Records of the Western AlIstralwll MllSClIlI1 Supplement No. 57: 139--150 (1999).

Preliminary report on the biostratigraphy of new placoderm discoveries inthe Hervey Group (Upper Devonian) of central New South Wales

Gavin C. YoungDepartment of Geology, The Australian National University, Canberra, ACT 0200;

email: gyoung@geology.anu.edu.au

Abstract - Re-mapping during 1994-96 of the Upper Devonian on the Parkesand Grenfell 1:100,000 map sheets was the first systematic field examinationof this area by an experienced vertebrate palaeontologist. This resulted in thediscovery of some 40 new fossil localities (mostly fish) at various horizonswithin the Hervey Group. Two new fish occurrences, in sediments associatedwith the underlying Dulladerry Volcanics, require an age revision from Earlyto late Middle Devonian for the termination of Devonian volcanism in centralNew South Wales. The age of the Canowindra fossil fish fauna is uncertain,but may be as old as late Frasnian, rather than Famennian as was previouslyassumed. Correlations across the Lachlan River are clarified, and a newlocality for the sinolepid antiarch Grenfellaspis is reported from near the top ofthe Bumberry Formation. A fish fauna from higher in the sequence includesan antiarch with an armour 1-2 m long, of comparable size to similar materialfrom South China, which supports a latest Devonian Asian connection withEast Gondwana. Preliminary analysis of faunal content is used to propose aprovisional succession of six faunal zones, to improve the resolution on theLate Devonian part of the macrovertebrate zonation for East Gondwana.

INTRODUCTIONFossil fish remains were first reported from the

Hervey Group of central NSW. over 60 years ago(Hills 1932), but systematic studies of the fishfaunas have only recently gained momentum. Hills(1932, 1936) recorded isolated plates and spinesincluding placoderm remains (fragments of aphyllolepid; antiarchs Bothriolepis and Remigolepis)from the Hervey Range, southeast of Peak Hill(locality 2, Figure 18), and a "large bothriolepidplate" from the Jemalong Range, west of Forbes. Hedid not formally describe any of the material asexisting or new species. The first species describedwas the 'crossopterygian' (tetrapodomorphsarcopterygian) fish Canowindra grossi published byThomson (1973), based on a single specimen from aremarkable sandstone slab covered in fossil fishimpressions discovered near Canowindra over 50years ago (Fletcher 1956; locality 5, Figure 18).Much additional material has recently beenobtained from a new (1993) excavation of this site(Cribb 1996). Several new sarcopterygians havebeen described (Ahlberg and Johanson 1997;Johanson and Ahlberg 1997, 1998), and Johanson(1995, 1997a) has described the antiarch Remigolepisfrom the Canowindra fauna.

In the last 20 years two other significant fossil fishlocalities were found in the region, at RedcliffMountain about 15 km northeast of Grenfell, and atJemalong Gap, about 32 km southwest of Forbes.These are respectively 35 km southwest and 80 km

west of the Canowindra locality (Figure 1). TheGrenfelllocality (8, Figure 1B) was first reported tocontain the arthrodire Groenlandaspis (Ritchie 1975),but is now known to have a diverse fossil fishfauna. From this locality Ritchie et al. (1992)monographed a remarkable antiarch, Grenfellaspisbranagani, belonging to the endemic Chinese familySinolepidae, and Johanson (1997b) has recentlydescribed new species of the antiarchs Bothriolepisand Remigolepis. Various acanthodian andrhipidistian remains also occur in the fauna.Sinolepid antiarchs are a major placoderm fishgroup otherwise only known from the South andNorth China blocks, so this occurrence in theLachlan Fold Belt of eastern Australia hasbiogeographic significance (Young 1990; Rich andYoung 1996; Young and Janvier in press). TheForbes locality (3, Figure 1B) has produced thellmgfish Soederbergia, and a lower jaw of an earlyamphibian (tetrapod), Metaxygnathus, (Campbelland Bell 1977, 1982), plus undescribed remains ofthe placoderms Bothriolepis, Remigolepis,Groenlandaspis and Phyllolepis (the lung fishSoederbergia has also recently been found in theCanowindra fauna; Z. Johanson, pers. comm. 1997).

Until recently, geologic knowledge of the UpperDevonian sediments of central NSW. was basedon field mapping undertaken over 30 years ago byJ. ConoIly, who noted the occurrence of "fishplates" at various horizons, but based correlationsbetween different outcrops primarily on lithology

140

(Conolly 1965; Conolly et al. 1969). He referred theCanowindra fish site to the Mandagery Sandstonewithin the Hervey Group, but many of his other'fish plate' occurrences are doubtful. In 1989 theauthor and Dr Zhang Guorui, a fossil fish expertfrom the Institute of Vertebrate Palaeontology andPalaeoanthropology (Beijing, China), visited anumber of Conolly's putative localities but no fossilfish remains were found. Apart from fish, thelycopod plant Leptophloeum australe, and one marineinvertebrate locality (Etheridge 1901) were the onlyfossils recorded from the Hervey Group at thattime. Williams (1977) reported a new marineinvertebrate locality with lingulid brachiopods andbivalves (associated with 'fish plates') in thesouthern part of the Parkes Syncline, and Pickett(1993a) described a single specimen of a horseshoecrab Kasibelinurus from the Bumberry Syncline eastof Parkes (M,K, Figure 18).

Campbell and Bell (1977) were the first tocompare the fossil fish assemblages at differentlocalities in an attempt to integrate biostratigraphicevidence with the lithological correlationspreviously used (Table 1). They disagreed withConolly's (1965) correlation of the WeddinSandstone in the west with the MandagerySandstone further east. More recently, Ritchie et al.(1992) noted the significant differences in thecontent of the Grenfell fish fauna, and suggestedthat it was also considerably younger than indicatedby lithological correlation. The earlier correlationsof Conolly (1965) placed the Grenfell fauna at thetop of his Hunter Siltstone, and stratigraphicallybelow bluff-forming sandstones identified as theMandagery Sandstone, the same formation thatcontains the Canowindra fish site. In a recentdiscussion of problems in correlation raised by theage assessment of Ritchie et al. (1992), Pickett(1993a: 282) concluded that "there can be no doubtof the Famennian age of the Mandagery Sst and atleast some of its overlying formations". At the sametime, Young (1993) presented an analysis of themacrovertebrate fish faunas of East Gondwana, andattempted a preliminary faunal zonation whichrecognized 15 distinct assemblages through theDevonian, with the Grenfell fish fauna representingthe youngest known Devonian macrovertebrateassemblage from the non-marine sequences ofeastern Australia.

In 1994 the Australian Geological SurveyOrganisation collaborated with the N.s. W.Geological Survey under the National GeoscienceMapping Accord, to initiate a field mappingprogram to revise the Forbes 1:250,000 geologicalmap, the first edition of which was published in1972. This map sheet includes the Grenfell andForbes fossil fish localities, with the Canowindrasite just off its eastern margin on the adjacentBathurst 1:250,000 sheet (Figure lA). During 1994-

G.c. Young

96 the author was responsible for re-mapping ofConolly's (1965) Upper Devonian stratigraphy ofthe Hervey Group on two of the six 1:100,000 mapsheets (Parkes and Grenfell) covering this area. Thiswas the first systematic field examination by anexperienced vertebrate palaeontologist of theseUpper Devonian strata. The Parkes sheet includestype sections for Hervey Group formations (e.g.Mandagery Sandstone of the Canowindra fauna),and the locality for Pickett's (1993a) horseshoe crabKasibelinurus (T,K, Figure 1B). The Grenfell sheet,with over 400 krn2 of Hervey Group exposure, hadyielded the sinolepid Grenfellaspis fish fauna, asmentioned above.

As a result of the 1994-96 field survey some 40new fossil localities were discovered at varioushorizons within the Hervey Group. Most of thesecontain fossil fish, in some cases associated withplant remains, and at four localities (at least twodifferent horizons) fish occur with linguloidbrachiopods, and other invertebrates includingbivalve molluscs, suggesting a marine influence. Afew other localities yielded only plants orinvertebrates. In addition many examples of fossiltracks and trails were recorded, and other evidenceof biological activity is often seen at certain horizonsin the Hervey Group (e.g. abundant burrowing andbioturbation of many fine-grained horizons).

This large collection of new fossil material has notyet been studied in detail. Material from mostlocalities has been curated, and some has beenprepared and briefly studied. However apreliminary assessment of faunal content, based onobservations in the field, and/or duringpreliminary curation, has implications for thestratigraphy and correlation of the Hervey Group,and for macrovertebrate biostratigraphy andzonation as previously outlined (Young 1993, 1995;Young and Laurie 1996). The biostratigraphicinformation from these new fossil localities willprovide a completely independent dataset withwhich to test previous lithological correlations ofsequences in different areas, and in this paper Isummarize some preliminary results. It is stressedthat the mapping programme is not yet completed,and all biostratigraphic conclusions are provisional,requiring support from detailed taxonomic study ofthe new fossil assemblages.

Three aspects of Hervey Group stratigraphy onwhich the new fossil fish discoveries have had amajor impact are briefly summarized below: (a) thepreviously assumed Middle Devonian hiatus, theage of underlying volcanics, and the termination ofvolcanism in central N.S.W.; (b) the age of theCanowindra fish site; and (c) correlations across theLachlan River, and the age of the Grenfell fishfauna. The new biostratigraphic information is thensummarized in a series of five provisional faunalzones which are related to an update of the

New South Wales placoderm biostratigraphy 141

(

A

::f.?::·.·t~::·::·ij·:·::·: .--6 Canowindra1__ '" Sydney

-t--I-Grenfell-Canberra

'-

rForbes~-+--+--.J-4J

~-

*1 148.30E

.PEAK HILL

*2NARROMINE

1:250000

148.00E---4~."...,....,.-.-.~....,..,..~-- 33.00S

p:arne:$::··$h~~t:·:::::··::::

::::::::::::::::::::::::::·::::·::·?·:::::\::\:::::·::::~:R:::::::::::~::::..............PARKES·.:..·:

B

FORBES1:250000

BATHURST1:250000

* FOSSIL FISHLOCALITY

• COUNTRY TOWN • COWRA

20 km33.30S

1 A. of central NSW. including fossil localities mentioned in the text, shown against the 1:250,000 mapsheet The four relevant geological maps are Narromine (N), Dubbo (D), Eathurst (E), and Forbes B.Parkes and Grenfell 1:100,000 map sheets (eastern third of the Forbes 1:250,000 sheet) showing importantfish fossil (numbered) and other localities in the Hervey Group. 1, Tomingley; 2, Gingham Gap, HerveyRange; 3, Jemalong Gap, Jemalong Range; 4, Nanami Road; 5, Canowindra fish site; 6, Merriganowryquarry; 7, Fragar's property; 8, Redcliff Mountain; K, Kasibelmztrlls locality of Picket! (1993a), EumberrySyncline; T, Mandagery-Eumberry-Eurow Formation type sections, Parkes Syncline; M, marine lingulidlocality, southern Parkes SYncline (Williams 1977); L, lingulid locality at Nyrang Creek (Etheridge 1901).

142 G.c. Young

Table 1 Stratigraphic nomenclature for the Hervey Group, after Conolly (1965).

HERVEY PARKES GRENFELLSYNCLlNE SHEET SHEET

COOKAMIDGERABurrill Fm

undiff.undiff.

SUBGROUP Eurow Fm *

Caloma Sst. Bumberry Fm * Bumberry FmNANGAR Pipe Fm Pipe Fm

SUBGROUPMandagery Sst. Mandagery Sst.* Mandagery Sst.

BEARGAMIL Kadina Fm Kadina Fm Hunter Sist.SUBGROUP Clagger Sst. Peaks Sst.

* type section in the Parkes Syncline, Parkes Sheet

macrovertebrate zonation previously published(Young 1993, 1995). The following stratigraphicsubdivisions of the Hervey Group (oldest toyoungest) are used in the text, as defined orrecognized through the type sections in the ParkesSyncline by Conolly (1965): Kadina Formation,Mandagery Sandstone, Pipe Formation, BumberryFormation, Eurow Formation. Existing stratigraphicnomenclature is summarized in Table l.

Conolly (1965) mapped the Hervey Group largelyby recognizing as formations the prominent strikeridges of sandstone separated by valleys ofrecessive finer-grained units. However, as pointedout by Sherwin (1973, figure 7), the prominence ofsuch features depends largely on overall relief, andis often an unreliable basis for correlation. A goodexample is seen in the western limb of theBumberry Syncline east of Parkes, on the main roadto Molong, where high relief has obscured thenormally recessive Pipe Formation. This is theimportant fossil locality (K, Figure 1B) whichyielded the lycopod Leptophloeum australe, and thesingle specimen of the horseshoe crab Kasibelinurus(Pickett 1993a). According to existing geologicalmaps this locality lies well within the 'MandagerySandstone' of Conolly (1965). New detailed fieldwork, supported by airborne geophysical surveys,shows that lithological correlation failed over adistance less than 20 km from the type section inthe Parkes Syncline. In the Bumberry Syncline thePipe Formation is clearly recognized onradiometries and lithology, and the MandageryFormation is much less extensive than mapped byConolly (1965). Much of the 'Mandagery' ofprevious maps is in fact the Bumberry Formation,including the sandstones beneath the dam wall ofLake Endeavour, which were illustrated by Conolly

(1965, plates 2-3) as sedimentary structures fromthe Mandagery Formation.

NEW RESULTS

Middle Devonian hiatus and termination ofvolcanism in central N.S.W.

Two new fossil fish occurrences associated withvolcanics underlying the Hervey Group havechanged the accepted view of Devoniansedimentation and volcanism in central N.5.W.Previously, the Upper Devonian was assumed tounconformably overlie volcanics and sediments ofSilurian-Early Devonian age, with the MiddleDevonian represented by a hiatus in sedimentation(Conolly 1965; Packham 1969; Figure 2A). Webby's(1972, table 2) synthesis of the stratigraphy of theLachlan Fold Belt showed the Middle Devonian asa time of "major orogenesis with main folding,uplift and granite emplacement", with volcanismterminating in the Lochkovian ('Gedinnian') on theParkes Platform to the west, and as young asEifelian (Winburn Tuff) on the Capertee Rise to theeast. An unconformable relationship between theHervey Group and underlying Early Devoniansediments and volcanics is established in somesections (Webby 1972; Powell et al. 1980).

A major underlying volcanic unit in the mappedregion, the Dulladerry Volcanics, outcrops over alarge area at the intersection of four 1:250,000 sheets(Narromine, Dubbo, Forbes, and Bathurst; FigurelA), and has yielded a single fish plate, associatedwith plant remains (O'Sullivan 1992). The type areafor the Dulladerry Volcanics (first description bySavage 1968) is the large outcrop on thenorthwestern corner of the Bathurst sheet, which

New South Wales placoderm biostratigraphy 143

A B

W..JC hiatus

I---I-~1-----------".. A,.,,. A A.

,. ,., ".. A"" ,.,. A A A "'" "" A

... A A ... ,. ,. "'""" A A A,.,. A

A A A A A "" ".A A A""" A ,.

A ,. A A A A ,.

,. A ... A ... '" ""... ... ... ".. ,. " A

,. ... ,. ,. h ,. A

A '" A"..,. ,. '"... ,. ,. ,. .... ,. ,.A A ... ,. A '" A

A A,.", A ... A­

A A A ,. A A AA ... A A A ,. A

... '" A"""" ,. ....... "'" ... ... ... A A

A ,.. '" '" ,. '" ,.... h A A ,. A A

"" ,. A A ,. '" A""'AAAAAA

... A A ,. A ,. A,. A A ,. ... A A

,. A A ,. A ... A"" ,. A A ,. A A

'" ... ,. ,. A ,., ,.... "" ,. ,. ,. ,. ,.A""""AA"''''AAA.AAAA

"" ,. ,. A ,. ,. A

,. "" A""'" ,. "",., "" ,. ... ,. ,., tI'o,. ,., ,., ,. ,. ,. ,.'" A ,. A A ... ,.,

A A A ,. A ,. ,.,

AAAAA,.,.

A A A A A A ...A A A ... A ,. A

,., ,. ,. A A ,., A

... '" '" ... ,., A ,.,., ,. A ,., ,. A

,. ,. ,. ,., ,., ,. ""

Eurow Fm Q. Q.::J ::J0 0

Bumberry Sst. r::r: r::r:C' C'> >

Pipe Fm w .............. w> ------- ....... >r::r: r::r:

Mandagery Sst. w W::I: ::I:

Kadina Fm

w................«..J

z~Zo>wc

SILURIAN I ? ?

Figure 2 Previous (A) and revised (B) stratigraphy resulting from the discovery of a fish/plant assemblage of probableGivetian/Frasnian age in sediments associated with the Dulladerry Volcanics.

extends onto the southeastern part of theNarromine sheet, where it underlies the HerveyGroup on the western limb of the Hervey Syncline(Sherwin 1996). At the northern end of the HerveyRange, near Tomingley (1, Figure 18), an epithermalsystem within the Oulladerry Volcanics (Geopeko'sGlen-Isla prospect) was studied by O'Sullivan(1992), who recorded a "plate fragment from thetrunk shield of an unidentified Oevonianarthrodire" from a sedimentary interbed ofmudstone, conglomerate and chert (sinter) enclosedwithin coarsely phyric and flow-banded rhyolites.The associated plant remains were provisionallyidentified by Pickett (1993b) as 'Protolepidodendron'yalwalense, a form previously recorded fromsoutheastern New South Wales (e.g. Fergusson et al.1979).

Upon examination in Canberra, it was clear thatthe fish specimen found by O'Sullivan was in facta dermal bone impression of a phyllolepidplacoderm, a group first recorded fromsoutheastern Australia by Hills (1929), andsubsequently found there at many localities(Young, in Fergusson et al. 1979; Long 1984, 1989;Ritchie 1984), as well as in Queensland, theGeorgina and Amadeus Basins of central Australiaand in southern Victoria Land, Antarctica (Young1985, 1988a 1988b, 1989). This placoderm indicatesan age of late Middle Oevonian or younger (Young1993).

In early 1993 another new Oevonian fish localitywas discovered by Or A. Ritchie in a quarry atMerriganowry, about 120 km southeast ofTomingley (6, Figure 18). This outcrop, previously

144 G.c. Young

AMERRlGANOWRY SECTION

undiff.

Mandagery Sst.

BBELUBULA RIVER SECTION

undiff.

EurowFm

BumbenyFrn

Kadina Fm

MandagerySsl.

Canowindra fish faunaMerriganowry fish/plant

horizon

•PeaksSst

r - ~------------....-- .....I&--_------tI.~ /-.........__.........-.,,-I

Pipe FmHunter Sltst.

: I-- ~ .......~-----....., ....... I- -~...:!t:::::__=_

IIIIIII

Grenfellaspis fish fauna

Figure 3 Comparisons of two sections through the M-V Devonian Hervey Group, with suggested correlations acrossthe Lachlan River. Formation names are those used by Conolly (1965), and indicate his correlations. A,Merriganowry section, south of the Lachlan River; B, Belubula River section, including the Canowindrafossil fish site. The Grenfellaspis fish fauna has not been located in the Merriganowry section. L, assumedapproximate stratigraphic horizon of the lingulid locality at Nyrang Creek, 8 km along strike to the north ofthe Canowindra fish site (L, Figure IB).

mapped as Canowindra Volcanics (Silurian), alsocontains abundant remains of a new phyllolepid,associated with similar lycopod stems ofProtolepidodendron type, and branching stems likethose referred to Praeramunculus alternatiramus byMcLoughlin and Long (1994). This associationindicates a comparable age to the Tomingleyoccurrence. The locality is a black shale depositclosely associated with underlying volcaniclasticsincluding rhyolite blocks and possible flows. Theshale includes slumped beds and graded interbedsof rhyolitic sand up to about 30 cm thick, suggestingdeposition in a lake of sufficient size and depth forturbidity currents to be generated, with waterdepths of tens to hundreds of metres, perhaps anelongate deep lake such as are associated withstrike-slip faulting (P. O'Brien, pers. comrn. 1994).The flow-banded rhyolites in the underlyingvolcanics at this locality are typical lithologies for theDulladerry Volcanics (0. Raymond, pers. comrn. 1995).

On the western limb of the Hervey Syncline thecontact between volcanics and the overlying HerveyGroup is often faulted, so it is not clear from field

evidence whether there was a depositional break,but if so, it was probably of short duration (L.Sherwin, pers. comm.). At Merriganowry, fieldmapping in the surrounding area has shown thesandstone ridges about 1 km to the northwest tohave the same dip and strike as the black shale inthe quarry, and there is evidently a conformablesequence from the underlying volcanics andMerriganowry Beds in continuous section to thewest, where they underlie strata mapped as PeaksSandstone (Table 1).

Previously the Dulladerry Volcanics was dated asno younger than Early Devonian, based onconodonts from near Murga (Molong 1:100,000sheet) on the eastern limb of the Eugowra syncline,which were determined by Pickett (1979) to indicatea late Lochkovian age. The conodont sample andage assessment were updated by Pickett (1993b) tothe slightly younger early Pragian sulcatusConodont Zone. The field relationship of thisconodont locality is now suspect, being interpretedas a possible faulted block of older sediments(?Garra Formation) within the rhyolite, whilst a

New South Wales placodenn biostratigraphy

provisional SHRIMP isotopic age determination of376 ± 4 Myr from the middle flow-banded rhyoliteunit of the Dulladerry Volcanics in the northwestcorner of the Bathurst sheet (0. Raymond pers.comm. 1994) is entirely consistent with the fish andplant evidence (Young 1993, 1995). The consistentevidence from two localities 120 km apart indicatesthat, at least in some localities, volcanism in centralNSW. continued until the late Middle Devonian(Figure 2B), as is the case with the Middle DevonianComerong Volcanics and Boyd Volcanic Complexin the eastern part of the Lachlan Fold Belt (Eden­Comerong-Yalwal volcanic zone; Fergusson et al.1979).

Age of the Canowindra fish faunaThe Canowindra fish site occurs about 8 km along

strike to the south of a locality from whichEtheridge (1901) described the inarticula tebrachiopod Lingula gregaria, recorded as "NyrangCreek, about five miles west of Canowindra" (L,Figure 1). R.K. Jones (1982, and pers. comm.) foundthe same lingulids at another locality in the vicinity(the type locality is not precisely known), a creekbed from which the author has also collected fishremains (Bothriolepis sp., antiarch indet. (coarseornament); ?phyllolepid fragments, probableosteichthyan cycloid scales). No section has beenmeasured here, but it lies to the west (i.e. higher inthe sequence) than the main strike ridge of thewesterly-dipping Mandagery Sandstone, ininterbedded fine-grained sandstones and redsiltstones probably representing the transitional bedsto the overlying Pipe Formation. The Canowindrafish site also lies on the western slope of theMandagery strike ridge, but occurs in sandstones,and thus presumably represents the upper beds ofthe Mandagery Sandstone (Figure 3B).

Young (1993: 216) argued that in the HerveyGroup of central NSW. "there is no confirmedoccurrence of fish remains older than the marinedeposits forming the basal part of the sequence".This now seems incorrect if the lingulid horizon atNyrang Creek is assumed to indicate brackish waterconditions approximating stratigraphically themore extensive marine horizon farther east, whichhas been equated with the late Frasnian sea levelhigh documented by conodonts etc. (Pickett 1972,1993a). Jones, Turner and Fordham (in press)consider an alternative younger age (?Famennian)for a conodont assemblage from Gap Creek nearOrange, but this seems unlikely for the MandagerySandstone, which is near the base of the UpperDevonian sequence in the Canowindra area. Otherlocalities within the Pipe Formation support anolder (?late Frasnian) age. Conolly (1965: 56) statedthat lepidodendroid plant remains were the onlyfossils to be found in the Pipe Formation, butWilliams (1975) reported abundant broken fish

145

plates, and the new field work has confirmedabundant fish material in the valley south of theMandagery railway section, in horizons in the lowerpart of the fine-grained recessive unit (PipeFormation) overlying the type MandagerySandstone. The fauna includes the following:Bothriolepis sp., Remigolepis sp., Groenlandaspis sp.,phyllolepid indet., osteolepid scales, andholoptychiid scales. The Groenlandaspis represents adifferent species from the one in the Grenfellassemblage at Redcliff Mountain, which ischaracterized by unornamented dermal bones(Ritchie et al. 1992). Detailed taxonomic work hasnot been undertaken, but it is predicted that theseassemblages may include species in common withthe Canowindra fish fauna, on the assumption thatthese faunas occupy a similar stratigraphic interval,representing the transitional beds between theupper Mandagery Sandstone and the lower PipeFormation (Figure 3B). The change to fine-graineddeposition may have been the lithological result ofthe marine flooding event, which caused brackishwater conditions and the lingulid assemblage at theNyrang Creek locality. The fine-grainedGooloogong Beds on the Grenfell sheet also containlingulid brachiopods and placoderm plates (Colwell1974; M. Jones 1984), suggesting correlation eitherwith the Pipe Formation, or with the KadinaFormation in the Hervey Syncline, from whichSherwin (1996) has also recorded crinoid ossicleswhich indicate marine conditions.

The above discussion implies that theCanowindra fish fauna is likely to be late Frasnianin age, rather than Famennian as previouslysuggested (e.g. Young 1993, 1995). At several of thenew fossil localities fish remains are also associatedwith lingulid brachiopods, indicating brackishwater conditions. Their complete absence fromother horizons with similar lithologies suggests thatfish occurrences may have been related to marineinfluence in an otherwise predominantly fluvialsystem, and more than one episode of marineinfluence seems likely through the sequence.

Correlations across the Lachlan River and age ofthe Grenfellaspis fish fauna

Problems in correlation of the MandagerySandstone raised by Ritchie et al.'s (1992) ageassessment of the Grenfell fish fauna werediscussed by Pickett (1993a), who accepted thelithological correlations of Conolly (1965) toconclude that the Mandagery was of Famennianage.

Conolly (1965: 61) identified the MandagerySandstone in the Gooloogong-Grenfell area as the"first thick white sandstone formation overlying thered beds of the Beargamil Sub-Group", which "canbe traced along the strike along the eastern side ofthe Manildra region across the Lachlan River down

146 G.c. Young

Figure 4 New reconstruction of Grenjellaspis branagani Ritchie et al., 1992, based on the new articulated specimen fromthe Eugowra Syncline. Tail restoration based on study of specimens of the sinolepid Lilljiangolepis describedby Wang (1987).

the eastern side of the Grenfell-Gooloogong region".The northernmost outcrop south of the LachlanRiver in this area is the Merriganowry sectionmentioned above (6, Figure 18), where the sequencewas given by Conolly (1965, figure 11) as: PeaksSandstone, Hunter Siltstone, Mandagery Sandstone,and Bumberry Formation, overlain byundifferentiated Cookamidgera Sub-Group (Figure3A). About 9 km to the north, across the river flatsof the Lachlan and Belubula Rivers, is the mostsoutherly outcrop of Hervey Group north of theLachlan River. This sequence, which includes theCanowindra fish locality, is referred to here as theBelubula River section (Figure 3B). Comparing thetwo sections, a problem with previous stratigraphicnomenclature is evident, because the lowermost

prominent sandstone ridge north of the LachlanRiver was referred to the Mandagery Sandstone,and that south of the river was named the PeaksSandstone. However, the type sections for the PeaksSandstone (Peaks Creek, Redcliff Syncline) andoverlying Hunter Siltstone (Hunter Gully Creek,Gooloogong Anticline) are some 20 km away, in thewestern part of the outcrop, and are not continuousalong strike with the Merriganowry section. Likethe Pipe Formation to the north, the HunterSiltstone was identified by Conolly (1965: 61) as redsiltstones and shales with some fine-grained redsandstones which "forms prominent valleysbetween the strike ridges of the Mandagery andPeaks Sandstones".

The Grenfellaspis fish fauna occurs just beneath the

.11 i ~'!"'Hw ...... :t"f' - - - - - - --

HERVEY GROUP

B z~

~Vlo~

::r:2~~

'"'"0t;',.,oQ.~

30'"o·'"~[

ciQ"...~

'"0::r

'<

en III .c:: en III . .~"a~~"a&§:Q.Q)ltlQ.IIl(J)(J)

oS! "0 :;: oS! .!!: .!!! IIIOl::J)C::O-O.."­-"- ltl c:: ~ III 0..>.EQ)"-C::I1:I~.c::Q)Clen~"OOQ. Cl: ~ <:J c:: ";::_ .!!:.c:

c::­Q) 00'0...

<:J

•~ I<.i • ?! CD - - - a.....;..-. ....

c:: ~

Q) E(iiLL..~ :::::l 00"'"

~ ~ II?~7i i~(/) I... c::Q) .-.0"0E"O:::l Q)a:I~

KEY PLACODERMTAXA

Pipe Bothrlo/Remlgo faun

i Canowindrafauna

16 Bumberry large1 antiarch fauna

1-------

z :5 IGrenfell fauna I«zl--------ffi I Bumberry sinolepidl r-. -. -. -. -.~

~ :4 phyllolepid fauna ~~~~~;~~~~;I~

r------1 Hunter phyllolepid13 fauna ~I;:~;:~;:~;:~;~~.~

..... .....1 t·...····)·..·..·) ---l1 1--- ......--1 1-1 1-1I

~12zl~I

fE~ ~~;~Ig,:~~:ry J r'MerriganOWry~edS- - - -------

",",

",",

",

A

- I- 15. Grenfell fauna

Z Z« ----------« z Ul- "0Z Z 14. Worange Pt. fauna 'a.0

w Q)

:2 ---------- 0> « cw u.

13. Jemalong/Canowindra UlC fauna- W ...- ....----------I- Z 12. Twofold Bay fauna« ~ ---------- Ul ........J Z 11. Nettletons Creek fauna 'g. UlCl) ----------« 1O. Pambula River fauna

_"0

a: o·a.- ---------- .0) Q)u. 9.BindJFreestone Ck fauna Eoz ---------- CD=

8.Taggerty/Mt. Howitt fauna cx:>-- « ---------- s::.- 7.Tatong/S.Blue Range fauna Co- Z z ---------- I- 0 « 6. Aztec fauna !> i=

- dW W - -~ -_ c ?C C) - - --_ b •:E - ---~ •

360

370

380

Age[Ma]TOURN.

354

Figure 5 A. Alignments for Middle-Late Devonian macrovertebrate zones, modified from Young (1995), to incorporate new results from Hervey Group maping. B. Preliminaryvertebrate faunal succession for the Hervey Group, with distribution of key placoderm taxa shown on the right. (L, assumed Nyrang Creek lingulid horizon; seeFigures 1B, 3B).

.......'"-l

148

base of sandstone ridges mapped by Conolly asMandagery Sandstone, and alternative correlations(arrows in Figure 3) proposed by Ritchie et al. (1992)placed it at a higher level in the sequence than theCanowindra fish fauna. This interpretation is nowsupported by a new Grenfellaspis locality discoveredin 1996, which is the first on the northern side of theLachlan River, about 25 km downstream(northwest) of the Merriganowry section (4, Figure1B). The locality is a gravel quarry on the westernlimb of the southern part of the Eugowra Syncline,in red mudstone forming an upper recessive unitnear the top of the Bumberry Formation. This hasyielded the only known articulated specimen ofGrenfellaspis branagani (Figure 4). The BumberryFormation is overlain by the recessive EurowFormation, which forms the extensive alluvialvalley in the core of this syncline, extending fromNanami in the south to Mount Taylor in the north(see Conolly 1965, figure 8). Whilst lithologicalboundaries over this distance are probablydiachronous, general continuity of outcrop alongstrike leaves little doubt that this new fish horizonapproximates to the upper Bumberry-lower Eurowinterval in the type section in the Parkes Syncline.

MACROVERTEBRATE ZONATION

A generalized stratigraphic frameworkincorporating the issues just discussed is shown inFigure 58. Black shales of the Merriganowry Bedsform the base of the sequence overlying thevolcanics. Above this, the first of alternating coarseand fine-grained units that characterize the HerveyGroup is represented by the Mandagery Sandstoneand overlying fine-grained Pipe Formation. Anotional subdivision of five fining-upward cycles isshown for the upper part of the sequence. TheBumberry Formation in its type section (thickness707 m), comprises four cycles, each with pebblywhite sandstones at the base, interbedded redsiltstones and white sandstones in the middle part,and fine red siltstones and shales at the top (Conolly1965, figure 10). The overlying Eurow Formationwas interpreted by Conolly as a fifth cycle in whichfine-grained sediments predominated. This outcroppattern can be traced south from the type section inthe Parkes Syncline, and is also seen in the EugowraSyncline, which includes the new Grenfellaspislocality. North of the Lachlan River the upper partof the sequence is predominantly fine-grained(Eurow Formation), but south of the Lachlan theupper part is predominantly coarse-grained(Weddin Sandstone), the latter formation beingprobably equivalent to one or more of the coarse­grained units within the Bumberry Formation to thenorth.

Distribution of key placoderm taxa (right side ofFigure 5B), derived from a preliminary analysis of

G.c. Young

the new fossil localities, forms the basis for aprovisional succession of new fossil faunas (leftside, Figure 5B), with alignments to an updatedmacrovertebrate zonation shown in Figure SA (seeYoung and Turner in press).

The oldest of six assemblages is the phyllolepidoccurrence of the Merriganowry Beds. Fish remainsin the upper Mandagery - lower Pipe intervalprobably include species in common with theCanowindra fish fauna, as discussed above. Severalnew localities in the middle part of the HunterSiltstone northeast of Grenfell have yielded aphyllolepid assemblage (including Bothriolepis),which is distinguished from the Grenfell fauna(with Grenfellaspis) at the top of the Hunter Siltstone,which lacks phyllolepids. Detailed taxonomiccomparisons are required to show whether thisassemblage can be differentiated from theunderlying Bothriolepis/Remigolepis assemblage. Anoteworthy new faunal association is the possiblestratigraphic overlap of phyllolepids and sinolepidantiarchs, based on a new assemblage from theBumberry Syncline on the Parkes sheet. However,the putative sinolepid remains are fragmentary, sothis association needs confirmation with bettermaterial. The Grenfell fauna, in its type localityoccurring at the top of the Hunter Siltstone, isplaced in the upper part of Bumberry Formationequivalents, as demonstrated by the new discoverynorth of the Lachlan River reported above.

The youngest assemblage is recorded from anumber of localities in the Conimbla Syncline andadjacent areas of the eastern and northern parts ofthe Grenfell sheet. The original locality (7, Figure1B), discovered in the early 1980's, was mentionedby Pickett (1993a: 282) as a "locality high in theBumberry Fmn". No taxa have yet been described,but the assemblage is characterized by a largeantiarch (new taxon) with an armour 1-2 m long,associated with Bothriolepis and Groenlandaspisremains, the latter distinguished from the smoothspecies in the Grenfellaspis fauna by their dermalornament. Sinolepids are apparently absent, whilstthe large antiarch may be compared with materialof similar size from South China (Zhang Guoruipers. comm. 1989), perhaps additional evidencesupporting a latest Devonian Asian connection withEast Gondwana (Young 1990; Rich and Young 1996;Young and Janvier in press). At this stage, withdetailed mapping of this area still incomplete,structural complexity permits only the generalconclusion that the localities producing thisuppermost fish assemblage occupy a high levelwithin Bumberry/Eurow Formation equivalents.

ACKNOWLEDGEMENTS

Or Doone Wyborn (A.N.V.) is thanked forfacilitating my involvement in the A.G.5.0. Lachlan

New South Wales placodenn biostratigraphy

field mapping project. For collaboration andassistance in the field I am grateful to Dave Wallaceand Ollie Raymond of A.G.s.O., and John Pickett,Lawrence Sherwin and other members of themapping team of the N.s.W. Geological Survey.R.W. Brown (A.G.S.O.) provided invaluablepalaeontological technical support both in the fieldand the laboratory, and H.M. Doyle (A.G.s.O.)worked for many months on preliminarypreparation and curation of large collections. TonyWatson and Christian Thun (A.G.S.O.) alsoprovided technical support, Carole Burrow(University of Queensland) assisted on a field trip,and advice, discussion, and unpublishedinformation was provided by various colleaguesincluding Lance Black, Clinton Foster, ZerinaJohanson, Bob Jones, Phil O'Brien, John Pickett,Alex Ritchie and Lawrence Sherwin. Bill Platts(Geopeko), Noreen Morris (University ofNewcastle), and Mrs Pam O'Sullivan are thankedfor making the Tomingley fish plate available forstudy, and providing locality details. Alex Ritchie,Zerina Johanson, Monica Yeung and Bruce Loomeshave provided invaluable assistance and support atthe Canowindra and Merriganowry sites. JohnCasey and Neville Exon assisted with geologicalinterpretations in the Merriganowry quarry.Professor Richard Arkulus is thanked for provisionof facilities at the Geology Department, A.N.U.

REFERENCES

Ahlberg, P.E. and Johanson, Z. (1997). Secondtristichopterid (Osteolepiformes: Sarcopterygii) fromthe Upper Devonian of Canowindra, New SouthWales, Australia, and phylogeny of theTristichopteridae. Journal of Vertebrate Faleontology 17:653-673.

Campbell, KSW. and Bell, M.W. (1977). A primitiveamphibian from the Late Devonian of New SouthWales. Alcheringa 1: 369-38l.

Campbell, KSW. and Bell, M.W. (1982). Soederberghia(Dipnoi) from the Late Devonian of New SouthWales. Alcheringa 6: 143-149.

Colwell, ].W. (1974). The geology of the Wammba Range area,New South Wales. Unpublished BSc. (Honours) thesis,Australian National University, Canberra, ACT.

Conolly, J. (1965). The stratigraphy of the Hervey Groupin central New South Wales. Journal and Proceedings,Royal Society of New South Wales 98: 37-83.

Conolly, L Hall, L.R. and Rose, G. (1969). UpperDevonian series. In Packham, G.H. (ed.), The Geologyof New South Wales. Journal (1' the Society ofAustralia 16: 150-178.

Cribb, ]. (1996). A prize catch. Australian Geographic 43:99-115.

Etheridge, R. (1901). Lingula associated with Lepidodendron.Records (1' the Australian Museum 4: 119-120.

Fergusson, CL., Cas, R.A.F., Collins, W.J., Craig, G.Y.,Crook, K.A.W., Powell, CMcA., Scott, P.A. and

149

Young, G.C (1979). The Upper Devonian BoydVolcanic Complex, Eden, New South Wales. Journal ofthe Geological Society of Australia 26: 87-105.

Fletcher, H.O. (1956). Ancient armoured fishesdiscovered at Canowindra, NSW. Australian MuseumMagazine 12: 37-40.

Hills, ES (1929). The geology and palaeontography ofthe Cathedral Range and Blue Hills in north westernGippsland. Proceedings of the Royal Society of Victona41: 176-20l.

Hills, ES. (1932). Upper Devonian fishes from NewSouth Wales. Quarterly Journal of the Geological Societyof London 88: 850-858.

Hills, ES (1936). Records and descriptions of someAustralian Devonian fishes. Proceedings of the RoyalSociety of Victoria 48: 161-17l.

]ohanson, Z. (1995). New informations on jaw elementsof Remigolepis (Placodermi; Antiarchi) fromCanowindra, New South Wales, Australia (UpperDevonian). Geobios Memoire Special 19: 103-107.

]ohanson, Z. (1997a). New Remigolepis (Placodermi;Antiarchi) from Canowindra, New South Wales,Australia. Geological Magazine 134: 813-846.

]ohanson, Z. (1997b). New antiarchs (Placodermi) fromthe Hunter Siltstone (Famennian) near Grenfell, NewSouth Wales. Alcheringa 21: 191-217.

]ohanson, Z. and Ahlberg, P.E. (1997). Newtristichopterid (Osteolepiformes: Sarcopterygii) fromthe Mandagery Sandstone (Famennian) nearCanowindra, NSW, Australia. Transactions of the RoyalSociety of Edinburgh: Earth Sciences 88: 39-fJ8.

]ohanson, Z. and Ahlberg, r.E. (1998). A completeprimitive rhizodont from Australia. Nature 394: 569­573.

]ones, M. (1984). Aspects of the geology of the MountYambira area. Unpublished BSc. (Honours) thesis,Australian National University, Canberra, AC.T.

]ones, RK. (1982). The Late Devonian succession and fallllasof central and south-eastern New South Wales.Unpublished MSc. thesis, University of Sydney,Sydney, NSW.

]ones, RK., Turner, S. and Fordham, B.G. (in press). LateDevonian fauna from the Columbine Sandstone(Coffee Hill Member), Gap Creek, near Orange,central New South Wales. Courier ForschungsinstitutSenckenberg.

Long, ].A (1984). New phyllolepids from Victoria andthe relationships of the group. Proceedings of theLinnean Society of New South Wales 107: 263-308.

Long, ].A (1989). Phyllolepid placoderm fishes from theLate Devonian of eastern Victoria. In Le Maitre, RW.(ed.), Pathways in Geology. Essays in honour of EdwinSherbon HIlls: 52-58, B1ackwell Scientific Publications,Australia.

McLoughlin, S. and Long, ].A. (1994). New records ofDevonian plants from southern Victoria Land,Antarctica. Geological Magazine 131: 81-90.

O'Sullivan, R (1992). The geology and genesis of a Devonianepithermal system in the Tomingley area. UnpublishedBSc. (Honours) thesis, University of Newcastle,Callaghan, NSW.

150

Packham, G.H., ed. (1969). The geology of New SouthWales. Journal of the Geological Society of Australia 16:1-654.

Pickett, J.W. (1972). Late Devonian (Frasnian) conodontsfrom Ettrema, New South Wales. Journal andProceedings of the Royal Society of New South Wales 105:31-37.

Pickett, J.W. (1979). Marine fossils and the age of theDulladerry Rhyolite. Unpublished PalaeontologicalReport 79/23. Geological Survey of New South WalesReport GS 1979/326.

Pickett, J.W. (1993a). A Late Devonian xiphosuran fromnear Parkes, New South Wales. Association ofAustralasian Palaeontologists, Memoir 15: 279-287.

Pickett, J.W. (1993b). Plant fossils from the DulladerryRhyolite. Unpublished Palaeontological Report 93/04.Geological Survey ofNew South Wales Report GS 1993/125.

Powel!, CMcA., Fergusson, CL. and Williams, AJ.(1980). Structural relationships across the Lambianunconformity in the Hervey Range-Parkes area,NSW. Proceedings of the Linnean Society of New SouthWales 104: 195-210.

Rich, T.H. and Young, G.C (1996). Vertebratebiogeographic evidence for connections of the east andsoutheast Asian blocks with Gondwana. AustralianJournal of Earth Sciences 43: 625-634.

Ritchie, A (1975). Groenlandaspis in Antarctica, Australiaand Europe. Nature 254: 569-573.

Ritchie, A (1984). A new placoderm, Placolepis gen. novo(Phyllolepidae) from the Late Devonian of New SouthWales, Australia. Proceedings of the Linnean Society ofNew South Wales 107: 321-353.

Ritchie, A, Wang, S.T., Young, G.C and Zhang, G.R.(1992). The Sinolepidae, a family of antiarchs(placoderm fishes) from the Devonian of South Chinaand eastern Australia. Records of the AustralianMuseum 44: 319-370.

Savage, N.M. (1968). The geology of the Manildradistrict, New South Wales. Journal and Proceedings ofthe Royal Society ofNew South Wales 101: 150-169.

Sherwin, L. (1973). Stratigraphy of the Forbes-Bogan Gatedistrict. Records of the Geological Survey of New SouthWales 15: 47-1Ol.

Sherwin, L. (1996). Narrornine 1:250000 geological map.Explanatory notes. Geological Survey of New SouthWales. i-viii, 1-104.

Thomson, K.S. (1973). Observations on a new rhipidistianfish from the Upper Devonian of Australia.Palaeontographica A143: 209-220.

Wang, S.T. (1987). A new antiarch from the EarlyDevonian of Guangxi. Vertebrata PalAsiatica 25: 81-90.(Chinese with English Summary)

G.c. Young

Webby, B.D. (1972). Devonian geological history of theLachlan Geosyncline. Journal of the Geological Society ofAustralia 19: 99-123.

Wil!iams, AJ. (1975). Geology of the Parkes-Bumberryregion, NSW. Unpublished BSc. (Honours) thesis,Australian National University, Canberra, A.CT.

Williams, AJ. (1977). InSight into lingulid evolution fromthe Late Devonian. Alcheringa 1: 401-406.

Young, G.C (1985). New discoveries of Devonianvertebrates from the Amadeus Basin, centralAustralia. Bureau of Mineral Resources Journal ofAustralian Geology and Geophysics 9: 239-254.

Young, G.C (1988a). Antiarchs (placoderm fishes) fromthe Devonian Aztec Siltstone, southern Victoria Land,Antarctica. Palaeontographica A202: 1-125.

Young, G.C (1988b). New occurrences of phyllolepidplacoderms from the Devonian of central Australia.Bureau of Mineral Resources Journal of AustralianGeology and Geophysics 10: 363-376.

Young, G.C (1989). The Aztec fish fauna of southernVictoria Land - evolutionary and bioge'lgraphicsignificance. In: Crame, J.A. (ed.), Origins andevolution of the Antarctic biota. Geological Society ofLondon Special Publication 47: 43-62.

Young, G.C (1990). Devonian vertebrate distributionpatterns, and cladistic analysis of palaeogeographichypotheses. In McKerrow, WS. and Scotese, CR.(eds), Palaeozoic biogeography and palaeogeography.Geological Society of London Memoir 12: 243-255.

Young, C.C (1993). Middle Palaeozoic macrovertebratebiostratigraphy of Eastern Gondwana: 208-251, InLong, J.A (ed.), Palaeozoic vertebrate biostratigraphy andbiogeography, Belhaven Press London, U.K.

Young, G.c. (1995). Timescales 4. Devonian.Biostratigraphic charts and explanatory notes. SecondSeries. AGSO Record 1995/33: 1-47.

Young, C.C and Laurie, J.R., eds (1996). An AustralianPhanerozoic timescale: 1-279, Oxford University Press,Melbourne, Vic.

Young, G.C and Janvier, P. (in press). Early-MiddlePalaeozoic vertebrate faunas in relation to Gondwanadispersion and Asian accretion. Final results volume,IGCP 321 'Gondwana dispersion and Asian accretion'.Balkema.

Young, G.C and Turner, S. (in press). Devonianrnicrovertebrates and marine-nonmarine correlationin East Gondwana: overview. Courier Forschungs­institut Senckenberg.

Manuscript received 3 March 1998; accepted 18 August 1998.