, from , is emphasised by the fact that sample TOOM, a peraluminous A-type rhyolite from the Tolmie Igneous Complex, has zero inheritance and yet also contains distinct, rounded cores in many of its zircon crystals. 12. Magma chemistry and temperature are important factors in zircon solubility, and the rates of magma generation, transport and emplacement govern the survival of entrained zircon grains. All these factors played roles in determining the degrees of inheritance in the zircon populations of the Devonian granitic magmas. In general, the higher the Zr content in a rock, the higher will be its proportion of inherited zircon, but mechanical segregation of zircon during any differentiation process does not affect the proportion of inherited grains. This general rule suggests that the degree of partial melting of the magma source rocks and the efficiency of crystal entrainment are critical in governing zircon inheritance patterns. 13. Given the presence of ca 1400 Ma zircons in some Itype rocks in the Bendigo Zone, the edge of the Selwyn Block may lie somewhat further west of the geophysically constrained limit. However, as suggested by Clemens (2020), many of the granitic magmas that intruded the Stawell and Bendigo zones are probably derived from an undocumented Proterozoic block or slivers of Proterozoic rocks lying beneath the known Paleozoic rocks of the region.
supplied the analyses of the zircon crystals in sample E27-2, which was originally provided by Dr Bill Birch (Museum Victoria). Dr David Moore (Monash University) alerted us to the existence of published zircon data for the Paleozoic metasediments of the Stawell, Bendigo and Melbourne structural zones. JDC also acknowledges many instructive e-mail exchanges with David. We also thank Fernando Bea and David Moore for careful and helpful reviews of the original manuscript.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Data availability statement
Files of sample preparation techniques, analytical data, descriptions of zircon crystals and illustrations of igneous and metamorphic inheritance proportions are included as supplemental data.