Magnetic polarity stratigraphy of a series of pliocene and pleistocene vertebrate fossil localities from southeastern Au...

MISSING IMAGE

Material Information

Title:
Magnetic polarity stratigraphy of a series of pliocene and pleistocene vertebrate fossil localities from southeastern Australia
Physical Description:
viii, 157 leaves : ill. ; 28 cm.
Language:
English
Creator:
Whitelaw, Michael J., 1960-
Publication Date:

Subjects

Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1990.
Bibliography:
Includes bibliographical references (leaves 146-154).
Statement of Responsibility:
by Michael J. Whitelaw.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001683423
notis - AHZ5394
oclc - 25034731
System ID:
AA00003763:00001

Full Text












MAGNETIC POLARITY STRATIGRAPHY OF A SERIES OF PLIOCENE AND
PLEISTOCENE VERTEBRATE FOSSIL LOCALITIES FROM SOUTHEASTERN
AUSTRALIA
















BY

MICHAEL J. WHITELAW


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1990













ACKNOWLEDGMENTS

I would like to thank my parents, Tom and Elaine

Whitelaw, for their support, understanding and tolerance

during all of my student days. I would never have come this

far without their help. I am sorry that my father did not

survive to see me finish the Ph.D. This dissertation is

dedicated to his memory.

I would also like to thank the rest of my family, Peter

and his wife Bronwyn, Tom Jr. and April, Andrew and Robbo,

and Chris. All, at some stage or other, were pressed into

service "in the interests of science" and gladly gave

logistical and physical help on many occasions (even if they

were not always entirely sure why they were doing it). I

would also like to thank my US family Coy, Carol and Dave

Laws for their help, support and friendship.

I would like to thank my supervisor, Dr. Bruce

MacFadden, for his support, academically, financially and as

a friend, throughout the project. He is responsible for

making it possible for me to come and study in the United

States. I trust that history will not judge him too

harshly.

I would like to thank the other members of my

committee, Dr. Neil D. Opdyke, Dr. David Hodell, Dr. Douglas

S. Jones, Dr. S. David Webb and Dr. Ronald G. Wolff, for

their help, support and encouragement throughout my stay at

ii









UF. Two other people who deserve special mention are my

paleo-parents back in Australia. I would like to thank Drs.

Pat and Tom Rich who started me off in palaeontology and

have maintained their support to this day. During my

several returns to Oz they provided monetary, logistical and

academic help as well as their friendship.

I would like to thank the other graduate students with

whom I worked, studied, and suffered. In particular, I

would like to mention Dan Bryant, Vic DiVenere, Teresa

Hawthorne, Ken Gilland, Greg Mead, George Houston, Richard

Hulbert, Dave "Mullet" Lambert and Matt Joeckel.

I would also like to thank a band of prominent

paleontologists from around the country who expressed

interest in this study and gave support in the form of

useful discussions and manuscript reviews. They include Dr.

Dick Tedford, Dr. Mike Woodburne, Dr. Bill Turnbull and Dr.

Ernie Lundelius and several anonymous reviewers.


iii










TABLE OF CONTENTS

1page

ACKNOWLEDGEMENTS................................... ii

ABSTRACT.... ...................... ................... vii

CHAPTERS

1 INTRODUCTION.................................... 1

Mammalian Biostratigraphy in Australia.......... 1
Problems in Australian Paleontology ............. 5
Australian Stages............................... 8
Purpose of this Study ........................... 12
Ancillary Studies............................. 13

2 MAGNETIC POLARITY STRATIGRAPHY OF THE HAMILTON
LOCAL FAUNA AND FORSYTH'S BANK FOSSIL
VERTEBRATE SECTIONS........................... 15

Introduction................................. ... 15
Geologic Setting................................ 17
Paleomagnetic Procedures and Results............ 20
Materials and Methods......................... 20
Hamilton Section Results....................... 21
Forsyth's Bank Section Results................ 23
Correlation to the Geomagnetic Timescale and
Conclusions................................... 24

3 MAGNETIC POLARITY STRATIGRAPHY OF THE PARWAN,
COIMADAI AND BULLENGAROOK FLOW SECTIONS....... 26

Introduction.......................... .......... 26
Previous Work and Local Fauna Description....... 28
Geologic Setting................................ 31
Boxlea............................................... 31
Parwan........... ................ .......... ...31
Coimadai........................... ........ ... 32
Bullengarook Flow.............................. 35
Paleomagnetic Procedures and Results............. 35
Field and Laboratory Procedures ............... 35
Results ................... .. ..... ............ 36
Parwan Section ................................ 36
Coimadai Section.............................. 39
Bullengarook Flow ............................. 40
Magnetic Polarity Stratigraphy and Correlation
to the Timescale.............................. 41
Conclusion...................................... 44









4 MAGNETIC POLARITY STRATIGRAPHY AND MAMMALIAN
FAUNA OF THE DOG ROCKS LOCAL FAUNA............. 47

Introduction............ ....................... 47
Previous Work................................... 49
Geologic Setting................................ 51
Micro-invertebrate Analysis................... 53
Paleomagnetic Procedures and Results............. 54
Magnetic Polarity Stratigraphy................. 57
Correlation to the Timescale.................. 58
Conclusion................................... 62

5 MAGNETIC POLARITY STRATIGRAPHY OF THE DUCK PONDS
AND LIMEBURNER'S POINT VERTEBRATE FOSSIL
FAUNAS......................................... 64

Introduction.................................... 64
Geologic Setting............................ .... 66
Paleomagnetic Procedures and Results............ 70
Correlation to the Timescale..................... 73
Conclusions..................... .............. 77

6 MAGNETIC POLARITY STRATIGRAPHY OF THE
FISHERMAN'S CLIFF AND BONE GULCH VERTEBRATE
FOSSIL FAUNAS....... ......................... 78

Introduction........... ....... ................. 78
Previous Work.................................. 79
Geographic and Geologic Setting............... 82
Paleomagnetic Procedures and Results............ 87
Magnetic Polarity Stratigraphy and Correlation
to the Timescale.............................. 91
Conclusions.................................... 94

7 SUMMARY AND CONCLUSIONS........................ 96

Introduction................................... 96
Description and Age Constraints of Fossil Local
Faunas.......................... ... .... ....... 99
Otway Basin
Hamilton Local Fauna......................... 99
Forsyth's Bank................................ 102
Nelson Bay Local Fauna......................... 103
Port Phillip Basin
Parwan Local Fauna............................ 104
Coimadai Local Fauna......................... 108
Boxlea Local Fauna............................ 110
Hines Quarry Local Fauna..................... 112
Dog Rocks Local Fauna........................ 114
Duck Ponds Local Fauna......................... 117
Limeburner's Point Local Fauna.................. 119










Murray Basin
Fisherman's Cliff Local Fauna................. 120
Bone Gulch Local Fauna......................... 124
Discussion...................................... 125

APPENDICES....................................... ..... 132

1 PROCESSED MAGNETIC DATA......................... 132

2 ISOTHERMAL REMANENCE MAGNETIZATION DATA......... 139

3 STEREO PLOT AND REVERSAL TEST OF CLASS I SITES.. 144

REFERENCES............................................ 146

BIOGRAPHICAL SKETCH................................... 155














Abstract of Dissertation Presented to the Graduate School of
the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

MAGNETIC POLARITY STRATIGRAPHY OF A SERIES OF PLIOCENE AND
PLEISTOCENE FOSSIL VERTEBRATE LOCALITIES FROM SOUTHEASTERN
AUSTRALIA

By

Michael J. Whitelaw

December 1990

Chairman: Dr. Bruce J. MacFadden
Major Department: Geology


A series of magnetostratigraphic studies is presented

from localities in Victoria and southwestern New South

Wales, Australia. These localities were examined in order

to improve the age resolution of known Pliocene and

Pleistocene fossil vertebrate sites in southeastern

Australia. Where possible, sections that contained the

fossil-bearing horizons were sampled directly. In cases in

which the original locality was unavailable, sections that

could be reliably correlated to the original were used as

substitutes. This study revealed that the following age

constraints can now be used for these southeastern

Australian local faunas: Forsyth's Bank: (Kalimnan Stage and

> 4.47 Ma); Hamilton: Gilbert Chron (4.46 + 0.01 4.47 Ma);

Boxlea: Gilbert Chron ? (> 4.10 Ma); Parwan: Gilbert Chron

(4.10 4.24 Ma); Coimadai: Late Gilbert Chron (3.64 + 0.01


vii









- 3.88 Ma); Fisherman's Cliff: Gauss Chron (2.92 2.47 Ma);

Bone Gulch: Early Matuyama Chron (2.47 1.88 Ma); Dog

Rocks: Early Matuyama Chron (2.47 2.03 + 0.03 Ma); Duck

Ponds: Late Matuyama Chron (< 1.66 Ma and probably > 0.98

Ma); Limeburner's Point: Late Matuyama or Brunhes Chron (<

0.98 Ma), Hines Quarry: Late Matuyama or Brunhes Chron (<

0.98 Ma); and Nelson Bay: Late Matuyama Chron (1.66 0.73

Ma).

Comparision of the above age ranges and the

distribution of taxa within the local faunas was made to

determine whether it was possible to construct a mammalian

biostratigraphy for the Australian Plio-Pleistocene. Whilst

many taxa have generic first appearance datum planes in this

suite of local faunas, the occurrence of these FADs are

largely restricted to only a few intensively studied

localities. This period of Australian mammal history is

also dominated by relative stasis in evolutionary rates and

fragmentation of the biota by paleoclimatic and geographic

endemism. The combination of these factors, together with

the general paucity of the vertebrate record, has hampered

the establishment of a reliable mammalian biostratigraphy

although the present studies represent the first attempt

toward that goal in Australia.


viii













CHAPTER 1
INTRODUCTION

Mammalian Biostratigraphy in Australia


The first serious mammalian biostratigraphic study in

Australia was carried out in 1831 in the Wellington Caves of

New South Wales. A local colonist, Mr. George Ranken, found

bones in one of the caves and guided an expedition led by

the famous explorer Major T. L. Mitchell to the area (Rich

et al., in press). On reaching the cave, Ranken lowered

himself down to a ledge and then, by fixing the rope to a

projecting portion of rock, proceeded to lower himself to

the next ledge. Australian biostratigraphy made an

inauspicious start when the "projection" gave way, and

Ranken discovered that he had tied the rope to a giant bird

femur (probably of the family Dromornithidae). Ranken

survived the fall and, together with Major Mitchell,

proceeded to collect the first significant fossil fauna from

the Australian continent (Rich et al., in press, after

Mitchell, 1838, p. 362).

Early vertebrate paleontology in Australia was

characterized by an inability or unwillingness of workers to

describe fossil material "in-house." Material was routinely

sent to Europe for description and publication by the

experts, notably French and English anatomists, in a manner

that smacks of intellectual subservience of the colonial










paleontologists to their European counterparts. This

situation led to a distinctly northern-hemisphere bias when

questions of Australian faunal evolution and biostratigraphy

were considered. The comments made by the renowned British

comparative anatomist, Sir Richard Owen, who dominated early

work, illustrate this point. He considered that "it was

necessary to search Britain's secondary [Mesozoic]

formations to find specimens analogous to Australia's recent

marsupial fossil forms" (Rich, 1982, p. 14). The

recognition of the unique character of mammalian evolution

on the island continent was not to be fully accepted until

the centre of studies was shifted to Australia.

The second phase in the development of Australian

vertebrate paleontology began in the 1850s and continued

into the next century. During this time the State Surveys,

Museums and Royal Societies all came into being. This stage

represented a period of transition away from dependence and

towards intellectual parity with European colleagues.

Notable workers during this period include J. T. Gregory,

Fredrick McCoy, Gerard Krefft, Robert Etheridge Jr., E. C.

Stirling, A. H. Zietz and Charles DeVis. The first

expeditions into the Lake Eyre region, notably to Lake

Callabonna, were made and led to the discovery of the first

complete skeletons of the enigmatic Diprotodon, and other

members of the Australian megafauna. Large numbers of

fossils were also recovered from the Darling Downs and

Chinchilla areas of Queensland. Many new fossil sites were









discovered through construction and mining operations

induced by expansion of the colonial population. Included

in this category are the Hamilton (1860s), Duck Ponds

(1875), Parwan (1892), Limeburner's Point (1895) and

Coimadai (1897) local faunas from Victoria, all of which are

discussed in this study (Fig. 1.1).

The first half of the twentieth century represents a

relative lull in Australian vertebrate paleontology. This

is demonstrated by the fact that none of the above-named

Victorian local faunas were seriously begun to be described

until the late 1950s. To underline the point, formal

descriptions of the Coimadai and Limeburner's Point local

faunas are only just going "in press" (Turnbull, pers comm.,

1990).

Australian vertebrate paleontology began a renaissance

in the late 1950s and 1960s that has continued to the

present. Joint expeditions between Australian and largely

American groups were led by people such as E. S. Hills, R.

A. Stirton, R. H. Tedford, D. Ride and E. D. Gill; these

resulted in the discovery of many new sites in the Lake Eyre

Basin and other areas. Other major contributors include W.

D. Turnbull, E. L. Lundelius and M. O. Woodburne whose

papers are commonly cited in this study. Many of the above-

named workers, their students, and a home-grown crop of

endemic (dinki-di) paleontologists have carried on the work.

The discovery or formal description of the Hamilton,

Forsyth's Bank, Fisherman's Cliff, Bone Gulch, Dog Rocks,























































Figure 1.1.


Location map of all mammalian local fauna
localities and geologic sections examined in
this study.





5


Nelson Bay and Hines Quarry Local Faunas (Fig. 1.1), which

are discussed in this study, have all occurred during this

period.

Marsupial biostratigraphy and evolution also advanced

during the renaissance of the 1960s. It began with the

seminal work of Ride (1964) who produced the first

phylogenetic synthesis of Australian fossil marsupials. It

was followed by the introduction of three important new

techniques. These are the biomolecular studies pioneered by

Kirsch (1968, 1977) and continued in a myriad of forms by

others today (see Archer, 1987 for a review); the 40K-40Ar

isotopic dating of basalts, particularly in southeastern

Australia by McDougall et al. (1966) and Aziz-ur-Rahman and

McDougall (1972); and phylogenetic systematics. These

studies have led to great advances in the understanding and

interpretation of marsupial evolution and biostratigraphy

(Archer, 1982, 1984, 1987; Flannery, 1990; Marshall, 1981;

P. V. Rich, 1982; T. H. Rich et al., in press; Woodburne et

al., 1985).


Problems in Australian Paleontology


In marked contrast to the abundant and well-studied

faunas from the northern hemisphere, there are several

problems that confront the mammalian paleontologist working

in Australia. The most important is the lack of material

available for study. This is largely because the dominant

geomorphic processes active on the continent were erosional









during the Mesozoic and Cenozoic. Terrestrial vertebrate

fossils have either not been preserved or have been

destroyed by the deep soil profiles that developed over much

of the country. This problem is compounded by the

remarkably small number of paleontologists working in

Australia, a function of the fiscal restraints imposed by a

limited tax-paying population. For example, today there are

only about 20 vertebrate paleontologists throughout

Australia actively engaged in research on fossil mammals of

that continent. The extent of the problem was highlighted

by the recent discovery of the beautifully opalized

Steropodon galmani Archer et al. (1985). This monotreme is

the first and only Australian mammal known from the Mesozoic

Era after 150 years of active paleontological research (it

was sold to the Australian Museum by an opal miner who had

no idea what he had but needed the cash!). The oldest

marsupial fauna is the geographically isolated Geilston Bay

Local Fauna (Tasmania) of late Oligocene age which remains

enigmatic because of poor preservation. Essentially, the

mammalian vertebrate record does not show any depth until

the middle Miocene, which is represented by localities from

the Lake Eyre Basin, Alcoota, Bullock Creek and Riversleigh.

Late Miocene, Pliocene, and Pleistocene localities are

better represented, but are still few in number and quality

relative to North American standards.

The other major problem confronting Australian

paleontologists is the lack of stratigraphic and chronologic









controls for many terrestrial vertebrate localities. A

general lack of tectonic activity and topographic relief

means that most fossil sites, when found, occur within

relatively thin stratigraphic sections. Correlation between

sites is difficult and often depends on the faunal stage-of-

evolution method as a means of relative age estimation.

Apart from terrestrial faunas that have marine tie-ins,

stratigraphic correlation to dated basalts has been one of

the few methods available that allow development of age

constraints independent of the faunas themselves. In

eastern and southeastern Australia the situation is

alleviated to a certain extent by the presence of a

widespread suite of basalt flows. These basalts originate

from a hot-spot source the continent has been passing over

as it has drifted north from Antarctica during the Cenozoic

(Wellman and McDougall, 1974; White et al., 1988). Many of

these basalts have been isotopically dated and thus provide

a stratigraphic basis for developing age constraints at some

fossil sites. However, of the isotopic dates currently

published, only one (at Hamilton) has been established for a

site where the basalt is in visible contact with the bone

bearing horizon.

Another method that previously has received little

attention in Australia is magnetic polarity stratigraphy.

The southeastern Australian basalt complexes were prominent

in the research that allowed development of the geomagnetic

timescale in the late 1950s and 1960s (Green and Irving,










1958; McDougall et al., 1966; Aziz-ur-Rahman, 1971).

However, apart from two studies done in the Murray Basin

(Bowler, 1980; An et al., 1986), the potential applications

of this method for dating fossil localities have been

ignored.



Australian Stages



Due to the poor record of Cenozoic terrestrial

vertebrates in Australia there are no equivalents to the

North American Land Mammal Ages (Woodburne, 1987). As a

result of this situation, paleontologists working in

Australia have made use of the invertebrate stage system as

an alternative in locations where mammalian fossils have

been found associated with marine units. This situation

occurs with the Forsyth's Bank, Hamilton, and Nelson Bay

local faunas of the Otway basin, and the Dog Rocks Local

Fauna of the Port Phillip Basin, all of which are discussed

in this study (Fig. 1.1).

Fortunately, these mammal localities are near the type

sections that define the widely used Neogene Australian

stages which occur within the Otway and Port Phillip Basins,

or the nearby Gippsland Basin. These local stages are used

in preference to European stages because of doubts and

controversy in correlations between the two (Abele et al.,

1988). The first invertebrate stage system was proposed by

Pritchard and Hall (1902), and later underwent major









additions and modifications after Singleton (1941), Crespin

(1943), Carter (1959), Wilkins (1963), and Ludbrook and

Lindsay (1969). The current status of these Neogene stages

is presented in Abele et al. (1988) (Fig. 1.2).

The Neogene stages mentioned in this study are, from

oldest to youngest, the Batesfordian, Balcombian,

Mitchellian, Cheltenhamian, Kalimnan and Werrikooian stages.

The first two and the last one are defined on the basis of

foraminiferal assemblages and the middle three on molluscan

assemblages. The Batesfordian through Mitchellian Stages

span a range of early through late Miocene (Abele et al.,

1988) (Fig. 1.2). These stages are not directly relevant to

this study, with the exception that they are represented in

units below the Dog Rocks fossil bearing horizon at

Batesford Quarry.

The Pliocene was originally represented by the Kalimnan

Stage of Singleton (1941). It was proposed for the

Gippsland Basin where it was underlain by the Mitchellian

Stage. The Cheltenhamian Stage was originally proposed for

a Pliocene/Miocene section located at Black Rock in the Port

Phillip Basin (Singleton, 1941). Wilkins (1963) later

redefined it as a stratigraphic interval that separated the

Mitchellian and Kalimnan Stages of the Gippsland Basin,

essentially replacing the lower half of the Kalimnan Stage as

defined by Singleton (1941). This redefinition has been

accepted by Abele et al. (1988) although the basis for the

subdivision is unclear. Currently, the Mitchellian is








TIME (Ma)
0.0


1.8


3.3-


5.0.-









10.5 -







15.0 -


EPOCH


PLEIST-
OCENE


LATE
PLIOCENE

EARLY
PLIOCENE


LATE
MIOCENE









MIDDLE
MIOCENE



EARLY
MIOCENE
i--^-


AUSTRALIAN STAGE


WERRIKOOIAN


KALIMNAN
CHELTENHAMIAN

MITCHELLIAN


BAIRNSDALIAN



BALCOMBIAN
BATESFORDIAN


Figure 1.2. Relationships between local Australian Stages
and the Geologic Timescale.









regarded as late Miocene, the Cheltenhamian as straddling

the Miocene-Pliocene boundary and the Kalimnan as Pliocene

(probably largely early Pliocene) (Abele et al., 1988) (Fig.

1.2). In this study Kalimnan Stage sediments are

encountered in the Forsyth's Bank, Hamilton and Dog Rocks

local fauna sections.

A hiatus occurs between the Pliocene Kalimnan Stage and

the very late Pliocene-early Pleistocene Werrikooian Stage

(Fig. 1.2). The Werrikooian Stage type section is in the

lower part of the Whalers Bluff Formation of the Otway Basin

(Abele et al., 1988) and has a temporal equivalent in the

Newer Volcanics unit which is encountered at the top of the

Dog Rocks section.

Australian invertebrate stages, as used in a mammalian

biochronological context, were most recently reviewed in the

land mark paper of Woodburne et al. (1985). This paper

presented a synopsis of current knowledge of the continental

fossil mammal record of Australia and New Guinea, describing

the record both in terms of localities and phyletic groups.

With the exception of the Limeburner's Point and Nelson Bay

local faunas, the Woodburne et al. (1985) paper discusses

all the local faunas examined in this study. The local

fauna ages they present are based on stratigraphic

relationships with inter-tonguing marine sequences; or on

previously established 40K/40Ar dates for basalts related to

fossil localities by superposition arguments, interpolations

or long distance correlations. The research presented in









this dissertation compliments the Woodburne et al. (1985)

paper by using a dating technique which permits direct

analysis of fossil localities, enhancing many of the

previously established local fauna ages and redefining

others, in the process.

The utilization of invertebrate stages to define mammal

assemblage ages is not new, but was traditional the case

elsewhere in the world. However, as better chronologies

have been developed on other continents modern mammalian

biostratigraphy has tended to replace older invertebrate

terminologies. Thus the early Miocene Burdigalian and late

Miocene to Pliocene Pontian have been replaced by,

respectively, the Orleanian and Turolian in Europe (Savage

and Russel, 1983). Although the Australian sequence is not

far enough along for a similar transition to occur, the

development of a terrestrial, land-mammal based

biochronology will evolve as independent chronologies, such

as this study presents, continue to be established.







Purpose of this Study


The purpose of this study is to constrain as tightly as

possible the ages of a series of fossil mammal faunas from

southeastern Australia. The primary method of achieving

this goal has been through magnetostratigraphic studies of









fossil bearing sections which are then calibrated to the

geomagnetic timescale by stratigraphic correlations to

established magnetostratigraphic sections, isotopically

dated basalts, or marine sections with good biochronologic

control. The local faunas were then examined within this

temporal framework to see if they might provide the basis

for a mammalian biostratigraphy for the Australian Pliocene

and Pleistocene.


Ancilliary Studies



Whilst collecting magnetic samples at the fossil

localities discussed in this report, an interest was also

maintained in collecting and describing two of the local

faunas from Victoria. The collection of the Dog Rocks Local

Fauna was largely conducted prior to my arrival at Florida.

The work consisted of hand sieving 80 metric tonnes of

material for a total yield of 3.5 kg of bone fragments and

approximately 250 teeth. Work on describing the fauna is

currently in progress and a total of some 22 taxa are now

recognized (Whitelaw, 1989). Over the last three years I

have continued to add to the Nelson Bay Local Fauna. Work

involved prospecting along the cliff edge for macrofaunal

elements and sieving of matrix in the Southern Ocean surf in

order to recover the microfauna. The sieving has produced a

hitherto little known microfauna including rodents,

dasyurids and a new Ektopodontid (Whitelaw, 1990e). The









fauna is now being described and currently totals 20

mammalian taxa.

Consideration was also given to the possible use of the

87Sr/86Sr at localities that included marine units in order

to develop a strontium isotope stratigraphy for the southern

Australian coast. This was to be done with a marine core on

which chronologic controls had been developed using

magnetostratigraphy. To this end, a core taken from 15 km

off the coast of Portland (BMR 53) was obtained. Fossil

invertebrate material was collected from several sites in

Victoria, including the Dog Rocks, Nelson Bay, Forsyth's

Bank, Lake Tyers and Bunga Creek localities on the

assumption that the strontium method would be viable (Fig.

1.1). Seventy-six closely spaced samples were collected

from BMR 53 for magnetostratigraphic analysis. All samples

were of normal polarity indicating that the base of the core

was no older than the Brunhes Chron. Unfortunately, this

part of the strontium curve is essentially flat and

therefore, unsuitable for resolution by strontium dating.

Furthermore, analysis of shell material from most localities

showed that carbonate recrystallization and contamination

was a major problem. Therefore the strontium study was

discontinued and magnetostratigraphy became the main method

employed in deriving chronologic constraints for the local

faunas discussed in this report.













CHAPTER 2
MAGNETIC POLARITY STRATIGRAPHY OF THE HAMILTON LOCAL
FAUNA AND FORSYTH'S BANK FOSSIL VERTEBRATE SECTIONS



Introduction



The Hamilton Local Fauna is the most productive

Pliocene vertebrate fauna in southern Australia. It is

located approximately 7 km west of Hamilton, on the south

bank of the Grange Burn, about 100 m downstream from a

small waterfall (370 43' S, 1410 57.3' E) (Fig. 2.1).

Vertebrate fossils were first recovered from the area in

the 1860's (see Gill [1955] for a good reference list of

early discoveries). The majority of the fauna was

recovered by Turnbull and Lundelius (1970) and by Rich

during field seasons in 1978-80. This material is

currently being described and re-evaluated (Flannery et

al., 1987 and in press; Turnbull, Rich and Lundelius,

1987a-c) with the faunal list including some 28 mammalian

fossil taxa (Rich et al., in press).

The fossil bearing strata are overlain by an

isotopically dated basalt (K-Ar age of 4.46 + 0.01 Ma;

Turnbull et al., 1965) and underlain by a Kalimnan Stage

macro-invertebrate fauna (Abele et al., 1988). These

stratigraphic relationships give the Hamilton Local Fauna































































Figure 2.1. Locality and Geology of the Hamilton Local
Fauna and the Forsyth's Bank Fauna Sections.









the most well constrained age of any vertebrate fauna in

Australia, and therefore allows it to be used as a

benchmark locality for the study of mammalian evolution.

The results of this study further constrain the age of

the locality by using magnetic polarity stratigraphy to

generate a lower age constraint for the vertebrate fauna.

The Forsyth's Bank fauna consists of a single

specimen, a ramus of Protemnodon sp. which was found in

1933 (Gill, 1953). The ramus was recovered from the bank

of the Grange Burn, approximately 8 km west of Hamilton

(370 43.7' S, 1410 56.7' E) (Fig. 2.1). It was found in

the Grange Burn Formation, a marine carbonate which

yields a Kalimnan Stage (ca. 5.0-3.5 Ma) molluscan fauna

(Ludbrook, 1973). This fossil is one of the oldest

Tertiary vertebrates known from Victoria. In order to

further constrain the age of this locality, paleomagnetic

samples were collected from a section at Forsyth's Bank

in an attempt to generate a magnetic polarity

stratigraphy for correlation to the timescale.


Geologic Setting


The Hamilton section is characterized by a 1.3 m

thick fossil soil that contains the Hamilton Local Fauna,

which is underlain by the marine Grange Burn Formation

and overlain by a basalt flow (Fig. 2.2). The basalt is

approximately 1.5 m thick and has been isotopically dated

by the 40K-40Ar method at 4.46 + 0.01 Ma (Rich et al., in








VGP
LATITUDE


HAMILTON
SECTION


HEIGHT
(m)
2.0



1.5



1.0-


SITE 0.
CLASSIFICATION
CLASS I
0 CLASS III
REJECTED


-90


1.


,vvvv vv
,BasaltN
"I vv Z I? I ,J
.JV'DateV
4.46 +V\
'0 v~a
O.10~vMa'i
rvvvvVvV
v V VV VV V
VVVVVVVVv
VVVYVVVVV\

Paleosol
Horizon
A


SITE


CORRELATION TO
GEOMAGNETIC
TIMESCALE


\106
-105
-104


103 HAMILTON
S LOCAL
102 FAUNA
101


-100-
Paleosol
5- Horizon
B


0 Creek
Grange Bed
Burn Fmj
FORSYTH'S q
BANK /
SECTION
5 I..E=E /


1.0



0.5



0


/I


f\
/


Age
(Ma)
3.97

4.10

4.24 W

4 4n (I-


I


Lvi


4.47 z
<
4.57 z
-J
2

4.77


ca.
5.00




5.35


5.53


Figure 2.2. Stratigraphy of the Hamilton and Forsyth's
Bank Sections and Correlation to the
Geomagnetic Polarity Timescale.


. r


I









press; corrected from an original date of 4.35 + 0.01 Ma

after Turnbull et al., 1965, using the revised 40K-40Ar

decay constants in Steiger and Jager, 1977). The same

basalt produced a normal polarity from a single sample

(GA 1141) that was collected for paleomagnetic analysis

(McDougall et al., 1965). This determination is useful

but a polarity produced from a single sample is not

considered sufficient to warrant the establishment of a

magnetic polarity stratigraphy (McElhinny, 1973).

The basalt is underlain by a duplex soil which

contains a grey to blue silty sand in the 'A' horizon and

a 'B' horizon characterized by the presence of abundant

carbonate nodules. Softwoods, in growth position, and

large numbers of fossil teeth and rare bone fragments

have been recovered from the 'A' horizon (Abele et al.,

1988).

The paleosol is underlain by the marine Grange Burn

Formation which contains a molluscan fauna indicative of

the Kalimnan Stage (Ludbrook, 1973). The Grange Burn

Formation was not sampled at this section but

approximately 1 km downstream, at the Forsyth's Bank

locality, where approximately four meters of the unit are

exposed (Fig. 2.2). At Forsyth's Bank the unit is

characterized by the occurrence of rich molluscan coquina

lenses within a flaggy, shelly marl.










Paleomagnetic Procedures and Results





Materials and Methods

As is standard paleomagnetic procedure, three

separately oriented hand samples (A, B and C) were

collected from each of seven sites at the Hamilton Local

Fauna section and each of four sites at the Forsyth's

Bank section (Fig. 2.2). Samples were cut into standard

2.5 cm cubes for analysis. All samples were analyzed in

the Paleomagnetics Laboratory at the University of

Florida, which contains a Superconducting Technology

cryogenic RF-driven SQUID magnetometer (Goree and Fuller,

1976) in a shielded room which attenuates the ambient

field to ca. 200 nT (see Scott and Frohlich, 1985, for

similar design details).

The A sample from each site was treated with a

stepwise thermal demagnetization regime of 0-6300 C (16

steps) in a Schonstedt thermal demagnetizer. The B

sample from each site was treated with a stepwise AF

demagnetization regime of 0-100 mT (11 steps) on a

Schonstedt AF demagnetizer. The AF demagnetization

regime failed to completely demagnetize some sediment

samples; therefore, the B samples were further treated

with a thermal demagnetization regime identical to that

applied to the A samples, as also were the C samples.









All basalt samples were subjected to a stepwise AF

demagnetization regime of 0-100 mT (11 steps).

Hamilton Section Results

The basalt samples collected from all three sites in

the Hamilton section exhibited identical demagnetization

characteristics. A stable characteristic component over

a demagnetization range of 40-90 mT was isolated (Fig.

2.3a). Two sites produced three samples, each with

concordant directions and R values >2.62 (after Fisher,

1953) and are categorized as Class I normal polarity

sites after Opdyke et al. (1977). The remaining basalt

site produced two samples with concordant directions and

is categorized as a Class III normal site. In the

Hamilton sediment samples, a stable characteristic

component was generally isolated over a demagnetization

range of 0-5500 C (Fig. 2.3b). Three sites produced

three samples each with concordant directions and R

values >2.62 and are categorized as Class I normal

polarity sites. One site produced three samples with

similar characteristic directions but an R value of

2.60. It is categorized as a Class III normal polarity

site.

Isothermal remanent magnetization (IRM) experiments

were performed on samples from two sites and both reached

a saturation plateau by 100 mT, but then continued to

increase slowly (Fig. 2.3d). This suggests the presence



















N (UP)


N (UP)


INTE
500-

400-

300

200

100

0


NSITY
Jr/Jo












0 1.0 2.0 3.0
TREATMENT (A/m)


Figure 2.3.


E
INTENSITY
40- Jr/Jo


30-


20-


10


0 1.0 2.0 3.0
TREATMENT (A/m)


200-

150-


Paleomagnetic data plots. (A) AF
demagnetization Zijderfeld for Hamilton Site
100.1; (B) AF/Thermal demagnetization
Zijderfeld for Forsyth's Bank Site 200.2 (C)
AF demagnetization Zijderfeld for Hamilton
Site 104.2; (D,E,F) Isothermal
remanence saturation plots for Hamilton
(Sites 102 and 106) and Forsyth's Bank (Site
201), respectively.









of at least two carriers of the natural remanent

magnetism (NRM), a low coercivity mineral which is

probably the major contributor to the NRM, and a high

coercivity mineral which adds a lesser component. The

IRM data, together with unblocking spectra obtained from

demagnetization studies suggest that magnetite is

probably the low coercivity mineral and that goethite may

be the high coercivity component. IRM experiments

carried out on a sample of basalt from this locality

produced greater than 80% saturation by 100 mT (Fig.

2.3e). This, along with AF demagnetization

characteristics, indicates that magnetite is probably the

main carrier of the NRM in the basalts from this section.

Forsyth's Bank Section Results

The interpretation of the Forsyth's Bank section

magnetostratigraphy is unclear. Two antiparallel

components appear to be preserved in site 203 (Fig.

2.3c). The first is a normal polarity overprint that was

successfully removed by 3500 C. The second, which

appears to be the stable characteristic component for all

three samples is of reversed polarity. These samples

give a Fisher R >2.62 and the site is characterized as a

Class I reversed polarity site. One sample from site 202

displays similar decay characteristics, but all other

samples produce demagnetization trends characterized by

straight decays to the origin and normal polarities.

These samples were very hard and extremely difficult to









prepare, due to apparent carbonate recrystallization, and

the normal polarity component isolated in these samples

may be an overprint that was introduced during

recrystallization. In many samples the depositional

remanent magnetism (DRM) has been lost or obscured. Site

203 may preserve the DRM but a single Class I site is a

tenuous basis on which to propose a magnetostratigraphy.

Therefore, the magnetic polarity stratigraphy of the

Forsyth's Bank section remains problematic.

IRM experiments carried out on samples taken from

Forsyth's Bank achieved greater than 80% saturation by

100 mT (Fig. 2.3f). This indicates the presence of a low

coercivity mineral, probably magnetite, as the carrier of

the NRM.


Correlation to the Geomagnetic Polarity Timescale and
Conclusions


The entire Hamilton section is characterized by a

single zone of normal polarity (Fig. 2.2). This

polarity, in addition to the revised 40K-40Ar age of 4.46

+ 0.01 Ma for the basalt which caps the section,

correlates well with the 4.40-4.47 Ma normal polarity

event of the Gilbert Chron. Tree stumps found in situ in

the A horizon of the fossil soil have been burnt and

indicate that the soil is probably contemporaneous with

the flow (Turnbull et al., 1965). Therefore, the

continued normal polarity below the basalt suggests that









the entire sequence was formed during the same normal

magnetic polarity event. The age of the Hamilton Local

Fauna is then constrained between 4.46 + 0.01 and 4.47

Ma.

The magnetic polarity stratigraphy of the Forsyth's

Bank locality was not resolved. The age of this locality

is currently constrained to fall within the Kalimnan

Stage (ca. 5-3.3 Ma), on the basis of its molluscan

fauna, and to be older than 4.46 + 0.01 Ma, the age of

the overlying basalt. If the suspected DRM of this

section is indeed of reversed polarity, the age of this

locality may be constrained to either the 4.47-4.57 Ma

reverse polarity event or to the overlap of the 4.77-5.35

Ma reversed polarity event and the base of the Kalimnan

Stage.













CHAPTER 3
MAGNETIC POLARITY STRATIGRAPHY OF THE PARWAN, COIMADAI
AND BULLENGAROOK FLOW SECTIONS


Introduction


Magnetic polarity stratigraphy has been successfully

used at several Plio-Pleistocene fossil localities in

southeastern Australia, including the Nelson Bay Local

Fauna (MacFadden et al., 1987), the Dog Rocks Local Fauna

(Whitelaw, 1989) and the Fisherman's Cliff and Bone Gulch

local faunas (Whitelaw, 1990a). Results from three

exposures, which include a section from Parwan, an

outcrop of the Coimadai Dolomite, and the Bullengarook

(basalt) flow are presented in this study. In addition,

the implications for the ages of the Boxlea, Parwan and

Coimadai local faunas are discussed. This suite of local

faunas occurs in close proximity to each other and

appears to form a superimposed temporal series of early

Pliocene age. Previous age determinations are based upon

the assumption that the local fauna sections, or their

stratigraphic correlatives, are overlain by the

Bullengarook flow which has produced 40K-40Ar ages of

3.31 + 0.01 Ma and 3.64 + 0.01 Ma (McKenzie et al., 1983)

(Fig. 3.1). Results presented here indicate that this

assumption is incorrect for the sections that contain the

Parwan and Boxlea local faunas.















* A T F. ALL

BJ ULLENGAROOI
SI SECTION


.UATERNARY

| QIUATERNARY (UNDIFF)

TERTIARY

31 o 40' TERTIARY (UNOIFF)
BACCHUS
A-RSH. NEWER VOLCANIC

4.tto.a1 1 e
IV INTRABASALT SEDIMENTS

~ WERRISEE FORMATION (UNOIPF.)

BOXLEA "C
LOCAL of If COIMADAI DOLOMITE
FAUNA

S MADOINGLEY COAL SEAM

PARWAN PALEOZOIC
SECTION
FAN A CAL PALEOZOIC (UNOIFF.)

F K-Ar DATE

ERUPTION CENTRE


0 1 1 2
A 37 km










Figure 3.1. Locality and Geologic Map of the Bacchus

Marsh Area.









Previous Work and Local Fauna Descriptions


The Boxlea Local Fauna was recovered from the

overburden of an open-cut brown coal mine located near

the mouth of the Parwan Creek, 1.5 km east of Bacchus

Marsh (370 41.5' S, 1440 27' E) (Fig. 3.1). The fauna

includes Propleopus, Vombatus, Trichosurus and small

macropodids (Woodburne et al., 1985).

The Parwan Local Fauna was recovered during the

excavation of a railway cutting 2.0 km southeast of

Bacchus Marsh and 1.5 km west of the Parwan Railway

Station in 1882 (370 41.6' S, 1440 27.3' E) (Fig. 3.1).

A note curated with this material describes the location

of the find as "240 feet east of the west end of the

railway cutting, west of Parwan Station; 14 inches above

the Older basalt flow." The fossils were recovered from

interbasalt sediments at this location, and include

Sarcophilus, a vombatid, a phalangerid and rodents

(Woodburne et al., 1985).

The Coimadai Local Fauna was discovered during

operations in Alkemade's Quarry in 1897 (Officer and

Hogg, 1897-8), while mining lacustrine dolomitic

limestones for mortar in the 1890s. The quarry is

situated approximately 8 km northwest of Bacchus Marsh

(370 37' S, 1440 29.5' E), but is now submerged under the

waters of the Merrimu Reservoir (Fig. 3.1). I visited

this site in 1985 when the reservoir was partly drained,









for repairs and enlargements, and Alkemade's Quarry and

its associated kiln complex was briefly above water

level. The site was prospected and sampled for

vertebrate and micro-vertebrate remains but no additional

material was recovered. C. W. De Vis in Appendix A of

Officer and Hogg (1897-8) described the recovery of 22

bones and molds from this site and identified Phascolomys

parvus, Halmaturus dryas, H. anak, H. cooperii and

Nototheridae in the fauna. The quality of preservation

of this material is poor and Woodburne et al. (1985)

question De Vis's identifications. This material has

been prepared and redescribed by Turnbull, Lundelius and

Tedford (in press) and the revised local fauna currently

includes Euowenia, Zyqomaturus, Vombatus (near V.

hirsutus, Vombatus ("Phascolomys parvus") and four

macropods which include Kurrabi, Protemnodon, Troposodon

and Macropus.

Currently, the age of all three local faunas is

based on their presumed stratigraphic relationships to

the Bullengarook basalt flow. 40K-40Ar dates of 3.31 +

0.01 Ma and 3.64 + 0.01 Ma have been produced for this

flow at a locality, marked by a spectacular waterfall, 15

km north of Bacchus Marsh (Roberts, 1984) (Fig. 3.1).

The Bullengarook flow can be continuously traced for 15

km south of the waterfall until it is interrupted by a

1.2 km wide valley that contains the Lerderderg and

Werribee Rivers and the Parwan Creek, just east of









Bacchus Marsh (Fig. 3.1). Woodburne et al. (1985)

describe the Boxlea Local Fauna as occurring in beds of

the Rowsley Formation which "lies beneath a weathered

basalt flow that, near Parwan, is separated by inter-

basalt sediments from the overlying Bullengarook flow."

They describe the Parwan Local Fauna as occurring in the

same interbasalt sediments within 1 km of Boxlea (Fig.

3.1). This description indicates that they correlate the

Bullengarook flow on the north side of the

Lerderderg/Werribee/Parwan Valley to the flow that lies

over the fossil localities on the south side. Therefore,

the age given for these two localities is older than 3.31

+ 0.1 or 3.64 + 0.1 Ma, the age of the presumed southern

extension of the Bullengarook flow.

The age of the Coimadai Local Fauna is also based on

its stratigraphic relationship to the Bullengarook flow.

At Alkemade's Quarry, the lacustrine dolomites which

contain the fauna are overlain by a layer of ash. This

may be correlated to similar sections to the west that

directly underlie undoubted Bullengarook flow basalts

(Coulson, 1924). Based on this correlation, Woodburne et

al. (1985) have suggested an age of greater than 3.64 Ma

for the Coimadai Local Fauna.










Geologic Setting

Boxlea

The exposure of the relevant stratigraphic section

at Boxlea is currently in a flooded part of the Boxlea

coal mine. Therefore, description of this site is

dependent on an original note stored with the material at

the Museum of Victoria and descriptions of the coal mine

by Thomas and Baragwanath (1949) and Abele et al. (1988).

The fossils were recovered from the beds of the Rowsley

Formation that conformably overlie the Werribee

Formation, which contains the Maddingley Seam brown coals

of early Miocene age. A note found with the Boxlea

material curated in the Museum of Victoria describes it's

location as "Sandy clay 10-14 feet above coal". In

describing the locality Woodburne et al. (1985)

summarized it as occurring in overburden sediments above

the coal seam and below a weathered basalt which, near

Parwan, is separated by interseam sediments from the

Bullengarook flow. The weathered basalts and part of the

overburden are still visible in the southern wall of the

cut but since no description of the fossil site (relative

to these strata) exists, the exact location of the bone

bearing horizon is indeterminate.

Parwan

The Parwan Local Fauna was recovered in 1882, during

the excavation of a northwest-southeast trending railway










cutting. The northwestern half of the cutting reveals

two basalt flows which are separated by sediments

consisting of a graded sequence of coarse quartz gravels

overlain by yellow-brown silty clays (Fig. 3.2). The

older flow outcrops on the northeast side of the track 80

m northwest of a bridge over the cutting. On the basis

of the site description included with the fossils, it is

probable that they were collected from the gravel unit on

the opposite side of the track to the older flow. Except

for 1.2 m of inter-basalt sediments located around the

bridge foundations, the younger flow occupies the entire

southeastern half of the cutting.

Coimadai

The geology of Alkemade's Quarry from which the

Coimadai Local Fauna was recovered has been well

documented (Officer and Hogg, 1897-8; Fenner, 1918;

Coulson, 1924; Keble, 1925; Gill, 1964; Roberts, 1984;

Woodburne et al., 1985; Abele et al., 1988). The fauna

came from the lacustrine Coimadai Dolomite, one of the

upper members of the Werribee Formation. In 1985 the

quarry was partially submerged, but an original section

illustrated by Officer and Hogg (1897-8, Plate IX, Fig.

I) was located. It consists of an exposure of the upper

part of the Coimadai Dolomite which is conformably

overlain by a 15 cm layer of white ash followed by a

series of sands and gravels. The dolomite exhibits a

variety of textures including both laminated units and
















BULLENGAROOK VGP
FLOW LATITUDE
SECTION
_JA -90 0
VV 7ISITE r----


PARWAN VGP
COMPOSITE LATITUDE
SECTION
0 +90
vvv v SITE
5.0 VVVWVV
SVVVVVV
VVVVVVV
VVVVVVV
VVVVVVV
BASALT 111 -
VVVVVVVV
VVVVVVV
vv-V,, 110
4.0-vvvvvvvv

/VVVVVVV 109V,
vvvvvvv 109
VVVVVVVV
VVVVVVVV

INTER- -108
BASALT
3.0- vEDIMENTS
107
ernf l


A
2.0 SECTION
B

PARWAN
LOCAL
FAUNA
1.0 ,,, ,
VVVVVVVV
VVVVVVvV
VVVVVVVVV

VVVVV'
BASALT
VVVVVVV
V VVVVVV
vVVVVVVV
0
HEIGHT
(m) C


-105--

-104

-103


-102
-101

-100


SITE
LASSIFICATION
CLASS I
0 CLASS III
REJECTED


Figure 3.2. Stratigraphic Column and Magnetic Polarity
Results for the (A) Bullengarook, (B)
Coimadai and (C) Parwan Sections.


i









nodule beds, and is interspersed with lenses of quartz

gravels and sands. The section is slumped to form a

small monocline, probably as a result of karstification

that led to localized collapse and sagging of the

overlying sediments (Coulson, 1924). Identical sequences

of horizontally lying strata were found in other parts of

the quarry. These have been correlated to small outcrops

of the Coimadai Dolomite to the south and southwest,

which are conformably overlain by a thin layer of ash and

the basalts of the Bullengarook flow (Coulson, 1924).

Field sampling for this study was conducted in 1989,

at which time the expansion work on the Merrimu Reservoir

was complete and the fossil site totally flooded.

Therefore a section of the Coimadai Dolomite with a

lithology identical to that observed at Alkemade's Quarry

was chosen as a representative magnetostratigraphic

substitute for the fossil site. Located above a clay

quarry on the east side of the Bacchus Marsh-Gisborne

Road and 4 km southwest of Alkemade's Quarry (Fig. 3.1),

this section is characterized by a basal unit of white

clays, sands and gravels which is overlain by 2 m of the

Coimadai Dolomite (Fig. 3.2). The dolomite has a

massive, silty texture at its base which grades upwards

into a thick nodule bed. It is conformably overlain by a

thin layer of white ash which appears to be identical to

the ash observed at Alkemade's Quarry, and is capped by









basalt of the Bullengarook flow that exceeds a thickness

of 3 m at this locality.

BullenQarook Flow

As the interpretation of the age of the three fossil

localities has been stratigraphically linked to the

Bullengarook flow, either in this or previous studies, it

was considered necessary to examine its magnetic

polarity. Consequently, a third section located 50 m

upstream from the waterfall where the 40K-40Ar dates of

3.31 + 0.01 and 3.64 + 0.01 Ma (McKenzie et al., 1983)

had been obtained, was selected for study (Fig. 3.2). At

this site the Bullengarook flow occupies the paleovalley

of the Bullengarook River (Keble, 1945) and is

undoubtedly a single flow. The basalt section exposed at

the falls is in excess of 150 m thick and is

characterized by a thick basal layer of columnar basalt

which changes upwards to massive and then blocky units.


Paleomagnetic Procedures and Results



Field and Laboratory Procedures

In the field, three separately oriented hand samples

(A, B and C) were collected from every site at the three

sections studied. These were prepared and analyzed in

the Paleomagnetics Laboratory at the University of

Florida. This facility contains a Superconducting

Technology cryogenic magnetometer (Goree and Fuller,









1976), a Schonstedt spinner magnetometer and Schonstedt

AF and thermal demagnetizers. The samples were analyzed

in the shielded room laboratory which attenuates the

ambient field to ca. 200 nT (Scott and Frohlich, 1985,

provide similar design details).

All basalt samples were subjected to stepwise AF

demagnetization over a range of 0-100 mT (at 10 mT

intervals) (Fig. 3.3a,c,d). For each site collected from

sedimentary rock units, one sample was subjected to

stepwise thermal demagnetization over a range of 0-6300 C

(13 steps), and one was subjected to stepwise AF

demagnetization over a range of 0-50 mT (at 5 mT

intervals) and 60-100 mT (at 10 mT intervals) (Fig.

3.3b,e). Some of the samples from both Parwan and

Coimadai were not completely demagnetized by the AF

treatment. These were further treated by a thermal

demagnetization regime which successfully demagnetized

the samples. Based on the demagnetization results on the

first two samples a thermal demagnetization regime was

used on the third sample.

Results

Parwan Section

Twelve sites were examined from the Parwan section,

with three sites from the lower basalt, six from the

interbasalt sediments and three from the upper basalt

(Fig. 3.2). For the basalt of the lower flow, a well




























D .-.---- E


C


C N(UP
20 In


1IOmT 9
W.

30 mT

S(OOWN)






SXD-3 A/m

b6ld'CE






DOWN)


INTENSITY
120- Jr/Jo

90-

go- f

30


0 1.0 2.0 3.0
TREATMENT A/m


INTENSITY
Jr/Jo
1200-





400



0 1.0 2.0 3.0
TREATMENT A/m


10


5


INTENSITY
Jr/

w00-

00-

00

0-
OF ________


0 1.0 2.0 3.0
TREATMENT A/m


Figure 3.3. Paleomagnetic data plots. (A) AF
demagnetization Zijderfeld plot of Parwan
basalt (Site 102); (B) Thermal
demagnetization Zijderfeld plot of Parwan
Site 106; (C) AF demagnetization Zijderfeld
plot of Bullengarook flow (?) basalt at
Coimadai Section (Site 205); (D) AF
demagnetization Zijderfeld plot of
Bullengarook flow basalt Site 300); (E)
Thermal demagnetization Zijderfeld plot of
Coimadai (Site 201); (F) Isothermal Remanence
Saturation plot of Parwan Site 105;
(G) Isothermal Remanence Saturation plot
of Coimadai Site 200; (H) Isothermal
Remanence Saturation plot of Bullengarook
flow Site 301.










defined, stable component was isolated over a range of

50-90 mT. Two of the three sites from the lower basalt

produced three samples with statistically significant

demagnetization characteristics (R >2.62, after Fisher,

1953). These are classified as Class I sites after

Opdyke et al. (1977) (Fig. 3.2). The third site produced

two samples with concordant directions and is categorized

as a Class III site.

The interbasalt sediments were not fully

demagnetized by AF demagnetization, suggesting that a

high coercivity component is present as part of the

natural remanent magnetism (NRM). However, thermal

demagnetization treatments were highly successful and

produced stable characteristic components over a range of

150-6600 C (Fig. 3.3b). Three sites produced three

samples each with statistically significant concordant

directions and are classified as Class I sites (Fig.

3.2). One site produced an R value >2.62, but only two

samples with concordant directions, and is therefore

classified as a Class III site. Two sites produced two

samples each with concordant directions but R values

<2.62 and were rejected.

A stable component was isolated over a range of 10-

80 mT for the upper basalt. Two sites produced three

samples each with statistically significant

demagnetization characteristics and are categorized as

Class I sites (Fig. 3.2). Statistically significant and









concordant directions could not be isolated from the

third site and it was rejected from this analysis.

One basalt and three sediment samples were chosen

for isothermal remanent magnetization (IRM) studies. The

basalt sample reached 80% saturation by 120 mT indicating

that a low coercivity mineral, probably magnetite, is the

carrier of the natural remanent magnetism (NRM). The

sediment samples failed to saturate by 350 mT or produced

erratic saturation curves (Fig. 3.3f). These results

suggest that the NRM is being carried by minerals of both

high and low coercivity in different parts of the

section. This conclusion is further supported by the

failure of the AF demagnetization technique to completely

demagnetize some sediment samples, presumably because

they contained a mineral with a high magnetic coercivity.

The spectrum of unblocking temperatures, AF

demagnetization behaviour, and IRM acquisition curves

suggest that hematite and magnetite are the high and low

coercivity minerals, respectively, that carry the NRM in

these sediment samples.

Coimadai Section

Seven sites were studied from the Coimadai section,

four from the Coimadai Dolomite and three from the

overlying Bullengarook flow (Fig. 3.2). The Coimadai

Dolomite was successfully treated with a thermal

demagnetization regime (Fig. 3.3e). One site produced

three samples each with statistically significant










directions and is categorized as a Class I site (Fig.

3.2). Two other sites produced two samples each with

concordant directions and are classified as Class III

sites. One site failed to produce a statistically

significant mean direction and was rejected from the

study.

A stable characteristic component was isolated from

the basalts over a range of 20-90 mT (Fig. 3.3c). One

site from the Bullengarook flow produced three samples

each with concordant directions and is classified as a

Class I site (Fig. 3.2). The other two sites produced

two samples each with concordant directions which were in

agreement with those of the Class I site. However, they

both produced R values <2.62 and have been rejected from

this study.

Sediment samples from two sites were selected for

IRM studies. Greater than 80% saturation was achieved by

80 mT in both samples indicating the presence of a low

coercivity carrier of the NRM (Fig. 3.3g). This,

together with the unblocking spectra of the AF and

thermal demagnetizations suggest that magnetite is

probably the dominant carrier of the NRM in samples from

this locality.



Bullengarook Flow

Four sites were collected from the Bullengarook

(waterfall) section. Samples from these sites were









successfully treated with an AF demagnetization regime

over a range of 10-80 mT (Fig. 3.3d). All four sites

produced three samples each with a statistically

significant mean directions and are categorized as Class

I sites (Fig. 3.2).

Samples from two sites were selected for IRM studies

and both reached greater than 80% saturation by 40 mT,

thereby indicating the presence of a low coercivity

mineral (Fig. 3.3h). This, together with AF

demagnetization results suggests that magnetite is the

dominant carrier of the NRM in the Bullengarook flow

basalts.

Magnetic Polarity Stratigraphy and Correlation to the
Timescale

The magnetic polarity stratigraphy of the entire

Parwan section is characterized by a single zone of

normal polarity (Fig. 3.4). As the contact between the

overlying flow and the interbasaltic sediments is

paraconformable, it is likely that both units were formed

during the same polarity interval. The lower basalt,

having the same polarity, was either formed during the

same normal polarity event or during an earlier one.

The Coimadai section is characterized by a single

zone of reversed magnetic polarity (Fig. 3.4). This

suggests that the Coimadai Dolomite and the overlying

Bullengarook flow were formed during the same interval of

reversed polarity. Therefore, the interpretation of a













COIMADAI
SECTION


HEIGHT
(m)


CORRELATION TO THE
MAGNETIC
)OK POLARITY
I TIMESCALE


TIME(Ma)

S3.08


PARWAN
SECTION
.64'
*VVV VVV
vVVVVVVVV
VVVVVVVVV
VVVVVVVV
vvvgvvvv
.88 vvvvvvv
v8 v9111 BT
vvv vvvv
vvvvvvvv
VVVVVVVV
,97 110- vvvvvvvv
K-Ar Date
.10 109- 4:1.fd.61
VV .vv*vv
vvvv


4.24 108-


4.40 107-
4.47
9 J 106-
105-

\
104-

103-


\
102-
101-

100


INTER-
BASALT
SEDIMENT!


PARWAN
LOCAL
FAUNA


5.0




1.0





3.0




2.0


VVVVVVVv
vvvyvvvv
=VVVVVVV
,VVVVVVVV
,VVVVVVVV
'gvvvvvvv

,VV VVVV

HEIGHT
(m)


Figure 3.4.


Correlation of magnetostratigraphy to the
Geomagnetic Polarity Timescale.


-q


i









reversed magnetic polarity for the Bullengarook flow, at

the Coimadai section, is in agreement with the reversed

polarity established for the basalt from the Bullengarook

waterfall section. With this constraint the date of 3.31

0.01 Ma (McKenzie et al., 1983) that was obtained for

the Bullengarook flow should be rejected as it would

indicate that the Bullengarook flow should carry a normal

polarity. The second McKenzie et al. (1983) 40K-40Ar

date of 3.64 + 0.01 Ma falls within a zone of reversed

polarity and is here accepted as approximating the age of

the Bullengarook flow. Using the magnetic polarity

timescale of Berggren et al. (1985) and an age of 3.64 +

0.01 Ma for the basalt, the Bullengarook flow must have

occurred within the 3.40-3.88 Ma reversed polarity zone

of the Gilbert Chron. As the Coimadai Dolomite occurs

within the same reversed polarity zone, and is overlain

by the 3.64 + 0.01 Ma old Bullengarook flow, a probable

age of between 3.64 + 0.01 and 3.88 Ma is indicated for

the Coimadai Dolomite and the Coimadai Local Fauna (Fig.

3.4).

The presence of a normal polarity for both of the

basalts at the Parwan section indicates that a

correlation of these flows across the

Lerderderg/Werribee/Parwan Valley to the reversed

polarity Bullengarook flow is untenable. However, the

presence of two eruption centers, one 5.1 km to the

southeast and one 6.6 km to the southwest, suggest










another possible source for the Parwan section basalts.

Basalts from these eruption centers appear to have flowed

north towards, and probably over, the Parwan locality as

part of a complex series of flows. A 40K-40Ar date of

4.12 + 0.04 Ma (corrected from 4.03 + 0.04 (Aziz-ur-

Rahman and McDougall, 1972) using the decay constants of

Steiger and Jager, 1977; Dalrymple, 1979) has been

produced for a basalt belonging to one of these flows, at

a locality 3.25 km northwest of the Parwan section (Fig.

3.1). The flow sampled at this locality was also found

to be of normal magnetic polarity (Aziz-ur-Rahman, 1971)

indicating that it probably falls within the 4.10-4.24

normal polarity Cochiti Subchron. This polarity matches

that found at both of the Parwan locality basalts.

Therefore, an age of 4.10-4.24 Ma is proposed for the

Parwan section and the Parwan Local Fauna (Fig. 3.4).

Conclusion


The magnetic polarity stratigraphies derived for the

Parwan, Coimadai and Bullengarook sections indicate that

the Bullengarook flow does indeed overlie the Coimadai

Dolomite but that there is no correlative of the same

flow at Parwan. The Coimadai Dolomite, being conformably

overlain and having the same reversed polarity as the

Bullengarook flow, is interpreted as having been formed

during the same magnetic polarity interval. On this

basis, the age of the Coimadai Dolomite and the Coimadai










Local Fauna is constrained to an age of 3.64-3.88 Ma

(Fig. 3.4).

The difference in magnetic polarities between the

Bullengarook flow and the Parwan section basalts suggests

that the Woodburne et al. (1985) date for the Parwan and

Boxlea local faunas should be re-evaluated. The

Bullengarook flow, and its associated 40K-40Ar dates,

must be rejected as a possible correlative for the Parwan

section basalts. Eruption centers to the south have

produced flows which are more likely candidates for

correlation to the Parwan section basalts. A 40K-40Ar

date of 4.12 + 0.04 Ma has been produced for a basalt

from one of these flows and it falls within a normal

polarity zone of 4.10-4.24 Ma. This date, and the

polarity it implies, would correlate well to the magnetic

polarity of either basalt at Parwan. Therefore, an age

of 4.10-4.24 Ma is suggested for the entire Parwan

section and the enclosed Parwan Local Fauna (Fig. 3.4).

Although it was not possible to sample a section at

the Boxlea Coal Mine, it is possible to make some new

inferences regarding the age of the Boxlea Local Fauna.

The Boxlea Local Fauna occurs in beds of the Rowsley

Formation that conformably overlie brown coals of the

Werribee Formation of early Miocene age and underlies a

basalt flow that is correlated to the lower flow at the

Parwan section which was assigned an age of older than

3.64 Ma (Woodburne et al., 1985). This age was based on






46



an incorrect correlation to the Bullengarook flow and is

therefore invalid. Based on the magnetic polarity age

suggested for the Parwan section, in this paper, the age

of the Boxlea Local Fauna must be revised downwards to an

age of older than 4.10 Ma.













CHAPTER 4
MAGNETIC POLARITY STRATIGRAPHY AND MAMMALIAN FAUNA OF THE
DOG ROCKS LOCAL FAUNA


Introduction



The lack of well dated faunal assemblages in

Australia has been a major obstacle to paleontological

attempts at interpreting the unique fossil record of this

continent. Along with the Nelson Bay Local Fauna (MacFadden

et al., 1987), the Dog Rocks Local Fauna is one of only a

few late Pliocene to early Pleistocene vertebrate

assemblages known in Australia with absolute

geochronological control currently available (Rich et al.,

1982; Woodburne et al., 1985). Located 7.0 km NW of

Geelong, Victoria at 380 6.5' S, 1440 17' E (Fig. 4.1), it

consists of a suite of bone fragments and isolated teeth

collected by field crews from the National Museum of

Victoria from 1975 to 1986. The assemblage is loosely

disseminated through a sedimentary fissure fill unit within

the Moorabool Viaduct Sand. At present, it is only found

within units exposed in the southeast wall of the Australian

Portland Cement Limited (A.P.C.L.) quarry at Batesford,

Geelong (Fig. 4.1).








48







---A AAA .AA^AA^^ AAAAAAAAAAAAAAAAAA
V AAAAAAAAA AAAAAAAAAAAAAAAA
S 'AAAAAAAA AAAAAAAAAAAAAA _
SAAAAAA AA A AAAAAAAAAA
AAAAAAAAAAA AAAAAAAA
AAAAAAAAAAAAAAAA AAAAAAAA
3 -- AAAAAAAAAAAAAAAA AAAA AA
S. \ AAAAAAAAAAAAAAA AA AA
AAAAAAAAAAAAAAAAA AAAAAAA AUSTRALIA
A AAAAAAAAAAAAAAAAA AAAAAA
C AAAAAAtAAAAAAAAAAAAA A


"AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAA
S ---AAAAAAAAAAAAAAAAAAAAAAAAAAAA
-- AAAAAAAAAAAAAAAAAAAAAAAAAAA
S I -- AAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAA
S AAAAAAAAAAAAAAAAAAAA^^^^^^^^^^^^^^^^AAAAA VICTORIA
.+ -- -%AAAAA AAAAAAAAAAAAAAAA
+ "V AAAA AAAAAAAAAAAAAA
S- --AAAAAAAAAAA MELBOURNE

ARi -- sAA AAAAAA AA /'
+'v DOG ROCKS AAAAAAAA A/ A


.. .,,vv v LOCAL FAUNA AAAAAA: X V V A
-A AAAAA142E E 14 AAAAAA
AT ES4 R -I_ AAAA^ A GEELO -
--" -- -. AAAAA A A '



;' \- ,-.\ ". ^ "' 'D-- O\ R ^ ^ ^^., A .- 0 -0* :
Lw / AAAAA





vvv V S ] ..- 'A AA AA A4 A .2-' 14 E_ m 4t. O'
VV W,'_ V A AAAA A -

/ /"' -..A .d.-
^AAA AAA A
vvvv V AAAAAAA A







v AVVV VVVVIUV V V IADD AASAAAAAA..AAAA
V VVVAAAAA AA A A A A A











N E W E R F Y A N S F ^ ^ ^ A A ^ ^ ^ ORA NAT E

NEWTERN TvMOIv, 0UA ^^^^^S F 0 UR
and: Geology A A GeA e long Ar e .
VC V AV AAAAAAA AAAA -.-



AIAA LV NW VIFADTFAAND SA;: A- GR4f
.1/. .~AAAA AAAA
VVVVVV VVVV AAKAAAAAAAAA AAAAAAA.1AAAAA











VOLCANIC I --I FORMATION AOR!
VV VVVVVVV AAAAAAAAAA AAA
VV A vv vv vA : : I (~ l AAAAAAAAAAAA .A .1.1.1AAAAAA

%vv vvvvvv v

vvvvvvv Av -


O OUATERNARYv GEELORNG 5UARRY-, 1A A
ALUVUMVVVVVDVVVVV AA S AN SCALE (













vvvv NEWER BATSFORD
VOLCANIC 2 LIMESTONE










Figure 4.1. Locality of the Dog Rocks Local Fauna Section
and Geology of the Geelong Area.









This local fauna consists of some 29 identifiable

vertebrate taxa (Table 4.1). The Macropodidae dominate the

assemblage but other groups include Diprotodontidae,

Vombatidae, Dasyuridae and the Muridae. Although this local

fauna is poorly preserved, the combination of its taxonomic

diversity and the possibility of a well constrained late

Pliocene date makes it an important biostratigraphic

reference assemblage for the Australian Neogene. Based on

stratigraphic relationships and an overlying basalt, the age

of the site has been stated to be in the range of 2.0-4.0 Ma

(Woodburne et al., 1985). In an effort to narrow this

range, sediment and basalt samples suitable for

paleomagnetic analysis were collected and treated at the

University of Florida Paleomagnetics Laboratory. The

results of this work, together with an assessment of the

chronological control on this site, are the purpose of this

paper.


Previous Work


The present interpretation of the geology of the

Geelong area was established by Bowler (1963) and is

summarized, with minor modifications, by Abele et al.

(1988). Bowler identified and described the units of

interest in this report; namely 1) the middle Miocene

Batesford Limestone, the target of the A.P.C.L. quarrying

operations, 2) the late Miocene-Pliocene Fyansford Formation

and 3) the Pliocene Moorabool Viaduct Sand (MVS), the latter










two being the main overburden units in the quarry. He

considered these units to be of shallow to marginal marine

origin; however, he also noted the presence of leaves and

other organic material towards the top of the MVS,

indicating at least the possibility of local sub-areal

deposition. These are overlain by late Pliocene flood

basalts which were examined at sites elsewhere in the

Geelong area. These sites produced an average date of 2.01

+ 0.13 Ma and were found to carry a reversed magnetic

polarity (Aziz-ur-Rahman and McDougall, 1972).

The Batesford Quarry is well known (Singleton, 1941;

Bowler, 1963; Rich, 1976) for its rich middle Miocene

invertebrate and vertebrate marine faunas. It is the type

section for the Batesford Limestone and the type locality

for the Batesfordian stage (Bowler, 1963). However, the

terrestrial vertebrate fauna was unknown before 1975 when a

pair of diprotodontid mandibles of Zygomaturus sp. were

discovered in an overburden dump (Rich, 1976). The origin

of these mandibles was traced back to a series of

sedimentary fissure fills occurring within the MVS and

periodic prospecting of this site has continued to produce

additional important material. Mining activities along this

face led to the destruction of the unit but the A.P.C.L.

management kindly saved approximately 80 metric tons of the

fissure fill sediments. This material was seived and picked

by the author during 1985-86 for a total yield of 3.5 kg of









bone fragments and approximately 250 recognizable teeth.

The vertebrate taxa recognized are listed in Table 4.1.


Geologic Setting


The geology of this site is well exposed in the

southeast wall of the quarry. It is characterized by the

presence of four units, each separated by a distinct

depositional hiatus (Fig. 4.2). The basal unit is the

middle Miocene (Batesfordian Stage) Batesford Limestone and

is characterized by a white to cream colored soft carbonate

matrix consisting of fragmented reef corals and bryozoans

together with a rich invertebrate shell fauna, shark teeth,

and rare cetaceans. A change from oxidizing to reducing

environments in the shallow Balcombian seas of the area

(Bowler, 1963; Abele et al., 1988) led to the deposition of

the dark and often pyritic Fyansford Formation. It consists

of clays which contain fragments of the Batesford Limestone,

an indication of active marine erosion and reworking within

the depositional basin, and are represented by up to 25 m of

vertical exposure within the Batesford Quarry.

Continued shallowing of the seas led to diachronous

north to south erosion and a faces change characterized by

the deposition of the argillaceous Moorabool Viaduct Sand

during Cheltenhaminan to Kalimnan times (Fig. 4.2). Most of

the MVS facies appear to be of marine origin but some of the

upper units contain leaf material and organic debris and are

















HEIGHT





30.0


SITE
LOCATIONS
"z
4
ow
206 o
K-Ar DATE <
2.03rO.13 Eco
Mao


vv "NEWER V

6VOLCANiiCS



FISSURE
FILL
SAGE=? -







..'M00RABOOL
v'"IIVIADUCT-7



























_FYANSFORD-
-FORMATION










CATESF0R0E-
LUMEfST qNEZ_


Figure 4.2. Stratigraphic column of the Dog Rocks Local
Fauna Section (Batesford Quarry).


106
309
GI oorovala
crossafo m.s
308
105
307

306



303
104

304

103



-107







305




302

102


301

101


25.0-







20.0


10.0






5.0






0-


4









probably sub-areal (Bowler, 1963). Vertical exposure of

this unit in the quarry wall is variable but averages

approximately 15 m.

Occasional sedimentary fissure fill units, which yield

the Dog Rocks Local Fauna, cut through the Moorabool Viaduct

Sand. They are characterized by grey, sulphurous clays

which contain large amounts of reworked material from all of

the underlying units as well as carbonaceous organic

material and occasional bone fragments. As they were

apparently formed under sub-areal conditions, they are

thought to be of Werrikooian age or younger (Woodburne et

al., 1985).

Two thin tongues of basalt, representing a flow of the

late Pliocene Newer Volcanics, overlie the sedimentary

units. The Newer Volcanics are dated from several locations

around the Geelong area (Aziz-ur-Rahman and McDougall,

1972). The average age of three of these sites, taken from

a group of quarries approximately 3.0 km south of Batesford

Quarry, is used to infer a date of 2.03 + 0.13 Ma

(corrected, using the decay constants of Faure (1986)) for

the age of the flow at the Batesford Quarry.


Micro-invertebrate Analysis


Micro-invertebrate faunas were recovered from samples

collected during the paleomagnetic work. The highest

productive sample in the sequence, site 309, yielded a

microfauna consisting of abundant benthonic and rare









planktonic foraminifera. Forms ranging from middle Miocene

to middle Pliocene in age were recognized, indicating

considerable reworking within the sequence. However, the

planktonic species Globorotalia crassaformis, which has a

first appearance datum of 4.0 Ma, (Kennett and Srinivasan,

1984) is present (Dr. David Hodell, pers. comm.) which

constrains the maximum age of the bone unit to less than 4.0

Ma.


Paleomagnetic Procedures and Results


Paleomagnetic samples were taken from 17 sites on the

southeast wall of Batesford Quarry during September and

November 1986. These include one site in the overlying

basalt, 11 sites in the MVS and five from the Fyansford

Formation, all of which are in close proximity to the fossil

unit. An attempt was made to take samples on both sides of

unit contacts and at mid points within units with three to

four separately oriented hand samples taken at each site.

As this was the working face, access was limited by the

condition and stability of the quarry wall which

necessitated collection along two roughly vertical sections

instead of a single continuous section (Fig. 4.2).

Individual samples from each site were subjected to

alternating field (AF) and thermal demagnetization in the

Paleomagnetics Laboratory at the University of Florida.

This facility is housed in a shielded room (see Scott and

Frohlich, 1985, for similar design details) which attenuates










the ambient field to ca. 200 gammas. Stepwise AF and

thermal demagnetization was carried out over a range of 0-50

mT (at 5 mT intervals) and a range of 0-625o C (with 13

steps) respectively, using Schonstedt AF and thermal

demagnetizers. In some cases an AF demagnetization series

was followed by selected thermal demagnetization steps in an

attempt to remove high coercivity remanent overprints.

Remanent magnetism of each sample was measured with a

Superconducting Technology cryogenic magnetometer (Goree and

Fuller, 1976) or, if magnetizations were too strong, e.g.

the basalts, a Schonstedt spinner magnetometer.

The demagnetization characteristics of these samples

was often unpredictable. However, a trend towards a stable

characteristic component in the range of 440-500o C emerged

from the successfully treated samples (Fig. 4.3a,b). Many

samples did not react well to these techniques and seven

sites were rejected for this analysis. Six sites produced

two samples each of concordant directions and are

categorized as Class III sites after Opdyke et al. (1977).

Four sites produced three samples each of statistically

significant demagnetization characteristics and produced R

values of >2.62. They are categorised as Class I sites

(Fig. 4.4). Isothermal remanent magnetization saturation

experiments were carried out on 11 samples from different

sites. Greater than 80% saturation was reached in all

samples by 140 mT but no samples saturated completely (Fig.

4.3c). This is interpreted to indicate the presence of at













A
N IP)

.OnA/n. v







0 500"C ~12 MlA
4002 C 300/ C



NRM
S(OOWN)






B
NILP
90-oAA.


30 mT< ~---^ "soo
0 T6 T


200 T AA
O OmTT



Nam
H

S (COWN)

C
INTENSITY (j/j)
1400


1200

1000oo


800


600


400


200


0 to ZD 30
TREATMENT (Am-I)






Figure 4.3. Paleomagnetic data plots; (A) Thermal
demagnetization Zijderfeld plot (Site 309); (B)
AF demagnetization Zijderfeld plot (Site 309);
(C) Isothermal remanence saturation plot (Site
305).









least two carriers of the NRM. The IRM results, together

with thermal demagnetization trends, indicate the presence

of both low and high coercivity magnetic minerals which are

here interpreted as magnetite and goethite, respectively.



Magnetic Polarity Stratigraphy

The magnetic polarity stratigraphy of this section is

characterized by a single zone of reversed polarity which

extends from the base of the Fyansford Formation, through

the Moorabool Viaduct Sand, and the overlying flow of the

Newer Volcanics. As each of these sedimentary units are

separated by erosional unconformities, it is possible that

this reversed signature may not represent a single polarity

zone from the time scale. The reversed polarity of the

basalt agrees with the extrapolation to this site by Rich

(1976) and Woodburne et al. (1985) from previous work in the

Moorabool and Barwon River valleys (Aziz-ur-Rahman and

McDougall, 1972). The magnetostratigraphic data tends to

confirm the K-Ar date of 2.03 + 0.13 Ma for the basalt at

Batesford Quarry and suggests that it occurs within the

early Matuyama (2r) Chron (2.04-2.12 Ma) of Harland et al.,

(1982). Given these data, and the lower age constraint of

the G. crassaformis FAD of 4.0 Ma, the magnetic signature

may be correlated with the standard time scale in either of

two ways. It may occur within the early Matuyama (2r)

Chron, giving the fossil unit an age of 2.04-2.48 Ma or it










may occur in the late Gilbert (2Ar) Chron, giving the

assemblage a possible age of 3.40-3.88 Ma. In either case,

the date of the fissure fill unit is constrained by the

magnetostratigraphy to a lower maximum age of 3.88 Ma.


Correlation to the Timescale


Prior to this report the age constraints on the Dog

Rocks Local Fauna were based on the dissection of the marine

MVS, of Cheltenhamian to Kalimnan age, by the apparently

sub-areal fissure fill sediments in which it occurs, and by

the interpolated K-Ar date on the Newer Volcanics which

overly it. This would appear to place the age of the site

and its local fauna in a range of 2.0-4.0 Ma (Woodburne et

al., 1985).

The lower age constraint is based on the presence of

the Cheltenhamian-Kalimnan macro-invertebrate fauna (Abele

et al., 1988) and the marine/sub-areal faces change that

occurs within the MVS (Bowler, 1963). In some sections of

the Geelong area the change is characterized by a

discontinuity of unknown duration and in other areas it is

diachronous. Undoubted sub-areal MVS, apart from the

fissure fill sediments, are not recognized at the Batesford

Quarry site. However, based on Bowler (1963), the basal age

could conceivably range from the early Kalimnan, when sub-

areal deposits were first known in the area. The micro-

invertebrate data confirm this evaluation and place an









approximate date of 4.0 Ma on the marine sediments

underlying the Dog Rocks Local Fauna.

A further constraint on the lower age limit of this

fauna is represented by the general characteristics of the

vertebrate fossil assemblage. The stage of evolution of the

macropodidines and the presence of rodents suggest a late

Pliocene to early Pleistocene date (Woodburne et al., 1985).

The rodents are particularly important in this sequence.

The oldest murids in Australia are known from the Chinchilla

and Bluff Down localities (Woodburne et al., 1985) and

appear to place the entry of this group into the continent

at about 4.5 Ma. However, the presence and relative

diversity of the Dog Rocks murids suggest that this site

must significantly postdate the first appearance of the

group on the Australian mainland. Similarly, the presence

of the Hamilton fauna (Turnbull and Lundelius, 1970) some

200 km to the west, an assemblage directly overlain by a

basalt dated at 4.47 Ma (Turnbull, Lundelius and McDougall

1965), which contains a diverse fauna of some 26 taxa but no

murids, limits the apparent age of the Dog Rocks Local Fauna

to less than 4.47 Ma.

The date on the overlying basalt is based on indirect

evidence and requires some discussion. This basalt is one

of a series of flows of the Newer Volcanics that filled the

paleovalleys of the Moorabool and Barwon Rivers during the

late Pliocene. They are dated using samples taken from

three localities, one each from the Geelong Quarries Ltd.





























































Figure 4.4. Correlation of Dog Rocks Section
Magnetostratigraphy to Magnetic Polarity
Timescale.









site, where at least three separate flows are exposed, and

the Mobile and Fyansford Quarry sites where at least two

flows are exposed in each (Fig. 4.1). The exposure of the

flow at the Batesford Quarry lies approximately 2 km north

of the former and approximately 3 km north of the latter but

cannot be traced to either quarry because of Quaternary

alluvial cover (Fig. 4.1). As a result, it is not possible

to observe which flow, if any, is common to both Batesford

and the other quarries. Aziz-ur-Rahman and McDougall (1972)

give K-Ar dates from 'a single flow in the Moorabool

Valley'. They cite petrological evidence and the occurrence

of a reversed magnetic polarity at each site as

corroborative evidence that they sampled within the same

flow. All later authors have accepted the average date for

the three quarry sites of 2.03 + 0.13 Ma (Aziz-ur-Rahman and

McDougall, 1972, corrected) as a minimum date for the

Moorabool Valley basalts and as the age of the Batesford

Quarry flow. The presence of a reversed magnetic polarity

in the Batesford Quarry basalt, as revealed in this study,

provides the first concrete evidence which corroborates this

assumption (Fig. 4.4). I conclude that a date of 2.02-2.42

Ma (Chron 2r after Harland et al., 1982) is likely for the

basalt and that the minimum date of 2.03 + 0.13 Ma for the

age of the Dog Rocks Local Fauna is valid.

An alternate maximum date for this fauna, as

constrained by the foraminiferal and paleomagnetic data

could also be ca. 3.88 Ma. However, the presence of a










relatively diverse rodent fauna suggests that the younger

age is more likely for this assemblage.




Conclusion


Australian paleontology suffers greatly from a lack of

vertebrate sites with absolute chronological control.

Although the Dog Rocks material is of relatively poor

quality, its faunal diversity (29 identified taxa) and its

age, as constrained by the results of this report, make it a

site of considerable importance in the Australian record.

Given the presence of the rodent fauna, a late Pliocene age

is considered most likely for this site. Therefore, an age

of 2.03-2.48 Ma is proposed for the Dog Rocks Local Fauna.











TABLE 4.1

Currently Recognized Vertebrate Taxa From the
Dog Rocks Local Fauna


Class Osteichthyes
Teleostei, indet
Class Amphibia
Anura, indet.
Class Reptilia
Squamata, indet.
Class Aves
Aves, indet.
Class Mammalia
Subclass Marsupialia
Family Dasyuridae
Dasyurus sp.
Antechinus sp.
Superfamily Syndactyla
Family Peramelidae
Perameles sp.
Isoodon sp.
Pseudocheirus sp.A
Pseudocheirus sp.B
Family Macropodidae
Macropus cf. qiqanteus
Macropus cf. fuliginosus
Macropus cf. irma
Macropus sp. A
Macropus sp. B


Family Macropodidae (cont.)
Protemnodon cf. anak
Protemnodon sp. A
Sthenurus sp.
Wallabia cf. bicolor
Wallabia sp. A.
Family Potorooidae
Potorous (Bettongia)
Tropsodon sp.
Family Phalangeridae
PhalanQer sp.
Family Vombatidae
Vombatus ursinus
Phascolonus sp.
Family Diprotodontidae
Zvyomaturus sp.
Family Petauridae
Petaurus sp.
Subclass Eutheria
Order Rodentia
Family Muridae
Pseudomys sp. A
Pseudomys sp. B
Murid indet. A
Murid indet. B













CHAPTER 5
MAGNETIC POLARITY STRATIGRAPHY OF THE DUCK PONDS AND
LIMEBURNER'S POINT VERTEBRATE FOSSIL FAUNA SECTIONS


Introduction


The Duck Ponds and Limeburner's Point local faunas

were both discovered in the last century, the former in

1875 during the excavation of a railway viaduct at Lara

(formerly Duck Ponds, 38 2' S, 1440 24.4' E)(Smyth R.

Brough, 1876) and the latter in 1895 at a lime kiln works

at Limeburner's Point (formerly Point Galena, 380 10' S,

1440 23' E) (Pritchard, 1895) (Fig. 5.1). The Duck Ponds

Local Fauna includes Thylacoleo carnifex, Diprotodon cf.

D. longiceps, Protemnodon cf. P. anak and Macropus titan

and was found in fluviatile sediments underlying the Lara

Limestone (Wilkinson, 1972) and overlying a flow of the

Newer Volcanics (Fig. 5.2). The Limeburner's Point Local

Fauna was recovered from a freshwater limestone which is

either an equivalent or correlative of the Lara Limestone

(Fig. 5.2). The Limeburner's Point Local Fauna is

undescribed but is under study by Turnbull at the Field

Museum in Chicago. It currently includes Diprotodon

longiceps, Sminthopsis orientalis and Sarcophilus

(Turnbull, pers. comm., 1990). On the basis of local

stratigraphy and the included mammal assemblages both

faunas are considered to be of early-middle Pleistocene











































































Figure 5.1.


,, "vvvV vvvvvv K-Avvv
vv \ lvvvvvvvvvvK-vsvA',
vvv ----- VVVVVVVVVV Ivy
V V V V V ----- --"-"----"----IVV V V V V V V V V
SVvvvvV -V'v 6+0.03
V--VVVV ----vvv --v v v v v
vvv 0 -LA V v ---- V1. vvvvv
vvvvv vvvvvvvvvvv
Melbourne VVV DUCK PONDS vvvvvvvvvv
Geelong LOCAL FAUNA v vvvvv v". v
3S vV^ v v ^vvvvvvvvvvvv
vv v vvvvvvvvv, vv%
^3S S I vvvvv vvvv vvv5^5i K^'-^-- ~ /vvvvvvvvvvvtvv
O 50 vvvvvvvv -- vvvvvvvvvvvvv
142* 14 E 146'vvvvvvvv vvv vvvvvvvvvvvvvvv
VVVVVVVVVVVVLVVVVVVV VV VVVVVVVVVVVVVVVV
VVVVV VV VVVVVV VVVV ----- VVVVVVVVVVVVVVVVV
vvvvvvvvvvvvvvvvvvV vvvvVVVt vv vvvvvvvvvv
'VVVVVVVVVVVVVVVVVV VVVVV-V VVVVVVVVVVVVVVVVV
VVVVVVVVVVVVVVVVVV VVVVVV --- VVVVVVVVVVVVVVVVVVVV
VVVVVVVVVVVVVVVVVV VVVVVV ---- VVVVVVVVVVVVVVVVVVVV
vvVVVVVVvvvvv VVV VVVVVV VVVVVVVVVVVVVVVVVvv
,VVVVVVVVVVVVVVVV VVVVVV VVVVVVVVVVVVVVVVVV V
VVVV VVVVVVVVVV VVVVVV VVVVVVVVVVVVVVVV
,VVVVVVVVVVVVV VVVV VVVVVVVVVVVVVVVVVV
VVVVVVVVVVVVV VVVVVV VVV VV VVVVV
SVVVVVVVVV VVVV
vvvvvvvvvvvvvvvvv`vvvvvvvv VV V vvvvvv
,vvvvvvvvvvvvv vvvvv vBvvvvvvv vvvvvvvv
*V vvVv vvvvvvv vvv vvVVV v*-

vvvvvvvvvvvv
rvvvvvvvvvvv /< t 1 /VV *O' 'Ky"

,vvvvvvvvvvvvv Limeburner's
vvvvvvvvvv y
vy Bay any
vv vvv Pt. Lillias
vv

: Corio Geelong
vv Bay Outer
v .*'*** Harbour
v Pt. Henry
'vvI i LIMEBURNER'S
vvv^^ POINT
vvvy LOCAL FAUNA
' GEELOPNGS----
..:. vvvv

VV :-Q VVV
AAAAAAAA VV, V
.AAAAAAAAAA'
AA AAAA AA
AA ----- -





FO1RECENT 1 0 1 2

03 NV2 VO LCANICS km


0 2 NV1 VOLCANICS

01 SMOORABOOL VIADUCT
Q1SANDS


Location and Geologic Map for the
Limeburner's Point and Duck Ponds local
fauna sections and the Limeburner's Bay
geologic section.









age (Wilkinson, 1972). Woodburne et al. (1985) correlate

a basalt that occurs at the base of the Duck Ponds

section to a flow exposed 5 km to the northeast of Lara

(380 0.6' S., 1440 28.3' E), which has a 40K-40Ar date of

1.66 + 0.03 Ma (Aziz-ur-Rahman and McDougall, 1972) (Fig.

5.1). Therefore, Woodburne et al. (1985) propose an age

of less than 1.66 + 0.03 Ma for the Duck Ponds Local

Fauna. Recently, magnetic polarity stratigraphy has been

highly successful in resolving and refining the ages of

several southern Australian mammal faunas (MacFadden et

al., 1987; Whitelaw, 1989). This study reports on a

magnetic polarity stratigraphy developed in the Geelong

area and its implications for the age of the Duck Ponds

and Limeburner's Point Local Faunas.


Geologic Setting


The Duck Ponds Local Fauna was recovered at a depth

of 6.4 m below the level of Hovell's Creek, in the

foundations of the Lara Railway Viaduct (Wilkinson,

1972). Access to the site is no longer possible, but a

complete description of the geology based on bores taken

near the viaduct is given by Keble (1945) after Daintree

(1863), and summarized by Wilkinson (1972) (Fig. 5.2).

The oldest unit in this section is a flow of the Newer

(Upper) Volcanics Basalt (lava plain stage) which is

described in Keble (1945) as soft decomposed basalt.

Similar basalts outcrop on the nearby Werribee Plain as















DUCK
PONDS
SECTION
(KEBLE, 1945)
DEPTH
(m) O


6.0-


LIMEBURNER'S VGP
POINT LATITUDE
SECTION


0 +90


/ POINT L.F.










02 LIMESTONE

S02 COMPACT LIMESTOI

W 02 CLASTICS

NV1 BASALTS

01 CLASTICS


SITE
CLASSIFICATION
CLASS I
O CLASS III




LIMEBURNER'S
BAY
SECTION
SITE
vv 3.0
106-
vvvvv
105 .
104- .
103 -2.0
---2.0-

102-

101 -
1.0

1000
E
vvvvvv
v vvvv
vv vy
HEIGHT
(m)


Figure 5.2.


(A) Stratigraphic column of the Duck Ponds
(Lara) section (after Keble (1945)); (B)
Stratigraphic column and Magnetic Polarity
Results for the Limeburner's Point section;
(C) Stratigraphic column and Magnetic
Polarity Results Limeburner's Bay section.









Nvl basalts (Spencer-Jones, 1970). The basalt is

overlain by fluviatile sediments that consist of muds,

clays, sands and gravels which Keble (pp. 30, 1945)

called 'early flood plain deposits'. Within the section,

at a depth of 20 feet (6.4 m) Keble (pp. 30, 1945)

describes the occurrence of a 'bed of quartz gravel and

rotten shells overlying a stiff clay'. Wilkinson (1972)

identified the same matrix on some of the bones in the

Duck Ponds collection and concluded that the Duck Ponds

Local Fauna was derived from the same horizon.

The fluviatile sediments are overlain by 16 feet

(4.9 m) of the Lara Limestone which Keble (pp. 30, 1945)

subdivided into a basal 'soft rubbly limestone', a

'compact limestone, containing freshwater shells', and an

upper rubblyy limestone'. The Lara Limestone is overlain

by 1.3 m of soft sandy loam. This sedimentary sequence

is described as Q2 on the 1:63,360 geologic map (Spencer-

Jones, 1970), which uses a Q1, Q2, Q3 notation to

characterize the oldest to youngest Quaternary sediments

of the area (Fig. 5.1).

Since the original Duck Ponds section is now buried

under the viaduct, an indirect approach to dating the

section was used. An attempt to refine the basal age

constraint for the fauna was made by sampling a section

of Q1 sediments which antedate the Q2 suite at Duck

Ponds. A section located on the west side of

Limeburner's Bay, into which Hovell's Creek flows, was









chosen (Fig. 5.1). It contains a 2 m exposure of Q1

sediments consisting of sands, silts and clays, that are

overlain by a 3.5 m thick flow of Nvl basalts and

underlain by a second Nvl basalt which outcropped at and

below water level, towards the south end of the section

(Fig. 5.2).

The Limeburner's Point Local Fauna was collected

from a section located on Limeburner's Point

approximately 80 m west of some abandoned lime kilns, on

the south side of the Inner Harbour of Corio Bay (Fig.

5.1). The best description of the Limeburner's Point

section is by Keble (1945, after Daintree (1863). Keble

describes the lower 11.3 m (37 feet) of the 21.3 m (70

foot) cliff section. He notes a 2.1 m (7 foot) thick

freshwater limestone that outcrops at and extends several

feet above and below sea level and that contained

freshwater shells that 'are identical with those obtained

in the well in the Duck Ponds (Lara) Limestone' (pp. 31,

Keble, 1945). The Limeburner's Point Local Fauna, which

includes Diprotodon longiceps, was recovered from this

limestone and, both Daintree (1863) and Keble (1945)

considered it to be contemporaneous with the Duck Ponds

(Lara) Limestone. Spencer-Jones (1970) concurred with

this assessment and classified both the Limeburner's

Point limestone and the Lara Limestone with its

underlying sediments, as Q2 outcrops (Fig. 5.1). Most of

the section has been obscured since Keble's (1945)









report, but outcrops of the freshwater limestone are

preserved at sea level and four paleomagnetic sites were

collected (Fig. 5.2).


Paleomagnetic Procedures and Results


At the Limeburner's Bay and Limeburner's Point

sections, seven and four sites, respectively, each with

three separately oriented hand samples (A, B and C) were

collected. These samples were prepared and analyzed in

the Paleomagnetics Laboratory at the University of

Florida. This facility contains a Superconducting

Technology cryogenic magnetometer (Goree and Fuller,

1976), a Schonstedt spinner magnetometer and Schonstedt

AF and thermal demagnetizers. Samples were measured in

the shielded room laboratory which attenuates the ambient

field to ca. 200 nT (Scott and Frohlich, 1985, provide

similar design details).

The Limeburner's Bay basalts were all subjected to

stepwise AF demagnetization carried out over a range of

0-100 mT (at 10 mT intervals) (Fig. 5.3). For both

localities, one sample from each site was treated with a

stepwise thermal demagnetization regime over a range of

0-6100 C (16 steps) (Fig. 5.3) and one by a stepwise AF

demagnetization regime over a range of 0-50 mT (10

steps). In both sections the AF demagnetization

treatment did not always successfully isolate a stable

component. Therefore, they were further treated by a















N(UP) E

5250 C
5000 C
4500 C
50 1000C
mT
-40
mT
{ 30 mT

5NRM v
S(DOWN) H


N(UP)


A/m


Figure 5.3. (A) AF zijderfeld plot Limeburner's Bay
basalt (Site 104); (B) Thermal zijderfeld
plot from Limeburner's Bay (Site 103); (C)
AF/Thermal zijderfeld plot from Limeburner's
Bay (Site 102); (D) IRM plot from
Limeburner's Bay (Site 103); (E) Thermal
zijderfeld plot from Limeburner's Point (Site
203); (F) AF/Thermal zijderfeld plot from
Limeburner's Point (Site 203); (G)
IRM plot from Limeburner's Point (Site 201).









stepwise thermal demagnetization regime over a range of

0-610 C which successfully demagnetized them (Fig. 5.3).

As the thermal treatment was the most successful

demagnetizattion method, the third sample from each of

the sediment sites was treated by this method.

Isothermal remanent magnetization (IRM) experiments

were carried out on four representative samples from the

Limeburner's Bay section and two samples from the

Limeburner's Point section (Fig. 5.3). Greater than 80%

saturation was reached by 140 mT in all four of the

Limeburner's Bay samples and by 80 mT in both

Limeburner's Point samples. In both cases the presence

of a low coercivity mineral as a dominant carrier of the

natural remanent magnetism (NRM) is indicated. These

data, together with thermal and AF demagnetization

characteristics from both the sediment and basalt samples

suggest that magnetite is the dominant carrier of the NRM

in both sections.

The magnetic polarity stratigraphy of the

Limeburner's Bay section is characterized by a single

zone of reversed magnetic polarity. A stable

characteristic component was isolated for all the

sediment sites over a range of 150-5900 C. All four

sites produced three samples each with concordant

directions; two sites have statistically significant mean

directions with R values >2.62 (after Fisher, 1953) and

are categorized as Class I reversed polarity sites (after









Opdyke et al., 1977) and two produced R values <2.62 and

are categorized as Class III reversed polarity sites

(Fig. 5.2). Samples from all three basalt sites in the

section produced a stable characteristic component over a

range of 30-100 mT after a low coercivity, viscous normal

overprint was removed. All three sites produced three

samples each with statistically significant mean

directions (R values >2.62) and are categorized as Class

I reversed polarity sites (Fig. 5.2).

The magnetic polarity stratigraphy of the

Limeburner's Point locality is characterized by a single

zone of normal magnetic polarity. All four sites

produced three samples each with stable characteristic

components and concordant directions over a range of 100-

4500 C. Three of the four sites have statistically

significant mean directions with R values >2.62 and are

categorized as Class I normal polarity sites (Fig. 5.2).

One site produced an R value of <2.62 and is categorized

as a Class III normal polarity site.


Correlation to the Timescale


A correlation to the geomagnetic polarity timescale

for the Limeburner's Point, Limeburner's Bay and Duck

Ponds sections is dependent upon three stratigraphic

interpolations. A basal age for these sections is

provided by the age of the Moorabool Viaduct Sands, which

underlie the area (Spencer-Jones, 1970). Magnetic










polarity stratigraphy of an exposure at Batesford Quarry

indicates that the Moorabool Viaduct Sands has an age of

1.88-2.47 Ma (Whitelaw, 1989). Therefore, the age of the

Duck Ponds and Limeburner's Point sections and their

enclosed local faunas must be younger than 2.47 Ma. A

second constraint is indicated by Woodburne et al.

(1985), who recognize basalts exposed on the Werribee

Plain as correlatives of the basalt in the Duck Ponds

section. Aziz-ur-Rahman and McDougall (1972) determined

a 40K-40Ar age of 1.66 + 0.03 Ma (corrected from 1.62 +

0.03 Ma by Woodburne et al., 1985) and a reversed

magnetic polarity for one of these basalts some 6 km

northeast of the Duck Ponds locality (Fig. 5.1).

Woodburne et al. (1985) use this isotopic date as a

maximum age for the Duck Ponds Local Fauna (Fig. 5.4).

This is in general agreement with Wilkinson (1972), who,

on the basis of the fauna, suggested an early to middle

Pleistocene age for the Duck Ponds Local Fauna. Thirdly,

on the basis of lithologic similarity and a shared fossil

assemblage, the Q2 outcrop of freshwater limestone at

Limeburner's Point is considered to be contemporaneous

with the Lara Limestone which occurs in the Duck Ponds

section (Daintree, 1863; Keble, 1945; Wilkinson, 1972)

(Fig. 5.2).

If the above correlations are correct, then it is

possible to construct a composite stratigraphy that

further constrains the age of the Duck Ponds Local Fauna.













DUCK LIMEBURNER'S
PONDS POINT
SECTION SECTION
DEPTH
(m)


CORRELATION
TO MAGNETIC
POLARITY
TIMESCALE
O.0.0 Ma


' "98. LIMEBURNER'S
SECTION
'E BAY
SITE
1\ 3.0
106%vvv
\ vvy v
105- vw w
.66 v 104- vvvvv
1.88 \103 2.0
102- -

11 -1.0

2.47 100
0

LIMESTONE HEIGHT
(m)
COMPACT LIMESTONE


| 02 CLASTICS


NV1 BASALTS

01 CLASTICS







Figure 5.4. Composite stratigraphic section and
correlation to the timescale for the
Limeburner's Bay, Limeburner's Point and
Duck Ponds Sections.









This composite section consists of a basal unit of the

Werribee Plains basalt (Nvl) overlain by the Limeburner's

Bay (Q1) and Limeburner's Point (Q2) sections (Fig. 5.4).

The resultant magnetostratigraphy may be characterized by

a lower zone of reversed polarity (Nvl) overlain by a

second, or part of the same, reversed polarity zone as

preserved in the Q1 sediment sequence. This is overlain

by a normal polarity zone, as preserved in the Q2

freshwater limestones (Fig. 5.4). Given a basal age

constraint of 1.66 + 0.03 Ma, there are only two possible

correlations of this sequence to the geomagnetic polarity

timescale. Using the timescale of Berggren et al. (1985)

the age of the Ql sediments must fall in one of the upper

Matuyama Chron reversed zones of 1.66-0.98 Ma or 0.91-

0.73 Ma and the age of the Q2 freshwater limestones must

fall within the Jaramillo Normal Subchron of 0.98-0.91 Ma

or the Brunhes Normal Chron of 0.73-0 Ma (Fig. 5.4).

The presence of a reversed polarity in the Q1

sediments supports the suggestion by Woodburne et al.,

(1985) of a basal age of 1.66 + 0.03 Ma for the Duck

Ponds Local Fauna. The polarity of the Q2 fluviatile

sediments, which contain the Duck Ponds Local Fauna,

cannot be uniquely resolved from the available data but

it is overlain by the Lara Limestone, which by

correlation to the Limeburner's Point section, has a

normal polarity. Therefore, the age of the Duck Ponds









Local Fauna remains constrained to less than 1.66 Ma

(Fig. 5.4).

The Limeburner's Point Local Fauna lies within Q2

sediments and stratigraphically above the Duck Ponds

Local Fauna. Therefore, it too is constrained by a basal

age of 1.66 0.03 Ma. The Limeburner's Point Local

Fauna also falls within a zone of normal polarity further

constraining the age to either 0.91-0.98 or 0-0.73 Ma.


Conclusions


The use of magnetic polarity stratigraphy to date

the Duck Ponds and Limeburner's Point Local Faunas is

dependent on two correlations. The first is a proposed

interpolation between the basal basalt in the Duck Ponds

section and a dated basalt (1.66 + 0.03 Ma) located 5 km

to the northeast on the Werribee Plain (Woodburne et al.,

1985). The second is a correlation, based on similar

lithologies and shared fossil assemblages, between the

freshwater limestone exposed at Limeburner's Point and

the Lara Limestone, as described in the Duck ponds

section. If these correlations are correct, then the age

of the Duck Ponds Local Fauna must be younger than 1.66

Ma and probably, older than 0.98 Ma. The age of the

Limeburner's Point Local Fauna is constrained to be

younger than 0.98 Ma.













CHAPTER 6
MAGNETIC POLARITY STRATIGRAPHY OF THE FISHERMAN'S CLIFF
AND BONE GULCH VERTEBRATE FOSSIL FAUNAS


Introduction


The Fisherman's Cliff and Bone Gulch Local Faunas

are, respectively, the second and third oldest mammalian

fossil assemblages known in the Murray Basin of New South

Wales, Australia (Woodburne et al., 1986). The current

faunal list for Fisherman's Cliff contains 22 mammalian

taxa including Diprotodon sp., Lasiorhinus sp. and

members of the Macropodidae, Dasyuridae and a large

number of murids (Crabb, 1975; Rich et al., 1982). The

current faunal list for Bone Gulch contains seven

mammalian taxa and includes Thylacoleo sp., Phascolonus

sp. cf. P. magnus, and members of the Diprotodontidae,

the Macropodidae and the Muridae (Rich et al., 1982).

These localities are units of a suite of five

geographically localized mammalian fossil faunas that

form a temporal sequence in the Murray Basin. Together

with the older Sunlands fauna and the younger faunas from

Frenchman's Creek and Lake Victoria, they document the

evolution of mammalian groups from the Pliocene to the

Recent in this region of Australia. In order to

understand the rates of evolutionary change and faunal

dynamics displayed in these assemblages, it is vital to









place them in a well constrained chronologic framework.

This study presents magnetic polarity stratigraphies for

sections at Fisherman's Cliff and Bone Gulch, and

discusses chronologic age constraints for the Fisherman's

Cliff and Bone Gulch Local Faunas.


Previous Work


The original age for the Fisherman's Cliff Local

Fauna, which is contained within the Moorna Sand, was

based on the geomorphic studies of Gill (1973) and on the

biostratigraphic relationships proposed by Marshall

(1973). The latter suggested a late Pliocene to

Pleistocene age for the fauna on the basis of the

presence of Glaucodon ballaratensis and Protemnodon cf.

P. otibandus. Later, based on the high diversity of

murids in the Fisherman's Cliff Local Fauna Crabb (1975),

suggested a "lower" (early) Pleistocene age. Crabb's

faunal age revision was indirectly called into question

by Bowler (1980) based on a magnetostratigraphy developed

at Chowilla, some 75 km west of Fisherman's Cliff (Fig.

6.1). The section at Chowilla contains the Karoonda

Surface and part of the Blanchetown Clay, two regionally

important stratigraphic markers, which were found to

occur within the Gauss Chron (Bowler, 1980; An et al.,

1986). Using these data and a correlation based on the

common occurrence of the Karoonda Surface and the

Blanchetown Clay at both the Chowilla section and

























































Figure 6.1. Location and Geologic Map of the Fisherman's
Cliff and Bone Gulch fossil mammal
localities and extent of Lake Bungunnia.









Fisherman's Cliff, Woodburne et al., (1985) proposed a

Gauss Chron age for the Fisherman's Cliff Local Fauna.

However, this proposal was never tested by direct

sampling of the section in which the mammal fossils

occur.

The Bone Gulch Local Fauna was described by Marshall

(1973), but he did not regard the mammals as being

biochronologically useful. The fauna is contained within

the Blanchetown Clay, and based on the stratigraphic work

of Firman (1965) and Gill (1973), Marshall (1973)

assigned an age of late Pliocene or early Pleistocene to

the fauna. Woodburne et al. (1985), based on the work of

Bowler (1980), correlated the exposed section of the

Blanchetown Clay at Bone Gulch to a section of the same

formation at Chowilla, and proposed an age 'probably

limited to the Matuyama and a position in the earlier

part of the chron, close to the Pliocene-Pleistocene

boundary' (Woodburne et al., 1985 pp. 350).


In 1986 a detailed magnetic polarity stratigraphy,

which used an exposure on the Murray River, at Chowilla,

and a core taken from Lake Victoria, was published (An et

al., 1986). The An et al. (1986) stratigraphy and

magnetostratigraphy functions as a reference section for

the area and allows direct stratigraphic and magnetic

polarity correlations between it and the Fisherman's

Cliff and Bone Gulch sections. The lower part of this










stratigraphy may be correlated to the Fisherman's Cliff

locality through the common occurrence of the Karoonda

Surface (Gill, 1973), a regionally important

stratigraphic marker and to both the Fisherman's Cliff

and Bone Gulch localities through the common occurrence

of the Blanchetown Clay. The purpose of this report is

to test these correlations by establishing magnetic

polarity stratigraphies for the Fisherman's Cliff and

Bone Gulch localities, which may then be compared to the

section studied by An et al. (1986) at Chowilla. A

successful correlation to the Chowilla section will allow

the establishment of chronologic age constraints for the

Fisherman's Cliff and Bone Gulch Local Faunas.


Geographic and Geologic Setting


Both the Fisherman's Cliff and Bone Gulch Local

Faunas were derived from the predominantly lacustrine

Lake Bungunnia series, which consists of the basal Moorna

Formation, the Chowilla Sand and the overlying

Blanchetown Clay. The Fisherman's Cliff Local Fauna was

recovered from the riverine/lacustrine Moorna Formation

from a cliff and associated gullies located on the north

(New South Wales) bank of the Murray River, 22.5 km west

of Wentworth (340 7' S, 1410 39' E) (Fig. 6.1). The Bone

Gulch Local Fauna was recovered from a series of low

lying gulches, formed in the Blanchetown Clay, located on

the north side of the Murray River, approximately 1 km









west of Fisherman's Cliff (340 7' S, 1410 37' E) (Fig.

6.1).

The Lake Bungunnia paleo-lake developed in the Lake

Victoria Syncline as a result of tectonic shifts of the

Pinaroo block and subsequent damming of the Murray-

Darling River system in the middle Pliocene (An et al.,

1986). Sedimentary deposits of Lake Bungunnia cover an

area of some 68,000 km2 in the Murray Basin (Fig. 6.1)

and overlie the regressive marine Parilla Sand. Much of

the post mid-Pliocene sedimentation history of the Murray

Basin, and the Lake Bungunnia sequence in particular, is

characterized by interdigiation and strong lateral and

vertical facies variation. Consequently, correlation of

time and stratigraphic units has been difficult and/or

tenuous. This situation is somewhat alleviated by the

presence of regionally extensive pedoderms or surfaces,

which are preserved as soil horizons, silcretes or

disconformities. These are used as horizon markers over

large areas of the basin (Gill, 1973). A major horizon,

the Karoonda Surface, is present at both the Chowilla and

Fisherman's Cliff sections and acts as an important

stratigraphic benchmark over much of the region.

The base of the Fisherman's Cliff section is formed

by the Moorna Formation (Fig. 6.2), the type section of

which is located at the west end of Fisherman's Cliff

(Gill, 1973). It is described as "an assortment of

mostly unoxidized riverine deposits, gravels and silts









. ." (Gill, 1973 p. 40). At the sampling locality the

Moorna Formation extends down below mean river level so

that its total thickness is unknown, however a section

9.8 m thick was measured (Fig. 6.2). From river level,

the Moorna Formation grades from reddish-brown, silty

clays, containing sand stringers, to massive, medium to

fine-grained quartz sands which are, in turn, overlain by

laminated and crossbedded, coarse grained, quartz sands.

Minor conglomeratic lenses occur towards the top of the

formation. The Moorna Formation exhibits strong lateral

gradation along the cliff face, ranging from poorly

sorted gravels and coarse sands at the western end to

laminated clayey silts at the eastern end. The former

are interpreted to be channel deposits and the latter

flood plain deposits (Gill, 1973). The contact of the

Moorna Formation with the overlying Chowilla Sand is

marked by a disconformity identified as the Karoonda

Surface (Gill, 1973).

The Chowilla Sand is represented by 0.8 m of sand at

Fisherman's Cliff (Fig. 6.2). The Chowilla Sand and the

overlying Blanchetown Clay heavily interdigitate

throughout the area with the former rising through the

latter, either dividing or capping it in many sections.

The thicknesses of the Blanchetown Clay and the Chowilla

Sand vary from 4.6 m and 3.6 m, respectively, at the

western end to 7.6 m and 0.9 m at the eastern end. Crabb

(1975) questions the stratigraphic validity of the









HEIGHT BONE GULCH
HEIGHT
(m) Sne SECTION
2.0 Number
104------- -
103 Blanche-
102- -town-
101 -:Cday:-
0- 100---
FISHERMAN'S
CLIFF
SECTION
18.0
114
113
16.0- 112
111- lanche-
--t owf--
110-
14.0- 09 ay


108 -
12.0 108
107
:Chowila-
106 :-'Saend:-. Kar
10.0 105 Sur
104
204
203
8.0 202 Mooma
115-
201- Fm

6.0- 200- ^w

103 -
102
4.0 101-
100 -
Cliff Base -i-
2.0- Re.


0 River Level


VGP
LATITUDE

90P ~~~ 90

St


900


oonda
face











-


+900












-


SITE
CLASSIFICATION
CLASS I
O CLASS III
REJECTED


Figure 6.2. Stratigraphy and Magnetic Polarity Results
of the Fisherman's Cliff and Bone Gulch
Local Fauna Sections.









Chowilla Sand and the Moorna Formation, both of which he

considers to be intraformational units within or faces

variants of the Blanchetown Clay. However, Crabb (1975)

continued to use the "Moorna Formation" to describe the

stratigraphic unit from which he recovered the

Fisherman's Cliff Local Fauna and the name has persisted

in recent literature. Therefore, I will continue to

follow the original interpretation of Gill (1973) until

the situation is clarified.

The Blanchetown Clay as exposed at Fisherman's Cliff

is characterized by an 8.0 m section of red-green mottled

clays (Fig. 6.2) which exhibit a popcorn like texture

when weathered. These grade into red, sandy clays

containing numerous phosphatic nodules towards the top of

the Fisherman's Cliff section.

At Bone Gulch the entire section consists of

approximately 3.5 m of the Blanchetown Clay exposed in

the walls of a gully. This outcrop occurs

stratigraphically higher within the Blanchetown Clay than

the section at Fisherman's Cliff, but its lithology is

closely comparable with the red-green mottled clays from

that locality. Small lenses of sand, which may be

equivalent to the Chowilla Sand, occur at the base of a

few gullies in the Bone Gulch area. The thickest

vertical section exposed at Bone Gulch is 2.0 m and was

sampled for magnetic polarity stratigraphy (Fig. 6.2).









Paleomagnetic Procedures and Results


A minimum of three separately oriented

paleomagnetic samples were collected from each of 21

sites at the Fisherman's Cliff section and each of five

sites from the Bone Gulch section (Fig. 6.2). Sampling

was conducted at the east end of the main face at

Fisherman's Cliff, in a section where it achieves its

greatest height of approximately 18.3 m. Samples were

collected along a westward traverse, which started on the

bank of the Murray River and moved up the face of the

cliff. A total of twelve sites were collected from the

Moorna Formation at the Fisherman's Cliff section (Fig.

6.2), one from the Chowilla Sand and eight from the

Blanchetown Clay. Samples from Bone Gulch were collected

from a 2.0 m section located in the first of a series of

gullies which incise the Blanchetown Clay, approximately

100 metres from the present position of the north bank of

the Murray River.

Samples were trimmed into 2.54 cm cubes and

impregnated with sodium silicate. Sodium silicate is

helpful in preventing sample collapse, particularly

during high temperature thermal demagnetization, and

laboratory experiments indicate that it carries a

negligible magnetic signal.

One sample from each Fisherman's Cliff site was

subjected to stepwise thermal demagnetization over a









range of 0-6300 C (16 steps) (Fig. 6.3b). A second

sample was subjected to stepwise alternating field (AF)

demagnetization over a range of 0-99 mT (5 mT intervals)

(Fig. 6.3a). The third sample was treated by whichever

method proved most effective for the first two. All

samples from the Bone Gulch section were treated by a

combination of both AF and thermal treatments.

All samples were demagnetized using Schonstedt AF

and thermal demagnetizers. Remanent magnetism was

measured with a Superconducting Technology cryogenic

magnetometer (Goree and Fuller, 1976) in the

Paleomagnetics Laboratory at the University of Florida.

This facility is housed in a shielded room which

attenuates the ambient field to ca. 200 nT (see Scott and

Frohlich, 1985, for similar design details).

The demagnetization characteristics of the

Fisherman's Cliff samples improve upwards through the

section. A stable magnetic component could not be

isolated for six sites (102, 106, 109, 110, 111 and 114),

and these were rejected, site 108 was rejected because of

sample collapse. Other sites produced stable

characteristic components over a range of 200-5250 C in

the thermal samples, and 0-70 mT in the AF samples (Figs.

6.3a,b). Nine sites produced three samples each of

statistically significant demagnetization characteristics

(R values >2.62, see Fisher, 1953). These are classified

as Class I sites (Fig. 6.2) (after Opdyke et al., 1977).









Five sites produced two samples each with concordant

directions and are classified as Class III sites. Of the

21 originally sampled sites, 14 were ultimately used to

interpret the magnetic polarity stratigraphy (Fig. 6.2).

Stepwise AF demagnetization was carried out over a

range of 0-100 mT (at 5 mT intervals) and followed by

thermal demagnetization over a range of 100-450o C (9

steps) on the Bone Gulch samples. The demagnetization

characteristics are dominated by a normal polarity

overprint which was resistant to AF demagnetization but

susceptible to thermal treatment. For four of five sites

a stable characteristic component was isolated over a

range of 100-4500 C. Sites 102, 101 and 100 produced

three sample each of reversed polarity and statistically

significant mean directions. These are categorized as

Class I reversed polarity sites (after Opdyke et al.,

1977) (Fig. 6.2). Site 103 produced two samples with

concordant directions and is categorized as a Class III

reversed polarity site. Site 104 failed to produce

demagnetization trends with concordant directions and was

rejected from the study.

Isothermal remanent magnetization (IRM) saturation

experiments were carried out on selected samples from 14

Fisherman's Cliff sites and three Bone Gulch sites.

Greater than 80% saturation was reached in all samples by

120 mT and in most cases, by 100 mT (Figs. 6.3c,d).









































S(OOWN)

INTENSITY




0-




S-


100-


0. 1.0 2.0
TREATMENT (A/m)


INTENSITY


. TREATMENT (A/m)


Figure 6.3. Vector demagnetization and isothermal
remanent acquisition diagrams for selected
samples from the Fisherman's Cliff (FC) and
Bone Gulch (BG)sections. (A) AF
demagnetization for FC sample 107.1; (B)
Thermal demagnetization for FC sample 113.2
(C) Isothermal remanence saturation plot
for FC sample 115.1. (D) Isothermal
remanence saturation plot BG sample 103.1.


N(UP)


C








104


54


1


0









These experiments, together with thermal demagnetization

results, suggest that magnetite is the major carrier of

the NRM in all samples from Fisherman's Cliff and Bone

Gulch. In several samples from both localities, IRM

acquisition continued slowly after reaching the

saturation plateau. This is interpreted to indicate the

presence of a second magnetic mineral which may also be

carrying part of the NRM. This mineral is characterized

by a high coercivity and a low blocking temperature (Fig.

6.3b) and is here interpreted to be goethite.

Magnetic Polarity Stratigraphy and Correlation to the
Timescale


The magnetic polarity stratigraphy of Fisherman's

Cliff is characterized by a short reversed zone (sites

100 and 101) that occurs at the base of the cliff in the

Moorna Formation. This is overlain by a zone of normal

polarity that extends up through the remainder of the

section. The horizon that produced the Fisherman's Cliff

Local Fauna is located in the Moorna Formation between

the base of the section and the Karoonda Surface and

therefore, falls within the zone of normal polarity (Fig.

6.2). The magnetostratigraphy of the Bone Gulch section

is characterized by the presence of a single zone of

reversed polarity that occurs within the Blanchetown

Clay.

As previously noted, the Karoonda Surface is present

at a Murray River cliff section at Chowilla (An et al.,









1986). An et al. (1986) indicate that the Karoonda

Surface and part of the Blanchetown Clay, in the Chowilla

section, occur within a normal polarity zone of the Gauss

Chron (2.47-2.92 Ma, after Berggren et al., 1985) with

the upper part of the Blanchetown Clay occurring in a

reversed polarity zone identified as the Matuyama Chron

(Fig. 6.4). The common occurrence of the Karoonda

Surface at both the Fisherman's Cliff and Chowilla

sections allows stratigraphic and magnetostratigraphic

correlations which may be used to establish the age of

both the Fisherman's Cliff and Bone Gulch Local Faunas.

The Karoonda Surface at Fisherman's Cliff, as marked by

the disconformity between the Moorna Formation and the

Chowilla Sand is also found in a normal polarity zone

which is here interpreted to be the Gauss Chron (2.47-

2.92 Ma) (Fig. 6.4). This polarity zone extends down

through the fossiliferous horizon and therefore restricts

its age to that of the Gauss Chron. The presence of a

short reversed polarity zone at the base of the section

is interpreted as being the Kaena Event (2.92-2.99 Ma).

The exposure of the Blanchetown Clay at Bone Gulch

is stratigraphically higher than that seen at Fisherman's

Cliff but is still considered to lie towards the base of

the formation (Woodburne et al., 1985; Crabb, 1977).

However, since it is younger than the Fisherman's Cliff

exposure and carries a reversed magnetic polarity, the

age of the Bone Gulch exposure is therefore constrained