Taylor, B., Fujioka, K., et al., 1992Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 126

11. GLASS INCLUSIONS WITH MICROGLOBULES IN PLAGIOCLASE AND PYROXENEPHENOCRYSTS OF VOLCANIC ROCKS FROM THE BONIN ARC, LEG 1261

Teruo Watanabe,2 Tadao Hirama,3 Makoto Yuasa,4 Shoichi Terada,2 and Kantaro Fujioka5

ABSTRACT

The chemical composition of glass inclusions in phenocrystic Plagioclase and pyroxene from Sites 792 and 793, drilled duringOcean Drilling Program Leg 126 in the Bonin Arc, is examined. Immiscible liquid, which is preserved as glass inclusions withunmixed textures in Plagioclase, is observed in a high-magnesian andesite, which suggests an important role of liquid immiscibilityin the fractionation of high-magnesian andesite. In other andesitic rocks (SiO2 = 57-60 wt%), such unmixed textures of glassinclusions in calcic Plagioclase with a similar percentage of An (around 80%) is not found. The degree of fractionation and mixingof liquid are inferred from the glass composition in pyroxene.

INTRODUCTION

Komatsu and Yajima's (1970) electron microprobe analysis (EMPA)of the chemical composition of glass inclusions in phenocrysts ofvolcanic rocks revealed the existence of rock series and residue thatresulted from the crystallization of trapped liquid in host phenocrysts.Anderson (1973) examined glass inclusions and discussed their sig-nificance in terms of the volatile components of magma. Anderson(1976) also studied glasses in matrix and glass inclusions of volcanicrocks and stressed the important role of glasses as a tool for the studyof magma mixing. Watson (1976) estimated the amount of liquidcomposition in a South Atlantic basalt on the basis of composition ofglass inclusions. Later, Roedder (1979) discussed comprehensivelythe origin and significance of glass inclusions during the crystal-lization of phenocrystic crystals. He also described the microglobulesof an immiscible sulfide phase in glass inclusions.

Primary glass inclusions in the phenocrystic crystals reflect apartial compositional trend of evolved magmatic liquid. For example,the compositional variation of glass inclusions plotted on a CaO-MgO-Al2O3 diagram indicate the trend produced by mineral frac-tionation (Watson, 1976). Also, lower Mg/(Mg + Fe) and higher TiO2

in inclusions indicate the existence of olivine fractionation (Dunganand Rhodes, 1978) or ferromagnesian mineral crystallization(Clocchiatti and Massare, 1985). C1/K2O ratios in glass inclusionsindicate the degree of differentiation (Anderson, 1982), and theCaO/Na2O ratio in glass inclusions of olivine can be used to infer theratio in the primary magma (Falloon and Green, 1986). Dungan andRhodes (1978) presented evidence for magma mixing on the basis oftheir study of residual glasses and melt inclusions in basalts fromDeep Sea Drilling Project (DSDP) Legs 45 and 46.

The study of volatiles in glass inclusions will reveal the gascomposition in magma, which in turn will contribute to the under-standing of the mechanism of eruption and the effect of the gascomposition on the atmosphere (Saitoh and Kusakabe, 1989). Assummarized by Shinohara (1990), total volatiles increase with SiO2

contents in melt inclusions, sulfur decreases with the SiO2 contents,and Cl contents may reach a maximum at intermediate SiO2 contents(around 60%). Thus, the accumulation of data on the occurrence and

Taylor, B., Fujioka, K., et al., 1992. Proc. ODP, Sci. Results, 126: College Station,TX (Ocean Drilling Program).

2 Department of Geology and Mineralogy, Hokkaido University, Sapporo 060, Japan.3 Department of Earth Sciences, Hokkaido University, Sapporo 060, Japan.4 Geological Survey of Japan, Tsukuba, Ibaraki 305, Japan.5 Ocean Research Institute, University of Tokyo, 1-15-1 Minamidai, Nakano, Tokyo

164, Japan (present address: Japan Marine Science and Technology Center, 2-14 Nat-sushima, Yokosuka, Kanagowa 238, Japan).

composition of glass inclusions is important for understanding thechemistry of magma and the crystallization process.

We describe in this paper the preliminary results from our studyof the glass inclusions, mostly <0.5 mm in diameter, in Plagioclaseand pyroxene phenocrysts of arc-volcanic rocks. Microglobules arecommonly found in the glass inclusions, but we have not yet foundfluid inclusions in the glass inclusion samples that are described inthis paper and were collected from Holes 792E and 793B in the BoninArc area.

TECTONIC SETTING

The locations of Sites 792 and 793 are shown in Figure 1. Site 792is approximately 100 km from Hichijo-jima to the southeast. Site 793is located to the northeast of Torishima. Tectonically, Site 792 issituated on the northern extension of the frontal-arc high of Bonin Arc,and Site 793 is located in the forearc basin (Honza and Tamaki, 1985).From the samples obtained during Leg 126, we selected four pieces(Samples A, B, C, and D) that commonly contain glass inclusions withmicroglobules in Plagioclase and pyroxene phenocrysts. Sample A isfrom Hole 793B and the other samples are from Hole 792E. Samples A,B, and C were collected from the acoustic basement underlyingOligocene rocks, and Sample D is from upper Oligocene rocks richin volcanic pebbles.

DESCRIPTION OF ROCKS

A brief description of the four samples, based on the Leg 126Initial Reports and our study, are presented here.

Sample A

Sample 126-793B-105R-1, 74-78 cm (porphyritic andesite), ispart of a thick pillow lava. It has the same lithology of analyzed rockslisted in the Initial Reports (Sample 126-793B-105R-1,127-131 cm).The composition of the latter is shown in Table 1 and is that ofhigh-magnesian andesite of the arc-tholeiite series. The analyzed rockfrom the basement at Site 793 was situated around 4555 m below sealevel (mbsl). The rock contains phenocrysts of orthopyroxene(approximately 5%, 0.5-3 mm, euhedral), clinopyroxene (10%—15%,2-6 mm, euhedral, stellate clusters), and Plagioclase (approximately2%, 0.5-1.0 mm, euhedral).

Amygdules, 2-20 mm in diameter, are filled with zeolite andnative copper (Taylor, Fujioka, et al., 1990). Zeolite minerals areconfirmed as mordenite and dachiardite by X-ray diffraction analysis.Maghemite is contained in the rock.

171

T. WATANABE ET AL.

Honshu, Japan

34°N

° v-~

Table 1. Chemical composition of volcanic rocks from Sites792 and 793.

140°E 142°

Figure 1. Bathymetric map of the Izu-Bonin Arc between the Shikoku Basinand Izu-Bonin Trench and location of Sites 792 and 793 (after Taylor, Fujioka,et al, 1990).

Sample B

Sample 126-792E-74R-1, 70-72 cm, is an andesitic hyaloclasticbreccia that contains plagioclase-orthopyroxene-clinopyroxene an-desite and plagioclase-clinopyroxene-quartz-dacite fragments in ahydrothermally altered glassy matrix. The andesitic hyaloclastiteoverlies a porphyritic andesite layer, and the andesitic blocks arethought to be compositionally equivalent to the porphyritic andesite,the composition of which is listed in Table 1 (Sample 126-792E-74R-2, 53-57 cm). Plagioclase is the most predominant phenocrysticmineral in the sample, which was located at 2672 mbsl (water depth,1787.7 m).

Sample C

Sample 126-792E-71R-1, 107-112 cm (porphyritic andesite), con-sists of part of a porphyritic andesite unit intercalating thin hyaloclas-tic layers. Under the microscope, the petrographical characteristics ofSample C are equivalent to an andesite analyzed by X-ray fluores-cence that was described in the Initial Reports (Table 1, Sample 126-793B-71R-1, 9-13 cm). The andesite consists of phenocrysts ofPlagioclase (35%, 0.2-5 mm, euhedral, zoned, and glomeropor-phyritic clots), clinopyroxene (5%, 0.1-5 mm, euhedral), orthopy-roxene (10%, 0.1-5 mm, euhedral, mostly replaced by smectite), andmagnetite (1%, 0.1-0.5 mm, subhedral). The sample was situatedabout 2599 mbsl (water depth, 1787.7 m).

SampleHoleCore, sectionInterval (cm)

Major elements (wt%):SiO2

TiO2

A12O3

FeOMnOMgOCaONa2OK2OP2O5

LOI

Total

Trace elements (ppm):ZnYRbSrBaVNbNiCrZnCuCe

A793B

105R-1127-131

57.000.28

13.079.260.158.53

10.431.820.440.012.57

100.99

28.89.63.8

12717.3

251.50.4

36.6249.2

82.77.2—

B792E74R-253-57

59.910.54

17.507.170.053.418.212.570.200.041.44

99.60

39182.4

16628

2420.660

7939

7

C792E74R-19-13

57.480.54

18.926.900.153.289.672.310.230.111.01

99.59

4020

1.417631

2580.99

128634

2

D792E47R-115-20

50.140.80

20.946.760.405.56

10.173.411.231.077.04

100.48

4122

217420

2560.7

10139326

2

Notes: Data are from Taylor, Fujioka, et al. (1990). LOI = loss on ignitionand dash (—) = below detection level.

Sample D

Sample 126-792E-47R-3, 1-6 cm (volcanic-lithic conglomerateand granule-rich vitric sandstone) consists of part of an upper Oligo-cene strata. The sedimentary unit consists up to 20% volcanic pebbles.One of the pebbles displays the basaltic composition shown in Table 1(Sample 126-792E-47R-1, 15-20 cm).

ANALYTICAL TECHNIQUEFOR GLASS ANALYSIS

The chemical composition of glass inclusions was obtained byEMPA at the Analytical Laboratory of the Department of Geology andMineralogy, Hokkaido University. Accelerating voltage was 15 kV,beam current was 2 × I0"8 Å, and beam size was 2 µm in diameter onpericlase for small glass inclusions. In this analytical condition, lossof sodium and potassium counts was very significant. Therefore, weestimated sodium and potassium contents on the basis of Ono et al.'s(1976) method (i.e., the contents were inferred from decreasingcurves of counts per second as shown in Fig. 2). Calibration curvesfor sodium (Fig. 3) and potassium between the weight percentage andcounts were obtained by standard glass analysis. As already men-tioned by Ono et al. (1976), loss of sodium was most significant forglass with high SiO2 contents. However, a 10-µm beam for Iow-Siθ2

glass (around 55%) did not generate any distinct loss of sodium andpotassium. The composition of glass inclusions often yielded lowtotals, as listed in Tables 2-5. By means of energy-dispersive-systems(EDS) analysis, we were able to determine that no other majorelements except for chlorine (<0.5%) were present. Sulfur was notpositively detected. Hence, we infer that the glass contains only a fewpercent of water and very small amounts of chlorine. The very fineparts of inclusions were examined by EDS at the Section of HistoricalEarth Dynamics of the Department of Geology and Mineralogy,

172

200 η

150

100

50

0

cps

Nain Plagioclase

Sample A

0

200cps

10 15 20 25 30

in orthopyroxene

Sample A

250

200 ]

150

100 \

50

0

Na

in orthopyroxene

Sample A(beam size =

10 15 20 25

150 η

100

in Plagioclase

Sample B

GLASS INCLUSIONS IN PHENOCRYSTS

in Plagioclase

Sample B

350

300

250

200

150

100

50-I

0

Λ.Na

in Plagioclase

Sample C

0 10 15 20

in clinopyroxene

Sample D

in clinopyroxene

Sample D(beam size = QµmΦ)

10

Figure 2. Representative graphs showing intensity variation (counts per second) of sodium and potassium with time (seconds)

for glass inclusions in Sample A, B, C, and D.

173

T. WATANABE ET AL.

100 200 300

Na cps400

2.2150 160 170 180

Na cps190 200

Figure 3. Graphs showing calibration line of sodium contents (wt%) for thesodium counts (cps). The upper graph is for acidic glass, the lower graph forintermediate glass. Potassium calibration line accidentally coincides with thatof sodium.

Hokkaido University. In this case, accelerating voltage was 10 or 20 kVand beam current 2 × I0"9 Å. In the glass inclusions, microglobuleswere observed under the microscope in all of the samples. They arethought to have been bubbles filled with a gaseous phase, but they arenow mostly filled by polishing powder or dust under the microscope.Therefore, it is impossible to estimate their composition.

OCCURRENCE AND CHEMICAL COMPOSITIONOF GLASS INCLUSIONS

Sample A

Plagioclase contains many irregularly shaped (including rectangu-lar, square, and round) glass inclusions (Plate 1, Figs. 1, 2, and 4). Weanalyzed a few glass inclusions in Plagioclase phenocrysts. Most of theinclusions solidified to brown glass and sometimes they had darkquench crystals. The inclusions often showed unmixed textures, acomplicated pattern of dark- and light-colored areas such as myrme-kite (Plate 1, Figs. 1 and 2). Each area was too fine to analyze.Therefore, one analysis of a glass inclusion (glass in Plagioclase,Table 2) included both unmixed parts.

By means of semiquantitative EDS analysis, it was revealed that onepart of the sample (white area in Plate 1, Fig. 2) contains high Mg, Fe,and low Si as compared with the dark part, which has a somewhat similarcomposition to the glass inclusions in the Plagioclase of Samples B andC. The lack of mixing may indicate immiscible liquids in Plagioclasephenocrysts (Philpotts, 1981, 1982). Around the unmixed glass inclu-sions, small, clear silicic glass inclusions with microglobules occur (Plate 1,Fig. 1). These inclusions seem to have been trapped later than the formerinclusions, showing an unmixed texture. The mechanism that forms twokinds of inclusions is not fully understood.

Approximately oval-shaped, dispersed, and comparatively largeglass inclusions were observed in pyroxene (Plate 1, Fig. 3). Theinclusions often contain a single microglobule in each and are palebrownish in color. The glass composition (Table 2) is different fromnormal dacite and/or host andesite, which suggests that the glass is aresidue of trapped liquid during pyroxene crystallization. K2O occursselectively in glass inclusions.

The Mg/(Mg + Fe) ratios in the inclusions are variable, as thesewere affected during pyroxene crystallization, but the Ca/(Ca + Na)ratios are comparatively constant. The high ratio is equivalent to thatof a liquidus composition formed by the crystallization of calcicPlagioclase (An 80%-85%).

Sample B

Glass inclusions were commonly observed in Plagioclase phe-nocrysts; they are mostly oval-shaped or rectangular. They occurirregularly along the inner parts of Plagioclase and are arranged alonga crystal plane (Plate 1, Fig. 5). The range of An in the host Plagioclasevaries from 74% to 85%, and the An of the Plagioclase adjacent toinner glass inclusions is as low as 74%. Similar enhancement ofsodium around glass inclusions was described by Watson (1976). Asshown in Plate 1, Figure 5, a rhythmical zoning pattern was observedonly in the part of the crystal without glass inclusions-mostly theouter rim. A minute grain of clinopyroxene occurs within a glassinclusion. Glass inclusions are compositionally high in SiO2 and K2Ocontents (Table 3). Ca/(Ca + Na) ratios in the inclusions are variable(0.2-0.32), but the Mg/(Mg + Fe) ratios are rather constant (0.27-0.30). These occurrences and compositional characteristics of theglass inclusions suggest that they were trapped during the crystalgrowth stage of Plagioclase and represent a residual composition afterreaction with host Plagioclase.

Sample C

In this sample, glass inclusions were also found in the Plagioclase;oval-shaped and coarse inclusions occur mostly on the inner side ofthe Plagioclase, which suggests rapid crystal growth of the innerPlagioclase.

A microglobule in each inclusion occurred in the central part ofPlagioclase (Plate 2, Figs. 1 and 2) and can occupy over 20% of theinclusions in terms of volume. These inclusions are coarser than thosein the marginal inclusions and in pyroxene. This suggests that gaseousphases were comparatively predominant during the early crystal-lization stage of the Plagioclase. The composition is similar to that ofSample B.

Sample D

A coarse clinopyroxene crystal in Sample D includes glass inclu-sions that are arranged in inner and outer rows (Plate 2, Fig. 3). Theclinopyroxene is compositionally rather homogeneous, but adjacentto areas near the outer inclusions, the pyroxene is slightly rich in CaO(numbers 3,4, and 14 in Fig. 4). The pyroxene also contains anhedral

174

GLASS INCLUSIONS IN PHENOCRYSTS

Plagioclase inclusions. The compositions of the glass inclusions areclassified into two groups on the basis of their TiO2 and CaO contentsand Mg/(Mg + Fe) ratio (Fig. 5). The inner inclusions are high in TiO2

and low in CaO and Mg/(Mg + Fe), indicating evolved magmaticliquid, as compared with the outer inclusions. Although these com-ponents and their ratio may be influenced during the crystallizationof host clinopyroxene, their compositions seem not to have beensubstantially modified during crystallization. The glass inclusionsplot on an extension of a fractionation trend on the CaO-MgO-Al2O3

diagram (Fig. 6) of Watson (1976). The outer inclusions, which weretrapped at a later stage of pyroxene crystallization, are higher in CaOand Mg/(Mg + Fe) and lower in TiO2. This is a paradox for closed-system crystallization models, for the compositional changes in theresidual liquid caused by the crystallization of clinopyroxene andother ferromagnesian minerals would produce the opposite effect.

DISCUSSION AND CONCLUSIONS

As described by Philpotts (1982), silica-rich glass typicallyforms a continuous phase that encloses iron-rich globules. There-fore, unmixed glass as described in Sample A does not commonlyoccur. Immiscible liquid preserved in glass inclusions in a basaltas an exception was described by Philpotts (1981). He reportedthat magnesium-rich glass occurs in small spheres of clear silica-rich glass. The lack of mixing was interpreted as being a metas-table occurrence. However, immiscible liquid preserved as glassglobules in the mesostasis of volcanic glass is commonly observed,as described by Philpotts (1982).

The unmixed glasses indicate variable magma-mixing processes(Philpotts, 1982). However, unmixed glasses that occur in the Sample Ainclusions do not indicate magma mixing. This lack of mixing indi-cates that trapped glasses became differentiated as magnesium- andsilica-rich glass after Plagioclase crystallization. Such lack of mixingwas not observed in the andesite of Samples B and C. Thus, weconsider that liquid immiscibility plays an important role in thedifferentiation of a high-magnesian andesite. The groundmass inSample A forms a small-scale layering of light- and dark-coloredmatrix. This may also be evidence for the lack of mixing in glass.

The Ca/(Ca + Na) ratio of glass inclusions in pyroxene of SampleA is equivalent to that of the residue after calcic (An = around 80%)Plagioclase crystallization under low-pressure conditions (e.g., Fig. 7in Johannes, 1989, and so on). However, the ratio is not as high as

that of the glass inclusions in olivine phenocrysts from Tonga (Fallonand Green, 1986). This may indicate that in a high-magnesian andesitethe magma was rich in CaO and orthopyroxene crystallized earlierthan Plagioclase.

Occurrences of larger glass inclusions and comparatively largeglobules in them in the inner part of calcic Plagioclase from Sample Cand also Sample B suggest rapid growth of calcic Plagioclase in amagmatic liquid rich in gaseous components. The calcic compositionof the Plagioclase may result in part from the crystallization of a liquidrich in volatiles.

In a pyroxene of Sample D, outer glass inclusions suggest crystal-lization from a relatively unevolved magma. The compositions of bothinner and outer inclusions plot along a fractionation trend, as shown inFigure 5. Therefore, during the later crystallization stage of clinopy-roxene, the pyroxene was surrounded by less fractionated magma. Thiscould have resulted either from magma mixing or from gravitationalsinking of pyroxene to the bottom of the magma chamber.

In this chapter we have described the chemical composition ofglass inclusions in andesitic rocks. Some element ratios indicate thedegree of crystallization and provide information on the complicatedcrystallization processes.

The following conclusions result from our preliminary study:

1. Liquid immiscibility was observed in glass inclusions in pla-gioclase phenocrysts of a high-magnesian andesite, which suggeststhe important role of liquid immiscibility for the differentiation ofhigh-magnesian andesites.

2. Judging from the occurrence of glass inclusions with largerglobules, a higher vapor content is suggested during an early stage ofPlagioclase crystallization of andesite.

3. From an analysis of the glass inclusions of a clinopyroxene, at leasttwo different liquids appear to have been present as the crystal grew. Lessfractionated liquid was trapped in the outer portion of the pyroxene,indicating magma mixing before the later stages of crystal growth.

ACKNOWLEDGMENTS

The authors would like to express our sincere thanks to Dr. C. H.Langmuir, Lamont-Doherty Geological Observatory of ColumbiaUniversity, for his kind and valuable comments. We also would liketo extend our thanks to Dr. M. R. Fisk for his critical reading of ourmanuscript in an early stage, and Dr. J. Yajima, Geological Survey of

Atomic per cent Fe+Mn

Figure 4. Chemical composition of clinopyroxene plotted on a diopside-hedenbergite-enstatite-ferrosilite diagram. Numbers correspond to those in Table 5 andPlate 2, Figure 3.

175

T. WATANABE ET AL.

Table 2. Chemical composition of Plagioclase, orthopyroxene, and glass inclusions in Plagioclase and orthopyroxene in Sample A (Site 793,high-magnesian andesite).

Number

Major elements (wt%):SiO2

TiO2

A12O3

FeOMnOMgOCaONa2OK2O

Total

Mineral formula

Trace elements (ppm):SiAlTiFeMnMgCaNaK

Total

Ca/(Ca + Na)Mg/(Mg + Fe)

1

47.120.00

33.250.760.030.06

17.261.450.00

99.93

2.1701.8050.0000.0290.0010.0040.8520.1290.000

4.990

87

2

49.420.00

31.640.710.010.14

15.472.460.02

99.87

2.2651.7090.0000.0270.0010.0090.7600.2180.001

4.990

78

3

49.270.02

31.770.730.060.13

15.892.100.04

100.01

2.2561.7150.0010.0280.0020.0090.7800.1860.003

4.980

81

Plagioclase

4

48.320.01

32.630.710.000.11

16.651.680.01

100.12

O = 8

2.2141.7620.0000.0270.0000.0080.8180.1490.001

4.979

85

5

49.410.01

31.960.780.000.13

15.972.040.03

100.33

2.2551.7190.0000.0300.0000.0090.7810.1800.002

4.976

81

6

49.260.01

32.040.780.000.16

15.741.990.02

100.00

2.2541.7280.0000.0300.0000.0110.7710.1760.001

4.971

81

7

49.010.06

32.740.900.010.10

16.461.850.04

101.17

2.2231.7510.0020.0340.0000.0070.8000.1630.002

4.982

83

Orthopyroxene

55.120.080.84

15.350.40

26.311.920.040.00

100.06

0 = 6

1.9880.0360.0020.4630.0121.4150.0730.0020.000

3.991

80

Plagioclasein orthopyroxene

49.360.01

32.411.110.000.13

15.822.140.02

101.00

O = 8

2.2411.7350.0000.0420.0000.0090.7700.1800.001

4.978

75

Glassin Plagioclase

66.450.539.968.980.103.033.992.801.08

96.92

4438

Note: Chemical composition in numbers 1 through 7 shows the variation of a Plagioclase phenocryst from one side of the margin to the other through the central part (number 4).

Table 3. Chemical composition of Plagioclase and glass inclusions in Sample B (Site 792).

Number

Major elements (wt%):SiO2

TiO2

A12O3

FeOMnOMgOCaONa2OK2O

Total

Mineral formula

Trace elements (ppm):SiAlTiFeMnMgCaNaK

Total

Ca/(Ca + Na)Mg/(Mg + Fe)

1

49.530.00

31.560.650.020.10

15.822.540.00

100.22

2.2641.7010.0000.0250.0010.0070.7750.2250.000

4.998

78

2

47.040.00

32.860.510.000.05

16.981.590.01

99.04

2.1821.7970.0000.0200.0000.0030.8440.1430.001

4.990

85

3

49.140.00

32.030.610.040.05

15.972.380.01

100.23

2.2471.7270.0000.0230.0010.0030.7830.2110.000

4.995

79

4

50.050.05

31.590.590.030.03

15.162.900.03

100.43

2.2801.6960.0020.0220.0010.0020.7400.2560.001

5.000

74

Plagioclase

5

48.440.00

32.120.570.010.03

15.732.080.00

98.98

0 = 8

2.2391.7500.0000.0220.0000.0020.7790.1870.000

4.979

83

6

48.040.00

32.490.640.060.04

16.351.960.03

99.61

2.2141.7650.0000.0240.0020.0030.8070.1750.002

4.992

82

7

47.250.00

33.080.540.000.05

17.001.600.01

99.53

2.1811.8000.0000.0210.0000.0160.8410.1430.001

5.003

85

8

47.060.04

33.020.570.020.05

16.861.660.03

99.31

2.1791.8020.0010.0220.0010.0030.8360.1490.001

4.994

85

9

48.010.02

32.340.620.050.03

16.302.100.02

99.49

2.2161.7590.0010.0240.0020.0020.8060.1880.001

4.999

81

10

49.030.01

31.680.700.000.08

15.612.380.03

99.52

2.2561.7180.0000.0270.0000.0010.7700.2120.001

4.985

78

Note: Numbers 1-15 correspond to those in Plate 1, Figure 5.

176

Table 2 (continued).

Number

Major elements (wt%):SiO2

TiO2

A12O3

FeOMnOMgOCaONa2OK2O

Total

Mineral formula

Trace elements (ppm):SiAlTiFeMnMgCaNaK

Total

Ca/(Ca + Na)Mg/(Mg + Fe)

1

68.690.46

14.993.300.030.434.273.060.96

96.19

4419

Glass in orthopyroxene

2

67.640.40

15.293.790.030.494.642.950.83

96.06

4719

3

67.620.40

15.213.680.020.494.253.010.94

95.62

4419

4

67.410.38

14.525.070.05

' 0.984.453.020.87

96.75

4525

5

67.700.30

14.575.100.110.914.303.010.93

96.93

4624

GLASS INCLUSIONS IN PHENOCRYSTS

Table 4. Chemical composition of Plagioclase and glass inclusionsin Sample C (Site 792).

Number

Major elements (wt%):SiO2

TiO2

A12O3

FeOMnOMgOCaONa2OK2O

Total

Mineral formula

Trace elements (ppm):SiAlTiFeMnMgCaNaK

Total

Ca/(Ca + Na + K)Mg/(Mg + Fe)

1

49.100.01

31.740.480.030.05

15.622.420.02

99.47

2.2581.7210.0000.0180.0010.0030.7700.2160.001

4.988

78

Plagioclase

2

48.000.02

32.460.440.030.02

16.461.960.03

99.42

O =

2.2141.7650.0010.0170.0010.0020.8140.1760.002

4.992

82

3

46.830.03

33.080.580.010.15

17.361.370.00

99.41

= 8

2.1681.8050.0010.0220.0000.0100.8610.1230.000

4.990

88

4

48.690.01

31.840.600.020.15

15.862.070.01

99.25

2.2461.7310.0000.0230.0010.0100.7840.1850.001

4.981

81

Glass

5

77.670.688.673.580.120.841.803.811.11

98.28

1830

Note: Numbers 1-5 correspond to those in Plate 2, Figure 1.

Table 3 (continued).

Number

Major elements (wt%):SiO2

TiO2

A12O3

FeOMnOMgOCaONa2OK2O

Total

Mineral formula

Trace elements (ppm):SiAlTiFeMnMgCaNaK

Total

Ca/(Ca + Na)Mg/(Mg + Fe)

11

78.480.529.033.680.070.771.762.200.90

97.41

3027

12

75.910.46

10.083.290.070.722.144.460.91

98.04

2128

Glass

13

80.510.678.883.300.090.721.742.050.85

98.81

3228

14

79.360.529.443.160.100.751.773.050.93

99.08

2530

15

76.360.609.134.620.080.951.864.350.62

98.57

2027

Japan, for his kind comments. Dr. M. Nakagawa, Hokkaido Univer-sity, and Dr. K. Wada, Hokkaido University of Education, also gaveus useful suggestions. Encouragement and assistance from ProfessorY. Katsui, Professor S. Yui, Dr. J. Maeda, and Ms. M. Ujiie, HokkaidoUniversity, are also acknowledged.

REFERENCES

Anderson, A. T., 1973. The before-eruption water content of some highalumina magma. Bull. Volcanol., 37:530-552.

, 1982. Chlorine, sulfur and water in magma and oceans. Bull. Geol.Soc. Am., 85:1485-1492.

-, 1976. Magma mixing: petrological process and volcanological tool./. Volcanol. Geotherm. Res., 1:3-33.

Clocchiatti, R., and Massare, D., 1985. Experimental crystal growth in glassinclusions: the possibilities and limits of the method. Contrib. Mineral.Petrol, 89:193-204.

Dungan, M. A., and Rhodes, J. M., 1978. Residual glasses and melt inclusionsin basalts from DSDP Legs 45 and 46: evidence for magma mixing.Contrib. Mineral. Petrol, 67:417^31.

Falloon, T. J., and Green, D. H., 1986. Glass inclusions in magnesian olivinephenocrysts from Tonga: evidence for highly refractory parental magmasin the Tonga Arc. Earth Planet. Sci. Lett., 81:95-103.

Honza, E., and Tamaki, K., 1985. The Bonin Arc. In Nairn, A.E.M., Stehli, E G.,and Uyeda, S. (Eds.), The Ocean Basins and Margins (Vol. 7): The PacificOcean: New York (Plenum), 459-502.

Johannes, W., 1989. Melting of plagioclase-quartz assemblages at 2 kbar waterpressure. Contrib. Mineral. Petrol, 103:270-276.

Komatsu, M., and Yajima, J., 1970. Chemical composition of glass inclu-sions in the phenocrysts of some volcanic rocks. Proc. Jpn. Acad.,46:7672-7677.

Ono, K., Okumura, K., and Soya, T., 1976. Volcanic glass analysis by EPMA.Ann. Meeting Sanko-gakkai (Soc.) Abstr. Programs, 117.

177

T. WATANABE ET AL.

Table 5. Chemical composition of clinopyroxene and Plagioclase included in clinopyroxene and glass inclusions in Sample D (Site 792).

Number

Major elements (wt%):SiO2

TiO2

A12O3

FeOMnOMgOCaONa2OK2O

Total

Mineral formula

Trace elements (ppm):SiAlTiFeMnMgCaNaK

Total

1

52.530.422.439.000.32

15.9218.610.320.00

99.55

1.9480.1060.0120.2790.0100.8800.7400.0230.000

3.998

2

52.090.402.568.470.31

15.8518.750.240.00

98.67

1.9460.1130.0110.2650.0100.8830.7510.0180.000

3.997

3

' 52.570.332.617.000.28

16.1019.990.220.00

99.10

1.9470.1140.0090.2170.0090.8890.7940.0160.000

3.995

4

52.660.312.577.390.21

15.7520.24

0.240.00

99.37

1.9500.1120.0090.2290.0070.8690.8030.0230.000

4.002

5

52.310.372.039.000.28

15.9918.750.310.00

99.04

1.9530.0890.0110.2810.0090.8900.7500.0230.000

4.006

6

52.650.362.378.700.25

16.1319.210.240.00

99.91

1.9460.1030.0100.2690.0080.8880.7610.0170.000

4.002

Clinopyroxene

7

52.450.422.389.550.23

15.8118.110.280.00

99.23

1.9530.1050.0120.2970.0070.8780.7220.0200.000

3.994

8

52.440.412.129.430.33

16.0918.570.250.00

99.64

1.9480.0930.0120.2930.0100.8910.7390.0180.000

4.004

9

52.610.372.268.640.32

15.9918.640.280.00

99.11

1.9570.0990.0100.2690.0100.8860.7430.0200.000

3.994

10

52.210.542.59

10.050.34

15.8917.680.270.00

99.57

1.9410.1140.0150.3130.0110.8810.7040.0190.000

3.998

11

52.130.472.588.620.20

15.4719.290.240.00

99.00

1.9450.1140.0130.2690.0070.8600.7710.0180.000

3.997

12

52.540.422.129.070.25

16.0218.380.300.00

99.10

1.9570.0930.0120.2830.0080.8900.7340.0220.000

3.999

13

52.320.392.389.290.32

16.2118.340.270.00

99.52

1.9430.1040.0110.2890.0100.8970.7300.0190.000

4.003

14

51.670.423.607.230.22

15.7919.830.220.00

98.98

1.9190.1580.0120.2250.0070.8740.7890.0160.000

4.000

Note: Numbers 1-35 correspond to those in Plate 2, Figure 3.

Philpotts, A. R., 1981. Liquid immiscibility in silicate melt inclusions inPlagioclase phenocrysts. Bull. Mineral., 104:317-324.

, 1982. Compositions of immiscible liquids in volcanic rocks. Con-trib. Mineral. Petrol, 80:201-218.

Roedder, E., 1979. Origin and significance of magmatic inclusions. Bull.Mineral, 102:487-510.

Saitoh, G., and Kusakabe, M., 1989. Behavior of volatiles of magma duringvolcanic eruptions based on glass inclusion analysis: a review. Tech. Rep.Inst. Study Earth's Int. (ISEI), Ser. A, 27.

Shinohara, H., 1990. Behavior of volatiles in magma. Bull. Volcanol. Soc. Jpn.,Second Ser., Special Number, Basic Studies ofVolcanology, S99-S110.

Taylor, B., Fujioka, K., et al., 1990. Proc. ODP, Init. Repts., 126: CollegeStation, TX (Ocean Drilling Program).

Watson, E. B., 1976. Glass inclusions as samples of early magmatic liquid:determinative method and application to a South Atlantic basalt. J. Vol-canol. Geotherm. Res., 1:73-84.

Date of initial receipt: 2 January 1991Date of acceptance: 19 August 1991Ms 126B-126

178

GLASS INCLUSIONS IN PHENOCRYSTS

Table 5 (continued).

Number

Major elements (wt%):SiO2

TiO2

A12O3

FeOMnOMgOCaONa2OK2O

Total

Mineral formula

Trace elements (ppm):SiAlTiFeMnMgCaNaK

Total

Plagioclase

15

49.280.03

31.680.840.000.12

15.492.590.06

100.09

2.2571.7100.0010.0320.0000.0080.7600.2300.004

5.002

16

54.151.03

15.169.440.204.568.202.440.50

95.68

17

53.201.05

14.6510.180.254.988.512.280.38

95.48

18

53.560.98

15.0710.320.174.738.182.430.36

95.80

19

54.240.96

15.579.880.224.357.912.530.45

96.11

20

54.181.02

15.3210.370.204.418.122.780.42

96.82

21

53.290.92

14.8810.680.274.948.372.470.37

96.19

Glass in clinopyroxene

22

54.231.07

15.5010.690.174.237.862.950.41

97.11

23

55.280.90

16.068.720.224.468.522.550.52

97.23

24

53.860.93

15.1310.160.214.828.132.440.42

96.10

25

56.050.91

16.139.000.184.308.152.270.54

97.53

26

54.951.13

15.7710.590.253.437.383.000.52

97.02

27

55.031.17

15.4011.290.213.027.973.810.24

98.14

28

55.671.25

14.979.260.193.877.432.660.72

96.02

29

58.081.12

15.699.460.212.896.242.660.75

97.10

Table 5 (continued).

56.791.17

15.459.380.162.916.322.750.72

56.141.27

14.2411.460.193.546.942.680.56

54.891.13

13.8411.270.183.987.072.460.48

53.521.23

13.9011.420.204.077.272.210.46

56.081.27

14.2411.460.193.546.942.680.56

54.861.10

13.8810.780.194.107.292.280.39

Number 30 31 32 33 34 35

Major elements (wt%):SiO2

TiO2

A12O3

FeOMnOMgOCaONa2OK2O

Total 95.65 97.02 95.30 94.28 96.96 94.87

Mineral formula

Trace elements (ppm):SiAlTiFeMnMgCaNaK

Total

179

T. WATANABE ET AL.

oCD

LL

O

0.5 i

O•0.4-

0.3

• • •• •

••

Outerπ Inner

α αD

αD

D

α

0-9 1 1.1 1.2 1.3Tiθ2 (Wt%)

9.0

^ 8.0

O

O7.0

6.0

i Outer•

• • • •••

D

D

D

α

D

α

Inner

α

0.9 1.0 1.1 1.2 1.3Tiθ2 (Wt%)

0.6

CO

O 0.5 j_ç\j<O8 0.4

Outer

o.:

• D• • •

G

D

D

π

α

α

Inner

0.9 1.1 1.2 1.3TiO2 (wt%)

Figure 5. Glass composition from inner and outer inclusions in the clinopy-

roxene from Sample D.

180

GLASS INCLUSIONS IN PHENOCRYSTS

CaO

AI2O3

40% CaO

Figure 6. CaO-MgO-Al2O3 plot of glass inclusions in the clinopyroxene from Sample D. The arrow denotes the clinopyroxene fractionation line by Watson (1976).

181

T. WATANABE ET AL.

<δ

Plate 1. Occurrence of glass inclusions from Samples A and B. 1. Glass inclusion with unmixed texture observed under the microscope and silicic clear glassinclusions with microglobules (Sample A). 2. Secondary electron image of a glass inclusion with unmixed glass (Sample A). Light area is rich in Mg and Fe andpoor in Si as compared to the dark area. The sphere was a bubble and has silicic wall. 3. An orthopyroxene phenocryst that includes glass inclusions (SampleA). 4. A Plagioclase phenocryst that includes dark-colored glass inclusions (numbered 1-5) with unmixed texture (Sample A). 5. A Plagioclase phenocryst thatincludes glass inclusions (numbered 1-15; see Table 3) (Sample B).

182

GLASS INCLUSIONS IN PHENOCRYSTS

Plate 2. Occurrence of glass inclusions from Samples C and D. 1. A Plagioclase phenocryst that contains larger glass inclusions (numbered 1-5; see Table 4) inthe central part (Sample C). 2. Enlarged photomicrograph of the central part of a Plagioclase phenocryst shown in Figure 1 (Sample C). 3. A clinopyroxene crystalwith inner and outer rows of glass inclusions (numbered 1-35; see Table 5) (Sample D).

183