38. petrology of volcanic rocks from the fore-arc … · 2007-05-08 · 38. petrology of volcanic...

38. PETROLOGY OF VOLCANIC ROCKS FROM THE FORE-ARC SITES1

Arend Meijer, Elizabeth Anthony, and Mark Reagan, Department of Geosciences,University of Arizona, Tucson, Arizona

ABSTRACT

Late Eocene and Oligocene submarine lavas recovered along the Mariana arc-trench slope appear to represent anearly stage of arc volcanism along the Palau-Kyushu-Mariana trend. At Site 458, 90 km west of the Mariana Trench, theoldest lavas recovered were two pyroxene andesites with chemical and mineralogic characteristics generally associatedwith the arc tholeiite series. These are overlain by an unusual series of pyroxene (bronzite) andesites with relatively highMg, Ni, and Cr and very low Ti, Zr, Y, and REE contents. Lavas with similar characteristics from the Bonin Islandshave been called boninites. Because the Site 458 samples show a wider range of textures than is commonly attributed toboninites, the term "boninite series" is suggested for the purpose of discussion. These lavas probably derived from wetpartial melting of very depleted mantle materials overlying the Palau-Kyushu-Mariana subduction zone. In Hole 459B,50 km west of the trench, all the lavas recovered have the general chemical characteristics of arc tholeiite seriesandesites. Petrographically, volcanic cobbles recovered in sedimentary deposits on the lower trench slope at Sites 460and 461 (Holes 460, 460A, 461, and 461 A) 23 and 11 km, respectively, from the trench axis are similar to Hole 459Blavas. Chemical data for samples from these sites are compatible with this association.

All the volcanic rocks recovered show evidence of secondary alteration. Secondary minerals include clays, car-bonate, zeolite, cristobalite, quartz, tridymite, celadonite, chlorite/clay, Fe-hydroxides, and other minor phases. Inspite of the alteration, primary phases including pyroxenes, Plagioclase, Fe-Ti-oxides, and glass remain unaltered insome samples. This type of alteration is similar to that observed in typical seafloor volcanic sections in the major oceanbasins. No evidence of high P, low T metamorphic conditions has been uncovered.

The proximity of arc-related volcanic rocks to the trench axis is difficult to explain in terms of current models ofsubduction complexes. Possible causes include magma generation at shallow depths along the Eocene-Oligocene sub-duction zone and tectonic erosion of the lower trench slope.

INTRODUCTION

The igneous rocks recovered from the Mariana fore-arc region during Leg 60 drilling represent the most ex-tensive recovery of fore-arc "basement" rocks to dateby DSDP/IPOD. Holes 458 and 459B, respectively 90and 50 km due west of the Mariana Trench, provide uswith a rare opportunity to study the in situ "basement"rocks of an active oceanic fore-arc region. An under-standing of the nature and origin of such rocks is criticalto the development of realistic models for the evolutionof subduction complexes and, in the case of the Marianafore-arc sites, to the understanding of the evolution ofearly arc volcanism.

This chapter is devoted primarily to the presentationand discussion of petrographic, mineralogic, and geo-chemical data for samples from the fore-arc sites, withsome consideration given to their petrologic and tec-tonic implications.

HOLE 458The lower Oligocene(?) igneous section recovered

from this hole is composed mainly of fine- to medium-grained (0.1-0.5 mm max.), generally aphyric, high Mg(bronzite) andesites and andesites of the arc tholeiiteseries. Glassy chill rinds, presumably from pillowedflows, were also recovered at various levels in the hole.On the basis of shipboard observations, the igneous sec-

Initial Reports of the Deep Sea Drilling Project, Volume 60.

tion is subdivided into five petrographic units as shownin Figure 1. The essential difference among the units isthe proportion of fine/glassy to medium-grained mate-rial, although other criteria, such as the degree of altera-tion, overall color, modal mineralogy, and extent offracturing, were also considered. These criteria reflectphysical processes and are therefore of considerablevalue even though they may have no direct significancewith regard to the origin of the magmas from whichthese rocks were derived.

The density of fractures and veins observed in thecores recovered, combined with the rather low overallrecovery rate, suggests most of the igneous rocks fromthis hole were originally highly fractured (e.g, seephotos for Cores 41 and 48 in site chapter). Small-scaledisplacements were observed along high-angle fracturesin some cores, but it is not clear what significance thesehave with regard to regional stress fields to which thesubduction complex has been subjected.

Mineralogy and Petrography

Primary TexturesPetrographically, most of the rocks at this site are

similar to seafloor basalts as drilled by DSDP/IPOD inthe major ocean basins. Typical textures are shown inPlates 1 and 2. No schistose or other fabrics reflectingpenetrative deformation were found, although slicken-sided surfaces were observed along several "shear"zones (Fig. 1). Among the more significant petrographicfeatures observed were (1) unusual chill rind textures in

709

A. MEIJER, E. ANTHONY, M. REAGAN

Late Oligocene carbonate ooze

Early Oligocene silt and ash

Unit I — Highly vesicular glassy to fine-grainedpillowed basaltic andesite. Some glassy frag-ments bear highly magnesian opx and aredevoid of Plagioclase. Occasional fragmentsof limestone included wi th in pil low rinds.

Unit II — Several massive flows or sills ofmedium-grained augite•plagioclase basalticandesite separated by chilled zones wi thpi l low structure. Several zones wi th veryhigh fracture density occur within unit.

Unit III — Vesicular glassy to fine-grainedbasaltic andesite similar to but more altered

Unit IV - Vesicular, glassy to fine-grained,massive to pil lowed, augite•plagioclase basaltand basaltic andesite. Fe-Ti-oxides common.Glassy rocks bear abundant Plagioclaseunlike Unit I and II I glassy rocks. Rocks arehighly altered to clays and veined by greenclay and zeolites(s).

Unit V - Massive, fine-grained, augite plagioclasebasalt wi th evidence of pillow structure inupper and lower part of unit. Rocks in unitheavily altered to clay.

Figure 1. Igneous stratigraphy in Hole 458. Units I through V repre-sent petrographic units. Percentage recovery of each 9-meter coreinterval indicated by shading in column to right of depth designa-tion. Horizontal cross-hatching represents "sheared" intervals.

which abundant pyroxene occurs as the only primarycrystalline phase, (2) a very low average phenocrystabundance (<5 vol. %), (3) spherulitic intergrowths ofquartz and Fe-Ti-oxide ± Plagioclase, and (4) a highaverage vesicularity.

The unusual chill rind textures occur among the pil-lowed bronzite andesites. They consist of small euhedralto acicular (± "swallow tail") clinopyroxene microlites(Plate 1, Figs. 1,2, and 5) and scattered clinopyroxeneand orthopyroxene microphenocrysts (Plate 1, Fig. 3)set in a mesostasis of clear glass. Occasional brown ve-sicular "clots" occur in the glassy matrix. These arecomposed of very fine grained Plagioclase, pyroxene, ±an oxide phase and apparently represent localizedcenters of degassing (Plate 1, Fig. 5). The vesicles inthese clots are irregular and very small (0.05-0.1 mm).

The chill rinds also contain large spherical to ellip-soidal vesicles (0.5-1.5 cm) near their margins (Plate 1,Fig. 5). These are occasionally lined with zeolites andother secondary minerals (Plate 3, Fig. 4). The vesicu-larity of the rinds is often as high as 30 to 40 vol. % andimplies substantial volatile contents at time of eruption.The fine- to medium-grained samples show a wide rangeof vesicularity (5-45%) which is roughly correlated withthe inverse of their grain size. In general the vesicles are

not filled but are often lined with clay minerals ±cristobalite. The abundance and form of the vesicles(Plate 2, Fig. 1) in these samples are similar to those ininterarc basin lavas dredged from the Mariana Trough(see site chapters for 454 and 456).

The groundmass of most of the intergranular andsubophitic textured samples contain small "spherulites"comprised of submicroscopic fibers of a silica phase andminute stringers of an Fe-Ti-oxide phase (Plate 2, Figs.2 and 3). These "spherulites" are generally nucleated onprimary origin. During secondary alteration they are re-crystallized to "buttons" of cristobalite and coarser-grained Fe-Ti-oxide (see Plate 4, Fig. 4).

Primary Mineralogy

The primary (igneous) minerals in rocks from this siteinclude, in order of abundance, clinopyroxene, Plagio-clase, orthopyroxene, Fe-Ti-oxides, quartz(?), and abrown (Cr-rich) spinel. Clinopyroxene and rare ortho-pyroxene microphenocrysts occur in Petrographic UnitsI, II, III, and IV, although together they rarely con-stitute more than 3 vol. % of the rock. Plagioclase andclinopyroxene microphenocrysts occur in nearly equalproportions in Petrographic Unit V and Core 41 of UnitIV. Their combined abundance is generally also lessthan 2 to 3%.

Analyses of pyroxene microphenocrysts (0.2-0.4 mm)from various levels in the hole, obtained by microprobetechniques, are listed in Table 1 and plotted in Figure 2.Because clinopyroxene microphenocrysts are not sub-stantially altered in most of the rocks in this hole,the analyses probably represent primary compositions.Those from the upper part of the section (Cores 28, 30)are substantially more magnesian than those from UnitV (Core 47). This correlates with the higher whole-rockMgO contents of samples from the upper portion of thehole and implies higher eruption temperatures. Clinopy-roxene also occurs as the dominant groundmass pyrox-ene and as overgrowths and exolution lamellae in ortho-pyroxene.

The orthopyroxene phenocrysts found in rocks fromUnit I are of bronzite composition (Table 1; Fig. 2).Representative analyses of orthopyroxenes from theother units were not obtained either because of smallgrain size or alteration effects. The orthopyroxenes ofUnit I commonly display complex zoning or twinning,as shown in Plate 2, Figure 4. Orthopyroxene with com-positions and forms similar to those observed at this sitehave been described from the Bonin Islands (Kurodaand Shiraki, 1975), Papua (Dallwitz and others, 1966),and the southern Mariana Trench (Dietrich et al., 1978;Fig. 2). Rocks with these distinctive orthopyroxeneshave been given the name "boninite" (Johanssen, 1939;Cameron et al., 1979). Unlike the latter two localities,however, enstatite has not been found in rocks fromHole 458.

Plagioclase is largely unaltered in rocks in this holeand shows a distinct difference in habit in the upper andlower portions of the section. In fine-grained or glassyrocks in the upper portion, it is either absent altogether(Plate 1) or occurs as spherulitic quench crystals inter -

710

PETROLOGY OF FORE-ARC VOLCANICS

Table 1. Pyroxene analyses.

Sample(interval in cm)

SiO2

TiO2

A12O3

'FeO'MnOMgOCaONa2OK2OP2O5Cr 2 O 3

Total

"state"

458-28-1,136-139

51.970.233.199.670.17

17.3316.070.11n.d.n.d.0.03

98.77

f.g.b0.2

53.490.112.246.710.16

18.2118.340.13n.d.n.d.0.42

99.81

f.g.b0.4

458-30-1,45-49

53.20n.d.2.486.660.19

17.7919.130.110.040.38n.d.

99.98

f,m0.2

53.320.102.256.970 18

18.8917.510.120.030.03n.d.

99.40

f,m,b0.8

458-33-1,121-124

50.920.313.92

11.49n d13.0220.670.150.03

n.d.n.d.

100.50

f,o,b1.5

55.250.061.215.29n d18.0319.870.090.03n.d.n.d.

99.84

f,o,

----

Clinopyroxene

458-47-1,94-98

51.580.351.58

16.18n d13.8116.320.210.03

n.d.n.d.

100.07

b.p,0.5

51.84n.d.1.41

14.44n d13.5818.300.170.03n.d.n.d.

99.78

b,p,0.3

459B-66-3,22-26

52.30.321.91

14.5n.d.15.5715.940.210.03

n.d.n.d.

100.79

f,g,0.2

459B-60-1,46-48

50.660.233.036.94n.d.16.1018.620.280.08n.d.n.d.

95.95

e.g,0.2

54.440.121.763.95

n.d.18.2319.810.210.03n.d.n.d.

98.56

e,g,0.1

459B-65-1,17-21

51.330.203.335.52n.d.16.6019.820.140.03n.d.n.d.

96.95

e,m,0.3

50.120.442.13

10.69n.d.12.9219.680.230.03n.d.n.d.

96.24

e,m,0.6

459B-72-1,138-141

48.250.753.95

17.64n.d.11.8916.550.260.03n.d.n.d.

99.32

f,g,b,0.3

52.630.172.855.89n.d.16.9220.63

0.160.03n.d.n.d.

99.34

f,m,0.2

Orthopyroxene

458-43-2,34-37

55.660.040.96

10.570.19

30.351.560.000.020.010.42

99.77

f,s,0.4

458-44-1,61-68

56.960.000.548.030.29

31.701.540.020.030.01n.d.

99.11

f,P,1.0

Note: b= bladed, e = etched, f = fresh, g = groundmass, m = microphenocryst, o = ophitic or subophitic, p = phenocryst, s = skeletal, 0.2 = maximum dimension in mm.

Di

Φ Hole 458

• Hole 459B

En

Hd

ΦΦ

1—high Mg andesite,

vCape Vogel, Papua

2—andesites and

Dacites, Saipan

• Fs

Figure 2. Plot of pyroxene compositions for samples from Holes 458, 459B, Saipan and Papua, New Guinea. Saipan analyses by optical methodsfrom Schmidt (1957). Papua analyses by microprobe from Dallwitz and others (1965). See Table 1 for details of new analyses.

grown with clinopyroxene (Plate 3, Fig. 1). In fine-grained rocks of the lower portion, it occurs as pheno-crysts and as the main crystalline phase in the ground-mass (Plate 3, Fig. 2). This difference in habit is prob-ably a function of differences in magma chemistry, asdiscussed later. Coarser-grained rocks from the upperportion contain abundant, coarsely crystalline Plagio-clase intergrown with clinopyroxene (Plate 2, Figs. 2and 3; Table 2, Sample 458-33-1, 121-124 cm), suggest-ing that some factor (PH2o?) inhibited Plagioclase crys-tallization in the fine-grained rocks of this portion.Analyses of Plagioclase phenocrysts presented in Table 2range from An63 to An76. No albite was found in thesamples studied.

In the upper portion of the section, the presence orabsence and the size of the Fe-Ti-oxide phase are gen-erally a direct function of the crystallinity of the matrixin any particular rock. The pillow rinds have none,whereas the microdiabases have up to 5% of discrete

0.02 to 0.05 mm crystals in addition to the "stringers"in the quartz-plagioclase spherulites (Plate 2, Figs. 2and 3). In the lower portion of the section (Unit V) a dis-crete oxide phase is evident in even the finest-grainedsamples (Plate 3, Fig. 2), presumably due to the highertotal iron content of these rocks (see site chapter). Noanalyses of oxide phases were obtained for samplesfrom this hole.

A brown spinel phase is occasionally found as octa-hedra included in clinopyroxene microphenocrysts. Noanalyses were obtained. No sulfide phases were iden-tified in any of the samples from this hole, but flecks ofnative copper occur as part of a vein filling in a samplefrom Core 39 (458-39-1, 81-84 cm).

Secondary Mineralogy

The effects of alteration and diagenesis are evidentthroughout the cored section. Secondary minerals tenta-tively identified in thin section include various smec-

711


Table 2. Plagioclase analyses.


SiO2

TiO2A12O3

'FeO'MnOMgOCaONa2OK2OP 2O 5

Totala

Mol % An"state"

458-33-1,121-124

52.650.04

32.420.94n.d.0.14

15.062.790.07n.d.

104.1

75e,o,0.4

53.000.00

30.330.770.000.07

13.193.400.05n.d.

100.82

68f,s.

0.3

458-46-1,100-106

51.450.10

31.010.950.000.09

13.853.080.100.04

100.6

73

f,s,0.25

51.300.06

31.650.740.000.09

14.512.570.050.00

100.97

76

f.s,0.25

458-47-1,94-98

52.340.72

31.401.52n.d.0.11

13.263.660.10n.d.

103.11

66e,m,0.2

51.420.03

28.30.74n.d.0.08

12.413.920.06n.d.

96.95

63e,m,0.3

459B-60-1,46-48

49.190.04

26.64n.d.n.d.0.12

14.242.461.84n.d.

94.52

68

e,o,0.2

50.210.01

31.830.76n.d.0.19

15.552.540.06n.d.

101.16

77

e,o,0.5

459B-65-1,17-

48.090.11

34.250.890.010.06

17.061.800.040.00

102.29

k4

e,m,0.7

-21

51.950.06

29.411.08

n.d.0.20

13.273.990.06

n.d.

100.03

65e,s,

0.3

459B-66-2,22-26

53.760.01

31.361.07

n.d.0.12

13.683.760.05

n.d.

103.79

67

e,o,0.4

459B-71-1,

56.050.04

27.490.87

n.d.0.10

12.034.560.10

n.d.

101.25

59e,g,0.1

142-145

54.98n.d.

27.050.76n.d.0.06

10.415.350.08n.d.

98.68

52

e,g,0.4

53.520.10

27.871.28

n.d.0.17

12.524.170.07n.d.

99.69

62

e.g.0.2

459B-72-1,138-141

49.670.04

28.70n.d.n.d.0.17

13.153.300.0^n.d.

95.10

69

e.g.0.6

54.140.07

29.391.04

n.d.0.11

12.214.490.10

n.d.

101.55

60

e.g,0.2

Note: "State" symbols same as Table 1.a Poor totals due to small grain size and/or alteration.

tite/illite series clays, interlayered chlorite/clay, cela-donite, carbonate, zeolite, and minor hydrous iron ox-ides. Of the primary (igneous) phases, clinopyroxene,Plagioclase, Fe-Ti-oxide, and quartz(?) remain largelyunaltered, whereas groundmass orthopyroxene and in-terstitial glass have been largely replaced.

Although no effort was made to identify the variousclay minerals in these rocks, the variations in color sug-gest a number of different types are present. The colorsof clays which occur as vesicle linings vary from lightgreen or aquamarine in Unit I, to light gray in Units IIand III, to dark green in Unit IV and chocolate brown inUnit V. These variations presumably reflect differencesin composition and the oxidation state of iron. Clayslining veins are usually white to cream-colored and(rarely) apple green. Alteration of the glassy mesostasishas produced light olive green to brown clays. Pseudo-morphs after orthopyroxene (Plate 3, Fig. 5) generallycontain clays (interlayered chlorite-clay[?]) of slightlygreener color than the clays in the adjacent mesostasis.Clays rarely fill vesicles entirely except in Unit V. Plate3, Figure 3 shows a vesicle lined with a brown clay andfilled with green clay (celadonite?). A microprobe anal-ysis of the brown clay is listed in Table 3.

A carbonate phase is present in most of the coresfrom 28 to 35. In the fine-grained samples the carbonateoccurs as a vein filling, whereas in coarser-grained sam-ples (e.g., Cores 32, 33) it occurs as a vesicle and cavityfilling. The composition of this phase was not deter-mined.

Table 3. Analyses of secondary phases.


SiO2

TiO2

A12O3

'FeO'MnOMgOCaONa2OK2OP2O5

Totala

458-43-2,34-37

37.880.12

12.880.000.020.121.52

33.232.470.00

88.24

zeolite

458-16-1,100-106

46.750.00

11.0324.560.096.402.470.251.040.00

92.59

clay

38.490.24

10.5817.670.144.452.530.391.240.03

75.78

clay

459B-66-3,22-26

43.550.233.71

12.56n.d.4.90.080.186.84

n.d.

72.05

celadonite

459B-72-1,138-141

37.520.463.13

32.6n.d.5.040.650.472.7n.d.

82.56

stilpno-melane(?)

459B-71-1,142-145

98.500.020.280.12n.d.0.020.040.130.03n.d.

99.16

97.630.080.290.14n.d.0.070.040.140.06n.d.

98.18

cristobalite

1 Not including H2O, CO2, etc.

Zeolites occur sporadically in the igneous section as avein and vesicle filling (Plate 3, Fig. 4) primarily in theglassy chill rinds of pillowed flows. On the basis ofmicroprobe analyses (Table 3) and an X-ray pattern thedominant zeolite appears to be phillipsite.

In summary, the secondary mineral assemblages ob-served in samples from this hole are similar to those ob-served in altered seafloor basalts from the major oceanbasins (Bass and others, 1973; Bass, 1976; Scheideggerand Stakes, 1977) except that silica minerals are morecommon in rocks from this site. No consistent correla-tions between secondary mineral assemblages and depthhave been discovered, although carbonate appears to berestricted to the upper portion of the section. The rocksin this section have clearly undergone diagenesis, asevidenced by the secondary mineral assemblages andchanges in the whole-rock chemistry discussed in thefollowing.

Geochemistry

Major and trace element analyses of samples selectedfrom cores recovered from this hole (Tables 4-6, Woodet al., this volume) suggest the following: (1) Samplesfrom petrographic Units I through IV are generallyquartz normative, bronzite andesites enriched in refrac-tory elements such as Ni, Cr, and Mg and depleted in"immobile" large ion lithophile (LIL) elements includ-ing Ti, Zr, Y, and HREE, relative to andesites fromother areas. (2) Samples from petrographic Unit V arealso quartz normative but are substantially depleted inrefractory elements and enriched in LIL elements rela-tive to samples from overlying units. They have chemi-cal characteristics common to the arc tholeiite series. (3)Secondary alteration has modified the original concen-trations of most of the major elements and many traceelements, although comparison of whole-rock analysesof adjacent altered and unaltered material suggests first-order trends are preserved in spite of the alteration. And(4) on the basis of variations in the concentrations of"immobile" elements, the degree of differentiation de-creases upward in the section (i.e., with decreasing age).

The normative mineral abundances of altered andunaltered samples from this site are plotted in Figure 3along with data for samples from Guam and a site south

712


Figure 3. Normative mineral composition of Mariana fore-arc lavas. Samples from dredge site 1403,southeast of Guam from Sharaskin (in press). Details of Guam analyses to be published elsewhere.Closed curve contains analyses of unaltered samples from Holes 458 and 459B. Scatter of other 458and 459B analyses due primarily to secondary alteration. Normative mineralogy calculated assumingFe2O3 =0 .15 FeO.

of Guam dredged in 1976 by the Dimitri Mendeleev(anonymous, 1977; Sharaskin, in press). Analyses ofunaltered aphyric samples from Holes 458 and 459Bconsistently show on the order of 15 to 20% normativequartz (closed curve in Fig. 3), whereas analyses ofaltered samples from both holes show a large amount ofscatter. This scatter is believed to be due primarily tosecondary mobilization of components as discussed inthe following section.

Samples from Petrographic Units I through IV gen-erally have high Mg/Mg + EFe (atom) ratios (50-60)relative to other types of andesites. Because the magmasthese samples represent were probably modified bycrystal fractionation of Fe-Mg minerals after separationfrom their source materials (see discussion following),they may originally have had even higher ratios. Green(1976) has shown that melts with these characteristicscan be produced by high degrees (25-35%) of partial fu-sion of ultramafic material under hydrous conditions at1000 to 1200°C and 10 to 20 kbar.

A discussion of the petrogenesis of these lavas re-quires analyses representative of their original (igneous)chemical compositions. Such analyses are difficult toobtain because most of the samples in this hole havebeen chemically altered by secondary processes. Thematerial which appears most accurately to reflect orig-

inal igneous compositions is found in the chill rinds ofthe pillowed flows. Although all of the rinds recoveredfrom this hole show some degree of alteration, second-ary phases in the rinds can be separated from primary(igneous) material by the use of heavy liquids, magneticseparation, and hand-picking. We have obtained whatwe believe to be unaltered aliquots (50-500 mg) of pil-low rind fragments from several levels in the hole. Thealiquots are composed mainly of glass and quench crys-tals and are thought to closely approximate composi-tions of the lava at the time of consolidation. Portionsof each aliquot were fused on an Mo strip furnace underan Ar atmosphere at 1750°C (Brown, 1976) to producehomogeneous glasses which were analyzed for majorelements on a microprobe (see Appendix).

The results of the analyses are presented in Table 4.Values for loss on ignition, obtained from weight lossduring fusion in the Mo strip furnace, are accurate towithin ±0.20% absolute. Relative to the conventionalwhole-rock analyses reported by Wood et al., this vol-ume, the pillow rind analyses, although fewer in num-ber, show much less variation. This is particularly evi-dent in Figure 4, where both types of analyses are plot-ted relative to depth of recovery. The smaller variationsevident in the pillow rind analyses correlate very wellwith the variations in "immobile" trace element con-

713


SiO-• TiO

50 60

250

300

QE 350

400

450

10 5 10 0 5 0 2.0

ft

it- f•

~<T

v .

T

Figure 4. Major elements in wt.% vs. depth of recovery for samples from Hole 458. C.U. I through III are chemical units. Soliddots represent whole rock analyses of altered samples. Stars represent microprobe analyses of fused chill rind separates. Notedepletion and enrichment of elements in altered samples relative to unaltered chill rind separates. "FeO" represents total Fe asFeO.

centrations obtained by conventional whole-rock analy-ses of altered samples (compare with Fig. 5). Together,these data define at least three chemical units (C.U.):C.U. I includes petrographic Units I, II, and III; C.U. IIcorresponds to Petrographic Unit IV; and C.U. Ill cor-responds to Petrographic Unit V. The extent of differ-entiation evident within these units decreases system-atically upward in the section.

Analyses of primary glasses found in the chill rindsamples are also listed in Table 4. Compared to thewhole-rock (fusion) analyses, glass compositions di-rectly reflect the crystallization of groundmass pyroxene.

The chemical effects of secondary alteration areclearly evident in Figure 4. Comparison of microprobe(fusion) analyses and conventional analyses of altered

whole rocks suggests that SiO2, and to a lesser extentCaO, were depleted by the alteration processes whereasNa2O, K2O, and MgO were variably enriched. TheA12O3 data suggest depletion in the upper parts of thesection and enrichment in the lower part (C.U. III). Thiscrossover trend and the increased depletion of SiO2 andenrichment of MgO downward in the section probablyreflect higher alteration temperatures at depth. Theminor systematic enrichment of FeO and TiO2 in thealtered samples is mostly a result of constant volumedensity changes resulting from the alteration processes.As the mean atomic weight of a rock is decreasedthrough the addition of H2O and other light elements,the relative concentrations of FeOt and TiO2 are in-creased if they remain immobile.

Table 4. Whole-rock and glass analyses.


SiO2

T i < H

Al 2 O 3

FeO,M i l l )

MgO( aON a 2 OK2()P 2 O 5

rotal

L.O.I.Ni (ppm)Cr (ppmiY (ppm)Zr (ppm)Ba (ppm)CaO/TiO2

Al 2 θ3/TiO 2

T,O2/Zr

(glassy)458-28-l,d

140-142

56.10.28

13.78.390.146.119.591.970.440.02

97.67

—74

1897

2833

34.248.9

100

2a

(cryst.)458-34-1,

39-41

56.90.30

14.78.060.125.53

10.762.611.120.04

101.02

_

65200

9\04735.949.0

100

3b

Boni

458-28-1,136-139

57.990.23

14.488.03n.d.6.71

10.471.870.35n.d.

100.13

3.07_—

28 e

—45.563.0S2

4b

nite Series

458-30-1,45-49

57.880.22

13.638.14

n.d.6.34

10.391.830.35n.d

98.^8

3.61—

_27C

47.262.081

5C

458-30-1,45-49

57.640.33

15.507.530.135.049.731.78

0.350.15

98. IS

_

——

27C

_

29.547 0

122

6b

458-39-1,23-26

57.180.24

13.48

n.d.7.56

10.841.590.30n.d.

99.9"

3.89——

35°

45.256.283

7b

458-43-1,137-140

58.470.47

13.428.76n.d.6.108.902.540.39n.d

99.05

3.85——_

51 e

—18.928.692

8C

458-43-2,34-37

58.640.52

15.388.020.124.568.741.840.350.17

98.34

_——

51C

—16.829.6

102

9 a

459B-46-1,75-77

52.81.13

13.612.450.106.175.883.410.750.06

9"."4

_

2311

635.2

12.0174

10 b

459B-46-1,100-106

59.280.88

15.1610.61

n.d.3.067.633.100.35n.d.

100.07

4.81———

68C

—8."

17.2135

l ie 12b

Tholeiite Series

459B-46-1,100-106

59.321.03

14.1610.820.182.366.^41.630.310.10

96.65

_

———

68 e

—6.5

13.7151

459B-47-1,94-98

58.060.93

14.2710.70

n.d.3.597.983.230.36n.d.

99.12

3.12—__

69e

—8.6

15.31.35

13b

459B-65-1,34-38

59 300.90

13.9910.34n.d.3.43^643.040.28n.d.

9S.92

3.74——

68°—

15.5132

14b

459B-73-1,80-82

59.580.90

14.539.43n.d.3.417.883.040.32n.d.

99.08

4.07——

66 e

—8.8

16.1136

15

Arc Thol.

16

eiite SeriesJakes and White (1972)

Basalt

51.890.80

16.019.570.176.77

11.812.420.440.11

99.99

_

3050_

707514.820.0

114

Andesite

59.191.29

16.098.39

3.486.334,330.440.45

99.99

_

2015

70100

4.912,5

184

j* Analyses of altered samples by University of Birmingham, England (Wood el al., this volume)b Whole rock microprobe analyses of "unaltered" separates (see text). Accuracy and precision data in Appendi:c Microprobe analyses of interstitial glass in pillow rinds.d Numbers represent 1POD notation of leg-site-core-section (cm), e.g., 60-458-28-1, 140-142.e Interpolated from University of Birmingham data for samples from same core and section.

714


Samples from Petrographic Units I through III showonly minor variations in igneous chemistry but display awide range of textures. Examples of the textural dif-ferences are shown in Plates 1 and 2.

Data for the concentrations of Y, Zr, Cr, and Ni insamples (see Wood et al., this volume) from this hole areplotted in Figure 5. These elements appear to be "im-mobile" during low temperature processes (Smith andSmith, 1976) and are therefore of considerable signifi-cance with regard to the definition of original chemicaltrends. In terms of overall trends, Ni and Cr decreaseand Y and Zr increase downward in the section._B_ecauseY and Zr are essentially incompatible elements (D < 0.2)and Ni and Cr are compatible elements (D >2.0) in thissystem, these trends also indicate a general decrease indifferentiation index upward. This is opposite to thetrend expected for the differentiation and episodic erup-tion of a single batch of magma and therefore impliesthe existence of multiple magma batches. Some idea ofthe number and complexity of magma batches can bederived by recognition of the fact that, during differ-entiation of this type of magma, a residual liquid willhave higher Zr and Y and lower Ni and Cr abundancesthan its parent. If we assume the analyses plotted inFigure 5 represent liquid compositions with the young-est at the top and the oldest at the bottom, a line con-necting two cogenetic magmas (residual and parent) willhave a negative slope for Ni and Cr versus depth and apositive slope for Y and Zr. Inspection of the figure in-dicates this constraint is not often met, suggesting eitherthat many individual batches of magma were producedin the source region and erupted or that the magma

chamber feeding the eruptions was periodically replen-ished.

The distinctions between magma types erupted in thishole are particularly well displayed in a Zr versus TiO2diagram. Because both of these elements have similarlylow bulk distribution coefficients in this system, theirratio closely reflects the ratio of their parent materials.As shown in Figure 6, samples from C.U. Ill havehigher TiO2/Zr ratios than C.U. I and II samples. Com-bined with the other major element and trace elementdata already discussed, these data suggest the existenceof two magma types in Hole 458; the first type, repre-sented by C.U. Ill, may be classified as an arc tholeiiteseries magma. The second type represented by C.U. Iand II, is characterized by enrichment of refractoryelements and depletion of LIL elements relative to otherandesites. Because this type includes boninite-like rocks,it is tentatively called a "boninite series" magma (Mei-jer, 1980).

The low TiO2 and Zr contents and the low TiO2/Zrratios of the C.U. I and II samples are distinctive amongigneous rock series in general and suggest derivationfrom very depleted source materials (Coish and Church,1979; Meijer, 1980). That C.U. II samples have higherLIL element abundances than samples from C.U. I,combined with the fact that samples from both unitshave similar LIL element ratios, suggests that C.U. IIsamples could have differentiated from parent magmassimilar in composition to the C.U. I samples. This is nottrue for the C.U. Ill samples. Note, however, that C.U.II is older than C.U. I and thus cannot have differen-tiated directly from the C.U. I magmas.

Cr (ppm) Ni (ppm) Y (ppm)

0 100 200 300 0 50 100 0 10 20 30

Zr (ppm) TiO2 (wt.

0 20 40 60 80 100 0 1.0

250 •

300

E 350 •

400

450

J.

•

X

•

•

•

•

•

•

I 1 1 1 •

t

t

*

:

:

* •

Figure 5. "Immobile" trace elements vs. depth for Hole 458 samples. Dashed lines separate chemicalunits. Arrows represent predicted trend of element concentrations with fractional crystallization. Onlygeneral trend of arrows is significant.

715


1.5-

1.0 -

3 0 .5-

0 . 3 -

0.2"

• 458+ 459B©1403X Guam

TiO2/Zr=160

/

/

26 3**27•Af—

24..17 25

/ ©© ©

©©

2

1 I >λ

f/+5y"ji "I"

^ w

X

©

/A./

3

A/

13 /

_

20 30 50

Zr (ppm)

100

Figure 6. Plot of log TiO2 vs. log Zr. Arc tholeiite series lavas haveTiO2/Zr close to 160 whereas for boninite series lavas this ratio is< 100. Dashed lines connect analyses of samples that could becogenetic based on criteria discussed in text. Numbers next to sym-bols represent relative age sequence of analyses listed in each sitechapter, with #1 being the oldest sample. Only analyses of possiblecogenetic samples are numbered. Note that Samples 9 through 13from Hole 458, C. U. II, could be differentiates of samples fromC.U. I (e.g., 17-27) according to this diagram, because fractionalcrystallization has little effect on TiO2/Zr in these magmas.

For those samples which appear to be cogenetic,based on trends in their trace element concentrations,information on the phases that must be removed fromone to produce the other by fractional crystallizationcan be obtained from log plots of the concentration ofan incompatible element, such as Zr, with compatibleelements, such as Cr and Ni (Allegre et al., 1977). Forsamples from this hole (Fig. 7) the slopes of lines con-necting cogenetic pairs yield average Cr and Ni bulkdistribution coefficients of 2.2 and 2.3, respectively, formost samples from C.U. I and II and 8.3 and 4.6 forthose of C.U. III. These results are puzzling in the sensethat the C.U. I and II samples would logically be ex-pected to have higher Ni and Cr bulk distribution coeffi-cients than those of C.U. III. Experiments on the liq-uidus phase relationships in a sample from C.U. I,reported by Kushiro (this volume), indicate that olivineand clinopyroxene are near liquidus phases for this com-position up to 13 kbar under hydrous conditions. Ifthese two minerals controlled the fractionation behaviorof the C.U. I and II liquids, the bulk distribution coeffi-cients would have values of at least 5 to 10 for Ni and 3to 8 for Cr.

Several factors could be responsible for the lowvalues inferred for the Cr and Ni distribution coeffi-cients, including the following:

1) The pairs connected by dashed lines in Figure 7are not all cogenetic even though they display the re-

quired age and chemical relationships. As shown in Fig-ure 7, this argument would not explain the more generalL-shaped trend among the analyses even without theconnecting lines.

2) The values of distribution coefficients are T, P,and composition-dependent and are lower at higher Tand more mafic compositions. For instance, the distri-bution coefficients for Ni between olivine and mafic liq-uids vary greatly with the MgO content of the liquid(Hart and Davis, 1978). According to the results of Hartand Davis a change in MgO content of the liquid from10 to 3 wt.% can result in a threefold increase in D^;/liq.If the compositional dependence is of similar magnitudefor clinopyroxene/liquid and orthopyroxene/liquid, itcould explain much of the observed trends. The magni-tude of the bulk distribution coefficient for Cr wouldalso be influenced by the crystallization of a Cr-richspinel phase. Whether these effects can quantitativelyexplain the observed trends cannot be adequately evalu-ated by the available data.

3) Melting versus crystallization effects. Bougaultand others (1979) have shown that during partial melt-ing the Ni and Cr concentrations in melts producedfrom mantle materials are buffered by the mantle min-eral assemblage (i.e., the melt compositions lie alongtrends with slightly negative slopes in a diagram such asFig. 7). This suggests that the low distribution coeffi-cients inferred for the C.U. I and II samples may bean artifact of the derivation of each of the C.U. I and11 samples directly from mantle sources and that thehigher coefficients for C.U. Ill samples may repre-sent fractional crystallization at crustal levels. Unfortu-nately, the calculated concentrations of Ni and Cr formelts in equilibrium with mantle assemblages are on theorder of 300 to 400 ppm and 800 to 1200 ppm, respec-tively. This is substantially higher than the averagevalues of approximately 75 ppm and 140 ppm observedin the C.U. I and II samples.

The preferred interpretation of the trends in Figure 7is a combination of factors 2 and 3. That is, C.U. I andII samples probably reflect multiple magma generationevents in the mantle followed by fractional crystalliza-tion before eruption. Presumably, the C.U. Ill sampleswere derived from less Zr-depleted source materials andexperienced a higher degree of crystal fractionation thanthe C.U. I and II samples. Quantitative estimates of theproportions of phases crystallized from these magmascannot be derived because sufficient data on the compo-sitional dependence of distribution coefficients are notavailable at present.

HOLE 459B

The igneous rocks cored in this hole are predomi-nantly fine- to medium-grained (0.5-0.8 mm) ortho-pyroxene(?)-elinopyroxene-plagioclase andesites of thearc tholeiite series. The igneous section was subdividedinto four units on the basis of shipboard observations(Fig. 8). The occurrence of glassy pillow rinds at severallevels in the hole (e.g., Core 65, Section 1 and Core 73,Section 1) suggests the rocks at this site were erupted ina subaqueous environment. Compared to rocks fromHole 458, Hole 459B rocks are generally (1) less vesicu-

716


150

100

3 50

o

20

.•17.

2 4 " ^ i i 8 .

27* +

-

•

10

IT

ti\iIii

\ +

i\\i1

i

i

8+"*"\\

V\

\,-r

10

i

12

\\\\\\\\

+5

I

1

1111

I

1

\ |

1 -f\

λ\\•2

I\\\\\\\\\•\

\

\

\

\

\

\

\

\

\

+ ^\ \\ \

T•T-

3

1r

D 2 - f i r>

-

-

_

\\

150 -

30 50 100 150Zr (ppm)

Figure 7. Plot of log Cr vs. log Zr and log Ni vs. log Zr. Sample numbers and symbols explained in Figure 6. Two trends evident in both plots.Average bulk distribution coefficients derived from the slope of lines connecting possible cogenetic samples within each trend are shown.

lar (5-15%), (2) more oxidized, (3) more differentiated,(4) coarser grained, (5) more altered, and (6) more ex-tensively fractured.

Mineralogy and Petrology

Primary Textures

Like the Hole 458 samples, Hole 459B rocks are tex-turally similar to typical seafloor basalts and diabases.Most samples studied in thin section show intergranularto subophitic textures with the dominant variable beinggrain size. The groundmass is generally composed ofPlagioclase laths, pyroxene granules (predominantly cli-nopyroxene), Fe-Ti-oxide grains, quartz-feldspar inter-growths, and altered glass. Phenocrysts of Plagioclaseand clinopyroxene occur in some samples but generallycomprise less than 8% of the total volume. In thisrespect, the andesites from this hole are distinct fromthe typically porphyritic arc tholeiites series andesiteserupted in subaerial settings. Glassy chill rinds are rareamong the rocks recovered. Those that were studied inthin section (Samples 459B-65-1, 34-38 cm and 459B-73-1, 80-82 cm) consist mostly of clay and zeolite withmicrolites of primary Plagioclase (5-8%) and clinopy-roxene (2-3%).

One of the more distinctive features of the rocksfrom this hole is the common occurrence of spheruliticto micrographic intergrowths of quartz, feldspar, andFe-Ti-oxides. The spherulites (Table 6) occur in thefiner-grained samples and are essentially identical tothose observed in samples from Hole 458 (Plate 2, Figs.2, 3). In the coarser-grained samples, micrographic in-tergrowths of quartz and feldspar occur (Plate 4, Fig.1). The intimate association of these intergrowths withdiscrete Plagioclase crystals suggests a primary origin asa late stage crystallization product.

Primary Mineralogy

The primary minerals observed in rocks from thishole include, in order of abundance, Plagioclase, clino-pyroxene, Fe-Ti-oxides, quartz, and possible unidenti-fied accessory phases (e.g., apatite). Clay-filled pseudo-morphs with outlines typical of orthopyroxene occur inmany of the samples and suggest this mineral should beincluded.

Plagioclase is the most abundant crystalline phase inall the samples studied. It varies in habit from spheru-litic, to micrographic (Plate 4, Fig. 1), to lath-shaped(Plate 4, Fig. 2), with an average abundance of 30vol.%. Compositionally, it varies from approximately

717


Table 5. Analyses of accessory primary phases.

middle Eocene (reworked?) claystone,

silicified claystone, and chert.

Unit I — Vesicular, fine-grained, aphyric cpx-plagioclase basaltic andesite with pillowstructure in upper part. Lower part of unitmedium to coarse grained cpx-plagioclasebasalt/diabse. Mesostasis of most rocksaltered to clay zeolite and carbonate.

Unit II — Fine- to medium-grained, sparselyphyric cpx-plagioclase basalt. Vesicularity isvariable from 10-20%. Textures suggest unitis comprised of pillows and thin flows. Alter-ation extends to clinopyroxene microlites inaddition to the mesostasis of most rocks.

Unit III — Sparsely vesicular, medium- to coarse-grained quartz-clinopyroxene•plagioclase dia-base. Micrographic intergrowths of quartzand feldspar common. Fe-Ti-oxides 2-5%.Isolated quartz found in several sections.Mesostasis of most rocks altered to brownto green clays.

Unit IV — Fine-grained, sparsely phyric clino-pyroxene-plagioclase basaltic andesite. Similarto Unit II except for lower vesicularity(3-15%), more abundant Fe-Ti•oxides (2-7%),and substantially greater fracture density.

Figure 8. Igneous stratigraphy in Hole 459B. Units I through IV rep-resent petrographic units. Percentage recovery shown as in Figure1. Cross-hatched horizons represent heavily fractured zones.

An50 in the micrographic intergrowths to An84 formicrophenocrysts in the glassy samples. Typical micro-probe analyses are given in Table 2. Zoning is generallyrestricted to the outer rims of larger crystals. Normalzoning is the most common, although complex zoningwas observed in crystals from Unit III.

Clinopyroxene occurs as (1) subhedral to euhedralcrystals (0.3-0.5 mm) subophitically intergrown withPlagioclase (Plate 4, Fig. 3), (2) small (0.05-0.1 mm)granules in glassy samples, and (3) as spherulites inquenched samples. Typical compositions are listed inTable 1 and plotted in Figure 2.

An Fe-Ti-oxide phase is invariably present in samplesfrom this hole and usually occurs in two size popula-tions. One size group (0.03-0.04 mm) displays a blockyhabit whereas the other (<O.Ol mm) forms skeletalcrystals in the mesostasis. The microprobe analysislisted in Table 5 indicates that the blocky crystals aretitanomagnetites.

Secondary Mineralogy

Secondary minerals identified in samples from thishole include carbonates, zeolites (phillipsite), quartz,cristobalite, tridymite(?), Fe-hydroxides, celadonite,stilpnomelane(?), chlorite/clay(?), and smectite/illite

Siθ2TiO2

A12O3

"FeO"MnOMgOCaONa2OK 2 O

P 2 θ 5

Totala

91.670.014.322.27n.d.0.030.170.320.09n.d.

98.88

Qtz.

Sample459B-66-3,22-26 cm

96.81n.d.

2.993.75

n.d.0.110.841.230.15

n.d.

105.89

-Fe-Ti-Plag.

spherulite

1.9814.62

1.2669.25n.d.0.440.040.050.21n.d.

87.84

Fe-Ti-oxide

Sample459-68-1,18-22 cm

93.94n.d.0.323.38

n.d.0.040.070.190.07n.d.

98.01

Qtz.-Fe-Ti-Plag.

spherulite

a Poor totals due to small grain size and uncor-rected "FeO" in the case of Fe-Ti-oxide.

series clays. These minerals occur as alteration productsof the fine-grained glassy mesostasis and/or primarypyroxene and in veins and vesicles. The degree to whichprimary minerals have been altered varies erraticallydown the hole. In general, fine-grained samples withpilotaxitic to intergranular textures have been morethoroughly reconstituted than medium-grained sampleswith subophitic textures. For samples of a given grainsize, however, alteration increases with depth. In fine-grained samples from Unit I clinopyroxene is etched butonly slightly altered, whereas in similar samples fromthe bottom of Unit IV clinopyroxene is almost totallyaltered.

The carbonate phase occurs primarily as a replace-ment of the fine-grained mesostasis in samples fromUnit I. It is also found as a minor alteration product ofclinopyroxene. Zeolite occurs locally as a vesicle andvein filling in glassy chill rinds.

Cristobalite and possibly tridymite occur as ubiq-uitous secondary groundmass minerals. A typical "but-ton" of cristobalite surrounded by secondary clay min-erals is shown in Plate 4, Figure 4. Microprobe analysesof two "buttons" are listed in Table 3. The amounts ofA12O3, FeO, and Na2O in these analyses are typical ofsecondary cristobalite. Tridymite was tentatively identi-fied in the groundmass of several samples on the basisof crystal form and optic sign.

Sheet silicates including clays, celadonite, sericite,interlayered chlorite/clay, and stilpnomelane(?) occuras alteration products of glass, groundmass constitu-ents, primary mineral phases, and as linings of vugs,vesicles, and veins. In general, the degree of crystallinityof these minerals increases with depth in the hole. Apoorly crystalline stilpnomelane-like phase was ob-served in the lowermost core (Core 73, Section 3). Itsidentification is based on petrographic characteristicsand a microprobe analysis (Table 3). The interlayeredchlorite/clay minerals occur primarily in Unit IV, al-though they also occur as alteration products of ortho-pyroxene in the other units. Celadonite (Table 3) occurs

718


as an alteration product of the groundmass from Core68 downwards. It occurs in the oxidized parts of thegroundmass and is generally accompanied by Fe-hy-droxides. Because the formation of celadonite requiressubstantial amounts of K2O, its presence in abundanceimplies that these rocks were metasomatically altered.

The clay minerals in these rocks vary in color fromaquamarine near the top of the section to chocolatebrown near the bottom. A single thin section will oftencontain two or three differently colored clays. One coloroccurs as an alteration product of the glassy mesostasis,another is found replacing ferromagnesian minerals,and a third occurs as a lining in vugs and vesicles. Thesevariations presumably reflect a complex alteration his-tory as well as locally variable chemical gradients. Amicroprobe analysis of vesicles filling clays is listed inTable 3.

Geochemistry

The primary chemical characteristics of samples fromthis hole indicate they are andesites of the arc tholeiiticseries (Jakes and White, 1972). These characteristics in-clude (1) low Mg, Ni, and Cr and relatively high Fe con-centrations; (2) low K and "intermediate" Ti, Y, Zr,and REE concentrations; and (3) TiO2/Zr ratios of 150to 160. Primary concentrations and ratios of "mobile"trace elements other than K2O have not been obtainedbecause of the effects of secondary alteration.

The extent to which alteration has modified the pri-mary (igneous) compositions of samples from this site isevident in Figure 9, a plot of major element compositionversus depth for altered and "unaltered" samples. Un-fortunately, glassy pillow rinds which contain the

"unaltered" material are rare among the rocks cored inthis hole; only two samples were obtained. Microprobeanalyses of these two samples, obtained according to thetechniques outlined earlier, are very similar to theaverage arc tholeiite series andesite (Table 4; Jakes andWhite, 1972).

Comparison of the analyses of altered and unalteredsamples from the same core and section (Fig. 9) suggestthat SiO2, and to a lesser extent CaO, have been lost;MgO, Na2O, K2O, and minor "FeO" have been gainedduring the alteration; and A12O3 and TiO2 were not sig-nificantly changed. A particularly profound example ofthe effect of secondary alteration is provided by micro-probe analyses of interstitial glasses in Sample 459B-65-1, 34-38 cm, (Table 6). All the analyses are of palebrown glass showing some etching but no optically dis-tinguishable secondary minerals. The low totals reflectthe hydrated condition of these glasses. Measured con-centrations of K2O vary by at least a factor of 10 amongthe analyses, the lowest value (0.33) being close to theanalysis of "unaltered" material separated from thissample (compare analyses for Sample 459B-65-1, 34-38cm in Tables 4 and 5). Combined with the whole-rockdata reported in the site chapter and in Figure 9, thesedata indicate substantial amounts of K2O, and other"mobile" cations were probably added to the igneoussection at this site.

Evidence for primary variations in the chemistry oflavas cored in this hole, based on variations in the ("im-mobile") elements Ni, Cr, Y, and Zr (Wood et al., thisvolume), are shown in Figure 10. A distinct trend inchemistry is evident, the lowest cores being the most dif-ferentiated. Because these analyses are for sparsely

SiO* TiO, AUO MgO CaO

550'50 60 0.5 1.0

"ε 600

650

Nao0 K,0

1

LU

T

L U

Figure 9. Major elements vs. depth of recovery for samples from Hole 459B. Dots represent analyses of altered samples listed in site chapter. Starsare microprobe-fusion analyses of pillow rind separates. Chemical units as defined in site chapter. All analyses in wt.%.

719


Table 6. Analyses of altered glasses.£

SiO2

Tiθ2Al?Oλ

" F e O "M n OMgOCaONa2θK 2 O

P2θ5

Totalb

46.590.8

10.968.88n d2.075.412.450.33n.d.

77.48

59.481.02

13.9911.34n.d2.646.913.130.42n.d.

98.93C

Sample 459B-65-1, 34-38 cm

50.270.29

12.9910.68n.d4.211.740.552.56n.d.

83.30

53.610.47

15.565.870 193.921.100.773.160.02

84.68

51.550.618.96

17.810 165.062.410.683.230.02

90.51

48.590.659.74

14.51n d4.451.180.513.63n.d.

82.26

51.491.89

11.5214.70 264.061.310.773.690.02

89.72

49.732.28

10.7414.89n d3.731.510.653.88n.d.

87.4

53.560.67

10.4216.39n.d.4.821.300.6

4.19n.d.

91.96

a All glasses pale brown with only minor evidence of devitrification.b Not including volatiles.c First analysis × 1.277. Compare with Table 4 fusion analysis for Sample

459B-65-1, 34-38 cm.

phyric samples, and these elements are not much af-fected by low temperature alteration, these trends prob-ably represent trends in magma compositions and in-dicate that the youngest magmas were the least differen-tiated. This again precludes a model in which the lavasin this hole derived through progressive crystallizationof a single batch of magma and suggests that numerousseparate batches of magma underwent shallow-levelcrystal fractionation and erupted episodically. The con-sistency of TiO2/Zr ratios in the inferred parentmagmas does suggest that each batch could have derivedfrom the same or similar source materials.

The similarities between rocks in this hole andsamples from C.U. Ill in Holes 458 are striking. Likethe 458 samples, the chemical variations partly reflect

shallow level crystal fractionation of the micropheno-cryst assemblages (see Fig. 7).

HOLES 460, 460A, 461, AND 461AVolcanic rocks were recovered as clasts in clayey to

conglomeratic sediments cored in these holes. Becausethey are located near the base of the lower trench slope(23 km for 460 and 460A; 11 km for 461 and 461 A), thecoarse-grained sediments probably include fragmentsderived from outcrops at various depths along thisslope. Whether the rock types and proportions observedin the deposits are representative of the exposures alongthe trench slope is not known.

Petrography

Examination of 30 thin sections of volcanic clastsfrom these holes indicates that many are very similar intexture and mineralogy to the samples cored in Hole459B. Photomicrographs of typical samples are shownin Plate 5. The common occurrence of SiO2 polymorphs(cristobalite, tridymite, quartz) in the groundmass ofthese samples makes the correlation with Hole 459Brocks particularly convincing.

In addition to the Hole 459B-like samples, sampleswith higher-grade metamorphic mineral assemblageswere also recovered. Although these were not studied indetail, the available shipboard data indicate that chlo-rite, amphibole, epidote, and quartz are common con-stituents. Texturally, most of these samples are ofvolcanic derivation.

550

Cr (ppm)

0 100 200

600

650

Ni (ppm)

0 50 100

Y (ppm)

0 20 40

Zr (ppm)

40 60

TiO2 (wt.

0.5 1.0

m

-QCM

cα

CO

Figure 10. "Immobile" trace elements vs. depth for altered samples from Hole 459B asreported in site chapter.

720


A reasonable interpretation of the diversity in meta-morphic grade among rocks recovered from depositsalong the lower trench slope is that these deposits in-clude samples from deeper stratigraphic levels that thosepenetrated in Holes 458 and 459B.

In an attempt to discover whether any unusual rocktypes (e.g., blueschists, eclogite) might be exposed alongthe lower trench slope, two samples of gravel, each con-taining several hundred millimeter- to centimeter-sizedpebbles were examined under binocular microscope.

One sample (460A-10,CC) contained exclusively vol-canic fragments. Most of the fragments were similar tosamples from Holes 459B in vesicularity, Plagioclasecontent, color of clay minerals, etc. However, about10% of the fragments were light-colored tuffaceousrocks with 10 to 20% unaltered pyroxene phenocrysts.These could represent either units along the trench slopenot penetrated in Hole 459B or be products of youngvolcanic activity along the Mariana arc trend. There wasno evidence of high-pressure, low-temperature (e.g.,blueschist facies) metamorphic conditions among thepebbles in this sample.

The second sample (461A-2,CC) contained both vol-canic and plutonic fragments. The volcanic fragmentswere indistinguishable from those in the previous hole,but the plutonic fragments were almost exclusively gab-broic, including two-pyroxene gabbros, uralitic gabbros,and noritic gabbros. Metamorphic minerals observedwithin these fragments generally reflect high T, low Pgreenschist to amphibolite facies conditions. High-pres-sure, low-temperature assemblages common to the blue-schist facies were not observed.

The available chemical data (Table 7) reinforce theclose correspondence between the volcanic clasts fromthe lower trench slope sites and rocks from Hole 459B.Analyses of chips of "fresh" vitreous glass, presumablyfrom pillow rinds, recovered from the gravel samplesare presented in Table 7. Some analyses (e.g., Table 7,analyses 9-14) are very close in composition to analysesfrom Hole 459B, whereas others appear to represent lessfractionated liquids of the same series. Note that analy-ses in Table 7 are for natural glasses whereas those forHole 459B in Table 4 are for artifically fused glasses.

DISCUSSIONIn most current models of subduction zone com-

plexes, the igneous units are inferred to be fragments of

ocean crust ± mantle. These fragments are thought tobe either trapped as a result of the initiation of a newsubduction zone (see review by Seely, 1978) or accretedafterward (Dewey and Bird, 1970; Karig and Sharman,1975), although the details of the structural mechanismsgoverning these processes have not been worked out.The morphologic diversity of modern fore-arc regions(e.g., Dickinson and Seely, 1979) suggests the possibilityof more than one mode of origin.

The rocks recovered from the Leg 60 fore-arc sitesprovide some basis for interpretations regarding the ori-gin of igneous "basement" in this particular fore-arccomplex. The generality of these interpretations clearlyremains to be tested by future studies of other com-plexes.

Rocks with chemical and mineralogic characteristicscommonly ascribed to andesites of the arc tholeiiteseries (Jakes and White, 1972) were recovered from allthe Mariana fore-arc sites. In terms of the abundancesof REE and other LIL elements, these rocks are similar,for example, to young lavas erupted in the Izu (Masuda,1968) and Tonga (in Ewart et al., 1977) arcs. The maindifferences between these groups are in texture (aphyricvs. porphyritic) and environment of eruption (sub-marine vs. subaerial). This suggests that the Marianafore-arc tholeiites may be similar to the submarine por-tions of the younger volcanic arcs.

The rocks from the fore-arc sites at 18°N are alsochemically and mineralogically similar to rocks of equiv-alent ages exposed on the islands of Guam and Saipanin the southern Mariana fore-arc region. Here, lavaserupted in both subaerial and submarine environmentsare exposed (Cloud et al., 1956; Tracey et al., 1963). Be-cause hypersthene-augite andesites are common in thevolcanic sequences on both islands, they probably ori-ginated in a volcanic arc environment. Hypersthene-augite andesites have not been reported from otheroceanic environments (Macdonald, 1960).

If a volcanic arc origin is accepted for the Marianafore-arc tholeiites, the question arises as to why theselavas are currently exposed in such close proximity tothe trench (< 50 km). Answers to this question are likelyto involve both petrologic and tectonic factors. Petro-logic factors might include the likelihood that theselavas were erupted onto very young and therefore verythin oceanic lithosphere of the West Philippine Sea asdiscussed in Meijer (in press). Tectonic factors include

Table 7. Analyses of glassesa in Holes 460A and 461A.

Analysis

SiO2TiOA12O3

"FeO"MgOCaONa2OK2O

Total0

1

56.990.77

14.6911.143.438.002.950.29

98.24

2

56.860.84

14.1611.353.087.872.950.30

97.41

Hole 460A3

55.890.79

14.5511.013.428.012.910.30

96.88

4

57.030.77

14.2611.153.207.952.990.31

97.66

5

57.500.80

14.5611.143.077.612.970.31

97.96

6

56.560.83

14.6011.193.337.952.940.31

97.71

1

54.730.65

15.309.604.739.482.490.24

97.21

2

55.140.56

15.309.674.729.052.350.24

97.04

3

55.270.55

15.499.604.639.272.490.26

97.55

4

55.210.59

15.179.624.609.242.390.25

97.07

5

55.340.58

15.389.834.589.302.460.25

97.72

6

56.690.71

14.5411.043.397.932.810.30

97.42

Hole 461A7

56.540.81

14.4711.023.187.702.890.30

96.91

8

57.100.82

14.2811.182.897.543.040.31

97.17

9

59.040.77

14.4310.622.557.033.010.34

97.80

10

58.990.82

14.4010.422.457.033.080.35

97.53

11

59.870.92

14.2110.832.166.772.920.33

98.02

12

60.260.91

14.1810.462.176.973.190.35

98.48

13

60.260.74

14.0410.002.056.273.280.37

97.00

14

60.490.82

14.059.721.956.222.860.37

96.47

Note: Each analysis represents a separate fragment. Fragments 90-100% glass.a There are natural glasses most likely from pillow rind fragments.b Low totals reflect varying volatile contents of the glasses.

721


tectonic erosion of the lower trench slope by subductionand rifting of the fore-arc region parallel to the trench(Karig, 1974). Whatever the cause, the phenomenon isclearly not restricted to the Mariana fore-arc region, be-cause similar geometries are found in the New Hebridesarc (Esperitu Santo and Malekula: Colley and Warden,1974), the Tonga arc (Eua: Ewart and Bryan, 1972), theBonin arc (Tsuya, 1937), and other arc systems.

The tectonic and petrologic significance of the bo-ninite series lavas recovered from Hole 458 is more dif-ficult to ascertain because, to our knowledge, lavas ofthis type have not been recognized among the productsof recent (Pleistocene-Holocene) volcanism. Their chem-ical characteristics apparently reflect derivation fromvery depleted ultramafic source materials (Coish andChurch, 1979; Meijer, 1980). Because such source mate-rials are likely to be very refractory, it would requirespecial conditions to cause them partially to melt. Labo-ratory experiments (Green, 1976) suggest that these con-ditions include high H2O contents and relatively lowpressures (10-20 kbar) at 1100 to 1200°C.

The primary abundance of H2O is characteristicallyvery low in oceanic lavas (Moore, 1970; Delaney et al.,1978; Harris, 1979a) but can be quite high in volcanicarc lavas (Anderson, 1975; Harris, 1979b). This dichot-omy is logically ascribed to the influence of the hydratedocean crust subducted beneath volcanic arcs (Delanyand Helgeson, 1978). Dehydration of this crust shouldrelease H2O and other volatiles into the mantle overly-ing the crust, and these presumably could cause meltingof the mantle to produce arc magmatism.

The boninite series magmas would appear therefore,to be genetically related to subduction zones as well. Amultistage melting process of mantle overlying a sub-duction zone could produce the arc tholeiite series dur-ing the early stages of melting and the boninite series inlater stages as the source material is continuously de-pleted in LIL elements (Meijer, 1980). Each progressivestage presumably would require a greater amount of wa-ter to cause a given amount of partial melting.

On the basis of paleogeographic reconstructions, itappears the P, T, fHzO conditions required for the pro-duction of boninitic magmas may well have been at-tained along the late Eocene-Oligocene Palau-KyushuRidge from which the Mariana frontal arc was split inthe Oligocene time (32 Ma; Sutter and Snee, 1978). Ifsubduction was initiated beneath the ridge in the lateEocene, as suggested by the age of the oldest arc rocksexposed in the Mariana fore-arc region (Meijer, 1980), itwould have occurred beneath very young (0-8 Ma) andtherefore relatively hot West Philippine Sea lithosphere.The addition of H2O to this lithosphere from the sub-ducted ocean crust would probably have caused meltingat relatively shallow depths as required by the experi-mental data (Green, 1976). The relative rarity of bonin-itic rocks might therefore, be ascribed to the unusualthermal conditions required for their production.

On the basis of the data on secondary mineral assem-blages discussed earlier, it appears that the igneous crustof the Mariana fore-arc region at 18°N has undergone arelatively simple low-grade alteration of the type gener-ally found in seafloor basalts (e.g, Bass, 1976). This,

combined with the other data discussed, leads to theconclusion that the Mariana fore-arc crust was formedin an arc environment and has remained near its presentstratigraphic level since formation. No evidence for ac-creted slabs of Pacific Plate crust or high P, low Tmetamorphism has been uncovered.

CONCLUSIONSIn terms of physical form (reflecting eruptive mode),

petrographic textures, and extent of alteration, the vol-canic rocks recovered from the fore-arc sites are similarto volcanic rocks recovered from the major ocean ba-sins. This similarity does not extend to the mineralogyand chemistry. The fore-arc samples consist mineralogi-cally of two-pyroxene andesites and chemically of arctholeiite series andesites and bronzite andesites of whatmight be called a "boninite series." This divergence be-tween physical characteristics on the one hand and min-eralogical and chemical characteristics on the other willcomplicate recognition of the fore-arc rock types inolder orogenic belts, particularly in view of the alteredcondition of volcanic sequences in most of these belts.

On the basis of chemical and mineralogic charac-teristics, their correlation with arc sequences on Guamand Saipan, the low grade of alteration they have expe-rienced, and their youth relative to adjacent Pacificlithosphere, we conclude that the fore-arc rocks at 18°Nwere formed in a subduction zone environment. If val-id, this conclusion implies arc-related lavas were eruptedvery close (50-100 km) to an Eocene-Oligocene trench.It follows that fore-arc basement rocks may, at least insome instances, have no direct relation to the oceaniclithosphere subducted beneath an arc. Because the base-ment at 18°N now represents one of the most tho-roughly studied in situ fore-arc basement terrains, othermodes of origin postulated for these terrains will requiremore detailed analysis. In particular, the common as-sumption that most ophiolites represent fragments ofoceanic lithosphere accreted or trapped during subduc-tion may require modification.

In terms of origins, the volcanic rock sequences inHoles 458 and 459B reflect the eruption of progressivelymore primitive (i.e., less differentiated) magmas throughtime. In both holes, andesites of the arc tholeiite seriesrepresent the oldest rocks recovered. These were followedin Hole 458 by the eruption of a second, more primitiveseries, not genetically related to the first by simple crys-tal fractionation. Both series could have been derivedfrom source materials in the mantle overlying the Eo-cene-Oligocene subduction zone. The main require-ments for their production at shallow depths are an ade-quate supply of H2O and a high geothermal gradient inthe source mantle. Paleogeographic reconstructionssuggest both conditions could have been satisfied withinthe Eocene-Oligocene Palau-Kyushu-Mariana Ridgesubduction zone above which the lavas were erupted.

ACKNOWLEDGMENTS

We would like to express our gratitude to the Deep Sea DrillingProject and its parent organization the National Science Foundationfor providing the Leg 60 samples. We thank John Tarney and his co-

722


workers for providing chemical analyses to the shipboard scientificparty. Paul Damon and John W. Anthony kindly reviewed the manu-script and Arthur Roe performed several X-ray diffraction analyses.Our work was carried out as part of the Southeast Asia Tectonics andResearch Project (S.E.A.T.A.R.) funded by the Office of the Interna-tional Decade of Ocean Exploration at the National Science Founda-tion. The University of Arizona cleared the way for one of us (A.M.)to participate in the Leg 60 cruise, and this assistance is gratefullyacknowledged. Finally, we thank Sandra L. Meijer for her patience inthe typing and completion of the manuscript.

REFERENCES

Abbey, S., 1975. Studies in "standard samples" of silicate rocks andminerals. Part 4: 1974 edition of "usable" values. Can. Geol.Surv. Pap., 74-41.

Albee, A. L., and Ray, L., 1970. Correction factors for electron probemicroanalysis of silicates, oxides, carbonates, phosphates, andsulfates. Anal. Chem., 42:1409-1414.

Allegre, C. J., Trueil, M., Minster, J. F., et al., 1977. Systematic useof trace element in igneous process. Part I: Fractional crystalliza-tion processes in volcanic suites. Contrib. Mineral. Petrol.,60:57-75.

Anonymous, 1977. Initial report of the geological study of oceaniccrust of the Philippine sea floor. Ofioliti, 2:137-168.

Bass, M. N., 1976. Secondary minerals in oceanic basalt, with specialreference to Leg 34, Deep Sea Drilling Project. In Barker, P. F.,Dalziel, I. W. D., et. al., Init. Repts. DSDP, 36: Washington (U.S.Govt. Printing Office), 393-432.

Bass, M. N.,, Moberly, R., Rhodes, J. M., et al., 1973. Volcanic rockscored in the central Pacific, Leg 17, Deep Sea Drilling Project. InWinterer, E. L., Ewing, J. I., et al., Init. Repts. DSDP, 17:Washington (U.S. Govt. Printing Office), 429-503.

Bence, A. E., and Albee, A. L., 1968. Empirical correction factors forthe electron microanalysis of silicates and oxides. J. Geol., 76:382-403.

Bougault, H., Cambon, P., Corre, O., et al., 1979. Evidence for vari-ability of magmatic processes and upper mantle heterogeneity inthe axial region of the Mid-Atlantic Ridge near 22° and 36°N.Tectonophysics, 55:1-34.

Brown, R. W., 1976. A sample fusion technique for whole rock analy-sis with the electronic microprobe. Geochim. Cosmochim. Acta,41:435-438.

Cameron, W. E., Nisbet, E. G., and Dietrich, V. J., 1979. Boninites,komatiites and ophiolitic basalts. Nature, 280:550-553.

Cloud, P. E., Jr., Schmidt, R. G., and Burke, H. W., 1956. Geologyof Saipan, Mariana Islands. U.S. Geol. Surv. Prof. Pap., 280-A.

Coish, R. A., and Church, W. R., 1979. Igneous geochemistry ofmafic rocks in the Betts Cove Ophiolite, Newfoundland. Contrib.Mineral. Petrol., 70:29-39.

Colley, H., and Warden, A. J., 1974. Petrology of the New Hebrides.Geol. Soc, Am. Bull., 85:1635-1646.

Dallwitz, W. B., Green, D. H., and Thompson, J. E., 1966. Clinoen-statite in a volcanic rock from the Cape Vogel area, Papua. J.Petrol., 7:375-403.

Delaney, J. R., Meunow, D., and Graham, D. G., 1978. Abundanceand distribution of water, carbon and sulfur in the glassy rimsof submarine pillow basalts, Geochim. Cosmochim. Acta, 42:581-594

Delany, J. M., and Helgeson, H. C , 1978. Calculation of the thermo-dynamic consequences of dehydration in subducting oceanic crustto 100 kbar and >800°C. Am. J. Sci., 278:638-686.

Dewey, J. F., and Bird, J. M., 1970. Mountain belts and the newglobal tectonics. J. Geophys. Res., 75:2625-2647.

Dickinson, W. R., and Seely, D. R., 1979. Structure and stratigraphyof forearc regions. Am. Assoc. Petrol. Geol. Bull., 63:2-31.

Dietrich, V., Emmerman, R., Oberhansli, R., et al., 1978. Geochem-istry of basaltic and gabbroic rocks from the West Mariana Basinand the Mariana Trench. Earth Planet. Sci. Lett., 39:127-144.

Ewart, A., Brothers, R. N., and Mateen, A., 1977. An outline of thegeology and geochemistry, and the possible petrogenetic evolutionof the volcanic rocks of the Tonga-Kermadec-New Zealand islandarc. /. Volcan. Geotherm. Res., 2:205-250.

Ewart, A., and Bryan, W. B., 1972. The petrology and geochemistryof the igneous rocks from Eua, Tongan Islands. Geol. Soc. Am.Bull., 83:3281-3298.

Green, D. H., 1976. Experimental testing of "equilibrium" partialmelting of peridotite under water-saturated, high-pressure condi-tions. Can. Mineral., 14:255-268.

Harris, D. M., 1979a. Geobarometry and geothermometry of indi-vidual crystals using H2O, CO2, S, and major element concentra-tions in silicate melt inclusions: 2. The 1959 eruption of KilaueaVolcano, Hawaii. Geol. Soc. Am. Abstracts with Programs, 11:439.

, 1979b. Pre-eruption variations of H2O, S, and Cl in a sub-duction zone basalt. EOS, 60:968.

Hart, S. R., and Davis, K. E., 1978. Nickel partitioning betweenolivine and silica melt. Earth Planet. Sci. Lett., 40:203-219.

Jakes, P., and White, A. J. R., 1972. Major and trace element abun-dances in volcanic rocks of orogenic areas. Geol. Soc. Am. Bull.,83:29-40.

Johanssen, A., 1939. A Descriptive Petrography of the IgneousRocks (Vol. 1): Chicago (University of Chicago Press).

Kang, D. E., 1974. Tectonic erosion at trenches. Earth Planet. Sci.Lett., 21:209-212.

Karig, D. E., and Sharman, G. F., Ill, 1975. Subduction and accre-tion in trenches. Geol. Soc. Am. Bull., 86:377-389.

Kuroda, M., and Shiraki, K., 1975. Boninite and related rocks of Chi-chi-jima, Bonin Islands, Japan. Rep. Fac. Sci. Shizuoka Univ.,10:145-155.

Macdonald, G. A., 1960. Dissimilarity of continental and oceanicrock types. / . Petrol., 1:172-177.

Masuda, A., 1968. Geochemistry of lanthanides in basalts of centralJapan. Earth Planet. Sci. Lett., 4:284-292.

Meijer, A., 1980. Primitive arc volcanism and a boninite series. InHayes, D. (Ed.), The Tectonic and Geophysical Evolution of South-east Asian Seas and Islands: Geophysical Monograph 23: Wash-ington (American Geophysical Union), 269-282.

Moore, J. G., 1970. Water content of basalt erupted on the oceanfloor. Contrib. Mineral. Petrol., 28:272-279.

Scheidegger, K. F., and Stakes, D. S., 1977. Mineralogy, chemistryand crystallization sequence of clay minerals and altered tholeiiticbasalts from the Peru Trench. Earth Planet. Sci. Lett., 36(3):413-422.

Schmidt, R. G., 1957. Petrology of the volcanic rocks. Geology ofSaipan, Mariana Islands (Pt. 2): U.S. Geol. Surv. Prof. Pap.280-B, 127-175.

Seely, D. R., 1978. The evolution of structural highs bordering majorforearc basins. Am. Assoc. Petrol. Geol. Mem., 29:245-260.

Sharaskin, A. Ya., Dobretsov, N. L., and Soboler, N. V., in press.Marianites: The clinoenstatite bearing pillow lavas associated withophiolite assemblage of the Mariana Trench. Proc. Int. OphioliteSymp. Cyprus.

Smith, R. E., and Smith, S. E., 1976. Comments on the use of Ti, Zr,Y, Sr, K, P, and Nb in classification of basaltic magmas. EarthPlanet. Sci. Lett., 32:114-120.

Sutter, J. F., and Snee, L. W., 1978. K/Ar and 40Ar/39Ar dating ofbasaltic rocks from DSDP Leg 59. Eos, 59:1179.

Tracey, J. I., Schlanger, S. O., Stark, J. T., et al., 1963. General geol-ogy of Guam. U.S. Geol. Surv. Prof. Pap. 403-A.

Tsuya, H., 1937. On the volcanism of the Huzi volcanic zone with spe-cial reference to the geology and petrology of Izu and the southernislands. Bull. Earthq. Res. Inst., 15:215-357.

APPENDIXAnalytical Procedures

Phenocryst and bulk rock major element compositions were deter-mined using an ARL-SEMQ electronic microprobe. The acceleratingvoltage was 15 kv and the sample current (on benitoite) 20 to 30nanoamperes. Bence-Albee corrections (Bence and Albee, 1968; Albeeand Ray, 1970) were applied to all analyses.

Phenocryst and glass analyses were performed on polished thinsections. A beam size of 30 µm was used for analyses of glass and largephenocrysts and 5 µm for microphenocrysts. A natural basaltic glassfrom the Juan de Fuca Ridge (United States National Museum111240/52) was used as a calibration sample for all elements. Preci-sion was monitored by analyzing the Juan de Fuca Ridge glass at thebeginning and end of each day and by re-analyzing selected pheno-crysts on successive days. Comparison of values, using the equation(1st value - 2nd value/1st value) × 100, showed that analyses were

723


generally within a few percent. Accuracy was monitored by analyzingminerals of known chemical composition, usually Kakanu horn-blende, after calibration during each microprobe session. The resultsof typical analyses are presented in Table 1.

The hand-picked chill rind separates were analyzed in the follow-ing manner. Samples were crushed and 20 mg of the powder fused in aMo strip furnace at 1750°C under Ar at atmospheric pressure (seeBrown, 1977, for a discussion of fusion techniques). The fused glasseswere then mounted and analyzed using a 30-µm beam. Two to sixanalyses of each fused sample were average to minimize the effects ofcompositional heterogeneities in the glasses. The Juan de Fuca Ridge

glass was used as a calibration sample for all elements exceptpotassium and titanium, for which the Kakanui hornblende (USNM143965) was used. Accuracy was monitored by fusing and analyzingUSGS rock standards BCR-1 and W-l. The values obtained for all themajor elements are usually within 5% of the reported " d r y " values(Table 1). There has been some debate over whether the fusion tech-nique and analysis of glasses using the electron microprobe causevolatization and loss of some elements, particularly sodium. Althoughwe did observe lower sodium values in the fused USGS rock powdersthan are reported, the problem was not substantial, as is indicated bythe data reported in Table 7.

Table 1. Estimates of accuracy for mineral and fused glass analyses of standards.

SiO2

TiO 2

A12O3

FeO t

MnOMgOCaONa 2 OK 2OP 2 O 5

Total

Kakanuihornblende

12-18-79

41.274.48

15.0310.51n.d.12.7210.142.712.0

n.d.

98.87

Δ % a

2.235.090.873.76

0.631.554.232.44—

-

Kakanuihornblende

11-29-79

41.024.63

15.0610.68n.d.12.3110.322.661.98

n.d.

98.66

Δ %

1.611.911.072.2

3.830.192.313.42—

-

Kakanuihornblendeav. Δ % b

0.943.852.033.73

4.70.762.974.39

—

-

KakanuihornblendeReported0

40.374.72

14.910.920.09

12.810.32.62.05n.d.

99.08

BCR-18-30-79

54.622.23

14.2112.19n.d.3.377.013.111.53

n.d.

98.27

Δ %

1.270.452.970.73—

4.260.436.339.47

—

-

BCR-1av. Δ % d

1.154.584.204.60—

8.293.675.674.76

—

—

BCR-1Report ed e

55.322.24

13.812.18

—3.527.043.321.69—

99.21

W-l10-9-79

52.750.93

16.1310.27

n.d.6.71

11.012.290.61n.d.

100.7

Δ %

0.5713.897.822:19—0.60.366.024.69—

—

W-lav. Δ W

1.0910.957.022.92—5.981.447.825.29—

—

W-lReported6

53.051.08

14.9610.05

—6.67

11.052.160.64—

99.66

a Δ % = ([reported value - observed value>/reported value] × 100.b Average of 7 analyses.c Personal communication, Mr. Joe Nelen, Smithsonian Institution (USNM), 1978.d Average of 6 analyses.e Abbey (1975) corrected for volatile loss.f Average of two analyses.

724


Plate 1. Photomicrographs of bronzite andesites, Hole 458.

Figure 1. Sample 458-30-1, 101-104 cm. Clinopyroxene crystallites inchill rind of bronzite andesite. Bladed habit characteristic. Noteskeletal "cocks comb" quench crystals. Lack of Fe-Ti-oxides andPlagioclase notable. Glassy mesostasis is pale brown owing toslight devitrification. Fresh glass is clear. Width of field 0.78 mm,plane light.

Figure 2. Sample 458-30-1, 101-104 cm. Same as Figure 1 but undercrossed nicols.

Figure 3. Sample 458-30-1, 101-104 cm. Orthopyroxene (bronzite)and clinopyroxene phenocrysts in chill rind of bronzite andesite.Groundmass as in Figures 1 and 2. Width of field 7.75 mm,crossed nicols.

Figure 4. Sample 458-28-1, 142-146 cm. Outgassing "clots" in upperleft corner and bottom center. Extremely fine grained Plagioclasein clots with clinopyroxene and an Fe-Ti-oxide phase which pro-vides color contrast. Large vesicles characteristic of chill rinds inthese lavas. Width of field 7.75 mm, plane light.

Figure 5. "Swallow tail" microlite found in sediments overlying vol-canic section in Hole 458. Width of field approximately 0.25 mm.

725


Plate 2. Photomicrographs of bronzite andesite, Hole 458.

Figure 1. Sample 458-28-1, 9-13 cm. Fine-grained pillow interior(?)showing typical high vesicularity. Compare with "clots" in Plate1. Note vesicles are empty. Width of field 7.75 mm, plane light .

Figure 2. Sample 458-31-1, 2-7 cm. Quartz-Fe-Ti-oxide-plagioclase"spherulite" in intersertal textured 2-pyroxene andesite. This par-

ticular "spherulite" larger than most others observed. Mesostasissurrounding "spherulite" is dusty pale brown slightly devitrifiedglass. Width of field 0.78 mm, plane light.

Figure 3. Sample 458-31-1, 2-7 cm. Same as Figure 2 but crossednicols. Note "spherulite" nucleated on Plagioclase crystal.

Figure 4. Sample 458-28-1, 142-146 cm. Complex zoning of ortho-pyroxene in bronzite andesite. With of field 0.78 mm, crossednicols.

726


Plate 3. Spherulites, Fe-Ti-oxides, and alteration minerals in igneousrocks, Hole 458.

Figure 1. Sample 458-28-1, 9-13 cm. Spherulitic Plagioclase withclinopyroxene. Width of field 1.94 mm, crossed nicols.

Figure 2. Sample 458-46-1, 100-106 cm. Clay-filled vesicle in Unit5 chill rind. Adjacent to vesicle wall is a layer of light clay which issharply bounded by brownish green clay in interior. Remainder ofphoto shows Plagioclase microlites and microphenocrysts in palag-onitized and unpalagonitized glass. Width of field 3.8 mm, planelight.

Figure 3. Sample 458-43-2, 34-37 cm. Zeolite-filled vesicles in chillrind of bronzite andesite. Note zeolites are superimposed on clayvesicle lining. Analysis of zeolite in lower left vesicle in Table 5.Analysis of skeletal orthopyroxene crystal just below upper leftvesicle in Table 1. Most of photo pale brown to clear glass withclinopyroxene microlites. Analysis of glass in Table 4. Width offield 3.8 mm, plane light.

Figure 4. Sample 458-33-1, 58-60 cm. Pseudomorph of clay afterorthopyroxene in subophitic textured andesite. Note lack of altera-tion of clinopyroxene in upper center and mantling orthpyroxenepseudomorph. Remainder of photo Plagioclase, quartz, and inter-stitial clay. Width of field 0.78, crossed nicols.

727


Plate 4. Quartz-alkali feldspar intergrowths and textures of basalts,Hole 459B.

Figure 1. Sample 459B-66-3, 3-4 cm. Micrographic intergrowth ofquartz and feldspar in quartz-bearing tholeiitic andesite or quartzdiabase. Note discrete quartz grains in upper left. Minor clinopy-roxene in bottom center. Interstices between blocky feldspars linedwith celadonite and clay. Width of field 1.94 mm, crossed nicols.

Figure 2. Sample 459-71-1, 142-145 cm. Fe-Ti-oxide granules in in-tersertal texture characteristic of Petrographic Unit 5. Otherphases include Plagioclase, clinopyroxene, and altered mesostasis.Width of field 0.15 mm, plane light.

Figure 3. Sample 459B-73-1, 6-9 cm. Typical diktytaxitic texture ofPlagioclase laths in Petrographic Unit 4. Most of interstitial andclinopyroxene material altered to clay. Note presence of Fe-Ti-oxide grains. Width of field 7.75 mm, plane light.

Figure 4. Sample 459B-62-1, 44-46 cm. Clinopyroxene-plagioclaseandesite. Width of field 7.75 mm, crossed nicols.

Figure 5. Sample 459B-67-2, 106-109 cm. Cristobalite "buttons" inorthopyroxene-clinopyroxene andesite. Note skeletal and anhedralFe-Ti-oxide grains. Cristobalite surrounded by devitrified meso-stasis. Width of field 0.6 mm, plane light.

728


Plate 5. Photomicrographs of volcanic rocks from the MarianaTrench, Holes 460, 460A, 461, and 461 A.

Figure 1. Sample 460-8-1, 148-150 cm. Highly vesicular pyroxene an-desite(?). Compare Plate 2, Figure 1. Width of field 1.94 mm,plane light.

Figure 2. Sample 460A-11-1, 20-26 cm. Pyroxene-plagioclase ande-site(?). Mesostasis is brown glass slightly devitrified. Vesicles lined

with palagonite-like substance. Width of field 1.94 mm, planelight.

Figure 3. Sample 461-3-1, 62-64 cm. Quartz-rich pyroxene ande-site(?). Some primary clinopyroxene and Plagioclase but most ofphoto comprised of secondary phase including quartz, clay, andchlorite. Width of field 1.94 mm, crossed nicols.

Figure 4. Sample 461A-2-1, 85-86 cm. Vesicular pillow rind. Palebrown unaltered glass with clay-filled vesicles and fractures. Singlemicrophenocrysts of clinopyroxene (lower center) and Plagioclaseupper left. Width of field 1.94 mm, plane light.

729

38. petrology of volcanic rocks from the fore-arc … · 2007-05-08 · 38. petrology of volcanic...

Documents