29. VOLCANIC ROCKS OF NATURALISTE PLATEAU, EASTERN INDIAN OCEAN,SITE 264, DSDP LEG 28

A. B. Ford, U.S. Geological Survey, Menlo Park, California

ABSTRACT

The upper part of a pile of andesitic volcanogenic sediments un-derlying Upper Cretaceous chalk was sampled on the NaturalistePlateau at DSDP Site 264. Material recovered from the pile is entire-ly volcanic and is composed mainly of rock fragments that probablyrepresent coarse clasts in the sequence. Minor conglomerate in whichpebbles of volcanic rock are embedded in a highly altered tuffaceousmatrix was recovered at two levels. In general, high alteration makesinterpretation of original compositions uncertain. The volcanic clastsrange widely in composition from rhyolite to probable andesite orbasaltic andesite of an apparent calcalkaline andesitic suite unlikeany volcanic suite now exposed on nearby parts of southwestAustralia. Rhyolitic rocks show apparent ignimbritic textures,suggesting that volcanism in the unknown source volcanic complexwas at least in part subaerial. The source terrain must have beennearby, possibly even the plateau which has been only superficiallysampled itself. Petrologic and chemical characteristics of the rockssuggest a volcanic island-arc source or perhaps a continental marginvolcanic and orogenic belt, now submerged, related to an early stageof development of the eastern Indian Ocean.

INTRODUCTION

Naturaliste Plateau is a submerged, table-like prom-ontory extending about 500 km westward towardBroken Ridge from Cape Leeuwin in southwestern Aus-tralia. The lithologic character of Cretaceous or oldermaterials on the plateau is poorly known; therefore, the

112°E

origin of the plateau and its setting in the early history ofthe eastern Indian Ocean is highly conjectural.

The plateau has been penetrated at only two sites ofthe Deep Sea Drilling Project: Site 258 of Leg 26 on thenorthern rim, and Site 264 of Leg 28 on the southernedge (Figure 1). At Site 258, drilling terminated in lowerCretaceous glauconitic sand and muddy silt below Al-

116°E

34°S

36°S

AUSTRALIA

Figure 1. Map of Naturaliste Plateau showing locations ofDSDP Sites 258 (Leg 26) and 264 (Leg 28). Bathymetric contoursin fathoms.

821

A. B. FORD

bian clay without reaching basement rocks (Davies,Luyendyk, et al., 1974). At Site 264, coring bottomed inhighly varied volcanogenic rocks below Cenomaniannannofossil chalk (Site 264 Report, this volume). Com-pared with overlying calcareous sediments, thesevolcanic materials are well lithified and form a strongacoustic-basement reflector at about 170 meters belowbottom at the drill site.

The plateau has been sampled locally by dredging.Dredge hauls from the steep north scarp by USNSEltanin Cruise 55 in 1972 yielded manganese slabs, someof which adhered to deeply weathered crystalline con-tinental rocks (Heezen and Tharp, 1973). The characterof the dredge samples suggested that the plateau is asubsided part of the Australian continent rather than ex-ceptionally thick oceanic crust as earlier proposed.

One of the purposes of drilling at Site 264 was to in-vestigate the continental or oceanic character of base-ment terrain from samples obtained in situ. This reportdescribes the petrographic and chemical characteristicsof the volcanic materials from the basal unit at the site.The possible tectonic setting of these volcanics in theevolution of this part of the eastern Indian Ocean isbriefly discussed. Much remains unknown about thisdeposit, as it was penetrated only superficially andrecovery was very poor. Nonetheless, as it containslithologic types previously unreported for this region,the materials recovered provide data critical to inter-pretation of tectonic history of the Naturaliste Plateauand nearby areas.

OCCURRENCE AND AGEVolcanic materials of highly varied mineralogy, com-

position, and texture occur at Site 264 below bottomdepth by at least 170.7 meters. Recovery of material wasvery poor. Broken and abraded rock fragments less thanabout 8 cm long were mainly recovered from corecatchers to a depth of 215.5 meters at the bottom of thehole. All the materials are altered to some extent, someso strongly that, other than being volcanic, their originallithologic character is obscured.

The contact of the volcanic unit with overlying clay-rich nannofossil and foram chalk is probably uncon-formable. As the actual contact was not observed in anysingle length of core, some of the comparatively soft in-tervening chalk, possibly as much as about 6 meters, wasprobably washed away under high pumping pressures.The downward increase of induration of the calcareoussediments is evidently not related to the underlying ig-neous materials here described, for the gradual transi-tion from calcareous ooze to chalk occurs nearly 90meters above the contact, and the lowest chalkrecovered is only semilithified and shows no evidence ofmetamorphic recrystallization. It is unlikely thereforethat any of the igneous-rock fragments at the contactrepresent a sill. Although the contact is probably uncon-formable, the absence of distinguishable reworkedvolcanic detritus in the basal chalks suggests thatcurrent velocities and local relief were minimal duringinitial carbonate sedimentation on the surface of thevolcanic pile.

The age of the volcanogenic sediment pile and thevolcanism is uncertain. The lowest chalk immediatelyabove the deposit contains a range of Cenomanian toCampanian microfossils (see paleontology report forSite 264, this volume). There is some evidence ofreworking of the microfossil assemblage by bottomcurrents. No lower age limit for the volcanic deposit isavailable and the high degree of alteration of all volcanicmaterials recovered precludes age determination byradiometric dating. Volcaniclastic sediments that in-clude minor volcanic conglomerate from the NinetyeastRidge are dated paleontologically as Paleocene or older(von der Borch, Sclater, et al., 1974). However, in theabsence of reported petrochemical data, correlationwith the volcanic sediments here described is uncertain.

NATURE OF THE VOLCANIC DEPOSITPetrogenetic interpretation of the volcanic deposit is

severely limited by the poor recovery of materials at Site264. Most recovered material consists of angular to sub-angular broken fragments of volcanic rock of a singlelithology. These fragments could be interpreted asrepresenting flow units within the section. But polymic-tic rocks with coarsely clastic texture were obtained attwo widely separated levels; one in Core 11 near the topof the section, the other in Sample 14, CC. Highlyvariable coring speed in this unit suggests that the entiredrilled "basement" section consists of inhomogeneousmaterial such as hard pebbles, cobbles, and boulders ina soft clayey to sandy matrix similar to that of thepolymictic rocks that appear to be volcanic con-glomerates (Figure 2). Accordingly, the rock fragmentsshowing a single lithology are considered likely to bepieces of very large clasts from which the matrix hasbeen washed away, and the entire drilled section is in-ferred to be the upper part of a pile of mainly volcanicsediments including conglomerate, tuff, and possiblybreccia. The presence within the section of flow units oreven thin sills or dikes cannot be ruled out. Moreover,the mechanism of sediment transport is uncertain; if bywater, the materials may have come some, but probablynot great, distance from a volcanic center; if by volcanicmudflow, the deposit could represent part of thevolcano structure itself.

ANALYTICAL METHODSMajor-element chemical analyses were made by Paul

L.D. Elmore using rapid-rock analytical methodsdescribed in U.S. Geological Survey Bulletin 1144-A,supplemented by atomic-absorption analysis. Minor-element abundances were determined by a variety ofmethods. Th and U were determined by H.T. Millard,P.J. Aruscavage, A. Bartel, and R.A. Zielinski using adelayed-neutron system. S was determined by L.Shapiro using the method of solubility in aqua regia. Auand Ag were analyzed by L.J. Schwarz by neutron-activation methods using fire assay for the radio-chemical separations as described in U.S. GeologicalSurvey Circular 599. Ce, Cr*, Eu, Hf, Nd, Sb, Ta, Th*,and Yb* were determined by L.J. Schwarz by neutron-activation methods. The reported radioactivation and

822

VOLCANIC ROCKS

2 cm

2 cm

Figure 2. Volcanic conglomerates from Site 264. Sub-bottom depth 111 meters, near top of volcanic unit(Core 11, Section 2). Largest two pebbles are moderateyellowish-brown (10YRS5/2), highly altered probable an-desite. Dark pebbles in upper center and lower left cor-ner are dark yellowish-brown (10YR4/2) altered and de-vitrified glass with abundant albitized microlites. Matrixin general is moderate yellowish-brown (10YR5/4) withlighter calcite-rich patches. Light veinlets are mostly cal-cite. Subbottom depth 206 meters (Sample 14, CC).Large pebble on right is pale brown (5YR5/2) highly al-

tered probable andesite with a few albitized plagioclasephenocrysts. Large clast on left is grayish olive-green(5GY4/2) altered probable andesite containing a fewchlorite- and calcite filled amygdules. Matrix in center isdusky yellow-green (5GY5/2) and is composed mostlyof secondary calcite and altered volcanic granules, sandand silt, and rare fragments of detrital biotite.

radiochemical results are averages using differentphotopeaks and/or several counting runs; those in twoor more significant figures including at least the tenthsdigit have an overall accuracy of ± 10%, and the othersare assumed to have an accuracy to ±50% owing to lowcounting rate, heavy interference, and/or concentrationbelow detection limit. All other minor elements weredetermined by R.E. Mays using quantitative spectro-graphic analysis. Spectrographic results are reported totwo significant figures and have an overall accuracy of±15% except that they are less accurate near the limitsof detection where only one digit is intended. Cr, Th,and Yb, in different sample splits, were analyzed byboth neutron activation (indicated by *) and spectro-graphic or other methods.

PETROLOGY AND GEOCHEMISTRY

Alteration EffectsThe igneous rocks from Site 264 are generally highly

but variably altered as shown by their common reddishto yellowish-brown color, the abundance of secondaryminerals and replacement textures seen in thin section,and such chemical indicators as generally high oxidationindex [100×Fe2O3 / 7Fe2O3 + FeO)] and high H2O andCO2 content. Most rocks, though highly altered, containrelict minerals and textures that clearly indicate theirvolcanic parentage. The uncertain nature andmagnitude of chemical transfer and oxidation duringalteration obscures the nature of the original rock andmagma types, as might be inferred directly fromchemical analyses.

Alteration doubtless includes the effects of low-temperature submarine alteration (halmyrolysis) in-volving seawater interactions. In some rocks, theseeffects may be superimposed on higher temperaturedeuteric or metamorphic effects suggested by thepresence of greenschist-facies assemblages includingalbite, epidote, and chlorite. As some rocks apparentlywere formed above sea level, effects of subaerialweathering may be present.

Submarine alteration results in critical chemicalchanges that accompany mineralogic alteration. Underconditions of high CO2 activity and free access of sea-water, particularly under a cover of carbonate sedi-ments as at Site 264, mafic minerals are attacked fasterthan plagioclase (Bass et al., 1973). Loss of Ca is chieflyfrom pyroxenes. In the Site 264 rocks destruction ofprimary mafic minerals is complete, and apparently atleast some released Ca has combined with CO2 to formthe locally abundant carbonate minerals. Plagioclase isvariably replaced by albite, forming secondary patchyand normal zoning; in some rocks, plagioclase appearsto be little altered. Albitization suggests the gain of Na

823

A. B. FORD

and K. Chemical changes may include changes inminor-element content, even in the so-called "re-fractory" elements such as Zr, P, and Ti, by the destruc-tion of the mafic minerals in which they were mainlycontained. Sr and Sc, whose content is highest in the Site264 rocks with the freshest plagioclase, appear to havebeen lost during albitization. Owing to the wide com-positional range from probable rhyolitic to andesitic ormore mafic rock, however, minor-element variation inthis suite cannot be uniquely tied to alteration alone.According to Bass et al. (1973), the usual net chemicalchange produced by submarine alteration, at least inmafic rocks, is an increase in the apparent "alkalinity"of the major elements and a decrease in that of the traceelements.

Volcanic Conglomerate

The two recovered samples of volcanic conglomerateconsist of subangular to subround polymictic volcanicpebbles and granules embedded in a yellowish-brown toyellowish-green, highly altered, clayey to sandy or tuf-faceous matrix (Figure 2). Matrix and pebbles are exten-sively replaced by calcite, to a lesser extent by quartz,chlorite, clays, and iron oxides, and, rarely, by epidote.Calcite and quartz veinlets intricately cut the matrixand, locally, the pebbles. Plagioclase microlites and rarephenocrysts in the pebbles are almost entirely albitizedand show only ragged interior patches of calcicplagioclase. Ca was contributed to form the calcite andminor epidote. Alteration masks the original mafic min-erals in the pebbles and no identifiable pseudomorphsare present. The pebbles appear, by their color and lackof quartz phenocrysts, to be highly altered andesites.The pebbles are much more highly altered than most ofthe single-lithology clast fragments described below andthis may account for preservation of matrix in the con-glomerate.

The high degree of pebble and granule rounding seenin the specimens of Figure 2 suggests transport by water.However, the lithologic character, apparently entirelyvolcanic, and the extremely poor sorting of the materialsuggests transport by volcanic mudflow, in which casethe rocks could actually represent part of the volcaniccomplex itself rather than a transported-sediment de-posit. A postdepositional age for at least some of the al-teration is indicated by the calcite veinlets cutting bothmatrix and pebbles (Figure 2).

Other Volcanic Rocks

Most of the rock fragments recovered from Site 264show a single lithology, and, owing to ambiguity in in-terpreting their occurrence in the volcanic deposit, aredescribed separately from the apparent conglomeraticrocks. Seven specimens (determined by thin-sectionstudy to include the principal variants) were selected forwhole-rock chemical analysis. Chemistry of the rocks ishighly varied; the range in SiC>2, for example, is about46%-72% (Table 1). Although the freshest appearingsamples were selected, alteration has doubtless affectedto some degree the bulk chemistry of major and minorelements in all rocks. Attempts to evaluate these changesare probably illusory but seem to show that the rocks

have taken on more alkaline characteristics. Owing totheir generally fine grain size and extensive alteration,the rocks are classified on the basis of bulk chemistryrather than petrography.

Brief petrographic descriptions follow the analyses ofTable 1. Most rocks are porphyritic. The Siθ2-rich rocksare devitrified glasses showing well-developed perliticfracturing and shard-like structures (Figure 3) thatclosely resemble features of glassy welded tuffs il-lustrated in Ross and Smith (1961). As the samplesanalyzed are altered, the powder densities, measured bypycnometer, do not correspond very closely withaverage densities of the rock types indicated, particular-ly of the more mafic rocks. The rocks are classified ac-cording to Streckeisen (1967), but owing to commonlyextensive alteration their nomenclature is somewhat un-certain. Nomenclature problems of rocks in theandesitic field are discussed in detail below.

The possible effects of alteration involving chemicaltransfer on the calculation of norms is evident—for ex-ample, the transfer of alkalis inward and the expulsionof Ca. Such changes, though possibly extensive, cannotbe accurately assessed here because no fresh rocks areavailable for comparison. The generally high content ofH2O and CO2 is probably not a primary characteristic ofthese rocks, but the amount that has been added is un-known. Fe oxidation can be a highly significant factoraffecting normative mineralogy (Coombs, 1963), par-ticularly in rocks as highly oxidized as these. Oxidationindexes (Table 2) for all rocks are high, generally muchhigher than 40, which, at least for submarine-weatheredbasalt, is considered by Cann (1971) to be the limitabove which alteration is accompanied by majorchanges in element abundances. In an attempt toevaluate at least some of the possible changes duringalteration, CIPW norms have been calculated from bothoriginal analyses and those adjusted by normalizing freeof volatiles and recalculating Fβ2θ3 in excess of 1.5% asFeO (Table 3). The results indicate a large range of un-certainty in original normative compositions. This rangeof uncertainty, indicated by the lines connecting unad-justed compositions to those adjusted for secondaryalterations (Figure 4), chiefly results from Fe oxidation.

All rocks from Site 264 are Q normative before andafter adjusting analyses. The chief effects of reducing Feoxidation are a decrease in Q and an increase inHy(Table 3). Although normative C, which indicates amolar excess of AI2O3 over Na2θ + K2O + CaO, isoften suggestive of altered material, the small amountsin these rocks are not uncommonly high for freshandesites (Chayes, 1969). Adjusted analyses (Table 3)show loss of C in several rocks; disappearance of C ac-companies the appearance of Di in the two samples withthe least Siθ2.

Normalized amounts of Q, Or, and Ab + An plottedin Streckeisen's (1967) ternary classification of volcanicrocks indicate that the Site 264 rocks range widely fromrhyolite to andesite (Figure 4). Adjustment for altera-tion results in migration of plots away from the Q cornertoward the feldspar join. The amounts of migration,however, are insufficient to affect the rock nomen-clature. All Ab is arbitrarily allotted to plagioclase; if

824

VOLCANIC ROCKS

TABLE 1Major-Element Analyses of Igneous Rocks at Site 264

11, CC 12,CCC 13-1,1' 13-1, 3C 13-1,4' 13,CCα) 15,CCë

Plot symbol B D G

SiO2

A12°3Fe2°3FeO

MgO

CaO

Na2OK2°H2O+H2°"TiO2

P2°5MnO

co2Total

47.217.8

10.7

0.52

1.6

8.6

3.2

1.0

2.3

2.0

2.1

0.27

0.14

2.3

99.7

45.817.1

9.6

0.60

1.3

10.2

3.0

1.1

2.1

1.6

2.1

0.26

0.12

4.9

99.8

72.012.1

3.0

0.84

0.56

0.24

0.86

7.7

1.4

0.46

0.52

0.06

0.04

0.03

99.8

57.412.2

8.4

2.6

1.8

3.5

2.4

3.8

3.0

1.3

1.5

1.2

0.17

0.70

100.0

70.312.8

3.6

1.3

0.76

0.86

2.3

5.2

1.4

0.45

0.67

0.16

0.09

0.01

100.1

54.815.0

6.9

3.8

3.1

2.5

2.6

3.0

3.9

2.1

1.9

0.30

0.09

0.03

100.0

54.016.3

4.3

4.9

4.7

5.6

2.3

1.9

2.8

1.3

1.5

0.25

0.11

0.04

100.0

Rock name

Density

Color

Andesiticrock

2.69

Lightbrown

(5YR5/5)

Andesiticrock

2.73

Lightbrown

(5YR5/5)

Rhyolite

2.59

Mediumgray(N5)

Rhyodacite

2.64

Mediumdk gray

(N4)

Rhyodacite

2.66

Mediumdk gray

(N4)

Rhyodacite

2.66

Dk green-ish-gray

(5GY4/1)

Andesiticrock

2.69

Mediumgray(N5)

Note: Plot symbols are used in Figures 4-8.aHolocrystalline, nonporphyritic rock containing about 50% little-altered zoned plagioclase, 5% magnetite,

and 45% alteration products. Plagioclase forms fine-grained laths, locally very weakly aligned, generallybetween 0.2 and 0.5 mm long. It is mostly sodic labradorite, AN5j_55, of intermediate ordering degree(Figure 7), commonly zoned to more albitic rims of about AN27-38- Alteration products are chiefly trans-lucent hydrous iron oxides in veinlets, films, and patches, with lesser calcite, chlorite, and serpentine-likeminerals and traces of epidote, that pervade entire rock. Primary mafic minerals entirely altered. Probablya flow rock but could be hypabyssal.

"Lighology similar to Sample 11, CC.cSlightly altered, devitrified, vitrophyric glass showing fine lenticular and fragmental structures having

shard-like appearance. Perlitic fractures well developed. Rare patches of nondevitrified, pale brown glass.Contains an estimated 5% phenocrysts, some cumulophyric, of altered albite (An4> to about 2.5 by 1.5mm in size. Embayments around phenocryst margins are filled by mesotasis material, which also occurs insmall, rounded inclusions, indicating resorption. Alteration products include mainly translucent ironoxides, minor chlorite, and traces of epidote, the epidote in albite phenocrysts. The mode includes severalpercent of magnetite, probably primary. Coarse layered structure is conspicuous; several large, indistinctmasses appear to be collapsed pumice fragments. Structures show evidence of compaction against pheno-crysts. Shards show evidence of plasticity and probable welding. The rock is a rhyolitic probable welded tuff.

dDevitrified, nearly opaque glass containing abundant lithophysae to about 2 mm across, consisting ofgranular quartz and minor opaque oxides. Rare blocky phenocrysts of albitized plagioclase (An5) to about1.5 by 1.0 mm in size, set in very dark, turbit, microlite-bearing matrix of devitrification products. Cut byveinlets to 0.1 mm wide of quartz, calcite, chlorite, epidote, and opaque oxides. No evidence of shardstructure. Probably a flow rock.

e Altered, devitrified vitrophyre, similar to but much more highly altered and recrystallized than Sample13-11. Contains about 10% albitized blocky plagioclase phenocrysts with patchy relict cores of about An4 η.Contains several percent of lithophysae consisting of rounded masses of granular quartz set in a highly tur-bid recrystallized matrix of fine quartz and alkali feldspar. Vague, ghost-like, fine lenticular structures rβrsemble shards of Sample 13-11. Local faint concentric structures appear to be relict perlitic fractures. Therock is probably an altered rhyodacitic welded tuff.

1 Highly altered fine-grained holocrystalline rock with rare albitized plagioclase phenocrysts to about 3.0 by1.5 mm. Turbid-appearing groundmass composed of albitized plagioclase laths to about 0.8 by 0.1 mm insize, quartz, chlorite, iron oxides, and rare epidote. Epidote occurs in albitized phenocrysts also. Norecognizable mafic-mineral pseudomorphs, and no evidence of original glass. Probably a flow rock.

SLighology similar to Sample 13, CC(1).

825

A. B. FORD

500µ

Figure 3. Slightly altered, devitrified, vitrophyric glass of Sample 264-13-1, 1 cm showing periltic cracks andflu-idal shard textures. Plane light.

TABLE 2Selected Characteristics of Rocks from Site 264

NormativeAn percent

(plagioclase)

OpticalAn percent

(plagioclase)

Oxidationindex

FeO (total)Mgo

Differen-tiationindex

Maficindex

11,CC

49.3

53

95.4

6.3

45.0

87.5

12, CC

41.4

53

94.1

7.1

47.2

88.7

13-1,1

7.7

4

78.1

6.3

89.8

87.3

13-1,3

20.1

5

76.4

5.6

67.7

85.9

13-1,4

14.0

47

73.5

6.0

85.2

86.6

13,CC(1)

31.8

64.5

3.2

60.1

77.5

15, CC

57.1

40

46.7

1.9

45.0

66.2

partial allotments were made to the orthoclase corner,varying according to the degree of albitization in eachsample, several would appear to be considerably morealkalic. Nomenclature would generally not be affectedexcept that the rhyolite (c in Figure 4) might therebybecome "alkali rhyolite," probably an inappropriateterm as the suite of rocks as a whole seems to have sub-alkaline characteristics. The differentiation index

(Thornton and Tuttle, 1960) ranges from about 45 forandesitic rocks to near 90 for rhyolite (Table 2).

Nomenclature of the three rocks in the plagioclasecorner of the ternary classification (Figure 3) is uncer-tain. Different criteria could class them as eitherandesitic or basaltic. The unusually low SiCh content ofSamples 11, CC and 12, CC is in the characteristic rangefor basalt, but in both samples recalculation on a

826

VOLCANIC ROCKS

TABLE 3CIPW Weight Norms of Analyzed Rocks from Site 264

11, CC 12, CC 13-1,1

Norms Calculated From Unadjusted Oxide

Q

C

Or

Ab

An

En

F S jMt

Hm

11

Ru

Ap

CC

• (Hy)

11.2

1.8

6.0

27.7

26.9

4.0

-

10.9

1.4

1.3

0.65

5.3

14.7

4.4

6.6

25.8

18.2

3.2

-

9.7

1.5

1.3

0.62

11.3

36.6

2.1

45.7

7.3

0.61

1.4

1.3

2.0

0.99-

0.14

0.06

13-1,3

s

24.3

2.3

22.7

20.5

5.1

4.5

4.6

5.3

2.8

—

2.8

1.6

13-1,4

34.7

2.2

30.8

19.5

3.1

1.9

2.5

1.8

1.2

-

0.38

0.02

Norms Recalculated After Adjusting Analyses for Alteration

Q

cOr

Ab

An

Di

Hy

Mt

11

Ap

' Wo

En

L F s

En

Fs

1.5

-

6.4

29.3

33.8

4.3

1.2

3.4

3.2

9.3

2.3

4.3

0.69

0.67

-

7.1

28.0

33.1

8.8

2.2

7.0

1.3

4.1

2.4

4.4

0.68

36.2

2.1

46.5

7.4

0.81

—

1.4

2.0

2.2

1.0

0.14

17.2

0.68

23.8

21.5

10.1

—

4.7

13.5

2.3

3.0

3.0

33.5

2.2

31.4

19.8

3.2

-

1.9

3.7

2.2

1.3

0.38

13,CC(1)

19.5

3.7

18.1

22.4

10.4

7.8

7.1

2.0

3.6

-

0.72

0.07

14.6

3.9

18.9

23.5

11.1

-

8.2

12.5

2.3

3.8

0.76

15, CC

13.8

0.9

11.3

19.7

26.2

11.8

3.2

6.3

-

2.8

-

0.60

0.09

10.9

0.91

11.7

20.3

27.3

-

12.2

10.5

2.2

2.9

0.62

volatile-free basis raises SiCh to a value slightly ex-ceeding 50%. Optical composition of plagioclase inSamples 11, CC and 12, CC (Table 2; Figure 5) and nor-mative An in Sample 15, CC (Table 2) exceed Anso, arequirement for basalt in some classifications. However,sodic labradorite is not uncommon in andesitic rocks(Chayes, 1969; Streckeisen, 1967). All three rocks con-tain abundant Q before norm recalculation from ad-justed analyses, and at least minor Q after. Q amountsare within the range reported in andesite, and averagenormative color index approximates the average valueof 21.3 for Cenozoic andesite (Chayes, 1969). Minor-element abundances (Table 4) more closely approxi-mate andesitic than basaltic compositions, particularlyNi and Co and the ratio Ni/Co as reported by Fleischer(1968). Although uncertainties exist, the three rocks arehere classed as andesitic chiefly on the basis of their un-usually low MgO (Thompson, 1973) and low color index(Streckeisen, 1967). Streckeisen subdivides the andesiticfield at 10% Or into trachyandesite, which would includesamples 12, CC and 15, CC, and true andesite, whichwould include only Sample 11, CC. The termtrachyandesite seems inappropriate for rocks with theselow K2θ/Na2θ ratios.

Secondary chemical changes possibly include in-troduction of K during hydration and devitrification ofthe siliceous glasses. High K2θ/Na2θ ratios of rhyoliteand rhyodacite (Table 1) may have resulted from Kmetasomatism during alteration of glass (Lipman,1965).

The subalkaline character of the volcanic suite fromSite 264 is displayed in the alkali-silica plot (Figure 6)which shows, for comparison, the alkalic and subalkalicfield dividers of Macdonald (1968), based on oceanictholeiitic and alkalic suites, and of Irvine and Baragar(1971), which includes calcalkaline suites. The fields ofIrvine and Baragar are more appropriate for the rockshere described. Adjusting analyses for alteration reducestheir apparent "alkalinity" in this diagram.

The suite of volcanic rocks from the NaturalistePlateau shows characteristics that appear to belongmore to the calcalkaline suite than to the tholeiitic suite.Although considerably scattered, most samples plot onan FMA diagram (Figure 7) within the calcalkaline fieldof Irvine and Baragar (1971, fig. 2). A typicalcalcalkaline suite, however, should fall in an almoststraight band from near midway of the FeO-MgO jointoward the alkali corner (Green and Ringwood, 1968).

827

A. B. FORD

Ab+An65 90

Figure 4. Plots of normative Q, Or, and Ab + An in vol-canic rocks from Site 264, using the ternary classifica-tion of Streckeisen (1967). Compositions of unadjustedanalyses shown by open circles. Lines connecting unad-justed compositions to those adjusted for secondary al-terations show the probable uncertainty of original rockcompositions (see text discussion). Letter symbols in-dicate sample number in Table 1.

The displacement here toward the FeO corner suggestsFe enrichment, a characteristic feature of a tholeiiticseries (Jakes' and Gill, 1970). Whether the apparent Feenrichment is the result of accumulative processes is un-known, but no textural evidence for this is preserved inthe altered, Fe-rich rocks. Jakes' and Gill (1970) dis-tinguish tholeiitic and calcalkaline series in island-arcvolcanic suites; plots in their fig. 2 of K2O against S1O2for Site 264 rocks are clearly more calcalkaline thantholeiitic. The rocks, however, have unusually highmafic indexes [100 × (FeO + Fe2θ3) / (FeO + Fe2O3 2MgO), in weight percent] for a calcalkaline series (seeTable 2). Plots, not here shown, of the ratio FeO(total)/MgO against Siθ2 in Miyashiro (1974, fig. 1)suggest that the Site 264 rocks belong to an island-arctholeiitic series.

The possible calcalkaline affinity of this volcanic suiteappears to be indicated by relations of total alkalis andCaO to Siθ2 in the diagram of Peacock (1931) shown inFigure 8. Like Figure 6, Figure 8 shows reduction in ap-parent "alkalinity" when analyses are adjusted foralteration. For unadjusted analyses, the peacock"alkali-lime index"—the Siθ2 value at which Na2θ +K2O = CaO in a suite of related rocks—is about 53 inthe alkali-calcic field of Figure 8; this value increases toabout 57 in the calcalkaline field, when analyses arerecomputed on a volatile-free basis. The sharp increasein K2θ/Na2θ ratios to greater than unity in Site 264rocks more silicic than andesite is a common feature ofthe calcalkaline series (Green and Ringwood, 1968). Us-ing different criteria, therefore, these rocks might becategorized either as calcalkaline or tholeiitic on thebasis of major oxides.

Minor-element concentrations, given in Table 4, showconsiderable scatter in plots (not here presented) againstmajor oxides. Variation of minor-element concen-trations in the suite is a consequence of majordifferences in original rock types and of secondaryalterations, but the contribution of either factor alonecannot be assessed in the absence of fresh material. Asmight be expected, high U and Zr are associated withhigh S1O2 whereas high Ni and Cu correlate with lowS1O2. Ni and Cu do not seem closely related to MgO.High Sc and Sr are apparently related to abundance oflittle-altered plagioclase. Except for high Cr in Samples11, CC and 12, CC, the minor-element concentrationsdo not differ greatly from those of calcalkaline volcanicsuites tabulated by Taylor (1969).

Plagioclase compositions were determined opticallyby the universal-stage method of Uruno (1963), in whichKohler's angles XAX and Y Y are measured across(010) on stereogram plots of optical directions of adjoin-ing albite twins. After determining compositions andaverage ordering degree from several such measure-ments (Figure 5), additional determinations were madeby the faster method of rotating the crystal axis aparallel to the microscope axis and using, or inter-polating between, the appropriate high- or low-T extinc-tion curves of Tobi (1963). Optical plagioclase com-positions listed in Table 2 are average determinations ofthe major parts of crystals and do not include zonedrims or groundmass plagioclases, which are usuallymuch more sodic. They therefore do not correspondclosely with calculated normative compositions.

INTERPRETATION

Character of the Basement

The Naturaliste Plateau is considered by Johnstone etal. (1973) to be a submerged part of the continentalLeeuwin-Naturaliste Block, about 670 m.y. old, ofsouthwestern Australia, and is separated from theArchean Yilgarn Block to the east by the Darling fault.Continental-type crystalline rocks dredged from thenorth scarp of the plateau (Heezen and Tharp, 1973)support the suggestion of a continental foundation.

The plateau also appears to be a physiographic exten-sion of Broken Ridge in the Indian Ocean to the westand conceivably could share similarities of basementstructure with that feature. The nature of Broken Ridgeis controversial. Whereas Dietz and Holden (1971) con-sider it to be of simatic origin, McKenzie and Sclater(1971) suggest, on the basis of seismic characteristics,that it is a continental fragment once joined toKerguelen Plateau at an edge of some part ofGondwanaland, postulating that Broken Ridgeseparated from Kerguelen Plateau, along with Australiafrom Antarctica, by Eocene spreading from the South-east Indian Ridge. Recent studies, however, indicatethat Kerguelen is a simple oceanic island consistingmainly of tholeiitic to alkalic basalt and trachyte with noevidence of any continental materials (Watkins et al.,1974). If Broken Ridge is indeed related to Kerguelen, ittherefore is more likely to be oceanic than continental incharacter.

828

VOLCANIC ROCKS

80

Figure 5. Optical plots of plagiocalse in Samples 11, CCand 12, CC showing composition and intermediate ordering degree indiagram of Uruno (1963). Vertical axis, XAX Kohler angle; horizontal axis, YAY Kohler angle. An content shown on soliddiagonal lines.

The term "basement" is relative. The volcanicmaterials described can be considered to make up thebasement of the section of Cenozoic carbonatesediments on this part of Naturaliste Plateau and theyform the acoustic basement in seismic profiler records.Direct examination of these basement rocks indicatesthat they have chemical and petrographic charac-teristics much more like continental or island-arc vol-canic rocks than oceanic rocks. For example, rhyolitesof oceanic regions, unlike the Site 264 rhyolitic rocks,commonly have pantelleritic compositions (McBirney,1969). Drilling at Site 264, however, penetrated thevolcanic deposit only superficially without reaching thebasement on which it in turn rests. The sampled sectioncontains no apparent debris reworked from this base-ment, the nature of which can only be inferred asprobably continental on the basis of the nature of theoverlying volcanic rocks. Although such an inferencemight be invalid if the volcanoclastic debris came from a

distant volcanic source, apparent lack of mixing withnonvolcanic detritus suggests a local source. Moreover,the interpretation of a continental, including island-arc,foundation is supported if this deposit consists largely ofmudflow breccias and tuffs representing part of thevolcano structure itself as appears possible from theavailable evidence.

Setting in Eastern Indian Ocean and the Question ofSource Terrain

The recovery of Cretaceous or older subalkalineandesitic volcanic rocks on Naturaliste Plateau is signifi-cant because such rocks have not been known in thearea. Consisting of the basalt-andesite-dacite-rhyoliteseries, the andesitic suite characterizes volcanism inisland arc and continental margin orogenic zones(Green and Ringwood, 1968) that are now commonlypostulated to be related to converging plate margins(Mitchell and Reading, 1971). If these generalizations

829

A. B. FORD

TABLE 4Minor Elements in Site 264 Volcanic Rocks

AgAu a

BBaCe

Co

Cr b

CuEu

Hf

LaNd

NiPb

SSb

ScSr

Ta

ThT h b

U

VY

Y \Yb b

Zr

11, CC

0.29,0.320.3, 0.3

N20

7027.0

32

300243

801.6

3.8

N4022

42

N20

0

1.144

3000.4

NA10.33

200

50

2

3.3

140

12, CC

<O.Ol0.6

N207021.5

32

340251

801.6

3.6N40

14

36

N20

0

0.544

3000.6

NA10.29

180

50

2

3.3

120

13-1, 1

0.20, 0.220.3, 0.6

N20900134.4

58

1422

2.5

12.6120

63

N2

N20

00.2

4

34

1.0

10.910.8

0.7822

60

36.4

300

13-1,3

0.170.4

N20900

84.8

18

302611

3.6

7.0

N4051

N2

N20

3000.4

16120

0.6

3.034.60.74

110

110

6

1.8

200

13-1,4

0.160.8

N201400

125.8

8201513

2.7

14.0100

59N2

40

1000.2

6

1201.0

NA12.2

NA

2060

35.8

300

13,CC(1)

0.03, 0.030.3, 0.2

N20320

60.4

285534

131.7

5.8N40

3411

N20

100

0.324

1200.5

NA6.8

NA

180

60

33.9

170

15, CC

0.130.3

N20280

38.2

19120

7524

1.4

3.8

N402214

N20

5000.2

32

2400.4

4.833.60.54

180

50

2

3.0

130

Note: See text for analysts and methods. N = not detected at value shown; NA = notanalyzed.

All values in ppm except for Au, which is given in ppb.

Analysis by neutron activation.

, 4

~\ r

Irvine and Baragar (1971)—//

50 60SiO

2 (weight percent)

70

Figure 6. Total alkalis (Na2θ and K2O) - silica plot, inweight percent, of volcanic rocks of Site 264. For ex-planation of plot symbols, see Figure 4.

Figure 7. FMA diagram, in weight percent, showing plotsof volcanic rocks at Site 264. For sample represented byletter, see Table 1.

830

VOLCANIC ROCKS

12-

4 •

2 0

alkalic

XB

X \<?<9

"V.

• ~~B A

alkal i-calcic

|

calc-alkalic

calcic

r-adjusted ~~"——-—____ E cI i i i •X

44 50 53 57 60

SiO2 (weight percent)

70

Figure 8. Variations of weight percent CaO and alkalis re-lated to Siθ2 in the volcanic rock suite from Site 264,indicating alkali-lime index (see text) of about 53. Aparallel but different set of curves, not shown, indicatesan index of about 57 for oxide amounts normalizedvolatile-free. Subdivisions of alkalic to calcic fields arethose of Peacock (1931). For sample represented by let-ter symbol, see Table 1.

are valid and if the rocks recovered from Site 264 areassumed to represent part of the actual volcano com-plex, then this area might be part of an old island-arcsystem related to an early stage of eastern Indian Oceandevelopment. If the volcanic debris consists of mostlytransported material, then this area at the time of ac-cumulation probably lay near an island-arc or continen-tal andesitic complex.

If the source must be sought elsewhere, the calc-alkaline andesitic fingerprint should identify the ter-rain. The nearest onland volcanic rocks of possibly simi-lar age belong to the Bunbury Basalt, dated originally byEdwards (1938) as Tertiary but now believed to beNeocomian (Johnstone et al., 1973). The restrictedrange of tholeiitic basalt compositions of the Bunbury(Edwards, 1938) rules it out as a potential source, atleast for the section sampled at Site 264. Recent offshoredrilling at the northern end of Perth Basin penetrated athick volcanic pile consisting chiefly of rocks of ap-parent alkalic compositions (E. M. Kemp, written com-munication, 1973); its age is unknown and com-positions of the rocks are probably unlike those at Site264.

Possible source terrains, even if originally near, mightbe distant if the McKenzie and Sclater (1971) concept ofEocene fragmentation of this part of Gondwanaland isvalid. All central and eastern Indian Ocean coastlines,according to Fairbridge (1965), are of the "Atlantic" orpassive type except for the arc-trench system adjoiningSoutheast Asia. Reconstructions of these parts ofGondwanaland are highly varied (Figure 9), except forthe generally accepted, simple meridional closure of thesoutheast Indian Ocean normal to its axial ridge. Thesoutheast Indian Ocean closure juxtaposes thesouthwest corner of Australia, and in most reconstruc-tions presumably the Naturaliste Plateau, with westernWilkes Land, Antarctica, an area near known centers ofTertiary alkalic volcanism (Watkins et al., 1974) butwhere Cretaceous or older calcalkaline volcanism is notknown. Southwest Australia and presumably theNaturaliste Plateau in the Dietz and Holden (1971)

Old sea f loor^>

Antarctica

Figure 9. Diagrammatic sketches showing suggested latePaleozoic reconstructions. Ridd (1971); Dietz and Hol-den (1971); Veevers et al. (1971). Dotted lines in a and c1000-meter isobath. C=Ceylon; NP=Naturaliste PlateauCross-hatched area in a, overlap of Southeast Asia withAustralia.

reconstruction would lie between Antarctica and aregion of old sea floor (Figure 9b), but the recent dis-covery of the young age of Wharton Basin (von derBorch, Sclater, et al., 1974) argues against this possibili-ty. Volcanic sources that would support Veevers et al.(1971) in their placement of the Naturaliste Plateaubetween Ceylon, India, and Western Australia (Figure

831

A. B. FORD

9c) are not known. Moreover, this placement seems un-likely if the Site 264 volcanic rocks are relatable to anandesitic island-arc belt along a convergent plate bound-ary.

The nearest known region of island-arc andesiticvolcanism is Southeast Asia. The Site 264 volcanicmaterials might support Ridd's (1971) fitting of south-western Australia near Borneo (Figure 9a). His inclu-sion of Southeast Asia as part of Gondwanaland, basedon stratigraphic similarities and the fragmentarypresence of a Glossopteris flora, is strongly questionedby Griffiths and Burrett (1973) on the basis of majordifferences in fauna and paleoclimate. The geometric fitof Southeast Asia to Western Australia, moreover, is ex-tremely poor and it seems fortuitous that all of theorogenic volcanic arc fragmented away, leaving no traceon Australia or India. And, if the Naturaliste Plateauwere part of a volcanic arc adjoining southwestAustralia, a record of this should be seen in Australianrocks. The absence of any such known record suggests atransform fault separating these two regions, but howsuch a conjectural fault relates to nearby oceanic frac-ture zones is unknown. The results of drilling at Site 264suggest that early chapters in eastern Indian Oceanhistory include a phase of convergent plate activity in aregion heretofore considered only as extensional by sim-ple rifting of Mesozoic Gondwanaland. However, theearly record is exceedingly fragmentary. How this possi-ble convergent plate activity, which is suggested by theandesitic compositions, fits into the regional tectonichistory is not now known.

ACKNOWLEDGMENTSI am indebted to all fellow shipboard members of Leg 28 of

D/V Glomar Challenger, including ship's crew, drilling crew,DSDP technical staff, and scientific staff, and, for shore-basedstudies, to analysts of the U.S. Geological Survey. In par-ticular, I am indebted to P.J. Barrett, who helped on ship-board studies of the Site 264 volcanic rocks, and to R.L.Christiansen, who made valuable comments on themanuscript.

REFERENCESBass, M.N., Moberly, R., Rhodes, J.M., Shih, Chi-yu, and

Church, S.E., 1973. Volcanic rocks cored in the centralPacific, Leg 17, Deep Sea Drilling Project. In Winterer,E.L., Ewing, J.I., et al., Initial Reports of the Deep SeaDrilling Project, Volume 17: Washington (U.S. Govern-ment Printing Office), p. 429-503.

Cann, J.R., 1971. Major element variations in ocean-floorbasalts: Roy. Soc. London Phil. Trans., ser. A., v. 268, p.495-505.

Chayes, F., 1969. The chemical composition of Cenozoicandesite. In McBirney, A.R., (Ed.), Proc. AndesiteConf.: Oregon Dept. Geol. Min. Indus. Bull. 65, p. 1-11.

Coombs, D.S., 1963. Trends and affinities of basaltic magmasand pyroxenes as illustrated on the diopside-olivine-silicadiagram: Mineral. Soc. Am. Spec. Paper 1, p. 227-250.

Davies, T.A., Luyendyk, B., et al., 1974. Initial Reports of theDeep Sea Drilling Project, Volume 26: Washington (U.S.Government Printing Office).

Dietz, R.S. and Holden, J.C., 1971. Pre-Mesozoic oceaniccrust in the eastern Indian Ocean (WhartonBasin)?: Nature, v. 229, p. 309-312.

Edwards, A.B., 1938, Tertiary tholeiitic magma in westernAustralia: J. Roy. Soc. Western Australia, v. 24, p. 1-12.

Fairbridge, R. W., 1965. The Indian Ocean and the status ofGondwanaland. In Sears, M. (Ed.), Progress inoceanography, v. 3: New York (Pergamon Press), p. 83-136.

Fleischer, M., 1968. Variation of the ratio Ni/Co in igneousrock series: J. Washington Acad. Sci., v. 58, p. 108-117.

Green, T.H. and Ringwood, A.E., 1968. Genesis of thecalcalkaline igneous rock suite: Contrib. Mineral. Petrol.,v. 18, p. 105-162.

Griffiths, J. and Burrett, C, 1973. Were South-east Asia andIndonesia parts of Gondwanaland?: Nature Phys. Sci., v.245, p. 92-93.

Heezen, B.C. and Tharp, M., 1973. USNS Eltanin Cruise55: Antarctic J. U.S., v. 8, p. 137-141.

Irvine, T.N. and Baragar, W.R.A., 1971. A guide to thechemical classification of the common volcanicrocks: Canadian J. Earth Sci., v. 8, p. 523-548.

Jakes, P. and Gill, J., 1970. Rare earth elements and the islandarc tholeiitic series: Earth Planet. Sci. Lett., v. 9, p. 17-28.

Johnstone, M.H., Lowry, d.C, and Quilty, P.G., 1973. Thegeology of southwestern Australia—a review: J. Roy. Soc.Western Australia, v. 56, p. 5-15.

Lipman, P.W., 1965. Chemical comparison of glassy andcrystalline volcanic rocks: U.S. Geol. Surv. Bull. 1201-D.

Macdonald, G.A., 1968. Composition and origin of Hawaiianlavas: Geol. Soc. Am. Mem. 116, p. 477-522.

McBirney, A.R., 1969. Andesitic and rhyolitic volcanism oforogenic belts. In Hart, P.J. (Ed.), The earth's crust and up-per mantle: Am. Geophys. Union, Geophys. Monogr. 13,Washington, p. 501-507.

McKenzie, D. and Sclater, J.G., 1971. The evolution of the in-dian Ocean since the Late Cretaceous: Roy. Astron. Soc.Geophys. J., v. 25, p. 437-528.

Mitchell, A.H. and Reading, H.G., 1971. Evolution of islandarcs: J. Geol., v. 79, p. 253-284.

Miyashiro, A., 1974. Volcanic rock series in island arcs and ac-tive continental margins: Am. J. Sci., v. 274, p. 321-355.

Peacock, M.A., 1931. Classification of igneous rock series: J.Geol. v. 39, p. 54-67.

Ridd, M.F., 1971. Southeast Asia as a part ofGondwanaland: Nature, v. 234, p. 531-533.

Ross, C.S. and Smith, R.L., 1961. Ash-flow tuffs: their origin,geologic relations, and identification: U.S. Geol. Surv.Prof. Paper 366.

Streckeisen, A.L., 1967. Classification and nomenclature of ig-neous rocks: Neues Jahrb. Mineral. Abhandl., v. 107, p.144-214.

Taylor, S.R., 1969. Trace element chemistry of andesites andassociated calc-alkaline rocks. In McBirney, A.R. (Ed.),Proc. Andesite Conf., Oregon Dept. Geol. Min. Indus.Bull. 65, p. 43-63.

Thompson, R.N., 1973. One-atmosphere melting behaviourand nomenclature of terrestrial lavas: Contrib. Mineral.Petrol., v. 41, p. 197-204.

Thornton, C.P. and Tuttle, O.F., 1960. Chemistry of igneousrocks—[Pt.] 1, differentiation index: Am. J. Sci., v. 258, p.664-684.

Tobi, A.C., 1963. Plagioclase determination with the aid of theextinction angles in sections normal to (010)—A criticalcomparison of current albite-Carlsbad charts: Am. J. Sci.,v. 261, p. 157-167.

Uruno, K., 1963. Optical study on the ordering degree ofplagioclases: Tohoku Univ. Sci. Rept., 3rd ser., v. 8, p.171-220.

Veevers, J.J., Jones, J.G., and Talent, J.A., 1971. Indo-Australian stratigraphy and the configuration and dispersalof Gondwanaland: Nature, v. 229, p. 383-388.

832

VOLCANIC ROCKS

von der Borch, C.C., Sclater, J.G., et al., 1974. Initial Reports Watkins, N.D., Gunn, B.M., Nougier, J., and Baksi, A.K.,of the Deep Sea Drilling Project, Volume 22: Washington 1974. Kerguelen: continental fragment or oceanic(U.S. Government Printing Office). island?: Geol. Soc. Am. Bull., v. 85, p. 201-212.

833