Palaeozoic within-plate volcanic rocks in Nova Scotia (Canada) reinterpreted: isotopic constraints on magmatic source and palaeocontinental reconstructions

1. Introduction

Major and trace element concentrations and ratios of mod-ern basaltic rocks (see, e.g. Pearce & Cann, 1973;Winchester & Floyd, 1977) have been used extensively todistinguish between plate margin and within-plate basalts.Tectonic models for the c. 630–360 Ma evolution of NovaScotia (Canada) and their genetic context within theAppalachian orogen are largely based on geochemistry ofthe volcanic rocks and tectono-stratigraphic analysis ofinterbedded sedimentary rocks. This has indicated twomain tectonic settings: rifted magmatic arc for mostNeoproterozoic basalts (Murphy et al. 1990; Murphy,Keppie & Hynes, 1991; Dostal, Keppie & Murphy, 1990;Keppie & Dostal, 1991) and within-plate, continental riftfor all Palaeozoic basalts (Keppie, Dostal & Zentilli,1979; Keppie & Dostal, 1980; Dostal, Keppie & Dupuy,1983; Murphy et al. 1985; Murphy, Keppie & Hynes,1991) except the early Silurian basalts in Cape BretonIsland which fall in a volcanic arc field (Barr & Jamieson,1991). The Palaeozoic rocks are especially important tocontinental reconstructions because they originated onpart of the Gondwanan margin that was subsequentlytransferred to Laurentia (see, e.g. Keppie et al. 1996).

Thus Palaeozoic volcanism took place at various stagesduring the detachment and transfer process. The inter-pretation of these Palaeozoic volcanic rocks are reassessedherein in light of new geochemical data and recent tectonic models.

The Palaeozoic volcanic rocks are either associatedwith a marine transgression (Cambrian–early Ordovicianand middle Ordovician–early Silurian units) or with localrifting within a mainly transpressional continental envi-ronment (Devonian units). The Palaeozoic volcanic suitesare invariably bimodal and the felsic rocks are generallyinferred to have been produced by anatexis of the crustdue to rising basaltic magma (e.g. Keppie & Dostal,1980).

Neodymium isotopes in mafic igneous rocks may pro-vide information on the nature of the mantle source(DePaolo, 1988, and references therein). On the otherhand, neodymium isotopic data in felsic igneous rockspotentially record the character of the directly underlyingcrustal source and, if the crystallization age of theigneous rocks is known, provide constraints on the ageand nature of the continental basement source at the timeof igneous activity (DePaolo, 1988). This technique hasconsiderable potential in characterizing the basement

Geol. Mag. 134 (4), 1997, pp. 425–447. Printed in the United Kingdom © 1997 Cambridge University Press 425

Palaeozoic within-plate volcanic rocks in Nova Scotia (Canada) reinterpreted: isotopic constraints on magmatic

source and palaeocontinental reconstructions

J. D. KEPPIE*, J. DOSTAL†, J. B. MURPHY‡ & B. L. COUSENS§

* Instituto de Geologia, Universidad Nacional Autonoma de Mexico, 04510 Mexico, D.F., Mexico† Department of Geology, Saint Mary’s University, Halifax, Nova Scotia, B3H 3C3, Canada

‡ Department of Geology, St. Francis Xavier University, Antigonish, Nova Scotia, B2G 2W5, Canada§ Department of Earth Sciences, Carleton University, Ottawa, Ontario, K1S 5B6, Canada

(Received 17 August 1996; accepted 6 February 1997)

Abstract – Palaeozoic volcanism in the Avalon Terrane of northern Nova Scotia occurred during three timeintervals: Cambrian–early Ordovician, late Ordovician–early Silurian and middle–late Devonian. In theMeguma Terrane of southern Nova Scotia, Palaeozoic volcanism is limited to the middle Ordovician.Geochemical data show that most of these volcanic rocks are bimodal, within-plate suites. Initial εNd signa-tures range from +5.4 to –1.9 in the rhyolites and +6.8 to +2.7 in the basalts, a difference attributable to theabsence or presence, respectively, of a significant crustal component. The data and regional tectonic settingsof the Avalon and Meguma terranes suggest that the volcanism was generated in three different within-platesettings: (1) Cambrian–early Ordovician volcanism related to thermal decay of late Proterozoic arc magma-tism during transtensional deformation; (2) middle Ordovician–early Silurian volcanism during sinistral tele-scoping between Laurentia and Gondwana where extensional bends in the Appalachians produced rifting;and (3) Devonian volcanism resulting from lithospheric delamination during dextral transpression and tele-scoping. In each setting, active faults served as conduits for the magmas. Nd isotopic data indicate that thesource of the Palaeozoic felsic volcanic rocks is isotopically indistinguishable beneath southern and northernNova Scotia and did not substantially change with time. This crustal source appears to have separated fromthe mantle during the Proterozoic, a conclusion consistent with the hypothesis that the Palaeozoic rocks inNova Scotia were deposited upon a late Proterozoic oceanic–cratonic volcanic arc terrane. The Nd data, whencombined with published faunal, palaeomagnetic and U–Pb isotopic data, suggest that the Avalon Terranewas peripheral to Gondwana off northwestern South America during Neoproterozoic and early Palaeozoictimes.

beneath different tectonostratigraphic terranes, and hasrecently been applied in the northern Appalachians whereeach terrane has a distinct neodymium signature (Barr &Hegner, 1992; Fryer et al. 1992; Whalen et al. 1994a,b;Kerr, Jenner & Fryer, 1995). However, Nd isotopic signa-tures may be influenced by mixing between coeval maficand felsic magmas, and published data on the coevalmafic magmatism is sparse. In this paper we also presentNd isotopic data for both mafic and felsic Palaeozoiclavas, and assess their role in providing constraints onmantle and crustal evolution in Nova Scotia.

2. Geological setting

Volcanic rocks form an important part of the Palaeozoicgeological record in Nova Scotia. Cambrian–earlyOrdovician, late Ordovician–early Silurian and Devonianvolcanic rocks are restricted to northern Nova Scotia,whereas middle Ordovician volcanic rocks are only pre-sent in southern Nova Scotia (Fig. 1a). Nova Scotia lieswithin the southeastern margin of the northernAppalachians, and has traditionally been subdivided into

two Palaeozoic tectonostratigraphic terranes: theMeguma and Avalon terranes. The Avalon Terrane iscomposed of several Proterozoic terranes that were amalgamated during the latest Proterozoic and were overstepped by a latest Proterozoic–early Palaeozoicplatformal sequence containing an unique Avalonian andRhenish-Bohemian fauna (Boucot, 1975; Theokritoff,1979; Keppie, 1982, 1985, 1989; Landing, 1991).Outliers of this overstep sequence occur in most majorfault blocks in northern Nova Scotia, where they overlie avariety of late Proterozoic arc-related volcanic and tur-biditic metasedimentary rocks and gneisses. A notableexception is the middle Proterozoic Blair River Complexin northwestern Cape Breton Island, where the oldestPalaeozoic strata are late Devonian.

In marked contrast to the Palaeozoic rocks of theAvalon Terrane, the Meguma Terrane consists of 6–20+km of Cambrian–early Ordovician continental rise tur-bidites derived from the present southwest (Schenk,1970). These turbidites, which constitute the MegumaGroup are overlain by 2.3–4.5 km of middle Ordovicianto early Devonian subaerial to shallow marine, bimodalwithin-plate volcanic rocks, and sediments containing aRhenish-Bohemian fauna (Boucot, 1960; Keppie &Dostal, 1980). The Liscomb Complex, exposed in agneiss dome, was intruded into the Meguma Group during the late Devonian (c. 372 Ma) and has beeninferred to represent remobilized basement of theMeguma Terrane (Kontak et al. 1990; Clarke, Chatterjee& Giles, 1993). Granulite facies metasedimentary andmeta-igneous xenoliths in late Devonian lamprophyredykes in the southern Meguma Terrane may also repre-sent samples of the basement (Eberz et al. 1991).Intrusion of the Liscomb Complex and lamprophyredykes was approximately synchronous with intrusion oflarge volumes of granitic magma (e.g. South MountainBatholith). The Avalon and Meguma terranes share a similar Devono-Carboniferous stratigraphy.

The tectonic evolution of Cape Breton Island is contro-versial. Barr & Raeside (1989) divided the island intofour Precambrian–Palaeozoic terranes (Mira, Bras d'Or,Aspy and Blair River, Fig. 1a) that they did not believehad been amalgamated until Devono-Carboniferoustimes. They correlated these terranes with the AvalonTerrane (Mira), with the central mobile belt of theAppalachians (Bras d'Or and Aspy) and with theautochthonous Grenvillian cratonic margin of ancientNorth America (Blair River). More recently, Lin (1993)and Lynch, Tremblay & Rose (1993) have shown that theOrdovician–Silurian Aspy rocks lie unconformably uponlate Proterozoic Bras d'Or rocks, implying that they werepart of a single terrane by Ordovician time. Helmstaedt &Tella (1973) reported pebbles of Bras d'Or lithologies inthe Cambro-Ordovician overstep sequence, an observa-tion which suggests that the Mira and Bras d'Or terraneswere either adjacent (White et al. 1994), or amalgamated(Keppie & Dallmeyer, 1989) during Cambrian time. Barr& Hegner (1992) used neodymium isotopic data from

426 J. D. K E P P I E A N D OT H E R S

Figure 1. (a) Map of the northern Appalachians showing tec-tonic subdivision (modified from Keppie & Dallmeyer, 1989)and terranes in Cape Breton Island according to Barr & Raeside(1989; BR = Blair River; A = Aspy; B = Bras d'Or; M = Mira);(b) latest Proterozoic and Cambrian–early Ordovician igneousrocks in Nova Scotia; (c) middle Ordovician and Silurian volcanic rocks in Nova Scotia; and (d) Devono-Carboniferousvolcanic rocks in Nova Scotia.

Proterozoic and Palaeozoic felsic rocks to suggest thatdifferent magma sources lay beneath the various terranesin Cape Breton Island throughout Proterozoic and earlyPalaeozoic times. However, comparison with neodymiumdata published for other regions of the northernAppalachians suggests that further analysis and reinter-pretation may be necessary (cf. Whalen et al. 1994a,b;Kerr, Jenner & Fryer, 1995).

The above data, along with new data from volcanicrocks presented in this paper, may allow discriminationbetween the various terrane models that have been pro-posed for Cape Breton Island. They may also be used totest the various models proposed for the tectonic relation-ship between the Meguma and Avalon terranes. Thesemodels are: (1) that the two terranes presently rest on dif-ferent Precambrian basements (Sable and Avalon, respec-tively: Keen et al. 1991), in which case they might havedistinct neodymium signatures; (2) that the MegumaTerrane was thrust upon the Avalon Terrane during theDevonian (Keppie & Dallmeyer, 1987), in which casetheir basements may have distinct pre-Devonianneodymium signatures that became similar after thrust-ing; (3) that the Lower Palaeozoic rocks of the MegumaTerrane were deposited upon the Avalon Terrane (Keppie& Dostal, 1991), in which case they should have the sameAvalonian basement; and (4) that the Lower Palaeozoicrocks of the Meguma Terrane were deposited onPalaeozoic oceanic lithosphere off northwest Africa(Schenk, 1970, 1981), in which case the igneous rocks ofthe Meguma Terrane should have northwest African Ndisotopic signatures which typically have higherinitial143Nd/144Nd ratios than the Avalon Terrane.

3. Palaeozoic volcanic suites in Nova Scotia

Palaeozoic volcanism in the Nova Scotian part of theAvalon Terrane occurred at several intervals: (1)Cambrian–early Ordovician bimodal volcanic rocks inthe Bourinot and McDonalds Brook/Iron Brook groups ofcentral Cape Breton Island and the Antigonish Highlands,respectively (Fig. 1b); (2) late Ordovician–early Silurianbimodal volcanic rocks at the base of the Arisaig Groupin the Antigonish Highlands and in the Money PointGroup and Sarach Brook Metamorphic Suite in the CapeBreton Highlands (Fig. 1c); and (3) middle–lateDevonian bimodal volcanic rocks in the Fisset BrookFormation of western Cape Breton Island and the McArasBrook Formation adjacent to the Antigonish Highlands.In the Meguma Terrane, volcanic rocks only occur in themiddle Ordovician part of the White Rock Formation(Fig. 1d).

3.a. Cambrian–early Ordovician volcanic suites (Fig. 1b)

Cambrian to early Ordovician rocks in the AntigonishHighlands occur in two groups (Iron Brook andMcDonalds Brook) that are lateral facies equivalents ofone another (Keppie & Murphy, 1988). The Iron Brook

Group consists of predominantly continental to shallowmarine sedimentary rocks that contain typical AvalonianAcado-Baltic fauna (Landing & Murphy, 1991). The pre-dominantly volcanic McDonalds Brook Group consistsof continental clastic rocks overlain by a bimodal vol-canic suite (Murphy et al. 1985). An early Cambrian agefor the volcanic rocks is indicated by palaeontologicaland stratigraphic data from interlayered fossiliferousstrata (Landing & Murphy, 1991).

In the Boisdale Hills of Cape Breton Island, middleCambrian–early Ordovician rocks of the Bourinot Groupunconformably to disconformably overlie rocks of thedeformed Precambrian George River Group, pebbles ofwhich occur in the basal Cambrian (Hutchinson, 1952;Weeks, 1954; Helmstaedt & Tella, 1973). Bimodal mafic and felsic flows and tuffs form a major component(200–600 m) of these rocks (Murphy et al. 1985).Although a middle Cambrian age for most of these vol-canic rocks is indicated by fossils in interbedded sedimentary rocks, a rhyolite sample from the northernend of the outcrop has yielded a 505 ± 3 Ma concordantU–Pb zircon age indicating that volcanism continued intoearliest Ordovician time (White et al. 1994).

3.b. Middle Ordovician volcanic suites (Fig. 1c)

Along the northwestern margin of the Meguma Terrane,the Meguma Group is overlain by 2300–4500 m of middle Ordovician–early Devonian metasedimentaryrocks and bimodal flows and tuffs (White Rock and Torbrook formations and equivalents) that lie discon-formably upon Tremadocian rocks of the MegumaGroup. They were deposited in subaerial, shoreface and shallow marine environments, with a shoreline lying to the present northwest (Lane, 1976). Generally, thesequartzites pass upwards through Silurian slates, siltstonesand quartzites, with minor interbedded volcanic rocks in some sections (Keppie & Dostal, 1980).Preliminary U–Pb zircon data from rhyolites in the WhiteRock Formation yield an age of c. 462 Ma (T. E. Krogh,pers. comm.).

3.c. Late Ordovician–early Silurian volcanic suites (Fig. 1c)

In the Cape Breton Highlands, the Sarach BrookMetamorphic Suite and Money Point Group consist ofmafic–intermediate-felsic pyroclastic rocks and flowsinterlayered with metasedimentary rocks. According toBarr & Jamieson (1991), these volcanic rocks erupted in a volcanic arc tectonic setting. U–Pb zircon analyses have yielded ages of 433+7–

–4 Ma and 427.5 ± 2 Ma(Dunning et al. 1990; Keppie, Dallmeyer & Krogh, 1992,respectively).

In the Antigonish Highlands, the late Ordovician toearly Silurian Arisaig Group consists of basal mafic andfelsic flows and tuffs (Murphy, Keppie & Hynes, 1991)overlain by a thick sequence of intertidal to shallowmarine siliciclastic sediments (Boucot et al. 1974). Field

Isotopic constraints on Palaeozoic volcanism, Canada 427

evidence (Boucot et al. 1974; Murphy, Keppie & Hynes,1991) indicates that extrusion of the volcanic rocks waspenecontemporaneous with the deposition of the fossil-iferous early Silurian (Llandoverian) Beechill CoveFormation.

3.d. Middle–late Devonian volcanic suites (Fig. 1d)

Devonian volcanic rocks occur in the McAras BrookFormation (Antigonish Highlands) and Fisset BrookFormation (Aspy and Blair River terranes of Barr &Raeside, 1989). They are mainly of lava flows (70–800m thick) intercalated with continental red conglomeratesand sandstones (Dostal, Keppie & Dupuy, 1983).

4. Geochemistry

4.a. Analytical methods

Additional trace elements and Nd isotopic ratios weredetermined in representative samples selected from thecollections previously investigated by Keppie & Dostal(1980); Keppie, Dostal & Zentilli (1979); Dostal,Keppie & Dupuy (1983); Murphy et al. (1985) andMurphy, Keppie & Hynes (1991), to which the reader isreferred for more information. These samples wereanalysed by inductively coupled plasma-mass spectro-metry for the rare-earth elements (REE), Hf and Th(Tables 1 and 2), and some were also re-analysed formajor and certain trace elements (Rb, Sr, Ba, Zr, Nb, Y,Cr, Ni) by X-ray fluorescence. Precision and accuracy are discussed in Dostal, Baragar & Dupuy (1986) andDostal, Dupuy & Caby (1994). In general, they are better than 5 % for major elements and 2–10 % for traceelements.

Neodymium isotopic ratios (143Nd/144Nd) and the con-centrations of Sm and Nd were determined by isotopedilution and thermal ionization mass spectrometry atCarleton and Memorial Universities (Table 3). The143Nd/144Nd ratios were normalized to 146Nd/144Nd =0.7219. At Carleton, the La Jolla Nd standard analysed aspart of every run yielded average 143Nd/144Nd value of 0.511864 ± 0.000020 which fall within the error of theScripps best estimate of 0.511858 ± 0.000006 (J. D.MacDougall, pers. comm., 1991). At Memorial, the LaJolla Nd standard analysed as part of every run yielded anaverage 143Nd/144Nd value of 0.511862 ± 0.000016. Initialratios were calculated using available radiometric ages.Epsilon values (εNdt) were calculated based on the aboveages and a modern 143Nd/144NdCHUR = 0.512638,147Sm/144NdCHUR = 0.1967 and λSm = 6.54 ×10–12/yr.Assuming a 1 % error in the 147Sm/144Nd ratio and a pre-cision of ± 0.00002 in the measured 143Nd/144Nd ratios(based on duplicate analyses), these epsilon values haveerror limits of ± 0.6 epsilon units. Depleted mantle modelages (TDM) were calculated assuming values of143Nd/144Nd = 0.513114 and 147Sm/144Nd = 0.213, formodern depleted mantle.

4.b. Results

All the Palaeozoic volcanic rocks of Nova Scotia havebeen affected to varying degrees by secondary processes,including regional metamorphism and alterations whichmodified the chemical composition of these rocks. Forexample, secondary mobility of alkali metals is indicatedby the wide range of Na, K and Rb contents in the basalts(Tables 1 and 2). However, the concentrations of mostmajor elements, high-field-strength elements (HFSE),rare earth elements, Th and transition elements, wereprobably not significantly remobilized and are thought toreflect the primary magmatic distribution (Murphy et al.1985; Murphy, Keppie & Hynes, 1991; Dostal, Keppie &Dupuy, 1983).

Most of the analysed suites are distinctly bimodal.Mafic rocks have SiO2 contents of < 55 % (on LOI-freebasis). According to geological and geochemical criteria,including various discrimination diagrams such as thoseof Pearce & Cann (1973), the mafic rocks (Keppie &Dostal, 1980; Murphy et al. 1985; Murphy, Keppie &Hynes, 1991; Dostal, Keppie & Dupuy, 1983) can be sub-divided into two major groups: (1) within-plate and (2)arc-related volcanic rocks.

4.b.1. Within-plate rocks4.b.1.a. Basalts and mantle sources

The intra-plate group includes the basaltic rocks ofCambrian and Devonian age and some of the rocks of Ordovician and Silurian age (White Rock and Arisaigformations). According to classification schemes(Winchester & Floyd, 1977) based upon immobile ele-ments (e.g. Zr/TiO2 vs. SiO2 and Nb/Y vs. Zr/TiO2), themafic rocks are mainly subalkalic basalts, although somerocks appear to be transitional to alkalic basalts(McDonalds Brook Group and White Rock Formation)(Keppie & Dostal, 1980).

Within each suite, basalts of a comparable Nb/Y ratio(Table 2) have similarly-shaped chondrite-normalizedREE patterns with abundances increasing with the degreeof differentiation. Light REE (LREE) increase fromtholeiitic to alkalic lavas (as measured by Nb/Y ratio)whereas the heavy REE decrease or remain about thesame (Figs 2,4,5). Such trends are present even within asingle suite (Figs 4,5; e.g. White Rock and McAras Brookformations). The McDonalds Brook and White Rockalkalic basalts also have higher contents of other incom-patible trace elements. In contrast to the incompatible ele-ments, the abundances of compatible elements (Cr, Ni,Sc, V) in the various groups overlap and most of theirvariations are likely to be related to fractional crystalliza-tion (see, e.g. Dostal, Keppie & Dupuy, 1983). Evenprimitive basalts have MgO, Cr and Ni abundances thatare too low to represent primary melts of average uppermantle (Maaloe & Aoki, 1977; Hart & Davis, 1978).

Cambrian. Basalts of this age include those of theBourinot and McDonalds Brook groups. In the Bourinot



Tabl

e 1.

Maj

or a

nd tr

ace

elem

ent c

ompo

sitio

n of

repr

esen

tativ

e Pa

laeo

zoic

Rhy

olite

s fr

om N

ova

Scot

ia

Cam

bria

nO

rdov

icia

n–Si

luri

anD

evon

ian

Bou

rino

tM

cDon

alds

Bro

okSa

rach

Bro

okM

oney

Poi

ntW

hite

Roc

kM

cAra

s B

rook

Fiss

er B

rook

Low

land

Cov

e

Sam

ple

281

283

516

425

307

SB-1

1221

736

738

675

677

854

856

781

783

779

780

SiO

2(%

)70

.29

70.4

469

.10

61.6

072

.80

75.1

374

.37

73.8

075

.40

74.2

875

.19

73.8

274

.51

75.2

070

.79

74.9

073

.50

TiO

20.

380.

420.

580.

550.

330.

190.

250.

230.

220.

490.

170.

230.

180.

270.

300.

280.

35A

l 2O3

13.8

814

.08

15.4

017

.50

13.9

012

.73

13.2

113

.50

13.8

012

.56

14.0

011

.81

11.1

812

.69

14.2

411

.70

12.4

0Fe

2O3*

2.21

3.32

3.04

8.90

2.78

3.77

1.80

0.64

0.78

4.06

1.91

2.25

2.94

1.58

2.07

3.18

3.67

MnO

0.04

0.02

0.27

0.35

0.34

0.01

0.08

0.02

0.02

0.04

0.04

0.01

0.01

0.04

0.03

0.05

0.08

MgO

0.32

0.55

0.56

0.77

0.83

0.15

0.61

0.19

0.19

0.94

1.16

0.73

0.76

0.41

0.50

0.11

0.46

CaO

1.73

1.02

0.38

0.37

1.62

0.15

1.40

1.07

1.03

0.79

0.76

0.36

0.36

0.24

0.14

0.63

0.59

Na 2O

4.74

3.66

2.43

3.87

1.59

6.71

3.86

4.83

5.90

3.72

0.86

0.95

0.57

1.91

1.57

2.03

3.21

K2O

4.05

3.83

5.40

3.82

2.25

0.29

2.95

2.94

2.21

1.91

4.12

7.61

7.21

7.94

10.0

06.

604.

90P 2O

50.

070.

080.

130.

170.

070.

020.

070.

030.

030.

420.

540.

050.

050.

030.

040.

020.

04L

OI

1.77

1.97

1.71

2.58

3.52

0.35

0.87

1.38

1.20

1.22

1.80

2.60

2.64

0.80

0.79

0.42

0.80

Tota

l99

.48

99.3

999

.00

100.

4810

0.03

99.5

099

.47

98.6

310

0.78

100.

4310

0.55

100.

4210

0.34

101.

1110

0.47

99.5

010

0.00

Rb

(ppm

)88

154

204

106

5611

79-

-11

515

519

017

523

229

314

312

4B

a28

136

910

5357

837

225

540

442

300

149

168

228

234

1172

1251

218

203

Sr84

102

131

112

108

3220

3-

-17

774

33-

115

103

7054

Ta0.

910.

94-

--

1.44

0.82

0.72

0.71

1.12

1.38

2.66

2.40

1.16

1.49

2.09

2.40

Nb

15.5

18.4

11.0

61.0

10.0

29.3

11.0

8.8

7.2

18.9

18.8

64.0

52.9

13.0

17.7

47.3

54.4

Hf

5.39

7.05

--

-4.

942.

493.

503.

932.

212.

7714

.00

11.3

83.

594.

0611

.51

12.2

4Z

r27

538

323

357

820

223

311

017

219

897

9776

260

813

917

864

068

4Y

3152

2735

2453

1514

1427

3390

9818

2178

93T

h11

.79

8.80

--

-16

.00

16.6

418

.30

18.7

05.

357.

9523

.21

12.0

517

.56

21.0

118

.12

19.0

8L

a16

.65

29.3

923

.20

64.5

27.8

061

.94

28.8

311

.52

10.4

516

.15

13.7

913

7.9

89.1

616

.57

13.4

213

5.8

120.

4C

e37

.40

65.3

653

.40

143.

063

.00

93.0

752

.36

22.4

019

.71

35.3

031

.78

264.

217

8.7

33.6

427

.00

285.

026

1.8

Pr4.

868.

72-

--

13.2

05.

312.

322.

054.

123.

7332

.34

22.7

54.

043.

3332

.90

29.5

7N

d19

.13

35.6

023

.50

56.5

028

.00

47.5

517

.84

8.12

7.33

15.8

113

.94

118.

080

.50

14.8

412

.33

125.

711

3.6

Sm4.

328.

114.

728.

775.

178.

553.

131.

601.

683.

793.

8420

.67

15.5

83.

152.

8721

.63

21.0

6E

u0.

821.

661.

322.

311.

280.

520.

710.

320.

340.

680.

430.

800.

640.

540.

581.

091.

47G

d4.

328.

53-

--

8.09

2.47

1.77

1.69

3.98

4.49

16.2

014

.22

2.85

2.98

17.7

419

.14

Tb

0.72

1.38

0.91

1.32

0.85

1.37

0.38

0.30

0.32

0.73

0.88

2.63

2.60

0.45

0.52

2.60

2.89

Dy

5.38

9.48

--

-9.

792.

672.

302.

435.

306.

2718

.32

19.3

53.

233.

6416

.94

19.6

5H

o1.

101.

85-

--

2.00

0.52

0.49

0.57

0.95

1.19

3.73

3.90

0.70

0.78

3.39

3.87

Er

3.33

5.57

--

-6.

061.

711.

761.

832.

733.

3411

.46

11.3

92.

062.

5710

.28

11.2

5T

m0.

510.

81-

--

0.96

0.29

0.28

0.32

0.37

0.48

1.71

1.67

0.35

0.41

1.46

1.65

Yb

3.41

5.47

3.86

4.59

3.48

6.56

1.82

2.08

2.30

2.37

2.83

11.7

110

.67

2.19

2.78

9.47

10.6

5L

u0.

530.

880.

660.

730.

591.

020.

310.

320.

390.

350.

411.

711.

580.

340.

431.

571.

63N

b/Y

0.50

0.35

0.41

1.7

0.42

0.55

0.73

0.63

0.51

0.70

0.57

0.71

0.54

0.72

0.84

0.61

0.58

(La/

Yb)

n2.

963.

263.

648.

524.

845.

729.

603.

362.

754.

132.

957.

145.

064.

592.

938.

696.

85

Maj

or a

nd tr

ace

elem

ent c

ompo

sitio

ns o

f ana

lyse

d sa

mpl

es o

f the

ear

ly S

iluri

an A

risa

ig G

roup

are

repo

rted

in M

urph

y et

al.

(199

4).


Tabl

e 2.

Maj

or a

nd tr

ace

elem

ent c

ompo

sitio

n of

repr

esen

tativ

e Pa

laeo

zoic

bas

alts

from

Nov

a Sc

otia

Cam

bria

nO

rdov

icia

n–Si

luri

anD

evon

ian

Bou

rino

tM

cDon

alds

Bro

okSa

rach

Bro

okM

oney

Poi

ntA

risa

igW

hite

Roc

kM

cAra

s B

rook

Fiss

er B

rook

Low

land

Cov

e

Sam

ple

288

295

375

376

8011

1222

740

731

GO

2G

O2-

1268

068

582

983

678

678

577

277

3

SiO

2(%

)48

.60

49.1

044

.60

43.9

148

.28

47.9

353

.90

50.0

042

.95

40.0

648

.19

50.6

551

.25

45.3

844

.09

41.0

948

.70

48.2

0T

iO2

1.89

2.54

2.98

3.45

2.58

1.51

1.37

2.69

2.78

3.82

1.85

2.47

1.40

2.36

1.87

2.23

2.86

1.76

Al 2O

315

.40

15.3

015

.41

15.9

314

.21

17.1

615

.20

14.0

016

.09

15.5

615

.01

13.7

716

.62

15.8

716

.78

17.5

713

.60

15.7

0Fe

2O3*

9.94

14.3

211

.61

15.7

314

.06

11.3

69.

9311

.41

14.6

017

.20

12.7

511

.90

8.82

14.5

013

.75

11.9

115

.33

10.3

0M

nO0.

260.

280.

390.

530.

160.

250.

190.

660.

480.

450.

140.

200.

770.

440.

520.

590.

240.

21M

gO9.

035.

576.

076.

345.

046.

143.

874.

157.

507.

505.

638.

547.

064.

188.

195.

784.

805.

14C

aO4.

563.

445.

934.

418.

449.

037.

118.

948.

634.

6511

.02

5.87

4.84

8.12

2.18

6.05

7.56

7.68

Na 2O

5.20

4.97

4.56

4.16

3.84

1.87

4.26

3.39

2.73

3.27

2.53

3.58

5.33

3.33

1.85

4.23

2.92

3.21

K2O

0.25

0.27

0.19

0.39

1.20

1.05

1.60

1.00

0.32

0.10

0.39

0.23

0.29

1.31

4.58

1.25

1.76

1.40

P 2O5

0.30

0.38

1.02

1.00

0.42

0.24

0.40

1.29

0.47

1.39

0.25

0.35

0.21

0.80

0.26

0.32

0.79

0.39

LO

I4.

703.

926.

154.

021.

073.

351.

631.

513.

546.

201.

921.

593.

492.

834.

858.

652.

554.

81

Tota

l10

0.13

100.

0998

.91

99.8

799

.30

99.8

999

.46

99.0

410

0.09

100.

2099

.68

99.1

510

0.08

99.1

298

.92

99.6

710

1.06

98.8

4

Rb

(ppm

)28

248

1036

29-

-9

57

104

2216

082

-36

Ba

153

8621

629

792

226

519

224

454

133

504

561

256

568

1858

6459

338

4Sr

507

363

833

696

201

421

--

535

489

541

513

836

682

189

167

-31

8Ta

0.34

0.31

3.75

3.19

0.46

0.37

0.51

0.70

0.74

0.85

0.87

1.67

0.74

1.28

0.43

0.44

0.87

0.69

Nb

6.1

8.2

82.1

76.0

10.2

8.5

12.2

15.2

15.5

17.5

23.0

40.3

14.6

30.9

8.3

10.1

20.8

15.7

Hf

3.38

4.73

8.18

7.64

3.98

3.58

5.27

5.56

3.10

3.65

2.50

3.78

3.42

5.39

3.24

3.68

6.41

4.93

Zr

153

222

409

378

186

170

256

259

131

168

118

169

160

275

160

194

322

239

Y27

3927

2735

2834

6022

3718

1926

3028

2747

33T

h0.

750.

786.

525.

833.

383.

896.

453.

330.

981.

151.

822.

463.

473.

000.

740.

703.

383.

78L

a8.

8414

.26

58.5

744

.36

17.1

820

.17

34.2

129

.50

14.0

025

.07

14.9

719

.48

13.4

933

.70

9.71

14.8

430

.21

23.2

0C

e24

.20

36.1

912

2.85

98.3

440

.51

45.0

779

.01

74.4

033

.53

60.1

332

.78

42.3

630

.75

77.1

924

.91

35.8

971

.03

53.1

1Pr

3.50

5.15

14.3

11.7

95.

435.

689.

9510

.12

4.59

8.62

4.05

5.39

3.81

9.63

3.52

4.67

9.17

6.64

Nd

16.6

224

.71

55.4

246

.83

24.9

724

.58

41.7

247

.53

20.6

240

.13

16.8

521

.99

15.6

039

.01

16.8

921

.25

39.3

227

.38

Sm4.

456.

556.

798.

956.

385.

648.

4411

.85

5.22

10.2

04.

085.

293.

908.

154.

705.

319.

086.

41E

u1.

472.

142.

922.

551.

911.

982.

423.

722.

004.

811.

321.

491.

172.

691.

581.

752.

841.

84G

d4.

857.

198.

137.

766.

935.

687.

7613

.20

5.18

10.1

44.

455.

214.

317.

595.

475.

789.

636.

73T

b0.

761.

111.

051.

041.

060.

851.

071.

850.

761.

310.

640.

730.

691.

000.

800.

861.

421.

00D

y5.

367.

946.

366.

337.

155.

517.

0012

.41

4.92

8.02

3.96

4.35

5.07

6.34

5.49

5.49

9.63

6.83

Ho

1.08

1.57

1.05

1.09

1.35

1.04

1.28

2.36

0.86

1.46

0.71

0.72

0.99

1.15

1.10

1.06

1.88

1.30

Er

2.99

4.33

2.82

2.89

3.90

2.91

3.73

6.63

2.43

3.67

1.88

1.92

2.99

3.28

3.24

3.05

5.29

3.72

Tm

0.43

0.63

0.37

0.37

0.54

0.42

0.51

0.93

0.33

0.48

0.24

0.22

0.42

0.47

0.44

0.46

0.74

0.51

Yb

2.85

3.95

2.32

2.24

3.47

2.72

3.62

5.45

1.97

2.95

1.50

1.45

2.75

2.93

2.77

2.90

4.64

3.41

Lu

0.39

0.55

0.34

0.32

0.51

0.40

0.53

0.84

0.30

0.43

0.22

0.18

0.41

0.47

0.44

0.40

0.74

0.54

Nb/

Y0.

230.

213.

02.

80.

290.

300.

360.

250.

700.

471.

32.

10.

561.

00.

300.

370.

440.

48(L

a/Y

b)n

1.88

2.19

15.3

12.0

3.00

3.49

5.73

3.28

4.30

8.50

6.05

8.14

2.97

6.97

2.12

3.10

3.95

4.12

Group, the mafic rocks are tholeiitic basalts. The REEpatterns (Fig. 2a) are only moderately enriched in lightREE with (La/Yb)n ~ 2–2.5. The mantle-normalizedincompatible element patterns show a gradual increasefrom Lu to La and then a decrease towards Th (Fig. 2b).Rb and Ba abundances in these rocks have been affectedby secondary alteration, and therefore we consider the

high Rb/La and Ba/La to be an artifact of alteration ratherthan evidence for a subduction component. Initial εNd val-ues vary slightly between +3.1 and +3.9, which are sig-nificantly lower than the value for depleted mantle at thistime (c. +8–8.5) (Table 3; Fig. 3). The basalts of theMcDonalds Brook Group are transitional to alkalic with adistinct enrichment of LREE (Fig. 2c) and (La/Yb)n


Table 3. Neodymium isotope data

Sample Type Sm Nd 143Nd/144Ndm Age (Ma) εNdt TDM

McDonalds Brook Formation375 Bas 10.9 58.8 0.51255(1) 530 +4.0 851376 Bas 10.5 51.9 0.51257(2) 530 +3.7 911307 Rhy 3.4 21.9 0.51257(1) 530 +5.0 679516 Rhy 3.8 18.8 0.51249(1) 530 +2.0 1005

Bourinot Group285 Bas 6.2 22.9 0.51276(2) 515 +4.6 932288 Bas 5.9 21.5 0.51273(1) 515 +3.9 1032293 Bas 7.8 29.0 0.51274(1) 515 +4.3 950295 Bas 10.1 36.7 0.51269(2) 515 +3.1 1151281 Rhy 3.8 15.9 0.51237(1) 515 -1.9 1466283 Rhy 8.4 37.2 0.51252(1) 515 +1.6 1070

White Rock Formation680 Bas 4.5 18.9 0.51271(2) 460 +4.5 797681 Bas 4.5 18.6 0.51264(1) 460 +3.0 953685 Bas 6.7 28.0 0.51269(1) 460 +4.1 835686 Bas 4.0 16.6 0.51284(2) 460 +6.8 569694 Bas 9.3 41.2 0.51272(1) 460 +5.0 716675 Rhy 4.5 18.0 0.51259(2) 460 +1.6 1150676 Rhy 4.4 16.7 0.51258(2) 460 +1.1 1320677 Rhy 4.0 13.4 0.51260(2) 460 +0.2 1837

Arisaig GroupG2 Bas 6.1 24.6 0.51273(2) 430 +4.3 843G2-12 Bas 11.5 44.7 0.51278(1) 430 +5.1 767425-4a Rhy 4.5 24.8 0.51243(1) 430 +0.8 98584-12 Rhy 2.6 13.6 0.51245(1) 430 +0.7 1016

Sarach Brook Metamorphic Suite8011 Bas 7.2 28.0 0.51279(2) 430 +5.3 7381222 Bas 8.5 36.8 0.51264(1) 430 +3.1 893SB-1 Rhy 7.5 42.3 0.51266(1) 430 +5.4 6071221 Rhy 3.3 18.8 0.51261(3) 430 +4.5 660

Money Point Group731 Bas 12.1 47.7 0.51281(2) 428 +5.8 677740 Bas 10.3 48.2 0.51270(2) 428 +4.8 690736 Rhy 1.7 7.9 0.51269(2) 428 +4.8 684738 Rhy 1.5 6.5 0.51269(2) 428 +4.0 821

McAras Brook Formation829 Bas 6.4 25.4 0.51271(1) 380 +3.5 894836 Bas 9.5 45.5 0.51265(1) 380 +3.6 746854 Rhy 22.0 123.9 0.51242(1) 380 0.0 856856 Rhy 16.8 88.5 0.51240(1) 380 -0.7 909

Fisset Brook (Cheticamp)785 Bas 5.8 23.3 0.51277(2) 375 +4.9 715786 Bas 4.8 19.4 0.51279(1) 375 +5.2 693781 Rhy 3.7 17.4 0.51248(1) 375 +0.2 1034783 Rhy 3.1 13.5 0.51246(2) 375 -0.7 1202

Fisset Brook (Lowland Cove)772 Bas 10.5 44.5 0.51264(2) 375 +2.7 900776 Bas 7.6 33.0 0.51268(2) 375 +3.7 794779 Rhy 22.7 127.3 0.51248(1) 375 +1.1 787780 Rhy 22.5 118.9 0.51245(1) 375 +0.3 883

Bas = basalt; Rhy = rhyolite. Sm and Nd concentrations determined by isotope dilution, expressed in weight parts per million. 143Nd/144Ndnormalized to 146Nd/144Nd = 0.72190. Precision of concentrations ± 1% and 2-sigma uncertainty in 143Nd/144Ndm (m = measured) indicated in brackets(value is the ± in fifth decimal place). The average obtained for the La Jolla standard over the analysis period is 0.511864 ± 0.000020 (2-σ). εNd

t

calculated assuming modern 143Nd/144NdCHUR = 0.512638 and 147Sm/144NdCHUR = 0.1967. Estimated uncertainty in εNdt = + 0.60 ε units. TDM depleted

mantle model ages calculated assuming modern depleted mantle with 143Nd/144Nd = 0.53114 nad 147Sm/144Nd = 0.213.

ranging between 12 and 16. The mantle-normalizedincompatible element plots peak at Nb (Fig. 2d). InitialεNd values in these alkaline basalts range from +3.7 to+4.0, and are thus indistinguishable isotopically from theBourinot tholeiites (Table 3; Fig. 3).

Middle Ordovician. The bimodal White RockFormation suite was sampled around Nictaux and CapeSt. Mary (Fig. 1c). The Nictaux basalts are tholeiitic totransitional whereas those of Cape St. Mary are transi-tional to alkalic (Keppie & Dostal, 1980). This isreflected in their REE patterns (Fig. 4a) which are dis-tinctly enriched in LREE in the Cape St. Mary rocks (that is, REE pattern shows moderate slope, Fig. 4a) with(La/Yb)n ratio around 20 as opposed to the Nictaux rockswhich have a shallower sloping REE pattern and a ratiobetween 4 and 9. Like the Cambrian alkalic basalts, theirmantle-normalized patterns also peak at Nb (Fig. 4b) andbasalts of the White Rock Formation have the highestNb/La of all the basaltic rocks in this study. Initial εNdranges from +3.0 to +6.8 (Table 3; Fig. 3). These are significantly lower than contemporary depleted mantlevalues (c. 8–8.5), indicating either that they have beenderived from different sources or that primary magmaswere variably contaminated by a crustal component. Thelack of obvious Th enrichment and the lack of Nb and Tidepletion in basalts with lower initial εNd, argues againstcrustal contamination.

Late Ordovician–early Silurian. Basalts of the ArisaigGroup are transitional alkalic-tholeiitic rocks with aNb/Y ratio ranging between 0.5 and 0.9 and high TiO2(> 2 %) even for rocks with high Mg#. The REE patterns(Fig. 4c) display an enrichment of LREE and fractiona-tion of HREE with a (La/Yb)n ratio of about 4–5.5. Thesebasalts have slight positive Eu and Ti anomalies. The


Figure 2. Rare earth and trace element compositions of isotopically analysed Cambrian basalts in Nova Scotia. Rare earth elementsnormalized using chondrite values of Sun (1982), and trace elements normalized using mantle values of Sun & McDonough (1989).

Figure 3. εNdt versus crystallization age for the Palaeozoic vol-canic samples in Nova Scotia (solid symbols = felsic; open sym-bols = mafic); data for McDonalds Brook and Arisaig samplesfrom Murphy et al. (1995). Dashed lines define the lower limitfor within-plate, felsic volcanic rocks in the Meguma andAvalon terranes in Nova Scotia constructed by drawing a linethrough most negative initial εNd values and their respective TDMmodel ages. Depleted mantle evolution line assumes εNd0 = +9.3and 147Sm/144Nd = 0.213 (average of DePaolo, 1988).

basaltic rocks have low Th/La ratio (< 0.1) suggestingthat they were not significantly affected by crustal conta-mination. The mantle normalized patterns of the basaltspeak at La or Nb (Fig. 4d). Overall, these basalts are verysimilar to the Bourinot tholeiites. Initial εNd ranges from+4.3 to +5.1, which are slightly higher than those ofBourinot Group basalts but are significantly lower thancontemporary depleted mantle values (Table 3; Fig. 3).

Devonian. The basaltic rocks of McAras BrookFormation are from two nearby sites (McAras Brook andBallantynes Cove, Fig. 1d) that differ in composition. Theformer are tholeiitic basalts with (La/Yb)n rangingbetween 2 and 7, whereas the latter are alkalic basaltswith high Nb/Y and (La/Yb)n ~18–20 (Fig. 5a). However,initial εNd is very similar at +3.5–3.6 (Table 3; Fig. 3), andthus both tholeiitic and alkaline lavas were derived fromisotopically similar sources. The mantle-normalizedincompatible element patterns (Fig. 5b) for both sites aresimilar to those of the other Palaeozoic suites which peakat Nb.

The basaltic rocks of the Fisset Brook Formation werealso sampled at two sections, Lowland Cove and FissetBrook (Fig. 1d). The basaltic rocks are tholeiitic and theirREE patterns are moderately fractionated and character-ized by a LREE enrichment and slight fractionation ofHREE, with (La/Yb)n ~ 4 in the Lowland Cove samples(Fig. 5c) and 2–3 in the Fisset Brook basalts (Fig. 5e).The mantle-normalized patterns of the basalts of FissetBrook peak at La with a gradual decrease towards both

Lu and Th (Fig. 5d). Initial εNd ranges only slightly from+5.2 to +4.9 (Table 3; Fig. 3). However, the basalts ofLowland Cove exhibit a gradual increase from Lu to Thwith negative Nb and Ti anomalies (Fig. 5f). These char-acteristics, along with the lower initial εNd values in thesebasalts (+2.7 to +3.7) (Table 3; Fig. 3), are consistent withan involvement of continental lithosphere in their evolu-tion (Dostal & Dupuy, 1984), most probably by crustalcontamination.

Petrogenesis. In general, compositional differenceswithin and between the basalts may reflect either differ-ences in the source composition, or the percentages ofmelting and fractional crystallization processes, or thedegrees of contamination with coeval felsic lavas or con-tinental crust. As there is no geochemical evidence forextensive crustal contamination in the sample suites(other than the Lowland Cove suite), this process will notbe considered further. Thus, within individual suites, thevariations of major and trace elements which correlatewith Fe/Mg are probably due to fractional crystallization.Differences in the slope of the REE patterns and in ele-mental ratios such as Zr/Y and Ti/Zr among the rocks ofthe same age (such as those observed in the White RockFormation) can be explained either by different degreesof melting of a garnet-bearing source or by derivationfrom compositionally distinct sources. These alternativesmay be distinguished by radiogenic isotope ratios.

No strong correlations between incompatible elementratios and Nd isotope ratios are observed either within or


Figure 4. Rare earth and trace element compositions of isotopically analysed middle Ordovician–Silurian basalts in Nova Scotia.Normalizing values as in Figure 2.

between the basalt groups, so it is difficult to constraintheir source compositions. However, their mantle-normalized incompatible element abundance patternspeak at Nb and smoothly decrease from Nb to more com-patible elements such as Y and heavy REE. This shape ischaracteristic of alkaline basaltic lavas from oceanicislands (OIB; see, e.g. Halliday et al. 1988), or of transi-tional to alkaline basalts from the Basin and RangeProvince of the southwestern United States (Kempton,Dungan & Blanchard, 1987; Menzies, Kempton &Dungan, 1985). In contrast, many tholeiitic continentalflood basalts with a significant subcontinental litho-spheric or crustal component, such as those of ColumbiaRiver and Karoo, exhibit a negative Nb anomaly (Hooper& Hawkesworth, 1993; Ellam & Cox, 1991; Ellam,Carlson & Shirey, 1992).

The mantle sources of the Cambrian through Silurianbasalts varied considerably in initial εNd, from +3 to +7.

Without evidence for a large thermal anomaly accompa-nying volcanism in Nova Scotia during this period, amantle plume origin for these basalts is unlikely.Alternatively, these basalts could be melts of a heteroge-neous upper asthenosphere, composed largely of adepleted mantle matrix (εNd c. +8) but containing incom-patible-element-enriched enclaves or veins (variable ini-tial εNd> +3)(see, e.g. Cousens et al. 1995). Mixing ofmelts from both enclaves and matrix would produce aspectrum of incompatible element patterns and initial εNdvalues. In this model, the more alkaline incompatible-ele-ment-rich basalts should have the lowest initial εNd, whichis indeed the case for the White Rock Formation.However, basalts of the McAras Brook Formation differstrongly in alkalinity but show little variation in initialεNd. This is also true of the Cambrian basalts of theBourinot and McDonalds Brook groups. A potentialexplanation of this non-correlation may be that variable


Figure 5. Rare earth and trace element compositions of isotopically analysed Devonian basalts in Nova Scotia. Normalizing values asin Figure 2.

pressure and extent of melting have a much greaterimpact on the composition of continental rift basalts than they do on oceanic basalts (for which the heteroge-neous mantle model works well). For example, smalldegree melts of depleted mantle in the garnet stabilityfield would produce alkalic lavas with high initial εNd,whereas large degree melts of the enriched veins orenclaves would be transitional in composition with lowinitial εNd. The lack of a subduction signature in theCambrian basaltic geochemistry suggests that theasthenospheric mantle wedge contaminated with fluidsfrom the latest Neoproterozoic subducting slab wasrapidly carried away. It also indicates that a subcontinen-tal lithospheric mantle heavily fluxed by subduction com-ponents is an unlikely source of the Cambrian basalticlavas.

4.b.1.b. Rhyolites and crustal basement compositions

The felsic lavas in the Cambrian through Silurian suitesmay have originated by fractional crystallization fromcoeval basaltic magmas, or by crustal anatexis. Nd isotopes provide an indication as to their origin.

Cambrian. Whereas basalts of the Bourinot andMcDonalds Brook groups vary significantly in compo-sition, their rhyolitic members are compositionally verysimilar. All Cambrian rhyolites have (La/Sm)n = c. 3, flatmiddle-to-heavy REE, and small negative Eu anomalies(Fig. 6a, c). They are also highly depleted in Sr, Ti, andNb, but are enriched in Th (Fig. 6b, d). Isotopically, initial

εNd values for the McDonalds Brook rhyolites range from+5.0 to +2.0, whereas the Bourinot rhyolites vary from+1.6 to –1.9 (Table 3; Fig. 3).

Middle Ordovician. Rhyolites of the White RockFormation are less LREE-enriched and have higher Gd/Luthan the Cambrian rhyolites (Fig. 7a). Both groups of rhyolites have negative Nb, Sr, Eu, and Ti anomalies, butare highly enriched in Th (Fig. 7b). Note, however, thatsample 675, with initial εNd = +1.6, has smaller Sr, Eu andTi anomalies than does sample 677 with an initial εNd of+0.2. It would therefore appear that increased fractiona-tion, probably accompanied by crustal contamination islinked to lower initial εNd values.

Late Ordovician–early Silurian. Rhyolites of theArisaig Group more closely resemble those of theBourinot Group than those of the White Rock Formation,having similar LREE enrichment and flat middle–heavyREE patterns (Fig. 7c). Sample 84-12, however, has apositive Eu anomaly and lacks a large negative Sr anom-aly (Fig. 7d), distinguishing it from all other rhyolitesfrom the Avalon Terrane. Both Arisaig samples exhibitnegative Nb anomalies and initial εNd values of +0.7 and+0.8, which fall in the middle of the range shown by thoseof the Bourinot Group (Table 3; Fig. 3).

Devonian. Rhyolites of the McAras Brook Formation(Fig. 8a) and the Fisset Brook Formation at LowlandCove (Fig. 8e) are highly LREE-enriched, with promi-nent negative Eu anomalies and (Gd/Lu)n> 1. They alsohave relatively large depletions in Nb, Sr and Ti (Fig. 8b, f),depletion of the latter two elements being consistent with


Figure 6. Rare earth and trace element compositions of isotopically analysed Cambrian rhyolites in Nova Scotia. Normalizing values asin Figure 2.

a highly fractionated felsic magma (explaining theirabnormally high incompatible element contents).Rhyolites of the Fisset Brook Formation at Fisset Brookare distinctly less LREE-enriched and have similar Euanomalies (Fig. 8c). They also show the depletions in Nb,Sr and Ti typical of all the Nova Scotia rhyolites, but areextraordinarily enriched in Th and Ba (Fig. 8d).Regardless of the significant trace element differencesbetween these Devonian felsic rocks, initial εNd valuesvary over the small range of +1.1 to –0.7 (Table 3; Fig. 3),suggesting that their sources were isotopically similar.

Petrogenesis. Compared to the basalts, the REE pat-terns of the rhyolites (typically > 66 % SiO2) have flatterHREE and most also have a distinct Eu anomaly. Mostrhyolites also have a more pronounced enrichment oflight REE. In comparison with associated basalts, themantle-normalized incompatible trace element patternsof the rhyolites are characterized by negative Nb and Tianomalies. The rhyolites also have a more distinct enrich-ment of strongly incompatible trace elements such as Rb,Ba and Th. With the exception of the CambrianMcDonalds Brook Group and the Silurian Arisaig Group,initial εNd values of the rhyolites from the bimodal,within-plate volcanic suites, are generally lower thanthose of the associated basalts: –1.86 to +2.98 versus+2.72 to +6.79, respectively. Initial εNd values of rhyolitesand basalts from the Cambrian McDonalds Brook Groupshow complete overlap, whereas those of the SilurianArisaig Group show partial overlap.

The lack of intermediate rocks, the high proportion offelsic to mafic compositions, the differences in the distribu-tion of incompatible trace elements, and the distinction ininitial εNd values are all inconsistent with derivation of somefelsic rocks by closed system fractionation crystallization.However, they can be accounted for by crustal anatexis aspreviously suggested for most of these suites (Keppie,Dostal & Zentilli, 1979; Keppie & Dostal, 1980; Dostal,Keppie & Dupuy, 1983; Murphy et al. 1985; Murphy,Keppie & Hynes, 1991). The range of initial εNd values inboth the mafic and felsic rocks of a given suite is bestexplained in terms of either various degrees of mixingbetween magma derived from a depleted mantle source andanatexis of a crustal source, or through melting of inhomo-geneous mantle and crustal sources. These crustal sourcesmay have had εNd values as low as –1.9, which is the lowestinitial εNd value of any of the rhyolites (assuming it includesno significant mantle component). On the other hand, over-lapping initial εNd values in basalts and rhyolites of theCambrian McDonalds Brook and Silurian Arisaig groupswould allow for some of the felsic rocks to have been frac-tionated from the basalts. Conversely, the data may merelysuggest that the source regions had similar εNd values.

4.b.2. Possible arc-related rocks

The late Ordovician–early Silurian Money Point Formationand Sarach Brook Metamorphic Suite in the Cape BretonHighlands have been classified on the basis of immobile


Figure 7. Rare earth and trace element compositions of isotopically analysed middle Ordovician–Silurian rhyolites in Nova Scotia.Normalizing values as in Figure 2.

trace elements as arc-related suites by Barr & Jamieson(1991). However, these two suites are generally stronglyaltered and display variable degrees of metamorphism sothat deciphering their original composition is difficult. Ourdata indicate that the mafic rocks are subalkaline basalts andsubordinate basaltic andesites with < 55 % SiO2 that havetholeiitic affinities. REE patterns of both suites (Fig. 9) arecharacterized by LREE enrichment with the (La/Yb)n ratiosranging from 3 to 6 for the basalts of Money Point Group(Fig. 9a) and from 3 to 4.5 for the Sarach BrookMetamorphic Suite (Fig. 9b). The mantle-normalizedincompatible element patterns (Fig. 9c, d) are enriched instrongly incompatible elements and display distinct nega-tive Nb anomalies accompanied in most samples by deple-tion in Ti. Although these patterns are characteristic ofvolcanic arc rocks, all of these features (includingTiO2> 1.5 %) are also typical of continental tholeiites thateither interacted with or were derived from subcontinental

lithosphere. Examples of such tholeiites include theColumbia River basalts (e.g. USGS standard BCR-1),basalts of the western Great Basin (Menzies, Leeman &Hawkesworth, 1983; Fitton et al. 1988), or certainMesozoic basalts occurring along the northeastern marginof North America (such as USGS standard rock W-1).

The associated felsic rocks, which are predominantlyrhyolites, have a contrasting trace element distribution.Their REE profiles are enriched in LREE but show flat oreven positive slopes for HREE and negative Eu anomalies(Fig. 10a, b). The mantle-normalized profiles of the rhyo-lites show a distinct Th enrichment accompanied by Tiand minor Nb depletion (Fig. 10c, d). These geochemicaldifferences suggest that the felsic rocks are not derived by closed system fractional crystallization of the associ-ated mafic rocks, but probably include a crustal compo-nent in their genesis. Initial εNd values in both basalts andrhyolites of the two Ordovician–Silurian suites overlap.


Figure 8. Rare earth and trace element compositions of isotopically analysed Devonian rhyolites in Nova Scotia. Normalizing values asin Figure 2.

5. Geodynamic models for Palaeozoic magmageneration

5.a. Interpretation of Nova Scotia data

The foregoing analysis indicates that Palaeozoic basalticmagmatism in Nova Scotia is within-plate, with the pos-sible exception of the late Ordovician–early Silurian vol-canism in the Cape Breton Highlands which could be ofeither arc-related or intra-continental tholeiitic affinities.In view of the unresolved affinity of the latter volcanicsuite, they are omitted from further consideration.Although all of the remaining Palaeozoic volcanic suitesare within-plate, their postulated tectonic setting, basedon the complete geological record, reveals several within-plate settings.

5.a.1. Cambrian–early Ordovician intra-plate magmatism

The contrasting tholeiitic-alkalic Cambrian–earlyOrdovician magma types may be equated with differ-ences in the degree and depth of partial melting; relativelyhigh degrees of partial melting at lower pressures yield tholeiites, whereas lower degrees of partial meltingat higher pressures produce the alkalic basalts. This con-trast appears to be related to the age of the underlyingmagmatism; tholeiitic magmatism in central Cape BretonIsland occurs in areas of 580–550 Ma arc magmatism,whereas alkalic lavas in the Antigonish Highlands occurin areas where arc magmatism ceased at c. 600 Ma (Fig.1). Modelling of the thermal decay of a similar-sized

plutonic complex to that of central Cape Breton Island(e.g. Lesquer, Bourmatte & Dautria, 1988) suggests thatit would take up to 80 Ma for the thermal anomaly to dis-sipate (Fig. 11). Temperatures high enough to producebasaltic partial melt in this decaying thermal anomalywould be located near the base of the lithosphere and inthe immediately underlying asthenosphere.

Keppie (1982) and Keppie & Murphy (1988) proposedthat the Cambro-Ordovician rocks in the Avalon Terraneof Nova Scotia were formed during a transtensional stagethat immediately followed the cessation of northwest dip-ping subduction which produced a late Proterozoic(680–550 Ma) rifted magmatic arc (Keppie & Dostal,1991; Dostal et al. 1995). However, the lack of a subduc-tion signature in the mantle-derived Cambrian basaltssuggests that their source was not in the lithospheric sub-duction-modified mantle, but rather was in normalasthenospheric mantle. Deep Cambrian faults active dur-ing transtensional rifting (Keppie & Murphy, 1988) arelikely to have intersected potential partial melts at differ-ent depths, both at the top of the asthenosphere beneaththe axis of the decaying thermal anomaly, and deeperwithin the asthenosphere in off-axis regions (Fig. 11).

A similar model may be applicable to Cambrian vol-canism elsewhere in the Avalon Terrane: (a) alkalicCambrian volcanism occurs in an area of easternNewfoundland where arc magmatism ended at c. 600 Ma(Greenough & Papezik, 1985); and (b) tholeiiticCambrian magmatism in southern New Brunswick


Figure 9. Rare earth and trace element compositions of isotopically analysed late Ordovician–early Silurian basalts in Cape BretonIsland. Normalizing values as in Figure 2.

occurs in a zone of 560–530 Ma plutonism (Barr et al.1993; Greenough, McCutcheon & Papezik, 1985).Palaeomagnetic and faunal provinciality data suggest thatthe thermal decay of this magmatism took place along amargin of Gondwanan during regional transtension fol-lowing cessation of peripheral arc subduction (Nance &Murphy, 1994; Keppie et al. 1996).

5.a.2. Middle Ordovician–early Silurian intra-platemagmatism

The middle Ordovician–early Silurian basalts show typical within-plate geochemical characteristics whichalso lack a subduction signature suggesting a source innormal asthenospheric mantle. This is also evident intheir εNd isotopic signatures which show an increaseddepleted mantle component relative to the Cambrianbasalts. In the Antigonish Highlands, the volcanic rocks are overlain by sediments of the Arisaig Group that have Laurentian–Baltic geochemical signaturesimplying that their deposition was post-accretionary(Murphy et al. 1996). This is consistent with the proposalof Chandler, Loveridge & Currie (1987) that the early Silurian volcanic rocks (including the White Rock Formation in the Meguma Terrane) form part of an overstep sequence across the entire AppalachianOrogen.

Structural data indicate that eruption and intrusion ofthe Money Point Group in northern Cape Breton Island

and the early Silurian Kingston dyke complex in theAvalon Terrane of southern New Brunswick took place inreleasing bends of a sinistral transtensional regime(Keppie, Dallmeyer & Krogh, 1992; Doig et al. 1990).Rifting in such an environment would have led to thin-ning of the lithosphere and a rise of the asthenospherebeneath rifts (Fig. 11), while active faulting would havefacilitated the ascent of magma to the surface.

New continental reconstructions provide a context forsinistral accretion (Keppie et al. 1996). The eruption ofmiddle Ordovician–early Silurian within-plate volcanicrocks in Nova Scotia is therefore interpreted to haveoccurred during sinistral telescoping after the initialaccretion of the Gander and Avalon terranes withLaurentia (Keppie, 1993, and references therein).

5.a.3. Devonian intra-plate magmatism

The Devonian basalts have a typically within-plate geo-chemical signature derived from the asthenospheric man-tle. This is consistent with the lack of a subductionsignature in all Devonian basalts except of those fromLowland Cove. The Devonian volcanic rocks form part ofa mainly continental sequence that was deposited in intra-montane basins following considerable uplift and erosion(Lynch, 1992; Keppie & Dallmeyer, 1995). In early–lateDevonian times, the northern Appalachians experiencedextensive telescoping that thickened the lithosphere dur-ing dextral transpression (Keppie, 1993; Lynch, 1992;


Figure 10. Rare earth and trace element compositions of isotopically analysed late Ordovician–early Silurian rhyolites in Nova Scotia.Normalizing values as in Figure 2.

Keppie & Dallmeyer, 1995). These authors have inferredthat this telescoping was followed by delamination thatproduced extensive, but short-lived magmatism and gravitational collapse. In this scenario, the asthenosphericmantle flowing laterally to replace the sinking lithos-pheric root would be the source of basaltic melts (Fig. 11). Active faulting (e.g. Minas and Hollow faults)during this period would once again have acted as conduits through which the magmas could reach the surface.

6. Tectonic implications

6.a. Crustal source

The lower limit of Nd isotopic signatures of the Avaloniancrustal source for felsic lavas may be derived by construct-ing lines through the most negative initial εNd values andtheir corresponding TDM model ages from unequivocallyAvalonian suites, that is, those areas underlain byCambro-Ordovician overstep sequences (Fig. 12). Theseinclude those overlying the type Avalon successions in


Figure 11. Geodynamic models for the genesis of Palaeozoic basalt suites: (b) Cambrian–early Ordovician; (c) middle Ordovician; (d)Devonian. See text for discussion.

southeastern Newfoundland (Kerr, Jenner & Fryer, 1995),southeastern Cape Breton Island, the AntigonishHighlands, and the central and eastern CaledonianHighlands of southern New Brunswick (Whalen et al.1994a). An upper limit may be defined by the depletedmantle curve because felsic magmas fractionated fromcoeval mantle-derived mafic rocks may have Nd signa-tures similar to depleted mantle values. TDM ages would,in this case, be similar to the time of their extrusion. Thisrange of isotopic compositions in unequivocallyAvalonian areas allows comparison with other potentiallyAvalonian areas that may lack preservation of the overstepsequence. The inclusion of these areas in the AvalonTerrane has been based primarily upon correlation of lateProterozoic units and events (e.g. Keppie et al. 1991).They include the Burgeo Terrane of southernNewfoundland, central Cape Breton Island (the Bras d'Orand Aspy terranes of Barr & Raeside (1989)), and theBrookville block/terrane lying north of the CaledonianHighlands in southern New Brunswick (Whalen et al.1994a). It is clear that the εNd signature from each of theseareas is indistinguishable from unequivocal Avalonian sig-natures (cf. Barr & Hegner, 1992) (Fig. 12). Furthermore,there appears to be a general spatial trend towards morenegative εNd signatures from east–southeast towest–northwest in Newfoundland, Cape Breton Island,and southern New Brunswick. This suggests a transitionfrom a more primitive middle–late Proterozoic source inthe east–southeast to a mixed source that includes oldercontinental components towards the west–northwest. InCape Breton Island, this transition has been interpreted interms of a late Proterozoic subduction zone dipping north-west beneath an oceanic–cratonic arc (Dostal et al. 1995).More primitive εNd signatures in the Antigonish Highlandshave been similarly attributed by Murphy et al. (1995) toits back-arc position during the late Proterozoic subduc-tion. These signatures contrast with those of the

Laurentian Grenville Province, but are comparable withjuvenile late Proterozoic belts in Gondwana (e.g. theTocatins; Pimentel & Fuck, 1992; Fig. 13), an observationthat led Nance & Murphy (1994) to place the AvalonTerrane peripheral to Gondwana off northwestern SouthAmerica. The εNd values for igneous rocks in the BlairRiver Complex in northwestern Cape Breton Island ploteither in the area of overlap between Grenville and Avalonfields (c. 1.2 Ga Lowland Cove syenite), or within theAvalon field above this area (Devonian felsic dyke andflow) (Fig. 13). Thus, the Nd data presented here do notresolve the Laurentian or Avalonian affinities of the BlairRiver Complex. Furthermore, the primitive εNd signaturesof the Devonian igneous rocks that intrude and overlie theBlair River Complex and also occur throughout CapeBreton Island, give no indication that LaurentianGrenvillian basement was thrust beneath western CapeBreton Island during closure of the Iapetus Ocean as proposed by Stockmal et al. (1987).

6.b. Avalon–Meguma relationships

The Avalon and Meguma terranes are presently separatedby a major fault (the Minas Fault) which permits severaldifferent interpretations of the original relationshipsbetween them: a thrust contact (Keppie & Dallmeyer,1987), a depositional contact (Keppie & Dostal, 1991), oran oceanic suture (Schenk, 1970, 1981). Comparison of


Figure 12. Lower limit of εNdt vs. crystallization age for igneousrocks (constructed as in Fig. 3) from different parts of theAvalon Terrane and correlatives in Atlantic Canada rocks. Datasources include Barr & Hegner (1992); Fryer et al. (1992);Whalen et al. (1994a); Kerr, Jenner & Fryer (1995) and thispaper.

Figure 13. Lower limit of εNdt vs. crystallization age forAvalonian igneous rocks compared to (1) igneous and sedimen-tary rocks and xenoliths (Mb = metabasaltic; Ms = metasedi-mentary) in Devonian intrusions from the Meguma and Ganderterranes (Clarke & Halliday, 1985; Clarke, Halliday &Hamilton, 1988; Clarke, Chatterjee, & Giles, 1993; Eberz et al.1991; Kerr, Jenner & Fryer, 1995); (2) Grenville-aged rocks ofeastern and southern Ontario and northern New York state(Dickin & McNutt, 1989; Dickin, McNutt & Clifford, 1990;Marcantonio et al. 1990; Daly & McLelland, 1991); (3) igneousrocks in and on the Blair River Complex (triangles: Barr &Hegner, 1992); (4) plutonic rocks from West Africa (data pointand evolution envelope from Allegre & Ben Othman, 1980) and(5) the Tocatins Province, Brazil (Pimentel & Fuck, 1992).

the εNd-time signatures of the felsic magmas in theMeguma Terrane (White Rock Formation and uncontami-nated Upper Devonian plutonic rocks) reveals that theytoo had an isotopic source indistinguishable from thatbeneath the Avalon Terrane. The Proterozoic TDM modelages of these middle Ordovician rhyolites and lateDevonian felsic plutons suggests that the Cambro-Ordovician Meguma Group was not deposited on oceanic

lithosphere. Instead, it is consistent with its deposition oncontinental crust isotopically indistinguishable from thatof the Avalon Terrane (Keppie & Dostal, 1990). Indeed,the depositional setting of the White Rock Formation isconsistent with the presence of a shoreline on the north-western side of the Meguma Terrane in middleOrdovician–Silurian times (Lane, 1976). Contaminationof the Devonian intrusives by incorporation of xenoliths


Figure 14. Global palinspastic reconstruction for (a) the early Cambrian (modified from McKerrow, Scotese & Brasier, 1992) showingMeguma (M), Avalon (A) and Gander (G) terranes as a peninsula off northwestern South America with hypothetical current directionsthat would derive Meguma sediments from northwest Africa and Avalon–Gander sediments from northwest South America; and (b) themiddle Silurian (modified from Van der Voo, 1988) showing accretion of the Meguma–Avalon–Gander peninsula to Laurentia.

moved their εNd signature towards that of the hostMeguma Group metasediments (Halliday et al. 1988).Unlike the White Rock Formation, the more negative εNdsignature of these metasediments is inconsistent with anAvalonian crustal source. Their source regions, and theirextent of tectonic transport are therefore unclear.However, the data are consistent with detrital zircon agesfrom the Meguma Group that indicate a mixture ofProterozoic and Archaean sources (550–675, 2000 and3000 Ma: Krogh & Keppie, 1990). Comparison of thesedetrital U–Pb ages with potential source areas suggeststhat the Meguma Group may have been derived fromnorthwest Africa (Reguibat and Man cratons: Krogh &Keppie, 1990), as proposed by Schenk (1970, 1981).

Similar εNd features to those in the Meguma Terranemay be observed in the Gander Zone, a continental riseprism lying on the other side of the Avalon Terrane fromthe Meguma Terrane (Kerr, Jenner & Fryer, 1995).However, the Gander Group contains a wider range ofdetrital zircon populations (530–600, 1017, 1112, 1215,1347, 1514, 2062 and 2717 Ma: O’Neill, 1991). This agedistribution is comparable to the detrital zircon popula-tion in a Neoproterozoic III volcano-sedimentary samplefrom the Antigonish Highlands (610–630, 1000,1160–1200, 1520–1550, 1835, 1950–2000 and c. 2600Ma: Keppie & Krogh, 1990). On the basis of these detri-tal ages, a potential source for both the Gander Group andthe Neoproterozoic sediments in the Avalon Terrane ofthe Antigonish Highlands is provided by the southwestGuyana craton and the Arequipa massif of South Americawhere orogenic belts with ages of 500–600 Ma(Brasiliano), 900–1100 Ma (Sunsas), 1250–1450 Ma(Rhondonian), 1500–1750 Ma (Jurena-Rio Negro) and1900–2250 Ma (Maroni-Itaciaunas) have been recog-nized (Teixeira et al. 1989).

6.c. Significance for palaeocontinental reconstructions

The εNd basement signatures and detrital zircon agesplace constraints on palinspastic reconstructions of theMeguma, Avalon and Gander terranes during earlyPalaeozoic times. It is proposed that the Avalon Terrane atthis time formed a peninsula off northwestern SouthAmerica with the Meguma and Gander terranes formingcontinental rise prisms on the northern and southernsides, respectively (Fig. 14). Cambrian transcurrent faultsin Nova Scotia that acted as conduits for magmatism, arebelieved to have been part of a regional tectonic environ-ment that dispersed the Avalon Terrane westward from aposition peripheral to Gondwana (northwestern SouthAmerica) into a promontory or chain of islands (Keppieet al. 1996). Structural data indicate that the ExploitsTerrane began to be accreted to the Gander Terrane in lateCambrian–early Ordovician times (Keppie, 1993),thereby providing a younger limit on the transition from apassive to an active margin. The early Cambrianpalinspastic base map published by McKerrow, Scotese& Brasier (1992) predicts current directions that would

be compatible with the Avalon Terrane in this positionand with derivation of Meguma Group detritus fromnorthwestern Africa and Gander–Avalon detritus fromnorthwestern South America (Fig. 14). On the middleSilurian palinspastic base map of Van der Voo (1988),eastern Laurentia lies relatively close to northwesternSouth America, and it is necessary to rotate theMeguma–Avalon–Gander peninsula clockwise to accom-modate it in the intervening space (Fig. 14). On thisreconstruction, an alternative origin for the Blair RiverComplex, which is generally assumed to have aLaurentian provenance (Barr & Raeside, 1989) becomespossible, namely the Grenville-aged Sunsas Orogen thatruns along the southwestern margin of the Guyanan cra-ton. This proposition can be tested when precise U–Pbdata become available from the Sunsas Orogen.

The proposed palinspastic reconstruction also suggests a model for the transfer of theMeguma–Avalon–Gander–Exploits peninsula fromSouth America to Laurentia. An essential element in this puzzle may be the Precordilleran Terrane in westernArgentina, which contains a Grenvillian basement overlain by a Cambro-Ordovician sequence identical in stratigraphy and fauna with that of eastern Laurentia(Ramos et al. 1986; Keppie, 1993). This terrane appears to have been derived from eastern Laurentia and was transferred to South America by lateOrdovician–Silurian times, at about the same time as theMeguma–Avalon–Gander terranes were accreted toLaurentia (Keppie, 1993; Astini, Benedetto & Vaccari,1995). The palinspastic reconstructions of the major cratons by Dalziel (1991) suggest a possible mechanism(Fig. 15). During most of the Palaeozoic, Laurentiamoved anticlockwise around Gondwana with a phase of


Figure 15. Relative positions of Laurentia and Gondwana fromCambrian to Devonian (500–390 Ma) showing the transfer ofthe Meguma–Avalon–Gander terranes from northwestern SouthAmerica to eastern Laurentia and the transfer of thePrecordillera from Laurentia to southwestern South America.Modified from Dalziel (1991). Abbreviations as in Figure 14,except AA = Arequipa-Antofalla.

clockwise rotation during the late Ordovician and earlySilurian. As it did so, it would have been possible for thePrecordillera and Meguma–Avalon–Gander peninsulas tohave rotated clockwise and, in so doing, become trans-ferred from one craton to the other (Fig. 15). The episodeof clockwise rotation coincides with the separation of theAvalon–Meguma–Gander terranes from Gondwana andtheir sinistral accretion to Laurentia (Keppie et al. 1996).

Acknowledgements. This project was supported by NaturalSciences and Engineering Research Council (NSERC) CanadaOperating and Lithoprobe grants of JBM and JD and UniversityCouncil Research grants at St. Francis Xavier University, theCanada–Nova Scotia Mineral Development Agreement, and theDepartment of Natural Resources. JDK acknowledges the sup-port of the James Chair of Pure and Applied Sciences, St.Francis Xavier University. We are grateful to G. Lang Farmer,David Rowley, R. Damian Nance and an anonymous reviewerfor constructive criticism of an earlier version of this paper.BLC thanks John Blenkinsop for his support and allowingunlimited access to clean laboratory and mass spectrometerfacilities at Carleton University, and Keith Bell, Tony Simonettiand Keiko Hattori for invaluable help and advice.

References

ALLEGRE, C. J. & BEN OTHMAN, D. 1980. Nd–Sr isotopic rela-tionship in granitoid rocks and continental crust develop-ment: a chemical approach to orogenesis. Nature 286,335–42.

ASTINI, R. A., BENEDETTO, J. L. & VACARRI, N. E. 1995. TheEarly Palaeozoic evolution of the Argentine Precordilleraas a Laurentian rifted, drifted, and collided terrane – a geodynamic model. Geological Society of AmericaBulletin 107, 253–73.

BARR, S. M. & HEGNER, E. 1992. Nd isotopic compositions offelsic igneous rocks in Cape Breton Island, Nova Scotia.Canadian Journal of Earth Sciences 29, 650–7.

BARR, S. M. & JAMIESON, R. A. 1991. Tectonic setting andregional correlation of Ordovician–Silurian rocks of theAspy terrane, Cape Breton Island, Nova Scotia. CanadianJournal of Earth Sciences 28, 1769–79.

BARR, S. M. & RAESIDE, R. P. 1989. Tectono-stratigraphic ter-ranes in Cape Breton Island, Nova Scotia: Implications forthe configuration of the northern Appalachian orogen.Geology 7, 822–5.

BARR, S. M., BEVIER, M. L., WHITE, C. E. & DOIG, R. 1993.Magmatic history of the Caledonia (Avalon) terrane ofsouthern New Brunswick based on U–Pb (zircon)geochronology. New Brunswick Department of NaturalResources and Energy, Mineral Resources, InformationCircular 93-1, 10–17.

BOUCOT, A. J. 1960. Implications of Rhenish Lower Devonianbrachiopods from Nova Scotia. 21st InternationalGeological Congress Report 12, 129–37.

BOUCOT, A. J. 1975. Evolution and Extinction Rate Controls(Developments in Palaeontology and Stratigraphy, 1).Elsevier Scientific Publ. Co., 427 pp.

BOUCOT, A. J., DEWEY, J. F., DINELEY, D. L., FYSON, W. K.,HICKOX, C. F., MCKERROW, W. S. & ZIEGLER, A. M. 1974.Geology of the Arisaig Area. Geological Society ofAmerica Special Paper 139, 191 pp.

CHANDLER, F. W., LOVERIDGE, D. & CURRIE, K. L. 1987. The ageof the Springdale Group, western Newfoundland, and correlative rocks – evidence for a Llandovery overlap

sequence in the Canadian Appalachians. Transactions ofthe Royal Society of Edinburgh: Earth Sciences 78, 41–9.

CLARKE, D. B., CHATTERJEE, A. K. & GILES, P. S. 1993.Petrochemistry, tectonic history, and Sr–Nd systematics ofthe Liscomb Complex, Meguma Lithotectonic Zone, NovaScotia. Canadian Journal of Earth Sciences 30, 449–64.

CLARKE, D. B. & HALLIDAY, A. N. 1985. Sm/Nd isotopic inves-tigation of the age and origin of the Meguma Zonemetasedimentary rocks. Canadian Journal of EarthSciences 22, 102–7.

CLARKE, D. B., HALLIDAY, A. N. & HAMILTON, P. J. 1988.Neodymium and Strontium isotopic constraints on the ori-gin of the peraluminous granitoids of the South MountainBatholith, Nova Scotia, Canada. Chemical Geology 73,15–24.

COUSENS, B. L., ALLAN, J. F., LEYBOURNE, M. I., CHASE, R. L. &VAN WAGONNER, N. 1995. Mixing of magmas fromenriched and depleted mantle sources in the northeastPacific: West Valley segment, Juan de Fuca Ridge.Contribributions to Mineralogy and Petrology 120,337–57.

DALY, J. S. & MCLELLAND, J. M. 1991. Juvenile MiddleProterozoic crust in the Adirondack Highlands, GrenvilleProvince, northeastern North America. Geology 19,119–22.

DALZIEL, I. W. D. 1991. Pacific margins of Laurentia and EastAntarctica–Australia as a conjugate rift pair: evidence andimplications for an Eocambrian supercontinent. Geology19, 598–601.

DEPAOLO, D. J. 1988. Neodymium isotope geochemistry: anintroduction. New York: Springer-Verlag.

DICKIN, A. P. & MCNUTT, R. H. 1989. Nd model age mapping ofthe southeast margin of the Archean foreland in theGrenville Province of Ontario. Geology 17, 299–302.

DICKIN, A. P., MCNUTT, R. H. & CLIFFORD, P. M. 1990. Aneodymium isotope study of plutons near the GrenvilleFront in Ontario, Canada. Chemical Geology 83, 315–24.

DOIG, R., NANCE, R. D., MURPHY, J. B. & CASSEDAY, R. P. 1990.Evidence for Silurian sinistral accretion of the Avaloncomposite terrane in Canada. Journal of GeologicalSociety of London 147, 927–30.

DOSTAL, J., BARAGAR, W. R. A. & DUPUY, C. 1986. Petrogenesisof the Natkusiak continental basalts, Victoria Island,N.W.T. Canadian Journal of Earth Sciences 23, 622–32.

DOSTAL, J., DUPUY, C. & CABY, R. 1994. Geochemistry of theNeoproterozoic Tilemsi belt of Iforas (Mali, Sahara): acrustal section of an oceanic island arc. PrecambrianResearch 65, 55–69.

DOSTAL, J. & DUPUY, C. 1984. Geochemistry of the NorthMountain basalts, Nova Scotia, Canada. ChemicalGeology 45, 245–61.

DOSTAL, J., KEPPIE, J. D., DUPUY, C. 1983. Petrology and geo-chemistry of Devono-Carboniferous volcanic rocks inNova Scotia. Maritime Sediments and Atlantic Geology 19,59–71.

DOSTAL, J., KEPPIE, J. D. & MURPHY, J. B. 1990. Geochemistryof Late Proterozoic basaltic rocks from southeastern CapeBreton Island, Nova Scotia. Canadian Journal of EarthSciences 23, 619–31.

DOSTAL, J., KEPPIE, J. D., COUSENS, B. L. & MURPHY, J. B. 1995.550–580 Ma magmatism in Cape Breton Island (NovaScotia, Canada): the product of NW-dipping subductionduring the final stage of amalgamation of Gondwana.Precambrian Research 76, 93–113.

DUNNING, G. R., BARR, S. M., RAESIDE, R. P. & JAMIESON, R. A.1990. U–Pb zircon, titanate, and monazite ages in the Bras


d'Or and Aspy terranes of Cape Breton Island, NovaScotia: implications for magmatic and metamorphic history. Geological Society of America Bulletin 102,322–30.

EBERZ, G. W., CLARKE, D. B., CHATTERJEE, A. K. & GILES, P. S.1991. Chemical and isotopic composition of the lowercrust beneath the Meguma lithotectonic zone, NovaScotia,: evidence from granulite facies xenoliths.Contributions to Mineralogy and Petrology 109, 69–88.

ELLAM, R. M., CARLSON, R. W. & SHIREY, S. B. 1992. Evidencefrom Re–Os isotopes for plume-lithosphere mixing inKaroo flood basalt genesis. Nature 359, 718–21.

ELLAM, R. M. & COX, K. G. 1991. An interpretation of Karoopicrite basalts in terms of interaction between asthenos-pheric magmas and the mantle lithosphere. Earth andPlanetary Science Letters 105, 330–42.

FITTON, J. G., JAMES, D., KEMPTON, P. D., ORMEROD, D. S. &LEEMAN, W. P. 1988. The role of lithospheric mantle in thegeneration of late Cenozoic basic magmas in the westernUnited States. Journal of Petrology Special LithosphereIssue, 331–49.

FRYER, B. J., KERR, A., JENNER, G. A. & LONGSTAFFE, F. J. 1992.Probing the crust with plutons: regional isotopic geochem-istry of granitoid intrusions across insular Newfoundland.Current Research. Newfoundland Department of Minesand Energy, Report 92-1, 119–39.

GREENOUGH, J. D. & PAPEZIK, V. S. 1985. Petrology and geo-chemistry of Cambrian volcanic rocks from the AvalonPeninsula, Newfoundland. Canadian Journal of EarthSciences 22, 1594–1601.

GREENOUGH, J. D., MCCUTCHEON, S. R. & PAPEZIK, V. S. 1985.Petrology and geochemistry of Cambrian volcanic rocksfrom the Avalon Zone in New Brunswick. CanadianJournal of Earth Sciences 22, 881–92.

HALLIDAY, A. N., DICKIN, A. P., FALLICK, A. E. & FITTON, J. G.1988. Mantle dynamics: a Nd, Sr, Pb and O isotopic studyof the Cameroon Line volcanic chain. Journal of Petrology29, 181–211.

HART, S. R. & DAVIS, K. E. 1978. Nickel partitioning betweenolivine and silicate melt. Earth and Planetary ScienceLetters 40, 203–19.

HELMSTAEDT, H. & TELLA, S. 1973. Pre-Carboniferous struc-tural history of S.E. Cape Breton Island, Nova Scotia.Maritime Sediments 9, 88–99.

HOOPER, P. R. & HAWKESWORTH, C. J. 1993. Isotopic and geo-chemical constraints on the origin and evolution of theColumbia River basalt. Journal of Petrology 34, 1203–46.

HUTCHINSON, R. D. 1952. The stratigraphy and trilobite faunas of the Cambrian sedimentary rocks of Cape BretonIsland, Nova Scotia. Geological Survey of Canada,Memoir 262.

KEEN, C. E., KAY, W. A., KEPPIE, J. D., MARILLIER, F., PE-PIPER, G.& WALDRON, J. W. F., 1991. Deep seismic reflection datafrom the Bay of Fundy and Gulf of Maine: tectonic impli-cations for the northern Appalachians. Canadian Journalof Earth Sciences 28, 1096–1111.

KEMPTON, P. D., DUNGAN, M. A. & BLANCHARD, D. P. 1987.Petrology and geochemistry of xenolith-bearing alkalicbasalts from the Geronimo Volcanic Field, southeastArizona; evidence for polybaric fractionation and implica-tions for mantle heterogeneity. In Mantle metasomatismand alkaline magmatism. (eds E. M. Morris and J. D.Pasteris), pp. 347–70. Geological Society of AmericaSpecial Paper no. 215.

KEPPIE, J. D. 1982. Tectonic Map of Nova Scotia. Scale1:500,000. Nova Scotia Department of Mines and Energy.

KEPPIE, J. D. 1985. The Appalachian collage. In The CaledonideOrogen – Scandinavia and related areas (eds D. G. Geeand B. A. Sturt), pp. 1217–26. Chichester: Wiley.

KEPPIE, J. D. 1989. Northern Appalachian terranes and their accretionary history. In Terranes in the Circum-Atlantic Paleozoic Orogens (ed. R. D. Dallmeyer), pp.159–92. Geological Society of America Special Paper no. 230.

KEPPIE, J. D. 1993. Synthesis of Palaeozoic deformationalevents and terrane accretion in the Canadian Appalachians.Geologische Rundschau 82, 381–431.

KEPPIE, J. D. & DALLMEYER, R. D. 1987. Dating transcurrentterrane accretion: an example from the Meguma andAvalon Composite terranes in the northern Appalachians.Tectonics 6, 831–47.

KEPPIE, J. D. & DALLMEYER, R. D. 1989. Tectonic map ofPrecambrian and Palaeozoic terranes in Circum-AtlanticPhanerozoic Orogens. Scale 1:5000000. InternationalGeological Correlation Programme Project #233, RR#3,Wolfville, Nova Scotia, Canada.

KEPPIE, J. D. & DALLMEYER, R. D. 1995. Late Palaeozoic colli-sion, delamination, short-lived magmatism, and rapiddenudation in the Meguma Terrane (Nova Scotia, Canada):constraints from 40Ar/39Ar isotopic data. CanadianJournal of Earth Sciences 32, 644–59.

KEPPIE, J. D., DALLMEYER, R. D. & KROGH, T. E. 1992. U–Pband 40Ar/39Ar mineral ages from the Cape North area,northern Cape Breton Island: Implication for docking ofthe Avalon Composite Terrane. Canadian Journal of EarthSciences 29, 277–95.

KEPPIE, J. D. & DOSTAL, J. 1980. Palaeozoic volcanic rocks ofNova Scotia. International Geological CorrelationProgramme Project 27: Caledonide Orogen. Proceedingsof Virginia Polytechnic Institute and State UniversityMemoir 2, 249–56.

KEPPIE, J. D. & DOSTAL, J. 1991. Late Proterozoic tectonicmodel for the Avalon Terrane in Maritime Canada.Tectonics 10, 842–50.

KEPPIE, J. D., DOSTAL, J., MURPHY, J. B. & NANCE, R. D. 1996.Terrane transfer between Eastern Laurentia and WesternGondwana in the Early Palaeozoic: constraints on globalreconstructions. In Avalonian and Related peri-Gondwanan terranes of the circum-North Atlantic (eds R.D. Nance and M. D. Thompson), pp. 369–80. GeologicalSociety of America Special Paper no. 304.

KEPPIE, J. D., DOSTAL, J. & ZENTILLI, M. 1979. Early Silurianvolcanic rocks at Arisaig, Nova Scotia. Canadian Journalof Earth Sciences 16, 1635–9.

KEPPIE, J. D. & KROGH, T. E. 1990. Detrital zircon ages fromLate Precambrian conglomerate, Avalon CompositeTerrane, Antigonish Highlands, Nova Scotia. GeologicalSociety of America Abstracts with Programs 22, 27–8.

KEPPIE, J. D. & MURPHY, J. B. 1988. Anatomy of a pull-apartbasin: an example from the Cambro-Ordovician rocks inthe Antigonish Highlands of Nova Scotia. MaritimeSediments and Atlantic Geology 24, 123–38.

KEPPIE, J. D., NANCE, R. D., MURPHY, J. B. & DOSTAL, J. 1991.Northern Appalachians Avalon and Meguma terranes. In The West African Orogens and Circum-AtlanticCorrelatives (eds R. D. Dallmeyer and J. P. Lecorche),pp. 315–33. Heidelberg: Springer.

KERR, A., JENNER, G. A. & FRYER, B. J. 1995. Sm–Nd isotopicgeochemistry of Precambrian to Palaeozoic granitoidsuites and the deep-crustal structure of the southeast mar-gin of the Newfoundland Appalchians. Canadian Journalof Earth Sciences 32, 224–45.


KONTAK, D. J., CHATTERJEE, A. K., REYNOLDS, P. H. & TAYLOR, K.1990. 40Ar/39Ar geochronological study of metamorphicand igneous rocks of the Liscomb Complex, southernNova Scotia. Nova Scotia Department of Mines andEnergy Report 90-1, 33–40.

KROGH, T. E. & KEPPIE, J. D. 1990. Age of detrital zircon andtitanite in the Meguma Group, southern Nova Scotia,Canada: clues for the origin of the Meguma Terrane.Tectonophysics 177, 307–23.

LANDING, E. 1991. A unified uppermost Precambrian throughLower Cambrian lithostratigraphy for Avalonian NorthAmerica. Geological Society of America, Abstracts withPrograms 23, 56.

LANDING, E. & MURPHY, J. B. 1991. UppermostPrecambrian(?)–Lower Cambrian of Mainland NovaScotia: Faunas, depositional environments and strati-graphic revision. Journal of Palaeontology 65, 382–96.

LANE, T. E. 1976. Stratigraphy of the White Rock Formation.Maritime Sediments 12, 87–106.

LESQUER, A., BOURMATTE, A. & DAUTRIA, J. M. 1988. Deepstructure of the Hoggar domal uplift (Central Sahara, southAlgeria) from gravity, thermal and petrological data.Tectonophysics 152, 71–87.

LIN, S. 1993. Relationship between the Aspy and Bras d'Or "ter-ranes" in the northeastern Cape Breton Highlands, NovaScotia. Canadian Journal of Earth Sciences 30, 1773–81.

LYNCH, J. V. G. 1992. Extensional collapse of the AppalachianOrogen: evidence from the Cape Breton Highlands.Geological Association of Canada Abstracts 17, A69.

LYNCH, G., TREMBLAY, C. & ROSE, H. 1993. Compressionaldeformation and extensional denudation of Early Silurianvolcanic overlap assemblages in western Cape BretonIsland, Nova Scotia. Geological Survey of Canada, Paper93-1D, 103–10.

MAALOE, S. & AOKI, K. 1977. The major element compositionof the upper mantle estimated from the composition oflherzolites. Contributions to Mineralogy and Petrology 63,161–73.

MARCANTONIO, F., DICKIN, A. P., MCNUTT, R. H. & HEAMAN, L. M. 1990. Isotopic evidence for the crustal evo-lution of the Frontenac Arch in the Grenville Province ofOntario, Canada. Chemical Geology 83, 297–314.

MCKERROW, W. S., SCOTESE, C. R. & BRASIER, M. D. 1992.Early Cambrian continental reconstructions. Journal of theGeological Society, London 149, 599–606.

MENZIES, M. A., KEMPTON, P. & DUNGAN, M. 1985. Interactionof continental lithosphere and asthenospheric melts belowthe Geronimo volcanic field, Arizona, USA. Journal ofPetrology 26, 663–93.

MENZIES, M. A., LEEMAN, W. P. & HAWKESWORTH, C. J. 1983.Isotope geochemistry of Cenozoic volcanic rocks revealsmantle heterogeneity below western USA. Nature 303,205–9.

MURPHY, J. B., CAMERON, K., KEPPIE, J. D., DOSTAL, J. &HYNES, A. J. 1985. Cambrian volcanism in Nova Scotia,Canada. Canadian Journal of Earth Sciences 22, 599–606.

MURPHY, J. B., KEPPIE, J. D., DOSTAL, J. & HYNES, A. J. 1990.The geochemistry and petrology of the Late PrecambrianGeorgeville Group: a volcanic arc-rift succession in theAvalon terrane of Nova Scotia. Geological Society ofLondon, Special Publication 51, 383–93.

MURPHY, J. B., KEPPIE, J. D., DOSTAL, J. & COUSENs, B. L. 1995.Repeated lower crustal melting beneath the Antigonishhighlands, Avalon Composite Terrane, Nova Scotia: Ndisotopic evidence and tectonic implications. GeologicalSociety of America Special Paper 304, 109–20.

MURPHY, J. B., KEPPIE, J. D., DOSTAL, J., WALDRON, J. W. F. &CUDE, M. P. 1996. Geochemical and isotopic cha-racteristics of Early Silurian clastic sequences inAntigonish Highlands, Nova Scotia, Canada: constraintson the accretion of Avalonia in the Appalachian–Caledonide Orogen. Canadian Journal of Earth Sciences33, 379–88.

MURPHY, J. B., KEPPIE, J. D. & HYNES, A. J. 1991. Geology ofthe Antigonish Highlands. Geological Survey of Canada,Paper no. 89-10, 114 pp.

NANCE, R. D. & MURPHY, J. B. 1994. Contrasting basement signatures and the palinspastic restoration of peripheralorogens: example from the Neoproterozoic Avalonian–Cadomian Belt. Geology 22, 617–20.

O'NEILL, P. P. 1991. Geology of the Weir's Pond area,Newfoundland (NTS 2E/1). Newfoundland Department ofMines and Energy, Geological Survey Branch Report no.91-3, 144 pp.

PEARCE, J. A. & CANN, J. R. 1973. Tectonic setting of basic vol-canic rocks determined using trace element analyses.Earth and Planetary Science Letters 19, 290–300.

PIMENTEL, M. M. & FUCK, R. A. 1992. Neoproterozoic crustalaccretion in central Brazil. Geology 20, 375–9.

RAMOS, V. A., JORDAN, T. E., ALLMENDINGER, R. W.,MPODOZIS, C., KAY, S. M., CORTES, J. M. & PALMA, M.1986. Palaeozoic terranes of the central Argentine–ChileanAndes. Tectonics 5, 855–80.

SCHENK, P. E. 1970. Regional variation of the flysch-likeMeguma Group (Lower Palaeozoic) of Nova Scotia, com-pared to recent sedimentation off the Scotian shelf.Geological Association of Canada, Special Paper 7,127–53.

SCHENK, P. E. 1981. The Meguma zone of Nova Scotia: a remnant of western Europe, South America or Africa?Canadian Society of Petroleum Geologists, Memoir 7,119–48.

STOCKMAL, G. S., COLMAN-SADD, S. P., KEEN, C. E., O’BRIEN, S. J.& QUINLAN, G. M. 1987. Collision along an irregular mar-gin: a regional plate tectonic interpretation of the CanadianAppalachians. Canadian Journal of Earth Sciences 24,1098–1107.

SUN, S. S. 1982. Chemical composition and origin of the Earth'sprimitive mantle. Geochimica et Cosmochimica Acta 46,179–92.

SUN, S. S. & MCDONOUGH, W. F. 1989. Chemical and isotopicsystematics of oceanic basalts: implications for mantlecompositions and processes. In Magmatism in the oceanbasins (eds A. D. Saunders and M. J. Norry), pp. 313–45.Geological Society of London, Special Publication no. 42.

TEIXEIRA, W., TASSINARI, C. C. G., CORDANI, U. G. &KAWASHITA, K. 1989. A review of the geochronology of theAmazonian craton: tectonic implications. PrecambrianResearch 42, 213–27.

THEOKRITOFF, G. 1979. Early Cambrian provincialism and bio-geographic boundaries in the North Atlantic region.Lethaia 12, 281–95.

VAN DER VOO, R. 1988. Palaeozoic palaeogeography of NorthAmerica, Gondwana, and intervening displaced terranes:comparisons of palaeomagnetism with palaeoclimatologyand biogeographical patterns. Geological Society ofAmerica Bulletin 100, 311–24.

WEEKS, F. J. 1954. Southeast Cape Breton Island, Nova Scotia.Geological Survey of Canada, Memoir 277.

WHALEN, J. B., JENNER, G. A., CURRIE, K. L., BARR, S. M.,LONGSTAFFE, F. J. & HEGNER, E. 1994a. Geochemical andisotopic characteristics of granitoids in the Avalon Zone,


southern New Brunswick: possible evidence for repeateddelamination events. Journal of Geology 102, 269–82.

WHALEN, J. B., JENNER, G. A., HEGNER, E., GARIEPY, C. &LONGSTAFFE, F. J. 1994b. Geochemical and isotopic (Nd,O, Pb) constraints on granite sources in the Humber andDunnage zones, Gaspesie, Quebec, and New Brunswick:implications for tectonics and crustal structure. CanadianJournal of Earth Sciences 31, 323–40.

WHITE, C. E., BARR, S. M., BEVIER, M. L. & KAMO, S. 1994. Arevised interpretation of Cambrian and Ordovician rocks inthe Bourinot belt of central Cape Breton Island, NovaScotia. Atlantic Geology 30, 123–42.

WINCHESTER, J. A. & FLOYD, P. A. 1977. Geochemical discrimi-nation of different magma series and their differentiationproducts using immobile elements. Chemical Geology 20,325–43.


Palaeozoic within-plate volcanic rocks in Nova Scotia (Canada) reinterpreted: isotopic constraints on magmatic source and palaeocontinental reconstructions

Documents