THE NATIONAL ACADEMIES PRESS

This PDF is available at [Link] SHARE

   

Generation of Basaltic Magma (1976)

DETAILS

281 pages | 5 x 9 | PAPERBACK

ISBN 978-0-309-02504-1 | DOI 10.17226/19924

CONTRIBUTORS

GET THIS BOOK Yoder, H. S.; National Academy of Sciences

FIND RELATED TITLES

SUGGESTED CITATION

National Research Council 1976. Generation of Basaltic Magma. Washington, DC:

The National Academies Press. [Link]


Visit the National Academies Press at [Link] and login or register to get:

– Access to free PDF downloads of thousands of scientific reports

– 10% off the price of print titles
– Email or social media notifications of new titles related to your interests
– Special offers and discounts


Distribution, posting, or copying of this PDF is strictly prohibited without written permission of the National Academies Press.
(Request Permission) Unless otherwise indicated, all materials in this PDF are copyrighted by the National Academy of Sciences.

Copyright © National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

)..,\"" )/
(). <.)-

...
H. S. YODER, JR .
GEOPHYSICAL LABORATORY
Carnegie Institution of Washington
Washington, D.C.

Generatio~f
Basaltic ~a

NATIONAL ACADEMY OF SCIENCES

Washington, D.C. 1976

NAS·NAE
NOV 9 1976
LIBRARY
Copyright National Academy of Sciences. All rights reserved.
Generation of Basaltic Magma

Library of Congress Cataloging In Publication Data

Yoder, Hatten Schuyler, Jr., 1921-

Generation of basaltic magma.

Based on a lecture series presented at Dartmouth College, May 5-9, 1975, and at the
Geophysical Laboratory ofthe Carnegie Institution of Washington, June 23-27, 1975.
I. Basalt-Addresses, essays, lectures. 2. Magmatism-Addresses, essays, lectures.
I. Title.
QE462.B3Y6 552' .2 76-29672

International Standard Book Number 0-309-02504-4

Available from
Printing and Publishing Office
National Academy of Sciences
2101 Constitution Avenue , N.W .
Washington, D.C. 20418

Printed in the United States of America

Order from 80 79 78 77 76 10 9 8 7 6 5 4 3 2 I
National Technical
Information Service,
Springfield, Va.
22161
Order NoX B1.1 q-3 b~
Copyright National Academy of Sciences. All rights reserved.
Generation of Basaltic Magma

Preface

A bequest was made by Dr. Arthur L. Day to the National Academy of

Sciences "for the purpose of advancing studies of the physics of the
earth." The Arthur L. Day Trust Fund Committee* established the
Arthur L. Day Prize and Lectureship with the expectation that the re-
cipient would prepare a publication, preferably a book, giving a "com-
prehensive summary and synthesis of current knowledge in his field."
It was the writer's good fortune to be chosen by the Selection
Committeet as the first recipient of the Arthur L. Day Prize and
Lectureship, and the opportunity to prepare a book on a subject of his
own choosing was indeed a welcome prospect. It had long been his
desire to take the time to think through the details of the process of
magma* generation. Most studies assume the existence of magma and
then consider its transport and crystallization; its actual formation is
rarely treated at length.
The principal problem was finding sufficiently long periods of time in
which to analyze the data and synthesize the views. The bits and pieces
*Trust Fund Committee: William W. Rubey, chainnan; Richard M. Goody; Merle A.
Tuve; J. Tuzo Wilson.
tSelection Committee: Herbert Friedman, chainnan; Jute G. Charney; Allan Cox; Julian
R. Goldsmith; 0. G. Villard.
iAs used by geologists, magma can be defined as a naturally occurring, mobile liquid
within the earth that may contain suspended crystals or rock fragments as well as
dissolved or ex solved gases and that can fonn a rock, crystalline to glassy, on intrusion
or extrusion as a Java.

iii

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

iv Preface
were pulled together over a 2-year period and presented as an inte-
grated lecture series at Dartmouth College on 5-9 May 1975 and at the
Geophysical Laboratory of the Carnegie Institution of Washington on
23-27 June 1975. The writer is indebted to Professor Robert W.
Decker, Chairman of the Department of Earth Sciences at Dartmouth
College, for organizing the visit and to the students and staff for their
thought-provoking questions-and answers. The staff of the Geophysi-
cal Laboratory provided a stimulating and critical audience for what
were intended to be provocative discussions. Their response, substan-
tiated by their willingness to undertake some of the suggested research
problems, was most gratifying.
The book is directed to those who would like a comprehensive view
of the generation of basaltic* magma at the site of origin.
The writer's purpose was to bring to a focus the many problems
amenable to theoretical analysis and experimental investigation. It is
evident that direct field study is not possible now, but it cannot be
considered as impossible in the near future. The Magma Tap Program,
being developed by the Energy Research and Development Adminis-
tration and the United States Geological Survey in an effort to meet
critical energy needs, may present opportunities for the direct observa-
tion of at least auxiliary magma chambers. The large uncertainties in
the deductions presented here will be evident to the reader, and it
would be an understatement to say the points of view are somewhat
controversial. If these views stand for any length of time, then the
attempt to stimulate new researches on the subject will have been for
naught.
The reader will gain an impression of the tenuous nature of the
evidence on which major concepts about magmas in the earth are
based. The postulated models have become accepted as the reality
instead of the lattice of assumptions they are. Authoritarianism domi-
nates the field, and a very critical analysis of each argument is to be
encouraged. The structure of facts is minimal-the reader should
continuously ask, "What is the evidence?" It is recommended that he
withhold judgment on an issue until completing the book. Persuasive
arguments in one section may be countered with equal vigor in another
*Basalt is a mafic, extrusive igneous rock composed chiefly of plagioclase (usually
labradorite) and clinopyroxene. Varieties may contain in addition olivine. ortho-
pyroxene, nepheline, and quartz, individually or in restricted combinations. The rock
may be glassy, fine-grained, or porphyritic. Apatite and magnetite are common acces-
sories. Basalt is in some places intrusive in the form of dikes; its intrusive equivalent.
however, is generally called diabase (dolerite) or gabbro. The average chemical com-
positions of the common varieties of basalt have been given by Nockolds (1954),
Manson (1976), and Chayes (1975).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Preface v
section. Not all arguments can be presented, and some are omitted,
although referenced, because they are based on experiments since
shown to be inaccurate or on assumptions now believed to be unwar-
ranted.
The book is organized about the what, where, how, why, and when
of magma production. The lectures included an examination of the
critical parameters and principles involved in the generation of the
most common magma, basaltic in composition,* that reaches the
earth's surface. Constraints are placed on (a) the nature of the parental
material and the location of melting in the earth; (b) the melting
process, including the heat requirements and sources; (c) the me-
chanics of magma accumulation, segregation, and rise; (d) the physical
chemistry of the evolving magma; (e) the tectonophysics of melting;
and (f) the energetics, periodicity, and duration of magmatic events.
The substance of the lectures has been expanded somewhat to meet the
needs of students having only a general course in petrology.
There is no attempt to discuss subsidiary or auxiliary magma cham-
bers and the associated differentiation processes taking place there, the
eruptive mechanics, or the morphological expression of volcanic prod-
ucts. These subjects have been adequately dealt with in texts on
volcanology and petrology.
No apologies are made for the breadth of subject matter covered in
the book. The writer is obviously not an expert in all the subjects
discussed, but who is? An appreciation and awareness of the many
exciting facets of magma generation are sought.
It is a pleasure to acknowledge the labors of Miss Marjorie Imlay,
who typed drafts and made alterations cheerfully and with patience;
Mr. A. David Singer, who organized the drafting, photography, and
assembly of the many figures; Mr. William C. Hendrix, who drafted all
the figures; and Miss Dolores Thomas, who edited the copy submitted
to the National Academy of Sciences and collated the references. An
early draft was reviewed by Drs. P. M. Bell (Chapters 1-6) and Bj!11rn
Mysen and Mr. H. Richard Naslund. More advanced drafts were
reviewed completely or in part by Dr. Nicholas T. Arndt, Dr. Felix
Chayes, Prof. R. W. Decker (Chapters 9 and 10), Prof. JohnS. Dickey,
Dr. Hugh C. Heard (Chapter 9), Dr. Albrecht Hofmann, Prof. Hans G.
Huckenholz, Dr. T. Neil Irvine, Dr. lkuo Kushiro, Prof. John B.
Lyons, Prof. Alexander R. McBirney, Prof. S. Anthony Morse, Dr.

*Basalts and related rocks (gabbros. amphibolites, and eclogites) comprise about 42.5
percent of the volume of the crust of the earth, according to Ronov and Yaroshevsky
(1969. p. 49. Table 7). The base of the crust is taken by them to be theM (Mohorovi~ic)
discontinuity.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

vi Preface
Dean C. Presnall (Chapter 6), Dr. Eugene Robertson, Prof. Jean-Guy
Schilling, Dr. Jan A..Tullis (Chapter 9), and Dr. Danielle Velde. The
many hours invested by these kind friends have led to a much im-
proved manuscript. The debates with these and other specialists and
generalists too numerous to list have been most enjoyable and reward-
ing. As every author knows, the writing of a book on a new
subject is a most exhilarating learning experience. All contributors are
thanked for their help and tuition.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

CONTENTS

INTRODUCTION

Volcanic Products-A Renewable Resource,

Systematic Studies Needed-Violent Volcanoes, I
Observed Crystallized Magma Chamber-Transported
Magmas, 2
Models of Magma Sources-A Prologue, 2
Balloon/Soda-Straw Model, 3; Geophysical Model, 4; Crystal-Mush
Model, 4; Benioff-Zone Model, 5; Depth-of-Generation Model, 7; Depth
of-Separation Model, 8; Explosive-Diatreme Model, 9; Conclusion, II

2 PARENTAL MATERIALS 12
Principal Requirement-Yield Basaltic Magma, 12
Parent-Daughter Test, 13
Assumptions for Test, 15; Basalts Equilibrated to Surface Conditions, 16;
General Types of Source Materials, 18
Basaltic or Gabbroic Source, 19
Eclogitic Source-High-Pressure Equivalent of Basalt, 20
Amphibolitic Source, 22
Olivine-An Important Source-Rock Constituent?, 24
Meteoritic Source, 24
Important Phases in Meteorites, 25; Phase Incompatible with Olivine
at High Pressure, 27; Meteorites Related to Peridotite?, 29
Gamet Peridotites from the Upper Mantle, 30
Key Position of Gamet Peridotite, 31; Will Gamet Peridotite Yield
Basaltic Magma?, 32
vii

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

viii CONTENTS

Meteorites Do Not Yield Basaltic Magma, 34

Phase Equilibria Support for Gamet Peridotite Source, 36
Source of Potassium?, 37
Seismic Velocities of Gamet Peridotite Suitable?, 38
Monte Carlo Test of Physical Parameters, 41
Composition(s) of the Mantle Not Determined, 42
Gamet Peridotite Accepted as Source-Tentatively, 42

3 DEPTH OF MELTING 44
Origin of Protoearth: Cool Accumulation, Differentiated
Melt, or Continuum of Both?, 44
Crust Formation through Volcanism, 46
Impacting, Deep Surficial Turnover, and Partial Melting, 48
Magma-A Residuum from Earth Formation?, 49
Seismic Evidence-Auxiliary Magma Chambers, 49
Low-Velocity Zone-Evidence of Partial Melting?, 51
Partially Melted Zone-Experimental Evidence, 52
Attenuation of Seismic Waves, 55
Summary, 55

4 , MELTING PROCESSES 57
Multitude of Mechanisms, 57
Melting Behavior of Critical Minerals, 57
The Mechanisms, 60
Stress Relief, 61
Tension, 61; Compression, 61
Thermal Rise to Cusp in Melting Curve, 63
Compositional Cusp, 63 ; Phase-Change Cusp, 64
Convective Rise, 65
Isentropic Rise, 65; Isenthalpic Rise, 66
Parameter Perturbation, 66
Thermal Conductivity Perturbation, 66; Density Perturbation, 68
Mechanical Energy Conversion to Heat, 71
Thrust Faulting, 72; Subducting Plate, 72; Propagating Crack, 73;
Regenerative Feedback, 74; Tidal Dissipation, 75
Compositional Change, 76
Diffusion, 76; Volatile Fluxing, 78; Other Mechanisms, 86
Phenomenological View of Melting, 86

5 THERMAL ENERGY REQUIREMENTS FOR MELTING 88

Enthalpy of Melting, 88
Measurement, 89; Previous Estimates, 90

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

CONTENTS ix

Enthalpy of Melting of Minerals and Rocks, 90

Values Available for End-Member Minerals, 90; Model System for
Basalt, 90; Improved Model System for Basalt, 92; Model System for
Eclogite, 94; Model System for Gamet Peridotite, 95
Application of Enthalpy of Melting, 96
Local Heat Production, 96; Adiabatic Rise, 101

6 PHYSICOCHEMICAL CONSTRAINTS ON MELTING 105

Methods of Initial Melt Withdrawal, 105
Batch Melting, 107; Fractional Melting, 107; Subsequent Melt With-
drawal, 109
Important Principles and Questions, 109
Proportion of Phases in Parental Material, ItO; Disaggregation of
Parental Material, 112; Amount of Liquid versus Temperature, 112
Zone Melting, 115
Disequilibrium Melting, 118

7 COMPOSITION OF MAJOR BASALTIC MAGMAS 120

Framework of Igneous Petrology, 120
Basalt Tetrahedron, 120
Illustration of Phase Equilibria, 121
Flow Sheet, 125
"Parental" Magmas, 128
Need for Flow Sheet for High Pressures, 131
Magma Separation, 133

8 OLIVINE IN THE PARENTAL MATERIAL 134

Olivine from Initial Melt, 134
Absence of Olivine in Basalts at High Pressures, 134
Melting Relationship of Basalt to Gamet Peridotite, 139
Melting Relationship of Eclogite to Gamet Peridotite, 139
Melting of Eclogite, 139
Orthopyroxene Reaction Relation, 142; Olivine Reaction Relation, 143
Basalt Composition Changes during Rise, 145
Olivine-Incompatible or Complementary?, 146
Evidence from Nodules, 147
First Phase Consumed in Melt, 148
Rare-Earth Element Data, 150
Measure of Degree of Partial Melting?, 150; Measure of Heterogeneity
or Proportion of Phases?, 154; Rare Earths in Gamet Peridotite, 158
General Remarks and Summary, 159

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

X CONTENTS

9 TECTONOPHYSICS OF MELTING 162

Location of Initial Liquid, 162
Accumulation of Liquid, 163
Homogeneity of Liquid, 165
Analogy to Glacial Water, 166
Role of Deformation, 166
Localization of Liquid, 167
Response to Stress, 167; Coalescence, 168
Volume Change of Melting, 170
Large Increase in Volume, 170; Tumescence, Rupture, or Plastic Aow?,
170; State of Stress in Earth, 171 ; Adhesive Strength of Crystals, 172
Escape from the Container, 172
Plastic Envelope, 173; Time Dependence, 175; "Kneadinlt Out" the
Uquid, 175; Solubility under Stress, 177; Loss in Strength of Container, 178;
Magmafracting, 179
Volcanism Produces Some Earthquakes, 180

10 ENERGETICS AND PERIODICITY 183

Energetics, 183
Partition of Energy, 183; Energy in Earthquakes, 186
Forces Causing Magma Rise, 186
Overburden Squeeze, 186; Buoyancy, 188; Importance of Viscosity, 189
Periodicity , 191
Paucity of Dated Basalts, 191; Spacing of Volcanoes, 192; Age and
Spacing Related?, 194
Convective Heat Transfer Models, 195
Joly-Cotter Model, 195; Tikhonov et al. Model. 196
Role of Volatiles, 196
Role of Subsidiary Reservoir, 199
Runaway Mechanisms Potentially Periodic, 200

II SUMMARY OVERVIEW 201

What Is Melted to Yield Basaltic Magma?, 201
Where Does Melting Occur?, 203
How Does Melting Begin?, 203
Why Do Basaltic Melts Predominate?, 204
What Factors Determine the Kinds of Basalt?, 205
How Does the Melt Form and Aggregate?, 206
What Are the Forces Causing Magma Release?, 207
Why Are Volcanic Eruptions Periodic?, 208
Conclusion, 209

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

CONTENTS xi
Appendix: Conversion Factors 211
References 213
Author Index 237
Subject Index 243
System Index 265

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

1 Introduction

VOLCANIC PRODUCTS-A RENEWABLE RESOURCE

A volcanic eruption is an exciting and terrifying event. It is impossible
to transmit in words the total experience of an eruption; each person
has different indelible memories of the sights, sounds, vibrations, and
emotions. Volcanic eruptions are classed as a geologic hazard. Funding
for present-day volcanological studies capitalizes on the emotional fear
of the hazard with its resultant loss of life and destruction of property.
Only rarely is an eruption recognized for its value-the replenishment
of the earth's surface with new material. Farmers appreciate the rich
soil formed from some volcanic ashes, and, with suitable crushing
equipment, even some new lava flows can become arable within
several years. Volcanic products are indeed a principal renewable
resource of fundamental importance to mankind.

SYSTEMATIC STUDIES NEEDED-VIOLENT

VOLCANOES
Present efforts in volcanological studies are devoted to the short-term
events that precede an eruption so that prediction can be made as to the
place, time, and intensity of the outbreak. Unfortunately, these studies
are made mainly on volcanoes of the quiescent type (e.g., Kilauea,
Hawaii), whereas the greatest danger lies with the explosive type (e.g.,
Lassen Peak, California). Systematic study of the violent-type erup-
tions should be organized and initiated before another catastrophe in a

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

2 GENERATION OF BASAL TIC MAGMA

populated area demands attention. Remote or telemetered sensing

devices (e.g., seismic, thermal, tilt, exhalation, photographic) should
be installed now at those volcanoes near densely populated areas,
particularly in the western United States (e .g., Mount St. Helens;
Crandell et al., 1975).

OBSERVED CRYSTALLIZED MAGMA CHAMBERS-

TRANSPORTED MAGMAS
As important as the present-day, short-term events are to man, it is
necessary to examine the origins of the large volumes of relatively
homogeneous lava that issue at the surface. For most purposes it is
sufficient to accept the idea that lavas arise from shallow (2-6-km)
reservoirs. These reservoirs, after cooling, have been exposed by
erosion, and their dimensions can be deduced on the basis of careful
field studies-the very substance of geology. Such rock masses, recog-
nized by their characteristics as crystallized from molten or partially
molten material, are described as intrusive; that is, the magma was
produced somewhere else and was moved, irrespective of the style of
emplacement, into the environment in which it is found. Even large
batholiths, usually siliceous, multiple, and not visibly floored, are
considered intrusive: the magma had moved from its site of origin.
Only rarely is the evidence sufficiently clear to lead to the conclusion
that small portions of basic or ultrabasic rocks were melted in place
(Dickey, 1970; Menzies, 1973). On the other hand, many siliceous
rocks exhibit features suggesting in situ melting, that is, anatexis, or,
more broadly, palingenesis.

MODELS OF MAGMA SOURCES-A PROLOGUE

Because there are few directly applicable data available and the mea-
sured parameters commonly lack the sensitivity for the investigator to
draw unambiguous conclusions, many different models of magma
sources have been proposed. Some of these will be reviewed. There is
some risk in commenting on current models of magma generation even
though they point up the need for a more thorough examination of the
problem. The reader should bear in mind that they are presented out of
context and are used only to illustrate some of the principal questions
that remain to be answered. It is not appropriate for the writer to
criticize them in detail; the reader, however, is not discouraged from
that exercise. The models are reviewed for the purpose of arousing the
reader's scientific curiosity and to set the stage for what follows.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Introduction 3

/_Main vent, partially blown away

FIGURE 1-1 Cross section of a typical volcano (after Rogers and Adams, 1966, p. 159,
Figure 8-3). Shape, size, and position of magma chamber are generally unknown. (With
permission of Harper & Row, Publishers.)

BALLOON/SODA-STRAW MODEL

One of the simplest models describing the source region of magma is

shown in Figure 1-1. The "balloon/soda-straw" model, no doubt
presented to nearly every freshman geology class, implies a point heat
source and ellipsoidal isotherms.* Although Rogers and Adams (1966,
p. 159, Figure 8-3) indicated that the shape, size, and position were
generally unknown, the impressions gained by their relative scales are
important. For example, if the vent is I km in diameter, the magma
chamber center is about II km below the summit and the volume of the
oblate spheroid chamber is about 90 km 3 , enough for 900 lava flows of
0.1 km 3 in volume. The stored thermal energy in the chamber is about
42 x 1026 erg, five times the total energy of the single, most violent
volcanic eruption known. t In the absence of a mechanism for refilling,
*Menard (1969, p. 138) gave a similar model with a more spherical chamber situated at
much greater depths, just above the asthenosphere, and an appropriately long " soda
straw."
tBased on ll£ = V·p (Cpl:iT + llllm) . where V = 90 x 10 1 • cm 3 , p = 2.78 glcm 3 .1:1T =
1250"C, llllm = 87.5 cal/g, C,, = 0.25 cal/g°C (I cal = 4. 186 x 107 erg).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

4 GENERATION OF BASALTIC MAGMA

it is presumed that the magma would differentiate and deplete with

time. Therefore, the chamber obviously could not yield large volumes
of homogeneous magma over an extended period of geologic time. Is it
possible to drain the chamber completely by successive collapses after
each withdrawal? Or is it more likely that the liquid is distributed in a
plastic sponge of crystals that deforms as the liquid is squeezed out?
What are the forces that cause the magma to rise in the pipe? One must
also ask, "What is the meaning of the line bounding the magma
chamber?" Is it an isotherm, the solidus, a compositional boundary, a
structural discontinuity, a solid-phase change, a diffusion limit, or
some other limiting parameter? The difference in the level of the
magma in the main vent and that in the subsidiary cone implies a
different density of magma; perhaps the magma in the subsidiary cone
contains more crystals and therefore is denser on the average.

GEOPHYSICAL MODEL

A more sophisticated, scaled model based on geophysical data col-

lected in Hawaii was given by Eaton and Murata (1%0) and is shown in
Figure 1-2. The depth of generation was assumed to be that of the
deepest earthquake swarms recorded under Kilauea. Eaton and
Murata believed that "Such activity appears to mark the zone from
which magma is collected and fed into the system of conduits .... "
The arrows indicate the presumed direction of magma flow into a
cavity or at least into a zone of higher permeability and porosity if
cavities or pores can be sustained at such depths. The authors empha-
sized the movement of magma as the cause of quakes but did not relate
the magma generation to seismic energy release. The source of the
magma was not identified, nor was the thermal regime outlined.

CRYSTAL-MUSH MODEL

At least two additional major ideas are illustrated by the model in

Figure 1-3 (Wyllie, 1970, p. 6, Figure 1), based on the description of
Green and Ringwood (1%7, pp. 164-167). First, the arrows indicate the
derivation from an unknown source of a crystal-liquid mush presumed
to have an average density less than that of the layer in which it was
produced. The question arises as to the ratio of crystals to liquid
required for the mush to detach itself from the country rock and move
upward. What is the velocity of movement, and is the pressure release
sufficient to offset the cooling rate? The second important concept is
the depth of separation of liquid from the crystal-liquid mush. No

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Introduction 5

Vp

---:::::::::==:::----::_::::o__..--..---"llr---.
__-______ --~2
km/Mc

___-:::::;::::::::
Vp Oenlity
kmllec 9lcc
1.03
!5.3 4.3 2.3
6.8 Crust 6.6 2.84

Mollotovlcic Oisconllnuity 8.2!5 Mantle 7.9 3.27

-Density
2.77 Qlcc

--:r. ,
--------~~~---------
~/---
0 200

FIGURE 1-2 Schematic cross section of an idealized Hawaiian volcano according to

Eaton and Murata (1960, p. 930, Figure 5). (With permission. Copyrighted 1960 by the
American Association for the Advancement of Science.)

chemical constraints on where the separation of the liquid takes place

are apparent in this model, and Wyllie presumed that the liquid was
basaltic irrespective of the depth of separation. Physical constraints,
on the other hand, may dominate the process of liquid separation. A
zone of shear or a layer of lower density may fix the depth of
separation.

BENIOFF-ZONE MODEL

A very imaginative model (Figure 1-4) of the source region of magma

was developed by Coats (1962). He had one of the earliest clear
pictures of the present-day theory of plate tectonics, emphasizing the
Benioff zone (Benioff, 1954) as one of the principal environments of
magma generation. He believed that there existed a zone in peridotite,
just above 100 km, containing "pockets" of basalt, olivine basalt,
eclogite, or preferably vitreous material of unspecified origin that were

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

6 GENERATION OF BASALTIC MAGMA

remelted to yield basaltic liquids or, by the addition of water and
materials from the thrust zone, to yield andesite. The quantity of
crustal material added to the eruptible basaltic magma determined the
extent of "differentiation" along with the basalt-andesite-rhyolite
trend. Coats did not specify how the material of basaltic composition
initially became molten. What, indeed, were the physicochemical
controls that produced the same magma fractionation trend irrespec-
tive of the source of materials? His model also illustrates auxiliary

Lava flaw

Sills

Stratiform intrusion

M. discontinuity

Magma
separation

Magma generation
(Basalt)

FIGURE 1-3 Schematic representation of magma generation resulting from diapiric

uprise of mantle according to WyUie (1970, p. 6, Figure 1). Stippled areas are crystalline
ultramafic material; black areas are interstitial basic liquid or crystalline basalt or gabbro.
(With permission of the Mineralogical Society of America.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Introduction 7

Distance, km

FIGURE 1-4 Diagrammatic cross section through crust and upper mantle of a gener-
alized island arc according to Coats ( 1962, p. 107, Figure 9), showing only the first stage
of development. (With permission . Copyrighted by the American Geophysical Union.)

magma chambers elongated vertically, but the reasons for their exis-
tence and elongation were not described. Do the elongated shapes
suggest diapirs, or do they result from long-term assimilation of the
country rocks? The diagram illustrates basaltic magma originating
within a narrow depth range, in contrast to the next model discussed.

DEPTH-OF-GENERATION MODEL

Kuno (1967, p. 108, Figure 10) presented a model (Figure 1-5) that
relates the kind of magma (deduced from the belts of rocks observed at
the surface) to the locus of earthquakes in the Benioff zone. The
earthquakes are present-day events, whereas the volcanoes increase in
age from the trough toward the continent (Shimozuru, 1963). The
model was constructed in response to a debate with Yoder and Tilley
(1962), who believed that Kuno's earlier figure (1959, p. 73, Figure 12)
indicated that each magma type was derived from different layers of
characteristic parental material. Kuno derived directly, without inter-
vening auxiliary chambers, alkali basalt from the deeper layers. high-
alumina basalt from the intermediate depths, and tholeiitic magma from
the shallower depths. Attention is called to his interesting characteriza-
tion of the magma as pockets or droplets, depending on the scale
factor, that accumulate into a converging crack system . The localized
nature of the melt is also to be noted, even though the thermal regime

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

8 GENERATION OF BASALTIC MAGMA

Ptlclflc Oc-
Yotuoo-
Okl-dito 011-toke toke Huzl ...._
0-tlmo SE

Mont le

Hlelt-olumlno basalt moemo

Alkali olivine basalt moemo

FIGURE 1-5 Cross section of central Honsyo from the lzu Islands to Oki-dngo Island
showing depths of generation of different basaltic magmas according to Kuno (1967,
p. 108. Figure 10). who assumed that the mantle was homogeneous and believed that
present-day average earthquake foci were applicable. (With permission of Academic
Press.)

was not defined. A modification of the model was later suggested by

Tarakanov and Leviy (1968). who proposed that each magma origi-
nated from special "plastic" layers alternating with "hard" layers.
They emphasized the layered structure of the Benioff zone rather than
the more widely held view of it as a continuous plate subducted into the
mantle.

DEPTH-OF-SEPARATION MODEL

In the same paper. Kuno (1967. p. 108. Figure II) described the ideas
of Yoder and Tilley ( 1962) in terms of the same environment of magma
generation. Although they did not relate magma generation specifically
to the Benioff zone. they emphasized the importance of the depth of
separation of magma, presumably in auxiliary chambers. from the
parental material. as illustrated in Figure 1-6. The magma consisting of
crystals and liquid was believed by Yoder and Tilley to originate at
relatively uniform depths. A zone having low seismic velocities had
been described by Gutenberg (1926). and among the possible interpre-
tations it was considered to be the result of partial melting. Because the
liquid in the crystal-liquid mush changes composition as it rises to the
surface in response to the change in physicochemical relations, dif-

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Introduction 9
ferent magma types result, depending on the depth at which the liquid is
separated from the mush.
Neither the Kuno model nor his version of the Yoder and Tilley
model provides a reasonable explanation for the appearance of alkali
basalt as the last event in the formation of some volcanic islands (e.g.,
Hawaii). It does not seem appropriate for the last magma to come from
the greatest depth, usually after a considerable lapse of time, and be
required to penetrate the entire volcanic edifice. Note should be taken
of those islands (e.g., Lanzarote, Canary Islands; Jan Mayen; Mull,
Scotland) where alkali basalts form the shield and the last eruptives are
tholeiitic. Although alkaline basalts occurring well behind an island arc
do not appear to be restricted in rising from a deep source, those
alkaline rocks recently discovered in the zone closest to the associated
trench (Arculus and Curran, 1972) may encounter some difficulty in
rising to the surface if indeed they are derived from a region directly
under the volcano.

EXPLOSIVE-DIATREME MODEL

A composite model was proposed by McGetchin (1975, p. 40) to

describe the structure of explosive diatremes and other volcanic fea-

HansJD Pttt:lfit: Ot:Hn

Okl-dOQa On-take YatiJ9CI- Huzl Hakane
take '0"-slma
SE
0 ~////: 'i//////~Crust-; ':; ~~
: ~/./~//~ V//:
Mantle
~
.::;::.-:.::.~:·:

- -
400-\~----------~----------~Jl''-----~----~'l~--~---JJ
Alkali olivine basalt mavma Hl9h- alumina basalt Tholeiite mavma
mavma

FIGURE 1-6 Cross section of central Honsyii from the Izu Islands to Oki-dOgo Island
showing depths of generation of primary magma (stippled area) and sites of liquid
separation (irregular black mass on vertical feeder) after Yoder and Tilley's hypothesis
as interpreted by Kuno (1967, p. 108, Fig. II). Not related to present-day earthquake
foci in Benioff zone by Yoder and Tilley ( 1962). (With pennission of Academic Press.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

10 GENERATION OF BASALTIC MAGMA

tures (Figure 1-7). He emphasized the importance of gases diffusing (A)

from depth, thereby fluxing an irregular region (B) in a heterogeneous
layered mantle. McGetchin shared the view of Jackson and Wright
(1970, p. 425, Figure 6), who showed refractory layers intercalated with
more fusible layers in a schematic section for Hawaii. The boundary of
the region is presumably related to the solidus of the volatile-fluxed
rock [McGetchin has agreed with this interpretation (personal com-
munication, 1975)]. It is interesting to speculate whether the crack (C)
penetrated the molten region or whether it was fortuitously intercepted

FIGURE 1-7 Composite cross section of volcanic systems accord-

ing to McGetchin (1975. p. 40, Figure IV-7), showing diffusion of
volatiles (A), site of magma generation {B), dike tapping magma
chamber (C). auxiliary reservoir (D), and explosion crater(£). (With
permission of Los Alamos Scientific Laboratory.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Introduction ll

by an apophysis of the magma chamber. McGetchin has shown an

auxiliary magma chamber (D), but it pertains only to quiescent erup-
tions and is not part of the explosive diatreme system. Compare the
size of the surface expression of the diatreme (£) and the melt region,
recalling that much material in a diatreme is composed of debris from
the layers penetrated.

CONCLUSION

The purpose in presenting these models is to illustrate the range of

ideas and to give an overview of some of the issues to be examined.
Other models will be introduced as appropriate .

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental
2 Materials

PRINCIPAL REQUIREMENT-YIELD
BASALTIC MAGMA
To begin the analysis of the problem of magma generation, it is
necessary to identify the parental material to be melted because
seismic evidence, reviewed below, indicates that, for the most part, the
mantle is crystalline. The principal requirement is that the rock or
rocks yield a liquid, without crystals, of basaltic composition or a
precursor crystal-liquid mush from which such a liquid may be derived
at or near the surface of the earth . Basalt has appeared in great volume
throughout geologic time and over most of the earth's surface. It is
observed today extruding occasionally in an all-liquid condition, and
older basalts are found in the glassy state, indicating that they too were
once entirely molten. In a broad sense, older and present-day basalts
are very similar in composition. Many other major rock types can be
related by crystal fractionation, for example, to basaltic compositions
(Bowen, 1928), and therefore basalt is usually considered "parental"
to other "daughter" magmas.* Of the major rock types, then, basalt

*But not all igneous rocks need have a basaltic parentage. Andesites, for example, may
not necessarily be derivative from basalt; they could form directly from more primitive
material (Yoder, 1969, p. 85).
12

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 13

oo- 0---~~---o-o--o--.
0

FIGURE 2-1 Plot correlating liquidus temperature with iron enrichment in natural
basic rocks determined in the laboratory. Solid dots are liquidus temperatures; open
circles mark the temperature where aU major phases begin to precipitate together. From
Tilley et a/. (1964, p. 95, Figure 23). (With permission of the Carnegie Institution of
Washington.)

should consequently have a higher liquidus* temperature than the

daughter products. Because of the great importance of the relationship
of liquidus temperature to composition, some details supporting the
parent-daughter principle will be given .

PARENT-DAUGHTER TEST
In Figure 2-1 is a plot correlating experimental determinations of
liquidus temperatures with iron enrichment in natural rocks. The liquidus
temperatures of rocks ranging in iron enrichment from 0.5 to nearly

*The liquidus is the maximum temperature of saturation of a solid phase, the primary
phase, in a liquid phase for a given bulk composition. Above that temperature the system
is completely liquid. In other words, it is the temperature at which the first crystal begins
to precipitate from a liquid under equilibrium conditions. In a system of variable
composition, the liquidus is the locus of temperature-composition points representing the
maximum saturation of a solid phase in the liquid phase.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

14 GENERATION OF BASAL TIC MAGMA

1.0 can be represented approximately by a straight line. The

approximation of a straight line is supported by the determined phase
relations in the diopside-forsterite-albite-anorthite system (Yoder and
Tilley, 1962, p. 395), where the principal variation is in the plagioclase
solid solution. A basaltic magma having an iron enrichment of0.5 could
presumably yield by fractionation all the liquids with successively
higher iron-enrichment values. That is to say, a magma with a higher
liquidus temperature is a possible parent to daughter magma having a
lower liquidus temperature. Conversely, it would be physicochemi-
cally unlikely, within reasonable bounds of P 02 , for a magma having a
relatively high iron-enrichment value to yield a liquid with a lower
iron-enrichment value. The reader may wish to verify this principle by
considering the forsterite-fayalite system (Figure 2-2). Similar argu-

;
!1500
:
!

Olivine

[Link] percent

FIGURE 2-2 Phase equilibria in the forsterite-fayalite system at I atm [after Bowen
and Schairer (1935, p. 163, Figure 7)] with an estimate of the stability of iron (Fe•), a
product of the incongruent melting of the iron-rich olivines. (With permission of the
American Journal of Science.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 15
ments can be presented for liquidus temperature variations with some
other major components (e.g., AI 20 3/Ca0 or MgO).

ASSUMPTIONS FOR TEST

Most magmas having an iron enrichment less than 0.5 have accumu-
lated a crystalline phase, commonly olivine, and, therefore, experi-
mentally determined liquidus temperatures are unduly high and not
applicable. The group of magmas represented by rocks such as picrites,
ankaramites, and oceanites are not found, after natural or dip quench-
ing, wholly in the glassy state. Obviously, one cannot apply the
parent-daughter test to rocks formed by the accumulation of crystals.
Yoder and Tilley (1962) assumed that the rocks investigated were at
one time all liquid and that volatiles were not an important constituent.
(As will be seen below, volatiles would lower the liquidus temperature
and displace the liquid to a lower derivative status. Rocks having lost
their volatiles would melt at anomalously high temperatures in the
laboratory.)
The influence of the accumulation of crystals in a magma on the
subsequent liquidus temperature measured in the laboratory can be
readily appreciated by examining a simple ternary system. In Figure
2-3 the composition P initially crystallizes olivine, which is joined by
anorthite when the composition of the liquid reaches L with decreasing
temperature. If olivine accumulates in the liquid L, the bulk composi-
tion of the accumulated layers might range from P through A to
forsterite, depending upon the proportion of trapped liquid. The liquidus
temperatures of the crystallized cumulates when measured in the
laboratory would be anomalously high relative to that of the parental
material P. Remelting such cumulates (Wager eta/., 1960) would, of
course, produce the same series of derivative magmas as the parental
material. Rare olivine-rich liquids such as A (representing picrites or
komatiites,* for example) cannot always be dismissed as remelted
cumulates, but may be the subsequent products of liquids generated
directly by more advanced partial melting of parental materials if the
temperature or other conditions are appropriate . In order to obtain the

*Thin peridotitic lava flows occur in the Archean Komati formation of Barberton
Mountain Land. Republic of South Africa (Viljoen and Viljoen. 1969a.b,c). Quench
textures (e.g.. spinifex texture) indicate that these ultramafic rocks were essentially all
liquid, and the quench textures have been reproduced experimentally from natural
samples (Green eta/.. 1975). The l)ormative olivine content ranges from 55 to 75 wt%.
Similar rocks have been found in Ontario and Quebec . Canada (Pyke eta/. , 1973). and
western Australia (Nesbitt. 1971).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

16 GENERATION OF BASALTIC MAGMA

Fonterite

Olopalde Welljht percent

FIGURE 2-3 Phase equilibria in the forsterite-diopside-anorthite system [after Os-

born and Tait (1952. p. 419. Figure 5)). Note change of liquidus temperature as liquid L.
derived from bulk composition P. accumulates forsterite. If23 percent forsterite accumu-
lates in the liquid. the composition of liquid plus accumulated crystals becomes A. and
the liquidus temperature is raised IOO"C above that of the parental material. (With
permission of the American Journal of Science.)

necessary high liquidus temperatures, greater than usual depths of

generation are envisaged where the mineralogy of the parental material
may involve special high-pressure phases (see Figure 4-2) that yield on
melting a composition more closely related to such ultrabasic rocks.

BASALTS EQUILIBRATED TO SURFACE CONDITIONS

Yoder and Tilley (1962, pp. 382-383), after examining a large variety of
natural basalts at I atm, noted that all major phases (olivine,
clinopyroxene, orthopyroxene, and plagioclase) can appear as primary
silicate phases on the liquidus. They also found that the temperature at

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 17
which the primary silicate appears lies within a narrow range irrespec-
tive of the kind of primary silicate phase. In addition, all major silicate
phases appear, with rare exceptions, within a small temperature inter-
val, about sooc; that is to say' these phases begin crystallizing together
at about the same temperature. Yoder and Tilley interpreted these
observations to mean that most basalts are eutecticlike* and therefore
are products of fractional crystallization. It appears that basalts in a
broad sense have equilibrated to the conditions at the earth's surface.
In other words, after separation the composition of the liquid has been
adjusted successively-presumably by crystallization and removal of
appropriate phases-on its rise to the surface so that it is approxi-
mately in chemical equilibrium at about I atm.
Only rare lavas are believed to have compositions inherited from
depth in the mantle. Those alkali basalt lavas containing nodules that
have phases (e.g., orthopyroxene) incompatible with the magma at low
pressures may exhibit such inheritance. In addition, Kushiro and
Thompson (1972, p. 405, Figure 23; see also Kushiro, 1973a, p. 215,
Figure 1) showed experimentally that one abyssal olivine tholeiite
exhibited eutecticlike behavior at 7.5 kbar and in one interpretation
presumed that the liquid had been separated from its parent at a depth
of about 25 km. They assumed that no crystallization of the liquid took
place en route to the surface and that the composition of the liquid is
that of the eutectic of plagioclase peridotite at the determined pressure.
These observations are considered to be clear evidence of the
physicochemical controls on basaltic magmas. Basalts are not just
random products, nor are they themselves primary liquids.
To the best of the writer's knowledge, glass with an iron enrichment
of less than 0.5 is exceptionally rare. A few occurrences of such rocks
(e.g., komatiites) having unusual quench textures are believed by some
to indicate that the rocks were once all liquid. One "fresh" sample
having 6.63 H 20+ has an iron enrichment value of 0.263 (Green et al.,
1975). One possible origin for such rare liquids involves the presence of
volatiles, a compositional effect presented in Chapter 4, pp. 78-86.
*The eutectic is the lowest melting point of any ratio of components in a system. The
proportion of crystaUine phases to liquid at the eutectic can be changed by the addition or
subtraction of heat without change in temperature. The eutectic is the intersection of the
saturation surfaces of the participating crystalline phases in equilibrium with liquid at the
lowest melting point. A liquid of the eutectic composition will yield on crystallization
under equilibrium conditions all participating crystalline phases at the eutectic tempera-
ture if adequate heat is removed. A mixture of participating phases of the eutectic
composition will yield on melting under equilibrium conditions all liquid if adequate heat
is supplied at the eutectic temperature.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

18 GENERATION OF BASAL TIC MAGMA

GENERAL TYPES OF SOURCE MATERIALS

The rock types considered as potential source materials for basaltic

magma are (a) those that are essentially of basaltic composition and (b)
those that are alleged to yield a basaltic magma on partial melting. The

t
Boyd and Eniiond (1959)
I 1
I I 120

I I
I I
• I I 100
•t-f
I I
I .., I ...~
I ll
~I 80 E
a;

. ._,.,
Eclogite 2
I I ~
I I ~
Q.
0
...
------ I 60
...,...,
--------- ? I
lO~ I

---- ------T
-- .-
------

--
.
"'o''o"
~co~''-~ 01 ---~
I
1 Liquid .
..;
~
------ Kenne1y (1959) ------
.-.-*-----X X XX xt
I
88 e/o
1
40 t
...-
...----
------
Basalt 1
I
/Xtla+L/
1
I
20 t\i
i
,.._
I I II
~
I I
I I
0 1600
Temperature, •c.

FIGURE 2-4 Pressure-temperature relations determined experimentally for a natural

eclogite by Yoder and Tilley ( 1962, p. 498, Figure 43), a natural basalt converted to glass
by Kennedy (1959), and a tachylite by Boyd and England (1959b, p. 88). The symbols
mark the conditions of the experiments that define each region . The question mark
denotes a region of complex transformations not elucidated by the experiments. Note the
restriction of basalts to depths less than 50 km. (With permission of the Oxford
University Press.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 19
first group includes basalt, gabbro, tachylite, amphibolite, hornblen-
dite, and eclogite. Recognition of the importance of olivine in the
parental material leads to a consideration of the second group, which
includes meteorites, plagioclase peridotites, spinel peridotites, and
garnet peridotites. Each of these rock types will be examined in tum.

BASALTIC OR GABBROIC SOURCE

The simplest source of basaltic liquid is a crystalline parent of basalt or
gabbro. Obviously, it would be necessary to melt the entire parental
material to recover the bulk composition desired. Because of crystal
settling, fractionation, and the effect of nonhydrostatic stresses, it is
not likely that complete melting of a material would ever be achieved in
situ. If that basalt or gabbro composition is fortuitously at a depth
where it corresponds to an invariant (e.g., a eutectic) point, or nearly
so, then partial melting will yield the same liquid composition as
complete melting (see, for example, phase relations of abyssal olivine
tholeiite at 7.5 kbar described by Kushiro, 1973a, p. 215, Figure 1).
Such a parental material was suggested as early as 1858 by von Cotta
and later by W. L. Green ( 1887).
Instead of considering a wholly crystalline basalt or gabbro, some
petrologists suggested starting with a liquid of basaltic composition, a
residuum of the processes that generated the earth itself. Daly (1925a)
proposed the existence of a glassy layer (tachylite), a possibility that
has the advantage of requiring essentially no enthalpy of melting. He
supposed that the vitreous basalt occurred at a depth of 60 km
(= 17 kbar), where the temperature was believed to be above the
melting temperature of crystalline basalt, perhaps 1300°C. He argued
that pressure increased the viscosity of the liquid so that it was a
rigid material and responded seismically as a glass. At the tempera-
tures believed to exist in the mantle, however, tachylite <Vv = 6.5
km/sec) does not have the appropriate seismic velocities to be con-
sidered as a parental material (Adams and Gibson, 1926; Birch and
Bancroft, 1942). * Furthermore, it is unlikely that a glass would remain
vitreous at the range of temperatures and pressures within the upper
*Bridgman (see Daly, 1933, p. 190) obtained a lower compressibility (and therefore a
higher velocity is obtained on calculation) on a Hawaiian tachylite. but it contained 20
percent crystals, whereas the tachylite used by Adams and Gibson (1926, p. 276)
contained only 3 percent crystals. [Adams and Williamson (1923) gave useful formulae
for calculating velocity from the compressibility, {3, and the density, p, where Poisson's
ratio is assumed to be 0.27: V,, = 13.13(106{3p)- 112 ; V, = 7.37(10 6{3p)- 112 .]

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

20 GENERATION OF BASALTIC MAGMA

mantle. All these suggestions merely delay consideration of the ques-

tion of the origin of the layers themselves, but there are even more
serious objections to having crystalline basalt or gabbro or their glassy
equivalents as parental material.

ECLOGITIC SOURCE-HIGH-PRESSURE
EQUIVALENT OF BASALT
Yoder and Tilley ( 1957) found that basalt (and gabbro) are stable only at
shallow depths (Figure 2-4), so it is clear that these rocks could be a
source for only relatively shallow magmas. It was demonstrated further
that all basalts investigated could be converted into eclogite. The
eclogites formed experimentally consisted of a simple mineral as-
semblage, predominantly clinopyroxene and garnet. Some of these
eclogites also contained small amounts of hypersthene, quartz, or
kyanite, individually or in various combinations. Conversion of the
mineral assemblage characteristic of basalt to that of eclogite has been
substantiated by experiments on combinations of the simple end-
member minerals. In addition to the reactions of anorthite with forster-
ite and enstatite, discussed below, the breakdown of albite (Birch and
LeComte, 1960) and the interaction of clinopyroxene and plagioclase
(Kushiro, 1969a, p. 187, Figure 7; Akella and Kennedy, 1971, p. 162,
Figure 2) yielded minerals critical to the edogite assemblage at ele-
vated pressures. The idea that eclogite might be the high-pressure form
of gabbro was first put forth by Fermor (1913).
The possibility of eclogite as parental material for basaltic liquids
does not present the same difficulty as a pre-existing basalt and gabbro
source, where the entire rock must be melted to make basaltic magma.
As shown in Chapter 8, eclogites melt over a very narrow temperature
range (see Figures 2-4 and 5-5), and each melt fraction yields in general
a basaltic composition (Figures 8-5 and 8-6). Whereas there is no
objection, disregarding physical problems, to the melting of a rock in
toto, there would be objection to the melting of that rock if it had a
higher melting temperature than its host rock! For example, the model
of Coats (Figure 1-4) includes the possibility of pockets of eclogite in
peridotite as a magma source, and Ringwood (1969, p. 12, Figure 5)
also illustrated magma issuing from eclogite in peridotite below 100
km. As will be seen below, at certain depths eclogite would melt at a
higher temperature than its host, for example, garnet peridotite, and
therefore would more likely represent a residuum than a source for the
liquid.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 21

X 0 0 0
Liquid+ Gas

Amphibolite

X X )(

X )( X
I

500 600
Olivine

700
--
Tholeiite~

800 900 1100 1200 1300

Temperature, oc
FIGURE 2-5 Projection of the pressure-temperature relations of a natural olivine
tholeiite-water system determined by Yoder and Tilley (1962, p. 449, Figure 27). Note
the restriction of basalt to low pressures (shallow depths) and the broad field of stability
of amphibolite. Symbols mark the conditions of the experiments used to define each
significant change in assemblage. x = crystalline; ® = crystals + liquid; 0 = liquid .
(With permission of the Oxford University Press.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

22 GENERATION OF BASALTIC MAGMA

AMPHIBOLITIC SOURCE
Because amphibolites have essentially the composition of basalt plus
water, they too have been proposed as a source rock for basaltic
magmas at high water pressures. Wagner (1928) considered this rock
type to be the major parent of basaltic magmas. Yoder and Tilley (1962)
demonstrated that amphibolites could indeed be readily made from
basalts and gabbros by adding water at high pressures. Furthermore,
basaltic liquids in the presence of water can be quenched in the
laboratory entirely to amphibole (Figure 2-5), and it is interesting to
note that Lacroix (1917) suggested hornblendite as the equivalent of
basalt plus water. Unfortunately, amphiboles have their pressure-
temperature limits (Gilbert, 1%9). Lambert and Wyllie (1970) showed
(Figure 2-6) that amphibole was no longer stable at pressures in excess
of about 30 kbar (=100 km).
Not a single residual amphibole has ever been recorded in the entire
Hawaiian Island chain even though Tuthill (I %9) suggested it as the
key mineral in the source rock of those basaltic magmas. Jackson and
Wright (1970, p. 414), on the other hand, found amphibole not uncom-
mon in nodules within alkaline basalts, but it was not present in the
basalts themselves. A genetic connection has not yet been established
between the amphibole-bearing nodules and the basalts. There is no

Temperature, "C Temperature, •c

FIGURE 2-6 A. Generalized diagram for the melting interval of basaltic material, dry
(diagonal ruling) and in the presence of excess water (stippled). The dotted curves are
estimated geotherms for the regions under oceans (lower temperatures) and under shield
areas (higher temperatures).
B. Estimated melting interval of basaltic material in the presence of 0.1 wt% water.
Basic liquid is produced only within the ruled area. Hb = hornblende. From Lambert and
Wyllie (1970, p. 765, Figures I and 2). (With permission. Copyrighted by the American
Association for the Advancement of Science.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 23

•
I~
.r20
...
:s
••
~

1200
Temperature, •c

FIGURE 2-7 Experimentally determined stability limits of am-

phiboles in equilibrium with excess H 20 formed from the following
natural samples:(/) kaersutite megacryst, Kakanui. New Zealand;
(2) brown hornblende mylonite, St. Paul's Rocks, equatorial mid-
Atlantic ridge; (J) olivine nephelinite. Oahu, Hawaii; (4) kaersutite
eclogite, Kakanui, New Zealand; (5) high-alumina olivine tholeiite,
Soay, Scotland; (6)alkali olivine basalt, Hualalai, Hawaii; (7) "alkali
olivine basalt," synthetic mixture; (8) olivine tholeiite, Kilauea,
Hawaii; (9) quartz tholeiite. Picture Gorge. Oregon. Summarized by
Merrill and Wyllie (1975, p. 564. Figure 7). (With permission of the
Geological Society of America.)

evidence of explosive action. such as accompanies volatile release at

low pressures. to indicate that amphiboles had broken down on the
way to the surface. The breakdown of a wide variety of amphiboles at
pressures less than about 30 kbar has since been substantiated by other
studies (Figure 2-7). In addition, an amphibolite or hornblendite source
of basaltic composition begs the question in the same way as basalt or
gabbro does. It is concluded that amphibole does not play a major role
in the generation of the most common basaltic magmas. but its pres-

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

24 GENERATION OF BASALTIC MAGMA

ence at depths <I 00 km may influence the fractionation trends (Bowen,

1928, p. 269-273). Of all the named rock types, only eclogite is still
considered a potential source of basaltic magmas, but it lacks what
appears to be a key phase, namely olivine .

OLIVINE-AN IMPORTANT SOURCE-ROCK

CONSTITUENT?
Glassy basalts with the highest liquidus temperatures at I atm have
olivine on the liquidus . It appears , therefore , that olivine would proba-
bly be a key phase in the parental material. That is, if a derivative
magma has olivine on the liquidus, olivine should be an important
phase in the parent. This principle does not imply that because olivine
is an important phase it is also the most abundant phase. Magnesian
olivine has a very high melting temperature; therefore, in systems
containing forsterite the olivine usually has a very large liquidus field .
This effect usually implies that the composition of liquid at the begin-
ning of melting of the system is displaced by a large amount from
forsterite itself. The concept can be illustrated with a simple ternary
diagram such as Figure 2-3 . If the bulk composition of the parent has
olivine on the liquidus (point A or P), one can expect that a liquid of the
eutectic composition E removed from the parent would also precipitate
olivine on cooling under the same pressure. Alternatives to this con-
cept are presented below . Other arguments in support of the idea that
olivine is an important phase in the parental material are given in the
following sections.

METEORITIC SOURCE
Silicate-rich meteorites have long been believed to be representative of
material comprising the earth's interior. Chondrites, the most abundant
(90 percent), are relatively homogeneous in major components . They
are probably parts of the solar system and not stray debris from outside
it. Their age (4.5 b.y .) is taken to be that of the earth . Some of the isotope
ratios, however, are not appropriate , and the abundances of heat-
producing elements (K , U , Th) are not wholly consistent with an earth
made of chondrite. The K/U is about 8 x 104 for chondrites compared
with 104 for most terrestrial rocks (Birch, 1965). The 87 Sr/86 Sr is
excessively high in chondrites, 0.755, compared with that of terrestrial
basalts,0.705 (Gast, 1960), or oceanic basalts, 0.7026(Hart, 1971). Some
have believed that achondrites are more appropriate representatives of
the earth's interior (Washington, 1925, p. 357; Lovering and Morgan, ·

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 25
1964), a view that is supported by recent isotopic data on oxygen
(Clayton and Mayeda, 1975). Others have thought that carbonaceous
chondrites are more representative because they are the least differen-
tiated meteorites and have more appropriate abundances of the more
volatile components (Reed eta/., 1960, p. 136).

IMPORTANT PHASES IN METEORITES

All three types of chondrites include members in which olivine is a

predominant phase. There are also members rich in orthopyroxene. A
petrologist would call such rocks peridotites or, more specifically,

Temperature, •c

FIGURE 2-8 Pressure-temperature plane for the I : I composition

(molecular ratio) of anorthite and forsterite . Symbols indicate the
conditions of the experiments: x = crystaUine; ® = crystals
+ liquid; 0 = liquid. An = anorthite; Cpx = clinopyroxene; Opx =
orthopyroxene; Sp = spinei; Gr = garnet;Fo = forsterite;L =liquid.
From Kushiro and Yoder (1966. p. 340. Figure I). (With permission
of the Oxford University Press.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

26 GENERATION OF BASALTIC MAGMA

)(
I
kr+ll
I
I
I
)( I

Gr+QHPX
)(
)( )( 'II,
)( 1
·~
?
,,• ,I
•i •I
,,,,
I -fSp+l
~'-XIs+l
I I
)( I I
)(

An+Opx+Cpx+Qz x I
1 '.,1
I
lI l l
I

'/
?I

18oo 1800

FIGURE 2-9 Pr~ssure-temperature plane for the I : 2 composition

(molecular ratio) of anorthite and enstatite. Symbols indicate the
conditions of the experiments: x = crystaUine; ® = crystals
+ liquid; 0 = liquid. Abbreviations as in Figure 2-8; and En =
enstatite; Px = pyroxene; Qz = quartz; XIs = crystals. From
Kushiro and Yoder (1966, p. 349, Figure 3). (With permission of the
Oxford University Press.)

harzburgites. Some members also contain clinopyroxene, but none

contains garnet* of a composition appropriate for the formation of
basaltic liquids. Most importantly, some chondrites contain plagio-
clase; therefore, they appear to be low-pressure assemblages. They
cannot be products of the complete disruption of another large planet;
there is, however, the remote possibility that they came from the

*The garnet in the AUende meteorite, for example, is grossularite-rich , whereas the
garnets from ultrabasic rocks believed to have been derived from great depths are
predominantly pyrope-almandine in composition . Very rare nodules of grospydite from
Siberian kimberlite pipes consist of grossularite + clinopyroxene + kyanite (Sobolev et
al .. 1968).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 27
breakup of the outer shell of a large planet. Presumably, meteorites are
condensed matter not incorporated in the aggregation of the primordial
planets of the solar system. What then is the evidence that silicate-rich
meteorites are low-pressure assemblages? The key relationship is the
coexistence of plagioclase with olivine or orthopyroxene.

PHASE INCOMPATIBLE WITH OLIVINE AT HIGH PRESSURE

Anorthite and forsterite react at relatively low pressures (8 kbar), as

illustrated in Figure 2-8 (Kushiro and Yoder, 1966, p. 340), to form
orthopyroxene + clinopyroxene + spinel. It was also shown that
anorthite and enstatite are restricted to relatively low-pressure envi-
ronments. They react to orthopyroxene + clinopyroxene + quartz at
about 15 kbar (Figure 2-9). The presence of albite in a plagioclase
would raise the pressure limit of stability (Emslie and Lindsley. 1969,
p. 109; K. E. Windom, personal communication, 1975), and iron in an
olivine would lower the pressure limit of stability (Green and Hibber-
son, 1970; Emslie, 1971, p. 155). If meteorites are indeed debris from a
disrupted planet, it would have to have been much smaller in diameter
than the moon. which has a pressure of 50 kbar at its center. On the

TABLE 2-1 Comparison of Chemical Composi-

tion (wt%) of Meteorite (Adjusted) with that of
Peridotite
Allende. Salt Lake Crater,
Mexico, Hawaii.
Oxide Meteorite" Peridotite"
Si02 46.50 48.26
Ti02 0.22 0.22
Al.o, 4.56 4.90
cr.o, 0.70 0.25
FeO 9.91 9.94
MnO 0.24 0. 14
MgO 33.31 32 .51
CaO 3.59 2.98
Na.O 0.60 0.66
K20 0.04 0.07
P.o. 0.32
- -- O.o7
TOTAL 99.99 100.00
FeO/(FeO + MgO) 0.229 0.234
"Aller subtraction or the following elements by Seitz and Kushiro ( 1974. p. 954.
Table 1): 0.. 4.73 wt% ; Fe. 19.2 wl%; Ni. 1.45 wl% ; S. 2.21 wt%.
'Kuno and Aoki (1970, p. 276, Table 2, No. 24) recalculated after subtraction
orH,O' .

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

28 GENERATION OF BASAL TIC MAGMA

Garnet Peridotite

!::0
• Crystals
f +
CL

_,-'
.,., Spinel Peridotite
.,~ Liquid

-----
Ploviocla.. Peridotite

Temperature, ec

FIGURE 2-10 Generalized assemblage diagram for various perido-

tites based on experimental data in Figure 2-8. Depth scale based on
average density of 3.2 g!cm". SFC = surface.

basis of Fe/Mg part1t1oning between olivine and orthopyroxene,

silicate-rich meteorite assemblages, even though brecciated, generally
appear to be in equilibrium.* The olivine of meteorites, ranging in
composition from FoH,, to Fo7., would be suitable as a constituent of a
parental material capable of yielding a residue having a modal olivine of
Fo00 , but meteorites would not be suitable for yielding liquids with
normative Fooo. [The reader may again wish to refer to the forsterite-
fayalite phase diagram of Bowen and Schairer (1935, p. 163, Figure 7)
revised in Figure 2-2.] The prevailing oxidation state would be ex-

*Van Schmus and Wood ( 1967. p. 757. Table 2. i) have based a classification of
chondrites on the degree of equilibration of olivine and orthopyroxene; however. the
samples displaying a lack of homogeneity are believed to be few (8 . Mason. personal
communication, 1976).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 29
tremely important, and meteorites are known to have a very wide range
of oxidation.

METEORITES RELATED TO PERIDOTITE?

The compositions of silicate-rich meteorites can be adjusted on reason-

able grounds so that their norms* are comparable to those of peridotite
found on earth (Table 2-1). It is necessary to remove metallic Fe, Ni,
and S, as well as oxygen (except from some reduced chondrites). (If
meteorites are representative of the material condensed to form the
earth, then the Fe, Ni, and S are presumably now in the core.) The
existence of meteorites rich in olivine provides major support for the
argument that peridotite is the parental material in the earth . On the
other hand, the presence of plagioclase and the absence of garnet, a
major contributor to melts of basaltic composition at high pressure,
would appear to constitute serious objections to the acceptance of
silicate-rich meteorites per se as mantle material unless metamor-
phosed. On the basis of the experiments summarized in Figures 2-8 and
2-9, it was demonstrated that meteorites are low-pressure assemblages,
and it now remains to be shown that the meteorites are related to
various other types of peridotite found in the earth. Figure 2-10, based
on the work of Kushiro and Yoder (1966) and MacGregor (1968),
illustrates schematically the relationship of plagioclase peridotite to
spinel peridotite and garnet peridotite.
Spinel peridotites are found as large intrusions in the central parts of
fold belts, as nodules in alkali basalts, and as small intrusions in large
fault zones. Gamet peridotites are commonly found in fold belts
subjected to high-grade metamorphism and as nodules in alkali basalts
and kimberlite pipes. Gamet peridotites have a relatively narrow
range of composition and are of worldwide occurrence. It would
appear from Figure 2-10 that garnet peridotites are stable only at
depth in the upper mantle. One can obtain an appreciation of just how
deep in the mantle these garnet peridotites are derived by examining
the compositions of the coexisting minerals.

*The norm is the calculated mineral composition of a rock in terms of hypothetical

standard mineral end members similar, especially in basalts, to those of the actual modal
minerals that occur or would form on complete crystallization from the same chemical
constituents at relatively low pressures in the absence of volatiles. The norm was devised
by Cross, Iddings, Pirsson , and Washington in 1902 and is commonly referred to as the
CIPW system, after the first letter of the last name of each author. The method of
calculation may be found in Washington (1917, pp. 1162-1164), Holmes (1921 , pp.
427-431), Johannsen (1939, pp. 88-92), and other petrological works.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

30 GENERATION OF BASALTIC MAGMA

Ga net

Lherzolite

HarzburQitt

FIGURE 2-11 Nomenclature tetrahedron for assemblages of olivine, clinopyroxene,

orthopyroxene, and garnet. Rock names underlined lie within the tetrahedron.

GARNET PERIDOTITES FROM THE UPPER MANTLE

The principal minerals of garnet peridotites are olivine, orthopyroxene,
clinopyroxene, and garnet. The compositions of these phases may be
displayed in a tetrahedron, and appropriate names may be assigned to
their various combinations (Figure 2-11). The compositions of the
coexisting pyroxenes change as a function of temperature , and the
solvus relating these changes, determined experimentally by Davis and
Boyd (1966) and Boyd (1970), is relatively insensitive to pressure
(Mysen, 1976a). In addition, the alumina content of the orthopyroxene
in the presence of garnet is a function of pressure,* and that relation-
ship has been determined experimentally by MacGregor (1974). The
relationship of the alumina content of orthopyroxene to pressure is also
sensitive to temperature, so that the temperature must be estimated first
from the two-pyroxene solvus. These indirect indicators of pressure
*The alumina content of orthopyroxene in the presence of spinel does not appear to be
sensitive to pressure (Obata, 1975, p. 2, Figure I; Fujii and Takahashi , 1976).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 31
and temperature* are considered to be relatively insensitive to minor
elements (e.g., Ti: Akella, 1974).
With this information, Boyd and Nixon (1973) have tested the
pyroxenes in nodules of garnet peridotite from kimberlite pipes in
which the nodules are thought to have been carried up rapidly from
great depths in explosive events. The pressure and temperature de-
duced from each nodule are assumed to define the conditions under
which the nodule last equilibrated. Also in the pipes is a large array of
other assemblages related to garnet peridotite: for example, harzbur-
gite, lherzolite, websterite, dunite, and rarely eclogite. Similar nodules
are also found in alkali basalts but are exceptionally rare in tholeiites.
The results that Boyd and Nixon obtained from nodules in the pipes of
northern Lesotho, shown in Figure 2-12, indicate that the nodules
record conditions of very high pressures and temperatures of equilibra-
tion. Metamorphic and metasomatic effects would tend to decrease the
magnitude of these values. Although the interpretation of the abrupt
change in the curvature of some of the pyroxene geotherms is being
debated (e.g., Mercier and Carter, 1975; Goetze, 1975), the range of
physicochemical conditions registered by the nodules from the upper
mantle is of great importance.

KEY POSITION OF GARNET PERIDOTITE

It is convenient to illustrate the assemblages found in the nodules

schematically in the Ca0-Mg0-A1 20a-Si02 system (Figure 2-13). The
diagram brings to light the key position of garnet peridotite (Fo + En +
Di + Py) relative to the eclogite assemblage (Di + Py), which provides
most of the components for basaltic liquids, as well as other peripheral
assemblages recorded mainly as nodules from kimberlite pipes. Several
questions arise: (a) Are these assemblages representative of the pri-
mary mantle itself? (b) Are they the cumulates from liquids produced
on partial melting of the mantle? Or (c) are they the residua from the
melting of the mantle? It does not appear to be possible at present to
reach an unambiguous decision on the basis of chemistry alonet;
however, other tests are presented below (pp. 107-109).
*Other estimates of the temperature or pressure of formation of garnet peridotite can be
made using the Fe-Mg partitioning between the following mineral pairs after further
calibration:
Opx-Cpx (Kretz. 1961)
Opx-Gr (Wood and Banno, 1973)
01-Cpx (Obata eta/., 1974; Powell and Powell. 1974)
Cpx-Gr (Raheim and Green. 1974; Akella and Boyd, 1974)
tAn example of the graphical techniques for relating parental material. primary liquid,
residua, and cumulates is given by O'Hara eta/. ( 1975, p. 585 ff.).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

32 GENERATION OF BASALTIC MAGMA

1600

Pyroxene Geotherm

FIGURE 2-12 Estimates of temperatures and depths of equilibra-

tion of lherzolite nodules (granular. open circle; sheared. filled circle)
primarily from kimberlites of northern Lesotho. Temperatures were
estimated from the diopside solvus of Davis and Boyd ( 1966) using the
Cai(Ca + Mg) ratio of the natural diopsides. Pressures were esti-
mated from the data of MacGregor ( 1974) using the raw Al2 0,. content
of the natural enstatites. From Boyd and Nixon ( 1973. p. 432. Figure
1). (With permission of the Carnegie Institution of Washington.)

WILL GARNET PERIDOTITE YIELD BASALTIC MAGMA?

If the reader accepts the conclusions that (a) garnet peridotite has the
appropriate mineralogy for the conditions of the upper mantle, (b) the
bulk composition is consistent with the source materials forming the
earth, and (c) the rock is prominent among the recovered deep-seated
samples, it is necessary then to ask whether garnet peridotite is
capable of yielding material of basaltic composition.
Unfortunately, only one garnet lherzolite and one spinel lherzolit~
have been studied at a series of pressures and temperatures in the

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 33

FIGURE 2-13 The Ca0-Mg0-AI2 0 3 -Si02 system, illustrating the

position of the eclogite assemblage (Py-Di tie lines), representing
basaltic compositions. and its relationship to the presumed parental
garnet peridotite assemblage (Fo + En + Di + Py). One type of solid
solution each is iUustrated for Di and Py. but none of the solid
solutions for other end-member compounds are shown . Stippled
planes are major ternary assemblages separating quaternary as-
semblages. which are Di + Py + Ky + Qz ; Di + Py +
Ky +Cor; Di + Py +Cor+ Sp; Di + Py + En+ Qz ; Di +
Py +En+ Fo;andDi + Py + Fo + Sp . Fo = forsterite;En = ensta-
tite; Di = diopside; Py = pyrope ; Sp =spinel; Cor= corundum;
CaTs= Ca-Tschermak's molecule; Ky = kyanite; Gross = grossula-
rite; Wo = wollastonite ; Qz = quartz. End-member compounds plot-
ted in mol percent. After O'Hara and Yoder (1967. p. 69. Figure 1).

laboratory. Ito and Kennedy (1967) found that a garnet lherzolite

transformed to an assemblage consisting of olivine + clinopyroxene
+ orthopyroxene + spinel at pressures Jess than 20 kbar. A spinel
lherzolite assemblage persisted at pressures less than 20 kbar, accord-
ing to the experiments of Kushiro eta/. (1968a), and transformed to
garnet-bearing assemblages above 20 kbar. The run times may have
been too short for equilibrium to be attained or the investigated
temperatures too high for the solidus of basalt to be reached, but the
main observation is that no plagioclase was seen. In view of the data of
Figures 2-8 and 2-9, the lack of plagioclase reaction relations at I atm.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

34 GENERATION OF BASAL TIC MAGMA

and the lack of unequivocal data on the natural rocks, the absence of
plagioclase suggests that (a) these lherzolites do not have the requisite
composition to produce a basalt or (b) the basaltic components had
already been removed-that is, the rock was depleted of basaltic
components by partial melting, for example, and is therefore a re-
siduum.
Alternatively, the normative An and Ab of the potential plagioclase
at 1 atm may be incorporated modally as tht. calcium Tschermak's
molecule and jadeite molecule, respectively, in clinopyroxene.
Clinopyroxenes may contain up to 26 wt% normative An at 1 atm,
according to Yoder and Tilley (1962, p. 406, Table 19). The normative
An content of the garnet lherzolite investigated by Ito and Kennedy
(1967, p. 521, Table 1, No.2) is 4.7 wt%, whereas the spinel lherzolite
investigated by Kushiro eta/. (1968a, p. 6024, Table 1) had a normative
An content of 10.2 wt%. Furthermore, a synthetic peridotite close to
Nockolds' (1954) average with 8 percent normative An yielded in the
crystal+ liquid region a glass ofbasaltic composition by chemical analy-
sis (Reay and Harris, 1964). Plagioclase was not observed as a phase in
those experiments at 1 atm . Assuming that equilibrium was achieved,
there appears to be a dilemma: How can one derive basalt, whose
principal phase is plagioclase , from a rock composition that does not
precipitate plagioclase at low pressures, especially 1 atm?
A sheared garnet lherzolite was partially melted (10-25 percent
liquid) by Kushiro (1973b, p. 297, Table 85), and the analyzed glass was
basaltic in normative character in spite of the fact that no plagioclase
was observed (Table 2-2). It appears that partial melting of some
undepleted peridotites can yield a liquid of basaltic composition if the
liquid is separated from the crystals. If the liquid is not separated from
the crystals. then the peridotite will not necessarily yield plagioclase,
which presumably remains occult in clinopyroxene.

METEORITES DO NOT YIELD BASALTIC MAGMA

Some peridotites appear to yield liquids of basaltic composition on
partial melting,* yet closely related chondrites, which have aggregated

*Peridotites that retain all the constituents required to form and yield a liquid of basaltic
(or eclogitic) composition are often referred to as undepleted. A peridotite from which
have been extracted some of or all the constituents required to form a liquid of basaltic
composition is considered by some investigators to be depleted. The applicability of the
tenns to a given peridotite depends on the past history of the rock, depth, temperature,
and composition(s) of the basaltic liquid to be extracted . The tenns are ambiguous and
should be explicitly defined if used.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 35
TABLE2-2 Compositions (wt%)ofGiasses Formed by Partial Melting
of a Natural Gamet Lherzolite (Kushiro, 1973b)
Eutectic
Chilled Composition
Starting 10 kbar. .15 kbar, 20 kbar, Bronzite at 40 kbar.
Oxide Material" 1375"C 1450"C 1475"C Gabbro• Di. Py.,Fo,En, 7

SiO. 44.54 50.3 51.4 50.3 51.33 50.16

AI.O. 2.80 12.4 14.1 12.3 13.69 11.89
TiO. 0.26 1.07 1.24 1.10 1.01
cr.o3 0.29 0.44 0.31 0.38 n.d.
Fe<Y 10.24 10.5 10.0 10.4 10.17
MnO 0.13 0.19 0.21 0.18 0.18
MgO 37.94 9.62 8.07 9.97 8.94 25.78
CaO 3.32 12.5 11.2 12.0 11.60 12.17
Na.O 0.34 0.110" 1.07" 1.49" 1.84
K.o 0.14 0.44 0.76 0.65 0.50
NiO n.d.' 0.07
- - -
n.d. 0.11 n.d.
TOTAL 100.00 98.3 98.4 98.8 99.871 100.00
FeO 0.213 0.522 0.553 0.511 0.532
MgO + FeO
Q 2.02 3.97
or 0.83 2.60 4.49 3.84 2.95
ab 2.88 6.n 9.05 12.6 15.57
an 5.70 29.0 31.4 25.0 27.62 32.44
di 8.54 27.1 19.8 28.3 24.56 21.75
hy 12.59 28.2 26.8 21.5 24.17 26.35
ol 68.54 4.96 2.47 19.46
cr 0.43 0.65 0.46 0.56
il 0.49 2.03 2.36 2.09 1.92
100.00 98.37 98.33 98.95 99.26 100.00
'"1'baba Putsoa. Northern Lesotho (Boyd and Nixon, 1973), specimen no. 1611. Recalculated on anhydrous basis.
"'Mutkox intrusion (Irvine. 1970).
'Total iron as FeO.
4J>.t of Na lost durinc analysis.
"n.d. ,. not determined.
1nctudinc p,o,. H.<>'.

to form the protoearth, do not yield liquids with appropriate iron-

magnesium ratios. The bulk sample of the unusual carbonaceous
chondrite from Allende was melted 25 percent at an oxygen fugacity
between that of the iron-wOstite and wOstite-magnetite buffers by
Seitz and Kushiro (1974, p. 955), and the melt was analyzed. The melt
is similar to some lunar ferrobasalts and has FeO/(FeO + MgO) equal
to 0.797. It is not sufficiently magnesian for the formation of common
terrestrial basalts; at a lower oxygen fugacity, however, the liquid
should be more magnesian because of the precipitation of a metallic
phase. Wager (1958, p. 38), after comparing the average analyses of94
chondrites and a weighted average of 4 peridotites, concluded on the
same basis that chondritic meteorites were not suitable mantle mate-

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

36 GENERATION OF BASALTIC MAGMA

rial. The writer believes that this conclusion should be considered as

tentative until the effects of P 02 are well established.

PHASE EQUILIBRIA SUPPORT FOR GARNET

PERIDOTITE SOURCE
Some support is gained from relevant phase diagrams for the conclu-
sion that on partial melting garnet peridotite will indeed yield a basaltic
composition. Only one example will be given here (Figure 2-14);
however, others are presented in Chapter 6. The anhydrous system

Oiopside

Orthopyroxene

Forsterite Enstotite Quortz

Weight percent

FIGURE 2-14 . The system forsterite-diopside-quartz at 20 kbar under anhydrous

conditions (dashed lines) and under excess H 20 conditions (solid lines). The anhydrous
invariant point Opx + Cpx + Qz + L is not shown. Adapted from Kushiro (1969b,
p. 282-283 , Figures 6 and 7). [The same system at 10 kbar was given by Warner ( 1973.
p. 938, Figure 4).] Invariant points A, B. and R. respectively, are analogous to andesite,
basalt, and rhyolite. (With permission of the American Journal of Science . )

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 37
diopside-forsterite-silica displays two invariant points at 20 kbar
(Kushiro, 1969b). The system contains three of the phases pertinent to
garnet peridotites, and the eutectic* composition, B, is representative
of an olivine-hypersthene normative basalt. (The second invariant
point, involving orthopyroxene + clinopyroxene + quartz + liquid,
has not been determined experimentally at 20 kbar and therefore does
not appear in Figure 2-14.)
The same system at 20-kbar water pressure also exhibits two in-
variant points. The invariant point A, applicable to the phases pertinent
to garnet peridotite, has a normative composition appropriate for
andesite. [The principal arguments relating to the production of basalt
under anhydrous conditions and andesite under hydrous conditions
from peridotite were outlined by Yoder ( 1969).] For some compositions
continued melting above the temperature of A would yield liquids on
the olivine-clinopyroxene boundary curve that would be analogous to
basalt. The important observations are that the entire range of compo-
sitions in diopside-enstatite-forsterite will produce, on melting, initial
liquids of basaltic composition under anhydrous conditions, whereas
more extensive melting is needed to produce liquids of basaltic compo-
sition under hydrous conditions.

SOURCE OF POTASSIUM?
One of the several deficiencies of the simple model of garnet peridotite
as represented by forsterite + enstatite + diopside + garnet is the
absence of a K20-bearing phase. The most likely candidate for the
K2 0-bearing phase is phlogopite because of its wide range of stability
and its presence in some nodules believed to come from the upper
mantle. Some investigators (e.g., Oxburgh, 1964, p. 6) also suggest
amphibole as a source of potassium in the mantle on the basis of the
observation of O'Hara (1961, p. 251, sample X282) that almost all the
potassium in a metamorphosed peridotite gneiss was contained in an
amphibole!

*The temperature of the eutectic 8 is 1640 ± IO"C, and the temperature of the beginning
of coprecipitation on the diopside-forsterite join of diopside solid solution and forsterite
solid solution is 1635 ± IO"C at 20 kbar. Although these two temperatures are within the
limits of error of each other, a maximum temperature must occur on the boundary curve
between the two points. The maximum temperature marks the intersection of the
boundary curve with a line connecting forsterite solid solution with the composition of
the diopside solid solution having the highest thermal stability in equilibrium with olivine.
That diopside solid solution will not necessarily lie on the diopside-enstatite join.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

38 GENERATION OF BASALTIC MAGMA

Although there is some debate about the primary or secondary

nature of phlogopite, the latter would imply a still deeper source for
potassium. There is greater concern about the actual amount of
phlogopite present because of the heat-producing aspects of the potas-
sium and the water content of the mica. If phlogopite is to make the
major contribution of potassium to the partial melt, then there must be
a sufficient amount of phlogopite to persist as a phase throughout the
melting interval and successive melt withdrawals. Unfortunately, an
amount of phlogopite of the order of I percent is considered to be
excessive with regard to radiogenic heat production, based on the K 2 0
content of chondrites (Hurley, 1968a,b) as a limit.*
In Figure 2-15 are displayed the deduced liquidus relations among
phlogopite, forsterite, and enstatite. The liquid at the beginning of
melting of this assemblage, A, is greatly enriched in potassium, and
phlogopite would be the first phase to be consumed. t On the other
hand, the low K 2 0 content of oceanic basalt may reflect the absence or
negligible amounts of phlogopite in the parental material (i.e., a de-
pleted peridotite) and suggests a source of potassium from within one
of the major phases (e.g., clinopyroxene).

SEISMIC VELOCITIES OF GARNET PERIDOTITE

SUITABLE?
If garnet peridotite is the source rock for basaltic magmas, then the
physical properties of the rock should be consistent with those
properties deduced for the mantle of the earth. Adams and Williamson
(1925, p. 245, footnote 9) related the seismic velocities of the primary
and secondary waves to the compressibility and density of the rocks:

where Ks = the adiabatic bulk modulus, that is, the inverse of the
compressibility, 11{3; p = density; V v = velocity of the primary wave;
and v. = velocity of the secondary wave. It appears that the seismic
velocities measured in the field could give a restrictive set of solutions
to the kinds of phases and their proportions in the rocks if the

*The minimum K20 content of the mantle was 0.017 percent if the mantle supplied the
40Ar of the atmosphere (Birch, 1951).
t According to the determined phase relations in the Fo-Qz-AnsoAboo-H20 system with
10 percent KA1Si 30 8 at 15 kbar(Kushiro, 1974, p. 245, Figure 22), amphibole, a possible
alternative source of potassium, would also be consumed first at a reaction point
involving forsterite, enstatite, amphibole, liquid, and vapor.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 39
Forsterite

Weight percent

FIGURE 2-15 Projection of the system forsterite-kalsilite-quartz at 20 kbar under

excess H20 . Invariant points are exaggerated in forsterite content: They probably lie
within a few percent of the baseline. Deduced from the relations presented by Luth ( 1967,
p. 3CJ7, Figure6f)for 3 kbar. (Leucite is not stable at 20kbar.)The assemblage forsterite +
enstatite + phlogopite begins to melt at the temperature of A, about IIJO"C (Modreski and
Boettcher, 1972, p. 859, Figure 2), and the liquid has the composition of A. The bulk
composition X (near the forsterite-enstatite join) is analogous to that of a portion of mica
peridotite. (With permission of the American Journal of Science.)

compressibility and density of the rocks and minerals were measured in

the laboratory.
To illustrate the application of the Adams and Williamson relation-
ship, selected values of Vv and p of the principal minerals of garnet
peridotite are compared with the range of the same parameters be-
lieved to exist in the mantle (Figure 2-16). These carefully selected
values appear to be reasonable until the ranges of these properties in
eclogites with different Fe/Mg and the extent of alteration are
examined. Birch (1970) clearly illustrated how sensitive the bimineralic

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

40 GENERATION OF BASALTIC MAGMA

9.0r------------------------------,

8.5 Fo

u
•
Ill

'E
~

~
>
).__-fl------""' Gr
8 .0 ;tC
Px
x/
~I
I
I

~X
I
I
I

3.5 4 .0
Density, o/cc

FIGURE 2-16 Selected values of primary wave velocity . Vp . and

density for the major minerals in garnet peridotite (open circles).
range of those values in the upper mantle (rectangle). and measured
values for various eclogites ( x) given by Birch ( 1970). Iron content of
eclogites increases with density. Both orthopyroxene and
clinopyroxene are represented by Px. Fo = forsterite; Gr =garnet.

rock eclogite is to Fe/Mg (Figure 2-16); the denser rock is the more
iron-rich. Eclogite. which should lie on the join clinopyroxene-garnet.
appropriate for the mantle, would have to be iron-rich. Other published
values for pyroxenite and harzburgite are too variable for a reasonable
analysis. One investigator, Anderson (1970, p. 89), on the basis of his
selected values , estimated that the mantle should consist of 45-75
percent olivine, 25-50 percent pyroxene. and less than 15 percent
garnet. Garnet peridotite and some eclogites are possible rock types in
the upper mantle; however, either the seismic data are too ambiguous
for a definitive conclusion or undepleted and unaltered rocks have not
yet been studied.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 41

MONTE CARLO TEST OF PHYSICAL PARAMETERS

On the basis of parameters including the two elastic velocities, density.
periods of the earth's free oscillations, dispersion of surface waves,
and mass and moment of inertia of the earth, Press (1968, 1970) made a
random test by a Monte Carlo method using successive constraints to
produce 18 possible models of the suboceanic upper mantle. He found
(Figure 2-17) that an eclogite facies model (-50 percent eclogite) was
generally acceptable between 80 and 150 km with a transition zone to
300 km. below which a peridotite model was favored. Press believed

Density, c;a/cc
3.0 3.2 3.4 3.6 3.8 4.0
0

50

100 • /Upper
Bound
•
~
.... 150 •
~ 200
,g
~
i250
Q.

0•
300

350

400

FIGURE 2-17 Various successful density models for the sub-

oceanic upper mantle (from Press. 1969. p. 174. Figure 1). Bounds
define range assigned in the Monte Carlo selection . Dots are density
values for the eclogite model. and squares are those for the pyrolite
model of Clark and Ringwood (1964). (With permission . Copyrighted
by the American Association for the Advancement of Science.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

42 GENERATION OF BASALTIC MAGMA

that ''eclogite'' is the high-pressure crystalline product derived by the

partial melting (15 percent) of peridotite that has not been wholly
depleted of eclogitic (basaltic) components. His model is gravitation-
ally unstable, and he emphasized that the considerable latitude in
properties may not necessarily be due to heterogeneity in the mantle.
For example, density contrasts may be stabilized by thermal gradients.

COMPOSITION(S) OF THE MANTLE

NOT DETERMINED
What, then, is the composition of the mantle? Because of the lack of
precise information on the kind, abundance, and composition of the
minerals, the bulk composition of the mantle from place to place
cannot be specified. The variation of Pb and Sr isotopes in basaltic
magmas is at present interpreted as resulting from the melting of
masses of different composition or age (Gast et at .. 1964; Hofmann and
Hart, 1975; Brooks et at .. 1976). According to Erlank and Kable (1976,
p. 287), element abundance ratios such as K/Rb, K/Ba, and Zr/Nb also
lead to the conclusion that the chemical differences between some
basalts cannot be attributed to fractionation and that different source
regions are involved.
The existence of large masses in the ·mantle unique in age or
composition implies that the mantle endured several remeltings,
perhaps early in the earth's history when the heat flow was several
times greater than at present (see Figure 5-8). The rise of diapiric
masses, for example, may have then been commonplace, and the
ascent of some of these masses may have terminated at moderate
depths when they encountered the base of a depleted, less dense layer.
These arrested and cooled masses, now nearer the surface, may be the
source of magma released on a later cycle of reheating (see Chapter
10). A discussion of the heterogeneity of the mantle is deferred until the
principles of the melting process are examined (see Chapter 6).

GARNET PERIDOTITE ACCEPTED AS

SOURCE-TENTATIVELY
On the basis of the above arguments, it is concluded that garnet
peridotite is the preferred source of basaltic liquids because:

1. The occurrences of garnet peridotite are appropriate to deep-

seated environments.
2. The close compositional relationship to meteorites supports its
derivation from material accumulated in the primordial earth.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Parental Materials 43
3. Partial melts of natural samples at high pressure have basaltic
compositions.
4. The mineral assemblage was found experimentally to be stable at
high pressures and temperatures.
5. It has appropriate densities and seismic velocities.

For these reasons most of the following arguments on magma genera-

tion are based on the minerals in this rock type . The reader should keep
in mind that this is a working conclusion, not well established, and it
will be tested in succeeding sections.
To give the reader a physical image of the kind of material envisaged
in the mantle, a hand specimen of garnet peridotite is displayed in Plate
I (p. 44a). The appearance of the rock has given rise to the "plum-
pudding" concept of the mantle, where garnet and clinopyroxene , the
potential basalt "plums," sit in a "pudding" of olivine and orthopy-
roxene . It is likely that under the pressure and temperature conditions
of the mantle the rock where stable would not exhibit the beautiful
colors displayed in the plate but would be nearly white hot .

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Depth of
3 Melting

As regards the source of the magmas. it is clear that a funda-

mental inquiry would take us back into a region of specula-
tion from which we had already turned away as unprofitable.
It does not appear, however, that ignorance concerning the
condition of the interior of the Earth as a whole need materially
embarrass the pursuit of the object immediately before us.
Harker (1909, p. 33)

Much of the debate about where melting takes place in the mantle rests
heavily on notions of the origin of the earth. Because the composition
of the earth and the distribution of rock types with depth depend on the
nature of the earth's formation, it is critical to know (a) whether the
earth accumulated relatively cool and was heated up by the decay of
radioactive elements and by conversion of gravitational energy to
thermal energy or (b) whether it was initially molten and differentiated
with cooling.

ORIGIN OF PROTOEARTH: COOL ACCUMULATION,

DIFFERENTIATED MELT, OR CONTINUUM OF BOTH?
These major concepts can perhaps be best described with a schematic
two-component system. Although the triple point involving liquid for
any simple igneous rock forming silicate has not yet been determined in
the laboratory, the relations are probably close to those displayed in
44

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Depth of Melting 45
Figure 3-lA for forsterite-H 2 • * If the effective pressure during the
condensation of the earth was about w- 2 bar, the earth may have
passed through a molten stage, depending on the bulk composition. For
example, composition X in Figure 3-IB would, on cooling, pass
through a liquid region. On the other hand, a composition such as Y,
highly dilute with respect to potential silicates, would not encounter
liquid on cooling. If the effective pressure on aggregation was approx-
imately t0- 6 bar (Figure 3-IC), the gas would have condensed directly
into crystalline material without encountering a liquid phase for any
composition. The eutectic, not shown in Figure 3-IC, would probably
lie at a temperature of the order of - 260°C. A most useful set of
practical experiments to be performed would no doubt include the
system forsterite-hydrogen or diopside-hydrogen at appropriate pres-
sures as an analogue of the condensation of a star. Identification of the
species in the gas phase would be very helpful to those using ther-
modynamic data to calculate the phase relations.
The two alternatives displayed in Figure 3-IA are not necessarily the
extreme pressure limits, and, in fact, a range of pressures and composi-
tions probably was effective during the cooling and aggregation of the
protoearth. The relationship of the rate of aggregation to the cooling
rate may have had considerable influence on the texture of the accumu-
lated material. For example, if the gas cloud cooled first at w- 2 bar
(Figure 3-IB, composition X) to liquid droplets, which then crystallized
before aggregation, the texture might be similar to that of achondrite.
The chondrules could have formed by the rapid cooling or quenching of
the droplets. On the other hand, if the gas cloud cooled directly to
crystalline aggregates (Figure 3-IB, composition Y; or Figure 3-IC,
composition Z), the texture might be akin to that of an achondrite.
Another factor of considerable importance is the time required to
collect or sweep up the condensing materials. If the sweep-up time was
short, the individual masses would be small and the heating on impact
would be preserved. On the other hand, if many orbits were necessary
for aggregation, large masses were assembled, and the accumulation
was slow, the heating on impact might be dissipated. In the first

*For another analysis of the problem wherein solar abundances are taken into considera-
tion and point A is believed to lie at several hundred bars, the reader should consult
Wood (1963, p. 156, Figure 3). Because the planets in the solar system are probably of
different compositions, the present solar abundances do not appear to be appropriate for
the initial composition of the specific gaseous region from which the earth was formed.
Thermodynamic calculations of the ''triple" point depend on knowledge of the gaseous
species; however, the species in gases of suitable compositions existing prior to the
formation of a protoearth have not been ascertained experimentally.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

46 GENERATION OF BASAL TIC MAGMA

.,...
0
,g

•.,.,...
:t 1o-2
f
a..
10-4 Crystal

Gas
I0-6

1890°
Temperature, •c-

instance, a relatively homogeneous protoearth would be formed,

whereas in the second instance, the protoearth formed would be
heterogeneous to a degree dependent on the size of the aggregated
masses prior to impact.

CRUST FORMATION THROUGH VOLCANISM

It will be assumed for the present purposes that the pressure through-
out the condensing toroidal cloud was indeed low and that the earth
mainly accumulated from a relatively cool gas directly into crystals
without an intervening molten stage except at the center of the toroid.
Birch (1965) thought that conversion of the gravitational potential
energy on core formation would raise the temperature of the mass
I600°C on the average; however, the efficiency of the conversion might
have been low and much of the heat might have dissipated so that only

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Depth of Melting 47

8 c P•lo· 6 bar

Gas

Gas
t z
Liquid+ Gas

Fosterite + Gas Fosterite + Gas

To eutectic :
Fo Composition Fo Composition Hz

FIGURE 3-1 A. Schematic pressure-temperature diagram for the high-temperature

portion of the forsterite-H. system. assuming congruent relations on all univariant
curves. A = the invariant point for forsterite; C.P. = critical point for forsterite; C.E.P.
= first critical end point of the binary system. Lowercase letters refer to points occurring
in the constant pressure sections (marked with light arrows) displayed in Figures 3-18
and 3-IC. Heavy arrows indicate that the curves continue. Heavy dashed curves about A
are metastable extensions of equilibrium curves (heavy lines) for phases of forsterite
composition .
B . Schematic temperature-composition diagram for the high-temperature portion of
the forsterite-H 2 system at a total pressure equal to 10-2 bar. Lowercase lettered points
are identified in Figure 3-IA. X and Y, respectively. are compositions that illustrate
possible formation of chondri tic and achondritic textures on cooling.
C. Schematic temperature-composition diagram for the high-temperature portion of
the forsterite-H 2 system at a total pressure equal to 10-• bar. Lowercase lettered point is
identified in Figure 3-IA . Z is a composition illustrating the cooling behavior of an
earth-forming gas cloud at very low pressure and relatively low temperature.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

48 GENERATION OF BASALTIC MAGMA

the core and a portion of the lower mantle remained or became molten.
The ambient temperature on initiation of condensation of the principal
phases in plagioclase peridotite would be well below 1250°C. That
temperature is the eutectic for anorthite + forsterite + diopside
+ enstatite at I atm. according to Schairer and Yoder (1970), p. 212,
Figure II), and the vaporus eutectic at I0- 6 bar, for example, would
be at much lower temperatures. The crust, formed at the earliest stage
by volcanic processes, was probably andesitic in composition because
of the degassing of the mantle, a process that is believed to have led
to the formation of the oceans. Andesite is thought to have been
derived by the hydrous partial melting of the mantle (see invariant
point A. Figure 2-14), according to the views expressed by Yoder
(1969). Water presumably was trapped in hydrous minerals, as a free
phase, or was the oxidation product of trapped primordial hydrogen,
the atmosphere in which the earth aggregated.

IMPACTING, DEEP SURFICIAL TURNOVER, AND

PARTIAL MELTING
Crust formation implies that there is a region at some depth below the
surface depleted of crustal materials. On the basis of present knowl-
edge of lunar history it is likely that the earth endured surticial melting
to a depth of the order of 50 km as a result of the great impacting that is
believed to have culminated at about 4 b.y. (Teraet a/., 1974). Whereas
the average depth of melting may have been about 50 km, a few areas
may have suffered impact melting to a depth as great as 200 km,
comparable to the radius of the largest of impacting bodies. The
question arises as to whether sufficient time has elapsed for all the
material melted on impact to have crystallized. In other words. are
there pockets of m:)Jten material residual from the impacting, which
must have been more severe than that observed on the lunar surface?
The turnover of crust and upper mantle materials was presumably
very thorough as a result of impacting. It should be recalled that the
oldest rocks already contain quartz-bearing sediments and are enclosed
by granitic* gneisses, about 3.75 b.y. old (Moorbath eta/., 1972). The
quartz was most likely derived from the churned andesitic (average 15
percent modal quartz) crust. It is indeed remarkable that there is no
evidence yet found on earth of impactite equivalent to the lunar
regolith. If plate tectonic processes were operating during, as well as

*Granitic melts are not a possible partial melting product of peridotite. but are presumed
to be (a) a derivative of basaltic or andesitic melts separated from their parental material
or (b) the partial melting product of sedimentary materials. The origin of rhyolites and
granites is in itself a story worthy of a separate volume.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Depth of Melting 49
after, Precambrian times, the crust might have been recycled several
times. Erosion must have been very efficient and crustal subduction
complete to remove all traces.

MAGMA-A RESIDUUM FROM EARTH FORMATION?

There are obviously too many options available to provide a unique
sequence of events describing the origin of the earth from the few facts
known. The reader should keep in mind the possible existence of
residual pockets of magma, at least in early Precambrian time, resulting
from (a) an initially molten earth, (b) an earth reheated by gravitational
energy conversion, (c) a deeply impacted surface, and (d) short-lived
radioactive isotope decay. The dimensions, of undefined aspect ratio,*
of these pockets, if existent today, are presumed to be less than the
wavelength of seismic body waves now received at the surface, that is,
several tens of kilometers (Shimozuru, 1963, p. 192). The cooling rate,
dependent on the temperature of the ambient rocks, of such small
pockets was probably sufficient that those residual magmas became
crystalline by the end of the Archean. Of course, there is no way of
knowing what the original dimensions of trapped magmas were.

SEISMIC EVIDENCE-AUXILIARY MAGMA CHAMBERS

One must depend mainly on seismic evidence for indications of the

depth of generation of contemporary melts. Magnetic and electrical
resistivity measurements are not sufficiently specific to delineate a
magma body, but they do provide useful confirmatory evidence. Broad
areas of relatively high temperature have been deduced by Porath et al.
(1970, pp. 258-259) by correlation with high electrical conductivity in
the mantle and by Morgan (1972, p. 206, Figure 4) by correlation with
high isostatic gravity anomalies. Intuitively, one would expect the
thermal expansion resulting from near solidus temperatures to produce
low gravity; however, the upward flow of denser, as well as hotter,
mantle material into the top of the [Link] crust may produce a net
increase in gravity.
In general, the crust and mantle transmit both primary and secon-
dary seismic waves at all depths, and therefore those regions are con-
sidered to be predominantly solid. This view is supported by data on
tidal deformation and the fundamental oscillations of the earth. Changes
in the velocity and attenuation of seismic waves in certain regions,
*The aspect ratio is the ratio of minor to major axes of an ellipsoidal inclusion. The
aspect ratio of liquid inclusions greatly influences the rigidity and attenuation in the host
rock (Walsh. 1969).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

50 GENERATION OF BASAL TIC MAGMA

however, suggest that some areas may not be completely solid. For
example, Kubota and Berg ( 1967), after the studies of Gorshkov ( 1958),
succeeded in outlining areas in the Katmai region, Alaska, where the
shear wave was attenuated, presumably by liquid. They outlined
shallow magma reservoirs using four seismic stations. By observing a
large number of earthquakes, Kubota and Berg noted ray paths along
which seismic stations did not receive the shear wave from specific
earthquake centers. The "screening effect" led to the deduction of the
boundaries of the magma chamber. The vertical distribution of the
chambers deduced is shown in Figure 3-2. These chambers are con-
sidered by the writer to be auxiliary (transported magmas) and not
primary (site of generation).
Gorshkov (1958) concluded that the chamber under Kliuchevsky
volcano, Kamchatka, was between 50 and 75 km deep, and he believed
that its shape was that of a convex lens or perhaps a triaxial ellipsoid.
More recent studies by Fedotov and Tokarev (1974) indicated "that the
roof of the zone of intense fractional melting and magma generation'' is
at a depth of 50-60 km; however, the "main zone of magma generation
is probably located within the depth interval from 120 to 200-250
km .... "Eaton eta/. (1975) outlined a partially molten region beneath
Yellowstone National Park, Wyoming, extending from near the surface
to about 100 km, on the basis of a 10 percent reduction in P-wave
velocity. The diameter of the region showing seismic wave attenuation
is about 30 km near the surface, expanding to about 60 km at 60 km

0 sw NE
f
-
•• 10
~ 20
o<?I
I
I

0
~30
-:4o
.1:

0
50~--------------------------------~

FIGURE 3·2 Vertical distribution of magma chambers (enclosed areas) and possible
conduits (dashed lines) along the Katmai, Alaska, volcanic range according to Kubota
and Berg (1967, p. 202, Figure 16). The two deepest chambers outlined may consist of
many smaller molten pockets, according to an alternative suggestion of Shimozuru
(1963). Volcanoes (triangles) and seismic stations (arrows) are noted. The fourth seismic
station is at Kodiak, about 150 km to the southeast of the range. Vertical and horizontal
scales are about equal. (With permission of the Bulletin Volcanologique . )

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Depth of Melting 51
depth. On the basis of teleseismic P-wave arrival time residuals,
Steeples and Iyer (1976) concluded that anomalously hot rock, or
possibly partially molten rock, lies at a depth of more than 7 km and
less than 40 km beneath the Long Valley, California, caldera. The
velocity in contrast with the surrounding material is lower by more
than 5 percent and is probably in the range 10-15 percent. Although the
areal extent is poorly constrained, a large amount of detailed geological
and geophysical evidence supports their hypothesis of a residual auxil-
iary magma chamber. Sanford et al. (1973) interpret the large ampli-
tudes of reflected shear waves (both [Link] and SJS) from a surface at a
depth of about 18 km under Socorro, New Mexico, as due to the
presence of· 'hot material of low rigidity.'' The region is known for its
high heat flow, hot springs, and recent basaltic eruptions.

LOW-VELOCITY ZONE-EVIDENCE OF
PARTIAL MELTING?
If the Alaskan and Kamchatka chambers are auxiliary and not the site
of generation, then it is necessary to look at broader-scale features,
unfortunately with great loss of resolution. The increase in data from
the study of surface waves in the past 10 years, however, has greatly
improved the understanding of the deeper structures. The change of
seismic velocities with depth according to Nur ( 1974, p. 302) is given in
Figure 3-3. The region of diminished velocities extends from about 70
to 150 km. In the low-velocity zone, discovered by Gutenberg (1926),
the amplitude, a parameter difficult to measure with precision, de-
creases strongly (Figure 3-4). The profiles given by Nur are not
necessarily worldwide, and some authors believe the low-velocity zone
is absent over large areas of the oceans and continental shields
(Knopoff, 1972; Jordan, 1975). Some data on vertical shear-wave and
Love-wave dispersion indicate a fundamental difference between the
mantle under oceans and continents: that debate is still in progress.
The velocity reversal, where observed, may be due to (a) a high-
temperature gradient, (b) a solid-phase change, (c) a compositional
change, or (d) partial melting.
The constraints of heat flow and potential phase changes observed in
the laboratory support the first two possibilities. A change of olivine
composition from Fo00 to Fo70 could account for the velocity change,
according to Birch (1969, p. 34). The 5 percent drop in velocity could
imply the presence of about 7 percent liquid if its Vv = 2.5 km/sec.
Machado (1974, p. 258-259) noted that the observed rigidity of the
low-velocity zone can be accounted for if the value for crystalline

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

52 GENERATION OF BASALTIC MAGMA

Velocity, km/sec
4 5 6 7 8

f FIGURE 3-3 Approximate

~ 100 velocities of the primary,
•e V P• and secondary, v•. waves
0
in the upper mantle, accord-
:ii200
L ing to Nur (1974, p. 302,
Q. Figure 4.10) .
•
0300

peridotite is reduced by the presence of 5 percent fluid inclusions.

Kubota and Berg (1%7, p. 212) explained the increase in Poisson's
ratio for the region under Katmai by assuming that a minimum of 8
percent of the volume of the upper mantle is molten.

PARTIALLY MELTED ZONE-EXPERIMENTAL

EVIDENCE
The interpretation of the low-velocity zone as a region of partial
melting is supported by the experiments of Spetzler and Anderson
( 1%8) and Anderson and Spetzler ( 1970). They determined the velocity
of both longitudinal and shear waves in cylinders of ice containing
irregular pockets of NaCI brine. A large drop in velocity was observed
with very small amounts of melt present. They also calculated the
influence of the aspect ratio of the droplets of melt on the velocity
(Figure 3-5) according to the Eshelby-Walsh theory (Eshelby, 1957;
Walsh, 1%9).
By using vitreous silica rods to transmit the compressional waves
through melt to a transducer, Murase and Suzuki (1966) and later
Murase and McBimey (1973, p. 3572) obtained a value of2-3 km/sec in
the temperature range of l100-1500°C for molten basalts. In both
studies an abrupt drop in velocity was observed as melting began.* It is
evident from Figure 3-5 that thin, intergranular films of a very small
amount of melt could account for the low-velocity zone. Birch (1969),
using olivine and basaltic glass inclusions as a model, found that
*There are many opportunities for experimental work in seismology. The present need
appears to be for a high-temperature transducer to measure velocities and attenuation in
silicate melts at appropriate mantle conditions. Progress has already been made with
some ceramic electronic components capable of withstanding soo•c (J. R. Rowley.
personal communication. 1975).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Depth of Melting 53
2.5
•
p\ s
2 \
,.,
:::J
!::
-a
1.5
\
.,
\
~.
-
E
~ \
0 \
E \ •
-
~

·.:
0
0' .5 .,. \
\
•
0
.J
\
0

•
-.5 '------1---....&..--~---'------'
0 500 1,000 1,500 2,000 2,500
Distance, kilometers
FIGURE 3-4 Worldwide average seismic amplitudes for primary,
P, and secondary, S, waves at various recording distances, according
to Gutenberg (1959). Amplitude variation supported by primary
waves from nuclear explosions (solid dots) originating in New
Mexico and Nevada (Anderson, 1962, p. 55). Drop in amplitude
from 100 to 1,000 km is attributed to the low-velocity zone. (With
permission of Scientific American, Inc.)

FIGURE 3-5 Variation of velocity

of primary (compressional) wave, VP,
and secondary (shear) wave, V,, at
1200"C with melt concentration as-
suming different aspect ratios, a, as
defined in the inset figure, of drop-
lets, according to Anderson and
Spetzler (1970, p. 63, Figure 2). The
major and minor axes of the eUipsoi-
dal droplets are L and d, respectively.
Ruled region is the low-velocity zone.
(With permission of the Elsevier Sci-
0.1 I entific Publishing Company.)
Melt concentration, percent

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

...
..... . . .
FIGURE 3-6 The Q structure beneath Japan (A) and South America (B). Dots indicate regions of high Q (I ,000-3,000); diagonal lines,
intermediate Q values (300-500); and cross-hatching, low Q values (50-100). After Sacks and Okada (1974, p. 215, Figure 5). (With permission
of the Elsevier Scientific Publishing Company.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Depth of Melting 55
spherical droplets and 6 percent melt could explain the low-velocity
zone.
Because V. 2 = JL/p, where JL is the shear modulus (or rigidity), a
density increase could also account for the drop in shear velocity.
Press (1969) has already shown (Figure 2-17) that density increases in
the same zone; the density change would not, however, account for the
threefold increase in attenuation (see below) according to Nur (1974,
p. 303).

ATTENUATION OF SEISMIC WAVES

The attenuation, damping, or dissipative function of seismic waves is
usually expressed in terms of a dimensionless quantitity, Q, defined as
21T times the ratio of the stored energy to the dissipated energy per
cycle. It is related in an unknown way to the viscosity. If the viscosity
is low, it is likely that the region is hot and possibly partially melted.
Sacks and Okada (1974) determined Q for sections across Japan and
the Chile-Peru region of South America (Figure 3-6). If Q values of 50
to 100 imply melting, then the source of magma under Japan is at
depths of 300~00 km , whereas the source under Chile-Peru is at about
400 km. The inferred melt region under Japan is above a subducting
plate characterized by very high Q values (I ,000-3 ,000), whereas under
Chile-Peru the subducting plate abruptly terminates at the inferred melt
region. Sacks and Okada noted the absence of low Q values in the
vicinity of known active volcanoes and suggested that "there must be a
narrow pipe transporting the magma from depths." The aseismicity in
the 350-500-km regit?n beneath Chile-Peru supports the view that the
temperature is so high that the rocks cannot store strain energy and are
unable to fracture elastically. Earthquakes are known to occur at
depths as great as 700 km .

SUMMARY
The seismic evidence indicates that magma may be generated in some
broad regions extending generally from 50 to 170 km in depth. In other
regions restricted to specific structural environments, the source of
magma may be as deep as 300~00 km. The seismic data are insuffi-
cient for independent determination of the shape of the chamber, the
distribut_ion of the liquid, and the extent of melting. It is difficult to
distinguish whether the deduced magma chamber is the site of genera-
tion or merely the region in which liquid is separated from a trans-
ported, partially molten mush . Much of the debate relating magma

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

56 GENERATION OF BASALTIC MAGMA

generation to tectonics will probably falter on this point in the absence

of distinguishing criteria. Many geophysicists prefer a model of a
magma chamber that is an oblate spheroid, elongated horizontally, to
represent the partially molten chamber containing up to 8 percent
liquid in droplets ranging in shape from thin films to spheres. (The form
of the dispersed liquid will be discussed further in Chapter 9.) The field
geologist, whose opinions are based mainly on investigations of cooled
and eroded chambers of transported magma, prefers models of magma
chambers that are more irregular in shape with the base undefined, a
lenticular body, or an ex tended layer.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting
4 Processes

MULTITUDE OF MECHANISMS
Now that the reader has a rough idea of what is melting and where it
melts, an effort will be made to present the various ideas expressed on
how the. melt is generated . An exceptionally large number of mecha-
nisms have been proposed, each having some advantages-and usually
several disadvantages. Some mechanisms are specific to certain struc-
tural environments, whereas others are of general applicability. The
reader will find it useful to consider each mechanism in relation to the
various regimes of plate tectonics. The mechanisms involve changes of
pressure, temperature, composition, position, and stress, as well as
contrasts in various other properties. A change in practically every
known parameter of consequence has been nominated as a cause of
melting. Not all the mechanisms will be described; those chosen,
however, will provide the framework for substituting other parameters.

MELTING BEHAVIOR OF CRITICAL MINERALS

First, it is necessary to look at the melting behavior of the four critical
end-member minerals in garnet peridotite because several of the
processes depend on the character of its melting curve. Historically, it
was thought that the melting curve of a substance would become
asymptotic to a fixed value with increasing pressure (Figure 4-IA),
reach a maximum (Figure 4-lB), terminate in a critical point (Figure
57

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

58 GENERATION OF BASALTIC MAGMA

A B

p p

T T

c Critical
D
Point

p p

T T

FIGURE 4-1 Schematic representation of various concepts of sili-

cate melting with increasing pressure: A, Asymptotic approach to
limiting melting temperature. B, Melting temperature attains a
maximum. C, Melting curve terminates at a critical point. D, Con-
tinuous rise of melting temperature within observational range.

4-IC), or continue indefinitely (Figure 4-ID). These views were super-

seded as data at higher pressures became available.
Forsterite, Mg2Si04 (Figure 4-2), melts with a gradient of 4. 77°C/kbar
up to 50 kbar, according to Davis and England (1964, p. 1115, Figure 2).
Akimoto (1972) found that the melting curve is terminated by a phase
change to a {3 phase, an orthorhombic structure termed a "modified"
spinel by Morimoto et al. (1970; see also Moore and Smith, 1970), at
pressures near 125 kbar at about 1000°C and near 115 kbar at 800°C.
Diopside, CaMgSi20 6 , melts (Figure 4-2) with an initial gradient of
13°C/kbar up to 5 kbar (Yoder, 1952) that decreases to 7 SC/kbar at 50
kbar (Boyd and England, 1963, p. 321, Figure 6). At I atm diopside
melts incongruently to a clinopyroxene solid solution relatively rich in
MgSi03 , and the liquid becomes enriched in CaSi03 • The melting curve

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 59
in Figure 4-2 marks the disappearance of the last crystal. The pressures
to which the incongruent melting character of diopside persists are not
known. Arguing from the behavioi of the germanate analogue,
Ringwood (1966, p. 381, Table 5) suggested that the melting curve is
probably terminated by the breakdown of diopside to garnet and
ilmenite structures.
Enstatite, MgSi03 , melts (Figure 4-2) with an initial gradient of
12.8°C/kbar that decreases to 6°C/kbar at about 50 kbar (Boyd eta/.,
1964, p. 2104, Figure 1). The very-low-pressure region is complicated
by incongruent melting to forsterite + liquid and polymorphism to
protoenstatite, but the nature and position of the inversion curve are
being debated (see Kushiro eta/., 1968a; Chen and Presnall, 1975). The
melting curve is probably terminated by a breakdown to ,8-M~Si04 and
stishovite (Ringwood, 1970, p. 128) at pressures in excess of200 kbar.
Pyrope, MgaAI 2 Sia0 12 , melts (Figure 4-2) with a gradient of 14.1°C/
kbar above 21.6 kbar and 1510°C (Boyd and England, 1959a, p. 84,
Figure 1). At lower pressures pyrope is represented by a succession of

Temperature, "C

FIGURE 4-2 Melting behavior of the four major end-member minerals in garnet
peridotite. The high-pressure breakdown products of diopside represent structure types
and not mineral compositions. Solid lines indicate portions of curves determined experi-
mentally. Dashed lines are extrapolations based on preliminary data, iron-rich members,
or analogues as noted in text.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

60 GENERATION OF BASAL TIC MAGMA

assemblages, which include forsterite + cordierite + spinel at 1 atm.

Pyrope is the only garnet end member not stable at 1 atm, and it has the
largest change in melting temperature with pressure. It is stable to very
high pressures (<300 kbar), where it breaks down to a mixture of
MgSi03 (perovskite structure) plus corundum (Liu, 1974).
The curves outlined in Figure 4-2 illustrate the general concept that
the melting temperatures of principal silicates rise with increasing
pressure and are terminated with a phase change or reaction. Evalua-
tions of solid solution effects, especially those involving Fe, suggest
that these four critical phases in garnet peridotite will probably be
stable to the greatest depths now being considered as possible sites
for magma generation (i.e., -400 km). It is, however, more important
to know the stability of the assemblage of minerals, and those limits
are discussed in Chapter 6.

THE MECHANISMS
Some of the proposed mechanisms for bringing about the melting of
crystalline parental material are outlined as follows.

Stress relief
Tension (Yoder, 1952)
Compression (Uffen, 1959)
Thermal rise to cusp in melting curve
Compositional cusp (Buddington, 1943; Hess, 1960)
Phase-change cusp (Presnall et al., 1973)
Convective rise
Isentropic rise (Verhoogen, 1954)
Isenthalpic rise (Waldbaum, 1971)
Property perturbation
Thermal conductivity (Lubimova, 1958; McBimey, 1963)
Density (Grout, 1945; Ramberg, 1972)
Mechanical energy conversion to heat
Thrust faulting (Nutting, 1929; DeLury, 1944)
Subducting plate (Toksoz et al., 1971; McKenzie and Sclater,
1968)
Propagating crack (Griggs, 1954; Griggs and Handin, 1960)
Regenerative feedback (Shaw, 1969; Anderson and Perkins, 1975)
Tidal dissipation (Shaw, 1970)
Compositional change
Diffusion under pressure differential (Bell and Mao, 1972)
Addition of volatiles (Yoder and Tilley, 1962; Bailey, 1970)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 61
Radioactive heat production
Internal (Joly, 1909)
External (Holmes, 1915)
Residual liquids from protoearth
Primordial condensed gas (Rittmann, 1962)
Residual liquids squeezed out on crystallization (Chamberlin and
Salisbury. 1905)
Exothermic chemical reactions
Surficial (Jaggar, 1917)
In chamber (Daly, 1911; Day and Shepherd, 1913)

No measure of importance should be attached to the order of their

presentation.

STRESS RELIEF

TENSION

The first melting curve of any common rock-forming silicate measured

under high pressures in the laboratory gave rise to a mechanism
involving relief of stress . Yoder (1952) studied the melting behavior of
diopside to the then very high pressure of 5 kbar at temperatures up to
1500°C in an argon-gas, high-pressure apparatus. He believed that the
pressure on the source region, then considered to be at or near the base
of the crust, could be relieved, for example, by gentle arching or
faulting of the overlying rock . On release of pressure, the source has a
temperature in excess of that required for melting at the reduced
pressure , and melting takes place. His case for a 5-kbar release is
reproduced in Figure 4-3 . The heat for melting was supplied from the
source region itself (that is, the process is adiabatic), and about 6
percent melt is produced in the model described. The existence of a
5-kbar stress in the mantle is not now considered tenable, as discussed
in Chapter 9.

COMPRESSION

The stress-release concept was embellished by Uffen (1959) and sup-

ported by Uffen and Jessop (1963), who considered the effects of
compression rather than tension. Uffen's model is reproduced in
Figure 4-4. Cyclical compressions and failures are required to take
place in the presence of continuously rising temperature. Eventually.
one failure is sufficient to reduce the pressure to that below the melting

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

62 GENERATION OF BASALTIC MAGMA

10

PERIDOTITE

FIGURE 4-3 Magma generation model of Yoder (1952, p. 372, Figure 5), based on
stress relief under tension. Dashed curve with arrows shows effective pressure release
and subsequent temperature change as magma generated moves to the surface. (With
permission of the University of Chicago Press.)

curve, and melting takes place. Field support for this version of the
stress-release concept may be derived from the work of Fahrig and
Wanless (1963). They presented a map of Canada (Figure 4-5) showing
swarms of dikes parallel to the compressive fold axes of the various
regions. Apparently, elastic rebound after a period of crustal compres-
sion and heating resulted in a low-pressure belt where basaltic magmas

Time---

FIGURE 4-4 Magma generation model of Uffen (1959, p. 118, Figure 2) based on
stress relief with a compressive failure in region of rising temperature. (With permission.
Copyrighted by the American Geophysical Union.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 63

MILES
300
I

I
I

GSC

FIGURE 4-5 Diabase dike swarms of the Canadian Shield postdating compressive
mountain-building stage in provinces. Map compiled by the Geological Survey of Canada
(GSC) . Fahrig and Wanless ( 1963. p. 936. Figure 2). Encircled numbers refer to localities
of dikes having measured ages in minions of years recorded in associated rectangles.
[With permission of MacmiUan (Journals) Ltd.]

intruded. Ages determined on dikes in the swarms noted in Figure 4-5

postdate the age of the provinces in which they occur.

THERMAL RISE TO CUSP IN MELTING CURVE

COMPOSITIONAL CUSP

Hess (1960) developed the idea of Buddington (1943. p. 137) that

melting could be achieved by depressing the crustal layers of the earth
into a warmer region. In Hess's model. illustrated in Figure 4-6.
subsequent rise of temperature in two basalt layers is sufficient to nick
the melting curve. and partial melting takes place at C and D . Hess
emphasized partial melting rather than complete melting and pointed

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

64 GENERATION OF BASALTIC MAGMA

Temperature, oc-

10
FIGURE 4-6 Compositional cusp and de-
pression model for magma generation of
Hess (1960, p. 180. Figure 36). Curves A
and B . respectively. refer to the solidus of
the rock types underlined before and after
Initial displacement. Re-establishment of thermal
ondfinol \
temperature\ equilibrium (temperatures on dashed curve
curve \
return to initial values) results in partial
Olivine Basalt\ melting at C and D . (With permission of the
\~...ljl----. Geological Society of America.)
\
\
\
\
~\
\
T8f111181:Gture---\
curve following \
displacement of \
crust from A-£1 \

out that the continental crust is probably more heterogeneous than a

simple layered structure . In his view. the requirements for enthalpy of
melting merely delay the rise of temperature. It is evident that in this
model a considerable amount of downwarping would be required to
initiate melting. As in the stress-release hypotheses, close proximity of
the geothermal gradient curve to the melting curve is necessary.

PHASE-CHANGE CUSP

Whereas Hess appealed to appropriate changes in bulk composition to

bring about interception of the geothermal gradient with the melting
curve, Presnall eta/. ( 1973) believed that the cusps arising in the solidus
of a single bulk composition , because of change in phase under
anhydrous conditions, are more likely conditions for melting to begin.
With increasing temperature. melting will occur first at the lowest
temperature on the solidus , which they indicated is at depths where a
change of phase occurs . The changes in slope of the solidus for the
anorthite + forsterite reactions at about 1335°C and 8.7 kbar as well as
1410°C and 17.2 kbar, shown in Figure 2-8, would be likely analogous
points for the temperature on a rising geotherm to meet that of the
solidus. Strong support for liquid generation at such invariant points is
given in Chapter 7.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 65

CONVECTIVE RISE

ISENTROPIC RISE

Instead of depressing the source material into a region of appropriate

temperatures for melting, Verhoogen (1954) suggested raising the
source material by solid-state convection (Figure 4-7) into a region
appropriate for melting. Because of the large vertical and horizontal
transport of heat by mass movements, Verhoogen (1973) did not accept
the concept of a geotherm based on a conductive heat model. He
argued that a temperature gradient less than a few degrees per kilome-
ter would induce convection. Material starting at S rises on an adiabat
(0.2° to OSC/km), and melting begins at M, the solidus of the material.
As melting continues, heat is absorbed in the process and the mass
cools until M 1 is reached, where liquid segregates, according to Ver-
hoogen. The liquid rises to the surface and has a temperature of M 2 •
Neither the degree of melting achieved nor the cause of segregation of
liquid is specified. The heat requirements, density contrasts, and
minimum depth of origin are discussed in Chapter 5.
A convecting system is a delicate balance between convective heat
transfer and conductive heat losses through the boundaries of the
mass. Shaw (1965, p. 145) considered such a system a "transient
condition." He pointed out that when heat transfer is greater than heat
losses, heating of the marginal material occurs, the temperature gra-
dient diminishes, and convection dies out. Alternatively, if the heat
transfer is less than the heat losses, the magma cools, viscosity
increases, crystallization begins, the temperature gradient diminishes,
and convection dies out. Quantitative analysis of the problem would be
a most difficult, but rewarding, task. The adiabatic rise of material is
commonly assumed in magma generation models; the assumption
should be considered as tenuous.

FIGURE 4-7 Magma generation model of

Verhoogen (1954; 1973, p. 516. Figure I)
based on the convective rise of matter from
S along an adiabat that intersects the sol-
idus. where melting begins at M . Liquid
s:
segregation takes place at M, and rises on
i
Q another adiabat to the surface, where the
liquid has the temperature M 2 • (With per-

' mission of the Geological Society of

America.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

66 GENERATION OF BASALTIC MAGMA

ISENTHALPIC RISE

Waldbaum (1971) suggested an isenthalpic rather than an isentropic

rise of a mass to achieve melting conditions. The irreversible adiabatic
decompression is given by the Joule-Thomson (1854) equation

dTI = V(Tap-l)
dP H Ct'

where V = volume, ap = isobaric thermal expansion, Cp = isobaric

heat capacity, and H = enthalpy. The isenthalpic adiabatic rise for a
garnet peridotite would be about -23°C/kbar, according to Wald-
baum's mineral values, compared with about 0.5°C/kbar for an
isentropic adiabatic rise. Ramberg (1971) and Shaw and Jackson (1973,
p. 8648-8649) pointed out that the equation does not include a term for
the effects of gravity,
Vpgdh,
dP
where h = height in gravitational field. The imposition of hydrostatic
equilibrium reduces the equation to the isentropic expression. On the
other hand, the Joule-Thomson equation is applicable to extrusion of
magma through a narrow constriction. [Attention was also called to
possible geophysical applications of the Joule-Thomson effect by
Adams as early as 1922 and by Bowen in 1928 (p. 183).]

PARAMETER PERTURBATION

THERMAL CONDUCTIVITY PERTURBATION

Several processes of magma generation depend on unusual perturba-

tions in one or more properties. Two examples involving either thermal
conductivity or density are presented.
Lubimova (1958) noted the general decrease in thermal conductivity
with temperature (Figure 4-8) and suggested that with further increase
in temperature radiative processes should cause an increase in thermal
conductivity. The thermal conductivity was believed to have a
minimum of 0.003 cal/cm sec deg at a depth of 50-100 km. Such low
thermal conductivity of the upper mantle reduces the rate at which heat
escapes from the interior of the earth. McBimey (1963, p. 6327)
realized that the temperature would rise in that region and the "heat
accumulation" might be sufficient to bring about partial melting. His

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 61

...:11
...
•..
Dunite and Orthapyraxenite
...eiO
'
...
0

": 9
2

...
..."'
c:
::: 7 Oz
0
e
: 6 Diabase and Gabbro
1-

Db a=-=====
_ob.
G1 Obz
5
Gz
Gs

Temperature, •c
FIGURE 4-8 Change of thermal conductivity of ultrabasic and
basic rocks with temperature. From Birch and Clark ( 1940, p. 549
and 550, Figures 4 and 5). D,. D 2 , and D,. are dunites; 0, and 0 2 are
orthopyroxenites; Db, Db 2 • and Db,. are diabases; and G ,. G 2 • and G,.
are gabbros. (With permission of the Amerkan Journal of Science. )

interpretation of Lubimova's statements was supported by the experi-

mental determinations of McQuarrie (1954) and Powell (1954). Confir-
mation of a minimum in the effective thermal conductivity between
1050°C and 1300°C (Figure 4-9) was obtained by Kawada (1966) for a
lherzolite and a dunite and by Murase and McBimey (1970, 1973) for
two basalts. (It would be desirable to determine experimentally
whether the minimum persists at high pressures.)
McBimey's melting process is probably initiated by high heat flow
resulting from a small increase in thermal conductivity produced by an
increase in stress (McBimey, 1963, p. 6325), as shown in Figure 4-10. The
stressed zone channels the heat flow into a lower temperature zone in
the range 1050-1300°C, where the conductivity is low. When the "heat
accumulation" is sufficient, the temperature of the ambient rocks,
already near the beginning of melting, is raised and melting begins.
There is a "runaway" character to McBimey's minimum thermal

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

68 GENERATION OF BASALTIC MAGMA

l
¥ FIGURE 4-9 Thermal conductivity
change in lherzolite and dunite (Kawada,
1966, p. 1084, Figures 5 and 6) and two
basalts (Murase and McBimey, 1973,
p. 3586, Figure 28) with temperature. Note

f the thermal conductivity minimum between

10500 and 1300'C in both ultrabasic and basic
rocks. (With permission of the Earthquake
J Research Institute, Tokyo University, and
the Geological Society of America.)

conductivity melting process (McBirney, 1963, p. 6325). As the lower-

temperature rocks are heated, the thermal conductivity decreases and
the temperature rises. With increase in temperature, the thermal con-
ductivity decreases further. The temperature rises until the minimum
thermal conductivity is reached. In this way, a thermal trap is gener-
ated for magma production.

DENSITY PERTURBATION

A contrast in density in a gravity field may also bring about a chain of

events leading to melting. Ramberg ( 1972) considered the behavior of a
single buoyant layer a few kilometers in thickness with a density
contrast, and suitable viscosity contrast, at a depth of I ,000 km. The
process is initiated by a perturbation in the form of a half wave in the
layer, extending it to shallower depths (Figure 4-11). The material at
the crest of the wave is accelerated upward because of the density
contrast with its new surroundings. The detachment characteristics of
the mass, a diapir, have been demonstrated by Grout (1945) and
Whitehead and Luther ( 1975) by tank experiments in which inverted
drops with long tails were observed. Many of Ramberg's (1972) cen-
trifuge experiments with greater body forces, as well as some of those
of Whitehead and Luther, exhibit a massive connection of the diapir to
the buoyant layer. According to Ramberg, a thermally isolated mass

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 69

Sfc~------------------------------------,

5or------------------------

..... 100
!e
!!
:.
IE
~
•
0
150

200

250~----------~----~----------------~

FIGURE 4-10 Interpretation of McBimey's (1963) model for

magma generation resulting from the thermal conductivity minimum
exhibited in Figure 4-9. A small increase in thermal conductivity.
resulting from abnormal stresses. channels the heat flow from depth .

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

70 GENERATION OF Bf\SALTIC MAGMA

Temperature, °C
0 400 800 1200 1600 2000 2400 2800
0

100

200

300

400

..
• 500
•e
!!
:i 600
..:
Q.
•
Q
700

800

900

1000

1100

1200

FIGURE 4-11 Magma generation model of Ramberg (1972, p. 58, Figure 10) based on
perturbation in buoyant stratum, diapiric rise of a solid mass, and pressure-release
melting. (With permission of the Elsevier Publishing Company.)

rises as a solid until the pressure is sufficiently reduced for melting to

begin at a depth of about 300 km. Ramberg considered the adiabatic
decompression adequate to achieve complete melting when the mass
reaches about 80 km. At the maximum rate of rise predicted by
Ramberg, it would take about 400,000 years for the mass to travel the
I ,000 km to the surface if the viscosity is 10 22 poises, the density
contrast is 0.3 g/cm3 , and the initial layer thickness is about 10 km.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 71
Phase changes, including the appearance of liquid, would no doubt
accelerate the process at shallower levels as the density contrast
increases, whereas cooling of the mass and higher viscosity of the
ambient rocks would retard the process.
The diapiric model rests heavily on the dynamic instability of a
buoyant layer. If the diapirs were to originate at shallower depths than
those considered by Ramberg, the low-velocity layer is an obvious
nominee. Contrarily, Press (1970) found the density to be high in that
region, and the instability indicated is for the sinking, not the rise, of a
mass. In addition, the reader is reminded of the rarity of gabbroic
batholiths. A detailed analysis of the heat losses from a diapir would be
valuable in determining its potential as a magma source at various
depths en route to the surface ((f. Cawthorn, 1975, p. 116).

MECHANICAL ENERGY CONVERSION TO HEAT

Another group of mechanisms for generating magma depends on the
conversion of mechanical energy into heat. These mechanisms have
been called shear melting, frictional melting, tidal deformation melting,
viscous heating, and stress-difference melting. The idea stems from
Nutting (1929), who calculated the force required to move one surface
of rock over another in the absence of deformation and grinding. The
work done per unit area is presumed to be converted entirely into heat:
dq = k F dx, where dq = heat generated, k = coefficient of friction,
F = force, and dx = displacement. None of the energy is released as
seismic waves. For example, if the force (F) pressing two rock layers
together is I kbar, the displacement (dx) = I em, and the coefficient of
friction (k) = 0.20, then the heat generated (dq) is 4.8 cal/cm 2 • This
amount of heat is sufficient to raise the temperature of a layer I em
thick, having a specific heat of0.30 cal/g/deg and a density of2.5 g/cm 3 ,
about 6.4°C.
Experimental evidence has been obtained indicating that the temper-
ature produced along sliding surfaces will approach the melting point of
minerals. Bowden eta/. (1947) observed local transient temperatures
up to I000°C at moderate loads and speeds, using glass, quartz, and
other transparent solids through which the transmitted infrared radia-
tion could be measured. Most of the corroborative experiments were
carried out on metals that are amenable to a greater variety of tech-
niques (e.g., Bowden and Thomas, 1954). Whereas the true contact
area is critical to heat generation, Reitan ( 1968) also emphasized the
need for strain rates greater than J0- 14/sec.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

72 GENERATION OF BASAL TIC MAGMA

THRUST FAULTING

DeLury (1944) applied Nutting's idea to a 1-km overthrust of an outer

layer of the earth 60 km thick, which he believed would develop
enough heat to warm and melt 200 m at the base of the overthrust. He
noted that the Nutting formula does not apply after melting starts and
the friction is diminished, and he further emphasized that the speed of
deformation and insulating qualities of the rock are important.

SUBDUCTING PLATE

With the development of plate tectonics, the DeLury application was

investigated in great detail. For example, Toksoz et a/. ( 1971) calcu-
lated the temperature as a function of time in a subducting plate, taking
into account phase changes. radioactive heating, adiabatic compres-
sion, and shear-strain contributions (Figure 4-12). They believed that
with a velocity of transport of 8 cm/yr and a high value of shear-strain
heating, melting would begin to take place along the upper surface of a
slab 100 km thick at a depth as shallow as 180 km. These magmas were

Horizontal distance, kilometers

100 200 300 400 500 600 700 800 900
0

1001=----

2001,----
.
~ 300
•e
0
~ 400

FIGURE 4-12 Magma generation model of Toksoz et a/. (1971. p. 1125. Figure 7)
based on a high value of shear-strain heating along the upper surface of a 100-km-thick
slab subducted into the mantle. The temperature distribution is calculated at time= 13
m.y. with a spreading rate of 8 em/yr. The stippled areas indicate zones of phase changes.
The zone of shear-strain heating is alleged to occur along a narrow zone ( -10 km thick)
at the top of the slab. Beginning at a depth of about 180 km. melting initiates at the top
surface of the slab. In the presence of water. the zone of melting is :!:: 100 km from the top
surface of the slab. (With permission. Copyrighted by the American Geophysical Union.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 73
probably derived from a significant portion of the upper part of the
descending slab, consisting of metamorphosed sediments and oceanic
basalt; that is, they are probably not basaltic. Temperatures more
appropriate for the production of basaltic magma are not achieved until
about 200 km, a depth greater than that necessary for magma produc-
tion in nearby mantle at higher temperature!
In an unsuccessful attempt to explain the high heat flow then
believed to occur behind the trench in the sea floor above the subduct-
ing plate, McKenzie and Sclater ( 1968, p. 3177) estimated the heat flow
due to stress heating. Using the same velocity of subduction (8 cm/yr)
used by Toksoz eta/.. they obtained a value of 0.6/Lcal/cm 2/sec, which
is 1/25 that estimated for the heat-flow contribution due to shear by
Toksoz eta/. (1971, p. 1117). In view of the uncertainties, these values
are in reasonable agreement.
The principal attraction of frictional melting is that it appears to fulfill
the need for magma production directly below a line of active vol-
canoes paralleling a trench. The lower temperatures relative to adja-
cent mantle are believed to be adequate if water is available-a point
considered below (pp. 78-86). Factors that raise doubts as to the
effectiveness of frictional heating as the main cause of magma produc-
tion associated with the Benioff zone include the presence of anhy-
drous basalts along the active volcanic line, the increase in age behind the
trench, the primitive character of the strontium isotopes of some of the
extrusions, the range of composition of layers from quartz- to
nepheline-normative, the reduction of friction with initiation of melt-
ing, the loss of energy due to seismic events. and the increased lateral
heat conduction with depth. The subducting slab is a heat sink and not
a likely place to generate magmas.

PROPAGATING CRACK

The shear-melting hypothesis was proposed by Griggs (1954) to ac-

count for deep earthquakes, the release of stored elastic energy. The
idea was expanded by Griggs and Handin (1960, p. 360-364) and Griggs
and Baker (1969), who equated Starr's (1928) calculation for energy
released from a propagating crack to the energy required to melt a
mass. The aspect ratio of a circular flaw having the appropriate energy
for melting is

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

74 GENERATION OF BASALTIC MAGMA

where r = radius of flaw, d = thickness of mass around flaw melted,

p, = shear modulus, Ml = enthalpy of melting and T = shear stress.
For p, = 3 x 10 11 dyne/cm 2 , p = 3.3 g/cm 3 , Ml = 100 cal/g = 4.2 x 1011
dyne cm/g, and T = 100 bars= 108 dyne/cm 2 , the aspect ratio, rid, is
5.5 x 105 • In order to initiate shear melting the flaw must obviously
have an exceptionally high aspect ratio. Griggs and Han din ( 1960)
thought that (a) weak tabular minerals such as mica, (b) a lenticular
fluid inclusion, or (c) an instability in structural elements generated by
non-Newtonian viscous flow-i.e., creep (see Chapter 9)-would serve
as a focal region for initiating shear melting.

REGENERATIVE FEEDBACK

Melting as the result of shear requires that the layer of rock have low
heat conductivity so that the heat will not be dissipated. In addition,
the heat must be concentrated to provide the large enthalpy of melting.
The concentration is presumed to take place because of the nonlinear
relationship between viscosity and temperature. In mantle materials
the viscosity, TJ, is experimentally dependent on temperature (Stocker
and Ashby, 1973; Kirby and Raleigh, 1973) according to the relation-
ship

where a varies inversely as the activation energy in rough approxima-

tion to the Arrhenius equation (Anderson and Perkins, 1975, p. 118).
Because of this relationship, a rise in temperature lowers the viscosity
with a resultant increase in strain rate. This sequence, in turn, results in
a larger temperature increase, which again increases the strain rate,
and so on until the energy source is exhausted. The "runaway"
character of this process was described by Griggs and Handin ( 1960, p.
361) and labeled by Orowan (1960, p. 341) an "avalanchelike" process.
A rigorous analysis of several cases was carried out by Shaw (1969,
1973), who used the Gruntfest (1963) theory describing "thermal
feedback." The adiabatic case, illustrating a sudden temperature rise
with time, is given in Figure 4-13. With a constant stress of 100 bars
acting on a solid at its beginning-of-melting temperature and having an
initial viscosity of TJo = 10 21 poises, a runaway condition is reached in
about 3.5 m.y.
The regenerative nature of the process, even in large bodies, was
emphasized by Anderson and Perkins (1975), who noted its similarity
to the exothermic conditions achieved in explosives. The size of the

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 75
60

100 50

40

~
~ CD

.§ 60
n
30 CD
~
....
I
3
~
40 20

20 10

0
2 3
Millions of years

FIGURE 4-13 Magma generation model of Shaw (1969, p. 528,

Figure 8) bas~d on accelerated heating resulting from a constant
stress on crystalline rock initially at its minimum melting point.
Shear heating produces a rise in temperature, which lowers the
viscosity , thereby increasing the strain rate. The sequence of events
results in a rapid rise in temperature. Model calculated for a 200"C
melt range, an initial viscosity of 1021 poises, a linear relationship of
viscosity with crystal-liquid fraction, and a constant stress of 100
bars. (With permission of the Oxford University Press.)

source region or schlieren zone through which the shear is converted

into the enthalpy of melting for a Hawaiian volcano was estimated by
Shaw (1973, p. 1516) to be about 5 x 10" km 3 (50 x 75 x 140 km). Such
a mechanism for melting no doubt requires mantle deformation closely
related to plate motions .

TIDAL DISSIPATION

Shaw (1970) proposed that the tidal energy dissipated in the solid earth,
-10 19 erg/sec, could also be converted to thermal energy . On the basis
of his enthalpy of melting of 100 cal/g and his estimate of tidal power,
about 30 km 3/yr of rock at the melting temperature could be converted

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

76 GENERATION OF BASAL TIC MAGMA

to magma by that form of shear melting. That amount of magma is

about three to five times the present annual production of lava. In view
of Swanson's (1972; personal communication, 1975) estimate that, for
present-day Hawaiian eruptions, only roughly 50 percent of the magma
reaches the surface, present-day magma could be produced by the
complete conversion of tidal energy.

COMPOSITIONAL CHANGE
A major group of processes for melting rocks depends on change of
bulk composition. Consider a rock at a temperature somewhat below
its beginning of melting. If the composition of the rock is changed, by
the addition or subtraction of material, to a new composition whose
solidus lies below the same temperature, then melting will ensue. The
change of composition can be achieved by metasomatic processes such
as diffusion and volatile fluxing.

DIFFUSION

Bell and Mao (1972) used diffusion in a pressure gradient to obtain a

more favorable composition for melting. They have suggested differen-
tial stress coupled with differential diffusion rates as critical conditions
for magma generation. They chose a binary system, R-S, with four
compounds related by one peritectic and two eutectic points to illus-
trate the process (Figure 4-l4A ). The chemical potential p, and the
concentration X for both components are shown in Figure 4-l4B in a
pressure gradient P 0 - P 1 • Before diffusion, the composition repre-
sented by m in Figure 4-l4A would consist of the compounds II and III.
If the diffusion of component R greatly exceeded that of component S
in response to the pressure gradient, monomineralic layers would
develop.
The chemical potentials and concentration after diffusion are given
in Figure 4-14C. Because the temperature of point m is above the
temperature of the eutectics c and d at P 0 , melting would begin at the
interface of layers III and IV as well as I and II. Bell and Mao (1972)
emphasized that the two derivative liquids generated from a single
parent cannot be related to the parental material by fractionation trends
and that the two contrasting magmas do not have a simple relationship
to each other. The bimodal character of igneous rocks (Daly, 1925b;
Chayes. 1963) could be accounted for by the process suggested by Bell
and Mao (1972).
Diffusion in Hofmann's (1974) view is limited even when a melt is

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 77

s- -R
m mn I

R s
P.R P.s

B XR Xs

n+m D+m
Po PI Po PI
P.s
R s
Xs

c P.R

I n m D7 XR 1 n m Il7
p p
FIGURE 4-14 A. Schematic phase diagram for a portion of the system
R-S with four compounds I, II, Ill, and IV. Two eutectics, having the
compositions c and d, and a peritectic are exhibited. The composition m
consists of compounds II and III. P 0 = liquidus at initial pressure; P 1 =
liquidus at higher pressure.
B. The chemical potential, 1-£, and concentration, X, of R and S compo-
nents as a function of pressure before diffusion.
C. The chemical potential, 1-£, and concentration, X, of R and S after
diffusion, assuming the diffusion of R greatly exceeds that of S . Four
monomineralic zones are produced. Melting begins at the interface of zones
I and II and at the interface of zones III and IV according to BeU and Mao
(1972, p. 418, Figure 28). (With permission of the Carnegie Institution of
Washington.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

78 GENERATION OF BASAL TIC MAGMA

..
I:ii

..i
::0

f
Q.

1400 1600 1800

Temperature, •c
FIGURE 4-15 Melting curves of diopside (Di) without volatiles
(Yoder, 1952; Boyd and England, 1963), with H20 (Yoder, 1965,
p. 87, Figure 12; Eggler, 1973, p. 458, Figure 17), and with C02
(Eggler, 1973, p. 458, Figure 17). (With permission of the Carnegie
Institution of Washington.)

present to "the scale of a few hundred meters," and therefore the

diffusion-melting process would require an extensive zone of shearing
to produce large volumes of magma. The more rapid diffusion of
volatiles, however, is another matter.

VOLATILE FLUXING

Volatiles have a considerable influence on melting behavior. Both H2 0

and C02 , the principal volatiles in the earth, reduce melting tempera-
tures well below that for melting in the absence of volatiles. A repre-
sentative measure of the relative influence of H2 0 and C02 on the
melting temperatures can be seen in Figure 4-15 (see also Figure 10-12)
for the major rock-forming mineral diopside. One may easily estimate
the effect of adding either H2 0 or C02 to a rock near the beginning-
of-melting temperature.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 79
It is not necessary that the volatile be present in amounts in excess of
that required to saturate the liquid. Although only one system, albite-
water, has been studied in sufficient detail to document the principles,
that system is thought to be representative of a wide variety of
rock-forming minerals in the presence of volatiles. The effect of
reducing the water pressure below the total pressure is displayed in
Figure 4-16. An important observation is that the beginning of melting

10

•...
0
j
:ii
5
i
~
•
••...
II. 4

Temperature, •c

FIGURE 4-16 Pressure-temperature projection of the albite-

H20 system with the liquidus indicated for various water pressures
(after Burnham and Davis. 1974. p. 936. Figure 19). The tempera-
tures a and b are displayed in an isobaric section in Figure 4-17.
(With permission of the American Journal of Science . )

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

80 GENERATION OF BASALTIC MAGMA

remains the same whether or not H2 0 is available in excess of that

required to saturate the initial liquid. The liquidus, on the other hand, is
lowered below the anhydrous liquidus as the water pressure is in-
creased. The temperature-composition section for a total pressure of 5
kbar is given in Figure 4-17. The points a and b are identified in Figure
4-16. If the amount of water present is 0.5 percent, the beginning of
melting is the temperature of b and the water content of the liquid is
about 10 percent, but only 5 percent liquid is produced. The amount of

Llquld+Gaa

lblte +Liquid

AlbHe+Gaa

&OOL-~~~~~~~~~--~--~~~~±-__.
Albite weitht [Link]

FIGURE 4-17 Estimate of the albite-H,O system at 5 kbar total

pressure. The temperature of a is from Birch and LeComte ( 1960); b
and care from Yoderet al. (1957. p. 208. Figure 38) with correction
by Morse (1970. p. 234. Figure 2; Errata). The composition of dis
from Davis (1972. p. 31. Figure lA). (With permission of the Car-
negie Institution of Washington.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 81
water present, in effect, determines the proportion of liquid to crystals.
If 2.5 percent H20 is present, then 25 percent of the albite + H 20
mixture will initially melt at the temperature b. As the temperature
increases above b , the liquid contains decreasing amounts of H 20.
The melting of a hydrous phase introduces a variety of complexities.
A temperature-composition section is shown in Figure 4-18 for the
anhydrous constituents of the phlogopite ( = forsterite + kal-
silite + leucite) and water system (Yoder and Kushiro , 1969). The
compositional region the writer believes to be most pertinent to the
melting of rocks in the mantle, where the H 20 content is presumed to
be very low, is unfortunately not well determined. (Additional hydrous
systems in the absence of a free gas phase should be investigated, but
the rates of reaction are slow and equilibrium is difficult to establish.)
The diagram implies unique melting relations different from those
where the H 20 content exceeds that of the hydrous crystalline phase. If
the higher beginning-of-melting temperature is a general effect in
assemblages consisting of anhydrous and hydrous phases only (without
a free gas phase), then present melting data on minerals in the presence
of excess H 20 may not be pertinent to melting in the mantle. For this
reason, melting in the mantle may not be as pervasive as the relations
in Figure 4-20 suggest. As will be seen later, the presence ofphlogopite
in the mantle is believed to be necessary because of its potassium and
water content. None of the major minerals in garnet peridotite is known
to carry sufficient potassium to yield basaltic compositions on melting
(see Kushiro, 1973b, p. 295, Table 84), and phlogopite has the largest
range of stability known for a hydrous mineral. On the other hand,
there are limitations to the amount of modal phlogopite because of the
heat production constraints on the potassium content of the mantle.
When both C02 and H 20 are present, both the liquidus and solidus
are dependent on the proportions of the volatiles. In Figure 4-19 is
shown the influence of various proportions of C02 and H20 on the
solidus and liquidus of the diopside system. The effect was also
observed in a study of a natural garnet peridotite by Mysen (1973,
p. 473).
Some investigators believe that there is always a gas phase present in
the mantle because of the small amount of C02 required for saturation
of magma. For example, only 4.8 ± 1.0 percent C02 (Eggler et a/.,
1974, p. 227) is required to saturate a diopside melt at 30 kbar and
1625°C, whereas 21.5 ± 1.0 percent H 20 (Hodges, 1974, p. 253) is
required at the same pressure and 1265°C. Most workers, however,
have the opinion that there is not a free gas phase and the volatiles are
stored in various hydrous and carbonate minerals. Most of the hydrous

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

82 GENERATION OF BASALTIC MAGMA

P= 10 Kb
\ \
\ \
\ \
\ L \
\ \ L+G
\
\
\\ 0

\ \
\ \

Fo + L

•
\

•
\\
(d)
\

------------ --
130
Fo + L+ G

II II CS> CS> CS>

B II

Ph+ Fo +G

11000 5 40
Phlogopite --+HzO
K20 • 6MgO•Aiz03•6SiOz
Weight percent

FIGURE 4-18 Temperature-composition projection at 10 kbar for the silicate-rich

portion of the K20 · 6Mg0 · Al20 3 • 6Si02-H 20 join, illustrating the complex melting
relations of the hydrous phase phlogopite in the absence of a free gas phase. Ph =
phlogopite; Fo = forsterite; Lc = leucite; Ks = kalsilite; L = liquid; G = gas.
Temperature a is the beginning of melting in the presence of excess H20; b, the beginning
of melting in the absence of a gas phase; c, the maximum thermal stability of phlogopite;
and d, the minimum liquidus in the absence of a gas phase. The melting relations for the
assemblage Ph + Fo + Lc + Ks + are closely analogous to the assumed relations in the
mantle. From Yoder and Kushiro (1969. p. 576, Figure 4). (With permission of the
American Journal of Science.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 83
8

7
0 .0 0.4 lo.e
I
6
I

5
"'
~

0
D
0
~
4
~

."'"'"
~

CL
3

1250 1300 1350 1400 1450 1500

Temperature, °C

FIGURE 4-19 Melting of diopside in the presence of C02 and H20

and in the absence of volatiles. The solidus is raised as C0 2/
(C02 + H20) in the vapor, indicated on each curve, increases ac-
cording to Rosenhauer and Eggler (1975, p. 475, Figure 47). (With
permission of the Carnegie Institution of Washington.)

minerals break down at temperatures below melting, and the volatiles ,

if retained, are eventually consumed in the melt; only a few hydrous
minerals are stable into the melt region. Phlogopite, for example, is not
an uncommon constituent in ultrabasic rocks. In addition, there is the
possibility that garnets and pyroxenes store water as hydrogamet and
hydropyroxene by way of substitution of 4H+ -Si 4 +. At pressures less
than about 30 kbar, various amphiboles are stable. Polymorphs of
calcite and dolomite are stable to very high pressures. The melting
curve of calcite (Yoder, 1975b, p. 892, Figure 6; Irving and Wyllie,
1975, p. 38, Figure 2) is about 100° below that of diopside . The paucity

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

84 GENERATION OF BASALTIC MAGMA

of diopside relative to enstatite in the mantle may be attributed to a

small degree to the reaction

diopside + 2 forsterite + 2 C02 = 4 enstatite + dolomite,

experimentally determined by Eggler (1975, p. 472, Figure 46). The

concept that the earth is continuously degassing was supported in a
detailed analysis by Rubey (1951). If the major oceans were formed by
degassing primarily during the Precambrian, then there should be vast
regions in the mantle depleted in volatiles. The observation of great
volumes of relatively anhydrous basalt on the present ocean floor
supports that view.
Over the years, the water content of the mantle has been estimated at
various values ranging from 13 wt% (the H 20 content of serpentine) to
zero. In general, opinion about water in the mantle has swayed from
excess H 20 to H 20 deficiency (free vapor absent and water insufficient
to saturate a magma), but with the increased role assigned to C02 the
idea of a free vapor phase has been revived. Others hold to the concept
of a relatively anhydrous mantle with volatiles, especially H 20, being
added to the magma in the crust while stored in an auxiliary magma
chamber. The hydration of a magma by diffusion, however, is probably
limited to within a few tens of meters of the boundary (Shaw, 1974, p.
156). The oxygen and hydrogen isotopes present strong evidence that
H20 circulating at the periphery of shallow chambers is absorbed for
the most part after crystallization of the magma (Taylor, 1974, p. 856).
Although the criteria for identifying primordial H 20 are not well
defined, anomalously high 3 He/ 4 He in seawater (Clarke et a/., 1969),
in pillow basalt glass rims (Lupton and Craig, 1975), and in volcanic
gases (Tolstikhin, 1975) has led to the view that primordial 3 He is being
released from the mantle. The implication is that there may be a
continuous flux of primordial H20 from the undepleted mantle. Al-
though most of the volatiles associated with basalt and andesite are
considered to be of crustal origin, those volatiles in kimberlite appear
to offer an opportunity to test a mantle origin.
The degassing of the earth, if that is indeed the source of the oceans,
must have resulted in extensive metasomatism of the upper mantle.
One cannot help being impressed by the data of Nakamura and Kushiro
(1974, p. 258) on the composition of gas in equilibrium with forsterite
and enstatite. The Si02 content of the gas is of the order of 20 percent
(wt) at 15 kbar and 1150-1350°C! Although the presence of C02 may
reduce the solubility of Si02, such an ascending gas would greatly
enrich the upper horizons in enstatite relative to olivine. The other

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 85
constituents in the gas would produce an array of other changes,
perhaps including the deposition of secondary phlogopite. It is no
wonder that most of the nodules retrieved from basalts and especially
from kimberlites appeared to have been soaked in a corrosive juice and
metasomatized. As will be seen below, volatiles may also play a major
role in determining the kind of magma that reaches the surface of the
earth.
The principles outlined above can be applied to the melting of
peridotite. Kushiro et al. (1968a) and Mysen and Boettcher (1975)
studied the melting of natural spinel lherzolite in the presence of H20
with and without C02 (Figure 4-20). In the presence of an excess of
H 20, the solidus is curve F and a large partial melt region would exist
below about 50 km under the oceans and below about 80 km under the
Precambrian shields, assuming the geotherms of Clark and Ringwood
(1964) are applicable. With increasing amounts of C02 (curves E. D.
and C), the solidus is raised and melting begins at successively greater

40

35

30 100

c
~
25
~~ ~
•
75 ::r
~·
0

r.:,•o"
.Q
!! 20 ~
:. :-.•'f) 0
3
~
..
••
~~
50 ~
e
•
a:•
25

Temperature, •c
FIGURE 4-20 Melting relations of a natural spinel lherzolite. Curve A is the solidus
under anhydrous conditions (Kushiro et a/., 1968a, p. 6026, Figure I); curve B, the
estimated solidus with C02 ; curves C, D, and E, respectively, the solidus with H2 0/
(C02 + H.O) = 0.25, 0.50, 0. 75; curve F. the solidus with only excess H 2 0 (Mysen and
Boettcher, 1975, p. 542, Figure 68). Geotherms from Clark and Ringwood (1964, p. 53,
Figure 2). (With permission. Copyrighted by the American Geophysical Union and
Oxford University Press.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

86 GENERATION OF BASAL TIC MAGMA

depths under both oceanic and shield regions. The results with H 20
have been verified for the most part by Kushiro (1970), Green (1973),
and Millhollen eta/. (1974).
The excessive amount of melting in the presence of an excess of
water was an embarrassment to the early experimenters applying their
results on minerals to magma production in the mantle. Some of the
conclusions drawn were that (a) water was not present in excess in the
mantle-it was not a free phase; (b) the assumed geothermal gradient
may be too high in temperature for the upper mantle; (c) peridotite may
not be the parental material in the mantle; (d) the mantle was depleted
of volatiles during the early Precambrian to form oceans; and (e) other
volatiles may reduce the effectiveness of water. The present preferred
view is that the mantle is not uniformly or entirely depleted of volatiles,
and whereas water is the principal volatile, other volatiles such as C02
reduce the activity of water. Given the present range of estimated
geotherms, which do not intersect the anhydrous solidus of peridotite,
it seems clear that some small amount of volatiles must be present if the
low-velocity zone now observed is caused by partial melting. The
amount and proportion of volatiles will determine the amount of melt
now existent.

OTHER MECHANISMS

One of the more important mechanisms for the production of melt is

radioactive heating, as will be discussed in the following chapter. The
role of exothermic chemical reactions was considered important by
Daly (1911), Day and Shepherd (1913), and Jaggar (1917); the present
writer, however, supports the view of Holmes (1927) that such reac-
tions are probably not of fundamental importance until the magma has
accumulated and is rising to the surface.

PHENOMENOLOGICAL VIEW OF MELTING

Before leaving the various mechanisms for producing a melt, 1t 1s
appropriate to mention the potential opportunities for investigating the
physics of melting of the crystals that make up the rocks. The change
from a crystalline state to a liquid state is considered by some inves-
tigators to result from the breaking of the weakest bonds in a crystal.
As the melting point of a crystal is approached, increasing amounts of
energy are absorbed (see Figure 5-1) by the crystal through vibrational
stretching of the bonds (Lindemann, 1910). When the stretching ex-
ceeds a critical value, the bond breaks and melting ensues. The

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Melting Processes 87
resulting masses of substructural units dispersed in a low-strength
medium appear to have exotic compositions.
Positional disordering and other mechanisms for increasing the en-
tropy are favored by some investigators in describing the change from
the crystalline to the liquid state. The liquid state is believed to be
quasi-crystalline and is characterized by loss of long-range ordering.
Such models are amenable to statistical analysis; at best, however, they
are somewhat artificial, according to Ubbelohde (1965, p. 201). On the
basis of the Tolland and Strens (1972) model for electrical conduction,
Strens (personal communication, 1975) described melting as the critical
stage where a three-dimensional net of dislocations obtains and the
masses of multiple unit cells fail in strength.
The structure of diopside was determined with precision at a series
of temperatures up to I000°C (Cameron eta/., 1973; Finger and Ohashi,
1976), and there were no significant distortions of the structures with
increasing temperature. The vibrational ellipsoids do not intersect
when their changes are projected to the melting "point," and there is
no obvious indication that the structure is about to fail because of
mechanical instability. Hazen (1975, p. 171) concluded that the melting
points of olivines may be determined by structural constraints. He
observed that forsterite, hortonolite, and fayalite all have similar cell
constants, hence a maximum cell volume, at their respective melting
points. Determination of the crystal structure in the region of anoma-
lous increase in enthalpy prior to melting and of the liquid structure just
above the melting point is a goal almost within reach of present
technology. A satisfactory theory is needed to explain the phenomenon
of melting and to account for the enthalpy of melting.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Thermal Energy
5 Requirements
for Melting

Volcanic heat represents only a very small fraction of the heat

radiated by the earth .
. . . the real problem of volcanoes is not so much to find a
suitable source of energy as to provide ways and means by which
rather insignificant amounts of heat can be focussed on relatively
insignificant volumes of the crust.
Verhoogen (1946. p. 747)

ENTHALPY OF MELTING
The thermal energy required to convert crystalline rock to magma, the
enthalpy of melting, is often referred to as "heat of melting," "heat of
fusion," or "latent heat of fusion" even though heat is defined as the
energy transferred as a result of a temperature difference (Zemansky,
1937, p. 48).* Why is it important to know the amount of energy required
to bring about melting? First, the enthalpy may be a guide to the distribu-

• Heat is not a quantity describing a state of a system. nor is it the stored energy inside a
system . In Professor G. Tunell's view (personal communication. 1975). "The amount of
heat that would be transmitted in a strictly isothermal expansion or compression'' can be
calculated even though heat does not flow from one body to another in the absence of a
temperature difference . He has considered the enthalpy of melting " a limit value
approached when the temperature difference between surroundings and a pure. simple
compound in equilibrium with its liquid is very small." There is some basis. therefore.
for using the common expressions "heat of melting," "heat of fusion." or "latent heat of
fusion" in describing the thermal energy required for melting a rock.
88

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Thermal Energy Requirements for Melting 89

tion of heat-producing sources. Second, it is a constraint on the mecha-
nism of magma formation . Third, enthalpy is an important factor in
determining the amount of time required to produce a melt, and the rate
of magma production may relate to the periodicity of eruption (see
Chapter 10). The amount of energy required to bring a rock to the melt-
ing temperature may be considerable, depending on the initial tempera-
ture, and some estimates for simple systems are given below. During
heating to the melting temperature, a rock may undergo several endother-
mic metamorphic reactions . The total energy consumed in these reac-
tions may far exceed the energy required for melting. For most purposes
the ambient temperature is presumed by petrologists to be already
below but near the beginning of melting, so the principal concern is the
source of energy for the melting process itself.

MEASUREMENT

From a practical standpoint the observation of the enthalpy of melting

in a differential thermal apparatus is made over a temperature interval,
and the energy absorbed appears as an anomalous heat capacity
(Figure 5-l). The area under the peak, when properly calibrated, is a
measure of the enthalpy of melting. The enthalpy of melting is usually
measured more accurately in a calorimeter by observing the difference
in energy released by dropping a mass of crystals and a separate mass
of glass of equivalent composition, respectively , from temperatures
just below and just above the melting point into a relatively cool,
thermally insulated bath. The rise in temperature of the bath or the
change in proportion of crystals and liquid in a suitable bath mixture is
used to evaluate the energy released by each state of the material.

t
!. FIGURE 5-1 Schematic represen-
tation of the differential thermal
~ analysis of a silicate undergoing con-
•
Q.
gruent melting. The area under the
E
!! peak, enclosed by extrapolating the
:! baseline, is proportional to the en-
c thalpy of melting. T,. = melting tem-
! perature.

=
0 ~--------------------~~----------------~
Tm
Ttmptroturt -

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

90 GENERATION OF BASAL TIC MAGMA

PREVIOUS ESTIMATES

Early measures of the enthalpy of melting, ![Link], were made, for

example, by Baros (1893), who obtained a value of24 cal/g (1200°C) for
a diabase using the cooling-curve method. Vogt (1904, p. 55) c~n­
sidered this value to be too low because the diabase did not crystallize
completely on cooling. He preferred a value of 90-100 cal/g, based on
the enthalpy of melting of minerals such as diopside, anorthite, aker-
manite, and fayalite. Bowen (1928, p. 314-315) assumed that the value
100 cal/g was roughly that for basalt, presumably relying on the values
for diopside and anorthite obtained from calculations of the freezing-
point depression as well as measured values for the constituent miner-
als (e.g., White, 1909, p. 486, last footnote). He later found support for
using this value even for peridotites when Bowen and Schairer (1935,
p. 207) estimated that the !l.H m for pure forsterite was exactly I 00 cal/g
on the basis of calculations of the freezing-point depression in the
forsterite-fayalite system. The value 100 cal/g has been used in many
present-day calculations of the thermal factors involved in the genera-
tion of basaltic magma.

ENTHALPY OF MELTING OF MINERALS AND ROCKS

VALUES AVAILABLE FOR END-MEMBER MINERALS

Although it is not yet possible to deal effectively with iron-bearing

natural rocks in high-temperature calorimeters because of oxidation
and crystallization problems, estimates of the enthalpy of melting can
be made using data on constituent end-member minerals. The values
now available are given in Table 5-l (Yoder, 1975a, p. 515).

MODEL SYSTEM FOR BASALT

The calculated enthalpy of melting for the "eutectic" composition,

Di 58 A~ 2 (wt%), of the diopside-anorthite system (Figure 5-2), neglect-
ing solid solution, phase changes, and heat of mixing, is 77.8 cal/g.
Using values for the heat content of the end-member phases from
Robie and Waldbaum (1968, p. 221 and 226), the heat requirements to
melt compositions in the diopside-anorthite system can be displayed in
a plot of the change in enthalpy, Hr-H2 98 , versus composition (Figure
5-3). Such a diagram can be used as a model for some aspects of the
melting of basalt.
The data in Figure 5-3 illustrate how small !l.H m is relative to the total

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Thermal Energy Requirements for Melting 91

TABLE 5-l Enthalpy of Melting"
End-Member
Mineral tJI,. (kcallmol) Formula wt t:JI,. (callg)

Forsterite 29.3 (est)'' 140.70 208.2

Fayalite 22.0:, 203.78 108.1
Clinoenstatite 14.7 100.38 146.4
Diopside 18.5 216.52 85.4
Pyrope'' 33.2'1 403.08 82.4
Anorthite•· 18.7'1 278.14 67.2
High albite 13.5. 262.15 51.7
High sanidine 14.7 278.24 52.8
"From Robie and Waldbaum ( 1968) unless otherwise noted.
•Bradley (1962).
ro. J. Kleppa and T. V. Charlu (unpublished data. 1975).
• .1H. at 7000C.

heat required to bring the assemblage to the temperature of the

beginning of melting. Beginning at temperatures assumed for the upper
mantle, the amount of heat required for partial melting is still relatively
small compared with that required to raise the temperature of the rock
to the beginning of melting. The figure also indicates how the heat

1500

~ 1400
i:J
\ Anorthite+ Liquid
\
..
~ 1300
Q.
E
\Diopside 55 +liquid ...;;::1~------------.........j 1274o
\ - - - - - - - - Diopsiden +Anorthite+ Liquid
{!! 1200 ,------------------- 1222°
'I

1100
i Diopsiden +Anorthite+ Tridymite

20 30 40 50 60 70 80 90 Anorthite
Weight percent
FIGURE 5-2 The diopside-anorthite system at I atm (after Osborn , 1942; Clark eta/.,
1962; Yoder, 1965). The depression of the solidus for diopside-rich compositions is due
primarily to solid solution of the Ca-Tschermak's molecule , CaA [Link] 20 6 , in diopside and
resultant enrichment of silica in the liquid. [The extensive solid solution (25 wt%)
indicated by Hytonen and Schairer (1961, p. 136), on the basis of ceU dimensions and
optical observations, is considered to be suspect because of metastability.) (With per-
mission of the Carnegie Institution of Washington.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

92 GENERATION OF BASALTIC MAGMA

FIGURE 5-3 Plot of Hr-H 2 .,. ver-

sus composition in the diopside-
anorthite system, neglecting solid so-
lutions, phase changes, heat of mix-
ing, and pressure effects. Tempera-
ture in degrees K. The value of 411,.
for diopside measured at 973°K was
used and not corrected to the melting
temperature. Modified after Yoder
(1975a, p. 516, Figure 76). (With
permission of the Carnegie Institution
of Washington.)

Mole percent

requirements change as a function of composition in the isothermal

region and through the crystal + liquid regions. Although basalts are
eutecticlike at the liquidus because all major phases appear within a
small temperature interval (Yoder and Tilley, 1962), these phases,
owing to solid solution, continue to crystallize together from liquid
over a more extended range of temperature . The temperature range is 1
also a function of the degree of fractionation of crystals and liquid. It is
instructive, therefore, to examine melting in a system involving a solid
solution series such as forsterite-fayalite. Figure 5-4 was constructed
for the most part from the thermodynamic data given by Robie and
Waldbaum (1968) and the phase-equilibrium data of Bowen and
Schairer (1935). It is seen that, with melting, heat is consumed at a
greatly increased rate but without a thermal arrest as for a eutectic
system.

IMPROVED MODEL SYSTEM FOR BASALT

The model system for basalt can be improved by choosing the

piercing-point composition at I atm, [Link] 7•5 (wt%), in the

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Thermal Energy Requirements for Melting 93

diopside-anorthite-forsterite system, studied by Osborn and Tait
(1952). The enthalpy of melting can be calculated from the end-member
minerals, assuming that there is no solid solution or heat of mixing. The
contribution of the requisite portion of forsterite to the enthalpy of
melting is estimated from the calculated value of Bradley ( 1962) be-
cause all techniques tried have failed to produce a glass of forsterite
composition for calorimetric measurement. The calculation yields a
value of 85.6 cal/g for the piercing-point composition. To obtain an
estimate of the heat of mixing, the enthalpy of melting of the piercing-
point composition was measured in a high-temperature calorimeter at
700"C by 0 . J . Kleppa and T.V. Charlu (see Yoder, 1975a). They
obtained a value for the enthalpy of melting of 18.25 ± 0.14 kcal/mol
based on an average molecular weight of 229.38, or 79.6 cal/g. If the
estimate for the enthalpy of melting of forsterite is acceptable and the
effects of solid solution can be neglected, the heat of mixing, obtained
by difference, is about -6.0 cal/g, a small but not negligible value.
For a model basalt with a more appropriate feldspar composition, the
piercing point of the system diopside-forsterite-anorthites0 -albiteso
(Yoder and Tilley, 1962, p. 396) may be taken. The composition of this
point is Di2s.sFo4.5An33.sAb33.s (wt%), and its calculated !lHm at 1 atm is
73.5 callg. The addition of iron to the system would tend to lower the
enthalpy of melting but by an unknown amount.

- __ _g_zQ9:..-

- --~'Q9:..-
zooo•_-
LIQUID
--------- FIGURE 5-4 Plot of Hr-H,98 ver-
sus composition in the forsterite-
fayalite system, neglecting heat of
mixing, incongruent melting, and
pressure effects. Temperature in de-
grees K. From Yoder(l975a, p. 517,
Figure 77). (With permission of the
Carnegie Institution of Washington .)

Mole percent

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

94 GENERATION OF BASALTIC MAGMA

WeiQht percent

FIGURE 5-5 The diopside-pyrope system at 30 kbar according to O'Hara and Yoder
(1967, p. 74, Figure 3). x = crystalline; ® = crystals + liquid; 0 = liquid. Sp =
spinel. (With permission of Oliver and Boyd.)

MODEL SYSTEM FOR ECLOGITE

In a similar way the enthalpy of melting for the "eutectic" composition

of the system diopside-pyrope at 30 kbar (Figure 5-5) can be calcu-
lated, neglecting solid solution, heat of mixing, and the presence of
orthopyroxene. For this composition, Di34 Py66 (wt%), the enthalpy of
melting calculated for I atm is 83.4 cal/g. To obtain the enthalpy of
melting at 30 kbar it is necessary to use the relationship

--
o!lH\ = !l V mO - !laD,
oP T
where !l V m is the volume change on fusion and !la is the difference
between the coefficients of thermal expansion of the solid and liquid
phases (see Maal~~te, 1973, p. Ill). Because of the small value of the
coefficients of thermal expansion (- I0-4/deg), the last term may be
neglected. For a !l V m of about 0.049 cm 3 , * the change of enthalpy of
melting resulting from a pressure increase of 30 kbar is about 35 .I cal/g, a
substantial increase! The value for the enthalpy of melting of Di34 Py66
at 30 kbar is therefore about 118.5 cal/g. The melting of the diopside-

* Based on the following densities measured at I atm and not corrected for pressure or
temperature:
Pyrope: Pvt ... = 3.031 g!cm• (B. 0. Mysen, unpublished data, 1975)
Pcr>Stot = 3.582 g!cm 3 (Skinner, 1956, p. 428)
Diopside: Pvt ... = 2.846 g!cm• (Larsen, 1909, p. 271)
Pcr,.tat = 3.275 g!cm 3 (ADen and White, 1909, p. 14)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Thermal Energy Requirements for Melting 95

pyrope system may be used as a model for some aspects of the melting
of eclogite.

MODEL SYSTEM FOR GARNET PERIDOTITE

Garnet peridotite is herein considered the most likely parent of basaltic

magma, so an estimate of its enthalpy of melting is useful. The piercing
point or "eutectic"* of the system diopside-forsterite-pyrope at 40
kbar (Figure 5-6) has the composition Di47 Fo.,Py47 (wt%), and the esti-
mated Mlm at 1 atm is 91.4 caVg. Davis and Schairer (1965, p. 125) esti-
mated that the composition of the invariant point for the diopside-
enstatite-forsterite-pyrope system was Di41En3Fo3Py47 (wt%) at 40
kbar. The Mlm at 1 atm for that reaction point, estimated from the end-
member phases, is 89.5 caVg. It is not possible to calculate accurately the
pressure effect because of lack of knowledge of the volume change of
forsterite on meltingt; however, a provisional value of the enthalpy of
melting at 40 kbar is about 135.4 cal/g. The invariant point is suffi-
ciently close in composition to represent the formation of basaltic
liquid. Its norm is Di22 Fo20 An32En2s (wt%). The ![Link] values, when
corrected for the pressure at the site of generation, may be as much as
25 percent higher than the 100 caVg value used by Bowen (1928) for
basalts. On the other hand, the pressure effect may be counterbalanced
to some degree by the presence of water, which would be expected to
reduce !l.H m greatly.

*The piercing point is the intersection of a ternary join, for example, with a univariant
curve. The univariant curve is the locus of composition of liquids in equilibrium with
the three participating crystalline phases over a range of temperature. If the range of
temperature is small. the piercing point is commonly designated as an approximation by
the term "eutectic." The univariant curve, or piercing-point curve. joins the eutectic of
the multicomponent system. The determination of a piercing point is difficult because
equilibrium is not always readily attained at the lowest melting temperatures of a system.
Once recognized in a join, the piercing point is an indication that the composition of one
or more of the phases cannot be represented wholly by the chosen components.
t A 11 V m for forsterite can be calculated from the melting-curve slope using the
Clausius-Clapeyron equation. If dT/dP = 4.77°C/kbar. T., = 2163°K. and 11H., = 208.2
callg. then 11 V m = 0.019 cm"/g. The 11 V"' calculated from the melting-curve slope is
usually considerably less than that based on measured values of the density of glass (or
liquid) and crystals (see discussion by Yoder. 1952. p. 369 ff). Using the index of
refraction offorsterite glass (n = 1.634), estimated by Bowen and Schairer (1935. p. 207,
Table VII) by extrapolation from synthetic iron-rich compositions and the specific
refractive energies for MgO and Si0 2 of Larsen and Berman ( 1934). the density of glass is
3.123 glcm 3 • The measured density of forsterite is 3.223 glcm 3 , and therefore the
11Vm is approximately 0.010 cm 3/g by this method. This disagreement will be resolved
when a glass of forsterite is obtained or a direct measurement of 11V m is made.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

96 GENERATION OF BASAL TIC MAGMA

Forsterite

Olivine

Diopside 169()- Pyrope

Welght percent

FIGURE 5-6 The liquidus fields of a portion of the forsterite-diopside-pyrope system

at 40 kbar. The point A is a piercing point of the univariant curve Di-Py-Fo-L leading
from the invariant reaction point Di-Py-Fo-En-L at a temperature of 1680 ± IO"C.
according to Davis and Schairer ( 1965. p. 124. Figure 35). The two temperatures
( 1670 ± IO"C and 1680 ± IO"C) are within the error of measurement. and the temperature
gradient may be reversed between the piercing point and the invariant point. (With
permission of the Carnegie Institution of Washington.)

APPLICATION OF ENTHALPY OF MELTING

To gain an appreciation of the role of the enthalpy of melting, two
mechanisms will be examined in detail: (a) consumption of local heat
production and (b) adiabatic rise.

LOCAL HEAT PRODUCTION

Consider a layer with a fixed radioactive heat production at 200 km

under a continent (Figure 5-7). The heat flow at 200 km is assumed to
be 0.60 X w-s cal/cm2/sec (Sclater and Francheteau, 1970). [Jordan
(1975), on the other hand, preferred a much lower value of0.21 x w-s
cal/cm2/sec at 200 km in order that the temperature at that depth be less
than 1200°C, which he believed to be appropriate for the partial melting
of a peridotite containing 0.1 percent H20.] Radioactive heat produc-
tion changes with time; it is therefore necessary to keep changes of
heat flow in mind when considering magma generation during various
geologic periods. Lee (1967) illustrated (Figure 5-8) the heat production

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Thermal Energy Requirements for Melting 97

Heat production Heat flow
alo13 eal/em:S see al06 eal/em2 see
SFC
8
Granite 3.0
Basalt
0.6
·w'ft
A 1.06
0.82

40
M. discontinuity lf 0.63

100 ~
Ultrabasie rocks -
(depleted)
0.02

••e
1l'
0
~ 200
0 .60

Garnet
peridotite
300 ~
( undepleted)
-
0.1

400
1f 0 .40

FIGURE 5-7 Heat production and heat flow for a model of a

continental shield assuming a constant heat flux at a depth of 400
km. After Holmes (1945, p. 481. Figure 254) and Sclater and Fran-
cheteau (1970, p. 525. Figure 12). SFC = surface. (With permission.
Copyrighted by the Ronald Press Company .)

from important isotopes as a function of time.* The major contributors

are uranium, potassium, and thorium, and their average abundances
in relevant rocks are given in Table 5-2.
There is no obvious correlation between magma generation and the
*The diagram of Lee (1967) was reproduced by Lubimova (1969, p. 65, Figure I) but
contains numerous original drafting errors. Dr. Lee (personal communication, 1976)
reconfirmed his calculations. and the curves drafted correctly are reproduced in Figure
5-8 with his kind permission.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

TABLE 5-2 Average Abundances of Heat-Producing Elements and Heat Production from
Radioactivity in Basic and Ultrabasic Rocks and Model Compositiona
Average Abundances of the
Radioactive Elements Total
(ppm) Heat Production
from Radioactivity
Rock Group Reference u Th .oKb Density (callsec/cm3)<
Basaltic rocks Holmes (1965) 0.7 3.0 I. I 2.91 1.20x I0- 13
Oceanic Engel and 0. 16 0. 15 0. 14 3.00 0.17 X J0- 13
~
tholeiites Engel (1964);
Tatsumoto
eta/. (1965)
Ultrabasic Holmes (1965) 0.013 0.05 0.001 3.33 0.02 X J0- 13
rocks
Lherzolites Wakita eta/. (1967) 0.019 0.05 0.007 3. 15 0.03 X J0- 13
Pyrolite II Ringwood (1958) 0.059 0.25 0.09 2.82 0.)0 X J0- 13
"In part from Sclater and Francheteau (1970. p. S24. TableS).
-K = 0.0119 percent total K (Nier. 19S0).
cPresent heat production in caVg/sec (MacDonald. 19S9. p. 1969): U: 2.2S x 10-•; "'Th: 0.63 x IO-•; .. K: 0.70 x 10-•.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Thermal Energy Requirements for Melting 99

lo' yr before present

FIGURE 5-8 Heat generation by importanl radioactive isotopes with geological time
(after Lee, 1967; personal communication, 1976).

abundance of radioactive elements. Magmas are not abnormally

radioactive, nor are regions high in radioactivity particularly more
active volcanically. As indicated in Table 5-2, samples of potential
mantle materials have a very low content of radioactive elements. Of
course, it is not known to what extent these samples have already been
depleted of these elements.
A cubic centimeter of undepleted garnet peridotite just under the
depleted zone, already divested in part of potential liquids of basaltic
composition, generates about w-• 4 cal/cma/sec. If the entire heat
production is retained, it would take 82 m.y. to supply enough energy
to melt just 5 percent of the rock at 200 km (P = 64 kbar) assuming that
it is already at the beginning-of-melting temperature, the enthalpy of
melting is 164 caVg, and the density is 3.2 glcm:~. Obviously, it would
take considerably longer if heat losses were appreciable or the degree
of melting were greater. The time required for melting would also be

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

}()() GENERATION OF BASALTIC MAGMA

longer in the depleted zone and shorter in the lower crust provided the
temperature were raised sufficiently for melting to begin.
Retention of local heat production would influence the heat flow . In
order to maintain the heat flow at the surface, the rocks at the site of
magma generation would require either (a) a higher than normal
radioactive element content or (b) a perturbation in the heat flow from
rocks below or adjoining. The higher radioactive element content of
basalt relative to its alleged source rocks is presumably the result ofthe
preference for those elements to partition into the basaltic melt ( 12: 1,
based on Table 5-2). The experimental data of Shimizu and Kushiro
(1973, p. 271) suggest an enrichment in the liquid relative to the
starting material for several elements, including potassium, of 5: 1 for a
-20 percent partial melt, tholeiitic in major element composition, of
garnet lherzolite at 1450°C and 15 kbar. The implication is that basalt is
the product of a small amount of partial melting.
McBirney ( 1967) considered how an initial thin layer of melt would
increase in thickness as additional heat was absorbed. Because the
thermal gradient is presumed to be initially tangent to the melting curve
(Figure 5-9) and the heat flow is from below, the zone of melting will
spread downward at a greater rate than it spreads upward. The temper-
ature of rocks below their melting point will be raised more rapidly than
that of rocks already at their melting point (Hess, 1960, p. 181) in that
the enthalpy of melting is many times the specific heat. The result is a
decrease in the thermal gradient below the .layer and an increase in the
thermal gradient above the layer, as illustrated in Figure 5-9. If the rock
is homogeneous laterally, the melt zone will tend to develop in the form
of a horizontal sheet. As the proportion of melt increases, convective
processes may reduce the difference in rate of advancement of the
boundaries. The downward migration of the melt zone will proceed
while the melted portion remains close to the solidus. These factors
have important implications for the kind of basalt derived from the
parental material. Diffusion within the liquid as a result of the develop-
ing pressure and temperature gradients will also influence the kind of
basaltic liquid.
The development of magma in a temperature and pressure gradient
leads to one form of "zoning melting" (Pfann, 1966, p. 254 ff.). The
ultimate effect, after partial melting has reached an advanced stage and
convection sets in, is to concentrate the more refractory phases con-
taining elements such as Mg and Ca at the high-temperature, high-
pressure end of the gradient and the more readily fusible phases
containing elements such as Fe and Na at the low-temperature, low-
pressure end of the gradient (see Chapter 6). Harris ( 1957) believed that
the potassic basalts in particular are the result of concentrating

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Thermal Energy Requirements for Melting 101

Temperature -

-•
~

Q.

0
•
ogN--+--
::
l -·
·- 'gE
~o

:::
0
Q.

Geotherm after
partial meltin9
(at t 1)

FIGURE 5-9 Preferential spreading of the partial melt zone

downward with time (t0-+t 1) as a result mainly of the enthalpy of
melting. Geothermal gradient increases (in deglkm) above the partial
melt zone, is close to the beginning-of-melting curve within the
partial melt zone, and decreases below the partial melt zone. After
McBimey (1967, p. 31, Figure 9). (With permission of Ferdinand
Enke Verlag.)

potassium in a zone-melting process; he pointed out, however, that

other expected elements, such as Sr, are not observed to- concentrate
to the same degree. Vinogradov et a/. (1971) attributed the major
differentiation in the mantle to zone melting on the basis of the
Joly-Cotter model presented in Chapter 10.

ADIABATIC RISE

A second mechanism to be analyzed from the standpoint of thermal

energy is the generation of magma through the adiabatic rise of a hot

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

102 GENERATION OF BASALTIC MAGMA

Temperature, OC
500 1000 1500 2000
0

FIGURE 5-10 Eclogite rising adiabaticaUy (dashed curve) begins

to melt at 160 km (point A). Thennal energy within the mass is
consumed for the enthalpy of melting, and the temperature of the
partially melted mass changes (dotted line) within the crystal +
liquid region. Additional heat is consumed through the pyroxenite
region as the remaining eclogite minerals are converted to the phases
in basalt. The mass reaches the surface with a temperature at or near
the liquidus of basalt. From Yoder (1975a, p. 518, Figure 78). (With
pennission of the Carnegie Institution of Washington.)

mass (Figure 5-10). It is· useful to determine the extent of melting a

mass at A will undergo as it rises under convective or buoyant forces. If
the rise is adiabatic and sufficiently rapid that melting does not take
place, the temperature of the mass will decrease slightly according to
the equation

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Thermal Energy Requirements for Melting 103

The adiabatic gradient is about 0.4°C/km (dashed line) when
a= 20 X I0-6/deg, T = I900°K, g = 980 cm/sec2 , and Cp = 0.28 callg/
deg = II. 7 x I06 dyne-cm/g/deg. The energy within the mass due to
superheating is available for the enthalpy of melting, !lHm, in an
amount equal to
Hr0 - Hrm = Cp(Ta- Tm),

where T m is the beginning-of-melting temperature and Ta is the adia-

batic temperature after rising to a given pressure level. If the average
!lHm = 120 cal/g and the average effective Cp = 0.28 caVg/deg, eclogite
rising from 160 km will arrive at the surface just at or near the liquidus.
In the region from 65 to 50 km, reactions take place as those phases of
the eclogite facies remaining in the liquid are converted to those of the
basalt facies. The reaction may merely be expressed by the change of
composition of pyroxene, for example, if garnet has been consumed in
the partial melt. The enthalpy of reaction for the conversion of crystal-
line eclogite to crystalline basalt is endothermic and has a value of about
10.7 cal/g, based on dT/dP = 145°C/kbar (Yoder and Tilley, 1962),
!lV = 0.051 cm3/g (pbasalt = 2.85glcm3 ; Pecloldte = 3.33g/cm3), and T =
1275°C.*
The assumption of ideally adiabatic conditions is, of course, not
realistic, nor can it be assumed that crystal settling, differentiation, in-
ternal radioactive heating, and other processes have not taken place. An
all-liquid magma could be achieved at any stage by separation of the
liquid produced, and the composition would be basaltic if the material
rising was at or near the eclogite "eutectic." It would be an unlikely
event under any circumstances for eclogite to rise, however, because
of its high density relative to host rocks such as peridotite. If the
material rising were garnet peridotite, the initial liquids presumably
would be closely related to basalts in composition; but if that magma,
separated at an early stage, rose adiabatically, it would arrive at the
surface greatly superheated-also an unlikely event because only very
rarely are lavas free from phenocrysts. Heat losses en route or perhaps
during storage in an auxiliary chamber may counter that objection.
A more likely event, especially in the midocean environment, is the
separation of liquid at relatively shallow depths from peridotite after
suitable mineralogical changes from garnet- to spinel- or even
plagioclase-bearing varieties. The opportunity for superheating would
be minimal and the heat losses greater. The separation of tholeiitic

• Oxburgh and Turcotte ( 1970, p. 1672) estimated the heat of transformation of eclogite
to gabbro to be 13 caVg using the slope of Green and Ringwood ( 1967).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

104 GENERATION OF BASALTIC MAGMA

magma, characteristic of the midocean environment, at shallow depths

is consistent with the physicochemical model of Yoder and Tilley
(1962) wherein the kind of basalt is dependent for the most part on the
depth of separation.
It is evident that the adiabatic rise of a rock at or near its beginning of
melting must commence at considerable depths if that is the sole
mechanism for achieving complete melting. As will be seen in the next
chapter, the loss of crystals en route is apparently necessary to achieve
the observed lava compositions, and a shallower point of initiation of
the beginning of melting would be more likely . The above discussion
serves to emphasize the importance of the enthalpy required for
melting.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Physicochemical
6 Constraints on Melting

The melting process is constrained by the physicochemical relations of

the components of the natural rock systems in accordance with princi-
ples that can be illustrated by simpler systems. Although the number of
variables in natural processes far exceeds those that can be identified
and controlled by experiment, the major variables can be successfully
investigated in the laboratory, and a close approach to the natural
systems can be achieved.

METHODS OF INITIAL MELT WITHDRAWAL

One of the requirements of a realistic melting process is the generation
of large volumes of magma of relatively uniform composition. Esti-
mates of the volumes of lava extruded in some of the great lava floods
are given in Table 6-1. On the Columbia River plateau the average rate
of extrusion was 0.02 km3/yr, and on the Deccan plateau the discharge
rate was about 0.025 km 3/yr-not excessive rates relative to present-
day emissions of about 2 km3 /yr from all 490 active volcanoes plus
approximately 5 km 3/yr along midocean ridges. There appear to be four
major physicochemical means of achieving such large volumes of
relatively homogeneous magma, the first under essentially equilibrium
conditions and the others under nonequilibrium conditions: (a) batch
melting, (b) fractional melting, (c) zone melting, and (d) disequilibrium
melting.
105

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

TABLE 6-1 Estimates of the Volume and Duration of Large Eruptions

Estimated Average Volume Duration
Region Reference Area (km 2 ) Thickness (km) (km 3 ) (m.y.)
Karroo lavas,
Republic of South Africa,
Rhodesia, Botswana,
and Lesotho Cox (1970) 2,000,000 -0.7 1,400,000 20-100

-
~
Siberian Plateau,
Soviet Union Lure and Masaitis (1964); Nesterenko and
Almukhamedov (1973) 2,500,000 0.36 900,000 130
Parana Plateau,
S. Brazil, Paraguay ,
Argentina, and Uruguay Cordani and Vandoros (1967) 1,200,000 0.65 780,000 30
Deccan Plateau,
Western India Subramanian and Sahasrabudhe (1964) 500,000 -I 500,000 10-20
North Australia volcanics,
Australia Dunn and Brown (1969) 400,000 -I 400,000
Columbia River Plateau,
United States A. C. Waters (in Kuno , 1969) 220,000 0.9 195,000 10

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Physicochemical Constraints on Melting 107

BATCH MELTING

The forsterite-diopside-pyrope system at P = 40 kbar (Figure 6-1)

serves as a useful model to illustrate the first two methods of magma
production. Consider the bulk composition X with 60 percent forster-
ite, and assume the system is ideally ternary. As the temperature is
raised, melting begins at 1670°C at the eutectic composition, E. At that
temperature, 30 percent liquid is produced, measured by the lever
rule.* The residue, R, has the mineralogy of garnet dunite. The liquid
produced isothermally can be removed in at least two different ways. It
can be removed as one batch or it can be removed as rapidly as it is
produced. The rate at which energy is supplied, discussed in Chapter 5,
is one of the controlling factors . At any stage during isothermal
removal, the liquid composition will be identical with the eutectic
composition except for the trace elements. t

FRACTIONAL MELTING

The principles of fractional melting have been presented by Presnall

(1969) and Roeder (1974), and, as Bowen (1928, p. 31) pointed out, it is
not simply the reverse of fractional crystallization. Consider the re-
moval from the site of generation of a few percent liquid as it is
formed.:J: The composition of the remaining rock is no longer X but is
depleted of a few percent of the components in composition E, that is,
X'. Because X' is in the same three-phase triangle, the new composi-
tion also begins to melt at the temperature of E, and the liquid has the

* The lever rule gives the proportion of liquid to crystals by a line constructed from the
eutectic composition, E, through the bulk composition, X, to its intersection with the join
representing the residual phases, R . The proportion of the line from X toR represents the
amount of liquid, and the proportion of the line from E to X is the amount of residual
phases.
t The definition of a trace element is obscure because under one set of conditions an
element may act as a major element (e.g., Fe in fayalite) and under other conditions,
behave as an element in trace amounts (e.g., Fe in quartz). It may be useful to restrict
the term trace element to a region of behavior of any element that satisfies Henry's
Jaw, that is, when its activity is proportional to its concentration.
:j:At first glance the removal of a few percent liquid seems to be an unlikely process; that
it can occur, however, is abundantly demonstrated by layered intrusions. The develop-
ment of monomineralic layers at the base of a crystallizing magma must involve similar
processes in which the last vestige of interstitial liquid is squeezed or diffused from the
cumulate. The same processes apparently are effective on a millimeter scale; for
example, Yoder and Tilley (1962, p. 532, Plate IOC), in an experiment on eclogite at high
pressures and temperatures, observed an accumulated garnet layer devoid of
clinopyroxene that was overlain by quench clinopyroxene.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

108 GENERATION OF BASALTIC MAGMA

1790·~----------~---------L~~L--------L----------~17~
rn ~~ ~
[Link] percent

FIGURE 6-1 The forsterite(Fo)-diopside(Di)-pyrope(Py) system at P = 40 kbar (after

Davis and Schairer, 1965, p. 124, Figure 35). E is the piercing-point composition. The
temperature of the beginning of melting is assumed to be sufficiently close to 167CY'C that
the behavior is eutecticlike. X is a bulk composition considered analogous to a peridotite,
and X' and X" are successive residual compositions resulting from the fractional melting
of X. B is the composition of liquid produced from the fractional melting of R. The dotted
line is a construction line to iUustrate the lever rule, and the dashed line is the locus of
compositions of liquids produced by melting of X at temperatures indicated by the
isotherms. (With permission of the Carnegie Institution of Washington.)

composition of E. Remove another few percent liquid, and the compo-

sition of the residue becomes X Each small batch of liquid has the
1 1
•

same composition, until the composition of the residual crystals

reaches that of R. The residual composition R, completely exhausted of
the Di component, cannot produce any more liquid of the composition
E at that temperature, and melting stops. The temperature must be
raised to 1770°C before melting can be resumed, and then the liquid will
have the composition B. No liquids of intermediate compositions can
be derived if the removal of each small batch from the site of generation
is complete. In the continuous-withdrawal model, fractional melting

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Physicochemical Constraints on Melting 109

would produce two unique batches of liquid having the compositions E
and B. The final residue would be dunite-or it too could be melted if
the requisite high temperatures could be attained, thereby yielding a
third unique magma batch.

SUBSEQUENT MELT WITHDRAWAL

The events following the production of 30 percent liquid in a single

batch remaining in contact with its residuum are quite different.
Further addition of energy now gradually raises the temperature, and
the liquid compositions change along the boundary curve between
forsterite and garnet, because all the diopside has been consumed, until
the composition A is reached. At this point the lever rule indicates that
46 percent liquid has been produced, and all the pyrope is consumed in
the liquid. The liquid compositions now leave the boundary curve
along a line constructed through the bulk composition toward forster-
ite. Complete melting is achieved when the temperature reaches the
liquidus of the bulk composition X. A succession of liquids could
thereby be produced having a range in composition between the initial
and final liquid as long as the residuum remains in equilibrium with the
liquid. Liquids withdrawn at any stage of the process would retain their
olivine-hypersthene normative character.*
Batch melting and fractional melting produce the same kind of liquid
initially, but subsequent liquids are quite different. The proportions of
forsterite, diopside, and pyrope are sufficiently close to those of
natural garnet peridotites to warrant careful consideration of the model
system. The norm of the "eutectic" is Fo20 Di22 An32 En26 , a reasonably
close approach to a pyroxene-rich basaltic composition. The model
system emphasizes the large amount of liquid of a single composition
produced initially and the importance of knowing when and how the
liquid is removed from the residuum.

IMPORTANT PRINCIPLES AND QUESTIONS

The simple phase-equilibrium diagram in Figure 6-1 illustrates several
other important principles relevant to magma generation and raises
some pertinent questions.

* Those liquids having compositions to the Py side of a line from Fo to Di,Py, (mol)
would also contain Cor, as well as An + En + Fo in the norm. In Figure 6-1, a line from
Fo through X and A to the Di-Py side line marks the boundary of corundum-normative
liquids. To the left of that line, the norm is Di + An + En + Fo.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

110 GENERATION OF BASALTIC MAGMA

I. How does the proportion of phases in the starting material influ-

ence the kind and amount of liquid?
2. At what proportion of liquid does the residuum disaggregate?
3. What proportion of liquid is produced in natural peridotite as a
function of temperature?
4. Is the melting of natural peridotite eutecticlike?
5. Which phase of garnet peridotite is consumed first in the liquid,
clinopyroxene or garnet?

Each question will be addressed in order, and the need for additional
experimental data will be evident.

PROPORTION OF PHASES IN PARENTAL MATERIAL

Consider individually in Figure 6-1 a bulk composition of 90 percent

forsterite, one of 90 percent clinopyroxene, and one of 90 percent
garnet and appropriate amounts of the remaining phases. Each bulk
composition will begin to melt at the same temperature, and it will yield
an initial batch of liquid of the same composition whether by batch or
fractional melting. The main differences arise after the eutectic liquid is
attained, depending on which phase is consumed first. As each phase is
consumed in batch melting, the liquid changes composition along one
of the three boundary curves. The nature of the change in magma
composition should therefore reflect the effects of the absent phase. In
fractional melting, the loss of a phase-no matter how minor-will cause
melting to cease temporarily in the present model. As illustrated in
Figure 6-1, an abrupt and discontinuous change of magma composition
takes place with each loss of phase. (In cases of complex solid solution,
the loss of a phase may cause an abrupt but continuous change in
magma composition.)
The above discussion emphasizes the identical nature of the initial
liquid irrespective of the proportions of the phases in the starting
material . On the other hand, the proportions of the phases will greatly
influence the trace element content of successive batches. Because the
partitioning of the trace elements into the liquid is not in the proportion
of the major phases, one can expect successive batches of liquid to
have different trace element contents even though the major element
content remains the same. Depletion schemes of trace elements, espe-
cially the rare earths (see Chapter 8), on partial melting have been
investigated by Schilling and Winchester (1967, p. 277, Figure 9).
The apparent mystery that surrounds trace amounts of elements
stems from the drastic changes in partition coefficients at low concen-

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Physicochemical Constraints on Melting Ill

trations. The partition coefficients must be measured at a series of
concentrations to determine the species in the melt and the activity
coefficients needed to obtain the thermodynamic equilibrium constant.
The thermodynamic equilibrium constant is dependent on such inten-
sive parameters as temperature and pressure, but not on the concentra-
tions of the constituents (Mysen, l976b). The drastic changes in
partitioning between crystals or between crystals and liquid are prob-
ably due to site restrictions in the competing crystal structures. The
common belief that trace elements are not buffered because they are
not a required constituent of a phase does not seem to b~ valid.
Partitioning in itself is a buffering process and when calibrated will no
doubt yield quantitative information on the processes of magma gener-
ation. The observed large variance in the abundances of trace elements
gives a considerable advantage in interpretation over the small vari-
ance in the abundances of major elements in recording the initial stages
of melting. The difficulty in the interpretation of trace element abun-
dances from derived rocks is the same as for major element
abundances-the composition of the parental material is not well
defined.
The differences in isotopic proportions of elements such as strontium
between oceanic island and ridge basalts described by Hedge and
Peterman (1970) and by Hart et a/. (1973) cannot be ascribed to
fractional crystallization or fractional melting. [No evidence for the
fractionation of strontium isotopes in nature has ever been reported
(Faure and Powell, 1972, p. 21).] They must, therefore , be due to
differences in source, perhaps produced at the time of accretion or as a
result of later events. If the earth underwent a melting stage, the long
cooling period may have resulted in a continuum of layers or perhaps a
random aggregation of diapirs having different rubidium contents or
ages of crystallization, or both. The higher 87Srf86Sr in oceanic island
basalts relative to ridge basalts would suggest an older source, which
presumably is also deeper if crystallization of the earth progressed
outwardly from the inside. On the other hand, the difference may
reflect a source for oceanic island basalts with a higher rubidium
content-that is, a source that has not been depleted of its alkalies. In
either case, in the view of Hofmann and Hart (1975), the composition
of the mantle appears to be heterogeneous both horizontally and
vertically. The Pb isotope content of these basalts also supports the
view that the magmas came from sources having separate and indepen-
dent U/Pb systems (Gast et a/., 1964; Tatsumoto, 1966; Sun et a/.,
1975).
For the above reasons, the writer believes that the upper mantle may

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

112 GENERATION OF BASAL TIC MAGMA

be quite heterogeneous in the proportion of phases, but not in the kinds

of phases, so that liquids close to basalt in composition will generally
be the product of the essentially anhydrous melting of garnet peridotite
below about 50 km. The concept of petrographic provinces (Judd,
1886), now well established, suggests that the minor mineral content
(e.g., phlogopite, amphibole, apatite) and trace element content vary
but that the major mineral assemblage at the site of magma generation
is about the same worldwide.

DISAGGREGATION OF PARENTAL MATERIAL

Continued melting will eventually lead to the collapse of the bridging

structure of the crystals. No data are available concerning the dis-
aggregation of parental material under the conditions of load or stress
within the mantle; some limits can be placed on the amount of melting,
however, from theory and field observations at the surface. The
effective viscosity of a suspension of uniform rigid spheres, based on a
relationship provided by Einstein (1906, 1911), was given by Roscoe
(1952) as

TJerr = TJnuld( I - 1.35C)- 2·5,

where C is the concentration of spheres. When C is 0.446, the effective

viscosity has increased one order of magnitude. Without regard for
grain shape or change of viscosity with temperature, it appears, there-
fore, that collapse of an equigranular rock structure begins to take
place when the rock is half melted.
Specimens collected from a drill hole in the lava lake at Makaopuhi,
Hawaii, were measured for their glass content after quenching by
Wright et al. (1968). The data are plotted in Figure 6-2. The crust
formed when the liquid content was reduced to 45 wt%.
These arguments lead to the view that about 45 percent melting would
lead to the collapse of the bridging structure of an undisturbed cystalline
mass. Thus, it appears that considerable melting would be required to
detach a crystal mush as a diapir under undisturbed conditions. The
extent of melting that must obtain before detachment under disturbed
conditions, the more likely circumstance, has not yet been ascertained.

AMOUNT OF LIQUID VERSUS TEMPERATURE

The amount of liquid produced as a function of temperature in a batch

process where the liquid remains in equilibrium with the residuum is

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Physicochemical Constraints on Melting 113

Gloss content, volume percent

FIGURE 6-2 Glass content of samples coUected at various tem-

peratures from Makaopuhi lava lake. Solid line, driU core from crust
of lava lake. Crosses, melt coUected through driU hole and
quenched. Inverted T, pahoehoe toe having maximum temperature
measured with an optical pyrometer. Black dot, liquidus tempera-
ture of pumice measured in the laboratory. Crust formation occurs
when the liquid content is reduced to 50 vol% (= 45 wt%). From
Wright et al. (1968, p. 3199, Figure 12). (With permission. Copy-
righted by the American Geophysical Union.)

shown in Figure 6-3 for the model system forsterite-diopside-pyrope at

P = 40 kbar. This thermal behavior is compared with the modal data of
Scarfe et al. (1972, p. 473) obtained by the point-counting technique on
the melting of natural garnet peridotite at I atm. Spinel instead of
garnet was observed at the lowest temperature (1200°C) investigated,
where the glass content was "nil." The initial stages of melting are
eutecticlike , and an abrupt change in the amount of melting with
temperature takes place upon the loss of the clinopyroxene and or-
thopyroxene.
The melting of a garnet peridotite was also studied at about P = 20
kbar by Ito and Kennedy (1967, p. 524, Table 2), who estimated the
proportion of glass after quenching. A plot of their data is also given in
Figure 6-3. No garnet is reported in any of the runs near 20 kbar, and
unfortunately the data are insufficient at pressures where garnet is
stable in the presence of liquid. The data of Ito and Kennedy may be
interpreted as a continuous smooth curve similar to the liquidus of the
simplified forsterite-fayalite system or with cusps where phases are
consumed. Although the latter interpretation of their data at P = 20

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

114 GENERATION OF BASALTIC MAGMA

2000r-~--,-~r--r--,-~r--r--~-.,-~

1700

/
1600
//
0 //ibar
0
.;
z!:!
~
///
..
Q.
E
1500

~
1400

I0000~~10---2~0--3~0--~40~-5~0--6~0---7L0--8~0---9L0~100

Percent liquid

FIGURE 6-3 Percentage of liquid produced on melting to various

temperatures. Curve at I bar is for a natural garnet peridotite
(determined by Scarfe et al., 1972, p. 473, Figure 1). Curve at 20
kbarwas constructed from the dataoflto and Kennedy (1967, p. 524,
Table 2) on a natural garnet peridotite. Curve at 40 kbar is deduced
from Figure 6-1 : solid curve assumes eutectic behavior, and optional
dashed branch of curve iUustrates quaternary melting behavior.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Physicochemical Constraints on Melting 115

kbar is preferred, the melting is not eutecticlike. Melting related to a
continuous series of solid solutions is implied. The few data obtained
by Ito and Kennedy at P = 40 kbar suggest a return to eutecticlike
melting where almost 20 percent liquid appears as a result of the
consumption of both clinopyroxene and garnet in a narrow temperature
interval.
It is difficult to determine the amount of glass in an experimental run;
nevertheless, such data are most useful in reconstructing the
physicochemical relations for a natural rock. Additional experiments
for the purpose of accurately determining the proportion of liquid as a
function of temperature and pressure are clearly needed. The use of
radioactive tracers, beta-track counting, and the scanning electron
microscope may aid in collecting these data. In this way it may be
possible to reconstruct the phase diagrams applicable to natural rocks
under conditions similar to those at their site of generation. (It is
perhaps not out of place to encourage the prospective experimenter to
choose his natural specimens or synthetic mixtures with due regard for
their characteristics as representative of mantle material that will yield
liquids of basaltic composition, i.e., undepleted.)
Because of the great importance the writer attaches to the eutectic
and eutecticlike behavior of major basaltic magma types, the next
chapter is devoted to the supporting arguments for questions 4 and 5
above.

ZONE MELTING
A nonequilibrium process for accumulation of magma related to frac-
tional crystallization was suggested by Harris ( 1957) and by Shimazu
(1959), based on the principles of zone melting outlined by Pfann (1952,
1966). A liquid layer or partially melted layer is presumed to move
upward in the mantle by simultaneous melting and mixing of the roof
rocks and crystallization and deposition at the floor of the magma. It is
assumed that there is relatively little heat loss until near-surface
conditions are attained, the mantle is relatively homogeneous, there is
no diminution in volume of liquid, and the temperature of the beginning
of melting of the rocks encountered decreases with decrease in pres-
sure. The melting of the cooler roof rocks is accomplished through
convection (cf. convective heat transfer models, Chapter 10) by
adiabatic heat transfer from the hotter base of the molten layer. The
enthalpy of melting required for melting the roof would be slightly less
than the enthalpy of crystallization released at the floor because of the
pressure effect. The effectiveness of the heat transfer is governed by

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

116 GENERATION OF BASALTIC MAGMA

the permeability, if any, and the thermal conductivity of the roof rocks.
A newly formed magma layer would begin to move upward as soon as
the proportion and total volume of melt had reached a sufficient
amount for convection to become effective.
The unique aspect of zone melting is the amount and composition of
liquid that is transferred as the assimilation-deposition process ad-
vances. Consider a partially molten zone 10 km thick (Figure 6-4A), the
top of which is at 200 km depth in a homogeneous upper mantle of
garnet wehrlite. For simplicity , it will be assumed that the garnet
wehrlite consists of FoeJ)i2oPy 20 so that the composition lies simply on
the line joining Fo and the eutectic composition at 40 kbar (see Figures
6-1 and 6-5). In addition, it is assumed that the phase relations of the
system remain unchanged over the pressure interval considered-a
most unlikely circumstance. As the melt zone at T0 moves a small
distance (dz) upward, forsterite is precipitated at the base and with the
liquid e is mixed the new partially melted layer, dz in thickness and
having the bulk composition of the initial material Y. With thorough
mixing, dz is incorporated into the mush, whose bulk composition
becomes Y' as forsterite precipitates. Because forsterite is being re-
moved and only Y added, the liquid e increases in amount as indicated by

A B
SFC

M. Oltcontinuity I FIGURE~ A. Schematic section

z'~ through crust and upper mantle , as-
50 t sumed for simplicity to be of garnet
wehrlite composition, showing the

.•.
E
position of a newly formed partially
melted zone (Z) prior to the onset of
convection and subsequent rise
0
~ 100 Gorntt through the mantle.

.
Wohrllto
& B. Composition-versus-depth dia-
A gram for the simplified system Fo-
0
Cpx,.,Gr,., having a eutectic composi-
tion E. The mantle has the composi-
tion Y (dashed line) prior to zone
150
melting and variable composition
(solid line) after a single pass of zone
melting.

200
··•· Matmo
····•·
z
t Fo Y E

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Physicochemical Constraints on Melting 117

2100

2000

-i

Fo + L
1900 3
"0
•...
1800~
-...
Q
c

0
(")

To ···················~t···········-·······
I ! 1700
I :I
I
I
I :
Fo +ICpx + Gr
I :I 1600
I

Fo
WeiQht percent

FIGURE 6-5 Schematic T -X diagram for the system Fo-Cpx,.Gr00 based on the
Fo-Di-Py system of Davis and Schairer (1965, p. 124, Figure 35) for P = 40 kbar. T0 is
the initial temperature of the partially melted zone rising through the mantle (Figure
6-4.4) by zone melting. f is the composition of liquid in equilibrium with forsterite at T. Y
is the initial composition of the melting zone at the onset of zone melting. Y ' is the
composition of the advanced partially melted zone resulting from the mixing of about 7
percent liquid f with the initial partially melted zone.

the lever rule. The process of melting and crystallization proceeds with
continued increase in volume of the liquid e.
In the meantime, below the base of the displaced melt layer, the
composition of the liquid must change accordingly by precipitating
forsterite. With continued movement upward of the molten zone the
liquid cooling at the base approaches the composition E (Figure 6-4B)
by crystallizing forsterite. At the temperature of E the eutectic liquid
crystallizes all the participating phases in the proportions of the initial

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

118 GENERATION OF BASAL TIC MAGMA

bulk composition Y. Therefore, when the base of the molten layer

reaches a depth of about 170 km the composition of the material
crystallizing out at the temperature of E will be exactly equal to that of
the material being melted at the top of the molten layer; that is, Y, the
original bulk composition. No further enrichment of the liquid is
achieved until the last melt zone is reached.
When the heat losses prohibit the advance of the molten layer, the
accumulated liquid crystallizes, presumably just below the M discon-
tinuity. The last portion of the zone to crystallize has the composition
of the eutectic. The width of the zone of enrichment near the surface is
equal to the width of the initial melt zone. Successive waves of later
melt zones would increase the width of the zone of enrichment.
Zone melting would eventually produce a melt of the eutectic com-
position equal in volume to that produced by batch melting or frac-
tional melting. Unfortunately, the process would be terminated if a
refractory layer was encountered. Furthermore, the initiation of the
rise of a zone requires melting to an extent considerably greater than
that envisaged in the mantle. For maximum efficiency, the rise of the
molten zone must take place at a rate that ensures complete mixing in
the liquid and complete removal of the crystals from reaction with the
liquid. The close approach to equilibrium attained within the times
available for magma generation and movement appears to relegate the
zone-melting process to a minor role. The life of a magma zone 20 km
in thickness generated at 100 km in the upper mantle is of the order of
10 m.y., according to Shimazu (1961, p. 193).

DISEQUILIBRIUM MELTING
O'Nions and Pankhurst (1974, p. 624/f.) suggested that the constituent
minerals of garnet peridotite enter the melt in proportions other than
those determined by equilibrium melting relations. This model was
proposed to account for the variations of strontium isotopes in rocks
obtained from the mid-Atlantic ridge. Because the minerals are as-
sumed to contain different 87 Sr/86Sr, the rate at which they contribute to
the magma would affect that ratio in the magma and its subsequent
fractionation products. O'Nions and Pankhurst attributed the differ-
ences in 87Srf86 Sr observed in rocks derived from what appeared to
them to be a major single magma source under the mid-Atlantic ridge to
disequilibrium melting.
The rapid response of garnet peridotite and eclogite of a wide range
of grain size to melting conditions at high pressures in the laboratory
would rule out such disequilibrium melting. Only one common

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Physicochemical Constraints on Melting 119

igneous-rock-forming mineral that melts with great difficulty under
anhydrous conditions is known to the writer; that is albite, and it is not
anticipated as a phase in rocks of normal basalt composition at high
pressures. If anhydrous albite diabases are indeed an igneous product,
then one might consider them as a possible candidate for disequilibrium
melting, but even then equilibrium melting may be achieved on the
order of several months. There is no laboratory experience yet on the
melting rate of large megacrysts now being recovered with the eclogite
and garnet peridotite nodules from kimberlite pipes. Nevertheless,
using measured diffusion coefficients, a grain I em in diameter in a
partially molten rock would equilibrate by volume diffusion with a melt
in less than a million years (Hofmann and Hart, 1975).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

7 Composition of Major
Basaltic Magmas

FRAMEWORK OF IGNEOUS PETROLOGY

The key to the most abundant basaltic magma compositions is the
composition of the invariant points and related univariant curves
controlling the physicochemical behavior of the natural chemical sys-
tems at each pressure level where magma separation takes place. The
fact that basalts are restricted in mineral content and composition
worldwide indicates that they are not random products of the various
processes described. It is most unlikely that a specific basalt composi-
tion occurs because of the complete melting of a pre-existing mass of
exactly the desired composition or is always the product of a specific
degree of melting. The repetitious extrusion of tholeiitic basalt, for
example, in many successive flows. especially in the ocean environ-
ment, cannot depend on specific flow rates of magma, heat losses, or
proportion of crystals to liquid. There appears to be a memory to
magma generation so that flow after flow of similar composition is
extruded over the world and through large intervals of time. The
principal control on the restriction of magma types is believed to be the
pressure-temperature-dependent phase relations.

BASALT TETRAHEDRON
The concept of a simple model for a wide range of basalt types (Yoder
and Tilley, 1957, pp. 156-161; 1961. pp. 106-113) arose from a detailed
experimental study of natural basalts and simple systems of principal
120

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Composition of Major Basaltic Magmas 121

end-member minerals that make up basalt. The investigation of a
number of relatively simple systems combining the early-crystallizing
minerals of rocks with late-crystallizing alkali aluminosilicates had
established the nature of residual liquids from the crystallization of
relatively complex magmas. The diversity of rocks produced in the late
stages of crystallization appears to have evolved from very similar
source magmas. The basalts, representing these source magmas, con-
sist primarily of feldspar and clinopyroxene with small amounts of one
or more of the minerals olivine, orthopyroxene, nepheline, quartz, and
iron ores. It was important, then, to investigate systems that deal with
these principal phases. Because of the large number of components
involved, it was expedient to limit the composition of the phases in the
preliminary studies. For experimental simplicity, the iron-bearing end
members of these phases were temporarily set aside. The small
amounts of K20, Ti02, and MnO found in the basaltic magmas were
neglected as well, however, without losing significant phases. If these
limitations are applied, the experimental major basalts can be repre-
sented for the most part in a normative tetrahedron comprised of
larnite(La)-nepheline(Ne )-forsterite(Fo)-quartz(Qz), as displayed in
Figure 7-1.
Experimental data on the principal planes were collected mainly by
Schairer and Yoder (1964, pp. 65-74) at I atm (Table 7-1). The system
involves five oxides (Na20-CaO-MgO-AI203-Si02), but a close ap-
proximation to the phase relations can be made by reducing the
effective components to four. An expanded view of the subtetrahedra
is given in Figure 7-2. The relationship of each of these subtetrahedra
to basaltic magma types is given below.
An analogous tetrahedron based on anorthite can also be constructed
(Table 7-2); however, contiguous subtetrahedra are not necessarily
compatible where nepheline is a component (cf Schaireret al., 1968, p.
467, Figure 67). A consistent set of relations can be obtained if, in
addition, nepheline is replaced by the calcium-Tschermak's molecule
CaAI 2Si06 , and the tetrahedron reduces to a portion of Ca0-Mg0-
AI203-Si02 \Schairer and Yoder, 1970). An evaluation of the effect of
plagioclase in place of albite in the expanded basalt tetrahedron may be
gained by examining those systems relating albite and anorthite to the
other end-member components (Table 7-3).

ILLUSTRATION OF PHASE EQUILIBRIA

On the basis of many detailed experiments (Table 7-1), the composi-
tions of liquids in equilibrium with one or more crystalline phases were

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

122 GENERATION OF BASALTIC MAGMA

La

Ne

Mol percent

FIGURE 7-1 Normative basalt tetrahedron of Yoder and Tilley (1962, p. 350, Figure I)
expanded by Schairer and Yoder (1964, p. 65 , Figure I) to include melilite-bearing
assemblages. Ruled planes are close to, but because of solid solution, not exactly the
thermal divides (see flow sheet of Figure 7-6) determined by experiment.

Ab =albite Fo = forsterite Mer = merwinite Ra = rankinite

Ak = akermanite Jd =jadeite Mo = monticellite Sm = soda melilite
Di = diopside La= Iamite Ne =nepheline Wo = wollastonite
En = enstatite Qz =quartz
Related end-member minerals in parentheses are not stable at I atm. (With permission of
Oxford University Press and the Carnegie Institution of Washington.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Composition of Major Basaltic Magmas 123

TABLE 7-1 Systems within the Expanded Basalt Tetrahedron
Studied at I atm, Including Joins Stable Only at High Pressure
System" References
Ab-Ak-Di Schairer and Yoder ( 1964, p. 69, Fig. 5)
Ab-Ak-Ne Schairer and Yoder (1964, p. 70, Fig. 6)
Ab-Ak-Wo Schairer and Yoder (1964, p. 68, Fig. 4)
Ab-Di-En Schairer and Morimoto (1959, p. 115, Fig. 16)
Ab-Di-Fo Schairer and Morimoto (1958, p. 213, Fig. 24)
Ab-Di-Ne Schairer and Yoder (1960a, p. 278, Fig. 2)
Ab-Di-Wo Yoshiki and Yoshida (1952, p. 167, Fig. I)
Ab-Di-Qz Schairer and Yoder (1960a, p. 278, Fig. 2)
Ab-En-Fo Schairer and Yoder (1961, p. 142, Fig. 35)
Ab-En-Qz Schairer and Yoder (1961. p. 142, Fig. 35)
Ab-Fo-Ne Schairer and Yoder (1961, p. 142, Fig. 35)
Ab-Ne-Wo Foster (1942, p. 165, Fig. 6)
Ab-Qz-Wo Not done (two-component joins done)
Ak-Di-Fo Ferguson and Merwin (1919, p. 109, Fig. 12)
Ak-Di-Ne Schairer and Yoder (1964, p. 66, Fig. 2); Onuma and Yagi (1967,
p. 238, Fig. 4)
Ak-Di-Wo Ferguson and Merwin (1919, p. 109, Fig. 12); Schairer and Bowen
(1942, p. 730, Fig. 5; p. 741, Fig. 6)
Ak-Fo-Ne Not done (two-component joins done)
Ak-Ne-Wo Schairer and Yoder (1964, p. 67, Fig. 3)
Ak-Sm Schairer et al. (1965, p. 96, Fig. 17); Yoder (1973. p. 154, Fig. II)
Di-En-Fo Kushiro ( 1972a, p. 1263, Fig. I)
Di-En-Qz Kushiro ( 1972a, p. 1263, Fig. I)
Di-Fo-Ne Schairer and Yoder (1960b, p. 70, Fig. 18)
Di-Jd Schairer and Yoder (1960a, p. 278, Fig. 2)
Di-Ne-Wo Schairer and Yoder (1964, p. 71, Fig. 7)
Di-Qz-Wo Ferguson and Merwin ( 1919, p. 109, Fig. 12)
Fo-Jd Mao and Schairer (1970, pp. 221-222)
•Compositions of the components. arranged alphabetically (symbols as in caption of Figure 7· 1):

Ab: NaAISi,O. Jd: NaAISi,Oa

Ak: Ca,MgSi,O, Ne: NaAISiO,
Di: caMgSi,o. Qz: SiO,
En: MgSiO, Sm: NaCaAISi,O,
Fo: M~oSiO, Wo: CaSiO,

determined. To illustrate the technique for determining the composi-

tion and course of liquids, the Ab-Di-En-Qz subtetrahedron will be
used. Polymorphic changes will be ignored for simplicity. Because of
plagioclase and pyroxene solid solutions, the system is in fact quinary,
but it can be treated for all practical purposes as quaternary. The
subtetrahedron can be laid out from the Di apex, and the ternary
phase-equilibrium data displayed (Figure 7-3). The ternary invariant
points are labeled a-f. The "quaternary" boundary curves and the

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

124 GENERATION OF BASALTIC MAGMA

Wo

Ne

FIGURE 7-2 Subsolidus tetrahedra within the expanded basalt tetrahedra as deter-
mined at I atm by experiment. Arrows indicate course of liquids to determined invariant
points lettered as in Figure 7-6. Because of solid solutions in clinopyroxene, B may be a
reaction point with liquids tending toward G.

invariant points A and B are presented in a generalized diagram (Figure

7-4). Because the "quaternary" invariant point B, where PI, Qz, Enss *,
and Di•• coexist with liquid, lies within the subtetrahedron representing
the four crystalline phases, it is eutectic. On the other hand, "quaternary"
invariant point A, where Fo, PI, En 55 , and Diss coexist with liquid, does
not lie within the subtetrahedron representing its four crystalline
phases; hence, it is a reaction point. The four crystalline phases in
equilibrium with liquid at point A form the contiguous subtetrahedron
(see Figure 7-2). Whereas all proportions of the phases in the subtet-
rahedron Ab-Di-En-Qz begin to melt at the "quaternary" invariant
point B, all proportions of phases in the adjoining subtetrahedron
Ab-Di-En-Fo begin to melt at the "quaternary" reaction point A, the

*The subscript ss refers to solid solutions of the designated phase.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Composition of Major Basaltic Magmas 125

TABLE 7-2 Additional Systems Studied at I atm, Required to Con-
struct an Expanded Basalt Tetrahedron Based on Anorthite.
System References
Ak-An-Di de Wys and Foster (1958, p. 741, Fig. 2)
Ak-An-Fo Yang et at. (1972, p. 167, Fig. 4)
Ak-An-Ne Not done (two-component joins done)
Ak-An-Wo Rankin and Wright (1915, p. 40, Fig. 8)
An-Di-En Hytonen and Schairer (1960, p. 71, Fig. 19)
An-Di-Fo Osborn and Tait (1952, p. 419, Fig. 5)
An-Di-Ne Schairer et at. (1968, p. 469, Fig. 69)
An-Di-Qz Clark et at. (1962, p. 66, Fig. 6)
An-Di-Wo Osborn (1942, p. 761, Fig. 5)
An-En-Fo Andersen (1915, p. 440, Fig. 10)
An-En-Qz Andersen (1915, p. 440, Fig. 10)
An-Fo-Ne Not done (two-component joins done)
An-Jd Mao and Schairer (1970, p. 221-222)
An-Ne-Wo Gummer (1943, p. 515, Fig. 7)
An-Qz-Wo Rankin and Wright (1915, p. 40, Fig. 8)

composition of which lies outside the subtetrahedron representing the

participating phases. Because of the complexity of such four-
component liquidus diagrams and the desirability of examining the
results in conjunction with those of several contiguous tetrahedra, a
scheme was devised to display the course of liquids in a simple way.

FLOW SHEET
The ternary and quaternary invariant points can be displayed in two
dimensions without regard to spatial orientation, as in Figure 7-5. The
lowercase letters represent the assemblages in Figure 7-3, and the

TABLE 7-3 Systems Relating Albite and Anorthite to Other End-

Member Components of the Expanded Basalt Tetrahedron
System References
Ab-An-Ak Schairer and Yoder (1969, p. 104, Fig. 15)
Ab-An-Di Bowen (1915, p. 178, Fig. 12); Kushiro (1972a, p. 1263, Fig. I)
Ab-An-En Not done (two-component joins done)
Ab-An-Fo Schairer and Yoder (1967, p. 206, Fig. 2)
Ab-An-Ne Schairer (1957, p. 232, Fig. 35); see also Yoder (1968b, p. 477,
Fig. 77)
Ab-An-Qz Schairer (1957, P·. 232, Fig. 35); see also Yoder (1968b, p. 477,
Fig. 77)
Ab-An-Wo J. S. Griffith (Ph.D. thesis, University of Chicago, unpublished)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

126 GENERATION OF BASALTIC MAGMA

Di

FIGURE 7-3 The system diopside(Di)-albite(Ab)-enstatite(En)-quartz(Qz) laid out to

show liquidus relations of the boundary systems at I atm. The piercing points and
invariant points are labeled with lowercase letters and also appear in Figures 7-4 and 7-5.
The sources of the data are given in Table 7-1 .

uppercase letters, those in Figure 7-4. The direction offalling tempera-

ture on the univariant curves connecting the invariant points is shown
by arrows. For this reason , the diagram in Figure 7-5 is described as a
flow sheet (Schairer, 1942, 1954).
On the basis of the vast amount of data collected for all the subtet-
rahedra in Figure 7-1 , a flow sheet can be constructed for the petrolog-
ically relevant part of the normative tetrahedron. This flow sheet,
incorporating the example, is given as Figure 7-6. It is seen that the
more primitive liquids, those obtained at the highest beginning-of-
melting temperatures, originate from the assemblages 01-Di-En-Ab,
01-Di-Ab-Ne, and 01-Di-Ak-Ne . The liquids generated at the begin-
ning of melting have compositions that lie outside their respective
subtetrahedra! They are all reaction points and involve reactions of

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Composition of Major Basaltic Magmas 127

olivine with liquid. (The importance of this observation is detailed in
Chapter 8.) The arrowheads in Figure 7-2 indicate the appropriate
subtetrahedron in which the invariant liquid composition lies for crystal-
line phases in the contiguous subtetrahedron at the arrow tails. The
rock names relating to the mineral assemblages are presented in an
identical flow sheet in Figure 7-7, corresponding to the univariant lines
and invariant points of Figure 7-6.
Olivine tholeiite (A), represented by the assemblage Fo-Di-En-Ab,
passes by reaction of olivine with liquid through hypersthene basalt
and eventually to tholeiite (B). This sequence of magmas is separated
from the nepheline basanite (F) sequence by a thermal maximum or

Oi

FIGURE 7-4 Schematic assembly of the diopside(Di)-albite(Ab)-enstatite(En)-

quartz(Qz) system shown in Figure 7-3 to illustrate the probable location of the "quater-
nary" invariant points, A and B . Because of the plagioclase and solid solutions in
pyroxenes that lie outside the tetrahedron. the system is quaternary only to a first
approximation. Polymorphism is neglected for simplicity. "Ternary" invariant points are
marked with triangles, and "quaternary" invariant points with black dots. Some of the
sides of the orthopyroxene volume are stippled to aid in visualizing the three-dimensional
relations.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

128 GENERATION OF BASALTIC MAGMA

·r---
Oi En Fo

·r"...
Oi En Qz

~ 1145!10• A AbOiEn B
01A0 •
Fo AB 01 AB 01
EN FO EN QZ

.1.098tl0"
Ab En Fo
.t......
AbEn Qz

FIGURE 7-5 Flow sheet of the " univariant" and "invariant" liquids in the system
diopside(Di)-albite(Ab)-enstatite(En)-quartz(Qz) displayed in Figures 7-3 and 7-4.
Lowercase letters represent "invariant" points in Figure 7-3; uppercase letters, those in
Figure 7-4. Positions of "invariant" points are similar to those in Figure 7-4; however,
no other spatial relations are to be inferred . Arrows indicate direction of falling tempera-
tures. The temperature of A must be< 1098" ± IO"C > 8, and the temperature of 8 must
be< 1058 ± S'C .

divide at 1 atm (Yoder and Tilley, 1962, p. 398) represented by the

olivine basalts bearing minor occult nepheline or hypersthene. The
nepheline basanites pass by reaction of olivine with liquid through
nepheline tephrite and eventually to wollastonite-nepheline tephrite
(E). This sequence of magma is separated from the olivine-melilite
nephelinites by another thermal maximum represented by olivine
nephelinites. The olivine-melilite nephelinites (C) pass by reaction of
olivine with liquid through melilite nephelinite and wollastonite-
melilite nephelinite (D), and by reaction of melilite with liquid eventu-
ally to wollastonite-nepheline tephrite (£).

"PARENTAL" MAGMAS
It would appear that three "parental" magmas (olivine tholeiite,
nepheline basanite, and olivine-melilite nephelinite) are required to
derive the magma sequences observed at 1 atm. As will be shown
below, these "parental" magmas can be derived from a single parental
material by various means involving high pressure with or without
volatiles. The important observation, and the main purpose in present-
ing this detailed argument, is that the major magma types observed at

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Composition of Major Basaltic Magmas 129

(Ab-Fo-Ozl lAb- Fo-Ne) (Ak·DI·Ne)

c ~[Link]• -·[Link]· -~[Link]•

01 Diu Neu Diu Diu Diu

PI Ot 01 ... Mel Mel

prPr
0
p;~:u 0/Ntu
01
(Di·Fo-Ozl Diu Diu Diu Mel
[Link]. 01 A 01 F Ness Cll45tl0" Neu ,1178:t5•
b Pr PI
[Link]>-
01 01
(~k-DI-Ne)
Diu
PI Diu
:~ ~:r
Pr
/
.....
[Link] 5" PJ
-,· Pr ~· [Link]·
fI
(Ab-Fo-Oz)
Tr •
Tr (Di-Fo-Ozl
Diu
PI
Tr
d [Link].
Diu
~.,
/ .....
Diu Neu

~0
PI
WO
~0
N.
Temp?~: Diu
PI
Diu
E950:t5• Neu 0[Link]" M:f [Link]·

(Ab·Oz~Wo) Wo wo Wo Wo 1
[Link]• (Ak-Di-Wol

Diu Neu Mel

Tr PI Neu
wo wo wo
__ [Link]• -~[Link]• ~ 1171 :1:5·

(01- Oz·Wo) (Ab·Nt·Wo) (Ak·Ne ·Wo)

FIGURE 7-6 Flow sheet determined experimentaUy for the petrologicaUy important
portions of the expanded basalt tetrahedron in Figure 7-1 (after Schairer and Yoder,
1964, p. 72, Figure 8; 1967, p. 209, Figure 5). Abbreviations as in Figure 7-1; also, Mel=
melilite; PI = plagioclase; Pr = protoenstatite; Tr = tridymite; ss = solid solution.
Uppercase letters are quaternary "invariant" points. Lowercase letters are ternary
.. invariant" points from Figure 7-5. (With permission of the Carnegie Institution of
Washington.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

130 GENERATION OF BASALTIC MAGMA

Olivine Olivine
Norite Melililite

Olivine Nepheline Olivine-melilite

Lherzolit A Tholeiite Olivine F Basanite Olivine C Nephelinite
Bosolt Nephelinite

Hypersthene
Basalt

Melilite
B Tholeiite Nepheline Nephelinites
TIJ).hrite

Wollastonite -nepheline Wollastonite -melilite

G E Tephrite VtbHostonite D Nephelinites
Nephelinites

FIGURE 7-7 Rock nomenclature diagram corresponding to the flow sheet in Figure
7-6. For the most part, names apply to extrusive rock types. Invariant point G may be
represented in nature by some metamorphic rocks. The reaction points A. F. and C
involve loss of olivine. From Schairer and Yoder (1964, p. 73, Figure 9). (With permis-
sion of the Carnegie Institution of Washington .)

the surface ofthe earth are closely* represented by "invariant"t points

or "univariant" curves. They obviously are derivative liquids con-
trolled mainly by the phase-equilibrium relations. It is concluded that:

*The normative proportions of Cenozoic basalts determined by Chayes (1972, p. 293,

Figure 4) support this thesis. It is important to document these conclusions further with
modes of basalt in the same manner as has been done with granites (Tuttle and Bowen.
1958. p. 75, Figure 38). The electron microscope is now a convenient tool for such a
tedious but rewarding investigation of the fine-grained basalts.
tThe assemblages labeled as " invariant" actually melt over a small range of tempera-
ture. Such points are not accurately determined because of the difficulty of observing the
first appearance of liquid in a four-phase assemblage. To the best of the writer's
knowledge, the greatest range of melting at the "invariant" points is displayed by C , for
which Schairer et a/. (1965, p. 97, Figure 29) observed a 40"C interval. That is.
Di.. + OJ + Mel + Ne,. coexisted with liquid as a univariant assemblage throughout
that temperature interval.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Composition of Major Basaltic Magmas 131

I. Igneous rocks are not random products of the melting of varying
mineral assemblages. In short, the flow sheet is the fabric of igneous
petrology and demonstrates that there is a rigorous control on the
generation of basaltic and nephelinic magmas.
2. The magmas that arrive at the surface have been equilibrated for
the most part at successive levels on the way to the surface or have
achieved equilibrium at near-surface conditions, either along conduits
or in auxiliary reservoirs.
3. The small number of invariant points at I atm involving olivine in
the 5-component system described, as well as in the systems K2 0-
Mg0-Aiz0a-Si0z (Schairer, 1954), Ca0-Fe0-AI 20 3-Si02 (Schairer,
1942), CaO-MgO-AI203 -Si02 (Schairer and Yoder, 1970, p. 212), and
Ne-Di-An-Fo-Ak (Schairer eta/.. 1968, p. 470), is a strong argument
for believing in a single unique assemblage, or at least not more than
two, as potential parental material for the generation of basalts. In
other words, in spite of the great range of compositions studied, the
number of beginning-of-melting points is exceptionally limited, and
these points, based on the abundance of basalt, are presumed to be close
in composition to the liquid(s) that yields basalt.

NEED FOR FLOW SHEET FOR HIGH PRESSURES

There is a clear need to ascertain the flow sheet for the same array of
bulk compositions at a series of pressures such asP = 10, 20, 30, and
40 kbar. * Only a few data are available, however, and there is as yet no
method for dealing adequately with large ranges of solid solution. (The
system Na2 0-Ca0-AI 20 3-Si02 can be illustrated by a flow sheet only
if it is assumed that albite and anorthite are separate, independent
phases-an obvious example of the shortcomings of the flow sheet for
systems involving continuous solid solution.) The direction of shift in
composition of some of the invariant points with pressure has been
indicated. For example, the invariant point for the assemblage Fo-
En-Ab (on the base of the iron-free normative basalt tetrahedron)
shifts with increasing pressure from quartz-normative through
hypersthene-normative to nepheline-normative compositions (Figure
*There are no theoretical reasons for choosing these pressure levels except that impor-
tant reactions in the synthetic and natural systems investigated occur near them. Because
of the advantages of correlating the work of many laboratories around the world, it is
hoped that these pressure levels, within the convenient working range for various
apparatus available today, will be found acceptable to the majority of investigators. The
exciting opportunities for research under conditions up to ., Mbar and 2500°C, attainable
in the diamond-anvil, laser-heated pressure cell (Bell and Mao, 1975; Mao and Bell,
1976), will no doubt lead to other preferred pressure levels as the important reactions are
identified.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

132 GENERATION OF BASALTIC MAGMA

7-8). A single bulk composition could therefore produce a wide array of

basaltic liquids, as suggested by Yoder and Tilley (1961, 1962), pro-
vided the liquids were separated at different pressures. In general, it
appears that the nepheline-normative basaltic magmas are derived at
greater depths than the hypersthene or quartz-normative magmas.
Volatiles produce similar effects, depending on their composition.
For example, Kushiro (1972b, p. 316, Figure 3) demonstrated that at
P = 20 kbar, H20 shifts the same invariant point described above to
more siliceous compositions, whereas Eggler (1974, p. 216, Figure I)
showed that C02 shifts the composition of that invariant point to more
nephelinitic compositions (Figure 7-9).

Forst.-lte
Weloht percent

FIGURE 7-8 Shift of the forsterite (Fo)-enstatite,.. (En,.) liquidus boundary and in-
variant points (f) with pressure in the forsterite-nepheline-quartz system. Data at I atm
(solid line;/0 ) from Schairer and Yoder (1961, p. 142. Figure 35). Estimates (dashed lines)
at 10. 20. and 30 kbar from Kushiro ( 1968. p. 625. Figure 4). Tie lines are illustrated for 30
kbar. (With permission of the Carnegie Institution of Washington and the American
Geophysical Union .)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Composition of Major Basaltic Magmas 133

Nepheline

Enstatite 11

Fortterite Enstatite Quartz

Weight percent

FIGURE 7-9 Shift of the forsterite (Fo~nstatite.. (En..) liquidus boundary and in-
variant points with H 20 and C02 from volatile-absent (Dry) positions at 20 kbar in the
forsterite-nepheline-quartz system. From Eggler (1974, p. 216, Figure 1). (With permis-
sion of the Carnegie Institution of Washington.)

MAGMA SEPARATION
It is essential, therefore, to specify not only the depth of separation of
magma but also the amount and kinds of volatiles present in the magma
in order to fix the probable magma type and eventual trend of deriva-
tive magmas. It is not useful to attempt at this time to outline the
possible flow sheets at various pressures without considerably more
data than are now available. Comprehensive studies at high pressures
are now under way, and a new array of problems is emerging. One of
these problems involves the presence or absence of the key phase
olivine, and because of its importance full discussion is given in the
next chapter.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the
8 Parental Material

OLIVINE FROM INITIAL MELT

In Chapter 2 arguments were presented for believing that olivine is the
dominant phase in the parental material of basaltic magmas. If olivine
is in the parental assemblage and the system is eutecticlike, then liquids
removed at the beginning of melting should reprecipitate olivine on
cooling. The liquid removed from the eutectic at I atm of the
forsterite-diopside-anorthite system, presented in Figure 2-3, would
crystallize olivine as well as anorthite and diopside. Similarly. in the
forsterite-diopside-pyrope system presented in Figure 5-6, the liquid
removed from the "eutectic" at 40 kbar would crystallize olivine as
well as clinopyroxene and garnet. If, on the other hand, the system is at
a pressure where the beginning of melting is a reaction point in which
olivine is the phase reacting with liquid, then the liquid removed from
that reaction point will not crystallize olivine at temperatures below
that of the reaction point. Furthermore, the absence of olivine in the
beginning-of-melting assemblage at one pressure does not preclude its
appearance from similar bulk compositions at another pressure.

ABSENCE OF OLIVINE IN BASALTS AT HIGH

PRESSURES
In the light of these principles, the results obtained on the reported
behavior of natural basalts at the high pressures where they are
presumed to have originated or separated from their parental material
134

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 135

are surprising. The melting relations of an olivine tholeiite were studied
up to 30 kbar by Thompson (1972, p. 409, Figure 24), who found the
liquidus phase to be clinopyroxene above 10 kbar and garnet above 32
kbar (Figure 8-1). No olivine was found in any of the assemblages at
any temperature at pressures above 13 kbar! In another olivine
tholeiite, investigated by Cohen et a/. (1967, p. 496, Figure 4), the
liquidus phase was clinopyroxene above 10 kbar and garnet above
about 28 kbar (Figure 8-2). No olivine was observed in any of the
assemblages above approximately 10 kbar! A nepheline-normative
olivine basanite subjected to a range of conditions by Arculus (1975, p.
514, Figure 74A) had olivine on the liquidus up to only 21 kbar (Figure
8-3). The highest pressure at which olivine was observed in any
assemblage was 26 kbar! Other less complete studies support the
observation that olivine is absent in assemblages obtained from natural
basalts at high pressures. Eclogites, the high-pressure equivalent of
basalt according to Yoder and Tilley (1962), also fail to produce olivine
at high pressures.
If olivine is an essential liquidus phase in the generation and fraction-
ation of basaltic magma, then the above data would appear to put
some severe constraints on its depth of origin or separation. The
olivine tholeiite from New Mexico would have to originate at depths
less than 30 km at temperatures between 1200° and 1250°C. Separation
of an olivine tholeiite liquid might be appropriate under such conditions
in the midocean ridge environment, but its generation at 30 km is not
considered a likely event under continental or abyssal oceanic areas.
The basanite from Australia would have to originate at depths less than
75 km at temperatures between 1225° and 1325°C. It would be difficult
to explain a genetic relationship of eclogite and garnet peridotite
nodules to their host basanites and alkali basalts if the depth of origin
was limited to 75 km. The nodules could be unrelated to the basaltic
liquids and merely samples of shallower and cooler parts of the mantle
ripped off as the magma erupted. The high temperatures and high
pressures deduced from the chemical composition of coexisting miner-
als in the nodules and the chemical relationships between the nodules
and the lavas do not support that view (Beeson and Jackson, 1970).
At least three other possible conclusions may be drawn from these
observations on basalts and eclogite:

I. Basalts and eclogite are not derived from an olivine-bearing

parent.
2. Basalt magmas are altered in composition on the way to the
surface.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

136 GENERATION OF BASAL TIC MAGMA

40

Gr+L

35
J •
Cpx+Gr+Oz+So ?/

30
I •

...
~ 25
.&l
.2
~
.;
...
......
~

20
!
0..

15 Liquid

10

1100 1200 1300 1400 1500 1600

Temperature, •c
FIGURE 8-1 Pressure-temperature diagram for an olivine tholeiite
from Snake River, Idaho, determined by Thompson (1972, p. 409,
Figure 24). (With permission of the Carnegie Institution ofWashington.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 137

40

35

30

i:3
:: 20
e
Cl.

15

10

1100 1200 1300 1400 1500 1600

Temperature, "C
FIGURE 8-2 Pressure-temperature diagram for an olivine
tholeiite from San Juan Mountains. New Mexico, determined by
Cohen et al. (1967. p. 496, Figure 4). (With permission of the
American Journal of Science .)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

138 GENERATION OF BASALTIC MAGMA

40

35

30

0
...
SJ 25
E
=-
•...
~

~
..•• 20

15

10
OI+Cpx+Sp+PI+L

1300 1400
Temperature, "C

FIGURE 8-3 Pressure-temperature diagram for a basanite from

Mt. ShadweU. Victoria, Australia, determined by Arculus (1975,
p. 514, Figure 74A ). (With permission of the Carnegie Institution of
Washington.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 139

3. A reaction relationship in the melt, wherein olivine is consumed,
occurs at high pressures.

MELTING RELATIONSHIP OF BASALT TO

GARNET PERIDOTITE
If basalt is indeed the partial melting product of garnet peridotite, then
the liquidus of basalt, after conversion to eclogite, should be clos~ to
the solidus of the parental material. Liquidus results for an array of
basalts are compared with solidus results for an array of peridotites in
Figure 8-4. Some of the peridotites would be appropriate parental
material for some of the basalts; others, however, are 50-100° apart.
The discrepancies, aside from experimental difficulties, may indicate
that:

I. The basaltic liquid fractionated after separation from its source,

and the liquidus temperature is now lower.
2. Certain compositional characteristics, such as the iron-
magnesium ratios, are not necessarily appropriate for an equilibrium
relationship.
3. Volatiles in appropriate concentrations may have been present in
either the peridotite or the basaltic liquid, or both, and may have
escaped.
4. The natural materials are sufficiently altered to cause spurious
results.

MELTING RELATIONSHIP OF ECLOGITE TO

GARNET PERIDOTITE
Because bimineralic eclogites form one binary join in quaternary garnet
peridotite, another test may be devised to ascertain the relationship of
eclogitic, hence basaltic, magmas to garnet peridotite. If the garnet
peridotite system is eutecticlike, then eclogite should begin to melt at a
higher temperature than garnet peridotite. On the other hand, if olivine
or orthopyroxene, or both, exhibit a reaction relation with liquid, then
eclogite should begin to melt at a lower temperature than garnet
peridotite. It is necessary, therefore, to examine the melting relations
of eclogite in detail.

MELTING OF ECLOGITE
The compositions of natural eclogites are confined to a narrow band
in the Af-C-Fm diagram (Figure 8-5). These data are interpreted to

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

140 GENERATION OF BASAL TIC MAGMA

01- Tholeiite7J
F/Mc0.60

40
Gr-Peridotite-:
F/M•0.15 : IIft
...
Ecl~ite0
F/M• .68 _.· t
~
Cl
s:a
01-Tholeiite--;/ .
F/M•0.67 ( hI Gr6o Cpx4o
F/M•0.46
~
~
.; 30
0
"'
%: I
f . I

;r·h/.
I
~
"':> £.
......=- ~
0
l (/)

j
20
I ~~
I
I
f/1
I I fl I

10
' \
\
\
/ / 'f-Gr-Lherzolite

I
I F/M•O.I5

\
I
\
I
0
500 1000 1500 2000
Temperature, •c

FIGURE 8-4 Comparison of the liquidus of eclogites and basalts converted to eclogites
with the solidus of possible parental materials (garnet peridotite, garnet lherzolite, and
spinel lherzolite). FIM = FeO/(FeO + MgO)
Sp-lherzolite: Kushiro et a/. ( 1968a) Gr110Cpx..,: Ito and Kennedy (1974)
Olivine tholeiite: Cohen et a/. (1967) Olivine tholeiite: Thompson (1972)
Gr-peridotite: Ito and Kennedy ( 1967) Gr-Cpx: O'Hara and Yoder (1967)
Eclogite: Yoder and TiUey ( 1962) Gr-lherzolite: Kushiro ( 1973b).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 141

CoO MgO+FeO+MnO(+ Ti<lz)

Mol percent
FIGURE 8-5 Projection of analyzed eclogites and their constituent minerals on the
Af-C-Fm diagram (after Yoder and Tilley, 1962, p. 476, Figure 36).

G, G' , g: clinopyroxene, garnet, and rock of Glenelg, Scotland, eclogite.

D, D ' , d: clinopyroxene, garnet, and rock of Loch Duich, Scotland, eclogite.
S, S', S", s: clinopyroxene, garnet, kyanite, and rock of Silberbach, Germany,
kyanite eclogite .
H, H' , H" , h: clinopyroxene, garnet, orthopyroxene, and rock of Salt Lake crater,
Oahu, hypersthene eclogite. (The arrow from H" indicates composition adjusted for
admixed clinopyroxene.)
X= composition of rock for which constituent minerals were analyzed.
Circle with number= composition of basalt analyzed by Yoder and Tilley (1962,
p. 361-362, Table 2).
Circle labeled Haleakala = composition of basalt from Yoder and Tilley (1962, p. 417,
Table 23, No. 4).

(With permission of Oxford University Press.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

142 GENERATION OF BASAL TIC MAGMA

define the boundary curve between the two major phases in the rock,
as illustrated in Figure 8-6. Their close proximity to such a boundary
curve suggests that eclogites are themselves derivative products of a
more primitive material, presumably garnet peridotite. The principal
question then becomes: How are the two major phases of garnet
peridotite, olivine and orthopyroxene, eliminated so that its partial
melt crystallizes only garnet and clinopyroxene (i.e., eclogite)?

ORTHOPYROXENE REACTION RELATION

A study of the diopside-pyrope system at 30 kbar by O'Hara and

Yoder (1967, p. 74, Figure 3) provided an explanation for the elimina-
tion of orthopyroxene (Figure 5-5). At about 1620°C, orthopyroxene
reacts with liquid to form garnet and clinopyroxene. The relationship is
better displayed in isothermal sections at nearby temperatures (Figure
8-7). The switching of the tie lines from orthopyroxene and liquid at
1630°C (Figure 8-7A) to those connecting coexisting clinopyroxene and
garnet at 1615°C (Figure 8-7B) demonstrates the reaction relationship.
Experiments using natural minerals yielded the same results. In addi-
tion, the reaction relationship persists to 40 kbar, according to Davis
(1964, p. 167, Figure 60). These results indicate that the separation of
volatile-free magmas of basaltic composition from garnet peridotite
would have to take place at depths where the pressure is less than
about 50 kbar. At and above that pressure, the garnet peridotite
system appears to be clearly eutectic, and olivine would crystallize out
of separated liquids.

FIGURE 8-6 Schematic interpreta-

tion of melting of garnet and
clinopyroxene solid solutions iUus-
trated in Figure 8-5. X = bulk compo-
sition. L, and L. are projections of
the liquid compositions at tempera-
tures T, and T2 , respectively, where
T, < T2 • Pressure constant. From
Yoder and TiUey (1962, p. 505, Fig-
ure 47). (With permission of Oxford
University Press.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 143

Alz03
-Groaularite

B
.... .........
........Cpx+Opx
...............
FIGURES-7 Isothermal sections fora part ofthe CaSi03-MgSi03-
A~03 plane at 30 kbar. A, 1630"C. B. 16J5•c. Switching of the tie
lines for orthopyroxene + liquid at 1630"C with the tie lines for
clinopyroxene + garnet at 16J5•c indicates that the reaction or-
thopyroxene + liquid = clinopyroxene + garnet has taken place.
From O'Hara and Yoder (1967. p. 75, Figure 4). (With permission
of Oliver and Boyd.)

OLIVINE REACTION RELATION

The search for the olivine reaction relationship in garnet peridotite

assemblages was unsuccessful for several years until a study by
Kushiro (1968, p. 629, Figure 8) of a composition on the join
enstatite-Ca-Tschermak's molecule, pursued for different purposes,
exhibited an olivine reaction relationship in the pressure range 17-26
kbar. Because of the importance of this reaction to the formation of
olivine-free eclogite, a wider range of compositions was investigated by
Kushiro and Yoder (1974, p. 266-267). The change of the liquidus
surface of a portion of the MgSiOa-CaSi03 -A I203 system with pres-
sure was determined (Figure 8-8). The liquidus field of olivine at 20
kbar in the chosen plane, its presence clearly establishing the reaction
relation, diminishes in size with increasing pressure. At 30 kbar only
the orthopyroxene reaction relationship is exhibited. Furthermore,
Davis and Schairer (1965, p. 124, Figure 35) showed that the system

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

144 GENERATION OF BASAL TIC MAGMA

Alz~

FIGURE 8-8 Change of liquidus surface with pressure for a por-

tion of the CaSi03 -MgSi03 -Ait03 system. Presence offorsterite on
liquidus at P = 20 and 26 kbar implies a reaction relationship of
forsterite with a Si01 -rich liquid. Field of pigeonite deleted for sim-
plicity. Data from O'Hara and Yoder (1967), Kushiro (1968). and
Kushiro and Yoder (1974, p. 268, Figure 39). (With permission of
the Carnegie Institution of Washington.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 145

forsterite-diopside-pyrope was close to eutecticlike at 40 kbar. The
rarity of olivine-bearing eclogite nodules whether or not genetically
related, supports the view that few magmas are derived from depths as
great as 130 km where garnet peridotite becomes eutectic and olivine
may precipitate from separated liquids.
A possible origin of eclogite from garnet peridotite is elucidated by
these experiments. If allowances are made for the differences in
composition between natural magmas and the simplified systems
studied, magmas formed by the partial melting of garnet peridotite
and separated at depths less than about 90 km ( = 28 kbar) will exhibit
both the olivine and the orthopyroxene reaction relationships. The
liquids so produced may yield bimineralic eclogite cumulates or
continue to shallower depths and yield olivine-bearing basalts. Fur-
thermore, eclogites would melt at lower temperatures than garnet
peridotite at pressures less than 28 kbar and at higher temperatures
above that pressure, where a eutecticlike relationship exists in the
garnet peridotite. O'Hara and Yoder (1967, p. 101), for example,
observed that at 30 kbar, eclogite melts at a higher temperature than
garnet peridotite and concluded that the garnet-clinopyroxene as-
semblage constitutes a thermal divide between forsterite-bearing as-
semblages and quartz-bearing assemblages. The sequence of reactions
remains to be shown; that is, does orthopyroxene or olivine react ~ut
first with decreasing pressure? If olivine reacts out first in the CaO-
MgO-Al203-Si02 system, then the liquids will become enriched in
silica directly, the composition of the liquid passing to the Si02-rich
side of the plane Cpx-Opx-Gr (see Figure 2-11). If orthopyroxene
reacts out first, the initial trend is toward larnite-normative liquids, the
composition of the liquid passing to the larnite-rich side ofCpx-01-Gr.

BASALT COMPOSITION CHANGES DURING RISE

The failure of olivine basalts to retain olivine when converted to
eclogite appears to confirm an olivine reaction relation. Another expla-
nation, however, may account for the absence of olivine. Suppose a
liquid of basaltic composition is derived from a parent composed
predominantly of olivine; then olivine should crystallize from the
separated liquid. As the liquid rises to the surface, the proportions and
composition of the precipitating phases may be expected to change in
response to the drop in pressure and temperature. On the basis of the
results on simple systems (discussed immediately below), olivine is the
principal phase to precipitate along with other phases if the magma is to
arrive at the surface with a eutecticlike composition and contain

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

146 GENERATION OF BASAL TIC MAGMA

phenocrysts. That is to say, the magma has no superheat. The magma

that appears at the surface is therefore not the composition originally
in equilibrium with the parent but is a derivative. This point has been
stressed by O'Hara (1965; 1968a, p. 686) and is his principal objection
to considering tholeiitic rocks as representative of primary magmas.
On these grounds it is necessary to restore the crystals lost on the way
to the surface before the relation of basalt to its presumed parent,
garnet peridotite, can be ascertained.
In this light, the failure to observe olivine in the high-pressure
assemblage experimentally obtained from basalts crystallized at the
surface can now be examined with the aid of two simple diagrams used
by Yoder and Tilley (1962, p. 500, Figures 44a and 44c) to illustrate the
variation of fractionation courses at high and low pressure. In Figure
8-9A are given the system Ne-Fo-Si02 and the joins at I atm and 30
kbar. If a parental liquid X were fractionated at successively lower
pressures, the final composition would probably be in a position such
as f. Because the forsterite field expands continuously with decreasing
pressure, only forsterite crystallizes from the liquid. The parental
assemblage would be Fo + En + Jd, and the derivative assemblage at
low pressure would be Fo + En + Ab. The recrystallization of Y at 30
kbar would yield the assemblage En + Jd + Qz; that is, no Fo would
be observed. The system Fo-CaTs-Si02 , shown in Figure 8-9B, can be
treated in a similar way. The parental assemblage X gives way at low
pressures to Fo + En + An after fractionating to Y. Recrystallization
of Y at high pressures results in an assemblage devoid of Fo. The loss of
olivine in basalt at high pressure could therefore be explained by the
predominance of plagioclase over olivine + hypersthene, the olivine
being consumed at high pressures initially through the reactions of
An + Foand An + En, investigated by Kushiroand Yoder(l966), and
of Ab + Fo, investigated by Yoder and Tilley (1962).

OLIVINE-INCOMPATIBLE OR COMPLEMENTARY?
It is easily seen that the addition of Fo to the composition Y in both
cases presented in Figure 8-9 must be done with considerable caution.
It must first be demonstrated that Y was indeed fractionated from a
Fo-bearing assemblage. The pyrolite model of Ringwood (1962 and
subsequent versions) was devised by adding approximately three or
four parts dunite to one part basalt on the assumptions that the dunite
was the only residuum and that the basalt was the complementary
primary magma from the partial melting of an olivine-bearing parent
(i.e., pyrolite). The derivative nature of basalts (and eclogites) has

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 147

CoTs
Ne (Grossulorilt)

1 aim
30 Kb
8

Fo En En
(Pyropt)

FIGURE 8-9 A. The joins in the Forsterite (Fo)-nepheline (Ne)-silica (Si02 ) system at
I atm and 30 kbar (after Yoder and TiUey, 1962). Point X represents schematically the
composition of a critical liquid formed at high pressure, and Y represents its derivative
liquid at low pressure, assuming removal of olivine only. The small-dash line is a
construction line for determining the course of liquid by removal of olivine.
B. The joins in the forsterite(Fo)-calcium-Tschermak's-molecule(CaTs)-silica(Si02 )
system at I atm and 30 kbar after Yoder and TiUey ( 1962). Liquid compositions X and Y
bear the same relationship as in A . End members of the garnet join (long-dash line)
grossularite-pyrope project from Al 20, on CaTs and En, respectively, in the Ca0-
Mg0-AI203-Si02 system. From Yoder ( 1974, p. 265, Figure 37A and 378). (With permis-
sion of the Carnegie Institution of Washington .)

already been demonstrated by Yoder and Tilley ( 1962) and emphasized

by O'Hara (1968b). Thus, the addition of olivine to a derivative rock
with which it may not be compatible at high pressures does not appear
to be a suitable model for deriving primary magmas.*

EVIDENCE FROM NODULES

That olivine and orthopyroxene are the principal phases in nodules
brought up by some alkali basalts supports the belief that these
minerals are important phases in the parental material of basalts. In
addition, Huckenholz (1965, 1966, 1973) observed relic orthopyroxene
in the nepheline-bearing olivine basalts and basanites of the Hocheifel
area, Germany. The preservation of the nodules and relics suggests that
*These arguments do not conflict with a model based on the addition of olivine to an
olivine-bearing, quartz-normative tholeiite derived by the separation of magma from a
plagioclase peridotite at low pressures ( = depths < 15 km) where the olivine reaction
relation is operative.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

148 GENERATION OF BASALTIC MAGMA

such magmas were brought directly to the surface from the mantle
source region. The nodules contain at least one phase, orthopyroxene,
with which the host rock is incompatible at low pressures. In high-
pressure experiments Kushiro (1964, p. 110, Figure 27B) found that
enstatite and nepheline can coexist; therefore, a nodule with or-
thopyroxene could be compatible with a nepheline-normative liquid at
certain depths.* The magma was effectively separated from most of the
parental material at high pressures, at temperatures below both the
olivine and orthopyroxene reaction points, or at the latter reaction
point.
The olivine-free eclogite nodules are probably derived as cumu-
lates in the 60-85-km region where the required reaction relationships
exist under anhydrous conditions, according to the experiments of
Kushiro ( 1968). Alternatively, some eclogites could be recycled deriva-
tive basalts. For example, the metamorphosed oceanic basalts in a
subducted plate might be an adequate source for those eclogites.

FIRST PHASE CONSUMED IN MELT

With the data now presented, it is appropriate to return to the question
of which of the two least abundant major phases in garnet peridotite is
consumed first in a partial melt. It was observed in Figures 8-2, 8-3, and
84 that clinopyroxene was on the liquidus of basaltic rocks at inter-
mediate pressures and garnet was on the liquidus at the highest
pressures investigated. These observations are supported by the shift
of the liquidus boundary between 30 and 40 kbar for the diopside-
pyrope system (Figure 8-10): the boundary curve (Figure 8-6) shifts
toward clinopyroxene with increasing pressure. For a bulk composi-
tion with clinopyroxene and garnet in the proportions Cpx50Gr50-
*The relationship was believed to have been confirmed in high-pressure experiments on
synthetic compositions simulating alkali basalt (Green and Ringwood, 1967), picritic
basanite, and picritic nephelinite (Bultitude and Green, 1971), respectively, having
nonnative nepheline equal to 2.2, 9.4, and 10.6 wt%. However, a natural basanite
(Ne = 10.8 wt%) studied by Ito and Kennedy (1968, p. 186, Figure 3), two natural
basanites (Ne = 7.8 and 11.9 wt%) investigated by Arculus (1975), and a synthetic
composition simulating olivine nephelinite (Ne = 15.3 wt%; Bultitude and Green, 1971)
did not yield orthopyroxene at high pressures. The effects of bulk composition, H 2 0 and
CO. from parts in the high-pressure apparatus, iron loss to the sample container,
oxidation, and metastability may account for these differences. The appearance of
orthopyroxene in nepheline-nonnative compositions at high pressures was erroneously
attributed to the presence of H.O by Bultitude and Green (1971), contrary to the results
on simple hydrous systems. Eggler ( 1973) demonstrated that the effect was due primarily
to the presence of co•. Experiments on a series of natural nepheline-nonnative rocks
under more carefully controlled conditions are warranted.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 149

Pyrope
Wtioht percent

FIGURE 8-10 The liquidus of the diopside-pyrope system at 30 kbar (after O'Hara and
Yoder, 1967, p. 74, Figure 3), ignoring a field of enstatite, and at 40 kbar (after Davis
1964, p. 167, Figure 60). The liquidus phase at and near pyrope at 30 kbar is spinel. The
shift of the coprecipitation curve for garnet and clinopyroxene with pressure is illus-
trated. (With permission of Oliver and Boyd and the Carnegie Institution of Washington.)

Cpx35Gr65 , it is apparent that at the lowest pressures where garnet peri-

dotite fonns a stable assemblage, garnet is consumed first, and with
increasing pressure clinopyroxene is consumed first, with partial melt-
ing. These shifts in the boundary curve have considerable influence on
the normative character of the liquid. Some of the principal normative
minerals in an analyzed clinopyroxene and garnet from a garnet
peridotite are plotted in Figure 8-11. The shift in normative character
of a potential melt resulting from changing proportions of
clinopyroxene and garnet in the liquid is evident. With increasing
garnet contribution to the melt, for example, the normative
hypersthene and anorthite content increases, whereas the normative
diopside content decreases within a reasonable range such as
Cpx7oGrao-CPXaoGr7o (wt%).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

150 GENERATION OF BASALTIC MAGMA

FIGURE 8-11 Norms calculated for mix-

tures of analyzed coexisting clinopyroxene
and garnet from a garnet peridotite from
South Africa (O'Hara and Yoder, 1967.
p. 104, Figure 9). C = corundum.

RARE-EARTH ELEMENT DATA

MEASURE OF DEGREE OF PARTIAL MELTING?

The rare-earth element analyses normalized to the average of 20 chon-

drites (see Appendix) for three basalt types in Hawaii are plotted in Fig-
ure 8-12. Similar rare-earth element patterns were obtained by Arculus
and Shimizu(1974, p . 558, Figure 211)fora suite ofbasanitoids and alkali
olivine basalts ranging in normative composition from 8.99 percent Ne
to 4.80 percent Hyp. These curves were interpreted by Schilling (1971)
to be indicative of the degree of melting, the top curve representing a
small amount of partial melting and the bottom curve, a more advanced
stage of partial melting. To accommodate the major element oxide
compositions and thermal divides, however, Schilling (1966, p. 79) also
indicated that the magmas would have had to form at different depths or
pressures. These conclusions must be tempered with the assumption
that the composition of the liquid from which the rocks crystallized did
not change en route to the surface. An examination of the normative
character of these three rocks does not support the view that they were
produced by varying degrees of partial melting from a single parent at

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 151

the same pressure. The critical normative contents in weight percent
are:
Normative Component
Rock 01 Hyp Ne La
Melilite nephelinite 18.35 - 21.58 7.40
Ankaramite 23.88 - 3.12 -
Olivine tholeiite 6.75 28.84 - -

If the melilite nephelinite is the product of a very small amount of

melting, then it is not likely that larger amounts of melting will yield an
oliyine tholeiite with such a small amount of olivine unless the pressure
is greatly reduced or H20 plays a significant role. Both H20 and
reduced pressure tend to enlarge the field of olivine, thereby reducing
the olivine content of liquids on its boundary curves and surfaces. On
the other hand, the ankaramite , assuming it is not a cumulate, might be
the product of a greater amount of melting at the same pressure of the
same parent that produced the melilite nephelinite, because of the
increase in olivine content. Increased melting at the same pressure of a
rock in which olivine is the most abundant phase would progressively
lead to an increase in the olivine content of the liquid. It appears that
these rocks cannot be related simply to a parental material solely as
products of different degrees of melting at the same pressure. The
series of rocks studied by Arculus and Shimizu (1974) are much more
restricted in normative character and exhibit a small but progressive
increase in olivine content from 20.45 percent to 24.38 percent with
moderate deviations.
A theoretical analysis by Schilling and Winchester (1%7) of the
sequence of rare-earth element changes with degree of partial melting
yields the patterns shown in Figure 8-13. Although these calculated
curves have not been correlated with the normative character of the
deduced liquids, the curve for the relative rare-earth composition of
liquid at I percent melting, L(l%), would approximately express the
character of the data obtained for melilite nephelinites. The curve
L(50%) is similar to that obtained for Hawaiian tholeiites. In Schilling's
view the rare-earth element data constitute the strongest argument that
the degree of melting influences the normative character of the melt.
Green (1971, p. 709-710, Figures l and 2) apparently accepted this
view and proposed that within the pressure range 15-35 kbar the
normative character changes from lamite to nepheline to hypersthene
merely by increasing the proportion of melt at constant pressure with

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

152 GENERATION OF BASAL TIC MAGMA

1000

-e 1oo
ci.

-
ci.

[-~
~
___.--:_ ' - • ............._~ 0

~..........·~
10 1 !t~
~

La Ce Pr Nd Pm Sm Eu Gd Tb Oy Ho Er Tm Yb Lu
57 (Yt) 71
Atomic number
FIGURE 8-12 Abundances of rare-earth elements relative to 20
chondrites in a melilite nephelinite (open circles: normative
La = 7.40, Ne = 21.58), ankaramite (X: normative Ne = 3.12), and
olivine tholeiite (black dots: normative Hyp = 28.84) from the
Hawaiian Islands plotted on a logarithmic scale as a function of
atomic number. From Schilling and Winchester ( 1967, p. 269, Figure
2). (With permission of John Wiley and Sons, Inc.)

as little as 0.01 percent H 2 0! In all the experimental studies of the

simplest to the most complex synthetic systems, little phase-
equilibrium evidence has been found to support this view. No single
bulk composition has yet been shown to produce a complete series of
La-Ne, 01-Ne, 01-Hyp, and Hyp-Qz normative liquids merely by
increasing the degree of melting over a moderate range (e.g., 30
percent) or, in fact, over any range of melting. It was demonstrated
with the forsterite-diopside-pyrope system that the degree of melting
does influence the composition of liquid, but within narrow normative

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 153

bounds. The proportion of volatiles and the pressure, on the other
hand, greatly influence the normative character of the melt (Mysen,
1975). Yet the presence of a gas phase does not appear to greatly
influence the partitioning of the rare-earth elements (Cullers et al.,
1970), nor does the accompanying shift of "eutectic" composition
(Schilling, 1975, p. 1467, Figure 9).
It seems fortuitous that the required degree of melting could be
achieved each time to produce successive batches of the same compo-

10

0
....
(.)

Relative atomic number

FIGURE 8-13 Theoretical relative concentration (C/C 0 )
(logarithmic scale) of the 14 rare-earth elements (0 =lanthanum;
14 = lutecium) in melts (L) and residua (R) resulting from the partial
melting of primary material having the illustrated composition for
1 percent, 10 percent, and 50 percent melting. From Schilling and
Winchester(l967, p. 277. Figure 9). (With permission of John Wiley
and Sons, Inc.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

154 GENERATION OF BASALTIC MAGMA

sition. The only potential control of degree of melting identified is the

ill-defined proportion of crystals to liquid that leads to the physical
disaggregation of the parental material. Schilling (1966, p. 80) believed
that the degree of melting could be determined by the abrupt change in
heat requirements attending the loss of a phase on partial melting. The
sharp rise in liquidus temperature, particularly that of olivine, may
place a temporary limit on the degree of melting. On fractional melting,
the loss of a major phase, however, greatly changes the subsequent
liquid composition, and a basaltic liquid may not be a likely product.
Furthermore, the occurrence within the Nazca Plate of nepheline-
normative basalts and olivine tholeiites with the same rare-earth ele-
ment pattern (Schilling and Bonatti, 1975) casts doubt on the concept
that the degree of melting is the principal control on the composition of
liquid produced. For that region, Schilling and Bonatti (1975) found it
necessary to assume that the two magma types originated at different
eutectic compositions but as products of approximately the same
degree of melting.

MEASURE OF HETEROGENEITY OR PROPORTION OF PHASES?

What, then, does produce the systematic variation in the patterns of

abundance of rare-earth elements for the major basaltic rock types if
the degree of partial melting is not always the important factor? One
factor is the abundance of the rare earths in the parental material. A
heterogeneous mantle would satisfy the variance, but would the vari-
ance be systematically correlated with the many other factors that
determine magma type? A more appropriate dependence appears to be
on the proportion of phases taking part in the partial melting. Represen-
tative distribution coefficients for the rare-earth elements between
crystal and liquid are given in Figure 8-14. The representative partition
coefficients for the rare-earth elements between "crystal" and "liq-
uid"* for the major phases in garnet peridotite were summarized by
Helmke and Haskin (1973, p. 1520, Figure 3). The partition coefficients
between synthetic clinopyroxene and liquid as well as synthetic garnet
and liquid were determined, respectively, by Masuda and Kushiro
(1970, p. 44, Table 2) and Shimizu and Kushiro (1975, p. 413, Table 1).
It is evident that the light rare-earth elements will always be enhanced
in the liquid relative to the residuum. On the other hand, the heavy
rare-earth elements may be depleted or enhanced in the liquid, depend-
ing on whether garnet persists as a phase during the melting process.
*The "crystals" are phenocrysts from natural rocks that are presumed to be in equilib-
rium with "liquid'' represented by the groundmass or calculated residual liquids.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 155

d
.....
.!
~
~
.li!

i
0
c:

~
0
0.1
IL

La Lu
!17 71

FIGURE 8-14 Representative partition coefficients for the rare-

earth elements between natural crystals (olivine, orthopyroxene,
clinopyroxene, and garnet) and deduced liquid compiled by Helmke
and Haskin (1973, p. 1520, Figure 3), between synthetic
clinopyroxene (Di.0 En 20) and glass at 20 kbar (Masuda and Kushiro,
1970, p. 44, Table 2), and between synthetic garnet (Py.,Gross,")
and glass at 30 kbar (Shimizu and Kushiro, 1975, p. 413, Table 1).
Cs =concentration of rare-earth element in the residual crystal;
C 1• =concentration of rare-earth element in coexisting liquid.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

156 GENERATION OF BASAL TIC MAGMA

Eu!Ktic: C,.311 Gru

Start1n9 Material: CllauGru

FIGURE 8-15 Calculated rare-earth element concentrations in liq-

uid, C,_, generated by I percent, 10 percent, or 50 percent partial
melting of a hypothetical eclogite eutectic composition (Cpx..,Gr._,)
having 10 times the normalized average chondrite abundances of
rare earths (see Appendix). The partition coefficients for synthetic
clinopyroxene-liquid and garnet-liquid given in Figure 8-14 were
used, and pressure difference was neglected.

It is evident from Figure 8-14 that major changes in the rare-earth

element pattern will result from variations in the proportion of
clinopyroxene and garnet. Variation in the content of olivine, the
presumed dominant phase, will have little influence on the form of the
rare-earth element patterns, and yet it will greatly affect the overall
abundance of rare-earth elements in the liquid. Consider, then, a
simple binary system, clinopyroxene-garnet, with a eutectic at
Cpx3sGrss (cf. Figure 5-5). The experimental partition coefficients
given in Figure 8-14, although they were determined at different
pressures and may not be compatible, can be used to calculate the
rare-earth element contents of liquids according to the relationship
given by Schilling (1975, Appendix 1):

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 157

where

i rare-earth element
j phase
concentration of rare-earth element in liquid
CL=
concentration of rare-earth element in parental material
Cp=
y =
weight fraction of liquid (i.e. degree of melting)
K= partition coefficient between residual phase and liquid
E eutectic proportion in weight percent of phase j
X fraction of residual phase remaining in equilibrium with liquid
in weight percent

The rare-earth element pattern of the liquid appears to depend on not

only the degree of melting, but also the initial abundances, the propor-
tion of phases in the parental material, and the proportion of phases
contributing to the liquid. If the eclogite (Cpx + Gr) undergoing melt-
ing has an initial rare-earth element abundance 10 times that of the
average chondrite, then the rare-earth element patterns calculated for
various degrees of partial melting of the eutectic composition are as
presented in Figure 8-15. If a constant degree of melting of I percent is
assumed, then the influence of various initial proportions of phases is
as displayed in the calculated rare-earth patterns of Figure 8-16. * It is
clear that changing the proportion of critical phases is as important as
the degree of melting (Schilling, 1975, p. 1466). Gast (1968, p. 1071) was
of the opinion that the relative concentrations of rare-earth elements
were "quite insensitive to the phase that melts." Aside from the fact
that all phases must participate in the melt, the conclusion appears to
be in error because it was based on calculated changes dominated by
olivine and orthopyroxene content. Even though the more abundant
phases in the parental material have the greatest influence on the
rare-earth composition of the melt, the changes in character of the
*At first glance, it may seem unreasonable to assign the same initial rare-earth abun-
dances to differing proportions of the same minerals. Many eclogites do appear to have
chondriticlike abundances, and one need only look at the chondrites themselves for
examples of material having the same rare-earth abundances yet varying greatly in the
proportions of phases.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

158 GENERATION OF BASALTIC MAGMA

I'[Link]
Eutecllc ' CpanGr15

Rore•eorth element

FIGURE 8-16 Calculated rare-earth element concentration in liq-

uid, C~-o generated with I percent partial melting of different propor-
tions of clinopyroxene and garnet in a hypothetical eclogite having
I0 times the normalized average chondrite abundances of rare earths
(see Appendix). The eutectic composition of the system was as-
sumed to be Cpx3 .,Gr6 ., . The partition coefficients for synthetic
clinopyroxene-liquid and garnet-liquid given in Figure 8-14 were
used , and pressure difference was neglected.

rare-earth patterns of the melt reside primarily in the contributions

from clinopyroxene and garnet (Figure 8-14). There is a clear need to
determine the effect of the variables experimentally and evaluate their
influence on the rare-earth element pattern. Nevertheless, the rare·
earth elements appear to record information not yet identified among
the major element relations .

RARE EARTHS IN GARNET PERIDOTITE

The distribution of rare earths in a hypothetical garnet peridotite and its

melts and residues has been computed by Schilling (1975, p. 1469,

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 159

Figure llb ). He assumed that the garnet peridotite consisted of
01570px11Cpxt2Gr14 and that the eutectic composition was Ol3 0px3 -
Cp~7Gr47 (after Davis and Schairer, 1965). One model yielded the
results shown in Figure 8-17 for 10 percent and about 30 percent*
partial melting. The principal observation is that the heavy rare earths
are retained in the residuum after 10 percent partial melting, whereas
the light rare earths are depleted after the same amount of partial
melting. This model demonstrates that garnet is present initially,
persists with small amounts of partial melting, and is consumed with
somewhat larger amounts of partial melting. Schilling pointed out that
no natural ultramafic rock has yet been found that exhibits the implied
enrichment of heavy rare earths representative of residuum from very
small amounts of partial melting of a garnet-bearing parent.

GENERAL REMARKS AND SUMMARY

In conclusion, the absence of olivine in basalts held at the high
pressures at which they are believed to be derived from garnet perido-
tite may be due, on the one hand, to the derivative nature of the
basaltic magmas. The primary magmas, partially melted and separated
from garnet peridotite, change composition en route to the surface and
result in a liquid composition that, when recrystallized at high pres-
sure, does not contain olivine. On the other hand, eclogite, equivalent
in composition to basalt, requires both an olivine and an orthopyroxene
reaction relationship with liquid if it is derived from garnet peridotite at
high pressures. The orthopyroxene reaction relationship was demon-
strated by O'Hara and Yoder (1967), and the olivine reaction relation-
ship was demonstrated by Kushiro and Yoder (1974). An eclogite
liquid, brought to the surface and crystallized as basalt, would there-
fore also be derivative.
The few data available on the liquidus phase of basaltic compositions
at high pressure suggest that clinopyroxene is consumed first in partial
melting of garnet peridotite at high pressures, whereas garnet is con-
sumed first at relatively moderate pressures. These conclusions are
based on the assumption that H 20 does not play a significant role in
basaltic magma generation .
The rare-earth-element data may yield information on the degree of
melting, the proportions of phases in the starting material, the presence
or absence of garnet, and the extent of heterogeneity in the upper
mantle. Experimental evaluation of the principal variables is needed
*Clinopyroxene is consumed after 25 .5 percent melting (Schilling. personal communica·
tion, 1976).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

160 GENERATION OF BASALTIC MAGMA

10

.,,.,... ,.,... -----

.,-
/
/
/
i /
Iu 1.0 /
~
/
/

0.1

La Ce Pr Nd Pm Sm Eu Get Tb Oy Ha Er Tm Yb Lu
57 (Y) 71
Atomic: number

FIGURE 8-17 Computed enrichment factor for rare earths in a

garnet peridotite assemblage (0l,,,Opx 17Cpx, 2 Gr,.) melting at the
eutectic composition of Davis and Schairer ( 1965). given as Ola-
OpxaCpx.7Gr.7. The enrichment factor is the concentration of a rare-
earth element in the liquid relative to its concentration in the
assumed primary mantle. The liquid compositions for 10 percent
(solid line) and about 30 percent (dashed line) melting (L, 0 and L:oo)
were taken from measurements on natural basalts from the Mid-
Atlantic Ridge. The rare-earth abundances for the primary material
were calculated from the observed composition using assigned parti-
tion coefficients. assumed eutectic composition. and assumed pro-
portion and kind of phases entering the melt. The rare-earth compo-
sitions of the residua are obtained by difference. From Schilling
(1975. p. 1469. Figure lib). (With permission of John Wiley and
Sons. Inc.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Olivine in the Parental Material 161

before unambiguous conclusions can be reached. In general, trace ele-
ments indicate that the peridotites in hand are depleted in varying
degrees in the constituents of basalt. Furthermore, some basalts,
especially the oceanic basalts, appear to be the product of secondary
or tertiary melting events or the later product of fractional melting.
Other basalts, the alkali types, appear to be the product of primary
melting at an appropriate invariant point or an early product of frac-
tional melting.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Tectonophysics
9 of Melting

LOCATION OF INITIAL LIQUID

Whatever the parental material was prior to the beginning of melting, it
went through one or more metamorphic reactions as the temperature
was raised. Garnet peridotite can be expected to have the granular
texture of a metamorphic rock (Plate 2, p. 44b). The interfacial
angles between the grains are approximately 120° on the average,
according to Vernon ( 1968; 1970), but vary because of the anisotropy of
the crystals (Figure 9-1). "The points of intersection" of three grains in
thin section mark a potential channel . or vein. The junction of four
grains is the possible site of a tetrahedral pore. The grain contacts can
be viewed as local binary, ternary, and quaternary systems; yet the
rock begins to melt at the temperature appropriate for the whole rock
system.
Melting involves all phases, and there obviously must be communi-
cation between adjacent and more distant neighboring grains by diffu-
sion processes. The very first melt may actually begin within inclusions
or involve minor phases (e.g., apatite) interstitial to the major phases.
Fluid inclusions tend to migrate to crystal boundaries up thermal
gradients, and liquid-vapor inclusions tend to migrate down thermal
gradients (Anthony and Cline , 1970, 1971, 1972a,b). As a result of the
failure of the rock to reach textural and chemical equilibrium during the
metamorphic stage preceding melting, the crystals may also retain
chemical zoning or impurities at their periphery that would tend to
162

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Tectonophysics of Melting 163

FIGURE 9-1 Measured interfacial angles

and dips of cleavage (where not approxi-
mately vertical) of hedenbergite aggregate in
a high-grade metamorphic rock from Broken
Hill, Australia. From Vernon (1968, p. 8,
Figure 6). (With permission of the Oxford
University Press.)

1 [Link] 1

lower the melting temperature at the grain boundary relative to the

interior of the grain . It is believed that pervasive melting begins at
those grain junctions where all phases are in contact. In this sense
some models illustrate the intial melt as isolated droplets. With an
increase in thermal energy the first liquid provides communication to
other grains . If the rock is mineralogically homogeneous, all grains will
be in equilibrium with that first liquid.

ACCUMULATION OF LIQUID
The interior of an initial liquid droplet, if that form ever obtains, is
subjected to an additional pressure resulting from the surface tension.
According to Coble and Burke (1963, p. 206), the pressure is given by

where y = surface tension and r = radius of curvature of the droplet.

For ceramic materials, y equals I ,000 erg/cm 2 ; therefore, a [Link]
droplet is under an additional confining pressure of 40 bars by virtue of
its size and shape . The external hydrostatic pressure is thereby acting
to reduce the size of the drop. The implication is that the liquid does
not wet the contiguous solid surfaces well. On the other hand, if the
liquid does wet the solid surfaces well, then the liquid will be drawn

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

164 GENERATION OF BASAL TIC MAGMA

into the re-entrant angles between the contiguous solid surfaces along
which melting is taking place (Smith, 1948, p. 20, Figure 5). The result
will be a "liquid-bonded aggregate" in the terms of the ceramicists
(Allison et al., 1959, p. 517). All grains would then be initially coated
with a thin fdm of liquid. The wettability of different crystals will
differ and in different crystallographic directions. The physical proper-
ties of a rock undergoing initial melting will no doubt differ greatly de-
pending on how the liquid is distributed. In the writer's view, a three-
dimensional network of liquid film is the preferred description of the
initial melt, in contrast to an array of isolated droplets of variable
aspect ratio.
There are no experimental observations on the succeeding events in
the melting of a garnet peridotite; some insight may be gained, how-
ever, by examining the results obtained on a quartz-feldspar gneiss.
Mehnert et al. (1973) described the initial anatectic stages on the basis

FIGURE 9-2. Effect of temperature on the amount and distribution of melt in a natural
quartz-feldspar gneiss held at P~~:o = 2 kbar for 24 hours at each temperature. Width of
section. approximately I mm. From Mehnert et a/. ( 1973, p. 168, Figure 1). (With
permission of E. Schweizerbart'sche Verlagsbuchhandlung.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Tectonophysics of Melting 165

of hydrothermal experiments at a series of temperatures, pressures,
and times. One set of experiments at constant water pressure (2 kbar)
and time (1 day) shows the development of liquid with increasing
temperature (Figure 9-2). At about 680°C very thin films of glass, about
1 p.m or less in thickness, can be seen locally, particularly at the
contacts of the·three principal phases. At successively higher tempera-
tures, the films increase in thickness uniformly, irrespective of the
composition of the adjoining crystals. Concentration of the low-
temperature melting residua and impurities along grain boundaries in
the final stages of the original crystallization of the rock no doubt
contributes greatly to the early development of a film of liquid around
all the grains of the rock on remelting (see McLean, 1957, pp. 100-107).
It appears that migration along cracks, cleavages, grain boundaries,
and other discontinuities occurs readily. The initial partial melting of
granite under a water pressure less than the total pressure also yields
an almost continuous network of films, according to Brace et al. ( 1968).
Further support of the "liquid-bonded aggregate" concept of initial
rock melting is also found in the study of simple oxide systems. A
synthetic mixture in the CaO-MgO-AI203 system was held at a suita-
ble temperature by Clark et al. (1953, pp. 32-33) so that 10 percent melt
was produced. Single-crystal grains of periclase were observed to be
separated by films of liquid. They particularly noted the absence "of
separate 'lakes' of liquid in a continuous solid matrix."

HOMOGENEITY OF LIQUID
As melting progresses, the films unite,* and the crystals and crystal
aggregates form irregular relics. Mehnert et al. (1973) carried out
electron microprobe studies of the composition of the film and con-
cluded
that the initial melts produced at relatively low temperatures are rather constant in
composition, but heterogeneous with respect to inclusions of sub-microscopic crystalline
relics. At higher temperatures, these relics vanish but the melt is heterogeneous in
another sense, i.e., concentration gradients of the respective elements can be observed
across the melt from one mineral contact to the other.

Such gradients may be eliminated in experiments of longer duration and

under higher water pressures where mixing and flow may be effected.
According to Butler ( 1961 ), the natural fusion of an arkose by a dolerite
plug has produced a melt (now observed as a glass) that reveals no
*The high probability of intersection of isolated droplets and the formation of a network
structure has been demonstrated by Haller (1965).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

166 GENERATION OF BASAL TIC MAGMA

inhomogeneity of the potash content even though there is no textural

evidence to indicate mechanical mixing.

ANALOGY TO GLACIAL WATER

Some appreciation for the nature of the veining of a rock undergoing
melting may be gained from the presence of water in glacial ice, even
though the pressure effect and volume change are of opposite sign to
those for silicates. Nye and Frank (1973) emphasized the three-
dimensional aspects of the channels formed at three-grain intersections
in temperate ice. Deformation and recrystallization, in the view of
Lliboutry (1971), tend to close off capillary intergranular channels and
make glacier ice, especially tim, "practically impermeable ." On the
other hand, Shreve (1972) pictured deformation as aiding the passages
to expand and contract. The larger channels gradually increase in size
because of the heat generated by viscous dissipation and the heat
transported by the flow of melt having higher temperatures. He vi-
sualized a network of passages tending in time to become "arbores-
cent" like a river system. The main " streams" trend in the general
direction of flow-upward in the case of magmas. The same picture on
a large scale was presented by Wilshire and Shervais (1975 , p. 270,
Figure 14) in describing the anastomosing feeders to the main conduits
of a volcano (Figure 9-3).

ROLE OF DEFORMATION
As in glaciers, deformation of the parental material plays a major role
in the collection of magma. The disposition of the melt will be influ-

M. Olacontlnulty

FIGURE 9-3 Schematic diagram il·

lustrating anastomosing feeders to
main conduits in a volcanic edifice.
After Wilshire and Shervais (1975.
p. 270, Figure 14). (With permission
of Pergamon Press, Ltd .)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Tectonophysics of Melting 161

FIGURE 9-4 Comparison of average pressures acting on a unit of rock under a

hydrostatic load pressure P with small (200 bars) incremental changes in horizontal
stresses. Equal horizontal compressive stresses of the same sign (A) and of opposite sign
(8).

enced by the shear stresses acting on the parental material, especially if

flaws are present. The mean stress lf in a rock is equal to
I/3(<T1 + <T2 + <T3 ), where <T is a principal stress. If the vertical stress
equals P and the horizontal stresses are both P + 200 bars, the mean
stress is P + 133 bars (Figure 9-4A ). In the anhydrous melting of rocks
the dT/dP is positive; therefore, an increase in the mean stress will
cause melting to stop. There appears to be a dilemma: Any unbalanced
nonhydrostatic stress tending to move magma will result in freezing the
magma! One possible explanation is that the deviatoric, horizontal
stresses are of opposite sign (Figure 9-4B), the rock can maintain its
mean stress, and melting continues. On this basis every principal
compression will be accompanied by a principal dilatation. Another
possible explanation is that the strain rate is sufficient for the nonhy-
drostatic stress (200 bars) to generate heat, and melting ensues, provided
the viscous heating exceeds the heat conduction away from the site.
Still another explanation is that freezing is controlled by the rate at
which the enthalpy offreezing can be conducted away. If heat conduc-
tion is slow relative to the strain rate, the magma will retain its fluidity
and can move in response to the applied stress.

LOCALIZATION OF LIQUID

RESPONSE TO STRESS

Analyses by Nye (1967) and Savage (1969, p. 119) illustrate how a

stress acting on a rock would cause fluid to collect in lenses parallel to

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

168 GENERATION OF BASAL TIC MAGMA

FIGURE 9-5 Response of two liquid-filled

lenses perpendicular to the maximum and
P+cr minimum principal stresses acting on a unit
of rock under an average lithostatic pres-
~ tP-cr sure. Total fluid pressure decreases in lenses
perpendicular to the minimum stress; and, in
an anhydrous rock. melting is enhanced in
those lenses. After Savage (1969, p. 119,
Figure 4). (With permission of Elsevier Sci-

t
entific Publishing Company.)

the principal stress (Figure 9-5). The flow of melt to regions of

relatively lower pressure is appropriate because of the positive dT/dP,
but what then is the flow of hydrous magma? In the presence of water,
it would seem that the magma would flow into lenses perpendicular to
the maximum compressive stress because higher pressure enhances
melting. A detailed analysis of the hydrous case is required; present
opinion, however, is that the mechanical state of a flaw or lens (open or
closed) is more important than the physicochemical effect on the
melting itself.

COALESCENCE

Some nodules retrieved from kimberlite pipes are extremely sheared-

almost mylonitized (Figure 9-6). According to Weertman (1972),
shearing is most effective in coalescing the initial droplets in a
magma. On the basis of his earlier analysis of gas-bubble coalescence
in a deforming glacial ice mass (Weertman, 1968), he illustrated (Figure
9-7) "bubbles" coalescing as a result of differential movement. In the
light of numerous high-pressure experiments on minerals associated
with garnet peridotite in solid-media apparatus, alleged to maintain
hydrostatic conditions but in fact having a shear component, the initial
liquid is only very rarely observed to be dispersed as droplets or
"bubbles." On the other hand, application of the concept to the
coalescence of "pools" of magma in an upper mantle undergoing shear
over large distances and through a thick zone may be more appropri-
ate. Shaw (1969) assigned an important role to "viscous failure and
flow" in the accumulation of magma. He described (1969, p. 533) the
melt fraction as being "virtually kneaded from the crystalline source."
The separation of phases attributed to mechanical deformation, how-

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Tectonophysics of Melting 169

FIGURE 9-6 Sheared garnet peridotite, No. 1611. Thaba Putsoa, Lesotho. Width of
section is approximately 12 mm. Photograph courtesy of Dr. F. R. Boyd.

ever, may result from other processes. The common segregation from
an initial homogeneous rock of quartz and feldspar into layers alternat-
ing with layers of the remaining ferromagnesium minerals has been
cited as the result of deformation during metamorphic differentiation.
The observed dissimilar rheology of quartz and feldspar casts doubt on
this concept (see also Shelley, 1974). Means and Williams (1974)
attributed such metamorphic differentiation in a salt-mica mixture to
enhanced solubility, due to stress concentrations, and redeposition.

FIGURE 9-7 Coalescence of melt globule

resulting from differential relative motion of
individual globules. A. differential move-

0 ment of two globules; B. intersection and

coalescence of two globules ; and C,
spheroidization of two globules. After
Weertman (1972. p. 3531, Figure 1). (With
A B c permission of the Geological Society of
America.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

170 GENERATION OF BASAL TIC MAGMA

VOLUME CHANGE OF MELTING

LARGE INCREASE IN VOLUME

In the absence of plastic flow, shearing usually results from multiple

fracturing, and fracture produces dilatancy-that is, an increase in the
volume occupied by the rock . Melting also produces a large volume
change. The amount of volume increase for relevant end-member
minerals is given in Table 9-1. For example, the volume increase on
melting the eutectic composition of the forsterite-diopside-pyrope
system at 40 kbar is about 16 percent. Correction for compressibility
and thermal expansion would no doubt reduce that value somewhat,
but it remains clear that a large volume change must be accommodated
as melting takes place. If the requisite volume is not available, the
porosity being negligible, the pressure will rise and melting will stop.
Although there are no data for the volume changes taking place under
hydrous melting, it would appear that such melting may be enhanced if
confined.

TUMESCENCE , RUPTURE , OR PLASTIC FLOW?

To accommodate the volume change on melting, a region could

undergo tumescence . The inflation could lead to a rise of tens of meters
at the surface, depending on the volume of melt and structural compen-
sations. Alternatively, the volume change may be accommodated by
plastic flow . If heating continues and the local stress increases further,
the resultant long-range stress difference may lead to plastic flow, and

TABLE 9-1 Volume Change on Melting at I atm

End-Member Mineral p, xtl p, Glass t!V (cm 3/gm) %Change
Forsterite 3.223 3.035" 0 . 01~ 6.2
Fayalite 4 .068 3.764" 0.020" 8. 1
Clinoenstatite 3.183 2.735 0.051 16.4
Diopside 3.275 2.846 0.046 15. 1
Pyrope 3.582 3.031 0.051 18.2
Anorthite 2.765 2.700 0.010 2.7
High albite 2.605 2.382 0.036 9.4
High sanidine 2.597 2.400 0.036 8.2
'Calculated from estimated AV and p,.,.
'Estimated usiflll measured dTidP = 4 .77'/kbar (Davis and England, 1964, p. 1114), melting temperature~ 189CY'C.
P.u. and All. = 208.2 callg.
'Estimated usi1111 measured dT!dP = 6.5"/kbar (Undsley. 1967, p. 230), melting temperature= 120S'C, P.•·
and
All. = 108. 1 calls.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Tectonophysics of Melting 171

the requisite volume will be provided by adjacent porous regions or by
means of tectonic movements. The pressure developed when melting
takes place at constant volume appears, at first glance, to be consider-
able. Fordiopside the volume change on melting is 0.046 cm 3 /g, and the
specific volume is 0.305 cm 3/g. Using the measured crystal compressi-
bility {3 = (1/V0 ) (aV l!lP), which is 1.07 x J0- 6 /bar for diopside (Adams
and Williamson, 1923), the potential pressure developed on melting is
140 kbar! As shown below, such circumstances never obtain.

STATE OF STRESS IN EARTH

Rocks obviously cannot withstand such high local pressures, which

would cause a large deviatoric stress. Bullen (1947) estimated, on the
basis of the energy release in large earthquakes, that the maximum
stress in rocks did not exceed I kbar. Kaula (1963) said that the form of
the geoid measured from satellite data indicated the earth could sustain
stress differences of only 0.2 kbar. Birch (1964) noted that the stress
differences associated with gravity anomalies of 200 mGal were about
0.5 kbar at depths of 50 to 100 km. Brace (1964) measured the tensile
strength of diabase in the laboratory and found it to be only 0.4 kbar,
whereas Roberts (1970, p. 297) thought that the tensile strength of the
upper mantle was about 0.5 kbar and that the stress difference for shear
fracture would therefore be 4 kbar. Jeffreys (in Runcorn, 1967, p.
433~35) argued for a strength of I kbar at the surface and 0.1 kbar at
600 km. [The deepest earthquakes are recorded at 700 km (Griggs and
Handin, 1960, p. 359).]
It is evident that the pressures developed on melting will in fact
deform or break the confining rock, but will the melting take place in
view of the positive dT/dP? Under constant volume conditions, the
temperature-pressure gradient is defined by

~!IV=!'
where {3 = compressibility and a = thermal expansion. For diopside
the dT/dP I v = 45°/kbar, and for olivine, 34°/kbar.* The constant-vol-
ume slope greatly exceeds that of the melting curve, and melting will

*Values of compressibility and thermal expansion at 2o•c used are:

Diopside: fJ = 1.07 x 10-6 /bar (Adams and Williamson, 1923)

a = 24 x 10-6/deg (Kozu and Ueda, 1933)
Olivine: fJ = 0.82xlo-•tbar (Bridgman , 1948)
a = 24 x J0- 6/deg (Skinner, 1962)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

172 GENERATION OF BASALTIC MAGMA

take place if the system is confined to a fixed volume. It would be

interesting to know the effect, if any, of stress on the melting curve: no
data on any relevant systems are known to the writer. In a pioneering
study, Coe and Paterson ( 1969) have demonstrated that the a-{3 quartz
transition is raised by uniaxial compression with or without a confining
pressure. That technique appears to be available for investigating a
melting curve under similar conditions.

ADHESIVE STRENGTH OF CRYSTALS

In view of the importance of the tensile strength of rocks to many

economic endeavors such as mining, it is surprising that a greater effort
has not been made to understand why dissimilar minerals adhere to
each other. The coherent and incoherent boundaries formed on exsolu-
tion of similar structures are now being studied (Yund, 1974), yet only
one study has been made to evaluate the adhesive strength of dissimilar
crystalline center forces. A quartz-feldspar interface was pulled apart
in the absence of a confining pressure with a force of 60 to 100 bars,
according to the tests of Savanick and Johnson (1972). They pointed
out that, judging from the properties of the broken surfaces, the forces
responsible for the adhesion operated over only a portion of the
interfacial area. A more extensive study should be undertaken involv-
ing more parameters.

ESCAPE FROM THE CONTAINER

The present view of the initial liquid, then, is of drops at four-phase
intersections, veins along three-phase boundaries, and films between
grains. If the rock is stressed, the liquid aggregates along those bound-
aries perpendicular to the least compressive stress. The problem then
seems to be how to assemble the aggregates into an anastomosing
system of rivulets eventually leading into a feeder dike.
It is held by some that the container could be viewed as a three-
dimensional sieve holding the liquid in the interstices, the liquid being
released by breaking the wires in the sieve. Cracking of a brittle
container seems to be inappropriate to a model of a magma chamber at
very high temperature and presumably in a plastic state. Ductile
fracture, wherein the solid undergoes marked deformation before
fracture, is frequently observed in the failure of metals and yet not well
understood (Biggs, 1970, p. 1227): that process may be a suitable
description of the failure of a magma chamber. The apparent contradic-
tion between earthquakes suggesting brittle failure and magmas indicat-

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Tectonophysics of Melting 173

ing fluid or plastic behavior is often resolved by noting that a brittle
response obtains at relatively low temperatures under short-term,
low-magnitude (-100 bars) stresses, whereas plastic response occurs in
the same material at relatively high temperatures under long-term,
exceptionally low-magnitude stresses. Analogy is often made to the
behavior of pitch because it flows like a liquid under its own weight,
yet, when hit sharply with a hammer, breaks with a conchoidal
fracture.
How, then, is the liquid to be moved from or through the hot plastic
region surrounding a magma chamber into the cooler brittle region?
Observation of a moving lava flow where cracks develop at the frontal
crust and fresh lava oozes out at temperatures as low as 900°C (Spry,
1962, p. 214) is adequate proof that liquid can escape through the
plastic region under a suitable force.

PLASTIC ENVELOPE

Plasticity is defined as continuous deformation without limit, relatively

independent of pressure and without fracture, under a reasonably
constant stress that exceeds a critical value (Figure 9-8). The mecha-
nisms involving intracrystalline plasticity depend on the diffusion of
material in various ways. Specific mechanisms dominate in various
temperature regions, stress conditions, and strain rates. The bound-
aries of the regions of deformation mechanisms are diffuse because of
such factors as grain size and water pressure. A deformation mecha-
nism map (Ashby, 1972) has been prepared for olivine (Stocker and
Ashby, 1973, p. 409) and is reproduced as Figure 9-9. At low tempera-

Plastic

t FIGURE 9-8 Description of the re-

sponse of a rock to increasing stress
depending on the observed strain for
a constant strain rate.

Strain-

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

174 GENERATION OF BASALTIC MAGMA

6r-~--~--~--~--~~--~--~--~--~----~--~

5
--------~~~~-----------
Dislocation glide

-I

200 400 600 800 1000 1200 1400 1600 1800

Temperature, •c
FIGURE 9-9 Steady-state deformation map for polycrystaUine olivine (Fo••-••> illus-
trating regions (outlined by heavy lines) of dominant deformation mechanisms as a
function of the stress (u) and temperature for various strain rates (light lines). High-stress
dislocation creep regime occurs above the heavy dashed line. N-H = Nabarro-Hening.
Grain diameter = I mm. Total hydrostatic pressure = 10 kbar. Activation volume for
diffusion = 50 cm3/mol (cf. atomic volume = 31 cm3/mol). From Stocker and Ashby
(1973, p. 409, Figure 7). (With permission. Copyrighted by the American Geophysical
Union.)

tures and strain rates, the dominant deformation mechanism is "dislo-

cation glide"; at intermediate temperatures and strain rates, it is
"dislocation creep"; and at high temperatures and strain rates, it is
lattice diffusion, "Nabarro-Herring creep." At temperatures some-
what below the melting temperature and at slow strain rates, the
mechanism is grain boundary diffusion, "Coble creep." If the most
likely conditions around a magma chamber are relatively high tempera-

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Tectonophysics of Melting 175

tures, slow strain rates, and low stress, the mechanism will be either
Coble or Nabarro-Herring creep; if the stress is high, however, dislo-
cation creep becomes important.

TIME DEPENDENCE

Strain rates for surface displacements are generally in the range w- 12 to

10- 15/sec, according to H. C. Heard (personal communication, 1975),
based on geodetic measurements on faults (Whitten, 1956), rebound
rates from crustal unloading (Crittenden, 1966), sea-floor displacement
rates (Heirtzler et al., 1968), and crustal shortening in orogenic regions
(Gilluly, 1970). Unfortunately, most laboratory experiments are re-
stricted to strain rates greater than I0- 8/sec, and extensive extrapola-
tion is required for application of the results to the earth. If the
effective viscosity and the stress of the plastic region are known, an
estimate of the strain rate can be made through the relation

TJ = a€,
where TJ is the viscosity; a, the stress; and €, the strain rate. For
T'/ = 10 10 poises and a= 0.5 kbar (= 5 x 108 dyne/cm 2 ), the strain rate
is about 5 x I0- 14/sec. For an average upper mantle with TJ = 1022
poises and a= 0.5 kbar, the strain rate is 5 x I0- 15/sec, in accord with
surface displacement rates due to isostatic rebound.
The response of the contained liquid to the strain rate in the host
rock is equally important. The effective viscosity of basaltic magma at
the surface is about 500 poises at the liquidus (Shaw et al., 1968).
Preliminary experiments by Kushiro et al. (in press) suggest that the
viscosity is about 30-50 poises at 1375-1400°C and 20 kbar! For such
an effective viscosity the liquid would literally squirt out of a contain-
ing rock under stresses of a few bars after an interconnecting liquid
network is established. Furthermore, dissolved volatiles lower the
viscosity markedly for tholeiitic melts but only negligibly for alkali
olivine basalts (Scarfe, 1973, p. 101).

''KNEADING OUT'' THE LIQUID

The plastic deformation of lherzolite has been studied by Carter and

Ave'Lallemant (1970) at a series of temperatures (Figure 9-10). Their
data show that at temperatures within the proximity of the solidus,
lherzolite behaves plastically under very small stress differences. Pre-
liminary data by the same authors show that even lower stress differ-

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

176 GENERATION OF BASALTIC MAGMA

20~~~-r~~~~-r~~

15
FIGURE 9-10 Deformation (E) of a
lherzolite under a confining pressure
of 10-20 kbar and a strain rate of
7.8 x 10-•tsec as a function of differ-
ential stress and temperature. From
Carter and Ave'LaUemant (1970.
p. 2193, Figure 7A). (With permission
of the Geological Society of America.)

1280°C
0~~~5~~Et~IO~~~I5
E, percent

ences are required if water is present. It appears that liquid could

readily flow along intergrain boundaries until it reaches a region that
can sustain brittle failure. In view of the ease of movement of liquid
through the "kneading out"* process, it is reasonable to suppose that
magma is more commonly extracted by fractional melting than by
batch melting. In such cases it would be necessary to store the magma
in auxiliary reservoirs to make large volumes available because the
initial stages of aggregation depend on relatively slow diffusional
processes. The behavior of lherzolite also suggests that mass move-
ment of a mush, once formed, would quickly bring about liquid
separation (Sleep, 1974). The source area appears to be best described
initially as a region mainly crystalline with a continuous network of
films of liquid, and yet it is difficult to abandon completely the notion of
magma-filled caverns, albeit multiple and deformable, for auxiliary
reservoirs.
If the spacing of volcanoes is controlled by the size and shape of the
source region (see Chapter 10), the degree of melting of the source rock

*The "kneading" process was described by Chamberlin and Salisbury (1909, pp. 579 and
629-633) as the result of stresses arising from lunar and solar tides. Harker (1909, p. 323
If.) described a similar process for the expulsion of residual magma under mechanical
stress as the "straining off" or "squeezing out" of liquid. Later. the terms "wine-press
differentiation," "filter pressing," and "filtration differentiation" were applied to the
removal of residual liquids. The terms would be equally appropriate for initially formed
liquids.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Tectonophysics of Melting 177

is small, and the liquid has a small viscosity, then the magma that feeds
a specific vent flows from considerable distances. The nature of the
flow is probably much more complex than flow of the Poiseuille type
(Darcy's law) or of the Knudsen type, depending on the size of the
capillaries. To the best of the writer's knowledge, the analysis of the
flow of fluids through a stressed, porous, heterogeneous solid at high
pressures and temperatures has not as yet been undertaken.

SOLUBILITY UNDER STRESS

Plastic flow involves recovery processes such as solution and re-

precipitation as well as deformation mechanisms (twinning, gliding, and
volume diffusion). The solubility of a crystal under a compressive
stress is greater than it is under only hydrostatic pressure. This concept
was first suggested (but not necessarily demonstrated) by Thomson
(1862, p. 396), who applied a piston load to NaCl crystals immersed in a
saturated solution. The compaction of the salt crystals was interpreted
by Thomson to indicate that the salt was dissolved away at the stressed
faces and deposited at the free faces. The circumstances of the Thom-
son experiment are analogous to those for crystals immersed in a
magma saturated with those crystals.*
Unfortunately, consideration was not given to the effects of anneal-
ing recrystallization in the absence of a saturated salt solution, nor was
it suggested how the stressed face, isolated from the salt solution by the
piston face and bottom of the container, communicated with the
ambient salt solution. A similar experiment was performed by H. K.
Mao and P.M. Bell (personal communication, 1974) with gypsum in a
saturated solution at very high pressure in a diamond cell, and no
solution and redeposition were observed. The "Thomson effect" was
qualitatively discussed by Sorby (1863, 1879) and applied to geological
problems by Riecke (1895, 1912). A thorough review of the ther-
modynamics of recrystallization of stressed solids in the absence of a
saturated solution was given by Paterson (1973). Although there ap-
pears to be a sound theoretical basis for enhanced solubility under
stress (Mao and Bell, 1971), definitive experimental demonstration has
not been achieved.

*The solubility curve is the same curve that describes the melting behavior of a crystal
under hydrous conditions. The· curve is continuous or discontinuous, respectively ,
depending on the absence or presence of critical end-point phenomena. In other words,
the solubility curve describes the behavior for dilute solutions, whereas the melting curve
describes the behavior for concentrated solutions.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

178 GENERATION OF BASAL TIC MAGMA

LOSS IN STRENGTH OF CONTAINER

The presence of a melt within a rock also brings about a great loss in
strength. In general, an increase in confining pressure produces an
increase in strength for a wide range of rock types. The effect of
confining pressure on the strength of basalt (diabase) and peridotite is
given in Figure 9-11 ; these curves are typical in form for most rocks
(Mogi, 1966).
With the formation of the first droplets of liquid in a rock, a pore

22

20

18

16

e0
Q 14
0
:i
~
0 12
~
10
~ 10
••
t~ 8
u

0 4
Confinln9 pressure, kilobora

FIGURE 9-11 Compressive strength of diabase (Brace, 1964,

p. 138) and peridotite (Mogi, 1965, p. 371) as a function of confining
pressure.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Tectonophysics of Melting 179

6

.,.... 4 f\ : 0 .5P,
0
.&J
E
:.ii 3
.,.,; ~ •0. 9~

-•...
( /)
2

OL---~--~~--._--~ ____.___ ~--~

0 2 4 6 8 10 12 14
Strain, percent
FIGURE 9-12 Stress-strain curves showing the effect of pore
fluid (H 20) on the wet strength and mode of deformation of a rock
(Solenhofen limestone) at various ratios of pore-fluid pressure (P1) to
confining pressure (P,). T = 20"C ; effective porosity = 5.3 percent;
grain size= 5.20 #Lm ; u ,. = 3 kbar. From Rutter ( 1972, p. 21, Figure
4A). (With permission of the Elsevier Scientific Publishing
Company.)

pressure is generated . In effect, the pore pressure counters the confin-

ing pressure, thereby reducing the strength of the rock (Figure 9-12).
The concept was demonstrated in detail by Hubbert and Willis (1957),
Hubbert and Rubey (1959), and Rubey and Hubbert (1959) and experi-
mentally investigated by Heard (1960), Handin et al. (1963), Raleigh
and Paterson (1965), and Heard and Rubey (1966).

MAGMAFRACTING

Application of the pore-pressure principle to partial melting appears to

have been made first by Mogi (1966), and to hydrothermal ore deposits
by Phillips (1972). The propagation of cracks by increasing the pore
pressure artificially, hydrofracting, has been used for many years,
since 1948, for the purpose of increasing the permeability in oil reser-
voirs. It is appropriate to apply the term magmafracting to the effec-
tive decrease in strength of a rock by virtue of the generation of magma

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

180 GENERATION OF BASALTIC MAGMA

and the process of crack propagation by magma in the brittle failure
region.
The propagation of the crack is again parallel to the maximum
principal stress at depth; however, it is generally perpendicular to the
least principal stress under relatively near-surface conditions. The
common occurrence of vertical or inclined dikes at depth and the
shallow formation of sills supports this view. The rate of propagation of
cracks in polycrystalline materials at high temperatures is under cur-
rent investigation, especially by ceramics engineers (Wachtman, 1974).
Theories that include chemical reaction enhancement terms have been
evolved for slow crack propagation (Wiederhom and Bolz, 1970). Of
particular interest is the experimental correlation of acoustic emissions
with crack velocity in materials such as polycrystalline corundum at
high temperatures (Evans et al., 1974). Comparison of such emissions
with earthquake foreshocks may be appropriate.

VOLCANISM PRODUCES SOME EARTHQUAKES

Dilatancy has been recorded in pre-earthquake episodes lasting 1-3
months and is potentially a tool for earthquake prediction (Nur, 1972).

M. discontinuity

50

.!• 100
• ••

..
Plostic zone
I
:. 150
•••
••
I!
A
---
0• 200

250

300~--------------------------------------------~

FIGURE 9-13 Plastic zone between lithosphere and asthenosphere as visualized by

Anderson (1962. pp. 52-53). Black dots are earthquake foci . Stippled areas are magma.
Arrows indicate relative motions of regions. (With permission of Scientific American.
Inc.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

·:··-.
Pacific Plate

-
00

Antarctic Plate Antarctic Plate

FIGURE 9-14 Location of volcanoes (heavy dots) relative to earthquake zones (light dots). Plate boundaries with offsets
by transform faults are shown by solid and dashed lines; subduction zones, by line with triangles. Arrows mark motions of
plates. Geological Museum of London (1972, p. 13, Figure 21). ("Crown copyright. Institute of Geological Sciences
illustration, reproduced by permission of the ControUer, Her Majesty's Brittanic Stationery Office." Fee paid.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

182 GENERATION OF BASAL TIC MAGMA

The increase in pore volume was believed by Nur to provide access for
ground water, but could equally well provide access for magma. The
presence of liquid decreases the strength of the rock, and rupture takes
place. It would appear that magma may invade a region quiescently
several months before its presence is announced by rupture. Near the
surface, such dilatancy may contribute to the observed tumescence of
the volcanic edifice (Mogi, 1958; Eaton, 1959, 1962).
Where conditions are appropriate for magmafracting, the aggrega-
tion of magma will, no doubt, produce earthquakes. It is the writer's
view that earthquakes associated with volcanism are the result, not the
cause, of volcanism. This view rests on the theses developed above:
(a) the thermal regime outside the boundary of the magma chamber,
marked by its solidus, suggests plastic rather than brittle behavior of
the confining rocks; and (b) the aggregation of the magma into rivulets
and cracks occurs through the process of magmafracting in the region
of brittle behavior.
In Anderson's view (1962), the plastic zone is not just a local region
around a magma chamber but is a broad (60-250 km thick), ill-defined
region of low velocity marking the transition from the lithosphere to the
asthenosphere (Figure 9-13). The temperature in the plastic zone ap-
proaches the melting point of the rock, and magma appears where
zones of disturbance are marked by earthquakes. There is indeed
strong correlation of earthquakes with volcanic eruption (Figure 9-14).
It is evident that the deformational structures of a region control the
position of the volcano at the surface. The regional fabric of the
prestressed environment exerts the major control on the location of new
volcanoes; Fiske and Jackson (1972) demonstrated that the position of
subsequent volcanoes may be the result of the gravitational stresses in
pre-existing edifices. There is considerable debate on whether or not
earthquakes associated with volcanism record the strain released as
magma rises to the surface. At present, the issue is not resolved;
however, the commentary in the next chapter on the energetics of
volcanism may be of some value .

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Energetics
10 and Periodicity

ENERGETICS

PARTITION OF ENERGY

The energy released in a volcanic eruption is both thermal and kinetic.

Estimates of the amounts released have been made by summing the
contributions by volcanic earthquakes, potential energy due to lifting
of the lava, thermal energy released from the flows, losses due to heat
conduction and gas expansion, thermal-spring activity, and kinetic
energy of ash and pumice eruptions. Estimates of the total energy for
the largest eruptions, calculated by Yokoyama (1957), are listed in
Table 10-1 . Only about half the magma generated reaches the surface in
Hawaii, according to D. A. Swanson (personal communication, 1975),
and the wave energy received at teleseismic distances is probably only
a few percer.t of the total yield (Bath, 1966, p. 161). The inferred
relations for Mount Rainier, Washington, prepared by Fiske et a/.
(1963), illustrate why a large proportion of the magma produced does
not reach the surface (Figure 10-1 ).
The partition of released energy in explosive events may be crudely
related to possible kinetic-thermal conversions in the alleged genera-
tion of magmas by seismic events. According to an analysis by Gault
and Heitowit (1963), the energy in a hypervelocity impact in basalt is
about 19-23 percent thermal, the remainder being distributed between
comminution and ejecta, with a negligible amount lost to compaction.
183

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

---
~ --

·.. ::::::--~--
... __:::::.: --- 7i
-
--
--·-- ~-

-----
7,; '' "
. . \ ~I',-,:-',''',:\
- ' , I _.. '

~-·
~~,P ~I , , I',,, '-,....,.'1~
- ,._,,,,I ;I
• ' - \ I ';;:,.(_

~;e''Jp'}/1', , , :- ~~ ' - I / I
1
1 _, \/' I ~' , ' _, ~­
~ ~,
~.:;-::_;•-[Link]',111-, -~ - 1: / ,-~,-~,--·'1~',-,
' //\ I / / ' I I I ' / \ \ I \
, -- I,I '\ '' -., , '/\ \
/ - ' .....

~\ ,-
,,,_ ~~-- _,,,,'''-1' ' " I ) ' /

,',--·- /;' - , - , _ , , ' ' I " ' _ .... /'''-,/-~

/ /

,'/l'~'"""i'-1",.,1~'1'"'1-'-1/\1 ·. . . ~~·,._<.·
\ _.-' \ \ - \ -.. I I ' I I - - I I I I I I - I \ '

....
' _1,::.1' -,-~-- _ _, 7; _,11'-1''
,-,-1 ,_1
I _ 1- 1 , 1 '
/':· /

....... _
',--,-'"_"-,-"-----' _,,,
I I I 1 I I 1 ' ' " 1 1 I ' ' - / '
/1 I \ ' -
-- I - _ , I \
/- ~
-
~
I 1-
_,, /
' ,- ' / I ,. - \ I - I - - I \ / I/ I / ;- \ / I/ - I

, ... ,, ,,,_ ,.... ,'.

I _,- I / /. I - -I ' '__I I I / ' I - \ I - II'/ \ I- I .,'..,_

( \ / ' / ' \ / ' - ' - - - .:: 1- /

FIGURE 10-1 Cross section of Mount Rainier, Washington, illustrating the large volume of magma (Tg) that penetrates the volcanic edifice as
sills (Td,) and dikes and does not erupt. The contemporaneous erupted rocks are labeled T•. T., T, and TJ are various Tertiary formations of
volcanic products, including mudflows and siltstones. From Fiske et al. (1963, p. 51, Figure 38). (With permission of the Director of the U.S.
Geological Survey.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Energetics and Periodicity 185

TABLE 10-1 Estimate of Total Energy Re-
leased in Dated Volcanic Eruptionsa
Energy
Released
Volcano Year (erg)

Tambora 1815 8.4 X 1018

Sakurajima 1914 4.6 X JOU
Bezymianny 1955- 1966 2.2 X 1015
Krakatoa 1883 ca. l x 1015
Asama 1783 8.8 X 1014
Fuji 1707 7. 1 x 1014
Capelinhos 1957 4 X 1014
Sakurajima 1946 2. 1 X 1014
Kilauea 1952 1.8 X 1014
Torishima 1939 9.7 X 1013
Komagatake 1929 5.6 X 10'3
Miyakeshima 1940 4.8 X 1013
Bandaisan 1888 ca. l x 1023
Pematang Bata 1933 4.5 X 10'1
Una-Una 1898 1.8 X 1021
Mihara 1954 1.3 X 10'1
A renal 1968 I X 1022
Adatarasan 1900 6.4 X 1011
Asama 1938 4.0 X 1011
Mihara 1912 6.3 X JOIO
Tokachidake 1926 2.8 X 1010
Showa Sin-zan 1944 1.4 X JOIO
Kusatsu-Shirane 1932 1.6 X 10 18
•After Yokoyama (1957. p. 96. Table V) and Macdonald (1972. p. 60. Table 4-3).

The thermal energy in a contained underground nuclear blast would be

expected to be somewhat greater. Butkovich (1974) estimated the
amount of melt generated by an underground nuclear explosion to be
about I ,000 metric tons of melt per kiloton of energy yield in a rock
with a density of 3.0. The energy released by 1-100 megatons of TNT
explosive is about equal to that of large volcanic eruptions (1022 -102:.
erg). If the energy yield was I 05 megatons ( 1027 erg), the amount of melt
produced would only be about 36 km 3 • Such an energy yield is equal to
that in an earthquake measuring 8.5 on the Richter scale, so it is not
likely that shock waves of moderate earthquakes will produce large
volumes of magma, even if the energy conversion is as great as that in a
nuclear explosion.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

186 GENERATION OF BASALTIC MAGMA

ENERGY IN EARTHQUAKES

The energy released in earthquakes can be estimated from the empiri-

cal relation of Gutenberg and Richter (1949, p. 19):

log E = 12 + 1.8M,

where E = energy released in ergs and M = magnitude of earthquake as

defined by Richter (1935).* A magnitude-6 earthquake releases about
1023 erg, whereas the greatest earthquakes on record, M = 8.6 (e.g.,
near Honshu, Japan), released 1028 erg. The rough equality of the
energy released in a volcanic eruption and in a moderate earthquake is
presumed to be fortuitous. The efficiency of conversion of mechanical
energy into thermal energy is no doubt low, and earthquakes are
related primarily to the movement of magma rather than its generation.
The relatively close association in time of some larger earthquakes with
eruption supports this view (Figure 10-2).
Several attempts have been made to relate surface fault length to the
magnitude of an earthquake (Tocher, 1958; lida, 1959; Press, 1965).
Such an empirical relation might be useful in estimating the propaga-
tion of a magma towards the surface. Unfortunately, the wide varia-
tions in stress drop, source dimensions, and the complexities of
measuring the parameters of the fault itself have not yielded a reliable
relationship (Brune and Allen, 1967, p. 511).

FORCES CAUSING MAGMA RISE

OVERBURDEN SQUEEZE

What, then, are the forces that cause magma to rise to the surface?
Two of the major processes are (a) overburden squeeze and (b)
buoyancy. The main force probably results from the difference in
pressure generated by the rock column and that generated by the

•The constants of the equation have changed with time and not all the data fit a
linear fonn. Some seismologists prefer the equation

logE= 11 .8 + 1.5M.

based on Gutenberg and Richter (1956). It should be evident that the calculation
of energy from seismic waves wiD undergo further revision as the parameters of a
shock are more fuUy understood.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Energetics and Periodicity 187

Guotemolo El Salvador Nicoroouo Costa Rico

18
•...
0
r
...
0
•;. <
0
0 ;;
! 17 Q

• "'
c;·
...•
0 10
'0
a
"'
~ 0.
~16
0
•
9
.a
c

0
0
..J
8

7
."'0
15
6

300 600
Distance, kilometer•

FIGURE 10-2 Smoothed energy released (left scale) for the period
1963-1971 from intermediate-depth (70-150 km) earthquakes ver-
sus distance along the Central American arc . Vertical bars are
orders-of-magnitude estimates of tons of volcanic products erupted
(right scale) during same period at active volcanoes. From Carr and
Stoiber (1973, p. 333 , Figure 2). (With permission of Bulletin Vol-
cano/ogique. )

magma column. The hydrostatic pressure at the base of a column may

be calculated from
p = pgx,
where ji = average density, g = acceleration due to gravity, and
x = depth. For Prock = 3.00 and Pmagma = 2.78 , the hydrostatic pressure
varies as shown in Figure I0-3A. At 60 km the pressure at the base of
the rock column is 18.0 kbar and that at the base of the magma column
is 16.7 kbar. The effect was displayed by Holmes (1945, p. 478), and a
similar diagram is given in Figure I0-3B . To equalize the pressure at the
base of the columns, the magma must rise 4.7 km (= 15,340 feet)* above
the surface . The highest volcanoes (not necessarily basaltic) rarely
*The highest continental extinct volcano in the world at present is Cerro Aconcagua,
Argentina, with an elevation of 6.96 km (22,834 feet). It is more than 11.83 km
(42,824 feet) above the Peru-Chile trench. The highest volcano measured from its
submarine base (in the Hawaiian trough) to its peak is Mauna Kea, Hawaii, with a
total elevation of 10.20 km (33 ,476 feet) .

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

188 GENERATION OF BASALTIC MAGMA

PretiUte, kilobort

50

60

Botoltic 1110111110
p. 2.78

FIGURE 10-3 A. Relationship, of hydrostatic pressure to depth in a column of magma

(average density, ;s = 2.78) and a column of rock (jj = 3.00). B, Schematic cross section
of a volcano. illustrating the rise of magma (M) above the reference surface (SFC)
necessary to equalize the pressure of the rock load on the magma chamber and the
magma column.

exceed this height. The implication that magmas are generated or

separated at a depth of about 60 km cannot be ignored; however, there
are many factors, most of which have not been evaluated quanti-
tatively, that would have to be considered before such a limit on magma
generation were applied.

BUOYANCY

An additional force in causing magma to rise to the surface is buoy-

ancy. The velocity of rise of a sphere of melt is given by Stokes' law:
_ 2ga 2 (p, - P2)
"- 97Jmed
'

where g = acceleration due to gravity, a = radius of sphere of melt,

p, =density of medium, P2 =density of melt, and 7Jmed =viscosity of
the medium. For a sphere of melt with a = I km (10 5 em), p, = 3.4
g/cm 3 , P2 = 2.8 g/cm 3 , g = 103 cm/sec 2 , and 7Jmed = 10 19 poises, then
v = 0.13 x 10-6 em/sec. The velocity is about 4 crnlyr, the average rate
of plate spreading at the midocean ridges. The implication is that

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Energetics and Periodicity 189

buoyancy could account for the rise of sufficient magma to satisfy the
present annual worldwide production.
The buoyant rise of magma has been related to the growth of salt
domes (e.g., Bayly, 1968, p. 117). Salt domes have long been con-
sidered products of a diapiric process (Nettleton, 1936); a new analysis
by R. 0. Kehle (personal communication, 1975), however, revives the
suggestion that overburden squeeze (Barton, 1936) may be responsible
for their puncturing overlying sediments. The pressure generated by
the lithostatic loading of a delta, for example, is transmitted through
underlying plastic salt to that part of the salt bed covered only by thin
sedimentary rocks, presumably in deeper water. The density contrast
between the deltaic material and ocean water is apparently sufficient to
generate a pressure difference adequate for extruding the more thinly
covered salt as domes.
Stokes' law can also be used to calculate the velocity required to
bring up nodules observed in alkali basalts. If the effective viscosity of
the magma is 103 poises and the maximum radius of a nodule is 2.5 em,
then v = 0.83 em/sec or about 3 m!hr. The rate of rise required to
support the nodules in the magma is not great compared with the
velocities observed for lava flows.

IMPORTANCE OF VISCOSITY

The viscosity of magma is obviously a most important parameter in

many aspects of its aggregation and flow. The viscosity coefficient is
the ratio of the shear stress to the strain rate, and when the ratio is a
constant, the magma has the behavior of a Newtonian fluid (Figure
10-4). When the ratio is not a constant, the magma exhibits non-
Newtonian behavior. If a critical value of the shear stress must be

1
II)
FIGURE 104 Schematic relation-
ship between shear stress and strain
rate for Bingham, pseudoplastic, and
Yield
stress Newtonian magmas.

Strain rote.-.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

190 GENERATION OF BASAL TIC MAGMA

achieved before flow can take place, the magma is classed as a
Bingham fluid. The behavior of a Bingham fluid is like that of house
paint: It can be spread easily with a brush, but stays essentially in place
without flowing when the brushing action stops. Shaw et a/. (1968)
found that the response of the lava in the Makaopuhi lava lake, Hawaii,
was closest to that of a Bingham fluid with critical shear stresses of 700
and 1,200 dyne/cm 2 (0.7 x 10-3 and 1.2 x 10-3 bar, respectively) at
two different depths in the lake. The respective viscosities at the two
positions were 7.5 x 103 and 6.5 x 103 poises. The energy required to
move magma of the Bingham type will be obviously greater than that to
move the Newtonian type. The yield stress also varies with composi-
tion, according to Hulme (1974, p. 378). He indicated that basaltic
magmas yield at a much lower value of stress than andesitic magmas
(Figure 10-5). The loss of volatiles may have led to an overestimate of

K. (measured)

40 45 50 55 60
Si02, weight percent

FIGURE 10-5 Yield stress of lava versus SiO. content calculated from depth of flow,
slope, and density . Crystal and gas content and temperature are unknown.
E = Etna, 1966 ML = Mauna Loa, 1942
H = Hekla, 1947 OS =O-Shima, 1951
K =Kilauea, measured (113o•c; -25% P = Paricutin, 1945-1946
crystals; 2-5 percent gas bubbles) T = Teide
Ml = Mare Imbrium TdC =Tristan da Cunha
After Hulme (1974, p. 378, Figure 16). (With permission of The Royal Astronomical
Society.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Energetics and Periodicity 191

the viscosities measured by Hulme on the basis of levee height, ground
slope, and lava density.
Emphasis has been placed on the eutecticlike character of basaltic
liquids, and it is not difficult to accept the idea of Feild and Royster
(1918) that viscosity is minimum at a eutectic. With disregard for the
structural change of liquid with composition, one might argue that the
eutectic composition should have the maximum viscosity because it
has the lowest temperature. Other studies (Herty et a/., 1930), how-
ever, suggest that the nature of the liquidus surface has no effect on the
viscosity. A high yield stress in a liquid of the eutectic composition
would indeed make the removal of small amounts of melt difficult.

PERIODICITY
The periodicity of eruption is one of the keys to predicting the outbreak
of a volcano. Some workers have assumed that volcanic activity is a
random process in both space and time (Verhoogen, 1946, p. 770;
Wickman, 1966, p. 291 ). Other investigators believe there are specific
patterns to volcanic activity in space and time. The great outpourings
of basalt were associated by Du Toit (1937, pp. 94, 174) with the
periodic breaking up of the lithospheric plates. On a shorter time scale,
one investigator (Dubourdieu, 1973) examined the records of active
volcanoes throughout the world using data from the sixth century
forward and [Link] that volcanic eruptions have the same fre-
quency as seismic activity, approximately 4 years. McBimey et a/.
(1974) presented evidence for a possible 5-m.y. cycle in the Oregon
Cascades and suggested that the "same episodes occurred in unison
over a large part of the earth."

PAUCITY OF DATED BASALTS

The major difficulty in assessing periodicity is the small number of

volcanoes studied. A small sample would tend to result in the descrip-
tion of volcanism as intermittent. episodic, or spasmodic. Some ap-
preciation of the worldwide frequency of magma production may be
gained by examining the ages of igneous rocks. Older basalts are dated
with great difficulty, and only few data are available; however, the ages
recorded in minerals from all rocks, although subject to some correc-
tion, may serve the purpose (Figure 10-6A). A crude periodicity of
magma generation is evident, and it is assumed that basaltic magmas
are also generated in large amounts over specific time periods. An
estimate by Engel eta/. ( 1965) of the periodicity and volume of basalts

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

192 GENERATION OF BASALTIC MAGMA

FIGURE 10-6 A. Frequency of mineral .. age" from 0 to 3000 m.y. before present from
worldwide localities. compiled by Gastil ( 1960. p. 5, Figure 1). (With permission of the
American Journal of Science.)
B. Volume of continental tholeiitic basalt extruded worldwide, estimated from geolog-
ical mapping by Engel eta/. ( 1965, p. 729. Figure 4). (With permission of the Geological
Society of America.)

based on geologic mapping is given in Figure 10-68. Menard (1964, p.

87 ff.) called attention to the 10,000 or more volcanoes in the Pacific
Basin, most of which lie in the western or southwestern sector,
emphasizing their maximum growth in periods during the late
Mesozoic through early Cenozoic and again from the late Tertiary to
the present. The duration of a single volcano is probably in the range
1-5 m.y.; however, a magma source may be of much longer duration,
according to the heat- and mass-transfer calculations of Shimazu
(1961).

SPACING OF VOLCANOES

Some volcanoes, especially in island arcs, also appear to exhibit a

periodicity in spacing. Marsh and Carmichael (1974, p. 1202) plotted
the distance between neighboring volcanic centers along the Aleutian
Islands, Alaska Peninsula, and Cascades and obtained a spacing of
about 70 km (Figure 10-7). They attributed the spacing, after Ramberg
(1972), directly to the growth of perturbations on a gravitationally
unstable layer deformed in a sinusoidal fashion (Figure 10-8). The size
of the conduits is probably of the order of 25 km in diameter (Koyanagi
and Endo, 1971).
Based on laboratory experiments, Ramberg (1972, p. 52) believed
that the critical feature causing the episodic character is that the rate of

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Energetics and Periodicity 193

FIGURE 10-7 Histogram of distance between pairs of volcanic

centers in the Aleutian Islands, the Alaska Peninsula, and the
Cascades. From Marsh and Carmichael (1974, p. 1202, Figure 8).
(With permission. Copyrighted by the American Geophysical
Union.)

ascent of the perturbation in the buoyant layer increases at a rate in

excess of a linear rate with the thickness of the layer. In other words,
as the buoyant layer thickens with continued heating, for example, the
rate of rise increases and the sinusoidal form amplifies rapidly. Mter
the diapir is detached by some unspecified process, there is a period of
quiescence until sufficient heating produces a renewed thickening of
the layer undergoing melting, and the process repeats.
Spacing of volcanoes may be controlled at depth by the size and
shape of the source region of each vent. For an extreme example, the
volume of the Mauna Loa, Hawaii, edifice is about to:> km 3 , and the
source region will be 3.3 x 105 km 3 if the magma type was generated
from 30 percent partial melting. If the magma was produced from 5
percent partial melting, then 2 x 106 km 3 of parental material will be

•, iI iI
I I
iI II FIGURE 10-8 Schematic cross section of

~:! ~;I
mantle, iUustrating gravitational stabiliza·
tion of a low-density melt by forming equaUy
IJI
II
montle
~~ II
spaced diapiric conduits. From Marsh and
I I I I II Carmichael (1974, p. 1202, Figure 9). (With

--J,~~~
movmo
permission. Copyrighted by the American
Geophysical Union.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

194 GENERATION OF BASAL TIC MAGMA

involved. The shape of the source region is obviously unknown;

however, if the magma is withdrawn from a cylindrical region 50 km
thick, then the nearest volcano of equivalent dimensions will be about
90 km away for 30 percent partial melting and about 225 km away for 5
percent partial melting. Although no overlap is assumed, continued
melting of an already partially depleted region is more likely. The
average spacing determined by Marsh and Carmichael, 70 km, implies
that cylindrical source areas would have to be 8. 7 or 52 km thick for
partial melting of 30 percent or 5 percent, respectively, for volcanoes
with an average volume of 104 km 3 • Marsh and Carmichael (1974, p.
1204, Figure 10) envisaged a horizontal, cylindrical source region of
small diameter supplying the vents. Cylindrical source areas are not
considered to be very realistic, and the reader may wish to try other
dimensions and geometric shapes of source regions more compatible
with his preferred mechanism of magma generation.

AGE AND SPACING RELATED?

The evolution of the Hawaiian Ridge also appears to be episodic in

space and time. Jackson et al. (1972) have shown quantitatively that
the .Hawaiian volcanoes increase in age from east to west and are
spaced in a systematic way (Figure 10-9). An early idea attributed the
southeastward extension of the Hawaiian chain to a slowly propagating
fracture that intercepted new regions of undepleted mantle undergoing

.
•
0
5

~ 4
~ Nihoo Nilhau
g 3 -o o-
.
~

1- 2
~

Distance from Kilauea, kilometers

FIGURE 10-9 Age (K-Ar) of tholeiitic volcanism versus distance from Kilauea for the
Hawaiian chain. Lines connect ages of adjacent shields. From Jackson et a/. (1972,
p. 608, Figure 3, inset). (With permission of the Geological Society of America.)

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Energetics and Periodicity 195

FIGURE 10-10 Schematic cross section of volcanic island chains

illustrating generation of a new volcano as a result of the motion of
the oceanic plate relative to the magma source. After Wilson (1963,
p. 93). (With permission of Scientific American, Inc.)

partial melting (Betz and Hess, 1942, p. 109 ff.). With the advent of
plate tectonics, Wilson (1963) suggested that sequential eruptive activ-
ity was due to the passage of the Pacific Plate over a "hot spot" in the
mantle from which the magma was derived (Figure 10-10). Jackson and
Wright (1970, p. 426) argued that the hot spot would soon be depleted,
and the relative constancy of magma composition could not be ex-
plained. Therefore, they suggested that the source as well as the plate
moved and accepted the idea of a slowly propagating fracture. The hot
spot was viewed by Morgan (1971) as a "plume" of hot material rising
from the mantle. Jackson et a/. (1972) defined a "melting spot,"
without regard to process, of about 300 km in diameter from which
tholeiitic magma could be supplied to the moving plate as the spot itself
progressed.*
Many of the seamount volcanoes in the Pacific Ocean appear to
occur in groups of about 10 to 100, and the center of volcanism usually
migrates along a lineation (Menard, 1964, pp. 76-79). The sequential
character of eruption is considered a general feature, and relative
motion, whether mainly in a propagating mantle source or mainly in the
lithosphere plate, seems to play a major role.

CONVECTIVE HEAT TRANSFER MODELS

JOLY-COTTER MODEL

One of the earliest ideas on the periodic nature of magmatism was

proposed in 1923 by Joly (1930, pp. 192-193) and elucidated by Cotter
"The terms "plume" and "melting spot" are closely analogous in concept to the terms
"ascending column" and "hot spot" of Holmes (1945, p. 410, 483).

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

196 GENERATION OF BASALTIC MAGMA

(1924). Joly considered the parental material to be crystalline basalt, its

melting curve being intersected by the geothermal gradient at a certain
depth. As the basalt melts, a quiescent period ensues because of the
time required to obtain the enthalpy of melting. After reaching an
"advanced stage" of melting, the weakened material breaks down
from "the disruptive action of tidal forces" and convection begins. As
the melt rises adiabatically, its temperature is above the melting point
appropriate for the reduced pressure, and the liquid begins "to melt
away and to erode the solid rock above it. ... " The molten zone is
advanced upward as the descending cooled melt crystallizes at the
bottom of the liquid mass. The rise of the molten zone is terminated
mainly by heat losses to the surface due to thermal conduction from the
molten zone. After a sufficient period of radioactive heating, thermal
conditions are restored, the basalt is again brought to the melting
temperature at depth, and the cycle is repeated. Joly's theory does not
depend on the starting material's being basaltic, but it would fail if a
more refractory layer intervened. If the starting material were garnet
peridotite, any zone melting process would have to be inefficient so
that basaltic fractions would remain after each successive pass.

TIKHONOV ET AL. MODEL

Joly's theory for thermal cycles was reconsidered by Tikhonov eta/.

(1970), who emphasized the high heat transfer in the molten layer itself.
In one numerical model, they suggested that radioactive heating of the
earth produced the first primordial melt at a depth of 460 km about
1.8-2.4 b.y. ago (Figure 10-11). The number of periods is four, and the
mean duration is about 300-500 m.y. With different heat-transfer rates
and thermal conductivities, a greater number of periods were obtained
in other models. With suitable choice of parameters, a model could be
constructed to fit the seven (?) periods of worldwide volcanism begin-
ning about 3.5 b.y . ago with a mean interval of about 400 m.y.

ROLE OF VOLATILES
Volatiles have always been high on the list of factors believed to be
responsible for the periodic character of volcanism. The development
of pressure in magmas as a result of crystallization was considered by
Morey (1922), on the basis of his studies in the K20-Si02-H20 system,
to be the cause of the paroxysmal eruptions, especially of salic mag-
mas. The key point of his argument was that in the presence of an
excess of volatiles the pressure rises with cooling along the univariant

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Energetics and Periodicity 197

Tlmt, 109 yr before prtttnt
0 I 2 3

FIGURE 10-11 Melting periods in the mantle as a function of

depth, according to the mathematical treatment of Tikhonov eta/.
(1970, p. 329, Figure 3). The calculation involves a solution for two
moving boundaries where one interface is melting and the other is
crystallizing. Intensive heat transfer is assumed in the melted layer.
The lower interface "catches up" with the upper interface at 130-
150-km depth. The four periods vary in mean duration from 300 to
500 m.y. (With permission of Elsevier Scientific Publishing Com-
pany.)

curve describing the equilibria between crystals, liquid, and gas (Figure
10-12). It is necessary, therefore, for a magma to reach a condition of
univariance before high pressure can develop and "break" the
chamber. The resealing of the chamber after eruption and continued
cooling again causes the gas pressure to build up, and the chamber is
"fractured" with the extrusion of the magma. The repetition of these
events was believed to be one of the main causes of periodic eruption
with relatively short repose times. The large number of components in
a magma makes the attainment of such a univariant condition unlikely
(Yoder, 1958). An additional difficulty is that high-pressure phases
such as pyrope (Figure 10-12) and jadeite (Boettcher and Wyllie, 1969,
p. 899, Figure 9) have a positive dT/dP in the presence of excess H2 0.
That is, pyrope and jadeite cool along the hydrous melting curve with
decreasing pressure. [It should be recalled that melting of the common

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

198 GENERATION OF BASAL TIC MAGMA

40~--~--~~--~--~~--~--~----~--~----~--~

Temperature, •c
FIGURE 10-12 Effect of H20 on the melting of the principal end-member minerals in
garnet peridotite.
Forsterite (Fo)-H2 0: Kushiro and Yoder (1969, p. 155, Figure 52).
Enstatite (En)-H20: Kushiro et a/. (1968b, p. 1690, Figure 2); coexisting forsterite
ignored.
Diopside (Di)-H20 : Yoder (1965, p. 87, Figure 12); Eggler (1973, p. 458, Figure 17).
Pyrope (Py)-H 2 0: Mysen (1975; unpublished data).

silicates in the presence of a fixed partial pressure of water generates a

liquidus with a positive dT/dP, whereas the solidus remains the same as
in the presence of excess water (see Figures 4-16 and 4-17).] This
relationship is also observed at high water pressures in natural spinel
lherzolites (Figure 4-20).
Explosive volcanism is more likely the result of a magma reaching
saturation with regard to volatile components. Saturation may be

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Energetics and Periodicity 199

reached as a result of crystallization, the gas content of the magma
increasing as crystals form. (If hydrous phases are formed, the water
content of the magma may increase or decrease, depending on the
initial amount of water.) In addition, Yoder (1965) illustrated with the
experimental results on the diopside-anorthite-H 2 0 system that re-
lease of pressure may also cause an undersaturated magma to reach
saturation. The liberated gas phase, requiring a much larger volume
than when dissolved, acts as the propellant in the explosive ejection of
a magma exsolving gas. The vesiculation of the magma may take place
quite independently of the crystal content, yet crystals may in fact
accelerate the vesiculation process. It is not necessary, therefore, for a
magma to reach a condition of univariance for explosive eruption.
Most basaltic magmas arrive at the surface quiescently, and the volatile
content is exceptionally low. An explosive basalt (?) eruption such as
that at Tambora, East Indies, in 1815 is indeed a rarity, and there
volatiles acquired by the magma from the ground and sea were con-
sidered by some investigators to be the cause. It seems most unlikely
that ground water or sea water would be absorbed by a hot magma-
the water was probably a primary constituent of the magma. On the
whole, the role of volatiles in the periodic eruption of basaltic lavas
seems to be minor.

ROLE OF SUBSIDIARY RESERVOIR

The periodicity of eruptions within a single volcano suggests a process

dependent on the refilling of a reservoir. Hekla in Iceland exhibits
sudden bursts of activity, with changes of magma composition reflect-
ing depletion of a reservoir (Figure 10-13). Mt. Vesuvius in Italy has
similar behavior, and it also appears to have a constant rate of magma
production (Wickman, 1966, p. 361). In other volcanoes having a
subsidiary reservoir (e.g., Kilauea, Hawaii) the periodicity may be
dependent more on the events taking place at the site of magma
generation than on the physical properties of the subsidiary reservoir.
It is not known whether all volcanoes have a subsidiary reservoir that
governs the periodicity or whether the periodicity depends on the rate
of production of magma at the source. The critical parameters have not
yet been identified, but present opinion relates the short-term periodic-
ity to the physical properties of the volcanic edifice that undergoes
tumescence and rupture. Shaw and Swanson (1970, pp. 287-288) note,
however, that the "viscous collapse" of a magma chamber does not
appear to be rapid enough to maintain constant fluid pressure.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

200 GENERATION OF BASALTIC MAGMA

65~~--------~---------r--------~--------~

1947

I
1766 1845 I
I
1~93
16~
.. ......
.·. .·
..· .: ...·
·tlII
I
II

I
I
I
11.

...· ...:-

50~~--------~--------~--------~------~
1600 1700 1800 1900 2000
Year
FIGURE 10-13 Relation of Si02 content of initial and final analyzed products of Hekla.
Iceland. eruptions (solid lines). Quiescent intervals are represented by dashed lines.
Dotted lines indicate periodicity based on prehistoric tephra layers and extrapolation.
From Thorarinsson (1954. p. 43. Figure 7). (With permission of Societas Scientarium
lslandica.)

RUNAWAY MECHANISMS POTENTIALLY PERIODIC

Mention has been made of the runaway character of shear melting and
melting associated with a minimum in thermal conductivity. In each
case the heating increases at an accelerating rate as the local viscosity
or thermal conductivity decrease. The heating "runs away" until the
energy source is depleted, convective heat transfer exceeds the pro-
duction, or the liquid is removed from the generation region. After
some period of quiescence the runaway character may again be ini-
tiated and the process repeated.
In summary, several mechanisms have been proposed that may
become periodic locally; however, only the Joly-Cotter scheme, suita-
bly adjusted for garnet peridotite, appears to provide for worldwide
magmatic events with long-range periodicity. The catastrophic rup-
tures that lead to plate separations may result from the mantle convec-
tion evolving from worldwide magma generation. For this reason the
principle of the Joly-Cotter model may be best suited to present-day
concepts of periodic mantle dynamics.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

11 Summary
Overview

The opportunities for research on the problems of basaltic magma

generation are indeed manifold. In view of the great need for funda-
mental data and methods to collect the necessary data, it is useful to
present an overview of the present state of knowledge. Whereas an
attempt to provide an evaluation at this stage of uncertainty may yield a
prejudicial view, a framework is helpful in planning research. Adequate
tests of conclusions soon indicate more appropriate directions to
pursue.

WHAT IS MELTED TO YIELD BASALTIC MAGMA?

The seismic data indicate that the crust and mantle are crystalline for
the greater part. The composition and mineralogy of that crystalline
material depend considerably on how the earth formed. If the earth
formed by the relatively cool accretion of condensed matter, differen-
tiated relative to other planets and heterogeneous, then one may
visualize a mantle consisting of masses, of a wide range of size, age,
and composition, welded together by metamorphic processes. Melting
due to the conversion, albeit inefficient, of gravitational energy to
thermal energy may have occurred only in the core and lower mantle.
On the other hand, the crust and mantle may have resulted from the
differentiation of a molten mass, initially homogenized by convection.
201

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

202 GENERATION OF BASALTIC MAGMA

The randomly heterogeneous earth and the continuously layered

homogeneous earth represent two extreme distributions of mantle
material, and many variations within these extremes have been as-
sumed. Many of the subsequent concepts of magma generation im-
plicitly and explicitly depend on the assumptions regarding the origin of
the earth.
It was concluded that the assemblage of minerals in the mantle most
likely to yield magma of basaltic composition is garnet peridotite; this
mineral assemblage could vary widely in the proportions of phases but
must necessarily remain restricted in the kinds of major phases pres-
ent. The distinct petrographic provinces, that is, the variability in
composition between regions, are attributed in part to the presence or
absence of minor phases such as phlogopite as well as variation in the
type of solid solution within the major phases of the mantle. Another
major factor in the development of a petrographic province is no doubt
the tectonic environment and the associated mechanism(s) for magma
generation. It is not known whether parts of the upper mantle also
differ in age as well as composition. The absence on the earth of a
historical record from 4.6 to 3.8 b.y. leads one to depend on the
interpretation of processes recorded on the moon. The severe impact-
ing believed to climax at about 4.0 b.y. may have churned up the crust
and mantle to the depths from which magmas are now believed to be
derived. That is, early vertical differentiation or inhomogeneity in the
upper mantle and crust may have been wiped out by the homogenizing
effects of impacting.
It is not possible to specify which garnet peridotite is most represen-
tative of the upper mantle because the samples available are believed to
be more or less depleted of basaltic constituents. Adequate tests of the
degree of depletion have not been carried out; however, relatively
lower Cr20JA120 3 and higher FeO/MgO may be suitable guides for
relatively undepleted parental material (Dickey and Yoder, 1972; Boyd
and Nixon, 1975, p. 452). Furthermore, the relationship between
derivative eclogitic liquids-that is, liquids that would crystallize to
eclogite mineralogy-and garnet peridotite is obscure because the
eclogite samples are not clearly distinguishable as cumulates, liquids,
or remelted basalts. Attempts to relate basalts, as representative
derivative liquids, to garnet peridotite fail because most basaltic mag-
mas are believed to have undergone considerable change en route to
the surface. Some ambiguity is introduced, therefore, when their
relations at high pressure are analyzed. In addition, experimental data
are lacking on the relevant partition coefficients of key elements at
various pressures.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Summary Overview 203

WHERE DOES MELTING OCCUR?

The drop in velocity and increase in attenuation of seismic waves
suggest that some broad regions of the earth's mantle are in a partially
(5~ percent) molten state at depths of 70 to 150 km. Other regions
exhibit possible molten regions at 300-400 km, whereas some old
continental cratons and the older oceanic regions show no evidence of
partial melting at depth. The shallowest regions of magma are pre-
sumed to occur under the midocean ridges, yet there is no seismic
evidence. The initial site of generation may have been several hundred
kilometers below, the liquid content of the magma increasing en route
upwards.
Estimates of the local geothermal gradient are related to the
beginning-of-melting curves of mineral phases common to garnet
peridotite, and their intercept signifies a possible zone of melting.
Unfortunately, it is necessary to specify the volatile content and
composition in order to fix an appropriate beginning-of-melting curve.
Vast regions of the mantle would be molten if H20 were present in
amounts in excess of that needed to saturate the liquid. Because the
seismic evidence indicates only small amounts of partial melt where
melting ensues, the volatile content is assumed to be low. The ratio of
C02 to H 2 0, for example, could be varied to achieve the desired
amount of melting, but the depth range of melting is still considerable.
The source of volatiles in the mantle and the extent of depletion in the
early stages of ocean formation become important issues.
The concept of a geotherm is not a persuasive constraint after the
processes of melting are examined. The ideas of convection and
buoyancy, for examples, focus on the adiabat. The rise of a mass to
lower pressures without loss of heat was shown to be adequate for
melting. The depths where melting could begin appear to be confined not
by fixed relationships between a geotherm and the melting curve, but
by the intersection of the family of adiabats from the local thermal
gradient with the melting curve.

HOW DOES MELTING BEGIN?

A wide array of mechanisms have been proposed for the initiation of
melting. Concepts involving the relief of tensional or compressive
stresses require stresses greater than those believed to exist in the
mantle. For mechanisms depending on the conversion of mechanical
energy to heat, a broad range of heat-producing capacity is assumed.
Whatever the capacity, these mechanisms appeared to fail with the

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

204 GENERATION OF BASALTIC MAGMA

onset of melting. Brittle failure and high friction are not the properties
of a rock mass near its melting temperature.
Places of abrupt change in composition or phase change are con-
sidered reasonable regions for melting to begin. The addition of vol-
atiles to rocks is found to be a suitable method for initiating melting at
temperatures considerably below those for melting in the absence of
volatiles. Perturbations in specific properties such as density or ther-
mal conductivity may lead to the initiation of melting. Because of the
decrease in thermal conductivity and viscosity with temperature, both
properties provide a runaway character to some methods. Although
each of these methods has its advantages, the most acceptable pro-
cesses still appear to be radioactive heating and adiabatic rise.
The enthalpy of melting was found to be relatively small compared
with the total amount of heat required to bring a rock up to the melting
temperature. New estimates ofthe enthalpy of melting, when corrected
for pressure, are greater than previously indicated. Millions of years
are required to produce small amounts of melt even if the entire local
radioactive heat production is consumed. Alternatively, the heat re-
quired for melting could be obtained from the adiabatic rise of a hot
mass. It was found that the depth of initiation of the rise would have to
be as great as 160 km for complete melting of eclogite, a material
equivalent in composition to basalt. On the other hand, the release of
liquid before complete melting would not require such great depths of
initiation. It would, however, be necessary for the eclogite to melt in a
eutecticlike fashion and for all fractions to be basaltic in composition.
The rise of eclogite is a most unlikely event because of its high density,
so a parental material such as garnet peridotite is preferred. The partial
melting of garnet peridotite yields a range of basaltic liquids, the type
being dependent on depth of separation, amount and proportion of
volatiles, and other factors. The adiabatic rise of garnet peridotite to
shallow regions with appropriate metamorphic changes, fractional
melting, and separation of liquid appears to account best for the
observations in the midocean ridge environment.

WHY DO BASALTIC MELTS PREDOMINATE?

Large volumes of relatively homogeneous basaltic magma have been
produced throughout geologic time. The melting process must there-
fore yield a uniform product consistently and repetitively. The
physicochemical properties of garnet peridotite outlined in the labora-
tory suggest that a substantial amount of liquid may be derived from it
at the beginning-of-melting temperature. The liquid may be removed in

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Summary Overview 205

one batch or fractionally as it is formed . Zone melting, which requires
time-dependent, nonequilibrium processes, does not appear to be
appropriate for the extended times involved in magma generation.
Fractional melting is the more likely process because the local dynamic
motions of the mantle readily displace a liquid from the site of origin.
Fractional melting at an invariant point, which tends to be the lowest
temperature (and pressure) for melting to begin, yields a liquid of
relatively uniform composition in major components but not in trace
elements. The proportion of phases in the parental material does not
influence the major element composition of the liquid, but it greatly
influences its trace element content. Fractional melting of the parental
material ceases with the loss of a phase, resuming again when the
temperature reaches that of the beginning of melting of the diminished
assemblage. The liquids withdrawn at each invariant point may be
collected into uniform batches, differing uniquely from each other in
bulk composition without intermediate compositions. It was demon-
strated that the first batch generated at high pressures at the lowest
temperature invariant point for the garnet peridotite system was, in a
broad sense, basaltic in composition. It is for this reason that basalts
are the most common magma type on earth.

WHAT FACTORS DETERMINE THE KINDS OF

BASALT?
A study of the thermochemical relations of the major end-member
minerals in basalts at I atm reveals that there are at least three invariant
points involving forsterite separated by thermal maxima. These in-
variant points are analogous to the three major types of basalt. The
eutecticlike behavior in the laboratory of natural basalts at I atm
reflects the character implied by the analogous invariant points. The
major rock types derived by crystal fractionation from basalt can be
related to the univariant curves leading to lower temperatures and
other subordinate invariant points. These observations demonstrate
that basalts and their derivatives result from controlled physicochemi-
cal behavior.
At high pressures all basalts convert to eclogite, and a new set of
phase relations obtains. These relations have not yet been determined
in detail; present evidence suggests, however, that a single parental
assemblage could yield the various liquids of basaltic composition with
suitable changes in pressure or volatile composition. In general, high
pressure and C02-rich volatiles produce alkali basalts, whereas low
pressure and H20-rich volatiles produce tholeiitic varieties. The varia-

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

206 GENERATION OF BASALTIC MAGMA

tion in trace elements may be attributed to different proportions of

phases in the parental assemblage, the degree of melting, and the
presence of various minor phases. It is essential to determine the depth
of separation of the liquid from the parental material and the amount and
ratios of the volatile constituents. Other factors, such as the partial
pressure of oxygen, also influence the nature of the liquid derived from
the partial or fractional melting of garnet peridotite.
The absence of olivine in eclogites produced at high pressures from
basalts would appear to cast doubt on the derivation of basaltic liquids
from an olivine-bearing parental material such as garnet peridotite. The
absence of olivine can be accounted for either by the removal of phases
from the liquid during its ascent to the surface or by the existence of a
reaction relation involving olivine and liquid. An additional reaction
relation involving orthopyroxene and liquid is necessary to account for
the production of an eclogitic liquid from garnet peridotite. Both
reaction relations have been demonstrated in the laboratory. The
sequence of reactions is believed to involve the removal first of olivine
and then of orthopyroxene, with decreasing pressure; however, the
exact compositional trend of the liquids as a result of these reactions at
high pressures in the absence of volatiles has not yet been established.
Removal of olivine tends to increase the normative quartz content of
the liquid, whereas removal of orthopyroxene tends to increase the
normative nepheline or larnite content of the liquid. There is the
prospect of generating larnite-normative liquids (potential melilite
nephelinites at low pressures) if, with decreasing pressure, the or-
thopyroxene reacts out first.
The proportions of garnet and clinopyroxene in garnet peridotite are
a measure of the degree of depletion of the parental material of its
basaltic fraction. From the viewpoint of partial or fractional melting,
the order in which clinopyroxene or garnet is consumed in the liquid
influences the kind of basaltic liquid produced. Phase-equilibrium
studies suggest that clinopyroxene may be the first phase consumed in
the partial melting of garnet peridotite at high pressures, whereas
garnet is consumed first at moderate pressures for compositions ap-
propriate to basalts. The norms of natural garnets and clinopyroxenes
from garnet peridotites clearly illustrate the great contrast in chemical
contribution of these phases to the melt.

HOW DOES THE MELT FORM AND AGGREGATE?

The first melt most likely appears at the common junction of all phases.
Because all phases are in equilibrium with the initial liquid, the liquid is

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Summary Overview 207

not confined to the initial pore. The liquid is presumed to wet all
crystals and permeates all grain boundaries, forming a "liquid-bonded
aggregate ." Furthermore, the concentration of impurities and the
low-temperature melting residua occurring at the grain boundaries of
the major phases facilitate the melting process. A significant drop in the
velocity of transmission of seismic waves occurs at this stage. The
aggregation is aided by the thermal gradient and shear stresses on the
host rock. In general, the liquid aggregates preferably in regions
perpendicular to the principal minimum stress and may be kneaded out
under very low stresses.
The region of magma generation is essentially defined by the solidus
of the rock undergoing melting. Because the solidus varies with pres-
sure, the boundary will not necessarily be an isotherm. The nature of
the heat source of the magma, whether external heat transfer or
internal heat production, will determine, to a large extent, the initial
size and shape of the melting region. The boundary , sharply defined by
the appearance of liquid, may be diffuse in regard to some physical
properties, yet sharp in its response to seismic wave transmission. A
plastic envelope no doubt surrounds the chamber as a result of heat
transfer.
Melting takes place with a large increase in volume . Space must be
provided for the volume increase by tumescence of the region, filling of
pore spaces in adjoining host rock, or rupture of the host rock. If space
is not provided, the magma will freeze because the melting temperature
generally increases with pressure. The presence of melt lowers the
strength of the containing rock, and flaws in the plastic envelope are
expanded by a magmafracting process. Earthquakes may accompany
the magmafracting process as the magma enters the brittle region
beyond the plastic envelope.
Differential stress tends to raise the mean stress on a volume of rock.
The melt does not freeze under the small increase in pressure because
(a) the ratio of compressibility to thermal expansion is favorable, (b)
the low thermal conductivity of the host slows the rate of release of the
enthalpy of melting, or (c) the presence of volatiles may enhance the
melting under an increase in pressure.

WHAT ARE THE FORCES CAUSING MAGMA

RELEASE?
The main force causing magma release probably results from the
pressure of the rock overburden. The loss of strength of the partially
molten rock initiates the magmafracting process, and the liquid first

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

208 GENERATION OF BASAL TIC MAGMA

migrates through the plastic envelope and then propagates fractures to

the surface. The density difference between the containing rock and
the anastomosing magma column results in a pressure difference. The
gravitational force of the rock column on the magma, exceeding that of
the magma column itself, causes the magma to rise.
If an appropriate viscosity of the mantle is assumed, the rate of rise
of small (1-km-diameter) liquid spheres by buoyancy could account for
the present worldwide production. Because heat losses from the
magma en route to the surface and extensive lateral penetration of the
crust by sills terminate many potential extrusions, large liquid masses
must be periodically released from the zone of partial or fractional
melting. Depending on the rate of rise, the masses may remain intact in
batholithic proportions or disintegrate into a series of smaller masses.
Overburden pressure and buoyancy appear to be adequate forces to
account for the rise of magma; however, they are probably not the
primary forces driving the large-scale convection in the mantle envis-
aged by some investigators. Tidal forces may trigger volcanic action,
but the major disruptions that initiate the worldwide geological revolu-
tions have not been identified.

WHY ARE VOLCANIC ERUPTIONS PERIODIC?

Since the beginning of the earth's history, there have been at least
seven major periods of igneous activity based on mineral "age."
Individual volcanoes are themselves periodic and appear to have a finite
lifetime of the order of 1-5 m. y. The great floods of basalt have appeared
in a relatively short time-span of a few to 20 m.y. The accurate dating
of older basalts is yet to be perfected, and many more measurements
are needed before major conclusions can be drawn.
The Joly-Cotter mechanism of convective heat transfer, unsatisfac-
tory in several details, appears to provide the only available long-time-
scale process for the igneous activity associated with geological revolu-
tions. The short-period eruptions of individual volcanoes may be
closely related to the continuous filling of a subsidiary reservoir that
swells and breaks periodically. The thermal effect on some properties,
such as viscosity and thermal conductivity, leads to runaway condi-
tions that may be periodic as long as the energy source persists. The
recovery time may be long because of the slow rate of heating and the
relatively large enthalpy of melting. Recurring explosive eruptions may
be related to the volatile saturation of a magma as a result of cooling or
pressure release.
The mechanisms of magma generation have, no doubt, changed with

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Summary Overview 209

geologic time. The early crust, no remnant of which has been iden-
tified, was probably andesitic because of the hydrous melting attending
the degassing of the earth to form the oceans. The more recent increase
in alkali basalt generation during the last 250 m.y. (Engel et al., 1965, p.
729) also suggests a style of magma generation different from that
producing the midocean upwelling of magma now considered essential
to plate tectonic theory. Poldervaart (1962, p. 10) was unable to find
any record of alkali basalts older than 1 b.y . and concluded that
tholeiite is therefore the ' 'primitive'' magma. The decline in production
of tholeiitic magmas may be reflecting the general decrease in antici-
pated heat production due to decay of radioactive elements.

CONCLUSION
The e:xciting questions now amenable to experiment ensure an expand-
ing effort in the study of magma generation. There is much for the
physicist, thermodynamicist, and physical chemist to ponder, provided
the questions are framed within the constraints deduced by the field
geologist, petrologist, geophysicist, and mineralogist. Recognition of
the relevant parameters is indeed a critical contribution to the problem.
Skepticism of the model approach to earth problems is warranted
because many key parameters have not been included. One can readily
sense a new revolution in geologic thought coming close on the heels of
the plate-tectonic awakening.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Appendix: Conversion Factors

ENERGY LENGTH
I joule = 10 erg
7 I km = 105 cm
I erg = 0.23901 X J0- 7 cal
I cal = 4. 1840 X 107 erg PRESSURE
I erg = I g cm 1/sec 1
I bar = to• dyne/cm1
FORCE I bar = to• glcm/sec1
I mm Hg = I torr = 1.33322 x J0-3 bar
I dyne = I g cm/sec 1
TIME
GRAVIMETRIC FACTORS
I yr = 3.1536 x 107 sec
X 0.89981 =FeO
X 1.1113 = Fts03 VELOCITY
x o.m3t =Fe
X 0.69944 =Fe I km/yr = 3.171 x to-s em/sec
X 0.83015 =K

VISCOSITY
GRAVITATIONAL ACCELERATION
I poise = I glcm/sec
Nonnal g.. = 978.032 cm/sec 1
I Gal = I em/sect

RARE-EARTH NORMALIZATION

Average rare-earth element composition in parts per miUion of chondrites used in

normalizing abundances of those elements in rocks and minerals

Schmitt et al. (1963, 1964) Haskin et al. (1968)

Element (average of 20) (average of 9)
La 0.30 ± 0.06 0.330 ± 0.013
Ce 0.84 ± 0.18 0.88 ± 0.01
Pr 0. 12 ± 0.02 0.112 ± 0.005
Nd 0.58 ± 0. 13 0.60 ± 0.01
Sm 0.21 ± 0.04 0. 181 ± 0.006
Pm
Eu 0.074 ± 0.015 0.069 ± 0.001
Gd 0.32 ± 0.07 0.249 ± 0.011
1b 0.049 ± 0.010 0.047 ± 0.001
Dy 0.31 ± 0.019
Ho 0.073 ± 0.014 0.070 ± 0.001
Er 0.21 ± 0.04 0.200 ± 0.005
Tm 0.033 ± 0.007 0.030 ± 0.002
Yb 0.17 ± 0.03 0.200 ± 0.007
Lu 0.031 ± 0.005 0.034 :!: 0.002
211

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

References

Adams, L. H. (1922) Temperature changes accompanying isentropic , isenergic, and

isenkaumic expansion, J . Wash . Acad. Sci., 12, 407-411.
Adams, L. H., and R. E. Gibson (1926) The compressibilities of dunite and of basalt
glass and their bearing on the composition of the earth, Proc. Natl. Acad. Sci.
U.S.A ., 12, 275-283 .
Adams, L. H., and E. D. Williamson (1923) The compressibility of minerals and rocks
at high pressures, J. Franklin lnst., 195, 475-529.
Adams, L. H., and E. D. Williamson (1925) The composition of the earth's interior,
Smithson . Rep. 1923, 241-260.
Akella, J. (1974) Solubility of Al20 0 in orthopyroxene coexisting with garnet and
clinopyroxene for compositions on the diopside-pyrope join in the system CaSi03-
MgSiO:r-Al203. Carnegie lnst. Washington Yearb . 73, 273-278.
Akella, J., and F. R. Boyd (1974) Petrogenetic grid for garnet peridotites, Carnegie
Inst. Washington Yearb. 73 , 269-273.
Akella, J., and G. C. Kennedy (1971) Studies on anorthite + diopside00-heden-
bergiteoo at high pressures and temperatures, Am. J . Sci .. 270, 155-165.
Akimoto, S.-1. (1972) The system Mg0-Fe0-Si02 at high pressures and temperatures-
phase equilibria and elastic properties. In A. R. Risema, ed., The upper mantle,
Elsevier Publishing Company, New York, p. 161-187.
Allen, E. T., and W. P. White (1909) Diopside and its relation to calcium and magnesium
metasilicates, Am. J . Sci., 27. 1-47.
Allison, E. B., P. Brock, and J. White (1959) The rheology of aggregates containing
a liquid phase with special reference to the mechanical properties of refractories at
high temperatures, Trans. Br. Ceram. Soc., 58, 495-531.
Anderson, D. L. (1962) The plastic layer of the earth' s mantle, Sci. Am. , 207, 52-59.
Anderson, D. L. (1970) Petrology of the mantle, Mineral. Soc. Am. Spec. Pap . 3,
85-93.
Anderson , D. L., and H. Spetzler (1970) Partial melting and the low-velocity zone,
Phys. Earth Planet. Inter., 4, 62-64.
Anderson, 0 . (1915) The system anorthite-forsterite-silica, Am. J . Sci .. 39, 407-454.

213

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

214 References
Anderson, 0. L., and P. C. Perkins (1975) A plate tectonics model involving non-
laminar asthenospheric flow to account for irregular patterns of magmatism in the
southwestern United States, Phys. Chern. Earth, 9, I 13-122.
Anthony, T. R., and H. E. Cline (1970) The kinetics of droplet migration in solids
in an acceleration~ field, Philos. Mag., 22, 893-901.
Anthony, T. R., and H. E. Cline (1971) Thermal migration of liquid droplets through
solids, J . Appl. Phys., 42, 3380-3387.
Anthony, T. R., and H. E. Cline (1972a) The thermomigration of biphase vapor-liquid
droplets in solids, Acta Metal/., 20, 247-255.
Anthony, T. R., and H. E. Cline (1972b) The interaction of migrating liquid inclusions
with grain boundaries in solids. In Hsun Hu, ed., Nature and behavior of grain
boundaries, Plenum Publishing Corp., New York, p. 185-201.
Arculus, R. J. (1975) Melting behavior of two basanites in the range 10 to 35 kbar and
the effect of Ti02 on the olivine~iopside reactions at high pressures, Carnegie
/nst. Washington Yearb. 74, 512-515.
Arculus, R. J., and E. B. Curran (1972) The genesis of the calc-alkaline rock suite,
Earth Planet. Sci. Lett., 15, 255-262.
Arculus, R. J., and N. Shimizu (1974) Rare earth elements in a suite of basanitoids and
alkali olivine basalts from Grenada, Lesser Antilles, Carnegie lnst. Washington
Yearb . 73, 553-560. .
Ashby, M. F. (1972) A first report on deformation-mechanism maps, Acta Metal/., 20,
887~97.
Bailey, D. K. (1970) Volatile flux, heat-focusing and the generation of magma. In G.
Newall and N. Rast, eds.,Mechanism of igneous intrusion, Gallery Press, Liverpool,
p. 1n-186.
Barton, D. C. (1936) Mechanics of formation of salt domes with special reference to
Gulf Coast salt domes of Texas and Louisiana. In D. C. Barton and G. Sawtelle,
eds., Gulf Coast Oil Fields; a Symposium on the Gulf Coast Cenozoic, American
Association of Petroleum Geologists, Tulsa, Oklahoma, p. 20-78.
Barus, C. (1893) High-temperature work in igneous fusion and ebullition, chiefly in
relation to pressure, U.S. Geol. Surv. Bull. 103, 51 p.
Bath, M. (1966) Earthquake energy and magnitude, Phys. Chern. Earth, 7, 115-165.
Bayly, B. (1968) Introduction to petrology, Prentice-Hall, Inc., Englewood Cliffs, New
Jersey, 371 p.
Beeson, M. H., and E. D. Jackson (1970) Origin of the garnet pyroxenite xenoliths at
Salt Lake Crater, Oahu, Mineral. Soc . Am. Spec. Pap. 3, 95-112.
Bell, P. M., and H. K. Mao (1972) The hypothesis of melting at stress dislocations in
the earth, Carnegie lnst. Washington Yearb. 71, 416-418.
Bell, P. M. , and H. K. Mao (1975) Laser optical system for heating experiments and
pressure calibration of the diamond-windowed, high-pressure cell, Carnegie lnst.
Washington Yearb . 74, 399-402.
Benioff, H. (1954) Orogenesis and deep crustal structure-additional evidence from
seismology, Geol. Soc. Am. Bull., 65, 385-400.
Betz, F., Jr., and H. H. Hess (1942) The floor of the North Pacific Ocean, Geogr.
Rev., 32, 99-116.
Biggs, W. D. (1970) Fracture. In R. W. Cann, ed., Physical metallurgy, North-Holland
Publishing Company, Amsterdam, p. 1199-1232.
Birch, F. (1951) Recent work on the radioactivity of potassium and some related
geophysical problems, J. Geophys. Res., 56, 107-126.
Birch, F. (1964) Mega-geological considerations in rock mechanics. In W. R. Judd,

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

References 215
ed., State of stress in the earth's crust, American Elsevier Publishing Company,
Inc., New York, p. 55-80.
Birch, F. (1965) Speculations on the earth's thermal history, Geol. Soc. Am. Bull., 76,
133-153.
Birch, F. (1969) Density and composition of the upper mantle: first approximation as an
olivine layer. In P. J. Hart, ed., The earth's crust and upper mantle, Geophys.
Monogr., Am. Geophys. Union, 13, 18-36.
Birch, F. (1970) Interpretations of the low-velocity zone, Phys. Earth Planet. Inter. ,
3, 178-181.
Birch, F., and D. Bancroft (1942) The elasticity of glass at high temperatures, and the
vitreous basaltic substratum, Am. J. Sci., 240, 451-490.
Birch, F., and H. Clark (1940) The thermal conductivity of rocks and its dependence
upon temperature and composition, Am. J. Sci., 238, 529-558.
Birch, F., and P. LeComte (1960) Temperature-pressure plane for albite composition,
Am. J. Sci., 258, 209-217.
Boettcher, A. L., and P. J. WyUie (1969) Phase relationships in the system NaAISi04 -
Si~-H20 to 35 kilobars pressure, Am. J. Sci., 267, 875-909.
Bowden, F. P., M. A. Stone, and G. K. Tudor (1947) Hot spots on rubbing surfaces
and the detonation of explosives by friction, Proc. R. Soc. London, Ser. A, 188,
329-349.
Bowden, F. P., and P. H. Thomas (1954) The surface temperature of sliding solids,
Proc. R. Soc. London, Ser. A, 223, 29-40.
Bowen, N. L. (1915) The crystallization of haplobasaltic, haplodioritic, and related
magmas, Am. J. Sci., 40, 161-185.
Bowen, N. L. (1928) The evolution of the igneous rocks, Princeton University Press,
Princeton, New Jersey, 332 p.
Bowen, N. L., and J. F. Schairer (1935) The system MgO-FeO-SiOt> Am. J. Sci., 29,
151-217.
Boyd, F. R. (1970) Gamet peridotites and the system MgSi03 -CaSi03 -AI20 3 • Mineral.
Soc. Am. Spec. Pap. 3, 63-75.
Boyd, F. R., and J. L. England (1959a) Pyrope, Carnegie lnst. Washington Yearb.
58, 83-87.
Boyd, F. R., and J. L. England (1959b) Quartz-coesite transition, Carnegie lnst.
Washington Yearb. 58, 87-88.
Boyd, F. R., and J. L. England (1963) Effect of pressure on the melting of diopside.
CaMgSiz08 , and albite, NaAISi3 0 8 , in the range up to 50 kilobars, J. Geophys. Res ..
68, 311-323.
Boyd, F. R., J. L. England, and B. T. C. Davis (1964) Effects of pressure on the melting
and polymorphism of enstatite, MgSi03 , J. Geophys. Res .. 69, 2101-2109.
Boyd, F. R., and P. H. Nixon (1973) Structure of the upper mantle beneath Lesotho,
Carnegie lnst. Washington Yearb. 72, 431-445.
Boyd, F. R., and P. H. Nixon (1975) Origins of the ultramafic nodules from some
kimberlites of northern Lesotho and the Monastery Mine, South Africa. Phys.
Chern. Earth, 9, 431-454.
Brace, W. F. (1964) Brittle fracture of rocks. In W. R. Judd, ed., State of stress in the
earth's crust, American Elsevier Publishing Company, Inc., New York, p. 111-178.
Brace, W. F., W. C. Luth, and J.D. Unger (1968) Melting of granite under an effective
confining pressure (abstract), Geol. Soc. Am. Spec. Pap. 1/5, 21.
Bradley, R. S. (1962) Thermodynamic calculations on phase equilibria involving fused
salts, Part II, Solid solutions and application to the olivines, Am. J . Sci., 260,
SS0-554.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

216 References
Bridgman, P. W. (1948) Rough compressions of In substances to 40,000 kg!cm 1 , Proc.
Am. Acad. Arts Sci., 76, 71-87.
Brooks, C., D. E. James, and J. R. Hart (1976) Ancient lithosphere: Its role in young
continental volcanism, Science /93, 1086-1094.
Brune, J. N., and C. R. ADen (1967) A low-stress-drop, low magnitude earthquake with
surface faulting: the Imperial, California, earthquake of March 4, 1966, Bull. Seismol.
Soc. Am., 57, 501-514.
Buddington, A. F. (1943) Some petrological concepts and the interior of the earth,
Am. Mineral., 28. 119-140.
Bullen, K. E. (1947) An introduction to the theory of seismology, Cambridge University
Press. Cambridge, 276 p.
Bultitude, R. J .. and D. H. Green (1971) Experimental study of crystal-liquid relation-
ship at high pressures in olivine nephelinite and basanite compositions, J. Petrol., 12,
121-147.
Burnham, C. W., and N. F. Davis (1974) The role of H20 in silicate melts, II, Thermo-
dynamic and phase relations in the system NaAISi3 0 8 -H2 0 to 10 kilobars, 700" to
IIOO"C, Am. J. Sci., 274, 902-940.
Butkovich, T. R. (1974) Rock melt from an underground nuclear explosion, Lawrence
Livermore Laboratory, ucRL-51554, Technical Information Document 4500, uc-35,
National Technical Information Service, Springfield, Virginia.
Butler, B. C. M. (1961) Metamorphism and metasomatism of rocks of the Moine series
by a dolerite plug in Glenmore, Ardnamurchan, Mineral. Mag .. 32, 866-897.
Cameron, M.. S. Sueno, C. T. Prewitt, and J. J. Papike (1973) High-temperature crystal
chemistry of acmite, diopside, hedenbergite, jadeite, spodumene, and ureyite,
Am . Mineral .. 58, 594~18.
Carr, M. J., and R. E. Stoiber (1973) Intermediate depth earthquakes and volcanic
eruptions in Central America, 1961-1972, Bull. Volcano/., 37, 326-337.
Carter. N. L., and H. G. Ave'Lallemant (1970) High temperature flow of dunite and
peridotite, Geol. Soc. Am. Bull .. 81. 2181-2202.
Cawthorn, R. G. (1975) Degrees of melting in mantle diapirs and the origin of ultra-
basic liquids, Eanh Planet. Sci. Lett., 27, 113-120.
Chamberlin, T. C., and R. D. Salisbury (1905) Geology, Vol. II. Earth history, 2d ed ..
Henry Holt & Co., New York, 692 p.
Chamberlin, T. C .. and R. D. Salisbury (1909) Geology, Vol. /, Geologic processes
and their results, 2d ed., Henry Holt & Co., New York, 684 p.
Chayes, F. ( 1963) Relative abundance of intermediate members of the oceanic basalt-
trachyte association, J. Geophys. Res .. 68, 1519-1534.
Chayes, F. (1972) Silica saturation in Cenozoic basalt, Philos. Trans. R. Soc. London.
Ser. A, 271, 285-296.
Chayes, F. (1975) Average composition of the commoner Cenozoic volcanic rocks.
Carnegie lnst. Washington Yearb. 74 , 547-549.
Chen, C.-H., and D. C. Presnall (1975) The system MgzSi04 -SiO, at pressures up to 25
kilobars, Am. Mineral .. 60, 398-406.
Clark, P. W.. J. H. Cannon, and J. White (1953) Further investigations on the sintering
of oxides, Trans . Br. Ceram. Soc .. 52, 1-49.
Clark, S. P., Jr .. and A. E. Ringwood (1964) Density distribution and constitution of the
mantle, Rev. Geophys., 2, 35-88.
Clark, S. P., Jr., J. F. Schairer, and J. de Neufville (1962) Phase relations in the
system CaMgSi 20 8 -CaAI2Si08-Si02 at low and high pressure, Carnegie lnst. Wash-
ington Yearb. 61, 59~.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

References 217
Clarke, W. B., M. A. Beg, and H. Craig (1969) Excess 3 He in the sea: evidence for
terrestrial primordial helium, Earth Planet. Sci. Lett., 6, 213-220.
Clayton. R. N., and T. K. Mayeda (1975) Genetic relations between the moon and
meteorites, Proc. Sixth Lunar Sci. Conf, Geochim. Cosmochim . Acta (Suppl. 6),
2, 1761-1769.
Coats, R. R. (1962) Magma type and crustal structure in the Aleutian arc. In G. A.
Macdonald and H. Kuno, eds., The crust of the Pacific basin, Geophys. Monogr.,
Am. Geophys. Union, 6, 92-109.
Coble, R. L., and J. E. Burke (1963) Sintering in ceramics. In J. E. Burke, ed., Progress
in ceramic science, MacmiUan Company, New York, p. 197-251.
Coe, R. S., and M. S. Paterson (1969) The cr-{3 inversion in quartz: a coherent phase
transition under nonhydrostatic stress, J. Geophys. Res., 74 , 4921-4928.
Cohen, L. H., K. Ito, and G. C. Kennedy (1967) Melting and phase relations in an
anhydrous basalt to 40 kilobars, Am. J. Sci., 265, 475-518.
Cordani, U. G., and P. Vandoros (1967) Basaltic rocks of the Parana Basin. In J. J.
Bigarella, R. D. Becker, and I. D. Pinto, eds., Problems in Brazj/ian Gondwana
geology, International Symposium on the Gondwana Stratigraphy and Palaeontology,
1st, Argentina, 1967, Curitiba, Brazil, p. 207-231.
Cotter, J. R. (1924) The escape of heat from the earth's crust, Philos . Mag ., 48,
458-464.
Cox, K. G. (1970) Tectonics and vulcanism of the Karroo Period and their bearing
on the postulated fragmentation of Gondwanaland. InT. N. Clifford and I. G. Gass,
eds., African magmatism and tectonics, Hafner Publishing Company, Darien,
Connecticut, p. 211-235.
Crandell, D. R., D. R. Mullineaux, and M. Rubin (1975) Mount St. Helens volcano:
recent and future behavior, Science, 187, 438-441.
Crittenden, M. D., Jr. (1967) Viscosity and finite strength of the mantle as determined
from water and ice loads, Geophys. J. R. Astron. Soc., 14, 261-279.
Cullers, R. L., L. G. Medaris, Jr., and L. A. Haskin (1970) Gadolinium: distribution
between aqueous and silicate phases, Science, 169, 580-583.
Daly, R. A. (1911) The nature of volcanic action, Proc. Am. Acad. Arts Sci., 47,
48-122.
Daly, R. A. (1925a) Relation of mountain building to igneous action, Proc. Am. Philos .
Soc., 64, 283-307.
Daly, R. A. (1925b) The geology of Ascension Island, Proc. Am. Acad. Sci., 60,
1-$0.
Daly, R. A. (1933) Igneous rocks and the depths of the earth, 2d ed., McGraw-HiD
Book Company, New York, 598 p.
Davis, B. T. C. (1964) The system diopside-forsterite-pyrope at 40 kilobars, Carnegie
1nst. Washington Yearb. 63, 165-171.
Davis, B. T . C., and F. R. Boyd (1966) The join Mg,Siz08 -CaMgSizOe at 30 kilobars
pressure and its application to pyroxenes from kimberlites, J. Geophys. Res .• 71,
3567-3576.
Davis, B. T. C., and J. L. England (1964) The melting of forsterite up to 50 kilobars,
J. Geophys. Res., 69, 1113-1116.
Davis, B. T. C., and J. F. Schairer (1965) Melting relations in the join diopside-
forsterite-pyrope at 40 kilobars and at one atmosphere, Carnegie lnst. Washington
Yearb. 64, 123-126.
Davis, N. F . (1972) Experimental studies in the system NaAISi3 0 8 -H20, Part I, The
apparent solubility of albite in supercritical water; Part II, The partial specific volume
of H,O in NaAISi3 0a melts with petrologic implications, unpublished Ph.D. thesis,
Pennsylvania State University, University Park, Pennsylvania.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

218 References
Day, A. L. , and E. S. Shepherd (1913) Water and volcanic activity, Geol. Soc. Am.
Bull., 24, 573...{,06.
DeLury, J. S. (1944) Generation of magma by frictional heat, Am. J. Sci. , 242, 113-129.
de Wys, E. C., and W. R. Foster (1958) The system diopside-anorthite-Akermanite,
Mineral. Mag ., 3/, 736-743 .
Dickey, J. S., Jr. (1970) Partial fusion products in Alpine-type peridotites: Serrania
de Ia Ronda and other examples, Mineral. Soc. Am. Spec. Pap. 3, 33-49.
Dickey, J. S., Jr., and H. S. Yoder, Jr. (1972) Partitioning of chromium and aluminum
between clinopyroxene and spinel, Carnegie Inst. Washington Yearb. 71,
384-392.
Dubourdieu, G. G. (1973) On the four-year seismic period, private publication, trans-
lated from the French by Noel Lindsay .
Dunn, P. R., and M. C. Brown (1969) North Australian plateau volcanics, Spec . Pub/.
Geol. Soc. Aust., 2, 117-122.
Du Toit, A. (1937) Our wandering continents, Oliver and Boyd, Edinburgh, 366 p.
Eaton, J. P. (1959) A portable water-tube tiltmeter, Bull. Seismol. Soc. Am., 49,
301-316.
Eaton, J. P. (1962) Crustal structure and volcanism in Hawaii. In G. A. Macdonald
and H. Kuno, eds., The crust of the Pacific basin, Geophys. Monogr., Am. Geophys.
Union, 6, 13-29.
Eaton, J. P., R. L. Christiansen, H. M. lyer, A. M. Pitt, D. R. Mabey, H. R. Blank,
Jr., I. Zietz, and M. E. Gettings (1975) Magma beneath Yellowstone National Park,
Science, 188, 787-796.
Eaton, J. P., and K. J. Murata (1960) How volcanoes grow, Science, 132, 925-938.
Eggler, D. H. (1973) Role of COt in melting processes in the mantle, Carnegie Inst.
Washington Yearb. 72, 457-467.
Eggler, D. H. (1974) Effect of C02 on the melting of peridotites, Carnegie Inst. Wash-
ington Yearb . 73 , 215-224.
Eggler, D. H. (1975) Peridotite-carbonatite relations in the system Ca0-Mg0-Si02-
C02, Carnegie lnst. Washington Yearb . 74, 468-474.
Eggler. D. H., B. 0 . Mysen, and M. G. Seitz (1974) The solubility of C02 in silicate
liquids and crystals, Carnegie Inst. Washington Yearb . 73, 226-228.
Einstein, A. (1906) Eine neue Bestimmung der Molekuldimensionen, Ann . Phys.
(Leipzig) , /9, 289-306.
Einstein, A. (1911) Berichtigung zu meiner Arbeit: "Eine neue Bestimmung der
Molekllldimensionen," Ann. Phys. (Leipzig), 34, 591-592.
Emslie , R. F. (1971) Liquidus relations and subsolidus reactions in some plagioclase-
bearing systems, Carnegie Inst . Washington Yearb . 69, 148-155.
Emslie, R. F., and D. H. Lindsley (1969) Experiments bearing on the origin of an-
orthositic intrusions, Carnegie Inst. Washington Yearb . 67, 108-109.
Engel, A. E. J .. and C. G. Engel (1964) Composition of basalts from the mid-Atlantic
Ridge, Science , /44 , 1330-1333.
Engel, A. E. J., C. G. Engel, and R. G. Havens (1965) Chemical characteristics of
oceanic basalts and the upper mantle, Geol. Soc. Am. Bull. , 76, 719-733.
Erlank, A. J., and E. J. D. Kable (1976) The significance of incompatible elements in
Mid-Atlantic Ridge basalts from 45" N with particular reference to Zr/Nb, Contrib.
Mineral. Petrol.• 54, 281-291.
Eshelby, V. D. (1957) The determination of the elastic field of an ellipsoidal inclusion,
and related problems, Proc. R. Soc. London, Ser. A, 24/, 376-396.
Evans, A. G., M. Linzer. and L. R. Russel (1974) Acoustic emission and crack propaga-
tion in polycrystalline alumina, Mater. Sci. Eng .• /5, 253-261.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

References 219
Fahrig, W. F., and R. K. Wanless (1963) Age and significance of diabase dyke swarms
of the Canadian shield, Nature (London) , 200, 934-937.
Faure, G., and J. L. Powell (1972) Strontium isotope geology, Springer-Verlag, New
York, 188 p.
Fedotov, S. A. , and P. I. Tokarev (1974) Earthquakes, characteristics of the upper
mantle under Kamchatka and their connection with volcanism (according to data
collected up to 1971), Bull. Volcano/., 37, 245-257.
Feild, A. L. , and P. H. Royster (1918) Slag viscosity tables for blast-furnace work,
U.S . Bur. Mines Tech . Pap. 187, 38 p.
Ferguson, J. B., and H. E. Merwin (1919) The ternary system Ca0-Mg0-Si02 ,
Am. J . Sci., 48, 81-123 .
Fermor, L. L. (1913) Preliminary note on garnet as a geological barometer and on an
infra-plutonic zone in the earth's crust, Rec. Geol. Surv. India , 43, Pt. I , 41~7.
Finger, L. W. , and Y. Ohashi (1976) The thermal expansion of diopside to 8000C and
a refinement of the crystal structure at 7000C, Am. Mineral., 61 , 303-310.
Fiske, R . S., C. A. Hopson, and A. C. Waters (1963) Geology of Mount Rainier
National Park, Washington, U.S . Geol. Surv. Prof. Pap. 444, 93 p.
Fiske, R. S., and E. D. Jackson (1972) Orientation and growth of Hawaiian volcanic
rifts: the effect of regional structure and gravitational stresses, Proc . R . Soc. London,
Ser. A , 329, 299-326.
Foster, W. R. (1942) The system NaAISi30 8-CaSi03-NaAISi04 , J. Geol., 50, 152-173.
Fujii, T ., and E. Takahashi (1976) On the solubility of alumina in orthopyroxene
coexisting with olivine and spinel in the system MgO-AI203-SiOt. Mineral. J ., 8,
No. 2, 122-128.
Gast, P. W. (1960) Limitations on the composition of the upper mantle, J . Geophys.
Res. , 65, 1287-1297.
Gast, P. W. (1968) Trace element fractionation and the origin of tholeiitic and alkaline
magma types, Geochim. Cosmochim . Acta, 32, 1057-1086.
Gast, P. W., G. R. Tilton, and C. Hedge (1964) Isotopic composition of lead and
strontium from Ascension and Gough Islands , Science , /45, 1181-1185.
Gastil, G. (1960) The distribution of mineral dates in time and space, Am. J . Sci ..
258, 1-35.
Gault, D. E., and E. D. Heitowit (1963) The partition of energy for hypervelocity
impact craters formed in rock, Proc . Sixth Hypervelocity Impact Symp .. Cleveland,
Ohio, 2, 419~56 .
Geological Museum of London (1972) The story of the earth , Institute of Geological
Sciences, Geological Museum, London, Her Majesty's Stationery Office, 36 p.
Gilbert, M. C. (1969) Reconnaissance study of the stability of amphiboles at high
pressure, Carnegie lnst. Washington Yearb . 67, 167-170.
Gilluly, J. (1972) Tectonics involved in the evolution of mountain ranges. In E. C.
Robertson, ed., Nature of the solid earth , McGraw-Hill Book Company, New
York, p . ~39.
Goetze, C. (1975) Sheared lherzolites: from the point of view of rock mechanics,
Geology, 3, 172-173.
Gorshkov, G. S. (1958) On some theoretical problems of volcanology, Bull. Volcano/. ,
19, 103-113.
Green, D. H. (1971) Composition of basaltic magmas as indicators of conditions of
origin: application to oceanic volcanism, Philos . Trans . R. Soc . London, Ser. A ,
268, 101-125.
Green, D. H. (1973) Contrasted melting relations in a pyrolite upper mantle under
mid-oceanic ridge , stable crust and island arc environments, Tectonophysics , 17,
285-297.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

220 References
Green, D. H., and W. Hibberson (1970) The instability of plagioclase in peridotite
at high pressure, Lithos, 3, 209-221.
Green, D. H., I. A. Nicholls, M. Viljoen, and R. Viljoen (1975) Experimental
demonstration of the existence of peridotitic liquids in earliest Archean magmatism,
Geology, 3, 11-14.
Green, D. H., and A. E. Ringwood (1967) The genesis of basaltic magmas, Contrib .
Mineral. Petrol., 15, 103-190.
Green, W. L. (1887) Vestiges of the molten globe, Part 1/, The earth's surface features
and volcanic phenomena, Hawaiian Gazette Publishing Company , Hawaii, 337 p.
Griggs, D. T. (1954) High pressure phenomena with applications to geophysics. In L. H.
Ridenour, ed., Modern physics for the engineer, McGraw-Hill Book Company. New
York, 499 p.
Griggs, D. T ., and D. W. Baker (1969) The origin of deep-focus earthquakes. In H. Mark
and S. Fembach, eds .• Properties of matter under unusual c·onditions, lnterscience
Publishers. New York , p. 23-42.
Griggs, D., and J. Handin (1960) ObseJVations on fracture and a hypothesis of earth-
quakes. In D. Griggs and J. Handin, eds., Rock deformation, Geol. Soc. Am. Mem .
79, 347-364.
Grout, F. F. (1945) Scale models of structures related to batholiths, Am. J. Sci. , Daly
Vol., 243A , 260-284.
Gruntfest, I. J. (1963) Thermal feedback in liquid tlow; plane shear at constant stress,
Trans. Soc. Rheol., 7, 195-207.
Gummer, W. K. (1943) The system CaSiO,.-CaAI2 Si 20.-NaAISi04 , J. Geol., 51 ,
503-530.
Gutenberg, B. (1926) Untersuchungen zur Frage, bis zu Welcher tiefe die erde Kristal-
lin ist, Z. Geophys., 2, 24-29.
Gutenberg, B. (1959)Physics of the earth's interior, Academic Press, New York, 240 p.
Gutenberg, B., and C. F. Richter (1949) Seismicity of the earth and associated
phenomena, Princeton University Press, Princeton, New Jersey , 273 p.
Gutenberg, B., and C. F. Richter (1956) Magnitude and energy of earthquakes, Ann .
Geofis., 9, 1-15.
Haller, W. (1965) Rearrangement kinetics of the liquid-liquid immiscible microphases
in alkali borosilicate melts, J . Chern. Phys., 42, ~93 .
Handin, J., R. V. Hager, Jr., M. Friedman, and J. N. Feather (1963) Experimental
deformation of sedimentary rocks under confining pressure: pore pressure tests,
Bull. Am. Assoc. Pet. Geol., 47, 717-755.
Harker, A. (1909) The natural history of igneous roc·ks, Methuen & Company, London,
384 p.
Harris, P. G. (1957) Zone refining and the origin of potassic basalts, Geochim .
Cosmochim. Acta, 12, 195-208.
Hart, S. R. (1971) K, Rb, Cs, Sr. and Ba contents and Sr isotope ratios of ocean
floor basalts, Philos. Trans . R. Soc. London , Ser. A, 268. 573-587.
Hart, S. R., J. G. Schilling, and J. L. Powell (1973) Basalts from Iceland and
along the Reykjanes ridge: Sr isotope geochemistry, Nature (London), Phys . Sci.,
246 , 104-107.
Haskin, L. A., M. A. Haskin, F. A. Frey, and T. R. Wildeman (1968) Relative and
absolute terrestrial abundances of the rare earths. In L. H. Ahrens, ed., Origin
and distribution of the elements , Pergamon Press, New York, p. 889-912.
Hazen, R. M. (1975) Effects of temperature and pressure on the crystal physics of
olivine, Ph.D. thesis, HaJVard University, Cambridge, Massachusetts, 264 p.
Heard, H. C. (1960) Transition from brittle fracture to ductile tlow in Solenhofen
limestone as a function of temperature, confining pressure, and interstitial fluid

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

References 221
pressure. In D. Griggs and J. Handin, eds., Rock deformation, Geol. Soc. Am. Mem.
79, 193-244.
Heard, H. C. , and W. W. Rubey (1966) Tectonic implications of gypsum dehydration,
Geol. Soc. Am. Bull., 77, 741-760.
Hedge, C. E., and Z. E. Peterman (1970) The strontium isotopic composition of basalts
from the Gordo and Juan de Fuca rises, northeastern Pacific Ocean, Contrib.
Mineral. Petrol., 27, 114-120.
Heirtzler, J. R. , G. 0. Dickson, E. M. Herron, W. C. Pitman III, and X . Le Pichon
(1968) Marine magnetic anomalies, geomagnetic field reversals, and motions of the
ocean floor and continents, J . Geophys. Res., 73, 2119-2136.
Helmke, P. A., and L. A. Haskin (1973) Rare-earth elements, Co, Sc and Hf in the
Steens Mountain basalts, Geochim. Cosmochim. Acta, 37, 1513-1529.
Herty, C. H ., Jr., F. A. Hartgen, J . A. Heidish, K. Metcalfe, F. G. Norris, and
M. B. Royer (1930) Temperature-visc<'sity relations in the lime-silica system,
Carnegie lnst. Techno/., Coop. Bull. Min . Metal/. Invest. , 47, 27 p.
Hess, H . H . (1960) Stillwater igneous complex, Montana, Geol. Soc. Am. Mem. 80,
230 p.
Hodges, F. N. (1974) The solubility of H20 in silicate melts, Carnegie lnst. Washington
Yearb. 73, 251-255.
Hofmann, A. (1974) Strontium diffusion in a basalt melt and implications for Sr isotope
geochemistry and geochronology, Carnegie lnst. Washington Yearb . 73, 935-941.
Hofmann, A. W. , and S. R. Hart (1975) An assessment of local and regional isotopic
equilibrium in a partially molten mantle, Carnegie lnst. Washington Yearb . 74,
195-210.
Holmes, A. (1921) Petrographic methods and calculations, Thomas Murby & Co.,
London, 515 p.
Holmes, A. (1915) Radioactivity and the earth's thermal history, Geol. Mag., 2,
60-70, 102-111.
Holmes, A. (1927) Some problems of physical geology and the earth's thermal history,
Geol. Mag. , 64, 263-278.
Holmes, A. (1945) Principles of physical geology, The Ronald Press Company, New
York, 532 p.
Holmes, A. (1965) Principles of physic·al geology, Th . Nelson and Sons, Ltd.,
London, 1288 p.
Hubbert, M. K., and W. W. Rubey (1959) Role of fluid pressure in mechanics of
overthrust faulting, I. Mechanics of fluid-filled porous solids and its application to
overthrust faulting, Geol. SoC'. Am. Bull., 70, 115-166.
Hubbert, M. K., and D. G. Willis (1957) Mechanics of hydraulic fracturing, Trans.
AIME, 210, 153-168.
Huckenholz, H. G. (1965) Der petrogenetische Werdegang der Klinopyroxene in den
tertiiiren Vulkaniten der Hocheifel (Parts I and II) , Beitr. Mineral. Petrogr., II ,
138-195, 415-448.
Huckenholz, H. G . (1966) Der petrogenetische Werdegang der Klinopyroxene in den
tertiiiren Vulkaniten der Hocheifel (Part III), Contrib. Mineral. Petrol., 12, 73-95.
Huckenholz, H. G . (1973) The origin of fassaitic augite in the alkali basalt suite of the
Hocheifel area, western Germany, Contrib . Mineral. Petrol., 40, 315-326.
Hulme, G . (1974) The interpretation of lava flow morphology , Geophys . J . R . Astron .
Soc., 39, 361-383.
Hurley, P. M. (1968a) Absolute abundance and distribution of Rb, K and Sr in the
earth, Geochim. Cosmochim . Acta, 32, 273-283.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

222 References
Hurley, P. M. (1968b) Correction to: Absolute abundance and distribution of Rb, K and
Sr in the earth, Geochim. Cosmochim . Acta, 32 , 1025-1030.
Hyt6nen, K., and J. F. Schairer (1960) The system enstatite-anorthite-diopside,
Carnegie lnst. Washington Yearb. 59, 71-72.
Hytonen, K., and J. F. Schairer (1961) The plane enstatite-anorthite-diopside and its
relation to basalts, Carnegie lnst. Washington Yearb . 60, 125- 141.
Iida, K. (1959) Earthquake energy and earthquake fault, J . Earth Sci. Nagoya Univ .,
7, 98-107.
Irvine, T. N. (1970) Crystallization sequences in the Muskox intrusion and other
layered intrusions, I, Olivine-pyroxene-plagioclase relations, Geol. Soc. S . Afr.
Spec. Pub/. I, 441~76 .
Irving, A. J. , and P. J. Wyllie (1975) Subsolidus and melting relationships for calcite,
magnesite and the join CaC03 -MgC03 to 36 kb, Geochim. Cosmochim. Acta, 39,
35-53.
Ito, K. , and G. C. Kennedy (1967) Melting and phase relations in a natural peridotite
to 40 kilobars, Am. J . Sci. , 265, 519-538.
Ito, K., and G. C. Kennedy (1968) Melting and phase relations in the plane tholeiite-
lherzolite-nepheline basanite to 40 kilobars with geological implications. Contrib.
Mineral. Petrol .• /9, 177-211.
Ito, K., and G. C. Kennedy (1974) The composition of liquids formed by partial melting
of eclogites at high temperatures and pressures, J . Geol., 82, 383-392.
Jackson, E. D.. E. A. Silver, and G. B. Dalrymple (1972) Hawaiian-Emperor
chain and its relation to Cenozoic circumpacific tectonics, Geol. Soc. Am. Bull., 83,
601-617.
Jackson, E. D., and T. L. Wright (1970) Xenoliths in the Honolulu volcanic series,
Hawaii, J. Petrol.. II , 405~30.
Jaggar, T. A. (1917) Volcanic investigations at Kilauea, Am. J. Sci., 44, 165-220.
Johannsen, A. (1939)A descriptive petrography of the igneous rocks, Vol. I, University
of Chicago Press, Chicago, 318 p.
Joly, J. (1909) Radioactivity and geology, Archibald Constable and Company, Ltd.,
London, 287 p.
Joly, J. (1930) The surface-history of the earth , Clarendon Press, Oxford, 211 p.
Jordan, T. H. (1975) The continental tectosphere, Rev. Geophys. Space Phys., 13,
No.3. 1-12.
Joule, J. P., and J. Thomson (1854) On the thermal effects of fluids in motion-No. II,
Proc. R . Soc. London , No.7, 127-130.
Judd, J. W. (1886) On the gabbros, dolerites and basalts of Tertiary age in Scotland
and Ireland, Q. J. Geol. Soc. London , 42, 49-97.
Kaula, W. M. (1963) Elastic models of the mantle corresponding to variations in the
external gravity field , J. Geophys. Res., 68, 4967~978.
Kawada, K. (1966) Studies of the thermal state of the earth, The 17th paper: Varia-
tion of thermal conductivity of rocks, Pt. 2, Bull. Earthquake Res. lnst. Tokyo
Univ., 44, 1071-1091.
Kennedy, G. C. (1959) The origin of continents, mountain ranges, and ocean basins,
Am. Sci. , 47, 491-504.
Kirby, S. H., and C. B. Raleigh (1973) Mechanisms of high-temperature, solid-state
flow in minerals and ceramics and their bearing on the creep behavior of the mantle.
Tectonophysics, /9 , 165-194.
Knopoff, L. (1972) Observation and inversion of surface-wave dispersion, Tectono-
physics, 13, 497-519.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

References 223
Koyanagi, R. Y., and E. T. Endo (1971) Hawaiian seismic events during 1969, U.S.
Geol. Surv. Prof Pap. 750-C, CI58-CI64.
Kozu, S., and J. Ueda (1933) Thermal expansion of diopside, Proc. Imp. Acad.
(Tokyo), 9, 317-319.
Kretz, R. (1961) Some applications of thermodynamics to coexisting minerals of variable
compositions. Examples: orthopyroxene-clinopyroxene and orthopyroxene-garnet,
J. Geo/., 69, 361-387.
Kubota, S., and E. Berg (1967) Evidence for magma in the Katmai volcanic range,
Bull. Volcano/., 31, 175-214.
Kuno, H. (1959) Origin of Cenozoic petrographic provinces of Japan and surrounding
areas, Bull. Volcano/., 20, 37-76.
Kuno, H. (1967) Volcanological and petrological evidences regarding the nature of the
upper mantle. InT. F. GaskeU, ed., The earth's mantle, Academic Press, New York,
p. 89-110.
Kuno, H. (1969) Plateau basalts. In P. J. Hart, ed., The earth's crust and upper
mantle, Geophys. Monogr., Am. Geophys. Union, 13, 495-501.
Kuno, H ., and K.-1. Aoki (1970) Chemistry of ultramafic nodules and their bearing on
the origin of basaltic magmas, Phys . Earth Planet. Inter., 3, 273-301.
Kushiro, I. (1964) Coexistence of nepheline and enstatite at high pressures, Carnegie
lnst. Washington Yearb. 64, 109-112.
Kushiro, I. (1968) Compositions of magmas formed by partial zone melting of the
earth's upper mantle, J. Geophys. Res., 73, 619-634.
Kushiro,l. (1969a) Clinopyroxene solid solutions formed by reactions between diopside
and plagioclase at high pressures, Mineral. Soc. Am. Spec. Pap. 2, 179-191.
Kushiro, I. (1969b) The system forsterite-diopside-silica with and without water at
high pressures, Am. J. Sci., Schairer Vol. , 267A, 269-294.
Kushiro, I. (1970) Stability of amphibole and phlogopite in the upper mantle,
Carnegie Inst. Washington Yearb. 68, 245-247.
Kushiro,I. (1972a) Determination of liquidus relations in synthetic silicate systems with
electron probe analysis: the system forsterite-diopside-silica at I atmosphere, Am.
Mineral., 57, 1260-1271.
Kushiro, I. (1972b) Effect of water on the compositions of magmas formed at high
pressures,}. Petrol., 13, 311-334.
Kushiro, I. (1973a) Origin of some magmas in oceanic and circum-oceanic regions,
Tectonophysics, 17, 211-222.
Kushiro, I. (1973b) Partial melting of garnet lherzolites from kimberlite at high pres-
sures. In P. H. Nixon, ed., Lesotho kimberlites, Lesotho National Development
Corporation, Maseru, Lesotho, p. 294-299.
Kushiro, I. (1974) The system forsterite-anorthite-albite-silica-H 2 0 at IS kbar and the
genesis of andesitic magmas in the upper mantle, Carnegie Inst. Washington
Yearb. 73, 244-248.
Kushiro, 1., Y. Syono, and S. Akimoto (1968a) Melting of a peridotite at high pres-
sures and high water pressures, J. Geophys. Res., 73, 6023~9.
Kushiro, 1., and R. N. Thompson (1972) Origin of some abyssal tholeiites from the
mid-Atlantic ridge, Carnegie lnst. Washington Yearb. 71, 403-406.
Kushiro, I., and H. S. Yoder, Jr. (1966) Anorthite-forsterite and anorthite-enstatite
reactions and their bearing on the basalt-eclogite transformation, J. Petrol., 7.
337-362.
Kushiro, 1., and H. S. Yoder, Jr. (1969) Melting of forsterite and enstatite at high
pressures under hydrous conditions, Carnegie lnst. Washington Yearb. 67,
153-158.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

224 References
Kushiro, 1., and H. S. Yoder, Jr. (1974) Formation of eclogite from garnet lherzolite
liquidus relations in a portion of the system MgSi0 3 -CaSi03 -Al2 0 3 at high pressures,
Carnegie lnst. Washington Yearb. 73, 266-269.
Kushiro, 1., H. S. Yoder, Jr., and B. 0 . Mysen (In press) Viscosity of basalt and
andesite melts at high pressures.
Kushiro, I.. H. S. Yoder, Jr., and M. Nishikawa (1968b) Effect of water on the melting
of enstatite, Geol. Soc . Am . Bull., 79, 1685-1692.
Lacroix, A. (1917) Sur Ia transformation de quelques roches eruptives basiques en
amphibolites, C. R. Acad. Sci .. 164, 969-974.
Lambert, I. B., and P. J. WyUie (1970) Low-velocity zone of the earth's mantle:
incipient melting caused by water, Science, 169, 764-766.
Larsen, E. S. (1909) Relation between the refractive index and the density of some
crystallized silicates. Am. J . Sci .• 28. 263-274.
Larsen, E. S .• and H. Berman (1934) The microscopic determination of the nonopaque
minerals, U.S. Geol. Surv. Bull. 848. 266 p.
Lee, W. H. K. (1967) Thermal history of the earth, Ph.D. thesis in planetary and
space physics, University of California at Los Angeles. 344 p.
Lindemann, F. A. (1910) Ober die Berechnung Molekularer Eigenfrequenzen, Phys.
Z .. II. 609-612.
Lindsley, D. H. (1967) Pressure-temperature relations in the system Fe0-Si02 ,
Carnegie lnst. Washington Yearb . 65, 226-230.
Liu, L. (1974) Silicate perovskite from phase transformations of pyrope-gamet at high
pressure and temperature, Geophys. Res. Lett .. I, 277-280.
Lliboutry. L. (1971) Permeability, brine content and temperature of temperate ice,
J. Geo/., /0, 15-29.
Lovering, J. F., and J . W. Morgan (1964) Uranium and thorium abundances in stony
meteorites, J. Geophys. Res .. 69, 1979-1988.
Lubimova, E. A. ( 1958) Thermal history of the earth with consideration of the variable
thermal conductivity of its mantle, Geophys. J. R. Astron. Soc., / , 115-134.
Lubimova. E. A. (1969) Thermal history of the earth. In P. J. Hart, ed .• The earth' s
crust and upper mantle , Geophys. Monogr., Am. Geophys. Union, /3, 63-77.
Lupton , J. E .• and H. Craig (1975) Excess 3 He in oceanic basalts: evidence for
terrestrial primordial helium , Earth Planet. Sei. Lett., 26, 133-139.
Lure, M. L., and V. L. Masaitis (1964) Principal features of the geology and petrology
of the trap formations of the Siberian Platform , Int. Geol. Congr., 22nd Session,
Plateau Basalts, Nauka. Moscow.
Luth, W. C. (1967) Studies in the system KAISiO.-MBtSiO.-Si02 -H 20 : I. Inferred phase
relations and petrologic applications, J . Petrol .• 8, 372~16.
Maal~. S. (1973) The significance of the melting interval of basaltic magmas at
various pressures, Geol. Mag .. I /0, 103-112.
Macdonald, G. A. (1972) Volcanoes, Prentice-Hall . Inc .• Englewood Cliffs, New
Jersey, 510 p.
MacDonald, G. J. F. (1959) Calculations on the thermal history of the earth, J.
Geophys. Res .• 64, 1967-2000.
MacGregor, I. D. (1968) Mafic and ultramafic inclusions as indicators of the depth of
origin of basaltic magmas, J. Geophys. Res .• 73, 3737-3745.
MacGregor. I. D. (1974) The system MgO-AI20,.-Si02: solubility of AltO• in enstatite
from spinel and garnet peridotite compositions, Am. Mineral .. 59, I 10-119.
Machado, F. (1974) The search for magmatic reservoirs. In L. Civetta, P. Gasparini.
G. Luongo, and A. Rapolla, eds., Physical volcanology, American Elsevier Scientific
Publishing Company, Inc., New York, p. 255-273.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

References 225
Manson, V. (1967) Geochemistry of basaltic rocks: major elements. In H. H. Hess
and A. Poldervaart. eds .• Basalts: the Poldervaart treatise on rocks of basaltic
composition, Vol. I, lnterscience Publishers, New York. p. 215-269.
Mao, H. K .• and P. M. Bell (1971) Generation of magma along faults and stress
dislocations in the earth, Carnegie Jnst. Washington Yearb . 70. 229-233.
Mao, H. K .. and P. M. Bell (1976) High-pressure physics: the 1-megabar mark on the
ruby R, static pressure scale. Science, /9/, 851-852.
Mao, H. K .• and J. F. Schairer (1970) Quenching experiments in the system jadeite
(NaAISitOs)-forsterite (MgtSiO.) and jadeite (NaAISitOs)-anorthite (CaAI 2 Si 20.),
Carnegie Jnst . Washington Yearb. 68. 221-222.
Marsh . B. D.• and I. S. E. Carmichael (1974) Benioff zone magmatism. J . Geophys.
Res .. 79, 1196-1206.
Masuda, A .. and I. Kushiro (1970) Experimental determination of partition coefficients
of ten rare earth elements and barium between clinopyroxene and liquid in synthetic
silicate systems at 20 kilobar pressure. Contrib . Mineral. Petrol .. 26, 42-49.
McBimey, A. R. (1963) Conductivity variations and terrestrial heat-flow distribution,
J . Geophys. Res .. 68, 6323-6329.
McBimey, A. R. (1967) Genetic relations of volcanic rocks of the Pacific Ocean,
Geol. Rundsch .. 57. 21-33.
McBimey, A. R.• J. F. Sutter, H. R. Naslund. K. G. Sutton, and C. M. White (1974)
Episodic volcanism in the Central Oregon Cascade Range, Geology, 2, 585-589.
McGetchin. T. R. (1975) Solid earth geosciences research activities at LASL, July !-
December 31, 1974, Prog. Rep. LA-5956-PR. Los Alamos Scientific Laboratories.
U.S. Energy Research and Development Administration. 93 p.
McKenzie. D. P .• and J. G. Sclater (1968) Heat flow inside the island arcs of the
northwestern Pacific, J . Geophys. Res .. 73 . 3173-3179.
McLean. D. (1957) Grain boundaries in metals. Oxford University Press. Clarendon.
346 p.
McQuarrie, M. (1954) Thermal conductivity: VII . Analysis of variation of conductivity
with temperature for Al20 3 • BeO , and MgO . J . Am. Ceram . Soc .. 37. 91-95.
Means, W. D.• and P. F. Williams (1974) Compositional differentiation in an experi-
mentally deformed salt-mica specimen. Geology, 2. 15-16.
Mehnert, K. R.• W. Busch. and G. Schneider (1973) Initial melting at grain boundaries
of quartz and feldspar in gneisses and granulites. Neues Jahrb. Mineral. Monatsh ..
165-183.
Menard. H. W. (1964) Marine geology of the Pacific , McGraw-Hill Book Company,
New York, 271 p.
Menard, H. W. (1969) The deep-ocean floor. Sci. Am .. 221, 127-142.
Menzies, M. (1973) Mineralogy and partial melt textures within an ultramafic-mafic
body, Greece, Contrib. Mineral. Petrol.. 42. 273-285.
Mercier. J.-C .• and N. L. Carter (1975) Pyroxene geotherms. J . Gl·ophys. Res. , 80.
3349-3362.
Merrill, R. B.• and P. J . Wyllie (1975) Kaersutite and kaersutite eclogite from Kakanui.
New Zealand-water-excess and water-deficient melting to 30 kilobars . Geo/. Soc .
Am. Bull., 86, 555-570.
Millhollen, G. L., A. J . Irving, and P. J. Wyllie (1974) Melting interval of peridotite with
5.7 per cent water to 30 kilobars, J . Geol., 82. 575-587.
Modreski, P. J., and A. L. Boettcher (1972) The stability of phlogopite +enstatite at
high pressures: a model for micas in the interior of the earth, Am . J . Sci .. 272,
852-869.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

226 References
Mogi, K. (1958) Relations between the eruptions of various volcanoes and the
defonnations of the ground surfaces around them, Bull. Earthquake Res. lnst.
Tokyo Univ. , 36, 99-134.
Mogi, K. (1965) Defonnation and fracture of rocks under confining pressure, (2)
Elasticity and plasticity of some rocks, Bull. Earthquake Res. lnst. Tokyo Univ ., 43,
349-379.
Mogi, K. (1966) Pressure dependence of rock strength and transition from brittle
fracture to ductile flow, Bull. Earthquake Res. lnst. Tokyo Univ ., 44, 215-232.
Moorbath, S., R. K. O' Nions, R. J. Pankhurst, N. H. Gale, and V. R. McGregor
(1972) Further rubidium-strontium age detenninations on the very early Precambrian
rocks of the Godthaab District, West Greenland, Nature (London), Phys. Sci. , 240 ,
78-82.
Moore, P. B., and J. V. Smith (1970) Crystal structure of /3-M~SiO. : crystal-chemical
and geophysical implications, Phys. Earth Planet. Inter., 3, 166-177.
Morey . G. W. ( 1922) The development of pressure in magmas as a result of crystalliza-
tion, J . Wash. Acad. Sci. , /2 , 219-230.
Morgan, W. J. (1971) Convection plumes in the lower mantle, Nature (London) , 230,
42-43 .
Morgan, W. J. (1972) Deep mantle convection plumes and plate motions, Bull. Am.
Assoc. Pet. Geol. , 56 , 203-213 .
Morimoto, N., S.-1. Akimoto, K. Koto, and M. Tokonami (1970) Crystal structures of
high pressure modifications of Mn2Ge04 and COtSi04 , Phys. Earth Planet. Inter. ,
3, 161-165.
Morse, S. A. (1970) Alkali feldspars with water at 5 kb pressure, J. Petrol., /1 ,
221-251.
Murase, T., and A. R. McBirney (1970) Thennal conductivity of lunar and terrestrial
igneous rocks in their melting range, Science, 170, 165-167.
Murase, T., and A. R. McBirney (1973) Properties of some common igneous rocks
and their melts at high temperatures, Geol. Soc. Am. Bull. , 84, 3563-3592.
Murase, T .• and T. Suzuki (1966) Ultrasonic velocity of longitudinal waves in molten
rocks, J. Fac. Sci. Hokkaido Univ. , Ser. 7, 2, 273-285 .
Mysen, B. 0 . (1973) Melting in a hydrous mantle: Phase relations of mantle peridotite
with controlled water and oxygen fugacities. Carnegie lnst. Washington Yearb.
72, 467-478.
Mysen, B. 0 . (1975) Solubility of volatiles in silicate melts at high pressure and
temperature: the role of carbon dioxide and water in feldspar, pyroxene, and feld-
spathoid melts, Carnegie lnst . Washington Yearb . 74, 454-468.
Mysen, B. 0 . (1976a) Experimental detennination of some geochemical parameters
relating to conditions of equilibration of peridotite in the upper mantle, Am. Mineral.,
6/, 677~3 .
Mysen, B. 0. (1976b) Partitioning of samarium and nickel between olivine, ortho-
pyroxene, and liquid: preliminary data at 20 kbar and 1025•c. Earth Planet. Sci.
Lett., 3/, 1-7.
Mysen, B. 0 ., and A. L. Boettcher (1975) Melting of a hydrous mantle: I, Phase
relations of natural peridotite at high pressures and temperatures with controlled
activities of water, carbon dioxide and hydrogen, J . Petrol. , /6, 520-548.
Nakamura, Y.• and I. Kushiro (1974) Composition of the gas phase in M~Si04 -
Si02-H20 at 15 kbar, Carnegie lnst. Washington Yearb . 73, 255-258.
Nesbitt, R. W. (1971) Skeletal crystal fonns in the ultramafic rocks of the Yilgarn
block. western Australia; evidence for an Archean ultramafic liquid, Spec. Pub/.
Geol. Soc. Aust., 3, 331-347.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

References 227
Nesterenko, G. V., and A. I. Almukhamedov (1973) The geochemistry of differentiated
traps (Siberian Platform), Nauka, Moscow, 198 p.
Nettleton, L. L. (1936) Fluid mechanics of salt domes. In D. C. Barton and G. Saw-
telle, eds., Gulf Coast Oil Fields; a Symposium on the Gulf Coast Cenozoic,
American Association of Petroleum Geologists, Tulsa, Oklahoma, p. 79-108.
Nier, A. 0. (1950) A redetermination of the relative abundances of the isotopes of
carbon, nitrogen, oxygen, argon, and potassium, Phys. Rev., 77, 789-793.
Nockolds, S. R. (1954) Average chemical compositions of some igneous rocks, Geol.
Soc. Am. Bull., 65, 1007-1032.
Nur, A. (1972) Dilatancy, pore fluids, and premonitory variations of t,ltv travel
times, Bull. Seismol. Soc. Am., 62, 1217-1222.
Nur, A. (1974) Tectonophysics: the study of relations between deformation and forces in
the earth. In Advances in rock mechanics, Vol. I, Part A, National Academy of
Sciences, Washington, D.C., p. 243-317.
Nutting, P. G. (1929) Deformation and temperature, J. Wash. Acad. Sci., /9, 109-115.
Nye, J. F. (1967) Theory of regelation, Philos. Mag., /6, 1249-1266.
Nye, J. F .• and F. C. Frank (1973) Hydrology of the intergranular veins in a
temperate glacier. In Symposium on the hydrology of glaciers, Assoc. Int.
Hydrologie Scientifique Pub/. 95, 157-161.
Obata, M. (1975) The solubility of [Link] in orthopyroxenes in spinel and plagioclase
peridotites and spinel pyroxenite. In Extended abstracts: International Conference on
Geothermometry and Geobarometry, Pennsylvania State University, University
Park, Pennsylvania, p. 1-2.
Obata, M., S. Banno, and T. Mori (1974) The iron-magnesium partitioning between
naturally occurring coexisting olivine and Ca-rich clinopyroxene: an application of the
simple mixture model to olivine solid solution, Bull. Soc. Fr. Mineral. Cristallogr.,
97, 101-107.
O'Hara, M. J. (1961) Zoned ultrabasic and basic gneiss masses in the early Lewisian
metamorphic complex at Scourie, Sutherland, J. Petrol., 2, 248-276.
O'Hara, M. J. (1965) Primary magmas and the origin of basalts, Scott. J. Geo/., /,
19-40.
O'Hara, M. J. (1968a) Are ocean floor basalts primary magma? Nature (London), 220,
683-686.
O'Hara, M. J. (1968b) The bearing of phase equilibria studies in synthetic and natural
systems on the origin and evolution of basic and ultrabasic rocks, Earth-Sci. Rev., 4,
69-133.
O'Hara, M. J., M. J. Saunders, and E. L. P. Mercy (1975) Garnet-peridotite, primary
ultrabasic magma and eclogite; interpretation of upper mantle processes in kimber-
lite, Phys. Chem. Earth, 9, 571-604.
O'Hara, M. J., and H. S. Yoder, Jr. (1967) Formation and fractionation of basic
magmas at high pressures, Scott. J. Geo/., 3, 67-117.
O'Nions, R. K., and R. J. Pankhurst (1974) Petrogenetic significance of isotope and
trace element variations in volcanic rocks from the mid-Atlantic, J. Petrol., /5,
603-634.
Onuma, K., and K. Vagi (1967) The system diopside-akermanite-nepheline, Am.
Mineral., 52, 227-243.
Orowan, E. (1960) Mechanism of seismic faulting. In D. Griggs and J. Handin, eds.,
Rock deformation, Geol. Soc. Am. Mem. 79, 323-345.
Osborn, E. F. (1942)The system CaSi03 -diopside-anorthite,Am. J. Sci., 240, 751-788.
Osborn, E. F., and D. B. Tait (1952) The system diopside-forsterite-anorthite, Am. J.
Sci., Bowen Vol., 413-433.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

228 References
Oxburgh, E. R. (1964) Petrological evidence for the presence of amphibole in the
upper mantle and its petrogenetic and geophysical implications, Geo/. Mag., 101,
1-19.
Oxburgh, E. R., and D. L. Turcotte (1970) Thermal structure of island arcs, Geol.
Soc. Am. Bull., 81, 1665-1688.
Paterson, M. S. ( 1973) Nonhydrostatic thermodynamics and its geologic applications,
Rev. Geophys. Space Phys., II, 355-389.
Pfann, W. G. (1952) Principles of zone melting, J. Met., 4, 747-753.
Pfann, W. G. (1966) Zone melting, 2d Ed. , John Wiley & Sons, Inc., New York,
310 p.
Phillips, W. J. (1972) Hydraulic fracturing and mineralization, J. Geol. Soc. London,
218. 337-359.
Poldervaart, A. (1962) Aspects of basalt petrology, J . Geol. Soc . India, 3, 1-14.
Porath, H., D. W. Oldenburg, and D. I. Gough (1970) Separation of magnetic varia-
tion fields and conductive structures in the western United States, Geophys. J. R.
Astron. Soc ., 19, 237-260.
Powell, M., and R. Powell (1974) An olivine-clinopyroxene geothermometer, Contrib.
Mineral. Petrol., 48, 249-263.
Powell, R. W. (1954) Thermal conductivities of solid materials at high temperatures, ·
Research (London), 7, 492-501.
Presnall, D. C. (1969) The geometrical analysis of partial fusion, Am. J. Sci., 267,
1178-1194.
Presnall, D. C., T. H. O'Donnell, and N. L. Brenner (1973) Cusps on solidus curves
as controls for primary magma compositions: a mechanism for producing oceanic
olivine tholeiites of uniform composition, Abstr. with Programs (Geol. Soc. Am .),
5, 771-772.
Press, F. (1965) Dimensions of the source region for small shallow earthquakes
(unpublished VESIAC report).
Press, F. (1968) Earth models obtained by Monte Carlo inversion,J. Geophys. Res., 73,
5223-5234.
Press, F. (1969) The suboceanic mantle, Science, 165, 174-176.
Press, F. (1970) Earth models consistent with geophysical data, Phys. Earth Planet.
Inter., 3, 3-22.
Pyke, D. R., A. J. Naldrett, and 0 . R. Eckstrand (1973) Archean ultramafic flows in
Munro Township, Ontario, Geol. Soc. Am . Bull., 84, 955-918.
RAheim, A., and D. H. Green (1974) Experimental determination of the temperature
and pressure dependence of the Fe-Mg partition coefficient for coexisting garnet
and clinopyroxene, Contrib. Mineral. Petrol .. 48, 179-203.
Raleigh, C. 8., and M. S. Paterson (1965) Experimental deformation of serpentinite
and its tectonic implications, J . Geophys. Res., 70, 3965-3985.
Ramberg, H. (1971) Temperature changes associated with adiabatic decompression
in geological processes, Nature (London) , 234, 539-540.
Ramberg, H. (1972) Mantle diapirism and its tectonic and magmagenetic consequences,
Phys. Earth Planet. Inter., 5, 45-60.
Rankin, G. A., and F. E. Wright (1915) The ternary system CaO-AI20a-Si02, Am. J.
Sci., 39, 1-79.
Reay, A., and P. G. Harris (1964) The partial fusion of peridotite, Bull. Volcano/. ,
27, 115-127.
Reed, G. W., K. Kigoshi, and A. Turkevich (1960) Determinations of concentrations
of heavy elements in meteorites by activation analysis, Geochim. Cosmochim .
Acta, 20, 122-140.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

References 229
Reitan, P. H. (1968) Frictional heat during metamorphism, Lithos, / , 151-163, 268-274.
Richter, C. F. (1935) An instrumental earthquake magnitude scale, Bull. Seismol. Soc.
Am. , 25, 1-32.
Riecke, E. (1895) Uber das Gleichgewicht zwischen einem festen, homogen deformirten
Korper und einer flussigen Phase, insbesondere uber die Depression des Schmelz-
punktes durch einseitige Spannung, Ann. Phys . (Leipzig), 54, 731-738.
Riecke, E . (1912) Zur Erniedrigung des Schmelzpunktes durch einseitigen Zug oder
Druck, Zentralbl. Mineral. Geol. Palaeontol., 3, 97-104.
Ringwood, A. E. (1958) Constitution of the mantle, 3, Geochim. Cosmochim. Acta, 15,
195-212.
Ringwood, A. E. (1962) A model for the upper mantle, J . Geophys. Res., 6/, 857-867.
Ringwood, A. E. (1966) Mineralogy of the mantle. In P. M. Hurley, ed. , Advances
in earth science , MIT Press, Cambridge, p. 357-399.
Ringwood, A. E. (1969) Composition of the crust and upper mantle. In P. J . Hart, ed.,
The earth's crust and upper mantle, Geophys. Monogr., Am. Geophys. Union, /3 ,
1-17.
Ringwood, A. E. (1970) Phase transformations and the constitution of the mantle,
Phys. Earth Planet. Inter., 3, 109-155.
Rittmann, A. (1962) Volcanoes and their activity, John Wiley & Sons, Inc., New
York, 305 p.
Roberts, J. L. (1970) The intrusion of magma into brittle rocks. In G. Newall and N.
Rast, eds., Mechanism of igneous intrusion , Gallery Press, Liverpool, p. 287-338.
Robie, R. A., and D. R. Waldbaum (1968) Thermodynamic properties of minerals and
related substances at 298. 15°K (25.0"C) and one atmosphere (1.013 bars) and at higher
temperatures, U.S . Geol. Surv. Bull. 1259, 256 p.
Roeder, P. L. (1974) Paths of crystallization and fusion in systems showing ternary
solid solution, Am. J . Sci., 274, 48-60.
Rogers, J. J. W., and J . A. S. Adams (1966) Fundamentals of geology, Harper and
Row, New York, 424 p.
Ronov , A. 8 ., and A. A. Yaroshevsky (1969) Chemical composition of the earth's
crust. In P. J . Hart, ed., The earth's crust and upper mantle, Geophys. Monogr.,
Am. Geophys. Union , /3, 37-57.
Roscoe, R. (1952) The viscosity of suspensions of rigid spheres, Br. J. Appl. Phys. , 3,
267-269.
Rosenhauer, M., and D. H. Eggler (1975) Solution of H20 and C02 in diopside melt,
Carnegie lnst. Washington Yearb. 74, 474-479.
Rubey, W. W. (1951) Geologic history of sea water, Geol. Soc. Am. Bull., 62 ,
1111-1147.
Rubey, W. W., and M. K. Hubbert (1959) Role of fluid pressure in mechanics of
overthrust faulting, II, Overthrust belt in geosynclinal area of western Wyoming
in light of fluid-pressure hypothesis, Geol. Soc. Am. Bull., 70, 167-205.
Runcorn, S. K. ( 1967) International dictionary of geophysics, Pergamon Press , Oxford,
1728 p.
Rutter, E. H. (1972) The influence of interstitial water on the rheological behaviour
of calcite rocks, Tectonophysics, 14, 13-33.
Sacks, I. S., and H. Okada (1974) A comparison of the anelasticity structure beneath
western South America and Japan, Phys. Earth Planet. Inter., 9, 211-219.
Sanford, A. R., 0 . Alptekin , and T . R. Topposada (1973) Use of reflection phases on
microearthquake seismograms to map an unusual discontinuity beneath the Rio
Grande rift, Bull. Seismol. Soc. Am. , 63, 2021-2034.
Savage, J . C. (1969) The mechanics of deep-focus faulting, Tectonophysics, 8, 115-127.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

230 References
Savanick, G. A., and D. I. Johnson (1972) Adhesion at crystalline interfaces in rock,
U.S. Bur. Mines Rep. Invest. 7709, 16 p.
Scarfe, C. M. (1973) Viscosity of basic magmas at varying pressure, Nature (London),
Phys. Sci., 241, 101-102.
Scarfe, C. M., D. K. Paul, and P. G. Harris (1972) Melting experiments at one
atmosphere on two ultramafic nodules, Neues Jahrb. Mineral. Monatsh., 469-476.
Schairer, J. F. (1942) The system Ca0-Fe0-AI,03 -Si0,: I, Results of quenching
experiments on five joins, J. Am. Ceram. Soc., 25, 241-274.
Schairer, J. F. ( 1954) The system K 1 0-Mg0-AI1 0a-Si02 : I, Results of quenching experi-
ments on four joins in the tetrahedron cordierite-forsterite-leucite-silica and on the
join cordierite-muUite-potash feldspar, J. Am. Ceram . Soc., 37, 501-533.
Schairer, J. F. (1957) Melting relations of the common rock-forming oxides, J. Am.
Ceram. Soc., 40, 215-235.
Schairer, J. F., and N. L. Bowen (1942) The binary system CaSi03 -diopside and the
relations between CaSiOa and akermanite, Am. J. Sci., 240, 725-742.
Schairer, J. F., and N. Morimoto (1958) Systems with rock-forming olivines,
pyroxenes, and feldspars, Carnegie lnst. Washington Yearb. 57, 212-213.
Schairer, J. F., and N. Morimoto ( 1959) The system forsterite-diopside-silica-
albite, Carnegie Inst. Washington Yearb. 58, 113-118.
Schairer, J. F., C. E. Tilley, and G. M. Brown (1968) The join nepheliJle-diopside-
anorthite and its relation to alkali basalt fractionation, Carnegie lnst. Washington
Yearb. 66, 467-471.
Schairer, J. F., K. Yagi, and H. S. Yoder, Jr. (1962) The system nepheline-diopside,
Carnegie Inst. Washington Yearb. 6/, 96-98.
Schairer, J. F., and H. S. Yoder, Jr. (1960a) The nature of residual liquids from
crystallization, with data on the system nepheline-diopside-silica, Am. J. Sci.,
Bradley Vol., 258A, 273-283.
Schairer, J. F., and H. S. Yoder, Jr. (1960b) The system forsterite-nepheline-
diopside, Carnegie Jnst. Washington Yearb. 59, 70-71.
Schairer, J. F., and H. S. Yoder, Jr. (1961) Crystallization in the system nepheline-
forsterite-silica at one atmosphere pressure, Carnegie Jnst. Washington Yearb.
60, 141-144.
Schairer, J. F., and H. S. Yoder, Jr. (1964) Crystal and liquid trends in simplified
alkali basalts, Carnegie Inst. Washington Yearb. 63, 65-74.
Schairer, J. F., and H. S. Yoder, Jr. (1967) The system albite-anorthite-forsterite
at I atmosphere, Carnegie lnst. Washington Yearb. 65, 204-209.
Schairer, J. F., and H. S. Yoder, Jr. (1969) The join albite-anorthite-akermanite,
Carnegie lnst. Washington Yearb. 67, 104-105.
Schairer, J. F., and H. S. Yoder, Jr. (1970) Critical planes and flow sheet for a portion
of the system Ca0-Mg0-AI2 0 3 -Si02 having petrological applications, Carnegie lnst.
Washington Yearb. 68, 202-214.
Schairer, J. F., H. S. Yoder, Jr., and C. E. Tilley (1965) Behavior of melilites in the
join gehlenite-soda melilite-akermanite at one-atmosphere pressure, Carnegie Inst.
Washington Yearb. 64, 95-100.
SchiUing, J.-G. (1966) Rare earth fractionation in Hawaiian volcanic rocks, Massa-
chusetts Institute of Technology thesis, 390 p.
Schilling, J.-G. (1971) Sea-floor evolution: rare earth evidence, Philos, Trans. R.
Soc. London, Ser. A, 268, 663-706.
Schilling, J.-G. (1975) Rare-earth variations across "normal segments" of the Reyk-
janes Ridge, 6<r-53°N, Mid-Atlantic Ridge, 29"S, and East Pacific Rise, 2°-19"S,

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

References 231
and evidence on the composition of the underlying low-velocity layer, J. Geophys.
Res., 80, 1459-1473.
Schilling, J.-G., and E. Bonatti (1975) East Pacific ridge (2°S-19"S) versus Nazca
intraplate volcanism: rare-earth evidence, Earth Planet. Sci. Lett., 25, 93-102.
Schilling, J.-G., and J. W. Winchester (1967) Rare-earth fractionation and magmatic
processes. In S. K. Runcom, ed., Mantles of the earth and terrestrial planets,
Interscience Publishers, New York, p. 267-283.
Schmitt, R. A., R. H. Smith, J. E. Lasch, A. W. Mosen, D. A. Olehy, and J.
Vasilevskis (1963) Abundances of the fourteen rare-earth elements, scandium, and
yttrium in meteoritic and terrestrial matter, Geochim. Cosmochim. Acta, 27,
577~22.
Schmitt, R. A., R. H. Smith, and D. A. Olehy (1964) Rare-earth, yttrium and scandium
abundances in meteoritic and terrestrial matter-11, Geochim. Cosmochim. Acta,
28, 67-86.
Sclater, J. G., and J. Francheteau (1970) The implications of terrestrial heat flow
observations on current tectonic and geochemical models of the crust and upper
mantle of the earth, Geophys. J. R. Astron. Soc., 20, 509-542.
Seitz, M. G., and I. Kushiro (1974) Melting relations of the Allende meteorite,
Science, /83, 954-957.
Shaw, H. R. (1965) Comments on viscosity, crystal settling, and convection in granitic
magmas, Am. J. Sci., 263, 120-152.
Shaw, H. R. (1969) Rheology of basalt in the melting range, J . Petrol., /0, 510-535.
Shaw, H. R. (1970) Earth tides, global heat flow, and tectonics, Science, /68, 1084-
1087.
Shaw, H. R. (1973) Mantle convection and volcanic periodicity in the Pacific; evidence
for Hawaii, Geo/. Soc. Am. Bull., 84, 1505-1526.
Shaw, H. R. (1974) Diffusion of H 20 in granitic liquids: Part I, Experimental data;
Part II, Mass transfer in magma chambers. In A. W. Hofmann, B. J. Giletti, H. S.
Yoder, Jr., and R. A. Yund, eds., Geochemical transport and kinetics, Carnegie
Inst. Washington Pub/. 634, 139-170.
Shaw, H. R., and E. D. Jackson (1973) Linear island chains in the Pacific: result of
thermal plumes or gravitational anchors.? J. Geophys. Res., 78, 8634-8652.
Shaw, H. R., and D. A. Swanson (1970) Eruption and flow rates of flood basalts. In
E. H. Gilmour and D. Stradling, eds., Proceedings of the Second Columbia River
Basalt Symposium, EWSC Press, Cheney, Washington, p. 271-299.
Shaw, H. R., T. L. Wright, D. L. Peck, and R. Okamura (1968) The viscosity of
basaltic magma; an analysis of field measurements in Makaopuhi Lava Lake, Hawaii,
Am. J. Sci., 266, 225-264.
Shelley, D. (1974) Mechanical production of metamorphic banding-a critical appraisal,
Geol. Mag., Ill, 287-292.
Shimazu, Y. (1959) A thermodynamical aspect of the earth's interior-physical
interpretation of magmatic differentiation process, J. Earth Sci., Nagoya Univ., 7,
1-34.
Shimazu, Y. (1961) Physical theory of generation, upward transfer, differentiation,
solidification, and explosion of magmas, J. Earth Sci., Nagoya Univ., 9, 185-223.
Shimizu, N., and I. Kushiro (1973) Trace element content of liquid formed by partial
melting of a garnet lherzolite at high pressures: a preliminary report, Carnegie /nst.
Washington Yearb. 72, 270-272.
Shimizu, N., and I. Kushiro (1975) The partitioning of rare earth elements between
garnet and liquid at high pressures: preliminary experiments, Geophys. Res. Lett., 2,
413-416.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

232 References
Shimozuru, D. (1963) Geophysical evidence for suggesting the existence of molten
pockets in the earth' s upper mantle, Bull. Volcano/., 26 , 181-195.
Shreve, R. L. (1972) Movement of water in glaciers, J . Geol., Jl, 205-214.
SkiMer, B. J. (1956) Physical properties of end-members of the garnet group, Am.
Mineral. , 41, 428-436.
Skinner, B. J. (1962) Thermal expansion of ten minerals, U.S. Geol. Surv. Prof. Pap.
450D, 109-112.
Sleep, N. H. (1974) Segregation of magma from a mostly crystalline mush, Geol.
Soc. Am. BuU., 85, 1225-1232.
Smith, C. S. (1948) Grains, phases, and interfaces: An interpretation of microstructure,
Trans . AIME, 175, 15-51.
Sobolev, N. V., Jr.• I. K. Ku~netsova, and N. I. Zyuzin (1968) The petrology of
grospydite xenoliths from the Zagadochnaya kimberlite pipe in Yakutia, J . Petrol., 9,
253-280.
Sorby, H. C. (1863) Uber Kalkstein-Geschiebe mit Eindrucken, Neues Jahrb . Mineral.
Geol. Palaeontol., 801~7 .
Sorby, H. C. (1879) The anniversary address of the President, Proc. Geol. Soc.
London , 35, 56-95.
Spetzler, H. , and D. L. Anderson (1968) The effect of temperature and partial melting
on velocity and attenuation in a simple binary system, J . Geophys. Res. , 73,
605 1-(,()6().
Spry, Alan (1962) The origin of columnar jointing, particularly in basalt flows, J . Geol.
Soc. Aust. , 8, 191-216.
Starr, A. T. (1928) Slip in a crystal and rupture in a solid due to shear, Proc. Cambridge
Philos. Soc., 24, 489-500.
Steeples, D. W., and H. M. lyer (1976) Low-velocity zone under Long Valley as
determined from teleseismic events, J . Geophys. Res., 8/, 849-860.
Stocker, R. L. , and M. F. Ashby (1973) On the rheology of the upper mantle, Rev.
Geophys. Space Phys .. 2, 391-426.
Subramanian, A. P .• and Y. S. Sahasrabudhe (1964) Geology of greater Bombay and
Aurangabad-EUora-Ajanta area,/nt. Geol. Congr. , 22nd Session, New Delhi, India,
Guide to Excursions Nos. A-13 and C-10, p. 1-12.
Sun, S.-S .• M. Tatsumoto, and J .-G . Schilling (1975) Mantle plume mixing along the
Reykjanes Ridge axis: lead isotopic evidence, Science, /90, 143-147.
Swanson, D. A. (1972) Magma supply rate at Kilauea Volcano, 1952-1971, Science,
175, 169-170.
Tarakanov, R. z.. and N. V. Leviy (1968) A model for the upper mantle with several
chaMels of low velocity and strength. In L. Knopoff. C. L. Drake, and P. J. Hart.
eds., The crust and upper mantle of the Pacific area, Geophys. Monogr., Am.
Geophys. Union , /2, 43-50.
Tatsumoto, M. (1966) Genetic relations of oceanic basalts as indicated by lead isotopes,
Science , 153, 1094-1101.
Tatsumoto, M., C. E. Hedge, and A. E. J. Engel (1965) Potassium, rubidium,
strontium, thorium and uranium in oceanic tholeiitic basalt, Scienee, /50 , 886-888.
Taylor, H. P., Jr. (1974) The application of oxygen and hydrogen isotope studies to
problems of hydrothermal alteration and ore deposition, Econ . Geol., 69, 843-883.
Tera, F .• D. A. Papanastassiou, and G. J. Wasserburg (1974) Isotopic evidence for a
terminal lunar cataclysm, Earth Planet. Sci. Lett., 22, 1-21.
Thompson, R. N. (1972) Melting behavior of two Snake River lavas at pressures up to
35 kb, Carnegie lnst. Washington Yearb . 71 , 406-410.
Thomson, J. (1862) On crystallization and liquefaction , as influenced by stresses tending

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

References 233
to change of form in the crystals, Proc. R. Soc . London, 1/, 473-481. [Reproduced
in Phil. Mag .. 24, 395-401 ( 1862).]
Thorarinsson, S. (1954) The tephra-fall from Hekla on March 29th 1947. InT. Einarsson,
G. Kjartansson, and S. Thorarinsson, eds .. The eruption of Hek/a 1947-1948,
Vol. II, No. 3, Societas Scientiarum lslandica, Reykjavik, 68 p.
Tikhonov, A. N., E. A. Lubimova, and V. K. Vlasov (1970) On the evolution of
melting zones in the thermal history of the earth, Phys. Earth Planet. Inter., 2,
326-331.
Tilley, C. E., H. S. Yoder, Jr., and J. F. Schairer (1964) New relations on melting of
basalts, Carnegie Inst. [Link] Yearb. 63, 92-97.
Tocher, D. (1958) Earthquake energy and ground breakage, Bull. Seismo/. Soc. Am .,
48, 147-153.
Toksoz, M. N., J. W. Minear, and B. R. Julian (1971) Temperature field and geo-
physical effects of a downgoing slab, J. Geophys . Res., 76, 1113-1138.
Tolland, H. G., and R. G. J. Strens (1972) Electrical conduction in physical and
chemical mixtures; application to planetary mantles, Phys. Earth Planet. Inter .. 5,
380-386.
Tolstikhin, I. N. (1975) Helium isotopes in the earth's interior and in the atmosphere:
a degassing model of the earth, Earth Planet. Sci. Lett., 26, 88-96.
TuthiU, R. L. (1969) Effect of varying [o 2 on the hydrothermal melting and phase
relations of basalt (abstract), Eos, 50, 355.
Tuttle, 0 . F., and N. L. Bowen (1958) Origin of granite in the light of experimental
studies in the system NaAISi 30 8- KAISi30s-SiO.-H.O, Geo/. Soc. Am. Mem. 74.
153 p.
Ubbelohde, A. R. ( 1965) Melting and crystal structure, Clarendon Press, Oxford, 325 p.
Uffen, R. J. (1959) On the origin of rock magma, J . Geophys. Res., 64, 117-122.
Uffen, R. J ., and A. M. Jessop ( 1963) The stress release hypothesis of magma formation,
Bull. Volcano/. , 26, 57~.
Van Schmus, W. R., and J. A. Wood (1967) A chemical-petrologic classification for
the chondritic meteorites, Geochim. Cosmochim. Acta, 31, 747-765.
Verboogen, J. (1946) Volcanic heat, Am. J. Sci., 244. 745-771.
Verboogen, J. (1954) Petrological evidence on temperature distribution in the mantle of
the earth, Trans. Am. Geophys. Union, 35, 85-92.
Verboogen, J. (1973) Possible temperatures in the oceanic upper mantle and the forma-
tion of magma, Geol. Soc. Am . Bull .. 84, 515-522.
Vernon, R. H. (1968) Microstructures of high-grade metamorphic rocks at Broken Hill.
Australia, J. Petrol., 9, 1-22.
Vernon, R. H. (1970) Comparative grain-boundary studies of some basic and ultrabasic
granulites, nodules and cumulates, Scott. J. Geo/., 6, 337-351.
Viljoen, M. J., and R. P. Viljoen (1969a) The geology and geochemistry of the lower
ultramafic unit of the Onverwacht Group and a proposed new class of igneous
rock. In The Upper Mantle Project, Geo/. Soc. S. Afr. Spec. Pub/. 2, 554!5.
Viljoen, M. J., and R. P. Viljoen (1969b) Evidence for the existence of a mobile
extrusive peridotitic magma from the Komati Formation of the Onverwacht Group.
In The Upper Mantle Project, Geol. Soc. S. Afr. Spec. Pub/. 2, 87-112.
Viljoen, M. J., and R. P. Viljoen (1969c) Evidence for the composition of the primitive
mantle and its products of partial melting from a study of the rocks of the Barberton
Mountain land. In The Upper Mantle Project, Geo/. Soc. S. Afr. Spec. Pub/. 2.
275-295.
Vinogradov, A. P., A. A. Yaroshevsky, and N. P. llyin (1971) A physicochemical

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

234 References
model for element separation in the differentiation of mantle material, Philos.
Trans. R. Soc. London, Ser. A, 268, 409-421.
Vogt, J. H. L. ( 1904) Die Silikatschmelzlosungen, II, Uberdie Schmelzpunkt-Erniedrigung
der Silikatschmelzlosungen, Videnskabs-Selskabets Skrifter, /, Math.-Naturv. Klasse
(Christiania), No. I, 235 p.
von Cotta, B. (1858) Geologische Fragen, J. A. Engelhardt, Freiberg, 344 p.
Wachtman, J. B., Jr. (1974) Highlights of progress in the science of fracture of ceramics
and glass,J. Am. Ceram. Soc., 57, 509-519.
Wager, L. R. (1958) Beneath the earth's crust, Adv. Sci., 15, 31-45.
Wager, L. R., G. M. Brown, and W. J. Wadsworth (1960) Types of igneous cumulates,
J. Petrol., I, 73-85.
Wagner, P. A. (1928) The evidence of the kimberlite pipes on the constitution of the
outer part of the earth, S. Afr. J. Sci., 25, 127-148.
Wakita, H., H. Nagasawa, S. Uyeda, and H. Kuno (1967) Uranium, thorium and
potassium contents of possible mantle materials, Geochem. J., I, 183-198.
Waldbaum, D. R. (1971) Temperature changes associated with adiabatic decompression
in geological processes, Nature (London), 232, 545-547.
Walsh, J. B. (1969) New analysis of attenuation of partially melted rock, J. Geophys.
Res .. 74, 4333-4337.
Warner, R. D. (1973) Liquidus relations in the system Ca0-Mg0-Si02-H2 0 at 10 kb
PHt<J and their petrologic significance, Am. J. Sci., 273, 925-946.
Washington, H. S. (1917) Chemical analyses of igneous rocks, U.S. Geol. Surv. Prof.
Pap. 99, 1201 p.
Washington, H. S. (1925) The chemical composition of the earth, Am. J. Sci., 9,
351-378.
Weertman, J. (1968) Bubble coalescence in ice as a tool for the study of its deforma-
tion history, J. Glacio/., 7, 155-159.
Weertman, J. (1972) Coalescence of magma pockets into large pools in the upper
mantle, Geol. Soc. Am. BuU .. 83, 3531-3532.
White, W. P. (1909) Melting point determination, Am. J. Sci., 28, 453-489.
Whitehead, J. A., Jr., and D. S. Luther (1975) Dynamics of laboratory diapir and
plume models,J. Geophys. Res., 80, 705-717.
Whitten, C. A. (1956) Crustal movement in California and Nevada, Trans. Am. Geophys.
Union, 37, 393-398.
Wickman, F. E. (1966) Repose period patterns of volcanoes: I, Volcanic eruptions re-
garded as random phenomena; II, Eruption histories of some East Indian volcanoes;
III, Eruption histories of some Japanese volcanoes; IV, Eruption histories of some
selected volcanoes; V, General discussion and a tentative stochastic model, Ark.
Mineral. Geo/., 4, 291-367.
Wiederborn, S. M., and L. H. Bolz (1970) Stress corrosion and static fatigue of glass,
J. Am. Ceram. Soc., 53, 543-548.
Wilshire, H. G., and J. W. Shervais (1975) At-augite and Cr-<liopside ultramafic
xenoliths in basaltic rocks from western United States, Phys. Chern. Earth, 9,
257-272.
Wilson, J. T. (1963) Continental drift, Sci. Am., 208, 86-100.
Wood, .9. J., and S. Banno (1973) Garnet-orthopyroxene and orthopyroxene-
clinopyroxene relationships in simple and complex systems, Contrib. Mineral.
Petrol., 42, 109-124.
Wood, J. A. (1963) On the origin of chondrules and chondrites, Icarus, 2, 152-180.
Wright, T. L., W. T. Kinoshita, and D. L. Peck (1968) March 1965 eruption of

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

References 235
Kilauea volcano and the formation of Makaopuhi lava lake, J. Geophys. Res. , 73,
3181-3205.
WyUie, P. J. (1970) Ultramafic rocks and the upper mantle, Mineral. Soc. Am. Spec.
Pap. 3, 3-32.
Yang, H.-Y., J. F. Salmon, and W. R. Foster (1972) Phase equilibria of the join
akennanite-anorthite-forsterite in the system Ca0-Mg0-AI10 3 -Si01 at atmospheric
pressure, Am. J. Sci., 272, 161-188.
Yoder, H. S., Jr. (1952) Change of melting point of diopside with pressure, J. Geol., 60,
364-374.
Yoder, H . S., Jr. (1958) Effect of water on the melting of silicates, Carnegie lnst.
Washington Yearb . 57, 189-191.
Yoder, H. S., Jr. (1965) Diopside-anorthite-water at five and ten kilobars and its
bearing on explosive volcanism, Carnegie Jnst . Washington Yearb . 64, 82-89.
Yoder, H . S., Jr. (1968) Experimental studies bearing on the origin of anorthosite,
N.Y. State Mus . Sci. Serv. Mem. 18, 13-22.
Yoder, H . S., Jr. (1969) Calcalkalic andesites: experimental data bearing on the origin
of their assumed characteristics. In A. R. McBimey, ed., Proceedings of the
Andesite Conference, Oreg. Dep. Geol. Miner. Ind. Bull. 65, 77-89.
Yoder, H . S., Jr. (1973) Melilite stability and paragenesis, Fortschr. Mineral., 50,
140-173.
Yoder, H . S., Jr. (1974) Gamet peridotite as a parental material for basaltic liquids,
Carnegie lnst. Washington Yearb. 73, 263-266.
Yoder, H. S. , Jr. (1975a) Heat of melting of simple systems related to basalts and
eclogites, Carnegie Jnst . Washington Yearb. 74 , 515-519.
Yoder, H. S., Jr. (1975b) Relationship of melilite-bearing rocks to kimberlite: a
preliminary report on the system akennanite-C02 , Phys. Chern. Earth , 9, 883-894.
Yoder, H- S., Jr., and I. Kushiro (1969) Melting of a hydrous phase: phlogopite,
Am. J . Sci. , Schairer Vol., 267A , 558-582.
Yoder, H. S., Jr., D. B. Stewart, and J. R. Smith (1957) Ternary feldspars, Carnegie
lnst. Washington Yearb. 56, 206-214.
Yoder, H. S ., Jr., and C. E. TiUey (1957) Basalt magmas, Carnegie lnst. Washington
Yearb . 56, 156-161.
Yoder, H. S., Jr., and C. E. TiUey (1961) Simple basalt systems, Carnegie lnst.
Washington Yearb . 60, 106-113.
Yoder, H. S .• Jr. , and C. E. TiUey (1962) Origin of basalt magmas: an experimental
study of natural and synthetic rock systems, J. Petrol., 3, 342-532.
Yokoyama, I. (1957) Energetics in active volcanoes, 2nd paper, Bull. Earthquake Res.
lnst. Tokyo Univ., 35, 15-97.
Yoshiki, B. , and R. Yoshida (1952) Composition of low-alkali glass, J . Am. Ceram.
Soc., 35, 166-169.
Yund, R. A. (1974) Coherent exsolution in the alkali feldspars. In A. W. Hofmann, B. J.
Giletti, H. S. Yoder, Jr., and R. A. Yund, eds., Geochemical transport and kinetics,
Carnegie Jnst . Washington Pub/. 634, 173-183. •
Zemansky, M. W. (1937) Heat and thermodynamics, McGraw-Hill Book Company,
New York, 388 p.

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Author Index

Adams, J. A. S., 3, 229 Benioff, H., 5, 214

Adams, L. H., 19, 38, 66, 171,214 Berg, E., 50, 52, 223
Akella, J., 20, 31,213 Berman, H., 95, 224
Akimoto, S.-1., 33, 34, 58, 59, 85 , 140, Betz, F., Jr., 195, 214
213, 223, 226 Biggs, W. D., 172, 214
Allen, C. R., 186, 216 Birch, F., 19, 20, 24, 38-40, 47, 51 , 52,
Allen, E. T., 94, 213 67, 80, 171,214,215
Allison, E. B., 164, 213 Blank, H. R., Jr. , 50, 218
Almukhamedov, A. I., 106, 227 Boettcher, A. L., 39, 85, 197, 215, 225,
Alptekin, 0 ., 51,229 226
Anderson, D. L., 40, 52, 53, 180, 182, Bolz, L. H., 180, 234
213, 232 Bonatti, E., 154,231
Anderson, 0 ., 125, 213 Bowden, F. P., 71,215
Anderson, 0 . L., 60, 74, 214 Bowen, N. L., 12, 14, 24, 28, 66, 90, 92,
Anthony, T. R., 162, 214 95, 107, 123, 125, 130, 215 , 230, 233
Aoki, K.-1. , 27, 223 Boyd, F. R., 18, 30-32, 35 , 58, 59, 78,
Arculus, R. J., 9, 135, 138, 148, 150, 151, 169, 202, 213, 2 is, 217, Plate 2 (p. 44b)
214 Brace, W. F., 165, 171, 178, 215
Arndt, N. T., v Bradley, R. S., 91, 93,215
Ashby, M. F., 74, 173, 174, 214, 232 Brenner, N. L., 60, 64, 228
Ave'Lallemant, H. G. , 175, 176, 216 Bridgman, P. W., 19, 171,216
Brock, P., 164, 213
Bailey, D. K .. 60, 214 Brooks, C., 42, 216
Baker, D. W., 73 , 220 Brown , G. M., 15, 121, 131, 230,234
Bancroft, D., 19, 215 Brown, M. C. , 106, 218
Banno, S., 31 , 227, 234 Brune, J. N., 186, 216
Barton, D. C., 189, 214 Buddington, A. F., 60, 63. 216
Barus, C., 90, 214 Bullen, K. E .. 171, 216
Bath, M., 183, 214 Bultitude, R. J., 148, 216
Bayly, B., 189, 214 Burke, J. E., 163, 217
Beeson, M. H .. 135, 214 Burnham , C. W., 79, 216
Beg, M. A., 84, 217 Busch, W., 164, 165, 225
Bell, P. M. , v, 60, 76, 77, 131, 177, 214, Butkovich, T. R., 185, 216
225 Butler. B. C. M., 165, 216

237

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

238 Author Index

Cameron, M. , 87, 216 Eaton, J. P. , 4, 5, 50, 182, 218
Cannon, J. H ., 165, 216 Eckstrand, 0 . R., 15, 228
Cannichael, I. S. E., 192-194, 225 Eggler, D. H., 78, 81, 83, 84, 132, 133,
Carr, M. J., 187, 216 148, 198, 218, 229
Carter, N. L., 31, 175, 176, 216, 225 Einstein, A., 112, 218
Cawthorn, R. G., 71,216 Emslie, R. F . , 27, 218
Chamberlin, T . C . , 61, 176,216 Endo, E. T. , 192, 223
Charlu, T. V ., 91, 93 Engel, A. E. J., 98, 191, 192, 209, 218,
Charney, J. G., iii 232
Chayes, F. , iv, v, 76, 130, 216 Engel, C. G., 98, 191, 192, 209, 218
Chen, C .-H., 59, 216 England, J. L. , 18, 58, 59, 78, 170, 215 ,
Christiansen, R. L., 50, 218 217
Clark, H., 67, 215 Erlank, A. J., 42, 218
Clark, P. W., 165, 216 Eshelby, V. D. , 52, 218
Clark, S. P., Jr., 41, 85, 125,216 Evans, A. G ., 180, 218
Clarke, W. B., 84, 217
Clayton, R. N., 25, 217 Fahrig, W. F . , 62, 63, 219
Cline, H . E. , 162, 214 Faure, G . , Ill, 219
Coats, R. R., 5-7, 20, 217 Feather, J. N., 179, 220
Coble, R. L., 163, 217 Fedotov, S. A., 50, 219
Coe, R. S., 172, 217 Feild, A. L. , 191 , 219
Cohen, L. H., 135, 137, 140, 217 Ferguson, J . 8., 123, 219
Cordani, U . G ., 106, 217 Fermor, L. L., 20, 219
Cotter, J. R., 195, 196, 217 Finger, L. W. , 87, 219
Cox, A., iii Fiske, R. S., 182-184, 219
Cox, K . G ., 106, 217 Foster, W. R., 123, 125, 218, 219, 235
Craig, H ., 84, 217, 224 Francheteau, J., ~98. 231
CrandeU, D. R., 2, 217 Frank, F. C., 166, 227
Crittenden, M. D. , Jr. , 175, 217 Frey, F. A., 211,220
Cross, C . W. , 29 Friedman, H. , iii
Cutlers, R. L., 153, 217 Friedman, M. , 179, 220
Curran, E. 8 ., 9, 214 Fujii, T., 30, 219

Dalrymple, G. 8 ., 194, 195, 222 Gale, N.H . , 48, 226

Daly, R. A., 19, 61, 76, 86,217 Gast, P. W., 24, 42, Ill , 157, 219
Davis, B. T. C. , 32, 58, 59, 95, 96, 108, Gastil, G ., 192, 219
117, 142, 143, 149, 159, 160, 170, 215, Gault, D. E. , 183, 219
217 Geological Museum of London, 181,219
Davis, N . F . , 79, 80, 216, 217 Gettings, M. E., 50, 218
Day, A. L., iii, 61, 86, 218 Gibson, R. E. , 19, 213
Decker, R. W. , iv, v Gilbert, M. C., 22, 219
DeLury, J. S., 60, 72, 218 GiUuly, J ., 175, 219
de Neufville, J. , 91, 125,216 Goetze, C. , 31,219
de Wys, E. C., 125, 218 Goldsmith, J. R., iii
Dickey, J. S., Jr., v, 2, 202, 218 Goody, R. M., iii
Dickson, G. 0 ., 175, 221 Gorshkov, G . S., 50, 219
Dubourdieu , G. G., 191.218 Gough, D. 1., 49, 228
Dunn, P. R., 106, 218 Green, D. H . , 4, 15, 17, 27, 31, 86, 103,
Du Toit, A . . 191.218 148, 151,216,219,220, 228
Green, W. L., 19, 220
Griffith, J . S. , 125
Griggs, D . T ., 60, 73, 74, 171, 220

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Author Index 239

Grout. F. F .• 60, 68, 220 Jackson. E. D.• 10. 22. 66. 135, 182, 194.
Gruntfest, I. J., 74, 220 195, 214. 219. 222. 231
Gummer, W. K., 125, 220 Jaggar. T. A., 61, 86. 222
Gutenberg, B .• 8, 51, 53, 186, 220 James. D. E., 42. 216
Jeffreys. H., 171
Hager, R. V .• Jr., 179,220 Jessop. A.M .• 61,233
Haller, W., 165. 220 Johannsen, A .• 29, 222
Handin, J., 60, 73. 74, 171, 179, 220 Johnson, D. I., 172. 230
Hare, P. E., Plate I (p. 44a) Joly. J., 61, 195, 196. 222
Harker, A .• 44. 176, 220 Jordan, T. H., 51, 96. 222
Harris. P. G., 34. 100. 113-115.220,228. Joule, J.P., 66, 222
230 . Judd. J. W., 112, 222
Hart. J. R., 42. 216 Julian. B. R .• 60. 72, 73. 233
Hart, S. R., 24. Ill. 119, 220. 221
Hartgen, F. A .• 191.221 Kable. E. J. D .• 42, 218
Haskin, L. A., 153-155, 211. 217, 220, Kaula, W. M.• 171, 222
221 Kawada. K., 67. 68. 222
Haskin, M.A., 211.220 Kehle, R. 0 .. 189
Havens, R. G .• 191. 192. 209,218 Kennedy. G. C.. 18. 20, 33, 34. 113-115.
Hazen, R. M., 87, 220 135. 137, 140. 148, 213. 217, 222
Heard, H. C., v. 175, 179, 220. 221 Kigoshi. K .. 25, 228
Hedge, C. E., 42, 98, Ill, 219.221.232 Kinoshita, W. T .• 112. 113. 234
Heidish, J. A .• 191,221 Kirby, S. H., 74. 222
Heirtzler, J. R., 175, 221 Kleppa, 0. J .• 91, 93
Heitowit, E. D.• 183, 219 Knopoff. L., 51 , 222
Helmke. P. A., 154, 155. 221 Koto, K., 58, 226
Hendrix, W. C., v Koyanagi. R. Y .• 192, 223
Herron, E. M., 175, 221 Kozu, S .• 171,223
Herty, C. H .• Jr.• 191. 221 Kretz, R., 3 I. 223
Hess, H. H., 60, 63, 64, 100, 195, 214. Kubota, S., 50, 52, 223
221 Kuno, H., 7-9, 27, 98, 106. 223. 234
Hibberson. W., 27. 220 Kushiro. I., v, 17, 18, 20. 25-27. 29,
Hodges, F. N .• 81, 221 33-38, 59, 81' 83-86. 100. 123, 125.
Hofmann. A. W., v,42. 76,111.119,221 132, 140. 143, 144, 146, 148, 154, 155.
Holmes, A .• 29, 61, 86, 97. 98. 187. 195, 159, 198,223.224-226.231,235
221 Kuznetsova, I. K., 26, 232
Hopson, C. A.• 183. 219
Hubbert, M. K., 179, 221. 229
Huckenholz, H. G., v, 147, 221 Lacroix. A .• 22, 224
Hulme, G., 190, 221 Lambert, I. B., 22, 224
Hurley, P.M., 38, 221, 222 Larsen, E. S., 94. 95, 224
Hytonen, K., 91. 125, 222 Lasch, J. E., 211,231
LeComte. P .• 20, 80, 215
Iddings, J. P., 29 Lee, W. H. K., 96, 97, 99, 224
lida. K., 186, 222 Le Pichon, X., 175. 221
llyin, N. P., 101, 233 Leviy, N. V., 8, 232
Imlay, M., v Lindemann, F. A .• 86. 224
Irvine, T. N., v, 35, 222 Lindsley, D. H., 27, 170,218.224
Irving, A. J., 83, 86. 222, 225 Linzer, M., 180, 218
Ito, K., 33, 34, 113-115. 135, 137, 140, Liu, L., 60. 224
148, 217, 222 Lliboutry, L., 166, 224
lyer, H. M., 50, 51, 218,232 Lovering. J. F .• 24, 224

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

240 Author Index

Lubimova, E. A .• 60, 66, 67, 97, 196. Mullineaux. D. R .• 2. 217
197. 224. 233 Murase. T .• 52, 67. 68. 226
Lupton. J. E .. 84, 224 Murata, K. J .• 4, 5, 218
Lure . M. L.. 106, 224 Mysen, B. 0 ., v, 30. 81. 85, 94, Ill , 153,
Luth. W. C . . 165, 215.224 198, 218. 224. 226
Luther. D. S .. 68, 234
Lyons. J. B .. v Nagasawa. H .. 98. 234
Nakamura, Y .. 84. 226
Maalee. s.. 94. 224 Naldrett, . . J .. 15. 228
Mabey, D. R .• 50.218 Naslund, H. R., v, 191, 225
MacDonald, G. A., 185, 224 Nesbitt. R. W.. 15. 226
MacDonald. G. J. F .. 98. 224 Nesterenko. G. V .• 106. 227
MacGregor. I. D., 29, 30. 32. 224 Nettleton. L. L.. 189, 227
Machado. F .. 51. 224 Nicholls. I. A.• 15. 17. 220
Manson. V .. iv. 225 Nier, A. 0 .. 98. 227
Mao. H. K .. 60. 76. 77. 123. 125. 131. Nishikawa. M .. 198. 224
177.214.225 Nixon. P. H .. 31. 32. 35,202. 215
Marsh. B. D .. 192-194. 225 Nockolds. S. R.. iv. 34, 227
Masaitis. V. L.. 106. 224 Norris. F. G .. 191. 221
Mason. B .• 28 Nur. A .• 51. 52. 55, 180. 227
Masuda. A .• 154. 155. 225 Nutting. P. G . . 60. 71. 72. 227
Mayeda. T. K .. 25. 217 Nye. J . F .. 166. 167. 227
McBimey, A. R., v, 52. 60. 6fHi9. 100.
101. 191. 225. 226 Obata. M .. 30. 31. 227
McGetchin. T . R .• 9-11. 225 O'Donnell, T. H .. 60. 64. 228
McGregor. V. R.• 48. 226 Ohashi. Y .• 87. 219
McKenzie. D. P .. 60, 73. 225 O'Hara. M. J .. 31. 33. 37. 94. 140. 142-
McLean, D .• 165. 225 147. 149. 150. 159, 227
McQuarrie, M., 67. 225 Okada. H .. 54. 55 . 229
Means. W. D .. 169.225 Okamura. R.• 175. 190. 231
Medaris. L. G .. Jr .. 153. 217 Oldenburg. D. W .. 49. 228
Mehnert. K. R.• 164. 165. 225 Olehy. D. A .. 211. 231
Menard. H. W .• 3, 192. 195, 225 O'Nions. R. K . . 48. 118. 226. 227
Menzies. M .• 2. 225 Onuma. K .. 123. 227
Mercier. J.-C .. 31. 225 Orowan. E.. 74. 227
Mercy. E. L. P .. 31. 227 Osborn. E. F .. 16. 91. 93. 125. 227
Merrill. R. B.. 23. 225 Oxburgh. E. R . . 37. 103. 228
Merwin. H . E .• 123. 219
Metcalfe. K .. 191. 221 Pankhurst. R. J .. 48. 118. 226. 227
Millhollen. G. L.. 86. 225 Papanastassiou. D. A .. 48. 232
Minear. J. W .• 60, 72. 73. 233 Papike. J . J . . 87. 216
Modreski. P. J .• 39, 225 Paterson. M.S .. 172. 177. 179.217. 228
Mogi, K., 178, 179, 182. 226 Paul. D. K .. 113. 114. 230
Moorbath, S .. 48. 226 Peck, D. L., 112. 113, 175, 190. 231. 234
Moore, P. B., 58, 226 Perkins. P. C .. 60. 74. 214
Morey, G. W .• 196, 226 Peterman. Z. E.. Ill. 221
Morgan, J . W., 24. 224 Pfann . W. G . . 100. 115. 228
Morgan, W. J ., 49, 195, 226 Phillips. W. J . . 179. 228
Mori, T., 31. 227 Pirsson. L. V., 29
Morimoto, N .. 58, 123, 226, 230 Pitman, W. C.. III. 175. 221
Morse. S. A., v, 80, 226 Pitt, A.M., 50, 218
Mosen, A. W. , 211.231 Poldervaart. A .. 209. 228

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Author Index 241

Porath. H .• 49, 228 Schilling, J.-G., vi, 110, Ill, 150-154.
Powell, J. L.. Ill, 219,220 156, 158-160, 220, 230-232
Powell, M., 31, 228 Schmitt, H. H., Plate I (p. 44a)
Powell, R. W., 31, 67.228 Schmitt, R. A .• 212, 231
Presnall. D. C., vi, 59, 60, 64, 107, 216, Schneider. G., 164, 165, 225
228 Sclater. J. G., 60, 73,96-98, 225,231
Press. F., 41. 55, 71, 186, 228 Seitz, M.G., 27, 35, 81,218,231
Prewitt, C. T., 87, 216 Shaw, H. R., 60, 65. 66, 74, 75, 84, 168,
Pyke, D. R .. 15. 228 175, 190, 199, 231
Shelley, D .• 169, 231
Shepherd, E. S. , 61, 86, 218
RM!eim, A., 31. 228
Shervais, J. W., 166. 234
Raleigh. C. B .• 74, 179, 222, 228
Shimazu, Y., 115, 118, 192, 231
Ramberg, H., 60, 66, 68. 70, 71. 192, 228
Shimizu, N .. 100, 150, 151, 154, 155,214.
Rankin, G. A., 125. 228
231
Reay, A .. 34. 228
Shimozuru, D .. 7, 49, 50, 232
Reed, G. W .. 25. 228
Shreve, R. L .. 166, 232
Reitan. P. H .. 71. 229 Silver, E. A., 194, 195, 222
Richter, C. F., 186, 220, 229
Singer, A. D., v
Riecke. E .. 177, 229
Skinner, B. J .. 94, 171, 232
Ringwood, A. E., 4, 20, 41. 59, 85. 98.
103, 146, 148, 216, 220, 229 Sleep, N.H., 176, 232
Smith, C. S .. 164, 232
Rittmann, A., 61. 229
Smith. J. R., 80, 235
Roberts, J. L.. 171.229
Smith, J. V .. 58, 226
Robertson, E .. vi
Smith, R. H., 211, 231
Robie. R. A .. 90-92. 229
Sobolev, N . V., Jr., 26.232
Roeder. P. L.. 107. 229
Sorby, H. C.. 177, 232
Rogers, J. J. W.. 3, 229
Spetzler, H .. 52. 53, 213. 232
Ronov. A. B .. v, 229
Spry, Alan. 173. 232
Roscoe, R .. 112. 229
Starr, A. T., 73, 232
Rosenhauer, M .. 83. 229
Steeples, D. W., 51, 232
Rowley. J. R .. 52
Stewart, D. B .. 80, 235
Royer, M. B .. 191. 221
Stocker, R. L., 74, 173, 174, 232
Royster. P. H .. 191. 219
Stoiber, R. E., 187. 216
Rubey. W. W .. iii. 84. 179. 221. 229
Stone. M.A., 71. 215
Rubin. M., 2. 217
Strens, R. G. J .. 87, 233
Runcom. S. K .. 171.229
Subramanian. A. P. , 106. 232
Russel. L. R .. 180. 218
Sueno, S .. 87. 216
Rutter. E. H .. 179. 229
Sun. S.-S .. Ill. 232
Sutte~J. F., 191,225
Sacks. I. S .. 54. 55. 229 Sutton. K. G., 191, 225
Sahasrabudhe. Y. S .. 106. 232 Suzuki, T .. 52. 226
Salisbury. R. D .. 61. 176.216 Swanson, D. A .. 76. 183, 199. 231. 232
Salmon. J. F .. 125. 235 Syono, Y .. 33. 34. 59. 85, 140. 223
Sanford. A. R .. 51. 229
Saunders. M. J .. 31. 227
Savage. J. C .. 167. 168. 229 Tait. D. B .. 16. 93. 125. 227
Savanick. G. A .. 172. 230 Takahashi. E .. 30. 219
Scaife. C . M .. 113. 114. 175. 230 Tarakanov. R. Z .. 8, 232
Schairer. J. F .. 13. 14. 28. 48. 90-92. 95. Tatsumoto. M .. 98, Ill. 232
96. 108. 117. 121-123. 125. 129-132. Taylor, H. P .. Jr .. 84. 232
143. 159. 160. 215-217. 222. 225. 230. Tera. F .. 48. 232
233 Thomas. D., v

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

242 Author Index

Thomas, P. H . , 71, 215 Wagner, P. A., 22, 234
Thompson, R.N., 17, 135, 136, 140,223, Wakita, H., 98, 234
232 Waldbaum, D. R. , 60, 66,90-92,229,234
Thomson, J., 66, 177, 222, 232 Walsh, J. B., 49, 52, 234
Thorarinsson, S., 200, 233 Wanless, R. K .• 62, 63, 219
Tikhonov, A. N., 196, 197, 233 Warner, R. D.• 36, 234
Tilley, C. E., 7-9, 13-18, 20-22, 34, 60, Washington, H. S., 24, 29, 234
92, 93, 103, 104, 107, 120-122, 128, Wasserburg, G. J., 48, 232
130-132, 135, 140-142, 146, 147, 230, Waters, A. C., 106, 183, 219
233, 235 Weertman, J ., 168, 169, 234
Tilton, G. R., 42, Ill, 219 White, C. M., 191, 225
Tocher, D., 186, 233 White, J., 164, 165, 213, 216
Tokarev, P. I., 50, 219 White, W. P., 90, 94, 213, 234
Tokonami, M., 58,226 Whitehead, J. A., Jr., 68, 234
Toks3z, M. N., 60, 72, 73, 233 Whitten, C. A .• 175, 234
Tolland, H. G., 87, 233 Wickman, F. E., 191, 199, 234
Tolstikhin, I. N., 84, 233 Wiederhom, S. M., 180, 234
Topposada, T. R., 51, 229 Wildeman, T. R., 211, 220
Tudor, G. K., 71, 215 Williams, P. F., 169, 225
Tullis, J. A., vi Williamson, E. D., 19, 38, 171.213
Tunell, G., 88 Willis, D. G., 179, 221
Turcotte, D. L., 103, 228 Wilshire, H. G., 166, 234
Turkevich, A., 25, 228 Wilson, J. T., iii, 195, 234
Tuthill, R. L.. 22, 233 Winchester, J. W., 110, 151-153,231
Tuttle, 0. F., 130, 233 Windom, K. E., 27
Tuve, M. A., iii Wood, B. J., 31, 234
Wood, J. A., 28, 45, 233, 234
Ubbelohde, A. R., 87, 233 Wright, F. E. , 125. 228, 234
Ueda, J., 171, 223 Wright, T . L., 10, 22, 112, 113, 175, 190,
Uffen, R. J. , 60-62, 233 195, 222. 231
Unger, J. D., 165, 215 Wyllie, P. J., 4, 5, 6, 22, 23, 83, 86, 197,
Uyeda, S., 98, 234 215, 222, 224, 225, 235

Vandoros, P., 106, 217 Yagi, K., 123, 227. 230

Van Schmus, W. R .• 28, 233 Yang, H.-Y., 125, 235
Vasilevskis, J., 211, 231 Yaroshevsky, A. A., v, 101,229, 233
Velde, D., vi Yoder, H. S .• Jr., 7-9, 12-18, 20-22.
Verhoogen, J., 60, 65, 88, 191,233 25-27,29,33,34.37.48,58,60-62, 78,
Vernon. R. H .• 162, 163, 233 80-83, 90-95, 102-104, 107. 120-123.
Viljoen, M. J., 15, 17, 220, 233 128-132, 135, 140-147, 149, 150, 159,
Viljoen, R. P., 15, 17, 220, 233 197-199, 202, 218, 223, 224, 227. 230.
Villard, 0 . G., iii 233, 235
Vinogradov, A. P .. 101. 233 Yokoyama, I., 183, 185, 235
Vlasov, V. K ., 196, 197,233 Yoshida, R., 123, 235
Vogt, J. H . L., 90, 234 Yoshiki, B., 123, 235
von Cotta, B., 19, 234 Yund, R. A., 172, 235

Wachtman, J . B., Jr., 180, 234 Zemansky, M. W., 88, 235

Wadsworth, W. J ., 15, 234 Zietz, I., 50, 218
Wager, L. R., 15, 35, 234 Zyuzin, N. I., 26, 232

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Subject Index
The student of petrology may find it useful to formulate his own views on the concepts
and principles summarized by a word or phrase and define the terms in The Subject Index
prior to scholastic examination.

Accretion Albite-forsterite reaction, 146

rate of, 45 Aleutian Islands, Alaska
sweep-up time, 45 spacing of volcanoes, 192, 193
Achondrite Alkali basalt
mantle material, 24 depth of origin, 205
texture, 45, 46 effect of C02 on melt generation, 205
Adams-Williamson relationship, 38, 39 fractional melting product, 161
Adhesive strength, 172 incompatible phases in, 17
Adiabatic compression, 72 invariant melt product, 161
Adiabatic gradient, 70, 103 nodules in. 17, 29, 189
equation, 102 occurrence during geologic time, 209
Adiabatic process rare-earth elements, I50
definition, 61 simulated, 148
Adiabatic rise, 65, 101-103, 196,203,204 viscosity, volatile effect on , 175
depth of initiation, 104 Allende meteorite
Aggregate oxygen pressure effects on, 35, 36
liquid-bonded, 164, 165, 207 partial melt analysis, 35
Aggregation Amphibole
dependence on diffusion , 176 absence in Hawaiian lavas, 22
in source area, In breakdown, 23
of liquid, 207 consumed in initial melts, 38
of magma, 181 K20-bearing phase, 37
role of deformation, 166 present in nodules, 22
under stress, 167-169, 172 role in basalt generation, 23
Albite stability limits, excess H20, 23
enthalpy of melting, 91 Amphibolite, v
melting rate, I 19 melting interval, 21, 22
volume change on melting, 170 pressure-temperature relations, 21
Albite diabase, 119 relationship to basalt, 21. 22
243

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

244 Subject Index Amp-Bas

source material, 19 energy ofhypervelocity impact in, 183

stability limits, 22 enthalpy of melting, 90
Anatexis, 2, 164 equilibrated, 16, 17
Andesite equivalent to eclogite, 18, 20
hydrous partial melt product, 36, 37, eutecticlike, 17, 92
48 genetic relationship to nodules, 22
parentage, 12 Hawaiian, 150, 152
volatile source, 84 heat production, 97
Andesitic magma inherited composition, 17
yield stress, 190 initial melt, 131
Anhydrous basalt intrusive form, iv
ocean floor, 84 iron enrichment, 14
Ankaramite, 15 liquid, 12
rare-earth element abundances, 152 liquidus temperature, 13
Annealing recrystallization, 177 melting at invariant point, 19
Anorthite melting interval, excess H 2 0, 22
enthalpy of melting, 91 melting interval, 0.1 wt.% H 20, 22
volume change on melting, 170 most common magma, v
Anorthite-enstatite reaction, 26, 27, 146 Nazca Plate, 154
Anorthite-forsterite reaction, 64, 146 nodules in, 85
effect of albite, 27 norms, 130
effect of iron, 27 not primary liquid, 17
pressure-temperature relations, 25, 27 occurrence, 12
Arkose olivine content at high pressures, 135
natural fusion of. 165 pockets of, 5
Arrhenius equation, 74 pressure-temperature relations, 18
Aseismicity, 55 principal minerals, 16, 17, 120-121
Aspect ratio relationship to amphibolite, 21, 22
definition, 53 relationship to eclogite, 205
droplets, 52 relationship to garnet peridotite, 146,
of flaws for shear melting, 73, 74 202
Assimilation, 7 relationship to other rock types, 12
zone melting, 116 restricted by water pressure, 21
Asthenosphere, 3, 180, 182 source material, 19, 64
terrestrial, 35
thermal conductivity, 66, 68
B
variation of isotopes, 42
Barberton Mountain Land, 15 varieties, iv
Basalt volume of, 12
absence of olivine at high pressures, volume in crust, v
159 Basalt, alkali
advanced hydrous melt product, 37 depth of origin, 7, 8, 9
age, 191, 192 forms shield, 9
anhydrous partial melt product, 36, 37 island arc occurrence, 9
average compositions, iv last eruptive on island, 9
changed en route to surface, 202 Basalt. high-alumina
complete melting of, 19 depth of origin, 7, 8, 9
definition. iv Basalt tetrahedron, 120-123
degree of partial melting. 100 anorthite analogue, 121. 125
depth restriction. 18. 20 CaTs analogue, 121
derivative. 146, 159 plagioclase analogue. 121. 125
differentiation of, v subsolidus subtetrahedra, 124

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Bas~ Ia Subject Index 245

Basaltic magma Canary Islands, 9
changes en route to surface, 135 Carlxmaceous chondrite
compositional changes, 145, 159 mantle material, 25
depth of separation, 142 volatile content, 25
homogeneity, 204 Carbon dioxide
major types, 120-133 in diopside melt, 81 '
most common type, 205 magma saturation, 81
predominance, 204 Cascades, Washington
relationship to garnet peridotite, 32, spacing of volcanoes, 192, 193
139 Caverns (see Reservoir), 176
viscosity, 175 Cavities in mantle, 4
Basanite, 147, 148 Chamber
depth of separation, 135 apophysis, II
pressure-temperature diagram, 138 auxiliary, v, 8, 10, II, 49, 103
Basanitoids, 150 boundary, 182
Batch melting, 107, 205 collapse, 4
compared to zone melting, 118 crystallized, 2
Batholith, 2 depth of, 3, 50
rarity of gabbroic, 71 elongated vertically, 7
Beginning of melting, 67, 74, 89, 91, 101, failure of, 172
108, 126, 131 magma escape from, 172
absence of olivine, 134 model of, 172
hydrous, 80, 81 plastic envelope, 207
variation with pressure, 132 pressure on, 187
variation with volatiles, 132 refilling, 3
whole rock involved, 162 shape, 3, 10, 50, 55, 56, 180
Benioff zone, 5, 7, 8, 9, 73 size, 3
layered structure, 8 subsidiary (see auxiliary), v
Beta-M82Si02 , 58, 59 vertical distribution, 50
Beta-track counting, 115 viscous collapse, 199
Brittle fracture, 172, 176, 180 Chamber rupture
Broken Hill, Australia, 163 role of gas pressure, 197
Bronzite gabbro, 35 Chemical potential
Bulk modulus (see Compressibility), 38 in pressure gradient, 76, 77
Buoyancy, 186, 188, 189, 203, 208 Chemical zoning
Buoyant layer, 68, 70 in crystals, 162
rate of rise, 193 Chili-Peru region, 55
Chondrite
c abundance, 24
Calcite melting curve, 83 classification, 28
Calcium-Tschermak's molecule, 91 compared with basalts, 24
potential anorthite, 34 homogeneity, 28
Caldera, 51 isotopic composition, 24
Calorimetric methods mantle material. 24, 35
cooling-curve, 90 mineral content, 25
differential thermal analysis, 89 partial melt Fe/Mg ratio, 35
drop-calorimeter, 89 related to peridotite, 25
freezing-point depression, 90 texture, 45
Calorimetry Chondrules
iron-bearing rocks, 90 quenched liquid droplets, 45
Canadian shield CIPW System (see Norm), 29
map of, 63 Clausius-Ciayeron equation, 95

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

246 Subject Index Cli-Def

Clinoenstatite (see Enstatite), 91 Corundum

Clinopyroxene crack propagation in, 180
crystal-liquid REE partition coeffi- Corundum normative liquids, 109
cients, 155 Crack propagation, 60, 73, 208
melting relations with Gr. 142 by magmafracting, 179
nonn, 150 in corundum, 180
nonnative An content, 34 rate of, 180
quench, 107 related to volcano age, 194, 195
Coble creep, 174 Crack systems
Columbia River plateau lava, 106 converging, 7
Compositional cusp, 60 Crater (see Vent), 10
Compressibility, 170 Creep, 74
relationship to density, 38 Crust
Compression andesitic, 48, 209
cyclical, 61 base of defined, v
Compressive failure, 62 crystallinity, 49, 50
Compressive strength of rock depression of, 63, 64
confining pressure effect on, 178 derivation from mantle, 48
Conchoidal fracture, 173 oldest, 48
Condensation recycled, 48
of primordial gas, 45 subduction of, 49
temperature, 47 turnover of, 48
Conductive beat model, 65 Crust formation
Conductivity in lava lake, 113
electrical, 49 through volcanism, 47-48
Conduit, 131 Crystal
diapiric, 193 adhesive strength, 172
in ice, 166 Crystal-liquid mush, 112, 116, 176
network, 164-166 Crystal-liquid partition coefficients
of volcano, 55 rare-earth elements, 154
pressure in, 187 Crystal physics
size of, 192 melting, 86
Constant volume Crystal settling, 19, 103
melting under, 171 Crystal sponge, 4
Convection, 203 Crystal structure
primary driving force, 208 changes near melting point, 86, 87
solid state, 65 mechanical instability, 87
transient nature of, 65 Cumulate
volume of melt required, 116 liquidus, 15
Convective rise olivine, 15
isenthalpic, 60 remelting of, 15
isentropic, 60, 65 Cusp
Conversion factors, 211 compositional, 60, 63, 64, 204
Coprecipitation garnet peridotite melting, 113, 114
clinopyroxene-garnet, 149 phase-change, 60, 64, 204
diopside-forsterite, 37
major phases, 13
Core of earth
D
elements in, 29
melting in, 201 Darcy's Law, 177
Core of earth formation Deccan plateau lava, 106
temperature rise in, 47 Deformation mechanism map, 173, 174

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Def-Dyn Subject Index 24 7

Deformation mechanisms, 177 Diatreme, 9, II
Degassing of mantle, 84, 209 material in, II
Degree of depletion size of vent, II
measure of, 206 Diffusion, 76
Degree of melting, 55, 65, 120 differential rates, 76
disaggregation of parental material, equilibration time, 119
154 in liquid, 100
estimate based on REE, 150, 152, 159 limitations, 76
influence on magma norm, 15~152 role in deformation mechanisms, 173
liquidus temperature relationship, 154 under pressure differential, 60
products, 151. 152 Diffusional creep, 174
source material, 176 Diffusion melting, 78
Density Dike (see Conduit), 10, 172
contrasts of, stabilized thermally, 42 age of, 63
eclogite, 41 diabase, 63
peridotite, 4 I. 43 orientation control, 180
perturbation of, 60, 68, 204 swarm. 62
relationship to compressibility, 38 Diopside (see Clinopyroxene)
Depletion, 4 breakdown assemblage, 59
in peridotite. 161 CO.-H 2 0 saturation in melt, 81,83
measurements of, 202 compressibility. 171
of rare earths, 110 constant volume P-T change, 171
of trace elements. 110 content in mantle, 84
Depth density of crystal, 94
magma generation. 49 density of glass, 94
Depth of separation enthalpy of melting. 91
basalt type. 104 germanate analogue. 59
midocean environment, 103 incongruent melting of. 59
Derivative liquids. 76 melting behavior. 58, 59
Derivative magma. 133 melting with volatiles, 78
eclogite, 142. 159 Dislocations
Diabase, iv, 63 role in melting. 87
compressive strength under confining Diversity
pressure, 178 of igneous rocks, 121
tensile strength. 171 Dolerite. iv
thermal conductivity. 66 Dolomite
Diapir in mantle. 84
aggregation in mantle. Ill stability, 83
detachment. 68. 112 Droplet, 163
heat losses. 71 aspect ratio. 52, 164
primordial frequency. 42 coalescence. 168. 169
rate of rise. 70, 193 in garnet peridotite, 168
salt dome. 189 internal pressure, 163
shape, 7, 70 intersection probability, 165
size vs. rate of rise, 208 shape, 53
Diapiric rise, 70. 71. 208 Ductile fracture. 172
experimental demonstration. 68 Dunite, 30, 109
initiation, 118 residuum. 146
periodicity. 196. 197 thermal conductivity. 66. 68
process described, 193 Dynamic instability
termination of, 42 of buoyant layer. 71

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

248 Subject Index Ear-Ent

mineral assemblage, 20, 141
E olivine content at high pressures, 135
Earth (see Protoearth) olivine-free, 143, 148, 206
condensation pressure of. 45 origin of. 145
cool accumulation of. 44 pockets in peridotite, 20
differentiated. 44 pockets of, 5
distribution of rock types in. 44 pressure-temperature relations. 18
early historical record. 202 range of composition. 139. 141
form of geoid. 171 rare-earth abundance during melting.
fundamental oscillations. 41. 49 calculated, 156. 157
initially molten. 44. 45 recycled basalt. 148
origin of. 44. 201. 202 related reactions. 20
principal volatiles. 78 relationship to garnet peridotite. 33.
state of stress in. 171 139, 145. 159
surficial melting of. 48 residuum. 20
Earthquake seismic velocity variation. 39. 40
energy release. 171 source material. 19
greatest on record. 186 Eclogite-basalt reaction. 18. 102. 103
magnitude related to energy release. pressure-temperature diagram. 18
186 volume change. 103
maximum depth. 55. 171 Eclogite facies
present-day. 7. 8. 9 density constraint. 41
related to magma movement. 182. 186 Elastic rebound. 62
related to volcanism. 180. 182 Endothermic reactions
source dimension. 186 metamorphic. 89
Earthquake energy Energy
related to magma generation. 184. 186 partition of. 183
Earthquake prediction Energy release
based on dilatancy. 180 seismic, 4
Earthquake zone Enrichment factor
related to volcanic zone. 180. 181 rare-earth elements. calculated. 160
Eclogite. v. 30. 204 Enstatite
adiabatic rise of. 102. 103 breakdown assemblage. 59
alteration of. 39 content in mantle. 84
beginning of melting. 102. 103 enthalpy of melting. 91
chemical analyses projected. 141 hydrous melt curve. 198
crystallinity. 19 incongruent melting of. 59
cumulates. 145 volume change on melting, 170
degree of partial melting required for. Enthalpy of crystallization
42 (see Enthalp~ of melting). 115
density. 103 Enthalpy of freezing
derivative product. 142. 146. 159. 202 (see Enthalpy of crystallization). 167
equivalent to basalt. 18. 20 Enthalpy of melting. 64. 204
experimental synthesis. 20 Ab-An-Di-Fo system. 93
Fe/Mg variation in. 39. 40 An-Di-Fo system. 93
gravitational stability. 42 anomalous heat capacity. 89
high-pressure form of gabbro. 20 anorthite--diopside ··eutectic'' com-
liquid, 20 position. 90
liquidus, 102. 103 basalt. 90
mantle material. 40 constraint on magma generation. 88
melting range, 20 definition. 88
melting relation to peridotite. 20 Di-En-Fo-Py system. 95

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Ent-Fra Subject Index 249

Di-Fo-Py system, 95 Fe/Mg ratios
diopside-anorthite system, 92 parent vs. derivative magma, 139
diopside-pyrope system. 94 Ferrobasalts, lunar, 35
forsterite-fayalite system, 92, 93 Film of liquid, 172
garnet peridotite, 95 Filter pressing (see Kneading out), 176
guide to heat sources, 88 Filtration differentiation (see Kneading
pressure effect on, 94 out), 176
related to anomalous crystal structure. Flaw, 167, 168
87 shear melting, 74
related to rate of magma production, Floods of lava
89 duration, 208
water effect on, 95 related to plate breakup, 191
Enthalpy of reaction Flow
eclogite to basalt.. 103 plastic, 170, 177
Environment of island arc, 7 Flow of liquid
Equilibrium constant, Ill Darcy type, 177
Eruption distance from site of generation, 177
duration. 106 intergrain boundary. 176
explosive, I. 9, II Knudsen type, 177
geologic hazard. I Poiseuille type, 177
mechanics, v under stress, 168
most violent, 3 Flow sheet
periodicity, 89, 191-195 at high pressures, 131, 133
prediction of. I basalt tetrahedron. 129
quiescent, I. II, 199 described, 125-126
recent, 51 shortcomings, 131
total energy. 3 Fluid inclusion (see Droplet), 52
Eruptive cycles, 191 migration, 162
Eshelby-Walsh Theory, 52 Forsterite
Etna, Sicily density of crystal. 95
yield stress of magma. 190 density of glass, 95
Eutectic enthalpy of melting, 90, 91
definition. 17 glass, 93
described, 95, 124 hydrous melt curve, 198
pressure effect on, 132 melting behavior. 58, 59
Eutecticlike melting, 113-115 phase change. 58
Exothermic chemical reactions, 61, 86 pressure effect on stability field. 146
Explosive volcanism structure near melting temperature. 87
role of volatiles, 198, 199. 208 triple point, 44, 45, 46
volume change of melting. 95. 170
Fractional crystallization. 17. I07. Ill.
115
Fractional melting. 107. Ill, 154. 204
F
compared to zone melting, 118
Fayalite Fractionation. 14. 19, 42, 76. 92
enthalpy of melting, 91 of isotopes, Ill
structure near melting temperature. 87 role in amphibole, 23-24
volume change on melting. 170 role of olivine, 135
Fault Fracture
length related to earthquake mag- brittle. 172, 176, 204
nitude. 186 conchoidal, 173
Feeder (see Conduit), 166. 172 ductile, 172

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

250 Subject Index Fra-Geo

multiple, 170 mantle material, 40, 202
product of magmafracting process, melting at I atm., 113, 114
179, 180, 182 melting at high pressure, 113, 114
Frequency of eruption, 191-195 metamorphism, 204
Friction mode in mantle estimated, 40
coefficient of, 71 mode vs. temperature, 113, 114
Frictional melting, 71, 73 occurrence, 29
evaluation of factors, 73 partial melt analyses, 34, 35
partial melt composition, 32, 36, 37, 43
primary mantle, 31
G principal minerals, 30
Gabbro, iv, v rare-earth abundances during melting,
depth restriction, 20 calculated, 159-160
source material, 19 relationship to meteorites, 42
thermal conductivity, 66 residua, 31
Gamet (see Pyrope), 60 seismic velocities, 38, 43
crystal-liquid REE partition coeffi- sheared, 34, 169
cients, 155 solidus, 140
in meteorites, 26 source of basaltic melt, 202
melting relations with clinopyroxene, source material, 19
142 specimen photograph, Plate I (p. 44a)
norm, 150 stability at depth, 28, 29
persistence during melting, 154, 159 stability limit of constituent minerals,
seismic velocity, 40 60
Gamet dunite, 107 texture, 162
Gamet lherzolite (see Gamet peridotite), Gamet peridotite, synthetic
32,33 eutectic composition, 35
chemical analysis, 35 Gamet wehrlite, 116
depleted of basaltic components, 34 Gas (see Gas phase)
reaction products, absence of plagio- composition in Ab-H 20, 80
clase, 33, 34 composition in mantle, 84
solidus, 140 diffusion from depth, 10
transformed to spinel peridotite, 33 primordial. 61
Gamet peridotite Gas cloud
basaltic magma, 139 cooling behavior, 45, 46
color at depth, 43 toroidal, 47
composition of coexisting minerals, 29 Gas fluxing (see Volatile fluxing), 10
constituent mineral melting behavior, Gas phase
58-60 effect on REE partitioning, 153
contiguous mineral assemblages, 30 presence in mantle, 81
cumulate, 31 silica content, 84
degree of depletion, 202 species identification, 45
enthalpy of melting, 95 Geotherm
eutecticlike, 142. 145 oceanic, 22. 85
field of stability. 43 shield, 22, 85
from deep-seated environment. 42 Geothermal gradient, 64, 196
granular, Plate 2 (p. 44b) changes with time, 100, 101
heat production, 97 Geobarometer, 30
initial stages of melting, 164. 165 Fe-Mg partitioning, 31
isenthalpic rise rate, 66 Geothermometer, 30
isentropic rise rate, 66 Ca-Mg partitioning, 30
key assemblage, 31 sensitivity to minor elements, 31

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Gla-Hua Subject Index 251

Glass (see Vitreous material, Tachylite) Heat balance
basalt, 12 convective, 65
detennination in experiments, 115 Heat capacity
dip quench. 15 anomalous, 89
iron enrichment, 17 Heat flow
layer in mantle, 19 at various depths, 96, 97
natural quench, 15 primordial rates, 42
quench product, 113 related to volcanism, 51
retention in mantle, 19 sea floor, 73
viscosity, 19 Heat generation
Glenelg, Scotland, 141 radioactive isotopes, 99
Globule (see Droplet), 169 Heating
Gneiss radioactive, 86
quartz-feldspar, 164 shear-strain, 72
Grain boundary stress, 73
coherent, 172 viscous, 71, 167
composition, 162-163 Heat offusion, 88
melting along, 165 Heat of melting, 88
Grain-boundary melting, 207 supplied adiabatically, 61
Granite (see Rhyolite), 64, 97 Heat of mixing, 93
initial melting of, 165 Heat producing elements
partial melt product, 48 abundances, 97, 98
Graphical techniques Heat production
relating parental material and melt related to magma production, 209
products, 31 Heat source
Gravity anomaly, 49, 171 conversion of gravitational energy, 44
Grospydite, 26 decay of radioactive elements, 44
Grossularite, 147 impacting, 45
Gruntfest Theory, 74 internal or external of magma, 207
point, 3
Heat transfer
convective, 115, 195, 196, 208
H
Heat transport
Haleakala, Hawaii, 141 by flow of melt, 166
Harzburgite, 26, 30, 40 mass movement, 65
Hawaii Hekla, Iceland, 199, 200
nodules from, 22 Henry's law, 107
Hawaiian eruptions Heterogeneity of magma
volume reaching surface, 76 variation of REE abundances, 154, 159
Hawaiian magma Hocheifel, Gennany, 147
yield stress, 190 Homogeneity
Hawaiian volcanoes mineralogical, 163
age-spacing relationship, 194 of liquid, 165
source region, 75 Honshu, Japan, 186
Heat Hornblendite
definition, 88 quench product, 22
from mechanical energy conversion, source material, 19
71 Hortonolite, 87
sink, 73 Hot spot, 195
viscous dissipation, 166 Hot Springs, 51
Heat "accumulation," 66, 67 Hualalai, Hawaii, 23

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

252 Subject Index Hyd-Kat

Hydrofracting Invariant point

related to magrnafracting, 179 magma generation, 64
Hydrogarnet, 83 Invariant point melting, 19
Hydropyroxene, 83 Invariant points
Hydrous melt curves number of, 131
end-member numerals, 198 Iron-bearing systems, 121
Hydrous melting, 168, 170 Iron enrichment
Hydrous minerals relation to liquidus, 13
stability in melt region, 83 lsenthalpic rise, 60, 66
Hydrous phases related to isentropic rise, 66
hydrogamet, 83 Isentropic rise, 60, 65
hydropyroxene, 83 Island arc
influence on magma H 20 content, 199 Alaskan peninsula, 7
Hypersthene alkali basalt occurrence in, 9
occult, 128 Japan, 8, 9
Hypersthene basalt, 127, 130 Isostatic rebound
rates of, 175
Isotopes
Ice differences in proportions, Ill
glacial, 168 helium, 84
temperate, 166 hydrogen, 84
Ice brine oxygen, 84
seismic velocity, 52 variation between magmas, 118
Igneous petrology
framework of, 120, 131 J
Igneous rocks Jadeite, 122, 123
bimodal character, 76 hydrous melt curve, 197, 198
Impact bodies Jadeite molecule
radius of, 48 potential albite, 34
Impact heating, 45 Jan Mayen Island, 9
Impacting Japan
climax, 202 seismic attenuation in, 54, 55
homogenization by, 202 Joly-Cotter model, 101, 195, 196, 200,
period of, 48 208
severity, 48 Joule-Thomson equation, 66
lmpactite, 48
Impact melting K
culmination, 48 K20-bearing phase
Inclusions amphibole, 37
related to initial melt, 162 clinopyroxene, 38
Inflation (see Tumescence), 170 phlogopite, 37, 81
Initial melt K20 content of liquid
distribution, 164, 165 homogeneity. 166
droplet, 163 K20 content of mantle
Interfacial angles between grains. 162, chondrite limitation, 38
163 minimum amount, 38
Intrusions Kakanui, New Zealand, 23
form of, 6 Kalsilite, 81, 82
"Invariant" Karroo lava, 106
described, 130 Katmai region, Alaska, 50, 51, 52

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Kil-Liq Subject Index 253

Kilauea, Hawaii, 4, 23, 199 Limestone
Kimberlite pore pressure vs. strength, 179
nodules from, 85 Lineation of volcanoes, 195
Siberian, 26 Liquid (see Magma, Melt)
volatiles in, 84 accumulation, 163-165
Kinetic-thermal energy conversion, 183 basaltic composition, 12
failure with onset of melting, 204 content related to magma rise, 203
Kliuchevsky volcano courses of fractionation, 123-129
Kamchatka, 50, 51 distribution in rock, 207
Kneading out liquid, 168, 175-177, 207 film, 164, 165, 176
Knudsen flow, 177 flow in stressed solid, 177
Kodiak, Alaska, 50 homogeneity, 165
Komati formation, 15 initial, 109
Komatiite "lakes," 165
liquid, 17 localization, 167-169
occurrences, 15 normative olivine content, 152
quench texture, 17 related to groundmass, 154
role of volatiles, 17 residual, 61, 176
separation, 204
squeezed out, 4
L
subsequent, 109
Latent heat of fusion, 88 trace element content, 205
Lava viscosity, 175
annual production, 76 viscosity related to structure, 191
discharge rates, 105 volume of, 117
floods, 106 Liquid content of magma
flow process, 173 temperature variation, 112-114
flow rate, 120 Liquid enrichment
lake, 190 in normative Hyp, 132
Lava lake in normative La, 145
crust formation, I 13 in normative Ne, 132, 148
Makaopuhi, Hawaii, 112, I 13 in normative Qz, 132, 145
Layer rare-earth elements, 154
buoyant, 68 Liquid inclusions (see Droplet), 52
monomineralic, 107 Liquid residuum
Layered intrusions, 107 primordial, 19
Layers in mantle Liquid segregation, 65
fusible, 10 Liquid separation
"hard," 8 by mass movement, 176
origin, 20 Liquidus
"plastic," 8 anhydrous, 80
refractory, 10 clinopyroxene, 148
Lesotho, 31, 32, 35 definition, 13
Leu cite, 8 I. 82 garnet, 148
stability limit, 39 of cumulates, 15
Lever rule, 107, 109, 117 olivine, 24
definition, 107 pahoehoe lava, I 13
Lherzolite, 30, 32 parent-daughter test, 14
plastic deformation, 175, 176 phase at high pressure, 135
thermal conductivity, 67, 68 pressure variation of, 148, 149

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

254 Subject Index Liq-Mag

pumice, 113 generation, iii
relationship to composition, 13 hydration, 84
relationship to iron enrichment, 13 intrusive, 2
volatile effect on, 15, 81 iron-enrichment, 15
Lithostatic load (see Rock load), 189 kinds of, related to depth, 7
Lithosphere, 180, 182 liquid content, 112-115
Loch Duich, Scotland, 141 major concepts, iv
Long Valley, California, 51 Newtonian fluid, 189, 190
Low-velocity zone normative change with pressure, 149,
absence of, 51 150
compositional change, 51 parental, 12
partial melting, 51, 52 penetration of volcano. 9
rigidity' 51 pockets of, 7
seismic velocity variation in, 52, 53 "pools," 168
solid-phase change, 51 primordial residuum, 49
source of magma, 8 proportion reaching surface, 183, 185
temperature gradient, 51 pseudoplastic fluid, 189
volatiles present, 86 rise, 4
separation, 148
shallowest depth of, 203
site of origin, iv, 2
M
storage, 176
Magma structural environment relationship,
aggregation, 182, 205 55, 57
amount reaching surface, 76 superheat, 146
annual production, 189 trace element variation factors, 206
batches, 108-109 transported, iii, 2, 55
Bingham fluid, 189, 190 viscosity, measured, 190
boundary, definition of, 4 volume erupted related to earthquake
boundary of, 10, 207 energy, 187
chamber shape, 180 Magmafracting
composition in subducted plate, 73 related to hydrofracting, 179
constancy of composition, 204 result of magma generation, 179
crystallization, iii Magmafracting process, 207
cumulate in, 15 Magma composition
daughter, 12 abrupt change of, 110
definition, iii constancy of, 195
density, 4, 5 continuous change, 110
depth of generation, 8, 9, 55 loss-of-phase effect on, 110
depth of separation, 4-6, 8, 9 Magma fractionation
depth of separation fixed by low- effect on liquidus, 139
density layer, 5 Magma generation
depth of separation fixed by shear abundance radioactive elements, 99
zone, 5 adiabatic rise, 101
derivative, 15 depth of, 50
differentiation, 4 depth of, relation to earthquake
direct observation. iv swarms,4
droplets of, 7 depth range, 7
equilibration of, 131 mechanisms, time variation of, 208
flow, 4, 177 period, 62
fluidity, 167 related to earthquake energy, 184, 186

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Mag-Mel Subject Index 255

Magma nonn depletion, depth of, 48
variation with degree of melting, 152. differentiation, 101, 201
153 gas composition, 84
volatile influence, 153 gas phase presence, 81
Magma production heterogeneity, 10, 42, Ill, 112, 201
frequency, 191. 192 homogeneous, 8
rate of, 199 hydrous melting of, 81, 86
Magma rise metamorphosed. 20 I
forces causing, 186-189, 207, 208 metasomatism, 84
liquid content change, 203 oceanic vs. continental, 51
Magma reservoirs plastic zone, 180
depth of, 50 "plum pudding" concept, 43
Magma separation remelting of, 42
factors involved, 133 seismic velocity, 5, 52
influence on magma type, 120 shear in, 168
Magma source state of stress, 61
duration, 192 suboceanic. 41
model of, balloon/soda-straw. 3 tensile strength. 171
model of, Benioff-zone, 5 turnover, 48
model of, crystal-mush. 4 volatile content, 203
model of, depth-of-generation, 7 volatile depletion, 84, 86
model of. explosive-diatreme. 9 water content, 84
model of, geophysical, 4 water flux, 84
single, 118 Mantle material
under volcano. 9 achondrite. 24
Magma Tap Program, iv carbonaceous chondrite, 25
Magma type characteristic property. 115
depth of generation, 9, 132 chondrite, 24
detennining factors. 204, 205 Mauna Loa, Hawaii
eutecticlike behavior. 115 volume of volcano, 193
from different layers, 7 Mechanical energy
number of invariant points, 154 converted to heat. 71. 203
principal control. 120 Mechanical instability
related to invariant points, 205 crystal structure. 87
role of volatiles, 85. 205 Megacrysts. 119
Magma withdrawal Melilite assemblages. 122
beginning of melting, 134 Melilite nephelinite
related to yield stress. 191 rare-earth element abundances. 152
Magma zone related to degree of melting. 151
life of, 118 Melt
Major phases amount generated in nuclear explo-
restriction in kinds of, 202 sion, 184
Makaopuhi, Hawaii, 112, 190 distribution in rock, 164
Mantle effect of stress on, 207
amount molten, 51. 52 films. 52
composition, 42 localized, 7
content of diopside, 84 migration of. 165
crystalline. seismic evidence, 12 phase contribution to. 206
crystallinity. 49. 50 primordial residuum, 48
degassing of, 84 seismic velocities. 52
density, 5, 41 source region. 61
structural state. 86. 87

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

256 Subjert Index Mel-Mol

viscosity, 175 under constant volume, 171
Melting under stress, 167
at composition cusp. 63 volatile effects on, 78. 81, 86
batch, 105. 107 volume change, 170
beneath volcano. 73 zone, 105. 115
complete, 104 Melting curve
compositional change. 76 related to solubility curve, 177
crystal physics, 86, 87 types of, 57, 58, 60
depth of. 44 Melting process
depth of. related to volatile content. density perturbation. 70
203 Melting processes, 57-4!,7
diffusion in pressure gradient, 76, 77 list of mechanisms. 60
disequilibrium, 105, 118 Melt withdrawal, 204, 205
equilibrium. 105 batch. 176
eutecticlike, 204 fractional, 176
fractional. 105. 107. 108 methods. 107-109
frictional. 71 small amounts. 107
grain-boundary. 163 subsequent, 109
in Benioff zone. 73 Melt zone
influence of stress on. 172 extension of. 100
influence on heat flow. 100 Metamorphic differentiation. 169
initial stages. 113. 114 Metamorphic rock. 130
initiation at cusp. 204. 205 Metasomatic processes. 76
in situ. 2 Metasomatism
in subducted plate. 72 in mantle. 84
in toto. 120. 204 Meteorite
in toto, physical problems. 20 absence of suitable garnet. 29
mechanisms for initiation, 203 Allende, 26
nonequilibrium. 105. 115 carbonaceous chondrite. 25
of hydrous phase, 81 composition. 27
participating phases. 117 condensed matter. 27
peridotite, 85 debris of disrupted planet. 27
phenomenological view. 86 low-pressure assemblage. 26, 27
physicochemical constraints. 105-119 mantle material, 24
radioactive heat. 99 metamorphosed mantle material. 29
rate in laboratory. 118 mineral equilibrium in. 28
rate of, 99 normative olivine after melting. 28
REE abundance changes. 156-158 oxidation range. 29
regenerative feedback. 74. 75 oxygen isotopes. 25
role of deformation, 166 related to peridotite. 27. 29
shear, 71. 72, 74 relation to planets. 26. 27
site of initiation, 206 source material. 19
spinel lherzolite, 85 Mica peridotite. 37
spot, 195 beginning of melting. 39
stress-difference, 71 Midocean Ridge. 204
tectonophysics of, 162-182 Minerals in mantle. 201
temperature of thermal conductivity Mixing
minimum, 66~9 in liquid. 118
theory of, 87 magma and roof rocks. 115. 116
thrust faulting, 72 Mohorovi~ic discontinuity. v. 5
tidal deformation, 71 Molten basalt
tidal dissipation. 75-76 seismic velocity. 52

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Moi-Ord Subject Index 257

Molten region
crack penetration, 10 0
depth of. 203 Oahu, Hawaii, 23
Molten zone, 116, 117, 118, 196 Ocean
related to geothermal gradient, 203 formation, 84, 86
Monastery Mine, South Africa Ocean floor
Monomineralic zones, 77 anhydrous basalt, 84
Monte Carlo Test, 41 Oceanic basalt
Moon low K 20 content, 38
pressure at center, 27 metamorphosed, 148
Mount Rainier. Washington, 183, 185 product of multiple meltings, 161
Mount Saint Helens, Washington, 2 Oceanite, 15
Mount Shadwell, Australia, 138 Olivine (see Forsterite)
Mount Vesuvius, Italy, 199 absence at high pressure, 134
Mull, Scotland, 9 coexistence with plagioclase, 25, 27
Multiple melting, 161 composition in meteorites, 28
Mush compressibility, 171
change in composition, 8 constant volume P-T change, 171
crystal-liquid, 4, 8, 12, 55 crystal-liquid REE partition coeffi-
detachment. 4 cients, 155
Mylonization, 168 cumulate, 15, 16, 116, 117
deformation mechanism map, 174
N dominant phase, 134
Nabarro-Herring creep, 174 effect of removal, 206
Nazca Plate, 154 fractionation, 146, 147
Ne-normative magma incompatible vs. complementary, 146
role of carbon dioxide, 132. 148 incongruent melting, 14
Nepheline in derivative magmas, 24
occult, 128 influence on REE abundance, 156
Nepheline basanite, 127, 128, 130 liquidus field extent, 24
Network of liquid, 175, 176 normative, 15
Nodule reaction with liquid, 126, 127, 139, 143,
assemblages. 31 159
equilibrium conditions, 31 seismic velocity, 40, 51
from alkali basalts, 31. 147 source material phase, 24
garnet peridotite. 31. 119 structure at melting point, 87
genetic relationship to basalt. 22, 135 thermal expansion. 171
grospydite, 26 Olivine eclogite
in kimberlite, 31 rarity, 145
metamorphic effects on. 31 Olivine-liquid reaction, 206
metasomatic effects on, 31. 85 at low pressures, 147
rarity in tholeiite, 31 pressure effect, 143, 144
rate of rise. 189 Olivine-melilite nephelinite, 128, 130
record of mantle conditions, 31 Olivine reaction relation, 145
sheared, 168, 169 Olivine tholeiite
Nomenclature tetrahedron depth of separation, 135
peridotite assemblages. 30 pressure-temperature diagram, 135-
Norm 137
calculation. 29 rare-earth element abundances. 152
definition. 29 Order-(!isorder
North Australian volcanics. 106 in melting process. 87

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

258 Subject Index Ort-Phl

Orthopyroxene Partial melting, 116, 117
aluminous, geobarometer, 30 compositional cusp, 63
coexistence with plagioclase, 26, 27 first phase consumed, 110, 115, 206
crystal-liquid REE partition coeffi- heat requirements, 91
cients, 155 relation to low-velocity zone, 8
effect of removal, 206 Partial melt region
incompatible phase, 17 under oceans, 85, 86
inherited, 147 under shields, 85, 86
reaction with liquid, 142, 143, 159, 206 Peridotite
relation toNe-normative magma, 148 composition, 27
Orthopyroxenite compressive strength under confining
thermal conductivity, 66 pressure, 178
O-Shima, Japan density constraint, 41
yield stress of magma, 190 depleted, 34
Overburden squeeze, 186-189 melting, 85
Oxygen isotopes parental material, 64, 86
in meteorites, 25 plagioclase, 17
Oxygen partial pressure related to meteorite, 29
effect on iron enrichment, 14 solidus related to basalt liquidus, 140
influence on magma type, 206 synthetic, 34
trace element depletion, 161
undepleted, 34
zone in mantle, 5
Peridotite gneiss, 37
p
Periodicity of eruption, 89, 191-195
Pacific plate, 195 Hekla, Iceland, 200
Pahoehoe related to production rate, 199
liquidus, 113 related to runaway conditions, 208
Palingenesis, 2 related to subsidiary reservoir, 199
Parana plateau lava, 106 Permeability
Parental magma, 128 effected by hydrofracting, 179
Parental material, 8, 12-43 magma chamber wall, 116
bulk compositions required, 132 of ice, 166
crystalline basalt, 196 of mantle, 4
disaggregation, 112 Petrographic province, 112, 202
key phase, olivine, 24 Phase
metamorphosed, 162 high-pressure, 16
normative olivine, 28 primary, 13, 16
number of assemblages, 131 temperature of appearance, 17
olivine content, 135 Phase-change cusp, 60
partial melting, 15 Phase diagram
peridotite, 86 construction from rock melting, 115
principal requirement, 12 Phase equilibria
products from partial melting, !51 basalt tetrahedron, 123, 125
proportion of phases, ItO Phase separation, 145
role of olivine, 134-148 by mechanical deformation, 168
solidus related to basalt liquidus, 139, Phenocrysts, 154
140 removal during magma rise, 145
Parent-daughter test, 13-16 Phlogopite
Partial melt consumed in initial melts, 38
phase consumed first, 148, 149, 159 heat-producing aspects, 38

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Phi-Pro Subject Index 259

K.O-bearing phase, 37, 81 Plate
liquidus relations, 38, 39 average spreading rate, 188
maximum thermal stability, 82 breakup related to magma generation,
melting, 81, 82 200
persistence during partial melt, 38 subducted, 8, 55, 72, 148
phase in ultrabasic rocks, 83 Plate tectonics
presence in mantle, 202 early recognition, 5
primary, 38 magma generation, 55, 56
range of stability, 81 melting processes, 57
secondary, 38, 85 Precambrian, 48
temperature-composition diagram, 82 processes, 48
water content, 38 Plume, 195
Physicochemical control, 17, 120, 205 Pockets
Picrite, 15 aspect ratio, 49
Picture Gorge, Oregon, 23 cooling rate, 49
Piercing point, 126 of basalt, 5
definition, 95 of eclogite, 5
Di-Fo-Py system, 96 of eclogite in peridotite, 20
Pigeonite, 144 of magma, 7, 50
Pillow basalt, 84 of residual magma, 49
Pitch Point-counting technique, 113
analogous to magma, 173 Poiseuille Flow, 177
Plagioclase Poisson's ratio, 19, 52
coexistence with olivine, 25, 27 Pore
coexistence with orthopyroxene, 26, at grain junctions, 162
27 Pore pressure, 178, 179
occult in pyroxene, 34 effect on rock strength, 179
presence in mantle material, 29 principle, 179
principal phase in basalt, 34 Pore volume
solid solution effect, 14 related to magma aggregation, 182
systems with other minerals, 125 Porosity, 170, 171
Plagioclase peridotite of mantle, 4
eutectic temperature, 48 Potassic basalts, 100
olivine-liquid reaction, 147 Potassium
source material, 19 content in mantle, 81
stability field, 28 source of, 37
Planets Precambrian shield, 85
different compositions, 45 Pressure rise
disruption, 26, 27 in cooling magma, 196, 197
Plastic deformation Primary magma, 50, 146, 159
of lherzolite, 175, 176 Primordial magma
Plastic flow, 170, 177 depth of, 196
Plasticity Principal stress
definition, 173 control of dikes and sills, 180
Plastic layers in mantle, 8 Production rate
Plastic region related to periodicity, 199
around chamber, 173 Propellant
effective viscosity, 175 explosive eruption, 199
Plastic sponge Proportion of phases
of crystals, 4 influence on REE abundances, 154,
Plastic zone 157, 159
in mantle, 180, 182 in parent material, 205

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

260 Subject Index Pro-Run

Protoearth depletion of, II 0
chondritic, 35 measure of degree of melting, 150, 159
cooling,45 minerals contributing to variation of,
homogeneity, 47 158
rate of aggregation, 45 normalization of, 211
Province Reaction point
petrographic, 112 described, 124
Pumice Reaction relation test, 139
kinetic energy release, 183 REE-rare-earth elements
liquidus, 113 Refractory layer, 196
Pyrolite model, 41, 146, 147 Regenerative feedback, 60, 74
Pyrope Region of magma generation, 207
breakdown assemblages, 59, 60 Regional stress
density of crystal, 94 effect on volcano position, 182
density of glass, 94 Relics
enthalpy of melting, 91 in partial melt, 165
hydrous melt curve, 197, 198 Remelting
melting behavior, 59 development of liquid film, 165
related to Fo-CaTs-Si02 system, 147 of cumulates, 15
volume change on melting, 170 Reservoir (see Chamber)
Pyroxene auxiliary, 176
seismic velocity, 40 depletion of, 199
Pyroxene geotherm magma, 50, 131
curvature change, 31, 32 role in periodic volcanism, 199
Pyroxene geothermometer, 30 shallow, 2
Pyroxenite, 102 Residua
eclogite, 20
REE enrichment, 159, 160
Q successive compositions, 108
Q (seismic wave attenuation) Residual magmas, 49
defined, 55 Resistivity
related to partial melting, 55 electrical, 49
structure, 54 magnetic, 49
Quenched liquid (see Glass, Vitreous ma- Rheology
terial, Tachylite), 15 related to metamorphic differentiation,
169
Rhyolite (see Granite)
R hydrous partial melt product, 36
Radioactive elements origin of, 48
content in basic rocks, 98 Richter Scale, 184
partitioning into melt, 100 Rigidity
Radioactive heat low-velocity zone, 51
major contributors, 97-99 Rise of magma
time dependency, 96, 97 produces earthquakes, 182
Radioactiveheating,61,96, 103,196,204 Rock load
Radioactive tracers, 115 related to magma rise, 187
Rare-earth elements Rock nomenclature
abundance calculated as function of flow sheet, 130
degree of melting, 151-153 Rock structure
abundance in basalt, 152 collapse of, 112
average chondrite, 211 Runaway mechanisms, 204

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Run-Spa Subject Index 261

Runaway mechanisms of melting Shear modulus, 74
periodic character, 200 related to density, 55
Runaway process Shear stress, 74
regenerative feedback, 74 Shield geotberm, 32
related to explosives, 74 Siberian plateau lava, 106
thermal conductivity, 66 Silberbacb, Germany, 141
Rupture (see Fracture), 182 SiD, 6
of chamber, 170 orientation control, 180
Rupture of volcano Si01 content
related to tumescence, 199 variation during eruption, 200
Si01 content of magma
s effect on yield stress, 190
St. Paul's Rocks, Atlantic, 23 Snake River, Idaho, 136
Salt dome, 189 Soay, Scotland, 23
Salt Lake Crater, Hawaii, 141 Socorro, New Mexico, 51
Sanidine Soda melilite, 122, 123
enthalpy of melting, 91 Solar abundances, 45
volume change on melting, 170 Solidus, 4, 10, 33, 65, 85, 101, 182
San Juan Mountains, New Mexico, 137 defines magma boundary, 207
Schlieren zone, 75 parental material, 139
Seamount volcanoes, 195 volatile effect on, 81
Seismic velocity Solubility
change of, 49, 51, 203 of silica in gas phase, 84
effect of melting on, 207 stress enhancement, 169, tn
formula for calculation, 19 theory for stress enhancement, 177
garnet, 40 Solubility curve
liquid inclusion influence on, 53, 55 related to melting curve, 177
olivine, 40 Source magmas, 121
pyroxene, 40 Source material (see Parental materials)
related to density, 55 age variation, 42
related to liquid content, 51 amphibolite, 22-24
tacbylite, 19 basaltic composition, 18
Seismic wave compositional differences, 42
amplitude, 51, 53 bomblendite, 22, 23
arrival time residuals, 51 meteorite, 24-29
attenuation, 49, 50, 52, 54, 55, 203 spinel peridotite, 19
damping, 55 tacbylite, 19
dissipative function, 55 types of, 18
energy estimate, 183 Source region
laboratory measurement, 52 distribution of liquid, 176
"screening effect," 50 Hawaiian volcano, 75
velocity reduction, 50, 51, 52 movement of, 195
Seismic wave length related to volcano spacing, 193
magma chamber dimensions, 49 shape, 176, 194
Sensing devices, 2 size, 176
Shear melting South America
basis of concept, 73 seismic attenuation, 54, 55
conversion of mechanical energy into Spacing of volcanoes
beat, 71 measurements, 192, 193
experimental demonstration, 71 related to age, 194
strain rate requirements, 71 related to source region, 176

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

262 Subject Index Spe-The

Specific heat, 71 Subducted plate (see Plate), 60
Spheroidization, 169 Subsidiary reservoir (see Reservoir)
Spinel lherzolite, 32, 33 filling related to productivity. 208
depleted of basaltic components, 34 Superheat, 103, 146
hydrous solidus, 85, 140, 198 Surface tension, 163
melting, 85 Surface wave, 51
normative An content, 34 dispersion, 41
reaction ·products, absence of plagio- Synthetic basalts, 121
clase,33 , 34 Synthetic systems (see System Index)
solidus, 140 related to natural rocks, 105
transformed to garnet peridotite, 33
Spinel peridotite
occurrence, 29 T
source material, 19 Tachylite
stability field, 28 crystallinity , 19
Spreading rate of plate Hawaiian , 19
average at midocean ridge, 188 pressure-temperature relations, 18
Star seismic velocity, 19
condensation analogue, 45 source material, 19
Stoke's law, 188, 189 Tambora, East Indies, 199
Strain rate, 167, 173-175 Tectonic environment, 202
Strength of rock Tectonophysics
effect of melting on, 207 of melting, 162-182
pore pressure effect on, 179 Tensile strength
Stress mantle, 171
compressive, 167 rocks, 172
deviatoric, 167, 171 Tephrites, 128, 130
dilatational, 167 Texture
effect on solubility, 169 achondrite, 45
effect on thermal conductivity, 66 chondrite, 45, 46
effect of volatile on , 175, 176 quench, 15
influence on quartz polymorphism, spinifex, 15
172 Thermal arrest, 92
mean, 167 Thermal conductivity
nonhydrostatic, 19, 167 basalt, 66, 68
principal, 167, 168 diabase, 66
shear, 167 dunite, 66, 68
tidal, 176, 196 gabbro, 66
yield, 190 lherzolite, 67. 68
Stress difference minimum, 66, 67. 68
measurements of, 171 orthopyroxenite, 66
Stress-difference melting, 71 perturbation of, 60, 66, 204
Stressed solids stress effect on. 66
recrystallization of, 177 temperature variation, 66, 67
Stress relief Thermal cycles
compression, 60, 61 , 62 theory of, 195, 196
tension, 60, 61, 62 Thermal divide, 128, 145, 150
Stress-strain nomenclature, 173 Thermal energy
Structural environment amount required for magma genera-
magma relationship to, 55, 57 tion, 88

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

The-Vol Subject Index 263

in nuclear explosion, 184 Transient temperatures, local, 71
released during eruption, 183 Transported magma, 50
stored, 3 Trapped magmas, 49
Thermal equilibrium, 64 Triple point
Thermal expansion, 170 of rock-forming silicate, 44, 45
relationship to gravity, 49 Tumescence, 170, 199, 207
Thermal gradients, 162 related to dilatancy, 182
Thermal maximum (see Thermal divide), Two-pyroxene solvus, 30
127
Thermal regime
of melt region, 7 u
Thermal rise, 61-64 Ultrabasic rocks, 15, 16, 97
Thermal trap, 68 garnet in, 26
Tholeiite
abyssal, 17, 19
depth of origin, 7, 8, 9, 205 v
depth of separation, 135 Vaporus eutectic, 48
effect of H 2 0 on, 205 Vein, 172
in oceanic environment, 120 melt conduit, 162, 166
last eruptive on island, 9 Vent
liquid, 20 explosion, 10
olivine-liquid reaction. 147 Vesiculation, 199
pressure-temperature diagram, 136. Viscosity
137 effective, 175
pressure-temperature relations, 21 estimated from levee height, 191
primitive magma. 209 glass, 19
stability field, hydrous, 21 of suspension of spheres, 112
synthetic assemblages, 127 related to liquidus, 191
viscosity. volatile effect on, 175 related to seismic wave attenuation, 55
Thomson effect, 177 temperature variation, 74
Thrust faulting, 72 volatile effect on, 175
melting mechanism. 60 Viscosity coefficient, 189
Tidal deformation Viscous failure
related to mantle crystallinity. 49 role in magma accumulation, 168
triggers volcanism. 208 Viscous flow
Tidal deformation melting. 71 non-Newtonian, 74
Tidal dissipation. 60, 75 Viscous heating, 71
Trace element Vitreous basalt (see Tachylite), 19
buffered in melt. Ill Vitreous material, 5, 7
definition, 107 Volatile fluxing. 10, 76, 78
depletion. 110 Volatiles
partitioning of, I 10-111 absorption by magma, 199
variation, 112 depletion in mantle, 84, 86
Trace element abundances effect on "invariant" point, 132
interpretation, Ill escape of, 139
Trace element partitioning induce melting, 60
conditions of formation from. Ill in excess. 80, 84. 86
crystal structure restrictions. Ill in excess, definition, 79
Transducer principal. 78
for seismic measurements, 52 related to periodic volcanism, 196-199

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

264 Subject Index Voi-Zon

relative influence on melting. 78

source of, 199. 203 w
Volcanic cone Water (see Volatiles)
subsidiary. 4 activity reduced, 83, 85, 86
Volcanic products content of mantle, 81, 86
morphological expression, v flux from mantle, 84
renewable resource. I free phase. 86
soil formed from. I glacial, 166
Volcanic zones magma saturation, 81
related to earthquake zones. 180. 181 primary magma constituent. 199
Volcanism primordial, 84
number of periods in earth's history. principal volatile, 86
208 saturation, 80
Volcano source. 84
Alaskan. 5-7 state in mantle, 48
conduit, 166 undersaturation. 79
duration of activity. 192. 208 Water deficiency
energy release, 183. 184 definition. 84
explosive, volatile release. 23 Water pressure
Hawaiian. 4, 5, 9, 10 less than total pressure. 79
highest. active and extinct. 187 Websterite, 30
in island arc. 7 Wehrlite, 30
lineation of centers. 195 Wettability, 164. 207
number active at present. 105 Wetting (see Wettability), 163
number in Pacific Basin, 192
y
quiescent period of. 200
regional control on position. 182 Yellowstone National Park. Wyoming,
related to plate motion. 195 50
related to seamounts. 195 Yield stress
sensing devices for. 2 Bingham fluid. 190
shield, 7 influence of Si02 content on. 190
spacing. 176, 192, 193
types. I z
vent, 3 Zone melting. 100, 101, 205
vent, level of magma in. 4 composition vs. depth. 116
Volume concentration of elements, 101
change in melting, 170. 207 process described. 116
change in shearing. 170 zone of enrichment. 118
Volume of lava Zone of enrichment
large occurrences. 105-106 zone melting. 118
related to periodicity, 191. 192 Zone of melting
present emissions, 105 migration of. 100

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

System Index
References only to oxide and mineral end-member systems are included . Experimental
studies on natural or simulated rocks are listed in the Subject Index. Mineral
end-member systems related to the expanded basalt tetrahedron are listed in Tables
7-1 (p. 123), 7-2 (p. 125), and 7-3 (p. 125). Abbreviations are alphabetical and as used
in text.

MINERAL END·MEMBER SYSTEM En-8 20, 198

Fa-Fo, 14, 93
Ab-8 20, 79, 80 Fo-8 2 , 46, 47
Ab-An-Di-Fo, 14 Fo-8 20, 198
AbsoAn50-Fo-Qz-HaO (with 10% Fo-Ks-Qz-8 20, 38, 39
KAISi 30 1 ), 38 Fo-Ne-Qz, 132, 133, 147
Ab-Di-En-Qz. 123, 126-128 Fo-Ne-Qz-C02 , 133
Ak-An-Di-Fo-Ne, 131 Fo-Ne-Qz-8 20, 133
An-Di, 91,92 Ph-8 20, 82
An-Di-Fo, 16, 134 Py-8 20, 198
An-Di-8 20, 199
An-Fo, 25 OXIDE SYSTEMS
CaTs-En, 143
CaTs-Fo-Si02 , 147 Naa0-Ca0-Ala03-Si0a. 131
Cor-Di-En, 143, 144 Naa0-Ca0-Mg0-Aia03-Si0a. 121,
Cpx50Gr50-Fo, 117 122
Di-C02 , 78, 81 K20-Si02-H 20, 196
Di-C02H20, 83 K20-Mg0-AI20 3-Si02 , 131
Di-Fo-Qz, 36 K20-Mg0-AI 20 3-Si02-H 20, 82
Di-Fo-Qz-8 20, 36, 37 Ca0-Mg0-Si02 , 165
Di-Fo-Py,96, 107,134,170 Ca0-Mg0-AI 20 3-Si02 • 31, 33, 131, 145,
Di-8 20, 78, 198 147
Di-Py, 94, 149 Ca0-Fe0-AI20 3-Si02 , 131

265

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Copyright National Academy of Sciences. All rights reserved.

Generation of Basaltic Magma

Copyright National Academy of Sciences. All rights reserved.