TEXTURAL AND MICROANALYSIS OF IGNEOUS ROCKS: TOOLS ...

TEXTURAL AND MICROANALYSIS OF IGNEOUS ROCKS: TOOLS FOR
UNDERSTANDING IGNEOUS PROCESSES
A Dissertation
Submitted to the Graduate School
of the University of Notre Dame
in Partial Fulfillment of the Requirements
for the Degree of
Doctor of Philosophy
by
William Scott Kinman, B.S.
Clive R. Neal, Director
Graduate Program in Civil Engineering and Geological Sciences
Notre Dame, Indiana
December 2006

This document is in the public domain.

TEXTURAL AND MICROANALYSIS OF IGNEOUS ROCKS: TOOLS FOR
UNDERSTANDING IGNEOUS PROCESSES
Abstract
by
William Scott Kinman
Mantle characterization is vital for understanding magmatism. The notion
that source characteristics are preserved transparently in primitive magmas from
mantle to eruption can be misleading. The crust acts a cool density filter leading
primitive magmas to pool, partially crystallize, and thus evolve. The details of
this evolution are seldom completely born out by whole-rock geochemistry. Dur-
ing crustal processing of magmas, crystal populations may be recycled between
geochemical reservoirs such as end-members involved in magma mixing, assimi-
lated country rock, or from variably evolved zones of a solidifying magma body.
Crystals act as physical vessels to carry compositional and temporal information
about magma evolution beyond whole-rock compositions. Textural and microana-
lytical approaches differ from whole-rock approaches, because they provide a way
to dissect crystal populations to reveal their chemical evolution as well as physical
details of their nucleation and growth. Retrieval of these types of information are
vital for understanding crustal magma evolution. The overarching goal of this
work is to better understand the physical manner in which magmas solidify, and
hence evolve, remains a fundamental problem in igneous petrology. I use crystal
size distributions to identify related crystal populations. I use EPMA, LA-ICP-
MS, and a microdrilling Sr isotope method to understand the provenance of select

William Scott Kinman
plagioclase crystals. This work is focused upon understanding the petrogenesis of
basaltic magmas. The Ontong Java Plateau is the size of Greenland yet basalt
compositions across the plateau vary little. I propose a thick complex magma
chamber system acted to buffer magma compositions, but at the plateau edge
magmas were subjected to less density filtration and thus show more diversity.
Highly depleted Detroit Seamount basalts formed during late Cretaceous inter-
action of the Hawaiian hotspot with a spreading center. Some of these basalts
contain evidence of magma mixing between unique melts from the hotspot source.
I see no role for OIB-MORB magma mixing. Plagioclase dominated crystalliza-
tion pre-dates peak crustal assimilation at Elan Ban, Kerguelen Plateau. Magma
chamber systems beneath each setting studied are complex and multichambered.
Future microanalytical work should focus on use of other isotope systems and use
of laser ablation methods.

DEDICATION
This work is dedicated to my wife for her unwavering support throughout my
education, to my parents for teaching me to work hard and be honest, and to my
son Cooper who helps me keep a smile on my face.
ii

CONTENTS
FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x
CHAPTER 1: INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Crustal Magma Chambers . . . . . . . . . . . . . . . . . . . . . . 6
1.2.1 Magma Chamber Processes: How Do Magmas Evolve? . . 7
1.2.2 Solidification Fronts and Partial Crystallization of Magmas 8
1.3 The Global Significance of Large Igneous Provinces (LIPs) . . . . 11
1.4 Crystal Size Distributions (CSDs): Identification and Examination
of Crystal Populations . . . . . . . . . . . . . . . . . . . . . . . . 13
1.5 Crystal Stratigraphy - Microanalysis of Crystal Populations . . . . 15
1.5.1 Plagioclase as a Target for Microanalysis of Basaltic Igneous
Rocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.6 Analytical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.6.1 Quantitative Textural Analysis: Crystal Size Distribution
Measurement - CSD . . . . . . . . . . . . . . . . . . . . . 17
1.6.2 Major Element Microanalysis: Electron Probe Microanalysis
- EPMA . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.6.3 Trace Element Microanalysis: Laser Ablation Inductively
Coupled Plasma Mass Spectrometry - LA-ICP-MS . . . . . 18
1.6.4 Microdrilling and Sr isotope microanalysis . . . . . . . . . 19
1.7 Project Introductions . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.7.1 The Ontong Java Plateau, SW Pacific Ocean . . . . . . . . 20
1.7.2 Detroit Seamount, Part of the Emperor Seamount Chain,
NW Pacific Ocean . . . . . . . . . . . . . . . . . . . . . . 22
1.7.3 The Kerguelen Plateau’s Western Salient - Elan Bank, Southern
Indian Ocean . . . . . . . . . . . . . . . . . . . . . . . 24
iii

CHAPTER 2: MAGMA EVOLUTION REVEALED BY ANORTHITE-
RICH PLAGIOCLASE CUMULATE XENOLITHS FROM THE ON-
TONG JAVA PLATEAU: INSIGHTS INTO LIP MAGMA DYNAMICS
AND MELT EVOLUTION . . . . . . . . . . . . . . . . . . . . . . . . 27
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2 Geologic Background of the Ontong Java Plateau . . . . . . . . . 30
2.3 Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.3.1 A Cogenetic Relationship Between Xenoliths? . . . . . . . 34
2.4 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.4.1 Analytical Methods . . . . . . . . . . . . . . . . . . . . . . 34
2.4.2 Data Quality . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.4.3 Choosing partition coefficients (D) . . . . . . . . . . . . . 37
2.4.4 Partition Coefficients for OJP Plagioclase Crystals . . . . 39
2.5 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.5.1 Plagioclase Zoning Patterns . . . . . . . . . . . . . . . . . 49
2.5.1.1 Malaita Xenolith(ML-X1) . . . . . . . . . . . . . . . . . 49
2.5.1.2 Site 1183 Xenoliths . . . . . . . . . . . . . . . . . . . . . 49
2.5.1.3 Site 1183 Phenocrysts . . . . . . . . . . . . . . . . . . . 50
2.5.1.4 Site 807 Xenolith (807-X1) . . . . . . . . . . . . . . . . . 52
2.5.2 Measured compositions . . . . . . . . . . . . . . . . . . . . 61
2.5.2.1 Major Elements . . . . . . . . . . . . . . . . . . . . . . . 61
2.5.2.2 Trace Elements . . . . . . . . . . . . . . . . . . . . . . . 61
2.5.3 Parental magma compositions . . . . . . . . . . . . . . . . 63
2.5.3.1 Major Elements . . . . . . . . . . . . . . . . . . . . . . . 63
2.5.3.2 Trace Elements . . . . . . . . . . . . . . . . . . . . . . . 64
2.5.4 Resorption Features in Site 1183 xenolith Crystals . . . . . 68
2.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
2.6.1 Inferred Chemical Magma Evolution . . . . . . . . . . . . 68
2.6.1.1 Major Elements . . . . . . . . . . . . . . . . . . . . . . . 68
2.6.1.2 An-rich OJP Plagioclase: Influence of Temperature and
Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
2.6.1.3 An-rich OJP Plagioclase: Influence of Water . . . . . . . 73
2.6.1.4 Trace Elements . . . . . . . . . . . . . . . . . . . . . . . 74
2.6.2 OJP Magma Chambers, Mush Layers, and Solidification
Fronts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
2.7 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . 84
CHAPTER 3: SHALLOW MAGMA EVOLUTION DURING LATE CRE-
TACEOUS HAWAIIAN HOTSPOT-RIDGE INTERACTION: INSIGHTS
FROM INTEGRATION OF CRYSTAL SIZE DISTRIBUTIONS AND
MICROANALYSIS OF PLAGIOCLASE . . . . . . . . . . . . . . . . . 86
iv

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.1.1 Magma Evolution in the Crustal Magma Chambers . . . . 87
3.2 The Emperor Seamount Chain . . . . . . . . . . . . . . . . . . . . 91
3.3 Detroit Seamount . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.3.1 Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.3.2 Volcanic history of Detroit Seamount . . . . . . . . . . . . 92
3.3.3 Sampling Strategy . . . . . . . . . . . . . . . . . . . . . . 93
3.4 Analytical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 94
3.4.1 Sample Preparation . . . . . . . . . . . . . . . . . . . . . . 94
3.4.2 Crystal Size Distribution (CSD) Measurement . . . . . . . 94
3.4.3 Electron Probe Microanalysis (EPMA) and Scanning Electron
Microscopy . . . . . . . . . . . . . . . . . . . . . . . . 95
3.4.4 Laser Ablation Inductively Coupled Plasma Mass Spectrometry
(LA-ICP-MS) . . . . . . . . . . . . . . . . . . . . . . 97
3.4.5 Major and Trace Element Data quality . . . . . . . . . . . 98
3.4.6 Choosing partition coefficients (D) . . . . . . . . . . . . . 98
3.4.6.1 Partition coefficients for DSM plagioclase crystals . . . . 100
3.5 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.5.1 Crystal Size Distributions . . . . . . . . . . . . . . . . . . 101
3.5.1.1 Site 1203 CSDs . . . . . . . . . . . . . . . . . . . . . . . 101
3.5.1.2 Site 884 CSDs . . . . . . . . . . . . . . . . . . . . . . . . 102
3.5.2 CSDs as Guides for Microanalysis . . . . . . . . . . . . . . 107
3.5.3 Petrography and Crystal Zoning Patterns of Microanalysis
Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.5.3.1 Site 1203 Unit 3 . . . . . . . . . . . . . . . . . . . . . . 107
3.5.3.2 Site 1203 Unit 14 . . . . . . . . . . . . . . . . . . . . . . 109
3.5.3.3 Site 1203 Unit 31 . . . . . . . . . . . . . . . . . . . . . . 111
3.5.3.4 Site 884 Unit 8 . . . . . . . . . . . . . . . . . . . . . . . 111
3.5.4 Major Elements . . . . . . . . . . . . . . . . . . . . . . . . 113
3.5.4.1 Site 1203 Unit 3 . . . . . . . . . . . . . . . . . . . . . . 113
3.5.4.2 Site 1203 Unit 14 . . . . . . . . . . . . . . . . . . . . . . 113
3.5.4.3 Site 1203 Unit 31 and Site 884 Unit 8 . . . . . . . . . . 114
3.5.4.4 Ti variations between crystal populations and Units . . . 114
3.5.5 Trace Elements . . . . . . . . . . . . . . . . . . . . . . . . 116
3.5.5.1 Measured Compositions . . . . . . . . . . . . . . . . . . 116
3.5.5.2 Inferred Parent Magma Compositions . . . . . . . . . . . 116
3.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
3.6.1 Evidence of Crystal Sorting . . . . . . . . . . . . . . . . . 120
3.6.1.1 Site 1203 Alkalic Basalts and Tholeiitic Sheet Flows . . 122
3.6.1.2 Site 1203 Pillow Basalts . . . . . . . . . . . . . . . . . . 122
3.6.2 Insights from Plagioclase Major Element Compositions . . 124
3.6.3 Crustal Magma Evolution: Insights from Trace Elements . 127
v

3.6.3.1 Crystal Population Origins . . . . . . . . . . . . . . . . . 127
3.6.4 The Crystal Record of Highly Evolved Magmas . . . . . . 133
3.6.5 Ba and Sr variations: Evidence for Polybaric Crystallization?134
3.7 Summary and Conclusions: Insights Into the Petrogenesis of Depleted
Detroit Seamount Basalts . . . . . . . . . . . . . . . . . . . 136
CHAPTER 4: CRUSTAL CONTAMINATION OF BASALTIC MAGMAS
REVEALED THROUGH MICROANALYSIS OF MAJOR, MINOR, AND
TRACE ELEMENTS AND Sr ISOTOPES IN PLAGIOCLASE: IMPLI-
CATIONS FOR ELAN BANK MAGMAS, KERGUELEN PLATEAU,
INDIAN OCEAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
4.1.1 Geologic Background of the Kerguelen Plateau and Elan Bank176
4.2 Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
4.3 Analytical Techniques . . . . . . . . . . . . . . . . . . . . . . . . 178
4.3.1 Sample Preparation . . . . . . . . . . . . . . . . . . . . . . 178
4.3.2 Electron Probe Microanalysis (EPMA) and Scanning Electron
Microscopy . . . . . . . . . . . . . . . . . . . . . . . . 178
4.3.3 Laser Ablation Inductively Coupled Plasma Mass Spectrometry
(LA-ICP-MS) . . . . . . . . . . . . . . . . . . . . . . 179
4.3.4 Major and Trace Element Data quality . . . . . . . . . . . 180
4.3.5 Microdrilling and Sr isotope microanalysis . . . . . . . . . 180
4.3.6 Choosing Partition Coefficients (D) . . . . . . . . . . . . . 181
4.3.6.1 Partition coefficients for Elan Bank Plagioclase Crystals 183
4.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
4.4.1 Petrography . . . . . . . . . . . . . . . . . . . . . . . . . . 185
4.4.2 Electron Probe Microanalysis (EPMA) - Major Elements . 188
4.4.3 LA-ICP-MS - Trace Elements . . . . . . . . . . . . . . . . 194
4.4.3.1 Scandium, Ti, V, and Pb . . . . . . . . . . . . . . . . . 194
4.4.3.2 Barium, Sr, and Rb . . . . . . . . . . . . . . . . . . . . 194
4.4.3.3 Yttrium and the Rare Earth Elements La, Ce, Pr, Nd,
Sm, and Eu . . . . . . . . . . . . . . . . . . . . . . . . . 198
4.4.4 Sr Isotope Microdrilling . . . . . . . . . . . . . . . . . . . 198
4.4.5 Inferred Parent Magma Compositions . . . . . . . . . . . . 202
4.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
4.5.0.1 Origin of Major Element Zoning and Resorption Surfaces 206
4.5.0.2 Insights from Trace Elements and Inverted Parent Magma
Compositions . . . . . . . . . . . . . . . . . . . . . . . . 208
4.5.0.3 Diffusive Redistribution of Trace Elements . . . . . . . . 210
4.5.0.4 Petrogenetic Insights from Plagioclase 87 Sr/ 86 SrI Ratios
and the Timing of Crustal Contamination at Elan Bank 213
vi

4.5.0.5 Where Did Plagioclase Dominated Partial Crystallization
Occur? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
4.6 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . 214
CHAPTER 5: SUMMARY, CONCLUSIONS, AND FUTURE WORK . . 215
5.1 Results and Conclusions . . . . . . . . . . . . . . . . . . . . . . . 215
5.1.1 The Ontong Java Plateau . . . . . . . . . . . . . . . . . . 215
5.1.2 Detroit Seamount, Emperor Seamount Chain . . . . . . . . 217
5.1.3 Elan Bank, Kerguelen Plateau . . . . . . . . . . . . . . . . 218
5.2 Recommendations for Future Work . . . . . . . . . . . . . . . . . 220
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
vii

FIGURES
1.1 Magma Chambers . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 The CSD Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Multiply Saturated Magmatic Solidification Front . . . . . . . . . 9
1.4 Global Distribution of Large Igneous Provinces . . . . . . . . . . 12
2.1 Ontong Java Plateau Map . . . . . . . . . . . . . . . . . . . . . . 29
2.2 OJP Xenoliths from ODP Site 1183: Hand Sample and Petrographic
Thin Section Images . . . . . . . . . . . . . . . . . . . . . 32
2.3 OJP Xenoliths from ODP Site 807 and Malaita: Hand Sample and
Petrographic Thin Section Images . . . . . . . . . . . . . . . . . . 33
2.4 ODP Site 1183 Stratigraphic Section and Locations of Xenoliths
Examined in This study . . . . . . . . . . . . . . . . . . . . . . . 36
2.5 Backscatter Electron SEM images of ODP Site 1183 Phenocrysts 38
2.6 Core to Rim Traverses of An (mol%) Content of Xenolith Plagioclase
Crystals and LA-ICP-MS Analysis Locations . . . . . . . . . 51
2.7 Measured Major and Trace Element Data Versus An (mol%) Content 62
2.8 Measured MgO Versus An (mol%) Content . . . . . . . . . . . . . 63
2.9 Trace Element Ratio Plots for OJP Plagioclase Parent Magmas . 66
2.10 Calculated Major and Trace Element Data Versus An (mol%) Content 69
2.11 A Schematic of a Possible OJP Magma Chamber System . . . . . 81
3.1 Map of the Hawaiian Ridge and Emperor Seamount Chain and
Bathymetric Map of Detroit Seamount . . . . . . . . . . . . . . . 89
3.2 Stratigraphy of igneous basement rocks recovered from ODP Sites
1203 and 884 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
3.3 Measurement and Interpretation of Crystal Size Distributions . . 96
3.4 Site 1203 CSD Results . . . . . . . . . . . . . . . . . . . . . . . . 103
viii

3.5 Site 884 CSD Results . . . . . . . . . . . . . . . . . . . . . . . . . 104
3.6 The CSD Diagram as a Guide for Microanalysis . . . . . . . . . . 108
3.7 Petrographic thin section photographs illustrating basalt textures 110
3.8 Backscatter Electron Scanning Electron Microscopy Images Showing
Plagioclase Major Element Zoning . . . . . . . . . . . . . . . 112
3.9 Major Element Compositions of Plagioclase Crystal Cores . . . . 115
3.10 Measured Trace Element Abundances Plotted Against Anorthite
Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
3.11 Parent Magma Ba/Sr and La/Sm Ratios Plotted Against Anorthite
Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
3.12 Population A and B Parent Magma La/Ce Vs. Sr Relative to Bulk-
Rock Compositions . . . . . . . . . . . . . . . . . . . . . . . . . . 129
3.13 Plagioclase Parent Magma La/Y Vs. Sr and Possible Source Affinities131
3.14 Plagioclase Parent Magma Sr Vs. Ba . . . . . . . . . . . . . . . . 135
4.1 Map of Indian Ocean and Kerguelen Plateau and Free-Air Gravity
Map of Elan Bank . . . . . . . . . . . . . . . . . . . . . . . . . . 175
4.2 ODP Site 1137 Igneous Basement Rocks . . . . . . . . . . . . . . 177
4.3 Plagioclase Sr Isotope Microdrilling . . . . . . . . . . . . . . . . . 182
4.4 Site 1137 Unit 4 and Unit 10 basalt petrography . . . . . . . . . . 186
4.5 Site 1137 Unit 4 and 10 Plagioclase Phenocryst Major An-Ab-Or
compositions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
4.6 Unit 4 An profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
4.7 Unit 10 An profiles . . . . . . . . . . . . . . . . . . . . . . . . . . 191
4.8 Microanalysis of Site 1137 Unit 4 Plagioclase Phenocrysts . . . . . 195
4.9 Microanalysis of Site 1137 Unit 10 Plagioclase Phenocrysts Part I 196
4.10 Microanalysis of Site 1137 Unit 10 Plagioclase Phenocrysts Part II 197
4.11 Measured Trace Element Abundances vs. An Part I . . . . . . . . 199
4.12 Measured Trace and Rare Earth Element Abundances vs. An Part II200
4.13 Relative Ranges of 87 Sr/ 86 SrI of Site 1137 Whole-Rock Samples and
Plagioclase Phenocrysts . . . . . . . . . . . . . . . . . . . . . . . 202
4.14 Trace Element and Sr Isotope Diagrams . . . . . . . . . . . . . . 209
4.15 Diffusive Redistribution of Sr and La . . . . . . . . . . . . . . . . 212
ix

TABLES
2.1 MAJOR AND TRACE ELEMENT DATA: PLAGIOCLASE PHE-
NOCRYSTS AND XENOLITH PLAGIOCLASE CRYSTALS . . 41
2.2 CALCULATED PARTITION COEFFICIENTS AND EQUILIB-
RIUM LIQUID COMPOSITIONS . . . . . . . . . . . . . . . . . . 54
2.3 RANGES OF TRACE ELEMENTS IN PLAGIOCLASE, WHOLE-
ROCK DATA, AND INFERRED PARENT MAGMAS . . . . . . 67
2.4 MELTS CRYSTALLIZATION MODELING OF AVERAGE OJP
BASALTIC GLASSES . . . . . . . . . . . . . . . . . . . . . . . . 71
3.1 CSD MEASUREMENT INFORMATION . . . . . . . . . . . . . 105
3.2 CSD RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
3.3 PLAGIOCLASE PHENOCRYST MAJOR AND TRACE ELEMENT
DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
3.4 PLAGIOCLASE TRACE ELEMENT PARTITION COEFFICIENTS
AND CALCULATED PARENT LIQUID COMPOSITIONS . . . 161
4.1 UNIT 4 PLAGIOCLASE MAJOR AND TRACE ELEMENT ABUN-
DANCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
4.2 UNIT 10 PLAGIOCLASE MAJOR AND TRACE ELEMENT ABUN-
DANCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
4.3 UNIT 4 AND 10 87 Sr/ 86 Sr DATA . . . . . . . . . . . . . . . . . . 201
4.4 UNIT 4 PLAGIOCLASE PARTITION COEFFICIENTS AND IN-
FERRED PARENT MAGMA COMPOSITIONS . . . . . . . . . 204
4.5 UNIT 10 PLAGIOCLASE PARTITION COEFFICIENTS AND
INFERRED PARENT MAGMA COMPOSITIONS . . . . . . . . 205
x

1.1 Introduction
CHAPTER 1
INTRODUCTION
Mantle source characterization is vital for understanding the origins of mag-
matism on planetary bodies within our solar system and to achieve a more com-
plete understanding of Earth’s geologic history. To accurately characterize mantle
source regions it is necessary to understand the extent to which magma compo-
sition changes between partial melting of the mantle and eruption at the surface.
The notion that source characteristics are preserved transparently in primitive
magmas from mantle source to eruption can be misleading. Indeed, O’Hara and
Herzberg [113] suggested that truly primitive basaltic magmas rarely escape to
the Earth’s surface, since the crust acts as a cool density filter to primitive mag-
mas ascending from hotter mantle source regions. Ponding, cooling, and partial
crystallization is favored as ascending magmas reach horizons of neutral buoyancy
within the shallow crust, which commonly occurs at depths of 4-8 km [97, 122].
Magma differentiation within the shallow crust may involve a combination of par-
tial crystallization (i.e., fractional, equilibrium, or in-situ), assimilation, and/or
magma mixing. A wide range of studies (e.g., [16, 38, 113]) have highlighted the
significance of shallow magma evolution in basaltic to rhyolitic magmatic systems.
Magma evolution in the shallow crust is complex. The details of this evo-
lution are seldom completely born out by whole-rock geochemistry. There has
1

thus been debate within the volcanology and petrology community regarding the
utility of whole-rock data alone for deconvolution of complex differentiation pro-
cesses. Whole-rock compositions are the sum of all of the processes and composi-
tional end-members that affected a given hybrid magma, which is what extrusive
igneous rocks commonly represent (e.g., [37, 98]). Textural and microanalytical
approaches differ from whole-rock geochemical approaches in essence because they
provide a way to dissect crystal populations to reveal their chemical evolution as
well as physical details of their nucleation and growth (e.g., [39, 104]). Retrieval
of these types of information are vital for understanding crustal magma evolu-
tion. During crustal processing of magmas, crystal populations may be recycled
between geochemical reservoirs such as end-members involved in magma mixing,
assimilated country rock, or from variably evolved zones of a solidifying magma
body [23]. Compositional data retrieved from distinct crystal populations and
individual zoned crystals thus provide insights into the timing and chemistry of
magma evolution, which helps to link changes in magma composition with phys-
ical processes. Indeed, the physical manner in which magmas solidify, and hence
evolve, remains a fundamental problem in igneous petrology (e.g., [70, 98]).
Individual crystals and related crystal populations (e.g., phenocrysts, micro-
lites, and/or xenocrysts) act as physical vessels to carry compositional and tempo-
ral information about magma evolution (e.g.,[37]). Charlier et al. [23] presented
an example that highlights the the wealth of petrogenetic information that can be
recovered using a microanalytical approach to understand magmatic processes. In
their examination of U-Th and U-Pb systematics of zircon crystals in Quaternary
age rhyolites from Taupo volcano, New Zealand, Charlier et al. [23] reported nu-
merous zircon crystals with cores inherited from the country rock (a zircon-bearing
2

A)
B)
19∞15' N
19∞20'
Mauna Iki
Kamakalauka
19∞25'
Kamakaiawaena
Summit Magma
Reservoir
Southwest Rift Zone
155∞20'
19∞30'
Kilauea
Caldera
Halemaumau
Keanakakoi
Kilauea Iki
Devil's Throat Pauahi
Aloi
East Rift Zone
Mauna Ulu Alae
155∞10'
Primary Conduit
Makaopuhi
Napau
Pu'u Kamoamoa
Pu'u O'o
0
Pu'u Kauka Heiheiahulu
Pu'u Kiai
Kalalua
Volcanic Shield
Oceanic Crust
Upper Mantle
3 km
155∞05'
30
35
40
25
0
20
15
3 km
Pacific Ocean
Figure 1.1. Graphic examples of relatively simple versus complex
magma chamber models for the A) Ontong Java Plateau (from Sano
and Yamashita [123]) constructed based primarily upon geochemical
data and B) Kilauea (from Ryan [122]) using seismic data. Which
model is more accurate?
3
Pu'u Kahaualea
10
155∞00' W
5
Depth (km)
0

Cambrian schist) including a crystal with a narrow ∼ 520 Myr core region and
∼ 300 Kyr rim grown in the rhyolitic host magma. These types of findings are of
profound significance, as they provide robust insights into the fundamental phys-
ical processes and chemical consequences of crustal assimilation. Although zircon
is a refractory mineral, this type crystal of inheritance is not limited to zircon and
may be exhibited by minerals that are stable on the basaltic liquidus over a wide
range of magmatic temperatures and pressures (e.g., plagioclase). Crystal popula-
tions and individual crystals not totally consumed by magmatic processes indeed
provide records of magma chemistry and magma chamber dynamics beyond what
can be learned through whole-rock studies alone.
The complexity of shallow magma evolution has also been highlighted by quan-
titative textural studies, most notably in crystal size distribution (CSD) studies
(e.g.,[67, 68, 96, 97]). Crystal size distributions yield information about crystal
nucleation and growth conditions as well as open system processes like crystal ac-
cumulation or magma mixing [20, 96]. Higgins [67] showed that the proportions of
mixing members can be well estimated from CSD data. Crystal size distributions
provide a means to qualitatively assess whether certain processes may have taken
place (e.g., magma mixing, crystal resorption/removal, or crystal accumulation),
as well as provide a quantitative method for determining magmatic residence times
and calculation of end-member mixing proportions [67].
A collective goal within the volcanology and petrology community is to under-
stand the timing and dynamics of shallow magma evolution. Meeting this goal
serves not only to answer fundamental scientific questions but also includes an im-
portant human factor. Magma evolution that takes place in shallow subterranean
chambers has a dramatic influence on when and how a volcano will erupt (e.g.,
4

Figure 1.2. (A) Relationship of CSD slope to crystal growth rate and
residence time. Slope = -1/Gτ where G is growth rate; τ = residence
time; n ◦ = nucleation density. (B) CSD showing mixture of two crystal
populations. (C) Plagioclase CSD of a ∼ 81 Ma pillow basalt sample
from Detroit Seamount where whole-rock data have indicated magma
mixing may have occurred.
5

explosive eruption vs. mild effusive eruption) as well as how seriously an eruption
will impact the surrounding human population e.g., [65]. This human factor has
been an impetus for extensive studies of active volcanoes like Mt. Vesuvius due to
the threat this volcano poses to the densely populated city of Naples, Italy. More
pertinent to the work described here is the fact that efforts to fully understand
and mitigate volcanic hazards has lead to creative new textural and microana-
lytical schemes to improve the accuracy of our view of how magmas evolve in
the shallow crust [105]. The overarching purpose of the work detailed in this
dissertation is to constrain the physical and chemical details of shallow magma
evolution in basaltic systems with an emphasis large igneous provinces (LIPs).
In this chapter I will introduce contemporary views of magma chambers, magma
dynamics, magma evolution, LIPs and their global significance, microanalytical
approaches, CSDs, and three focused research projects where I applied textural
and/or microanalytical approaches to better understand fundamental processes of
shallow basaltic magma evolution.
1.2 Crustal Magma Chambers
Rarely can one confidently identify detailed physical processes that
drive petrologic diversity. Part of the difficulty originates in the construction
of physical and chemical models; they often appear as fundamentally
different enterprises. – G.W. Bergantz [4]
Modeling magma chamber architecture and magma chamber processes is com-
plicated by our inability to directly observe these subterranean environments as
they are active. The architecture of magma chambers and/or magma chamber
systems are depicted drastically different, as suggested by [4], whether modeled
using chemical (i.e., using isotopic, major, and trace element data) or physical
6

(i.e., using textural, seismic, or other geophysical data) data. Magma chamber
architecture has a profound influence on the chemistry of evolving magmas [98].
Geochemists customarily envisage magma chambers as large vats of magma, where
processes such as fractional crystallization, magma mixing, and assimilation are
modeled within a physically simple context (e.g., Fig. 1.1a). For example, Fig-
ure 1.1a, from Sano and Yamashita [123], illustrates a relatively simple magma
chamber model for the Ontong Java Plateau (OJP) based only upon geochemical
data and petrographic observations. The magma chamber system of Kilauea vol-
cano, Hawaii is depicted much differently as a complicated interconnected network
of dikes, sills, and chambers [122]. This Kilauea model was constructed using seis-
mic data by Ryan [122] (Fig. 1.1b). Neither model is necessarily ill-constructed,
but each model was contrived to explain either physical observations or chemical
observations but not both. Chemical and physical magma chamber models must
be better integrated in order to generate more realistic (and accurate) models
of magma chamber architecture and how this architecture is related to magma
dynamics and the chemical evolution of magma.
1.2.1 Magma Chamber Processes: How Do Magmas Evolve?
Magma chamber processes such as partial crystallization, magma mixing, and/or
assimilation play a controlling role in shallow level differentiation and need to be
understood if a parental (or even primary) magma compositions are to be es-
timated. Magma chamber architecture influences the physical manner in which
differentiation processes occur (i.e., homogenous crystallization vs. solidification
front crystallization)[97]. The role of solidification fronts, crystal-mush piles, and
their interstitial liquids have increasingly recognized roles in magmatic differen-
7

tiation (e.g., [84, 98]; Fig. 1.3). Contemporary magma chamber models have
suggested solidification front processes have pervasive and significant effects on
the textural and chemical evolution of magmas (e.g., [97, 98]). Classic magma
chamber models tend to favor high rates of convection and bottom-up solidification
primarily by crystal sedimentation (e.g., [138]). However, growing evidence over-
whelmingly suggests crystallization occurs commonly along chamber walls, where
most heat is lost, and progresses inward by solidification front growth [97, 98, 102].
1.2.2 Solidification Fronts and Partial Crystallization of Magmas
In his discussion of solidification front dynamics, Marsh [98] suggested that
upon emplacement of any magma body, solidification front growth begins imme-
diately and thickens with the square root of time. Silicate magmas are commonly
multiply saturated favoring non-dendritic solidification front growth (e.g., Fig. 1.3
adapted from [98]). Crystal nucleation and initial growth take place within the
suspension zone, which is bounded outward by the capture front and inward by
the liquidus. Outward from the capture front, within the mush zone, the solidifi-
cation front has a crystallinity of >25% where crystals continue to grow but have
little chance of escaping the advancing front. Toward the magma chamber walls
the solidification front has a crystallinity > 50% and essentially behaves as a solid
(for a dedicated discussion of solidification fronts see [98]).
In a crystallizing magma, chemical components not incorporated into the grow-
ing mineral phases will accumulate in the melt adjacent to the growing crystals
particularly where non-turbulent crystallization environments exist. These envi-
ronments most plausibly exist within solidification fronts in crystal mush layers.
Mush layers, as described above, are rheological elements of solidification fronts
8

Figure 1.3. A simple schematic of a multiply saturated solidification
front. This solidification front morphology is favored over dendritic type
solidification fronts in most of magmatic systems. Figure adapted from
[98]
9

defined as having crystallinities between 25% and 50%-55% [98] (Fig. 1.3). Pro-
cesses occurring within mush-layers and their interstitial spaces are thus best dis-
cussed in terms of solidification front processes. Crystallization within the mush
zone produces evolved interstitial melts [84, 98]. Continued crystallization traps
these evolved liquids within the solidification front. Dependent upon reservoir
geometry and magma supply rate, solidification fronts may propagate inward to
a point where the chamber is filled largely with a crystal mush that has a net-
work of interstitial spaces filled with variably evolved melts [98]. Flushing or filter
pressing of the crystal mush frees these evolved melts that can: 1) mix with more
primitive magmas; 2) be erupted; or 3) be intruded. A filter pressing mecha-
nism has been suggested for off-summit (axial) Hawaiian eruptions, where axial
tholeiitic Kilauea magmas are olivine-clinopyroxene-plagioclase-saturated versus
olivine saturated in summit magmas (e.g., [71, 98]). Interstitial liquids in the
crystal mush may be flushed out by less evolved magma rising from lower levels
of the magmatic system.
The crystal-free interior of the magma chamber is not affected by the solidi-
fication front differentiation processes, though evolved liquids may attain enough
buoyancy to rise and mix with undifferentiated magmas in the magma chamber
interior [84, 85, 101]. Scouring of magma chamber margins by vigorous input
of new magma may free evolved packets of magma and crystals by solidification
front erosion (e.g., [98]). Phases that are below the liquidus of the main body
of magma can grow in interstitial spaces of solidification fronts and mush piles
that are thermally and mechanically insulated from the hotter magma chamber
interior [84]. Crystal debris, including sub-liquidus phases, may be fully or par-
tially resorbed by the hotter less evolved magma in the chamber interior, which
10

will change composition reflecting such resorption. If entrained crystals survive
they may be deposited by sedimentation elsewhere in the magmatic system ef-
fectively fractionating the residual magma, a process referred to as punctuated
differentiation by Marsh [98]. Shallow magma evolution in basaltic systems on
Earth (LIPs) is thus influenced by a balance of several processes that include: 1)
Formation of evolved melts within crystal-mush regions due to partial crystalliza-
tion and the return of these melts to the main magma body by buoyancy driven
flow or solidification front erosion; 2) Flushing of variably evolved melts from the
crystal-mush interstitial spaces (i.e, filter pressing mechanism); 3) Entrainment
of phenocrysts and other crystal debris during magma recharge; 4) Resorption;
5) Crystal settling and solidification front capture of crystals leading to magma
fractionation [84, 85, 98, 101]; 6) Assimilation; and 7) Magma mixing.
1.3 The Global Significance of Large Igneous Provinces (LIPs)
Large igneous provinces or LIPs are massive crustal emplacements of mafic
magma unrelated to normal ocean floor spreading [31] (Fig. 1.4). Magmas from
LIPs provide insight into regions of the mantle untapped during normal sea-floor
spreading [32]. Understanding LIP magma dynamics is important, particularly in
light of recent debate regarding the existence of mantle plumes and their common
association with LIPS (e.g., [48, 124]). Plume based models generally favor deep
mantle source regions for LIPS (e.g., [17]), possibly at the core mantle boundary
(e.g., [14, 31]), whereas alternative, non-plume models favor shallow fertile upper
mantle source regions referred to as “perisphere” [81]. Resolution of this debate
must first begin with a better understanding of primitive LIP magmas, including
their volatile content, temperature, and most importantly - the extent to which
11

12
Figure 1.4: Large Igneous Provinces around the globe; The three LIPs at the focus of this study are labeled: Hawaiian-
Emperor Chain, Ontong Java Plateau, and the Kerguelen Plateau. Map is adapted from [32].

these magmas chemically evolved in the shallow crust. Clues to these processes and
relatively primitive magma compositions may be best recorded in early liquidus
mineral phases [5]. Coffin ad Eldholm [32] suggested that some LIPs are emplaced
at rates greater than 10 5 km 2 m.y. −1 , which underscores the possibility that LIP
emplacement can have dramatic effects on the global environment. Large Igneous
Provinces are more buoyant than normal oceanic crust and are relatively resistant
to subduction, which indicates they may have an important role in the genesis of
continents throughout geologic time [32].
1.4 Crystal Size Distributions (CSDs): Identification and Examination of Crystal
Populations
The origins of phenocrysts are an often under-appreciated aspect of petrologic
investigations. It is becoming increasingly apparent that they commonly represent
a more complex formation than simple crystallization from their current host
magmas [98]. This complexity is well documented by CSDs, which measure the
number of crystals of a characteristic size per unit volume of rock (e.g., [20, 96,
98]). Higgins [67] used CSDs to identify two distinct populations of plagioclase in
porphyritic dacites from Kameni Volcano, Greece, which he interpreted to be a
result of mixing between two phenocryst-carrying magmas (Fig. 1.2b). Crystal size
distributions are customarily displayed on a log-normal plot of population density
(number of crystals per unit volume rock) versus crystal size (L) (Fig. 1.2) [20].
The utility of the CSD is best illustrated by considering a simple example,
emplacement of a thin lava flow where magma arrives at the surface in a wholly
liquid state. Crystal nucleation rate increases as the flow cools [18]. If nucleation
and growth continue uninterrupted, the final rock will yield a linear or near linear
13

distribution of crystal sizes (e.g., Fig. 1.2a) [68, 96]. However, if the magma arrives
at the surface carrying macroscopic crystals (i.e., phenocrysts), be they entrained
crystals from a mush pile or ripped up pieces of the conduit, the final rock will
yield a non-linear CSD (Fig. 1.2b) [18]. If the magma in a wholly liquid state
stalled and experienced partial crystallization en-route to the surface then lost
some of its newly grown crystals via crystal settling, then a downward deflection
of the CSD would be favored [99]. Crystals carried in the initial magma may vary
in size and can easily be interpreted as phenocrysts crystallized from the present
host magma [98]. Non-linear CSDs indicate dynamic and/or kinetic processes af-
fected crystallization [18, 96], which reflects the presence of more than one distinct
subterranean crystal nucleation and growth environment (e.g., Fig. 1.2b,c) [67].
In the case of a rock formed from a phenocryst-carrying magma, it will contain
at least two distinct populations of crystals. In the case of a non-linear CSD, the
crystals grown as the lava cooled and solidified at the surface reflect the most
recent conditions of nucleation and growth (e.g., the segments of the CSD with
steeper negative slope in Figs. 1.2b,c). Phenocrysts carried in the initial magma
(i.e., those represented by the shallower negative slope in Figs. 1.2b,c) reflect nu-
cleation and growth conditions in some subsurface environment. If a linear crystal
growth rate is assumed, CSDs can be used to estimate crystal residence times in
magmatic systems (Fig. 1.2a) e.g., [120]. The curvature and slopes of CSD seg-
ments provide a way to examine and narrow the possible physical processes that
affected a batch of magma [70].
14

1.5 Crystal Stratigraphy - Microanalysis of Crystal Populations
Advances in microsampling and technology have made it possible to chemically
and in some cases isotopically dissect igneous rocks, mineral by mineral (e.g.,
[22]). Detailed rim-to-rim isotopic, major, and trace element studies (i.e., crystal
stratigraphy) have commonly documented phenocryst-host rock disequilibrium
(e.g., [39, 40, 134, 135, 143]). Crystal stratigraphy studies have also provided
insights into the microenvironments surrounding crystals as they grow or where
growth has been disrupted (e.g., [59, 60]). Processes affecting magmatic crystal
growth are revealed on a crystal-by-crystal basis using this method, however its
power requires a statistically relevant number of crystals be analyzed. Analysis
of too few crystals provides only a limited, local-scale history of the rock. As
an example of this potential limitation, consider a basaltic magma body that
has experienced numerous magma mixing events consisting of liquid ± crystal
exchange. Lavas produced from the system will carry crystals from some or all
of the mixing events (e.g., [59, 139]). Like whole-rock analysis, documenting the
histories of too few crystals from the basalt will not adequately characterize all
of the processes (in this case mixing events) responsible for the final texture and
composition of the rock. This potential limitation is minimized via integration of
CSDs and crystal stratigraphy studies, whereby CSDs are used to identify crystal
populations with related nucleation and growth histories that then become targets
for microanalysis.
1.5.1 Plagioclase as a Target for Microanalysis of Basaltic Igneous Rocks
In this study plagioclase feldspar is a frequent target for major and trace el-
ement microanalysis. As noted by Bindeman et al. [8], plagioclase is ideal for
15

investigating magmatic evolution using crystal stratigraphy because 1) it is a
common igneous mineral in mafic through felsic systems; 2) it appears early as
a liquidus phase in basaltic systems and is stable on the liquidus for a relatively
long period of time because of its capability for solid solution; 3) minerals that
crystallize before plagioclase (e.g., olivine) do not significantly fractionate trace
elements so plagioclase compositions adequately reflect the composition of the
parental magma; 4) plagioclase contains measurable quantities of trace elements
from different geochemical groups (e.g., large ion lithophile elements - LILE and
rare earth elements - REE); 5) plagioclase has a highly polymerized crystal struc-
ture relative to olivine or clinopyroxene that leads to very slow diffusivities of
major and trace cations so magmatic-induced zonations are commonly preserved
(e.g., [13, 24–28, 56–58]). Plagioclase therefore commonly records events in the
evolution of a magma body that are evident as optical zonation or resorption
features. Such features have been previously studied through experimental work
(e.g., [77, 87, 107, 136]) and numerical modeling (e.g., [1, 86, 90]).
The simple fact that Sr is compatible in plagioclase makes it an ideal candidate
mineral for intra-mineral measurements of 87 Sr/ 86 Sr ratios via microdrilling, micro
Sr extraction and analysis by TIMS [38]. Micro Sr isotope studies have proven to
be fruitful at elucidating important processes during open system shallow magma
evolution, as 87 Sr/ 86 Sr ratios are not heavily influenced by closed system magmatic
processes and partial crystallization (e.g., [39]).
16

1.6 Analytical Methods
1.6.1 Quantitative Textural Analysis: Crystal Size Distribution Measurement -
CSD
Digital image mosaics of entire petrographic thin sections were captured in
cross polarized light using an automated microscope stage system manufactured
by Prior Scientific Instruments, Cambridge, UK. Photographs of each scanned
area were printed on photograph paper and overlain with tracing paper. Minerals
of interest (plagioclase and/or olivine) were outlined by hand using a pen with
a 0.1 mm diameter tip (which also corresponds to the smallest measurable crys-
tal size). The tracings were digitized via high resolution scanning with a flatbed
scanner. Crystal intersection areas were shaded gray, detected, and measured
according to a best fit ellipse routine using the freeware program UTHSCSA Im-
ageTool (http://ddsdx.uthscsa.edu/dig/itdesc.html). The best fit ellipse major
and minor axis results were input into the spreadsheet program CSDslice written
by Morgan and Jerram [106] to estimate the three-dimensional crystal habit. As
a thin section through an igneous rock will be a random 2D representation, the
plagioclase crystals will be represented as a variety of cross sections. The crystal
major axis data, the estimated 3D crystal habit, rock fabric, total area measured,
and an estimate of crystal roundness were input into the program CSDcorrections
version 1.37 written by Michael Higgins [68] to covert two-dimensional CSD data
to true three-dimensional CSDs as outlined by Marsh [96].
1.6.2 Major Element Microanalysis: Electron Probe Microanalysis - EPMA
The Electron Probe Microanalysis (EPMA) and SEM methods outlined in
this section are general descriptions of the methods used for quantitative wave-
17

length dispersive spectroscopy(WDS) and collection of compositionally contrasted
backscatter electron images throughout all of the work presented in this disser-
tation. Deviations from these general methods are noted with accompanying ra-
tionale for each change. Backscatter electron images and major element analyses
were performed using a JEOL JXA-8600 Superprobe electron microprobe at the
University of Notre Dame. Backscatter electron images were collected using a 1
µm beam, an accelerating voltage of 20 kV, and a probe current of 25-50 nA.
Microprobe analyses were performed using a 10 µm defocused beam, accelerating
voltage of 15 kV, a probe current of 20 nA, 15 second on-peak counting time, and
two background measurements per peak. Sodium was measured first to minimize
loss via volatilization. Microprobe data were corrected for matrix effects using
a ZAF correction routine. Data points near Fe-rich phases such as melt inclu-
sions and alteration-filled fractures were discarded. I utilized variety of mineral
calibration standards for mineral analyses and glass standards for melt inclusion
analyses, and I generally attempted to match mineral standards with the mineral
understudy (e.g., orthoclase and sanidine as Al, Na, K, and Si standards for anal-
ysis of plagioclase). I routinely measured major elements in mineral standards as
unknowns during analytical sessions to monitor calibration and data quality.
1.6.3 Trace Element Microanalysis: Laser Ablation Inductively Coupled Plasma
Mass Spectrometry - LA-ICP-MS
The LA-ICP-MS method described in this section is a general method descrip-
tion, and deviations from this general routine in terms of elements and instrument
parameters are noted. Scandium, Ti V, Rb, Sr, Y, Ba, La, Ce, Nd, Sm, Eu, and
Pb were measured in plagioclase crystals using a New Wave UP-213 UV laser
18

ablation system interfaced with a ThermoFinnigan Element 2 ICP-MS operated
in fast magnet scanning mode at the University of Notre Dame. I used a laser
frequency of 5 Hz, pulse energy of 0.02-0.03 mJ pulse −1 , 15 - 40 µm diameter pits
depending on crystal size, and helium as the carrier gas (∼ 0.7 l min −1 ) mixed
with argon (∼ 1.0 l min −1 ) before introduction to the plasma. Due to the transient
nature of the laser ablation signal, analyses were conducted in peak jumping mode
with one point quantified per mass. The LA-ICP-MS spots were coincident with
previous EPMA analyses, and Ca measured by EPMA was used as an internal
standard for each spot analysis, because its fractionation behavior is similar to
that of Sr, Ba, the REE [52]. The trace element glass NIST 612 was used as a
calibration standard for all laser ablation analyses. Although heterogeneity for
certain elements has been documented in the widely used NIST 612 glass i.e.,
[44], Eggins et al. [44] considered it a reliable calibration standard for the ele-
ments examined in this study. The analytical protocols of Longerich et al. [88]
were used for LA-ICP-MS data reduction within the LAMTRACE spreadsheet
data reduction program written by Dr. Simon Jackson or Macquarie University.
1.6.4 Microdrilling and Sr isotope microanalysis
Microdrilling for Sr isotope microanalysis was performed using a Merchantek
MicroMill at the University of Durham, UK after completion of EPMA and LA-
ICP-MS work. We used a highly tapered tungsten carbide drill bit for all drilling,
which provided < 50 µm diameter single drill pits. Microdrilling was performed
by setting precision scans consisting of grids and lines of drill points, which al-
lowed us to target zones of interest and to maximize the amount of Sr recovered
to ensure as accurate and precise 87 Sr/ 86 Sr measurements as possible. Prior to
19

microdrilling, a pre-cleaned square 4 cm x 4 cm piece of parafilm with a ∼ 2 cm
center hole was adhered to the polished sample surface. A droplet of ultrapure
water was then placed over the center hole and area to be drilled. The parafilm
was was used to minimize dispersion of the water droplet and to minimize sample
loss. Microdrilling was then performed within the water droplet, which generated
a slurry as sample material was removed. The slurry was removed using a mi-
cropipette and placed in a 3.5 mL Teflon beaker. The sample was then subjected
to a hotplate concentrated HF and HNO3 digestion. Strontium was extracted
using a scaled down ion exchange column method. Extracted Sr was loaded on
Re filaments and 87 Sr/ 86 Sr ratios were measured using a Finnigan Triton thermal
ionization mass spectrometer (TIMS). Davidson et al. [39] and Tepley et al. [135]
provide more extensive discussions about of the micro-Sr isotope method.
1.7 Project Introductions
1.7.1 The Ontong Java Plateau, SW Pacific Ocean
The Ontong Java Plateau is the worlds largest oceanic LIP spanning an area
of roughly 2.018 x 10 6 km 2 in the Southwest Pacific ocean [30, 32, 93, 94, 132, 133]
(Fig. 1.4). Tarduno et al. [129] and Mahoney et al. [93, 94] suggested that the
bulk of the OJP formed in a submarine eruptive environment relatively rapidly
around ∼122 Ma, with a minor event at ∼90 Ma [93, 94, 132, 133]. Most of the
OJP is presently submerged, exceptions being the southern plateau margin on
the islands of Malaita, San Cristobal (Makira), and Santa Isabel in the Solomon
Islands, where obducted segments now outcrop [116].
Despite its size, only three subtly different low-K tholeiitic basalt types have
been recovered from the OJP, either by field work or ocean drilling, which have
20

een referred to as Singgalo, Kwaimbaita, and Kroenke basalts [46, 131, 133]
(Fig. 1.4). The uniformity of basalt composition over the Greenland-size OJP is
remarkable [46, 110, 131, 133]. Three different basalt types have been recovered
from this vast LIP, and of these three types the Kwaimbaita basalt makes up >
90% of the OJP [110, 131, 133]. My study is focused on plagioclase cumulate
xenoliths from the Kwaimbaita basalt. Kwaimbaita basalt is the most widespread
OJP basalt type and the only one of the three that contains cumulate xenoliths,
and the magma chamber processes that produced this ubiquitous composition
were the most common and important during formation of the OJP.
Key OJP research questions include: What are the physical processes (i.e.,
magma chamber dynamics) of differentiation that led to the repeated production
of the same basalt type (Kwaimbaita basalt) over a Greenland-size geographic
area? Hypothesis 1: OJP magma chambers were crystal-mush dominated, where
differentiation was heavily influenced by residence time of magmas in the crys-
tal mush. Near steady state flow of melt into and out of a vast crystal-mush-
dominated magma chamber system is amenable to repeated production of the
same type of basalt over a wide area. Crystals will be isotopically similar to
their host basalt, but will contain greater ranges of trace element enrichments and
depletions than the whole-rock host basalt.
Hypothesis 2: Nucleation and growth of crystals occurred homogenously through-
out the magma chamber interior. Plagioclase crystals will contains similar isotopic
and trace element compositions as the whole-rock host basalt.
21

1.7.2 Detroit Seamount, Part of the Emperor Seamount Chain, NW Pacific
Ocean
Detroit Seamount is located near the northern terminus of the Emperor Seamount
Chain and formed when the Hawaiian hot spot was adjacent to a mid-ocean ridge
during the Late Cretaceous (76-81 Ma) [34, 43] (Fig. 1.4). Interaction of hot
spots and mid-ocean ridges (MOR) often have profound effects on the surround-
ing lithosphere [47, 62–64, 79, 119]. Geochemically anomalous hot spot and MOR
basalts are common in areas of hot spot-MOR interaction and are often linked
to unique mantle processes resulting from the interaction [79, 119]. Several re-
cent studies have noted incompatible trace element and isotopically depleted hot
spot basalts from Detroit Seamount e.g., [72, 79, 119]. Detroit Seamount basalt
compositions extend from N-type MORB-like to compositions intermediate be-
tween N-MORB and young Hawaiian tholeiitic basalts [72, 79, 119]. Frey et al.
[50] presented compositional evidence linking DSM basalts to the Hawaiian hot
spot and concluded that it is unlikely DSM basalts are simply MORB. Several
hypotheses have been put forward to explain the petrogenesis of depleted hot
spot basalts at DSM. Keller et al. [79] suggested that the rising Hawaiian plume
entrained MORB-source upper mantle to an extent that partial melting of the
MORB source component overwhelmed the hot spot source signature in DSM
magmas. Regelous et al. [119] favored an alternative model, where a depleted
component inherent to the Hawaiian plume was more apparent in DSM magmas
due to greater partial melting under a thinner lid of lithosphere near the spreading
center. Huang et al. [72] noted the significance of fractionation and accumulation
of olivine and plagioclase, and thus an implied of signifigance low pressure partial
crystallization (e.g., [130]) during the evolution of DSM magmas in the crust. If
22

one is then to relate DSM basalt chemistry to the nature of a mantle source region
or some mantle process, it is vital to understand the extent to which bulk DSM
magma compositions were influenced during their transit through the crust. Once
the extent of magma evolution in the shallow crust is constrained, understanding
the roles of mantle processes becomes less convoluted. I employ an integrated tex-
tural and microanalytical approach to constrain shallow magma evolution during
the formation of Detroit Seamount by 1) identification of crystal populations with
related physical growth histories and 2) understanding the provenance of these
crystals.
Hypothesis 1: Magma mixing occurred between OIB and MORB magmas,
as the Hawaiian plume was close to a MOR [95]. If this is the case, evidence
supporting magma mixing, such as curved CSDs (e.g., [67]) from mixed crystal
populations (e.g., [139]) should be apparent. If magma mixing did occur, and
the plume-ridge proximity was least when the Site 884 (∼81 Ma) basalts were
erupted. These basalts should contain a better record of mixing relative to Site
1203 (∼75 Ma) basalts. Plagioclase crystals grown after the mixing event will
reflect the trace element and isotopic composition of the hybridized melt and will
appear in textural equilibrium with the melt. Zones grown before the mixing
event will bear the trace element and isotopic compositions of respective mixing
end-members. Some crystals may contain compositional evidence of exposure and
growth end-member magmas.
Hypothesis 2: The ascending OIB plume punched through and entrained
MORB source rocks as well as MORB intrusive and extrusive rocks. Melting of
the source rocks prior to plagioclase crystallization would have led to MORB-like
signatures in initial growth zones in crystals from the oldest plagioclase popu-
23

lations. Reprocessing of intrusive and extrusive MORB materials by ascending
plume magmas would be evident in CSDs as crystal accumulation (assuming only
partial resorption of the debris; e.g., [98]) and the presence of partially resorbed
plagioclase crystals. Late crystallizing plagioclase would reflect growth from a
hybridized magma, whereas entrained crystals would generally have MORB sig-
natures.
Hypothesis 3. The melt source was entirely related to the Hawaiian hotspot.
Ascending hotspot magmas passed through a complex magma chamber system
where they picked up crystal debris leading to curved CSDs. There is no evidence
of distinct MORB or OIB end-members. Compositional variations are ascribed
to variations in melting of a single heterogeneous source and variations in shallow
magma chamber dynamics over time.
1.7.3 The Kerguelen Plateau’s Western Salient - Elan Bank, Southern Indian
Ocean
The submarine Kerguelen Plateau (KP) and Broken Ridge constitute the sec-
ond largest oceanic LIP on Earth and rise up to 4 km above the surrounding
Indian Ocean basin (Fig. 4.1). Current sampling of the Cretaceous portion of the
KP has been done via drilling at 11 drill sites during Ocean Drilling Program
Legs 119, 120, and 183. Prior to Leg 183 drilling a number of workers suggested
that continental crust was involved during the petrogenesis some KP basalts (e.g.,
[51, 92]). Drilling at Site 1137 on Elan Bank confirmed hypotheses of continental
crust involvement when clasts of garnet-biotite gneiss were recovered from a flu-
vial conglomerate unit (Unit 6 in Fig. 4.2) [33, 74]. Nicolaysen et al. [111] and
Frey et al. [49] suggested that during the break up of the Gondwana superconti-
24

nent fragments of old continental crust were stranded amongst the Indian Ocean
lithosphere including fragments within the KP crust.
One of the scientific objectives of Leg 183 drilling was to constrain the post-
melting evolution of Kerguelen magmas, of which crustal assimilation was clearly
an important process. In this study I explore the timing and nature of crustal con-
tamination of KP magmas by focusing on the compositional and isotopic record of
assimilation contained within zoned plagioclase phenocrysts in two basalts from
Leg 183 Site 1137 - Units 4 and 10. On the basis of whol-rock geochemistry, Ingle
et al. [75] placed the Unit 4 basalt in a relatively uncontaminated upper basalt
group and and the Unit 10 basalt in a relatively contaminated lower basalt group.
Examination of plagioclase phenocrysts from Units 4 and 10 provide snapshots of
the magmatic processes that were occurring when both when crustal contamina-
tion was significant and when it was not.
Hypothesis 1: Initial LIP magma emplacement in the crust was accompanied
by large amounts of crustal assimilation followed by armoring of magma chamber
walls. Crystallization occurred by solidification front growth. In this case early
formed plagioclase crystals or zones (i.e., cores) will contain the greatest signa-
ture of the crustal wall rocks, however the contamination signature in the crystal
cores may be indistinct if contamination significantly pre-dates plagioclase crys-
tallization. Late formed crystals or zones (i.e., rims) will contain less of a crustal
signature. Hypothesis 2: Crustal assimilation occurred progressively as crystals
25

nucleated homogenously and grew throughout the magma chamber interior. Early
formed crystals or zones (i.e., cores) will not necessarily bear the most contami-
nated signatures. Late formed crystals or zones (i.e., rims) will be equally or more
contaminated than early formed crystals.
26

CHAPTER 2
MAGMA EVOLUTION REVEALED BY ANORTHITE-RICH PLAGIOCLASE
CUMULATE XENOLITHS FROM THE ONTONG JAVA PLATEAU:
INSIGHTS INTO LIP MAGMA DYNAMICS AND MELT EVOLUTION
2.1 Introduction
Petrologic studies have historically relied upon whole-rock chemistry alone to
deduce evolutionary processes that modify magmas. Textural and microanalyti-
cal studies focused on plagioclase have been recently used to understand magma
chamber processes and the chemical evolution of basaltic to silicic magmas e.g.,
[5, 6, 16, 39, 135]. Plagioclase is well suited for microanalysis, as it contains
measurable quantities of select incompatible trace elements, including the light
rare earth elements (LREE). It is often preceded on the liquidus only by olivine,
which does not appreciably fractionate the REE or other incompatible trace ele-
ments in the residual liquid [5]. In this study I use major, minor, and trace element
abundances measured in plagioclase by electron probe microanalysis (EPMA) and
laser ablation ICP-MS (LA-ICP-MS) to invert chemical compositions of basaltic
parental (equilibrium) magmas from a large igneous province (LIP) to investigate
magma evolution. Accurate partition coefficient data are critical for this inversion.
Plagioclase partition coefficients must be applied carefully because anorthite (An
= 100*[Ca/(Ca+Na)]) content is a controlling factor of cation substitution [7, 10].
27

With accurate partition coefficients, inferred parent magma compositions provide
insight into magma evolution beyond what is revealed by whole-rock data alone.
I examine basaltic magma evolution and magma chamber process of the Ontong
Java Plateau (OJP), a mid-Cretaceous large igneous province (LIP) (Fig. 2.1).
The uniformity of basalt composition over the Greenland-size OJP is remarkable
[46, 110, 131, 133]. Three different basalt types have been recovered from this vast
LIP, and of these three types the Kwaimbaita basalt makes up > 90% of the OJP
[110, 131, 133]. My study is focused on plagioclase cumulate xenoliths from the
Kwaimbaita basalt (Figs. 2.2, 2.3). Kwaimbaita basalt is the only one of the three
OJP basalt types that contains the xenoliths, and the magma chamber processes
that produced this ubiquitous basalt type were the most common and important
during formation of the OJP. Previous studies have suggested shallow (< 6-8
km) fractional crystallization of olivine, plagioclase, and clinopyroxene was a ma-
jor mechanism of OJP magma differentiation [46, 103, 110, 121, 123, 133]. Sano
and Yamashita [123] presented a magma chamber model to explain the narrow
range of OJP basalt chemistry. An outstanding question in their study of OJP
basalt petrography and phase equilibria was the origin of An-rich (> An80) pla-
gioclase, which they demonstrated through experiments to be out of equilibrium
with Kwaimbaita basalt (Sano and Yamashita, [123]). I evaluate whether An-rich
crystals grew in an evolved water-rich boundary layer, which is the model favored
by Sano and Yamashita [123], or whether they formed by an alternate process
such as growth in a hotter more primitive magma. I also examine whether the
28

Figure 2.1. Predicted bathymetry of the Ontong Java Plateau (OJP)
(modified after Mahoney et al. [91]). The locations of Ocean Drilling
Program (ODP) and Deep Sea Drilling Project (DSDP) drill sites on the
OJP are shown with the basalt types recovered at each location.
Kw=Kwaimbaita basalt, Kr=Kroenke basalt, and Sg=Singgalo basalt.
29

uniform composition of large volumes of OJP basalt represent an averaging of
more varied compositions, which may be reflective of important magma chamber
processes. Finally, I examine how these processes relate to the spatial distribution
of known basalt types across the OJP.
2.2 Geologic Background of the Ontong Java Plateau
The Ontong Java Plateau is the worlds largest oceanic LIP spanning an area of
roughly 2.018 x 106 km 2 in the Southwest Pacific ocean [30, 32, 93, 94, 132, 133]
(Fig. 2.1). Tarduno et al. [129] and Mahoney et al. [93, 94] suggested that the bulk
of the OJP formed in a submarine eruptive environment relatively rapidly around
∼122 Ma, with a minor event at ∼90 Ma [93, 94, 132, 133]. Most of the OJP is
presently submerged, exceptions being the southern plateau margin on the islands
of Malaita, San Cristobal (Makira), and Santa Isabel in the Solomon Islands,
where obducted segments now outcrop [116] (Fig. 2.1). Despite its size, only
three subtly different low-K tholeiitic basalt types have been recovered from the
OJP, either by field work or ocean drilling, which have been referred to as Singgalo,
Kwaimbaita, and Kroenke basalts [46, 131, 133] (Fig. 2.1). Singgalo basalts are
isotopically distinct and overlie Kwaimbaita basalt on Malaita and ∼1,500 km to
the north at Ocean Drilling Program (ODP) Site 807 [93, 94, 110, 133] (Fig. 2.1).
Incompatible trace element abundances are slightly enriched and MgO abundances
lower (6-7.3 %) in Singgalo basalts relative to Kwaimbaita basalts (7-8 %) [46,
93, 94, 110, 131–133]. Kroenke basalts have higher MgO (8-11 wt.%) and lower
incompatible trace element abundances than Kwaimbaita basalts [46]. However,
these relatively primitive basalts are isotopically indistinguishable from the more
fractionated Kwaimbaita basalts and share the same ∼122 Ma age [46, 131]. Fitton
30

and Godard [46] and Tejada et al. [131] suggested Kroenke basalts are parental
to Kwaimbaita basalts, being related by olivine fractionation. On the basis of
current sampling Kroenke and Singgalo basalts represent minor components of
the overall OJP volume and have been recovered primarily from the margins of
the plateau [46] (Fig. 2.1).
2.3 Samples
The samples examined in this study were collected from outcrops along the
Singgalo River on Malaita and by drill core at ODP Sites 807 (Leg 130) and 1183
(Leg 192), which are shown in figure 2.1 with the basalt types recovered at each
location (see also Figs. 2.2, 2.3). All xenolith crystals and phenocrysts examined
in this study are from Kwaimbaita basalt and the ∼122 Ma eruptive event [21, 46,
93, 94, 131, 133]. Malaita is located along the southern margin of the plateau (see
Figs. 2.1, 2.3). Site 807 is along the northern margin of the OJP, and the xenoliths
are from Units C-G, which are equivalent to the Kwaimbaita basalts (Figs. 2.1,2.3).
Site 1183 is near the bathymetric high of the plateau, where 80.7 m of Kwaimbaita
pillow basalt was cored, representing 8 flow units [91] (Fig. 2.4). Plagioclase
cumulate xenoliths (Fig. 2.2) are found in all but Unit 1 of the cored section at
Site 1183 [91] (Fig. 2.4). Kwaimbaita basalt is found as massive and pillow basalt
flows, generally has a subophitic to intergranular texture, and < 2% (volume)
phenocrysts [91, 123, 133]. The dominant phenocrysts in Kwaimbaita basalts are
olivine and plagioclase with lesser amounts of clinopyroxene. I measured major
and trace elements in single large crystals from one Malaitan xenolith (host basalt
SGB-21, c.f., [133]; Fig. 2.3a-d), from three Site 1183 xenoliths (Units 5B, 6,
and 7; (see Fig. 2.2a-d), and from a single glomerocryst/xenolith from Site 807
31

Figure 2.2. Hand sample and thin section photographs of plagioclase rich
xenoliths hosted in Kwaimbaita basalt from ODP Site 1183. A) hand sample
of Unit 5B xenolith (192-1183A-59R2 112-117 cm piece #10A); B) Mosaic
photomicrograph of Unit 7 xenolith (192-1183A-64R2 116-120 cm piece
#9B). Note network of plagioclase crystals and labeled resorption feature.
All visible crystals are plagioclase; C) Photomicrograph showing partially
resorbed clinopyroxene and devitrified glass in a Unit 5B xenolith (adapted
from [91]); D) An altered olivine crystal along the margin of Unit 7 xenolith;
E) Photomicrograph of Unit 6 xenolith illustrating the xenolith interior and
exterior and a resorption feature (192-1183A-63R2 25-27 cm piece #4); F)
Photomicrograph of the margin of a Unit 6 xenolith. The xenolith exterior
(i.e., rim) and a clinopyroxene crystal are visible.
32

Figure 2.3. Hand sample and thin section photographs of plagioclase rich
xenoliths hosted in Kwaimbaita basalt from ODP Site 807 and Malaita. A)
A boulder of Kwaimbaita basalt (SGB-21) from along the Singgalo River,
Malaita exhibiting large plagioclase-rich xenoliths. Note rock hammer for
scale; B) Mosaic photomicrograph of the interior of Malaita xenolith ML-X1,
exhibiting oscillatory zoning; C) Rim of xenolith ML-X1 illustrating
destruction of the rim zones by alteration and a clinopyroxene crystal present
along the rim; D) Small melt inclusions within ML-X1 that run parallel to
zoning features; E) Mosaic photomicrograph of xenolith 807-X1 from ODP
Site 807 (130-807C-93R-1-137-139). Note olivine and clinopyroxene along the
xenolith margins; F) Altered olivine along the margin of 807-X1.
(see 2.3e). I also measured major and trace element data at select points in a
variety of phenocrysts from Site 1183 Units 5B, 6, and 7 basalts (Fig. 2.5a-d).
I did not examine any plagioclase phenocrysts from Kwaimbaita basalt SGB-21
from Malaita or the Site 807 Kwaimbaita basalt. Detailed descriptions of zoning
patterns of the xenolith plagioclase crystals and phenocrysts are provided in the
results section.
33

2.3.1 A Cogenetic Relationship Between Xenoliths?
Isotopic studies were not conducted on any of the plagioclase crystals in this
study. Tejada et al. [133] reported isotopic data for a single OJP plagioclase
xenolith, which was isotopically indistinguishable from its Kwaimbaita host basalt
(SGB-21, see sample description below). I therefore assume for this study that
the plagioclase-rich xenoliths, all in Kwaimbaita host basalt, are indistinguishable
from Kwaimbaita-type basalt.
2.4 Methods
2.4.1 Analytical Methods
Samples were prepared as polished thick sections (80-100 µm) with Bi-doped
epoxy to signal when the sample had been penetrated during LA-ICP-MS analy-
ses e.g., [76]. Bismuth was chosen as a dopant because naturally occurring Bi is
well below analytical detection limits in plagioclase. Backscatter electron images
and major element analyses were performed using a JEOL JXA-8600 Superprobe
electron microprobe at the University of Notre Dame. Backscatter electron im-
ages were taken using a 1 µm beam, an accelerating voltage of 20 kV, and a
probe current of 25-50 nA. Microprobe analyses were performed using a 5 µm
defocused beam, accelerating voltage of 15 kV or 20 kV depending upon samples,
a probe current of 20 nA, 15 second on-peak counting time, and two background
measurements per peak. I measured Na first to minimize loss via volatilization.
Microprobe data were corrected for matrix effects using a ZAF correction method.
Data points near Fe-rich phases such as melt inclusions and alteration-filled frac-
tures were discarded due to Fe fluorescence effects. Titanium was quantified by
EPMA. Strontium, Y, Ba, La, Ce, Nd, and Eu were measured in plagioclase
34

crystals using a New Wave UP-213 UV laser ablation system interfaced with a
ThermoFinnigan Element 2 ICP-MS operated in fast magnet scanning mode at
the University of Notre Dame. I used a laser frequency of 4 Hz, pulse energy of
0.02-0.03 mJ pulse −1 , 12 µm or 20 µm diameter pits depending on crystal size,
and helium as the carrier gas (∼ 0.7 l min −1 ) mixed with argon (∼ 1.0 l min −1 )
before introduction to the plasma. Analyses were conducted in peak jumping
mode with one point quantified per mass. The LA-ICP-MS spots were coincident
with previous EPMA analyses, and Ca measured by EPMA was used as an inter-
nal standard for each spot analysis. The trace element glass NIST 612 was used
as a calibration standard in all laser ablation analyses. Although heterogeneity
for certain elements has been documented in the widely used NIST 612 glass i.e.,
[44], Eggins et al. [44] considered it a reliable calibration standard for Sr, Ba, Y,
and the REE. The methods of Longerich et al. [88] and Norman et al. [112] were
used to reduce LA-ICP-MS data and calculate analytical uncertainty for each spot
analysis (Table 2.1).
2.4.2 Data Quality
The quality of EPMA data were monitored by major element and cation totals.
Data points where major element totals were greater than 101.5% or less than
98.5% are not used in further discussion nor are points with cation totals greater
or less than 20 ± 0.1 cations (based on 32 O). All laser ablation analyses were
collected in time-resolved mode so that signal from inclusions and alteration filled
fractures could be easily excluded. Results from laser ablation spots close to
fractures or that penetrated into underlying fractures were discarded. The rate of
laser ablation was tested using a 100 µm thick plagioclase wafer prior to sampling,
35

Figure 2.4. Stratigraphic section of basaltic basement cored at ODP Site
1183 consisting entirely of Kwaimbaita pillowed basalt flow units. The
locations of xenoliths are labeled with asterisks. The locations of xenoliths
examined in this study (59R2-X1, 63R2-X1, and 64R2-X1) are labeled.
36

which allowed optimization of operating conditions to yield desired sensitivity and
sampling depth (< 30 µm).
2.4.3 Choosing partition coefficients (D)
Blundy [9] concluded that inversion of magma compositions from mineral data
using appropriate D values to be a robust means of estimating parent magma com-
positions and suggested that D values derived from microbeam techniques were
more accurate than those derived from bulk crystal-matrix analyses. Bindeman
and Davis [6] noted that D values derived from doping experiments should be used
with caution as, for example, the DREE and DY in REE doped runs are 30-100%
higher than in undoped runs. Blundy and Wood [10] and Bindeman et al. [7]
showed that the dominant factors controlling the trace element Ds for plagioclase
are An content and crystallization temperature (pressure effects appeared to be
negligible). Bindeman et al. [7] applied this relationship to a wide variety of trace
elements and produced values for the constants “a” and “b” in their Equation 2
for calculating plagioclase partition coefficients via the expression:
RT ln(Di) = aXAn + b (2.1)
where R is the gas constant, T is temperature (Kelvin), and XAn is the mole
fraction of anorthite in the plagioclase. Using equation 2.1, Ginibre et al. [59]
noted that temperatures in the range of 850-1,000 ◦ C had only a small effect on
calculated melt concentrations for Sr and Ba (within their analytical uncertainty).
Likewise, Bindeman et al. [7] showed that variations ∼150 ◦ C produce < 10%
differences for particular partition coefficients, which were often within error of
the respective D values.
37

Figure 2.5. Backscatter electron images of plagioclase phenocrysts from Site
1183 basalts A) A type I normal zoned crystal (63R2P4, Unit 6). Note the
brighter, higher An core; B) A type II reverse zoned crystal with a darker,
lower An core (64R2P4, Unit 7); C) An olivine-plagioclase glomerocryst
(64R2P3, Unit 7). The plagioclase crystal has a sieve textured core; D) A
reverse zoned type II crystal that contains numerous melt inclusions
(64R2P1).
38

2.4.4 Partition Coefficients for OJP Plagioclase Crystals
I used the equations and constants of Bindeman et al. [7], derived from par-
titioning experiments run at natural (i.e., undoped) trace element concentration
levels at atmospheric pressure, to calculate D values for Sr, Y, Ba, La, Ce, Nd, and
Ti (see Table 2.2). I assumed a crystallization temperature of 1185 ◦ C to calculate
D values. Sano and Yamashita [123] reported initial crystallization of plagio-
clase at 1170 ◦ C from Kwaimbaita-type magmas and 1200 ◦ C from more primitive
Kroenke-type magmas. The use of 1170 ◦ C vs. 1200 ◦ C in D calculations yields <
4% differences in trace element concentrations of parent liquids, within the ana-
lytical error for most of my analyses. I neglect pressure effects since they are not
well documented for most elements, and where they were documented for DSr by
Vander Auwera et al. [137] they were negligible. Magmatic water can affect D val-
ues by decreasing the activity of trace element components in silicate melts [144].
However, Kwaimbaita basaltic glass samples have relatively low H2O contents
(generally ≤ 0.22 wt %; [121]), so we estimate the effect of water on OJP plagio-
clase D values to be negligible. The partitioning studies of Bindeman et al. [7]
were run in air such that their equations and constants for calculation of DEu and
DFe (hereafter DFeOT for total Fe as FeO) neglect the effects of magma oxygen fu-
gacity (f O2). Oxygen fugacity affects the partitioning of Fe and Eu in plagioclase
[142]. Trivalent Fe substitutes for Al 3+ in the tetrahedral site, Eu 2+ substitutes in
the plagioclase M cation site at low f O2, and Eu 3+ present at higher f O2 is incom-
patible [41, 89, 127]. I estimated a DEu ≈ 0.43 graphically from Blundy [9] (his
Fig. 3: DEu vs.∆QFM of basaltic magma), where it was demonstrated that DEu
increased systematically with decreasing f O2 in basalts. I estimate DFeOT ≈ 0.065
for OJP plagioclase. This value is an average from experimental crystallization
39

studies of Kwaimbaita basalt by Sano and Yamashita [123] at the QFM buffer,
which is the approximate f O2 condition of OJP basalt crystallization suggested
by Roberge et al. [121]. This estimate of DFeOT is supported by the results of
Phinney [117] who reported little change in DFeOT at f O2 below the QFM oxy-
gen buffer, while DFeOT increased steadily at higher f O2, which was attributed to
greater amounts of Fe 3+ in the melt. If generally constant partitioning of Mg be-
tween plagioclase and melt is assumed (i.e., using the method of Bindeman et al.
[7]), we calculate an average DMg value of 0.024 ± 0.005 for the OJP plagioclase
crystals considered in this study. The minor variation of DMg, using the Bindeman
et al. [7] method, over the large compositional range of plagioclase examined in
this study (An65−86) is in accordance with Longhi [89] and Bindeman et al. [7]
who suggested An content has minimal effect on Mg partitioning. However, the
average DMg values calculated using the equations and constants of Bindeman et
al. [7] are inconsistent with the experimental results of Sano and Yamashita [123],
whose results for OJP basalts reveal slightly larger values for Kwaimbaita (0.1
MPa DMg = 0.043; 190 MPa DMg = 0.055) and Kroenke basalts (0.1 MPa DMg
= 0.04; 190 MPa DMg = 0.06) at elevated pressures. This suggests crystallization
pressure may have an effect on DMg [123]. I use DMg ≈ 0.047 to invert parent
magma compositions of OJP plagioclase crystals, which is an average of eight DMg
values from the experimental results of Sano and Yamashita [123] conducted at
0.1 and 190 MPa. Phinney [117] reported an average basaltic plagioclase DMg
value of 0.044, which is consistent with my selected DMg value.
40

41
Crystal
Zone 1
An
(mol%)
ML-X1 xenolith
TABLE 2.1
MAJOR AND TRACE ELEMENT DATA: PLAGIOCLASE
PHENOCRYSTS AND XENOLITH PLAGIOCLASE CRYSTALS
SiO2 TiO2 Al2O3 FeO T 2 MgO CaO Na2O K2O P2O5 Total Sr Y Ba La Ce Nd Eu 3
Interior 85 47.5 0.02 31.9 0.54 0.22 17.4 1.71 0.07 0.17 100.8 172.4 0.21 2.31 0.24 0.59 - 0.25
Interior 85 48.0 0.01 32.4 0.55 0.23 17.5 1.67 - 0.13 100.5 161.3 0.18 2.99 0.17 0.25 0.37 0.34
Interior 85 47.7 0.02 31.4 0.52 0.23 17.3 1.65 0.05 0.17 100.5 176.9 0.19 1.93 0.20 0.37 0.39 0.35
Interior 86 47.6 0.01 32.2 0.56 0.23 17.7 1.63 0.04 0.13 100.4 187.0 0.13 2.18 0.25 0.36 0.76 0.42
Interior 84 48.0 0.01 31.7 0.56 0.25 17.4 1.79 0.09 0.13 100.4 172.7 - 2.90 0.39 0.36 0.38 0.47
Interior 86 47.7 0.03 32.6 0.56 0.23 18.1 1.57 0.03 0.17 99.8 185.3 0.43 3.12 0.30 0.54 0.23 -
59R2-X1 xenolith
Exterior 66 55.4 0.03 26.2 1.01 0.21 13.4 3.67 0.15 0.16 100.2 - - - - - - -
Exterior 75 50.6 0.04 29.6 0.72 0.32 16.1 3.98 0.03 0.13 100.5 225.7 0.19 4.97 0.43 0.65 0.49 0.49
Interior 81 49.3 0.02 30.1 0.65 0.26 17.5 2.32 0.03 0.18 100.4 - - - - - - -
continued...

42
Crystal
Zone
An
(mol%)
TABLE 2.1
Continued
SiO2 TiO2 Al2O3 FeO T MgO CaO Na2O K2O P2O5 Total Sr Y Ba La Ce Nd Eu
Interior 82 50.2 0.03 29.4 0.70 0.27 17.8 2.18 0.01 0.17 100.8 199.7 0.14 3.84 0.27 0.57 0.30 0.45
Interior 83 48.8 0.02 31.0 0.65 0.24 17.6 1.94 0.01 0.15 100.4 218.7 0.18 4.65 0.22 0.57 0.52 0.45
Interior 82 49.5 0.01 29.8 0.67 0.28 17.8 2.11 0.02 0.21 100.4 203.5 0.12 3.64 0.20 0.61 0.42 0.47
Center/core 83 48.8 0.01 30.7 0.66 0.25 18.0 1.98 0.02 0.14 100.5 206.0 0.19 5.05 0.31 0.58 0.64 0.34
59R2 phenocrysts P1-P4
P1 Rim 71 51.7 0.03 27.8 0.84 0.33 15.0 3.35 0.02 0.16 99.3 202.2 0.29 6.06 0.35 0.70 0.78 0.38
P1 Core 74 51.0 0.03 29.4 0.74 0.28 15.4 2.97 0.01 0.12 100.0 187.4 0.13 4.30 0.36 0.54 0.42 0.42
P2 Rim 70 51.4 0.04 29.2 0.84 0.33 14.8 3.44 0.02 0.12 100.3 207.1 0.22 5.26 0.39 0.71 0.46 0.47
P2 Core 73 51.2 0.03 29.2 0.75 0.34 15.5 3.13 0.02 0.15 100.3 195.4 0.34 4.56 0.46 0.52 0.47 0.39
P3
Center
P4
Center
69 51.0 0.03 29.4 0.85 0.34 14.3 3.50 0.03 0.14 99.6 - - - - - - -
67 52.1 0.03 28.6 0.89 0.37 13.7 3.77 0.03 0.15 99.7 - - - - - - -
continued...

43
Crystal
Zone
An
(mol%)
63R2-X1 xenolith
TABLE 2.1
Continued
SiO2 TiO2 Al2O3 FeO T MgO CaO Na2O K2O P2O5 Total Sr Y Ba La Ce Nd Eu
Exterior 79 47.6 0.03 31.2 0.74 0.28 17.40 2.32 0.02 0.22 99.9 178.1 0.27 3.04 0.20 0.43 0.51 0.22
Exterior 77 48.8 0.03 30.7 0.76 0.31 16.40 2.69 0.02 0.18 100.2 180.3 0.22 4.31 0.13 0.29 0.24 0.28
Exterior 81 48.2 0.02 31.0 0.70 0.25 17.50 2.24 0.01 0.20 100.1 168.9 0.34 3.41 0.22 0.33 0.04 0.22
Exterior 75 50.1 0.02 29.8 0.60 0.26 16.10 2.90 0.03 0.21 100.1 160.8 0.21 3.32 0.22 0.36 0.33 0.26
Interm. 85 48.3 0.01 31.7 0.65 0.23 18.10 1.79 0.01 0.19 100.9 175.0 0.21 2.73 0.19 0.38 0.34 0.26
Resorp. 81 48.9 0.03 30.6 0.62 0.26 17.50 2.26 0.02 0.20 100.3 171.9 0.10 2.91 0.26 0.27 0.32 0.26
Interm. 85 48.0 0.04 31.2 0.66 0.24 18.10 1.78 0.02 0.22 100.2 177.9 0.14 2.66 0.17 0.41 0.27 0.28
Interm. 85 48.0 - 31.3 0.67 0.23 18.30 1.82 - 0.20 100.5 171.80 0.18 2.91 0.13 0.34 0.17 0.39
Interm. 82 48.7 - 30.8 0.68 0.27 17.70 2.10 0.01 0.20 100.4 172.0 0.18 2.82 0.17 0.39 0.31 0.28
Interm. 83 48.5 0.01 31.2 0.63 0.26 17.50 1.91 0.02 0.21 100.3 170.9 0.13 2.98 0.16 0.39 0.33 0.22
Interm. 84 48.5 0.02 30.9 0.64 0.25 18.20 1.99 0.02 0.22 100.6 169.6 0.25 2.93 0.19 0.57 0.41 0.25
Interm. 85 48.2 0.02 31.2 0.62 0.25 18.20 1.77 0.01 0.19 100.5 170.4 0.15 2.88 0.19 0.50 0.22 0.35
continued...

44
Crystal
Zone
An
(mol%)
63R2 phenocrysts P1-P4
TABLE 2.1
Continued
SiO2 TiO2 Al2O3 FeO T MgO CaO Na2O K2O P2O5 Total Sr Y Ba La Ce Nd Eu
P1 Core 73 51.5 0.02 29.0 0.81 0.33 14.8 2.94 0.03 0.17 99.6 - - - - - - -
P1
Interm
P1
Interm
76 50.5 0.03 29.7 0.83 0.33 16.0 2.70 0.02 0.17 100.3 236.4 0.25 8.57 0.43 0.72 0.17 0.24
74 51.1 0.03 29.5 0.81 0.30 15.6 3.06 0.04 0.15 100.6 173.3 0.12 4.64 0.14 0.49 0.28 0.10
P1 Rim 70 52.5 0.07 28.4 0.97 0.35 14.8 3.43 0.20 0.18 100.9 - - - - - - -
P2
Center
77 50.0 0.03 30.0 0.82 0.31 16.3 2.75 0.02 0.16 100.4 177.6 0.22 4.87 0.31 0.55 0.09 0.32
P3 Core 73 51.1 0.03 29.7 0.95 0.53 14.2 2.90 0.03 0.14 99.6 - - - - - - -
P3 Rim 68 51.9 0.03 28.7 0.97 0.50 13.7 3.57 0.05 0.17 99.6 195.0 0.29 5.98 0.31 0.74 0.56 0.30
P4 Core 75 50.2 0.03 30.4 0.74 0.29 15.0 2.80 0.05 0.20 99.8 225.9 0.11 8.49 0.31 0.62 0.28 0.33
P4 Rim 67 52.5 0.03 28.5 0.95 0.40 13.5 3.67 0.05 0.12 99.7 161.7 0.13 3.63 0.33 0.69 0.33 0.20
continued...

45
Crystal
Zone
An
(mol%)
64R2-X1 xenolith
TABLE 2.1
Continued
SiO2 TiO2 Al2O3 FeO T MgO CaO Na2O K2O P2O5 Total Sr Y Ba La Ce Nd Eu
Exterior 71 51.7 0.03 29.4 0.99 0.28 14.9 3.29 0.03 0.16 100.7 157.1 0.13 2.70 0.22 0.37 0.45 0.35
Exterior 79 49.9 - 30.5 0.63 0.28 16.2 2.43 0.01 0.16 100.1 169.3 0.13 3.04 0.21 0.40 0.24 0.39
Interior 80 49.3 0.02 30.5 0.63 0.25 16.2 2.24 0.01 0.18 99.3 165.9 0.21 2.90 0.22 0.37 0.53 0.38
Interior 84 48.6 0.01 30.4 0.62 0.22 17.2 1.84 0.01 0.17 99.1 181.6 0.21 3.10 0.16 0.43 0.08 0.41
Interior 82 49.3 0.02 31.0 0.61 0.25 16.9 2.06 0.02 0.18 100.3 168.2 0.20 2.66 0.25 0.36 0.28 0.32
Interior 82 49.0 0.02 30.7 0.56 0.25 16.6 1.95 0.01 0.18 99.7 161.2 0.20 3.23 0.16 0.40 0.19 0.21
Interior 82 48.7 0.01 31.5 0.53 0.26 16.6 1.96 0.02 0.17 99.8 162.0 0.19 2.48 0.18 0.38 0.39 0.44
Interior 82 48.5 0.02 31.1 0.58 0.27 16.5 2.03 0.02 0.20 99.1 168.7 0.19 2.77 0.10 0.41 0.32 0.41
Resorp. 83 48.0 - 31.9 0.57 0.24 16.7 1.93 0.03 0.21 99.5 167.1 0.30 2.53 0.14 0.38 0.21 0.35
Interior 80 50.0 - 31.0 0.54 0.26 15.6 2.21 0.02 0.17 99.8 155.9 0.16 2.45 0.21 0.30 0.49 0.36
64R2 phenocrysts P1-P4
continued...

46
Crystal
Zone
An
(mol%)
TABLE 2.1
Continued
SiO2 TiO2 Al2O3 FeO T MgO CaO Na2O K2O P2O5 Total Sr Y Ba La Ce Nd Eu
P1 Core 69 52.6 0.02 29.2 0.53 0.22 14.4 3.61 0.02 0.15 100.6 186.7 0.17 4.06 0.18 0.40 0.31 0.39
P1
Interm.
P1
Interm.
78 50.7 0.02 30.2 0.67 0.29 16.3 2.56 - 0.14 100.7 172.6 0.13 2.88 0.10 0.33 0.20 0.18
74 51.5 0.03 29.1 0.67 0.29 15.1 2.99 - 0.15 99.7 170.0 0.12 5.38 0.15 0.41 0.17 0.25
P1 Rim 67 52.5 0.04 28.3 0.96 0.30 13.9 3.82 0.05 0.14 100.0 - - - - - - -
P2 Core 83 48.8 0.02 31.4 0.66 0.22 16.5 1.84 - 0.15 99.60 195.2 0.48 4.91 0.43 0.53 0.48 0.43
P2 Rim 79 50.0 0.03 29.8 0.67 0.28 16.2 2.42 0.02 0.13 99.6 - - - - - - -
P4 Core 66 53.2 0.02 27.8 0.50 0.23 13.5 3.83 0.03 0.10 99.2 - - - - - - -
P4
Interm.
P4
Interm.
79 48.2 0.04 30.3 0.68 0.25 16.1 2.34 0.01 0.18 99.6 167.9 0.38 2.63 0.18 0.70 0.85 0.27
74 51.0 0.04 28.8 0.73 0.30 14.8 3.26 0.03 0.17 99.9 182.7 0.38 4.81 0.32 0.72 0.44 0.21
P4 Rim 65 52.6 0.06 27.7 1.04 0.26 13.3 3.99 0.03 0.11 99.0 - - - - - - -
P3 Core 78 50.3 0.01 29.8 0.68 0.29 15.8 2.39 0.03 0.13 99.5 183.0 0.26 3.27 0.09 0.45 0.14 0.13
P3
Interm.
75 51.0 0.03 29.4 0.71 0.29 15.8 2.85 0.02 0.16 100.3 180.9 0.27 4.87 0.38 0.44 0.28 0.34
continued...

47
Crystal
Zone
An
(mol%)
807-X1 xenolith
TABLE 2.1
Continued
SiO2 TiO2 Al2O3 FeO T MgO CaO Na2O K2O P2O5 Total Sr Y Ba La Ce Nd Eu
Core 83 47.0 - 31.30 0.78 0.26 18.7 2.02 0.18 - 100.2 - - - - - - -
Core 84 47.0 - 31.50 0.65 0.22 18.9 1.88 0.14 - 100.2 - - - - - - -
Core 84 47.2 - 31.30 0.89 0.29 18.6 1.94 0.11 - 100.2 - - - - - - -
Interm. 83 47.4 - 31.30 0.71 0.22 18.6 2.02 0.15 - 100.5 - - - - - - -
Interm. 83 46.6 - 31.50 0.72 0.27 18.9 2.14 0.05 - 100.2 - - - - - - -
Interm. 83 47.1 - 31.30 0.76 0.27 18.9 2.06 0.11 - 100.5 - - - - - - -
Rim 74 49.7 - 28.90 0.68 0.28 16.2 2.99 0.19 - 98.9 - - - - - - -
Rim 71 50.8 - 28.30 0.79 0.32 15.6 3.41 0.16 - 99.3 - - - - - - -
Rim 74 50.1 - 29.10 0.73 0.29 16.0 2.97 0.20 - 99.4 - - - - - - -
Average detection limits
continued...

48
Crystal
Zone
An
(mol%)
TABLE 2.1
Continued
SiO2 TiO2 Al2O3 FeO T MgO CaO Na2O K2O P2O5 Total Sr Y Ba La Ce Nd Eu
12 µm pit - - - - - - - - - - - 0.33 0.14 0.10 0.07 0.03 0.05 0.06
25 µm pit - - - - - - - - - - - 0.12 0.06 0.07 0.04 0.03 0.03 0.03
NIST
614 4
measured
NIST
614 error
measured
NIST 614
recommended
- - - - - - - - - - - 44.7 0.76 3.25 0.73 0.81 0.73 0.73
- - - - - - - - - - - ±1.9 ±0.06 ±0.27 ±0.02 ±0.1 ±0.04 ±0.03
- - - - - - - - - - - 45.1 0.80 3.13 0.71 0.78 0.75 0.73
1 Interm. = Intermediate area between core and rim of largest discernible plagioclase crystal in each xenolith.
2 Fe reported as total FeO
3 Major elements in wt.% and trace elements in ppm
4 NIST 614 values reported are the average of ten measurements

2.5 Results
2.5.1 Plagioclase Zoning Patterns
2.5.1.1 Malaita Xenolith(ML-X1)
Plagioclase cumulate xenoliths from Malaita are round to subround, generally
1-8 cm in diameter, consist entirely of interlocking plagioclase crystals, and are
generally moderately altered (Figs. 2.3a-d). Alteration is evident as fracturing
and discoloration of xenolith crystals (e.g., Fig. 2.3). The xenolith I examined
from Malaita (ML-X1) is larger than xenoliths from Site 1183 or 807 (6.7 cm
diameter, roughly twice the diameter than the largest Site 1183 xenolith from Unit
7). ML-X1 contains zones of small devitrified melt inclusions oriented parallel to
oscillatory zoning and is completely altered along its rim (Figs. 2.3c,e). It also
contains small olivine (now completely altered to clay minerals) and clinopyroxene
grains along its margins (Fig. 2.3c). I collected major and trace element data from
the least altered interior section of ML-X1, which exhibited oscillatory zonation
(Figs. 2.3b, 2.6a).
2.5.1.2 Site 1183 Xenoliths
The three rounded xenoliths I examined from Site 1183 are texturally similar,
and Mahoney et al. [91] suggested they have a cumulate origin (Figs. 2.2a,b).
The Site 1183 xenoliths contain only sparse melt inclusions, are primarily oscil-
latory zoned, and are 0.5-3 cm in diameter [91] (Fig. 2.2). There are distinct
resorption/reaction rims on each of the crystals that are common to the entire
xenolith rather than individual crystals (Figs. 2.2e,f). I refer to these common
rim sections as xenolith exteriors (see Fig. 2.2e). All three Site 1183 xenoliths
contain clinopyroxene and completely altered olivine grains in similar proportions
49

along the xenolith margins (e.g., Figs. 2.2d,f). Similarly, all three xenoliths con-
tain patches of devitrified glass and minor clinopyroxene that appears to have
been almost completely melted out (Mahoney et al., 2001; Fig. 2.2c). The Site
1183 xenolith crystals I examined contain few resorption features, exceptions are
the core region of the Unit 7 xenolith crystal (Fig. 2.2b) and a small resorption
surface in the Unit 6 xenolith (Fig. 2.2e). Selected major and trace element data
collected along core to rim transects in each crystal are displayed in Table 2.2 (see
also Figs. 2.6b-d).
2.5.1.3 Site 1183 Phenocrysts
Sano and Yamashita [123] documented two types of zoned plagioclase phe-
nocrysts, Types I and II, in Kwaimbaita basalts from Site 1183 and other ODP
drill sites. Type I crystals are generally euhedral, exhibit normal zonation, and
have An73 to An78 cores (e.g., Fig. 2.5a). Type II crystals have complex zon-
ing, exhibit reverse or oscillatory zonations, and have cores in the An73 to An78
range that are commonly surrounded by An80 to An84 zones [123] (e.g., Fig. 2.5b).
Representative examples of the different zoning types are shown in Figure 2.5. I
observed similar zoning patterns in the Units 5B, 6, and 7 Kwaimbaita basalts
from Site 1183. Basalts from Units 5B and 6 contain primarily Type I normal
zoned phenocrysts, whereas the Unit 7 basalt I examined contains a mixture of
Type I normal and Type II reverse zoned phenocrysts (Table 2.1). Plagioclase phe-
nocrysts from Unit 7 exhibit other morphological differences from Units 5B and
6 phenocrysts. For example, 64R2P3 from Unit 7 is part of a plagioclase-olivine
glomerocryst and has a sieve textured core (Fig. 2.5c). Olivine was entirely re-
placed with clay minerals and was identified by its morphology. Melt inclusions
50

Figure 2.6. Major and trace element data were collected along a traverse
from interior to rim (i.e., exterior). Locations LA-ICP-MS analysis sites are
illustrated (labeled LA) along the measured core to rim An profile of each
plagioclase crystal. A) Interior section of xenolith crystal ML-X1, Malaita;
B) Xenolith crystal 59R2X1, ODP Site 1183 Unit 5B (192-1183A-59R2
112-117 cm piece #10A); C) Xenolith crystal 63R2X1, ODP Site 1183 Unit 6
(192-1183A-63R2 25-27 cm piece #4); D) Xenolith crystal 64R2X1, ODP
Site 1183 Unit 7 (192-1183A-64R2 116-120 cm piece #9B); E) Xenolith
crystal 807-X1, ODP Site 807 (130-807C-93R-1-137-139).
51

are more abundant in phenocrysts from Unit 7 relative to Units 5B and 6 (e.g.,
64R2P1; Fig. 2.5d). Unit 7 basalts also contain plagioclase-clinopyroxene glome-
rocrysts and partially resorbed strained clinopyroxene xenocrysts [91].
2.5.1.4 Site 807 Xenolith (807-X1)
Plagioclase xenoliths, although smaller and commonly in the form of glom-
erocrysts, were noted in Kwaimbaita basalts from ODP Site 807 [93, 94]. The
subround xenolith I examined from the Site 807 (807-X1) has a similar morphol-
ogy to the larger xenoliths from Site 1183 and Malaita but is more disaggregated
(Fig. 2.3e). Xenolith 807-X1 is composed primarily of interlocking plagioclase
52

grains and contains small clinopyroxene and completely altered olivine grains in
similar proportions along its margins (130-807C-93R-1-137-139; Unit 4G; 1525.8
mbsf; [82]; Figs. 2.3e,f). Major and minor element data only are reported in Ta-
ble 2.1 for selected points sampled along a transect from the core area to the rim
of a single large xenolith crystal (Fig. 2.6e).
53

54
Crystal Zone 5
ML-X1 xenolith
TABLE 2.2
CALCULATED PARTITION COEFFICIENTS AND EQUILIBRIUM
LIQUID COMPOSITIONS
An D Sr D Ti D Y D Ba D La D Ce D Nd D Eu D Mg D Fe Sr Ti Y Ba La Ce Nd Eu Mg# 6 Ba/Sr La/Y Sr/Ti
Interior 85 1.26 0.04 0.03 0.1 0.17 0.11 0.12 0.43 0.047 0.065 136.8 3331 8 22.08 1.41 5.52 - 0.58 0.3 0.16 0.18 0.04
Interior 85 1.24 0.04 0.03 0.1 0.17 0.1 0.11 0.43 0.047 0.065 130.5 1991 7.05 29.68 0.99 2.37 3.3 0.78 0.31 0.23 0.14 0.07
Interior 85 1.24 0.04 0.03 0.1 0.17 0.11 0.11 0.43 0.047 0.065 142.1 3664 7.51 18.91 1.19 3.49 3.39 0.81 0.32 0.13 0.16 0.04
Interior 86 1.23 0.04 0.03 0.1 0.17 0.1 0.11 0.43 0.047 0.065 152.3 2036 5.15 21.85 1.52 3.49 6.71 0.97 0.31 0.14 0.26 0.07
Interior 84 1.28 0.04 0.03 0.11 0.17 0.11 0.12 0.43 0.047 0.065 135 2147 - 27.01 2.32 3.4 3.29 1.1 0.33 0.2 - 0.06
Interior 86 1.21 0.04 0.02 0.1 0.17 0.1 0.11 0.43 0.047 0.065 153.6 4392 17.51 32.31 1.81 5.25 2.03 - 0.31 0.21 0.1 0.03
59R2-X1 xenolith
Exterior 66 1.99 0.06 0.06 0.24 0.2 0.14 0.16 0.43 0.047 0.065 - 2998 - - - - - - 0.18 - - -
Exterior 75 1.61 0.05 0.04 0.16 0.19 0.12 0.14 0.43 0.047 0.065 140.6 9313 4.76 30.42 2.33 5.26 3.6 1.13 0.32 0.22 0.49 0.02
Interior 81 1.39 0.04 0.03 0.13 0.18 0.11 0.12 0.43 0.047 0.065 - 5985 - - - - - - 0.3 - - -
continued...

55
TABLE 2.2
Continued
Crystal Zone An D Sr D Ti D Y D Ba D La D Ce D Nd D Eu D Mg D Fe Sr Ti Y Ba La Ce Nd Eu Mg# Ba/Sr La/Y Sr/Ti
Interior 82 1.35 0.04 0.03 0.12 0.17 0.11 0.12 0.43 0.047 0.065 148.2 7881 4.55 32.29 1.56 4.26 2.51 1.04 0.3 0.22 0.34 0.02
Interior 83 1.3 0.04 0.03 0.11 0.17 0.11 0.12 0.43 0.047 0.065 168.5 3953 6.58 41.83 1.29 5.2 4.38 1.05 0.29 0.25 0.2 0.04
Interior 82 1.33 0.04 0.03 0.12 0.17 0.11 0.12 0.43 0.047 0.065 152.6 2929 4.17 31.19 1.15 5.48 3.5 1.09 0.31 0.2 0.28 0.05
Center/Core 83 1.3 0.04 0.03 0.11 0.17 0.11 0.12 0.43 0.047 0.065 158.5 2461 6.79 45.33 1.79 5.36 5.42 0.79 0.29 0.29 0.26 0.06
59R2 phenocrysts P1-P4
P1 Rim 71 1.76 0.05 0.05 0.19 0.19 0.13 0.14 0.43 0.047 0.065 115 5117 6.45 31.46 1.81 5.44 5.38 0.89 0.3 0.27 0.28 0.02
P1 Core 74 1.63 0.05 0.04 0.17 0.19 0.12 0.14 0.43 0.047 0.065 114.8 6525 3.22 25.56 1.92 4.35 3.04 0.99 0.29 0.22 0.6 0.02
P2 Rim 70 1.79 0.05 0.05 0.2 0.19 0.13 0.15 0.43 0.047 0.065 115.4 8039 4.79 26.36 2.01 5.4 3.13 1.09 0.3 0.23 0.42 0.01
P2 Core 73 1.68 0.05 0.04 0.18 0.19 0.13 0.14 0.43 0.047 0.065 116.6 5962 8.2 25.85 2.42 4.14 3.35 0.9 0.33 0.22 0.3 0.02
P3 Center 69 1.84 0.05 0.05 0.21 0.2 0.13 0.15 0.43 0.047 0.065 - 3597 - - - - - - 0.3 - - -
P4 Center 67 1.97 0.06 0.05 0.24 0.2 0.14 0.16 0.43 0.047 0.065 - 2998 - - - - - - 0.31 - - -
continued...

56
TABLE 2.2
Continued
Crystal Zone An D Sr D Ti D Y D Ba D La D Ce D Nd D Eu D Mg D Fe Sr Ti Y Ba La Ce Nd Eu Mg# Ba/Sr La/Y Sr/Ti
63R2-X1 xenolith
Exterior 79 1.4 0.04 0.03 0.13 0.18 0.11 0.12 0.43 0.047 0.065 127.9 7685 8.64 24.02 1.14 3.83 4.14 0.52 0.29 0.19 0.13 0.02
Exterior 77 1.51 0.05 0.04 0.15 0.18 0.12 0.13 0.43 0.047 0.065 119.4 6671 6.28 29.48 0.71 2.47 1.83 0.65 0.31 0.25 0.11 0.02
Exterior 81 1.37 0.04 0.03 0.12 0.18 0.11 0.12 0.43 0.047 0.065 123.3 5807 11.18 27.88 1.29 2.94 0.33 0.5 0.28 0.23 0.11 0.02
Exterior 75 1.63 0.05 0.04 0.17 0.19 0.12 0.14 0.43 0.047 0.065 101.1 5033 5.39 20.65 1.2 2.91 2.47 0.62 0.32 0.2 0.24 0.02
Interm. 85 1.25 0.04 0.03 0.1 0.17 0.11 0.12 0.43 0.047 0.065 138 1435 7.41 25.45 1.08 3.24 2.58 0.61 0.27 0.18 0.15 0.1
Resorp. 81 1.38 0.04 0.03 0.12 0.18 0.11 0.12 0.43 0.047 0.065 124.9 6266 3.27 23.56 1.46 2.39 2.64 0.61 0.31 0.19 0.45 0.02
Interm. 85 1.25 0.04 0.03 0.1 0.17 0.11 0.12 0.43 0.047 0.065 142.3 9791 5.43 25.59 1 3.89 2.38 0.65 0.29 0.18 0.18 0.01
Interm. 85 1.25 0.04 0.03 0.1 0.17 0.11 0.12 0.43 0.047 0.065 137 - 6.79 27.84 0.78 3.15 1.45 0.9 0.27 0.2 0.12 -
Interm. 82 1.33 0.04 0.03 0.12 0.17 0.11 0.12 0.43 0.047 0.065 129.3 - 6.11 24.27 1.01 3.52 2.58 0.65 0.3 0.19 0.16 -
Interm. 83 1.29 0.04 0.03 0.11 0.17 0.11 0.12 0.43 0.047 0.065 132.2 3055 4.78 27.06 0.95 3.6 2.8 0.52 0.3 0.2 0.2 0.04
Interm. 84 1.27 0.04 0.03 0.11 0.17 0.11 0.12 0.43 0.047 0.065 133.5 4651 9.05 27.45 1.11 5.34 3.56 0.57 0.3 0.21 0.12 0.03
Interm. 85 1.24 0.04 0.03 0.1 0.17 0.11 0.11 0.43 0.047 0.065 137 4422 5.76 27.93 1.1 4.7 1.92 0.82 0.3 0.2 0.19 0.03
continued...

57
TABLE 2.2
Continued
Crystal Zone An D Sr D Ti D Y D Ba D La D Ce D Nd D Eu D Mg D Fe Sr Ti Y Ba La Ce Nd Eu Mg# Ba/Sr La/Y Sr/Ti
63R2 phenocrysts P1-P4
P1 Interm. 76 1.54 0.05 0.04 0.15 0.18 0.12 0.13 0.43 0.047 0.065 153.6 6571 6.69 56.62 2.36 5.97 1.31 0.57 0.3 0.37 0.35 0.02
P1 Interm. 74 1.65 0.05 0.04 0.17 0.19 0.12 0.14 0.43 0.047 0.065 105 5171 3.02 27.06 0.73 3.96 2.03 0.23 0.28 0.26 0.24 0.02
P2 Center 77 1.54 0.05 0.04 0.15 0.18 0.12 0.13 0.43 0.047 0.065 115.5 5876 6.09 32.24 1.68 4.59 0.67 0.75 0.29 0.28 0.28 0.02
P3 Rim 68 1.91 0.06 0.05 0.22 0.2 0.14 0.15 0.43 0.047 0.065 102 4513 5.58 26.69 1.55 5.47 3.67 0.7 0.36 0.26 0.28 0.02
P4 Core 75 1.62 0.05 0.04 0.17 0.19 0.12 0.14 0.43 0.047 0.065 139.8 5920 2.71 51.41 1.64 5.03 2.09 0.76 0.29 0.37 0.61 0.02
P4 Rim 67 1.96 0.06 0.05 0.23 0.2 0.14 0.15 0.43 0.047 0.065 82.7 4345 2.41 15.57 1.64 5.02 2.15 0.46 0.31 0.19 0.68 0.02
64R2-X1 xenolith
Rim 71 1.76 0.05 0.05 0.19 0.19 0.13 0.14 0.43 0.047 0.065 89.5 6605 2.96 14.05 1.16 2.82 3.09 0.83 0.23 0.16 0.39 0.0
1 Rim 79 1.46 0.04 0.03 0.14 0.18 0.12 0.13 0.43 0.047 0.065 115.9 757 3.75 22.06 1.18 3.47 1.91 0.91 0.32 0.19 0.31 0.15
Interior 80 1.51 0.04 0.03 0.13 0.18 0.11 0.12 0.43 0.047 0.065 117.4 4832 6.43 22.41 1.23 3.23 4.23 0.87 0.3 0.19 0.19 0.02
continued...

58
TABLE 2.2
Continued
Crystal Zone An D Sr D Ti D Y D Ba D La D Ce D Nd D Eu D Mg D Fe Sr Ti Y Ba La Ce Nd Eu Mg# Ba/Sr La/Y Sr/Ti
Interior 84 1.29 0.04 0.03 0.11 0.17 0.11 0.12 0.43 0.047 0.065 141.2 3288 7.72 28.37 0.92 3.93 0.68 0.96 0.28 0.2 0.12 0.04
Interior 82 1.35 0.04 0.03 0.12 0.17 0.11 0.12 0.43 0.047 0.065 124.9 5056 6.62 22.4 1.41 3.26 2.31 0.73 0.31 0.18 0.21 0.02
Interior 82 1.34 0.04 0.03 0.12 0.17 0.11 0.12 0.43 0.047 0.065 120.2 5862 6.9 27.39 0.92 3.64 1.57 0.48 0.29 0.23 0.13 0.02
Interior 82 1.33 0.04 0.03 0.12 0.17 0.11 0.12 0.43 0.047 0.065 121.6 3332 6.61 21.33 1.03 3.47 3.21 1.03 0.34 0.18 0.16 0.04
Interior 82 1.35 0.04 0.03 0.12 0.17 0.11 0.12 0.43 0.047 0.065 124.7 3900 6.41 23.1 0.58 3.64 2.63 0.94 0.33 0.19 0.09 0.03
Interior 83 1.32 0.04 0.03 0.12 0.17 0.11 0.12 0.43 0.047 0.065 126.2 - 10.22 22 0.83 3.43 1.76 0.81 0.31 0.17 0.08 -
Interior 80 1.43 0.04 0.03 0.13 0.18 0.11 0.13 0.43 0.047 0.065 109.2 - 4.83 18.55 1.15 2.6 3.88 0.83 0.34 0.17 0.24 -
64R2 phenocrysts P1-P4
P1 Core 69 1.87 0.06 0.05 0.21 0.2 0.13 0.15 0.43 0.047 0.065 100 4378 3.35 18.91 0.9 2.99 2.05 0.91 0.31 0.19 0.27 0.02
P1 Interm. 78 1.49 0.05 0.03 0.14 0.18 0.12 0.13 0.43 0.047 0.065 115.9 4822 3.73 20.18 0.57 2.8 1.58 0.41 0.32 0.17 0.15 0.02
P1 Interm. 74 1.66 0.04 0.04 0.17 0.19 0.13 0.14 0.43 0.047 0.065 102.6 6897 2.87 31.09 0.79 3.24 1.19 0.59 0.32 0.3 0.29 0.01
P2 Core 83 1.3 0.04 0.03 0.11 0.17 0.11 0.12 0.43 0.047 0.065 150.1 4125 17.18 44.03 2.5 4.87 4.03 0.99 0.26 0.29 0.15 0.04
continued...

59
TABLE 2.2
Continued
Crystal Zone An D Sr D Ti D Y D Ba D La D Ce D Nd D Eu D Mg D Fe Sr Ti Y Ba La Ce Nd Eu Mg# Ba/Sr La/Y Sr/Ti
P2 Rim 72 1.7 0.05 0.04 0.18 0.19 0.13 0.14 0.43 0.047 0.065 - 5216 - - - - - - 0.31 - - -
P4 Interm. 79 1.44 0.04 0.03 0.13 0.18 0.12 0.13 0.43 0.047 0.065 116.5 8180 11.37 19.58 1 6.04 6.75 0.62 0.28 0.17 0.09 0.01
P4 Interm. 74 1.66 0.05 0.04 0.17 0.19 0.13 0.14 0.43 0.047 0.065 110.2 7640 9.09 27.82 1.68 5.77 3.2 0.5 0.31 0.25 0.18 0.01
P3 Core 78 1.47 0.04 0.03 0.14 0.18 0.12 0.13 0.43 0.047 0.065 124.7 1576 7.51 23.57 0.5 3.89 1.1 0.31 0.31 0.19 0.07 0.08
P3 Interm. 75 1.58 0.05 0.04 0.16 0.18 0.12 0.13 0.43 0.047 0.065 114.2 6477 7.05 30.56 2.07 3.62 2.11 0.8 0.3 0.27 0.29 0.02
807-X1 xenolith
Core 83 - - - - - - - - 0.047 0.065 - - - - - - - - 0.26 - - -
Core 84 - - - - - - - - 0.047 0.065 - - - - - - - - 0.27 - - -
Core 84 - - - - - - - - 0.047 0.065 - - - - - - - - 0.26 - - -
Interm. 83 - - - - - - - - 0.047 0.065 - - - - - - - - 0.25 - - -
Interm. 83 - - - - - - - - 0.047 0.065 - - - - - - - - 0.29 - - -
Interm. 83 - - - - - - - - 0.047 0.065 - - - - - - - - 0.28 - - -
continued...

60
TABLE 2.2
Continued
Crystal Zone An D Sr D Ti D Y D Ba D La D Ce D Nd D Eu D Mg D Fe Sr Ti Y Ba La Ce Nd Eu Mg# Ba/Sr La/Y Sr/Ti
Rim 74 - - - - - - - - 0.047 0.065 - - - - - - - - 0.3 - - -
Rim 71 - - - - - - - - 0.047 0.065 - - - - - - - - 0.3 - - -
Rim 74 - - - - - - - - 0.047 0.065 - - - - - - - - 0.29 - - -
5 Interm. = Intermediate area between core and rim of largest discernible plagioclase crystal in each xenolith
6 Mg# = [Mg/(Mg+Fe)]

2.5.2 Measured compositions
2.5.2.1 Major Elements
The interior sections of xenolith crystals have greater An content (An80-An86)
than the xenolith exteriors and phenocrysts, which display a similar range of
An content (An65-An83) (see Table 2.1 and Figs. 2.7,2.8). Phenocrysts from Site
1183 Units 5B, 6, and 7 trend to lower An (An65−73) content than phenocrysts
reported by Sano and Yamashita [123] (An73−78) (Table 2.1). The phenocrysts
contain FeOT, MgO, and Ti within the ranges exhibited by xenolith interiors, but
do however extend to higher abundances of these elements (see Table 2.1). In
Table 2.1 I display representative major and trace element data of both xenolith
crystals and phenocrysts. Note that Ti was measured by EPMA. A spread sheet
of all major element data is available in an online archive.
2.5.2.2 Trace Elements
Selected trace element abundances measured in xenolith crystals and phe-
nocrysts are plotted against mole percent An of plagioclase in Figure 2.7 for sam-
ples from Malaita, Site 1183, and Site 807. With the exception of Eu, a general
negative correlation of element concentration with An content is apparent, which is
a likely result of the greater ease of trace element partitioning into the more elastic
structure of the relatively albite-rich plagioclase (e.g., [10]; Fig. 2.7). Phenocrysts
contain concentrations of La, Ce, and Nd that overlap the ranges exhibited by
xenolith interiors, but extend to higher abundances (see Table 2.1). Xenolith
crystals and phenocrysts contain similar Y and Eu abundances (Fig. 2.7).
61

Figure 2.7. Measured major and trace element data versus An content of
xenolith plagioclase crystals and plagioclase phenocrysts from: A) ML-X1
(Malaita) and 807-X1(130-807C-93R-1-137-139); B) ODP Site 1183 Unit 5B
xenolith 59R2X1 (192-1183A-59R2 112-117 cm piece #10A); C) ODP Site
1183 Unit 6 xenolith 63R2X1 (192-1183A-63R2 25-27 cm piece #4); D) ODP
Site 1183 Unit 7 xenolith 64R2X1 (192-1183A-64R2 116-120 cm piece #9B).
Resorption features examined in the Unit 6 and Unit 7 xenolith crystals are
noted in columns C and D. Average analytical uncertainties (analytical
precision relative 1σ) calculated for each segment of signal integrated for
each element are displayed as error bars in column A. Ti and FeOT measured
by EPMA, all others measured by LA-ICP-MS.
62

Figure 2.8. Measured MgO (wt %) abundance versus An content of all
xenolith plagioclase crystals and plagioclase phenocrysts examined in this
study. All data were obtained by EPMA. Magnesium partitioning is weakly
affected or unaffected by plagioclase composition e.g., [7]. I suggest low
pressure (depth range shallower than 7 km) crystallization of relatively
primitive OJP magmas can generate low MgO, An-rich plagioclase.
2.5.3 Parental magma compositions
2.5.3.1 Major Elements
Parent magmas of the An-rich xenolith interiors tend to have lower Ti at a
given Mg number (Mg/[Mg+Fe]) than phenocryst parent magmas. On a plot of
Mg number vs. Ti, phenocryst parent magmas overlap the Kwaimbaita whole-rock
basalt field, whereas the An-rich xenolith interiors trend below the fields defined
by OJP basalts (Fig. 2.9a; Table 2.3).
63

2.5.3.2 Trace Elements
It is difficult to accurately quantify the amount of error introduced by chosen
or calculated partition coefficients. Error bars displayed in figures 2.9 and 2.10,
which can be compared to analytical uncertainty displayed in figure 2.7, repre-
sent the combined potential error related to analytical uncertainty and calculation
of partition coefficients using uncertainty on constants reported by Bindeman et
al. [7]. Parent magmas of xenolith interiors typically exhibit lower La/Y and
Ba/Sr ratios and slightly higher Sr/Ti ratios than phenocryst parent magmas
(Tables 2.2and2.3; Fig. 2.9). Parent magmas of the An-rich xenolith interiors are
also enriched in Sr, Eu, and Y and relatively depleted in Ba, La, Ce, and Nd rela-
tive to phenocryst parent magmas (Tables 2.2, 2.3; Fig. 2.10). Parent magmas of
both xenolith crystals and phenocrysts have Sr, Ba, and Eu abundances that over-
lap published Kwaimbaita and Kroenke basalt whole-rock values, but generally
have lower La, Ce, and Nd abundances (see Table 2.3). Given the potential error
in calculated D values (compare error bars in Figs. 2.7 and 2.10), OJP plagioclase
parent magmas plot in a reasonably consistent manner in terms of La/Y, Sr/Ti,
and Ba/Sr ratios, Mg number, Eu and Sr, compared to the Kroenke, Kwaim-
baita, and Singgalo whole-rock basalt fields (Fig. 2.9; Table 2.3). For example,
the compositions of parent magmas of xenolith interiors from Malaita and Unit
7 of Site 1183 overlap Kroenke and Kwaimbaita basalt fields (see Fig. 2.9). Par-
ent magma compositions of xenolith exteriors and phenocrysts from all three Site
1183 Units and xenolith interior parent magmas from Units 5B and 6 consistently
trend outside of the fields defined by the whole-rock compositions (Fig. 2.9). Par-
ent magmas of the core and rim of phenocryst 63R2P4, the core of 59R2P1, the
rim of 59R2P2, and an exterior zone of xenolith 59R2X-1 are unique, because they
64

have much higher La/Y ratios than whole-rock data or parent magmas of all other
xenolith and phenocryst zones examined (see Figs. 2.9, 2.10). A zone within the
Unit 6 xenolith associated with a resorption feature has a similarly evolved parent
magma and is described in the section that follows.
65

Figure 2.9. Trace element ratio plots for parent magmas of OJP plagioclase
phenocrysts and xenolith plagioclase crystals. Error bars represent the total
potential error obtained by combining the uncertainty of LA-ICP-MS
measurements and that inherent in calculating the partition coefficients (D)
using uncertainties reported in Bindeman et al. [7] using standard error
propagation. Fields defined by published OJP whole-rock basalt data are
displayed [46, 133]. A) Ti vs Mg # [Mg + (Mg+Fe)] data show no distinct
differences between xenolith crystals and phenocrysts, except that An-rich
crystals tend to have lower Ti and trend to greater Mg#; B) Sr/Ti vs. La/Y:
note phenocryst parent magmas extend to higher La/Y ratios than higher
An xenolith crystals; C) La/Y vs. Ba/Sr: vectors defined by clinopyroxene
and plagioclase crystallization of relatively primitive Kroenke basalt are
shown as labeled arrows. For illustrative purposes, the head of the arrow on
the clinopyroxene vector corresponds to a crystallinity of ∼70%; D) La/Y vs.
Eu and vectors defined by clinopyroxene and plagioclase crystallization of
relatively primitive Kroenke basalt. Again, the head of the arrow on the
clinopyroxene vector corresponds to a crystallinity of ∼70%. Parent magmas
of xenolith crystals and phenocrysts from Site 1183 Unit 7 extend to lower
Eu and La/Y.
66

67
Sample An
(mol%)
Ranges in Published Whole-Rock and Glass Data
Kroenke - 8.9-
10.1
Kwaimbaita - 8.6-
12.5
Singgalo - 10.7-
13.5
TABLE 2.3
RANGES OF TRACE ELEMENTS IN PLAGIOCLASE,
WHOLE-ROCK DATA, AND INFERRED PARENT MAGMAS
FeO T MgO Ti Sr Y Ba La Ce Nd Eu Mg# Ba/Sr La/Y Sr/Ti
7.1-
13.6
5.3-
10.7
Ranges Measured in OJP Plagioclase Crystals
Xenolith Int. 80-86 0.52-
0.89
Xenolith Ext. 66-81 0.60-
1.01
Phenocrysts 65-83 0.50-
1.04
3410-
5790
4760-
10198
6.3-8.1 6775-
10611
0.22-
0.29
0.21-
0.32
0.22-
0.53
70.7-
194.3
33.4-
939.3
116.4-
207.1
0-221 155.9-
218.7
0-266 157.1-
225.7
41-441 161.7-
236.4
Ranges in Parent Magmas Calculated from OJP Plagioclase Crystals
Xenolith Int. - 8.0-
13.7
Xenolith Ext. - 9.2-
15.5
Phenocrysts - 7.7-
16.0
1 FeOT = Fe reported as total FeO.
4.7-6.1 0-9790 109.2-
168.5
4.4-6.8 0-9313 89.5-
140.6
4.7-
11.3
1576-
8180
2 Major elements in wt.% and trace elements including Ti in ppm.
82.7-
153.6
16.1-
18.8
19.9-
37.1
18.3-
31.8
0.10-
0.43
0.13-
0.34
0.11-
0.48
3.27-
17.51
2.96-
11.18
2.41-
17.8
4.65-
25.63
6.97-
36.27
15.6-
133.0
1.93-
5.05
2.70-
4.97
2.63-
8.57
18.55-
45.33
14.05-
30.42
15.57-
56.62
1.84-
2.48
2.44-
5.19
4.06-
6.76
0.10-
0.39
0.13-
0.43
5.19-
5.98
6.66-
13.69
3 Kwaimbaita, Kroenke, and Singgalo whole-rock and glass major and trace element abundances taken from [133] and [46].
4 Mg# = [Mg/(Mg+total Fe)].
0.09-
0.46
0.58-
2.32
0.71-
2.33
0.50-
2.50
11.0-
19.6
0.25-
0.63
0.29-
0.65
0.28-
0.74
2.37-
5.53
2.47-
5.26
2.19-
6.04
4.31-
5.23
5.58-
10.98
8.14-
14.10
0.08-
0.76
0.05-
0.51
0.09-
0.85
0.68-
6.71
0.33-
4.14
0.60-
6.75
0.59-
0.69
0.73-
1.34
1.02-
1.65
0.21-
0.47
0.22-
0.49
0.10-
0.47
0.48-
1.10
0.50-
1.13
0.36-
0.53
0.29-
0.44
0.27-
0.36
0.05-
0.29
0.03-
0.39
0.11-
1.14
0.11-
0.15
0.12-
0.19
0.18-
0.29
- - - -
- - - -
- - - -
0.25-
0.34
0.23-
0.32
0-1.09 0.21-
0.38
0.13-
0.29
0.16-
0.25
0.17-
0.37
0.02-
0.04
0.01-
0.15
0.01-
0.02
0-0.53 0.01-
0.10
0.11-
0.49
0.07-
0.68
0.01-
0.15
0.01-
0.08

2.5.4 Resorption Features in Site 1183 xenolith Crystals
A resorption feature in the Unit 6 xenolith crystal 63R2-X1, identified by tex-
ture, included a decrease in An content across the resorption surface towards the
rim (i.e., the direction of crystal growth) from An85 to An81 (see Figs. 2.2e, 2.9, 2.10).
There is no significant change in Mg number of parent magmas across this resorp-
tion surface, but there is a decrease in parent magma Ti content. The parent
magma of the An81 zone, which was the zone grown following resorption, had a
higher La/Y and Sr/Ti, although it had less Sr, Ti, and Y, and roughly the same
Ba/Sr ratio as the parent magma of the An85 zone (see Figs. 2.9, 2.10). A re-
sorption feature in the Unit 7 xenolith crystal 64R2-X1, also identified by texture,
included a slight increase in An content across the surface from An80 to An83 (see
Fig. 2.2b). The lower An parent magma (core region) in 64R2X-1 had a higher
La/Y and Sr/Ti, and again less Sr, Ti, and Y, with a similar Ba/Sr ratio and
Mg# as the An83 parent magma. The An83 zone was grown following resorption
(see Fig. 2.2b).
2.6 Discussion
2.6.1 Inferred Chemical Magma Evolution
2.6.1.1 Major Elements
In terms of An content, the cores of Site 1183 cumulate xenolith plagioclase
crystals are generally distinct from the phenocrysts in the host basalt. Sano and
Yamashita [123] suggested plagioclase in equilibrium with Kwaimbaita magmas
should be An73−78. I observed phenocryst zones as low as An65 and as high as An83,
whereas xenolith interiors were An80-An86 (Tables 2.2, 2.3; Figs. 2.9, 2.10). It
68

Figure 2.10. Calculated major and trace element abundances of magmas
that were parental to the studied plagioclase crystals plotted against the An
content of each crystal from: A) ML-X1 (Malaita) and 807-X1
(130-807C-93R-1-137-139); B) ODP Site 1183 Unit 5B xenolith 59R2X1
(192-1183A-59R2 112-117 cm piece #10A); C) ODP Site 1183 Unit 6
xenolith 63R2X1 (192-1183A-63R2 25-27 cm piece #4); D) ODP Site 1183
Unit 7 xenolith 64R2X1 (192-1183A-64R2 116-120 cm piece #9B).
Resorption features examined in the Unit 6 and Unit 7 xenolith crystals are
noted in columns C and D. Error bars for the y-axis represent the total
potential error obtained by combining the uncertainty of LA-ICP-MS
measurements and that inherent in calculating the partition coefficients (D)
using uncertainties reported in Bindeman et al. [7] using standard error
propagation. Error bars for the x-axis represent just analytical uncertainty.
69

should, however, be noted that in their experimental studies Sano and Yamashita
[123] observed crystallization of An82 plagioclase from Kroenke basalts (parental
to Kwaimbaita basalts) at higher temperatures (∼1200 ◦ C) and lower pressure (0.1
MPa). Although the Mg# of a residual magma decreases with increased olivine
fractionation, I see no evidence of olivine-dominated fractionation in xenolith or
phenocryst parent magmas (i.e., no systematic difference of parent magma Mg#
between the xenolith crystals and phenocrysts). The ranges of Mg# for parent
magmas of both xenolith crystals and phenocrysts overlap the range observed for
Kwaimbaita basalts (Fig. 2.9a; Table 2.3). Parent magmas of An65−79 phenocryst
zones from the three Site 1183 basalt units are compositionally similar with regard
to Mg number and Ti abundance (e.g., Fig. 2.9a). This first order observation
suggests they share a common parent magma, although some lower An phenocryst
zones (i.e., 63R2P4, 59R2P1, and 59R2P2) appear to have more evolved parent
magmas (Fig. 2.9). However, when individual units from Site 1183 are considered,
the Mg# of the parent magmas changes little with quite a large range in An
content (Fig. 2.10b-d), suggesting that the Mg# is buffered.
Parent magmas of the An-rich xenolith crystals and phenocrysts extend to
more primitive compositions than the Kwaimbaita whole-rock composition (Ta-
ble 2.2; Fig. 2.9a). This is contrary to the interpretation of Sano and Yamashita
[123] that An-rich zones grew in a relatively evolved and H2O-rich boundary layer.
It is thus necessary to elucidate other factors, in addition to growth in a water-rich
boundary layer, which favor crystallization of An-rich plagioclase.
70

TABLE 2.4
MELTS CRYSTALLIZATION MODELING OF AVERAGE OJP
BASALTIC GLASSES
Pressure (kbar) 0.1 0.7 0.9
Average Site 1183 Kwaimbaita basaltic glass
Temperature( ◦ C) 1184 1188 1198
An content 79 79 77
Average Site 1185 Kroenke basaltic glass
Temperature( ◦ C) 1200 1204 1216
An content 84 83 82
Glass 192-1185B-16R1, 64
Temperature( ◦ C) 1217 1222 1229
An content 85 84 84
Average Site 1187 Kroenke basaltic glass
Temperature( ◦ C) 1200 1206 1218
An content 84 83 82
1 Glass compositions are from [121]
2 Temperature at which plagioclase starts to crystallize
3 Anorthite content of the first-formed plagioclase
71

2.6.1.2 An-rich OJP Plagioclase: Influence of Temperature and Pressure
Experimental results have demonstrated that relatively albite-rich plagioclase
crystallizes at elevated pressure, and growth of An-rich plagioclase is favored in
high temperature magmas at lower pressure e.g., [3, 61, 123, 137]. I used the
MELTS program ([2, 55]) to test the likelihood of equilibrium crystallization of
An-rich (i.e., >An82) plagioclase at low pressure from relatively hot and primitive
OJP magmas. Major element, H2O, and other volatile data reported by Roberge
et al. [121] for unaltered Kwaimbaita and Kroenke basaltic glasses from ODP
Leg 192 were used as starting compositions for equilibrium crystallization mod-
eling using MELTS (see Table 2.4). Equilibrium crystallization modeling, which
was performed to estimate initial crystallization temperature and An content of
plagioclase, was run at 0.1 kbar, 0.7 kbar, and 1.9 kbar to also explore the influ-
ence of pressure on the composition of crystallizing plagioclase. These pressure
estimates adhere to previous models that indicate OJP basalts experienced par-
tial crystallization in the shallow crust (< 6-8 km depth) [46, 103, 121, 123]).
Results of MELTS modeling demonstrate average Kroenke basaltic glasses from
both Site 1187 and 1185, which are considered to be parental to Kwaimbaita
basalt e.g., [46, 123], crystallize An84 plagioclase at 1200 ◦ C and 0.1 kbar pressure
(see modeling results in Table 2.4). I also modeled crystallization of a single more
primitive Kroenke glass with greater MgO, CaO, Al2O3 (sample 1185B-15R-1
149 of Roberge et al., 2004), which crystallized An85 plagioclase at 1217 ◦ C and
0.1 kbar (Table 2.4). There indeed may be more primitive melts, currently un-
sampled, that would produce slightly An-richer initial plagioclase (i.e., to account
for An86 plagioclase from Malaita xenolith ML-X1). The results of the MELTS
numerical simulations suggest equilibrium crystallization a low pressure (i.e., 1
72

kbar), high temperature (>1200 ◦ C) crystallization environment favors An-rich
plagioclase growth from known compositions of magmas parental to Kwaimbaita
basalt.
2.6.1.3 An-rich OJP Plagioclase: Influence of Water
Sano and Yamashita [123] invoked a role for water in the genesis of low-MgO,
An-rich OJP plagioclase crystals, which they envisaged to form in a water-rich
(and presumably incompatible trace element-rich) crystal mush layer. Their in-
terpretation is consistent with the hydrous experimental data of Sisson and Grove
[126]. However, while the plagioclase observed by Sisson and Grove [126] grown
in H2O saturated parent magmas were low-MgO and An-rich, the crystals they
observed were more An-rich (generally An90) and lower MgO (< 0.2 wt.% MgO)
than OJP plagioclase crystals (< An90, and > 0.2 wt.% MgO) reported in this
study or by Sano and Yamashita [123] (Table 2.1). Plagioclase data from the
anhydrous experiments of Grove et al. [61] and Bartels et al. [3] produced a neg-
ative correlation of MgO with An content, which was also documented for OJP
plagioclase xenolith crystals and phenocrysts ( [123]; this study; Fig. 2.8). The
experiments of Bartels et al. (1991) were run at 10-20 kbar, and the experiments
of Grove et al. (1982) run at 10-3 kbar. The lower pressure experiments of Grove
et al. [61] produced crystals with MgO and An contents similar to natural OJP
plagioclase observed in this study and by Sano and Yamashita [123]. While I can-
not rule out the influence of H2O, MELTS modeling and previous experimental
studies provide evidence that crystallization of low MgO, An-rich OJP plagioclase
from low pressure, high temperature magmas can account for my observations
without a significant role for H2O. It is, however, impossible to determine un-
73

equivocally the roles of H2O, pressure, and temperature on the genesis of low
MgO, An-rich OJP plagioclase based solely upon major element data. Plagioclase
parent magma trace element compositions allow us to further test the contrasting
hypotheses that An-rich crystals grew from relatively primitive magmas versus
evolved H2O-rich magmas.
2.6.1.4 Trace Elements
The relatively low incompatible trace element abundances in the calculated
parental magmas, relative to the whole-rock basalt compositions, indicates that
crystal fractionation had not occurred to any significant degree and, thus, cannot
be invoked as a process to elevate the water content of the magma. Therefore,
if water-rich magmas were responsible for An-rich plagioclase crystallization, the
magma(s) would need to have been derived from a relatively water-rich source.
However, the data from Roberge et al. [121] indicate this was not the case. As
an alternative to growth in an H2O-rich magma, I suggest An-rich plagioclase,
whether xenolith or phenocryst, crystallized from relatively primitive magmas
in the shallow portions of the OJP magma chamber system. This interpreta-
tion reconciles the relatively primitive parent magma compositions (i.e., lower
incompatible element abundances at a given Mg#; see Fig. 2.9a) of the An-rich
portions of OJP xenolith crystals and phenocrysts. Extensive plagioclase crys-
tallization depletes a residual melt of Sr, as DSr in plagioclase is generally > 1.2
(see Table 2.2). In an environment where plagioclase is the dominant crystallizing
phase, early formed crystals will contain higher Sr/Ti and lower Ba/Sr ratios,
since DBa and DTi < 1 in plagioclase. In a confined, nonturbulent environment
components rejected by growing plagioclase crystals spawn local oversaturations
74

of components needed for later stage clinopyroxene crystallization e.g., [98]. As
clinopyroxene crystallizes the incompatibility of La in clinopyroxene relative to
Y results in relative depletion of Y and enrichment of La, which will yield a
higher La/Y residual liquid (e.g., Table 2.2; Fig. 2.9b-c). Since the trace elements
considered here are highly incompatible in olivine, crystallization of this phase
yields minimal change in the ratios of any of these elements c.f., [6, 7]. Trace
element enrichments and depletions not encountered in whole-rock basalt data
include: 1) An80−86 xenolith and phenocryst zones that were derived from parent
magmas with greater Sr, Y, and Eu, lower La/Y and Ba/Sr, and lower Ba and
LREE abundances extending to overall more primitive compositions than whole-
rock data (Fig. 2.9; Table 2.2). These compositional characteristics suggest that
parent magmas of the An-rich zones experienced less clinopyroxene and plagio-
clase crystallization than lower An zones (Fig. 2.9). The resorption feature in
the Unit 7 xenolith 64R2X-1 provides textural and chemical evidence of exposure
to more primitive magma, where the parent magma of the zone grown after re-
sorption is more An-rich, has lower La/Y, and has greater Sr and Y abundances
(Figs. 2.2b, 2.9, 2.10), and 2) An65−79 phenocryst and xenolith exterior zones with
lower Sr, Y, and Eu, higher La/Y and Ba/Sr, and higher Ba and LREE abun-
dance parental magma compositions that extend to more evolved compositions
than whole-rock data (Figs. 2.9, 2.10). Included in this group are a small number
of phenocrysts, a xenolith exterior zone, and a single resorption feature in the
Unit 6 xenolith that indicate crystallization from magma(s) that had experienced
relatively extreme plagioclase and clinopyroxene fractionation. In Figures 2.9a
and 2.9c vectors for clinopyroxene and plagioclase crystallization are shown as
arrows. Generation of the highest La/Y ratio I observed in a plagioclase par-
75

ent magma requires significant clinopyroxene fractionation (Fig. 2.9). These high
La/Y, relatively evolved liquids, were parental to portions of both xenolith crys-
tals (i.e., 63R2X-1 resorption feature) and phenocrysts. In figure 2.9c the head
of the arrow defining the clinopyroxene crystallization vector corresponds to ap-
proximately 70% clinopyroxene crystallization in terms of La/Y for the highest
MgO Site 1187 Kroenke basalt reported by Fitton and Godard [46]. However, the
amount of crystallization required to account for trace element variations may be
much less given the combined error of calculated partition coefficients (which is
larger for Y than La, Ba, Sr, or Ti), LA-ICP-MS analyses, crystallization tem-
perature, pressure, magmatic H2O, and uncertainties regarding bulk DY and DLa
for clinopyroxene (see error bars in Figs. 2.9, 2.10 and the captions Figs. 2.9,
and 2.10 for details on how the error bars were estimated). The observed major
and trace element variations require magmas both more and less evolved than
those represented by the whole-rock/groundmass. The evolution of these liquids
involved crystallization that was dominated early by plagioclase and olivine (as
suggested by whole-rock and basaltic glass studies) (e.g., Fig. 2.9), although as
noted above, evidence of olivine fractionation is difficult to identify from the pla-
gioclase trace element data presented here [103, 110, 121, 123, 131, 133]. Late
stage clinopyroxene crystallization also played a significant role in evolution of
OJP basaltic magmas, as indicated by variations in La/Y ratios between magma
parental to the high and low An zones. Crystallization of plagioclase from high
temperature, relatively primitive anhydrous magmas at low pressure was an im-
portant process. Conditions to produce the spectrum of relatively evolved and
76

primitive magma compositions recorded in OJP plagioclase are not fully met in
a homogenous magma body and require more confined environments such as a
crystal mush layers e.g., [84, 98, 123].
2.6.2 OJP Magma Chambers, Mush Layers, and Solidification Fronts
Sano and Yamashita [123] suggested zoned plagioclase, clinopyroxene, and
olivine crystals in Kwaimbaita basalts formed as they moved between crystal
mush boundary layers (solidification fronts) and the main magma body. Bound-
ary mush layers are rheological elements of multiply saturated solidification fronts
defined as having crystallinities between 25% and 50%-55% [98]. Crystallization
within the mush zone produces interstitial melts that are evolved relative to the
magma in the chamber interior, and continued crystallization traps these evolved
melts within the solidification front [84, 98]. Mineral phases with compositions
out of equilibrium with main magma body grow within the mush zone of the so-
lidification front where they are thermally and mechanically insulated from the
hotter magma chamber interior [84]. Early formed crystals that potentially con-
tain chemical signatures of relatively primitive magmas may also exist in the rigid
crust of the solidification front near the solidus. Solidification fronts thicken over
time due to crystal nucleation and growth but may also thicken by capturing
crystals carried within new batches of magma. Capture of crystals by the advanc-
ing solidification front is greatest along the magma chamber floor due to crystal
settling, and the thickness of these cumulates in any given magma chamber may
grow to be quite large if repeated transit and storage of crystal carrying magmas
occurs from other magma chambers [98]. Crystal debris may be mobilized from
mushy cumulate layers and solidification fronts by erosion, density and convec-
77

tion driven plumes, or stirring by new magma input [83, 98]. In the case of the
OJP magma chamber system, magma in the chamber interior(s) was the most
eruptible, had Kwaimbaita or Kroenke basalt-like composition(s), and made up
the most significant compositional and textural fractions of basalts erupted to
the surface. Disruption of solidification fronts released evolved material (liquid +
crystals) to mix with the main body of magma e.g., [98]. Mixing small fractions of
evolved material from disrupted solidification fronts into the main magma body
can lead to changes in basalt chemistry (Marsh, 1996). However, if the geometry
of the OJP magma chamber system was such that inputs from disrupted solid-
ification fronts were small relative to the size of the magma body, the input of
more evolved material would have had little affect on the bulk magma chemistry.
Evidence for this process is thus best preserved by allochthonous crystals e.g.,
[37, 98], much like the An65−73 phenocrysts with evolved parent magma composi-
tions and the An80−86 xenolith crystals with primitive magma compositions, which
are all out of equilibrium with the host Kwaimbaita basalt. Parent magma trace
element compositions of An65−73 phenocryst zones, which trend to much more
evolved compositions than whole-rock data, suggest some of these crystals grew
deep within solidification fronts that were eroded and mixed into the main magma
body. Crystals with similarly evolved compositions may also form in dikes and
conduits within the magma chamber system that become readily congested. In
these regions geometry and magma supply rate allow solidification fronts to prop-
agate inward and meet filling the space with a crystal mush that has a network
of interstitial spaces filled with variably evolved melts [98]. Within the interstices
of the crystal mush significant melt differentiation can occur where the mush is
infrequently flushed with new (primitive) melt. I suggest the xenolith crystals and
78

phenocrysts with relatively evolved parent magmas originated from these types of
environments (e.g., phenocryst 63R2P4; phenocryst zones 59R2 P1 core) or were
exposed to liquids flushed from these types of environments (e.g., 59R2P2 rim;
xenolith zones 59R2X1 and 63R2X1). Anorthite-rich plagioclase crystals from the
cumulate xenoliths grew from relatively primitive magmas that ascended and par-
tially crystallized in shallow magma chambers. Contraction of the clinopyroxene
stability field at lower pressures, coupled with evidence of less clinopyroxene frac-
tionation in parent magmas of An-rich zones and results of MELTS simulations,
suggest that crystallization of OJP magmas at low pressure was dominated by
plagioclase ± olivine. Farnetani et al. [45] also suggested that shallow crystal-
lization of OJP magmas was largely plagioclase and olivine with lesser amounts
of late crystallizing clinopyroxene. This relationship is favored by the textures of
Site 1183 xenoliths, where partially resorbed clinopyroxene crystals were present in
voids within interlocking plagioclase crystals (e.g., Fig. 2.2c). Given the size of the
xenoliths (Figs. 2.2a, 2.3a) and their apparent low porosity, they are likely pieces
of the deepest and earliest formed portion of the solidification front. The large
size and general round nature of plagioclase cumulate xenoliths in Kwaimbaita-
type basalt qualitatively suggest that magma delivery and eruption processes were
quite vigorous. Vigorous input of magma into the chamber promotes heavy ero-
sion of solidification fronts and deep stirring of cumulates along the chamber
floor. I suggest solidification front disruption was a ubiquitous process over large
vertical and lateral scales, which freed both more evolved phenocrysts and less
evolved xenoliths to mix with Kwaimbaita-like magma in the chamber interior.
The volume fraction of material from disrupted solidification fronts generally was
probably small (i.e., as suggested by < 2% phenocrysts by volume; Sano and
79

Yamashita, [123], and thus did not significantly change bulk magma chemistry,
unless extensive resorption of crystals occurred.
A laterally and vertically extensive system of interconnected dikes and sills
with regions dominated by liquid and other regions dominated with a crystal-liquid
mush can account for the zoning and compositional features recorded in OJP xeno-
lith and phenocryst plagioclase crystals (see Fig. 2.11). Oscillatory zoned crystals
may have formed in a quiet environment where feeDBack between the rate of crys-
tal growth and diffusion in a boundary layer at the crystal-melt growth interface
yielded small and frequent variations of melt composition near the growing crystal
[1, 115, 141] (Figs. 2.2, 2.3, 2.5). Oscillatory zoned crystals may have also formed
where small and frequent variations in melt composition and/or temperature were
produced by convection of the main magma body at the inward edge of the solid-
ification front e.g., [60]. Indeed, where magma recharge is frequent, solidification
front growth can be arrested or reversed placing deeper zones of the front in re-
peated contact with fresh primitive melt (Marsh, 1996). I have noted evidence
that magma recharge was both vigorous and frequent to have generated the most
of the volume of the OJP around ∼122 Ma. In the case of a shallow OJP magma
chamber, frequent and vigorous recharge would favor production of oscillatory-
zoned An-rich plagioclase crystallizing from relatively primitive parental magmas,
whereas a stagnant environment would favor production of oscillatory zoned An-
poor plagioclase crystallizing from more evolved parental melts. Thick cumulate
mush layers along the floors of OJP magma chambers were also likely important in
the widespread genesis of Kwaimbaita basalt (e.g., Fig. 2.11). An important factor
in basalt differentiation in the crystal-mush dominated portions of the magmatic
system would have been the residence time of melts within the vertically and lat-
80

Figure 2.11. This is a simplified schematic of the Ontong Java Plateau
magma chamber system consisting of a series of interconnected dikes and sills
similar to the mush-column magma chamber model of Marsh [98] in the 0 to
7 km depth range. Magma ascending magma chamber system that follows
the path (dashed line) X-X is subject to more density filtration (i.e., in the
center of the plateau). Magma that ascends along the path (dashed line)
Y-Y is subject to less density filtration, hence a wide variety of magma types
are erupted along this path (i.e., along the plateau margins). A) Due to
geometry, these areas quickly become congested via crystallization during
period of low magma throughput. Interstitial fluids in these regions, within
the congested crystal-liquid mush, experience greater degrees of
fractionation; B) Solidification front material may dislodge from the roof and
cascade as a plume of liquid + crystals with components both in and out of
equilibrium with the main magma body. Some of this material may escape to
be carried to the surface; C) Liquid-dominated regions of the magma
chamber, which are also the hottest portions of the system. In the OJP
magma chamber system this was filled with Kwaimbaita and/or
Kroenke-type magmas; D) Examples of crystal-mush dominated zones, where
within the mush evolved liquids are present. Crystal phases out of
equilibrium with the main magma body are also present in these zones,
where their growth environment is thermally and mechanically insulated
from the hot chamber interior e.g., [98]. E) Thick solidification fronts along
chamber floors (i.e., cumulate layers). These crystal-mush cumulate layers
thicken with the repeated input of crystal-bearing magmas.
81

erally extensive crystal-mush network (Marsh, 1996; Fig. 2.11). Near steady-state
supply of magma into the system (e.g., [120]) would have resulted in tightly con-
strained basalt compositions consistent with Kwaimbaita-type basalts that make
up the bulk of the OJP. Given the size of the OJP, persistent melting over the
lifetime of the causative thermal disturbance is amenable to near steady-state melt
infiltration of the mush-column-type OJP magma chamber system as illustrated
in Figure 2.11 e.g., [98]. Shallow OJP magma evolution was thus influenced by a
balance of several processes that included: 1) Formation of evolved melts within
crystal-mush dominated regions; 2) Flushing of variably evolved melts from the
crystal mush and homogenization of these melts within the main magma body;
3) Entrainment of phenocrysts and other crystal debris (e.g., An-rich xenoliths)
during magma recharge; 4) Resorption (partial or complete) of crystal debris; 5)
Crystal settling; 6) Input of new, primitive magma (see Fig. 2.11). Partial crystal-
lization of magmas at depths of 0-7 km, within the range suggested for the OJP,
coincide with zones of neutral buoyancy as discussed by Ryan [122]. Studies have
suggested the OJP crust contains significant gabbro and/or dolerite intrusions
that would have corresponded to zones of neutral buoyancy and a large shallow
magma chamber system during active OJP volcanism e.g., [45, 53, 54, 73, 110].
Widespread gabbro and dolerite intrusions are consistent with my interpretation
that the OJP magma chamber system consisted of interconnected crystal-mush-
rich magma chambers. Ryan [122] suggested as crust is thickened by erupted
lavas, the zone of neutral buoyancy will slowly migrate upward. Slow upward
migration of the magma chamber system would favor slow yet pervasive assim-
ilation of overlying rock, potentially seawater altered basalts in the case of the
OJP. Indeed, Michael [103] and Roberge et al. [121] suggested assimilation of sea-
82

water altered basalt or materials with a seawater brine component as the source
of widespread Cl contamination of OJP basaltic glasses. Additionally, Neal and
Davidson [108] invoked seawater altered basalt as the contaminant to the parental
magma that produced megacrysts entrained in post-OJP alnite pipes on Malaita,
Solomon Islands. However, I note that the assimilation of such a seawater-altered
component did not radically elevate the water content of the magma within the
different chambers. Magma chambers are more sheet-like near the Earths surface
[97]. The maximum vertical extent of interconnected sheet-like chambers would
have been where the supply of intruding magma was the greatest (i.e., toward the
center of the plateau; Fig. 2.11). The total integrated thickness of the dike and
sill system was thus less along the margins of the OJP, which allowed magmas
distinct from Kwaimbaita-type to ascend through the mush column and erupt
(see Fig. 2.11). Indeed, the greatest compositional diversity is observed along the
plateau margins e.g., [46, 91, 93, 94]. Along the thinner margins of the plateau
the magma chamber system and overlying rock would have acted as less of a
density filter than the thicker central portion of the plateau (Fig. 2.11). I cal-
culated average OJP magma densities using the KWare Magma program written
by Dr. Ken Wohletz using the glass compositions reported by Michael [103] and
Roberge et al. [121] at a temperature of 1180 ◦ C and a pressure of 1 kbar (KWare
Magma program is available at http://geont1.lanl.gov/Wohletz/Magma.htm). It
would have been possible for more (e.g., Kroenke-type: 2869 ± 3 kg/m 3 ) and
less dense (Singgalo-type: 2860 kg/m 3 and Kwaimbaita-type: 2861± 3 kg/m 3 )
magmas to ascend the mush pile and reach the surface along the margins of the
plateau where there was less overlying volume of intruded and extruded magma.
Indeed, Kroenke basalt flows are found to be stratigraphically above Kwaimbaita
83

asalt at Site 1185 of ODP Leg 192 [91]. Although the magma chamber system
may have been thinner along the margins of the plateau, variations in magma res-
idence time in the system did not preclude formation and eruption of Kwaimbaita
magmas along the OJP margins. Differences, however slight, in the densities of
Kroenke, Kwaimbaita, and Singgalo magmas would have lead to different rates
and patterns of magma ascent, pooling, and crystallization en route to the surface.
2.7 Summary and Conclusions
Anorthite-rich (An80−86) portions of plagioclase crystals from cumulate xeno-
liths and phenocrysts in host basalts were formed primarily by crystallization in
shallow (low pressure, 2-7 km depth) regions of the OJP magma chamber system
from more primitive magmas than those basalts sampled from the OJP that were
low in H2O and at relatively high temperature (liquidus temperature near 1200 ◦ C).
Evidence from this study shows that the role of H2O-rich evolved boundary layer
interstitial melts in the formation of An-rich plagioclase was, at best, minor. Par-
ent liquid compositions derived from OJP xenolith and phenocryst plagioclase
crystals contain a wider compositional record of magma evolution than that re-
vealed by OJP whole-rock basalt data. Cumulate xenoliths and phenocrysts gen-
erally are pieces of disrupted solidification fronts. Solidification fronts would have
been ubiquitous throughout the OJP magma chamber system. The OJP magma
chamber system was composed of interconnected dikes and sills and consisted of
regions dominated by liquid (magma chamber interior) and regions dominated by
crystal-liquid mush (along magma chamber floors and walls, within dikes and con-
duits). Crystals and crystalline debris were extensively recycled throughout the
OJP magma chamber system. This recycling accounts for the wide compositional
84

spectrum of magmas parental to plagioclase xenolith crystals and phenocrysts
with An content out of equilibrium with their Kwaimbaita host basalts. Ironi-
cally, this process also accounts for the dominance of a single magma composition
erupted on to the OJP, as it represents the homogenized body of magma in the
main part of the upper chamber. The compositions of melts periodically flushed
from the interstices of the crystal-mush network were affected by their residence
times in the mush. Depending upon the extent of the crystal-rich portions of the
magma chamber system, residence time may have had a large effect on overall
basalt chemistry. The shallow crustal OJP magma chamber system corresponded
with a zone of neutral buoyancy in the 0-7 km depth range, as in other volcanic
systems [122]. As OJP volcanism progressed the plateau gained elevation, and the
zone of neutral buoyancy gradually migrated upward leading to slow yet pervasive
assimilation of overlying seawater altered basalt. The greatest diversity of OJP
basalts is along the margins of the plateau where it is thinner. During formation,
magmas ascending into the marginal and thinner magma chamber system around
the plateau margins were subject to less density filtration, which allowed ascent
and eruption of relatively dense Kroenke magmas and slightly less dense Singgalo
magmas (see Fig. 2.1 for magma types and distribution across the OJP). Slight
variations in paths of magma ascent, depths of magma pooling and crystallization,
as well as rates and amounts of assimilation of seawater altered basalt would have
affected the isotopic and incompatible trace element characteristics of known OJP
basalts. I suggest that these physical factors may account for both the homogene-
ity of basalts in the central portion of the OJP and the relative heterogeneity of
basalts around the plateau margin.
85

CHAPTER 3
SHALLOW MAGMA EVOLUTION DURING LATE CRETACEOUS
HAWAIIAN HOTSPOT-RIDGE INTERACTION: INSIGHTS FROM
INTEGRATION OF CRYSTAL SIZE DISTRIBUTIONS AND
3.1 Introduction
MICROANALYSIS OF PLAGIOCLASE
Detroit Seamount is located near the northern terminus of the Emperor Seamount
Chain and formed when the Hawaiian hot spot was adjacent to a mid-ocean ridge
during the Late Cretaceous (76-81 Ma) [34, 43] (Fig. 3.1). Interaction of hot
spots and mid-ocean ridges (MOR) often have profound effects on the surround-
ing lithosphere [47, 62–64, 79, 119]. Geochemically anomalous hot spot and MOR
basalts are common in areas of hot spot-MOR interaction and are often linked
to unique mantle processes resulting from the interaction [79, 119]. Several re-
cent studies have noted incompatible trace element and isotopically depleted hot
spot basalts from Detroit Seamount (e.g., [72, 79, 119]). Detroit Seamount basalt
compositions extend from N-type MORB-like to compositions intermediate be-
tween N-MORB and young Hawaiian tholeiitic basalts [72, 79, 119]. Frey et al.
[50] presented compositional evidence linking DSM basalts to the Hawaiian hot
spot and concluded that it is unlikely DSM basalts are simply MORB. Several
hypotheses have been put forward to explain the petrogenesis of depleted hot
86

spot basalts at DSM. Keller et al. [79] suggested that the rising Hawaiian plume
entrained MORB-source upper mantle to an extent that partial melting of the
MORB source component overwhelmed the hot spot source signature in DSM
magmas. Regelous et al. [119] favored an alternative model, where a depleted
component inherent to the Hawaiian plume was more apparent in DSM magmas
due to greater partial melting under a thinner lid of lithosphere near the spreading
center. Huang et al. [72] noted the significance of fractionation and accumulation
of olivine and plagioclase, and thus an implied of signifigance low pressure partial
crystallization (e.g., [130] during the evolution of DSM magmas in the crust. If
one is then to relate DSM basalt chemistry to the nature of a mantle source region
or some mantle process, it is vital to understand the extent to which bulk DSM
magma compositions were influenced during their transit through the crust. Once
the extent of magma evolution in the shallow crust is constrained, understanding
the roles of mantle processes becomes less convoluted.
3.1.1 Magma Evolution in the Crustal Magma Chambers
The crust acts as a cool density filter to ascending hot mantle derived magmas
creating a tendency for rising magmas to stall [113]. Crustal magma chambers
are sites of extensive partial crystallization (fractional, equilibrium, or in-situ),
assimilation, and/or magma mixing, each of which are fundamental processes of
magma evolution. Over the lifetime of magmatism a magma chamber may be filled
with a variety of magma types including but not limited to more primitive, more
evolved, or end-member magmas involved in mixing. Crystals exchanged during
transfer, movement, or filling of magma or crystals grown after these events are
physical vessels that carry compositional information about otherwise inaccessible
87

magma compositions [23, 39, 98]. A sensible approach then for constraining the
processes of shallow magma evolution is to decipher the textural and compositional
record of these processes recorded in individual crystals e.g., [37].
In this work I employ an integrated textural and microanalytical approach to
constrain shallow magma evolution during the formation of Detroit Seamount by
1) identification of crystal populations with related physical growth histories and
2) understanding the provenance of these crystals. Textural and microanalyti-
cal studies focused on plagioclase have been used to elucidate basaltic to silicic
magma evolution processes (e.g., [5, 6, 16, 39, 135]). Plagioclase contains mea-
surable quantities of select incompatible trace elements, including the light rare
earth elements (LREE). When preceded on the liquidus only by olivine, which
does not appreciably fractionate the REE or other incompatible trace elements
[5], plagioclase crystals carry valuable compositional clues about primitive magma
compositions. In this study I use major, minor, and trace element abundances
measured in plagioclase by electron probe microanalysis (EPMA) and laser abla-
tion ICP-MS (LA-ICP-MS) to invert chemical compositions of parental (equilib-
rium) magmas. Accurate partition coefficient data are critical for this inversion.
Plagioclase partition coefficients must be applied carefully because anorthite (An
= 100*[Ca/(Ca+Na)]) content has a dominant influence on cation substitution
[7, 10]. Combined with accurate partition coefficients, inferred parent magma
compositions provide insight into magma evolution beyond what is revealed by
whole-rock data alone.
88

Figure 3.1. Map of the Hawaiian Ridge and Emperor Seamount Chain
(ESC) showing the locations and ages of selected seamounts in the ESC and
the location of Detroit Seamount and a bathymetric map of Detroit
Seamount showing the locations of ODP Sites 884 and 1203. Map is adapted
from [128]. Age data are from [43, 79].
89

Figure 3.2. Stratigraphy of basalts and volcaniclastic rocks cored at ODP
Site 1203 and 884 including the depth, rock types, flow types, and Unit
boundaries (both observed and inferred). Stratigraphic columns are adapted
from [128] and [118]. CSD only and CSD plus microanalysis sample locations
are labeled with white and gray stars, respectively.
90

3.2 The Emperor Seamount Chain
The Hawaiian Ridge and the Emperor Seamount Chain define a nearly con-
tinuous 6000 km long age progressive track of hot spot magmatism ranging in age
from ∼ 85 Ma at Meiji Seamount ([79]) to active volcanism on the Island of Hawaii
(Fig. 3.1). The northwest trending Hawaiian ridge meets the northerly trending
Emperor Seamount Chain south of Koko Seamount (48-49 Ma) [43] (Fig. 3.1).
Early plate reconstructions (e.g., [95]) and recent paleomagnetic studies (e.g.,
[34]) have indicated that the Hawaiian hot spot was close to a mid-ocean spread-
ing center during the formation of Meiji and Detroit Seamounts. Of the oldest
two seamounts, Detroit Seamount is the best sampled and has thus been a focus
of studies aimed at understanding the nature of the Late Cretaceous interaction
of the Hawaiian hot-spot with a MOR e.g., [72, 79].
3.3 Detroit Seamount
3.3.1 Samples
Detroit Seamount covers an area of ∼ 10,000 km 2 and likely represents the
remains of multiple coalesced shield volcanoes [43, 72]. Detroit Seamount was
sampled at multiple drillsites during Legs 145 and 197 of the Ocean Drilling Pro-
gram (ODP). Huang et al. [72] provided a thorough review and interpretation of
whole-rock isotopic and geochemical variations in basalts recovered from Detroit
Seamount. I focused on samples recovered from Sites 1203 and 884 (see Fig. 3.1).
Site 884 is on the northeastern flank of DSM where 87 m of igneous basement was
cored [118] (Fig. 3.1). Basalt samples cored at Site 884 represent 13 flow units
that consist of tholeiitic massive flows in the upper portion of the cored section
and highly plagioclase phyric pillow lavas in lower portion of the cored section
91

[118]. No volcaniclastic rocks were recovered at Site 884 [118] (Fig. 3.2). Keller
et al. [78] reported an 40 Ar- 39 Ar age of 81 ± 1 Ma for a single Site 884 basalt
sample. Site 1203 is located on the eastern flank of Detroit Seamount where 453
m of igneous basement was cored [128] (Fig. 3.2). Eighteen flow units were re-
covered at Site 1203 including tholeiitic pillow basalts, tholeiitic simple pahoehoe
flows, and alkalic compound pahoehoe flows [128]. A variety of volcaniclastic and
sedimentary rocks were recovered from Site 1203 [128] (Fig. 3.2). Duncan and
Keller [43] reported an average 40 Ar- 39 Ar age of 76 ± 1 Ma for four separate lava
flows from Site 1203. Basalts in lower portion of the cored section from Site 1203
include tholeiitic pillow basalts and tholeiitic simple pahoehoe flows but are dom-
inated by thick compound alkalic pahoehoe flows. Basalts in the upper portion
of the cored section from Site 1203 are all tholeiitic pillow basalts and tholeiitic
simple pahoehoe flows [128]. The alkalic to tholeiitic transition is between Units
18 and 19 (Fig. 3.2)
3.3.2 Volcanic history of Detroit Seamount
The tholeiitic compositions of Site 884 basalts are consistent with the shield
building stage of a typical Hawaiian shield volcano [29]. Drilling at Site 1203,
however, recovered a more extensive sequence of lavas that exhibit an upward
change from dominantly subaerial erupted alkalic basalt to submarine erupted
tholeiitic basalt, which Huang et al. [72] noted as compositionally consistent
with the transition from the pre-shield to shield stage of magmatism. The entire
Island of Hawaii is made up of five shield volcanoes and has been constructed
over ∼ 1 Myr [29]. The ∼ 5 Myr age difference between Site 1203 and 884 lavas
and the eruption environment differences between Site 884 and Site 1203 lavas
92

are thus inconsistent with a typical Hawaii-like emplacement. Huang et al. [72]
suggested Site 884 lavas were erupted closest to the spreading ridge axis and that
the proximity of the Hawaiian hotspot to the ridge at ∼ 81 Ma brought about
greater flow of plume mantle toward the ridge axis prior to mantle partial melting
[72, 119]. They suggest Site 1203 lavas were erupted away from the ridge axis
through hotspot-related lithosphere [72]. It cannot be ruled out that Sites 1203
and 884 sampled separate shield volcanoes (see discussion in [72]). Huang et al.
[72] suggested that all of the Site 1203 lavas came from single volcano, where melts
were segregated at lower mean pressure than Hawaiian volcanoes but higher mean
pressure than Site 884 melts. The sequence of rocks recovered from Site 1203
represent a more diverse record of eruptive styles relative to cored section from
Site 884, as well as several distinct periods of non-deposition or erosion (between
Units 5 and 6, Units 23 and 24, and Units 27 and 28) [128] (Fig. 3.2).
3.3.3 Sampling Strategy
A large proportion of the basalts from Site 1203 and Site 884 contain high
volumes of plagioclase phenocrysts [118, 128]. Regelous et al. [119] noted that this
has a diluting effect on whole-rock compositions. Given the size and volume of the
phenocrysts in the numerous plagioclase phyric basalts from Sites 1203 and 884, it
is apparent that whole-rock compositions of these lavas may be best characterized
as hybrids rather than true liquid compositions. Perhaps of greater significance is
that the plagioclase crystals, which have clearly influenced bulk-rock compositions
([119]), may hold clues to otherwise inaccesible facets of magma evolution.
I measured a total of 14 plagioclase CSDs and two olivine CSDs on 12 separate
lava flows from Site 1203 to examine temporal changes in the physical history of
93

magmatic crystallization. Using the CSD results I selected three Site 1203 pillow
lavas for microanalytical studies to examine crystal provenance. I measured six
plagioclase CSDs on two separate plagioclase megaphyric pillow basalts from Site
884 to examine the origins and histories of the cm size plagioclase phenocrysts in
these flows and to examine CSD variations within single flows.
3.4 Analytical Methods
3.4.1 Sample Preparation
Basalt samples were prepared as polished thin sections using a microdiamond
slurry and final grit size of 0.5 µm. Prior to analytical work the slides were rinsed
in ethanol in an ultrasonic bath to remove remnants of the polishing compounds
followed by three 30 minute rinses in ultrapure water in an ultrasonic bath. Sam-
ples were carbon coated for EPMA work. After completion of EPMA work the
carbon coats were removed with 3 µm and 1 µm diamond polishing clothes under
ethanol followed by three 30 minutes rinses in ultrapure water in an ultrasonic
bath. The surface of each slide was inspected under a reflected light optical mi-
croscope to confirm that all vestiges of the carbon coat had been removed prior
to LA-ICP-MS work.
3.4.2 Crystal Size Distribution (CSD) Measurement
Digital image mosaics of entire petrographic thin sections were captured in
cross polarized light using an automated microscope stage system manufactured
by Prior Scientific Instruments, Cambridge, UK. Photographs of each scanned area
were printed on photograph paper and overlain with tracing paper. Minerals of
interest (plagioclase and/or olivine) were outlined by hand using a pen with a 0.1
94

mm diameter tip (which also corresponds to the smallest measurable crystal size).
The tracings were digitized via high resolution scanning with a flatbed scanner.
Crystal intersection areas were shaded gray, detected, and measured according
to a best fit ellipse routine using the freeware program UTHSCSA ImageTool
(http://ddsdx.uthscsa.edu/dig/itdesc.html). The best fit ellipse major and minor
axis results were input into the spreadsheet program CSDslice written by Morgan
and Jerram [106] to estimate the three-dimensional crystal habit. The crystal
major axis data, the estimated 3D crystal habit, rock fabric, total area measured,
and an estimate of crystal roundness were input into the program CSDcorrections
version 1.37 written by Michael Higgins [68] to covert two-dimensional CSD data
to true three-dimensional CSDs as outlined by Marsh [96]. The areas and numbers
of crystals measured in each thin section are listed in table 3.1. Modal abundances
of plagioclase (and olivine for Site 1203 Units 11 and 16) were measured as the
percentage of the phase in the traced area, which Higgins [69] determined to be
an accurate method for estimation of modal abundance.
3.4.3 Electron Probe Microanalysis (EPMA) and Scanning Electron Microscopy
Backscatter electron images and major element analyses were performed us-
ing a JEOL JXA-8600 Superprobe electron microprobe at the University of Notre
Dame. Backscatter electron images were collected using a 1 µm beam, an acceler-
ating voltage of 20 kV, and a probe current of 25-50 nA. Microprobe analyses were
performed using a 10 µm defocused beam, accelerating voltage of 15 kV, a probe
current of 20 nA, 15 second on-peak counting time, and two background measure-
ments per peak. Sodium was measured first to minimize loss via volatilization.
Microprobe data were corrected for matrix effects using a ZAF correction rou-
95

Figure 3.3. A cross polarized light photomosaic (a) of a Detroit Seamount
basalt (Unit 3) and the trace and filled outlines (b) of plagioclase crystals. In
a steady state open system a magma in a wholly liquid state that experiences
simple nucleation and growth of crystals upon cooling will have a linear CSD
[96], but if physical history of crystal nucleation and growth may be more
complex and may include crystal accumulation, crystal sorting or loss,
textural coarsening, or magma mixing (c).
96

tine. Data points near Fe-rich phases such as melt inclusions and alteration-filled
fractures were discarded.
3.4.4 Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-
MS)
Strontium, Y, Ba, La, Ce, Nd, Sm, and Eu were measured in plagioclase
crystals using a New Wave UP-213 UV laser ablation system interfaced with a
ThermoFinnigan Element 2 ICP-MS operated in fast magnet scanning mode at
the University of Notre Dame. I used a laser frequency of 5 Hz, pulse energy of
0.02-0.03 mJ pulse −1 , 15 µm or 30 µm diameter pits depending on crystal size
(40 µm diameter pits were used for analysis of Site 884 plagioclase crystals), and
helium as the carrier gas (∼ 0.7 l min −1 ) mixed with argon (∼ 1.0 l min −1 ) before
introduction to the plasma. Due to the transient nature of the laser ablation sig-
nal, analyses were conducted in peak jumping mode with one point quantified per
mass. The LA-ICP-MS spots were coincident with previous EPMA analyses, and
Ca measured by EPMA was used as an internal standard for each spot analysis,
because its fractionation behavior is similar to that of Sr, Ba, Y, and the REE
[52]. The trace element glass NIST 612 was used as a calibration standard for
all laser ablation analyses. Although heterogeneity for certain elements has been
documented in the widely used NIST 612 glass (i.e., [44]), Eggins et al. [44] consid-
ered it a reliable calibration standard for Sr, Ba, Y, and the REE. The analytical
protocols of Longerich et al. [88] were used for LA-ICP-MS data reduction.
97

3.4.5 Major and Trace Element Data quality
The quality of EPMA data were monitored by major element and cation totals.
Data points where major element totals were greater than 101.5% or less than
98.5% are not used in further discussion nor are points with cation totals greater
or less than 20 ± 0.1 cations (based on 32 O). All laser ablation analyses were
collected in time-resolved mode so that signal from inclusions and alteration filled
fractures could be easily identified. Results from laser ablation spots close to
fractures or that penetrated into underlying fractures were discarded. The rate of
laser ablation was tested using a 100 µm thick plagioclase wafer prior to sampling,
which allowed optimization of operating conditions to yield desired sensitivity and
sampling depth (< 30 µm).
3.4.6 Choosing partition coefficients (D)
Blundy [9] noted that inversion of magma compositions from mineral data
using appropriate trace element D values is a robust means of estimating parent
magma compositions and suggested that trace element D values derived from
microbeam techniques were more accurate than those derived from bulk crystal-
matrix analyses. Blundy and Wood [10, 11] and Bindeman et al. [7] showed
that the dominant factors controlling the trace element D vaues for plagioclase
are An content and crystallization temperature. Bindeman et al. [7] applied this
relationship to a variety of major and trace elements and produced values for the
constants “a” and “b” in their Equation 2 (equation 3.1 below) for calculating
plagioclase partition coefficients via the expression:
RT ln(Di) = aXAn + b (3.1)
98

where R is the gas constant, T is temperature (Kelvin), i is the element of interest,
and XAn is the mole fraction of anorthite. Using equation 3.1, Ginibre et al.
[59] noted that temperatures in the range of 850-1,000 ◦ C had miniscule effect on
calculated melt concentrations for Sr and Ba (within their analytical uncertainty).
Likewise, Bindeman et al. [7] showed that variations of ∼150 ◦ C produce < 10%
differences for particular partition coefficients, which were often within error of
the respective D values.
Blundy and Wood [11] presented an alternative model for quantitative predic-
tion of D values based upon thermodynamic principles and lattice strain theory
(i.e., [15]). Specifically, they noted the dominant roles of cation size, lattice site
size, and elasticity of the lattice site related to cation substitution in minerals.
The elastic response of a lattice site to cation substitution is measured by Young’s
Modulus (E), and the value of E plag is dependent upon plagioclase An content
and the charge of the substituent cation [11, 12]. Equation 3.2 (Equation 2 of
Blundy and Wood [11]) is the basis of their predictive D model that I apply to
plagioclase,
Di = D z+
o
∗ exp
−4πEplagNA
ro
2 (r2 o − r2 1
i ) + 3 (r3 i − r3 o )
RT
(3.2)
where Di of a trace element (i) is a function of the strain compensated partition
coefficient for the divalent (D 2+
0 ) or trivalent (D 3+
0 ) substituent cation, Young’s
Modulus for the plagioclase M lattice site (E plag), NA (Avogadro’s Number), the
radius of the plagioclase M site (ro), the radius of the substituent cation (ri),
the gas constant (R), and crystallization temperature (T). For a more thorough
discussion of this model see Blundy and Wood [11, 12].
99

3.4.6.1 Partition coefficients for DSM plagioclase crystals
I used equation 3.2 and the parameterizations of Blundy and Wood [11, 12] to
calculate D values for Sr, Y, Ba, La, Ce, Nd, and Sm. (see Table 3.4). I assumed
a crystallization temperature of 1190 ◦ C to calculate D values. The validity of
this assumption was examined using the MELTS program [2, 55] to determine
the temperature at which plagioclase arrives on the liquidus during crystallization
of a typical Detroit Seamount tholeiitic basalt. I input Detroit Seamount tholei-
itic basaltic glass compositions reported by Huang et al. [72] into the MELTS
program. During equilibrium crystallization of the lowest MgO glass from Site
1203 (Unit 3), plagioclase appeared on the liquidus at 1170 ◦ C, at at 1190 ◦ C from
the highest MgO glass from Site 1203 (top of Unit 18), and at 1200 ◦ C from the
highest MgO glass from Site 884 (Unit 10). The use of 1170 ◦ C vs. 1200 ◦ C in D
calculations yields < 4% differences in trace element concentrations of parent liq-
uids, which is within the analytical error for my analyses. The effects of pressure
on plagioclase D values are not well documented for most elements, and where
they were documented for DSr by Vander Auwera et al. [137] they were negligible.
I used a crystallization pressure of 0.001 kbar for D calculations. I assume the
strain compensated (i.e., strain free) partition coefficient D 2+
0 in plagioclase is best
approximated by DCa (c.f., [12]), and to estimate DCa I used equation 3.1 and the
“a” and “b” constants reported for Ca by Bindeman et al. [7]. Estimation of D 3+
0
is less straightforward. I assumed D 3+
0 to be ∼ 7.5% of D 2+
0 , which was derived
from experimental data reported in Blundy [9]. I used equation 3.1 and the “a”
and “b” constants reported for Ti by Bindeman et al. [7] to estimate DTi
100

3.5 Results
3.5.1 Crystal Size Distributions
3.5.1.1 Site 1203 CSDs
Plagioclase CSDs of Site 1203 basalts are categorized as linear, gently curved
upward, or strongly kinked upward (Fig. 3.4). Of the 12 separate flows I sampled
from Site 1203, two flows: Units 14 and 31, have strongly upward kinked CSDs. A
number of flows have plagioclase CSDs that are gently curved upward, including
Unit 3 (top), Unit 8, Unit 18, and Unit 19. Each of these flows, with the exception
of Unit 19, are tholeiitic pillow basalts (Fig. 3.4). Unit 19 is a subaerial alkalic
basalt, is the youngest alkalic basalt sampled at Site 1203, and is the closest to
the alkalic to tholeiitic basalt transition. Three submarine tholeiitic sheet flows
were examined (Units 11, 16, and 24), and each have linear plagioclase CSDs. The
Unit 11 and Unit 16 tholeiitic sheet lobes have linear olivine CSDs with slopes
nearly identical to their corresponding plagioclase CSDs (Fig. 3.4, Table 3.2). I
sampled two site 1203 basalts near flow tops and bottoms. Unit 3, which is a ∼
25 m thick tholeiitic pillow basalt, and Unit 23, which is a ∼ 62 m thick alkalic
compound pahoehoe basalt flow. The plagioclase CSD of the top of Unit 3 is
curved gently upward (Fig. 3.4b) but otherwise has a nearly identical slope to the
CSD from the bottom of the flow. Samples from both the top and bottom of Unit
23 have linear plagioclase CSDs but the sample from the top of the flow has a
distinctly shallower slope (Fig. 3.4i). Below the alkalic to tholeiitic transition a
larger proportion of the flows I examined have linear CSDs (Fig. 3.4).
101

3.5.1.2 Site 884 CSDs
I measured plagioclase CSDs of samples taken from the tops, middles, and
bottoms of two thin tholeiitic pillow basalts (Unit 8: ∼ 4 m thick and Unit 10:
∼ 7 m thick) from Site 884 (Fig. 3.2). The top of Unit 8 is within error of being
linear, whereas samples from the middle and bottom of the Unit 8 pillow basalt
exhibit identical curved CSDs (Fig. 3.5). Samples from the top and bottom of
the Unit 10 pillow basalt have similar initial slopes (i.e., slope at smaller L), but
the sample from the top of the flow is curved more strongly upward. The CSD
of the sample from the middle of Unit 10 has a shallower initial slope and is the
most curved upward of the three sampled sections of Site 884 Unit 10 (Fig. 3.5;
Table 3.2).
102

Figure 3.4. Plagioclase and olivine CSDs of selected ODP Site 1203 basalts.
Basalts below the alkalic to tholeiitic transition, the boundary between Units
18 and 19, are more frequently linear. Units 14 and 31 have strongly kinked
CSDs. With the exception of Unit 1, all pillow basalts have non-linear CSDs.
103

Figure 3.5. Plagioclase CSDs of ODP Site 884 Unit 8 and 10 pillow basalts.
Plagioclase CSDs were measured on samples taken from flow tops, middles,
and bottoms. a) The top of Unit 8 has a linear CSD and the lowest modal
plagioclase abundance. b) The middle of Unit 10 has the most prominent
upward CSD curvature and the greatest modal plagioclase abundance.
104

105
TABLE 3.1
CSD MEASUREMENT INFORMATION
Sample ODP Site Unit Mineral Crystals Measured Area Measured (mm 2 ) Habit (S:I:L) Habit R 2
17R5-34-36 1203 1 plagioclase 1502 560.2 1-3.2-8 0.78
19R2-136-138 1203 3 top plagioclase 2356 100.4 1-3.2-10 0.74
20R3-4-10 1203 3 bottom plagioclase 3477 105.9 1-3.2-10 0.80
31R1-29-31 1203 8 plagioclase 721 516 1-3.2-10 0.69
32R3-85-87 PL 1203 11 plagioclase 842 585.2 1-2.9-7 0.80
32R3-85-87 OL 1203 11 olivine 462 585.2 1-1.5-1.8 0.92
36R4-92-95 1203 14 plagioclase 3388 72.3 1-4.0-9.0 0.88
37R3-10-13 PL 1203 16 plagioclase 1781 350.5 1-3.4-9 0.85
37R3-10-13 OL 1203 16 olivine 859 350.5 1-1.25-1.9 0.89
40R5-119-121 1203 18 plagioclase 2762 478.1 1-4.0-10 0.90
42R5-20-23 1203 19 plagioclase 1285 711.8 1-3.6-7 0.86
52R6-12-14 1203 23 plagioclase 3280 327.8 1-3.2-10 0.88
57R2-127-129 1203 23 plagioclase 2722 199.7 1-3.4-9 0.87
59R2-99-102 1203 24 plagioclase 2580 394 1-4.0-9.0 0.91
59R5-30-32 1203 26 plagioclase 1902 108 1-4.0-9.0 0.86
68R4-29-30 1203 31 plagioclase 2227 164.9 1-2.9-5.5 0.84
9R1-36-45 884 8 top plagioclase 2653 857.7 1-3.0-10 0.82
9R2-60-75 884 8 middle plagioclase 2406 861.5 1-3.2-8 0.83
9R3-1-12 884 8 bottom plagioclase 2940 789.7 1-3.2-8 0.85
10R2-31-43 884 10 top plagioclase 2711 805.1 1-3.4-7 0.84
10R3-60-74 884 10 middle plagioclase 2329 728.8 1-3.2-8 0.82
10R4-33-49 884 10 bottom plagioclase 2561 576.7 1-3.6-7 0.88
1 Crystal habit was estimated using the CSDslice program written by Morgan and Jerram [106]. S:I:L = short:intermediate:long axis lengths.

106
TABLE 3.2
CSD RESULTS
Sample ODP Site Unit Mineral CSD segment Slope (mm −1 ) σ (slope) ln n o (mm −4 ) σ (intercept) X plag X oliv
17R5-34-36 1203 1 plagioclase linear -1.95 0.03 3.57 0.04 0.13 –
19R2-136-138 1203 3 top plagioclase A -4.47 0.06 7.68 0.03 0.23 –
B -1.26 0.02 1.69 0.01 – –
20R3-4-10 1203 3 bottom plagioclase A -4.33 0.05 7.86 0.02 0.17 –
31R1-29-31 1203 8 plagioclase A -4.24 0.08 5.01 0.07 0.04 –
B -0.96 0.02 -0.49 0.01 – –
32R3-85-87 PL 1203 11PL plagioclase A -1.80 0.04 2.73 0.19 0.08 0.38
32R3-85-87 OL 1203 11OL olivine A -2.01 0.08 1.38 0.03 0.08 0.38
36R4-92-95 1203 14 plagioclase A -10.80 0.16 9.89 0.03 0.2 –
B -1.31 0.02 0.86 0.00 – –
37R3-10-13 PL 1203 16PL plagioclase linear -3.55 0.05 5.82 0.06 0.07 0.3
37R3-10-13 OL 1203 16OL olivine linear -4.07 0.12 3.34 0.07 0.07 0.3
40R5-119-121 1203 18 plagioclase A -3.71 0.06 6 0.04 0.11 –
B -0.91 0.02 -1.15 0.01 – –
42R5-20-23 1203 19 plagioclase A -3.90 0.06 4.44 0.05 0.14 –
B -1.10 0.02 -0.11 0.00 – –
52R6-12-14 1203 23 plagioclase linear -2.99 0.04 6.18 0.04 0.14 –
57R2-127-129 1203 23 plagioclase A -4.43 0.08 6.91 0.04 0.11 –
59R2-99-102 1203 24 plagioclase linear -4.36 0.09 6.12 0.05 0.07 –
59R5-30-32 1203 26 plagioclase linear -6.27 0.13 7.75 0.06 0.09 –
68R4-29-30 1203 31 plagioclase A -17.77 0.21 9.26 0.03 0.18 –
B -1.69 0.02 1.72 0.01 – –
9R1-36-45 884 8 top plagioclase A -1.36 0.02 3.02 0.03 0.17 –
9R2-60-75 884 8 middle plagioclase A -2.01 0.03 3.48 0.03 0.36 –
B -0.31 0.00 -4.32 0.04 – –
9R3-1-12 884 8 bottom plagioclase A -2.52 0.05 4.27 0.03 0.27 –
B -0.39 0.01 -3.8 0.03 – –
10R2-31-43 884 10 top plagioclase A -3.79 0.05 4.94 0.04 0.34 –
B -0.59 0.01 -1.94 0.02 – –
10R3-60-74 884 10 middle plagioclase A -2.12 0.03 3.83 0.03 0.43 –
B -0.37 0.01 -2.95 0.03 – –
10R4-33-49 884 10 bottom plagioclase A -4.07 0.07 5.57 0.04 0.24 –
B -0.95 0.02 -0.65 0.00 – –
1 Segment A of a curved or kinked CSD here refers to the small crystals that populate the steeper segment, whereas segment B refers to the larger crystals.
2 1σ errors were calculated following the protocol of Higgins [68].
3 Several samples listed contained sparse olivine phenocrysts but each less than 5% by volume unless otherwise noted.

3.5.2 CSDs as Guides for Microanalysis
I selected three samples from Site 1203 for detailed microanalysis of plagio-
clase phenocrysts based upon CSD curvature (Fig. 3.6). I chose to examine Site
1203 Units 14 and 31 due to the highly kinked nature of their CSDs. I also chose
one gently curved CSD sample, the top of Site 1203 Unit 3, for microanalysis to
compare with the results from the strongly kinked CSDs. I selected two large
phenocrysts, one each from the middle and bottom of Site 884 Unit 8 for micro-
analysis. I divided non-linear CSDs into two segments for targeted microanalysis
based upon the approximate position CSD slope change (Fig. 3.6a). A generaliza-
tion of the division of CSDs into separate segments, which are hereafter referred
to as Population A (smaller L) and Population B (larger L), is illustrated in fig-
ure 3.6a. The locations and sample names of each plagioclase crystal analyzed in
this study, as listed in tables 3.3 and 3.4, are labeled in figure 3.6b-e. The stereo-
logic correction used to convert 2D crystal sizes to 3D crystal sizes for generation
of true 3D CSD plots was applied to each crystal each crystal examined in order
to accurately sample CSD segments shown in figure 3.6b-e.
3.5.3 Petrography and Crystal Zoning Patterns of Microanalysis Samples
3.5.3.1 Site 1203 Unit 3
Unit 3 basalts have massive textures and large plagioclase phenocrysts (i.e.,
population B) that occur both as individual euhedral crystals and as portions of
rounded glomerocrysts (e.g., Fig. 3.7a). Both the large euhedral crystals and the
crystals present within glomerocrysts exhibit oscillatory zoning (e.g., Fig. 3.8d)
superimposed upn normal zonation ± distinct changes in An content of > 5 mole %
(e.g., Fig. 3.8c,d). Major element zonation is visible in backscatter electron images
107

ln (n) [mm -4 ]
ln (n) [mm -4 ]
10
5
0
-5
5
0
-5
Unit 8 middle
Unit 8 bottom
ln (n) [mm -4 ]
Site 884 Unit 8
}Crystals: PG1-PG5,PG-8-PG-13,Y3A,Y3C,Y3F,Y4A,
Y4D,Y4E,Y4F,Y5B,Y2B,Y2C,Y2D,
XLE1-XLE4,XLB1-XLB4,XLC1-XLC-5,
XLD1,XLD2,SB1
} Crystals:
Site 1203 Unit 14
-10
0 2 4 6
L [mm]
8 10
10
5
0
-5
}
population A
}
-10
0 2 4 6
L [mm]
8 10
A,B,C1,D1,E,X1,X2,Y3B,Y3D,Y3E,Y4G,
Y5A,Y2A,X4,C,B2,A2
population B
}
A)
Crystals: D-S1,D-S2,E-1,F,H,PL-A,PL-B,PL-C,PL-D,
PL-E,PL-F,PL-G,PL-H
B) C)
9R2A 9R3B }Crystals:
A,B,C,D-L,
J,X1,X4,X2-1
Site 1203 Unit 3
Unit 3 top
Unit 3 bottom
}Crystals: SPL3,SPL1,SPL,5,SPL2,SPL6,SPL7,SPL8,SPL9,
SPL10,Y1,Y2B,Y3,Y4A,Y4B,Y6C,Y6B
D) E)
A,Bsm,C,D,E,G,I,J,Y5L,
Y10,Y11
}Crystals:
Site 1203 Unit 31
0 2 4
L [mm]
6 8
Figure 3.6. a) Non-linear CSDs were divided into two regions based upon
the approximate position of CSD slope change. Crystals from each
population segment were then targets for compositional microanalysis. The
general locations of crystals listed in tables 3.1 and 3.2 are shown for b) Site
884 Unit 8; c) Site 1203 Unit 3; d) Site 1203 Unit 14; and e) Site 1203 Unit
31.
108

and is most apparent near population B crystal rims where there are changes of up
to > 10 mole % An. Population A and B crystals have thin (< 10 µm) rims that
are as low as An62. The majority of the population A crystals in Unit 3 samples
are unzoned or normally zoned (Fig. 3.7b). There is a small proportion of crystals
exhibiting reverse zonation that have chaotic zoning in their crystal cores that on
average have lower An than the rims (Fig. 3.8b). Small reverse zoned plagioclase
crystals are commonly present in clots of clinopyroxene and plagioclase crystals,
where the small clinopyroxene crystals also exhibit reverse zonation (see inset
image in Fig. 3.8b). Of the two Unit 3 basalt samples I examined, the sample
from the bottom of the flow has a 17% modal abundance of plagioclase and the
sample from the top of the flow has 23% modal plagioclase. I did not measure
olivine modal abundances in the Unit 3 samples but were estimated to be 3-5 %
by Tarduno et al. [128].
3.5.3.2 Site 1203 Unit 14
The Unit 14 basalt I examined has a massive texture and population B crystals
that occur as clusters of euhedral crystals and as occasional individual rounded
crystals (e.g., Fig. 3.7b). Population B crystals exhibit oscillatory zoning (e.g.,
Fig. 3.8e,f) superimposed upon normal zonation with distinct changes in An con-
tent near the rims, often a > 10 mole % An change (e.g., Fig. 3.8e,f). The majority
( 95%) of the population A crystals exhibit the zoning pattern displayed in fig-
ure 3.8g. This pattern, which is characterized by a change from An69−71 in the
core to a thin (2-5 µm) higher An zone then an outermost rim that is An69−71, is
similar to the zoning pattern observed along the rims of the population B crystals
(e.g., rims in Figs. 3.8e,f vs. Fig. 3.8g). The remaining proportion of the pop-
109

ulation A crystals are unzoned (An64−79) and normal zoned with cores up An82
(e.g., Fig. 3.8h). The single Unit 14 basalt sample I examined has a 20% modal
abundance of plagioclase. I did not measure olivine modal abundances in the Unit
14 sample and Tarduno et al. [128] did not note any olivine in this sample.
Figure 3.7. In each basalt chosen for microanalysis large phenocrysts (i.e.,
population B crystals) are present as glomerocrysts and individual crystals
that are commonly rounded or embayed. The field of view in each
photomicrograph is 2 cm. Large plagioclase phenocrysts are a) 0.5-2 mm in
Site 1203 Unit 3; b) 1-8 mm in Site 1203 Unit 14; c) 1-4 mm in Site 1203
Unit 31; d) up to 1.5 cm in the sample from the middle of Site 884 Unit 8;
and e) up to 1.5 cm from the bottom of Site 884 Unit 8.
110

3.5.3.3 Site 1203 Unit 31
The Unit 31 basalt I examined has a massive texture and population B crys-
tals that occur as rounded and euhedral individual crystals or glomerocrysts (e.g.,
Fig. 3.7c). Most population B crystals exhibit interior oscillatory zoning superim-
posed upon normal zonation with distinct changes in An content near the rims,
often a > 10 mole % An change (e.g., Fig. 3.8j). A small number of population B
crystals exhibit patchy disorganized major element zoning in their interiors with
An variations close to the limits of An content resolution (∼ 1-2 mol % An) of
backscatter electron images (see Fig. 3.8i). Population A crystals are unzoned and
normal zoned (Fig. 3.8k,l). The single Unit 31 basalt sample I examined has 18%
modal abundance of plagioclase. I did not measure olivine modal abundances in
the Unit 31 sample but was estimated to be < 1% by Tarduno et al. [128].
3.5.3.4 Site 884 Unit 8
The Site 884 Unit 8 basalts I examined have massive textures and and cm size
population B crystals that occur as rounded clusters of crystals and individual
euhedral or rounded crystals (e.g., Fig. 3.7d,e). The interiors of the population B
crystals exhibit oscillatory type zoning and normal zonation with subtle decreases
in An content of 1-4 mole % near the rims (e.g., Fig. 3.8m,n). Thin (5-10 µm)
low An rims are apparent in backscatter electron images of population B crystals
but were not sampled in detail. The basalt sample from the middle of this flow
has 36% modal plagioclase abundance, the sample from the bottom of the flow
has 27% modal plagioclase abundance, and the sample from the flow top has 17%
modal plagioclase abundance. I did not measure olivine modal abundances in the
Site 884 Unit 8 samples but they were estimated to be < 1% by Rea et al. [118].
111

Figure 3.8. a) The majority of Site 1203 Unit 3 population A crystals are normal zoned or
unzoned. b) There is a small proportion of population A crystals in Unit 3 that are reverse zoned
that are clustered with small reverse zoned clinopyroxene crystals (inset image). c) Unit 3
population B crystals exhibit oscillatory zonation superimposed upon a (d) normal zoning pattern.
e) Site 1203 Unit 14 population B crystals exhibit oscillatory zonation superimposed upon a (f)
normal zoning pattern. g) The majority of Unit 14 population A crystals exhibit minor low An
cores bounded by thin An-rich mantles then low An rims. h) A small proportion of Unit 14
population A crystals exhibit normal zoning. i) A small number of Unit 31 population B crystals
exhibit patchy disorganized interior zoning. j) Most Site 1203 Unit 31 population B crystals exhibit
oscillatory zonation superimposed upon a normal zoning pattern. k) A small proportion of Unit 31
population A crystals are (l) normal zoned, but the majority are unzoned. m,n) Site 884 Unit 8
population B crystals exhibit oscillatory zonation superimposed upon normal zoning.
112

3.5.4 Major Elements
Major element compositions measured by EPMA and Ti abundances measured
by LA-ICP-MS in plagioclase crystals from each Unit are reported in table 3.3
and illustrated in figure 3.9.
3.5.4.1 Site 1203 Unit 3
I noted variations in An content apparent in backscatter electron images that
correlate with zoning patterns. Unit 3 Population A crystal cores fall into two
groups in terms of An content. The higher An group (An74−79) overlaps the
population B cores (An78−81) (Fig. 3.9a). The cores of population B crystals
extend to higher An contents (up to An81) than population A crystals. There is
no obvious variation in Mg # (measured in plagioclase) between population A
and B crystals or with An content.
3.5.4.2 Site 1203 Unit 14
The cores of population A crystals from Unit 14 also fall into two groups in
terms of An content. The more An rich population A group have An77−82 cores
that overlap the compositions of population B cores, and the lower An population
A group have An64−76 cores. The cores of population B crystals have a higher
range of An content (An81−87), but a single population B crystal with a core
composition of An73−74 was observed. There is no obvious variation in Mg #
(Mg/Mg+total Fe measured in plagioclase) between population A and B crystals
or with An content, but a single Unit 14 crystal rim (An56) has the lowest overall
Mg # (Fig. 3.9d).
113

3.5.4.3 Site 1203 Unit 31 and Site 884 Unit 8
The cores of population A crystals from Site 1203 Unit 31 define a single
An67−77 group. The cores of Site 123 Unit 31 population B crystals are An-rich
(An80−86), define a single group of compositions, and do not overlap population A
cores. The cores of the two Site 884 Unit 8 population B crystals I examined are
An85−86. There is no obvious variation in Mg # (measured in plagioclase) between
population A and B crystals or with An content. Of all crystals examined with
An82−86 zones, the two Site 884 Unit 8 crystals tend to have the highest Mg #’s
(Fig. 3.9d).
3.5.4.4 Ti variations between crystal populations and Units
There is a strong negative correlation of Ti abundance with An content that
is apparent for crystals from each Site examined, however there seems to be min-
imal correlation at An contents greater than ∼ An72 (Fig. 3.9e). In the An80−87
range Unit 3 population B crystals have greater Ti abundances than population
B crystals from all other Units examined, and population B crystals from Site 884
Unit 8 have the lowest overall Ti abundance (Fig. 3.9e).
114

115
A)
An (mol %)
pop. A cores
pop. B cores
Site 1203 Unit 3
Ab 50 (mol %) Or 50 (mol %)
B)
An (mol %)
An (mol %)
pop. A cores
pop. B cores
Site 1203 Unit 14
Ab 50 (mol %) Or 50 (mol %)
C)
pop. A cores
pop. B cores
Site 884 Unit 8 pop. B cores
Site 1203 Unit 31
(and Site 884 Unit 8)
Ab 50 (mol %) Or 50 (mol %)
Mg #
Ti (ppm)
An (mol %)
Figure 3.9. a) The cores of Site 1203 Unit 3 population A and B crystals overlap, but
population A crystals extend to much lower An. b) Unit 14 and Unit 31 (c) population B crystals
extend up to An87, which is distinct from population A crystals in each Unit. d) There is little
obvious variation of Mg# with An content, between crystal populations, or between Units. e) A
strong correlation of Ti abundance with An content was observed, although correlations appears to
weaken considerably at An > 70. Correlation coefficient for Unit 3 Ti vs. An R2 = 0.79, Unit 14
R2 = 0.80, and Unit 31 R2 = 0.39.
D)
E)

3.5.5 Trace Elements
3.5.5.1 Measured Compositions
Trace element abundances measured by LA-ICP-MS in plagioclase crystals
from each Unit are reported in table 3.3 and illustrated in figure 3.10. Yttrium,
Sr, La, Nd, Sm, and Eu abundance exhibit little or no correlation with An content
in figure 3.10. More prominent negative correlations of Ba and Ce abundance with
An content are apparent if figures 3.10c,e, although these correlations appear to be
most significant at An contents < An70. In the An80−87 range, which is dominated
by the interiors of population B crystals from each Unit, Unit 14 population B
crystals trend to highest Sr abundance, Unit 31 crystals have the lowest Sr of the
three Site 1203 Units I examined, and the Site 884 Unit 8 crystals have lower
abundances Sr than all Site 1203 crystals (Fig. 3.10). In terms of Y, Sm, Nd, and
Eu all crystals in the An80−87 range overlap. It should be noted, however, that the
Site 884 Unit 8 crystals tend to have the lowest abundances of these elements and
Sm low enough to be below analytical detection limits. Lanthanum and Ce exhibit
similar patterns in crystal zones in the An80−87 range, where Unit 31 crystals are
the most La and Ce depleted of the Site 1203 samples and the Site 884 Unit 8
crystals have notably lower La and Ce abundances (Fig. 3.10d,e)
3.5.5.2 Inferred Parent Magma Compositions
I inverted the compositions of magmas in equilibrium with the crystals at dif-
ferent periods during their growth. The calculated parent magma compositions
and the partition coefficients used for inversion are listed in table 3.4. Rare earth
element, Ti, Ba, and Sr abundance in parent magmas of the four Units I examined
exhibit the following general trend of decreasing abundance: Unit 3 ≅ Unit 14 >
116

Figure 3.10. Strong negative correlations of are apparent for c) Ba and e)
Ce, although the correlation is weaker above ∼ An70. There little or weak
correlation of a) Y, b) Sr, d) La, f) Nd, g) Sm, and h) Eu with An content.
Unit 31 population B crystals have the lowest parent magma LREE
abundances of Site 1203 samples. Site 884 Unit 8 population B parent have
the lowest overall abundances of each element measured except Y. Yttrium
abundance of Site 884 population B magmas trends higher than any other
population B parent magma.
117

Unit 31 > Site 884 Unit 8 (Table 3.3). Selected trace element ratios are shown
plotted against An content in figure 3.11. It is difficult to estimate the amount
of error introduced to the inverted compositions via the choice of partition coeffi-
cients, and for this reason the error shown in figure 3.11 is analytical uncertainty
only. Parent magma Ba/Sr ratios of population A and B crystals from Unit 3
overlap at An > 70 and trend to higher ratios at An contents approaching An60
(Fig. 3.11a). Parent magmas of Unit 14 and 31 population A crystals trend to
higher Ba/Sr than parent magmas of population B crystals (Fig. 3.11a). Par-
ent magma Ba/Sr of population B crystals from Units 14, 31, and Site 884 Unit
8 overlap and are lower than Unit 3 population B parent magmas (Fig. 3.11a).
Population A crystals from Units 14 and 31 trend to the lowest overall La/Sm
ratios, but there are no other noteworthy La/Sm distinctions between Units or
crystal populations (Fig. 3.11b). Samarium was below analytical detection limits
in the Site 884 Unit 8 crystals. In terms of La/Y ratios, parent magmas from
populations A and B from Units 3, 14, and 31 overlap, but population A parent
magmas exhibit a wider range of La/Y (Fig. 3.13). I measured trace elements in
a single Unit 31 population A crystal, which has lower parent magma La/Y than
the population B crystals (Fig. 3.13). When the range of La/Y in parent magmas
of the Unit 3 and 14 population A magmas are considered, the separation of the
single Unit 31 population A crystal parent magma from parent magmas of the
population B crystals is probably an artifact of limited sampling. In summary,
the parent magmas of population B crystals from Unit 31 and Site 884 Unit 8 are
the most LREE and trace element depleted of all crystals sampled. There appears
to be little significant fractionation of REE between parent magmas of population
A and B crystals, but there are distinct differences in Ba and Sr content.
118

Figure 3.11. a) Population B parent magmas trend to lower Ba/Sr, which is
sensitive to plagioclase fractionation. In general population B crystals are
older and grew from less evolved melts. b) La/Sm ratios of population A and
B magmas are broadly similar with the exception of Unit 14 population A
parent magmas. These parent magmas trend to lower La/Sm, which may
indicate a unique source compositions relative to Unit 14 population B
crystals.
119

3.6 Discussion
3.6.1 Evidence of Crystal Sorting
In each instance where I measured multiple CSDs in a single flow, I observed
evidence of plagioclase sorting. For example, my observations that the top of the
Site 1203 Unit 3 pillow basalt has greater modal abundance plagioclase and an
upward curved plagioclase CSD are consistent with crystal sorting and concen-
tration (Fig. 3.4b). Plagioclase phenocrysts were concentrated near the top of
this ∼ 25 m thick pillow flow most plausibly by flotation. I observed evidence of
plagioclase sorting in Unit 8 from Site 884, where the plagioclase CSD from the
top of the ∼ 4 m thick flow is linear and has lower modal plagioclase, and CSDs
from the middle and bottom of the flow are nearly identical, curved upward, and
have greater modal plagioclase (Fig. 3.5a; Table 3.2). Faster cooling of the flow
top could have lead flowing lava to concentrate larger crystals toward the bottom
of the flow. Simple gravitational sinking of plagioclase can also explain concentra-
tion of large crystals near the flow bottom, however the density contrast between
basaltic lava and plagioclase would likely have minimized significant plagioclase
settling. Plagioclase sorting is also evident in the Unit 10 pillow flow (∼ 7m thick)
from Site 884 (Fig. 3.5a). The plagioclase CSD I measured from the center of the
flow displays the greatest upward curvature and highest modal plagioclase abun-
dance (Table 3.2), which is consistent with greater accumulation of plagioclase at
the flow center. This type of flow sorting of crystals was documented by Marsh
[98] on a larger scale in basaltic sills.
The two CSDs I measured on samples taken from near the top and bottom
of the ∼ 62 m thick Site 1203 Unit 23 alkalic compound pahoehoe flow appear
to fan (or rotate) about a point at low L (Fig. 3.4i). Two samples provide only
120

a limited picture of CSD variation across this thick flow, but CSD fanning of
co-magmatic lavas may indicate progressive deeper sampling of the solidification
fronts lining the magma chamber [146]. The deeper portions of solidification fronts
lining the magma chamber contain larger and older crystals, and as this region is
increasingly scoured by passing magma en-route to the surface the resulting lava
will have a flatter CSD slope (e.g., [146]). In the case of this single thick flow the
steeper CSD from the bottom of the flow is representative of the earliest and lowest
crystallinity magma emptied from the magma chamber, and as evacuation of the
magma chamber progressed and fed the flow, deeper portions of the solidification
fronts lining the chamber were eroded leading to a shallower sloped CSD near the
top of the flow (i.e., the terminus of magma chamber evacuation).
I conclude that flow sorting of plagioclase was a ubiquitous process during
the formation of Detroit Seamount and that measurement of one plagioclase CSD
per flow is inadequate for making robust interpretations based solely upon CSD
results. Crystal size distribution results do, however, provide insight into when
and where crystal accumulation or magma mixing most heavily influenced bulk
magma composition. Based upon the fact that everywhere I looked for evidence of
crystal sorting I found convincing evidence of this process, I suggest that sorting
was not limited to plagioclase. Indeed, the greater density of olivine relative to
basaltic melt and plagioclase may have accentuated olivine sorting and sorting
due to both flow sorting and gravitational settling. Indeed Tarduno et al. [128]
noted olivine cumulate picritic basalts at Site 1203
121

3.6.1.1 Site 1203 Alkalic Basalts and Tholeiitic Sheet Flows
Each of the alkalic lavas I examined are plagioclase phyric [128]. Close in-
spection of the single alkalic basalt with a curved CSD, Unit 19, reveals that the
upward curvature is largely due to a slope change at the smallest L (i.e., small-
est crystal sizes), which is consistent with a rapid change in cooling related to
eruption [96] (Fig. 3.4h). With the overall linearity of Unit 19 CSD considered,
each of the alkalic basalts I examined can be taken as having a linear CSDs. This
observation suggests an insignificant role for plagioclase removal or accumulation
during the petrogenesis of Site 1203 alkalic basalts. Submarine tholeiitic sheet
lavas have the lowest overall modal abundances of plagioclase as a group and all
have approximately linear CSDs (Figs. 3.4d,f,j).
The Unit 11 and 16 tholeiitic sheet lavas each contain > 30 % modal abundance
olivine and linear olivine CSDs. If I assume similar olivine and plagioclase growth
rates, then the similarity of olivine and plagioclase CSD slopes indicates that
they had similar residence times in magma chambers where crystallization was
dominated by olivine (Figs. 3.4d,f). The single tholeiitic sheet flow recovered
below the alkalic to tholeiitic transition (Unit 24) is devoid of large phenocrysts
and has a steeper CSD slope than the upper sheet flows (Units 11 and 16), which
indicates a short magmatic residence time and minimal partial crystallization.
3.6.1.2 Site 1203 Pillow Basalts
Each tholeiitic pillow basalt I examined except Site 1203 Unit 1 has a curved
or kinked plagioclase CSD (Figs. 3.4, 3.5), which indicates that removal or accu-
mulation of plagioclase was a significant process during the petrogenesis of Detroit
Seamount basalts and was most prevalent during the formation of pillow basalts.
122

The common occurrence of population B crystals in rounded and embayed crys-
tals in Site 1203 Units 3, 14, and 31 and Site 884 Unit 8 basalts is a qualitative
indication that these crystals were commonly out of equilibrium with their host
basalts, which is also consistent with plagioclase accumulation.
Huang et al. [72] specifically noted the significance of plagioclase accumulation
in Unit 14 and 31 lavas from Site 1203 (i.e. Fig. 13 of [72]). My observations
of strongly upward kinked plagioclase CSDs for these two Units (Fig. 3.4e,l) are
consistent with their assertions of plagioclase accumulation. Although I provide
seemingly independent evidence of plagioclase accumulation in these two lavas,
the exact origin of the non-linear CSDs of these lavas is not straightforward. Both
magma mixing and crystal accumulation can produce a non-linear CSDs (Fig. 3.3)
[67]. Higgins [67] discussed a number of explanations for curved CSDs, including
size dependent crystal growth rate, destruction of small crystals (i.e., fines destruc-
tion), changes in cooling rate, or changes in the dominant crystallizing phases.
Cashman and Marsh [20] and Marsh [99] presented sound arguments against size
dependent growth. Fines destruction and changes in cooling rate cannot be ruled
out based upon textural evidence alone. I have noted distinct compositional dif-
ferences between population A and B crystals from Site 1203 Units 14 and Units
31, which is evidence against these processes as noted by Higgins [67]. Huang et
al. [72] suggested a significant role for olivine and plagioclase fractionation during
DSM basalt petrogenesis, and there is little other evidence to indicate there was
a change in the dominant crystallizing phases. Although I cannot rule out minor
roles for any of these processes, I suggest accumulation and/or magma mixing are
responsible for the majority of the CSD curvature I have noted.
Higgins [67] noted that magma mixing would lead to compositional differences
123

etween the different crystal populations (i.e., see Fig. 3.6). Entrainment of early
formed crystals from the magma chamber floor would also lead to compositional
differences related to partial crystallization (see also [19, 20, 96]. For example,
early formed plagioclase crystals would likely contain higher abundances of Sr,
because as plagioclase crystallization progresses the residual melt becomes Sr de-
pleted. I suggest that DSM magmas, particularly those from Units 14 and 31,
incorporated significant amounts of plagioclase during recharge and/or eruption.
What is not is clear is the physical process by which this occurred. Was it via
mixing of magmas with distinct mantle sources? Was it by mixing of genetically
related magmas that had experienced different degrees of partial crystallization?
Was it via accumulation of older genetically related (or unrelated) crystals? I
chose samples from Site 1203 Units 3, 14 and 31 and Site 884 Unit 8 for a micro-
analytical dissection of the crystals that populate distinct segments of non-linear
CSDs to test the hypothesis that plagioclase accumulation affected the bulk com-
positions of these basalts, to explore the possibility of a magma mixing scenario as
an explanation for the non-linear CSDs and textures of DSM pillow basalts, and
to better understand the physical process of plagioclase incorporation by rising
DSM magmas.
3.6.2 Insights from Plagioclase Major Element Compositions
A dichotomy in the An contents of population A and B crystal cores is most
pronounced in the Unit 14 and 31 basalts, where population B crystal cores are
more An-rich than population A cores (Fig. 3.9). The combination of strongly
kinked CSDs (Fig. 3.4e,l) and the distinct core compositions between the two
populations supports the notion that population A and B crystals have separate
124

origins via accumulation or magma mixing. Unit 3 population A and B crystal
cores have overlapping An content, which suggests the two crystal populations
could have formed under similar conditions. There are, however, two groups of
Unit 3 population A crystals in terms of core An content. This grouping may
be an artifact of limited sampling, as the population A crystals from Units 14
and 31 span a similar range of An content. Experimental studies have shown
that in low H2O magmas relatively albite-rich plagioclase crystallizes at elevated
pressure, and growth of An-rich plagioclase is favored in high temperature magmas
at lower pressures [3, 61, 137]. Based strictly upon An content, this suggests that
population B crystals from Units 14 and 31 formed at lower mean pressures in
shallower magma chambers relative to population B crystals from Unit 3 and all
population A crystals.
The lack of variation in Mg# [Mg/(Mg+total Fe)] between crystal populations
is likely related to the poorly constrained manner in which Fe and Mg partition
into plagioclase (i.e., partitioning into more than one site under different condi-
tions; [7, 117]; Fig. 3.9d). The negative correlation of Ti abundance with An
content is undoubtedly related to enhanced Ti partitioning into the more elastic
structure of lower An plagioclase [11] (Fig. 3.9e). However, if each unit is consid-
ered individually, there is minimal correlation of Ti abundance with An content
at An > 70. Based upon this observation, I suggest Unit 3 population B crystals
grew from a magma that was slightly more Ti-rich than parent magmas of popu-
lation B crystals from Units 14 and 31 and much more Ti-rich than Site 884 Unit
8 population B parent magmas.
A working hypothesis based upon CSD results, mineral zoning, and major
element compositions is that the Site 884 Unit 8 and 10 magmas and Site 1203
125

Unit 14 and 31 magmas accumulated older plagioclase crystals. Based strictly
upon CSD results, petrography, plagioclase An variations, and the conclusions of
Huang et al. [72], I suggest that accumulation was the dominant mechanism for
generation of non-linear CSDs. Magmas ascending from depth accumulated crys-
tals from the margins of shallow magma chambers and/or conduits. This magma
experienced minor partial crystallization in a a deeper magma chamber and/or
partial crystallization related to cooling during ascent to form the population A
crystals. As the population B crystals were stirred up and accumulated they were
partially resorbed, which is consistent with my petrographic observations and the
whole-rock results reported by Huang et al. [72]. The Unit 3 magma experienced
minor plagioclase accumulation that was accentuated by flow sorting or plagio-
clase flotation. Unit 3 population A and B crystals appear to share a common
parent magma. At least two different parent magma compositions are recorded in
Units 14 and 31 each, which I suggest were from the same source but had expe-
rienced different degrees of plagioclase-dominated partial (fractional, equilibrium,
or in-situ) crystallization. My limited sampling of Site 884 Unit 8 reflects only one
unique parent magma composition (i.e., that of the population B crystals). Mea-
sured trace element abundances and inferred parent magma compositions allow
us to test this working hypothesis and constrain the parent magma compositions
of each crystal population as well as assess potential differences in source affinity
for each parent magma.
126

3.6.3 Crustal Magma Evolution: Insights from Trace Elements
3.6.3.1 Crystal Population Origins
I use Y as a proxy for the heavy REE (HREE), because Y 3+ has a ionic ra-
dius intermediate between Dy 3+ and Ho 3+ , and REE heavier than Eu are highly
incompatible in plagioclase and thus at or below LA-ICP-MS analytical detection
limits. Ratios of LREE such as La or Ce to Y provide insight into OIB (high
La/Y) vs. MORB (low La/Y) source affinities (see Figs. 3.13, 3.12). Trivalent
La and Ce 3+ have the largest ionic radii of the REE and are the two least com-
patible REE in a crystallizing basaltic assemblage (e.g., plagioclase + olivine ±
clinopyroxene). Lanthanum and Ce are difficult to fractionate from one another by
partial crystallization, and in this sense they provide insight into mantle source
affinity. If Unit 3 population A and B crystals indeed share a common parent
magma, as indicated by their major element similarities, then parent magmas of
these populations should have similar La/Ce and La/Y ratios.
Parent magmas of Unit 3 population A and B crystal cores and Unit 3 whole-
rock data reported by Huang et al. [72] have broadly similar Sr abundances,
La/Ce and La/Y ratios (Figs. 3.12a, 3.13), which supports my initial hypothesis
that population A and B crystals had compositionally similar parent magmas. The
extension of Unit 3 population B parent magmas to slightly higher Sr abundance
can be attributed to crystallization of these older crystals from less fractionated
magmas with a source shared with the population A crystals. The gentle upward
curvature of the plagioclase CSD from the top of Unit 3 supports this notion, as it
provides evidence of minor accumulation of larger (i.e., older) crystals. Accumu-
lation and partial resorption of crystalline debris (i.e., large rounded population
B crystals) was volumetrically minor, such that the resultant bulk-rock compo-
127

sition is similar to parent magmas of large and small crystals. There are three
population A crystal core parent magmas that have notably higher La/Y ratios
(but similar La/Ce to other population A and B parent magmas) that plot outside
Site 1203 tholeiitic basalt field and within a field defined by Mauna Kea shield
stage basalts (Fig. 3.13). Elevated La/Y in a residual magma can be generated
in a region where partial crystallization was dominated by clinopyroxene due to
the greater compatibility of Y relative to La in clinopyroxene. Based upon trace
element data alone it is difficult to determine whether these magmas represent
Hawaii-like end-member magma compositions or whether they came from some
isolated environment where clinopyroxene crystallization dominated. I explore the
latter possibility in the next section.
Unit 14 population A parent magmas consistently range to lower La/Ce, lower
La/Y, and lower Sr, which is consistent with derivation from a more depleted
source than population B magmas (Figs. 3.12b, 3.13). This compositional divi-
sion between population A and B parent magmas is not surprising considering
the kinked nature of the Unit 14 plagioclase CSD relative to the gentle upward
curvature of the Unit 3 CSD. If differences in trace element abundances between
Unit 14 population A and B parent magmas were strictly related to partial crys-
tallization, the lower Sr abundances of parent magmas of later formed crystals
would be expected but not coupled with lower La/Ce and La/Y. I do point out,
however, that while the Sr abundance in the low La/Ce - low La/Y population
A parent magmas is slightly greater than EPR N-MORB and Garrett Transform-
type basalts that were discussed by Huang et al. [72], the low La/Ce and La/Y
ratios are consistent with a relatively depleted mantle source.
On a plot of La/Ce vs. Sr, Unit 14 whole-rock compositions plot between low
128

La/Ce
La/Ce
La/Ce
O
X
Sr (ppm)
Figure 3.12. a) There is little distinction between Unit 3 crystal populations
and Unit 3 whole-rock La/Ce ratios. Population B crystals extend to higher
Sr and population A crystals to lower Sr, which is likely related to degree of
fractionation each parent magma experienced. b) Unit 14 population A and
B parent magmas are distinct from one another. Population A parent
magmas trend to distinctly lower La/Ce, which is consistent with derivation
from a relatively depleted source. A 50-50 mixture of the end-member
compositions shown as boxed X’s yields a hybrid shown as a black star. A
60-40 mixture of the boxed circle and boxed X (higher La/Ce and Sr) yields
a hybrid magma shown as an open star. This simple mixing model can
explain the intermediate bulk-rock compositions of Unit 14 basalts shown as
black triangles (bulk-rock data from [72]). c) Unit 31 Population B parent
fall within a narrow range of La/Ce and extend to higher Sr than the single
population A parent magma reported or Unit 31 whole-rock data [72].
129
A)
B)
C)
X

Sr - low La/Ce population A parent magmas and high La/Ce - high Sr population
B parent magmas, which is consistent with magma mixing (Fig. 3.12b). A simple
binary 50-50 mixture of two compositional end-members, each shown as a boxed
in X in figure 3.12b, yields a bulk-rock composition (shown as a black star in
figure 3.12b) similar to Unit 14 bulk-rock compositions reported by Huang et al.
[72]. The Unit 14 basalt sample I examined has 20% modal abundance plagioclase,
where ∼ 73% of the plagioclase is from population A and ∼ 27% of the plagioclase
is from population B. From this observation it seems unlikely that the bulk-rock
is the a 50-50 mixture. The range of population A parent magma La/Ce indicates
that not all of the population A crystals had similarly depleted parent magmas.
If I represent population A parent magmas as a single hybrid magma with the
La/Ce and Sr characteristics depicted in figure 3.12b as a boxed in circle, 60% of
this magma mixed with 40% of the same population B end-member used previ-
ously results in a bulk-rock composition shown by the open star in figure 3.12b.
Although this may be an excessivley simple model, it does better reconcile the
population A and B modal abundance data with the compositional data.
Based upon the simple model above, I suggest that the Unit 14 basalt is a
mixture of at least two end-member magmas, one slightly more depleted than the
other. I cannot, however, exclude a role for plagioclase accumulation and partial
resorption as was suggested by Huang et al. [72]. Marsh [98] suggested accumu-
lation and partial resorption of crystalline debris is a more common process than
generally recognized and certainly may have had a role during the genesis of the
magma mixing end-members. Indeed, the rounded nature of Unit population B
crystals indicates partial resorption of plagioclase occurred (Fig.3.7). It is appar-
ent from the large compositional range of the Unit 14 population A crystals that
130

division of the plagioclase crystals into two populations in this rock may be an
oversimplification. It also may suggest that crystals and crystalline debris was
recycled throughout the DSM magmatic system (e.g., [37, 98]).
Figure 3.13. The majority of population A and B parent magmas are within
analytical uncertainty of overlapping the field defined by tholeiitic basalts
from Sites 1203 and 884 Fields (data from [72, 119]). Select samples extend
to higher La/Y and plot within the field defined by Mauna Kea shield stage
basalts. Elevated La/Y can also be explained by derivation of the crystal
from an environment where late stage clinopyroxene crystallization was
significant, which would elevate La/Y (i.e., D La
cpx

The Unit 31 plagioclase CSD is kinked in manner similar to the Unit 14 basalt,
which may indicate another instance of magma mixing. I successfully measured
trace elements in only one Unit 31 population A crystal, which has lower Sr,
similar La/Ce and La/Y in its parent magma relative to Unit 31 population B
parent magmas. It is difficult to test this hypothesis due to a lack of population
A compositional data. Unit 31 whole-rock data has similar La/Ce and La/Y as
the population B parent magmas but is offset to lower Sr, which indicates that
they were derived from a similar mantle source (Fig. 3.12c). I suggest plagioclase
accumulation was an important process during the petrogenesis of the Unit 31
basalt, although I cannot rule out magma mixing as a process at least partially
responsible for the final texture and composition of this basalt.
The two crystals I examined from Site 884 Unit 8 have distinctly lower La/Y
ratios and Sr abundances in the calculated parent magmas than Site 1203 basalts
(Fig. 3.13). The core of the two population B crystals I examined have similar
La/Y ratios and greater Sr relative to bulk-rock samples reported by Regelous
et al. [119]. This is consistent with accumulation scenario, as was suggested for
Site 1203 Units 3 and 31, where the population B crystals were derived from
a magma similar to that which produced the population A crystals, which was
probably related to that represented by the bulk-rock composition. A depleted
source component is most apparent in Site 884 basalts, and may be best recorded
in the early formed population B crystals that clearly crystallized from a magma
less modified by fractional crystallization than the bulk rock (i.e., higher Sr).
132

3.6.4 The Crystal Record of Highly Evolved Magmas
Upon magma emplacement and the onset of conductive cooling mushy bound-
ary layers form along magma chamber margins and floors. Boundary mush layers
are rheological elements of multiply saturated solidification fronts defined as hav-
ing crystallinities between 25% and 50%-55% [98]. Crystallization within the so-
lidification front in energetically favorable, and in this relatively stagnant environ-
ment interstitial melts can evolve rapidly and significantly [84, 98]. Crystallization
within the mush zone produces interstitial melts that are evolved relative to the
magma in the chamber interior, and continued crystallization traps these evolved
melts within the solidification front [84, 98]. Mineral phases with compositions out
of equilibrium with main magma body can grow deep within solidification fronts
where they are thermally and mechanically insulated from the hot magma cham-
ber interior [84]. Capture of crystals by advancing solidification fronts is greatest
along magma chamber floors due to gravitational settling, and the thickness of
these cumulates in any given magma chamber may become quite significant if re-
peated input and storage of crystal carrying magmas occurs from deeper magma
chambers [98].
Crystal debris may be mobilized from mushy cumulate layers and solidification
fronts by erosion, density and convection driven plumes, or stirring by new magma
input [83, 98]. If the introduction of mushy crystal debris and evolved interstitial
liquid to the main magma body is small relative to the size of the magma body,
the effect on bulk magma compositions would be negligible. Crystals that survive
this movement of mushy debris into the main magma body thus may provide the
only evidence of this process. I suggest plagioclase crystals with elevated parent
magma La/Y and lower Sr grew deep within mush zones where crystallization
133

was dominated by late stage clinopyroxene crystallization. I suggest there was
minimal clinopyroxene crystallization outside of the mush layer interstitial spaces.
3.6.5 Ba and Sr variations: Evidence for Polybaric Crystallization?
Plagioclase parent magmas of the three Site 1203 basalts and the single Site
884 basalt I examined plot within fields defined by Detroit Seamount basalts in
Sr vs. Ba space (Fig. 3.14). This indicates there was not significant involvement
of extreme magma compositions such as the low Sr type magmas from Garrett
Transform fault (c.f., [72]). A number of linear trends are apparent in figure 3.14
that can be explained by simple fractional crystallization of olivine and plagio-
clase, which Huang et al. [72] suggested were the dominant crystallizing phases
in DSM tholeiitic basalts. The slope of each fractionation trend is related to the
proportions of plagioclase and olivine during crystallization. The steepest trends
correspond to environments where partial crystallization was dominated by olivine
(Fig. 3.14b, and the shallowest trends in figure 3.14 correspond to environments
where partial crystallization was dominated by plagioclase. Hash marks on the
model lines in figure 3.14b mark 10% increments of fractional crystallization. The
three steepest trends (i.e., olivine dominated) in figure 3.14 consist of Unit 3, 14,
and 31 population B parent magmas, although the majority of Unit 31 population
B parent magmas exhibit a single shallower plagioclase dominated fractionation
trend.
Increased pressure increases Al2O3 in magmas along the olivine gabbro cotectic
[66]. A physical manifestation of this relationship is a greater proportion plagio-
clase in the crystallizing assemblage at increased pressure. Experimental results
have demonstrated that relatively albite-rich plagioclase crystallizes at elevated
134

Sr (ppm)
Sr (ppm)
A)
B)
Site 1203
tholeiitic basalts
Site 884
tholeiitic basalts
*
*
*
*
Ba (ppm)
*
*
Site 883 and Site 1204
alkalic basalts
* PL:OL; Ba, Sr
PL(%):OL(%); Ba (ppm at start), Sr (ppm at start)
Figure 3.14. a) Population A and B parent magmas have Sr and Ba
abundance that, by majority, overlap the field defined tholeiitic basalts from
Sites 1203 and 884 (both whole-rock and glass sample data from [72, 119].
Several linear trends are apparent, which are consistent with progressive Sr
and Ba enrichment related to partial crystallization that was dominated by
olivine and plagioclase. b) The steepest trends correspond to environments
where crystallization was dominated by olivine, which is favored at lower
mean pressures. The steep trends are consist primarily of population B
parent magmas. The starting composition of each modeled trend is shown as
a black star. The shallower trends that consiste largely of population A
parent magmas correspond to environments where crystallization was
dominated by plagioclase. Higher pressure increases Al2O3 along the olivine
gabbro cotectic, which stabilizes greater plagioclase fractions at greater
pressure [66]. From this I suggest that polybaric partial crystallization was
significant during petrogenesis of Detroit Seamount lavas.
135
*
*

pressure, and growth of An-rich plagioclase is favored in high temperature mag-
mas at lower mean pressure (e.g., [3, 61, 137]). Population B crystals are An-rich
and have parent magmas that make up the steepest fractionation trends in Sr vs.
Ba space, whereas population A parent magmas constitute the shallowest trends
(Fig. 3.14). From this I suggest that population B crystals formed in shallow
magma chambers than population B crystals and were added to magmas carrying
population A crystals either by accumulation (i.e., Site 1203 Units 3 and 31; Site
884 Unit 8) or magma mixing (i.e., Unit 14). Each of the Units I examined from
Site 1203 exhibit evidence of partial crystallization in at least two distinct envi-
ronments, which can be easily explained by polybaric crystallization. Although
my modeling considered only olivine and plagioclase fractionation, small amounts
of clinopyroxene fractionation would not have significantly affected the slopes of
the fractionation trends, as Sr is incompatible and Ba is highly incompatible in
clinopyroxene.
3.7 Summary and Conclusions: Insights Into the Petrogenesis of Depleted De-
troit Seamount Basalts
There is an undeniable role for plagioclase fractionation and accumulation in
the petrogenesis of Detroit Seamount tholeiitic basalts. I see no evidence for a
similar role during the genesis of alkalic basalts (i.e., Site 1203 alkalic basalts
shown in Fig. 13 of Huang et al. [72]). I conclude that the Detroit Seamount
magma chamber system was complex and consisted of extensive mush zones and
multiple interconnected chambers. Evidence of mixing between relatively depleted
and enriched end-member magmas recorded in the hybrid Unit 14 basalt do not
contradict existing models seeking to explain the origin of the depleted component
136

in Detroit Seamount basalts. Unit 14 magma mixing does suggest, however, that
subtle variations in partial melting or source compositions do occur over relatively
short time spans but may be easily overlooked in whole-rock compositional studies.
I suggest this variation is consistent with subtle variations in partial melting of
a heterogeneous mantle source that consisted of relatively depleted and relatively
enriched components (cf. [50, 72, 119]) Subtle variations in partial melting are
consistent with enhanced mantle flow and complex mantle dynamics in the near
ridge hotspot environment as discussed by Pearce [114]. The nature of the depleted
component in Detroit Seamount basalts is prominent in tholeiitic basalts from Site
884, and perhaps most prominent in the cm-size plagioclase phenocrysts. These
large crystals have low Sr parent magmas suggesting they were less evolved than
their host basalt. A fruitful continuation of this research would be to conduct Sr
and Pb isotope studies on these large crystals by in-situ microdrilling (e.g., [39])
or plagioclase separation (e.g., [16]).
137

138
TABLE 3.3
PLAGIOCLASE PHENOCRYST MAJOR AND TRACE ELEMENT
DATA
Sample Zone 1 2
Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
3
Site 1203 Unit 3
A C B 80 0.63 0.15 0.03 0.06 2.20 15.91 0.19 47.15 33.81 100.13 0.19 275.2 – 393 0.14 10.2 0.28 0.76 0.33 0.05 0.23
A I B 74 0.67 0.15 0.06 0.06 2.88 15.04 0.19 48.82 32.75 100.62 0.18 – – 378 0.15 10.4 0.28 0.77 0.26 0.08 0.28
A I B 75 0.67 0.13 0.04 0.06 2.72 15.08 0.18 48.62 32.87 100.37 0.18 269.2 – 369 0.14 9.89 0.28 0.75 0.29 0.09 0.25
A I B 80 0.71 0.13 0.02 0.01 2.20 16.16 0.21 46.89 33.47 99.80 0.18 277.0 – 391 0.10 9.92 0.29 0.72 0.30 0.05 0.25
A I B 80 0.62 0.15 0.03 0.04 2.25 16.05 0.17 46.57 33.33 99.22 0.18 288.4 – 395 0.13 10.4 0.32 0.72 0.26 0.08 0.25
A I B 80 0.64 0.15 0.02 0.07 2.20 16.03 0.18 46.44 33.68 99.41 0.18 293.2 – 400 0.14 10.5 0.31 0.77 0.33 0.08 0.25
A I B 79 0.66 0.16 0.04 0.04 2.27 15.51 0.17 47.69 32.97 99.50 0.17 299.2 – 369 0.11 9.54 0.27 0.71 0.27 0.07 0.22
A I B 78 0.64 0.16 0.03 0.03 2.34 15.60 0.18 46.43 34.05 99.46 0.18 277.0 – 360 0.11 9.43 0.28 0.68 0.25 0.03 0.23
A I B 78 0.65 0.16 0.02 0.05 2.43 15.61 0.19 47.38 32.91 99.41 0.19 – – – – – – – – – –
A R B 70 0.77 0.13 0.05 0.05 3.20 13.91 0.22 49.15 32.06 99.55 0.18 – – – – – – – – – –
B C B 79 0.64 0.16 0.03 0.03 2.29 15.53 0.18 47.93 32.61 99.40 0.18 257.2 – 355 0.10 9.23 0.28 0.63 0.25 0.09 0.24
continued...

139
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
B I B 80 0.64 0.14 0.03 0.04 2.16 15.79 0.17 48.03 32.65 99.66 0.17 268.0 – 362 0.11 9.37 0.27 0.70 0.27 0.05 0.26
B I B 80 0.64 0.14 0.03 0.04 2.19 15.68 0.18 48.44 32.77 100.12 0.17 257.8 – 358 0.11 9.28 0.27 0.72 0.29 0.05 0.25
B I B 78 0.67 0.15 0.04 0.05 2.38 15.19 0.20 48.94 32.19 99.80 0.19 271.0 – 343 0.12 9.17 0.27 0.67 0.23 0.06 0.19
B R B 73 0.68 0.13 0.05 0.06 2.75 13.84 0.21 49.87 31.52 99.12 0.19 – – – – – – – – – –
C C B 81 0.65 0.12 0.04 0.03 2.09 16.17 0.18 47.29 33.86 100.44 0.18 296.2 – 366 0.11 9.14 0.29 0.70 0.35 0.04 0.25
C C B 80 0.62 0.15 0.03 0.00 2.22 15.77 0.18 47.21 33.46 99.65 0.19 321.3 – 363 0.10 9.61 0.29 0.69 0.25 0.06 0.22
C I B 79 0.61 0.13 0.03 0.03 2.26 15.79 0.17 46.96 33.34 99.33 0.18 314.7 – 362 0.10 9.41 0.29 0.67 0.23 0.05 0.23
C I B 77 0.65 0.11 0.05 0.04 2.45 15.43 0.19 47.78 32.86 99.56 0.18 321.9 – 360 0.13 9.47 0.29 0.71 0.25 0.08 0.25
C I B 71 0.67 0.11 0.06 0.06 3.20 14.12 0.23 49.34 31.70 99.48 0.21 375.3 – 336 0.09 9.85 0.24 0.63 0.24 0.06 0.21
C I B 78 0.66 0.11 0.04 0.03 2.43 15.33 0.18 47.66 32.61 99.05 0.17 398.7 – 373 0.11 10.7 0.31 0.64 0.33 0.06 0.25
C I B 77 0.66 0.11 0.05 0.05 2.51 14.98 0.19 48.20 32.47 99.21 0.18 343.5 – 353 0.08 9.8 0.25 0.71 0.25 0.05 0.22
C I B 78 0.62 0.12 0.04 0.04 2.36 15.37 0.18 47.93 32.54 99.20 0.19 313.5 – 354 0.14 9.34 0.32 0.68 0.27 0.06 0.24
C R B 69 0.67 0.11 0.04 0.05 3.37 13.67 0.22 49.62 31.29 99.04 0.2 410.1 – 334 0.11 10.3 0.25 0.68 0.24 0.06 0.25
continued...

140
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
D-L C B 80 0.69 0.16 0.03 0.03 2.30 16.45 0.18 48.49 33.04 101.37 0.17 272.2 – 373 0.12 9.84 0.30 0.67 0.29 0.04 0.26
D-L C B 81 0.66 0.15 0.04 0.04 2.12 16.40 0.19 48.51 33.31 101.43 0.18 – – – – – – – – – –
D-L I B 79 0.68 0.16 0.03 0.03 2.32 15.99 0.18 48.83 33.06 101.28 0.17 289.0 – 371 0.09 9.62 0.29 0.75 0.32 0.06 0.22
D-L I B 78 0.65 0.15 0.04 0.04 2.33 15.50 0.18 48.39 32.07 99.36 0.18 286.6 – 364 0.13 9.74 0.28 0.72 0.29 0.07 0.21
D-L I B 79 0.66 0.14 0.03 0.05 2.32 15.74 0.18 48.84 32.49 100.46 0.17 291.4 – 368 0.09 9.84 0.31 0.73 0.35 0.06 0.21
D-L R B 73 0.72 0.13 0.05 0.06 2.83 13.85 0.21 51.15 30.48 99.49 0.19 – – – – – – – – – –
D-S1 C A 75 0.67 0.14 0.05 0.02 2.71 15.05 0.22 48.86 32.57 100.30 0.2 – – – – – – – – – –
D-S1 C A 76 0.68 0.11 0.04 0.01 2.65 14.98 0.20 48.91 32.47 100.05 0.18 321.9 – 367 0.06 10.6 0.28 0.68 0.24 0.12 0.20
D-S1 I A 74 0.73 0.12 0.05 0.03 2.86 14.97 0.19 49.67 32.24 100.86 0.17 290.2 – 355 0.04 9.96 0.39 0.59 0.36 0.16 0.20
D-S1 I A 75 0.62 0.15 0.03 0.05 2.70 14.84 0.17 49.53 32.35 100.45 0.18 308.2 – 373 0.11 10.6 0.28 0.63 0.22 0.13 0.31
D-S1 I A 79 0.67 0.15 0.05 0.03 2.32 15.63 0.15 48.64 32.67 100.31 0.15 271.6 – 369 0.11 9.7 0.26 0.66 0.34 0.09 0.23
D-S1 I A 78 0.68 0.14 0.05 0.03 2.42 15.68 0.18 47.76 32.82 99.75 0.17 – – – – – – – – – –
D-S1 R A 68 0.73 0.12 0.06 0.07 3.48 13.81 0.21 49.76 30.92 99.17 0.18 402.3 – 372 0.07 13.2 0.33 0.87 0.36 0.08 0.22
continued...

141
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
D-S1 R A 66 0.75 0.13 0.07 0.08 3.75 13.12 0.22 50.88 29.97 98.97 0.18 446.0 – 364 0.14 12.8 0.34 0.63 0.24 0.08 0.24
D-S2 C A 78 0.61 0.12 0.03 0.04 2.34 15.28 0.19 48.17 32.00 98.79 0.19 272.2 – 363 0.14 9.99 0.22 0.75 0.26 0.08 0.18
D-S2 I A 73 0.67 0.12 0.04 0.04 2.98 14.49 0.22 49.32 31.55 99.43 0.2 344.7 – 365 0.04 9.71 0.30 0.68 0.09 0.13 0.32
D-S2 I A 78 0.66 0.12 0.03 0.06 2.43 15.33 0.20 48.36 32.18 99.39 0.19 323.7 – 377 0.10 10.1 0.26 0.70 0.23 0.09 0.25
D-S2 I A 79 0.71 0.13 0.04 0.07 2.26 15.37 0.18 47.79 32.25 98.79 0.16 297.4 – 363 0.08 9.48 0.25 0.81 0.28 0.12 0.25
D-S2 R A 62 0.90 0.12 0.08 0.09 4.05 12.09 0.20 52.54 28.76 98.83 0.15 530.0 – 386 0.12 22.8 0.32 0.84 0.29 0.09 0.26
E-1 I A 80 0.63 0.14 0.03 0.04 2.17 15.79 0.17 48.58 32.75 100.30 0.17 270.4 – 367 0.10 9.46 0.32 0.71 0.32 0.04 0.24
E-1 I A 80 0.65 0.15 0.03 0.04 2.16 15.68 0.18 48.02 32.40 99.31 0.18 263.8 – 359 0.09 9.3 0.24 0.68 0.28 0.06 0.22
E-1 I A 78 0.62 0.16 0.03 0.03 2.35 15.19 0.19 48.37 32.01 98.95 0.2 – – – – – – – – – –
E-1 I A 81 0.67 0.18 0.03 0.04 2.07 15.82 0.17 47.62 32.80 99.40 0.17 267.4 – 358 0.10 9.47 0.29 0.72 0.22 0.05 0.20
E-1 I A 80 0.63 0.15 0.03 0.05 2.09 15.75 0.17 47.69 32.68 99.23 0.17 288.4 – 379 0.11 9.7 0.28 0.72 0.32 0.06 0.26
E-1 R A 69 0.70 0.15 0.07 0.07 3.52 14.03 0.22 51.15 31.00 100.90 0.2 – – – – – – – – – –
F C A 80 0.70 0.12 0.02 0.01 2.15 15.79 0.18 48.11 33.69 100.76 0.16 280.0 – 358 0.09 8.75 0.31 0.69 0.28 0.04 0.21
continued...

142
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
F I A 81 0.63 0.15 0.02 0.04 2.13 16.09 0.17 47.78 33.43 100.44 0.17 285.4 – 367 0.10 8.91 0.34 0.66 0.25 0.05 0.24
F I A 73 0.63 0.13 0.04 0.03 2.97 14.35 0.23 49.71 31.98 100.06 0.22 305.2 – 342 0.13 9.64 0.30 0.57 0.32 0.05 0.23
F I A 80 0.64 0.15 0.02 0.07 2.15 15.93 0.16 47.48 33.83 100.41 0.16 266.2 – 363 0.12 8.79 0.32 0.68 0.31 0.07 0.24
F I A 78 0.61 0.14 0.03 0.08 2.38 15.53 0.19 47.94 33.22 100.12 0.19 269.2 – 355 0.11 8.97 0.30 0.66 0.31 0.07 0.21
F I A 79 0.64 0.11 0.02 0.02 2.34 15.56 0.18 48.36 33.22 100.45 0.18 304.0 – 367 0.14 9.81 0.27 0.71 0.31 0.05 0.25
F R A 64 0.77 0.10 0.06 0.08 3.97 12.72 0.22 52.29 30.22 100.43 0.18 565.9 – 428 0.16 26.3 0.43 1.02 0.33 0.06 0.35
H I A 77 0.64 0.14 0.04 0.04 2.50 15.08 0.20 48.88 32.21 99.73 0.2 285.4 – 362 0.09 9.89 0.28 0.70 0.29 0.07 0.22
H I A 78 0.62 0.12 0.03 0.02 2.34 15.48 0.19 47.60 32.63 99.04 0.19 276.4 – 360 0.16 9.29 0.28 0.73 0.28 0.06 0.25
H I A 81 0.64 0.15 0.05 0.03 2.04 15.64 0.18 48.76 32.88 100.36 0.18 285.4 – 369 0.11 9.87 0.25 0.66 0.26 0.08 0.23
H I A 80 0.60 0.18 0.03 0.03 2.16 15.78 0.18 47.92 33.04 99.91 0.19 245.8 – 364 0.13 9.13 0.29 0.68 0.28 0.04 0.25
H R A 64 0.70 0.12 0.06 0.07 3.79 12.57 0.21 50.58 29.72 97.82 0.19 – – – – – – – – – –
J C B 78 0.65 0.12 0.04 0.04 2.37 15.18 0.18 47.27 33.38 99.25 0.18 293.8 – 353 0.11 8.6 0.31 0.63 0.25 0.04 0.21
J I B 75 0.62 0.11 0.05 0.04 2.77 14.85 0.19 48.37 32.71 99.70 0.19 296.2 – 346 0.11 8.55 0.32 0.63 0.26 0.10 0.22
continued...

143
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
J I B 82 0.60 0.12 0.02 0.03 1.95 15.87 0.16 46.59 33.67 99.01 0.17 234.4 – 351 0.11 7.7 0.29 0.67 0.28 0.07 0.22
J I B 78 0.59 0.12 0.03 0.01 2.32 15.31 0.16 48.34 32.68 99.56 0.17 283.0 – 351 0.11 8.82 0.28 0.59 0.27 0.05 0.24
J I B 77 0.61 0.12 0.04 0.04 2.43 14.68 0.18 47.98 32.94 99.02 0.18 277.6 – 340 0.12 8.84 0.26 0.67 0.27 0.07 0.23
J R B 61 0.84 0.10 0.08 0.09 4.34 12.58 0.18 53.41 28.05 99.67 0.14 – – – – – – – – – –
PL-A C A 64 0.85 0.08 0.07 0.11 4.07 13.03 0.22 51.35 29.89 99.68 0.17 – – – – – – – – – –
PL-B C A 62 0.86 0.10 0.08 0.09 4.23 12.45 0.19 52.17 29.65 99.81 0.15 – – – – – – – – – –
PL-C C A 65 0.86 0.12 0.06 0.10 3.96 13.31 0.21 52.31 29.95 100.88 0.16 – – – – – – – – – –
PL-C R A 59 0.86 0.11 0.18 0.09 4.47 12.04 0.19 53.20 29.63 100.78 0.14 – – – – – – – – – –
PL-D C A 62 0.85 0.10 0.08 0.06 4.17 12.51 0.21 52.48 30.04 100.50 0.16 449.0 – 381 0.16 17.6 0.31 0.76 0.28 0.09 0.27
PL-E C A 64 0.85 0.08 0.06 0.09 3.99 13.13 0.20 51.43 30.37 100.22 0.15 – – – – – – – – – –
PL-G C A 64 0.68 0.10 0.07 0.09 3.98 12.72 0.25 50.57 30.60 99.06 0.22 398.1 – 354 0.10 14.8 0.44 0.74 0.30 0.10 0.26
PL-G R A 66 0.64 0.10 0.05 0.04 3.76 13.60 0.21 50.67 31.03 100.11 0.2 388.5 – 352 0.07 12.7 0.27 0.69 0.30 0.08 0.27
PL-F I B 80 0.71 0.15 0.04 0.03 2.13 16.01 0.17 48.00 33.49 100.72 0.16 225.4 – 355 0.13 8.64 0.24 0.67 0.25 0.12 0.23
continued...

144
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
PL-F C B 74 0.61 0.12 0.04 0.07 2.93 15.12 0.20 49.83 32.19 101.11 0.2 – – – – – – – – – –
PL-F R B 66 0.83 0.10 0.08 0.03 3.90 13.63 0.22 51.73 31.01 101.52 0.17 – – – – – – – – – –
PL-H C A 63 0.78 0.08 0.06 0.08 4.16 12.77 0.24 52.15 29.83 100.17 0.19 489.2 – 405 0.09 22.2 0.41 0.93 0.33 0.09 0.42
X1 C B 81 0.67 0.21 0.03 0.01 1.98 15.54 0.11 47.20 33.93 99.68 0.12 – – – – – – – – – –
X1 I B 75 0.70 0.17 0.05 0.05 2.66 14.35 0.14 49.30 32.60 100.03 0.14 – – – – – – – – – –
X1 R B 65 0.73 0.13 0.07 0.06 3.72 12.77 0.21 51.81 30.72 100.22 0.18 – – – – – – – – – –
X4 C B 79 0.58 0.17 0.04 0.04 2.16 14.75 0.17 49.12 32.84 99.88 0.18 – – – – – – – – – –
X4 I B 77 0.60 0.17 0.04 0.04 2.34 14.46 0.17 48.98 32.93 99.73 0.18 – – – – – – – – – –
X4 R B 68 0.69 0.14 0.07 0.05 3.39 13.10 0.20 52.05 30.45 100.14 0.18 – – – – – – – – – –
X2-1 C B 79 0.72 0.16 0.03 0.04 2.26 15.47 0.13 48.98 32.41 100.21 0.12 – – – – – – – – – –
X2-1 C B 79 0.71 0.16 0.04 0.04 2.26 15.42 0.13 49.19 32.32 100.28 0.13 – – – – – – – – – –
X2-1 I B 70 0.72 0.15 0.05 0.05 3.11 13.53 0.16 51.99 30.35 100.12 0.15 – – – – – – – – – –
X2-1 I B 77 0.75 0.17 0.05 0.05 2.40 14.81 0.14 50.03 31.82 100.22 0.13 – – – – – – – – – –
continued...

145
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
X2-1 R B 71 0.75 0.15 0.06 0.05 3.05 13.56 0.16 51.92 30.55 100.25 0.14 – – – – – – – – – –
Site 1203 Unit 14
A I B 87 0.55 0.16 0.03 0.03 1.43 17.54 0.14 45.00 34.94 99.82 0.16 193.0 – 394 0.16 7.3 0.32 0.69 0.34 0.06 0.25
A C B 82 0.60 0.15 0.04 0.02 2.03 16.62 0.17 47.23 34.06 100.93 0.18 – – – – – – – – – –
A I B 83 0.63 0.13 0.03 0.00 1.90 16.87 0.19 46.33 33.43 99.51 0.19 249.4 – 403 0.12 7.9 0.32 0.59 0.29 0.09 0.24
A I B 86 0.53 0.14 0.02 0.04 1.49 17.21 0.13 45.84 34.56 99.96 0.16 227.2 – 387 0.14 7.2 0.32 0.69 0.33 0.06 0.22
A I B 83 0.64 0.13 0.01 0.01 1.84 16.54 0.16 47.23 33.38 99.95 0.17 226.6 – 380 0.11 7.2 0.28 0.64 0.29 0.05 0.21
A I B 81 0.58 0.16 0.03 0.03 2.08 16.48 0.17 47.43 33.16 100.12 0.18 261.4 – 375 0.11 7.1 0.31 0.55 0.26 0.08 0.22
A I B 82 0.59 0.18 0.01 0.03 2.02 16.44 0.19 47.54 33.62 100.63 0.2 275.8 – 376 0.09 7.4 0.26 0.63 0.31 0.09 0.25
A R B 73 0.59 0.15 0.04 0.09 3.04 14.86 0.21 48.99 31.90 99.86 0.22 245.2 – 333 0.10 6.6 0.26 0.53 0.25 0.05 0.21
A R B 78 0.72 0.13 0.02 0.04 2.41 15.74 0.21 48.37 32.76 100.40 0.18 – – – – – – – – – –
A R B 67 0.81 0.17 0.06 0.00 3.77 13.86 0.29 51.11 30.39 100.47 0.22 – – – – – – – – – –
continued...

146
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
B C B 81 0.54 0.13 0.04 0.00 2.11 16.14 0.18 47.67 32.91 99.72 0.21 176.3 – 375 0.07 6.3 0.25 0.43 0.19 0.05 0.21
B I B 82 0.61 0.13 0.04 0.04 2.02 16.79 0.19 47.61 33.21 100.64 0.19 226.6 – 377 0.12 6.5 0.27 0.53 0.25 0.10 0.23
B I B 82 0.56 0.12 0.01 0.06 2.08 17.02 0.16 47.32 33.30 100.63 0.18 241.0 – 402 0.14 6.9 0.30 0.53 0.32 0.04 0.22
B I B 81 0.53 0.19 0.03 0.10 2.16 16.76 0.17 47.20 33.60 100.74 0.2 225.4 – 397 0.13 7.3 0.29 0.56 0.27 0.04 0.23
B I B 83 0.61 0.17 0.02 0.03 1.93 17.23 0.14 46.99 33.69 100.82 0.15 245.2 – 396 0.16 7.2 0.36 0.70 0.34 0.10 0.24
B I B 83 0.57 0.18 0.02 0.04 1.92 17.12 0.17 47.49 33.72 101.23 0.19 254.8 – 418 0.16 8.4 0.34 0.63 0.34 0.03 0.25
C1 C B 83 0.57 0.15 0.03 0.03 1.93 16.66 0.17 46.33 33.35 99.22 0.18 260.2 – 357 0.11 6.4 0.30 0.57 0.26 0.07 0.19
C1 I B 83 0.52 0.16 0.05 0.02 1.83 16.70 0.17 45.99 33.80 99.24 0.2 251.8 – 360 0.13 6.2 0.27 0.49 0.27 0.04 0.23
C1 I B 83 0.61 0.16 0.02 0.00 1.90 16.80 0.18 46.91 33.73 100.30 0.19 251.2 – 362 0.14 6.1 0.28 0.57 0.29 0.05 0.17
C1 I B 81 0.60 0.17 0.02 0.03 2.11 16.80 0.17 46.17 33.49 99.57 0.18 278.8 – 375 0.14 7.2 0.32 0.55 0.29 0.06 0.28
D1 C B 83 0.63 0.16 0.01 0.01 1.89 16.58 0.15 46.03 33.56 99.02 0.16 220.6 – 381 0.12 6.1 0.26 0.49 0.26 0.06 0.20
D1 I B 81 0.61 0.20 0.04 0.01 2.14 16.73 0.19 46.92 33.25 100.09 0.2 202.6 – 374 0.12 6.3 0.24 0.48 0.26 0.07 0.22
D1 I B 83 0.58 0.15 0.02 0.04 1.86 16.89 0.15 46.75 33.73 100.15 0.16 226.6 – 388 0.11 6.5 0.25 0.49 0.26 0.08 0.24
continued...

147
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
D1 I B 83 0.52 0.16 0.03 0.04 1.88 17.19 0.16 46.80 34.19 100.97 0.19 235.6 – 381 0.15 6.8 0.33 0.55 0.28 0.04 0.28
D1 R B 77 0.71 0.19 0.03 0.04 2.63 15.89 0.20 48.25 33.04 100.97 0.18 – – – – – – – – – –
E C B 82 0.58 0.16 0.02 0.04 1.98 16.65 0.19 47.08 32.80 99.50 0.2 271.6 – 390 0.14 7.4 0.26 0.53 0.28 0.06 0.23
E I B 72 0.60 0.12 0.04 0.08 3.03 14.54 0.20 48.76 31.98 99.37 0.21 259.6 – 343 0.11 6.7 0.32 0.56 0.30 0.03 0.18
E R B 73 0.65 0.13 0.05 0.05 3.05 14.82 0.21 49.83 31.48 100.28 0.2 – – – – – – – – – –
E R B 60 1.05 0.10 0.10 0.12 4.62 12.47 0.25 53.18 29.17 101.05 0.15 – – – – – – – – – –
PG1 C A 79 0.80 0.14 0.02 0.08 2.40 16.01 0.19 48.19 32.35 100.18 0.16 – – – – – – – – – –
PG1 R A 67 0.91 0.13 0.05 0.10 3.73 13.56 0.24 51.10 30.18 100.01 0.17 – – – – – – – – – –
PG2 CTR A 72 0.74 0.14 0.04 0.06 3.14 14.39 0.22 49.25 31.30 99.28 0.19 387.3 – 365 0.18 10.2 0.26 0.52 0.45 0.11 0.16
PG3 CTR A 66 0.88 0.13 0.07 0.06 3.75 13.51 0.24 51.56 30.20 100.39 0.17 – – – – – – – – – –
PG4 CTR A 68 0.76 0.14 0.06 0.05 3.61 13.83 0.21 50.70 30.48 99.84 0.18 – – – – – – – – – –
PG5 C A 70 0.77 0.13 0.08 0.04 3.34 14.29 0.20 49.68 31.02 99.55 0.17 – – – – – – – – – –
PG5 R A 66 0.94 0.11 0.06 0.08 3.75 13.59 0.24 50.73 30.70 100.20 0.17 – – – – – – – – – –
continued...

148
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
PG8 C A 68 0.89 0.13 0.08 0.10 3.68 14.03 0.26 50.27 30.53 99.96 0.18 – – – – – – – – – –
PG8 R A 65 1.00 0.14 0.07 0.08 3.94 13.16 0.27 51.50 30.09 100.24 0.17 – – – – – – – – – –
PG9 C A 70 0.73 0.13 0.06 0.06 3.36 14.48 0.21 50.01 31.00 100.04 0.19 370.5 – 376 0.10 11.9 0.27 0.60 0.35 0.14 0.27
PG9 R A 67 0.94 0.10 0.06 0.10 3.75 13.62 0.27 51.01 29.87 99.73 0.18 – – – – – – – – – –
PG10 CTR A 69 0.81 0.12 0.09 0.07 3.57 14.78 0.25 49.48 31.60 100.79 0.19 – – – – – – – – – –
PG11 CTR A 69 0.96 0.15 0.05 0.07 3.60 14.33 0.28 49.33 30.65 99.42 0.18 – – – – – – – – – –
PG12 CTR A 73 0.81 0.15 0.04 0.07 3.10 15.26 0.23 48.69 32.04 100.40 0.18 – – – – – – – – – –
PG13 CTR A 71 0.81 0.13 0.05 0.06 3.36 14.99 0.24 49.14 31.80 100.59 0.18 – – – – – – – – – –
X1 C B 85 0.58 0.28 0.05 0.02 1.64 16.61 0.18 46.71 34.19 100.27 0.2 – – – – – – – – – –
X1 I B 82 0.58 0.19 0.03 0.04 1.98 16.09 0.17 48.02 33.72 100.82 0.19 – – – – – – – – – –
X2 C B 83 0.57 0.21 0.04 0.03 1.85 16.64 0.16 47.94 33.23 100.67 0.18 – – – – – – – – – –
X2 I B 83 0.57 0.20 0.04 0.02 1.84 16.44 0.17 47.91 33.02 100.21 0.19 – – – – – – – – – –
X2 R B 75 0.67 0.18 0.03 0.05 2.79 14.96 0.22 50.52 31.38 100.80 0.2 – – – – – – – – – –
continued...

149
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
Y3A CTR A 82 0.69 0.16 0.06 0.09 1.76 14.39 0.26 51.38 30.99 99.79 0.22 – – – – – – – – – –
Y3B C B 81 0.68 0.16 0.04 0.05 1.93 15.08 0.25 50.01 31.67 99.88 0.22 – – – – – – – – – –
Y3B R B 73 0.68 0.18 0.04 0.06 2.96 14.79 0.22 51.29 30.72 100.94 0.2 – – – – – – – – – –
Y3C C A 73 0.67 0.17 0.05 0.06 2.98 14.73 0.24 50.37 30.75 100.02 0.22 – – – – – – – – – –
Y3C I A 75 0.65 0.18 0.05 0.05 2.75 15.13 0.23 50.66 31.23 100.93 0.21 – – – – – – – – – –
Y3D CTR B 74 0.69 0.21 0.04 0.04 2.87 15.04 0.23 50.63 31.20 100.94 0.2 – – – – – – – – – –
Y3E C B 73 0.73 0.15 0.06 0.08 2.97 14.90 0.26 51.06 30.43 100.64 0.22 – – – – – – – – – –
Y3E R B 68 0.81 0.17 0.05 0.08 3.59 14.07 0.27 52.41 29.33 100.78 0.2 – – – – – – – – – –
Y3F C A 74 0.70 0.19 0.04 0.08 2.98 15.45 0.23 49.13 31.47 100.27 0.2 – – – – – – – – – –
Y3F I A 80 0.68 0.19 0.04 0.04 2.30 16.73 0.20 47.55 32.59 100.33 0.18 – – – – – – – – – –
Y3F R A 67 0.87 0.16 0.07 0.10 3.74 13.70 0.26 51.89 29.77 100.55 0.19 – – – – – – – – – –
Y4A CTR A 77 0.67 0.21 0.04 0.04 2.56 16.06 0.21 49.08 31.85 100.74 0.2 – – – – – – – – – –
Y4D C A 73 0.70 0.18 0.05 0.06 3.05 15.31 0.23 49.51 31.29 100.39 0.2 – – – – – – – – – –
continued...

150
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
Y4D R A 67 0.94 0.18 0.07 0.11 3.59 13.55 0.28 52.27 29.50 100.49 0.19 – – – – – – – – – –
Y4E C A 71 0.68 0.14 0.05 0.07 3.40 15.00 0.25 50.38 30.57 100.54 0.22 – – – – – – – – – –
Y4E R A 56 1.09 0.13 0.11 0.13 4.96 11.51 0.14 54.30 28.01 100.40 0.09 – – – – – – – – – –
Y4F CTR A 69 0.65 0.18 0.06 0.06 3.55 14.72 0.25 50.46 30.79 100.72 0.23 – – – – – – – – – –
Y4G CTR B 71 0.68 0.18 0.04 0.09 3.40 15.01 0.25 50.02 30.83 100.51 0.22 – – – – – – – – – –
Y5A C B 85 0.56 0.23 0.02 0.04 1.69 17.89 0.16 46.29 33.61 100.50 0.18 – – – – – – – – – –
Y5A I B 73 0.65 0.17 0.05 0.04 3.13 15.18 0.23 49.59 31.14 100.20 0.22 – – – – – – – – – –
Y5A R B 66 0.80 0.17 0.07 0.09 3.86 13.80 0.26 51.19 30.06 100.30 0.2 – – – – – – – – – –
Y5B CTR A 72 0.70 0.19 0.05 0.05 3.20 15.39 0.26 48.71 31.07 99.65 0.23 – – – – – – – – – –
Y2A C B 72 0.70 0.19 0.05 0.05 3.23 14.91 0.24 49.48 30.75 99.61 0.21 – – – – – – – – – –
Y2A C B 72 0.62 0.16 0.05 0.06 3.20 15.16 0.23 49.37 30.82 99.66 0.22 – – – – – – – – – –
Y2B CTR A 66 0.85 0.18 0.06 0.07 3.90 14.07 0.30 50.48 29.69 99.61 0.22 – – – – – – – – – –
Y2C R A 71 0.77 0.19 0.06 0.04 3.32 14.79 0.25 49.41 30.77 99.63 0.2 – – – – – – – – – –
continued...

151
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
Y2C C A 68 0.75 0.23 0.16 0.08 3.62 14.17 0.24 50.21 30.27 99.73 0.2 – – – – – – – – – –
Y2D CTR A 68 0.80 0.15 0.07 0.06 3.69 14.44 0.26 50.59 30.36 100.45 0.2 – – – – – – – – – –
X4 C B 82 0.60 0.19 0.04 0.03 2.09 17.06 0.19 47.83 32.60 100.64 0.2 – – – – – – – – – –
X4 R B 74 0.58 0.21 0.04 0.08 2.97 15.70 0.22 48.70 31.41 99.91 0.22 – – – – – – – – – –
X3 C B 84 0.54 0.20 0.05 0.04 1.83 17.47 0.15 46.22 33.51 100.01 0.18 – – – – – – – – – –
X3 I B 83 0.57 0.23 0.04 0.03 1.99 17.50 0.16 47.92 32.41 100.87 0.18 – – – – – – – – – –
XLE1 C A 69 0.55 0.14 0.05 0.05 3.37 13.80 0.23 50.32 31.42 99.93 0.25 367.5 – 348 0.09 9.2 0.23 0.56 0.32 0.06 0.18
XLE2 C A 68 0.67 0.14 0.05 0.06 3.45 13.65 0.24 50.02 31.07 99.35 0.22 469.4 – 371 0.18 8.5 0.32 0.49 0.15 0.12 0.24
XLE3 C A 68 0.65 0.15 0.04 0.05 3.56 13.81 0.24 50.97 31.34 100.82 0.23 494.6 – 362 0.16 10.2 0.30 0.58 0.33 0.12 0.25
XLE4 C A 70 0.64 0.16 0.04 0.05 3.26 14.10 0.23 50.19 31.92 100.59 0.21 472.4 – 385 0.16 10.8 0.30 0.86 0.16 0.10 0.28
XLB1 C A 70 0.66 0.16 0.08 0.04 3.27 13.88 0.26 50.71 30.62 99.68 0.24 429.3 – 406 0.10 16.1 0.35 0.73 0.33 0.08 0.26
XLB1 C A 67 0.60 0.10 0.05 0.06 3.57 12.98 0.25 50.61 31.40 99.62 0.25 401.1 – 382 0.08 14.9 0.32 0.70 0.33 0.05 0.24
XLB2 C A 68 0.63 0.18 0.07 0.08 3.39 13.29 0.22 50.80 32.00 100.66 0.22 371.7 – 373 0.11 10.4 0.33 0.63 0.28 0.05 0.21
continued...

152
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
XLB2 C A 69 0.60 0.09 0.02 0.10 3.38 13.46 0.24 50.34 32.01 100.23 0.24 359.1 – 368 0.12 9.9 0.29 0.61 0.30 0.03 0.24
XLB3 C A 65 0.62 0.13 0.05 0.07 3.77 12.92 0.23 51.22 31.31 100.32 0.22 444.8 – 338 0.10 9.7 0.20 0.60 0.20 0.17 0.30
XLB3 C A 64 0.62 0.15 0.05 0.02 3.85 12.57 0.24 51.55 31.02 100.08 0.23 392.1 – 378 0.09 12.6 0.29 0.59 0.22 0.03 0.26
XLB4 C A 66 0.65 0.12 0.04 0.07 3.50 12.57 0.23 50.22 32.25 99.66 0.22 451.4 – 320 0.18 9.8 0.32 0.58 0.05 0.12 0.14
XLC1 C A 69 0.63 0.15 0.03 0.04 3.37 13.68 0.20 50.31 32.51 100.93 0.2 377.7 – 379 0.09 11.6 0.31 0.70 0.32 0.05 0.25
XLC1 C A 68 0.61 0.14 0.05 0.06 3.49 13.36 0.24 50.27 32.10 100.32 0.23 456.2 – 373 0.18 11.2 0.26 0.72 0.24 0.12 0.25
XLC2 C A 66 0.63 0.14 0.06 0.09 3.68 13.03 0.24 50.95 31.90 100.71 0.23 507.8 – 357 0.16 11.3 0.20 0.73 0.26 0.14 0.19
XLC2 C A 67 0.56 0.16 0.05 0.09 3.53 13.02 0.21 50.06 31.89 99.56 0.22 365.7 – 370 0.08 10.7 0.28 0.60 0.31 0.05 0.24
XLC3 C A 71 0.63 0.14 0.04 0.05 3.13 13.74 0.24 50.05 32.65 100.66 0.23 365.7 – 374 0.09 9.8 0.26 0.60 0.22 0.03 0.26
XLC4 C A 67 0.66 0.12 0.10 0.06 3.48 12.86 0.26 50.55 31.59 99.68 0.23 420.9 – 336 0.16 9.6 0.28 0.58 0.24 0.20 0.28
XLC5 C A 66 0.66 0.14 0.04 0.05 3.60 12.82 0.23 50.36 31.89 99.79 0.21 480.8 – 338 0.13 12.1 0.28 0.67 0.19 0.08 0.26
XLD1 C A 68 0.70 0.13 0.03 0.08 3.45 13.44 0.25 50.25 32.00 100.33 0.22 380.7 – 399 0.11 13.8 0.34 0.64 0.31 0.05 0.24
XLD1 C A 68 0.61 0.11 0.07 0.06 3.50 13.39 0.23 49.02 32.03 99.03 0.23 386.7 – 390 0.08 12.3 0.28 0.65 0.28 0.06 0.28
continued...

153
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
XLD2 C A 70 0.57 0.14 0.05 0.07 3.26 13.83 0.20 50.28 32.45 100.84 0.21 375.3 – 398 0.13 14.1 0.35 0.72 0.31 0.06 0.26
C C B 84 0.57 0.17 0.03 0.02 1.71 16.74 0.15 46.66 33.41 99.45 0.17 – – – – – – – – – –
C I B 80 0.58 0.16 0.02 0.01 2.26 16.01 0.19 47.42 32.86 99.51 0.2 – – – – – – – – – –
C I B 82 0.51 0.16 0.02 0.01 1.96 16.62 0.18 46.69 33.60 99.75 0.22 – – – – – – – – – –
C R B 74 0.69 0.15 0.05 0.07 2.91 15.01 0.21 49.40 31.69 100.17 0.19 – – – – – – – – – –
B2 C B 85 0.57 0.18 0.01 0.04 1.75 17.45 0.14 45.96 33.63 99.73 0.15 – – – – – – – – – –
B2 I B 84 0.55 0.15 0.02 0.00 1.76 17.14 0.19 46.13 33.80 99.72 0.21 – – – – – – – – – –
B2 R B 75 0.80 0.12 0.04 0.05 2.82 15.36 0.21 48.76 32.36 100.53 0.17 – – – – – – – – – –
SB1 C A 72 0.61 0.12 0.04 0.06 3.13 14.84 0.22 49.89 31.36 100.28 0.22 – – – – – – – – – –
SB1 I A 77 0.80 0.12 0.01 0.03 2.57 15.48 0.32 48.75 32.61 100.70 0.24 – – – – – – – – – –
SB1 R A 62 0.86 0.13 0.05 0.06 4.37 13.10 0.28 52.61 29.30 100.76 0.2 – – – – – – – – – –
A2 C B 84 0.55 0.15 0.02 0.03 1.74 17.00 0.17 46.54 32.99 99.19 0.2 – – – – – – – – – –
A2 C B 84 0.56 0.15 0.03 0.00 1.82 16.93 0.18 46.70 32.95 99.32 0.2 – – – – – – – – – –
continued...

154
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
A2 I B 84 0.59 0.21 0.00 0.03 1.86 17.27 0.17 46.43 33.40 99.97 0.19 – – – – – – – – – –
A2 R B 72 0.72 0.13 0.04 0.09 3.11 14.85 0.22 49.56 31.21 99.94 0.2 – – – – – – – – – –
A2 R B 63 0.95 0.10 0.06 0.12 4.19 12.86 0.34 52.68 28.96 100.26 0.22 – – – – – – – – – –
Site 1203 Unit 31
A C B 86 0.57 0.14 0.02 0.03 1.52 16.72 0.16 46.07 33.91 99.15 0.18 192.4 – 336 0.11 5.4 0.25 0.54 0.30 0.07 0.19
A C B 85 0.61 0.13 0.01 0.05 1.60 16.54 0.15 46.23 33.68 99.01 0.16 191.8 – 330 0.13 5.4 0.24 0.50 0.25 0.08 0.18
A C B 82 0.56 0.13 0.02 0.02 1.98 15.90 0.19 47.16 33.29 99.24 0.21 224.2 – 321 0.12 5.7 0.23 0.45 0.23 0.05 0.22
A I B 83 0.55 0.16 0.03 0.03 1.87 16.33 0.18 46.87 33.86 99.87 0.2 – – – – – – – – – –
A I B 79 0.56 0.14 0.03 0.05 2.28 15.48 0.22 48.01 33.06 99.84 0.23 – – – – – – – – – –
A R B 86 0.58 0.12 0.03 0.03 1.50 16.77 0.15 46.51 33.61 99.39 0.17 197.2 – 328 0.16 7.0 0.22 0.53 0.29 0.04 0.20
Bsm C B 85 0.56 0.15 0.02 – 1.63 17.35 0.17 46.30 34.43 100.61 0.19 – – – – – – – – – –
Bsm R B 72 0.55 0.14 0.10 0.07 3.12 15.05 0.25 49.58 32.19 101.03 0.26 – – – – – – – – – –
continued...

155
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
C C B 81 0.64 0.16 0.04 0.04 2.07 16.14 0.18 48.01 33.21 100.48 0.18 – – – – – – – – – –
C C B 81 0.60 0.17 0.03 0.03 2.11 15.92 0.18 47.97 33.55 100.56 0.19 241.0 – 341 0.10 6.4 0.29 0.57 0.21 0.05 0.22
C C B 82 0.63 0.18 0.03 0.04 1.99 16.23 0.17 47.38 33.81 100.47 0.17 – – – – – – – – – –
C C B 82 0.63 0.10 0.03 0.04 1.92 15.89 0.17 47.63 34.18 100.63 0.17 244.0 – 334 0.12 5.9 0.25 0.44 0.21 0.06 0.19
C I B 81 0.57 0.17 0.04 0.06 2.04 15.58 0.17 48.24 33.78 100.65 0.19 217.0 – 323 0.10 5.3 0.22 0.44 0.21 0.03 0.17
C I B 80 0.57 0.15 0.02 0.02 2.08 15.50 0.19 47.46 33.81 99.80 0.21 221.8 – 318 0.09 5.4 0.23 0.41 0.24 0.04 0.17
C I B 85 0.54 0.17 0.01 0.01 1.62 16.24 0.18 45.58 34.88 99.24 0.21 – – – – – – – – – –
C I B 87 0.52 0.17 0.02 0.02 1.33 16.68 0.18 44.97 35.25 99.13 0.21 191.8 – 345 0.12 5.6 0.24 0.45 0.27 0.06 0.21
C I B 84 0.59 0.16 0.03 0.04 1.78 16.76 0.19 47.32 33.83 100.69 0.2 225.4 – 357 0.14 6.3 0.28 0.56 0.28 0.05 0.20
C R B 82 0.63 0.17 0.03 0.02 1.94 15.98 0.18 47.70 33.84 100.49 0.18 – – – – – – – – – –
D C B 82 0.54 0.14 0.02 0.04 1.95 16.25 0.18 48.50 32.28 99.90 0.21 177.5 – 323 0.09 5.5 0.15 0.38 0.13 0.03 0.18
D C B 81 0.59 0.15 0.02 – 2.05 15.93 0.18 48.76 33.06 100.74 0.19 168.5 – 324 0.09 5.4 0.20 0.47 0.19 0.02 0.21
D C B 86 0.56 0.15 0.01 0.02 1.51 16.79 0.16 47.34 34.13 100.65 0.18 195.4 – 330 0.11 5.5 0.24 0.57 0.20 0.07 0.22
continued...

156
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
D I B 84 0.52 0.17 0.02 0.03 1.76 16.27 0.19 47.79 33.71 100.45 0.22 210.4 – 328 0.11 5.9 0.24 0.53 0.21 0.06 0.20
D I B 83 0.55 0.15 0.02 0.04 1.83 16.09 0.18 48.18 33.65 100.70 0.2 216.4 – 325 0.10 5.9 0.22 0.47 0.25 0.04 0.22
D R B 84 0.54 0.13 0.03 0.03 1.77 16.42 0.18 47.88 33.75 100.72 0.21 – – – – – – – – – –
E C B 84 0.61 0.16 0.03 0.02 1.78 16.89 0.17 47.18 33.95 100.79 0.18 – – – – – – – – – –
E I B 71 0.63 0.10 0.04 0.05 3.20 14.24 0.26 50.66 31.49 100.67 0.24 – – – – – – – – – –
E I B 82 0.61 0.13 0.02 0.01 1.96 16.60 0.21 47.23 33.76 100.52 0.21 214.6 – 338 0.14 6.2 0.24 0.47 0.27 0.04 0.24
E I B 84 0.64 0.15 0.02 0.03 1.78 16.47 0.20 45.90 34.12 99.32 0.2 253.6 – 336 0.10 8.4 0.26 0.48 0.22 0.08 0.22
E R B 72 0.66 0.11 0.21 0.05 2.96 14.12 0.25 49.33 31.85 99.54 0.23 204.4 – 346 0.08 6.6 0.21 0.52 0.24 0.06 0.18
G C B 80 0.58 0.17 0.02 0.02 2.22 16.51 0.23 46.63 34.09 100.46 0.23 223.6 – 351 0.12 6.5 0.25 0.51 0.25 0.04 0.22
G I B 85 0.60 0.12 0.00 0.02 1.63 16.73 0.18 45.70 34.28 99.27 0.19 232.0 – 330 0.10 6.7 0.21 0.52 0.24 0.06 0.21
I C B 83 0.63 0.15 0.03 0.03 1.89 16.43 0.20 46.82 33.79 99.97 0.2 229.6 – 327 0.13 6.6 0.25 0.51 0.25 0.03 0.20
I I B 83 0.61 0.12 0.03 0.02 1.79 16.22 0.20 46.28 34.01 99.27 0.2 220.6 – 318 0.11 6.3 0.22 0.51 0.27 0.07 0.24
I I B 82 0.59 0.15 0.03 0.02 1.86 15.82 0.19 46.62 33.24 98.53 0.2 – – – – – – – – – –
continued...

157
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
I R B 71 0.51 0.13 0.05 0.04 3.18 14.02 0.21 48.88 31.78 98.80 0.24 – – – – – – – – – –
J C B 84 0.58 0.10 0.03 0.04 1.70 16.76 0.20 46.24 34.70 100.34 0.21 241.0 – 353 0.14 6.0 0.26 0.51 0.29 0.04 0.22
SPL3 C A 70 0.73 0.11 0.05 0.05 3.29 14.15 0.26 49.56 31.11 99.31 0.22 – – – – – – – – – –
SPL1 C A 69 0.79 0.12 0.05 0.04 3.42 13.64 0.27 50.53 31.81 100.66 0.21 – – – – – – – – – –
SPL5 C A 67 0.77 0.12 0.06 0.09 3.62 13.34 0.23 49.96 31.01 99.20 0.19 – – – – – – – – – –
SPL5 R A 70 0.76 0.11 0.07 0.05 3.25 13.69 0.21 50.47 30.55 99.17 0.17 – – – – – – – – – –
SPL2 C A 71 0.72 0.12 0.05 0.09 3.17 14.14 0.23 49.13 31.68 99.32 0.2 – – – – – – – – – –
SPL6 R A 70 0.83 0.13 0.08 0.06 3.33 14.44 0.28 50.30 31.38 100.83 0.21 548.6 – 337 0.22 9.7 0.32 0.55 0.18 0.04 0.30
SPL7 C A 68 0.79 0.12 0.18 0.07 3.47 13.77 0.30 50.97 30.85 100.52 0.23 – – – – – – – – – –
SPL7 R A 68 0.81 0.12 0.05 0.09 3.57 13.73 0.26 51.18 31.15 100.95 0.2 – – – – – – – – – –
SPL8 C A 72 1.59 0.16 0.11 0.18 3.03 14.10 0.89 50.11 29.45 99.61 0.3 – – – – – – – – – –
SPL9 C A 68 0.90 0.10 0.12 0.09 3.49 13.83 0.40 49.97 30.65 99.57 0.26 – – – – – – – – – –
SPL9 R A 71 0.71 0.11 0.27 0.08 3.00 14.22 0.25 48.54 31.88 99.06 0.22 – – – – – – – – – –
continued...

158
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
SPL10 C A 67 1.02 0.14 0.10 0.13 3.73 13.70 0.41 50.56 30.11 99.91 0.24 – – – – – – – – – –
Y1 C A 77 0.61 0.15 0.08 0.06 2.37 14.71 0.23 50.30 32.27 100.78 0.23 – – – – – – – – – –
Y2B C A 74 0.76 0.14 0.20 0.06 2.54 13.92 0.29 50.43 30.73 99.07 0.23 – – – – – – – – – –
Y3 C A 73 0.80 0.16 0.05 0.06 2.88 14.47 0.28 50.80 31.35 100.86 0.21 – – – – – – – – – –
Y4A C A 74 0.90 0.14 0.06 0.07 2.59 13.34 0.34 51.92 29.82 99.18 0.23 – – – – – – – – – –
Y4B C A 74 0.88 0.14 0.06 0.06 2.60 13.70 0.26 51.21 30.57 99.49 0.19 – – – – – – – – – –
Y6C C A 72 0.81 0.16 0.09 0.08 2.87 13.67 0.32 51.85 30.73 100.59 0.24 – – – – – – – – – –
Y6B C A 73 0.56 0.15 0.09 0.06 2.81 14.04 0.23 50.31 31.03 99.29 0.25 – – – – – – – – – –
Y7A C A 73 0.77 0.16 0.06 0.06 2.88 14.20 0.57 50.61 31.35 100.65 0.36 – – – – – – – – – –
Y5L C B 84 0.58 0.18 0.02 0.03 1.75 16.55 0.18 47.81 33.81 100.93 0.2 – – 336 – 6.5 – – – – –
Y5L I B 80 0.60 0.16 0.03 0.04 2.18 15.74 0.21 48.63 32.64 100.23 0.22 – – 344 – 7.6 – – – – –
Y5L I B 83 0.56 0.18 0.02 0.01 1.80 16.48 0.18 47.75 33.88 100.87 0.2 – – 327 – 6.7 – – – – –
Y5L I B 84 0.55 0.17 0.02 0.04 1.65 16.35 0.19 47.87 33.25 100.10 0.21 – – 342 – 7.4 – – – – –
continued...

159
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
Y5L I B 81 0.57 0.17 0.03 0.02 2.03 15.39 0.21 49.10 32.60 100.13 0.22 – – – – – – – – – –
Y5L I B 84 0.56 0.18 0.02 0.02 1.71 16.42 0.18 47.69 33.74 100.53 0.2 – – 317 – 6.0 – – – – –
Y5L R B 71 0.70 0.13 0.06 0.06 2.92 12.89 0.28 53.49 29.88 100.41 0.24 – – – – – – – – – –
Y10 C B 82 0.55 0.16 0.26 0.02 1.65 15.44 0.20 48.81 33.52 100.61 0.22 – – 324 – 6.3 – – – – –
Y10 I B 82 0.56 0.16 0.02 0.01 1.94 15.89 0.18 48.22 33.54 100.54 0.2 – – 324 – 6.3 – – – – –
Y10 I B 83 0.57 0.18 0.02 0.02 1.77 16.16 0.19 47.84 33.99 100.72 0.2 – – 331 – 6.6 – – – – –
Y10 I B 82 0.55 0.19 0.02 0.03 1.86 15.98 0.19 47.97 33.68 100.48 0.21 – – 329 – 6.9 – – – – –
Y10 R B 73 0.61 0.17 0.08 0.06 2.73 13.78 0.23 50.35 31.25 99.25 0.23 – – – – – – – – – –
Y11 C B 84 0.54 0.18 0.02 0.03 1.74 16.20 0.19 47.74 33.62 100.25 0.21 – – 307 – 5.6 – – – – –
Y11 C B 84 0.52 0.18 0.02 0.03 1.65 16.23 0.18 47.54 33.60 99.95 0.21 – – 327 – 6.0 – – – – –
Y11 I B 84 0.53 0.20 0.02 0.03 1.67 16.15 0.19 47.23 33.63 99.64 0.21 – – 310 – 6.1 – – – – –
Y11 I B 83 0.57 0.19 0.03 0.03 1.80 15.79 0.19 47.74 32.94 99.27 0.2 – – – – – – – – – –
continued...

160
TABLE 3.3
Continued
Sample Zone Pop. An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Mg# Ti Rb Sr Y Ba La Ce Nd Sm Eu
(mol%)
Site 884 Unit 8
9R2A C B 86 0.50 – 0.01 – 1.54 17.7 0.2 46.1 34.2 100.25 0.24 153.2 0.07 201 0.16 2.9 0.12 0.27 0.21 0.13 0.17
9R2A R B 82 0.51 – 0.02 – 2 16.9 0.2 48 33.2 100.83 0.23 125.3 0.09 207 0.21 2.8 0.09 0.29 0.23 0.16 0.16
9R3B C B 85 0.52 – 0.01 – 1.61 17.2 0.18 46.2 34.9 100.62 0.21 191.8 0.12 207 0.11 3.5 0.14 0.26 0.14 0.15 0.17
9R3B R B 84 0.53 – 0.01 – 1.8 16.8 0.21 47.2 33.8 100.35 0.24 124.1 0.10 195 0.11 2.5 0.13 0.22 0.20 0.18 0.17
1 Interm. = Intermediate area between core and rim of largest discernible plagioclase crystal in each xenolith
2 Fe reported as total FeO
3 Major elements in wt.% and trace elements in ppm

161
TABLE 3.4
PLAGIOCLASE TRACE ELEMENT PARTITION COEFFICIENTS AND
CALCULATED PARENT LIQUID COMPOSITIONS
Sample Zone Pop. An DSr DBa DLa DCe DNd DSm DY Sr Ba La Ce Nd Sm Y
(mol%)
4
Site 1203 Unit 3
A C B 80 1.46 0.19 0.1 0.08 0.06 0.03 0.01 268.9 53.4 2.84 9.01 5.94 1.5 15.9
A I B 74 1.61 0.23 0.1 0.09 0.06 0.03 0.01 235.3 44.5 2.79 8.97 4.66 2.48 16.5
A I B 75 1.58 0.22 0.1 0.09 0.06 0.03 0.01 234 44 2.81 8.7 5.27 2.54 16.1
A I B 80 1.45 0.19 0.1 0.08 0.06 0.03 0.01 268.8 52.4 2.93 8.46 5.35 1.32 11.4
A I B 80 1.47 0.19 0.1 0.08 0.06 0.03 0.01 269.2 54 3.19 8.5 4.61 2.33 15
A I B 80 1.46 0.19 0.1 0.08 0.06 0.03 0.01 274.6 55.3 3.16 9.04 5.92 2.4 16.6
A I B 79 1.48 0.2 0.1 0.08 0.06 0.03 0.01 248.5 48.3 2.68 8.34 4.74 2 12.2
A I B 78 1.49 0.2 0.1 0.09 0.06 0.03 0.01 241 47.1 2.81 7.96 4.51 0.73 12.1
B C B 79 1.49 0.2 0.1 0.09 0.06 0.03 0.01 238.8 46.6 2.79 7.42 4.42 2.51 11.8
B I B 80 1.46 0.19 0.1 0.08 0.06 0.03 0.01 248.5 49.4 2.7 8.24 4.83 1.54 13.3
B I B 80 1.47 0.19 0.1 0.08 0.06 0.03 0.01 244.3 48.3 2.69 8.45 5.11 1.57 12.2
continued...

162
TABLE 3.4
Continued
Sample Zone Pop. An DSr DBa DLa DCe DNd DSm DY Sr Ba La Ce Nd Sm Y
(mol%)
B I B 78 1.51 0.21 0.1 0.09 0.06 0.03 0.01 226.7 44.6 2.68 7.84 4.07 1.71 13.7
C C B 81 1.44 0.18 0.1 0.08 0.06 0.04 0.01 254.7 49.5 2.93 8.24 6.26 1.24 13
C C B 80 1.47 0.19 0.1 0.08 0.06 0.03 0.01 247.2 49.8 2.92 8.08 4.43 1.6 11.3
C I B 79 1.47 0.19 0.1 0.08 0.06 0.03 0.01 245.5 48.3 2.91 7.87 4.13 1.53 12
C I B 77 1.52 0.21 0.1 0.09 0.06 0.03 0.01 236.9 45.6 2.87 8.33 4.43 2.2 14.7
C I B 71 1.69 0.26 0.1 0.09 0.05 0.03 0.01 198.3 37.6 2.33 7.25 4.33 1.97 9.9
C I B 78 1.52 0.21 0.1 0.09 0.06 0.03 0.01 245.7 51.7 3.06 7.48 5.86 1.68 12.2
C I B 77 1.54 0.21 0.1 0.09 0.06 0.03 0.01 228.8 45.7 2.52 8.26 4.48 1.43 8.8
C I B 78 1.5 0.2 0.1 0.09 0.06 0.03 0.01 235.5 46 3.21 8.01 4.78 1.81 15.8
C R B 69 1.74 0.28 0.1 0.09 0.05 0.03 0.01 191.9 37.1 2.39 7.85 4.48 1.88 12.6
D-L C B 80 1.47 0.19 0.1 0.08 0.06 0.03 0.01 254.5 51.2 3.04 7.92 5.19 1.25 13.6
D-L I B 79 1.48 0.2 0.1 0.08 0.06 0.03 0.01 250.7 49 2.89 8.86 5.75 1.83 11
D-L I B 78 1.5 0.2 0.1 0.09 0.06 0.03 0.01 243.4 48.5 2.78 8.4 5.16 1.91 14.8
continued...

163
TABLE 3.4
Continued
Sample Zone Pop. An DSr DBa DLa DCe DNd DSm DY Sr Ba La Ce Nd Sm Y
(mol%)
D-L I B 79 1.49 0.2 0.1 0.09 0.06 0.03 0.01 247.4 49.6 3.12 8.61 6.31 1.87 10.3
D-S1 C A 76 1.57 0.22 0.1 0.09 0.06 0.03 0.01 234.1 47.8 2.82 7.97 4.22 3.54 6.8
D-S1 I A 74 1.61 0.23 0.1 0.09 0.06 0.03 0.01 221.2 42.7 3.83 6.83 6.55 4.66 5.1
D-S1 I A 75 1.58 0.23 0.1 0.09 0.06 0.03 0.01 236.2 47.1 2.82 7.37 4 3.91 12.3
D-S1 I A 79 1.49 0.2 0.1 0.09 0.06 0.03 0.01 247.5 48.6 2.64 7.78 6.06 2.49 12.8
D-S1 R A 68 1.75 0.28 0.1 0.09 0.05 0.03 0.01 212 46.8 3.22 10.08 6.61 2.51 8.3
D-S1 R A 66 1.83 0.31 0.1 0.09 0.05 0.03 0.01 198.6 41.3 3.26 7.23 4.51 2.54 15.3
D-S2 C A 78 1.5 0.2 0.1 0.09 0.06 0.03 0.01 241.5 49.3 2.18 8.76 4.66 2.2 16.2
D-S2 I A 73 1.64 0.24 0.1 0.09 0.06 0.03 0.01 222.6 39.8 2.9 7.92 1.59 3.92 4.3
D-S2 I A 78 1.52 0.21 0.1 0.09 0.06 0.03 0.01 248.4 48.8 2.58 8.2 4.09 2.63 11.3
D-S2 I A 79 1.49 0.2 0.1 0.09 0.06 0.03 0.01 244.2 47.8 2.51 9.54 4.99 3.41 9.7
D-S2 R A 62 1.94 0.35 0.11 0.09 0.05 0.03 0.01 199.3 65.2 3.05 9.64 5.44 2.9 13.1
E-1 I A 80 1.46 0.19 0.1 0.08 0.06 0.03 0.01 251.6 49.7 3.23 8.42 5.62 1.12 12.1
continued...

164
TABLE 3.4
Continued
Sample Zone Pop. An DSr DBa DLa DCe DNd DSm DY Sr Ba La Ce Nd Sm Y
(mol%)
E-1 I A 80 1.46 0.19 0.1 0.08 0.06 0.03 0.01 245.7 48.7 2.48 8.06 5.03 1.78 10
E-1 I A 81 1.44 0.19 0.1 0.08 0.06 0.04 0.01 248.6 51.1 2.97 8.46 3.94 1.54 11.6
E-1 I A 80 1.45 0.19 0.1 0.08 0.06 0.03 0.01 262 51.9 2.8 8.51 5.68 1.78 12.5
F C A 80 1.45 0.19 0.1 0.08 0.06 0.03 0.01 246.3 46.3 3.13 8.1 4.92 1.26 10.8
F I A 81 1.44 0.19 0.1 0.08 0.06 0.04 0.01 254.2 47.8 3.51 7.82 4.37 1.51 11.6
F I A 73 1.64 0.25 0.1 0.09 0.06 0.03 0.01 208 39.3 2.97 6.67 5.71 1.54 14.4
F I A 80 1.45 0.19 0.1 0.08 0.06 0.03 0.01 250.3 46.8 3.23 8.03 5.56 2.08 14
F I A 78 1.5 0.2 0.1 0.09 0.06 0.03 0.01 236.3 44.3 2.97 7.73 5.48 2.14 12.8
F I A 79 1.49 0.2 0.1 0.09 0.06 0.03 0.01 245.8 49.1 2.67 8.37 5.48 1.39 15.8
F R A 64 1.89 0.33 0.1 0.09 0.05 0.03 0.01 226.7 79.5 4.09 11.78 6.17 1.93 17.1
H I A 77 1.54 0.21 0.1 0.09 0.06 0.03 0.01 235.5 46.5 2.82 8.14 5.12 2.15 10.3
H I A 78 1.5 0.2 0.1 0.09 0.06 0.03 0.01 240.4 46.2 2.79 8.6 4.91 1.71 18.7
H I A 81 1.44 0.19 0.1 0.08 0.06 0.04 0.01 256.2 53.2 2.53 7.78 4.58 2.14 13.2
continued...

165
TABLE 3.4
Continued
Sample Zone Pop. An DSr DBa DLa DCe DNd DSm DY Sr Ba La Ce Nd Sm Y
(mol%)
H I A 80 1.46 0.19 0.1 0.08 0.06 0.03 0.01 249.7 48 2.99 8 5.01 1.16 15
J C B 78 1.51 0.21 0.1 0.09 0.06 0.03 0.01 233.4 41.8 3.09 7.35 4.52 1.04 12.6
J I B 75 1.59 0.23 0.1 0.09 0.06 0.03 0.01 217.3 37.3 3.2 7.32 4.65 2.96 12.3
J I B 82 1.42 0.18 0.1 0.08 0.06 0.04 0.01 247.8 43.1 2.99 7.89 4.87 1.87 13.3
J I B 78 1.5 0.2 0.1 0.09 0.06 0.03 0.01 234.5 43.9 2.85 6.92 4.78 1.41 12.6
J I B 77 1.54 0.21 0.1 0.09 0.06 0.03 0.01 221.1 41.5 2.64 7.81 4.87 2.12 14.1
PL-D C A 62 1.93 0.35 0.1 0.09 0.05 0.03 0.01 197 50.5 2.95 8.72 5.21 2.98 17.5
PL-G C A 64 1.89 0.33 0.1 0.09 0.05 0.03 0.01 187.3 44.7 4.2 8.52 5.64 3.17 11.3
PL-G R A 66 1.81 0.3 0.1 0.09 0.05 0.03 0.01 194.5 42.1 2.62 7.93 5.53 2.43 8
PL-F I B 80 1.45 0.19 0.1 0.08 0.06 0.03 0.01 245.3 46.2 2.48 7.87 4.46 3.52 15
PL-H C A 63 1.92 0.34 0.1 0.09 0.05 0.03 0.01 211.4 65 3.89 10.76 6.16 3.08 10.2
Site 1203 Unit 14
continued...

166
TABLE 3.4
Continued
Sample Zone Pop. An DSr DBa DLa DCe DNd DSm DY Sr Ba La Ce Nd Sm Y
(mol%)
A I B 87 1.29 0.15 0.1 0.08 0.06 0.04 0.01 304.4 49.3 3.32 8.25 5.91 1.75 19.6
A I B 83 1.39 0.17 0.1 0.08 0.06 0.04 0.01 290.5 45.9 3.27 6.98 5.07 2.51 14.1
A I B 86 1.31 0.15 0.1 0.08 0.06 0.04 0.01 295.6 47.2 3.34 8.2 5.83 1.82 17
A I B 83 1.38 0.17 0.1 0.09 0.06 0.03 0.01 274.8 42.2 2.8 7.52 5.19 1.48 12.9
A I B 81 1.43 0.18 0.1 0.08 0.06 0.04 0.01 262.8 39.1 3.21 6.59 4.59 2.09 12.7
A I B 82 1.42 0.18 0.1 0.08 0.06 0.04 0.01 265.7 41.6 2.68 7.58 5.51 2.62 10.7
A R B 73 1.64 0.24 0.1 0.09 0.06 0.03 0.01 203.3 27 2.59 6.15 4.54 1.54 11.3
B C B 81 1.44 0.19 0.1 0.08 0.06 0.04 0.01 260.5 34.2 2.51 5.07 3.29 1.45 8.7
B I B 82 1.41 0.18 0.1 0.08 0.06 0.04 0.01 267.1 36.5 2.77 6.23 4.41 2.75 13.6
B I B 82 1.41 0.18 0.1 0.08 0.06 0.04 0.01 284.6 38.6 3.08 6.28 5.61 1.13 16.2
B I B 81 1.43 0.18 0.1 0.08 0.06 0.03 0.01 276.9 39.7 2.97 6.59 4.76 1.04 15
B I B 83 1.39 0.17 0.1 0.08 0.06 0.04 0.01 285.9 42.3 3.78 8.36 5.9 2.79 18.8
B I B 83 1.38 0.17 0.09 0.08 0.06 0.04 0.01 301.9 49.2 3.56 7.54 5.89 0.76 19.4
continued...

167
TABLE 3.4
Continued
Sample Zone Pop. An DSr DBa DLa DCe DNd DSm DY Sr Ba La Ce Nd Sm Y
(mol%)
C1 C B 83 1.4 0.17 0.1 0.08 0.06 0.04 0.01 255.5 36.7 3.08 6.8 4.51 2.08 12.7
C1 I B 83 1.38 0.17 0.1 0.08 0.06 0.04 0.01 260.8 36.6 2.72 5.75 4.7 1.03 15.5
C1 I B 83 1.39 0.17 0.1 0.08 0.06 0.04 0.01 261 35.8 2.91 6.85 5.01 1.47 16.7
C1 I B 81 1.42 0.18 0.1 0.08 0.06 0.04 0.01 263.2 39.6 3.33 6.54 5.07 1.62 16.9
D1 C B 83 1.39 0.17 0.1 0.08 0.06 0.04 0.01 274.2 35.6 2.63 5.78 4.56 1.68 13.6
D1 I B 81 1.43 0.18 0.1 0.08 0.06 0.04 0.01 261 34.1 2.44 5.71 4.54 1.85 14.2
D1 I B 83 1.38 0.17 0.1 0.08 0.06 0.04 0.01 281.1 38.5 2.54 5.8 4.62 2.15 12.5
D1 I B 83 1.38 0.17 0.1 0.09 0.05 0.03 0.01 276.4 40.1 3.24 6.34 5.07 1.09 17.3
E C B 82 1.41 0.18 0.1 0.08 0.06 0.04 0.01 277.4 41.7 2.66 6.27 4.98 1.56 16.7
E I B 72 1.65 0.25 0.1 0.09 0.05 0.03 0.01 208.1 26.9 3.17 6.45 5.37 0.76 12.1
PG2 CTR A 72 1.67 0.25 0.1 0.08 0.06 0.04 0.01 218.4 40.1 2.65 6.22 7.89 2.97 21.6
PG9 C A 70 1.71 0.27 0.1 0.09 0.05 0.03 0.01 220.2 44.7 2.64 6.99 6.4 4.37 11.1
XLE1 C A 69 1.73 0.28 0.1 0.09 0.05 0.03 0.01 200.6 33.4 2.25 6.48 5.86 1.73 9.8
continued...

168
TABLE 3.4
Continued
Sample Zone Pop. An DSr DBa DLa DCe DNd DSm DY Sr Ba La Ce Nd Sm Y
(mol%)
XLE2 C A 68 1.75 0.28 0.1 0.09 0.05 0.03 0.01 211.4 30 3.1 5.69 2.82 3.71 20.1
XLE3 C A 68 1.77 0.29 0.1 0.09 0.05 0.03 0.01 204.9 35.6 2.87 6.73 6.14 3.88 18.2
XLE4 C A 70 1.7 0.26 0.1 0.09 0.05 0.03 0.01 226 40.8 2.95 9.99 2.93 3.1 17.9
XLB1 C A 70 1.72 0.27 0.1 0.09 0.05 0.03 0.01 236.3 59.7 3.44 8.46 6.06 2.41 11.5
XLB1 C A 67 1.81 0.3 0.1 0.09 0.05 0.03 0.01 211.5 49.6 3.11 8.04 6.13 1.64 9.3
XLB2 C A 68 1.76 0.29 0.1 0.09 0.05 0.03 0.01 211.6 36.5 3.23 7.24 5.18 1.72 11.8
XLB2 C A 69 1.75 0.28 0.1 0.09 0.05 0.03 0.01 210.5 35.3 2.82 7.08 5.48 0.96 12.8
XLB3 C A 65 1.84 0.31 0.1 0.09 0.05 0.03 0.01 183.4 31 1.94 6.96 3.73 5.29 10.4
XLB3 C A 64 1.87 0.33 0.1 0.09 0.05 0.03 0.01 201.6 38.7 2.73 6.79 4.09 1.06 9.7
XLB4 C A 66 1.81 0.3 0.1 0.09 0.05 0.03 0.01 176.5 32.3 3.04 6.65 1.01 3.71 19.3
XLC1 C A 69 1.74 0.28 0.1 0.09 0.05 0.03 0.01 218 41.9 2.98 8.06 5.81 1.44 9.9
XLC1 C A 68 1.78 0.29 0.1 0.09 0.05 0.03 0.01 210.1 38.7 2.48 8.28 4.36 3.67 19.9
XLC2 C A 66 1.82 0.31 0.1 0.09 0.05 0.03 0.01 195.8 36.9 1.93 8.42 4.82 4.39 17.7
continued...

169
TABLE 3.4
Continued
Sample Zone Pop. An DSr DBa DLa DCe DNd DSm DY Sr Ba La Ce Nd Sm Y
(mol%)
XLC2 C A 67 1.8 0.3 0.1 0.09 0.05 0.03 0.01 205.9 36 2.71 6.89 5.65 1.65 9.1
XLC3 C A 71 1.7 0.26 0.1 0.09 0.05 0.03 0.01 220.6 37.5 2.58 6.98 3.95 1.05 10.2
XLC4 C A 67 1.8 0.3 0.1 0.09 0.05 0.03 0.01 186.4 32 2.7 6.64 4.47 6.23 17.4
XLC5 C A 66 1.82 0.3 0.1 0.09 0.05 0.03 0.01 185.9 39.7 2.73 7.79 3.48 2.47 14.7
XLD1 C A 68 1.76 0.29 0.1 0.09 0.05 0.03 0.01 226.4 48.4 3.3 7.41 5.73 1.68 12.6
XLD1 C A 68 1.78 0.29 0.1 0.09 0.05 0.03 0.01 219.4 42.4 2.72 7.47 5.1 1.98 9.1
XLD2 C A 70 1.72 0.27 0.1 0.09 0.05 0.03 0.01 232 52.4 3.39 8.39 5.71 1.75 14.8
Site 1203 Unit 31
A C B 86 1.32 0.15 0.13 0.11 0.07 0.05 0.01 254.1 34.9 1.95 4.96 4 1.42 9.5
A C B 85 1.34 0.16 0.13 0.11 0.07 0.05 0.01 246.5 34.3 1.86 4.54 3.32 1.74 11.3
A C B 82 1.42 0.18 0.13 0.11 0.07 0.05 0.01 225.9 31.6 1.81 4.08 3.1 1.13 11
A R B 86 1.32 0.15 0.13 0.11 0.07 0.05 0.01 248.7 45.3 1.79 4.82 3.88 0.92 14.4
continued...

170
TABLE 3.4
Continued
Sample Zone Pop. An DSr DBa DLa DCe DNd DSm DY Sr Ba La Ce Nd Sm Y
(mol%)
C C B 81 1.44 0.19 0.13 0.11 0.07 0.05 0.01 236 34.4 2.24 5.15 2.82 1.16 9.1
C C B 82 1.41 0.18 0.13 0.11 0.07 0.05 0.01 236.6 33.2 1.94 3.96 2.84 1.2 10.5
C I B 81 1.44 0.19 0.13 0.11 0.07 0.05 0.01 224 28.6 1.7 3.97 2.77 0.64 8.8
C I B 80 1.45 0.19 0.13 0.11 0.07 0.05 0.01 219.5 28.7 1.77 3.69 3.24 0.86 8
C I B 87 1.29 0.15 0.12 0.11 0.07 0.05 0.01 267.8 37.9 1.94 4.12 3.66 1.27 11.2
C I B 84 1.37 0.17 0.13 0.11 0.07 0.05 0.01 260.8 38 2.18 5.03 3.78 1.05 12.9
D C B 82 1.41 0.18 0.13 0.11 0.07 0.05 0.01 229.3 31 1.15 3.44 1.8 0.6 8.4
D C B 81 1.43 0.18 0.13 0.11 0.07 0.05 0.01 226 29.4 1.52 4.2 2.52 0.5 7.7
D C B 86 1.32 0.15 0.13 0.11 0.07 0.05 0.01 250.6 36 1.91 5.16 2.73 1.46 10
D I B 84 1.37 0.17 0.13 0.11 0.07 0.05 0.01 238.8 35.1 1.86 4.8 2.78 1.21 10.2
D I B 83 1.39 0.17 0.13 0.11 0.07 0.05 0.01 233.8 34.2 1.68 4.25 3.32 0.89 8.6
E I B 82 1.4 0.17 0.13 0.11 0.07 0.05 0.01 241.1 35.5 1.84 4.22 3.66 0.96 12.6
E I B 84 1.37 0.17 0.13 0.11 0.07 0.05 0.01 244.6 49.8 2.01 4.37 2.9 1.76 9.2
continued...

171
TABLE 3.4
Continued
Sample Zone Pop. An DSr DBa DLa DCe DNd DSm DY Sr Ba La Ce Nd Sm Y
(mol%)
E R B 72 1.67 0.25 0.13 0.11 0.07 0.04 0.01 207.1 25.8 1.56 4.54 3.31 1.3 7.2
G C B 80 1.45 0.19 0.13 0.11 0.07 0.05 0.01 242.2 34.7 1.9 4.57 3.42 0.85 10.4
G I B 85 1.34 0.16 0.13 0.11 0.07 0.05 0.01 246.3 42.2 1.7 4.71 3.17 1.32 8.8
I C B 83 1.4 0.17 0.13 0.11 0.07 0.05 0.01 234.3 37.8 1.93 4.56 3.41 0.62 11.7
I I B 83 1.38 0.17 0.13 0.11 0.07 0.05 0.01 230.2 37.1 1.76 4.58 3.67 1.4 10.2
J C B 84 1.35 0.16 0.13 0.11 0.07 0.05 0.01 260.7 36.9 2.01 4.6 3.87 0.82 12.3
SPL6 R A 70 1.71 0.27 0.13 0.11 0.07 0.04 0.01 197.5 36.4 2.39 4.88 2.5 1.03 18.3
SPL6 I B 83 1.37 0.17 – – – – – 245.9 38.9 – – – – –
SPL6 I B 84 1.46 0.19 – – – – – 235.4 39.6 – – – – –
SPL6 I B 81 1.38 0.17 – – – – – 237.6 39.8 – – – – –
SPL6 I B 84 1.35 0.16 – – – – – 252.9 45.6 – – – – –
SPL6 R B 71 1.44 0.19 – – – – – – – – – – – –
Y10 C B 82 1.36 0.16 – – – – – 232.7 36.5 – – – – –
continued...

172
TABLE 3.4
Continued
Sample Zone Pop. An DSr DBa DLa DCe DNd DSm DY Sr Ba La Ce Nd Sm Y
(mol%)
Y10 I B 83 1.4 0.17 – – – – – 231.5 36.3 – – – – –
Y10 I B 82 1.42 0.18 – – – – – 229 35.5 – – – – –
Y10 R B 73 1.38 0.17 – – – – – 240.4 39.2 – – – – –
Y11 C B 84 1.40 0.17 – – – – – 235.2 39.7 – – – – –
Y11 I B 84 1.37 0.17 – – – – – 223.9 33.7 – – – – –
Y11 I B 83 1.35 0.16 – – – – – 241.5 36.8 – – – – –
Y11 I B 83 1.36 0.16 – – – – – 228.1 37 – – – – –
4 Major elements in wt.% and trace elements in ppm

CHAPTER 4
CRUSTAL CONTAMINATION OF BASALTIC MAGMAS REVEALED
THROUGH MICROANALYSIS OF MAJOR, MINOR, AND TRACE
ELEMENTS AND Sr ISOTOPES IN PLAGIOCLASE: IMPLICATIONS FOR
ELAN BANK MAGMAS, KERGUELEN PLATEAU, INDIAN OCEAN
4.1 Introduction
Large Igneous Provinces (LIPs) are voluminous emplacements of basaltic magma
that occur over geologically short spans of time [32, 42]. The submarine Kergue-
len Plateau (KP) and Broken Ridge constitute the second largest oceanic LIP on
Earth and rise up to 4 km above the surrounding Indian Ocean basin (Fig. 4.1).
Studies of LIPs such as the expansive KP are important for several reasons includ-
ing: 1) LIPs sample portions of the mantle not tapped during normal mid-ocean
ridge spreading and help us better understand the broader mantle, 2) during the
Cretaceous LIP volcanism may have represented as much as 50% of the global
volcanic output, whereas LIPs currently represent < 10% of the output possibly
signifying large scale changes in mantle dynamics, 3) rapid formation of LIPs such
as the KP may have had catastrophic effects on the environment [32, 33]. Oceanic
LIPs such as the KP are among the least understood geologic features on the
planet, which is due in part to the practical difficulties of sampling these largely
submerged oceanic plateaus. Current sampling of the Cretaceous portion of the
173

KP has been done via drilling at 11 drillsites during Ocean Drilling Program Legs
119, 120, and 183.
Prior to drilling during Leg 183 a number of workers suggested that continen-
tal crust was involved during the petrogenesis some KP basalts (e.g., [51, 92]).
Drilling at Site 1137 on Elan Bank confirmed hypotheses of continental crust
involvement when clasts of garnet-biotite gneiss were recovered from a fluvial con-
glomerate unit (Unit 6 in Fig. 4.2) [33, 74]. Nicolaysen et al. [111] and Frey
et al. [49] suggested that during the break up of the Gondwana supercontinent
fragments of old continental crust were stranded amongst the Indian Ocean litho-
sphere including fragments within the KP crust.
One of the scientific objectives of Leg 183 drilling was to constrain the post-
melting evolution of Kerguelen magmas, of which crustal assimilation was clearly
an important process. In this study I explore the timing and nature of crustal con-
tamination of KP magmas by focusing on the compositional and isotopic record of
assimilation contained within zoned plagioclase phenocrysts in two basalts from
Leg 183 Site 1137 Units 4 and 10. Ingle et al. [75] placed the Unit 4 basalt in a
relatively uncontaminated upper basalt group and and the Unit 10 basalt in a rel-
atively contaminated lower basalt group. Examination of plagioclase phenocrysts
from Units 4 and 10 provide snapshots of the magmatic processes that were oc-
curring when both when crustal contamination was significant and when it was
not. From this work I suggest that Unit 4 and 10 plagioclase phenocrysts origi-
nated from the shallowest magma chambers of what was likely a complex system
of interconnected dikes and sills. I contend that crustal assimilation was not an
active process in these shallowest chambers but rather pre-dated plagioclase dom-
inated partial crystallization. Crustal assimilation occurred deeper in the crust
174

Figure 4.1. Map of the Indian Ocean showing the Kerguelen Plateau
(KP) and other features attributed to the Kerguelen plume, and an
inset free-air gravity map of Elan Bank and the location of ODP Site
1137. NKP is Northern KP, CKP is the Central KP, and SKP is the
Southern KP. Maps adapted from [33].
175

and over time crustal wall-rocks became armored, which minimized progressive
crustal assimilation.
4.1.1 Geologic Background of the Kerguelen Plateau and Elan Bank
The bulk of the 2 x 10 6 km 2 Kerguelen Plateau formed during the Cretaceous
[31, 32, 42]. Formation of the KP has been ascribed to surfacing of the Kerguelen
plume and is closely spaced in time with the break up of India and Australia at
∼ 132 Ma [80]. Broken Ridge was separated from the KP during the Early Ter-
tiary with the onset of spreading at the Southeast Indian Ridge [109] (Fig. 4.1).
The Kerguelen Plateau has been divided into distinct sections, which include the
Southern KP, Central KP, Northern KP, and Elan Bank [49] (Fig. 4.1). Heard and
McDonald Islands represent the most recent volcanism ascribed to the Kerguelen
plume, and the northerly trending Ninetyeast Ridge has been attributed to Ker-
guelen plume tail volcanism (e.g., [125]). Elan Bank is a salient that extends from
western edge of the KP and contains a number of seismic reflections consistent
with the presence of continental crust [33] (Fig. 4.5). The presence of continental
crust at Elan Bank was confirmed by drilling [33] (Fig. 4.5).
4.2 Samples
Unit 4 consists of a single ∼ 27 m thick variably vesicular subaerial plagioclase
phyric inflated pahoehoe flow [33]. I selected a single sample from a portion of
the Unit 4 flow that exhibited the least alteration and lowest vesicle content (183-
1137A-32R-05W-22-27; Fig. 4.2). Unit 10 is at the bottom of the Site 1137 cored
section. Unit 10 consists of the top of a subaerial massive plagioclase phyric basalt
176

Figure 4.2. The cored section of igneous rocks from ODP Site 1137.
The locations of each sample used in this study is shown as a boxed in
X. Figure adapted from [33].
177

flow [33]. I selected a single sample from the Unit 10 flow with minimal alteration
(183-1137A-46R-01W-20-35; Fig. 4.2).
4.3 Analytical Techniques
4.3.1 Sample Preparation
Basalt samples were prepared as polished thin sections using a microdiamond
slurry and final grit size of 0.5 µm. Prior to analytical work the slides were rinsed
in ethanol in an ultrasonic bath to remove remnants of the polishing compounds
followed by three 30 minute rinses in ultrapure water in an ultrasonic bath. Sam-
ples were carbon coated for EPMA work. After completion of EPMA work the
carbon coats were removed with 3 µm and 1 µm diamond polishing clothes under
ethanol followed by three 30 minutes rinses in ultrapure water in an ultrasonic
bath. The surface of each slide was inspected under a reflected light optical mi-
croscope to confirm that all vestiges of the carbon coat had been removed prior to
LA-ICP-MS work. Slides were re-polished and re-cleaned after LA-ICP-MS and
prior to Sr isotope microdrilling.
4.3.2 Electron Probe Microanalysis (EPMA) and Scanning Electron Microscopy
Backscatter electron images and major element analyses were performed us-
ing a JEOL JXA-8600 Superprobe electron microprobe at the University of Notre
Dame. Backscatter electron images were collected using a 1 µm beam, an acceler-
ating voltage of 20 kV, and a probe current of 25-50 nA. Microprobe analyses were
performed using a 10 µm defocused beam, accelerating voltage of 15 kV, a probe
current of 20 nA, 15 second on-peak counting time, and two background measure-
ments per peak. Sodium was measured first to minimize loss via volatilization.
178

During core to rim line scans a point was measured every 6 µm using a 5 µm
diameter beam. Microprobe data were corrected for matrix effects using a ZAF
correction routine. Data points near Fe-rich phases such as melt inclusions and
alteration-filled fractures were discarded.
4.3.3 Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-
MS)
Scandium, Ti V, Rb, Sr, Y, Ba, La, Ce, Nd, Sm, Eu, and Pb were measured in
plagioclase crystals using a New Wave UP-213 UV laser ablation system interfaced
with a ThermoFinnigan Element 2 ICP-MS operated in fast magnet scanning
mode at the University of Notre Dame. I used a laser frequency of 5 Hz, pulse
energy of 0.02-0.03 mJ pulse −1 , 30 µm diameter pits, and helium as the carrier
gas (∼ 0.7 l min −1 ) mixed with argon (∼ 1.0 l min −1 ) before introduction to the
plasma. Due to the transient nature of the laser ablation signal, analyses were
conducted in peak jumping mode with one point quantified per mass. The LA-
ICP-MS spots were coincident with previous EPMA analyses, and Ca measured by
EPMA was used as an internal standard for each spot analysis. The trace element
glass NIST 612 was used as a calibration standard for all laser ablation analyses.
Although heterogeneity for certain elements has been documented in the widely
used NIST 612 glass i.e., [44], Eggins et al. [44] considered it a reliable calibration
standard for the elements examined in this study. The analytical protocols of
Longerich et al. [88] were used for LA-ICP-MS data reduction.
179

4.3.4 Major and Trace Element Data quality
The quality of EPMA data were monitored by major element and cation totals.
Data points where major element totals were greater than 101.5% or less than
98.5% are not used in further discussion nor are points with cation totals greater
or less than 20 ± 0.1 cations (based on 32 O). All laser ablation analyses were
collected in time-resolved mode so that signal from inclusions and alteration-filled
fractures could be easily identified. Results from laser ablation spots close to
fractures or that penetrated into underlying fractures were discarded. The rate of
laser ablation was tested using a 100 µm thick plagioclase wafer prior to sampling,
which allowed optimization of operating conditions to yield desired sensitivity and
sampling depth (< 30 µm).
4.3.5 Microdrilling and Sr isotope microanalysis
Microdrilling for Sr isotope microanalysis was performed using a Merchantek
MicroMill at the University of Durham, UK after completion of EPMA and LA-
ICP-MS work. We used a highly tapered tungsten carbide drill bit for all drilling,
which provided < 50 µm diameter single drill pits. Microdrilling was performed by
setting precision scans consisting of grids and lines of drill points, which allowed
us to target zones of interest to maximize the amount of Sr recovered in order to
ensure as accurate and precise 87 Sr/ 86 Sr measurements as possible (see Fig. 4.3).
Prior to microdrilling, a square 4 cm x 4 cm piece of parafilm pre-cleaned in
ultrapure water with a ∼ 2 cm center hole was adhered to the polished sample
surface. A droplet of ultrapure water was then placed over the center hole and area
to be drilled. The parafilm was used to minimize dispersion of the water droplet
and to minimize sample loss. Microdrilling was then performed within the water
180

droplet, which generated a slurry as sample material was removed. The slurry was
removed using a micropipette and placed in a 3.5 mL Teflon beaker. The sample
was then subjected to a hotplate digestion with concentrated HF and HNO3.
Strontium was extracted using a scaled-down ion exchange column method. The
samples were run Re filaments with a TaF activator and 87 Sr/ 86 Sr ratios were
measured using a Finnigan Triton thermal ionization mass spectrometer (TIMS).
Davidson et al. [39] and Tepley et al. [135] provide more thorough discussions
about of the micro-Sr isotope method.
4.3.6 Choosing Partition Coefficients (D)
Blundy [9] noted that inversion of magma compositions from mineral data
using appropriate trace element D values is a robust means of estimating parent
magma compositions and suggested that trace element D values derived from
microbeam techniques were more accurate than those derived from bulk crystal-
matrix analyses. Blundy and Wood [10, 11] and Bindeman et al. [7] showed
that the dominant factors controlling the trace element D vaues for plagioclase
are An content and crystallization temperature. Bindeman et al. [7] applied this
relationship to a variety of major and trace elements and produced values for the
constants “a” and “b” in their Equation 2 (equation 4.1 below) for calculating
plagioclase partition coefficients via the expression:
RT ln(Di) = aXAn + b (4.1)
where R is the gas constant, T is temperature (Kelvin), i is the element of interest,
and XAn is the mole fraction of anorthite. Using equation 4.1, Ginibre et al.
[59] noted that temperatures in the range of 850-1,000 ◦ C had miniscule effect on
181

B)
C)
parafilm
splash guard
A)
tapered
tungsten
carbide bit
1 mm
1 mm
Figure 4.3. A) Drilling into a droplet of ultra-pure water. B) Unit 10
crystal 10-1C-A core drill site. Individual drill spots are ∼ 50 µm in
diameter and are arranged into finely controlled point and line scans.
C) A drill zone in Unit 10 crystal 10-1D-A. See figure 4.9 for more
details and results.
182

calculated melt concentrations for Sr and Ba (within their analytical uncertainty).
Likewise, Bindeman et al. [7] showed that variations of ∼150 ◦ C produce < 10%
differences for particular partition coefficients, which were often within error of
the respective D values.
Blundy and Wood [11] presented an alternative model for quantitative predic-
tion of D values based upon thermodynamic principles and lattice strain theory
(i.e., [15]). Specifically, they noted the dominant roles of cation size, lattice site
size, and elasticity of the lattice site related to cation substitution in minerals.
The elastic response of a lattice site to cation substitution is measured by Young’s
Modulus (E), and the value of E plag is dependent upon plagioclase An content
and the charge of the substituent cation [11, 12]. Equation 4.2 (Equation 2 of
Blundy and Wood [11]) is the basis of their predictive D model that I apply to
plagioclase,
Di = D z+
o
∗ exp
−4πEplagNA
ro
2 (r2 o − r2 i ) + 1
3 (r3 i − r3 o)
RT
(4.2)
where Di of a trace element (i) is a function of the strain compensated partition
coefficient for the divalent (D 2+
0
) or trivalent (D3+ 0 ) substituent cation, Young’s
Modulus for the plagioclase M lattice site (E plag), NA (Avogadro’s Number), the
radius of the plagioclase M site (ro), the radius of the substituent cation (ri),
the gas constant (R), and crystallization temperature (T). For a more thorough
discussion of this model see Blundy and Wood [11, 12].
4.3.6.1 Partition coefficients for Elan Bank Plagioclase Crystals
I used equation 4.2 and the parameterizations of Blundy and Wood [11, 12]
to calculate D values for Rb, Sr, Y, Ba, La, Ce, Pr, Nd, and Sm, as partitioning
183

of these elements are the better constrained through experimental studies (see
Tables 4.5 and 4.5). Estimates of Sc DSc are unrealistically low (< 0.001) using the
Blundy and Wood [11] method, which is most likely due to inadequate estimation
of the strain-free partition coefficient. The partitioning behavior of Fe and Eu are
strongly affected by oxygen fugacity f O2, and due to f O2 constraints I did not
attempt to estimate DFe or DEu values. Magnesium is known to substitute in more
than one plagioclase cation site, and based upon this complex partitioning I did
not calculate DMg [7]. Blundy and Wood [11] noted that predictive approaches are
inaccurate for Pb, which has a lone pair of electrons. I assumed a crystallization
temperature of 1150 ◦ C to calculate D values. The validity of this assumption
was examined using the MELTS program [2, 55] to determine the temperature
at which plagioclase arrives on the liquidus during crystallization of a typical
Elan Bank basalt sample. I input four upper group and lower group Site 1137
whole-rock basalt compositions with 4.4 wt% MgO, 5.9 wt% MgO, 6.0 wt% MgO,
and 7.3 wt% MgO reported by Shipboard Scientific Party [33] into the MELTS
program. During equilibrium crystallization of the lowest MgO basalt, plagioclase
appeared on the liquidus at 1080 ◦ C and at at 1180 ◦ C from the highest MgO
sample. The use of 1080 ◦ C vs. 1180 ◦ C in D calculations yields < 10% differences
in trace element concentrations of parent liquids, which is within the analytical
error for my analyses. The effects of pressure on plagioclase D values are not
well documented for most elements, and where they were documented for DSr by
Vander Auwera et al. [137] they were negligible. I used a crystallization pressure
of 0.001 kbar for D calculations. I assume the strain compensated (i.e., strain free)
partition coefficient D 2+
0 in plagioclase is best approximated by DCa (c.f., [12]),
and to estimate DCa I used equation 4.1 and the “a” and “b” constants reported
184

for Ca by Bindeman et al. [7]. I used equation 4.1 and the “a” and “b” constants
reported for Na by Bindeman et al. [7] to estimate D 1+
0 . Estimation of D 3+
0 is less
straightforward. I assumed D 3+
0 to be ∼ 7.5% of D 2+
0 , which was derived from
experimental data reported in Blundy [9]. I used equation 4.1 and the “a” and
“b” constants reported for Ti by Bindeman et al. [7] to estimate DTi.
4.4 Results
4.4.1 Petrography
Unit 4 plagioclase phenocrysts are typically euhedral and occur as clusters of
crystals but most prominently as individual crystals (Fig. 4.4a). Unit 10 plagio-
clase phenocrysts are large relative to Unit 4 phenocrysts and occur most com-
monly as clusters of euhedral to sub-rounded crystals (Fig. 4.4a). Plagioclase
phenocrysts from both Units 4 and 10 exhibit distinct zonation in cross polar-
ized light (e.g., Figs. 4.8, 4.9). The interiors of plagioclase phenocrysts frequently
exhibit oscillatory type zoning (e.g., crystal 10-1D-A; Fig. 4.9), and most have
obvious core and rim sections. Sections of zoned plagioclase crystals are divided
into three types: 1) cores, 2) intermediate zones, and 3) rims. Dissolution sur-
faces are abundant in crystals from both Units 4 and 10, and frequently enclose
crystal core regions (e.g., Fig. 4.10). Glass and partially crystallized inclusions are
present along some dissolution surfaces. It is important to note that the labeling
of cores and intermediate zones may be purely artificial, as geometric distortions
introduced during sectioning of basalt samples make it difficult to know when a
true crystal core is being sampled [140].
185

5 mm
5 mm
Figure 4.4. Cross polarized light petrographic thin section photographs
of the A) the Unit 4 basalt (183-1137A-32R-05W-22-27) examined in
this study, and B) the Unit 10 basalt (183-1137A-46R-01W-20-35)
examined in this study. Plagioclase phenocrysts in the Unit 4 basalt
occur as individual euhedral crystals and as glomerocrysts. Plagioclase
phenocrysts in the Unit 10 basalt occur most commonly as parts of
glomerocrysts.
186
A)
B)

187
TABLE 4.1
UNIT 4 PLAGIOCLASE MAJOR AND TRACE ELEMENT
ABUNDANCES
Sample Zone An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Sc Ti V Rb Sr Y Ba La Ce Pr Nd Sm Eu Pb
(mol%)
K C 69 0.62 0.12 0.18 0.05 3.47 14.29 0.16 51 31 100.95 2 642 3.5 0.37 829 0.33 99 0.92 2.2 0.21 0.77 0.07 0.59 0.76
K I 71 0.67 0.08 0.17 0.04 3.17 14.78 0.18 50.6 31.2 100.91 1.4 522 3.3 0.22 850 0.27 81 1.07 2.4 0.23 1.03 0.15 0.61 0.67
K R 66 0.67 0.05 0.21 0.09 3.78 13.7 0.18 51.5 30.8 100.99 1.9 565 3.1 0.17 839 0.17 79 0.95 2 0.2 0.68 0.09 0.59 0.6
L C 61 0.63 0.1 0.25 0.09 4.23 12.41 0.19 52.9 29.9 100.73 1.2 566 2.8 0.30 797 0.07 73 0.94 1.7 0.18 0.52 0.14 0.52 0.55
L I 62 0.68 0.1 0.27 0.1 4.23 12.8 0.2 52.7 29.7 100.84 1.4 630 3.4 0.42 804 0.33 84 0.77 1.7 0.16 0.48 0.14 0.63 0.67
L R 68 0.65 0.09 0.18 0.08 3.55 14.17 0.17 50.9 31 100.79 1.1 481 3.8 0.23 803 0.23 68 0.84 1.8 0.18 0.67 0.16 0.53 0.71
Q C 66 0.6 0.07 0.24 0.07 3.7 13.68 0.18 52 30.1 100.59 1.5 478 5.1 0.34 852 0.12 70 0.74 1.8 0.16 0.48 0.1 0.63 0.86
Q I 65 0.65 0.08 0.22 0.03 3.9 13.37 0.18 52.5 29.5 100.49 1.6 483 4 0.24 882 0.13 69 0.87 1.8 0.17 0.7 0.05 0.55 0.7
Q R 60 0.73 0.05 0.26 0.08 4.34 12.47 0.16 53.4 28.9 100.39 2.8 581 3.4 – 922 0.08 91 1.01 2 0.23 0.73 0.12 0.59 0.78
P C 66 0.65 0.1 0.21 0 3.63 13.34 0.18 52.1 30.2 100.48 2.1 445 3.8 0.34 856 0.14 68 0.69 1.5 0.16 0.48 0.07 0.51 0.89
P I 65 0.08 0.21 0 3.82 13.04 0.15 53 29.9 100.17 1.4 462 4.6 0.35 825 0.13 67 0.78 1.8 0.17 0.69 0.07 0.56 0.65
J C 68 0.6 0.04 0.2 0.08 3.6 14.43 0.2 51.3 30.4 100.92 1.5 499 4.4 0.30 924 0.09 71 0.86 1.8 0.2 0.61 0.08 0.58 0.44
J I 68 0.62 0.09 0.19 0.08 3.66 14.26 0.17 51.4 30.7 101.21 1.6 479 4.2 0.29 909 0.1 72 0.9 1.9 0.17 0.57 0.12 0.55 0.48
J R 62 0.79 0.08 0.25 0.07 4.23 12.91 0.19 53 29.4 100.9 1.8 580 3.9 0.34 925 0.09 98 1.05 1.9 0.2 0.68 0.05 0.73 0.64
N C 64 0.65 0.08 0.2 0.28 3.92 13.22 0.19 52.4 30 100.94 1.7 501 3.3 0.26 849 0.1 70 0.84 1.7 0.18 0.66 0.14 0.57 0.63
N R 61 0.8 0.06 0.26 0 4.23 12.71 0.19 52.9 29.7 100.88 1.5 845 6.9 0.59 1040 0.21 132 1.53 2.8 0.26 1.22 0.21 1.33 0.97
O C 67 1.24 0.07 0.2 0 3.57 13.57 0.18 51.1 31 100.97 1.2 429 3.8 0.31 874 0.13 69 0.84 1.8 0.14 0.68 0.1 0.61 0.54
O I 67 0.68 0.11 0.2 0.04 3.58 13.43 0.16 51.9 30.7 100.72 2.2 535 4.1 0.40 934 0.11 92 0.96 1.9 0.19 0.73 0.1 0.72 0.82
O I 57 0.71 0.06 0.3 0.09 4.6 11.6 0.16 54.2 29.1 100.73 1.5 660 3.5 0.51 905 0.1 91 0.92 2.1 0.21 0.8 0.1 0.55 0.53
O I 62 0.73 0.1 0.25 0.11 4.09 12.56 0.16 53 29.6 100.62 1.2 713 4.9 0.48 983 0.1 116 1.26 2.3 0.2 0.93 0.2 0.9 0.67
O R 54 0.87 0.09 0.33 0.13 4.83 10.67 0.38 55.2 28 100.46 2 785 7.2 0.69 917 0.13 125 1.15 2.1 0.2 0.72 0.15 0.98 0.69
D C 67 0.67 0.1 0.2 0 3.65 13.98 0.17 51.6 30.3 100.7 1.1 522 4 0.18 929 0.13 69 0.88 1.7 0.18 0.77 0.12 0.59 0.46
D R 64 0.6 0.08 0.21 0.07 3.96 12.9 0.19 52.1 30 100.13 1.5 534 4.5 0.34 906 0.12 71 0.88 1.6 0.17 0.56 0.1 0.64 0.48
A C 63 0.6 0.08 0.22 0.09 4.1 12.83 0.19 53 29.2 100.21 1.7 492 3 0.36 949 0.08 77 0.87 1.8 0.19 0.64 0.1 0.44 1.21
A I 64 0.65 0.06 0.22 0.08 4.07 13.28 0.19 52.7 29 100.24 2 519 3.6 0.40 929 0.09 84 0.82 2.1 0.2 0.64 0.07 0.57 0.53
A I 64 0.68 0.07 0.22 0.08 4.01 13.3 0.21 52.9 28.8 100.23 2.4 525 3.4 0.36 942 0.11 80 0.88 2.1 0.22 0.73 0.1 0.6 0.46
A R 64 0.71 0.08 0.23 0.09 3.85 12.84 0.17 52.8 28.9 99.66 2.2 474 3.9 0.33 836 0.1 73 0.74 1.8 0.16 0.6 0.12 0.6 0.46
C C 71 0.65 0.08 0.18 0.04 3.24 15.04 0.14 50.5 30.3 100.16 1.2 447 3.2 0.27 945 0.21 89 1.05 2.2 0.22 0.87 0.09 0.59 0.67
C I 66 0.6 0.07 0.2 0.09 3.8 13.78 0.18 52.3 29.2 100.15 1.5 502 3.5 0.31 904 0.15 74 0.99 1.9 0.18 0.78 0.11 0.55 0.64
C R 59 0.69 0.05 0.28 0.07 4.58 12.52 0.17 53.7 28.3 100.31 2.4 594 3.8 0.60 904 0.09 97 1 2 0.21 0.61 0.08 0.65 0.68
1 C = core; I = Intermediate Zone; R = rim
2 total Fe as FeO
3 [An = Ca / (Ca+Na+K)]*100

4.4.2 Electron Probe Microanalysis (EPMA) - Major Elements
The An-Ab-Or compositions of zones within Unit 4 and 10 plagioclase phe-
nocrysts that were targeted for trace element microanalysis are shown in figure 4.5.
The An content of Unit 4 plagioclase zones range from An54 to An71, and Unit
10 crystals span a narrower An57 to An67 range. Unit 4 plagioclase cores trend
to higher An content, and rims trend to lower An content and greater Or con-
tent (Fig. 4.5a). Intermediate zones are nearly identical to Unit 4 cores in terms
of An-Ab-Or composition (Fig. 4.5b). The lowest An zone sampled in a Unit 4
plagioclase was An54 along a crystal rim, whereas the highest An zones are an
An71 intermediate zone near a crystal core and an An71 core. Unit 10 plagioclase
cores and intermediate zones trend to lower An content than rim zones (Fig. 4.5b).
The highest An zones sampled in Unit 10 plagioclase phenocrysts are an An67 rim
zone and an An67 intermediate zone that is adjacent to a crystal rim, and the
lowest An zones sampled include two near-core An57 zones and an An58 crystal
core (Fig. 4.5b).
Core-to-rim EPMA point traverses reveal that each of the Unit 4 and 10 pla-
gioclase crystals analyzed in this study contain major element oscillations, which
is consistent with oscillatory zoning apparent in cross polarized light and backscat-
ter electron images (e.g., Figs. 4.9b,c). Half of the Unit 4 plagioclase phenocrysts
examined exhibit normal zoning patterns (Fig. 4.6b,d,g,h,j). Crystals from Units
4 and 10 have ubiquitous thin (5-10 µm thick) low An rims that are not obvious
for each crystal in figures 4.6 and 4.7. One Unit 4 crystal examined in this study
exhibits little significant core to rim change in An content (Fig. 4.6c). Four of the
ten Unit 4 plagioclase phenocrysts examined in this study exhibit reverse zoning
(Fig. 4.6a,e,f,i). Each of the Unit 10 plagioclase phenocrysts examined in this
188

Figure 4.5. A) An-Ab-Or compositions of Unit 4 plagioclase phenocryst
core, intermediate, and rim zones. Unit 4 rim zones trend to the lowest
An contents. B) An-Ab-Or compositions of Unit 10 plagioclase
phenocryst core, intermediate, and rim zones. Unit 10 plagioclase cores
trend to the lowest An contents of all zones sampled.
189

An (mol %)
An (mol %)
An (mol %)
An (mol %)
An (mol %)
crystal A
crystal C
crystal D
crystal J
crystal K
Unit 4
A)
B)
C)
D)
E)
crystal L
crystal N
crystal O
crystal P
crystal Q
1 mm 1 mm
core rim core rim
Figure 4.6. Core to rim profiles of ten Unit 4 plagioclase phenocrysts.
Other major and trace element data measured at various points along
the core to rim transects are presented in table 4.1.
190
F)
G)
H)
I)
J)

An (mol %)
An (mol %)
An (mol %)
An (mol %)
10-1D-A
10-1C-A
10-1C-B1
10-1C-B2
A)
B)
C)
D)
Unit
10
10-2A-C
10-1C-D
10-2A-D
2 mm 2 mm
core rim core rim
Figure 4.7. Core to rim profiles of seven Unit 10 plagioclase
phenocrysts. Other major and trace element data measured at various
points along the core to rim transects are presented in table 4.2.
191
F)
E)
G)

study exhibit their highest An at the crystal rim (Fig. 4.7). With the exception of
crystal 10-1C-B2 (Fig. 4.6d), each of the Unit 10 crystals exhibits reverse zonation
(Fig. 4.7). Crystal 10-1C-A exhibits complex zonation, but has an overall reversed
zoning pattern (Fig. 4.7b). In summary Unit 4 plagioclase phenocrysts are a mix-
ture of normal zoned crystals, reversely zoned crystals, and a few unzoned crystals.
Unit 10 plagioclase phenocrysts are are dominantly reversely zoned.
192

193
TABLE 4.2
UNIT 10 PLAGIOCLASE MAJOR AND TRACE ELEMENT
ABUNDANCES
Sample Zone An FeOT P2O5 K2O TiO2 Na2O CaO MgO SiO2 Al2O3 Total Sc Ti V Rb Sr Y Ba La Ce Pr Nd Sm Eu Pb
(mol%)
10-1C-A C 60 0.63 0.11 0.33 0.16 4.48 12.98 0.14 52.7 29.4 100.93 1.5 654 6 0.7 862 0.09 133 1.10 2.5 0.22 0.83 0.13 1.09 1.18
10-1C-A I 65 0.58 0.09 0.27 0.03 4.03 13.85 0.14 52 30.1 101.09 1.9 695 4.7 0.62 890 0.13 106 1.08 2.2 0.21 0.78 0.15 0.88 1.21
10-1C-A I 62 0.59 0.06 0.31 0.1 4.19 13.18 0.16 52.6 29.1 100.29 1.6 797 5.8 0.79 911 0.12 142 1.16 2.5 0.24 0.88 0.15 0.97 1.07
10-1C-A R 66 0.59 0.1 0.23 0.06 3.9 14.06 0.17 51.8 29.6 100.41 1.8 648 5.3 0.54 836 0.12 96.9 1.03 2.1 0.19 0.71 0.08 0.79 1.06
10-1C-B1 C 65 0.55 0.1 0.25 0.07 3.82 13.6 0.14 52.4 28.8 99.68 1.8 779 5.9 0.67 895 0.13 142 1.16 2.8 0.27 0.99 0.13 1.22 1.16
10-1C-B1 I 61 0.56 0.11 0.27 0.08 4.31 12.96 0.15 53.7 27.8 99.86 2.5 845 6.5 0.97 905 0.17 147 1.23 2.8 0.23 0.97 0.14 1.09 1.24
10-1C-B1 I 65 0.62 0.09 0.25 0.09 3.86 13.78 0.16 51.5 29.9 100.19 2.1 695 5.8 0.5 846 0.09 112 1.07 2.6 0.22 0.82 0.15 0.97 1.15
10-1C-B1 I 64 0.64 0.11 0.23 0.05 3.92 13.29 0.17 51.9 29.5 99.78 1.9 689 5.5 0.55 804 0.14 110 1.00 2.4 0.2 0.84 0.13 0.8 1.21
10-1C-B1 R 66 0.66 0.08 0.22 0.13 3.82 14.14 0.18 51.2 29.7 100.16 2.3 731 5.4 0.47 848 0.11 111 1.07 2.6 0.22 0.87 0.09 0.88 1.15
10-1C-B2 C 58 0.59 0.06 0.37 0.12 4.61 12.35 0.16 52.8 28.4 99.46 2.2 995 7.1 1.11 913 0.2 151 1.39 3.1 0.29 0.91 0.16 0.95 1.27
10-1C-B2 I 64 0.62 0.11 0.28 0.03 4.06 13.42 0.17 52.1 29 99.82 1.7 701 5.3 0.57 806 0.13 105 1.04 2.5 0.23 0.9 0.12 0.8 1.24
10-1C-B2 I 62 0.63 0.09 0.29 0.07 4.33 13.2 0.15 51.8 29.1 99.63 2.2 719 5.3 0.5 800 0.13 113 1.06 2.5 0.23 0.79 0.11 0.85 1.05
10-1C-B2 R 59 0.61 0.13 0.32 0.09 4.49 12.12 0.17 52.9 28.7 99.51 2.6 761 5.1 0.88 813 0.13 113 1.02 2.2 0.21 0.68 0.08 0.71 1.30
10-2A-D C 59 0.47 0.03 0.32 0.08 4.35 11.99 0.17 54.2 28.6 100.15 1 803 5 0.66 885 0.2 127 1.34 2.2 0.23 0.86 0.18 1.09 1.42
10-2A-D R 67 0.48 0.05 0.23 0.1 3.54 13.27 0.12 52.2 29.7 99.66 1.6 636 4.1 0.47 777 0.21 84.6 1.21 1.9 0.2 0.75 0.12 0.75 0.78
10-2A-C C 58 0.53 0.06 0.37 0.11 4.47 11.77 0.17 54.3 28.2 100 1 785 5 0.71 862 0.16 120 1.36 2.2 0.25 0.93 0.17 1.04 0.78
10-2A-C I 62 0.48 0.11 0.32 0.06 4.07 12.61 0.19 53.4 29.1 100.29 1.3 917 5.5 0.85 941 0.2 139 1.44 2.4 0.27 0.86 0.13 1.2 0.93
10-2A-C I 67 0.49 0.07 0.27 0.05 3.62 13.97 0.17 52 29.3 99.91 0.9 654 4.2 0.36 772 0.16 84 1.12 2 0.23 0.81 0.12 0.79 0.65
10-2A-C R 65 0.56 0.08 0.25 0.07 3.79 13.52 0.17 52.1 29.3 99.83 1.2 755 6.2 0.46 845 0.19 101 1.30 2.1 0.23 0.81 0.15 0.99 0.89
10-1C-D C 0 0.48 0.08 0.27 0.1 3.96 12.83 0.18 53 29.4 100.36 1.9 737 5.3 0.55 801 0.15 114 0.97 2.3 0.21 0.77 0.1 1.01 0.83
10-1C-D I 60 0.47 0.08 0.32 0.12 4.38 12.2 0.2 53.6 28 99.35 1.7 887 6.1 0.76 801 0.07 146 1.08 2.5 0.2 0.59 0.09 0.99 0.96
10-1C-D I 63 0.44 0.05 0.28 0.11 3.97 12.71 0.16 52.9 29.3 99.92 1.8 827 5.4 0.8 797 0.15 141 1.05 2.4 0.21 0.7 0.1 0.93 0.99
10-1C-D I 59 0.5 0.11 0.33 0.07 4.38 11.99 0.19 53.9 27.8 99.26 2.3 911 6.5 0.86 845 0.08 143 1.05 2.5 0.23 0.94 0.11 0.95 1.00
10-1C-D R 62 0.46 0.11 0.28 0.12 4.15 12.64 0.17 53.3 27.8 98.96 1.4 839 6.8 0.51 785 0.13 119 1.03 2.4 0.23 0.74 0.09 1.06 1.34
10-1D-A I 57 0.53 0.09 0.34 0.14 4.48 11.42 0.17 54.8 28.8 100.68 1.5 833 5.8 0.65 797 0.11 121 1.06 2.3 0.22 0.61 0.13 0.89 0.95
10-1D-A I 57 0.41 0.07 0.36 0.14 4.49 11.53 0.18 54.5 28 99.68 1.6 947 5.7 0.79 815 0.1 142 1.01 2.3 0.23 0.66 0.05 0.87 0.87
10-1D-A R 66 0.46 0.1 0.25 0.11 3.65 13.44 0.17 52.5 29 99.6 1.4 803 6.1 0.52 832 0.11 118 1.02 2.5 0.26 0.69 0.13 0.82 0.70
BCR-2G – – – – – 2.30 – – – – – – 33.1 – 402.2 46.9 335.4 33.5 706.9 26.2 52.6 6.9 28.7 6.7 2.0 10.0
– – – – – – ±0.36 – – – – – – ±0.74 – ±22.9±3.7 ±16.2±1.23±47.2±1.1 ±3.0 ±0.29±1.7 ±0.43±0.06±0.79
BCR-2
cert.
– – – – – 2.24 – – – – – – 32.6 – 407 47.5 337 32.5 684 25.3 53.6 6.8 28.6 6.7 2.0 10.34
1 C = core; I = Intermediate Zone; R = rim
2 total Fe as FeO
3 [An = Ca / (Ca+Na+K)]*100

4.4.3 LA-ICP-MS - Trace Elements
Anorthite content is a controlling factor of trace element substitution in pla-
gioclase (hence negative correlations with An content in Figs. 4.11 and 4.12; e.g.,
[10]), so comparisons of core, rim, and intermediate zones with similar An con-
tents are most informative. Given the narrow An range exhibited by Unit 4 and
10 crystals this manner of comparison is possible.
4.4.3.1 Scandium, Ti, V, and Pb
Negative correlations of Ti, V, and Pb abundance with An content are ap-
parent for crystals from both Units 4 and 10. There is little obvious correlation
of Sc abundance with An content for crystals from either Unit. (Fig. 4.11a-c,e;
Tables 4.4), 4.5). In terms of Sc abundance there is little distinction between Unit
4 and Unit 10 plagioclase crystals (Fig. 4.11a). For a given An content, Unit 10
phenocrysts have greater Ti, V, and Pb abundances than Unit 4 plagioclase crys-
tals (Fig. 4.11b,c,f). There is little distinction in Sc, Ti, V, and Pb abundances
between plagioclase zones from either Unit.
4.4.3.2 Barium, Sr, and Rb
. Crystals from Units 4 and 10 exhibit negative correlations of Ba and Rb with
An content and no correlation of Sr abundance with An content (Figs. 4.11d,e
and 4.12a). At a given An content, Unit 10 crystals have greater Ba and Rb
abundances than Unit 4 phenocrysts. Unit 4 and 10 phenocrysts have similar Sr
overlapping Sr abundance, but Unit 10 crystals tend to have lower Sr at a given An
content (Fig. 4.12a). Where Unit 10 phenocryst interiors (cores and intermediate
zones) and rims have overlapping An content, rims tend to have the lowest Ba
194

87
Sr/
86
SrI = 0.704937(10)
Ba/Sr = 0.08
Rb/Sr = 0.0004
crystal A
crystal J
87
Sr/
86
SrI = 0.704951(14)
Ba/Sr = 0.08
Rb/Sr = 0.0003
1 mm
87
Sr/
86
SrI = 0.704902(10)
Ba/Sr = 0.08
Rb/Sr = 0.0003
1 mm
87
Sr/
86
SrI = 0.704845(6)
Ba/Sr = 0.09
Rb/Sr = 0.0004
87
Sr/
86
SrI = 0.704849(12)
Ba/Sr = 0.07
Rb/Sr = 0.0002
87 86
Sr/ SrI = 0.704985(68)
Ba/Sr = 0.08
Rb/Sr = 0.0004
1 mm
A)
B)
crystal D
Figure 4.8. Diagrams of zoning features in the three of five Unit 4
plagioclase phenocrysts that were subjects of EPMA, LA-ICP-MS and
micro Sr isotope analyses. Thin dotted black lines mark the locations of
core to rim EPMA transects. 87 Sr/ 86 SrI results are shown for each
along with the measured 2σ error in parentheses. Measured Ba/Sr and
Rb/Sr ratios are shown for each zone examined. A) The core of crystal
A is bounded by a resorption surface that is marked with a dotted
white line. Anorthite content increases from ∼ An65 to > An70 (see
Fig. 4.6a). B)The core of crystal D is bounded by a resorption surface
with an a change from ∼ An63 to > An67 across the surface (see
Fig. 4.6c). C) Crystal J is a normal zoned crystal. There is a change in
the zoning pattern marked noted by a white dotted line but no obvious
resorption surface. (see also Fig. 4.6j).
195
C)

87 86
Sr/ SrI = 0.705722(8)
Ba/Sr = 0.12
Rb/Sr = 0.0007
Ba/Sr = 0.12
Rb/Sr = 0.0006
1 mm
87 86
Sr/ SrI = 0.705575(10)
Ba/Sr = 0.18
Rb/Sr = 0.0009
10-1D-A
87 86
Sr/ SrI = 0.705637(12)
Ba/Sr = 0.17
Rb/Sr = 0.0010
87 86
Sr/ SrI = 0.705536(8)
Ba/Sr = 0.14
Rb/Sr = 0.0006
87 86
Sr/ SrI = 0.705735(10)
Ba/Sr = 0.15
Rb/Sr = 0.0008
87 86
Sr/ SrI = 0.705685(6)
Ba/Sr = 0.16
Rb/Sr = 0.0009
Ba/Sr = 0.14
Rb/Sr = 0.0007
Ba/Sr = 0.18
Rb/Sr = 0.001
1 mm
1 mm
Ba/Sr = 0.15
Rb/Sr = 0.0006
87
Sr/
86
SrI = 0.705620(10)
Ba/Sr = 0.17
Rb/Sr = 0.0010
Ba/Sr = 0.15
Rb/Sr = 0.0008
A)
10-1C-A
B)
10-1C-D
Figure 4.9. Diagrams of zoning features in the two of five Unit 10
plagioclase phenocrysts that were subjects of EPMA, LA-ICP-MS and
micro Sr isotope analyses. Thin dotted black lines mark the locations of
core to rim EPMA transects. 87 Sr/ 86 SrI results are shown for each
along with the measured 2σ error in parentheses. Measured Ba/Sr and
Rb/Sr ratios are shown for each zone examined. A) Crystal 10-1C-A
contains four distinct zones that were sampled. These zones are visible
in the cross polarized light photo. These zones, which were each
sampled, are ∼An60 in the core, an ∼An65 intermediate zone, an ∼An60
intermediate zone, and an > An65 rim zone (see Fig. 4.7b). B) Crystal
10-1C-D contains abundant oscillatory zoning but has an overall reverse
zonation pattern and at least two zones, which were each sampled,
where An content is higher (up to An65 than the core An61−62 (see
Fig. 4.7f). C) Crystal 10-1D-A appears to be a fragment of a large
crystal with an oscillatory zoned interior. This crystal has an obvious
resorption surface near the rim that in accompanied by a change in
from An62−63 to > An66.
196
C)

87 Sr/ 86 SrI = 0.705542(8)
Ba/Sr = 0.14
Rb/Sr = 0.0008
87 Sr/ 86 SrI = 0.705510(8)
Ba/Sr = 0.11
Rb/Sr = 0.0006
2 mm
Ba/Sr = 0.15
Rb/Sr = 0.0009
2 mm
A)
10-2A-C
10-2A-D 87 Sr/ 86 SrI = 0.705572(10)
Ba/Sr = 0.14
Rb/Sr = 0.0007
Figure 4.10. Diagrams of zoning features in the three Unit 10
plagioclase phenocrysts that were subjects of EPMA, LA-ICP-MS and
micro Sr isotope analyses. Thin dotted black lines mark the locations of
core to rim EPMA transects. 87 Sr/ 86 SrI results are shown for each
along with the measured 2σ error in parentheses. Measured Ba/Sr and
Rb/Sr ratios are shown for each zone examined. A) Crystal 10-2A-C
has an ∼ An60 core bounded by a resorption surface and an An65−66
rim. Both the core and rim were sampled. B) Crystal 10-2A-D has an
∼ An60 core bounded by a dissolution surface and an An65−67 rim. Both
the core and rim were sampled.
197
B)

and Rb (Fig. 4.11d,e).
4.4.3.3 Yttrium and the Rare Earth Elements La, Ce, Pr, Nd, Sm, and Eu
The highest Y abundances were measured in Unit 4 crystal cores and inter-
mediate zones (Fig. 4.12b). There is no other obvious systematic variation in Y
abundance between phenocryst zones in crystals from either Unit. At a given An
content, Unit 10 phenocrysts have greater La, Ce, Pr, Nd, Sm, and Eu. Where core
and interior sections of Unit 4 phenocrysts overlap, they tend to have lower overall
REE abundances relative to the instances where there is no overlap (Fig. 4.12).
The opposite relationship holds true or Unit 10 phenocryst zones with overlap-
ping An, where core and intermediate zones tend to have greater REE abundances
(Fig. 4.12).
4.4.4 Sr Isotope Microdrilling
There are subtle variations in 87 Sr/ 86 SrI between crystals zones in Unit 4 and
10 crystals (Figs. 4.13, 4.8, 4.9, and 4.10; Table 4.3). Unit 4 crystals have much
less radiogenic 87 Sr/ 86 SrI than Unit 10 crystals (Fig. 4.13). The rim of Unit 4
crystal D bears the most radiogenic 87 Sr/ 86 SrI observed in a Unit 4 crystal and
exhibits a higher 87 Sr/ 86 SrI than all Site 1137 upper group whole-rock samples
reported by Ingle et al. [75] (Fig. 4.8b). The rim zone of crystal D is bounded by a
resorption surface and a change from An67 in the intermediate zone to An64 at the
rim. All other Unit 4 plagioclase zones overlap Site 1137 upper group whole-rock
87 Sr/ 86 SrI reported by Ingle et al. [75]. There is little other distinction in terms
of 87 Sr/ 86 SrI between core, intermediate, and rim zones (Fig. 4.13).
Site 1137 lower group whole-rock samples reported by Ingle et al. [75] extend
198

Sc (ppm)
Ti (ppm)
V (ppm)
An (mol %)
A)
B)
C)
Rb (ppm)
Ba (ppm)
Pb (ppm)
Unit 4 plagioclase cores
Unit 4 plagioclase intermediate zones
Unit 4 plagioclase rims
An (mol %)
Unit 10 plagioclase cores
Unit 10 plagioclase intermediate zones
Unit 10 plagioclase rims
Figure 4.11. Trace element abundances (in ppm) measured in core,
intermediate, and rim zones of Unit 4 and 10 plagioclase phenocrysts
plotted against anorthite. A) Sc, B) Ti C) V, D) Rb, E) Ba, and F) Pb.
199
D)
E)
F)

Sr (ppm)
Y (ppm)
La (ppm)
Ce (ppm)
An (mol %)
A)
B)
C)
D)
Pr (ppm)
Nd (ppm)
Sm (ppm)
Eu (ppm)
Unit 4 plagioclase cores
Unit 4 plagioclase intermediate zones
Unit 4 plagioclase rims
An (mol %)
Unit 10 plagioclase cores
Unit 10 plagioclase intermediate zones
Unit 10 plagioclase rims
Figure 4.12. Trace and rare earth element element abundances (in ppm)
measured in core, intermediate, and rim zones of Unit 4 and 10
plagioclase phenocrysts plotted against anorthite. A) Sr, B) Y C) La,
D) Ce, E) Pr, F) Nd G) Sm, and H) Eu.
200
E)
F)
G)
H)

TABLE 4.3
UNIT 4 AND 10 87 Sr/ 86 Sr DATA
Sample Zone
Site 1137 Unit 4 Plagioclase
87 Sr/ 86 SrM Rb/Sr
87 Sr/ 86 SrI
J C rep1 0.704940(12) – 0.704939
J C rep2 0.704965(14) – 0.704964
J C avg 0.704953(13) 0.0003 0.70495
J I 0.704903(10) 0.0003 0.704902
D C 0.704850(12) 0.0002 0.704849
D R 0.704987(68) 0.0004 0.704985
A C rep1 0.704941(12) – 0.704939
A C rep2 0.704936(8) – 0.704934
A C avg 0.704939(10) 0.0004 0.704937
A R 0.704847(6) 0.0004 0.704845
Site 1137 Unit 10 Plagioclase
10-1C-A C 0.705738(10) 0.0008 0.705735
10-1C-A I 0.705725(8) 0.0007 0.705722
10-1C-A I 0.705689(6) 0.0009 0.705685
10-2A-D C 0.705575(10) 0.0007 0.705572
10-2A-D R 0.705513(8) 0.0006 0.705510
10-2A-C C 0.705545(8) 0.0008 0.705542
10-1C-D I 0.705577(10) 0.0009 0.705575
10-1C-D I 0.705623(10) 0.0010 0.705620
10-1D-A I 0.705640(12) 0.0010 0.705637
10-1D-A R 0.705539(8) 0.0006 0.705536
1 A 107.7 Ma age of Site 1137 basalts from [42] was used to correct Sr isotope ratios for in-situ decay. I subscript denotes
initial 87 Sr/ 86 Sr ratio.
2 values in parentheses are errors reported as measured 2σ in the last two decimal places of 87 Sr/ 86 SrM.
3 C = core, I = intermediate zone, R = rim.
4 Rb/Sr measured at coincident microdrill site by LA-ICP-MS.
201

Unit 4
plagioclase rims
plagioclase intermediate zones
plagioclase cores
Site 1137 Upper Group whole-rock
Unit 10
plagioclase rims
plagioclase intermediate zones
plagioclase cores
Site 1137 Lower Group
whole-rock
0.704500 0.705000 0.705500 0.706000 0.706500
Figure 4.13. The range of 87 Sr/ 86 SrI of Unit 4 and Unit 10 phenocrysts
are shown relative to their respective host basalt groups as defined by
Ingle et al. [75]. Note that plagioclase phenocrysts from each unit are
similar to whole-rock 87 Sr/ 86 SrI reported by Ingle et al. [75]. Unit 10
plagioclase rims are less radiogenic but still overlap intermediate and
rim zones.
to more radiogenic 87 Sr/ 86 SrI than any Unit 10 plagioclase zone sampled in this
study (Fig. 4.13). Unit 10 plagioclase rims exhibit the least radiogenic 87 Sr/ 86 SrI
of the three zones (Fig. 4.13). The core of crystal 10-1C-A has the most radiogenic
87 Sr/ 86 SrI any Unit 10 crystal zone sampled (Fig. 4.9a), and the rim of crystal 10-
2A-D has the least 87 Sr/ 86 SrI of any Unit 10 crystal zone (Fig. 4.10b). In summary,
crystals from both Units 4 and 10 have 87 Sr/ 86 SrI similar to Site 1137 upper group
(less radiogenic) and lower group (more radiogenic) whole-rock samples (i.e., their
host basalts) reported by Ingle et al. [75].
4.4.5 Inferred Parent Magma Compositions
Plagioclase parent magmas from Units 4 and 10 have Ba/Sr ratios and Rb
abundances that overlap Site 1137 upper group and lower group basalts. (Fig. 4.14),
although this overlap is less apparent when a smaller subset of Site 1137 basalts
202

are considered (i.e., those where 87 Sr/ 86 Sr ratios were also measured (Fig. 4.14b).
Rare earth element abundances and ratios (including Y) in Unit 4 and 10 plagio-
clase parent magmas are distinctly lower than Site 1137 basalts (Tables 4.4, 4.5;
(Fig. 4.14c). The origins of these low abundances are explored below.
203

204
TABLE 4.4
UNIT 4 PLAGIOCLASE PARTITION COEFFICIENTS AND
INFERRED PARENT MAGMA COMPOSITIONS
Sample Zone An DRb DSr DY DBa DLa DCe DP r DNd DSm Rb Sr Y Ba La Ce Pr Nd Sm
(mol%)
K C 69 0.01 1.77 0.01 0.27 0.10 0.09 0.07 0.05 0.03 33.1 467.4 39.4 367.3 8.8 25.6 3.0 14.3 2.2
K I 71 0.01 1.70 0.01 0.25 0.10 0.09 0.07 0.05 0.03 20.0 499.7 31.9 327.6 10.4 27.3 3.3 18.9 4.8
K R 66 0.01 1.86 0.01 0.30 0.10 0.09 0.07 0.05 0.03 14.9 452.1 19.9 265.2 9.1 23.5 2.8 12.8 3.0
L C 61 0.01 2.00 0.01 0.35 0.11 0.09 0.07 0.05 0.03 25.6 397.6 8.2 208.4 8.8 19.9 2.6 10.0 4.9
L I 62 0.01 1.98 0.01 0.34 0.11 0.09 0.07 0.05 0.03 35.3 405.3 38.0 244.8 7.3 19.8 2.4 9.3 4.9
L R 68 0.01 1.79 0.01 0.28 0.10 0.09 0.07 0.05 0.03 20.6 448.0 26.5 246.0 8.0 20.3 2.6 12.4 5.0
Q C 66 0.01 1.85 0.01 0.29 0.10 0.09 0.07 0.05 0.03 30.0 461.2 13.6 238.6 7.1 20.1 2.3 9.0 3.3
Q I 65 0.01 1.89 0.01 0.31 0.11 0.09 0.07 0.05 0.03 20.8 465.9 15.4 221.7 8.3 21.1 2.5 13.3 1.6
Q R 60 0.01 2.02 0.01 0.36 0.11 0.09 0.07 0.05 0.03 0.0 456.4 8.8 254.1 9.5 23.2 3.4 14.1 4.0
P C 66 0.01 1.85 0.01 0.29 0.10 0.09 0.07 0.05 0.03 30.2 463.2 16.1 232.0 6.6 17.7 2.3 9.0 2.4
P I 65 0.01 1.90 0.01 0.31 0.11 0.09 0.07 0.05 0.03 30.7 435.2 14.8 216.3 7.4 20.1 2.4 13.1 2.4
J C 68 0.01 1.79 0.01 0.28 0.10 0.09 0.07 0.05 0.03 26.4 515.8 10.3 256.8 8.3 20.9 2.8 11.3 2.7
J I 68 0.01 1.81 0.01 0.28 0.10 0.09 0.07 0.05 0.03 25.4 502.8 11.6 255.6 8.6 21.8 2.5 10.6 3.8
J R 62 0.01 1.98 0.01 0.34 0.11 0.09 0.07 0.05 0.03 29.0 468.2 10.3 288.0 9.9 22.1 2.9 13.0 1.8
N C 64 0.01 1.90 0.01 0.31 0.11 0.09 0.07 0.05 0.03 22.9 446.4 11.8 223.1 8.0 19.6 2.6 12.5 4.5
N R 61 0.01 1.99 0.01 0.34 0.11 0.09 0.07 0.05 0.03 49.8 523.2 23.8 382.6 14.4 32.1 3.8 23.4 7.1
O C 67 0.01 1.83 0.01 0.29 0.10 0.09 0.07 0.05 0.03 27.3 478.7 15.2 240.9 8.1 20.3 2.1 12.6 3.3
O I 67 0.01 1.83 0.01 0.29 0.10 0.09 0.07 0.05 0.03 35.0 509.5 12.3 318.2 9.2 22.0 2.8 13.7 3.1
O I 57 0.01 2.12 0.01 0.40 0.11 0.09 0.07 0.05 0.03 41.8 426.6 11.0 229.6 8.6 23.9 3.1 15.7 3.4
O I 62 0.01 1.97 0.01 0.34 0.11 0.09 0.07 0.05 0.03 40.9 498.4 11.3 342.0 11.9 26.6 2.9 17.8 6.9
O R 54 0.01 2.23 0.01 0.44 0.11 0.09 0.07 0.05 0.03 55.1 411.6 14.8 282.2 10.7 24.2 3.0 14.4 5.6
D C 67 0.01 1.82 0.01 0.28 0.10 0.09 0.07 0.05 0.03 16.0 510.6 14.8 241.2 8.4 19.5 2.6 14.4 3.9
D R 64 0.01 1.93 0.01 0.32 0.11 0.09 0.07 0.05 0.03 29.6 470.3 13.4 220.0 8.4 17.9 2.4 10.6 3.3
A C 63 0.01 1.95 0.01 0.33 0.11 0.09 0.07 0.05 0.03 31.1 485.4 9.3 232.3 8.2 21.2 2.7 12.1 3.2
A I 64 0.01 1.93 0.01 0.32 0.11 0.09 0.07 0.05 0.03 34.0 482.2 10.4 259.1 7.8 23.9 2.9 12.2 2.4
A I 64 0.01 1.92 0.01 0.32 0.11 0.09 0.07 0.05 0.03 30.9 491.7 13.2 252.5 8.3 24.0 3.2 13.8 3.3
A R 64 0.01 1.91 0.01 0.32 0.11 0.09 0.07 0.05 0.03 28.2 437.0 11.9 229.7 7.0 20.8 2.2 11.4 4.0
C C 71 0.01 1.70 0.01 0.25 0.10 0.09 0.07 0.05 0.03 24.5 554.3 24.5 357.9 10.2 25.9 3.1 16.1 2.8
C I 66 0.01 1.85 0.01 0.30 0.10 0.09 0.07 0.05 0.03 27.0 487.4 17.0 250.3 9.4 21.5 2.6 14.6 3.6
C R 59 0.01 2.06 0.01 0.37 0.11 0.09 0.07 0.05 0.03 50.1 439.3 9.9 261.6 9.4 23.4 3.1 11.9 2.6
1 C = core; I = Intermediate Zone; R = rim
2 [An = Ca / (Ca+Na+K)]*100

205
TABLE 4.5
UNIT 10 PLAGIOCLASE PARTITION COEFFICIENTS AND
INFERRED PARENT MAGMA COMPOSITIONS
Sample Zone An DRb DSr DY DBa DLa DCe DP r DNd DSm Rb Sr Y Ba La Ce Pr Nd Sm
(mol%)
10-1C-A C 60 0.01 2.02 0.01 0.36 0.11 0.09 0.07 0.05 0.03 58.3 426.8 10.3 372.3 10.3 28.5 3.2 16 4.5
10-1C-A I 65 0.01 1.9 0.01 0.31 0.11 0.09 0.07 0.05 0.03 53.9 469.3 14.5 340.4 10.3 25.8 3 14.8 5
10-1C-A I 62 0.01 1.96 0.01 0.33 0.11 0.09 0.07 0.05 0.03 67.3 464.7 13.9 424.3 11 29.1 3.5 16.7 5
10-1C-A R 66 0.01 1.86 0.01 0.3 0.1 0.09 0.07 0.05 0.03 47 449.3 14.1 324.2 9.8 24.5 2.8 13.4 2.7
10-1C-B1 C 65 0.01 1.87 0.01 0.3 0.11 0.09 0.07 0.05 0.03 58.8 478.1 14.8 468.9 11 32.7 3.9 18.6 4.3
10-1C-B1 I 61 0.01 1.99 0.01 0.35 0.11 0.09 0.07 0.05 0.03 82.2 455.2 19.6 426 11.6 32.3 3.3 18.6 4.7
10-1C-B1 I 65 0.01 1.87 0.01 0.3 0.11 0.09 0.07 0.05 0.03 43.8 452.4 10.9 370.8 10.2 30.1 3.2 15.4 5
10-1C-B1 I 64 0.01 1.9 0.01 0.31 0.11 0.09 0.07 0.05 0.03 47.9 422.7 16.2 351 9.5 27.5 2.9 16 4.3
10-1C-B1 R 66 0.01 1.84 0.01 0.29 0.1 0.09 0.07 0.05 0.03 40.9 460 12.4 378.9 10.2 29.9 3.2 16.4 2.8
10-1C-B2 C 58 0.01 2.08 0.01 0.38 0.11 0.09 0.07 0.05 0.03 91.4 438.6 22.5 395.9 13 35.4 4.2 17.9 5.4
10-1C-B2 I 64 0.01 1.92 0.01 0.32 0.11 0.09 0.07 0.05 0.03 49.1 419.1 14.8 327 9.9 29 3.3 17.1 3.9
10-1C-B2 I 62 0.01 1.98 0.01 0.34 0.11 0.09 0.07 0.05 0.03 42.4 404.2 14.5 330.6 10 28.3 3.3 15.1 3.9
10-1C-B2 R 59 0.01 2.07 0.01 0.38 0.11 0.09 0.07 0.05 0.03 72.5 392.3 14.1 299.2 9.6 25.2 3.1 13.2 2.9
10-2A-D C 59 0.01 2.06 0.01 0.37 0.11 0.09 0.07 0.05 0.03 54.8 430.1 22.3 341.5 12.6 25.8 3.4 16.8 6.3
10-2A-D R 67 0.01 1.84 0.01 0.29 0.1 0.09 0.07 0.05 0.03 41 422.8 24.6 290.8 11.6 21.4 2.8 14 3.8
10-2A-C C 58 0.01 2.1 0.01 0.39 0.11 0.09 0.07 0.05 0.03 58.4 411.2 18.5 309.7 12.7 25.1 3.6 18.2 6
10-2A-C I 62 0.01 1.97 0.01 0.34 0.11 0.09 0.07 0.05 0.03 71.7 476.8 22.3 409.2 13.6 27.1 3.9 16.4 4.4
10-2A-C I 67 0.01 1.82 0.01 0.29 0.1 0.09 0.07 0.05 0.03 31.7 423.5 18.6 293.8 10.7 23 3.3 15.1 4
10-2A-C R 65 0.01 1.87 0.01 0.3 0.11 0.09 0.07 0.05 0.03 39.7 451.6 21.5 333.9 12.4 24.3 3.4 15.2 5
10-1C-D C 63 0.01 1.94 0.01 0.33 0.11 0.09 0.07 0.05 0.03 47.4 413.4 17 349.3 9.1 26.2 3 14.6 3.3
10-1C-D I 60 0.01 2.05 0.01 0.37 0.11 0.09 0.07 0.05 0.03 63 391 7.7 396.3 10.1 28.8 2.9 11.4 3.1
10-1C-D I 63 0.01 1.95 0.01 0.33 0.11 0.09 0.07 0.05 0.03 68.2 409.4 16.9 427.6 9.9 27.7 3.1 13.3 3.3
10-1C-D I 59 0.01 2.06 0.01 0.37 0.11 0.09 0.07 0.05 0.03 71 409.6 9.5 382.2 9.8 29.3 3.4 18.2 3.8
10-1C-D R 62 0.01 1.98 0.01 0.34 0.11 0.09 0.07 0.05 0.03 43 396.3 15 347.7 9.7 27.5 3.3 14.1 3
10-1D-A I 57 0.01 2.12 0.01 0.4 0.11 0.09 0.07 0.05 0.03 52.9 376.3 12.6 305.4 9.9 26.9 3.2 12 4.6
10-1D-A I 57 0.01 2.11 0.01 0.39 0.11 0.09 0.07 0.05 0.03 64.4 385.6 11.5 359.9 9.4 26.9 3.4 13 1.8
10-1D-A R 66 0.01 1.85 0.01 0.3 0.1 0.09 0.07 0.05 0.03 46 449.6 12.4 399.5 9.7 29.2 3.7 12.9 4.3
1 C = core; I = Intermediate Zone; R = rim
2 [An = Ca / (Ca+Na+K)]*100

4.5 Discussion
4.5.0.1 Origin of Major Element Zoning and Resorption Surfaces
Abrupt changes of 5 to > 10 mole % An and the presence of resorption sur-
faces amongst Unit 4 and 10 plagioclase phenocrysts indicates these crystals were
variably exposed to magmas with which they were at a state of disequilibrium
(e.g., Fig. 4.6, 4.7, Fig. 4.8, Fig. 4.9, Fig. 4.10). Injection of hot fresh magma
into a crystallizing magma chamber can lead to partial resorption and subsequent
growth of higher An zones, which is a feature that was noted for select crystals
from each Unit (e.g., Fig. 4.6a,e,f,p, Fig. 4.7). An alternative scenario is that mo-
bilization of crystals from mush piles within the magma chamber by convective
overturn, avalanches of wall and roof debris, or magmatic stirring generated these
features [83, 98]. Crystals growing in such mush layers are thermally and me-
chanically insulated from the hottest portions of the magma chamber and can be
out of equilibrium with magma in the chamber interior. Either of these processes
may have operated to generate some or all of the observed dissolution and zoning
patterns in Unit 4 and 10 plagioclase crystals. There are, however, systematic
differences in the dominant zoning patterns between the Unit 4 and 10 basalt
samples examined in this study.
The Unit 4 basalt examined in this study contains a fairly even mixture or
normal and reverse zoned crystals, which is a first order indication that crystals
from different portions of the magmatic system were entrained during or before
eruption. Deposition of plagioclase-rich crystal debris at one or more locations
(walls, floors, etc.) in the magmatic system prior to eruption of the Unit 4 basalt
is supported by the presence of plagioclase glomerocrysts, where after deposition
the crystals were partially compacted to later be mobilized as cognate clusters of
206

crystals (glomerocrysts: Fig. 4.4a). Despite the presence of glomerocrysts, most
Unit 4 plagioclase crystals are present as individual euhedral crystals. I suggest
that the normal zoned plagioclase phenocrysts bear a close genetic relationship
with the Unit 4 host basalt, whereas the reverse zoned crystals likely come from
disaggregated glomerocrysts that crystallized earlier and under different condi-
tions. The mixture of normal and reverse zoned crystals indicates that plagioclase
crystals were recycled throughout the magmatic system during the petrogenesis
of the Unit 4 basalt, which is significant in that the plagioclase crystals provide a
greater temporal record of magma evolution.
The majority of the plagioclase phenocrysts in the Unit 10 basalt examined in
this study exhibit reverse zonation, which indicates they may have initially grown
from a more evolved and/or crustally contaminated magma then exposed to a more
primitive magma just before or during eruption. A significant fraction of the large
plagioclase phenocrysts in the Unit 10 basalt sample examined in this study are
present within glomerocrysts (Fig. 4.4b). Individual Unit 10 plagioclase crystals
not present in glomerocrysts are generally sub-rounded and are probably from
disaggregated glomerocrysts. I suggest that a large amount of mushy plagioclase
dominated crystal debris was entrained prior to and during eruption of the Unit
10 basalt and is the primary source of plagioclase phenocrysts in this basalt.
Trace element and Sr isotope data provide a greater insight into these processes
and whether early formed Unit 4 and 10 crystals provide insight into temporal
variations in crustal assimilation.
207

4.5.0.2 Insights from Trace Elements and Inverted Parent Magma Compositions
Plagioclase parent magmas from Units 4 and 10 overlap whole-rock fields de-
fined by their respective Site 1137 upper and lower basalt groups in Ba/Sr vs Rb
space (Fig. 4.14). Whole-rock compositions tend to represent an integration of
all of the subsurface processes that were active during basalt petrogenesis, and a
number of studies have noted greater records of processes such as crustal assim-
ilation and exposure to primitive magmas within individual plagioclase crystals
[37, 135]. Although major element and zoning data indicate that the plagioclase
phenocrysts contain greater temporal records of magma evolution relative to the
whole-rock chemistry, this overlap in figure 4.14a indicates that peak crustal con-
tamination present in Units 4 and 10 pre-dated plagioclase crystallization.
Plagioclase parent magmas trend to lower La/Sm relative to their respective
host basalt groups (Fig. 4.14c), and inverted parent magma trace element and
REE abundances are generally lower than those of Site 1137 upper and lower group
basalts (Table 4.4 and Table 4.5). Low trace element abundances of plagioclase
parent magmas may indicate that plagioclase crystals grew from magmas that
were more depleted than their host basalts. This however seems unlikely given
the regular occurrence of reversely zoned plagioclase crystals in each basalt and the
overlap in figure 4.14a. An alternative origin for the low trace element abundances
of plagioclase parent magmas is the use of inaccurate partition coefficients and/or
trace element diffusion. It is difficult to gauge the accuracy of calculated partition
coefficients, but Bindeman et al. [7] and Blundy and Wood [10] pointed out that
An content is a dominant factor that affects cation substitution in plagioclase
and temperature has a minor affect. Although An variations and crystallization
temperatures are accounted for in partition coefficient calculations, it is possible
208

Ba/Sr
Ba/Sr
La/Sm
1.20
1.00
0.80
0.60
0.40
0.20
1.20
1.00
0.80
0.60
0.40
0.20
7.00
6.00
5.00
4.00
3.00
2.00
1.00
Unit 4 cores
Unit 4 interm. zones
Unit 4 rims
Unit 10 cores
Unit 10 interm. zones
Unit 10 rims
Site 1137 LG basalts
Site 1137 UG basalts
20 40 60 80 100 120
Rb
Unit 4 cores
Unit 4 interm. zones
Unit 4 rims
Unit 10 cores
Unit 10 interm. zones
Unit 10 rims
Site 1137 LG basalts
Site 1137 UG basalts
0.7045 0.705 0.7055 0.706 0.7065
Unit 4 cores
Unit 4 interm. zones
Unit 4 rims
Unit 10 cores
Unit 10 interm. zones
Unit 10 rims
Site 1137 LG basalts
Site 1137 UG basalts
87 Sr/ 86 SrI
0.7045 0.705 0.7055 0.706 0.7065
87 Sr/ 86 SrI
Figure 4.14. A) Inverted Unit 4 and 10 plagioclase parent magmas of
overlap fields defined by whole-rock data for Site 1137 upper and lower group
basalts. B) Inverted Unit 4 and 10 plagioclase parent magma Ba/Sr versus
87 Sr/ 86 SrI and Site 1137 basalt fields drawn using whole-rock data. C)
Inverted Unit 4 and 10 plagioclase parent magma La/Sm versus 87 Sr/ 86 SrI
and Site 1137 basalt fields drawn using whole-rock data reported by [74].
Plagioclase parent magma compositions overlap whole-rock basalt field well
in terms of 87 Sr/ 86 SrI but below or above whole-rock fields. The origin of
this anomaly is discussed in the text.
209

that other factors influenced the accuracy of calculated D values. For example,
Ba and Sr D values are probably the best constrained because they have been the
focus of more partitioning studies than lower abundance elements like the REE
(e.g., [7, 10]. My estimate of the trivalent strain-free partition coefficient (D +3
0 ;
c.f., [11]) is poorly constrained. A D +3
0 that is too low would lead to systematically
low estimates of Y and the REE (i.e., Table 4.4 and Table 4.5; Fig. 4.14c).
Trace element diffusion may have also played a role in generating the impres-
sion of low trace element and REE abundances in plagioclase parent magmas.
This possibility is explored in the next section.
4.5.0.3 Diffusive Redistribution of Trace Elements
Cherniak [27] and Cherniak and Watson [24] noted that at magmatic tem-
peratures divalent cations diffuse faster through the plagioclase structure than
trivalent cations. They also noted that diffusion occurs slower in An-richer pla-
gioclase. Given the intermediate An contents (∼ An54−71) of crystals examined
in this study, the possibility of trace element re-distribution via diffusion is rela-
tively high. Zellmer et al. [145] presented a test for diffusive equilibrium, which
as they noted does not lead to a flat compositional profile, but rather a profile
that mirrors the variation in trace element D values most heavily influenced by
An content. Equation 4.3 is a adaption of equation 2 of Zellmer et al. [145] that
describes partitioning of a trace element (n) between plagioclase compositions A
and B.
D A/B
n = CA n /CB n = DA n /DB n (4.3)
210

When diffusive equilibrium has been reached between two adjacent plagioclase
zones with An contents A and B the condition in equation 4.4 is met.
C A/B
n /D A/B
n = 1 (4.4)
This test was applied to each plagioclase crystal examined in this study for Sr
and La at texturally distinct transitions between intermediate zones and crystal
rims (Fig. 4.15). In case of Unit 4 crystal P, where an intermediate zone to rim
transition examination was not possible, a core-intermediate zone transition was
tested. The partition coefficient values listed in tables 4.4 and 4.5 were used for
the test. Two of the ten Unit 4 crystals (Q and P) are within error of the Sr
diffusive equilibrium line (i.e., equation 4.4; Fig. 4.15a), whereas only one Unit 4
crystal (P) is within error of the La diffusive equilibrium line (Fig. 4.15c). These
observations are consistent with the findings of Cherniak [27] that trivalent REE
diffuse slower in plagioclase than divalent cations like Sr.
Six of the seven Unit 10 crystals are within error of the Sr diffusive equilib-
rium line (Fig. 4.15b), whereas five of the seven crystals are within error of the
La diffusive equilibrium line (Fig. 4.15d). The greater proportion of diffusively
re-equilibrated plagioclase crystals in the Unit 10 basalt qualitatively suggests
that relative to the Unit 4 crystals, Unit 10 crystals were held at magmatic tem-
peratures for greater periods of time based upon the fact that diffusion occurs
more rapidly when crystals are held at magmatic temperatures (i.e., [27]). This
consistent with the working hypothesis that the majority of the Unit 4 plagioclase
phenocrysts bear a close genetic relationship with their host basalt, and the ma-
jority of the Unit 10 crystals were stirred up from a cumulate layer or other mush
pile.
211

C A/B / D A/B
C A/B / D A/B
Unit 4 O
A)
K
10-1D-A
L
Unit 10 B)
10-1C-A
Q
10-1C-B1
P
10-1C-B2
J
10-2A-C
N
10-1C-D
10-2A-D
D
A C
Sr
Sr
Figure 4.15. A) Two of ten Unit 4 crystals are within analytical
uncertainty (the propagated 2σ error when values are divided, e.g.,
Sr/Sr) of having reached Sr diffusive equilibrium, whereas one of ten
have reached C) La diffusive equilibrium. The diffusive equilibrium is
drawn based upon equation 4.4. B) Six of seven Unit 10 phenocrysts
are within error of having reached Sr diffusive equilibrium, whereas five
of seven crystal are within error of the La diffusive equilibrium line.
These results indicate La diffuses slower than Sr, which is consistent
with the findings of [27].
212
K
10-1D-A
L
10-1C-A
Q
10-1C-B1
P
10-1C-B2
J
10-2A-C
N
10-1C-D
O
10-2A-D
D
A
C
La
La
C)

4.5.0.4 Petrogenetic Insights from Plagioclase 87 Sr/ 86 SrI Ratios and the Timing
of Crustal Contamination at Elan Bank
Plagioclase crystals from Units 4 and 10 exhibit narrow ranges of 87 Sr/ 86 SrI
that overlap their respective Site 1137 upper and lower group basalts reported
by Ingle et al. [75], which indicates that crustal assimilation was not an ongoing
process during plagioclase dominated partial crystallization. Crustal rocks (e.g.,
garnet biotite gneiss clasts) recovered during drilling at Site 1137 are plausible
crustal assimilants and have distinctly radiogenic 87 Sr/ 86 SrI (> 0.784) [74]. If
these in fact were the types of crustal rocks assimilated, as favored by Ingle [75],
assimilation must have occurred prior to plagioclase dominated partial crystal-
lization. Core, intermediate zone, and rim 87 Sr/ 86 SrI variations are subtle, which
indicates that the Unit 4 and 10 plagioclase crystals grew from already contami-
nated parent magmas. This may also suggest that over time the magma chamber
walls became well armored, which minimized late stage pulses of greater crustal
assimilation.
4.5.0.5 Where Did Plagioclase Dominated Partial Crystallization Occur?
In their study of Kerguelen Archipelago lavas Damasceno et al. [36] sug-
gested that basalts that make up the Mont Crozier section passed through a
magmatic system that consisted of a complex arrangement interconnected con-
duits and chambers. They suggested that plagioclase-dominated partial crystal-
lization occurred primarily in shallow sub-volcanic magma reservoirs [36]. Elan
Bank magmas passed through a section of crust with both oceanic and continental
components (versus oceanic crust for Mont Crozier magmas) [36], and it is thus
probable Elan Bank magmas passed through similarly complex magmatic system.
213

Magma chambers tend to form at horizons of neutral buoyancy, and the depth
of the neutral buoyancy zone is closely linked to the density of the surrounding
country rock [122]. Continental crust is seldom lithologically uniform on a vertical
scale, and the formation of multiple magma chambers over time at different hori-
zons of neutral buoyancy is more probable within continental crust versus oceanic
crust. Indeed, Marsh [100] discussed such a magmatic system consisting of inter-
connected dikes and sills similar to the system inferred for Kerguelen Archipelago
by Damasceno et al. [36] that penetrated through Gondwanan crust in nearby
Antarctica [35], which he referred to as the Ferrar magmatic mush column. Marsh
[100] cited evidence of kinetic sieving and flow sorting of plagioclase and orthopy-
roxene which lead to segments of the magma chamber system to be filled with
large volumes of these minerals.
4.6 Summary and Conclusions
I propose that, much like the Ferrar magmatic mush column and Kergue-
len Archipelago examples, the magma chamber system below Elan Bank was a
complex system of interconnected dikes and sills. I suggest that the abundant
plagioclase phenocrysts In the Unit 4 and 10 basalts originated from the shal-
lowest magma chambers within this system where crustal contamination was an
insignificant process. Crustal assimilation was most prominent during the initial
emplacement of basaltic magma into Elan Bank crust, and variations in the crustal
contamination signature over time may be as much linked to re-mobilization and
partial resorption of crystal mush material formed directly following initial Elan
Bank magma emplacement as progressive erosion of crustal wall rock by repeated
injection of hot and primitive magmas into Elan Bank magma chambers.
214

CHAPTER 5
SUMMARY, CONCLUSIONS, AND FUTURE WORK
5.1 Results and Conclusions
5.1.1 The Ontong Java Plateau
This research focused upon understanding the physical processes (i.e., magma
chamber dynamics) of differentiation that led to the repeated production of the
same basalt type (Kwaimbaita basalt) over a Greenland-size geographic area?
Plagioclase-rich cumulate xenoliths consisting of An-rich (up to An86) plagioclase
crystals from provide insight into these processes beyond what has been learned
from whole-rock studies alone. Anorthie-rich sections of cumulate xenolith crystals
and phenocrysts in host basalts were formed primarily by crystallization in shal-
low (low pressure, 2-7 km depth) regions of the OJP magma chamber system from
more primitive magmas than those basalts sampled from the OJP that were low in
H2O and at relatively high temperature (liquidus temperature near 1200 ◦ C). Evi-
dence from this work shows that the role of H2O-rich evolved boundary layer inter-
stitial melts in the formation of An-rich plagioclase was, at best, minor. This is in
contrast the interpretations of a significant role for water by Sano and Yamashita
[123], which was based upon plagioclase major element observations. Parent liquid
compositions derived from OJP xenolith and phenocryst plagioclase crystals con-
tain a wider compositional record of magma evolution than that revealed by OJP
215

whole-rock basalt data. Cumulate xenoliths and phenocrysts generally are pieces
of disrupted solidification fronts. Solidification fronts would have been ubiquitous
throughout the OJP magma chamber system. The OJP magma chamber system
was composed of interconnected dikes and sills and consisted of regions dominated
by liquid (magma chamber interior) and regions dominated by crystal-liquid mush
(along magma chamber floors and walls, within dikes and conduits). Crystals and
crystalline debris were extensively recycled throughout the OJP magma chamber
system. This recycling accounts for the wide compositional spectrum of magmas
parental to plagioclase xenolith crystals and phenocrysts with An content out of
equilibrium with their Kwaimbaita host basalts. Ironically, this process also ac-
counts for the dominance of a single magma composition erupted on to the OJP,
as it represents the homogenized body of magma in the main part of the up-
per chamber. The compositions of melts periodically flushed from the interstices
of the crystal-mush network were affected by their residence times in the mush.
Depending upon the extent of the crystal-rich portions of the magma chamber
system, residence time may have had a large effect on overall basalt chemistry.
The shallow crustal OJP magma chamber system corresponded with a zone of
neutral buoyancy in the 0-7 km depth range, as in other volcanic systems [122].
As OJP volcanism progressed the plateau gained elevation, and the zone of neutral
buoyancy gradually migrated upward leading to slow yet pervasive assimilation of
overlying seawater altered basalt.
The greatest diversity of OJP basalts is along the margins of the plateau
where it is thinner. During formation, magmas ascending into the marginal and
thinner magma chamber system around the plateau margins were subject to less
density filtration, which allowed ascent and eruption of relatively dense Kroenke
216

magmas and slightly less dense Singgalo magmas (see Fig. 2.1 for magma types and
distribution across the OJP). Slight variations in paths of magma ascent, depths
of magma pooling and crystallization, as well as rates and amounts of assimilation
of seawater altered basalt would have affected the isotopic and incompatible trace
element characteristics of known OJP basalts. I suggest that these physical factors
may account for both the homogeneity of basalts in the central portion of the OJP
and the relative heterogeneity of basalts around the plateau margin.
5.1.2 Detroit Seamount, Emperor Seamount Chain
A key objective of this study was to elucidate the roles that shallow magmatic
processes played during the petrogenesis of depleted hotspot basalts at Detroit
Seamount when the Hawaiian hotspot was near a mid-ocean ridge [34]. I employed
an integrated approach to understand the physical growth histories and chemical
provenance of plagioclase crystals in four pillow basalts from Detroit Seamount.
This was done by using plagioclase crystal size distributions to identify unique
crystal populations and compositional microanalysis to better under stand their
origins.
I suggest that the Detroit Seamount magma chamber system was complex and
consisted of extensive mush zones and multiple interconnected chambers. There is
an undeniable role for plagioclase fractionation and accumulation in the petrogen-
esis of Detroit Seamount tholeiitic basalts. Evidence of mixing between relatively
depleted and enriched end-member magmas recorded in the hybrid Unit 14 basalt
do not contradict existing models seeking to explain the origin of the depleted
component in Detroit Seamount basalts. Unit 14 magma mixing does suggest,
however, that subtle variations in partial melting or source compositions do occur
217

over short time spans but may be easily overlooked in whole-rock compositional
studies. I suggest this variation is consistent with subtle variations in partial
melting of a heterogeneous mantle source that consisted of relatively depleted
and relatively enriched components (c.f., [50, 72, 119]) Subtle variations in partial
melting are consistent with enhanced mantle flow and complex mantle dynamics
in the near ridge hotspot environment as discussed by Pearce [114]. The nature
of the depleted component in Detroit Seamount basalts is prominent in tholei-
itic basalts from Site 884, and perhaps most prominent in the cm-size plagioclase
phenocrysts. These large crystals have low Sr parent magmas suggesting they
were less evolved than their host basalt. A fruitful continuation of this research
would be to conduct Sr and Pb isotope studies on these large crystals by in-situ
microdrilling (e.g., [39]) or plagioclase separation (e.g., [16]).
5.1.3 Elan Bank, Kerguelen Plateau
The main objective of this study was to elucidate the timing and dynamics of
crustal contamination of a basaltic magma. This was done via a compositional mi-
croanalytical dissection of plagioclase phenocrysts in two Elan Bank basaltic lavas
- one contaminated with crust (Unit 10), and the other relatively uncontaminated
(Unit 4; as reported by [75]).
Plagioclase crystals from Units 4 and 10 exhibit narrow ranges of 87 Sr/ 86 SrI
that overlap their respective Site 1137 upper and lower group basalts reported
by Ingle et al. [75], which indicates that crustal assimilation was not an ongoing
process during plagioclase dominated partial crystallization. Crustal rocks (garnet
biotite gneiss clasts) recovered during drilling at Site 1137 are plausible crustal
assimilants and have distinctly radiogenic 87 Sr/ 86 SrI (> 0.784) [74]. If these in fact
218

were the types of crustal rocks assimilated, as favored by Ingle [75], assimilation
must have occurred prior to plagioclase dominated partial crystallization. Core,
intermediate zone, and rim 87 Sr/ 86 SrI variations are subtle, which indicates that
the Unit 4 and 10 plagioclase crystals grew from already contaminated parent
magmas. This may also suggest that over time the magmatic system became well
buffered, which minimized late stage pulses of greater crustal assimilation.
Inverted parent magma compositions from both Units 4 and 10 are offset from
ranges of trace and REE abundances of their respective whole-rock basalt groups.
This indicates either poor choices of partition coefficients and/or that diffusion
of trace elements occurred within Unit 4 and 10 plagioclase crystals. Indeed, a
greater proportion of crystals from the Unit 10 basalt are diffusively re-equilibrated
relative to the Unit 4 crystals. This indicates that Unit 10 crystals were held at
magmatic temperatures for greater periods of time. This consistent with the
working hypothesis that the majority of the Unit 4 plagioclase phenocrysts bear
a close genetic relationship with their host basalt, and the majority of the Unit
10 crystals were stirred up from a cumulate layer or other mush pile.
I propose that, much like the Ferrar magmatic mush column and Kerguelen
Archipelago examples, the magma chamber system below Elan Bank was a com-
plex system of interconnected dikes and sills (c.f., [36, 100]. I suggest that the
abundant plagioclase phenocrysts In the Unit 4 and 10 basalts originated from
the shallowest magma chambers within this system where crustal contamination
was an insignificant process. Crustal assimilation was most prominent during
the initial emplacement of basaltic magma into Elan Bank crust, and variations
in the crustal contamination signature over time may be as much linked to re-
mobilization and partial resorption of crystal mush material formed directly fol-
219

lowing initial Elan Bank magma emplacement as progressive erosion of crustal wall
rock by repeated injection of hot and primitive magmas into Elan Bank magma
chambers.
5.2 Recommendations for Future Work
Once one embarks upon the type of work detailed in this dissertation it does
not take one long to make the realization that igneous rocks (and thus igneous
systems) are far more complex than they are commonly modeled. The notion
that mantle source characteristics are preserved in magmas from mantle to erup-
tion indeed may be true in some settings, but there is overwhelming evidence
that magma chamber systems are complex, ubiquitous, and tend to buffer the
compositions erupted magmas (e.g., [100]. Indeed, this level of magma chamber
complexity is apparent in the three studies described in this dissertation. The im-
portance of bulk-rock compositional studies should not be understated, however
an entire new level of insight into the genesis, evolution, and eruption of magmas
is on the horizon thanks to technological advances that make it possible conduct
accurate and precise compositional studies on the smallest of samples. Innovative
implementations of the new and advanced microanalytical techniques will bring
about a paradigm shift in igneous petrology and volcanology.
With increased use of techniques such as LA-ICP-MS to measure low abun-
dance elements within minerals comes an increased need for accurate partition
coefficients. More accurate partition coefficients will improve estimates of parent
magma compositions inverted from minerals such as plagioclase and will improve
the accuracy of trace element diffusion calculations within minerals, which has
proven a useful way to estimate magmatic residence time [145]. In plagioclase
220

trivalent REE diffuse slower that the more abundant divalent cations (i.e., Sr and
Ba). This indicates that investigation of REE diffusion in plagioclase may provide
a more extensive time record of magmatic residence.
Although plagioclase trace element studies have proven to be useful, perhaps
the most robust future petrologic insights will come from intramineral isotope
studies beyond Sr (including Pb, Sm-Nd for example). Mineral scale isotope
studies are limited by the amount of the element of interest in the mineral. For
example, the greater abundance of Sr in plagioclase versus clinopyroxene makes
Sr isotope microanalysis of plagioclase much less difficult. Improvement of ion-
ization efficiency in TIMS analysis can help alleviate some of this problem (e.g.,
discussion in [40]), as can instrumentation advances that improve signal intensity
at a given concentration (e.g., guard electrodes). One of the difficulties faced
with a microdrill-TIMS approach is the high incidence of fracturing, flaking, and
inadvertent sampling of undesired zones. Perhaps one of the surest ways to im-
prove the controlled removal of mineral material for isotopic analysis is through
the use of laser ablation [40]. Laser ablation based isotopic studies generally re-
quires interfacing the LA system with a multicollector ICP-MS. The use of a Ar
based plasma introduces a host of issues and interferences not encountered with
TIMS. For example Ar used in ICP-MS commonly contains trace Kr, of which 86 Kr
(17.37% abundance) represents a particularly problematic polybaric interference.
At present a variety of corrections are necessary to obtain close to meanigful
87 Sr/ 86 Srm by MC-LA-ICP-MS. If mineral standards were available for isotopic
analyses, quality control of MC-LA-ICP-MS work would be more straightforward.
221

BIBLIOGRAPHY
1. C. Allegre, A. Provost and C. Jaupart, Oscillatory zoning: a pathological
case of crystal growth. Nature, 294: 223–228 (1981).
2. P. Asimow and M. Ghiorso, Algorithmic modifications extending MELTS to
calculate subsolidus phase relations. American Mineralogist, 83(1127-1131)
(1998).
3. K. Bartels, R. Kinzler and T. G. and, High pressure phase relations of primitive
high-alumina basalts from Medicine Lake volcano, northern California.
Contributions to Mineralogy and Petrology, 108: 253–270 (1991).
4. G. Bergantz, Melt fraction diagrams: the link between chemical and transport
models. In J. Nicholls and J. Russell, editors, Modern methods of igneous
petrology: Understanding magmatic processes, volume 24 of Reviews in Mineralogy,
pages 239–258, Mineralogical Society of America (1990).
5. I. Bindeman and J. Bailey, Trace elements in anorthite megacrysts from the
Kurile Island Arc: A window to across-arc geochemical variations in magma
compositions. Earth and Planetary Science Letters, 169: 209–226 (1999).
6. I. Bindeman and A. Davis, Trace element partitioning between plagioclase
and melt: Investigation of dopant influence on partition behavior. Geochimica
et Cosmochimica Acta, 63: 2863–2878 (2000).
7. I. Bindeman, A. Davis and M. Drake, Ion microprobe study of plagioclasebasalt
partition experiments at natural concentration levels in trace elements.
Geochimica et Cosmochimica Acta, 62(7): 1175–1193 (1998).
8. I. Bindeman, A. Davis and S. Wickham, 400 m.y. of basic magmatism in
a single lithospheric block during cratonization: ion microprobe study of
plagioclase megacrysts in mafic rocks from transbaikalia, russia. Journal of
Petrology, 40: 807–830 (1999).
9. J. Blundy, Experimental study of a Kiglapait marginal rock and implications
for trace element partitioning in layered intrusions. Chemical Geology, 141:
73–92 (1997).
222

10. J. Blundy and B. Wood, Crystal-chemical controls on the partitioning of
Sr and Ba between plagioclase feldspar, silicate melts, and hydrothermal
solutions. Geochimica et Cosmochimica Acta, 55: 193–209 (1991).
11. J. Blundy and B. Wood, Prediction of crystal-melt partition coefficients.
Nature, 372(1): 452–454 (1994).
12. J. Blundy and B. Wood, Partitioning of trace elements between crystals and
melts. Earth and Planetary Science Letters, 210: 383–397 (2003).
13. J. Brady, Diffusion data for silicate minerals, glasses, and liquids. In
T. Ahrens, editor, A Handbook of Physical Constants, volume 2 of Reference
Shelf , pages 269–290, AGU (1995).
14. A. Brandon, M. Norman, R. Walker and J. Morgan, 186 Os/ 187 Os systematics
of Hawaiian picrites. Earth and Planetary Science Letters, 174: 25–42 (1999).
15. J. Brice, Some thermodynamic aspects of the growth of strained crystals.
Journal of Crystal Growth, 28: 249–254 (1975).
16. J. Bryce and D. DePaolo, Pb isotopic heterogeneity in basaltic phenocrysts.
Geochimica et Cosmochimica Acta, 68(4453-4468) (2004).
17. I. Campbell and R. Griffiths, Implications of mantle plume structure for the
evolution of flood basalts. Earth and Planetary Science Letters, 99: 79–93
(1990).
18. K. Cashman, Textural constraints on the kinetics of crystallization of igneous
rocks. In J. Nicholls and J. Russell, editors, Modern Methods of Igneous
Petrology, Understanding Magmatic Processes, volume 24 of Reviews
in Mineralogy, pages 292–305, Mineralogical Society of America (1990).
19. K. Cashman, Relationship between plagioclase crystallization and cooling
rate in basaltic melts. Contributions to Mineralogy and Petrology, 113: 126–
142 (1993).
20. K. Cashman and B. Marsh, Crystal size distribution (CSD) in rocks and the
kinetics and dynamics of crystallization II: Makaopuhi lava lake. Contributions
to Mineralogy and Petrology, 99: 292–305 (1988).
21. L. Chambers, M. Pringle and J. Fitton, Age and duration of magmatism
on the Ontong Java Plateau: 40 Ar/ 39 Ar results from ODP Leg 192. EOS
Transactions, 83: F47 (2002), Abstract V71B-1271.
223

22. B. Charlier, J. Davidson, O. Bachmann and M. Dungan, Mineral-scale Sr
and Pb isotopic variations as recorders of magma differentiation processes
in the Fish Canyon magmatic system, San Juan Volcanic Field, U.S.A. EOS
Transactions AGU, 84: V11F–05 (2003).
23. B. Charlier, C. Wilson, J. Lowenstern, S. Blake, P. V. Calsteren and J. Davidson,
Magma generation at a Large, Hyperactive Silicic Volcano (Taupo, New
Zealand) Revealed by U-Th and U-Pb Systematics in Zircons. Journal of
Petrology, 46(1): 3–32 (2005).
24. D. Cherniak, A study of strontium diffusion in plagioclase using Rutherford
backscattering spectroscopy. Geochimica et Cosmochimica Acta, 38: 5179–
5190 (1994).
25. D. Cherniak, Diffusion of lead in plagioclase and K-feldspar: an investigation
using Rutherford Backscattering and resonant nuclear reaction analysis.
Contributions to Mineralogy and Petrology, 120: 358–371 (1995).
26. D. Cherniak, Ba diffusion in feldspar. Geochimica et Cosmochimica Acta, 66:
1641–1650 (2002).
27. D. Cherniak, REE diffusion in feldspar. Chemical Geology, 193: 25–41 (2003).
28. D. Cherniak and E. Watson, A study of strontium diffusion in K-feldspar,
Na-K feldspar and anorthite using Rutherford backscattering spectroscopy.
Earth and Planetary Science Letters, 113: 411–415 (1992).
29. D. Clague and G. Dalrymple, The hawaiian-emperor volcanic chain: Part I,
geologic evolution. In R. Decker, T. Wright and P. Stauffer, editors, Volcanism
in Hawaii, volume P 1350, pages 5–73, U.S. Geological Survey Professional
Paper (1987).
30. M. Coffin. Personal Communcation.
31. M. Coffin and O. Eldholm, Scratching the surface: Estimating the dimensions
of large igneous provinces. Geology, 21: 515–518 (1993).
32. M. Coffin and O. Eldholm, Large igneous provinces: crustal structure, dimensions,
and external consequences. Reviews of Geophysics, 32: 1–36 (1994).
33. M. Coffin, F. Frey, P. Wallace and et al., Proceedings of the Ocean Drilling
Program, Initial Reports, 183. [CD-ROM], Available from: Ocean Drilling
Program, Texas A&M University, College Station TX 77845-9547, USA
(2000).
224

34. R. Cottrell and J. Tarduno, A Late Cretaceous pole for the Pacific plate:
implications for apparent and true polar wander and the drift of hotspots.
Tectonophysics, 362(321-333) (2003).
35. K. Cox, Flood basalts, subduction and the breakup of gondwanaland. Nature,
274: 47–49 (1978).
36. D. Damasceno, J. Scoates, D. Weis, F. Frey and A. Giret, Mineral chemistry
of mildly alkalic basalts from the 25 ma mont. crozier section, kerguelen
archipelago: Constraints on phenocryst crystallization environments. Journal
of Petrology, 43(7): 1389–1413 (2002).
37. J. Davidson, J. Hora, J. Garrison and M. Dungan, Crustal forensics in arc
magmas. Journal of Volcanology and Geothermal Research, 140: 157–170
(2005).
38. J. Davidson and F. Tepley, Recharge in Volcanic Systems: Evidence from
Isotope Profiles of Phenocrysts. Science, 275: 826–829 (1997).
39. J. Davidson, F. Tepley and K. Knesel, Isotopic fingerprinting may provide
insights into evolution of magmatic systems. EOS Transactions AGU, 79:
185,189,193 (1998).
40. J. Davidson, F. Tepley, Z. Palacz and S. Meffan-Main, Magma recharge,
contamination and residence times revealed by in situ laser ablation isotopic
analysis of feldspar in volcanic rocks. Earth and Planetary Science Letters,
184: 427–442 (2001).
41. W. Deere, R. Howie and J. Zussman, Rock-forming minerals: Framework
silicates feldspars, volume 4A. The Geological Society of London, second
edition (2001).
42. R. Duncan, A time frame for construction of the Kerguelen Plateau and
Broken Ridge. Journal of Petrology, 43(7): 1109–1119 (2002).
43. R. Duncan and R. Keller, Radiometric ages for basement rocks from the
Emperor Seamounts, ODP Leg 197. Geochemistry, Geophysics, Geosystems,
5 (2004), Paper number: 10.1029/2004GC000704.
44. S. Eggins and J. Shelley, Compositional heterogeneity in NIST SRM 610-617
glasses. Geostandards Newsletter, 26(3): 269–286 (2002).
45. C. Farnetani, M. Richards and M. Ghiorso, Petrological models of magma
evolution and deep crustal structure beneath hotspots and flood basalt
provinces. Earth and Planetary Science Letters, 143(81-94) (1996).
225

46. J. Fitton and M. Godard, Origin and evolution of magmas on the Ontong
Java Plateau. In J. Fitton, J. Mahoney, P. Wallace and A. Saunders, editors,
Origin and Evolution of the Ontong Java Plateau, volume 229 of Special
Publications, pages 151–178, Geological Society, London (2004).
47. J. Fitton, A. Saunders, P. Kempton and B. Hardarson, Does depleted mantle
form and intrinsic part of the Iceland plume? Geochemistry, Geophysics,
Geosystems, 4(3): doi:10.1029/2002GC000424 (2003).
48. G. Foulger, Making the evidence fit the plume. Geoscientist (2003).
49. F. Frey, M. Coffin, P. Wallace, D. Weis, X. Zhao, S. Wise, V. Wähnert,
D. Teagle, P. Saccocia, D. Reusch, M. Pringle, K. Nicolaysen, C. Neal,
R. Müller, C. Moore, J. Mahoney, L. Keszthelyi, H. Inokuchi, R. Duncan,
H. Delius, J. Damuth, D. Damasceno, H. Coxall, M. Borre, F. Boehm, J. Barling,
N. Arndt and M. Antretter, Origin and evolution of a submarine large
igneous province: the kerguelen plateau and broken ridge, southern indian
ocean. Earth and Planetary Science Letters, 176: 73–89 (2000).
50. F. Frey, S. Huang, J. Blichert-Toft, M. Regelous and M. Boyet, Origin of
Depleted Components in Lavas Related to the Hawaiian Hotspot: Evidence
From Hf Isotope Data. Geochemistry, Geophysics, Geosystems, 6(1): doi:
10.1029/2004GC000757 (2005).
51. F. Frey, D. Weis, A. Borisova and G. Xu, Involvement of continental crust in
the formation of the Cretaceous Kergeulen Plateau: new perspectives from
ODP Leg 120 Sites. Journal of Petrology, 43(7): 1207–1239 (2002).
52. B. Fryer, S. Jackson and H. Longerich, The design, operation and role of
laser-ablation microprobe coupled with an inductively coupled plasma - mass
spectrometer (LAM-ICP-MS) in the Earth sciences. Canadian Mineralogist,
33: 303–312 (1995).
53. A. Furomoto, D. Hussong, J. Campbell, G. Sutton, A. Malahoff, J. Rose and
G. Woollard, Crustal and upper mantle structure of the Solomon Islands as
revealed by seismic refraction survey of November-December 1966. Pacific
Science, 24: 315–332 (1970).
54. A. Furomoto, J. Webb, M. Odegard and D. Hussong, Seismic studies of the
Ontong Java Plateau, 1970. Tectonophysics, 34: 71–90 (1976).
55. M. Ghiorso and R. Sack, Chemical mass transfer in magmatic processes, IV.
A revised an internally consistent thermodynamic model for interpolation
and extrapolation of liquid-solid equilibria in magmatic systems at elevated
temperatures and pressure. Contributions to Mineralogy and Petrology, 119:
197–212 (1995).
226

56. B. Giletti, Isotopic equilibrium/disequilibrium and diffusion kinetics in
feldspars. In I. Parsons, editor, Feldspars and their Reactions. NATO ASI
Series, volume 421, pages 351–382, Kluwer Academic (1994).
57. B. Giletti and J. Casserly, Strontium diffusion kinetics in plagioclase
feldspars. Geochimica et Cosmochimica Acta, 58: 3785–3793 (1994).
58. B. Giletti and T. Shanahan, Alkali diffusion in plagioclase feldspar. Chemical
Geology, 139: 3–20 (1997).
59. C. Ginibre, G. Wörner and A. Kronz, Minor- and trace-element zoning in plagioclase:
implications for magma chamber processes at Parinacota volcano,
northern Chile. Contributions to Mineralogy and Petrology, 143: 300–315
(2002).
60. C. Ginibre, G. Wörner and A. Kronz, Structure and dynamics of the Laacher
See magma chamber (Eifel, Germany) from major and trace element zoning
in sanidine: A cathodoluminescence and electron microprobe study. Journal
of Petrology, 45: 2197–2223 (2004).
61. T. Grove, D. Gerlach and T. Sando, Origin of calc-alkaline series lavas at
Medicine Lake volcano by fractionation, assimilation, and mixing. Contributions
to Mineralogy and Petrology, 80: 160–182 (1982).
62. K. Haase, Geochemical constraints on the magma sources and mixing processes
in Easter Microplate (MORB) (SE Pacific): a case study of plumeridge
interaction. Chemical Geology, 182: 335–355 (2002).
63. P. Hall and C. Kincaid, Plume-ridge interaction and the geometry of melting.
EOS Transactions AGU, 79(45): 1006 (1998).
64. B. Hanan, J. Blichert-Toft, R. Kingsley and J. Schilling, Depleted Iceland
mantle plume geochemical signature: Artifact of multicomponent mixing?
Geochemistry, Geophysics, Geosystems, 1: doi:10.1029/1999GC000009
(2000).
65. G. Heiken, Will Vesuvius Erupt? Three Million People Need to Know. Science,
286(5445): 1685–1687 (1999).
66. C. Herzberg, Partial crystallization of mid-ocean ridge basalts in the crust
and mantle. Journal of Petrology, 45(12): 2389–2405 (2004).
67. M. Higgins, Magma dynamics beneath Kameni volcano, Thera, Greece, as
revealed by crystal size and crystal shape measurements. Journal of Volcanology
and Geothermal Research, 70: 37–48 (1996).
227

68. M. Higgins, Measurement of crystal size distributions. American Mineralogist,
85: 1105–1116 (2000).
69. M. Higgins, Closure in crystal size distribution (CSD), verification of CSD
calculations and the significance of CSD fans. American Mineralogist, 87:
160–164 (2002).
70. M. Higgins and J. Roberge, Crystal size distributions of plagioclase and
amphibole from Soufrière Hills volcano, Montserrat: evidence for dynamic
crystallization – textural coarsening cycles. Journal of Petrology, 44: 1401–
1411 (2003).
71. R. Ho and M. Garcia, Origin of differentiated lavas at Kilauea Volcano,
Hawaii; implications from the 1955 eruption. Bulletin of Volcanology, 50:
35–46 (1988).
72. S. Huang, M. Regelous, T. Thordarson and F. Frey, Petrogenesis of Lavas
from Detroit Seamount: geochemical differences between Emperor Chain
and Hawaiian Volcanoes. Geochemistry, Geophysics, Geosystems, 6(1): doi:
10.1029/2004GC000756 (2005).
73. D. Hussong, L. Wipperman and L. Kroenke, The crustal structure of the
ontong java and manihiki oceanic plateaus. Journal of Geophysical Research,
84: 6003–6010 (1979).
74. S. Ingle, D. Weis and F. Frey, Indian continental crust recovered from Elan
Bank, Kerguelen Plateau (ODP Leg 183, Site 1137). Journal of Petrology,
43: 1241–1257 (2002).
75. S. Ingle, D. Weis, J. Scoates and F. Frey, Relatiosnhip between the early Kerguelen
plume and continental flood basalts of the paleo-Eastern Gondwanan
margins. Earth and Planetary Science Letters, 197: 35–50 (2002).
76. T. Jefferies, S. Jackson and H. Longerich, Application of a frequency quintupled
nd:yag source (? = 213 nm) for laser ablation inductively coupled
plasma mass spectrometric analysis of minerals. Journal of Analytical Atomic
Spectrometry, 13: 935–940 (1998).
77. W. Johannes, J. Koepke and H. Behrens, Partial melting reactions of plagioclases
and plagioclase-bearing systems. In I. Parsons, editor, Feldspars and
their Reactions. NATO ASI Series, pages 161–194, Kluwer Academic (1994).
78. R. Keller, R. Duncan and M. Fisk, Geochemistry and 40 Ar/ 39 Ar geochronology
of basalts from ODP Leg 145 (North Pacific Transect). In D. Rea,
I. Basov, D. Scholl and J. Allan, editors, Proceedings of the Ocean Drilling
228

Program, Scientific Results, volume 145, pages 333–344, Ocean Drilling Program,
College Station, TX (1995).
79. R. Keller, M. Fisk and W. White, Isotopic evidence for Late Cretaceous
plume-ridge interaction at the Hawaiian hotspot. Nature, 405: 673–676
(2000).
80. R. Kent, M. Pringle, A. Saunders and N. Ghose, 40 Ar/ 39 Ar age of the rajmahal
basalts, india, and their relationship to the southern kerguelen plateau.
Journal of Petrology, 43: 1141–1153 (2001).
81. S. King and D. Anderson, Edge-driven convection. Earth and Planetary Science
Letters, 160: 289–296 (1998).
82. L. Kroenke, W. Berger, T. Janecek and et al., Proceedings of the Ocean
Drilling Program, Initial Reports, 130. Technical report, Available from:
Ocean Drilling Program, Texas A&M University, College Station TX 77845-
9547, USA (1991).
83. T. Kuritani, Boundary layer crystallization in a basaltic magma chamber:
evidence from Rishiri Volcano, Northern Japan. Journal of Petrology, 39:
1619–1640 (1998).
84. C. Langmuir, Geochemical consequences of in situ crystallization. Nature,
340: 199–205 (1989).
85. C. Langmuir, E. Klein and T. Plank, Petrological systematics of Midocean
ridge basalts: constraints on melt generation beneath ocean ridges.
In J. Phipps-Morgan, D. Blackman and T. Plank, editors, Mantle flow and
melt generation at Mid-ocean ridges, volume 71 of Geophysical Monograph,
pages 183–280, American Geophysical Union (1992).
86. I. L’Heureux and A. Fowler, Isothermal constitutive undercooling as a model
for oscillatory zoning in plagioclase. Canadian Mineralogist, 34: 1264–1288
(1996).
87. G. Lofgren, An experimental study of plagioclase crystal morphology:
isothermal crystallization. American Journal of Science, 274: 243–273
(1974).
88. H. Longerich, S. Jackson and D. Günther, Laser ablation inductively coupled
plasma mass spectrometric transient signal data acquisition and analyte
concentration calculation. Journal of Analytical Atomic Spectrometry,
11: 899–904 (1996).
229

89. J. Longhi, D. Walker and J. Hays, Fe and Mg in plagioclase. Proceedings of
the7th Lunar and Planetary Science Conference, L4: 1281–1300 (1976).
90. T. Loomis, Numerical simulations of crystallization processes of plagioclase
in complex melts: the origin of major and oscillatory zoning in plagioclase.
Contributions to Mineralogy and Petrology, 81: 219–229 (1982).
91. J. Mahoney, J. Fitton and P. Wallace, Proceedings of the Ocean Drilling Program,
Initial Reports, 192. [CD-ROM], Available from: Ocean Drilling Program,
Texas A&M University, College Station TX 77845-9547, USA (2001).
92. J. Mahoney, W. Jones, F. Frey, V. Salters, D. Pyle and H. Davies, Geochemical
characteristics of lavas from Broken Ridge, the Naturaliste Plateau
and southernmost Kerguelen Plateau: Cretaceous plateau volcanism in the
southeast Indian Ocean. Chemical Geology, 120: 315–345 (1995).
93. J. Mahoney, M. Storey, R. Duncan, K. Spencer and M. Pringle, Geochemistry
and age of the Ontong Java Plateau. In M. Pringle, W. Sager, W. Sliter and
S. Stein, editors, The Mesozoic Pacific: Geology, Tectonics, and Volcanism,
volume 77 of Geophysical Monograph, pages 233–262, American Geophysical
Union (1993).
94. J. Mahoney, M. Storey, R. Duncan, K. Spencer and M. Pringle, Geochemistry
and geochronology of Leg 130 basement lavas: Nature and origin of the
Ontong Java Plateau. In W. Berger, L. Kroenke, L.A.Mayer and et al., editors,
Proceedings of the Ocean Drilling Program, Scientific Results, volume
130 (1993).
95. J. Mammerickx and G. Sharman, Tectonic evolution of the North Pacific
during the Cretaceous Quiet Period. Journal of Geophysical Research, 93:
3009–3024 (1988).
96. B. Marsh, Crystal size distribution (csd) in rocks and kinetics and dynamics
of crystallization i. theory. Contributions to Mineralogy and Petrology, 99:
277–291 (1988).
97. B. Marsh, Magma chambers. Annual Review of Earth and Planetary Sciences,
17: 439–474 (1989).
98. B. Marsh, Solidification fronts and magma dynamics. Mineralogical Magazine,
60: 5–40 (1996).
99. B. Marsh, On the interpretation of crystal size distributions in magmatic
systems. Journal of Petrology, 39(4): 553–599 (1998).
230

100. B. Marsh, A magmatic mush column rosetta stone: The mcmurdo dry valleys
of antarctica. EOS Transactions AGU, 85(47): 497,502 (2004).
101. A. McBirney, B. Baker and R. Nilson, Liquid Fractionation. Part I: basic
principles and experimental simulations. Journal of Volcanology and Geothermal
Research, 24: 1–24 (1985).
102. A. McBirney and R. Noyes, Crystallization and layering of the Skaergaard
intrusion. Journal of Petrology, 20: 487–554 (1979).
103. P. Michael, Implications for magmatic processes at Ontong Java Plateau from
volatile and major element contents of Cretaceous basalt glasses. Geochemistry,
Geophysics, Geosystems, 1(1999GC000025) (1999), Paper number:.
104. A. Mock, D. Jerram and C. Breitkreuz, Using Quantitative Textural Analysis
to Understand the Emplacement of Shallow-Level Rhyolitic Laccoliths -
a Case Study from the Halle Volcanic Complex, Germany. Journal of Petrology,
44(5): 833–849 (2003).
105. D. Morgan, D. Chertkoff, D. Jerram, J. Davidson and L. Francalanci, What’s
in a Whole Rock Analysis? Integrating Crystal Size Distributions and Micro-
Scale Isotopic Variations at Stromboli Volcano. EOS Transactions AGU,
85(V12A-08) (2005).
106. D. Morgan and D. Jerram, On estimating crystal shape for crystal size distribution
analysis. Journal of Volcanology and Geothermal Research, 154: 1–7
(2006).
107. M. Nakamura and S. Shimakita, Dissolution origin and syn-entrapment compositional
changes of melt inclusions in plagioclase. Earth and Planetary Science
Letters, 161: 119–133 (1998).
108. C. Neal and J. Davidson, An unmetasomatized mantle source for the
Malaitan alnöite (Solomon Islands): Petrogenesis involving zone refining,
megacryst fractionation, and assimilation of oceanic lithosphere. Geochimica
et Cosmochimica Acta, 53: 1975–1990 (1989).
109. C. Neal, J. Mahoney and W. C. III, Mantle sources and the highly variable
role of continental lithosphere in basalt petrogenesis of the Kerguelen Plateau
and Broken Ridge LIP: results from ODP Leg 183. Journal of Petrology,
43(7): 1177–1205 (2002).
110. C. Neal, J. Mahoney, L. Kroenke, R. Duncan and M. Petterson, The Ontong
Java Plateau. In J. Mahoney and M. Coffin, editors, Large Igneous
Provinces: Continental, Oceanic, and Planetary Flood Volcanism, volume
231

100 of Geophysical Monograph, pages 183–216, American Geophysical Union
(1997).
111. K. Nicolaysen, S. Bowring, F. Frey, D. Weis, S. Ingle, M. Pringle, M. Coffin
and the Leg 183 Shipboard Scientific Party, Provenance of Proterozoic
garnet-biotite gneiss recovered from Elan Bank, Kerguelen Plateau, southern
Indian Ocean. Geology, 29: 235–238 (2001).
112. M. Norman, N. Pearson, A. Sharma and W. Griffin, Quantitative analysis of
trace elements in geological materials by laser ablation ICP-MS: instrumental
operating conditions and calibration values of NIST glasses. Geostandards
Newsletter, 20(247-261) (1996).
113. M. O’Hara and C. Herzberg, Interpretation of trace element and isotope
features of basalts: Relevance of field relations, petrology, major element
data, phase equilibria, and magma chamber modeling in basalt petrogenesis.
Geochimica et Cosmochimica Acta, 66(12): 2167–2191 (2002).
114. J. Pearce, Mantle Preconditioning by Melt Extraction during Flow: Theory
and Petrogenetic Implications. Journal of Petrology, 46(5): 973–997 (2005).
115. T. Pearce, Recent work on oscillatory zoning in plagioclase. In I. Parsons,
editor, Feldspars and their reactions, volume 421, pages 313–350, Kluwer
Academic (1994).
116. M. Petterson, C. Neal, J. Mahoney, L. Kroenke, A. Saunders, T. Babbs,
R. Duncan, D. Tolia and B. McGrail, Structure and deformation of north
and central Malaita, Solomon Islands: tectonic implications for the Ontong
Java Plateau—Solomon arc collision, and for the fate of oceanic plateaus.
Tectonophysics, 283: 1–33 (1997).
117. W. Phinney, Partition coefficients for iron between plagioclase and basalt as
function of oxygen fugacity: Implications for lunar anorthosites. Geochimica
et Cosmochimica Acta, 56: 1885–1895 (1994).
118. D. Rea, I. Basov, T. Janecek, A. Palmer-Julson and et al., Proceedings of
the Ocean Drilling Program, Initial Reports, 145. Technical report, Available
from: Ocean Drilling Program, Texas A&M University, College Station TX
77845-9547, USA (1993).
119. M. Regelous, A. Hofmann, W. Abouchami and S. Galer, Geochemistry of
lavas from the Emperor Seamounts, and the geochemical evolution of Hawaiian
magmatism from 85 to 42 Ma. Journal of Petrology, 44(1): 113–140
(2003).
232

120. R. Resmini and B. Marsh, Steady-state volcanism, paleoeffusion rates, and
magma system volume inferred from plagioclase crystal size distributions in
mafic lavas: Dome Mountain, Nevada. Journal of Volcanology and Geothermal
Research, 68: 273–296 (1995).
121. J. Roberge, R. White and P. Wallace, Volatiles in submarine basaltic glasses
from the Ontong Java Plateau (ODP Leg 192): implications for magmatic
processes and source region compositions. In J. Fitton, J. Mahoney, P. Wallace
and A. Saunders, editors, Origin and Evolution of the Ontong Java
Plateau, volume 229 of Special Publications, pages 329–257, Geological Society
(2004).
122. M. Ryan, Neutral buoyancy and the mechanical evolution of magmatic systems.
In B. Mysen, editor, Magmatic processes: physicochemical principles,
volume 1 of Special Publication, pages 259–288, The Geochemical Society
(1987).
123. T. Sano and S. Yamashita, Experimental petrology of basement lavas from
Ocean Drilling Program Leg 192: implications for differentiation processes
in Ontong Java Plateau magmas. In J. Fitton, J. Mahoney, P. Wallace and
A. Saunders, editors, Origin and Evolution of the Ontong Java Plateau, volume
229 of Special Publications, pages 185–218, Geological Society, London
(2004).
124. A. Saunders, Mantle plumes: an alternative to the alternative. Geoscientist
(2003).
125. A. Saunders, M. Storey, I. Gibson, P. Leat, J. Hergt and R. Thompson,
Chemical and isotopic constraints on the origin of basalts from the ninetyeast
ridge, indian ocean: results from dsdp legs 22 and 26 and odp leg
121. In J. Weissel, J. Peirce, E. Taylor and J. Alt, editors, Proceedings of
the Ocean Drilling Program, Scientific Results, volume 121, pages 559–590,
Ocean Drilling Program, College Station TX 77845-9547, USA (1991).
126. T. Sisson and T. Grove, Experimental investigations of the role of H2O in
calc-alkaline differentiation and subduction zone magmatism. Contributions
to Mineralogy and Petrology, 113: 143–166 (1993).
127. J. Smith and W. Brown, Feldspar Minerals, volume 1. Springer-Verlag
(1988).
128. J. Tarduno, D. Duncan, R. Scholl and et al., Proceedings of the Ocean
Drilling Program, Initial Reports, 197. [CD-ROM], Available from: Ocean
Drilling Program, Texas A&M University, College Station TX 77845-9547,
USA (2002).
233

129. J. Tarduno, W. Sliter, L. Kroenke, M. Leckie, J. Mahoney, R. Musgrave,
M. Storey and E. Winterer, Rapid formation of the Ontong Java Plateau by
Aptian mantle plume volcanism. Science, 254: 399–403 (1991).
130. R. Taylor, B. Murton and M. Thirlwall, Petrographic and geochemical variation
along the Reykjanes Ridge, 57 ◦ -59 ◦ N. Journal of the Geological Society
of London, 152: 1031–1037 (1995).
131. M. Tejada, J. Mahoney, P. Castillo, S. Ingle, H. Sheth and D. Weis, Pinpricking
the elephant: evidence on the origin of the Ontong Java Plateau from
Pb-Sr-Hf-Nd isotopic characteristics of ODP Leg 192 basalts. In J. Fitton,
J. Mahoney, P. Wallace and A. Saunders, editors, Origin and Evolution of
the Ontong Java Plateau, volume 229 of Special Publications, pages 133–150,
Geological Society, London (2004).
132. M. Tejada, J. Mahoney, R. Duncan and M. Hawkins, Age and geochemistry of
basement and alkalic rocks of Malaita and Santa Isabel. Journal of Petrology,
177: 361–393 (1996).
133. M. Tejada, J. Mahoney, C. Neal, R. Duncan and M. Petterson, Basement
Geochemistry and Geochronology of Central Malaita, Solomon Islands, with
Implications for the Origin and Evolution of the Ontong Java Plateau. Journal
of Petrology, 43(3): 449–484 (2002).
134. F. Tepley and J. Davidson, Mineral-scale sr-isotope constraints on magma
evolution and chamber dynamics in the rum layered intrusion, scotland. Contributions
to Mineralogy and Petrology, 145: 628–641 (2003).
135. F. Tepley, J. Davidson, R. Tilling and J. Arth, Magma mixing, recharge
and eruption histories recorded in plagioclase phenocrysts from El Chichón
Volcano, Mexico. Journal of Petrology, 41: 1397–1411 (2000).
136. A. Tsuchiyama, Dissolution kinetics of plagioclase in the melt of system
diopside-albite-anorthite, and origin of dusty plagioclase in andesites. Contributions
to Mineralogy and Petrology, 89: 1–16 (1985).
137. J. VanderAuwera, J. Longhi and J. Duchesne, The effect of pressure on DSr
(plag/melt) and DCr (opx/melt): implications for anorthosite petrogenesis.
Earth and Planetary Science Letters, 178: 303–314 (2000).
138. L. Wager and W. Deere, Geological investigations in East Greenland: Part
III. The petrology of the Skaergaard Intrusion, Kangerdluqssuaq, East
Greenland. Meddelelser om Grønland, 105: 1–16 (1939).
234

139. G. Wallace and G. Bergantz, Wavelet-based correlation (WBC) of zoned
crystal populations and magma mixing. Earth and Planetary Science Letters,
202: 133–145 (2002).
140. G. Wallace and G. Bergantz, Constraints on mingling of crystal populations
from off-center zoning profiles: A statistical approach. American Mineralogist,
89: 64–73 (2004).
141. G. Wallace and G. Bergantz, Reconciling heterogeneity in crystal zoning
data: An application of shared characteristic diagrams at Chaos Crags,
Lassen Volcanic Center, California. Contributions to Mineralogy and Petrology,
140: 98–112 (2005).
142. M. Wilke and H. Behrens, The dependence of the partitioning of iron and
europium between plagioclase and hydrous tonalitic melt on oxygen fugacity.
Contributions to Mineralogy and Petrology, 137: 102–114 (1999).
143. J. Wolff and F. Ramos, Assimilation yesterday, today and tomorrow. EOS
Transactions AGU, 83: F1406 (2003).
144. B. Wood and J. Blundy, The effect of H2O on crystal-melt partitioning of
trace elements. Geochimica et Cosmochimica Acta, 66: 3647–3656 (2002).
145. G. Zellmer, S. Blake, D. Vance, C. Hawkesworth and S. Turner, Plagioclase
residence times at two island arc volcanoes (Kameni Islands, Santorini, and
Soufriere, St. Vincent) determined by Sr diffusion systematics. Contributions
to Mineralogy and Petrology, 136: 345–357 (1999).
146. M. Zieg and B. Marsh, Crystal size distributions and scaling laws in the quantification
of igneous textures. Journal of Petrology, 43(1): 85–101 (2002).
This document was prepared & typeset with L ATEX2ε, and formatted with
nddiss2ε classfile (v1.0[2004/06/15]) provided by Sameer Vijay.
235