You are on page 1of 275

Introduction to

Textures and
Introduction to
Textures and
Second edition
A.J. Barker
Lecturer in Geology
University of Southampton, UK

Stanley Thornes (Publishers) Ltd

© 1998 A.J. Barker

The right of A.J. Barker to be identified as author of this work has been
asserted by him in accordance with the Copyright, Designs and Patents
Act 1988.

All rights reserved. No part of this publication may be reproduced or

transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording or any information storage and
retrieval system, without permission in writing from the publisher or
under licence from the Copyright Licensing Agency Limited. Further
details of such licences (for reprographic reproduction) may be obtained
from the Copyright Licensing Agency Limited, of 90 Tottenham Court
Road, London WIP OLP.

First edition published in 1990 by Chapman & Hall, London

Second edition published in 1998 by:

Stanley Thomes (Publishers) Ltd
Ellenborough House
Wellington Street
United Kingdom

98 99 00 01 02 / 10 9 8 7 6 5 4 3 2

A catalogue record for this book is available from the British Library.

ISBN 978-1-4615-7293-0 ISBN 978-1-4615-7291-6 (eBook)

DOl 10.1007/978-1-4615-7291-6

Typeset in Sabon 10.5/12.5 by Cambrian Typesetters, Frimley, Surrey


Technological advances over recent years, and examines interrelationships between deforma-
a wealth of new research providing refreshing tion and metamorphism. It includes an exten-
new interpretations on many metamorphic sive chapter on strain-related microstructures,
microstructures, encouraged me to embark on and others on the controversial but important
this considerably expanded and fully updated topics of porphyroblast-foliation relation-
second edition. ships, and shear-sense indicators. The penulti-
My aim in producing this book on metamor- mate chapter is on the topic of fluids and
phic textures and microstructures has been to vemmg, and in the final, completely
provide a detailed introduction to the thin- redesigned, chapter, the characteristic reaction
section description and interpretation of meta- textures and microstructures associated with
morphic rocks. Although primarily written for particular environments of metamorphism and
the advanced undergraduate student, it should specific P-T- t trajectories are discussed.
provide a useful first source of reference for This is done with reference to many of the
any geologist dealing with metamorphic rocks. classic areas of metamorphism from around
It is intended that the text should be both well- the world. Appendices giving mineral (and
illustrated and comprehensive, but at the same other) abbreviations (Appendix I), a glossary
time concise and affordable. of terms (Appendix II), and a list of key
The book is comprised of three parts. Part mineral assemblages for the major composi-
A provides an introduction to metamorphism tional groups of rocks at each metamorphic
and metamorphic rocks, and compared to the facies (Appendix III) form an integral part of
first edition has been enlarged to include more the book.
detail on the basic interrelationships between A comprehensive list of references is given at
equilibrium assemblages, mineral chemistry the end of each chapter. Wherever possible, the
and microstructures in the interpretation of intention has been to refer to specific references
metamorphic reactions. Part B introduces the relevant to individual topics covered in the
fundamental textures and microstructures of text, as well as referring to key review papers
metamorphic rocks, with emphasis on the on all the major subjects. It is important to
conditions and processes responsible for their recognise that this book can provide only a
formation. This section includes a chapter on glimpse into the complexities of textural and
fabric development (including the use of microstructural development of metamorphic
SEM), followed by chapters covering topics rocks, and so for a more comprehensive insight
such as crystal nucleation and growth, inclu- into a particular topic the cited references are
sions, intergrowths and retrogression. Part C strongly recommended. My knowledge of


metamorphic rocks has grown considerably tion to this fascinating area of study, and
during the writing of this second edition, and it unlocks a few of the secrets of the rocks that
is my hope that your knowledge and enthusi- you are studying.
asm will be enhanced by reading it. Without
doubt, the interpretation of metamorphic rocks
and their microstructures is a complex topic, Andy Barker
and there are still many unanswered questions. Southampton
I hope that this text provides a useful introduc- March 1997


Acknowledgement is given once again to all for developing and printing the photographs,
those postgraduate and undergraduate students and to Chris Forster, who converted many of
who over the years have contributed to what is the line drawings of the first edition into
now a most intriguing and diverse array of thin computerised images. All new diagrams were
sections. The study of this worldwide collection computerised and/or hand drawn by the
of metamorphic rocks has without doubt author. I should also like to thank those people
broadened my knowledge of their textural and who supplied thin sections, and those authors
microstructural features. It provided the back- who supplied original photographs, or permit-
ground that induced me to write the original ted me to reproduce diagrams from their publi-
edition, and in the seven years since initial cations. Also, thanks to Barbara Cressey for
publication, my enthusiasm has been main- advice and assistance with SEM work.
tained through the continued interest shown by I would also like to thank the many people
students in providing me with new and inter- who gave me feedback on the first edition, or
esting thin sections to look at. have spared some of their time to review
Once again, I would like to acknowledge the chapters of this new edition. The comments
classic works of Harker, Spry and Vernon, and suggestions have all been useful in shap-
whose books greatly stimulated my own interest ing and improving on early drafts of the text.
in metamorphic textures and microstructures. The people I would particularly like to thank
Additionally, I would like to acknowledge the are Kate Brodie, Giles Droop, Scott Johnson,
Metamorphic Studies Group, whose interesting Cees Passchier, Stephen Roberts, Doug
programme of conferences and field meetings Robinson, Joan Soldevila, Peter Treloar,
has, over the years, afforded me excellent discus- Rudolph Trouw and Ron Vernon. Last but by
sion with a truly international list of contribu- no means least, I offer many thanks to Linda,
tors and participants. Without doubt, this has for her continued patience and encourage-
maintained my enthusiasm for the subject and ment, especially during the final stages of
greatly enhanced my understanding of metamor- writing and production.
phic rocks and metamorphic processes.
Special thanks are extended to Barry Marsh AJB


Preface v

Acknowledgements Vll

Part A Introduction to metamorphism and metamorphic rocks 1

1 Environments and processes of metamorphism 3

1.1 Environments of metamorphism 3
1.1.1 Regionally extensive metamorphism 4
1.1.2 Localised metamorphism 6
1.2 The limits of metamorphism 7
1.3 An introduction to chemical processes of metamorphism 8
1.3.1 Equilibrium assemblages and the phase rule 8
1.3.2 The energy of the system 10
1.3.3 Reaction types 11
1.3.4 Reaction rates 14
1.3.5 Diffusion 15
1.3.6 Fluid phase 16
1.4 Physical processes acting during metamorphism 16
1.4.1 Volume changes during reaction 16
1.4.2 Deformation processes on the macro- and micro scale 16
1.4.3 Crystal defects and surface energy 17
1.5 Deformation-metamorphism interrelationships 18
References 18

2 Facies concept and petrogenetic grids 21

2.1 Metamorphic facies, grade and zones 21
2.2 Petrogenetic grids 23
References 25

3 Compositional groups of metamorphic rocks 27

3.1 Pelites 27
3.1.1 Medium-pressure 'Barrovian' metamorphism 28
3.1.2 Low-pressure assemblages 31
3.1.3 High-pressure assemblages 32


3.2 Metacarbonates and calc-silicate rocks 32

3.3 Quartzofeldspathic metasediments 33
3.4 Metabasites 33
3.5 Metamorphosed ultramafic rocks 36
3.6 Meta-granitoids 37
References 37

Part B Introduction to metamorphic textures and microstructures 39

Definition of texture and microstructure 39
Equilibrium and equilibrium assemblages 39
Reference 40

4 Layering, banding and fabric development 41

4.1 Compositional layering 41
4.2 Introduction to stress, strain and fabric development 42
4.3 Classification of planar fabrics in metamorphic rocks 43
4.4 Processes involved in cleavage and schistosity development 45
4.5 Processes involved in formation of layering in gneisses and migmatites 51
4.5.1 The nature and origin of gneissose banding 51
4.5.2 The nature and origin of layered migmatites 53
References 55

5 Crystal nucleation and growth 57

5.1 Nucleation 57
5.2 Growth of crystals 60
5.3 Size of crystals 61
5.4 Absolute growth times 65
5.5 Shape and form of crystals 65
5.6 Twinning 75
5.6.1 Introduction 75
5.6.2 Primary twins 76
5.6.3 Secondary twins 77
5.7 Zoning 79
References 83

6 Mineral inclusions, intergrowths and coronas 85

6.1 Growth of porphyroblasts to enclose residual foreign phases 85
6.2 Exsolution textures 89
6.3 Inclusions representing incomplete replacement 93
6.4 Symplectites 93
6.5 Coronas (of high-grade rocks) 98
References 99

7 Replacement and overgrowth 101

7.1 Retrograde metamorphism 101


7.1.1 Environments of retrograde metamorphism 101

7.1.2 Textural features of retrogression 103
7.1.3 Specific types of retrogression and replacement 105
7.2 Overgrowth textures during prograde metamorphism 110
References 112
Part C Interrelationships between deformation and metamorphism 115
8 Deformed rocks and strain-related microstructures 117
8.1 Deformation mechanisms 117
8.2 Inter- and intracrystalline deformation processes and microstructures 117
8.2.1 Defects 117
8.2.2 Dislocations 119
8.2.3 Creep mechanisms 123
8.2.4 Grain boundaries 125
8.2.5 Recovery 128
8.2.6 Recrystallisation 129
8.2.7 Crystallographic-preferred orientations 131
8.3 Fault and shear zone rocks and their microstructures 132
8.3.1 Deformation of quartzitic and quartzofeldspathic rocks 136
8.3.2 Deformation of mafic rocks 137
8.3.3 Deformation of carbonate rocks 139
8.3.4 Distinguishing between schists and mylonites 141
8.4 The influence of deformation on metamorphic processes 142
8.5 The influence of metamorphism on deformation processes 144
References 145
9 Porphyroblast-foliation relationships 149
9.1 Thin-section 'cut effects' 149
9.2 Porphyroblast growth in relation to foliation development 151
9.2.1 Recognition and interpretation of pre-foliation (pre-tectonic)
crystals 151
9.2.2 Recognition of syntectonic crystals 154
9.2.3 Recognition and interpretation of post-tectonic crystals 160
9.2.4 Complex porphyroblast inclusion trails and multiple growth
stages 160
References 161
10 Shear-sense indicators 163
10.1 Introduction 163
10.2 Vein asymmetry and sense of fold overturning 163
10.3 S-C fabrics, shear bands and mica-fish 165
10.4 Differentiated crenulation cleavages 168
10.5 Spiralled inclusion trails 169
10.6 Mantled porphyroclasts and 'rolling structures' 169
10.7 Strain shadows 171


10.8 Grain-shape fabrics and crystallographic preferred orientations 176

References 177

11 Veins and fluid inclusions 179

11.1 Controls on fluid migration and veining 179
11.2 Initial description and interpretation of veins 179
11.3 The 'crack-seal' mechanism of vein formation 183
11.4 Interpretation of fibrous veins 185
11.4.1 Syntaxial fibre veins 185
11.4.2 Antitaxial fibre veins 185
11.4.3 Composite fibre veins 186
11.4.4 'Stretched' (or 'ataxial') crystal fibre veins 187
11.5 Veins and melt segregations at high metamorphic grades 187
11.6 Fluid inclusions 187
References 194

12 Deciphering polydeformed and polymetamorphosed rocks 197

12.1 Polymetamorphism 197
12.2 Local and regional complications 198
12.3 P-T-t paths 199
12.3.1 Introduction 199
12.3.2 Orogenic metamorphism 203
12.3.3 Orogenic metamorphism with a subsequent thermal overprint 209
12.3.4 Granulite facies P-T-t paths 212
12.3.5 Blueschist facies P-T-t paths 215
12.3.6 Eclogite facies P-T-t paths 218
12.4 Final comments 220
References 221

Appendix I: Abbreviations 227

Mineral abbreviations 227
Additional abbreviations 228

Appendix II: Glossary 229

Appendix III: Key mineral assemblages 247

Zeolite facies 247
Sub-greenschist facies 247
Greenschist facies 248
Epidote-amphibolite facies 248
Amphibolite facies 248


Granulite facies 249

Eclogite facies 249
Blueschist facies 250
Albite-epidote hornfels facies 250
Hornblende hornfels facies 250
Pyroxene hornfels facies 251
Sanidinite facies 251

References for Appendices I-III and the plates 253

Index 255

Plates between 130-131

Part A

Introduction to
and metamorphic
Chapter one

and processes of

The process of metamorphism is one of change, rocks in question originated' (Bates & Jackson,
and within the mineral assemblage and texture 1980).
of a metamorphic rock is a memory of that
change. The transformations are brought about
1.1 Environments of metamorphism
by geological processes from the global plate
tectonic level to the more localised scale. In There are various environments in which meta-
view of this, recovering the memory of the morphism occurs (Fig. 1.1), the most major of
change locked in metamorphic rocks helps to which are linked to processes operating at
constrain geological processes well back into constructive and destructive plate margins, and
the Earth's history. are closely interrelated to igneous activity. Such
Locked within the mineralogies of metamor- metamorphism is thus of regional extent,
phic rocks is much information about changing although, following the approach of Miyashiro
p-T conditions. Locked within the textures of (1973, 1994) and Bucher & Frey (1994),
metamorphic rocks is further information on 'regional metamorphism' in the traditional sense
metamorphic process, and principally interac- linked to mountain-building processes is termed
tion with deformation, that ultimately records orogenic metamorphism. This is to distinguish it
plate tectonic movements. As metamorphic from several other types of metamorphism that,
rocks recrystallise in the solid state, they can, in being linked to plate tectonic processes, are of
favourable circumstances, record a memory of regional extent. There are other types of meta-
events operating over many millions of years. morphism that are of localised extent, linked to
This book concentrates primarily on recovering geological processes of a non-global character.
the memory of metamorphic processes locked The main styles of metamorphism can be
into the textures of these rocks. classified as follows (numbers are cross-refer-
In formal terms, metamorphism can be enced to Fig. 1.1):
defined as "The mineralogical, chemical and
Regionally extensive metamorphism (in relative
structural adjustment of solid rocks to physical
order of importance/abundance)
and chemical conditions which have generally
been imposed at depth below the surface zones (1) orogenic metamorphism (traditionally
of weathering and cementation, and which referred to as regional metamorphism);
differ from the conditions under which the (2) ocean-floor metamorphism;

Environments and processes of metamorphism



FIG. 1.1 Environments of metamorphism: (a) a schematic illustration of oceanic crust subducting beneath continental
crust at a convergent plate margin; (b) a schematic illustration of a mid-oceanic spreading ridge (divergent plate
margin). Different metamorphic environments are numbered as follows: (1) orogenic metamorphism; (2) ocean-floor
metamorphism; (3) subduction zone metamorphism; (4) burial metamorphism; (5) contact metamorphism; (6)
hydrothermal metamorphism; (7) shear-zone metamorphism; (8) shock metamorphism.

(3) subduction-zone metamorphism; 1.1.1 Regionally extensive metamorphism

(4) burial metamorphism.
Localised metamorphism Orogenic metamorphism
(5) contact metamorphism; Major belts of orogenic metamorphism
(6) hydrothermal metamorphism; (traditionally referred to as regional, or
(7) shear-zone metamorphism; dynamothermal, metamorphism) occur on all
(8) shock metamorphism. continents, and extend over distances of

Environments of metamorphism

hundreds to thousands of kilometres. Areas of and crustal thickening, which generates region-
such orogenic metamorphism include major ally elevated P-T conditions. The duration of
tracts of the Caledonian-Appalachian orogenic such orogenic metamorphism is estimated to be
belt, more recent and present-day orogenic of the order of 10-50 Ma (e.g. England &
mountainous zones such as the Alps, Rockies, Thompson, 1984; Bucher & Frey, 1994),
Andes and Himalayas, as well as vast areas although the metamorphic history experienced
within Precambrian cratonic blocks, such as by any given rock in such an environment is
those of Africa, Australia, India, Brazil, often recorded as a punctuated series of short-
Canada and Scandinavia. These places record lived events of less than 10 Ma duration during
the sites of former areas of continental thicken- a protracted history (e.g. Barker, 1994).
ing (i.e. mountain belts), and such orogenic
metamorphic belts account for the vast major- Ocean-floor metamorphism
ity of metamorphic rocks seen at the Earth's At mid-oceanic ridges (Fig. 1.1(b)), ocean-floor
surface. metamorphism is an important process. The
Orogenic metamorphism occurs during combination of high heat flow and sea water
active deformation, over a broad range of pres- percolating into fractured oceanic crust causes
sure (P) and temperature (T) conditions and metamorphism of the primary basalt assem-
variable geothermal gradients. At low P-T in blages. This occurs in the upper parts of the
the middle to upper crust, extensive slate belts oceanic crust, typically in the sheeted dyke
form, whereas in mid-crustal regions metamor- complex and above, but not in the gabbros and
phic belts are dominated by schist, marble, ultramafic rocks. Once metamorphosed, ocean-
amphibolite and quartzite. These metamorphic floor spreading transports such rocks away
rocks have generally experienced polyphase from the spreading ridge to be replaced by new
deformation and metamorphism in response to mafic material generated at the ridge, which is
lateral and vertical motions associated with in turn metamorphosed. This means that
convergent plate tectonic movements. In conse- although the metamorphism occurs as a
quence, these metamorphic rocks retain a localised style of hydrothermal metamorphism,
memory of these forces in the form of strong the continual spreading away from the ridge
planar, linear or combined planar-linear fabric and provision of fresh material to be metamor-
(e.g. cleavage and schistosity; Chapter 4) and phosed means that most of the ocean-floor
are usually extensively folded. At deeper levels crust has been metamorphosed. Rocks that
in the crust in the presence of hydrous fluids, have experienced ocean-floor metamorphism
high-grade gneisses form, and partial melting show little evidence of any foliation (except in
may occur to form rocks of mixed metamor- localised fault/shear zones), but commonly
phic and igneous appearance, termed exhibit extensive veining.
migmatites. In the absence of aqueous fluid,
rocks dominated by anhydrous mineral assem- Subduction-zone metamorphism
blages, such as granulites, develop, and at At convergent margins, the subduction of cold
deepest crustal levels eclogites can form. oceanic crust and overlying sediments against
Although there are areas of orogenic meta- an adjacent plate results in an environment of
morphism that have been interpreted in terms high shear strain and low geothermal gradient,
of high heat flow during extension-related so that rocks record a high-P/low- T imprint. To
crustal thinning (e.g. Weber, 1984; Sandiford preserve such high-P/low- T mineral assem-
& Powell, 1986), most orogenic metamor- blages requires rapid uplift, during which
phism is associated with collisional orogenesis process the rocks are often tectonically

Environments and processes of metamorphism

dismembered. This leaves fragmented areas of induced transformations in the country rock.
high-P/low-T metamorphism possessing a For a comprehensive insight into the subject of
strong tectonic fabric ('blueschists'), within a contact metamorphism, the edited volume of
complexly faulted zone. The circum-Pacific Kerrick (1991) provides an excellent starting
margins (e.g. California and Japan) preserve point.
some of the best examples of subduction-zone
(blueschist facies) metamorphism, but Hydrothermal metamorphism
remnants of blueschist facies metamorphism Ocean-floor metamorphism is a type of
have been identified from a growing number of hydrothermal activity giving rise to regionally
locations worldwide. extensive metamorphism. There are many
other examples of hydrothermal metamor-
Burial metamorphism phism (i.e. the interaction of hot, largely aque-
Burial metamorphism (Coombs, 1961) is the ous, fluids with country rock) of localised
term used to describe incipient metamorphism rather than regional extent. A hydrothermal
developed in thick basinal sequences in the fluid can originate from various sources,
absence of major deformation (Fig. 1.1(a)). including those of igneous or metamorphic
These rocks characteristically lack any form of origins. It is usually transported via fractures or
foliation, show incomplete mineralogical trans- shear zones, and at some distance either near to
formation and preserve many of their original or far from the source, the fluid interacts with
textural features. The type area for burial the host rock to cause hydrothermal alteration
metamorphism is the South Island of New or metamorphism. Many cases of such meta-
Zealand, but other areas displaying this style of morphism are intimately associated with
metamorphism have also been recognised. mineralising fluids and ore deposits.
Shear-zone metamorphism
1.1.2 Localised metamorphism
Faults and shear zones represent localised envi-
ronments of stress-induced dynamic metamor-
Contact (or thermal) metamorphism phism, where textural and mineralogical
The most common example of localised meta- transformations take place in an environment
morphism, that occurring in the immediate of high or very high shear strain and variable
vicinity of an igneous intrusion, is considered strain rates, and over a wide range of P-T
as a separate case and is referred to as contact conditions according to the depth within the
metamorphism. Unlike regional low-P/high-T crust. As described in Chapter 8 (Sections 8.4
examples of orogenic metamorphism, the & 8.5), the deformation processes may
contact metamorphic environment is usually enhance metamorphic transformations, while
one involving very limited synmetamorphic metamorphic processes may assist the deforma-
deformation, such that contact-metamorphosed tion of the rock concerned. Although at high
rocks commonly show little in the way of a crustal levels brittle fracturing dominates, at
foliation, unless the rock already possessed deeper levels where ductile deformation occurs
one, or experienced subsequent deformation. it is crystal-plastic processes that are the most
Concentrically arranged zones of contact meta- significant. Such processes involve grain-size
morphism define a thermal (contact) aureole reduction, and produce rocks with a strong
around the intrusion (Fig. 1.1(a)). These zones, planar and linear fabric, termed mylonites.
with their characteristic mineral assemblages Shear zones are commonly environments in
and textures, are the result of thermally which mineral assemblages formed under high

The limits of metamorphism

P-T conditions start to alter to lower-tempera- higher-grade metamorphic rocks, with their
ture assemblages in response to increased varied and interesting assemblages, provide a,
shear stress in the presence of a fluid. The much more attractive proposition. However,
presence of a fluid (H 2 0 ± CO 2 ) is usually technological advances in recent decades,
essential, since the lower-temperature alter- providing much greater resolution than the
ation assemblages typically comprise hydrous petrographic microscope (e.g. X-ray diffrac-
minerals such as micas, serpentine and chlo- tion, scanning electron microscopy (SEM) and
rite, or carbonate minerals such as calcite, tr~nsmlsslOn electron microscopy (TEM))
dolomite and ankerite. have shown that regular mineralogical trans-
formations do occur at low grade and that
Shock (or impact) metamorphism fruitful studies of low-grade metamorphic
This rare and unusual type of metamorphism rocks can also be made. Understanding of the
('dynamic metamorphism') is caused by mete- mineralogical changes and processes operating
orite impact, and consequently is an extremely in these rocks has advanced considerably in
short-lived event in a localised area. For a few recent years. For further details on low-
microseconds, the area of impact experiences temperature metamorphism, the various chap-
extreme P-T conditions, varying up to 1000 ters in the book edited by Frey (1987) are
kbar/5000°C. At the highest P-T conditions recommended.
the impacted rocks are vaporised, but at At the high-temperature end of the scale
slightly less extreme conditions quartz and there is overlap between the realm of metamor-
feldspar can melt to produce a highly vesicular phic rocks and that of igneous rocks. Again,
glass containing coesite and stishovite, the there is no precise boundary. At depth in the
high-P and extreme high-P polymorphs of Si0 2 Earth's crust, metamorphic rocks of suitable
(Fig. 7.1). Notable cases of impact metamor- composition can commence melting at temper-
phism are recorded at Ries Crater, Germany, atures as low as about 630°C, to produce gran-
and Meteor Crater, Arizona, USA. It is outside itoid lenses and layers interleaved with
the scope of this book to cover this type of high-grade metamorphic rocks. In other cases,
metamorphism in any detail, but for further most notably when an aqueous fluid is absent,
insight reference should be made to the publi- metamorphic rocks will continue to recrys-
cations by Englehardt & Staffler (1968), tallise in the solid state to temperatures of the
Grieve (1987), Bischoff & Staffler (1992) and order of 1000°C. In the upper mantle, solid-
White (1993). state metamorphic processes in ultramafic
rocks can take place at still higher tempera-
1.2 The limits of metamorphism tures. Since granitoid melt can form at temper-
atures as low as 630°C, it is clear that the
The precise limits of metamorphism within the realms of igneous and metamorphic rocks
Earth's crust and upper mantle are not sharply show considerable overlap in the range c.630o-
defined. At the low-temperature end of the c.ll00°C. The essential distinction is that
scale there is a rather blurred transition from metamorphic processes are dominantly solid-
diagenesis to metamorphism, over the range state. Even in migmatites, where there is
100-200°C. In the past, petrologists paid little considerable partial melting, as long as the
attention to the very low grade metamorphic bulk of material remains solid, the rock is
rocks because they are generally fine-grained considered to be metamorphic.
and difficult to examine using standard optical Since hydrothermal processes operate III
mlcroscopy. In contrast, coarser-grained geothermal regions around the world to

Environments and processes of metamorphism

produce metamorphic minerals, the low-pres- rocks, it is of course necessary to introduce

sure limit of crustal metamorphism can be some elementary principles and concepts relat-
considered as being about 1 bar. The upper ing to metamorphism. Abbreviations for miner-
pressure limit of crustal metamorphism is less als used throughout this book are listed in
easily defined, but on the basis of experimental Appendix I, and are primarily based on Kretz
and theoretical considerations it is estimated (1983) and Bucher & Frey (1994).
that certain coesite-bearing rocks indicate
stability at pressures of at least 30 kbar (3.0
1.3.1 Equilibrium assemblages and the
GPa). This indicates that crustal rocks can be
phase rule
buried to mantle depths in excess of 100 km
and subsequently be exhumed. The process A rock can be considered as a chemical system
allowing such metamorphism to develop is comprised of a number of chemical compo-
presently a stimulating area of the subject. nents, and for most rocks these can be consid-
The precise type and style of metamorphism ered in terms of oxide components. Pure
is controlled by a number of variables such as quartzite is a simple system comprised of just
temperature, pressure, whole-rock chemistry, one component (Si0 2), whereas a pelitic schist
fluid chemistry, fluid flux, strain rate and so is a much more complex system comprised of
on. Pressure is a fairly simple function of depth many components (Si0 2, A1 20 3 , K20, Na 20,
in the crust, and for approximately every 3 km FeO, MgO, MnO, Ti02, CaO and H 20). The
depth of burial, the litho static pressure different components are contained within
increases by 1 kbar (= 0.1 GPa). The tempera- various minerals of the system, which are
ture experienced by a given rock within the termed phases. As well as the solid phases
crust throughout its history is a more complex (minerals), there is often a fluid phase present
function of the geothermal gradient and geot- in the intergranular region during metamor-
herm for the given region. The conductive phism. For most metamorphic rocks this fluid
properties of different rocks in a given segment is dominantly a mix of H 20 and CO 2, with
of crust play an important role, and the precise small amounts of CH4 and N2 (high-density
nature of the geotherm can be affected and gas). Based on the components of the rock
disrupted by many factors, including thrusting, (system), and assuming enough time for reac-
uplift, erosion and localised intrusive activity. tions to proceed, the system will develop a
stable mineral (phase) assemblage, according to
the prevailing P-T-X(fluid) conditions. This is
1.3 An introduction to chemical processes the equilibrium assemblage. The equilibrium
of metamorphism assemblage will change as P-T conditions
The aim of this book is primarily the recogni- change. For example, during the heating of a
tion and interpretation of textures and mudrock by contact metamorphism, the clay
microstructural features of metamorphic rocks, minerals of the original assemblage break
rather than providing an in-depth petrology down to form micas such as muscovite, which
text. For a comprehensive treatment of meta- at higher temperatures is itself involved in reac-
morphism and metamorphic petrology, the tions to form andalusite; and at the highest
reader is referred to the texts of Yardley (1989), temperatures andalusite will be succeeded by
Philpotts (1990), Spear (1993), Bucher & Frey sillimanite. These phase changes can all occur
(1994), Kretz (1994) and Miyashiro (1994). without any chemical modification of the
However, in order to understand the textural system (i.e. the system is isochemical).
and microstructural features of metamorphic Therefore, according to the prevailing P- T

Chemical processes of metamorphism

conditions a given rock will have a specific 10

equilibrium assemblage. During prograde
(increasing PIT) metamorphism, assemblages 't
-~ -~N
close to equilibrium are usually maintained, 8 Ky
but during retrograde (decreasing PIT) meta- l1I ~
..Q + + ~
morphism, equilibrium is usually not reached.
~ ~

This is because the decline of P-T conditions ~ 6

Q. it ~ ~~
may be at a faster rate than that at which reac- 11.1
tions can proceed. It is this feature, in fact, that !!! Sil
allows metamorphic rocks to survive uplift to 4
the Earth's surface, where they are metastable.
From field and thin-section observation, and
with the knowledge of the chemistry of a given
rock, it is clear to all metamorphic geologists
that the more chemically complex a system is,
the more phases are present in the assemblage.
For example, a chemically complex rock such 200 300 400 500 600 700 800
as a pelitic schist may have a phase assemblage Temperature (OC)
comprising Qtz + PI + Ms + Bt + Grt + 11m, FIG. 1.2 The AI 20 r Si02-H 2 0 (ASH) system (based
whereas a pure quartzite from the same locality on data of Berman, 1988), showing stability fields for
and equilibrated to the same P-T conditions kaolinite, pyrophyllite and Al 2 SiO s phases (modified
will have an assemblage consisting solely of after Bucher & Frey, 1994). Invariant points are shown
by dots. Note that the mineral abbreviations used in
quartz. The Gibbs Phase Rule allows us to text, this diagram and others are given in Appendix I
evaluate the maximum number of phases that (based on Kretz, 1983; Bucher & Frey, 1994).
can exist in a particular system at equilibrium.
It enables assessment of whether an observed
assemblage considered to be at equilibrium (ASH). The stable assemblages in this system
satisfies the rule. The simplified phase rule can are linking petrographic, theoretical and experi-
be expressed as: mental information. Such diagrams are widely
F = C - P + 2, used by petrologists to assess conditions of
metamorphism, and graphically to depict the
where F is the number of degrees of freedom P- T histories of particular rocks based on
(or variance) of the system, C is the number of textural and mineralogical observation. The
chemical components and P the number of P-T conditions of bounding reactions for key
phases. The '2' in the equation relates to the minerals and assemblages are being increasingly
number of intensive or independent variables. refined based on new experimental work and
For metamorphic rocks the two intensive vari- improved thermodynamic data sets. On such
ables are pressure and temperature. Based on diagrams, an area between reaction lines, in
the above equation, the maximum number of which a specific assemblage is stable, is referred
phases in an equilibrium assemblage is C + 2 to as a divariant field (Fig. 1.2): that is, it has
(i.e. for equilibrium F must be ~ 0). two degrees of freedom. As long as P-T condi-
P- T phase diagrams show the stability fields tions stay within the region defined by the field,
of different mineral assemblages for a certain P and T can change independently of each other
chemical system, and in the case of Fig. 1.2 it is without reactions occurring and a change in
for one consisting only of A1 2 0 3 , Si02 and H 2 0 assemblage taking place. The reaction curves

Environments and processes of metamorphism

(lines) bounding divariant fields are termed Where 1l.Hr is pOSItIve, a reaction consumes
univariant curves (Fig. 1.2). The line separating energy, such that heat energy must be added to
the divariant fields of kyanite stability and silli- the system for the reaction to proceed.
manite stability is one such univariant curve, Prograde dehydration reactions in pelitic
and at points precisely on this curve both kyan- schists are good examples of this type of reac-
ite and sillimanite are in stable coexistence. tion, and are described as endothermic. The
Because univariant curves have an extra phase converse case is that of reactions that give off
that is at equilibrium (compared to divariant heat. These are said to be exothermic, and
fields), in order to satisfy the phase rule they include retrograde hydration reactions.
have one less degree of freedom (i.e. they are Whatever the case, the First Law of
univariant). In this situation, to stay on the Thermodynamics states that for any cyclical
univariant curve, a change in T has to be process, the work produced in the surround-
matched by an appropriate change in P, so ings is equal to the heat removed from the
only one out of T and P can change indepen- surroundings (i.e. conservation of energy in the
dently. The position at which univariant curves system). The second property, entropy (symbol
intersect is known as an invariant point (Fig. S, units J mol-1 K-l), is a term representing the
1.2). At such a position another phase is added degree of randomness in a system, and reflects
to the equilibrium assemblage and thus the the thermal energy of the system. As well as
assemblage has no degrees of freedom, and vanattons in entropy between individual
defines a point in P-T space. Changing either mineral phases, it is worth noting that for any
of the independent variables, T and P, would given material (e.g. H 2 0), the vapour phase has
change conditions off the invariant point and higher entropy than the liquid (water) phase,
the assemblage would no longer be at equilib- and the liquid phase a higher entropy than the
num. solid (ice) phase.
The total energy of the rock system, referred
1.3.2 The energy of the system to as the Gibbs free energy (G), is the sum total
of the energy contributed by each phase in the
Rocks, like all materials, have a certain energy system. The Gibbs free energy is defined as
content, referred to as the free energy of the
system, linked to features such as atomic vibra- G==E + PV- TS.
tion and bonding. The energy (units = Joules) Since
can be considered in various forms (thermal,
chemical, mechanical and electrical). Of these, H =E + PV,
chemical and thermal energy are of greatest then
interest to the petrologist. There are two thermo-
dynamic properties that are of fundamental
importance when considering reactions. The Because the Gibbs free energy varies according
first of these, enthalpy (symbol H, units J to the amount of a given material, it is stan-
mol-1 ), reflects the heat content of a phase or dard practice to normalise and speak in terms
system (H == E + PV, where E is the internal of molar Gibbs free energy. Values of molar
energy - a combination of heat and work - P is Gibbs free energy at the standard state (T =
the pressure and V the volume). Of greatest 298.15K and P = 105 Pa (1 bar)) have been
interest to metamorphic petrologists is the determined for most common rock-forming
change in enthalpy on reaction, i.e. the minerals, and are tabulated in publications
enthalpy of reaction (1l.HJ = Hproducts - Hreactants' such as those of Berman (1988) and Holland &

Chemical processes of metamorphism

Powell (1990). Where minerals form simple free energy. The reaction with the lowest acti-
two-component solid-solution series, such as in vation energy for nucleation will be that which
the case of plagioclase feldspar (from the sodic occurs most rapidly.
end-member, albite, to the calcic end-member, At constant P, T, the enthalpy change (ilH)
anorthite), the concept of chemical potential associated with a reaction is comprised of a
has to be considered. The chemical potential thermal energy component due to entropy
(~) of a component i in solution (units J mol-I) change (TilS) plus the change in free energy
is defined as the rate of increase of G of a solu- (ilG). As a reaction proceeds, the net entropy
tion of constant composition, at constant P, T, of the system always increases (Second Law of
when 1 mole of component i is added. More Thermodynamics) and ilG diminishes towards
complicated solid solutions exist for minerals zero. A fundamental concept is that at equilib-
with several end-member components. Such rium ilG = 0, and the relationship between
minerals include the Ca-Mg-Fe pyroxenes entropy and enthalpy is given as TilS =ilH (see
(end-members diopside CaMgSi20 6 , hedenber- Philpotts, 1990, for further details). From this
gite CaFeSi2 0 6 , enstatite Mg 2Si2 0 6 and and the preceding discussion, it is clear that
ferrosilite Fe2Si20 6 ) and the Ca-amphiboles knowledge of some of the fundamental aspects
(end-members tremolite, hornblende (sensu of thermodynamics is crucial to understanding
stricto), edenite, pargasite and tschermakite). In what drives metamorphic reactions, and conse-
such cases, where a number of end-member quently the petrographic features of metamor-
components (j components) are involved, the phic rocks.
molar Gibbs free energy (of a given phase) is
defined as
1.3.3 Reaction types
Prograde metamorphism gives rise to extensive
G p= CL~jnj)/np, (1.1) solid-state recrystallization accompanied by
i =1 various metamorphic reactions, as some miner-
where ~j is the chemical potential of the ith als become unstable and break down to form
component, is the number of moles of the ith new minerals in equilibrium with the prevailing
component and is the number of moles of the P-T conditions. In order to keep the energy of
phase. Thus for a system comprising m phases, the system at a minimum for a given set of
the Gibbs free energy can be expressed as conditions, reactions will always proceed by
the lowest energy route.
Gs = L
Gp,knp,k' (1.2)
Most reactions are of a univariant or discon-
tinuous character, and define a unique univari-
k=l ant curve in P-T space (which has one degree
where Gp,k is the molar Gibbs free energy of of freedom). For example, in the ASH system
the kth phase and np,k is the number of moles of Fig. 1.2, if we consider the case of progres-
of this phase. As P-T conditions change, so the sive increase in T, while maintaining P constant
values of Gibbs free energy for individual at 6 kbar, then the stable assemblage will shift
phases change, and thus G5 changes. For equi- from KIn + Qtz to Prl + Qtz, to Ky + Qtz and
librium, G5 must be maintained at the lowest ultimately to Sil + Qtz. Each of these changes
possible value for the P-T experienced. This is involves breakdown of one aluminous phase
achieved via reactions, where phases with and formation of another, at a precise point in
lower Gibbs free energy are formed at the P-T space defined by the various univariant
expense of unstable phases with higher Gibbs reaction curves. Other reactions are divariant

Environments and processes of metamorphism

(also known as continuous or sliding reac- where a single reactant produces a single
tions). In such cases, the reactants and products product that is chemically identical, but has
coexist over a range of P-T (divariant field) different crystallographic arrangement (e.g.
and the reaction proceeds by varying the Ky ~ Sil, Cal ~ Arg); and (ii) those reac-
composition and modal amounts of coexisting tions involving various chemically distinct
phases, until eventually one of the reactants is reactants and products (e.g. Ab ~ Jd +
exhausted. This is typical for reactions involv- Qtz). At equilibrium, solid-solid reactions are
ing phases that are solid solutions, and can independent of the fluid phase and, because of
vary their compositions by cation exchange this, are potentially useful indicators of P-T
with other phases in the system. The break- conditions. However, because tlG is often
down of chlorite and formation of garnet, by small (e.g. Ky ~ Sil), it means that the
the reaction ChI + Ms ~ Grt + Bt + H 2 0 is a kinetics of reaction are slow and metastable
good example. persistence of the reactant phase is common-
Metamorphic reactions can be classified into place. Thus some rocks may contain two
six main types: Al 2SiO s phases, seemingly at equilibrium, but
it is not safe to assume that they record peak
(i) solid-solid reactions (reactions involving
conditions of the appropriate univariant
solid phases only);
curve, since conditions significantly beyond
(ii) dehydration reactions (reactions that liber-
the curve may have been attained.
ate H 2 0);
Univariant curves have slopes defined by the
(iii) decarbonation reactions (reactions that
Clausius-Clapeyron equation:
liberate CO 2 );
(iv) oxidation-reduction reactions (reactions dP/dT = tlHITtl V = tlSltl V, (1.3)
that change the valence state of Fe-oxide
where tlH, tlS and tl V are changes in enthalpy,
entropy and volume on reaction. Therefore, the
(v) cation-exchange reactions (e.g. Fe-Mg
slope can be expressed in terms of change in
exchange between coexisting ferro-
entropy and change in volume. Since the up-
magnesian phases);
temperature side of reactions has higher
(vi) ionic reactions (reactions that are balanced
entropy and usually increased volume, most
by inferring involvement of ionic species
univariant curves have a positive slope.
derived from the fluid phase).
Types (i)-(iv) are single reactions involving the Dehydration reactions
conversion of a mineral or set of minerals to There are many hydrous mineral phases, and
another mineral or set of minerals (net-transfer consequently many prograde metamorphic
reactions), the fifth type (exchange reactions) reactions are dehydration reactions. Some
involve exchange of atoms between two or dehydration reactions are univariant reactions
more coexisting phases without producing new (e.g. Prl ~ And + 3Qtz + H 2 0), whereas
minerals, and the final type (ionic reactions) others are divariant (e.g. ChI + Ms ~ Grt +
involves ionic exchange between several differ- Bt + H 2 0). Most common dehydration reac-
ent reaction sites within the system in order to tions recorded by metamorphic rocks have
facilitate a particular transformation. positive tlS and tl V, and therefore have positive
slopes. At pressures exceeding a few kilo bars,
Solid-solid reactions most dehydration reactions are very steep;
Solid-solid reactions can be subdivided into: (i) indeed, they are often near to isothermal due to
those reactions that are phase transformations, very large tlS and small (but positive) tl V (see

Chemical processes of metamorphism

the reaction Prl ~ Al2SiO s + Qtz + H 20 in 3.0

Fig.1.2}. Dehydration reactions have much ..,
~ d
larger !l.S compared to solid-solid reactions, I
and consequently are always much steeper. At ~
pressures exceeding 10 kbar, H 2 0 becomes ~
highly compressed, such that dehydration reac- °CII
:I: 2.0
tions commonly give rise to negative !l.V and +CII
thus the slope of the univariant curve becomes 8 Cal Wo

negative. I
+ +
0.. Qtz CO2
Decarbonation reactions 1.0
This type of reaction (e.g. Cal + Qtz .;;==: Wo +
CO 2 ) involves the breakdown of carbonate
minerals and liberation of CO2, Like dehydra-
tion reactions, the univariant curves of de-
carbonation reactions over the normal range of
P-T conditions usually have steep, positive 300 500 700 900
slopes. The position of such a reaction in P-T T( °C)
space is strongly affected by the composition of FIG. 1.3 A fluid pressure (Pf) - temperature (T) -
the fluid phase (Fig. 1.3). In the wollastonite- composition (X) diagram, showing how the reaction
forming reaction illustrated, higher XC0 2 is a Cal + Qtz ~ Wo + CO2 is strongly influenced by the
consequence of the reaction and thus in turn fluid composition of the system. XC0 2 = 1.00 means
inhibits the reaction, such that higher tempera- 100% CO 2 fluid composition, whereas XeO? = 0
means 100% H 2 0 fluid composition (modified after
tures are required for the reaction to proceed to Greenwood, 1967).

Oxidation-reduction reactions most significant in relation to Fe-Ti phases. In

Iron within stable Fe minerals can occur in two metamorphic systems the availability of oxygen
different oxidation states, namely ferrous (Fe2+) is usually discussed in terms of oxygen fugacity,
or ferric (Fe3+). Ferrous iron (Fe 2+) shares two (f02)' or else the partial pressure of oxygen,
electrons in forming bonds with other atoms (P0 2 ). Oxygen fugacity is analogous to activity
(i.e. it is divalent), and silicate phases rich in in describing the effective concentration of a
ferrous ions are commonly green (e.g. olivine given component in a non-ideal solution such as
and hornblende). Ferric iron (Fe3+) is trivalent - a melt or a metamorphic fluid. Since magnetite
that is, it has three of its electrons bonded - (Fe30 4 ) and ilmenite (FeTi0 3 ) tend to be the
and minerals rich in ferric iron are commonly stable Fe-Ti oxides in most metamorphic rocks,
red or orange (e.g. haematite). whereas primary haematite is usually absent,
Reactions that involve a change in valency of reference to Fig. 1.4 indicates that reducing
Fe-Ti phases in the system are known as oxida- conditions are predicted under typical meta-
tion-reduction reactions. If a reaction causes morphic temperature ranges. In certain environ-
iron to increase its valency from Fe2+ to Fe3+, it ments of low-grade metamorphism oxidising
is termed an oxidation reaction, whereas in the conditions may prevail, but at deeper levels in
reverse case, reduction, the valency decreases. the crust (higher grades of metamorphism),
Many metamorphic minerals are susceptible to conditions tend towards moderate or strongly
redox (reduction-oxidation) reactions, but it is reducing.

Environments and processes of metamorphism

o Ionic reactions
Oxidising In natural rock systems, the process of metamor-
(no Fe 2+) phic reaction is complex, and may consist of one
-5 Fe3++ Fe 2+ or more of the reaction types operating simulta-
neously and involving many if not all phases in
N the system, as the rock attempts to minimise the
Q -10 ~0~xtfo
free energy of the system and maintain equilib-
CI> ~'ttCb rium. A classic example of this is the seemingly
~~ simple reaction concerning the breakdown of
-15 x 011-
~'ttx kyanite and formation of sillimanite, each with
the composition AI2Si0s- It might be expected
Reducing that with increasing temperature (Fig. 1.2), silli-
-20 (no Fe3+)
manite would form by atom-for-atom replace-
ment of kyanite. In reality, the evidence seen in
300 500 700 900 thin section does not support this, since the
Temperature ( °c ) normal observation at the sillimanite isograd is
for the sillimanite to have nucleated and grown
FIG. 1.4 The stability of different phases in the in an aggregate of Bt + Qtz crystals. Carmichael
Fe-Si-0 2 system at different values of oxygen fugacity
(1969) explained such observations by proposing
((OJ) over the temperature range 300-1000°C (modi-
fied after Gill, 1996). The top curve is generally that three intimately linked reactions take place
referred to as the haematite-magnetite (HM) buffering (Fig. 1.5) at different sites within the rock
reaction, and the lower curve as the quartz- matrix. He proposed that aluminium remains
fayalite-magnetite (QFM) buffer. largely immobile, but that other components are
transferred between reaction sites via an inter-
granular fluid. For further discussion of the
complexity of this sillimanite-forming reaction,
Cation-exchange reactions
see Foster (1991). The seemingly straightforward
This type of reaction involves ionic substitu- retrograde pseudomorphing of an andalusite or
tion between two or more phases in the kyanite porphyroblast by a fine-grained aggre-
system. Usually the cations are of the same or gate of white mica (sericite) is another example
similar charge, and possessing similar ionic of a reaction in which components not present in
radius. In metamorphic rocks partitioning of the reactant(s) (i.e. K+ and H 20) must be derived
Fel. and Mgl. between ferromagnesian miner- from elsewhere in the system to make the prod-
als such as garnet and biotite is a common uct(s), or alternatively an external source for the
example of such a reaction. Such changes components might be postulated, as in the case
occur in order to minimise the energy of the of metasomatic reactions.
system and maintain equilibrium during retro-
grade processes, but they do not involve the
breakdown or growth of minerals. Such reac-
1.3.4 Reaction rates
tions can be predicted based on theoretical/ The various reactions described above will
thermodynamic considerations, and while they give rise to different microstructural and
cannot be identified petrographically (except textural changes within a metamorphic rock,
perhaps with SEM back-scattered electron and these changes will be considered in
imaging), they can be determined by electron subsequent chapters. The precise reaction
microprobe analysis. sequence that occurs in a rock is a function of

Chemical processes of metamorphism

chemical potential gradient. In solids this

means the periodic jumping of atoms from one
site in the structure to another. If effective, the
process of diffusion will cause homogenisation
of zoned solid-solution minerals such as garnet
in an approach towards equilibrium (see
Chapter 5 for further details). The rate of diffu-
+Ab sion is measured as the flux (J) of atoms across
a unit area down a concentration gradient over
unit time. It is given by
dC (1.4)
FIG. 1.5 A schematic illustration of a Carmichael-type which is known as Fick's first law of diffusion,
ionic exchange reaction; in this case illustrating the
complexity of the kyanite ~ sillimanite transforma- where D is the diffusion coefficient (units
tion (based on Carmichael, 1969). The upper half of m2 s-l) of the diffusing material, C is the
the diagram shows a simplified representation of the concentration of the diffusing material and x is
interrelationships of the various phases in the assem- the direction of diffusion. Since the concentra-
blage, and the lower diagram summarises the various
aspects of the ionic exchange reaction. In the upper tion of diffusing material will vary with time, a
diagram, aligned rectangles and small 'circles' represent more useful equation is Fick's second law of
the Bt + Qtz + Ms + Ab matrix; the large stippled diffusion, given as
rectangle represents the pseudomorphic replacement of
kyanite by muscovite, with a cross-hatched remnant of dC = ~D dC . (1.5)
kyanite in the core; the large circle with dots around dt dx dx
the perimeter represents a garnet porphyroblast; and
the mesh of wispy lines represents a sillimanite (fibro- For most minerals over metamorphic tempera-
lite) aggregate mixed with biotite. tures, the diffusion of atoms within a crystal
lattice is an extremely slow process, with
bulk rock chemistry, fluid chemistry and the values of D generally less than 10-18 m2 s-1 (see
relative rates at which reactions occur. For a Freer, 1981, for data on diffusion in silicate
given prograde reaction, once the product or minerals). As well as depending on D and
products have nucleated, their growth is time, the process of diffusion within and
controlled by either diffusion or interface kinet- between materials is also strongly temperature
ics. The parameter with the slower rate is rate- dependent. The higher the temperature, the
controlling with respect to the reaction more rapid diffusion in silicates becomes, such
progress. As P-T conditions decline in associa- that at 300°C diffusion is fairly insignificant in
tion with uplift and erosion, the high-grade metamorphic processes, but from 650°C
minerals move outside their stability fields. If a upwards diffusion becomes increasingly
fluid is present, extensive retrograde metamor- important. Diffusion through the matrix of the
phism usually occurs (Section 7.1). rock via an intergranular fluid, while still a
slow process, is many orders of magnitude
faster than diffusion through the crystal
1.3.5 Diffusion lattice. Ildefonse & Gabis (1976) determined
Diffusion is the process by which atoms, mole- that the value of D for Si0 2 in an aqueous
cules or ions move from one position to another fluid at 550°C and 1 kbar (0.1 GPa) was on
within a fluid or solid, under the influence of a the order of 10-8 m2 S-I.

Environments and processes of metamorphism

1.3.6 Fluid phase from the likely precursor sedimentary and

igneous rocks (e.g. amphibolites generally
It is only in recent decades that petrologists have a basaltic or basaltic andesite chemistry).
have properly appreciated that it is not just
the mineral being replaced and the obvious
products of a texturally observed reaction that 1.4 Physical processes acting during
are actually involved in the reaction. More metamorphism
realistically, the final texture observed is the
end result of a complex series of replacement 1.4.1 Volume changes during reaction
and ion exchange reactions involving many
phases in the system, and especially involving In Section 1.3, the chemical processes involved
a grain-boundary fluid. This fluid is an effi- in metamorphic reactions were introduced, but
cient means of transfer of ions within a rock reactions also involve physical processes and
by diffusive processes. The presence of a fluid physical change. Since minerals have different
is essential for most metamorphic reactions to densities and many reactions involve liberation
proceed, and its chemistry has a strong of a vapour phase, it is unlikely that volume
control over precisely which reactions occur. will remain constant after a reaction. In many
All things considered, in most cases it is fluid cases (e.g. reactions involving solids only), the
and temperature that have the greatest influ- change in volume (L1 V) is usually small, and
ence on most reactions and on reaction rates. may be considered negligible, but in other cases
The precise process of movement of 'volatile' L1 V may be substantial. For example, most
and other ionic components in metamorphic prograde devolatilisation reactions show large
rocks is something that still requires further positive values for L1 V. If reaction rates are fast,
study. Modification of chemical zonation rapid increases (or decreases) in volume can
profiles in porphyroblasts such as garnet gives induce extensive fracturing and modification of
evidence for solid-state diffusion within meta- the pre-existing microstructure, especially in
morphism, while the nature of pressure-solu- low-pressure regimes. For example, the reac-
tion cleavage gives ample evidence for tion Cal + Qtz -7 Wo + CO 2 involves a volume
diffusive mass transfer via the fluid phase. In decrease of up to 35%, and serpentinisation of
the VICInIty of major intrusions and peridotite by breakdown of olivine may involve
hydrothermal systems there are often exten- 35-45% volume increase.
sive metasomatic changes in associated meta-
morphic rocks, while the focusing of fluids
1.4.2 Deformation processes on the
into fault and thrust zones gives rise to exten-
macro- and microscale
sive retrogression. Outside these environments
metamorphic petrologists often prefer, for At low P-T conditions of high crustal levels,
simplicity, to consider most rocks as essen- rocks behave in a largely brittle manner, involv-
tially isochemical ('closed') systems. Although ing fracturing on all scales (cataclasis). In
it is clear from the presence of veins that a contrast, under moderate and high P-T condi-
certain degree of element migration (mass tions the behaviour is essentially ductile,
transfer) must occur, in most cases this does involving power-law creep processes, such as
not appreciably modify the bulk rock chem- dislocation creep (Chapter 8). This is consid-
istry. Bulk rock analyses of metamorphic ered to be the dominant mechanism for most
rocks suggest that in the majority of cases metamorphic situations in the crust and upper
their chemistry is not significantly different mantle. At high temperatures and low strain

Physical processes acting during metamorphism

rates in the upper amphibolite facies, and 1.4.3 Crystal defects and surface energy
higher grades of metamorphism, rearrangement
of grains by diffusive transfer of atoms (diffu- When considering chemical processes in rela-
sional creep) can be a significant process facili- tion to mineral assemblages and reactions, a
tating ductile deformation. basic assumption in the analysis presented was
Somewhere between the two end-member that the minerals of each phase were perfect
situations of brittle and ductile deformation crystals. In the case of metamorphic rocks, it is
there is a macroscopic 'brittle-ductile' transi- clear from petrographic observation using stan-
tion, where rocks display features of both brit- dard microscopy, and especially from SEMI
tle and ductile conditions (Murrell, 1990). TEM studies, that metamorphic minerals are
This change in behaviour varies considerably essentially imperfect crystals, and that their
between rock types, and according to particu- boundaries with each other are imperfect. To
lar P, T and strain rate. The variation between gain a more complete understanding of the
rock types is due to the fact that the processes involved in the microstructural devel-
constituent minerals of each rock type have opment of metamorphic rocks, it is therefore
different mechanical properties. This means essential to have an understanding of defects in
that, for a given sequence of rocks, some rocks crystals and the energy of crystal surfaces.
may be experiencing broadly brittle deforma- Atoms at the crystal edge do not have all
tion at a time at which others are experiencing their bonds satisfied, and consequently the
broadly ductile deformation. On the scale of surface of the crystal is less stable, and has
an individual rock, some minerals will display excess energy, termed surface energy (or inter-
features of brittle deformation, whereas others facial energy). The contribution of surface
are undergoing ductile deformation. A good energy to the overall free energy of the system
example of this is the case of granitic may seem small, but in rocks of fine grain size
mylonites. At low metamorphic grades the (e.g. mylonites), where surface area is obvi-
feldspars form porphyroclasts displaying brit- ously greater, the contribution becomes signifi-
tle deformation, while quartz in the same cant, and consequently fine-grained aggregates
assemblage experiences extensive grain-size are generally more reactive. In all metamorphic
reduction, and overall ductile deformation. At rocks, the interfacial energy drives grain
higher temperatures, both quartz and feldspar growth (coarsening) and causes modification of
experience ductile deformation (Fig. 8.3). grain-boundary relationships and grain shape,
Other minerals such as hornblende also expe- in an approach towards the most stable config-
rience brittle deformation at low metamorphic uration of grains. This is achieved by minimis-
grades, but behave in a ductile manner at high ing the contribution of grain-boundary
metamorphic grades. Even the same minerals (interfacial) energy to the total free energy of
in the same rock may show major differences the system. The process of coarsening is most
in mechanical behaviour, since the heterogene- conspicuous in monomineralic rocks, but is
ity of rock deformation will lead to strain also apparent in many polymineralic assem-
partitioning on all scales, so that not only will blages.
high strain zones develop on a macro scale, but As well as surface irregularities and imper-
on the microscale there will be strain and fections causing increased free energy, natural
strain rate variations. This will produce differ- crystals also have internal imperfections, or
ent textural and microstructural features in defects. These defects have the combined effect
both monomineralic and polycrystalline of increasing the energy of the crystal, and the
assemblages. higher the number of defects, the greater is the

Environments and processes of metamorphism

instability of the individual grain. A three-fold over the years. Much of the discussion has
classification of crystal defects can be made, focused on inclusion trails of syntectonic
namely: (i) point defects (vacancies); (ii) line garnets, and the interpretation of whether or
defects (dislocations); and (iii) surface (grain not the porphyroblasts have rotated (Chapter
boundary) defects. Chapter 8 examines these in 9). The controls on porphyroblast nucleation
more detail, and introduces some of the charac- and growth within a metamorphic rock are
teristic deformation-related textural and many, and encompass a wide range of variables
microstructural features of deformed metamor- such as P-T conditions, fluid chemistry, bulk
phic rocks, and the processes responsible for rock chemistry, strain rate gradients and chemi-
such features. However, for more detailed cal potential gradients. In shear zones, bulk
information, the reader is referred to publica- strain and strain rates have an important
tions such as Poirier (1985) and the various control over the processes that operate and the
papers contained within Barber & Meredith structures observed. Increased dislocations in
(1990) and Boland & Fitzgerald (1993). crystals provide more sites to which fluids will
be attracted and at which reactions may occur.
Grain boundaries are shown to be crucial as
1.5 Deformation-metamorphism
sites of metamorphic reactions, as well as
controlling most textural and microstructural
The interrelationships between deformation changes. Despite this, there is still much to be
and metamorphism are fundamental to the learnt about the precise nature of grain-bound-
overall mineralogical and microstructural ary conditions, and the processes that operate.
evolution of a given rock. Deformation
processes accompanying metamorphism vary
as a function of the prevailing temperature, the
confining pressure, the strain rate and lithologi- Barber, D.J. & Meredith, P.G. (eds) (1990)
Deformation processes in minerals, ceramics and
cal factors (e.g. mineralogy, grain size, porosity rocks. Mineralogical Society Monograph No.1,
and permeability). All of these variables affect Unwin Hyman, London, 423 pp.
the rheological behaviour of a given rock. At, Barker, A.J. (1994) Interpretation of porphyroblast
high crustal levels (i.e. low P-T) and moderate inclusion fabrics: limitations imposed by growth
kinetics and strain rates. Journal of Metamorphic
strain rates, grain-boundary sliding and diffu- Geology, 12, 681-694.
sive mass transfer are the dominant processes, Bates, R.L. & Jackson, J.A. (1980) Glossary of
while at high strain rates disaggregation and geology. American Geological Institute, Falls
brecciation occur. At middle to low crustal Church, Virginia, 751 pp.
Berman, R.G. (1988) Internally-consistent thermo-
levels the significant increase in temperature dynamic data for stoichiometric minerals in
means that rocks deform very differently. At the system Na1 0-K1 0-CaO-MgO-FeO-Fe20r
these depths crystal-plastic processes associ- AI1 0 3-SiO r Ti0 2-H zO-CO z' Journal of Petrology,
ated with ductile deformation are most impor- 29,445-522.
Bischoff, A. & Stoffler, D. (1992) Shock metamor-
tant, with dislocation creep being the dominant phism as a fundamental process in the evolution of
mechanism in operation. planetary bodies: information from meteorites.
The relationship between porphyroblastesis European Journal of Mineralogy, 4, 707-755.
and deformation is an area of special interest, Boland, J.N. & FitzGerald, J.D. (eds) (1993) Defects
and processes in the solid state: geoscience appli-
since it can often provide the key to interpret- cations. Developments in Petrology No. 14 ('The
ing the crustal evolution of a region. The McLaren Volume'), Elsevier, Amsterdam, 470 pp.
interpretation of porphyroblast-foliation rela- Bucher, K. & Frey, M. (1994) Petrogenesis of meta-
tionships has proved to be a controversial topic morphic rocks. Springer-Verlag, Berlin, 318 pp.


Carmichael, D.M. (1969) On the mechanism of Reviews in Mineralogy No. 26, Mineralogical
prograde metamorphic reactions in quartz-bearing Society of America, Washington, DC, 847 pp.
pelitic rocks. Contributions to Mineralogy and Kretz, R. (1983) Symbols for rock-forming minerals.
Petrology, 20, 244-267. American Mineralogist, 68, 277-279.
Coombs, D.S. (1961) Some recent work on the lower Kretz, R. (1994) Metamorphic crystallization. John
grade metamorphism. Australian Journal of Science, Wiley, Chichester, 507 pp.
24,203-215. Miyashiro, A. (1973) Metamorphism and metamorphic
Engelhardt, W.V. & Stoffler, D. (1968) Stages of shock belts. George Allen & Unwin, London, 492 pp.
metamorphism in crystalline rocks of the Ries Basin, Miyashiro, A. (1994) Metamorphic petrology. UCL
Germany, in Shock metamorphism of natural mate- Press, London, 404 pp.
rials (eds B.M. French & N.M. Short). Mono Book Murrell, S.A.F. (1990) Brittle-to-ductile transitions in
Corporation, Baltimore. polycrystalline non-metallic materials, in Deforma-
England, P.C & Thompson, A.B. (1984) tion processes in minerals, ceramics and rocks, (eds
Pressure-temperature-time paths of regional meta- D.J. Barber & P.G. Meredith). Mineralogical
morphism I. Heat transfer during the evolution of Society Monograph No.1, Unwin Hyman, London,
regions of thickened continental crust. Journal of Ch. 5, 109-137.
Petrology, 25, 894-928. Philpotts, A.R. (1990) Principles of igneous and meta-
Foster, CT. (1991) The role of biotite as a catalyst in morphic petrology. Prentice Hall, Englewood Cliffs,
reaction mechanisms that form sillamanite. New Jersey, 498 pp.
Canadian Mineralogist, 29, 943-963. Poirier, J.-P. (1985) Creep of crystals: high-temperature
Freer, R. (1981) Diffusion in silicate minerals and deformation processes in metals, ceramics and
glasses: a data digest and guide to the literature. minerals. Cambridge University Press, Cambridge,
Contributions to Mineralogy and Petrology, 76, 260 pp.
440-454. Sandiford, M. & Powell, R. (1986) Deep crustal meta-
Frey, M. (ed.) (1987) Low temperature metamorphism. morphism during continental extension: modern
Blackie, Glasgow, 351 pp. and ancient examples. Earth and Planetary Science
Gill, R. (1996) Chemical fundamentals of geology, 2nd Letters, 79, 151-158.
edn. Chapman & Hall, London, 320 pp. Spear, F.S. (1993) Metamorphic phase equilibria and
Greenwood, H.J. (1967) Wollastonite: stability in pressure-temperature-time paths. Mineralogical
H 20-C02 mixtures and occurrences in a contact Society of America Monograph, Washington, DC,
metamorphic aureole near Almo, British 799 pp.
Columbia, Canada. American Mineralogist, 52, Weber, K. (1984) Variscan events: early Paleozoic
1669-1680. continental rift metamorphism and late Paleozoic
Grieve, R.A.F. (1987) Terrestrial impact structures. crustal shortening, in Variscan tectonics of the
Annual Review in Earth and Planetary Science, 15, North Atlantic region (eds D.H.W. Hutton & D.J.
245-270. Sanderson). Geological Society of London Special
Holland, T.J.B. & Powell, R. (1990) An enlarged and Publication No. 14, Blackwell Scientific, Oxford,
updated internally consistent thermodynamic dataset 3-22.
with uncertainties and correlations: the system White, J.C (1993) Shock-induced melting and silica
K20-Na20-CaO-MgO-MnO-FeO-Fe203-AI203- polymorph formation, Vredefort Structure, South
Ti0 2-Si0 2-C-H-Oz. Journal of Metamorphic Africa, in Defects and processes in the solid state:
Geology, 8, 89-124. geoscience applications (eds J.N. Boland & J.D.
Ildefonse, J.-P. & Gabis, V. (1976) Experimental study FitzGerald). Developments in Petrology No. 14
of silica diffusion during metasomatic reactions in ('The McLaren Volume'), Elsevier, Amsterdam,
the presence of water at 550 0 C and 1000 bars. 69-84.
Geochimica et Cosmochimica Acta, 40, 297-303. Yardley, B.W.D. (1989) An introduction to metamor-
Kerrick, D.M. (ed.) (1991) Contact metamorphism. phic petrology. Longman, London, 248 pp.

Chapter two

Facies concept
and petrogenetic

2.1 Metamorphic facies, grade and nised in metapelitic sequences from many areas
zones of orogenic metamorphism, but it is not the
only zonal sequence recognised. The Barrovian
Metamorphic rocks show vanatIOns in their sequence characterises medium-pressure
mineralogical and microstructural features that regimes, whereas in low-P, high-T regimes,
are linked to variations in P- T conditions of kyanite and almandine garnet are absent, while
metamorphism. At low P- T conditions, typical andalusite and cordierite are important zonal
minerals are hydrous and generally of low minerals.
density (e.g. ChI, Ms, Qtz and PI). With There are limitations to the assessment of
increasing P- T, less hydrous and anhydrous metamorphic conditions based on single zonal
minerals become dominant (e.g. Stt, Grt, Ky, minerals. Not least of these is the fact that
Pyx and Hbl). Similarly, with increasing P- T, particular minerals will only develop in rocks
grain size generally increases. There are a of a certain chemistry, and even within the
number of ways of monitoring these changes range of rocks with appropriate chemistry, the
(i.e. classification) to record changing condi- first appearance of a particular mineral will not
tions. be synchronous. The use of mineral assem-
The zonal scheme was the first to be intro- blages, rather than individual minerals, is a
duced, by Barrow (1893, 1912). In the Scottish much more reliable approach to evaluating
Highlands, he established that metamorphic P- T conditions in any detail.
zones could be mapped out based on differ-
ences in the mineral assemblages of pelitic
rocks. Zonal boundaries were defined by
TABLE 2.1 Differences between the characteristic
isograds representing the first appearance of mineral assemblages of metasediments in the metamor-
key index minerals. From lowest to highest phic aureoles around intrusions of the Oslo region
metamorphic grade, the order of these key (Goldschmidt, 1911) and Orijiirvi, Finland (Eskola,
zonal minerals, referred to as Barrovian zones,
is chlorite, biotite, garnet (almanditic), stauro- Oslo (Goldschmidt, 1911) Orijiirvi (Eskola, 1915)
lite, kyanite and sillimanite. It has subsequently Kfs + And Ms + Qtz
proved possible to relate these changes directly Kfs + Crd Bt + Qtz
to the P-T conditions experienced by the rocks. Kfs + Hy + An Bt + Hbl
The Barrovian zonal sequence has been recog- Hy At

Facies concept and petrogenetic grids

The now firmly established concept of facies may include rocks of quite different
metamorphic facies was introduced by Eskola whole-rock chemistry. Indeed, mineralogical
(1914, 1915) following his work in Orijarvi studies of a range of rock types in a given area
(Finland), and with due consideration of the often allow greater certainty in assigning a
observations of Goldschmidt (1911) from the particular metamorphic facies. Eskola (1920)
Oslo area of Norway. Eskola had come to proposed five metamorphic facies to define
realise that in order to achieve or approach different conditions of metamorphism, with a
chemical equilibrium the mineral assemblages further three facies added in 1939. The work
observed in metamorphic rocks changed as a of Coombs and co-workers in the early 1960s
function of the P- T conditions experienced. It (e.g. Coombs, 1961) introduced the 'zeolite'
was apparent to Eskola that for a given range and 'prehnite-pumpellyite' facies, such that by
of rock compositions and a particular range of this time ten distinct metamorphic facies had
P- T conditions certain characteristic mineral been recognised. There then followed a period
associations always occurred, and that as P- T during which, rather than simplifying meta-
conditions varied so these mineral associations morphism using the 'facies' concept, the litera-
varied. By comparison of the Orijarvi assem- ture became cluttered and confused, as various
blages with those recorded by Goldschmidt workers began to subdivide the original
(1911) from zones around intrusions in the 'facies' into two or more 'sub-facies' based on
Oslo region, Eskola noticed that there were slight mineralogical differences. It was not
consistent mineralogical differences between always easy to assess whether these slight vari-
the two areas (Table 2.1). Those assemblages ations were solely P- T controlled, or whether
from Orijarvi are dominated by hydrous they were controlled by some other factor such
phases such as biotite and muscovite, while as fluid composition. In any case, the so-called
those from the Oslo region contain much K- 'sub-facies' were really 'facies' by definition,
feldspar. Since broadly comparable rock types since they contained a distinctive set of miner-
were involved, Eskola concluded that although als characterising certain conditions of forma-
metamorphism in both areas was related to tion.
high-level intrusions, that at Orijarvi occurred There are problems with the facies concept as
at lower T than that at Oslo. This difference defined by Eskola. For example, he did not
occurred due to the former being associated address the role of the fluid phase in any detail.
with granite intrusions, whereas the latter was Eskola, and others, either considered the fluid
in association with small mafic bodies, to be unimportant or assumed that all fluid was
intruded at significantly higher temperature. aqueous, and thus that Plolal = Pfluid = PH o. This
With further work, Eskola developed the is certainly not the case, and for many ~inerals
'metamorphic facies' concept and, in a classic the XH20: XC0 2 ratio, and fluid salinity, are
paper (Eskola, 1920), defined the term 'meta- crucial to their fields of stability in P- T space.
morphic facies' to designate a group of rocks For carbonate and calc-silicate rocks the facies
characterised by a definite set of minerals, approach has very limited use, because of
which under the conditions that prevailed the strong influence of fluid chemistry on the
during their formation were in perfect equilib- phase assemblage present. However, for meta-
rium with each other. The mineral composi- pelites and metabasites most geologists still use
tions in the rocks of a given facies vary the facies concept and tentatively regard Pf =
gradually in correspondence with variations in PH20 (or close to it), accepting that in some
the chemical compositions of the rocks. Tilley instances there are variable degrees of dilution
(1924) emphasized that a given metamorphic by CO 2, CH4 and various salts. In addition, it

Petrogenetic grids

must be recalled that zonal and facies the facies boundaries are not always sharply
approaches were based solely on hand speci- defined, and that transitional assemblages are
men and basic petrographic analysis of thin the norm.
sections, and only qualitative ideas of the link Rather than adopting the facies concept,
between mineralogical changes and P-T condi- Winkler (1979) preferred simply to divide P-T
tions. Since then, there have been major space into large divisions of metamorphic
advances, such as the advent of the microprobe grade based on key mineral reactions in
and computerised thermodynamic data sets. common rocks. These divisions are designated
These advances have allowed stability fields of as very low, low-, intermediate- and high-
key minerals and assemblages to be defined grade metamorphism. They are largely temper-
ever more clearly. This means that the P-T ature-related divisions, since most of the
ranges represented by the mineral assemblages reactions on which they are based are much
defined by Eskola in developing the facies more strongly temperature controlled, and in
concept are now known in a much more quan- many cases nearly isothermal. Winkler empha-
titative basis. Accordingly, it is more common sised that a qualitative estimate of pressure
nowadays for petrologists to associate a facies should also be stated (e.g. 'high-grade and
with a range of P-T space rather than individ- high-pressure metamorphism' or 'high-grade
ual mineral associations. and low-pressure metamorphism' etc.) in order
In Fig. 2.1, P-T space has been divided into to give a better idea of the environment of
12 facies, based on experimental work, empiri- metamorphism. As such, this particular
cal observations and theoretical stability fields approach gives a convenient overview of meta-
of key minerals and assemblages based on morphic conditions, without having to be too
thermodynamic data sets. Between the zeolite specific. It provides a useful and straightfor-
facies and the greenschist facies is an area ward first-order definition of metamorphic
referred to as the sub-greenschist facies. This conditions based on initial field observations.
term has been used to encompass the Within the present text, the approach taken is
prehnite-pumpellyite, prehnite-actinolite and to use 'grade' to make a generalised qualitative
pumpellyite-actinolite facies that are referred description of metamorphic P-T conditions,
to in some publications. The reason for prefer- while using 'facies' when speaking more
ring to amalgamate under the general heading precisely. Key mineral assemblages associated
'sub-greenschist' facies is because recent work with major compositional groups of rocks for
has shown that the previously used each metamorphic facies are listed m
prehnite-pumpellyite, prehnite-actinolite and Appendix III.
pumpellyite-actinolite facies show partial or
complete overlap in P-T space as a function of
2.2 Petrogenetic grids
even small whole-rock controls. The facies
boundaries of Fig. 2.1 are of necessity, and for Experimental work has now reliably estab-
the sake of simplicity, drawn up for systems in lished the P-T stability fields of many minerals
which PH2 0 =Ptotol. In many cases the effects of and mineral assemblages. Combining this
CO 2 and other fluids on given reactions are information allows the construction of petroge-
still either totally unknown or poorly defined, netic grids. On such grids, the intersection of
so at the present state of understanding it univariant reaction curves at invariant points
would be difficult to draw them in any other enables us to define the bounding conditions
fashion. It should be noted that metastable for particular equilibrium assemblages. From
persistence of phases (Chapter 1) means that this, the metamorphic conditions experienced

Facies concept and petrogenetic grids

by a particular rock can be quantified in terms accurate assessment of metamorphic tempera-

of a specific P-T range. A few minerals - such tures based on individual minerals or assem-
as jadeite, kyanite and andalusite - allow blages is considerably easier than assessing
reasonable estimation of maximum- or mini- pressures. Petrogenetic grids for the major
mum-pressure conditions, but the majority of chemical systems corresponding to metabasites,
minerals and reactions are most senSItive to metapelites and other systems relevant to meta-
changes in temperature. For this reason, the morphic rocks are comprehensively dealt with


Temperature ('>C)

FIG. 2.1 A simplified diagram illustrating the positions of the different metamorphic facies in P-T space. The
facies abbreviations are as follows: A, amphibolite facies; AE, albite-epidote hornfels facies; B, blueschist facies;
E, eclogite facies; EA, epidote-amphibolite facies; Gra, granulite facies; Gs, greenschist facies; HH, hornblende
hornfels facies; PH, pyroxene hornfels facies; S, sanidinite facies; sGs, sub-greenschist facies; Z, zeolite facies. The
boundaries between GS and AE, and between EA, A and HH, are not shown as solid lines because the distinction
between these facies is not always clear. The lines radiating from a dot (invariant point) are the univariant curves
defining the stability fields of Al 2SiOs polymorphs, based on Salje(1986). Note that key assemblages of major
rock types for different metamorphic tacies are given in Appendix III.


by Spear (1993) and will not be evaluated in decades such grids are largely being defined on
this text (see Appendix III for key assem- the basis of internally consistent thermody-
blages). namic data sets (e.g. Berman, 1988; Holland
The greater the number of components in a & Powell, 1990). This approach enables the
system, the more restricted the stability field of calculation of all the univariant curves and
an individual phase or phase assemblage divariant fields of interest for a particular
becomes. As an example, one of the best chemical system, as long as reliable thermody-
known and most studied systems is the one- namic data are available for all phases of the
component Al2SiO s system. The stability fields system. For simple end-member systems the
of the three Al 2SiO s polymorphs (andalusite, data are very good, and the grids are well
kyanite and sillimanite) have been of consider- defined. However, for complex systems such as
able interest to petrologists and have been those of metapelites and metabasites, involving
extensively studied both on an experimental phases with complex solid solution between
and theoretical basis (Kerrick, 1990). The end-members (e.g. chlorite, micas, garnet and
stability fields for andalusite, kyanite and silli- amphiboles), and sliding reactions in natural
manite shown in Fig. 1.2 are based on the ther- systems with variable fluid compositions, it is
modynamic data set of Berman (1988). Note less easy to construct a wholly reliable petro-
how the addition of H 20 into the system genetic grid. Even so, important advances are
prevents the formation of andalusite and kyan- being made, and some of the important grids
ite at lower temperatures and gives rise to the constructed over the past ten years include
formation of pyrophyllite instead. The addition those of Will et al. (1990) for ultramafic rocks;
of further components gives rise to further limi- Connolly & Trommsdorff (1991) for meta car-
tations and an increasingly complex system, bonates; Frey et al. (1991) for metabasites;
more closely resembling pelitic rocks. Strens and for metapelites the grids of Spear &
(1967) examined the effect of small amounts of Cheney (1989), Powell & Holland (1990) and
Fe2 0 3 on the stability fields of the Al2SiO s poly- Symmes & Ferry (1992). For comprehensive
morphs, noting that the univariant lines sepa- treatment of metamorphic phase equilibria,
rating the stability fields of different phases the text of Spear (1993) is strongly recom-
become divariant zones over which two of the mended.
polymorphs coexist. Bulk rock and fluid chem-
istry variations significantly influence whether
or not a particular mineral or mineral associa- References
tion is present at a given metamorphic grade. Barrow, G. (1893) On an intrusion of
In consequence, the absence of a phase such as muscovite-biotite gneiss in the south-eastern
Highlands of Scotland, and its accompanying meta-
staurolite in a garnet-mica schist does not morphism. Quarterly Journal of the Geological
necessarily preclude the possibility of staurolite Society, 49, 330-358.
grade conditions having been attained. As a Barrow, G. (1912) On the geology of lower Dee-side
general rule, it is inadvisable to place too much and the southern Highland Border. Proceedings of
the Geologists Association, 23, 274-290.
emphasis on the observations from a single thin Berman, R.G. (1988) Internally-consistent thermody-
section, but much more meaningful to draw namic data for minerals in the system
conclusions from a group of thin sections of Na20-K20-CaO-MgO-FeO-Fe203-AI203-Si02-
related rocks. Ti0 2-H20-C0 2. Journal of Petrology, 29,
Originally, it was experimental data and Connolly, J.A.D. & Trommsdorff, V. (1991)
empirical observation that aided the construc- Petrogenetic grids for metacarbonate rocks: pres-
tion of petrogenetic grids, but over recent sure-temperature phase-diagram projection for

Facies concept and petrogenetic grids

mixed-volatile systems. Contributions to (Kp-FeO-MgO-AI20 3-Si02-H20). American

Mineralogy and Petrology, 108, 93-105. Mineralogist, 75, 367-380.
Coombs, D.S. (1961) Some recent works on the lower Salje, E. (1986) Heat capacities and entropies of
grades of metamorphism. Australian Journal of andalusite and sillimanite: the influence of fibroliti-
Science, 24, 203-215. zation on the phase diagram of Al2SiOs polymorphs.
Eskola, P. (1914) On the petrology of the Orijiirvi American Mineralogist, 71, 1366-1371.
region in southwestern Finland. Bulletin de la Spear, F.S. (1993) Metamorphic phase equilibria and
Commission Geologique de Finlande, 40. pressure-temperature-time paths. Mineralogical
Eskola, P. (1915) On the relations between the chem- Society of America Monograph, Washington, DC,
ical and mineralogical composition in the meta- 799 pp.
morphic rocks of the Orijiirvi region. Bulletin de Spear, F.S. & Cheney, J.T. (1989) A petrogenetic grid
la Commission Geologique de Finlande, 44. for pelites in the system Si02-AI20 3-FeO-
Eskola, P. (1920) The mineral facies of rocks. Norges MgO-K20-H20. Contributions to Mineralogy and
Geologisk Tiddskrift, 6, 143-194. Petrology, 101, 149-164.
Frey, M., de Capitani, C. & Liou, J.G. (1991) A new Strens, R.G.]. (1967) Stability of Al2SiOs solid solu-
petrogenetic grid for low grade metabasites. Journal tions. Mineralogical Magazine, 31,839-849.
of Metamorphic Geology, 9, 497-509. Symmes, G.H. & Ferry, ].M. (1992) The effect of
Goldschmidt, V.M. (1911) Die kontaktmetamorphose whole-rock MnO content on the stability of garnet
im kristianiagebiet. Vidensk. Skrifter. 1. Mat.- in pelitic schists during metamorphism. Journal of
Naturv. K., 11. Metamorphic Geology, 10,221-237.
Holland, T.J.B. & Powell, R. (1990) An enlarged and Tilley, C.E. (1924) The facies classification of
updated internally consistent thermodynamic data- metamorphic rocks. Geological Magazine, 61,
set with uncertainties and correlations: the system 167-171.
K20-Na20-CaO-MgO-MnO-FeO-Fe203-AI203- Will, T.M., Powell, R. & Holland, T.J.B. (1990) A
Ti02-Si02-C-H2-0l" Journal of Metamorphic calculated petrogenetic grid for ultramafic rocks in
Geology, 8, 89-124. the system CaO-FeO-MgO-AIPrSi02-C02-HzO
Kerrick, D.M. (1990) The Al2SiO s polymorphs. at low pressures. Contributions to Mineralogy and
Reviews in Mineralogy No. 22, Mineralogical Petrology, 105, 347-358.
Society of America, Washington DC, 406 pp. Winkler, H.G.F. (1979) Petrogenesis of metamorphic
Powell, R. & Holland, T.J.B. (1990) Calculated rocks, 5th edn. Springer-Verlag, New York, 348
mineral equilibria in the pelite system, KFMASH pp.

Chapter three

groups of meta-
morphic rocks

Metamorphic rocks display a wide range of CALCITEIDOlOMITE

chemical compositions and mineral assem-
blages, reflecting the variety of original rock
types that become metamorphosed. Metasedi-
mentary rocks can broadly be subdivided into
three main categories; namely, metapelites,
metacarbonates/calc-silicates and quartzofelds-
pathic metasediments. There is of course a
complete spectrum of rocks between these end-
members (Fig. 3.1). In addition to metasedi-
mentary rocks, there are three main
compositional groups of meta-igneous rocks;
namely, metabasites, metagranitoids and meta-
ultra basics. The mineralogical and chemical FIG.3.1 A ternary diagram illustrating the nomencla-
characteristics of these various metamorphic ture in common use for metasedimentary rocks based
rocks are now discussed, with emphasis on key on the relative proportions of quartz, micas and
minerals and assemblages. The mineral abbre- carbonate minerals.
viations used throughout this book are those
suggested by Kretz (1983), and extended by
Bucher & Frey (1994). For quick reference, Accordingly, this system is known in petrolo-
they are tabulated in Appendix I. gist's shorthand as KFMASH. In addition to
the major components of the system, minor or
trace amounts of CaO, NazO, Fe z0 3, MnO and
3.1 Pelites TiO z may also be present, and thus give rise to
Rocks of this category represent metamor- extra mineral phases other than those expected
phosed argillaceous sediments (mudstones). As for the strict KFMASH system. The ratio of
well as being moderately to highly siliceous, individual oxides varies from one rock to the
such sediments have a chemistry characterised next, and this complex chemistry gives rise to
by high Alz0 3 and KzO, variable FeO and the development of a large range of minerals
MgO and much Hz 0 in hydrous minerals. during metamorphism. The appearance of

Compositional groups of metamorphic rocks

certain of these minerals can be very important 1965, 1967). More recently, Massone &
for assessing metamorphic grade. Schreyer (1987) have extended Velde's work,
and established that for the assemblage Phe +
PhI + Kfs + Qtz there is a very strong, almost
3.1.1 Medium-pressure '8arrovian'
linear, increase of Si content per formula unit
of phengite with respect to pressure, and a
Pelitic rocks initiate as mudstones, which in the slight decrease in Si content per formula unit
diagenetic field are dominated by fine-grained with increasing temperature. This relationship
clay minerals, quartz, chlorite and minor detri- has also been calibrated for use as a geobarom-
tal feldspar. The dominant clay minerals have eter. In certain pelitic lithologies, minerals such
mixed-layer structure (e.g. illite-smectite). With as stilpnomelane, paragonite or pyrophyllite
increasing heating and burial, the mixed-layer may occur.
clay minerals become ordered with the onset of The increase in temperature from low green-
'metamorphic' conditions, indicated by the schist to upper greenschist facies conditions is
mineral illite at around 200°C. Over the marked by the incoming of biotite, by the reac-
approximate temperature range 200-280°C, tion ChI + Kfs ~ Ms + Bt. At 4 kbar
there are recognisable changes in illite structure (KFMASH system) this occurs at around 440°
(crystallinity), as recorded by XRD and TEM ::!: 20°C. This first appearance of biotite by the

methods. Ultimately, illite is replaced by reaction given is not strictly in true pelites, but
muscovite, the 2M mica-in reaction curve being more correctly relates to greywackes with detri-
located at about 270-280°C. tal K-feldspar. In true pelites the biotite-in
Chlorite may occur either as authigenic or isograd occurs at slightly higher temperatures,
detrital crystals. During low-grade metamor- by a continuous reaction involving the break-
phism it changes from a Type Ib to a Type lIb down of chlorite and the transformation of
structure, this change occurring at tempera- white mica from phengite to muscovite. The
tures of the order of 150-200°C. At around range of temperatures for first appearance of
300-325°C, there is a transition from sub- biotite is due to the continuous or divariant
greenschist facies into the lower part of the true nature of the reaction. In rocks with a high Fe:
greenschist facies. At low greenschist facies Mg ratio (i.e. high XFeO) the reaction will
conditions, although often fine-grained, the proceed at 420°C, whereas for rocks with low
mineral assemblage is recognisable in thin XFeO temperatures> 450°C may be needed
section. For pelites, the characteristic assem- before biotite forms.
blage is ChI + Ms (often phengitic) + Qtz + Ab In particularly Mn-rich rocks, spessartine-
(Appendix III). This assemblage typifies rich garnets may occur in rocks of the lower
temperatures of the order of 325-425°C almost greenschist facies, but for typical pelites of the
regardless of pressure. The muscovites present KFMASH system, porphyroblasts of almanditic
are usually phengitic, having significant Fe (or garnets do not occur until temperatures of the
Mg) substitution for Al in octahedral sites. As order of 460-S00°C. The experimental work of
the temperature rises these micas become more Hsu (1968) has shown that pure almandine
muscovitic, with the Al content increasing at does not form until even higher T is attained.
the expense of Fe. A regular relationship of As with a significant spessartine component, a
increasing Al substitution for Si in the octahe- significant grossular component in almandine
dral site with increasing temperature and garnet has the effect of lowering the tempera-
decreasing pressure has been noted, and cali- ture of first formation. The bulk rock chem-
brated for use as a geothermobarometer (Velde, istry clearly has a strong influence on the


composltlon of almanditic garnets that form, frequently occurs as porphyroblasts, and exists
and as such the garnet-in isograd covers a over a broad range of pressure conditions. It is
broad range of temperatures and will vary not uncommon in contact metamorphic aure-
from one area to the next. Although spessar- oles (Plate 6(b)), as well as occurring in region-
tine-rich garnets may occur in lower green- ally metamorphosed pelites. Petrographic
schist facies rocks, the presence of almanditic studies, experimental work and petrogenetic
garnets is an indication of upper greenschist grids constructed on the basis of internally
(epidote-amphibolite) facies conditions or consistent thermodynamic data sets have estab-
higher. The continuous reaction ChI + Ms --7 lished that at temperatures of the order of
Grt + Bt + H 2 0 is probably the most important 520-560°C over the normal range of P-T
reaction to form almanditic garnet. As temper- conditions (3-10 kbar) during orogenic meta-
ature rises, the Mg: Fe ratio of garnet steadily morphism (higher T with higher P) chloritoid
increases as a result of a cation-exchange reac- reacts out to produce staurolite-bearing assem-
tion with biotite. blages (Fig. 3.2). In view of this, the presence of
At about the same P-T conditions as for the chloritoid gives a good indication of the maxi-
formation of almanditic garnet, the small mum temperature experienced by pelites, and
proportion of plagioclase present in pelitic the incoming of staurolite gives a good estimate
rocks becomes more calcic, changing from of the mInImUm temperature conditions
albite in greenschist facies rocks to oligoclase in attained. Stable coexistence of the assemblage
epidote-amphibolite and amphibolite facies Grt + St + Cld, although only seen in certain
rock. In terms of anorthite component, the aluminous pelites, is a good indication of
change is typically from An o_s (albite) in low temperature conditions of about 550°C
greenschist facies rocks, to Ans_Io in upper (Bucher & Frey, 1994).
greenschist facies rocks, and then a jump to Over the temperature range 550-600°C,
An IS _2S (oligoclase) In rocks of the pelites in the KFMASH system with high
epidote-amphibolite and low amphibolite alumina content and X FeO > 0.5 have the
facies. This jump marks the 'peristerite gap' assemblage Grt + ChI + 5t + Ms + Qtz. Those
corresponding to a structural change in plagio- pelites with similar X FeO ' but lower alumina
clase feldspar. This change in feldspar type is lack staurolite, but maintain the ChI + Bt
not recognisable by standard optical association (i.e. Grt + ChI + Bt + Ms + Qtz).
microscopy, but can be readily confirmed by Above 600°C (almost irrespective of pres-
electron-microprobe analytical work. sure), the association Grt + ChI is no longer
Another mineral characteristic of greenschist stable and in Fe-rich pelites the assemblage
and epidote-amphibolite facies conditions is Grt + St + Bt + Ms + Qtz occurs, while in
chloritoid. However, its occurrence is strongly pelites with a higher Mg content, the assem-
influenced by bulk rock chemistry, and it has blage St + Bt + ChI + Ms + Qtz is seen. This St
been shown that it only occurs in highly alumi- + Bt association is crucial, and marks the start
nous pelites with a very low calcium content. In of mid-amphibolite facies conditions. By a
the pure FA5H system (not true pelites), it can combination of field and thin-section study, it
start to appear at sub-greenschist facies condi- is an isograd that is quite easily defined. The
tions (c. 230°C) by the reaction ChI + Prl ~ Grt + St + Bt association is diagnostic of
Cld + Qtz (Fig. 3.2), but in average pelites of pelites of the mid-amphibolite facies (Ky + 5t
the KFMA5H system chloritoid appears at + Bt [no Grt] in more aluminous pelites), but
temperatures of around 300°C, by reactions once upper amphibolite facies conditions in
involving the breakdown of chlorite. Chloritoid excess of 670°C are reached, staurolite is no

Compositional groups of metamorphic rocks


+ 0)
Q) :E ~
'- 0
:::J +
en "0


0 Sil
+ And
:E 0 Crdl He
200 300 400 500 600 700 800
Temperature (OC)
FIG. 3.2 A petrogenetic grid for the FeO- AI 20 r Si02-H 20 (FASH) system (modified after Bucher & Frey,
1994; constructed using data of Holland & Powell, 1990; details of reactions involving cordie rite and hercynite
not shown). Shaded field marks St + Qtz stability.

Metacarbonates and calc-silicate rocks

longer stable in pelites, and breaks down by 10

the reaction 5t ~ Grt + Bt + Ky/5il, to give KNFASH system
H2().Qtz 981uration
the characteristic association Grt + Bt +
Ky/5il. 8
At mid-amphibolite facies conditions (P of li!
the order of 5-10 kbar, T = 600-670°C), -'"
kyanite is the stable Al 2 5iO s polymorph, !!!
whereas in the upper amphibolite facies, and II)

passing into the granulite facies (typically P of Q.

the order of 5-10 kbar, T ~ 700°C), it is silli- 4

manite that is the stable Al 25iO s phase present
(Fig. 3.3). At temperatures around 650-660°C
a certain degree of in situ partial melting
('anatexis') commences in metasedimentary 500 600 700 800
sequences (H 2 0-saturated conditions) to Temperature (Oc )
produce a 'granitoid' melt fraction. The melt
fraction is usually represented by irregular FIG. 3.3 The P-T diagram for the KNFASH (K 2 0-
Na O-FeO-AI 2 0 r SiO -H 2 0) system, showing the
pods, discontinuous layers and segregations sta5ility fields of the Ar2 SiO s phases, the curve for the
parallel to the schistosity/gneissosity of the Ms + Qtz ;r=" Kfs + Als + H 2 0 reaction and the curve
rock. At higher temperatures the melt fraction defining the conditions under which anatexis
may be more extensive and may give rise to commences. The reaction curves shown are based on
the thermodynamic data set of Holland & Powell
the development of migmatites (Fig. 4.11). (1990) . For simplicity, the numerous other reactions in
For H 2 0-saturated pelitic rocks with quartz in this system are not depicted.
excess, an important reaction indicative of
very high temperature conditions (typically T (commonly var. chiastolite; Plate 1 (c)) are
~ 700-750°C at P = 4-7 kbar) is the decom- common in suitable pelitic lithologies.
position of muscovite and formation of K- Andalusite is a very good indicator of low-
feldspar by the reaction Ms + Qtz ~ Kfs + pressure conditions. It is unstable above P =
Al 2 5iO s + H 2 0 (Fig. 3.3). 4.5 kbar and most typical of P < 4kbar (Fig.
3.3). Above 515-550°C (depending on P and
3.1.2 Low-pressure assemblages bulk rock chemistry), porphyroblasts of
cordierite form, and the assemblage And + Crd
The succession of assemblages and reactions + Bt is common to many rocks. In other cases
observed in KFMA5H pelites under lower- staurolite (Plate 1 (a)) develops to give the
pressure conditions (T ~ 4.5 kbar), while show- common association of And + 5t + Bt. Precisely
ing some similarities to those developed during which assemblages develop is largely a function
medium-pressure 'Barrovian' metamorphism, of bulk rock chemistry and pressure.
also exhibits some notable differences. The For a suite of rocks being metamorphosed at
incoming of biotite is one of the first clearly P = 3.5 kbar, the andalusite-sillimanite reaction
recognisable mineralogical changes when line is crossed at about 650°C (Fig. 3.3).
moving into a contact metamorphic aureole (or Commonly, andalusite can exist meta stably
an area of regional high heat flow). The biotite- even once this line has been crossed, but above
in isograd indicates temperatures of about temperatures of 660-670°C it is common to
425°C or greater. At these temperatures and record fibrolitic sillimanite, as well as K-
above, porphyroblasts of andalusite feldspar development from muscovite.

Compositional groups of metamorphic rocks

Prismatic crystals of sillimanite tend to be more meta-carbonates and calc-silicate rocks. H 20

characteristic of higher temperatures. At and CO 2 are the dominant components in typi-
temperatures above 680°C, the assemblage Qtz cal metamorphic fluids, and the amount of
+ PI + Kfs + Bt + Sil is most characteristic of each is commonly expressed as a value between
pelitic rocks. In common with higher-pressure 0.0 and 1.0 of total fluid. For example, a fluid
regional metamorphic regimes, the highest- with XH20 = 0.8 and XC0 2 = 0.2 is a fluid
temperature parts of contact aureoles or belts that is 80% H 20 and 20% CO2 , The values of
of regional high heat flow are dominated by XH20 and XC0 2 show considerable variation
rocks comprising a significant proportion of in fluids of carbonate and calc-silicate rocks,
granitoid melt. and act as an important control over the
assemblages produced and the temperatures at
which certain reactions occur (Trommsdorff &
3.1.3 High-pressure assemblages
Evans, 1977; Walther & Helgeson, 1980). In
In the high-P/low-T blueschist facies terranes, general, high XC0 2 inhibits most reactions, so
pelites exhibit few reactions or assemblages of that for any given reaction higher T is required
note. At medium pressures they are charac- for it to occur (Fig. 1.3). The salinity of aque-
terised by much the same phase assemblage as ous fluids also has an important effect.
greenschist facies rocks. However, it is worth Generally speaking, the higher the salinity, the
noting that the temperature conditions of the higher is the temperature required for a given
blueschist facies are generally too low for the reaction to proceed.
formation of biotite. The brown, strongly Metamorphism of a pure CaC0 3 limestone
pleochroic sheet-silicate phase sometimes will give rise to recrystallisation of the calcite,
mistakenly identified as biotite is in fact usually to produce a fairly equigranular calcite marble,
stilpnome1ane (Figs 5.5 & S.13(a)). Hence the probably becoming coarser with increasing
characteristic phase assemblage of blueschist grade. Only under very high pressure condi-
facies pelites is therefore ChI + white mica tions will any mineralogical change be seen,
(often phengitic) + Ab + Qtz (+ Stp, Cal, Ep) + and under such conditions the polymorphic
opaques (often 11m or Py). While abundant in transformation of calcite to the denser poly-
metabasites at this grade, sodic amphiboles are morph, aragonite, will occur (Fig. 12.19).
notable by their absence from pelites. In the case of siliceous dolomitic lime-
stones (CaO-MgO-Si02-H20-C02 ; CMSHC
3.2 Metacarbonates and calc-silicate system), the increased variety of original rock
rocks chemistry means that many reactions are pos-
sible during progressive metamorphism. In
Sedimentary carbonates are chemically domi- low-pressure orogenic metamorphism, at tem-
nated by CaC0 3, MgC03 and Si02 , with peratures less than 500°C, in the presence of a
increasing amounts of Al2 0 3 and K20 as they H 2 0-rich fluid, talc should be a diagnostic
grade into pelites. The mineral assemblages mineral, but it is often only seen as a retro-
produced during progressive metamorphism grade phase, since on the prograde path it
are dependent not only on P, T and bulk rock breaks down by the tremolite-forming reaction:
composition, but additionally fluid composi- 2Tlc + 3Cal ~ Tr + Dol + H 20 + CO2 , For
tion has a profound effect on the assemblages fluid compositions in the range XH20 =
produced and the temperatures at which partic- 0.3-0.9, the tremolite-forming reaction in
ular reactions occur. This is true of all meta- siliceous dolomitic limestones is SDol + 8Qtz +
morphic rocks, but is especially true of H 2 0 ~ Tr + 3Cal + 7C0 2 • Since this


consumes H 20 and liberates CO 2, then in a extensively recrystallised to produce a more

closed, internally buffered system, the fluid stable granoblastic-polygonal aggregate (Fig.
composition is driven to increasingly higher 5.14(b)}. In immature sandstones, detrital miner-
XC0 2 • At low- to mid-amphibolite facies als give increased variety, although the metamor-
conditions (T = 500-670 o q, tremolite-bearing phic assemblage will usually be little more than
assemblages dominate, as long as XH 2 0 is Qtz + PI + Kfs + mica. There is a complete
~0.3. At upper amphibolite conditions and gradation from quartzites via psammites and
higher, diopside is the characteristic calc-silicate semi-pelitic schists to pelitic schists. However,
phase. At high XC0 2, it forms by a reaction since there are no standardised definitions for
involving the breakdown of Dol + Qtz, but at these rocks based on the relative proportions of
moderate to high XH20, it forms from the the main end-member components, no definite
breakdown of calcite + tremolite. boundaries have been marked on Fig. 3.1.
Assemblages containing forsterite + diopside The simple mineralogy of metamorphosed
require high-T conditions and characterise meta- quartzofeldspathic sediments generally makes
morphosed siliceous dolomites from the inner- them of little use in assessing metamorphic
most parts of contact metamorphic aureoles, and grade. However, during the metamorphism of
situations of very high T hydrothermal metamor- such rocks at blueschist facies conditions, a
phism. Wollastonite may be seen in certain meta- crucial change in mineralogy does occur. With
carbonates and calc-silicate rocks near to major increasing pressure, albite becomes unstable
intrusions, but is not associated with orogenic and reacts out to form jadeitic pyroxene +
metamorphism. It commonly forms by the decar- quartz. In meta clastic rocks, jadeite is never
bonation reaction Cal + Qtz = Wo + CO2, but pure but always contains at least 10% of other
the temperature at which this reaction occurs is pyroxene components, such as diopside, heden-
greatly influenced by the H 20: CO2 ratio of the bergite and aegirine. The purity of the jadeitic
fluid (Fig. 1.3). At total fluid pressure {Pf} in the pyroxene gives a good indication of the P
range 1-3 kbar, for XC0 2 = 0.13 the reaction conditions attained, with the purest jadeite
proceeds at about 520-550°C whereas for XC0 2 being indicative of the highest P.
= 1.0 (i.e. pure CO2 fluid) temperatures of Although their mineralogy can hardly be
700-800°C are required (Greenwood, 1967). described as excltmg, the textural and
Impure carbonates containing a significant microstructural changes observed during the
pelite component will have a chemistry with a metamorphism and deformation of quartzites
large proportion of A1 20 3• When metamor- and quartzofeldspathic rocks have attracted
phosed, Ca-Al silicates such as epidote, grossu- considerable interest over the decades. This is
lar, hornblende, idocrase and anorthite are largely due to the differences in behaviour of
commonly produced. The K2 0 component of quartz and feldspar during deformation, the
the original clays usually gives rise to simple slip systems of quartz and the simple
muscovite or phlogopite in the calc-schists and mineralogy, which makes it easier to isolate
calc-silicate rocks produced. and understand those processes in operation
under different experimental and geological
conditions (see Chapter 8 for more details).
3.3 Quartzofeldspathic metasediments
When metamorphosed, clastic sediments such as
sandstones and arkoses give rise to quartzitic
3.4 Metabasites
and quartzofeldspathic metamorphic rocks. Metabasic rocks represent metamorphosed
During metamorphism pure quartzites are basalts, dolerites, gabbros and basaltic

Compositional groups of metamorphic rocks

andesites. Their chemistry is typified by 45-55 include porosity/permeability, fluid composI-

wt% SiOz and dominant MgO, FeO, Fe z0 3 , tion (XH2 0 : XC0 2 ) and pressure, which at
CaO and Al z0 3 , with lesser amounts of NazO such high crustal levels may be at or close to
and Ti0 2, and very little K2 0. Consequently, a Phydrostatie, rather than Plithostatie' Although
varied assemblage forms during metamorphism complex, the general trend is one of most
of mafic rocks, dominated by silicates enriched hydrous zeolites at lowest T being replaced by
in the major components listed. Like meta- least hydrous zeolites at higher T. A combina-
pelites, metabasic rocks show mineralogical tion of experimental work and improved ther-
changes that have proved useful for assessing modynamic data sets has enabled the stability
metamorphic conditions. At low metamorphic fields for many of the key zeolites to be well
grades original igneous textures and micro- defined with reference to simple chemical
structures may be recognisable, but with systems. For example, in the CASH system,
increasing metamorphism and deformation heulandite-analcime represents temperatures of
such features are usually lost. Metabasites from the order of 100-200°C. It is succeeded by
environments of orogenic metamorphism or laumontite, representative of about 200-275°C.
subduction-zone metamorphism generally have The high-temperature zeolite, wairakite, is
a pronounced schistosity, while basic hornfelses common in the active geothermal system
and those rocks that have experienced granulite (Taupo Volcanic Zone) of the North Island of
facies or eclogite facies metamorphism New Zealand. It is indicative of temperatures
normally show a granoblastic structure. of the order of 250-400°C in the CASH
Pressures of less than 3 kbar and tempera- system, but is rarely encountered in other
tures of up to 300°C characterise areas of areas. In the complex chemical system of
burial metamorphism, very low grade orogenic natural metabasites (NCMFASH), the equating
metamorphism, ocean-floor metamorphism of specific temperature stability ranges for the
and metamorphism in geothermal regions. At various zeolites is less assured. In many areas it
such grades it is typical to observe mixed dis- is more usual to pass directly from laumontite
equilibrium assemblages containing high-T assemblages into rocks bearing the minerals
igneous minerals in association with low-T prehnite, pumpellyite and actinolite, along with
metamorphic minerals. This is because fluid ubiquitous quartz and albite. In simplified
access to the rock is very heterogeneous, NMASHINCMASH systems, it is possible to
depending strongly on the extent of fracturing recognise distinctive facies that appear to be
and/or primary porosity, such as in lava flow diagnostic of discrete P-T regimes. These have
tops. The zeolite group of minerals (Fig. been referred to as prehnite-pumpellyite,
5.12(c)) can prove useful as a means of docu- prehnite-actinolite and pumpellyite-actinolite
menting changes in metamorphic conditions. facies. However, in more complex systems
However, these only form in association with (NCFMASH) and natural metabasites, such
high XH2 0 fluids. If the fluids have moderate distinct 'facies' show substantial overlap. For
or high XC0 2, carbonate minerals form. this reason I use the term 'sub-greenschist
Various zeolite sequences have been docu- facies' for P-T conditions represented by
mented in the basaltic lava piles of east Iceland, assemblages with prehnite and/or pumpellyite
in various ophiolites worldwide and in the (Fig. 2.1), rather than attempting more detailed
geothermally active volcanic regions of New subdivision.
Zealand. Such zonal sequences are variable With increasing temperature, prehnite and
from one area to another, and are dependent pumpellyite become unstable and are replaced
on variables other than temperature. These by epidote group minerals, in the transition to

Metamorphosed ultramafic rocks

the greenschist facies. In the greenschist facies, temperature of metamorphism, there is a

(-325°C to 4 75°C), the characteristic assem- tendency for hornblende to become more Ti-
blage of metabasites is Act + ChI + Qtz + Ab + rich, and for the pleochroic scheme to change
Ep + Spn (Appendix III). The transition from from blue green - green under epidote amphi-
greenschist facies to amphibolite facies via the bolite facies conditions to green-brown in the
epidote-amphibolite facies of Eskola (1939) is upper amphibolite facies.
indistinct in most areas, and is marked by The boundary between hydrous amphibolite
as·semblages in which two Ca-amphiboles (acti- facies and anhydrous granulite facies is very
nolite and hornblende) coexist. In practice, this much transitional, but as the temperature rises
means that it is usually simpler to identify defi- the last remaining hydrous phase (hornblende)
nite greenschist facies rocks and definite reacts out to form pyroxenes. Although some
amphibolite facies assemblages, and to refer to hornblende persists in granulite facies rocks,
intermediate assemblages as transitional. The most has usually gone by 750-800°C. When
change from greenschist to amphibolite facies metabasites have been metamorphosed under
is marked by a number of simultaneously granulite facies or pyroxene hornfels facies
occurring reactions. The most notable of these conditions, in addition to the characteristic
are: (i) the actinolite to hornblende reaction in granoblastic structure (Section 5.5), they also
the Ca-amphibole solid-solution series by display a distinctive mineralogy consisting of
Tschermak substitution (MgSi (or FeSi) = orthopyroxene (hypersthene) + anorthitic PI +
Al'vAlvI exchange), coupled with chlorite break- Cpx (± Spl ± Grt). Under high-pressure gran-
down, and (ii) the simultaneously occurring ulite facies conditions (bordering eclogite
reaction involving breakdown of albite and facies) the anorthitic plagioclase becomes
epidote to form oligoclase, with Na released in increasingly unstable and ultimately reacts out.
this reaction incorporated in hornblende. The precise position of the plagioclase-out
As well as the miscibility gap in the Ca- reaction in mafic and ultramafic rocks is
amphiboles there is also a miscibility gap in the strongly dependent on the composition of the
case of the plagioclase (the 'peristerite gap'). parent rock. While it is generally absent from
The 'peristerite gap' in metabasic rocks at low pyroxene hornfels facies rocks, garnet is
P (2 kbar) has been studied in detail by common in granulite facies rocks, and becomes
Maruyama et al. (1982). They found that the increasingly represented at higher P. Again, the
transition zone consisting of 'peristerite pairs' + extent of the P-T conditions suitable for
Ep + ChI + Ca-amph (usually Act + Hbl) + Qtz garnets in metabasites/ultramafites of the gran-
+ Spn occurs between 370°C and 420°C. At ulite facies is strongly bulk-rock controlled.
higher T, this assemblage is succeeded by one The change from spinel-bearing to garnet-
in which actinolite is absent, chlorite persists bearing pyroxenites and lherzolites occurs
and oligoclase is present (PI(An zo_so )+ Hbl + ChI within the top granulite facies through eclogite
+ Spn + 11m. At temperatures of about 500°C facies and reflects increased pressure. The exact
and moderate pressures of orogenic metamor- position of this reaction varies according to
phism, amphibolites have the characteristic bulk rock chemistry, but the presence of garnet
assemblage Olg + Hbl + Ep + Rt (+ Qtz + Grt ± in such rocks is a good indication of very high
Bt). As upper amphibolite facies conditions are pressure.
attained, garnet becomes modally more impor- In the eclogite facies the characteristic
tant, and plagioclase becomes increasingly mineral association is Cpx(omphacite) + Grt,
calcic at the expense of epidote (which becomes commonly in equal porportions. Accessory
increasingly scarce or absent). With increasing phases often present include quartz, rutile and

Compositional groups of metamorphic rocks

kyanite. Green & Ringwood (1967) undertook CFMASH. In cases in which the fluid has a
a classic piece of experimental work examining significant CO 2 component, the systems
the transformation of gabbros into eclogites become MS-HC, CMAS-HC and CFMAS-
with increasing pressure and under virtually HC, and carbonate phases (e.g. magnesite)
anhydrous conditions. Although the precise form part of the mineral assemblage.
boundaries for the eclogite facies are not well Highly sheared and metamorphosed ultra-
defined on the P-T grid, such rocks are clearly mafic rocks occur as discontinuous pods and
representative of high-P/high-T conditions. lenses in fault and shear zones, as well as in
Metabasites from the environment of thick units at the base of many ophiolite slabs.
blueschist facies metamorphism have a distinc- Rocks from these environments often show
tive mineralogy dominated by sodic amphi- pronounced and pervasive ductile shear fabrics.
boles such as glaucophane or crossite. Fluid flow through the shear zones readily
Experimental work has shown that the phase leads to the retrogression of the primary
assemblages of blueschist facies rocks are igneous assemblage to give serpentinites. While
indicative of high-P/low- T conditions. In the serpentine is widely associated with metamor-
lower-pressure part of the blueschist facies, phosed ultramafic rocks, other Mg-rich miner-
with P of the order of 5-8 kbar and T = als such as anthophyllite, talc, magnesite and
200-350°C, the metabasite assemblage is typi- clinochlore are also common in certain ultra-
cally GIn + Ep (or Lws) + Spn + Ab + Qtz (+ mafites. The Fe component of ultramafites is
ChI + white mica + Stp + Cal). Higher-pressure re-utilised in the new growth of Fe-oxides, Fe-
blueschists may contain small amounts of bearing amphiboles and chlorites. None of
jadeitic pyroxene in addition to glaucophane. these minerals is especially useful for tightly
Such rocks lack albite, and if a carbonate phase defining P-T conditions, especially if the fluid
is present it is most likely to be aragonite rather composition is not well known.
than calcite. Many blueschist facies rocks are During metamorphism of simple MgO-Si0 2
transitional into the greenschist facies (Section ultramafites (harzburgites), it is well estab-
12.3.5). In addition to Na-amphiboles (glauco- lished that fluids rich in CO2 generate entirely
phane--crossite), such higher-temperature blue- different mineralogies compared to H 20-rich
schist facies rocks commonly contain garnet, fluids. In mixed fluids the precise H 2 0 : CO 2
and a second amphibole such as actinolite or ratio significantly controls the temperature at
the Na-rich sub-calcic hornblende known as which a given reaction occurs for given fluid
barroisite. pressure (Johannes, 1969). Serpentine group
minerals are indicative of very aqueous fluids
(i.e. very low XC0 2 ) and while often associ-
3.5 Metamorphosed ultramafic rocks
ated with metamorphism at temperatures less
Ultramafic rocks such as dunite, lherzolite, than 400°C, the extreme upper limit is shown
harzburgite and pyroxenite have an assemblage to be at about 480°C/1 kbar to 590°C/7 kbar
characterised by olivine and pyroxene, with the (Evans & Trommsdorff, 1970). Chrysotile is
proportions varying according to rock type. the stable serpentine group mineral below
Hydrated ultramafic rocks generally have a about 250°C, but at higher temperatures
simpler chemistry compared to metapelites and (greenschist-amphibolite facies) antigorite is
metabasites, and are dominated by MgO, Si0 2 , the characteristic serpentine phase. The upper
H 20 with CaO and A1 20 3, and so can be temperature limit for magnesite is somewhat
considered as a MSH to CMASH system. For similar to that for serpentine minerals,
hydrated lherzolites the system is more strictly although magnesite can form over a very


broad range of fluid composition from high nied by epidote group minerals. Under similar
H 2 0 to high CO 2 , conditions, clastic sediments will also retain
Talc forms over a wide range of fluid and much of their original texture, and thus it is
P-T conditions. The lower stability limit ranges easy to distinguish them from metagranitoid
from as low as about 310°C for highly aqueous rocks. However, at higher temperatures the
fluids at 1 kbar, to as high as 670°C if the fluid original rock microstructure becomes blurred,
is very H 2 0-rich (see Johannes, 1969, for and ultimately lost, as the quartz-
detailed reaction curves). The presence of feldspar-mica assemblages recrystallise to a
anthophyllite, enstatite or forsterite in meta- more stable arrangement. As temperature
morphosed ultramafics signifies a high-grade increases, the plagioclase becomes more calcic,
assemblage. Anthophyllite is stable at tempera- and at high grades of metamorphism muscovite
tures between about 500°C and 760°C, will ultimately break down to K-feldspar. At
depending on bulk rock chemistry, fluid these high grades it is virtually impossible,
composition and Pf' Forsterite and enstatite are from microscopic work alone, to distinguish
commonly part of the original igneous assem- rocks that were originally 'granites' from those
blage. The reaction curves determined by that were quartzofeldspathic sediments.
Johannes (1969) suggest that the lower stability Similarly, highly sheared or mylonitised
limit for these minerals is of the order of granites can be difficult to distinguish from
500-550°C at 2 kbar. mylonitised arkosic sediments. However, a
CMASH (CFMASH) systems of metamor- good understanding of the field relationships
phosed lherzolites, as well as containing and associated lithologies to each side of the
serpentine and talc, also have chlorite, Ca- shear zone will normally resolve the problem.
amphibole (tremolite) and/or Ca-pyroxene
(diopside) as key phases at greenschist and
amphibolite facies. Passing into the granulite References
facies, spinel forms, while chlorite and amphi- Bucher, K. & Frey, M. (1994) Petrogenesis of meta-
bole disappear, to give a typical assemblage of morphic rocks. Springer-Verlag, Berlin, 318 pp.
Eskola, P. (1939) Die metamorphen Gesteine, in Die
Fo + En + Cpx + Spl at low to moderate pres- Entstehung der Gesteine (eds T.F.W. Barth, C.W.
sure, and Fo + En + Cpx + Grt at high pressure, Correns & P. Eskola). Julius Springer, Berlin
bordering eclogite facies. (reprinted, 1960, 1970),263-407.
Evans, B.W. & Trommsdorff, V. (1970) Regional
metamorphism of ultramafic rocks in the Central
3.6 Meta-granitoids Alps; paragneisses in the system CaO-MgO-
Si0 2-H 20. Schweizerische Mineralogische und
The characteristic mineral assemblage of Petrographische Mitteilungen, 50, 481-492.
deformed and/or metamorphosed 'granitoid' Green, D.H. & Ringwood, A.E. (1967) An experimen-
tal investigation of the gabbro to eclogite transfor-
rocks is essentially the same as that for mation and its petrological implications.
deformed and/or metamorphosed quartzofelds- Geochemica et Cosmochimica Acta, 31, 767-833.
pathic sediments. At low metamorphic grades Greenwood, H.J. (1967) Wollastonite: stability in
in granitoid rocks that have experienced only H 20-C0 2 mixtures and occurrence in a contact-
metamorphic aureole near Salmo, British Columbia,
limited deformation, the original igneous Canada. American Mineralogist, 52, 1669-1680.
texture and mineralogy is usually recognisable. Holland, T.J.B. & Powell, R. (1990) An enlarged and
However, the feldspars commonly show signs updated internally consistent thermodynamic
of alteration. They frequently show partial dataset with uncertainties and correlations: the
system K20-Na 20-CaO-MgO-MnO-FeO- Fe 20 3-
replacement or complete pseudomorphing by a A120j-Ti02-Si02-C-H2-02' Journal of Meta-
fine aggregate of sericite, typically accompa- morphic Geology, 8, 89-124.

Compositional groups of metamorphic rocks

Hsu, L.c. (1968) Selected phase relationships in the Trommsdorff, V. & Evans, B.W. (1977) Antigorite-
system Al-Mn-Fe-Si-O-H: a model for garnet equi- ophicarbonates: contact metamorphism in
libria. Journal of Petrology, 9, 40-83. Valmalenco, Italy. Contributions to Mineralogy and
Johannes, W. (1969) An experimental investigation of Petrology, 62, 301-312.
the system MgO-Si0 2-H 2 0-C0 2 • American Journal Velde, B. (1965) Phengite micas: synthesis, stability and
of Science, 267,1083-1104. natural occurrence. American Journal of Science,
Kretz, R. (1983) Symbols for rock-forming minerals. 263, 886-913.
American Mineralogist, 68, 277-279. Velde, B. (1967) Si+ 4 content of natural phengites.
Maruyama, S., Liou, ].G. & Suzuki, K. (1982) The Contributions to Mineralogy and Petrology, 14,
peristerite gap in low-grade metamorphic rocks. 250-258.
Contributions to Mineralogy and Petrology, 81, Walther, ].V. & Helgeson, H.C. (1980) Description
268-276. of metasomatic phase relations at high pres-
Massone, H.]. & Schreyer, W. (1987) Phengite sures and temperatures: 1. Equilibrium activities
geobarometry based on the limiting assemblage with of ionic species in non-ideal mixtures of CO 2
K-feldspar, phlogopite, and quartz. Contributions and H 20. American Journal of Science, 280,
to Mineralogy and Petrology, 96,212-224. 575-606.

Part B

Introduction to
textures and
The chapters in this part aim to give a broad materials science groups all crystalline materi-
introduction to the description and interpreta- als together in terms of basic processes, it is
tion of the fundamental textures and logical to use a common descriptive terminol-
microstructures of metamorphic rocks. The ogy. In synthetic materials (e.g. metals)
significance of each will be emphasised and, 'texture' specifically relates to cases of
where relevant, current theories and controver- preferred orientation. It therefore makes sense
sies relating to their origin will be discussed. to use the term in a similar way when referring
Before doing this, however, it is first necessary to rocks, rather than perpetuate a more general
to define texture, microstructure and equilib- use of the term and promote possible confu-
rium in relation to metamorphic rocks. SIOn.

Definition of texture and microstructure Equilibrium and equilibrium assemblages

In many previous texts dealing with metamor- Assessing the degree of equilibrium attained by
phic rocks, there has been a more or less a particular metamorphic assemblage may not
synonymous usage of the terms TEXTURE and always be easy. The strict definition of equilib-
MICROSTRUCTURE to describe the shapes rium is that state of a rock system in which the
and arrangement of grains within the rock. phases present are in the most stable, low-
However, following the reasoning of Vernon energy arrangement, and where all phases are
(1976), it is advocated here that the term compatible with the given P, T and fluid condi-
MICROSTRUCTURE should be used to cover tions. This concept has been introduced in
all aspects of the microscopic arrangements Sections 1.3 &1.4.
and interrelationships between grains, while If the equilibration rate is rapid compared to
restricting the term TEXTURE to those the rate of change of P, T and/or fluid condi-
arrangements in which there is some preferred tions, then the rock will maintain equilibrium.
orientation. This more specific use of the term This may be true through certain parts of a
'texture' is favoured for a number of reasons. rock's P-T trajectory, but is seemingly not the
In particular, since the modern approach to case throughout the whole trajectory. The

Introduction to metamorphic textures and microstructures

many high-grade metamorphic assemblages pseudomorphs, and yet other phases of the
preserved in rocks at the Earth's surface high-grade assemblage appear entirely fresh. In
provide a clear indication of this lack of equili- such cases the complete assemblage is very
bration. It therefore seems that the rate of equi- much a disequilibrium assemblage. It contains
libration is much slower than the rate of more phases than would satisfy the Phase Rule
change in P-T during uplift, and that the and it contains phases representing both high-
assemblage observed records some earlier stage grade and low-grade metamorphic conditions.
of the P-T evolution. In most cases it is taken The degree of microstructural equilibrium also
to represent the assemblage at peak metamor- depends on the rate of change of the different
phic conditions. How close to equilibrium this variables. The most stable arrangement of
assemblage is depends as much as anything on grains is a polygonal aggregate of unstrained
how long the particular rock was held at or crystals with flat faces, whereas aggregates of
close to peak conditions, since the longer a strained crystals with irregular boundaries
particular set of conditions prevails, the more show a much lower degree of equilibrium.
time there is to equilibrate. In many rocks there
are signs of partial equilibration to lower P-T Reference
conditions in that some phases show signs of Vernon, R.H. (1976) Metamorphic processes. George
retrogression in the form of reaction rims or Allen & Unwin, London, 247 pp.

Chapter four

banding and

4.1 Compositional layering primary compositional banding, as well as

producing additional segregation of felsic and
The presence of LAYERING or BANDING mafic minerals. The combination of processes
in metamorphic rocks can be either primary described above, often synchronous with
or secondary in origin, and can occur on active shearing at elevated temperatures,
the macro-, meso- or microstructural scale. produces the characteristic inter banding of
In metamorphosed sedimentary and vol- quartzofeldspathic and mafic layers referred to
LAYERING is often still recognisable. This is ITY) or MIGMATITIC BANDING. Such
due to primary chemical differences that influ- banding is generally coarse (centimetre to
ence the nature of metamorphic mineral assem- decimetre scale) and while recognisable in
blages that can develop. Such differences are large hand specimens is not readily appreciated
often recognisable in thin section by virtue of in thin section. AUGEN GNEISS is a special
mineral assemblage or grilit size variations (Fig. name applied to rocks with a structure consist-
4.1). In low- to medium-grade pelitic and semi- ing of 'augen' (eyes), usually of feldspar (typi-
pelitic sequences it is often possIble to recognise cally K-feldspar), in a strongly foliated gneissic
fining-upwards sequences and to determine (typically Qtz-Bt-Ms) matrix (Fig. 4.3). This
'way-up' based on grain size variatiOns or subtle structure develops by a combination of defor-
changes in the proportion of quartz and phyl- mation and recrystallisation of original coarse-
losilicate minerals present (Fig. 4.2). grained quartz-feldspar-mica assemblages (e.g.
In high-grade gneisses and migmatites, metagranitoids) at low to moderate shear
original compositional layering still has a strains and moderate to high temperatures. At
strong influence on nature of banding higher shear strains mylonitic rocks are
observed, but at such high metamorphic produced (Chapter 8). In fact, there is a
grades recrystallisation is so extensive that complete transition between the appearance of
almost all primary features are obliterated. typical augen gneisses and granitic mylonites.
Diffusive mass transfer and in situ partial Although augen gneiss is most commonly
melting contribute to the enhancement of any developed in metamorphic rocks of original

Layering, banding and fabric development

FIG. 4.1 Compositional layering in greenschist facies semi-pelitic schist from Sierra Leone. Alternating layers are
richer and poorer in biotite with respect to quartz. Note also the weak schistosity trending oblique to the layering
from the top right to the bottom left. Scale = 1 mm (PPL).

gramtlc to granodioritic composltlOn, mafic 4.2 Introduction to stress, strain and

gneisses may also display augen structure. fabric development
A further example of compositional (miner-
alogical) banding seen in metamorphic rocks is The forces acting on a rock mass produce a set
the case of zoned or banded SKARNS. These of stresses. Since force (unit = newton (N)) is a
are often formed in contact metamorphic aure- vector quantity, the stresses can be defined in
oles and other situations in which hydrother- terms of direction and magnitude. The unit of
mal silica-saturated fluids have infiltrated stress (pascal (Pa) = 1 N m-2 ) is the same unit
carbonate (metacarbonate) rocks, or inter- as that of pressure, and although many meta-
banded carbonate and siliceous rocks. Bands of morphic petrologists talk of pressure in terms
calc-silicate minerals (skarns) form due to of bars and kilo bars, the SI unit is strictly the
metasomatic reaction between carbonate and pascal (1 bar = 105 Pa; 1 kbar = 0.1 GPa (100
silicate systems in mutual contact. The banding MPa)).
develops by diffusive migration of ions both When a body of rock (metamorphic or
towards and away from the reaction front: otherwise) is subjected to a system of superim-
some atoms such as Si are clearly sourced from posed stresses, it experiences a certain degree of
the silicate system, while others such as Ca and strain reflecting the stress imposed. Strain is
Mg are likely to have originated from the defined as the change in size and shape of a
marble. However, the origin of some elements body resulting from the action of an applied
present in skarn bands may be ambiguous, and stress field (Park, 1989). Strain can be recorded
it is quite likely that they have been introduced in terms of volume change, distortion or both.
with an exterq,ally derived fluid rather than The shape change may be non-rotational (co-
coming from immediately adjacent rocks. axial), as in the case of pure shear regimes, or

Planar fabrics in metamorphic rocks

FIG. 4.2 A field photograph of grading in pelitic/semi-pelitic rocks, Snake Creek, Queensland, Australia. The
pencil is 12 cm long; the younging direction (left to right) is shown by the symbol in the lower left corner. Platy
layers represent micaceous (originally fine-grained clay-rich) tops to cycles.

have a rotational component, where superim- By contrast, the quartz-mica matrix shows
posed stresses form a couple, as in the case of evidence of pronounced mineral alignment and
simple shear regimes. Three mutually perpen- deformation. It therefore follows that there
dicular principal strain axes, X, Y and Z, can must be a strain gradient between porphyrob-
be used to define a strain ellipsoid, where X ~ Y last and matrix (Fig. 9.7). The strain recorded
~ Z (i.e. X is the axis of maximum elongation, in a given metamorphic rock can be evaluated
Z the principal shortening direction and Y the on the basis of textural and microstructural
intermediate axis). Because rocks are highly observation and measurement. Features of
heterogeneous in character, their deformation is different stages in the evolution of the rock
a very uneven process. Different rock types may be recognised, but the nature of the strain
have different rheological properties and there- ellipsoid deduced is that of the bulk finite
fore behave very differently under a given set of strain ellipsoid (i.e. the sum total of all the
conditions. Even on the scale of an individual strain experienced).
rock, certain domains within the rock will be
weaker than others. As a consequence, rock
4.3 Classification of planar fabrics in
deformation is heterogeneous on all scales and
metamorphic rocks
this leads to pronounced strain partitioning.
When considering strain experienced by a As well as compositional layering or banding,
given metamorphic rock, we can examine most deformed metamorphic rocks exhibit
aspects of bulk strain, but on closer inspection some kind of preferred orientation of
microdomains of different strain can be identi- constituent grains. Such structures result from
fied. In a garnet-mica schist for example, the alignment of inequidimensional grains (i.e.
garnet porphyroblasts typically have sub- GRAIN-SHAPE FABRIC) or else alignment of
rounded to euhedral form, and with the excep- crystal structures (CRYSTALLOGRAPHIC
tion of examples showing brittle fracturing, PREFERRED ORIENTATION). In the vast
most garnets show no sign of being deformed. majority of deformed metamorphic rocks,

Layering, banding and fabric development

FIG. 4.3 A hand specimen of augen gneiss (Baltic Shield, Norway), comprising large K-feldspar crystals in a
biotite-rich matrix.

parallel alignment of elongate grains is visible tion and evaluation of the various cleavage
in hand specimen and therefore visible in thin types and other rock fabrics, the reader is
section. They may define a FOLIATION referred to Borradaile et at. (1982) and
(planar structure) or a LINEATION (linear Passchier & Trouw (1996).
structure) (Fig. 4.4). Many metamorphic rocks Parallel planes of preferred splitting in a
(e.g. schists and mylonites) are L-S rock are known as CLEAVAGE, and are widely
TECTONITES, and are comprised of a linear developed under conditions of low and
and planar component. The lineation generally medium metamorphic grade. Various types of
lies within the plane of the foliation, and cleavage have been recognised and described by
defines the maximum elongation direction (X- geologists, but by the mid-1970s it was appar-
direction) of the finite strain ellipsoid. The ent that existing classifications of rock cleavage
planar element lies parallel or very close to were rather confused. This was a consequence
parallel with the X-Y plane, and is perpendicu- of a plethora of terms being introduced into the
lar to the principal direction of bulk shortening literature, often poorly defined and used differ-
or compression. The high strains associated ently by different workers, and a mixture of
with ductile shear zones generate strongly foli- terminology based on morphological and
ated rocks termed MYLONITES and PHYL- genetic considerations. In 1976 a Penrose
LONITES. However, these rocks and their Conference on Cleavage attempted to resolve
fabrics will not be dealt with in this chapter, the problem and standardise the terminology in
since they are covered comprehensively in use. Although several classifications were
Chapter 8. The remainder of this chapter deals attempted, each had their problems.
with the classification of cleavage, and the Nevertheless, a subsequent paper by Powell
processes involved in cleavage and schistosity (1979) presented a purely morphological classi-
development, followed by a section on the fication of rock cleavage, which has consider-
processes responsible for banding and layering able merits. An important advantage is that it
in gneisses and migmatites. For further illustra- provides an objective description of all cleavage

Cleavage and schistosity development

cleavage' (Fig. 4.5(b)). For reviews of slaty

cleavage development, reference should be
made to Siddans (1972), Wood (1974) and
Kisch (1991).
Spaced cleavages can be subdivided on the
basis of whether or not there is a pre-existing
planar anisotropy to the rock. Those cleavages
FIG. 4.4 Schematic block diagrams to illustrate the post-dating an earlier planar anisotropy form
difference between S-tectonites (a) with a pronounced CRENULATION CLEAVAGES, and are further
foliation (planar texture), and L-tectonites (b) with a divided into 'discrete' and 'zonal' types (Fig.
pronounced lineation (linear texture). Metamorphic 4.6). Spaced cleavages that have developed
rocks such as schists and mylonites are generally L-S
tectonites, and have both a linear and planar compo- where there is no pre-existing anisotropy are
this heading come 'stylolitic', 'anastomosing',
'rough' and 'smooth' cleavages. In previous
types, without having any built-in genetic literature many of these types would have been
implications. Having established a clear loosely referred to as 'fracture cleavage', but
description of cleavage type based on morpho- this is not recommended, since this term has
logical features, any discussions of the cleav- genetic implications. The spacing of cleavage
age-forming mechanism can follow, without domains vanes from the scale of
fear of confusion about the type of cleavage centimetres/decimetres for anastomosing and
under consideration. stylolitic cleavages, down to 0.1 mm - 1 cm for
Research has established that most cleaved crenulation cleavages (Fig. 4.6(a)), and 0.01-0.1
rocks have a 'domainal' structure, comprising mm for domainal 'slaty cleavage' (Fig. 4.7).
domains with strong mineral alignment (cleav- Studies by scanning electron microscopy have
age domains) separated by domains with a shown that there is in fact a complete gradation
lesser degree of alignment (microlithons). from finely spaced cleavages into continuous
Scanning electron microscopy has established cleavages, with domainal features still recognis-
that even the most finely cleaved slates are able on the scale of 5-10 pm. At moderate and
commonly domainal. Such cleavages have been high grades of metamorphism where active
termed SPACED CLEAVAGES, while those deformation is involved, the preferential growth
with apparently no domainal features are and alignment of medium- and coarse-grained
termed CONTINUOUS CLEAVAGES. This inequant minerals (e.g. chlorite, micas and
represents the most basic subdivision of cleav- amphiboles) parallel to the X-Y plane of the
ages on a morphological basis. strain ellipsoid leads to the development of a
'Continuous' cleavages are those with a planar fabric termed schistosity.
penetrative appearance in hand-specimen. They
can be further subdivided into 'coarse' and 4.4 Processes involved in cleavage and
'fine', on the basis of mean grain size. Under schistosity development
the Powell classification, the term SCHISTOS-
ITY (Fig. 4.5(a)) would be classified as a When argillaceous sediments are deposited, a
'coarse continuous cleavage'. In sub-greenschist primary fabric may develop sub-parallel to
facies pelitic rocks of orogenic metamorphism, bedding due to the preferential alignment of
the fine-scale splitting referred to as SLATY minerals, especially phyllosilicates such as clays
CLEAVAGE is classed as a 'fine continuous and detrital micas. When the sediment becomes

Layering, banding and fabric development

FIG.4.5 (a) Schistosity, defined by aligned hornblende (dark crystals) plus feldspar and quartz (white). Schistose
amphibolite, Norway. Scale = 1 mm (PPL). (b) A fine continuous cleavage, pervasively developed in the Skiddaw
Slates. Contact aureole of the Skiddaw Granite, Lake District, England. Scale = 1 mm (PPL). Note the ghost-like
oval pseudomorphs of cordierite, and the fact that the cleavage (defined by biotite, muscovite and elongate
quartz) is slightly oblique (clockwise) to the horizontal compositional layering.

compacted and lithified during burial, this primary fabrics are insignificant and all major
primary fabric may be enhanced. This accounts fabrics developed are secondary in origin, and
for the fissile nature of undeformed and caused principally by deformation of the rock.
unmetamorphosed mudrocks such as shales. The relative importance of the main processes
However, for most metamorphic rocks, involved - namely, (i) mechanical rotation of

Cleavage and schistosity development

FIG. 4.6 (a) Zonal crenulation cleavage developed in a pelitic schist. The Sl fabric trends top left to bottom
right, and is overprinted by an S2 crenulation cleavage at 90° to this. The separation into phyllosilicate-rich
domains (P-domains), and quartz-rich domains (Q-domains) is clearly observed. Mica schist, Ox Mountains,
Ireland. Scale = 1 mm (XPL). (b) Discrete crenulation cleavage developed in a fine-grained schist. Dark pressure
solution seams defining S2 trend top to bottom and overprint a crenulated Sl fabric trending left to right. Semi-
pelitic schist/phyllite, unknown locality. Scale = 0.5 mm (PPL).

pre-existing grains, (ii) deformation and ductile combination of processes that operate (Knipe,
flow of individual crystals and (iii) dissolution 1981). The key factors controlling fabric devel-
and new mineral growth - has been and still is opment in metamorphic rocks are (i) rock
the focus of much debate with regard to cleav- composition, (ii) P- T conditions, (iii) stress
age formation. In many cases it is likely to be a orientation and magnitude, (iv) strain rate, (v)

Layering, banding and fabric development

FIG. 4.7 An SEM photograph (back-scattered electron image) of domainal 'slaty cleavage' in slate (Luss,
Scotland). Scale bar (lower centre) = 10 ~m. Note the textural similarity to Fig. 4.6(a), but the difference in scale.

the amount of fluid present and (vi) fluid lar to the principal shortening axis (Z) and thus
composition. In the model of Knipe (1981), it defines the X- Y plane of the finite strain ellip-
is envisaged that mechanical rotation domi- soid. As strain increases so the fabric intensi-
nates the initial stages of cleavage development, fies, and where there is progressive increase in
possibly accompanied by solution proceses and bulk strain the cleavage planes converge. The
grain-boundary sliding (Chapter 8). Later commonly observed phenomenon of cleavage
stages of cleavage development are interpreted refraction displays this feature well, especially
as a more complex and heterogeneous interac- in beds that show grading from psammite to
tion of deformation and metamorphic growth pelite. In Fig. 4.8, note the change in orienta-
processes. The development of new phyllosili- tion and curved nature of the main fabric as
cates (e.g. phengite and chlorite) synchronous the cleavage passes from the less strained, more
with deformation is an integral part of the psammltlc unit, into the more intensely
cleavage-forming process. deformed metapelitic layers in this metamor-
To form a recognisable cleavage generally phosed sequence of clastic sediments. Assuming
requires that rocks of appropriate composition regional strain rates during cleavage formation
have experienced at least 20-30% shortening. of the order of 10- 14 S-I, it has been estimated by
For a pronounced slaty cleavage, more substan- Paterson & Tobisch (1992) that formation of a
tial shortening is required. In many cases it has regional cleavage would take 2-4 Ma. At faster
been estimated that typical shortening associated strain rates, such as those operating in shear
with a regionally pervasive slaty cleavage is zones (e.g. 10- 12 S-I), an intense fabric could
around 60-75%. The fabric forms perpendicu- form in < 40 000 years.

Cleavage and schistosity development

As stated by Passchier & Trouw (1996), recently, Mancktelow (1994) presents evidence
cleavage differentiation by solution transfer from a range of classic areas to suggest that
depends on a substantial amount of intergranu- marked bulk volume change is not a prerequi-
lar fluid in order to be effective, and conse- site for the development of crenulation cleav-
quently is most significant as a process at low age. Clearly, the topic remains a matter of
metamorphic grades. During cleavage forma- current debate, with no unanimous view.
tion in sub-greenschist facies rocks there is good On the basis of detailed electron microprobe
evidence (e.g. marker veins and oolites) of studies and SEMffEM work, it is now clear
extensive dissolution and up to 50% volume that even in penetratively cleaved slates
loss (Fig. 4.9). The intensity of cleavage devel- distinctly recognisable microdomains are often
opment will be a key factor, but for pressure present. In crenulation cleavages, the existence
solution cleavage development, 30-50% of phyllosilicate-rich domains (P-domains), and
volume loss is a typical estimate given by many quartz-rich domains (Q-domains) has long been
researchers. At higher metamorphic grades the known. The development of this distinctive
evidence for substantial volume loss is less clear- 'domainal' or 'zonal' crenulation cleavage (Fig.
cut, and wide-ranging estimates have been 4.6(a)) commences with initial microfolding of
made. The majority of metamorphic petrolo- an earlier formed cleavage or schistosity to give
gists have always viewed volume loss at higher a series of gentle crenulations in the rock.
metamorphic grades to be very low (or zero), Depending on the relative importance of pure
because of limited fluid presence. However, shear and simple shear, the crenulations devel-
authors such as Bell & Cuff (1989) have oped can range from upright and symmetrical
suggested that as much as 50% volume loss to overturned and asymmetrical. The initial
may occur during differentiated crenulation stage is dominated by kinking and bending of
cleavage development by dissolution and solu- phyllosilicate minerals, but as cleavage develop-
tion transfer in phyllosilicate-rich rocks at ment progresses, instability of the pre-existing
greenschist and amphibolite facies. More phyllosilicates leads to the crystallisation of new

FIG. 4.8 Cleavage refraction in slates with varying quartz: phyllosilicate ratios (Bovisand Bay, Devon, England).
The lens cap is 50 mm in diameter and located on one of the more pelitic (phyllosilicate-rich) layers.

Layering, banding and fabric development

phyllosilicates parallel to the crenulation axial istry and mineralogy may remain largely the
surfaces. This becomes the dominant process of same, this redistribution of material by solution
fabric development, whereas the significance of transfer leads to significant changes on the
mechanical rotation of old grains is greatly domain scale. In other cases bulk rock chem-
diminished. The alignment of new phyllosilicate istry may be significantly modified by transfer
grains is the first clear sign of the newly devel- of more soluble material (e.g. silica) out of the
oping crenulation cleavage fabric. Further local system by pressure solution along cleavage
increase in strain tightens the hinge-angle of the surfaces. In such cases, rather than distinct P-
crenulations and leads to an intensification of domains developing, thin dark cleavage 'seams'
the new fabric by preferential nucleation of form (Fig. 4.9). These seams contain insoluble
phyllosilicates in the limbs (P-domains). At this carbonaceous (graphitic) material and phyllosil-
stage of development, significant strain and icates (Gray, 1979; Gray & Durney, 1979).
chemical potential gradients exist between Such spaced cleavage is known as 'discrete'
'hinge' and 'limb' regions of the crenulations crenulation cleavage (Fig. 4.6(b)). Displacement
(the limbs becoming sites of higher strain and of pre-existing compositional markers and veins
chemical potential compared to the hinges). gives the impression of microfaulting, but this is
This leads to soluble minerals of the limbs (e.g. not the case, the offset simply representing
quartz and calcite) preferentially entering into major volume loss of material. Compared to
solution, and being transferred down chemical 'zonal' crenulation cleavages which are
potential gradients, via grain-boundary fluid, to common in both greenschist and amphibolite
sites of deposition in the hinge regions. Gray & facies rocks, the development of 'discrete'
Durney (1979) established a mineralogical crenulation cleavage is usually restricted to
order in which, in terms of decreasing mobility greenschist and sub-greenschist rocks. During
by solution transfer, Cal> Qtz > Feld > ChI> Bt polyphase deformation and metamorphism, an
> Ms > opaques. Although the bulk rock chem- early fabric may be virtually obliterated, or

FIG. 4.9 Displacement and partial dissolution of marker horizons in crenulated phyllite. A quartz-rich horizon
(pale) defining the original lamination (left to right) is cut by vertical pressure solution seams (thin black lines)
related to a secondary crenulation cleavage. Area a-b shows displacement and partial dissolution, whereas c-d
shows more pronounced dissolution. Phyllite, unknown location, UK. Scale = 1 mm (PPL).

Layering in gneisses and migmatites

completely transposed by a later intense fabric. 4.5 Processes involved in formation of

In such cases, the only remaining evidence of layering in gneisses and migmatites
the early fabric may be in the form of inclusion
trails trapped in porphyroblasts, or in less
intensely deformed areas in the porphyroblast
4.5.1 The nature and origin of gneissose
strain shadows. The full approach to interpreta-
tion of porphyroblast-foliation relationships Gneisses and migmatites both reflect high-grade
and deciphering of polydeformed metamorphic metamorphism. Gneisses are characteristically
rocks is dealt with in Chapters 9 & 12. coarse-grained and comprise alternating felsic
Although mechanical rotation of grains and mafic layers on a scale of centimetres to
occurs to some extent in rocks of low meta- metres. They are strongly recrystallised, and
morphic grade, it is not an important process display a granoblastic texture. Compared to
in the development of schistose fabrics of the pronounced schistosity of schists, the fabric
higher-grade rocks. At these deeper levels in the of most gneisses is often weaker (Fig. 4.10),
crust, with elevated temperatures (and pres- especially in gneisses with a low mica content.
sures), crystal-plastic processes associated with Gneisses that had a sedimentary protolith are
ductile deformation become dominant in the referred to as paragneiss, whereas those that
modification of rock fabric. Dislocation creep were originally igneous rocks are referred to as
gives rise to grain boundary migration and orthogneiss.
intragranular movement along certain crystal- The origin of gneissose banding has been a
lographically controlled slip planes. The topic of considerable debate over the decades.
general process of recrystallisation, both during Even the distinction between schists and
and after deformation, plays an important role gneisses is not always clear, since schist
in the development of both grain-shape fabrics sequences often show compositional inter-
and crystallographic preferred orientations. layering on the centimetre to metre scale, and
Continual recrystallisation and equilibration of gneisses may have a schistose fabric (alignment
matrix minerals during regional metamorphism of inequant minerals) as well as displaying
will tend to enhance the dimensional preferred gneissic banding. An interlayered sequence of
orientation. During grain coarsening from slate pelitic schists, semi-pelitic schists and psam-
to schist, those crystals most favourably mites is commonly seen to pass gradationally
oriented with respect to the prevailing stress up metamorphic grade to take on a progres-
will grow more rapidly than those with less sively more gneissic or migmatitic appearance.
favourable orientation, and in so doing will So what are the key processes that give rise to
generate a pronounced alignment of grains to the development of gneissose banding? There
define a schistosity. The supply of material for will of course be some original anisotropy,
nucleation and growth of new grains involves probably some original compositional layering
diffusive mass transfer from reactants via a in the rock. Added to this, metamorphic
grain boundary fluid. The process of processes such as diffusional creep become
MIMETIC GROWTH, which involves growth significant as temperature increases. Compared
of new minerals in the direction of a pre-exist- to migmatites (discussed below), which largely
ing fabric, further enhances the fabric of the involve in situ partial melting, the transforma-
rock, and is common at moderate and high tions leading to the development of gneissic
temperatures as a late-stage process in orogenic banding are for the most part melt-absent
metamorphism. Further details on nucleation solid-state processes. This sounds simple
and growth are given in Chapter 5. enough, but in practice, as with the distinction

Layering, banding and fabric development

between schist and gneiss, the distinction foliated host rock. Lucas & St-Onge (1995)
between migmatite and gneiss is not always argue for 'granite' vein emplacement into layer-
clear, because there is a transition in both the parallel extension fractures, rather than rota-
appearance and the processes operating. For tion of pre-exIstmg oblique veins into
many gneisses, it has been suggested that meta- parallelism, on the basis that the layer-parallel
morphic differentiation in the solid state is one 'granitic' veins are considerably less strained
of the key processes (e.g. Robin, 1979). This is than the surrouding host rock. However, they
achieved by reaction and diffusion of material also point out that the 'gneissose' banding is
down chemical potential gradients. However, subsequently enhanced by later deformation.
others (e.g. Myers, 1978; Kmner et al., 1994) From the above discussion, it is apparent
would argue that diffusion is a much less that there are many possible models for the
important process in the development of gneis- origin of banded gneisses. Original composi-
sic banding, and that in many cases the band- tion layering is often an important element,
ing results from intense shear of original and intense shearing at high temperatures
compositional layering and early oblique dykes usually plays a part in enhancing the banding.
and veins. In some - or possibly many - cases, However, to advocate a single model to
gneISSIC banding may represent sheared account for all banded gneisses would be
migmatites! unwise, since other processes such metamor-
Lucas & St-Onge (1995) studied phic differentiation and emplacement of melt
Precambrian granulite facies banded rocks of may also be involved to some degree.
the Ungava Peninsula, Canadian Shield. They Therefore, before deciding on the likely origin
interpreted the cmldm-scale banding of of a particular banded gneiss it is vital that the
tonalites, quartz diorites and monzogranitic gneiss in question has been examined carefully
rocks in terms of broadly layer-parallel in both outcrop and thin section to assess the
emplacement of externally derived monzogran- available evidence for and against the various
ite veins into previously layered and strongly interpretations.

FIG. 4.10 A weak fabric of gneisses. Hbl-Bt gneiss, Ghana. Scale = 1 mm (XPL).

Layering in gneisses and migmatites

4.5.2 The nature and origin of layered selvages around layers and lenses of quartzo-
migmatites feldspathic leucosome is a characteristic
feature of certain migmatites (Fig. 4.12). The
A simple but useful summary of the main boundary between such melanosome selvages
processes that may operate during migmatisa- and the mesosome is commonly gradational.
tion is provided in Table 4.1 (after Ashworth, The development of these selvages results from
1985). One or more of these processes may melt segregation within migmatites, and is not
operate, and all have been suggested as seen in gneisses and lower-grade metamorphic
processes responsible for the formation of rocks. This relationship can be viewed on the
migmatites. Evidence suggests that in situ scale of an individual thin section, but most of
partial melting (anatexis) is responsible for the the key migmatite relationships are best deter-
formation of most migmatites, or at least that mined by field studies of individual exposures,
is the interpretation currently favoured by most followed by thin-section studies of specific
petrologists. On the basis that most (though igneous and metamorphic components of the
not all) migmatites involve melt, they can migmatite. In the presence of an aqueous fluid,
largely pe defined as coarse-grained heteroge- melting of high-grade schists and gneisses may
neous rocks, characteristically with irregular commence in favourable (quartz-feldspar-
and discontinuous interleaving of melt-derived mica) rocks at temperatures as low as 640°C,
leucocratic granitoid material (leucosome) and but extensive melting and migmatization typi-
residual high-grade metamorphic rock (restite), cally occurs at 670-750°C (upper amphibolite
also referred to as mesosome. The leucosome facies and the innermost part of contact aure-
originates by partial melting of the high-grade oles). Johannes & Gupta (1982) and Johannes
metamorphic rock, and in many migmatites an (1988) describe the migmatisation of a layered
accompanying dark coloured component paragneiss by progressive melting of individual
(melanosome) is also present. Of the many layers. Because of the high viscosity of the
structural types of migmatites recognised in the granitoid melt, in the absence of open frac-
classic work of Mehnert (1968), it is stromatic tures, the melt has difficulty migrating away
migmatites and those with schlieren texture from the area in which it formed. It tends to
that are relevant to this chapter on layering, accumulate at boundaries between layers, but
banding and fabrics in metamorphic rocks. will also migrate into localised areas of lower
Stromatic migmatites are those that have a pressure such as boudin necks and shear
pronounced layering (e.g. Maal0e, 1992). zones. In most cases the leucosome is repre-
These are common in migmatites that have sented as discontinuous layers and lenses.
experienced only moderate degrees of partial Once the melt has migrated to some degree, the
melting (metatexites). Schlieren texture is the boundary between leucosome and other
term used for migmatites that possess streaks portions of the migmatite is usually clear, but
or elongate segregations of non-leucosome in areas of melt generation the distinction
material (usually biotite-rich) in leucosome between the leucosome and the refractory
(Fig. 4.11). This is especially common in residue (the 'mesosome') often appears much
migmatites that have experienced extensive more diffuse and nebulous. Because it is a rock
partial melting and are leucosome dominated formed from a melt, the granitoid leucosome
(diatexites). The schlieren represent entrained often lacks any significant alignment of miner-
restite that has not been entirely separated als to define a fabric. It should also display a
from the melt. random arrangement of constituent mineral
The presence of biotite-rich melanosome phases, with little or no tendency to segregation

Layering, banding and fabric development

FIG. 4.11 A field photograph of schlieren-texture migmatite (St. Jacut, Brittany, France). The coin is 22 mm in

FIG.4.12 A field photograph showing Bt-rich melanosome selvages in metatexitic migmatite (St. Jacut, Brittany,
France). The lens cap is 50 mm in diameter.

TABLE 4.1 The processes involved in migmatisation (after Ashworth, 1985).

Process requires open Process does not require

system open system
Process requires presence of Magmatic injection Anatexis
a melt
Process does not require Metasomatism Metamorphic differentiation
presence of a melt


or clustering of particular mineral phases of a migmatite. Contributions to Mineralogy and

(Ashworth & McLellan, 1985). This is true of Petrology, 79, 14-23.
Kisch, H.J. (1991) Development of slaty cleavage and
well-preserved migmatites, but in a large degree of very low-grade metamorphism: a review.
number of cases the original migmatite Journal of Metamorphic Geology, 9, 735-750.
textures have been modified by subsequent Knipe, R.J. (1981) The interaction of deformation and
deformation and metamorphism. In such metamorphism III slates. Tectonophysics, 78,
cases, the leucosome may also display a strong Kroner, A., Kehelpannala, K.V.W. & Kriegsman, L.M.
fabric, and may show mineral segregation. (1994) Origin of compositional layering and mecha-
Many migmatites also display intense and nism of crustal thickening in the high-grade gneiss
irregular folding. There is often a lack of terrain of Sri Lanka. Precambrian Research, 66,
continuity of folds, with many appearing Lucas, S.B. & St-Onge, M.R. (1995) Syn-tectonic
intrafolial (rootless) or dismembered. Intense magmatism and the development of compositional
and irregular veining, including ptygmatic layering, Ungava Orogen (northern Quebec,
veins, is another feature of many migmatites. Canada). Journal of Structural Geology, 17,
For further insight into migmatites and the Maaloe, S. (1992) Melting and diffusion processes
processes responsible for their formation, the in closed-system migmatization. Journal of
text edited by Ashworth (1985) is strongly Metamorphic Geology, 10,503-516.
recommended. Mancktelow, N.S. (1994) On volume change and mass
transport during the development of crenulation
cleavage. Journal of Structural Geology, 16,
References 1217-1231.
Mehnert, K.R. (1968) Migmatites and the origin of
Ashworth, J.R. (ed.) (1985) Migmatites. Blackie, granitic rocks. Elsevier, Amsterdam.
Glasgow, 302 pp. Myers, J.S. (1978) Formation of banded gneisses by
Ashworth, J.R. & McLellan, E.L. (1985) Textures, in deformation of igneous rocks. Precambrian
Migmatites (ed. J.R. Ashworth). Blackie, Glasgow, Research, 6,43-64.
Ch. 5, 180-203. Park, R.G. (1989) Foundations of Structural Geology,
Bell, T.H. & Cuff, C. (1989) Dissolution, solution- 2nd edn. Blackie, Glasgow, 148 pp.
transfer, diffusion versus fluid flow and volume loss Passchier, C.W. & Trouw, R.A.J. (1996)
during deformation/metamorphism. Journal of Microtectonics. Springer-Verlag, Berlin, 289 pp.
Metamorphic Geology, 7,425-447. Paterson, S.R. & Tobisch, O.T. (1992) Rates of
Borradaile, G.J., Bayly, M.B. & Powell, C.McA. (eds) processes in magmatic arcs: implications for the
(1982) Atlas of deformational and metamorphic timing and nature of pluton emplacement and wall
rock fabrics. Springer-Verlag, Berlin, 550 pp. rock deformation. Journal of Structural Geology,
Gray, D.R. (1979) Microstructure of crenulation cleav- 14,291-300.
ages: an indicator of cleavage origin. American Powell, C.McA. (1979) A morphological classification
Journal of Science, 279, 97-128. of rock cleavage. Tectonophysics, 58, 21-34.
Gray, D.R. & Durney, D.W. (1979) Crenulation cleav- Robin, P.-Y.F. (1979) Theory of metamorphic segrega-
age differentiation: implications of solution-deposi- tion and related processes. Geochemica et
tion processes. Journal of Structural Geology, 1, Cosmochimica Acta, 43, 1587-1600.
73-80. Siddans, A.W.D. (1972) Slaty cleavage, a review of
Johannes, W. (1988) What controls partial melting in research since 1815. Earth Science Reviews, 8,
migmatites? Journal of Metamorphic Geology, 6, 205-232.
451-465. Wood, D.S. (1974) Current views of the development
Johannes, W. & Gupta, L. (1982) Origin and evolution of slaty cleavage. Annual Review of Earth and
Planetary Sciences, 2, 369-401.

Chapter five

nucleation and

5.1 Nucleation
Crystallisation of metamorphic rocks in
response to changing P-T conditions requires
crystallisation of minerals by nucleation and
growth. Homogeneous nucleation, involving
spherical nuclei with uniform surface energy,
randomly distributed throughout the host in
which they develop, may be relevant in chem-
istry and metallurgy, but is an inappropriate
description of nucleation in heterogeneous
polycrystalline metamorphic rocks. Hetero-
geneous nucleation describes non-random
nucleation on some pre-existing substrate, such
as new crystals preferentially nucleating at pre-
existing grain boundaries. Even the purest
monomineralic quartzite or marble has hetero-
geneities such as microfractures or small detri-
tal grains. Because of this there will always be
some places where new crystals are more likely
to nucleate, and other areas where nucleation is
less likely. In view of this, heterogeneous nucle-
ation is the most relevant way to consider
nucleation in metamorphic rocks. Bulk rock
chemistry also plays a key role in influencing FIG. 5.1 Metasediments in the contact aureole of the
the number and size of porphyroblasts that Corvock Granite, Mayo, Ireland, showing widespread
develop. An example from the aureole of the cordierite development. Due to bulk rock chemical
controls on nucleation and growth, porphyroblasts
late Caledonian Corvock granite (Co. Mayo, have preferentially developed to a larger size in the
Ireland) is shown in Fig. 5.1. In this case, pelitic layers compared to the semi-pelitic layers. The
cordierite is preferentially developed as larger lens cap is 45 mm in diameter.

Crystal nucleation and growth

porphyroblasts in pelitic compared to seml- energy, not considered in (5.1) is strain energy
pelitic horizons. (E) stored in elastically strained crystals. Some
With the exception of crystallisation of or all of this energy may be released during
minerals from a melt (e.g. migmatites) and reactions and recrystallisation, and thus reac-
development of minerals in fluid-filled fractures tions may commence earlier and proceed more
or cavities, nucleation and growth of minerals rapidly in aggregates in which strained crystals
in metamorphic rocks essentially occurs in the are present.
solid state. Classical nucleation theory, first Thermal fluctuations will cause embryos to
proposed by Gibbs (1878), has been developed change in size, but once re is exceeded, a
in more detail by various workers and was nucleus is formed and from this a crystal can
neatly summarised by Kretz (1994). It is grow. Authors such as Christian (1975) have
outside the scope of this book to give a detailed developed the classical theory further.
theoretical treatment of the topic but, instead, Considering the activation energy of formation
the key aspects of nucleation will be examined. of a nucleus (LiG*) in terms of LiG y and a, the
If we consider a simple solid-state phase relationship re = -2a/LiG is obtained and,

transformation in which phase A reacts to form when substituted into (5.1), gives
phase B, the classical nucleation theory
* 161t oJ
proposes that once the temperature of the reac- ~G =3 (~Gy' (5.2)
tion has been exceeded, micro domains
(embryos) of phase B will start to nucleate in Activation energy (LiG*) represents an energy
phase A. If it is assumed that the embryos are barrier that must be surpassed before atoms
spherical, the free energy of formation of an
embryo can be written as

(5.1) Surface energy

4lt r 2 0
where r is the radius of the embryo, LiG is the

Gibbs energy change (per unit volume of B) for

the reaction A ~ B (always < 0, because the
reaction always proceeds to lower G) and a (>
i 0 ~~--~~~------~__ radius (r)

0) is the surface (interfacial) energy (per unit l

area) of the A-B interface, which exists because
of surface tension between two phases in 4 3
aGe- '3 It r aG y
contact. Because LiG y is always negative, the + 4ltr 2 0
first term on the right is always negative, and
because a is always positive the second term on
the right is always positive. This means that,
depending on the magnitude of LiG there will

be some critical size (re ), corresponding to FIG. 5.2 The contributions of the Gibbs energy
change per unit volume (~Gvl of B for the reaction A
maximum LiG e (Fig. 5.2), above which further -7 B (lower curve), and the surface (interfacial) energy
growth of the embryo results in a progressively (a) per unit volume of the A-B interface (upper curve),
more stable state. LiG e at re is the activation to the energy of formation of a spherical embryo (r <
energy of nucleation (LiG*) or, in other words, rJ, a nucleus (r = rJ and a crystal (r > re). ~G is the
Gibbs free energy change, and ~G* is the activation
the energy that must be supplied to the system energy that must be overcome for the reaction to
for a nucleus to form. An additional source of proceed (modified after Kretz, 1994).


will transfer freely across the A-B interface and energy and stored elastic strain energy. The
thus promote growth of B. One important reason for preferential nucleation at sites such
implication of the relationships expressed in as grain boundaries can be understood in terms
(5 .2) is that the more the equilibrium position of a simple extension of the homogeneous
of transformation is overstepped (i.e. the more nucleation theory described above. If we take a
negative LiG v becomes), the smaller rc and LiG * spherical embryo of B formed at an A-A grain
become, so that (other things being equal) the boundary that is being consumed, the surface
easier it will be for a nucleus to form. energy term of (5.1) is modified to
Correspondingly, no nucleation will be possible 41tr2(J'A-B -1tr2(J'A-A (5.3)
at equilibrium, since at that point LiG v = 0 and
thus LiG * will be infinite. This modification is valid because the sum of
Because of the heterogeneous nature of poly- energy obtained from the A-A boundary being
crystal aggregates that constitute metamorphic destroyed contributes to the formation of the
rocks, nucleation of new phases is unlikely to embryo and can thus be subtracted from the
show an even distribution. Preferential sites for term for homogeneous nucleation. This means
nucleation include grain boundaries (Fig. 5.3), that for heterogeneous nucleation the surface
vein margins (Plate 8(e)), and on previously energy term is lower than for homogeneous
strained crystals. Nuclei preferentially develop nucleation and consequently LiG c is lowered,
at such sites because there are more loose and thus grain-boundary areas are more
bonds available to attach atoms, and the favourable sites for nucleation. Grain edges
elevated energy of these disordered areas facili- (triple-junctions) and grain corners are even
tates nucleation. The increased dislocation more favourable, as are strained grain bound-
density of strained crystals, particularly at their aries, which have the the added component of
margins, gives rise to an increase in surface stored elastic strain energy associated with

FIG. 5.3 Fibrolitic sillimanite preferentially nucleating at quartz grain boundaries. Connemara, Ireland. Scale =
0.1 mm (XPL).

Crystal nucleation and growth

dislocations. Coherent nucleation (e.g. oriented in which case an even distribution of crystal
intergrowths and topotactic replacement) is sizes might be expected for the new phase.
more favourable than incoherent nucleation, However, for some dehydration reactions the
since lattice-matching across the grain bound- nucleation rate diminishes and can be attrib-
ary reduces 0', and thus the surface energy term uted to a progressive depletion of favourable
in (5.3) is diminished and Lie' is likewise nucleation sites. A variation on this theme,
lowered. proposed by Carlson (1989), is that growing
Nucleation theory predicts that for the porphyroblasts develop a diffusion halo, within
nucleation of phase B from the breakdown of which nucleation is inhibited. Consequently,
phase A, a certain amount of temperature (or the potential for nucleation of new porphyrob-
pressure) overstepping of the A-B equilibrium lastic phases during the later stages of porphy-
boundary is necessary before significant nucle- roblastesis is limited.
ation rates occur (as discussed above in relation
to (5.2)). Once a certain finite amount of over-
5.2 Growth of crystals
stepping has been achieved, nucleation of phase
B will usually start abruptly. Temperature is the From the development of a stable nucleus,
main control, since it provides energy in the further addition of atoms marks the start of
form of heat to drive reactions. However, the growth of a crystal. From this point onwards,
precise amount of overstepping varies from one nucleation and growth become competing
case to the next. A combination of experimen- processes. Crystals formed in a fluid environ-
tal observation and theoretical calculations ment, such as a melt or a fluid-filled fracture or
suggests that for dehydration reactions the cavity, often display perfectly regular (planar)
amount of overstepping required is normally crystal faces. One of the principal mechanisms
< lOoC, but for solid-solid reactions the value by which these faces advance relates to the
in some instances may be as much as 100°e. emergence of screw dislocations (Frank, 1949;
Significant modification of local fluid chem- Griffin, 1950). These are a type of line defect
istry, either by local reaction or by infiltration that displaces part of a plane of atoms and
of some externally derived fluid, can also causes a step in the crystal face, as distinct
induce rapid nucleation. from edge defects, which mark the termination
The above discussion has concentrated on of a plane of atoms (Section 8.2). From these
the nucleation of a single product (B) from a screw dislocations a crystal growth spiral
single reactant (A), but the case of nucleation develops. In detail, the growth of natural crys-
in polyphase aggregates is understandably tals is more complex than this, with surface
more complex. If reaction rates are slow rela- imperfections and surface diffusive processes
tive to atomic mobility, the product(s) may not undoubtedly having a role. In solid-state trans-
form at the site of the reactant(s). This is formations, growth is largely controlled by
commonly seen in the case of the polymorphic interface processes, diffusion processes or a
transformation of andalusite to sillimanite, or combination of these. In the interface model
of kyanite to sillimanite (Section 1.3.3). Rather (Christian, 1975), the rate of advance of the
than the sillimanite nucleating on the precursor A-B interface in the transformation A ---7 B is
Al 2 SiO s phase, it is commonly seen to nucleate largely a function of the Gibbs energy change
at Qtz-Bt, and Qtz-Qtz boundaries (Fig. 5.3), (Lie) and activation energy (Lie''). At equilib-
at some distance from the reactant. rium, or close to it, the growth rate is approxi-
In the case of isothermal nucleation, the mately linear with respect to Lie (i.e. as Lie
nucleation rate could be considered constant, increases, the growth rate increases).

Size of crystals

In cases in which the observed reaction next. The classification that we shall adopt
microstructure exhibits reaction products that here considers rocks with matrix grain size
are chemically different from the reactants, it is <0.1 mm as fine-grained, those with grain size
likely that diffusion of atoms to and from the 0.1-1.0 mm as medium-grained and those
growth surface plays an important role. Indeed, with matrix grain size >1.0 mm as coarse-
it may be the rate-limiting process for crystal grained.
growth in such cases. This type of process is During initial prograde metamorphism,
crucial during exsolution (Section 6.2) and in reaction products are generally fine-grained,
the formation of symplectites (Section 6.4). and much of the original microstructure of the
However, during many metamorphic reactions rock remains clear (e.g. original clastic features
it is likely that a combination of interface and delicate compositional layering (bedding)
processes and diffusive processes are involved in sediments, phenocryst outlines and relation-
in growth. Since metamorphic reactions often ships in igneous rocks). However, with time,
proceed during rising or falling temperature, it and in response to increasing P- T conditions,
is possible that the rate-limiting process may the rock recrystallises further. This leads to an
change with time. Fisher (1978) gave a detailed overall coarsening of the matrix, and the origi-
evaluation of the rate-controlling mechanisms nal microstructural and mineralogical features
for crystal growth over the full range of meta- of the rock are largely or completely obliter-
morphic conditions. It was concluded that ated. This includes the loss of all compositional
'spherical' porphyroblasts pass from an initial layering (laminations) on a scale smaller than
reaction-controlled stage, through an interme- final matrix grain size.
diate grain-boundary diffusion-controlled stage Although monomineralic rocks such as
and finally to a heat-flow-controlled stage. marble and quartzite are typically equigranu-
Computer simulations by Sempels & Raymond lar, as are high-temperature rocks such as
(1980) and simulations compared with natural hornfels and granulite, other rocks such as
samples (Carlson, 1989, 1991) have provided pelitic schists and amphibolites commonly
additional insight into the microstructural and have a structure in which some minerals have
textural features developed during crystal grown to a much larger size than those of the
(porphyroblast) nucleation and growth in matrix. Such minerals are known as PORPHY-
polyphase aggregates. The influence of defor- ROBLASTS, and the structure is termed
mation on nucleation and growth of minerals is PORPHYROBLASTIC (Fig. 5.4). This is anal-
considered in Section 8.3. For further details on ogous to 'phenocrysts' and 'porphyritic struc-
the kinetics of heterogeneous reactions, espe- ture', used to describe a similar feature in
cially in relation to contact metamorphism, the igneous rocks. Many factors contribute to the
review by Kerrick et al. (1991) is highly recom- formation of this structure. They largely relate
mended. to the variable nucleation and growth rates of
constituent minerals of the rock. These in turn
are dependent on factors such as P, T, fluid,
5.3 Size of crystals
rock chemistry and a critical activation energy
It is important when describing any rock or that needs to be overcome for nucleation and
texture to note the overall grain size and varia- growth to occur. Since some minerals nucleate
tions in relative size between the constituent and grow more easily than others, this gives
mineral phases. The precise subdivisions of rise to an heterogeneous porphyroblastic struc-
fine-, medium- and coarse-grained metamor- ture. In particular minerals such as garnet and
phic rocks vary slightly from one author to the staurolite are almost always porphyroblasts,

Crystal nucleation and growth

FIG. 5.4 Porphyroblastic structure. Porphyroblasts of garnet enveloped by a schistose fabric comprised of Chi +
Ms + Bt + Qtz + Hm. Garnet-mica schist, Norway. Scale = 1 mm (PPL).

whereas minerals such as quartz are always Buntebarth & Voll (1991) of quartz coarsening
matrix phases. in quartzites within the Ballachulish contact
In single-phase aggregates (e.g. metals, metamorphic aureole, Scotland.
quartzite and marble) it has long been recog- While coarsening is readily apparent in
nised that crystals coarsen with both time and monomineralic aggregates, it does nevertheless
increased temperature. This coarsening process, also occur in bimineralic and polymineralic
also known as 'Ostwald ripening', occurs in assemblages. However, in such situations the
order to decrease the Gibbs free energy of the processes that occur and the overall kinetics of
system and thus produce a more stable configu- coarsening are considerably more complex. In
ration, closer to equilibrium. This is facilitated pelites for example, the lowest grade, and most
by reducing the total grain boundary surface fine-grained, rocks are slates. These are trans-
area and thus reducing the contribution of formed into phyllites and fine- to medium-
surface (= interfacial) energy to the total energy grained schists at greenschist facies conditions.
of the system. The coarsening is achieved by In turn, these are converted into medium- and
elimination or amalgamation of small grains by coarse-grained schists and gneisses at higher
grain-boundary migration. Although coarsen- metamorphic grades.
ing increases with both time and temperature, Factors such as the metamorphic fluid, diffu-
it has been shown both experimentally and sion rates, the grain-boundary energy (which
theoretically that the rate of coarsening dimin- affects the rate at which a boundary migrates)
ishes as time and temperature increase. Joesten and the rate of change of temperature will all
(1991) gives a comprehensive summary of the be significant in controlling the amount and
theory of grain coarsening and examples of rate of crystal growth. If the nucleation rate is
natural and experimental studies in contact high relative to the growth rate, numerous
metamorphism, including the study by nucleation sites are utilised at an early stage

Size of crystals

following the onset of reaction. This results in growing phase. The rate of interface migration
a fine-grained reaction product disseminated (i.e. the rate at which the crystal faces can
throughout the rock (Fig. 5.5), and may result advance), can also be rate-controlling. If the
in site saturation. The converse of this is when growing porphyroblasts exhaust the supply of
the ratio of nucleation rate to growth rate is reactants, growth will terminate either perma-
small. This results in rapid growth of just a few nently or until such a time as the matrix of the
early nuclei, and often involves the consump- rock is replenished in relevant ions. This
tion or inclusion of many small grains, thus replenishment may occur by external fluid
eliminating many potential nucleation sites in input or else by ionic release into the matrix
the form of grain boundaries, corners, and so system as a result of some other reaction (Fig.
on. The development of a limited number of 5.6). Incomplete diffusion of reactants in the
large porphyroblasts can be considered in this matrix may give a reaction halo around the
way (Fig. 5.4). porphyroblast (Fig. 5.7).
Once nucleated, the ultimate size of a given For a given rock, it might seem reasonable
porphyroblast will be a function of the growth to expect that following some time interval
rate and the time available for growth. The after porphyroblast nucleation and growth,
growth rate is strongly influenced by the rate of the earliest formed porphyroblasts (X) would
diffusive transfer of required elementslions to be the largest, and the most newly formed
and from the reaction site(s), as well as the rate porphyroblasts (Y) the smallest (Fig. 5.8(a)).
at which ions can be incorporated into the However, this assumes constant growth rates

FIG. 5.5 A high nucleation rate relative to the growth rate has produced thin layers with numerous small spes-
sartine garnets in this blueschist facies meta-ironstone. The acicular and lath-like crystals with variable orientation
are of stilpnomelane, developed as a late-stage overprint. The white background is quartz. Blueschist facies meta-
ironstone, Laytonville Quarry, California. Scale = 0.5 mm (PPL).

Crystal nucleation and growth

PorphyfobIaIt virtue of the faster growth rate, the

SIze later formed porphyroblast will become
larger than the earlier porphyroblast (Fig.
In a comparison of grain size variations in
basaltic hornfels and garnet-mica schists over a
range of metamorphic grades, Cashman &
Tme (t)
Ferry (1988) established that the rocks from
o Supply of reec:tants contact aureoles have linear crystal size distrib-
utions (i.e. the smallest sized crystals were most
• SWt of new growth
numerous, the largest crystals least numerous
and there was a linear relationship inbetween).
FIG. 5.6 A schematic illustration to show the influ-
ence of growth rate, time and supply of reactants on Such a relationship is indicative of continuous
porphyroblast size. For simplicity, constant growth nucleation and growth of crystals. By contrast,
rates are shown, but in practice many porphyroblasts regionally metamorphosed pelites have 'bell-
experience an initial phase of rapid growth followed by shaped' crystal size distributions. In other
a period of slower growth.
words, the most numerous crystals are not the
smallest grain size or the largest grain size, but
have some intermediate value. The data from
rocks of the chlorite zone through to the silli-
manite zone all show this relationship, but
rather than being a normal distribution are all
skewed towards the smaller sizes. This distribu-
tion of grain sizes in regional metamorphic
rocks is interpreted in terms of initial continu-
FIG. 5.7 A schematic illustration (width approxi- ous nucleation and growth of crystals followed
mately 1.5 mm) of reaction halo developed around a by a period of annealing after the cessation of
porphyroblast. This occurs due to incomplete diffusion nucleation. This annealing by Ostwald ripening
within the rock, the material required for porphyrob- leads to amalgamation of smaller crystals and
last growth only being derived from the local environ-
ment immediately around the porphyroblast, rather hence modification of the initial linear crystal
than being diffused through the whole rock. size distribution. The observed differences are
readily explained in terms of the differences in

for all porphyroblasts of the same phase, and

need not necessarily be true. For example, the (a) (b)
most favourable site for nucleation may not
necessarily be the most favoured site for Sile Siu
growth. If diffusion rates are variable within (d) (d)

the rock, perhaps as a result of small-scale

compositional variations, then early formed
porphyroblasts may grow more slowly than
later formed porphyroblasts, as a direct result FIG. 5.8 A schematic illustration of the relationships
of variations in the rate of supply and between growth rate, time and porphyroblast size: (a)
variation in porphyroblast sizes at a constant growth
removal of elements involved in the reaction. rate; (b) variation in porphyroblast sizes at different
This being the case, after a certain time, by growth rates.

Shape and form of crystals

thermal history of the two regimes. Contact small age differences (approximately 1-2 Ma)
metamorphism involves high temperatures for between cores and rims. However, on the
a short period of time, while regional metamor- basis of error bar overlap, they conceded that
phism involves high temperature followed by a their results were unable to resolve precisely
prolonged period of cooling. In the latter case, the small time interval indicated.
the rock is held at moderate to high P-T for Considering the various studies of absolute
some considerable time (several tens of Ma). A growth times for porphyroblasts, Barker
final point to be aware of is that the presence (1994) concluded that for garnets d.5 mm
of certain phases may inhibit Ostwald ripening diameter, in situations of orogenic metamor-
of other phases (e.g. graphite inhibits phism, growth times of d Ma and possibly
muscovite growth), so the relationships <0.1 Ma would seem a reasonable estimate.
outlined above are not always as simple as they For larger (e.g. 1-3 cm diameter) garnets, avail-
might seem. able estimates for growth time vary from d
Ma to as much as 5-10 Ma. Such conclusions
relating to porphyroblast growth times have
5.4 Absolute growth times
important implications for the interpretation of
By substituting what they considered to be porphyroblast-foliation relationships (Chapter
reasonable values for equilibrium temperature, 9). For further details regarding rate and time
growth rates and heating rates, Cashman & controls on metamorphic and tectonic
Ferry established that garnet growth times processes based on garnet chronometry, see
(garnets 0.1-1.2 mm in diameter) were in the Vance (1995).
range d00-40 000 years (i.e. geologically very
rapid), and that nucleation and growth
occurred at small ,1T. This compares
5.5 Shape and form of crystals
favourably with the simplified theoretical The form of the constituent minerals of
models of Walther & Wood (1984), which metamorphic rocks can be described as EU-
suggest that porphyroblast growth times during HEDRAL, SUBHEDRAL or ANHEDRAL.
regional metamorphism were probably of the EUHEDRAL crystals are those with good crys-
order of 104-10 5 years. A more recent evalua- tal form and well developed crystal faces (Fig.
tion by Paterson & Tobisch (1992), utilising 5.9(a)), while SUBHEDRAL crystals are less
estimates by Ridley (1986) for metamorphic well formed but have some well developed
mineral growth (2 x 10-5 cm ye l ) suggests faces (Fig. 5.9(b)) and ANHEDRAL crystals
that porphyroblasts formed during regional have irregular form with no well developed
metamorphism could attain lengths of 5 cm in crystal faces (Fig. 5.9(c)). The controlling
250 000 years. factors on whether a given mineral has eu-
Radiometric dating provides an additional hedral or anhedral form are many, but the
means of estimating porphyroblast growth influence of growth kinetics is paramount. The
times. Christensen et al. (1989) undertook development of euhedral porphyroblasts is
Rb-Sr core-rim dating of 3 cm diameter most favoured by conditions of unimpeded
garnets from Vermont (USA) and determined slow growth in an anisotropic medium,
growth times of 6-10 Ma, indicating a mean whereas anhedral crystals commonly reflect
radial increase of 1.1-1.7 mm Ma- 1 . Sm-Nd, rapid growth.
U-Pb and Rb-Sr dating of 1.0-1.5 mm Considering euhedral crystals in a little more
garnets in Caledonian schists from Norway detail, the shape and number of faces depends
(Burton & O'Nions, 1991) demonstrated on certain properties specific to the gIven

Crystal nucleation and growth



FIG. 5.9 (a) EUHEDRAL (or IDIOBLASTIC) porphyroblast of spessartine garnet. Porphyroblastesis was a late
event in this rock, the garnet clearly overgrowing Qtz-Chl-Serc veinlets, and a slaty matrix with fine-scale crenu-
lations. The rock is a pelite that has experienced low-grade regional metamorphism overprinted by later contact
metamorphism. Isle of Man, England. Scale = 0.5 mm (PPL). Note the star-like arrangement of inclusions in the
porphyroblast due to their concentration at interfacial boundaries (for further explanation, see Section 6.1, Fig.
6.4 & Plate l(d)). (b) A SUBHEDRAL porphyroblast of almandine garnet, enveloped by a well-defined regional
schistosity. Garnet-mica schist, Norway. Scale =0.5 mm (PPL).

Shape and form of crystals

mineral phase, and on the growth process distance from a given crystal face to the crystal
involved. The mineral will attempt to maintain centre (d) divided by the surface free energy (y,
the lowest-energy form, and this is controlled of that given face:
by the inherent surface energy, as well as the
lattice energy of the crystal. If all faces had the -d =-d = ... =-do = constant.
1 2
same surface energy, then numerous faces Yl Y2 Yo
would develop, and the crystal would approxi- This expression is now generally referred to as
mate to a sphere. However, in the ideal form of Wulff's theorem. In the schematic example
all minerals, certain faces have lower surface shown in Fig. 5.10, face B grows more rapidly
energy than others, and these preferentially than face A, and it can be seen that the slow-
develop in the equilibrium shape, despite giving growing lower-energy faces predominate.
a greater surface area per unit volume than a Surface defects increase the surface energy of a
sphere. A flat face has a lower energy than an given face such that a low-energy (slow-grow-
irregular surface because the number of ing) face may grow more quickly than
disturbed bonds is less, and of flat faces, those expected. Following original work by Harker
with the greatest density of atoms usually have (1939), metamorphic petrologists have recog-
lower surface energy than those with least nised that certain minerals have a greater
density. Those faces with highest surface energy tendency to develop euhedral form than others,
advance most rapidly, and form a proportion- and are often porphyroblastic. It has been
ately smaller part of the crystal surface. Curie established that a given mineral develops good
(1885) established that for a given crystal there crystal faces when in contact with certain
was a constant relationship between the phases, but not when in contact with others.


FIG. 5.9 (contd) (c) An ANHEDRAL porphyroblast of garnet in a biotite-quartz matrix, with poorly developed
schistosity. Garnet-mica schist, Connemara, Ireland. Scale = 0.5 mm (PPL).

Crystal nucleation and growth

From this, minerals have been arranged in a euhedral form against quartz, but is not
sequence termed the crystalloblastic series expected to develop euhedral form against
(Table 5.1), with those at the top (e.g. pyrite, garnet, except when favourably oriented with
garnet and staurolite) having the greatest respect to the garnet boundaries.
tendency towards euhedral form. The sequence Crystals with SKELETAL, and more rarely
reflects decreasing surface energy, and minerals DENDRITIC, form are sometimes encountered
higher in the sequence will always have a in metamorphic rocks. Both develop as a result
tendency to form euhedral faces against miner- of very rapid growth around a limited number
als lower in the sequence. Therefore, in a of nucleii. True dendritic crystals are uncom-
garnet-mica schist, the garnet will have a mon in geological examples, although crystal-
tendency towards euhedral form against both lites of dendritic olivine in volcanic glasses are
mica and quartz, whereas mica will develop good examples. In metamorphic rocks they are
especially rare because of their high surface
energy and thermodynamic instability.
However, records of dendritic calcite developed
during the epitaxial replacement of aragonite
have been noted. Skeletal crystals form by
rapid mineral growth along intergranular
boundaries under circumstances of
unfavourable nucleation. Truly skeletal
porphyroblasts are not especially common in
metamorphic rocks, but are most frequently
observed in quartz-rich lithologies or segrega-
tions (Fig. 5.11).
ACICULAR (needle-like) crystals (Figs
5.12(a)-(c)), FASCICULAR bundles (Fig.
5.13(a)), BOW-TIE arrangements (Fig. 5.13(b))
FIG. 5.10 A schematic illustration of how certain
faces of a crystal grow more rapidly than others. In this and SPHERULITIC aggregates all result from
example B grows more rapidly than A, and forms a predominance of growth over nucleation.
smaller proportion of the crystal surface. Acicular crystals develop from a single nucleus,
and may occur as scattered individual crystals
or clusters throughout the rock. They are often
TABLE 5.1 The crystalloblastic series of minerals concentrated in specific areas of favourable
(modified after Harker, 1939; Philpotts, 1990). The chemistry or nucleation (e.g. grain boundaries;
sequence reflects decreasing surface energy, and those
minerals higher in the sequence will always have a Fig. 5.3), but may also occur as radiating
tendency to form euhedral faces against minerals lower aggregates (e.g. zeolite minerals in amyg-
in the sequence. daloidal basalts (Fig. 5.12(c)) and tourmaline
Magnetite, rutile, sphene, pyrite, ilmenite
in certain hornfelses). A fascicular growth
Sillimanite, kyanite, garnet, staurolite, chloritoid, consisting of a bundle of rods or needles initi-
tourmaline ates from a single nucleus, but then branches
Andalusite, epidote, zoisite, forsterite, lawsonite slightly. Further divergence of needles will give
Amphibole, pyroxene, wollastonite
Muscovite, biotite, chlorite, talc, prehnite, stilpnomelane
rise to a 'bow-tie' structure. This is common in
Calcite, dolomite, vesuvianite (idocrase) amphiboles of 'garbenschiefer', calc-schists and
Cordierite, feldspar, scapolite meta-volcanic rocks (Fig. 5.13(b)).
Quartz Having considered free growth forms,

Shape and form of crystals

FIG.S.11 A "skeletal' porphyroblast of garnet formed by growth between quartz grain boundaries. Garnet-mica
schist, Norway. Scale = 0.5 mm (PPL).

largely in connection with porphyroblastesis, energy. Unless a grain has experienced appre-
let us now consider the various aspects of ciable internal strain, its boundaries against
mutual growth forms, in which adjacent crys- other grains are generally more atomically
tals impinge on each other. This is mostly disordered than its internal structure. In order
concerned with the process of matrix recrys- to become more stable there is a tendency to
tallisation and adjustment towards equilibrium, reduce the total area of grain boundaries and
but also relates to impingement of growing to develop more stable (regular) crystal faces.
porphyroblasts having nucleated close together. The degree to which an aggregate of grains in a
The lowest-energy system is the most stable metamorphic rock approaches stability is
one, and in order to maintain the lowest chemi- dependent on the time available for adjust-
cal free energy, mineral phases will tend to ments, and on prevailing conditions.
react in response to changing P, T and fluid Non-equilibrium impingement structures
conditions, and hence produce a more stable consist of a .range of grain sizes, irregular and
assemblage. During or after these mineralogical variable grain shapes, curved and irregular
changes, the shapes of grains and grain-bound- boundaries and multiple junctions. The total
ary arrangements will often become modified surface energy in low-grade schists is often
in an attempt to reduce grain-boundary energy. considerably higher than in higher-grade rocks
This energy is considerably smaller than the because of the irregular and often curved
chemical free energy, but is nevertheless signifi- nature of many of the grain boundaries (Fig.
cant and, throughout the microstructural S .14( a)). The minimisation of surface area
changes, the grain boundaries will become during recrystallisation, and the development
modified and rearranged to minimise this of GRANOBLASTIC-POLYGONAL (or

Crystal nucleation and growth



FIG.5.12 (a) ACICULAR (needle-like) crystals of stilpnomelane and deerite developed in a quartzitic layer within
a blueschist facies meta-ironstone sequence. Laytonville Quarry, California. Scale = 0.1 mm (PPL).
(b) RODDED and ACICULAR sillimanite crystals. Sillimanite gneiss, Broken Hill, Australia. Scale = 0.1 mm (PPL).

Shape and form of crystals

MOSAIC) microstructure compnsmg many 120° (:t100). The interfaces are largely regular
nearly planar grain boundaries (Fig. 5 .14(b)), is and planar, the regularity being achieved by
indicative of a high degree of stability, and is surface (interfacial) energy-driven grain-bound-
especially common at higher metamorphic ary migration. This arrangement minimizes
grades. It is commonly observed in quartzites surface area and thus the contribution of
(Fig. 5 .14(b)), hornfelses and granulites, and surface energy to the total free energy of the
the same microstructure is also recognised in system. However, at a triple-point, there will
metamorphosed massive sulphide ores (e.g. only be three angles of exactly 120° if the
Craig & Vaughan, 1994). Such a microstruc- surface energies on all boundaries are the same.
ture could be considered in terms of random This will be approximately true if the three
nucleation and growth during metamorphic grains meeting at the triple-point are all of the
crystallisation, but the fact that concentrations same mineral, but where different minerals are
or clusters of certain minerals occur in many involved, it will not. In a study of granulites
cases suggests that, even at high grades of from Quebec, Kretz (1966) established that the
metamorphism, pre-exlstmg microstructural dihedral angle (B) for clinopyroxene against
and mineralogical heterogeneities may have an two scapolite grains was 128° (s.d.=13°), and
important influence on the location of clinopyroxene against two plagioclase grains
favourable and unfavourable nucleation sites was 109 0 (s.d.=16°). Similar studies have also
for particular minerals. been made for other phases. Another point to
In polygonal aggregates, interfaces generally note in relation to polygonal aggregates - and
meet at triple-points with interfacial angles of for simplicity let us consider a monomineralic


FIG. 5.12 (contd) (c) Radiating acicular crystals of natrolite infilling a vesicle within altered basalt. Antrim,
Northern Ireland. Scale = 0.1 mm (XPL).

Crystal nucleation and growth



FIG. 5.13 (a) A fascicular bundle of stilpnomelane crystals developed in a quartzitic layer within a blueschist
facies meta-ironstone sequence. Laytonville Quarry, California. Scale = 0.5 mm (PPL). (b) A bow-tie arrangement
of actinolitic hornblendes on the schistosity plane of a hornblendic schist hand specimen from Troms, Norway.

Shape and form of crystals



FIG. 5.14 (a) Irregular quartz grain boundaries in greenschist facies semi-pelitic schist. Loch Leven, Scotland.
Scale = 0.5 mm (XPL) . (b) A granoblastic-polygonal aggregate of quartz in a blueschist facies quartzitic rock.
California. Scale = 0.5 mm (XPL). Note the straight grain boundaries and 120° triple-junctions between grains,
indicating a very stable arrangement.

Crystal nucleation and growth

aggregate such as quartz in a quartzite - is that grains that vary from three- to eight-sided in
if all grains are hexagonal in cross-section, 120° cross-section, 1200 triple-junctions can only be
triple-junctions are easily satisfied with straight maintained with curved boundaries (Fig. 5.15).
crystal boundaries, but if the aggregate contains Grains with less than six sides have convex-
outward boundaries, and phases with more
than six sides have convex-inward boundaries.
Curved grain boundaries are also required in
polymineralic equant aggregates. For example,
in a bimineralic aggregate of hexagonal grains,
where phase A is much more abundant than
phase B, grains of phase B entirely surrounded
by phase A will be convex-outward if BABA >
120° and convex-inward if BABA < 120°.
Certain aggregates of minerals develop a
DECUSSATE structure (Fig. 5.16). This is a
special type of granoblastic structure in which
crystals are subhedral, prismatic or flaky, and
randomly oriented, and have a strong crystal
FIG. 5.15 A schematic drawing of a recrystallised anisotropy (i.e. the surface energies of the differ-
polygonal aggregate of grains, to show that in order to
maintain 120 0 triple-junctions in three- to eight-sided ent faces are very different). It is generally asso-
crystals, a certain number of curved boundaries is ciated with monomineralic vein assemblages,
required. monomineralic amphibole- or mica-hornfelses

FIG. 5.16 Decussate structure exhibited by axinite crystals in a vein from the contact aureole of the Bodmin
Granite, Cornwall, England. Scale = 0.5 mm (XPL).


or granofelses, and monomineralic amphibole or quently do not become adjusted by quartz or

mica domains within polymineralic hornfelses feldspar impingement during growth. This
and granofelses. Following impingement, grain results in quartz/quartz interfaces generally
growth proceeds by grain-boundary migration meeting mica(OOl)/quartz interfaces at 90° (Figs
and engulfment of small grains by large ones, to 5.17(a) & (b)). By restricting grain growth of
produce a rational low surface energy aggregate quartz in certain directions, layer silicates will
of subhedral crystals. Equal-angle (120°) triple- influence the size and shape of the quartz grains
junctions between grains tend not to occur. (i.e. quartz will tend to be elongate parallel to
In bi-mineralic and polycrystalline aggregates, aligned micas; see Figs 5.17(a) & (b)).
equilibrium textures and microstructures are
more complex. In quartz-feldspar-mica aggre- 5.6 Twinning
gates (e.g. psammites and semi-pelitic schists),
anisotropic minerals such as mica will tend to 5.6.1 Introduction
dominate the microstructure. The planar 001 Twinned crystals are comprised of two or more
surfaces of micas are very stable, and conse- portions of the same crystal species in high-angle




FIG. 5.17 (a) A schematic illustration showing how micas influence the shape of quartz grains during recrystalli-
sation. (b) A natural example showing how the 001 faces of mica crystals have influenced the shape of quartz
crystals, to make them sub-rectangular. Semi-pelitic schist, Norway. Scale = 0.1 mm (XPL).

Crystal nucleation and growth

contact with each other, and with a rational

symmetry according to specific laws. They are
easily identified in thin section because, due to
their different optical orientation, different
portions of the twinned crystal go into extinc-
tion at different times, or show different inter-
ference colour. Since twin boundaries represent
high-angle contacts, they are usually sharply
defined. SIMPLE TWINS are those comprised
of two units (Plate 2(c)), while MULTIPLE or
POLYSYNTHETIC TWINS (Plate 2(e)) consist
of many. It is important to distinguish between FIG. 5.18 A schematic illustration of the differences
PRIMARY TWINS, which represent twins between (a) enclosed primary twins with abrupt or
stepped terminations, and (b) enclosed deformation
formed at the time of crystal growth, and twins with tapered terminations. (c) Conjugate sets of
SECONDARY TWINS such as those formed deformation twins. (d) An example of deformation
due to crystallographic inversion or in response twins concentrated at a bend in a crystal.
to stress during subsequent deformation.

amphibolite facies but are mostly untwinned.

5.6.2 Primary twins Where twinning is present, it is usually a simple
Primary twins (or growth twins) commonly twin on the Albite Law (Plate 2(c)), which
develop in minerals such as amphiboles, pyrox- suggests that twinning is difficult to produce
enes and feldspars. They can be further subdi- under the temperatures and stresses associated
vided into 'layer growth twins' and 'twinned with low- and medium-grade metamorphism.
nuclei'. For further details on the various types The effect of stress as a restraint on growth
of feldspar twinning, the reader is referred to twins can be argued with reference to
the classic reference on feldspars by Smith & cordierite. In this case, sector twinning (or
Brown (1988). 'sector trilling') in cordierite (see below, and
LAYER GROWTH TWINS (e.g. albite twins Fig. S.19(a)) is commonly encountered in
in feldspar) are those involving the addition of contact metamorphosed rocks where stresses
successive layers to the growing crystal faces in are very low, but is uncommon in cordierite
either the 'normal' or the 'twin' position. Partial schists and gneisses from terrains of orogenic
or enclosed primary twins have abrupt or metamorphism, where stresses at the time of
stepped terminations (Fig. S.18(a)), in contrast growth would be appreciably higher.
to deformation twins (see below), which gener- TWINNED NUCLEI represent a separate
ally have tapered terminations (Figs S.18(b)-(d) type of growth twin, and develop when crystals
& Plate 3(b)). Feldspar, amphibole and pyrox- nucleate in the twinned state. Examples of this
ene all commonly show primary twins in type of twin are the cruciform twins commonly
igneous rocks, but in metamorphic rocks observed in staurolite and chloritoid (Plate
growth twinning in these minerals is less 2(d)). ANNEALING TWINS represent a
common. This can be explained in terms of further type of growth twin, and form during
ongoing deformation at the time of growth, recrystallisation. They are produced when a
inhibiting twin development. Plagioclase migrating grain boundary of one crystal meets
feldspars, for example, are common in metaba- an adjacent crystal with its lattice suitably
SlC and pelitic rocks of the greenschist and oriented. A twin boundary is of lower energy


than a normal boundary and will be developed metamorphic aureoles, sector trilling is rarely
in preference. reported from orogenic metamorphic terrains.
However, interpenetrant sector twinning
certainly occurs in cordierite from some high-
5.6.3 Secondary twins
grade gneiss terrains. The reason for the rather
Two main types of secondary twin (those limited occurrence of sector trilling in
formed after crystal growth) can be recognised cordierite from orogenic metamorphic terrains
in minerals of metamorphic rocks. These are is probably related to the active shearing preva-
INVERSION (or TRANSFORMATION) lent in such environments at the time of
TWINS and DEFORMATION TWINS. cordie rite growth. For details relating to TEM
Inversion twins occur in certain minerals imaging of transformation-induced microstruc-
when changing metamorphic conditions give tures, the reader is referred to the excellent
rise to instability of the original crystal struc- review by Nord (1992).
ture. This ultimately leads to a change in crys- DEFORMATION TWINS are the other
tal habit. CROSS-HATCHED TWINNING of dominant type of secondary twin seen in meta-
microcline (Fig. S.19(b)) is one such example of morphic minerals. They are extremely common
this type of secondary twinning. For igneous in calcite (Plate 3(a)} and dolomite, but
rocks, the traditional interpretation of cross- frequently occur in certain other minerals (e.g.
hatched twinning in microcline has been in plagioclase (Plate 3(b)), and pyroxenes).
terms of inversion from higher-temperature Deformation twins are occasionally simple, but
monoclinic sanidine or orthoclase to lower- are most commonly polysynthetic twins or in
temperature triclinic microcline. In order to conjugate sets (Fig. 5.18 & Plate 3). Unlike
accommodate this crystallographic change 'enclosed' primary twins which have abrupt
from monoclinic to triclinic brought about by terminations, 'enclosed' deformation twins
increasing degrees of AI, Si ordering, a complex have tapered ends (Figs S.18(b)-(d)). Other
network of albite twins (± pericline twins) than this, it may be difficult in minerals such as
develops, to give the characteristic cross- plagioclase to decide whether the twins
hatched twinning of microcline. However, in observed are 'primary' or 'secondary'.
the case of metamorphic rocks, microcline has The mechanism responsible for deformation
been shown to increase at the expense of ortho- twinning has similarities with that required for
clase in rocks showing greatest deformation. translation 'gliding' (or 'slip'), in that both
Therefore, in addition to the control of temper- processes involve displacement of a layer in the
ature, this suggests that superimposed shear crystal lattice relative to its neighbour. However,
stress is another key factor that will cause there is a clear difference in terms of how this is
orthoclase or sanidine to invert to microcline. achieved (Fig. 5.20). Translation gliding (or
SECTOR TRILLING (also referred to as 'slip') involves individual layers of the lattice
'sector twinning' or 'sector zoning') in cordierite slipping past each other but coming to rest with
(Fig. 5.19(a)} has received considerable attention crystal portions on each side of the slip plane
in the literature (e.g. Kitamura & Yamada, being similarly oriented both before and after
1987). It is not a growth twinning feature, but is (Fig. 5.20(a)}. This means that unless some
now widely accepted as developing in response marker (e.g. a primary twin) has been offset,
to a transformation from metastable high- slip planes are virtually impossible to detect
temperature hexagonal cordierite to stable low- optically. Deformation twins, on the other
temperature orthorhombic cordierite. While hand, involve (in simple terms) shear of part (or
frequently observed in hornfelses of contact parts) of the crystal structure (Fig. S.20(b)} by

Crystal nucleation and growth



FIG.5.19 (a) Sector trilling (or sector twinning) in cordierite. This results from the inversion of high-temperature
hexagonal cordierite (indialite) to low-temperature orthorhombic cordierite during cooling. Cordierite Hornfels,
Skiddaw, England. Scale = 0.5 mm (XPL). (b) Cross-hatched twinning of microcline, caused by the inversion from
monoclinic sanidine or orthoclase to triclinic microcline. High-grade gneiss, Ghana. Scale =0.5 mm (XPL).


while twinning occurs over the full range of

(a) m!~ metamorphic temperatures, it will only develop
when the direction of shortening is at a high
angle to the c-axis of the crystal. Experimental
work by Rowe & Rutter (1990) has examined
FIG. 5.20 A schematic illustration of the processes of the possibility of assessing the magnitude of
lattice rearrangement in crystals undergoing deforma- palaeostresses based on the volume fraction
tion: (a) translation gliding; (b) deformation twinning.
and twinning incidence within calcite. They
record that at constant differential stress there
twin-glide rather than offset and slip. This is a linear increase in both twinning incidence
results in a reoriented (adjusted) structure that and volume fraction in relation to original
mirrors the original pattern (i.e. is in the grain size. For a rock of given grain size the
twinned position). Deformation twinning gives volume fraction of twinning is found to
homogeneous deformation in the twinned increase non-linearly with differential stress,
region, but the distribution of twins within a while the incidence of twinning increases
grain is generally heterogeneous (Plate 3(a)). In linearly with stress. These factors are seemingly
a polycrystalline aggregate, differences in the independent of temperature and strain rate.
crystallographic orientation of individual Bearing in mind grain-size considerations, these
grains relative to the superimposed shear stress results suggest that there is considerable poten-
will lead to some grains developing twins more tial for the assessment of palaeostresses in
readily than others. Nucleation occurs at posi- naturally deformed carbonate rocks based on
tions of strain concentration such as disloca- the abundance and orientation of deformation
tions and irregularities on grain boundaries. twins.
Twins will propagate along the favoured
plane(s) (e.g. {0112} in calcite, to form a
5.7 Zoning
continuous (and usually thin) twinned layer
(Plate 3 (a)). At temperatures above 200°C it is A crystal growing by incremental addition to
common to see two or three deformation twin its outer surfaces will be uniform if the units
sets in a single calcite grain. A section in added are of constant composition and orienta-
Shelley (1993) gives a useful summary of the tion. However, if there is a change in either the
principal twin sets that develop in calcite and source reservoir or in the element partitioning
dolomite, and how they can be measured using between the crystal and matrix minerals (due
a Universal Stage ('U-stage'). Recent work by to changing P and/or T) the material added will
Ferrill (1991) and Burkhard (1993) suggests probably change in composition. This will lead
that the geometry of deformation twins in to a crystal the structure of which persists from
calcite may have potential for indicating the core to rim, but which shows gradual or
temperature conditions at the time of deforma- abrupt, continuous or cyclic compositional
tion. zoning (Tracy, 1982). Zoned crystals in igneous
Experimental studies have established that rocks (e.g. plagioclase and pyroxene) can be
deformation twinning is easier in coarse- modelled in terms of incomplete reaction with
grained than in fine-grained rocks. In carbon- a residual magma of changing composition.
ate lithologies it has been shown that there is a Zoning in metamorphic minerals relates to
strong relationship between the orientation of changing composition of the matrix and fluid,
deformation twins in calcite, and the orienta- which in the case of a closed system is directly
tion of the principal stress axes. In calcite, related to the availability of reactants and the

Crystal nucleation and growth

degree of completion of a reaction. At any

given time it is only the rim of the zoned
mineral that is at equilibrium with the matrix.
The preservation of a zoned profile demon-
strates that the interior of the crystal has not
equilibrated under the changing conditions. As
such, this gives considerable potential for gain-
ing an insight into the early stages of the meta- 1mm

morphic history. This has been exploited by

Spear & Selverstone (1983) and others, who
have used variations in the chemistry of zoned 202M·

porphyroblasts and inclusions trapped within 1.3

them to give detailed evaluation of P-T-t

trajectories of individual rocks. A crucial 0.8
assumption to this approach is that the trapped
phases were in local equilibrium at the time of
trapping, and that post-trapping diffusive Mn
processes have not modified the chemistry of 0.2

the inclusion or host. The whole approach is OJ)

invalidated if such assumptions are not true, so

a cautious approach must always be exercised. FIG. 5.21 Chemical profiles (based on electron micro-
Prior to any interpretation of zoned minerals probe analyses) for a typical almanditic garnet in a low
(e.g. plagioclase) in rocks of low metamorphic amphibolite facies garnet-mica schist, from the
grade, the first step should always be to estab- Scandinavian Caledonides of arctic Norway. The
substitution of Fe for Mn from core to rim has given a
lish that they are not detrital phases. chemical zonation pattern typical of many low amphi-
Although not usually recognisable by optical bolite facies garnets, with a characteristic Mn-bell
means, one of the most widespread and exten- profile. Dots on the drawing of garnet analysed repre-
sively studied chemically zoned minerals of sent probed points. Chemical variations are given in
formula amounts of a particular element, on the basis
metamorphic rocks is almandine garnet. It o = 12 (see Fig. 5.22 for an SEM back-scatter image
largely develops from the breakdown of chlorite and X-ray element maps for the same garnet).
and, on the basis of electron microprobe studies,
SEM back-scatter imaging and X-ray element
mapping, it is found that rocks metamorphosed (e.g. pyroxenes (Plate 3(c)), tourmalines (Plate
in the approximate range 450-625°C commonly 3(d)) and amphiboles). Intracrystalline refractive
have garnets that show Mn-Fe zonation from index variation in minerals such as garnet had
core to rim. The characteristic zonation ('Mn- also enabled compositional zonation to be
bell' profile) is comprised of an Mn-rich core recognised.
and an Mn-depleted rim, with a converse The 'growth zoning' process introduced
pattern for Fe (Figs 5.21 & 5.22). At higher above involves reaction partitioning to give a
metamorphic temperatures (approximately > continuous or discontinuous change in compo-
625°C) such primary chemical zonation is wiped sition of material supplied to the growing crys-
out by diffusion throughout the crystal lattice tal surface. It also requires slow volume
(see below). Even before the advent of the elec- diffusion in order that the interior of the crystal
tron microprobe, chemical zonation was realised is isolated and does not equilibrate to new
by virtue of colour zoning in certain minerals conditions. A second type of zoning, known as


FIG. 5.22 (a) An SEM (back-scatter image) of the garnet shown in Fig. 5.21. (b-f) SEM X-ray element maps of
the same garnet, showing chemical variations in the garnet and the matrix. Mn-enrichment in the garnet core
shows up particularly well in (d), and the chlorite alteration (pale grey/white) at the garnet margins shows up well
in (e).

Crystal nucleation and growth

(a)'~n Mg (b)~Mn Mg (C),~Fe I

Ca I I Ca ca
Fe Fe Mn
rim core rim rim core rim rim core rim

FIG. 5.23 Chemical zonation in garnet porphyroblasts. (a) A typical Mn-bell profile of garnet porphyroblasts
that have grown largely at the expense of chlorite. (b) Garnet with an Mn-enriched rim: this commonly occurs as
the result of late-stage diffusion zoning or resorption during retrogression. (c) Garnets that grow under high-grade
metamorphic conditions, or that later experience high-grade metamorphic conditions, are largely unzoned: such
garnets are said to be homogenized and have flat profiles- this is the result of complete intracrystalline diffusion.

'diffusion zoning', may occur in some meta- diffusion for all elements becomes so impercep-
morphic minerals, given the right circum- tibly slow that zonation patterns do not equili-
stances. This is particularly common in garnet brate. At high temperatures, where diffusion is
porphyroblasts of many high-grade metamor- much faster, complete intracrystalline diffusion
phic rocks, and has only been recognised ('volume diffusion') occurs, and gives rise to
following detailed microprobe work. The homogenisation of earlier growth zoning
process involved is one of intra-crystalline profiles. The 'flat' chemical profiles characteris-
diffusion driven by disequilibrium and reaction tic of garnet porphyroblasts of the upper
between the crystal (porphyroblast) surface and amphibolite facies are related to this process
prevailing matrix conditions. Diffusion zoning (Fig. 5.23(c)}. In complex situations associated
is typically imposed on pre-existing crystals with polymetamorphism, discontinuous zona-
rather than being associated with growth. The tion profiles may be recognised, which result
crystal mayor may not have been zoned to from resorption followed by new overgrowths.
start with, and while normally associated with Within the leucosome component of
waning P-T conditions and retrogression it can migmatites, euhedral (idiomorphic) zoning in
also develop during 're-heating'. Diffusion plagioclase is sometimes encountered. This is
zoning involves redistribution of atoms in the interpreted in terms of unimpeded growth
crystal structure and change in the relative zonation in the melt prior to solidification and
abundance of certain atoms. It is usually recog- impingement with surrounding grains
nised by sharp apparently 'reverse' zonation in (Ashworth, 1985).
the outer parts of crystals. In the case of Oscillatory zoning is recorded in some solid-
garnets, Mn-enrichment at the margins is a solution metamorphic minerals, and is best
characteristic diffusion zoning feature (Fig. known from pyroxenes and garnets of skarns
5.23(b)}. This is commonly interpreted in terms (Plate 3(c)}. While the significance of such
of resorption of the garnet edge during retro- zoning in igneous phenocrysts and sedimentary
gression, associated with a biotite- or carbonate cements is well documented, its
cordierite-producing reaction. The extent of significance in metamorphic rocks has received
the resorption is dependent on the rate of cool- limited attention. In an SEM study using back-
ing and on diffusion rates. Since the degree of scattered electron imaging, Yardley et at.
resorption is often small, cooling rates must be (1991) document several examples of oscilla-
fast relative to diffusion. Diffusion rates vary tory zoning in metamorphic minerals. They
from one element to the next under given interpret such zoning as diagnostic of open-
conditions, but below a certain temperature system behaviour during metamorphism, and


suggest that it may provide useful evidence for dynamics of crystallization: III. Metamorphic crys-
metasomatic mineral growth in metamorphic tallization. Contributions to Mineralogy and
Petrology, 99, 401-415.
rocks. The fact that such zoning is preserved Christensen, ].N., Rosenfeld, ].L. & De Paulo, D.].
indicates sluggish intracrystalline diffusion (1989) Rates of tectonometamorphic processes from
through the minerals concerned. rubidium and strontium isotopes in garnet. Science,
As well as chemical zonation, textural zona- 244, 1465-1468.
Christian, ].W. (1975) The theory of transformations
tion may be recognised in many porphyroblasts. in metals and alloys. Pergamon Press, Oxford.
For example, garnet porphyroblasts often have a Craig, ].R. & Vaughan, D.]. (1994) Ore microscopy
heavily included core, but an inclusion-free rim and are petrography, 2nd edn. John Wiley, New
(e.g. Figs 5.22(a) & (b)). This may represent two York, 434 pp.
Curie, P. (1885) Sur la formation des cristaux et sur les
growth stages or, alternatively, may be inter- constantes capillaires de leur different faces. Soc.
preted as rapid initial growth followed by a Mineral. France Bull., 8, 145-150.
slowing of the growth rate. More complex Ferrill, D.A. (1991) Calcite twin widths and intensities
textural zonation occurs during polyphase meta- as metamorphic indicators in natural low-tempera-
ture deformation of limestone. Journal of Structural
morphism and deformation, and is discussed Geology, 13, 667-676.
more fully in Chapters 9 & 12. In many porphy- Fisher, G.W. (1978) Rate laws in metamorphism.
roblasts, breaks in chemical zonation commonly Geochimica et Cosmochimica Acta, 42,1035-1050.
coincide with textural breaks, but in other cases Frank, F.e. (1949) The influence of dislocations on
crystal growth. Discussions of the Faraday Society,
there may be no clear relationship. No. 5,48-54.
Gibbs, ].W. (1878) On the equilibrium of heteroge-
neous substances. Transactions of the Connecticut
References Academy III, 1875-1876, 108-248; 1877-1878,
Ashworth, ].R. (1985) Migmatites. Blackie, Glasgow, Griffin, L.]. (1950) Observation of unimolecular
302 pp. growth steps on crystal surfaces. Philosophical
Barker, A.]. (1994) Interpretation of porphyroblast Magazine, 41,196-199.
inclusion trails: limitations imposed by growth Harker, A. (1939) Metamorphism - a study of the
kinetics and strain rates. Journal of Metamorphic transformations of rock masses. Methuen, London,
Geology, 12, 681-694. 362 pp.
Buntebarth, G. & Voll, G. (1991) Quartz grain coars- Joesten, R.L. (1991) Kinetics of coarsening and diffu-
ening by collective crystallization in contact sion controlled mineral growth, in Contact meta-
quartzites, in Equilibrium and kinetics in contact morphism (ed. D.M. Kerrick). Mineralogical Society
metamorphism (eds G. Voll, ]. Topel, D.R.M. of America, Reviews in Mineralogy No. 26,
Pattison & F. Seifert). Springer-Verlag, Berlin. 507-582.
Burkhard, M. (1993) Calcite-twins, their geometry, Kerrick, D.M., Lasaga, A.e. & Raeburn, S.P. (1991)
appearance and significance as stress-strain markers Kinetics of heterogeneous reactions, in Contact
and indicators of tectonic regime: a review. Journal metamorphism (ed. D.M. Kerrick). Mineralogical
of Structural Geology, 15, 351-368. Society of America, Reviews in Mineralogy No.26,
Burton, K.V. & O'Nions, R.K. (1991) High-resolution 583-67l.
garnet chronometry and the rates of metamorphic Kitamura, M. & Yamada, H. (1987) Origin of sector
processes. Earth and Planetary Science Letters, 107, trilling in cordierite in Diamonji hornfels, Kyoto,
649-67l. Japan. Contributions to Mineralogy and Petrology,
Carlson, W.D. (1989) The significance of intergranular 97, 1-6.
diffusion to the mechanisms and kinetics of porphy- Kretz, R. (1966) Interpretation of the shape of mineral
roblast crystallization. Contributions to Mineralogy grains in metamorphic rocks. Journal of Petrology,
and Petrology, 103,1-24. 7,68-94.
Carlson, W.D. (1991) Competitive diffusion-controlled Kretz, R. (1994) Metamorphic crystallization. John
growth of porphyroblasts. Mineralogical Magazine, Wiley, Chichester, 507 pp.
55,317-330. Nord, G.L. Jr. (1992) Imaging transformation-induced
Cashman, K.V. & Ferry, ].M. (1988) Crystal size microstructures, in Minerals and reactions at the
distribution (CSD) in rocks and the kinetics and atomic scale: transmission electron microscopy (ed.

Crystal nucleation and growth

P.R. Buseck). Mineralogical Society of America, under the microscope. Chapman & Hall, London,
Reviews in Mineralogy No. 27, 455-508. 445 pp.
Paterson, S.R. & Tobisch, O.T. (1992) Rates and Smith, ].V. & Brown, W.L. (1988) Feldspar minerals.
processes in magmatic arcs: implications for the Springer-Verlag, Berlin.
timing and nature of pluton emplacement and wall Spear, F.S. & Selverstone, ]. (1983) Quantitative P-T
rock deformation. Journal of Structural Geology, paths from zoned minerals: theory and tectonic
14,291-300. applications. Contributions to Mineralogy and
Philpotts, A.R. (1990) Principles of igneous and meta- Petrology, 83, 348-357.
morphic petrology. Prentice-Hall, Englewood Cliffs, Tracy, R.]. (1982) Compositional zoning and inclu-
New Jersey, 498 pp. sions in metamorphic minerals, in Characterization
Ridley, ]. (1986) Modeling of the relations between of metamorphism through mineral equilibria (ed.
reaction enthalpy and the buffering of reaction ].M. Ferry). Mineralogical Society of America,
progress in metamorphism. Mineralogical Reviews in Mineralogy, 10, 355-397.
Magazine, 50, 375-384. Vance, D. (1995) Rate and time controls on metamor-
Rowe, K.]. & Rutter, E.H. (1990) Palaeostress estima- phic processes. Geological Journal, 30, 241-259.
tion using calcite twinning: experimental calibra- Walther, J.V. & Wood, B.]. (1984) Rate and mecha-
tion and application to nature. Journal of Structural nism in prograde metamorphism. Contributions to
Geology, 12, 1- 17. Mineralogy and Petrology, 88,246-259.
Sempels, ].-M. & Raymond, ]. (1980) Mathematical Yardley, B.W.D., Rochelle, C.A., Barnicoat, A.C. &
simulation of the growth of single crystals. Lloyd, G.E. (1991) Oscillatory zoning in metamor-
Computers and Geosciences, 6, 211-226. phic minerals: an indicator of infiltration metasoma-
Shelley, D. (1993) Igneous and metamorphic rocks tism. Mineralogical Magazine, 55, 357-365.

Chapter six

intergrowths and

Small solid-phase inclusions are commonly POIKILOBLASTIC structure, or sometimes

observed in metamorphic minerals. They can termed 'sieve' structure. It is analogous to
develop either by exsolution of a solute phase poikilitic structure of igneous rocks, with
during cooling, by enclosure of residual foreign which it should not be confused. The inclusions
phases during porphyroblast growth, or due to may either show no preferred orientation, may
incomplete pseudomorphing of an early phase be arranged with respect to the internal struc-
by a later one. In certain high-grade rocks two ture of the crystal (see below) or may be
or more phases may crystallise simultaneously oriented in relation to some pre-existing fabric
to give rise to distinctive symplectic inter- of the rock.
growths. In granulite facies rocks, the develop- Inclusions increase the total free energy,
ment of concentric coronas of one or more because they give rise to a greater surface area
phases around a core of another phase is a in poikiloblastic crystals (depending on the
commonly observed feature. The characteristics number and size of inclusions) and because of
of these various microstructures and the the inherent problem of fitting inclusions of a
processes involved in their development will different phase into the host porphyroblast. For
now be examined. this reason, poikiloblasts are less stable
compared to inclusion-free porphyroblasts of
6.1 Growth of porphyroblasts to enclose the same phase. The development of euhedral
residual foreign phases porphyroblasts with few inclusions probably
reflects slow growth, whilst the formation of
Many porphyroblasts have rather cloudy cores anhedral poikiloblasts is favoured by rapid
due to very fine inclusions, but in many cases growth.
(e.g. in garnet, cordierite, staurolite and horn- The included phases in porphyroblasts may
blende) the host porphyroblast contains abun- represent an 'inert' phase not involved in the
dant clearly discernible inclusions (e.g. quartz porphyroblast-forming reaction, but enveloped
and opaques) giving the host crystal a 'spongy' as the porphyroblast grows (e.g. zircon in
appearance (Plate l(a)). This is known as garnet). Alternatively, they may represent

Mineral inclusions, intergrowths and coronas

residual excess of a phase involved in the but may be overlooked because of their small
porphyroblast-forming reaction but overtaken size (typically < 50 JIm). However, in micas,
by the growth 'front' before it could be chlorites and certain other minerals, the pres-
completely utilised (e.g. quartz in garnet, ence of zircon inclusions may be recognised on
cordierite, andalusite, and so on). This occurs account of the pleochroic haloes that surround
because diffusion rates in the host are too slow them (Fig. 6.2). These haloes result from the
with respect to growth rate. In order to radioactive decay of small amounts of U and
minimise the surface (interfacial) energy, inclu- Th contained within zircon, which affects the
sions tend towards shapes with minimal structure of the host mineral. The size of the
surface area per unit volume. This is limited in pleochroic halo is proportional to the size of
some cases by their crystallographic structure. the zircon inclusion, but is commonly 10-
Quartz, for example, commonly occurs as 150 pm diameter. Inclusions of allanite and
fairly rounded to elliptical inclusions in most monazite may also show such haloes.
minerals, while phases such as sillimanite, Because of the two processes of ADSORP-
apatite and rutile form elongate inclusions with TION and ABSORPTION by which ions,
rounded ends (Fig. 6.1). Vernon (1976) empha- molecules or minerals may become attached to
sises that these 'ideal' inclusion shapes may not the growing front of a crystal, the presence of
necessarily be attained under low-grade meta- inclusions need not necessarily mean a high
morphic conditions, and that growth under energy or unstable situation. ADSORPTION is
significant stress can give strong preferred the loose physical bonding to the crystal
orientation in a given direction. surface of foreign material. The material is
Inclusions of zircon occur in many minerals, located at surface defects rather than forming

FIG. 6.1 Elongate rutile (dark) and tourmaline (light) inclusions in kyanite porphyroblast from kyanite schist.
Ross of Mull, Scotland. Scale =0.1 mm (PPL).

Growth of porphyroblasts

FIG. 6.2 Pleochroic haloes developed around zircon inclusions in biotite, as a result of radioactive decay of small
amounts of U and Th in zircon affecting the structure of the host mineral. Pelitic schist, Snake Creek, Queensland,
Australia. Scale bar = 200 JIm (PPL).

an integral part of the structure. The adsorbed 1991). In most other situations porphyroblasts,
material may either be enclosed as the crystal are considered to grow by matrix replacement.
grows or else accumulate in front of the ABSORPTION involves chemical bonding
advancing crystal face. Graphite (or amor- and integration of foreign material into the
phous carbon) has a strong tendency to become crystal structure. If the impurity has or devel-
adsorbed to porphyroblasts and is commonly ops a low-energy interface with the growing
seen concentrated at the edges of andalusite, porphyroblast, then during conditions of rapid
garnet, chloritoid and staurolite (Fig. 6.3). growth (which is probably most usual) the acti-
Harvey et al. (1977), Ferguson et al. (1980) vation energy of attachment is easily surpassed
and more recently Rice & Mitchell (1991) and the porphyroblast will build up around,
discuss the development of CLEAVAGE and eventually enclose, the material as an
DOMES at the faces of idioblastic porphyrob- inclusion. Because it is bounded by a low-
lasts. These domes, compnsmg graphite, energy immobile interface, the force of rejec-
muscovite and a low proportion of quartz tion will be small. It appears that certain
compared to the adjacent matrix, are inter- crystals (e.g. quartz) are more readily included
preted in terms of displacement of insoluble than others (e.g. micas). This suggests that
matrix graphite and muscovite during porphy- porphyroblast-quartz interfacial energies are
roblast growth in a bulk hydrostatic stress lower compared to porphyroblast-mica interfa-
field. Such domes, though rarely recorded, are cial energies. Alternatively, it may indicate that
the only reliable indicator of porphyroblast mica is being consumed in the porphroblast-
growth by displacement (Rice & Mitchell, forming reactions and simply that there is an

Mineral inclusions, intergrowths and coronas

FIG. 6.3 Fine-grained graphite ADSORBED to the surface of a staurolite porphyroblast. Graphitic staurolite
schist, Ghana. Scale = 0.5 mm (PPL).

excess of quartz in the system, such that even and chiastolite could be explained in terms of
though quartz may be involved in the porphy- preferential adsorption of impurities at certain
roblast-forming reaction, small remnant inclu- crystal faces. This general model has been
sions of quartz are inevitable. Within a advocated by subsequent workers, and is reiter-
sequence of schists it is more than coincidence ated here. Andalusite is an orthorhombic crys-
that the greatest density of quartz inclusions tal with (1l0)"(1I0) of 89°. It forms prismatic
generally occur in porphyroblasts from those crystals elongate parallel to C. Depending on
lithologies that are most quartz-rich. the thin-section cut, the basal sections of
The HOUR-GLASS structure frequently andalusite (var. chiastolite) show various inclu-
observed in chloritoid porphyroblasts (Plate sion patterns (Fig. 6.4 & Plate l(c)), the inclu-
1(b)) and the characteristic cross-like inclusion sions usually being of fine carbonaceous
arrangement of CHIASTO LITE (a textural material and quartz.
variety of andalusite) from contact metamor- Although less frequently encountered, rhomb-
phosed pelites (Plate l(c)) represent regular dodecahedral garnets with inclusions of quartz
geometrical patterns of inclusions arranged in concentrating at the interfacial boundaries are
relation to the host structure of the crystal also recorded (e.g. Powell, 1966; Atherton &
(Kerrick, 1990, pp. 302-310). Staurolite has Brenchley, 1972; Anderson, 1984; Burton, 1986;
similarly been shown to have regular arrange- Rice & Mitchell, 1991). These give spectacular
ment of inclusions in relation to crystal struc- examples of the same type of feature (Plate l(d)
ture in some instances. Frondel (1934) & Figs S.9(a) & 6.4). Two distinct types of
advanced the idea that the regular arrangement inclusions can be recognised, namely, Type 1
of inclusions seen in crystals such as chloritoid inclusions of quartz, Fe-Ti oxides and graphite,

Exsolution textures

Edge 101 long
110 ~110 cut

."-@ $




FIG. 6.4 The influence of the thin-section cut in relation to crystal orientation on the patterns seen in porphyrob-
lasts where inclusions are concentrated at interfacial boundaries. (a) and (b) are for different cuts of garnet, and
(c) is for different cuts of andalusite (var. chiastolite).

which were derived from the matrix and are amphiboles) representing non-ideal solid solu-
preferentially located along the interfacial tions (i.e. in which the enthalpy of mixing,
(sector) boundaries, and Type 2 inclusions, 6.Hmi x * 0) show unmixing as temperature
which are elongate or rodded quartz inclusions decreases. This immiscibility between solute
arranged perpendicular to the crystal faces. and solvent gives rise to the development of an
These Type 2 inclusions have tubular, or rodded interphase boundary and recognisable inclu-
form, and are not really inclusions, but inter- sions of the solute within the host mineral. The
growths formed simultaneously with garnet degree of ordering of solute atoms is dependent
growth. In all probability, growth was relatively on the rate of temperature reduction. If the
slow in order to develop such a well defined drop is relatively slow it allows increased ionic
crystallographically controlled arrangement. mobility, and the solute ions will become
In minerals which possess a strong crystal increasingly organized before finally separating
cleavage (e.g. amphiboles and micas) it is as discrete inclusions to give an EXSOLUTION
common for solid-phase inclusions to be prefer- texture within the host phase (Fig. 6.5). This
entially incorporated and aligned within the unmixing occurs in order to minimise the
cleavage planes. When studying thin sections, Gibbs free energy of the system by producing
care should be taken not to misinterpret such an exsolved phase and a chemically changed
well aligned inclusions as clear evidence for host phase that have a combined energy contri-
overgrowth of some earlier rock fabric, when in bution that is less than that of the original
fact they result from a crystallographic control. phase. On a T-X diagram (T = temperature; X
= composition between two end-member
components of a solution series), the solvus
6.2 Exsolution textures (Fig. 6.6) is a curve that separates the single
In igneous and metamorphic rocks, a number phase field (above) from the immiscible two-
of minerals (e.g. feldspars, pyroxenes and phase field (below). If a horizontal line of

Mineral inclusions, intergrowths and coronas

FIG. 6.5 Exsolved iron oxides along cleavage planes in hornblende. Epidiorite, Norway. Scale = 1 mm (PPL).



i~600 Ca/ ss + OoI ss



0.1 0.2 0.3 0.4 CaMg [C03[ 2


FIG. 6.6 The Cal-Dol solvus fitted to experimental data (modified after Anovitz & Essene, 1987). Arrowheads
show the final compositions, and the direction of the arrow gives the direction of the composition shift.
Experimental pressures are corrected to 2 kbar. The points near the dolomite limb represent natural dolomites
coexisting with calcite.

constant temperature is drawn across the temperature. It is apparent from Fig. 6.6 that,
diagram, the two points of intercept with the as temperature decreases, the compositional
solvus give the equilibrium compositions of the difference between the two phases (calcite and
phases resulting from phase separation at that dolomite) will steadily increase. The example

Exsolution textures

of calcite-dolomite intergrowths is well docu- oriented lamellar form. This is frequently

mented by Kretz (1988). Other minerals that observed in feldspars and pyroxenes of certain
commonly show exsolution include Ca,Mg,Fe- igneous and high-grade metamorphic rocks.
pyroxene, Ca,Mg,Fe-amphiboles, feldspars and The development of such a texture must either
spinels. have involved some primary ordering of the
On the basis of theory, experiment and impurity atoms in the host, which became
empirical observation, T-X diagrams are well more ordered during cooling or, alternatively,
established for many solution series relevant to may indicate restricted nucleation and growth
metamorphic phases. The systematic relation- along specific lattice planes. Throughout the
ships observed have often been developed as exsolution process the framework of the host
geothermometers (e.g. Cal-Dol; Anovitz & remains largely undisturbed, and the host
Essene, 1987). In natural situations, exsolution retains its original shape and size. There is
usually requires high diffusion rates and suffi- usually a strong relationship between the orien-
cient time for nucleation and growth of the tation of the exsolved inclusions and the
exsolved phases. Since these conditions are not atomic structure of the host (Fig. 6.5).
always satisfied, some phases formed above the Unmixing of alkali feldspar during cooling
solvus will persist metastably at lower tempera- from high temperatures gives rise to exsolution
tures. However, certain mineral phases seem to intergrowths of plagioclase (albite) and K-
have compositions that are so unstable below feldspar. This gives the characteristic
the solvus that even with rapid uplift they will PERTHITIC texture common to feldspars of
exsolve. many plutonic igneous and high-grade meta-
On the basis of exsolution intergrowths morphic rocks (Figs 6.7 & 6.8). The term
observed in metals and alloys, Christian (1965) PERTHITE strictly refers to feldspar inter-
discussed different modes of precipitation from growth, in which the dominant phase is K-
solid solutions. Most common mineral exsolu- feldspar with inclusions of plagioclase.
tion textures, particularly oriented inter-
growths (e.g. perthites) are considered to form
by a process of 'continuous precipitation', b
involving continuous draining of the solute
throughout the host crystal's lattice. This is a
distinctly different process from the model of
simultaneous crystallization of two or more
phases to give a symplectic texture.
The variations in size and shape of the
exsolved phase are a function of ionic mobility c d
(in turn dependent on temperature) and the
interfacial energy between the phases. The
exsolved phase often exists as abundant small
blebs (Fig. 6.5). These commonly show an even
distribution, suggesting homogeneous nucle-
ation and minimal change in lattice spacing
within the host. In other cases, concentration at
the margin of the host crystal is observed, FIG.6.7 Some common forms of perthite: (a) rodded
or string; (b) flame-like; (c) banded or lamellar; (d)
suggesting heterogeneous nucleation at grain braided or interlocking. The width of the box is
boundaries. Other exsolution textures have an roughly 0.2 mm in each case.

Mineral inclusions, intergrowths and coronas

FIG.6.8 A natural example of string perthite. Migmatite, Ghana. Scale =0.1 mm (XPL).

ANTIPERTHITE is the converse and consists tion is favoured as the dominant mechanism
of K-feldspar enclosed in plagioclase (typically relative to 'replacement', since the latter would
oligoclase-andesine). The term MESO- give varying proportions (Hubbard, 1965).
PERTHITE is given to exsolved feldspars with However, this is undoubtedly not the favoured
subequal volumes of intergrown K-feldspar interpretation in all examples. A recent study
and plagioclase. Although sometimes visible in by Pryer & Robin (1995) examined the origin
hand specimen, most perthitic and antiperthitic of flame perthite in deforming granites in the
intergrowths occur on the microscopic and Grenville Front Tectonic Zone, Canada. They
sub-microscopic scale. Alling (1938) classified reviewed the arguments for and against exsolu-
perthites in igneous rocks largely on the basis tion and replacement and, in their particular
of shape. Many of these types are recognised in example, concluded that flame perthite had
high-grade metamorphic rocks, and a selection developed due to retrograde metamorphic
of commonly observed types is shown in Fig. reactions involving the replacement of K-
6.7. Various processes have been suggested for feldspar by albite. In another study of flame
the formation of perthites (and antiperthites). perthites, Pas schier (1985) documented exam-
These are reviewed by Smith (1974), who ples in which the origin is attributed to stress
concluded that the relative importance of variations and the degree of deformation of K-
'replacement' and 'exsolution' are very much a feldspar grains. This certainly appears to be an
function of host rock and of the crystallo- attractive interpretation in the case of deform-
graphic and chemical properties of the feldspar ing high-grade gneisses, so it appears likely
itself. In cases in which separate grains in the that more than one process is responsible for
same sample show a constant proportion of the formation of perthite in metamorphic
host to inclusion, the 'exsolution' interpreta- rocks.


6.3 Inclusions representing incomplete CaAl 2Si20 S + 2(Mg,Fe)2Si04 =Ca(Mg,Fe)Si20 6 +

replacement 2(Mg,Fe)Si0 3 + MgAI20 4,
An + 01 =Di + Hy +Spl (6.2)
Of the many examples in which porphyrob-
lasts are packed with numerous small inclu- in metagabbros, caused by instability of PI + 01
sions (poikiloblastic), there are certain cases in due to increasing P or decreasing T. In this
which the inclusions do not represent over- case, the originallabradoritic plagioclase shape
grown matrix minerals or remnants from a is preserved, but the labradorite is transformed
porphyroblast-forming reaction, but in fact to andesine with a 'clouding' of minute blebs of
represent incomplete replacement of the host spinel. Another case involving plagioclase
phase by the small included phase(s). A good (illustrated by Shelley, 1993, p.159) involves
example of this type of reaction microstructure incoherent incomplete replacement of a large
is the 'gefullte plagioklas' (= stuffed plagio- plagioclase (in a basic metatuff) by amphibole
clase) commonly developed in Alpine meta- and quartz. A final example of a host phase
granodiorites and metatonalites (Angel, 1930; containing numerous inclusions relating to an
Ackermand & Karl, 1972). The microstructure incomplete prograde reaction is the paramor-
comprises subhedral to euhedral plagioclase phic replacement of andalusite by sillimanite.
packed with numerous tiny, euhedral/subhe- In this case, Kerrick & Woodsworth (1989)
dral and randomly oriented clinozoisite and illustrate examples of abundant new crystals of
muscovite inclusions (Fig. 6.9; see also Yardley sillimanite enclosed within single crystals of
et al., 1990, p. 77). This feature can be inter- andalusite as a result of the incomplete trans-
preted in terms of original (moderately An- formation. Rosenfeld (1969) and Vernon
rich) plagioclase of the igneous proto lith (1987) similarly describe excellent examples of
reacting with K-feldspar and H 2 0 to form the paramorphic replacement of andalusite by
more Ab-rich plagioclase and numerous inclu- sillimanite.
sions of clinozoisite, muscovite and quartz,
during Alpine greenschist/amphibolite facies 6.4 Symplectites
orogenic metamorphism. In effect, the An-
Two or more mineral phases frequently crys-
component of plagioclase has reacted with K-
tallise simultaneously in a metamorphic rock
and yet remain largely separated. For example,
4CaAl2Si20 S + KAlSi 3 0 s + 2H20 =
during the prograde metamorphism of pelites,
garnet and staurolite may develop contempora-
2Ca 2AI 3Si 30 12 (OH) + KAI 3Si 30 lO (OH)2 + 2Si0 2,
neously as porphyroblasts and yet show a wide
An + Kfs + HzO=Czo/Zo+ Ms +Qtz. (6.1) distribution throughout the rock because of
abundant and dispersed sites of nucleation. In
A key feature of this type of inclusion is that other cases, however, two or more phases can
some chemical components have to be able to show intimate intergrowth textures. These can
diffuse through the host grains. Another exam- form regular oriented arrangements, but more
ple in which included phases are directly usually occur as irregular and complex inter-
related to a complex reaction involving the growths. The most frequently encountered
host phase is the case of minute blebs of spinel types in metamorphic rocks consist of irregular
glVlng a 'clouding' in plagioclase (e.g. fine-grained mineral intergrowths, and are
Nockolds et aI., 1978, p. 355), as a result of termed SYMPLECTITES, or SYMPLECTIC
the reaction INTERGROWTHS.

Mineral inclusions, intergrowths and coronas

(often referred to as 'kelyphytic rims'), or asso-

ciated with the replacement of a primary phase
by a pair of secondary phases (Figs 6.10(b) &
(c)). In kelyphitic rims the intergrowths occur
on a range of scales from those visible by stan-
dard microscopy down to ultra-fine inter-
growths only visible by high-magnification
SEM work (Fig. 6.11).
Diffusion (especially grain-boundary diffu-
sion) is the main process involved in the forma-
tion of symplectites. Diffusion proceeds down
FIG. 6.9 A schematic illustration of plagioclase the chemical potential gradient, so the direction
packed with Czo + Ms inclusions as a result of the
and rate of diffusion is different for each
prograde reaction 4An + Kfs + 2H 2 0 = 2Zo/Czo + Ms
+ 2Qtz (based on Angel, 1930; Ackermand & Karl, element. In symplectites, the intergrowths
1972; Droop, pers. comm., 1997) in relation to meta- comprise elongate minerals and small blebs
tonalites from the Hohe Tauern, Austria). The field of arranged roughly perpendicular to the reacting
view is approximately 6 mm. interface. Slow diffusion of elements relative to
the rate of progress of the reacting interface is
Most symplectites are associated with high- held responsible for the formation of these
grade metamorphism, and are especially symplectites. In many cases it is slow diffusion
common in high-grade gneisses and granulites. of Al and Si that are probably responsible, but
Although some tri-mineralic intergrowths many other elements are also involved, some
occur, most symplectites consist of intimate diffusing towards the interface, and others
intergrowths of mineral pairs. In many cases moving in the opposite direction. The spacing
the development of symplectites is considered of the rods relates to the range of the slowest
the direct result of decompression during uplift diffusing species.
(Section 6.5). This gives rise to instability and Kretz et at. (1989), in a study of corona and
retrogression of the high-grade granulite assem- symplectite textures developed between olivine
blages. Commonly encountered symplectites and plagioclase in metagabbros from Quebec,
include: Canada, emphasised the complexity of diffusion
pathways (A-E in Fig. 6.12) and the fact that
(i) Crd + Qtz (or Crd + Opx + Qtz) as a retro-
the reaction zone is not entirely isochemical:
grade breakdown of Grt;
that is, some elements involved in the reaction
(ii) Opx + PI symplectites as retrogression of
have diffused to the reaction area from outside.
Hbl + Grt, or developed due to instability
The study of olivine-plagioclase coronas in
of Cpx + Grt + Qtz (Fig. 6.10(a) is based
metagabbros of the Adirondacks, USA, by
partly on a diagram of De Waard (1967),
Johnson & Carlson (1990), similarly concluded
which shows a symplectic intergrowth of
that to achieve a mass balance for the reactions
orthopyroxene (C) and plagioclase (D),
involved diffusional transport of material into
formed due to the instability of Grt(A) +
and out of the reaction bands occurred. Since
Cpx(B) + Qtz);
the width of a symplectite reaction zone
(iii) Opx + Spl symplectites formed by reaction increases with time, this means that the concen-
of 01 + PI. tration and chemical potential gradients across
Symplectites generally occur either at bound- the zone will diminish. This causes slowing
anes between reacting minerals (Fig. 6.10(a)) of diffusion rates and thus reaction rates, and



FIG. 6.10 (a) A schematic illustration of symplectite developed in the zone between two reacting phases. In this
illustration, A and B are at disequilibrium and are separated by a symplectic intergrowth of C and D. (b) A
schematic illustration of a symplectic intergrowth of phases A and B as total replacement (pseudomorph) after
some earlier phase. (c) A natural example, showing symplectic intergrowth of orthopyroxene (light) and magnetite
(black). Meta-norite, south-west Norwa y. Scale = 0.5 mm (PPL).

ultimately the reaction terminates when the plagioclase is known as MYRMEKITE (Fig.
slowest diffusing element(s) are unable to keep 6.13). The formation of myrmekite both in
pace with the reaction. The situation is not metamorphic and in igneous rocks has long been
helped if the symplectite is forming during cool- a subject of debate. The development of the vari-
ing, because the decline in temperature will also ous schools (:jf thought on its origin are reviewed
cause a slowing of diffusion rates. It is perhaps by Smith (1974) and Phillips (1974, 1980).
for these reasons that many symplectite-forming The present consensus of opinion regarding
reactions fail to go to completion and are myrmekite in metamorphic rocks generally
commonly preserved in granulite facies rocks. relates its development to the breakdown of K-
A specific type of symplectite comprising feldspar. This is founded on the common obser-
vermicular (worm-like) quartz intergrown with vation that myrmekitic intergrowths of

Mineral inclusions, intergrowths and coronas

FIG. 6.11 An SEM back-scattered image of symplectic intergrowths in a granulite facies mafic protomylonite
(after Brodie, 1995, courtesy of Blackwell Science). The instability of Grt (the principal reactant) in the presence
of Cpx has produced a symplectic intergrowth of PI (dark) + Opx (medium grey). Fine (1-3 /lm) 'strings' of Spl
(light grey) are intergrown with the Opx, but it is difficult to determine whether this is part of the original
symplectite or else produced during subsequent exsolution from the Opx (Brodie, 1995). Minor Fe-Ti oxides
(white) are also present. The symplectites of this example developed in dilatant zones, and their alignment defines
the original extension direction at the time of reaction. Scale bar = 100 Jim; field of view c. 340 /lm.

plagioclase and quartz are especially common

in retrogressed high-grade gneisses lacking K-
feldspar, and commonly in association with the
breakdown of sillimanite or kyanite to white
mica. The complete reaction can be written
simplistically as
Kfs + Al 2SiO s + H 2 0 ~ PI + Qtz + Ms. (6.3)

Ot This model has its roots in the proposals of
Becke (1908), but has been expanded in detail
FIG. 6.12 Complex PI-OI symplectites (based on Fig. by Phillips and co-workers. Phillips (1980)
5.15 of Kretz et al., 1989). Scale bar = 0.1 mm. describes and discusses the different



FIG. 6.13 (a) Myrmekite in augen gneiss from Tongue, Scotland. This microstructure comprises vermicular
quartz (white) intergrown with plagioclase. It is common in high-grade metamorphic rocks, and forms in associa-
tion with the retrogression of K-feldspar. Scale = 0.1 mm (XPL) .

tered type consists of relict alkali feldspar

surrounded by myrmekite and muscovite (Fig.
6.13(c)). Simpson & Wintsch (1989) demon-
strate that in many cases such myrmekite forms
in direct response to stress-induced K-feldspar
replacement, and that it preferentially develops
on the most strained margins of the feldspar
(see also Section 6.2 regarding the development
of flame perthite on strained feldspars). In
FIG 6.13 (b) A schematic illustration of bulbous
myrmekite as a total replacement for K-feldspar. Scale Al 2SiOs-free rocks (e.g. meta-granitoids),
= 0.1 mm. (c) A schematic illustration of a relict K- myrmekite is also common, and can be
feldspar crystal mantled by myrmekite . Scale =0.1 mm. explained in terms of a late-stage ion-exchange
reaction between K-feldspar and fluid (Droop,
pers. comm., 1997), namely,
morphologies of myrmekite, of which those
types most common to metamorphic rocks are KAlSi 30 g + Na+ -> NaAlSi 3 0 g + K+,
illustrated in Figs 6.13(b) & (c). 2KAlSi 30 g + Ca 2+ -> CaAl 2 Si 2 0 g + 4SiO z + 2K+,
Kfs fluid PI Qtz fluid.
The type termed 'bulbous myrmekite' (Figs
6.13(a) & (b)), in which the myrmekite appears
to invade adjacent alkali feldspar, is very A recent study by Brodie (1995) documents
common in high-grade (especially quartzofelds- how vermicular symplectites of orthopyroxene,
pathic) gneisses. The other commonly encoun- plagioclase and spinel have formed at the

Mineral inclusions, intergrowths and coronas

margins of garnets synchronous with deforma- As with symplectic intergrowths, the develop-
tion under granulite facies conditions in metaba- ment of corona structures in high-grade meta-
sites within a shear zone. Such symplectites morphic rocks results from the instability of peak
develop in rocks in areas of relatively low strain, assemblages in response to declining P-T condi-
such as at shear zone margins but, interestingly, tions during uplift. In all cases the 'core' phase
the symplectites are oriented parallel to the main reaches a state of disequilibrium with one or
foliation and regional stretching direction (Fig. more of the surrounding phases. The ensuing
6.11). Brodie (1995) concluded that the distribu- reaction produces a corona of some new phase
tion of symplectites within the rock indicate that or phases, thus forming a barrier isolating those
they only developed in dilatant areas, especially phases at disequilibrium. In common with
garnet grain boundaries and cracks in garnet. symplectites, the process involved is one of
Such observations from the Ivrea-Verbano zone, localised diffusive transfer of material between
Italy, are used as evidence for deep crustal exten- the reactants, with high temperatures allowing
sion localised along high-strain zones. reasonable reaction rates. As with symplectites,
the broadening of the corona decreases chemical
potential gradients and thus slows diffusion.
6.5 Coronas (of high-grade rocks)
Likewise, diffusion rates slow as temperature
Corona structures (Plates 2(a) & (b)) are espe- declines, and ultimately reaction progress is
cially common in high-grade gneisses, granulites terminated. The study of olivine-plagioclase
and eclogites. They consist of a core of one coronas by Johnson & Carlson (1990), utilising
mineral phase completely (or almost both transmitted light microscopy and SEM
completely) enclosed by a 'corona' of another back-scattered electron (atomic number contrast)
phase (or phases). The terms 'collar', 'atoll' and imaging, provides a particularly well documented
'moat' have also been used in the literature to example of coronas and the processes involved in
describe such features. However, the use of the their formation. The contribution by Rubie
term 'atoll' is not recommended, because it (1990) emphasises that corona-forming reactions
could cause confusion, as it has also been used occur at interfaces between specific mineral
with reference to 'atoll porphyroblasts' (e.g. phases, and gives evidence for complex multi-
atoll garnets), a specific microstructure in which component diffusion of elements derived from
the core of a chemically zoned porphyroblast adjacent domains. Additionally, Rubie (1990)
has reacted to form other minerals (e.g. micas). provides valuable discussion on possible
An example of a high-grade corona structure approaches towards evaluating the kinetics of
involves a garnet core with a plagioclase collar corona-forming reactions in eclogites.
(Plate 2(a)) and newly developed orthopyrox- Coronas and symplectites clearly provide
ene separating plagioclase from quartz. This is useful information about reactions, but with
a familiar decompression texture of many gran- the information now available to construct
ulites, and results from the reaction Grt + Qtz detailed petrogenetic grids it is also possible to
= Opx + Pl. Many other types of high-grade relate such reaction microstructures in terms of
coronas have been recorded, the minerals and the P- T path that the rock has followed. This
reactions being largely the same as those of is discussed more fully in Section 12.3.4.
symplectites. Coronas may be mono- or bi- A structure comparable to that of the
mineralic and in some cases are multiple. In coronas described above in relation to gran-
multi shell coronas it is common to find that ulites may also be encountered in certain types
certain 'shells' are represented by symplectic of migmatite. The so-called 'flecky' gneisses of
intergrowths. Loberg (1963), Russell (1969) and Ashworth


:: --: :.-~ -~--------------

... __ -
--:--:-~~:-:--::--~----------- -- - - - - - - -.. - ----
-- .. -
- - - ..-~-~~-~~---~-~.~~-:~:---------- --- - - - - ---
-- ~-- : -- --.. --- -.-. --"'-
----- : =- ..

FIG.6.14 A drawing of "flecky gneiss' (based on Fig. 50a of Loberg, 1963). Mafic cores (black) have an assem-
blage And + Bt + Qtz + PI, and are surrounded by a quartzofeldspathic leucosome corona, or selvage (white), that
separates the core from the mesosome matrix assemblage of Sil + Bt + Qtz + Feld. Scale bar = 1 cm.

(1985) result from high-temperature reaction Alling, A.L. (1938) Plutonic perthites. Journal of
processes, again involving diffusion of elements Geology, 46, 142.
Anderson, T.B. (1984) Inclusion patterns in zoned
down local chemical potential gradients, to garnets from Magemy, north Norway.
produce a coarse and irregular corona structure Mineralogical Magazine, 48, 21-26.
(Fig. 6.14). In the Loberg (1963) example from Angel, F. (1930) Dber Plagioklasefiillungen und ihre
Viistervik, Sweden, mafic cores of And + Bt + genetische Bedeutung. Mitt. Naturw. Ver.
Stein mark, 67, 36-52.
Qtz + PI are surrounded by quartzofeldspathic Anovitz, L.M. & Essene, E.]. (1987) Phase equilibria in
(leucosome) coronas or haloes. The mesosome the system CaCO r MgC0 3-FeCO y Journal of
forming the main body of the rock comprises Petrology, 28, 389-414.
Sil + Bt + Qtz + Feld. The mafic clots and coro- Ashworth, ].R. (1985) Introduction, in Migmatites (ed.
].R. Ashworth). Blackie, Glasgow, 1-35.
nas are considered to have developed during Atherton, M.P. & Brenchley, P.]. (1972) A preliminary
the polymorphic transformation sillimanite to study of the structure, stratigraphy and metamor-
andalusite, and to have involved Fe + Mg diffu- phism of some contact rocks of the western Andes,
sion inwards to the cores and K diffusion near the Quebrada Venado Muerto, Peru.
GeologicalJournal, 8, 161-178.
outwards into the leucosome (Fisher, 1970). By Becke, F. (1908) Dber myrmekit. Mineralogische und
considering reasonable diffusion rates, Fisher Petrographische Mitteilungen, 27, 377-390.
(1977) estimated that for core radii of about 5 Bowman, ].R. (1978) Contact metamorphism, skarn
mm, the segregations would form in about formation and origin of C-O-H skarn fluids in the
Black Butte aureole, Elkhorn, Montana. Ph.D.
0.066 Ma. thesis, University of Michigan.
Brodie, K.H. (1995) The development of orientated
References symplectites during deformation. Journal of
Ackermand, D. & Karl, F. (1972) Experimental studies Metamorphic Geology, 13,499-508.
on the formation of inclusions in plagioclases from Burton, K.W. (1986) Garnet-quartz intergrowths in
metatonalites, Hohe Tauern, Austria (lower temper- graphitic pelites: the role of the fluid phase.
ature stability limit of the paragenesis anorthite Mineralogical Magazine, 50, 611-620.
plus potash feldspar). Contributions to Mineralogy Christian, ].W. (1965) Theory of transformations in
and Petrology, 35, 11-21. metals and alloys. Pergamon Press, Oxford.

Mineral inclusions, intergrowths and coronas

De Waard, D. (1967) The occurrence of garnet in the Ducktown, Tennessee massive sulfides and adjoin-
granulite-facies terrane of the Adirondack ing portions of the Blue Ridge Province. Ph.D.
Highlands and elsewhere, an amplification and a thesis, University of Michigan.
reply. Journal of Petrology, 8,210-232. Nockolds, S.R., Knox, R.W.O'B. & Chinner, G.A.
Essene, E.]. (1982) Geologic thermometry and barome- (1978) Petrology for students. Cambridge
try, in Characterization of metamorphism through University Press, Cambridge, 435 pp.
mineral equilibria red. ].M. Ferry). Mineralogical Passchier, CW. (1985) Water-deficient mylonite zones
Society of America, Reviews in Mineralogy No.10, - an example from the Pyrenees. Lith as, 18,
153-206. 115-127.
Ferguson, CC, Harvey, P.K. & Lloyd, G.E. (1980) On Phillips, E.R. (1974) Myrmekite - one hundred years
the mechanical interaction between a growing later. Lithos, 7, 181-194.
porphyroblast and its surrounding matrix. Phillips, E.R. (1980) On polygenetic myrmekite.
Contributions to Mineralogy and Petrology, 75, Geological Magazine, 117,29-36.
339-352. Powell, D. (1966) On the preferred crystallographic
Fisher, G.W. (1970) The application of ionic equilibria orientation of garnet in some metamorphic rocks.
to metamorphic differentiation: an example. Mineralogical Magazine, 35, 1094-1109.
Contributions to Mineralogy and Petrology, 29, Pryer, L.L. & Robin, P.-Y.F. (1995) Retrograde meta-
91-103. morphic reactions in deforming granites and the
Fisher, G.W. (1977) Non-equilibrium thermodynamics in origin of flame perthite. Journal of Metamorphic
metamorphism, in Thermodynamics in Geology red. Geology, 13, 645-658.
D.G. Fraser). D. Reidel, Boston, Ch. 19,381-403. Rice, A.H.N. & Mitchell, J.I. (1991) Porphyroblast
Frondel, C (1934) Selective incrustation of crystal textural sector-zoning and matrix displacement.
forms. American Mineralogist, 19,316. Mineralogical Magazine, 55, 379-396.
Harvey, P.K., Lloyd, G.E. & Shaw, K.G. (1977). Rosenfeld, J.L. (1969) Stress effects around quartz
Arcuate cleavage zones adjacent to garnet porphy- inclusions in almandine and the piezothermometry
roblasts in a hornfelsed metagreywacke. of coexisting aluminium silicates. American Journal
Tectonophysics, 39,473-476. of Science, 267,317-351.
Hubbard, F.H. (1965) Antiperthite and mantled Rubie, D.C (1990) Role of kinetics in the formation
feldspar textures in charnockite (enderbite) from and preservation of eclogites, in Eclogite Facies
S.W. Nigeria. American Mineralogist, 50, Rocks red. D.A. Carswell). Blackie, Glasgow,
2040-2051. 111-140.
Johnson, CD. & Carlson, W.D. (1990) The origin of Russell, R.V. (1969) Porphyroblastic differentiation in
olivine-plagioclase coronas in metagabbros from fleck gneiss from Vastervik, Sweden. Geologiska
the Adirondack Mountains, New York. Journal of Foreningen Stockholm Forhandlingar, 91,217-282.
Metamorphic Geology, 8, 697-717. Shelley, D. (1993) Igneous and metamorphic rocks
Kerrick, D.M. (1990) The AlzSiO s polymorphs. under the microscope. Chapman & Hall, London,
Reviews in Mineralogy No. 22, Mineralogical 445 pp.
Society of America, Washington, DC, 406 pp. Simpson, C & Wintsch, R.P. (1989) Evidence for
Kerrick, D.M. & Woodsworth, G.]. (1989) Aluminium deformation-induced K-feldspar replacement by
silicates in the Mount Raleigh pendant, British myrmekite. Journal of Metamorphic Geology, 7,
Columbia. Journal of Metamorphic Geology, 7, 261-275.
547-563. Smith, J.V. (1974) Feldspar minerals: Volume 2,
Kretz, R. (1988) SEM study of dolomite microcrystals Chemical and textural properties. Springer-Verlag,
in Grenville marble. American Mineralogist, 73, Berlin, 690 pp.
619-631. Vernon, R.H. (1976) Metamorphic processes. George
Kretz, R., Jones, P. & Hartree, R. (1989) Grenville Allen & Unwin, London, 247 pp.
metagabbro complexes of the Otter Lake area, Vernon, R.H. (1987) Oriented growth of sillimanite in
Quebec. Canadian Journal of Earth Sciences, 26, andalusite, Placitas - Juan Tabo area, New Mexico,
215-230. USA. Canadian Journal of Earth Science, 24,
Loberg, B. (1963) The formation of a flecky gneiss and 580-590.
similar phenomena in relation to the migmatite and Yardley, B.W.D., MacKenzie, W.S. & Guildford, C
vein gneiss problem. Geologiska Foreningens (1990) Atlas of metamorphic rocks and their
Stockholm Forhandlingar, 85,3-109. textures. Longman Science & Technology, and John
Nesbitt, B.E. (1979) Regional metamorphism of the Wiley, New York.

Chapter seven

Replacement and

7.1 Retrograde metamorphism in specific zones. Active zones of deformation

such as fault and thrust zones are particularly
favoured (e.g. Beach, 1980). This is because
7.1.1 Environments of retrograde metamorphic fluids are channelled into these
areas as a result of Pf and chemical potential
Retrogression or retrograde metamorphism is a gradients, in addition to the fact that deforma-
process involving the breakdown of higher P-T tion processes aid metamorphic reactions and
assemblages in association with declining P-T often enhance bulk permeability (see Section
conditions. Since the majority of retrograde 8.4 for details). Fluids migrate to areas of low
reactions require hydration or carbonation Pf' and in regimes of extensional and strike-slip
(Table 7.1), the presence of a fluid phase is tectonics it is not surprising that fluids concen-
essential for these reactions to proceed. Many trate into dilatant fault zones. In addition to
high-grade metamorphic rocks seen at the the brecciation and cataclasis, it is widely
Earth's surface exhibit remarkably fresh assem- observed that such fault zones are intensely sili-
blages, despite the phases in the assemblage cified, dolomitized, sericitised, chloritised or
being well outside their stability fields. This show some other features of chemical or miner-
indicates either that fluids did not enter the alogical change indicative of high fluid flux.
rock to promote the reactions expected with Many such zones are important sites of
declining P-T conditions, or that uplift rates economic mineral deposits (e.g. shear zone
were far greater than reaction rates. hosted gold deposits such as Golden Mile,
Rocks of regional metamorphic and subduc- Kalgoorlie, Australia; Boulter et al., 1987). In
tion-related environments commonly display the case of thrust zones, the emplacement of a
late-stage retrogression associated with their 'hot' slab over cooler rocks may lead to
uplift history. This suggests that fluids, having convective circulation of fluids in the footwall
been driven off by the various devolatilisation block. These fluids will enhance reactions in
reactions associated with burial and prograde the thrust zone and lead to retrogression at the
metamorphism, re-enter rocks during uplift, to base of the hangingwall block.
induce retrogression. Although regional-scale Retrogression is also important in areas of
retrograde metamorphism can be observed (e.g. igneous intrusion and hydrothermal activity.
amphibolitised granulites), it is often found The heat associated with such areas promotes
that retrograde metamorphism is concentrated metamorphic reactions by providing the energy

Replacement and overgrowth

TABLE 7.1 Common retrograde reactions in the main compositional groups of metamorphic rocks.

Ultramafic rocks
Olivine ~ serpentine H 2 0-rich fluids
~ magnesite CO 2-rich fluids
Enstatite ~ anthophyllite
Opx and/or olivine ~ talc ± serpentine

Ca -plagioclase ~ Na-plagioclase + Ep/Zo/Czo Very common amphibolite facies ~ greenschist
facies retrogression (H 20-rich fluids).
~ zeolites Common in very low grade burial
metamorphism and ocean-floor metamorphism
~ sericite/muscovite In metabasites this usually requires significant K+
~ calcite CO -rich fluids.
~ scapolite Hydrothermal metamorphism CO 2 -rich fluids.
Clinopyroxene ~ hornblende/actinolite
Hypersthene ~ hornblende/actinolite
Hornblende ~ actinolite
~ chlorite
~ biotite Usually associated with significant
K+ introduction
Blue (Na-) amphibole ~ green (Ca-) amphibole (actinolite)
Garnet ~ chlorite
Ilmenite or rutile ~ sphene

Granitoid rocks
K-feldspar ~ sericite/muscovite/pyrophyllite
~ clay minerals (e.g. kaolinite)
Plagioclase ~ sericite (epidote group minerals)
Biotite ~ chlorite

Calc-silicate rocks
Forsterite ~ serpentine
Anorthite ~ epidote minerals (± sericite)
~ carbonate minerals
Diopside ~ tremolite-actinolite
Tremolite ~talc

Garnet ~ chlorite and/or biotite
Staurolite ~ sericite
~ sericite + chlorite
Andalusite, kyanite, ~ sericite/white mica
Cordierite ~ pinite (fine mix of sericite + chlorite)
Chloritoid ~ chlorite (± sericite)
Biotite ~ chlorite
Ilmenite ~ sphene

Retrograde metamorphism

necessary for reactions to proceed. Intrusion- recognised around porphyroblasts that are in
related fluids carry various ions in solution, disequilibrium with matrix conditions, either
and by their interaction with interstitial fluids due to shearing at lower-grade conditions or in
and minerals in the surrounding country rocks association with uplift and cooling. Continued
can induce both prograde and retrograde reac- disequilibrium and reaction may give rise to a
tions. The principal factors controlling the type partial or complete pseudomorphing of the
of alteration that occurs are the composition unstable phase (Plates 4(c)-(e)). A PSEUDO-
and mineralogy of the host rock, the composi- MORPH is defined as a crystal that has been
tion of the fluid and the temperature of the completely altered or replaced by another
fluid. The extent of infiltration is controlled by mineral or aggregate of minerals, and yet still
the nature of the host-rock permeability and by retains its original shape. It may be either (a) a
Pf. Upward-moving hydrothermal fluids from a single-phase single-crystal pseudomorph, (b) a
cooling magma (especially granite or granodi- single-phase multicrystal pseudomorph or (c) a
orite) invade the surrounding country rocks to multiphase multicrystal pseudomorph. Single-
cause 'metasomatic' alteration. METASOMA- phase multicrystal pseudomorphs are fre-
TISM is a process of alteration or chemical quently observed, especially cases involving the
modification involving enrichment of the rock replacement of a porphyroblast by an aggre-
in certain ions derived from some external gate of some hydrous phase as part of a rehy-
source. This influx of ions via a fluid phase dration reaction; for example, garnet to
induces various metamorphic reactions in the chlorite (Plate 4(c)) or olivine to serpentine
country rock, although much of the original (Plate 4(d)). Equally common are multiphase
microstructure of the rock may still be multicrystal pseudomorphs, particularly those
preserved. Potassic alteration (e.g. sericitisa- producing bi-mineralic hydrous assemblages;
tion) is a particularly common type of alter- for example, hornblende ~ actinolite + chlo-
ation in such situations, as is propyllitic rite, and staurolite ~ chlorite + sericite and
alteration, involving widespread chloritisation, garnet ~ chlorite + biotite (Plate 4(e)). The
and argillic alteration, giving rise to extensive fact that fine-grained multicrystal aggregates
development of clay minerals. Extensive tour- are common indicates high nucleation rate rela-
malinisation is a common feature of metasoma- tive to growth rate. This causes numerous
tism in the vicinity of granitoid intrusions (e.g. nucleation sites and a tendency towards site
Hercynian granites of south-west England). saturation. The degree of pseudomorphing of a
Later sections of this chapter deal with specific particular phase can vary considerably across
types of retrogression and replacement, but for the domain of a thin section (Plate 4(d)) and on
further details on the various types of the outcrop scale may be localised in certain
hydrothermal alteration, the reader should zones. This reflects the variable degree of fluid
refer to Thompson & Thompson (1996). infiltration and illustrates the important role
that fluids play in promoting reactions. In Plate
4(d), H 2 0 is required for olivine to alter to
7.1.2 Textural features of retrogression
serpentine. Those areas infiltrated by fluid
CORONAS and REACTION RIMS are a clear show complete pseudomorphing of olivine, and
sign of disequilibrium between certain phases yet in adjacent areas in which the fluid has only
in the assemblage. In pelites and metabasites of gained minimal access the olivines are almost
the amphibolite facies and lower grade, coro- entirely unaltered. The preservation of un dis-
nas (or reaction rims) of hydrous mineral torted 'soft assemblage' pseudomorphs in
phases are common (Plate 4(b)). They are best highly sheared rocks (Plate 4(c)) implies that

Replacement and overgrowth

OO L-__-----------{


FIG.7.1 A phase diagram for the Si0 2 system, show- FIG. 7.2 A drawing of a coesite inclusion in garnet in
ing the P-T stability fields (based on Zoltai & Stout, the process of transformation to a-quartz (based on
1984) for the various silica polymorphs, and emphasis- Fig. 3d of Chopin, 1984). The positive volume change
ing the polymorphic transformation coesite ~ a-quartz associated with this transformation causes characteris-
(arrowed) associated with decompression during uplift tic radial fracturing in the garnet. The solid line defines
of certain ultra-high-pressure rocks. T, Tridymite; C, the Si02 inclusion, comprising a remnant high-relief
cristobalite; L, liquid. coesite core (inside stippled area), surrounded by a-
quartz (unornamented). Scale bar = 100 ;tm.

pseudomorphing occurred after deformation. graphic structure to the mineral replacing it. In
Polymorphic transformations such as aragonite this case biotite is being replaced by chlorite,
to calcite, and coesite to a-quartz (e.g. Chopin, both of which are phyllosilicates. This type of
1984; Okay, 1995) occur when rocks formed transformation can represent either a prograde
at very high pressure undergo decompression or a retrograde feature. It is retrograde if
during uplift (Figs 7.1 & 7.2). biotite forming part of a high-temperature
Reaction rims (Plate 4(b)) and partial or assemblage is being replaced, but prograde if,
complete pseudomorphs (Plates 4( c )-( e)) are for example, metamorphic chlorite is in the
obvious disequilibrium textures. Other alter- process of replacing detrital biotite (Section
ation features include 'core replacement' and 7.2). Numerous XRD, SEM and TEM studies
'zone replacement'. Zonal alteration is espe- have been undertaken on such interlayered
cially common in igneous plagioclase crystals phyllosilicate phases and the transformation
(Plates 4(f) & (g)). It indicates that certain processes involved (e.g. Veblen & Ferry, 1983;
zones of the plagioclase are out of chemical Jiang & Peacor, 1991). The use of high-resolu-
equilibrium with the matrix fluid, and conse- tion TEM in such research on polysomatic
quently are more prone to alteration. In such intergrowths is reviewed by Allen (1992).
cases the fluid gains access along microfrac- When studying porphyroblasts it may not
tures or cleavage planes of the crystal. always be easy to decide whether fine-grained
Feldspars, amphiboles and micas are particu- mineral enclosures represent alteration products
larly prone to such alteration. Plate 4(h) is a or inclusions incorporated during growth.
good example of TOPOTAXY (or TOPOTAC- However, there are certain points to take into
TIC TRANSFORMATION), in which the account, and features to look for that should
mineral being replaced is of similar crystallo- make the decision-making easier. First, with

Retrograde metamorphism

fine-grained alteration products it is often diffi- ophiolite obduction, and in contact metamor-
cult to clearly discern grain boundaries by stan- phic aureoles with extensive aqueous fluid
dard microscopy, whereas inclusions, although infiltration. The addition of H 2 0 is essential
commonly fine-grained, usually have well- for serpentinisation to proceed, and in
defined boundaries. Second, alteration products general it takes place below 500°C, and
will usually be of minerals chemically similar to commonly at less than 350°C. If mass
the mineral they are retrogressing (e.g. sericite remains constant during serpentinisation of a
after K-feldspar), whereas included mineral peridotite a substantial volume increase
phases will often be quite different (e.g. rutile (35-45%) will occur. The basic reactions can
in garnet). be written as
2Mg2Si04 + 3Hp = Mg3Sips(OH)4 + Mg(OH)2 (7.1)
7.1.3 Specific types of retrogression and forsterite serpentine brucite
replacement and
3MgSiOJ + 2Hp = Mg 3Sips(OH) 4 + Si0 2. (7.2)
Serpentinisation enstatite serpentine silica
Serpentine is the most common alteration If the original olivine or pyroxene contains a
product of olivine, and serpentinisation of component of iron, magnetite will be an addi-
ultramafic rocks (Fig. 7.3) and forsterite tional product of the reaction. Experiments by
marbles (Plate 4(d)) has received considerable Wegner & Ernst (1983), which included a
attention in the literature (e.g. Peacock study of the hydration of forsterite to give
(1987). Serpentinisation occurs in environ- serpentine plus brucite (7.1), showed that the
ments such as ocean-floor metamorphism, reaction rate was appreciably faster when
shear zones developed during orogenesis or P(H 2 0) was increased from 1 to 3 kbar.

FIG. 7.3 Serpentinised ultramafic rock, showing a distinctive serpentine 'mesh'. Dawros, Connemara, Ireland.
Scale bar = 125 Jim (XPL).

Replacement and overgrowth

FIG. 7.4 Pyroxenes replaced by amphibole (actinolitic hornblende) as part of the process of uralitisation in meta-
igneous rocks within a Precambrian basement window. Troms, Norway. Scale bar = 125 !fm (XPL). The
surrounding minerals are plagioclase and minor quartz.

Ural itisation Chloritisation

Uralitisation is the term given to the replace- Extensive replacement of original assemblages by
ment of primary igneous pyroxenes by amphi- chlorite (i.e. chloritisation) involves major influx
boles (typically tremolite, actinolite or of aqueous fluids. Such retrogression is
hornblende). This replacement commonly commonly associated with hydrothermal alter-
occurs at the margins of plutons when late-stage ation of mafic rocks and greenschist facies shear
intrusion-related aqueous fluids interact with zones (Fig. 7.5). Aqueous fluids are always
already crystallised parts of the intrusion. A involved, but the precise nature of the reaction(s)
similar replacement process can also take place will be a function of bulk rock composition and
during orogenic metamorphism to produce a the chemistry of the infiltrating fluids. If biotite is
rock termed epidiorite. This rock type preserves one of the reactants the K+ liberated will enter
a coarse crystal aggregate with no preferred the fluid phase and may be used in the formation
orientation (i.e. igneous appearance), but has of sericite elsewhere in the rock.
original igneous pyroxenes replaced by actino-
lite or hornblende (Fig. 7.4). For such a rock to Sericitisation
form, low strain rates are required to preserve The alteration of K-feldspar, plagioclase,
the igneous appearance, and aqueous fluids are AlzSiO s polymorphs, staurolite and cordierite
required to produce the amphiboles. If strain to fine-grained aggregates of white mica
rates are higher a cleavage or schistosity would (termed 'sericite') is a common feature of retro-
form, the original igneous microstructural gressed quartzofeldspathic rocks and meta-
features would be lost and the rock would be pelites. This sericitic alteration is especially
termed a greenschist. common in hydrothermal environments and

Retrograde metamorphism

FIG. 7.5 Chloritisation of mafic rock in a greenschist facies shear zone, Syama, Mali. Dark areas =chlorite; light
grey to white speckled areas =calcite. Scale = 125 flm (XPL).

greenschist facies shear zones (Fig. 7.6). The Saussuritisation

basic form of the K-feldspar ~ sericite reaction
Saussuritisation involves the liberation of Ca
can be written as
from calcic plagioclase to leave a feldspar of
3 KAISips + 2H+ =KAI2[AISiPIO](OH)2+ 6Si0 2+2K+. more albitic composition 'dusted' with a fine-
K-feldspar sericite quartz grained aggregate of epidote group minerals (:j:
(7.3) Serc :j: Cal). This type of alteration is common
during greenschist facies retrogression and
Whenever the activity of H+ relative to K+ situations of low-pressure hydrothermal alter-
increases, the reaction will be driven to the ation. Although this type of alteration is often
right and sericite will start to form at the patchy, in zoned plagioclase crystals of igneous
expense of K-feldspar. In cases of extensive rocks it is common to see the saussuritic alter-
hydrogen-ion metasomatism, H+ infiltrates the ation concentrated in the more anorthitic core
system and Ca and Na are flushed out, to leave of individual crystals (Plate 4(f)), or in An-rich
a rock with an extensively modified chemistry zones (Plate 4(g)). To develop such specific
and a mineral assemblage largely devoid of zonal alteration, the fluid causing the alter-
plagioclase and enriched in sericite (:j: clay ation probably diffused through the crystal
minerals). If the fluid is enriched in K+ (potas- lattice, probably along cleavage or micro-
sium metasomatism), mafic igneous rocks and cracks, and gave rise to selective alteration of
metabasites may be completely transformed to those zones at disequilibrium with the fluid.
mica-dominated assemblages, with little or no
Sodic and sodic-calcic (Na-Ca)
plagioclase, amphibole or pyroxene. If temper-
ature conditions are below that of biotite
stability (approximately 400°C), the assemblage Na(-Ca) metasomatism has been recorded from
ChI + Serc will dominate. environments of sea-floor metamorphism (e.g.

Replacement and overgrowth



FIG. 7.6 (a) The sericitisation of staurolite schist, giving rise to total pseudomorphs after staurolite. [Stt-Grt
schist, Glen Lethnot, Scotland. Scale bar = 0.5 mm (XPL).j (b) The sericitic alteration of andalusite (var. chiasto-
lite), in the contact metamorphic aureole of the Skiddaw Granite, England. Scale bar =0.5 mm (XPL).

Seyfried et al., 1988), in associatlOn with metasomatism is characterised by exchange of

porphyry copper deposits around acid- interme- Na for Ca or K, and to a lesser extent Ca for Fe
diate arc intrusives (e.g. Carten, 1986), and and Mg. The most characteristic mineralogical
within regions of extensive fluid infiltration change that occurs is formation of new albite,
associated with Proterozoic granitoids (e.g. De and replacement of pre-existing feldspar by
Jong & Williams, 1995; Oliver, 1995). The albite (i.e. ALBITISATION). Other possible

Overgrowth textures



FIG. 7.7 Chlorite-mica stacks developed in Devonian slates near Siouville, in the outermost part of the
Flamanville Granite contact metamorphic aureole, Normandy, France. (a) Standard optical microscopy (XPL),
Scale = 125 pm. The chlorite-mica stack (centre) comprises intergrown chlorite (dark) and muscovite (light) as a
probable replacement for detrital biotite. The matrix of the rock comprises rounded grains of quartz (white/light
grey) in a fine groundmass of phyllosilicates. (b) An SEM (back-scattered electron) image showing enhanced detail
of the chlorite-mica stack (light grey/white = Chi; medium grey = muscovite). The image also shows greater detail
of the matrix, comprising 10-20 pm length chlorite laths (light grey), and 20-100 pm quartz grains (dark grey) in
an ultra-fine (<5 pm) groundmass of white mica and quartz. Scale = 100 pm.

Replacement and overgrowth


FIG. 7.7 (contd) (c) A detail of the central part of the chlorite-mica stack shown in (b), showing individual layers
to vary from 1 pm to 15 pm in thickness. Scale = 10 pm.

changes, depending on the nature of the fluid of detrital biotites by metamorphic chlorite
and the P- T conditions, include conversion of (e.g. Jiang & Peacor, 1994) and the develop-
feldspars to zeolites (ZEOLITISATION), ment of other CHLORITE-MICA STACKS
replacement by scapolite (SCAPOLITISATION) (Fig. 7.7) in late diagenesis to low greenschist
or conversion to epidote (EPIDOTISATION). facies metamorphism (e.g. Warr et at., 1993;
In all cases the origin of the fluid is considered Merriman et at., 1995). Other prograde reac-
to be either magmatic, derived from reactions tion textures include the overprint of blue glau-
in meta-evaporites, of direct sea-water origin, cophanic amphibole rimming green calcic
or of some combination of these sources. amphibole in metabasic rocks, and the
mantling of kyanite by coarse muscovite laths
7.2 Overgrowth textures during prograde (or Mc + Sill in pelites (Fig. 7.8). The latter is
metamorphism related to the prograde breakdown of kyanite
to sillimanite in the upper amphibolite facies.
Although the evidence of retrograde reactions The sillimanite is more commonly observed to
is commonly preserved, the evidence of nucleate at quartz grain boundaries (Fig. 5.3),
prograde reactions is far less frequently or intergrown with biotite, rather than as a
encountered. Even so, there are several notable direct overgrowth on kyanite. Reaction rims of
examples in which prograde reaction rims and coarse muscovite have similarly been observed
intergrowth textures have been recognised. around andalusite as part of the prograde reac-
One such example is the prograde replacement tion to form kyanite. Such reaction rims should

Overgrowth textures

FIG. 7.8 The prograde breakdown of kyanite (at extinction) to coarse muscovite relating to the upper amphibo-
lite facies transformation Ky ~ Sil. This example is from a kyanite gneiss from the Caledonides of north Norway.
Scale =0.5 mm (XPL).

FIG. 7.9 The direct overgrowth of one AI,2SiO S polymorph on another. In this case, from an A1 2SiO s-Qtz vein
from Snake Creek (Queensland, Australia), kyanite is overgrowing andalusite. Scale bar =0.5 mm (XPL).

Replacement and overgrowth

not be confused with fine-grained sencltlc tions at the atomic scale: transmIssIon electron
rims and pseudomorphs commonly associated microscopy (ed. P.R. Buseck). Mineralogical Society
of America, Reviews in Mineralogy No. 27,
with the retrogression of aluminium-silicate Mineralogical Society of America, Washington, DC,
phases (Fig. 7.6). Direct overgrowth of one Ch. 8, 289-333.
Al 2SiO s polymorph by another (Fig. 7.9) is Beach, A. (1980) Retrogressive metamorphic processes
rarely seen in metamorphic rocks, but more in shear zones with special reference to the Lewisian
Complex. Journal of Structural Geology, 2,
normally two (and sometimes three) poly- 257-263.
morphs can exist meta stably together. The Boulter, C.A., Fotios, M.G. & Phillips, G.N. (1987)
reason for the coexistence of one or more of The Golden Mile, Kalgoorlie; a giant gold deposit
the polymorphs is largely due to the small localized in a ductile shear zone by structurally
induced infiltration of an auriferous metamorphic
differences in Gibbs free energy between each fluid. Economic Geology, 82, 1661-1678.
of them. This commonly leads to the Carten, R.B. (1986) Sodium-calcium metasomatism:
metastable persistence of one of the phases chemical, temporal, and spatial relationships at the
outside its stability field. Yerrington, Nevada, porphyry copper deposit.
Economic Geology, 81, 1495-1519.
The term EPITAXIAL OVERGROWTH (or Chopin, C. (1984) Coesite and pure pyrope in high-
EPITAXY) is used to describe cases of coherent grade blueschists of the Western Alps: a first record
overgrowth, in which a new phase preferen- and some consequences. Contributions to
tially nucleates on a crystallographically similar Mineralogy and Petrology, 86, 107-118.
De Jong, G. & Williams, P.J. (1995) Giant metaso-
pre-existing phase. This occurs because struc- matic system formed during exhumation of mid-
tural similarities between substrate phase and crustal Proterozoic rocks in the vicinity of the
overgrowth phase minimise the surface energy Cloncurry Fault, northwest Queensland. Australian
and allow easier nucleation. The example of Journal of Earth Sciences, 42, 281-290.
Diella, V., Spalla, M.I. & Tunesi, A. (1992)
one type of amphibole preferentially nucleating Contrasting thermochemical evolutions in the
and growing on another (Plate 4(a)), is a good Southalpine metamorphic basement of the Orobic
example of such overgrowth. It can occur as Alps (Central Alps, Italy). Journal of Metamorphic
both a prograde and retrograde feature, Geology, 10,203-219.
Jiang, W.-T. & Peacor, D.R. (1991). Transmission
depending on the amphiboles involved and electron microscopic study of the kaolinization of
their interrelationships. muscovite. Clays and Clay Minerals, 39,1-13.
A final example of a prograde overgrowth or Jiang, W.-T. & Peacor, D.R. (1994). Formation of
replacement feature is the case of sillimanite corrensite, chlorite and chlorite-mica stacks by
replacement of biotite in low-grade pelitic rocks.
(fibrolite) + biotite as a prograde pseudomor- Journal of Metamorphic Geology, 12, 867-884.
phic replacement of garnet (e.g. Yardley, 1977; Merriman, R.J., Roberts, B., Peacor, D.R. & Hirons,
Diella et aI., 1992). Such matted aggregates or S.R. (1995) Strain-related differences in crystal
'knots' of fibrolitic sillimanite are often referred growth of white mica and chlorite: a TEM and
XRD study of the development of meta pelitic
to as FASERKIESEL. This replacement of microfabrics in the Southern Uplands thrust terrane,
garnet by biotite and fibrolitic sillimanite Scotland. Journal of Metamorphic Geology, 13,
produces pseudomorphs appreciably enriched 559-576.
in Al and depleted in Fe, indicating extensive Okay, A.I. (1995) Paragonite eclogites from Dabie
Shan, China: re-equilibration during exhumation?
local-scale diffusive mass transfer, in a complex Journal of Metamorphic Geology, 13, 449-460.
'Carmichael-type' ionic reaction (Section 1.3.3). Oliver, N.H.S. (1995) Hydrothermal history of the
Mary Kathleen Fold Belt, Mt. Isa Block,
Queensland. Australian Journal of Earth Sciences,
References Peacock, S.M. (1987) Serpentinization and infiltration
Allen, C.M. (1992) Mineral definition by HRTEM: metasomatism in the Trinity peridotite, Klamath
problems and opportunities, in Minerals and reac- province, northern California: implications for


subduction zones. Contributions to Mineralogy and Warr, L.N., Primmer, T.]., & Robinson, D. (1993)
Petrology, 95, 55-70. Variscan very low-grade metamorphism in south-
Seyfried, W.E., Jr., Bernt, M.E. & Seewald, ].S. (1988) west England: a diastathermal and thrust-related
Hydrothermal alteration processes at mid-ocean origin. Journal of Metamorphic Geology, 9,
ridges: constraints from diabase alteration experi- 751-764.
ments, hot-spring fluids, and composition of the Wegner, W.W. & Ernst, W.G. (1983) Experimentally
oceanic crust. Canadian Mineralogist, 26, 787-804. determined hydration and dehydration reaction
Thompson, A.].B. & Thompson, ].F.H. (1996) Atlas of rates in the system MgO-SiOz-HzO. American
alteration (a field and petrographic guide to Journal of Science, 283-A, 151-180.
hydrothermal alteration minerals). Geological Yardley, B.W.D. (1977) The nature and significance of
Association of Canada (Mineral Deposits Division), the mechanism of sillimanite growth in the
St. Johns, Newfoundland, Canada, 119 pp. Connemara Schists, Ireland. Contributions to
Veblen, D.R. & Ferry, ].M. (1983) A TEM study of the Mineralogy and Petrology, 65, 53-58.
biotite-chlorite reaction and comparison with Zoltai, T. & Stout, ].H. (1984) Mineralogy: con-
petrologic observations. American Mineralogist, 68, cepts and principles. Burgess, Minneapolis,
1160-1168. Minnesota.

Part C

deformation and
Chapter eight

Deformed rocks
and strain-

Various effects of deformation, recovery and higher temperatures and lower strain rates
recrystallisation can be seen in thin sections of promote ductile deformation. That is not to say
deformed metamorphic rocks. The following that the change in deformation style occurs
sections give an introduction to such features, simultaneously in all minerals. The different
how they form and how they can be recog- crystallographic properties of individual phases
nised. For further details, the texts and edited mean that under a given set of conditions some
volumes by Poirier (1985), Barber & Meredith phases may undergo plastic deformation while
(1990), Knipe & Rutter (1990), Boland & others deform in a brittle manner. Both horn-
FitzGerald (1993), Passchier & Trouw (1996) blende and feldspar deform in a brittle manner
and Snoke et at. (in press), plus the review by at low to moderate temperatures, whereas at
Green (1992) on TEM analysis of deformation high temperatures crystal-plastic processes
in geological materials, are all invaluable operate. However, in the case of quartz, defor-
sources of reference. mation by crystal-plastic processes is charac-
teristic over a much broader range of
8.1 Deformation mechanisms conditions. This means that in granitoid rocks
sheared at low temperatures, quartz will
When a rock undergoes deformation due to exhibit features of ductile deformation, while
some superimposed stress, the mineral feldspars show brittle fragmentation (Fig. 8.3).
constituents of that rock may deform in either
a brittle manner (Fig. 8.1) by fracturing (=
CATACLASTIC FLOW), or in a ductile fashion 8.2 Inter- and intracrystalline deformation
by crystal-plastic processes (Fig. 8.2). The way processes and microstructures
in which a particular mineral deforms is influ-
enced by many factors, but especially by
8.2.1 Defects
temperature and strain rate. As a general rule,
lower temperatures and higher strain rates Plastic deformation of crystals is facilitated by
favour brittle deformation of minerals, while lattice defects, of which there are three main

Deformed rocks and strain-related microstructures

FIG. 8.1 The early stages of brecciation in sub-greenschist oolitic ironstone, Gwna Group, Anglesey, Wales.
Scale bar = 1 mm (PPL).

FIG. 8.2 Mylonite (an example of crystal-plastic (ductile) processes). Mylonitised granite, Baltic Shield, Sweden.
The coin is 22 mm in diameter.

classes; namely, POINT DEFECTS, DISLOCA- atoms or molecules in the lattice (lNTERSTI-
TIONS and GRAIN BOUNDARIES. Point TIALS). Point defects may migrate through the
defects (Fig. 8.4) can be subdivided into those crystal lattice by diffusive processes involving
that represent vacant sites in the crystal lattice exchange with neighbouring ions in a manner
(VACANCIES), and those representing extra obeying Fick's laws of diffusion (Section 1.3.5).

Inter- and intracrystalline processes

FIG. 8.3 Protomylonite. Low-temperature deformation of this granitoid rock has produced grain-size reduction
of quartz by ductile shearing, while giving rise to brittle fragmentation of feldspars (centre) in the same assem-
blage. Troms, Norway. Scale bar = 1 mm (XPL).

0 -0 -0 -0 -0
DISLOCATIONS are thermodynamically
unstable linear defects along which some slip

- O-
_I O xO- O-
has occurred (see below for details). In chemi-
cally homogeneous material (e.g. pure ice, salt,
and so on), GRAIN BOUNDARIES are effec-
tively two-dimensional defects separating
grains the lattices of which are differently
0 - 0 - 0 - _I0 - 0
oriented. In polycrystalline rock, the grain-
boundary region is somewhat more complex,
- O- O - O - O x
but in essence can be viewed as a complex two-
dimensional defect (see below).
0 - -0 -0 -0

8.2.2 Dislocations
-0 -0 -0 -0 -
FIG. 8.4 Point defects in a crystal lattice. x, vacancies;
Dislocations are contained within the crystal i, interstitia Is.
structure, and concentrated at grain bound-
aries. Such defects distort the lattice of the crys-
tal and introduce an internal strain. There are dislocation, but elsewhere the lattice planes line
two main types of dislocations. These are up.
EDGE DISLOCATIONS (Fig. 8.S(a)), where The greater the number of dislocations, the
the crystal has an additional half lattice plane, greater is the internal energy (stored elastic
and SCREW DISLOCATIONS (Fig. 8.S(b)), energy), and consequently if such dislocations
where part of the crystal is displaced by a can be eliminated a certain amount of energy will
lattice unit, giving a twisted lattice at the line of be released. Each dislocation is characterised by

Deformed rocks and strain-related microstructures

(a) / / / / / / / / /
/ / / / / / / / / /y /
/ / / / / / / / / v /v /
Vv y/
! Y Vy
! ./ /
t V /y /
. .... .... .... ....
t VVy
Edge dislocation Screw dislocation

FIG. 8.S A schematic illustration of different types of dislocation within the crystal lattice: (a) edge dislocations;
(b) screw dislocations. Thick arrows define the Burgers vector for each dislocation; small arrows define the dislo-
cation loop or Burgers circuit for each example (see the text for further details).

a slip vector or BURGERS VECTOR, defining planes in the crystal lattice, with the slip direc-
the direction and amount of lattice displace- tion defined by the Burgers vector. In quartz, for
ment. In the schematic illustration (Fig. 8.5(b)), example, basal slip is often the dominant mode
a 'square' circuit is shown around the disloca- of deformation. Depending on factors such as
tion, defined by an equal number of atoms on grain orientation, temperature and strain rate, a
each side (small arrows in Fig. 8.5(b)). This given crystal may have more than one slip
loop or 'Burgers circuit' is not closed because system active at any time. If dislocations can
of the step caused by the dislocation. The propagate fairly freely during ductile deforma-
connecting line across this step (solid arrow in tion, then STRAIN SOFTENING processes will
Fig. 8.5(b)) defines the Burgers vector. operate. In mechanical terms strain softening is
Dislocations may split into two or more partial expressed as 'a reduction in stress at constant
dislocations, which show misfit in relation to strain rate, or increase in strain rate at constant
the crystal lattice adjacent to the slip plane. stress' (White et ai., 1980). GEOMETRIC
This surface defining the zone of mismatch SOFTENING is one such softening process, and
between the partial dislocations and the adja- involves grain size reduction via intracrystalline
cent ordered crystal lattice is termed a STACK- slip and grain reorientation. Not only are the
ING FAULT. Such microstructural features are new grains smaller, but they are also strain-free.
especially common in minerals with a large Since quartz largely deforms by basal slip, this
unit cell distance (e.g. orthopyroxene). They leads to pronounced crystallographic alignment
preferentially develop whenever the combined of quartz in strongly deformed quartzofelds-
energy contribution due to misfit of the partials pathic mylonites. The basal slip planes become
plus the energy of the stacking fault is less than reoriented to lie close to parallel with the shear
the strain energy due to the single dislocation plane. This alignment can be recognised in thin
(Green, 1992). section by use of the sensitive tint plate (see
Dislocations playa vital role in ductile defor- Sections 8.2.7 & 10.8 and Plates 7(a) & (b) for
mation of rocks. The various minerals present details).
in a deforming rock each have their own SLIP The intersection of different slip systems will
SYSTEMS, representing preferential slip (glide) lead to entanglement of migrating dislocations.

Inter- and intracrystalline processes

Such DISLOCATION TANGLES make further by a tangle, or some other obstruction such as
deformation of the crystal increasingly diffi- an inclusion, the migration of vacancies to the
cult, and greatly contribute to overall STRAIN dislocation plane may allow the dislocation to
(WORK) HARDENING during deformation move at right-angles to the dislocation plane
of rocks at low temperatures or else relatively and thus enable it to by-pass the obstruction
fast strain rates. Although the migration of by a process known as CLIMB. The combina-
dislocations along a slip plane may be impeded tion of dislocation glide and climb is referred



FIG. 8.6 (a) Undulose (or undulatory) extinction in quartz. Deformed quartz-rich lens in a shear zone, 'Pyrite
Belt', Spain. Scale = 1 mm (XPL). (b) Sub-grain development (some examples arrowed), in a deformed quartz-rich
lens. Pyrite Belt, Spain. Scale =0.1 mm (XPL).

Deformed rocks and strain-related microstructures



FIG. 8.6 (contd) (c) New grains (some examples arrowed), in quartzitic mylonite. Pyrite Belt, Spain. Scale = 0.5
mm (XPL). (d) Mortar ('core-and-mantle') structure developed in mylonite. Pyrite Belt, Spain. Scale = 0.5 mm

to as DISLOCATION CREEP, a process of schists and gneisses) UNDULOSE EXTINC-

increasing importance at higher temperatures, TION (= undulatory extinction) is a feature of
and one which gives much greater mobility to some or many grains. This arises due to distor-
dislocations (see Section 8.2.3 for details). tion or bending of the crystal lattice, giving a
In most metamorphic rocks developed in high concentration of dislocations and other
association with deformation (e.g. mylonites, defects. It is frequently seen in many pre- and

Inter- and intracrystalline processes


FIG. 8.6 (contd) (e) Ribbon quartz in quartz mylonite. Moine Thrust, N.W. Scotland. Scale =0.5 mm (XPL).

syntectonic minerals, especially quartz (Fig. two slip systems must have operated 10 the
8.6(a)), feldspar, olivine, kyanite and mica. In formation of the kink bands.
thin section, UNDULOSE EXTINCTION is
recognised by a zone of extinction sweeping
8.2.3 Creep mechanisms
across the crystal as the stage is rotated. In
quartz, undulose extinction is generally elongate CREEP processes are extremely important
sub-parallel to [0001] (Carter et al., 1964) as a during crystal-plastic deformation of rocks
result of heterogeneous slip on this plane during and many other materials. Creep experiments
distortion of the lattice. Increased bending leads on rocks and ceramics are mostly undertaken
to the development of discrete DEFORMATION at strain rates (e) of 10-9 to 10-4 S- I , but in
BANDS (kink bands) which have sharply defined natural rock, where features of creep are
high-angle boundaries compared to zones of undoubtedly recorded, the strain rates are
undulose extinction. They have significant crys- much slower, and typically in the range 10- 15 to
tallographic mismatch relative to the main crys- 10- 12 S- 1 During deformation experiments it is
tal, and may terminate either at grain boundaries observed that many materials have a period
or inside grains. A detailed study by Mawer & over which strain rate is constant, known as
Fitzgerald (1993) on kink band boundaries in STEADY-STATE CREEP. At constant applied
quartz of a quartz ribbon mylonite showed that load, the material being tested often behaves in
rather than being single high-angle boundaries a manner approximating to constant strain
apparently representing simple rotations of rate. This implies creep with no overall strain
{OOOl} (as seen with light microscope), I-211m hardening, and is achieved by any strain hard-
wide strips existed at kink boundaries (TEM ening that occurs being counterbalanced by
study), with an intermediate orientation relative processes of recovery and recrystallisation
to the two limbs of the kink. From the geometry (Barber, 1990). This type of behaviour,
observed, the conclusion drawn was that at least referred to as POWER-LAW CREEP, where

Deformed rocks and strain-related microstructures

e oc an is characteristic in ceramics and rocks identified, namely, DISLOCATION CREEP

at temperatures above OATm (Barber, 1990), and DIFFUSION CREEP. In rock deformation
where T m is the melting temperature (in it is of course likely that more than one creep
Kelvin) of the material concerned (note that mechanism may operate at any given time, but
TITm is referred to as the homologous temper- depending on particular conditions one process
ature). At stresses of geological interest (a = will usually assume dominance.
1-100 MPa), experimental studies have shown DISLOCATION CREEP characterises defor-
that at moderately high temperatures the mation of rocks at low to moderate bulk shear
behaviour of almost all rocks and minerals is stress and over a wide range of temperatures
dominated by power-law creep. For example, and strain rates. It is a slow process involving
the experimental research by Ranalli (1982) propagation of dislocations through crystals by
for olivine assemblages demonstrated a clear a combination of glide and climb. Dislocations
power-law relationship, with stress exponent may glide rapidly until an obstacle is met
(n) of 3, for a ~ 200 MPa. which temporarily or finally halts any further
Having introduced key behavioural aspects propagation. Minor obstacles may be over-
of creep, let us consider the main processes come by thermal agitation, while for larger
involved. Two main creep processes have been obstacles diffusion-controlled climb may be


pressure ,.

SOlut\ion\ 13 /

14 I r-------~~----~~
15 ~ /1 /-~
-,,/ I / creep
I 15 o
6 \:: I


FIG. 8.7 A deformation mechanism map for quartz modified by the addition of a pressure solution field. The
region of dashed strain rate contours represents the inhibition of pressure solution through a decrease in pore-
water concentration. Grain size diameter, d = 100 lim; V =22 cm3; O'is the differential stress (O'tCO'.33); contours
of -log strain rate (after Fig. 9 of Rutter, 1976). Note that for a larger grain size (e.g. 1 mm), the coble creep/pres-
sure solution field shows appreciable contraction to lower temperature and lower differential stress (see Fig. 7 of
White, 1976; courtesy of The Royal Society).

Inter- and intracrystalline processes

required before they can be negotiated or elimi- 500°C. Grain-boundary diffusion in the pres-
nated (Poirier, 1985). Dislocation creep is an ence of a fluid is termed PRESSURE SOLU-
important process in low- to medium- grade TION, and it is dominant during diagenesis and
(e.g. greenschist facies) shear zones, but also at low-grade metamorphism (Fig 8.7). The process
higher metamorphic grades. At lower-tempera- involves stress-induced solution transfer of
ture, high crustal levels, rock deformation is material down chemical potential gradients
dominated by brittle failure. The change from along grain boundaries. In this case, the super-
brittle failure to crystal-plastic creep processes imposed stress causes material to be taken into
is referred to as the brittle- ductile transition. It solution at high-solubility sites and transported
is not a sharply defined changeover, because to low-solubility sites where it is precipitated.
the changeover from brittle to plastic behav- The distance over which transport occurs can
iour varies from one mineral to the next, and vary considerably, such as very local transport
thus from one rock to the next. For quartz-rich from high- to low-stress boundaries of individ-
rocks the changeover from dominantly brittle ual grains, to transfer over greater distances,
behaviour to dominantly plastic behaviour is depositing material in veins (Chapter 11) or
often approximated to basal greenschist facies even taking material out of the local system.
conditions (e.g. Sibson, 1977, 1990). Solution transfer processes are important in
DIFFUSION CREEP is an important process the development of crenulation cleavages
at various strain rates, over a wide range of (Section 4.4). In Fig. 4.6(b), a discrete crenula-
geological temperatures and shear stresses. tion cleavage is defined by thin dark pressure
Under such conditions it provides the driving solution seams of insoluble carbonaceous mate-
force for grain-boundary sliding. It is especially rial. STYLOLITES (Fig. 8.8) are irregular,
important in rocks of small grain size, whereas serrated or jagged pressure solution surfaces,
in coarse-grained rocks dislocation creep domi- which are particularly common in massive
nates over a wider range of conditions (White, carbonate and quartzite units from diagenetic
1976). Diffusion creep can be subdivided into to greenschist facies conditions. They generally
NABARRO-HERRING CREEP and COBLE develop perpendicular to principal compressive
CREEP. NABARRO-HERRING creep is the stress and are commonly defined by thin (typi-
dominant process at high temperatures and low cally 0.5-3.0 mm) seams of dark insoluble
shear stress (Fig. 8.7). By a combination of material. Stylolites are often considered to
grain-boundary sliding and diffusive trans- develop during compaction in the early stages
port of matter through the crystal lattice of diagenesis (Park & Schot, 1968). However,
and along grain boundaries, metamorphic tectonic stylolites are also recognised. These
rocks deforming at high-temperature condi- cross-cut bedding, and may intersect and offset
tions (e.g. granulite facies) can experience tectonic features such as veins (Ramsay &
major microstructural transformations involv- Huber, 1983).
ing grain shape changes and rearrangement
without intergranular cracks opening up.
COBLE CREEP is fluid-absent grain-boundary
8.2.4 Grain boundaries
diffusion. It typifies lower-temperature condi- Grain boundaries are the regions of contact
tions and a wide range of shear stress condi- between adjacent crystals. To be classed as true
tions. Because of the lower-temperature grain boundaries, the crystallographic misori-
conditions, grain-boundary diffusion predomi- entation of adjacent grains must be greater
nates over lattice diffusion; indeed, lattice than 100. Low-angle (<10 0 boundaries) are
diffusion becomes extremely inefficient below termed sub-grains, and characterise recovery

Deformed rocks and strain-related microstructures

FIG. 8.8 Stylolite development in vein quartz. Ashanti, Ghana. Scale bar =5 mm.

processes (see below). Grain-boundary condi- connecting the various voids and tubules (Fig.
tions in polymineralic rocks are considerably 8.9). In addition, mineral inclusions are
more complex than those of chemically homo- commonly observed along grain boundaries.
geneous materials. Even in largely monominer- Work by Watson & Brenan (1987) and others
alic rocks such as quartzites and marbles, there has shown that, depending on the wetting char-
are likely to be some impurities. The role of acteristics of the fluid concerned, the grain-
grain boundaries as sites for nucleation and boundary fluid may form a totally connected
growth of new minerals has already been network or else be isolated at grain triple-junc-
discussed in Section 5.1. tions or as tubes and inclusions along the grain
In metallurgy, the structure of grain bound- boundary. Empty or fluid-filled tubes and 'ellip-
aries is relatively well studied, but the study of soidal' inclusions are commonly recorded along
boundaries between rock-forming minerals has grain boundaries in various minerals (e.g. Spiers
only received detailed attention since the early et at. (1990) for rocksalt; and Craw & Norris
1980s. White & White (1981) described the (1993) for vein quartz), suggesting that fluids
grain-boundary regions of deforming rocks as utilise grain-boundary regions and have a role in
disordered regions comprising tubules at grain grain-boundary processes.
triple-junctions, that can form an interconnected The lattice misorientation introduced by
network through the grain aggregate, isolated grain boundaries can be considered as the
microscopic voids (pm-scale) along grain bound- misorientation introduced by a planar array of
aries, and a thin (nm-scale) film of distorted dislocations. During deformation and meta-
crystal structure (possibly with fluid present) morphism, the configuration of grain bound-

Inter- and intracrystalline processes

o - void /Inclusion
t - tubule
f - fluid film

FIG. 8.9 A schematic illustration of the microstructure of grain-boundary regions (modified from Fig. 4 in White
& White, 1981). Fluid may reside as JIm-scale inclusions along grain boundaries (possibly connected by nm-scale
fluid film), and as larger tubules (t) at grain triple-junctions.

aries becomes modified. The two most impor- experiments, Walker et al. (1990) concluded
tant processes in operation are those of that in situations of low shear stress, grain-
GRAIN-BOUNDARY SLIDING and GRAIN- boundary sliding is the dominant deformation
BOUNDARY MIGRATION. mechanism in calcite polycrystal aggregates. It
GRAIN-BOUNDARY SLIDING is move- is achieved largely by dislocation processes, but
ment within the plane of the grain boundary. It with a certain degree of diffusive mass transfer.
is a process that typifies regimes of creep by GRAIN-BOUNDARY MIGRATION, in-
diffusional flow, and is generally accompanied volving movement normal to the plane of the
by grain-boundary diffusion of material via a grain boundary, is stress-induced but aided
fluid film (see above). It occurs under applied greatly by elevated temperatures. The principal
shear stress and is most prevalent in the defor- factors controlling the rate of grain-boundary
mation of fine-grained aggregates at elevated migration are temperature, lattice orientation
temperatures and low stresses. It can be consid- and minor phases or impurities within the
ered in terms of the physical movement of indi- aggregate of grains. Temperature is important
vidual grains past each other, and is largely because at higher temperatures diffusive
achieved by the climb and glide of grain- processes become more important and thus
boundary dislocations, often with an accompa- atomic rearrangement in the grain-boundary
nying degree of boundary migration. Analogue region is easier. Lattice orientation is also
experiments using octachloropropane (e.g. Ree, found to have an important influence on grain-
1994) provide some evidence for the nature of boundary migration, with those grains with the
grain-boundary sliding and void development, same or very similar orientation having the
but perhaps the best evidence for grain-bound- least mobile boundaries. The role of boundary
ary sliding in natural materials comes from migration during grain coarsening, which
experimental work such as that of Walker et al. minimises the total free energy by decreasing
(1990), on deformation of synthetic calcite the surface energy contribution, was discussed
polycrystal aggregates. On the basis of their in Section 5.3, and will not be reiterated here.

Deformed rocks and strain-related microstructures

The influence of minor phases in the aggregate been used to estimate the stress responsible for
is clear from Fig. 5 .17(b), where micas have rock deformation in natural examples, since it
'pinned' quartz boundaries and thus inhibited appears that with stress decrease the sub-grain
migration during recrystallisation. size remains stable and is not modified. This
being the case, sub-grain size should be repre-
sentative of maximum stress experienced by the
8.2.5 Recovery
mineral, and thus offers considerable potential
RECOVERY includes an important set of as a palaeopiezometer.
processes that decrease the stored elastic energy Because fluid inclusions represent imperfec-
of the system. A deformed crystal has increased tions within crystals, they increase the internal
internal energy relative to its undeformed state energy of such crystals. TEM studies (e.g.
due to dislocations contained within the lattice. Reeder (1992) for carbonates; and Bakker &
The internal energy increase is directly propor- Jansen (1994) for quartz), have shown that
tional to the dislocation density (= combined fluid inclusions often have a close association
length of dislocations per unit volume) of the with dislocations (Fig. 11.16), and since dislo-
crystal. The lowering of energy by elimination cations may be eliminated by propagation
or ordering of dislocations by propagation into towards grain boundaries, the elimination of
existing grain boundaries and voids coupled fluid inclusions in a similar manner has often
with migration and climb of randomly been suggested as part of the recovery process,
arranged dislocations into stable arrays but especially during recrystallisation. An
('walls') at a high angle to active glide planes is important aspect of the experimental study of
an important part of the recovery process. Such Bakker & Jansen (1994) was the recognition
ordering leads to the development of strain free that during recovery minute quantities of fluid
SUB-GRAINS in larger crystals (Fig. 8.6(b)). leaked from the micron-scale fluid inclusions,
SUB-GRAINS are defined as areas with without rupturing the original inclusion. TEM
misorientations of a few degrees relative to the studies showed that the leakage was facili-
parent grain, and separated from the parent tated by dislocations, which display numerous
grain by dislocation walls. Because of the small nm-scale 'bubbles' of leaked fluid along their
difference in optical orientation between sub- length. Such features are of course far too
grain and parent grain, such boundaries are not small to be recognised with standard optical
sharply defined. Whole grains can be converted microscopy. Bakker & Jansen (1994) argue
to a mosaic of sub-grains, but especially that bubble nucleation on dislocations makes
common is the concentration of sub-grains at an important contribution to recovery,
grain margins. Their size varies from those sub- because each bubble eliminates part of the
grains clearly visible by standard microscopy dislocation, and thus the elastically strained
(Fig. 8.6(b)) to those minute sub-grains only atoms around that part of the dislocation. In
visible by SEMITEM studies. They develop so doing, the internal energy of the crystal is
during primary creep after relatively little diminished. The experimental work of
strain, and are a clear sign that RECOVERY Gerretsen et al. (1993) also demonstrated that
processes have operated. Experimental studies dislocation generation accompanying re-equi-
by Pontikis & Poirier (1975) and Ross et al. libration of fluid inclusions plays an impor-
(1980) have demonstrated an empirical rela- tant role in the deformation of 'wet' synthetic
tionship between sub-grain size and superim- quartz.
posed stress, the sub-grain size decreasing as DEFORMATION LAMELLAE (Fig. 8.10)
the applied stress increases. This in turn has are narrow crystallographic ally oriented

Inter- and intracrystalline processes

FIG. 8.10 Deformation lamellae in quartz from deformed metaconglomerate. Cherbourg region, Normandy,
France. Scale bar = 125/Jm (XPL).

planar features of dO /lm width that are which slip systems may have operated, but
commonly seen in deformed quartz (e.g. that the presence of sub-basal deformation
Carter et at., 1964; Drury, 1993). They are lamellae in quartz can be interpreted in terms
visible with the light microscope, and until the of dynamic recovery of dislocations initially
mid-1970s, were considered to represent dislo- present in slip bands.
cation slip bands. On this basis, they were
used to evaluate the dominant slip systems
8.2.6 Recrystallisation
operating at particular strain rate and P-T
conditions. TEM studies by McLaren et al. RECRYSTALLISATION is the natural progres-
(1967) and others revealed that many of the sion from recovery processes, and minimises
observed defect substructures could be directly the energy of the system still further, by stress-
equated with the fine lamellae seen with induced grain-boundary migration. This amal-
an ordinary petrological microscope. gamates smaller grains into larger ones, so
Experimentally deformed quartz lamellae are reducing the surface energy of the system, and
typically defined by dislocation slip bands, eliminates dislocations within crystals to reduce
Brazil twins and zones of glass. However, the the internal energy. These changes create a
review by Drury (1993) points out that TEM more stable arrangement of grains and grain
studies of naturally produced lamellae usually boundaries, and thus decreases the energy of
show them to be defined by elongated sub- the system as a whole. Elevated temperatures
grains, sub-grain walls, and zones with vari- and the presence of a grain boundary fluid
able densities of dislocations and sub-micron greatly aid such transformations.
fluid inclusions ('bubbles'). In view of this, DYNAMIC RECRYSTALLISATION is the
Drury (1993) concluded that natural deforma- term used for recrystallisation synchronous with
tion lamellae cannot be used to determine deformation. An important component of

Deformed rocks and strain-related microstructures

dynamic recrystallisation is sub-grain rotation. It MORTAR STRUCTURE) is commonly

was described above how, as part of the recov- observed. This consists of large strained porphy-
ery process, dislocations migrate to form stable roclasts of quartz or feldspar surrounded by a
'walls' which define boundaries of sub-grains. fine-grained aggregate of recrystallised new
These sub-grains have lattice misorientations of grains (Fig. 8.6(d)). Similar features of dynamic
a few degrees relative to the parent grain, but if recrystallisation are also observed in certain
climb-accommodated dislocation creep contin- sheared amphibolites, eclogites and marbles. In
ues, the further addition of dislocations at the such cases, the porphyroclasts and fine-grained
sub-grain boundaries leads to gradual rotation aggregates are of hornblende, clinopyroxene and
of the sub-grain and increasing mismatch calcite respectively.
between sub-grain and parent grain. Once a Another mechanism by which NEW
high-angle contact (>10°) has been established, GRAINS are produced during dynamic recrys-
the term 'sub-grain' is no longer appropriate, tallisation involves grain-boundary migration
and the term NEW GRAIN is used. New grains to isolate lobes of irregular or serrated grain
are a characteristic feature of dynamic recrys- boundaries. Irregularly SUTURED or SER-
tallisation and, although similar to sub-grains, RATED boundaries mItIate by boundary
they can be distinguished by virtue of their sharp migration, causing BULGING of one grain into
contacts with adjacent grains (Fig. 8.6(c)). In its neighbour in order to eliminate areas with
cross-polarised light, this is clearly seen by high dislocation density. In such cases the grain
abrupt changes in interference colours between with the more stable, low dislocation density
adjacent grains, compared to the slight and margin always bulges into the grain with less
gradual changes associated with sub-grains. In stable high dislocation density boundary, so
sheared quartzites and quartzofeldspathic rocks reducing the internal energy of the system. This
CORE-AND-MANTLE STRUCTURE (= bulging of grain boundaries can ultimately give

FIG. 8.11 Bulging of grain boundaries in recrystallised vein quartz from Snake Creek, Queensland, Australia.
Scale bar = 0.5 mm (XPL).

.... Plate la .. Plate Ie .... Plate Ib .. Plate Id
<III Plate 2a ... Plate 2b

.. Plate 2c

Plate 2d ~

<III Plate 2e .. Plate 2f

<0lIl Plate 3a .& Plate 3b

T Plate 3c

Plate 3d ~

<0lIl Plate 4a T Plate 4b

~ Plate 4c .. Plate 4d

~ Plate 4e

Plate 4f~

~ Plate 4g ~ Plate 4h
.... Plate 5a • Plate 5b

T Plate 5c

Plate 5d ~

.... Plate 5e T Plate 5f

<III Plate 6a • Plate 6b

T Plate 6c

Plate 6d ~

<III Plate 6e T Plate 6f

~ Plate 7a ... Plate 7c ~ Plate 7b ... Plate 7d
..... Plate 8a .. Plate 8b

T Plate 8c

Plate 8d ~

..... Plate 8e T Plate 8f

Inter- and intracrystalline processes

rise to new grains, especially if the bulge devel- In the case of granular aggregates of more or
ops a narrow neck (Fig. 8.11). The neck less equant grains (e.g. quartz, calcite and
becomes the likely site for sub-grain wall devel- olivine), dislocation creep (Section 8.2.3) is an
opment, and if rotation occurs a new grain important process in changing crystal shape, as
with a high-angle boundary forms. For further well as orientation relative to neighbouring
details on dynamic crystallisation, the reader is grains. Deformation twinning will also playa
referred to the paper by Drury & Urai (1990). role in the development of both grain-shape
STATIC RECRYSTALLISATION occurs fabrics and CPOs. If an aggregate of grains
post-deformation and takes place in response with random crystallographic orientation exists
to elevated temperature conditions promoting prior to deformation, when subjected to a
further re-equilibration and energy reduction. certain amount of deformation (whether coax-
The principal way in which this is achieved is ial or non-coaxial), there will be a gradual
by lowering the surface energy of the system by change to some form of CPO. Different miner-
reducing total grain-boundary area. This als have different slip and deformation-twin
involves grain-boundary migration to smooth systems. For some minerals relatively few slip
out irregular grain boundaries within the systems operate, whereas for others numerous
aggregate, and to eliminate small grains by slip systems may operate, and thus the resul-
amalgamation. Elimination of dislocations and tant CPO after deformation will be less
other lattice defects is an integral part of the straightforward to interpret. It can be appreci-
process, and if the aggregate is completely ated that the nature of CPOs in polyphase
recrystallised, crystals showing evidence of aggregates is more complex because of the
strain (e.g. undulose extinction) should be interaction of crystals with different slip
absent. A polygonal-equigranular aggregate systems. The precise conditions at the time of
(e.g. Fig. 5.14(b)) is a prime example of the deformation will influence which slip systems
microstructural arrangement characteristic of are active, and as a general rule increasing
static recrystallisation. differential stress will increase the number of
active slip systems. The more intense and
8.2.7 Crystallog raphic-preferred prolonged the deformation is, the greater the
orientations tendency is for a more pronounced CPO.
However, this may not be entirely true, because
Crystallographic (lattice )-preferred orientations the degree of dynamic recrystallisation may
(CPOs) are a common feature of highly serve either to weaken or improve the CPO.
deformed rocks, and in many cases the CPO Jessell & Lister (1990) examined the influence
has a close relationship with the grain shape of temperature on quartz fabrics.
orientation. For example, biotite grains in It is outside the scope of this text to discuss
schists show a strong alignment of their {OOl} the full range of CPOs that have been identified
planes to define the schistosity, and since the in relation to different mineral aggregates and
{OOl) planes contain the long axes, the grain in different deformation regimes. However, for
shape orientation is often aligned, especially in further insight, useful overviews are given by
L-S tectonites (Section 4.3; Fig. 4.4). However, Law (1990) and Pas schier & Trouw (1996).
in other cases, most notably highly recrys- Passchier & Trouw (1996) also provide a
tallised mylonitic rocks, the original grain- useful description of the procedure for use of
shape fabric may have been largely lost during the universal stage (U-stage) in order to
recrystallisation processes, but a residual CPO measure quartz or calcite c-axes, the two
may still remain. minerals for which CPOs have been most

Deformed rocks and strain-related microstructures

extensively studied. Although CPOs have a related foliated rocks. As discussed in an earlier
number of potential uses for interpreting the section, the so-called "brittle-ductile' transition
structural evolution of a particular rock, the occurs at approximately 300°C (10-15 km
complexity of deformation often makes it diffi- depth) for quartz-rich rocks, which corre-
cult to unravel the story, especially if late-stage sponds broadly with the lower-temperature
recrystallisation has been strong. One of the boundary of greenschist facies conditions. In
most useful applications (especially quartz c- broad terms, this approximates to the base of
axes) has been in the interpretation of amount the shallow seismogenic zone of the crust (seis-
and sense of shear during non-coaxial deforma- mic-aseismic transition). It should be empha-
tion (Section 10.8 and Plates 7(a) & (b)). sised that the depth of transition will not be the
same for all rocks, but will vary according to
their rheological properties. Rutter (1986)
8.3 Fault and shear zone rocks and their
makes the important point that broad use of
the term 'brittle-ductile' transition can be
Rocks of fault zones and shear zones can be potentially misleading, since 'ductility' is not
subdivided into foliated types (i.e. mechanism dependent, but simply reflects
MYLONITES and PHYLLONITES) and non- substantial non-localised strain. He advocates
foliated types (termed CATACLASITES), in- expressions such as 'brittle-plastic' or 'cataclas-
cluding such rocks as FAULT BRECCIAS. tic-plastic' in order to give a precise indication
At high crustal levels localisation of high of the change in deformation mechanism asso-
strain rates gives rise to brittle faulting. With ciated with mode of failure transition in rocks.
the exception of fault gouge, which is often Wise et al. (1984) introduced a classification
well foliated, faulting at high crustal levels of deformed rocks in terms of strain rate versus
generates non-foliated fault rocks consisting of recovery rate. The rocks thus far described are
variable-size rock fragments in a finer grained all placed in the field of high strain rate and
matrix. Sibson (1977) draws the distinction low recovery rate. Rather than use the term
between cohesive and incohesive fault rocks. 'crush breccia' to describe coherent non-foli-
On the basis of the proportion of matrix, he ated fault rocks with dO% matrix, terms such
subdivides incohesive non-foliated fault rocks as 'silicified fault breccia' and 'carbonate-
into FAULT BRECCIAS (>30% visible frag- cemented fault breccia' (depending on the
ments) and FAULT GOUGE (dO% visible nature of the cement) are preferred. This is due
fragments). Cohesive non-foliated fault rocks to the fact that fault breccia is already widely
are also subdivided on the basis of matrix used in the literature for both coherent and
proportion and range from 'CRUSH BREC- incoherent rocks of this type.
CIAS' (0-10% matrix) through PROTOCATA- A special type of cataclasite formed by rapid
CLASITE (10-50% matrix), CATACLASITE fault movement at high crustal levels (e.g.
(50-90% matrix) to ULTRACATACLASITE earthquake-related) is known as PSEUDO-
(90-100% matrix). CATACLASIS is a process TACHYLITE (Fig. 8.12). This localised and
involving brittle fragmentation and rotation of relatively rare rock type generally occurs in
mineral grains. During the development of a narrow zones (mm-cm scale) and commonly
cataclasite this is accompanied by grain-bound- displays irregular mm-scale injection vein lets
ary sliding and diffusive mass transfer mecha- off the main surface. It comprises fine frag-
nisms (e.g. Lloyd & Knipe, 1992). At deeper ments in a dark glassy groundmass (black in
levels crystal-plastic processes operate in transmitted light). Sibson (1977) describes it as
ductile shear zones to generate mylonites and forming due to rapid movement inducing ther-

Fault and shear zone rocks

mal fragmentation and frictional melting (T nition by Lapworth in 1885 have been the focus
probably > 1000°C) under dry conditions at of attention for many structural and metamor-
depths greater than 1 km but less than about phic geologists. The rock 'mylonite' is best
10 km. For further details on pseudotachylite defined as a cohesive, foliated and usually
microstructures, see Maddock et ai. (1987) and linea ted rock produced by tectonic grain-size
Lin (1994) . reduction via crystal-plastic processes in narrow
MYLONITES are rocks of considerable zones of intense deformation. It contains abun-
tectonic significance, and ever since their recog- dant relict crystals, 'porphyroclasts', (10-50%



FIG. 8.12 (a) Pseudotachylite. Hetai mine, Guangdong, China. Scale bar = 0.5 mm (PPL) . (b) Pseudotachylite
cutting Qtz-Pl-Bt gneiss. Butt of Lewis, Scotland. Scale bar = 1 mm (PPL).

Deformed rocks and strain-related microstructures

of the rock) which characteristically are of 'granitoid mylonites', 'carbonate mylonites',

similar composition to the matrix minerals. It 'amphibolitic mylonites', and so on. Highly
is not a term restricted to a particular composi- sheared rocks dominated by phyllosilicate
tional range of rocks. Depending on the minerals are termed PHYLLONITES. These
observed mineralogy, it is thus possible to have are characterised by S-C fabrics formed



FIG. 8.13 (a) Protomylonite: single and polycrystalline quartz porphyroclasts surrounded by a fine-grained
matrix of quartz and sericite. Arran, Scotland. Scale = 1 mm (XPL). Note the serrated grain boundaries in the
centre of the photograph, the undulatory extinction exhibited by several porphyroclasts, and the deformation
bands displayed by the crystal in the top right of the photograph. (b) Mylonite: porphyroclasts of quartz and
feldspar in an ultra-fine-grained matrix of quartz and sericite. Ghana. Scale = 1 mm (XPL).

Fault and shear zone rocks


FIG. 8.13 (contd) (c) Ultramylonite. Abisko, Sweden. Scale = 1 mm (XPL). This quartzitic ultramylonite shows a
characteristic lack of porphyroclasts and ultra-fine grain size common to many ultramylonites. The quartz shows
very good crystallographic alignment (see Plates 7(a) & (b)).

synchronous with shearing. The intersection of fine-grained matrix. As a general rule, the
these fabrics produces a characteristic 'button recrystallised grain size decreases as differen-
schist' or 'oyster-shell' texture (see Section 10.3 tial stress increases. MYLONITE (sensu
for further details). stricto) (Fig. 8.13(b)) lies between these two
Mylonites are associated with thrusts and end-members and has 10-50% clasts. Those
shear zones, generally operating at deeper mylonites involving extensive recrystallisation
crustal levels than those responsible for the and mineral growth synchronous with shear-
development of cataclasites. At such depths ing are often referred to as BLASTO-
deformation is more ductile, and crystal-plastic MYLONITES.
processes predominate. Mylonites are generally In strongly deformed rocks such as
associated with relatively high strain rates mylonites, cataclasites and gneisses, it is
coupled with appreciable recovery rate. common to observe large relict crystals in a
Estimates of the rates of microstructural finer-grained matrix. This IS known as
changes in mylonites have been made by Prior PORPHYROCLASTIC MICROSTRUCTURE
et al. (1990), with particular reference to the (Figs 8.13(a) & (b)), and the relict crystals as
Alpine Fault Zone, New Zealand. PORPHYROCLASTS. It is important to appre-
Like cataclasites, MYLONITES (sensu lata) ciate the difference between porphyroclastic
can be classified in terms of matrix: porphyro- structure, which consists of large relict crystals,
clast ratios. PROTOMYLONITE (Fig. 8.13(a)) from porphyroblastic structure (Fig. 5.4) which
consists of abundant (50-90%) clasts, while at consists of large newly grown crystals.
the other end of the spectrum highly sheared Porphyroclasts are generally of the same
ULTRAMYLONITE (Fig. 8.13(c)) has dO% minerals as those present in the matrix.
clasts, and these are usually small, in a very However, the proportions of different phases

Deformed rocks and strain-related microstructures

present as porphyroclasts may differ with dislocation climb to be an effective recovery

respect to the matrix according to the mechan- mechanism. Instead, the principal mode of
ical behaviour of individual minerals under the recovery is grain boundary migration recrys-
prevailing conditions. Minerals such as tallisation (Hirth & Tullis, 1992). At increased
feldspar, kyanite, hornblende and garnet, that temperature or decreased strain rate, undulose
deform in a brittle manner over a wide range extinction is pronounced, and sub-grains are a
of conditions, commonly form porphyroclasts, prominent feature at grain boundaries. Under
while phyllosilicates and epidote minerals are these conditions, dislocation climb has become
often more important as matrix constituents. a more effective recovery mechanism, and
In less sheared rocks the porphyroclasts are dynamic recrystallisation takes place by sub-
generally larger and more angular, becoming grain rotation.
smaller and more rounded as deformation In experiments on quartz aggregates at
intensifies. Depending on the extent of recrys- temperatures of approximately 700°C and fast
tallisation, porphyroclastic rocks such as strain rates (10-6 S-I), recovery and recrystalli-
mylonites commonly display a mix of sation are seen, facilitated by dislocation creep
microstructural features indicative of deforma- processes. Some deformation lamellae occur,
tion (e.g. undulose extinction and deformation but the widespread development of small
bands), recovery (e.g. sub-grains and deforma- recrystallised new grains at boundaries of origi-
tion lamellae) and recrystallisation (e.g. new nal grains becomes a prominent feature.
grains and core-and-mantle structure). The Unrecrystallised grains show undulose extinc-
nature and attitude of porphyroclasts and their tion instead of sharp deformation bands, have
'tails' can be useful in determining shear sense sutured boundaries and become progressively
(Section 10.6). flattened as strain increases. Tullis (1990)
points out that at the pressures and strain rates
of most experiments dislocation creep
8.3.1 Deformation of quartzitic and
processes initiate at much higher temperatures
quartzofeldspathic rocks
than is the case for natural pressures and low
Experimental work such as that of Tullis et al. strain rates (e.g. 10-14 S-I). At high tempera-
(1973) and Hirth & Tullis (1992), and the tures (> 800°C), complete, or near complete,
detailed overview given by Tullis (1990), docu- recrystallisation is observed. Very few original
ments the progressive changes in mechanisms grains remain, most having been transformed
and microstructures of quartzitic and quart- to an aggregate of small recrystallised grains
zofeldspathic aggregates during deformation and sub-grains. Vernon (1976) describes how,
under various temperature, strain and strain rate during experiment, syntectonic recrystallisation
conditions. These transformations in deforma- of highly deformed quartzite (axial compres-
tion mechanisms and resultant microstuctures, sion > 50%) produces finely recrystallised
while focusing on quartz-rich rocks, are gener- RIBBON QUARTZ, which is identical to micro-
ally applicable to a wide range of materials. structures of many natural quartz-mylonites
At low temperatures, low strains, but fast (Fig. 8.6( e)}.
strain rates, deformed quartz aggregates The observations and conclusions from
display patchy undulose extinction and limited studies focusing on the progressive microstruc-
sub-grain development and recrystallisation at tural changes seen in natural quartz mylonites
grain boundaries. Under these conditions in as a product of ductile shearing and myloniti-
deforming quartz aggregates, dislocation sation (e.g. White et aI., 1980; Knipe, 1990)
production is too fast for diffusion-controlled have much in common with the findings from

Fault and shear zone rocks

experimental work. In the early deformation of of P-T conditions, strains and strain rates (e.g.
quartzofeldspathic mylonites, intracrystalline Brodie & Rutter, 1985; Skrotzki, 1990; Rutter
slip occurs and quartz develops undulose & Brodie, 1992; Lafrance & Vernon, 1993;
extinction. Increased ductile deformation leads Stiinitz, 1993). The next few paragraphs
to the development of sub-grains, deformation summarise the main observations and conclu-
bands and deformation lamellae in quartz, sions from these studies.
while large feldspar grains show brittle frag- The dominant minerals of mafic rocks vary
mentation and may show deformation twins according to P-T conditions, and consequently
(Section 5.6.3). Continued recovery and the nature of rock deformation and resultant
dynamic recrystallisation reduces internal dislo- microstructures varies according ro the specific
cation density by development of sub-grains assemblage and the conditions under which
and the formation of new grains. Serration and deformation occurs. Plagioclase is a major
new grain development at porphyroclast constituent of mafic rocks under all conditions
margins are widespread. If the temperature, or except the eclogite facies (Cpx + Grt assem-
the amount of strain, is high during deforma- blages), while hornblende is a key phase at
tion, elongation of grains will occur to give mid- to lower crustal conditions. Pyroxenes
ribbon quartz texture. assume dominance in the lower crust (granulite
facies), whereas towards the upper crust, green-
schist facies assemblages such as
8.3.2 Deformation of mafic rocks
Chl-Act-Ep-Ab-Qtz are widespread. In mafic
Over the past two decades there have been a and ultramafic rocks of the upper mantle,
number of important studies concerning the olivine-pyroxene assemblages dominate,
deformation processes and microstructures of whereas primary igneous rocks such as gabbro
deformed mafic rocks (Fig. 8.14) over a range have PI-Cpx assemblages.

FIG. 8.14 Mafic mylonite: white porphyroclasts are of plagioclase, while other porphyroclasts are of hornblende.
Lewis, Scotland. Scale bar = 0.5 mm (XPL).

Deformed rocks and strain-related microstructures

While an understanding of deformation from the Sesia Zone, western Alps, found that
processes prevalent under peak conditions is as deformation intensifies there is quite
important, it is also essential to recognise the contrasting behaviour between the mafic
microstructures and processes involved in the minerals and plagioclase. Irrespective of the
deformation of high P-T mafic assemblages at intensity of deformation, the mafic minerals
different conditions of P, T, stress and strain. deform by fracturing. Synchronous with the
At low-temperature conditions, minerals such deformation, the pyroxene and hornblende
as pyroxene, hornblende, plagioclase and crystals break down to actinolite, but also
garnet deform by cataclastic processes, involv- exhibit microfracturing. Fractured plagioclase
ing micro fracturing of individual grains. porphyroclasts are also observed, but the fine-
However, at elevated temperatures and/or grained products of plagioclase recrystallisation
strain rates, empirical observation and experi- and retrogression (albite and zoisite) are not
mental work has demonstrated that plagioclase fractured. These fine-grained aggregates are
is considerably more deformable (softer) than interpreted as playing a crucial role in initiating
hornblende, pyroxene or garnet. This means a change from deformation, largely by fractur-
that in highly sheared bi-mineralic or polymin- ing to bulk deformation predominantly by
eralic mafic assemblages, plagioclase often viscous flow. Stiinitz (1993) noted this change
shows evidence of considerable intracrystalline in mechanism in the moderately deformed
plastic deformation (facilitated by dislocation metagabbros with S-C fabrics, but recorded it
glide), whereas amphiboles and pyroxenes in as most pronounced in the more intensely
the same assemblage display much less mylonitic rocks. The shift to viscous flow
evidence for intracrystalline plasticity (e.g. deformation in the fine-grained albite-zoisite
Brodie, 1981; Rutter & Brodie, 1992). In terms aggregates also led to focusing of subsequent
of the rock microstructure, this means that deformation into zones enriched in such miner-
while the matrix and porphyroclast phases will als (i.e. formerly plagioclase-rich zones), and
be more or less identical, there will be a suggests that plagioclase retrogression, coupled
tendency for a dominance of resistant phases with development of chlorite-rich assemblages,
such as hornblende, pyroxene, and/or garnet as is a crucial part of the 'softening' process in the
porphyroclasts, while softer phases such as deformation of mafic rocks at greenschist facies
plagioclase will be more significantly repre- conditions.
sented in the flowing matrix (e.g. Rutter & Lafrance & Vernon (1993) examined gab-
Brodie, 1992). broic mylonites and ultramylonites deformed
In addition to mechanical processes, a under low- to mid-amphibolite facies condi-
number of mineralogical (chemical) changes tions. They recorded extensive recrystallisation
also take place as high-temperature mafic of plagioclase by grain-boundary migration,
minerals and Ca-plagioclase react to form more but noted that plagioclase-rich layers have a
stable lower-temperature phases. These retro- strong crystallographic preferred orientation.
grade reactions generally take place at green- Polygonal pyroxenes at the margins of larger
schist facies conditions in the presence of an pyroxene porphyroclasts, and the general
aqueous fluid. Indeed, the fluid presence is lack of sub-grains also indicates pronounced
essential for the formation of hydrous phases. high-temperature recrystallisation involving
The interrelationships between deformation crystal-plastic processes, but amphiboles
and metamorphism are discussed more fully in within the gabbroic mylonites show brittle
Sections 8.4 & 8.5. fragmentation, with a preference for intragran-
The study by Stiinitz (1993) of metagabbros ular fracturing along the {110} cleavage. Many

Fault and shear zone rocks

of the amphibolite facies metabasic mylonites Solnhofen limestone, or else on synthetic pure
studied by Lafrance & Vernon (1993) show calcite rock (e.g. Schmid et al., 1987; Walker et
pronounced (mm-scale) differentiation into al., 1990; Rutter et al., 1994; Rutter, 1995), in
layers rich in plagioclase and layers rich in order to establish the behaviour of carbonate
mafic minerals. Although dislocation creep is rocks and the dominant deformation mecha-
dominant in amphibolite facies mylonites, nism(s) over a broad range of conditions rele-
fracturing of amphiboles and solution-transfer vant to geological situations. As well as
processes also play an important role. Brodie showing variations according to different
& Rutter (1985) and Skrotzki (1990) have conditions of P, T, superimposed stress and
studied high-temperature (T > 650°C) amphi- strain rates, the behaviour of carbonate rocks is
bolitic mylonites in shear zones from the Ivrea shown to be strongly influenced by grain size.
Zone, north-west Italy. The TEM work of In fine- and ultrafine-grained carbonate
Skrotzki (1990) shows that hornblendes of the mylonites, the temperature increase from
shear zone comprise recrystallised grains, sub- 300°C to 700°C under experimental conditions
grains, free dislocations and abundant stacking marks a progressive change from grain-size
faults. These microstuctures indicate both insensitive crystal-plastic flow processes to
recovery and recrystallisation and are consis- grain-size sensitive superplastic flow (e.g.
tent with dislocation creep being the dominant Rutter et al., 1994). The experiments of Rutter
deformation process. The lack of any signifi- et al. (1994) found that although strong low-
cant difference in microstructures between temperature fabrics in fine-grained aggregates
porphyroclasts and matrix suggests wide- become weakened during high-temperature
spread dynamic recrystallisation. recrystallisation and crystal-plastic or super-
plastic flow, they do survive to some extent. In
8.3.3 Deformation of carbonate rocks their simple shear experiments on Solnhofen
limestone (grain size "" 4 pm) and Carrara
The study of deformed carbonate rocks (both marble (grain size "" 200 pm), Schmid et al.
natural and experimental) has been a subject of (1987) identified four distinct microfabric
interest for many decades, but has been the regimes. The first of these, the twinning regime,
focus of considerable research in recent years was observed from room temperature to 400°C
(e.g. Dietrich & Song, 1984; Schmid et at., in Solnhofen limestone, and at temperatures ~
1987; Wenk et al., 1987; Burkhard, 1990; 600°C in Carrara marble at shear stresses> 80
Walker et al., 1990; van der Pluijm, 1991; MPa. In these simple shear experiments, most
Rutter et aI., 1994; Rutter, 1995; Busch & van grains displayed a single set of e-twins, whereas
der Pluijm, 1995). With so much literature on in coaxial testing conjugate sets are often devel-
the subject, it is difficult to give an adequate oped (Schmid et aI., 1987). At higher tempera-
review of all the research findings that have been tures and/or lower strain rates, Schmid et al.
made, but it is hoped that the description below (1987) found that twinning was absent and
provides a useful summary of some of the obser- that a regime of intracrystalline slip prevailed,
vations regarding deformation mechanisms and and that original serrate grain boundaries
microstructures over a range of conditions. remain as such, with no evidence of grain-
boundary migration. Carrara marble
Experimental data commonly displayed core-and-mantle struc-
ture, with sub-grains in grain-boundary
A considerable amount of experimental work regions. The marked difference in temperature
has been undertaken on Carrara marble and defining the transition from twinning to

Deformed rocks and strain-related microstructures

intracrystalline slip (:::: 400°C in Solnhofen lime- modifying grain size. Where stresses are too
stone and:::: 700°C in Carrara marble, at labo- low for twinning to develop, grain-boundary
ratory strain rates) is directly related to the migration recrystallisation occurs, and leads to
grain size of the starting material. Thus, overall grain coarsening throughout the rock
coarser-grained calcite rocks display twinning (Rutter, 1995). This recrystallisation involves
to higher temperatures. In the ultra fine-grained the development of nuclei at grain-boundary
Solnhofen limestone, Schmid et al. (1987) bulges, and subsequent sub-grain rotation.
observed a gradual transition into a grain- Such behaviour typified experiments at
boundary sliding regime at high temperatures temperatures in the range 700-900°C, and
(700-900°C), but did not record this in the thus compares well with the findings of
Carrara marble. In this regime, the observed Schmid et al. (1987).
microstructural features include (a) straight With regard to coarser grain sizes, Walker et
grain boundaries in place of originally serrate al. (1990) suggest that at high temperatures,
boundaries, and (b) weak or completely absent grain-size sensitive flow may extend to coarse
grain-shape fabric. Both of these grain-bound- calcite aggregates (grain size > 1 mm), such
ary equilibration features will have involved that equigranular calcite aggregates of amphi-
grain-boundary migration, and there is bolite facies marbles may be no different from
evidence of grain growth (Schmid et al., 1987). calcite aggregates formed by static grain
Direct microstructural evidence for grain- growth.
boundary sliding was not recorded by Schmid
et al. (1987). However, it was inferred as the Natural examples
dominant mechanism in Solnhofen limestone at The work of Burkhard (1990) on deformed
high temperatures during simple shear, because micritic limestones from the Helvetic nappes,
of microstructural features virtually identical to Switzerland, provides a good example of the
those seen in the coaxial experiments of types of deformation mechanisms that operate
Schmid et al. (1977), where rheological and in carbonate rocks over the range 150-350°C.
microstructural considerations led the authors Despite large bulk strain and low temperature
to conclude that under the specific high- « 300°C), micritic limestones (grain size 3-6
temperature conditions grain-boundary sliding pm) generally lack any crystallographic
would be the dominant deformation mecha- preferred orientation (CPO), and show no signs
nism. During experiments on Carrara marble of recrystallisation. The principal deformation
at 800-900°C, significant grain growth process inferred to be operating at these condi-
occurred as a result of dynamic recrystallisa- tions is grain-boundary sliding, assisted by
tion. This was facilitated by grain-boundary diffusive-transfer processes. A single low-
migration, and represents the fourth regime temperature (d80°C) mylonitic fault rock
identified by Schmid et al. (1987). proved an exception to the rule and displayed
Rutter (1995) undertook a series of experi- substantial grain-size reduction (grain size 1-3
ments on Carrara marble, with a mean grain pm): this was interpreted as recrystallised sub-
size of 130:1:29 pm, a confining pressure of grains and thus indicative of substantial
200 MPa, a strain rate of 10-4 s-1 and temper- intracrystalline slip or creep. In the low green-
atures in the range 500-1000°C. At tempera- schist facies ('epizone') samples (T > 300°C),
tures of ::::600°C, deformation twinning was Burkhard (1990) noted increase in grain size
well developed in calcite, and twin boundary (6-10 pm) as the most notable feature. Together
migration has been identified as a key process with the weak grain-shape fabric and the vari-
involved during recrystallisation, without ably developed CPO, this suggests a significant

Fault and shear zone rocks

degree of dynamic recrystallisation. Schmid et indicating that solution-transfer processes

al. (1987), when considering the implications of were important during mylonitisation. Fluid-
their experimental work in relation to naturally assisted grain-boundary sliding is considered
deformed calcite rocks, reasonably inferred that as the dominant process. Moving further into
for low greenschist facies conditions or slightly the shear zone, and higher strains, the coarse
lower, intracrystalline twinning was the domi- mylonite passes with abrupt transition (over
nant mechanism. However, since calcite readily 1-2 cm) into fine-grained S-C mylonite and
anneals, the evidence of twinning is often lost then ultramylonite. The S-C mylonites have a
due to twin-boundary migration. Therefore, a well developed oblique shape fabric, and a
lack of observed twins in deformed calcite rocks pronounced CPO with c-axes perpendicular to
does not necessarily mean that twinning has not the shear plane. Ultramylonites are fine-
occurred. grained (20-30 pm), homogenous and with a
Studies by van der Pluijm (1991) and Busch shape fabric oblique to the shear plane. The
& van der Pluijm (1995) on the Bancroft shear development of shape fabrics and CPOs is
zone, Ontario, give a valuable insight into the attributed to dislocation creep processes, and
sequence of processes that operate and the grain-size reduction shown by the S-C
microstructural changes that occur from mylonites and ultramylonites is attributed to
proto lith to ultramylonite in a 15-20 m wide, rotation recrystallisation. A key point from the
upper greenschist facies (T approximately study of Busch & van der Pluijm (1995) was
450-500°C) shear zone in marble. Rapid that over very short distances within a shear
changes in intensity of shearing are observed, zone there is evidence that various deforma-
including the transition from protomylonite to tion mechanisms have operated, although at
ultramylonite over the scale of a thin section. any given time in the evolution of the shear
The proto lith has experienced low strain, facil- zone only one particular mechanism would be
itated by grain-boundary migration and the dominant.
production of deformation twins. The proto- In coarse-grained amphibolite facies
mylonite has a strong CPO and a well devel- marbles, deformation twinning and a lack of
oped porphyroclastic microstructure with any preferred crystallographic orientation is
abundant core-and-mantle structures as well typical. This suggests extensive recrystallisation
as undulose extinction and deformation bands. involving grain-boundary migration and
These microstructural features suggest that intracrystalline (twinning-dominated) slip as
dislocation creep has been an important mech- the dominant mechanisms.
anism, and that rotation recrystallisation
played a key role during dynamic recrystallisa-
8.3.4 Distinguishing between schists and
tion. The coarse mylonites comprise a dynami-
cally recrystallised aggregate of grains, of
similar size to the proto mylonite matrix Although protomylonites, with their distinctive
grains; some of the larger grains have a sub- porphyroclastic texture, and ribbon mylonites,
grain microstructure. The coarse mylonites with highly elongate quartz grains are readily
have a weak shape fabric and an almost identified, the distinction between certain
random crystallographic arrangement. This mylonites and schists may in some instances be
randomness is used by Busch & van der Pluijm less easy. Both rock types experience strong
(1995) as evidence for limited dislocation ductile deformation, and both possess a strong
creep. Cathodoluminescence reveals that most planar fabric (usually also with a linear compo-
grains have a secondary calcite overgrowth, nent). In strict terms, to call a rock a mylonite

Deformed rocks and strain-related microstructures

it must be demonstrated that there has been 8.4 The influence of deformation on
grain-size reduction. In the field, mylonitic metamorphic processes
fabrics may usually be distinguished from
schistose fabrics by virtue of their porphyro- During metamorphic reactions, the rate at
clastic texture, and by the fact that mylonites which heat is produced or consumed varies
grade laterally into undeformed or less according to the reaction kinetics and the
deformed rocks of similar composition. In enthalpy of the reaction (that energy evolved
narrow shear zones a recognisable change in when substances react). Prograde reactions are
orientation of the principle fabric may be visi- endothermic or, in other words, they absorb
ble in outcrop, but in other cases this may not heat from the surroundings. In the contact
be clear. Although mylonites develop in both aureole around intrusions it is fluids, and
narrow zones and broad crustal shears, schists energy in the form of heat, that drive metamor-
and schistose fabrics are always regionally phic reactions. However, in regional meta-
extensive. morphic environments where deformation
In thin section, mylonites typically contain accompanies metamorphism, and more partic-
angular or rounded porphyroclasts (e.g. quartz ularly in fault and shear zones where high
and feldspar) in a fine-grained or variable- strains and strain rates are concentrated into
sized matrix. Porphyroclasts show features of narrow zones, the influence of active deforma-
strain, recovery and recrystallisation, and may tion and mechanical energy on metamorphic
be polycrystalline. By contrast, schists processes can be significant. The role of defor-
normally contain rounded or euhedral mation during metamorphic transformations
unstrained porphyroblasts in a medium- to can be particularly significant in relation to
coarse-grained matrix. To distinguish fine- reaction kinetics (e.g. nucleation rate, growth
grained schists without porphyroblasts from rate and overall transformation rate). Rutter &
quartz-rich ultramylonites, it is necessary to Brodie (1995) identify three main ways in
examine the degree of crystallographic align- which deformation may enhance the rate of
ment of quartz. Initial microscopic examina- metamorphic transformations, namely: (1)
tion of the rock in Fig. 8.13(c) may suggest a grain-size reduction leading to increased
fine-grained psammite or semi-pelitic schist. surface area and thus more surface free energy
However, by insertion of a sensitive tint plate, to promote reaction; (2) production of strained
a pronounced crystallographic alignment of grains with high dislocation densities, which
quartz is seen, demonstrating that it is in fact have enhanced solubility relative to unstrained
an ultramylonite, albeit slightly recrystallised grains of the same phase; and (3) increased
(see Plates 7(a) & (b), which show the same temperature due to shear heating, where energy
sample). Use of the sensitive tint plate as a in the form of heat is produced by the release
quick check on the degree of crystallographic of strain energy, and as a result enhances meta-
alignment of quartz is highly recommended as morphic reactions.
a first approach to evaluating strain and defor- Grain-size reduction may also have the effect
mation processes, even if detailed Universal- of enhancing permeability, and fluid movement
stage work is not going to be undertaken. For through areas such as shear zones, but where
further details on the various applications of the dihedral angle between grains is >60 0 (e.g.
crystallographic fabric data in the study of carbonates), porosity will not be connected, so
strain paths and deformation processes in grain-size reduction alone is no guarantee of
rocks, the review by Law (1990) provides a increased permeability (Rutter & Brodie,
useful introduction. 1995). Shear zones are often documented as

The influence of deformation

areas of increased fluid flow, so a transient strain energies (elastic strain energy, defect
enhancement of permeability synchronous with energy and surface energy) and heat generated
shearing seems likely in many cases. There are during active deformation become sources of
a number of ways in which this could occur, energy for chemical work. This transformation
including grain-boundary sliding, which gener- of mechanical energy into chemical energy
ates grain-scale dilatancy, or cataclastic defor- occurs by the reaction of deformed crystals
mation, which gives dilation on a range of with the grain-boundary fluid. This fluid forms
scales. a vital link between deformational and chemi-
There are many ways in which deformation cal processes operating in a rock. It is crucial
may influence the sites of reactions. First, it for the operation of deformation processes
will influence the spatial distribution of those such as pressure solution, as well as having an
sites at which dissolution is most favoured. essential role in most metamorphic reactions.
Second, it will establish chemical potential When considering the controls on porphyrob-
gradients on a variety of scales and, third, it last nucleation in Chapter 5, the heterogeneous
will often increase bulk permeability and thus nature of rock materials was discussed. Because
aid diffusive mass transfer within the rock. rocks and rock sequences exhibit such strong
It is now well established that grain-bound- heterogeneity, it means that their deformation in
ary processes playa crucial role in both defor- response to applied stress will similarly be
mation and metamorphism. The concentration heterogeneous. This gives rise to strain partition-
of loose bonds and dislocations provides ideal ing, with some areas experiencing only low strain
sites for fluid-mineral interaction, leading to while others become highly strained. This parti-
reaction and nucleation of new phases. The tioning occurs from the macro scale right down
migrating grain boundary interfaces (which are to the micro scale, with significant strain varia-
highly disordered regions) are particularly tions developing around fold hinges, boudins,
favourable sites for reactions. The interconnec- porphyrohlasts and porphyroclasts. The develop-
tion of grain boundaries provides the necessary ment of strain and strain rate gradients produces
pathways for diffusive mass transfer of mater- dislocation density and thus chemical potential
ial via the fluid phase. Driven by chemical gradients. Fluctuations in chemical potential
potential gradients, this results in the redistrib- gradients will influence all reactions, and it is
ution of chemical components and promotes unlikely that chemical equilibrium will he
phase transformations within the rock. achieved in actively deforming rocks, especially
The deformation of rocks produces strained where stresses and strains are large.
crystals, giving rise to an increased defect Through its effect on the local activities of
density and thus enhancing intracrystalline particular aqueous species, deformation
diffusion. The increase in dislocation density at contributes to determining which minerals will
grain margins raises the surface energy and be dissolved and replaced. Sites of dissolution
provides more available free bonds. This preferentially develop in local areas of stress or
increase in surface energy causes a lowering of strain concentration. Where there is high differ-
the activation energy for nucleation, and means ential stress (e.g. crenulation limbs), solubility
that strained crystals will offer more favourable increases and leads to pressure solution.
sites for nucleation. Experimental work (Davis Because of this, authors such as Bell &
& Adams, 1965) has shown that strain result- Hayward (1991) have suggested that porphy-
ing from high shear stresses increases disloca- rohlasts (e.g. garnet) will not nucleate on
tion densities, and can increase reaction rates actively shearing crenulation limbs or other
by several orders of magnitude. The various situations of active shear.

Deformed rocks and strain-related microstructures

There are a number of ways in which defor- permeability, and thus enable greater fluid
mation can increase solubility. The first is by access. This in turn promotes reactions by way
increasing the concentration of dislocations in of the ions carried in solution and the catalytic
deformed minerals, and thus increasing the affect of the fluid.
lattice energy. Densities of dislocations in A portion of the work of deformation will
'tangles' and sub-grain walls can be large also be dissipated as heat. This too will
enough to increase solubilities by >10%. This contribute to increasing silicate solubility in
means that strained crystals of a mineral such the aqueous fluid, and if stress and strain
as K-feldspar are more soluble than unstrained rates are high enough, frictional heating can
crystals, and will thus more readily retrogress locally raise temperatures by as much as
(Wintsch, 1985). Pas schier (1985) documents a 1000°C. This induces frictional melting and
case in which the origin of flame perthites is pseudotachylite (Fig. 8.12) formation, and
attributed to the degree of deformation of K- has been related to cases of rapid brittle
feldspar grains, and Simpson & Wintsch faulting at high crustal levels (e.g. earth-
(1989) have demonstrated a link between quakes). The contribution of shear-heating in
myrmekite (in metamorphic rocks) and stress- ductile regimes is less well established.
induced K-feldspar replacement. Theoretical calculations suggest that a
While dissolution occurs in areas of high temperature increase of up to 150°C may be
stress and strain, the precipitation of dissolved possible. However, these are likely to be over-
material occurs in low-strain regions and areas estimates, since the calculations assume that
of extension. Pressure shadows around rigid all mechanical work is converted to heat, and
objects such as porphyroblasts (Fig. 5.4) and generally overlook the fact that synchronous
precipitation in extension fractures to give veins prograde reactions are endothermic, and thus
provide the most obvious sites, but crenulation consume most of the heat generated, and that
hinges are also important 'sinks' for material circulating fluids will be effective at transfer-
being actively dissolved from strongly sheared ring heat.
limbs (Fig. 4.6). Fibrolite is often observed to
be concentrated in zones of high shear strain, 8.5 The influence of metamorphism on
both on the micro- and the macroscale. This deformation processes
has led various authors (e.g. Vernon, 1987;
Wintsch & Andrews, 1988; Kerrick, 199m to The foregoing discussion has concentrated on
conclude that the development of fibrolite the influence that deformation has on meta-
aggregates is deformation-induced. The expla- morphic processes, but there are also various
nations developed by each author differ, but metamorphic processes that influence the rate
the basic link between fibrolite development and type of deformation. Brodie & Rutter
and zones of high shear has now been recog- (1985) produced a five-fold classification of
nised in many high-grade schists and gneisses. mechanistic interactions between metamorphic
A second way in which deformation trans-formations and deformability, recently re-
enhances solubility and general transformation evaluated (Rutter & Brodie, 1995) with refer-
kinetics is by the general process of grain-size ence to experimental and natural examples.
reduction. This is characteristic of mylonitisa- The effects of metamorphism on deformability
tion, and enhances solubility by increasing the are categorised as follows: (1) facilitation of
surface area of grain boundaries available for cataclasis due to elevated pore-fluid pressure in
reaction. This gives an overall increase in dehydration reactions or in response to melt-
surface energy, and may also increase bulk ing; (2) enhanced plasticity resulting from


transformation-induced volume changes; (3) propagate until such a time as P f subsides, but
development of fine-grained reaction products, may develop further if P f rises once more to
facilitating grain size sensitive flow processes; exceed rock strength. Therefore, it is apparent
(4) changes in plastic deformability of silicate that increasing P f due to devolatilisation reac-
minerals due to recrystallization and increased tions may aid cataclastic flow and increase
activity of pore fluid; and (5) promotion of deformability. As with grain-size reduction, this
diffusion creep via by the enhanced potential will significantly modify the rheological behav-
gradient of a reaction along the diffusion path iour. The release of fluid during dehydration
(Rutter & Brodie, 1995). reactions can also enhance grain-boundary
'Reaction enhanced ductility (plasticity), diffusion and may promote diffusion-accom-
(White & Knipe, 1978; Rubie, 1990) is a modated grain-boundary sliding (Rubie, 1990).
particularly important process that operates in Certain phase transformations, especially in
actively deforming rocks where metamorphic carbonate rocks, involve significant negative
reactions are simultaneously in progress. volume changes (up to 30%). These may have a
Metamorphic reactions give rise to increased significant, although probably short-lived, effect
ductility in a number of ways. First, they will on bulk rock porosity and permeability, which
produce small grains, and allow grain size in turn influences rock deformation. For exam-
sensitive flow processes such as grain-boundary ple, stress concentrations and additional void
sliding to operate (i.e. a change in deformation formation resulting from volume changes may
mechanism is induced). Second, reactions will induce rock failure by brittle fracturing. Other
aid 'strain softening' by producing soft, strain- effects may also occur, but our present knowl-
free grains. Third, retrograde reactions often edge of the contribution of volume changes
convert 'hard' phases such as feldspar to 'soft' during phase transformations is rather limited,
phases such as quartz, sericite and calcite, a and certainly requires further investigation.
process known as 'reaction softening' (White et
al., 1980). These reaction-related processes References
often make an important contribution to the
Bakker, R.]. & Jansen, ].B. (1994) A mechanism for
overall softening and enhanced ductility in preferential H 2 0 leakage from fluid inclusions in
mylonites (e.g. White et ai. (1980) and quartz, based on TEM observations. Contributions
Williams & Dixon (1982) for granitoid to Mineralogy and Petrology, 116, 7-20.
mylonites; and Brodie & Rutter (1985) and Barber, D.]. (1990) Regimes of plastic deformation -
processes and microstructures: an overview, in
Stiinitz (1993) for mafic mylonites), but if Deformation processes in minerals, ceramics and
prograde reactions take place this typically rocks (eds D.]. Barber & P.G. Meredith).
gives rise to overall hardening. Dynamic recrys- Mineralogical Society of Great Britain and Ireland,
tallisation during metamorphism aids deforma- Monograph Series No.1, Ch. 6, 138-178.
Barber, D.]. & Meredith, P.G. (1990) Deformation
tion processes by continually providing new processes In minerals, ceramics and rocks.
strain-free grains. Mineralogical Society of Great Britain and Ireland,
By liberating or consuming fluid, metamor- Monograph Series No.1, 423 pp.
phic reactions have a major influence on fluid Bell, T.H. & Hayward, N. (1991) Episodic metamor-
phic reactions during orogenesis: the control of
pressure (P f ) at the time of deformation. Where deformation partitioning on reaction sites and dura-
fluid production is fast with respect to diffusiv- tion. Journal of Metamorphic Geology, 9,
ity/permeability, the rapid fluid release may 619-640.
induce a local increase in P f (Chapter 11). If Boland, ].N. & FitzGerald, ].D. (eds) (1993) Defects
and processes in the solid state: geoscience applica-
the production of fluid is sufficiently rapid, tions. Developments in Petrology No. 14 ('The
hydraulic fracturing may occur. Fractures will McLaren Volume'). Elsevier, Amsterdam, 470 pp.

Deformed rocks and strain-related micrastructures

Brodie, K.H. (1981) Variation in amphibole and Hirth, G. & Tullis, J. (1992) Dislocation creep regimes
plagioclase composItIOn with deformation. in quartz aggregates. Journal of Structural Geology,
Tectonophysics, 78, 385-402. 14,145-159.
Brodie, K.H. & Rutter, E.H. (1985) On the relation- Jessell, M.W. & Lister, G.S. (1990) A simulation of the
ship between deformation and metamorphism, with temperature dependence of quartz fabrics, in
special reference to the behaviour of basic rocks, in Deformation mechanisms, rheology and tectonics
Metamorphic reactions (eds A.B. Thompson & D.C. (eds R.J. Knipe & E.H. Rutter). Geological Society
Rubie). Advances in Physical Geochemistry, 4, Special Publication No. 54, 353-362.
138-179. Kerrick, D.M. (1990) The AlzSiO s polymorphs.
Burkhard, M. (1990) Ductile deformation mechanisms Mineralogical Society of America, Reviews in
in micritic limestones naturally deformed at low Mineralogy No. 22, 406 pp.
temperatures (150-350°C), in Deformation mecha- Knipe, R.J. (1990) Microstructural analysis and
nisms, rheology and tectonics (eds R.J. Knipe & tectonic evolution in thrust systems: examples from
E.H. Rutter). Geological Society Special Publication the Assynt region of the Moine Thrust Zone,
No. 54,241-257. Scotland, in Deformation processes in minerals
Busch, J.P. & van der Pluijm, B.A. (1995) Calcite ceramics and rocks. (eds D.J. Barber & P.G.
textures, microstructures and rheological properties Meredith). Mineralogical Society of Great Britain
of marble mylonites in the Bancroft shear zone, and Ireland, Monograph Series No.1, 228-261.
Ontario, Canada. Journal of Structural Geology, Knipe, R.J. & Rutter, E.H. (1990) Deformation mech-
17,677-688. anisms, rheology and tectonics. Geological Society
Carter, N.L., Christie, J.M. & Griggs, D.T. (1964) Special Publication No. 54, The Geological Society,
Experimental deformation and recrystallization of London, 535 pp.
quartz. Journal of Geology, 72, 687-733. Lafrance, B. & Vernon, R.H. (1993) Mass transfer and
Craw, D. & Norris, R.J. (1993) Grain boundary microfracturing in gabbroic mylonites of the
migration of water and carbon dioxide during uplift Guadalupe Igneous Complex, California, in Defects
of the garnet-zone Alpine Schist, New Zealand. and processes in the solid state: geoscience applica-
Journal of Metamorphic Geology, 11,371-378. tions (eds J.N. Boland & J.D. FitzGerald).
Davis, B.L. & Adams, L.H. (1965) Kinetics of the Developments in Petrology No. 14 ('The McLaren
calcite-aragonite transformation. Journal of Volume'). Elsevier, Amsterdam, 151-167.
Geophysical Research, 70, 433-441. Law, R.D. (1990) Crystallographic fabrics: a selective
Dietrich, D. & Song, H. (1984) Calcite fabrics in a review of their applications to research in structural
natural shear zone environment, the Helvetic nappes geology, in Deformation mechanisms, rheology and
of western Switzerland. Journal of Structural tectonics (eds R.]. Knipe & E.H. Rutter). Geological
Geology, 6, 19-32. Society Special Publication No. 54, The Geological
Drury, M.R. (1993) Deformation lamellae in metals Society, London, 335-352.
and minerals, in Defects and processes in the solid Lin, A. (1994) Glassy pseudotachylite veins from the
state: geoscience applications (eds J.N. Boland & Fuyun fault zone, northwest China. Journal of
J.D. FitzGerald). Developments in Petrology No. 14 Structural Geology, 16, 71-84.
('The McLaren Volume'). Elsevier, Amsterdam, Lloyd, G.E. & Knipe, R.J. (1992) Deformation mecha-
195-212. nisms accommodating faulting of quartzite under
Drury, M.R. & Urai, J.L. (1990) Deformation-related upper crustal conditions. Journal of Structural
recrystallisation processes. Tectonophysics, 172, Geology, 14, 127-143.
235-253. McLaren, A.C., Retchford, J.A., Griggs, D.T. &
Gerretsen, J., McLaren, A.C. & Paterson, M.S. (1993) Christie, J.M. (1967) Transmission electron micro-
Evolution of inclusions in wet synthetic quartz as a scope study of Brazil twins and dislocations experi-
function of temperature and pressure; implications mentally produced in natural quartz. Physica Status
for water weakening, in Defects and processes in Solidi, 19,631-644.
the solid state: geoscience applications (eds J.N. Maddock, R.H., Grocott, J. & van Nes, M. (1987)
Boland & J.D. FitzGerald). Developments in Vesicles, amygdales and similar structures in
Petrology No. 14 ('The McLaren Volume'). Elsevier, fault-generated pseudotachylites. Lithos, 20,
Amsterdam, 27-47. 419-432.
Green, H.W. (1992) Analysis of deformation in geolog- Mawer, C.K. & FitzGerald, J.D. (1993) Microstructure
ical materials, in Minerals and reactions at the of kink band boundaries in naturally deformed
atomic scale: transmission electron microscopy (ed. Chewings Range Quartzite, in Defects and processes
P.R. Buseck). Mineralogical Society of America, in the solid state: geoscience applications (eds J.N.
Reviews in Mineralogy, No. 27, Ch. 11,425-454. Boland & J.D. FitzGerald). Developments in


Petrology No. 14 ('The McLaren Volume'). Elsevier, dynamic recrystallization of Carrara marble.
Amsterdam, 49-67. Journal of Geophysical Research, 100(B12),
Park, W.e. & Schot, E.H. (1968) Stylolites: their 24651-24663.
nature and ongm. Journal of Sedimentary Rutter, E.H. & Brodie, K.H. (1992) Rheology of the
Petrology, 38,175-191. lower crust, in Geology of the lower continental
Passchier, e.W. (1985) Water-deficient mylonite zones crust (eds D. Fountain, R. Arculus, R. & R. Kay).
- an example from the Pyrenees. Lithos, 18, Elsevier, Amsterdam, 201- 268.
115-127. Rutter, E.H. & Brodie, K.H. (1995) Mechanistic inter-
Pas schier, e.w. & Trouw, R.A.J. (1996) actions between deformation and metamorphism.
Microtectonics. Springer-Verlag, Berlin, 289 pp. Geological Journal, 30, 227-240.
Poirier, J.-P. (1985) Creep of crystals. Cambridge Rutter, E.H., Casey, M. & Burlini, L. (1994) Preferred
University Press, Cambridge, 260 pp. crystallographic orientation development during
Pontikis, V. & Poirier, J.-P. (1975) Phenomenological plastic and superplastic flow of calcite rocks.
and structural analysis of recovery-controlled creep, Journal of Structural Geology, 16, 1431-1446.
with special reference to the creep of single-crystal Schmid, S.M., Boland, J.N. & Paterson, M.S. (1977)
silver chloride. Philosophical Magazine, 32, Superplastic flow in fine grained limestone.
577-592. Tectonophysics, 43, 257-291.
Prior, D.J., Knipe, R.J. & Handy, M.R. (1990) Schmid, S.M., Panozzo, R. & Bauer, S. (1987) Simple
Estimates of the rates of microstructural changes in shear experiments on calcite rocks: rheology and
mylonites, in Deformation mechanisms, rheology microfabric. Journal of Structural Geology, 9,
and tectonics (eds R.J. Knipe & E.H. Rutter). 747-778.
Geological Society Special Publication No. 54, The Sibson, R.H. (1977) Fault rocks and fault mecha-
Geological Society, London, 309-319. nisms. Journal of the Geological Society, 133,
Ramsay, J.G. & Huber, M.1. (1983) The techniques of 191-213.
modern structural geology: Volume 1: Strain analy- Sibson, R.H. (1990) Faulting and fluid flow, in
sis. Academic Press, London. Fluids in tectonically active regimes of the conti-
Ranalli, G. (1982) Deformation maps in grain size- nental crust (ed. B.E. Nesbitt). Mineralogical
stress space as a tool to investigate mantle rheol- Association of Canada, Short Course No. 18, Ch.
ogy. Physics of the Earth and Planetary Interiors, 4,93-132.
29,42-50. Simpson, e. & Wintsch, R.P. (1989) Evidence for
Ree, J.-H. (1994) Grain boundary sliding and develop- deformation-induced K-feldspar replacement by
ment of grain boundary openings in experimentally myrmekite. Journal of Metamorphic Geology, 7,
deformed octachloropropane. Journal of Structural 261-275.
Geology, 16,403-418. Skrotzki, W. (1990) Microstructure in hornblende of
Reeder, R.J. (1992) Carbonates: growth and alteration mylonitic amphibolite, in Deformation mechanisms,
microstructures, in Minerals and reactions at the rheology and tectonics (eds R.J. Knipe & E.H.
atomic scale: transmission electron microscopy. (ed. Rutter). Geological Society Special Publication No.
P.R. Buseck). Mineralogical Society of America, 54,321-325.
Reviews in Mineralogy No. 27, Ch. 10,381-424. Snoke, A.W., Tullis, J. & Todd, V.R. (eds) (in press)
Ross, J.V., Ave Lallement, H.G. & Carter, N.L. (1980) Fault-related rocks: a photographic atlas. Princeton
Stress dependence of recrystallized grain and University Press, Princeton, New Jersey.
subgrain size in olivine. Tectonophysics, 70, 39-61. Spiers, e.J., Schutjens, P.M.T.M., Brzesowsky, R.H.,
Rubie, D.e. (1990) Mechanisms of reaction-enhanced Peach, e.]., Liezenberg, J.L. & Zwart, H.J. (1990)
deformability in minerals and rocks, in Deformation Experimental determination of constitutive parame-
processes in minerals, ceramics and rocks (eds D.J. ters governing creep of rocksalt by pressure solu-
Barber & P.G. Meredith). Mineralogical Society of tion, in Deformation mechanisms, rheology and
Great Britain and Ireland, Monograph Series No.1, tectonics. (eds R.J. Knipe & E.H. Rutter).
Ch. 10,262-295. Geological Society Special Publication No. 54,
Rutter, E.H. (1976) The kinetics of rock deformation 215-227.
by pressure solution. Philosophical Transactions of Stiinitz, H. (1993) Transition from fracturing to
the Royal Society of London, 283A, 203-219. viscous flow in a naturally deformed metagabbro,
Rutter, E.H. (1986) On the nomenclature of mode of in Defects and processes in the solid state:
failure transitions in rocks. Tectonophysics, 122, geoscience applications (eds J.N. Boland & J.D.
381-387. fitzGerald). Developments in Petrology No. 14
Rutter, E.H. (1995) Experimental study of the influ- ('The McLaren Volume'). Elsevier, Amsterdam,
ence of stress, temperature, and strain on the 121-150.

Deformed rocks and strain-related microstructures

Tullis, ].A. (1990) Experimental studies of deforma- Wenk, H.-R., Takeshita, T., Bechler, E., Erskine, B.G.
tion mechanisms and microstructures in quartzo- & Matthies, S. (1987) Pure shear and simple shear
feldspathic rocks, in Deformation processes in calcite textures: comparison of experimental, theo-
minerals, ceramics and rocks. (eds D.]. Barber & retical and natural data. Journal of Structural
P.G. Meredith). Mineralogical Society of Great Geology, 9, 731-745.
Britain and Ireland Monograph Series No.1, White, ].c. & White, S.H. (1981) On the structure of
190-227. grain boundaries in tectonites. Tectonophysics, 78,
Tullis, ].A., Christie, ].M. & Griggs, D.T. (1973) 613-628.
Microstructures and preferred orientations of White, S. (1976) The effects of strain on the
deformed quartzites. Bulletin of the Geological microstructures, fabrics, and deformation mecha-
Society of America, 84,297-314. nisms of quartzites. Philosophical Transactions of
van der Pluijm, B.A. (1991) Marble mylonite in the the Royal Society of London, 283A, 69-86.
Bancroft shear zone, Ontario, Canada: microstruc- White, S.H. & Knipe, R.]. (1978) Transformation- and
tures and deformation mechanisms. Journal of reaction-enhanced ductility in rocks. Journal of the
Structural Geology, 13, 1125- 1135. Geological Society of London, 135,513-516.
Vernon, R.H. (1976) Metamorphic processes. George White, S.H., Burrows, S.E., Carreras, J., Shaw, N.D. &
Allen & Unwin, London, 247 pp. Humphreys, F.]. (1980) On mylonites in ductile shear
Vernon, R.H. (1987) Growth and concentration of zones. Journal of Structural Geology, 2, 175-187.
fibrous sillimanite related to heterogeneous defor- Williams, G.D. & Dixon, ]. (1982) Reaction and
mation of K-feldspar-sillimanite metapelites. geometric softening in granitoid mylonites. Textures
Journal of Metamorphic Geology, 5, 51-68. and Microstructures, 4, 223-239.
Walker, A.N., Rutter, E.H. & Brodie, K.H. (1990) Wintsch, R.P. (1985) The possible effects of deforma-
Experimental study of grain-size sensitive flow of tion on chemical processes in metamorphic fault
synthetic, hot-pressed calcite rocks, in Deformation zones, in Metamorphic reactions (eds A.B.
mechanisms, rheology and tectonics (eds R.]. Knipe Thompson & D. C. Rubie). Advances in Physical
& E.H. Rutter). Geological Society Special Geochemistry, 4, 251-268.
Publication No. 54,259-284. Wintsch, R.P. & Andrews, M.S. (1988) Deformation
Watson, E.B. & Brenan, ].M. (1987) Fluids in the induced growth of sillimanite: 'stress' minerals revis-
lithosphere, 1. Experimentally-determined wetting ited.Journal of Geology, 96,143-161.
characteristics of CO 2-H 2 0 fluids and their impli- Wise, D.U., Dunn, D.E., Engelder, ].T., Geiser, P.A.,
cations for fluid transport, host-rock physical Hatcher, R.D., Kish, S.A., Odom, A.L. & Schamel,
properties, and fluid inclusion formation. Earth S. (1984) Fault-related rocks: suggestions for termi-
and Planetary Science Letters, 85, 497-515. nology. Geology, 12, 391-394.

Chapter nine


When studying a suite of metamorphic rocks, growth were considered, and in Section 6.1
especially those formed during orogenic meta- general aspects concerning mineral inclusions
morphism, the common occurrence of porphy- were discussed. This chapter concentrates on
roblasts and one or more tectonic fabrics the various aspects of porphyroblast-foliation
provides valuable information concerning relationships, providing an introduction to the
deformation-metamorphism interrelationships. characteristic microstructural features observed,
When integrated with pressure and tempera- and providing an overview of the various
ture estimates based on mineral assemblages points to take into consideration when making
and geothermobarometry, coupled with any interpretations. The chapter highlights particu-
age dates that may exist to date particular lar problem areas, where opinion is divided
events in the region studied, such microstruc- concerning the interpretation.
tural features can enable the construction of a
pressure-tern perature-time (P-T -t) path for
individual rocks. This in turn can give impor-
9.1 Thin-section 'cut effects'
tant insights into aspects of crustal evolution, Many metamorphic rocks exhibit planar
or metamorphic evolution in the vicinity of and/or linear structures defined by the
major igneous intrusions. Although the inter- preferred alignment of minerals, as a result of
pretations from porphyroblast-foliation rela- one or more deformation events. Since porphy-
tionships have excellent potential, the roblasts commonly overgrow these structures,
interpretation of porphyroblast inclusion trails the thin-section cut relative to these planar or
and their relationships with external matrix linear elements has a major influence on the
fabrics in the rock have always been types of inclusion patterns observed in porphy-
contentious topics. Different interpretations of roblasts. In order to interpret these inclusion
the same textural and microstructural features trails correctly, the main cut effects must be
have led to fundamental differences of opinion fully appreciated. Let us consider these with
concerning the processes responsible for partic- reference to the classical case of S-shaped inclu-
ular features observed and consequently the sion trails commonly seen in garnets.
overall interpretations can vary greatly between Traditionally interpreted as syntectonic rota-
different researchers. In Sections 5.1, 5.2 & tion of the porphyroblast with respect to the
5.3, aspects of porphyroblast nucleation and matrix foliation (e.g. Zwart, 1962; Spry, 1963),

Porphyroblast-foliation relationships

many so-called 'rotated' porphyroblasts have R

now been reinterpreted in terms of change in \
fabric orientation relative to porphyroblast
during growth, or as porphyroblast overgrowth
of crenulations, the porphyroblast not having R
rotated at all (e.g. Bell & Rubenach, 1983;
Bell, 1985). There is considerable debate on
this matter, which will be covered in Section
9.2.2, but at this stage discussion is confined to
the different porphyroblast-foliation relation-
ships that can be obtained for a single prophy-
b. ~
roblast or group of porphyroblasts according
to the orientation of the thin-section cut. FIG. 9.1 :)C and GD inclusion fabrics in garnet for
Powell & Treagus (1970) made a study of sections cut parallel to, and containing, the rotation
the 'cut effect' on inclusion patterns by axis (R) (modified after MacQueen & Powell, 1977,
Figs 5(a) & (b)). Scale = 0.5 mm.
comparison of model-generated patterns with
those observed in natural syntectonic garnet
porphyroblasts from Norway and Scotland. In
cases of equidimensional porphyroblasts such porphyroblast, whereas m neighbouring
as garnet, the inclusion pattern observed can be porphyroblasts the cut may have been at or
likened to a cross-section through a sphere, and close to the edge. As a result, there may be
those sections cut perpendicular to the 'rota- appreciable microstructural differences from
tion' axis (X-Z sections) will produce an S- one porphyroblast to the next. Cuts through the
shaped fabric. A similar, though not identical, centre will record the full history, whereas cuts
fabric is said to occur in sections cut oblique to close to the edge will only record the final
this orientation, but for sections cut parallel to, stages of porphyroblast development. This has
and containing, the rotation axis, J C, 11 and been verified by cutting serial sections through
aD patterns are observed (Fig. 9.1). More individual porphyroblasts. It has been shown
recently, Johnson (1993a) made an in-depth that in the case of porphyroblasts showing S-
analysis of variations in spiral-shaped inclusion fabrics, a more pronounced'S' is seen in cuts
trail geometries of natural garnets as a function through the centre compared to cuts near to the
of orientation of thin-section cut. He records rim (Fig. 9.2(a)). In porphyroblasts with a more
various inclusion trail geometries, including complex history of growth - let us say, two or
closed loop aD patterns. However, Johnson more textural zones - there will be more signifi-
(1993a) records that spiral axes are generally cant differences between 'centre' and 'rim' cuts
oblique to the main external foliation, and that (Fig. 9.2(b)). Because of the problems of inter-
closed loops result from non-cylindrical spiral- pretation caused by this type of 'cut effect', it is
shaped inclusion trails. necessary to identify carefully, and to base inter-
Another thin-section 'cut effect', examined pretations on those porphyroblasts cut
by Powell & Treagus (1970), Johnson (1993a) centrally. Unless you have good control on the
and others, is the case of parallel cuts passing cut of a specific porphyroblast in a given thin
through different parts of the porphyroblast. section, it may be advisable to place greater
When thin sections are made it is understand- emphasis on the porphyroblasts with the largest
able that in some cases the cut will have passed cross-sectional area. This approach is based on
through the centre or close to the centre of the the assumption that porphyroblasts with the

Porphyroblast growth in relation to foliation

(a) Single phase of syn-tectonic growth the principal foliation should also be made (S-

pCD0 CD ~

cut Centre·
cut 9.2 Porphyroblast growth in relation to
foliation development
(b) Multiple growth phases

-@) /if\"-
.... ~/
'-- ,/
- -Q-
9.2.1 Recognition and interpretation of
pre-foliation (pre-tectonic) crystals
cut Pre-foliation crystals are those that existed in
the rock prior to the onset of deformation, and
FIG. 9.2 Schematic illustrations of centre-rim cut the development of a continuous foliation (e.g.
effects in porphyroblasts: (a) single period of growth; cleavage or schistosity). This includes detrital
(b) multiphase growth. grains, relict igneous crystals and any meta-
morphic crystal that existed in the rock prior to
deformation (e.g. 'chlorite-mica stacks';
smallest radius (cross-sectional area) represent Section 7.2, Fig. 7.7). With the exception of
cuts closest to the edge of the crystaL However, cases in which there has been extensive matrix
in doing this, a certain amount of caution must recrystallisation, which makes relationships
be exercised, because not all porphyroblasts between crystal and matrix difficult to deter-
nucleate simultaneously, and those nucleated mine, 'pre-foliation' crystals are usually
late will often be smaller. In consequence, they enveloped by the main foliation of the rock.
will naturally have an included fabric that is Since pre-foliation crystals precede most if not
only representative of the later parts of the all deformation, they will commonly exhibit
tectonometamorphic history. Therefore, the signs of strain. Pre-foliation porphyroblast
simpler fabrics of smaller porphyroblasts need growth seems to be a relatively uncommon
not always be put down to a cut effect. In order phenomenon, or at least is rarely identified.
to obtain a full three-dimensional picture of However, Vernon et al. (1993) have described
porphyroblast-inclusion trails, many researchers examples of pre-foliation porphyroblast devel-
consider it necessary routinely to cut a set of opment in various cases of low-P/high- T
serial sections through a single porphyroblast, orogenic metamorphism.
and to make a range of thin-section cuts in One or more of the following features may
different orientations (e.g. Johnson, 1993a; be seen in pre-foliation crystals, but should in
Johnson & Moore, 1996). It has been common- no way be considered exclusive to pre-foliation
place to cut thin sections perpendicular to the crystals:
main fabric (schistosity/cleavage) and parallel to
the principal mineral elongation direction (P- (i) Undulose extinction. This occurs due to
sections), but for a better 3-D understanding of distortion of the crystal lattice (Section 8.2.2).
the inclusion geometry it is advisable to cut at It is especially common in quartz, but is also
least one section perpendicular to both the prin- seen in various other minerals (e.g. feldspar,
cipal fabric and mineral elongation lineation biotite and kyanite). Undulose extinction in
(N-section). Bell & Johnson (1989) further quartz porphyroclasts within a protomylonite
advocate that for complete 3-D understanding is shown in Fig. 9.3(a).
of the inclusion fabric, a section cut parallel to (ii) Fractured crystals and boudinage. Minerals

Porphyroblast-foliation relationships

such as feldspar and hornblende behave in a originated as a granite, is shown in Fig. 9.3(b).
brittle manner at low temperatures and/or high (iii) Strain (pressure) shadows. Although also a
strain rates (for details, see Sections 8.1, 8.3.1 feature of crystals formed synchronous with
& 8.3.2). A highly fractured microcline crystal foliation development, many pre-foliation crys-
in what is now a protomylonite, but which tals exhibit strain shadows. These are low-



FIG. 9.3 Some characteristic features associated with pre-tectonic crystals (these may also be seen for syntec-
tonic crystals). (a) Undulose extinction (in quartz) . Scale = 0.5 mm (XPL). (b) Fragmented crystals (microcline
in sheared granite). Scale = 1 mm (XPL).

Porphyroblast growth in relation to foliation

strain areas in which new minerals crystallise later generation of euhedral pyrite crystals in
preferentially. They result from strain partition- the centre of the photograph has developed
ing around rigid porphyroblasts or clasts (see after final deformation and shows no strain
also Section 10.7). The example in Fig. 9.3(c) shadows.
shows quartz strain shadows developed around (iv) Kinking. This is a common feature of
early framboidal pyrite. Note the fact that a detrital and early-formed minerals, especially



FIG. 9.3 (contd) (c) Pressure shadows (around framboidal pyrite). Scale = 0.5 mm. (d) Kinking (in early biotite
porphyroblast). Scale = 0.1 mm (PPL).

Porphyroblast-foliation relationships

phyllosilicate phases (see also Section 8.2.2). crystals. The porphyroblast-forming reaction
An early porphyroblast of biotite in a mica will involve one or more of the matrix miner-
schist, which is strongly kinked due to subse- als, and in general the reactant phases are not
quent deformation, is shown in Fig. 9.3(d). significantly represented in the included assem-
blage. The minerals most commonly present as
inclusion trails are quartz, opaque minerals
9.2.2 Recognition and interpretation of (e.g. ilmenite and graphite), epidote group
syntectonic crystals minerals and, to a lesser extent, amphibole,
During orogenic metamorphism, most porphy- mica, chloritoid and feldspar. The inclusion
roblasts are syntectonic, and although the trails are widely interpreted as representing the
majority of matrix minerals in deformed meta- matrix grain size at the time of inclusion, and
morphic rocks recrystallised or grew synchro- because matrix grain size shows a general
nous with deformation, it is often difficult to increase during prograde metamorphism it is
say unequivocally whether a given crystal is usual for the enclosed fabric to be of similar or
'pre-tectonic' or 'syntectonic'. This is because finer grain size than the external fabric (Plates
many crystals developing synchronous with 5&6).
deformation may also be strained to some For those porphyroblasts in which inclusion
degree, and may show one or more of the same trails are clearly defined, various types are
strain features listed above for pre-tectonic crys- commonly encountered. These range from
tals. However, there is one commonly observed straight inclusion trails that are sharply discor-
textural feature of syntectonic porphyroblasts dant with the external fabric (Plates 5(a) &
which is absent from pre-tectonic crystals. The (b)), to inclusion fabrics with a shallow or
feature in question is that of aligned inclusion pronounced'S' form (Plates 5(c) & (d)), and in
fabrics. The preservation of such fabrics, and some cases spirals of much greater complexity
the minerals defining the fabrics, are crucial to (Fig. 9.4).
the understanding of the early deformation and Discordance between porphyroblast inclusion
metamorphic history of a particular rock. trails and the external (matrix) fabric
Cases of straight porphyroblast inclusion trails
'Straight' and 'S-shaped' inclusion trails
(SJ sharply discordant (Plate 5(a)) with the
Not every syntectonic porphyroblast phase external fabric (S,) have traditionally been inter-
exhibits aligned inclusion trails, but they are preted in terms of static overgrowth of some
widely reported in minerals such as garnet, pre-existing foliation followed by a later defor-
albite, hornblende, staurolite and andalusite. mation event intensifying the matrix fabric and
Even in these minerals it is more common to causing the discordant relationships. With such
observe poikiloblastic crystals without any an interpretation the internal fabric would often
distinct alignment of inclusions. Particularly be considered as SI and the external fabric as S2.
common are examples in which the porphyrob- This interpretation has been advanced largely
last cores are heavily included, whereas the because of the assumption that porphyroblast
rims are largely devoid of inclusions (Fig. growth rates are slow relative to deformation
5.22(b)). Although this may represent more rates. Depending on the nature of the deforma-
than one growth stage, an alternative explana- tion partitioning in the early stages of fabric
tion is in terms of rapid initial growth causing development, it can prove difficult to reconcile
numerous inclusions, followed by a slowing of straight inclusion fabrics with porphyroblastesis
growth rate and consequently fewer included synchronous with foliation development. The

Porphyroblast growth in relation to foliation

FIG. 9.4 Garnet with an exceptionally spiralled inclusion fabric: Dikanas Schist, Vasterbotten, Sweden. This
figure is redrawn from a photograph by N.O. 01esen, and figured by Schoneveld (1977).

reason for this is that the attitude of the matrix earlier schistosity or cleavage, as has often been
fabric relative to the porphyroblast should proposed. With this in mind, 'straight' and
progressively change and thus generate a 'slightly curved' inclusion trails are equally as
curved or sigmoidal inclusion trail. However, it likely to represent porphyroblastesis synchro-
has now been demonstrated that most porphy- nous with foliation development as they are to
roblasts probably grow extremely rapidly in represent growth between fabric-forming
geological terms (i.e. in < 1 Ma, and possibly in events. Because of the difficulty in using S;-Se
< 0.1 Ma; Section 5.4). This means that at typi- relationships to determine with certainty
cal regional orogenic strain rates (e.g. 10-14 S-I) whether the porphyroblast has rotated with
there would often be no appreciable rotation of respect to the external fabric or vice versa,
porphyroblast or matrix fabric during growth, Pas schier et al. (1992) suggest that it is safest to
and thus 'straight' inclusion trails would be discuss Si-Se angular relationships in terms of
preserved. However, at faster strain rates (e.g. relative rotation between S; and So'
10-12 S-I) of shear zones, significant relative
rotation between porphyroblast and fabric The influence of crystallographic structure
could occur, and S-shaped inclusion trails When interpreting straight inclusion fabrics in
should be more widely developed (Barker, minerals with a strong cleavage or cleavages
1994). Considering periods of porphyroblaste- (e.g. micas, amphiboles and kyanite), always be
sis as rapid events punctuating protracted peri- aware of the fact that crystallographically
ods of foliation development, the continuation controlled inclusion trails can develop which
of shearing after cessation of porphyroblastesis have nothing to do with any pre-existing or
would explain the discordant S,-S, relationships developing tectonic fabric. These crystallo-
that are commonly observed in porphyroblastic graphically controlled 'straight' inclusion trails
schists. With this interpretation, it means that develop because, within minerals possessing a
the included fabric will commonly represent an cleavage, inclusions are often preferentially
early stage in the development of the external incorporated and aligned within the cleavage of
fabric, and need not be some entirely separate the crystal in order to minimise the increase in

Porphyroblast-foliation relationships

internal energy caused by incorporating an

inclusion into the lattice. In Chapter 6 it has
already been discussed how certain types
of crystallographic and growth-influenced

inclusion patterns develop in certain minerals, a b c
but because of their distinctive appearance
(Plate 1) these are unlikely to be confused with
tectonic fabrics.

Interpretations for and against porphyroblast

In cases of gentle or pronounced S-fabrics
d --

(Plates Sic) & (d)), the traditional interpreta-

tion, exemplified by Zwart (1962) and Spry
(1963), is to consider the garnet as a sphere
physically rotating during growth in response
to simple shear within the matrix. The inclu-
sion trail (defined by included matrix material)
FIG. 9.5 A model for the development of 'snowball'
is considered to initiate sub-parallel to the S-fabrics in porphyroblasts (especially garnet) (after
matrix fabric, but to become progressively Spry, 1963).
rotated as the porphyroblast itself physically
rotates relative to externally fixed reference co-
ordinates (Fig. 9.5). This model has been
advanced by many researchers (e.g. Rosenfeld,
1970; Schoneveld, 1977), and on the basis of
this interpretation various authors have
analysed S-fabrics in an attempt to assess the
amount of relative rotation between porphy- (a) (b) (c) ~

roblast and external matrix. Inclusion fabrics

suggesting rotations of up to 90 (Plate Sic))
0 FIG. 9.6 A model for the development of S-shaped
inclusion fabric in porphyroblast by transposition of
are very common, those up to 180 (Plate S(d))

the external foliation through 90° during incremental

are less common, and while fabrics apparently porphyroblast growth. The porphyroblast does not
indicating rotations up to 800 (Fig. 9.4) have
been recorded, these are extremely rare.
Wilson (1971) promoted a realistic alterna-
tive interpretation for gently curved and more be inadequate. Note that for a given sense of
pronounced S-shaped inclusion trails, advocat- shear the'S' or 'Z' developed in this model is
ing that sub-spherical porphyroblasts such as the reverse of that produced when modelling in
garnet do not rotate, but that S-shaped inclu- terms of physical rotation of porphyroblasts
sion trails develop as the matrix foliation (compare Figs 9.Se & 9.6c; see also Section
changes orientation relative to the growing 10.5). A detailed study by Johnson (1993b)
porphyroblast. This could adequately explain concluded that most if not all spiral-shaped
the common spirals with less than 90 relative
inclusion trails could be explained just as well
rotation (Fig. 9.6), but for porphyroblasts with in terms of porphyroblast rotation relative to
more pronounced spiralling this model would externally fixed reference co-ordinates as they

Porphyroblast growth in relation to foliation

could by external fabric rotation relative to

fixed porphyroblasts.
Bell & Rubenach (1983), and a series of
subsequent publications by Bell (1985) and co-
workers, have also presented a model advocat-
ing non-rotation of syntectonic porphyroblasts.
Their model is based on the fact that deforma-
tion in rocks is heterogeneous and results in
strain partitioning. This occurs on various
scales, such that during different stages of
progressive bulk non-coaxial deformation (e.g.
simple shear), different portions of the deform-
ing mass, from the macroscale right down to
the microscale, experience one of several defor-
mation types (Fig. 9.7). The style of deforma-
tion ranges from areas of no strain to areas
dominated by progressive shortening strain,
and areas of progressive shearing strain. The
difference in deformability between matrix and
porphyroblasts gives rise to strong deformation FIG. 9.7 (a) The distribution of deformation parti-
partitioning about the rigid porphyroblasts, tioning on a strain-field diagram constructed for the
and produces pronounced strain gradients close x-z plane, representing a block of rock which has
undergone non-coaxial progressive bulk inhomoge-
to their margins. This partitioning produces an neous shortening. 1, No strain; 2, progressive shorten-
ellipsoidal island of porphyroblast width that is ing strain; 3, progressive shortening plus shearing
protected from the effects of progressive shear- strain (after Bell, 1985). (b) Deformation partitioning
ing (Fig. 9.7(b)). For roughly equidimensional around a porphyroblast resulting from non-coaxial
progressive bulk inhomogeneous shortening. The
porphyroblasts, provided that they do not porphyroblast is outlined by the dashed line; the key to
deform, and provided that deformation parti- numbering is the same as in (a) (after Bell, 1985).
tioning is maintained, with perfect decoupling
between porphyroblast and matrix, they should
not rotate. With reference to the Mount Isa
region of Australia, Bell & Rubenach (1983)
successfully demonstrated how the variety of
common inclusion fabrics seen in syntectonic
porphyroblasts could be explained without the
need to invoke rotation. They interpreted the
various inclusion trails in terms of the timing of
porphyroblast nucleation and growth relative
to different stages in the development of an S2
crenulation cleavage (Fig. 9.8). In view of the
fact that most porphyroblasts grow very
rapidly once nucleated (Section 5.4), the inclu-
sion trails preserved represent a 'snapshot' of a FIG. 9.8 A sketch showing porphyroblasts that have
particular stage in textural and mineralogical overgrown various stages of development of a D2
evolution of the matrix. Inclusion trails with crenulation cleavage (after Bell, 1985).

Porphyroblast-foliation relationships

S-shaped form would represent porphyroblast well as from evidence provided in natural
growth at a time when the rock possessed a examples such as porphyroclasts (Fig. 10.16) in
strong crenulation fabric. Because a prominent rocks that have experienced moderate and high
crenulation cleavage can form with as little as shear strains (e.g. Mawer, 1987; Prior, 1987;
40-50% shortening, this might equally well Passchier, 1987; van den Driessche & Brun,
indicate nucleation early in the deformation 1987; Passchier & Sokoutis, 1993).
history or late-stage porphyroblastesis. Observations confirm that at suitably high
Although perfect decoupling between shear strains rigid inequidimensional objects
porphyroblast and matrix, and deformation can rotate through values at least up to
partitioning about equidimensional porphyrob- approximately 120°. However, if porphyroblast
lasts such as garnet, may theoretically suggest growth is geologically very rapid, most rotation
that they do not rotate during bulk simple should not occur during growth, and thus
shear, there is little evidence to suggest that rigid spiralled inclusion fabrics should not develop
elongate porphyroblasts (e.g. hornblende) (Fig.9.9, see also Barker, 1994). Many
would not be rotated. Many lines of evidence researchers would also argue that because there
suggest that during non-coaxial bulk shear is probably imperfect decoupling between
inequidimensional crystals experience a couple porphyroblast and matrix, then equidimen-
and consequently become rotated into the shear sional porphyroblasts such as garnet would
direction, without necessarily experiencing any also rotate.
internal deformation. Evidence for this has The non-rotation model of Bell and co-
been gained from shear-box modelling and workers has been the focus of considerable
from theoretical and computer modelling, as debate in recent years, with many aspects of

FIG. 9.9 Syntectonic hornblende porphyroblasts that have grown rapidly and in different orientations to over-
print an earlier stage of fabric development, and preserve straight inclusion fabrics. Continued shearing after the
short-lived phase of porphyroblastesis has rotated the two crystals into parallelism within the schistosity plane.
The larger crystal, which grew with its long axis at a high angle to the initial fabric, shows sharply discordant
Si-Se relationships. Hornblende mica-schist, Troms, Norway. Scale = 1 mm

Porphyroblast growth in relation to foliation

the model and interpretations called into ques- interpretation as a possibility, it is clearly diffi-
tion. The following paragraphs examine some cult to evaluate the validity of the non-rotation
of the main points of criticism, but for an model. Another point that makes the consis-
insight and full discussion of the main points tency of orientation of S, over a large geograph-
against the model see Pas schier et al. (1992), ical area all the more remarkable is that any
with a response by Bell et at. (1992), giving post-porphyroblastesis rigid block rotation
further comment in favour of the model and (e.g. by faulting) would cause rotation relative
answering the criticisms raised by Passchier to a fixed geographical reference frame. This
and co-workers. would cause rotation of Si from porphyroblasts
Fyson (1975, 1980), Bell (1985), Steinhardt of one block relative to Si in porphyroblasts of
(1989) and Bell et at. (1992) have argued that a neighbouring block.
the internal inclusion trails of porphyroblasts The porphyroblast non-rotation model
may maintain a remarkable consistency of developed by Bell and co-workers since the
orientation, both on the scale of a thin section mid-1980s has certainly renewed interest and
and over large regions within an orogenic belt. stimulated considerable discussion concerning
Consistency of Si-S, relationships on the scale the interpretation of porphyroblast-foliation
of a thin section are commonly reported, but relationships. The interpretation of 'shallow-S'
depending on your preferred model they can and certain other common porphyroblast inclu-
either be interpreted in terms of rotation of the sion trails in terms of porphyroblast over-
porphyroblasts by a similar amount relative to growth of specific stages of crenulation
Sc (e.g. Oleson, 1982), or by rotation of S, ret~­ development (after Bell & Rubenach, 1983)
tive to 'stationary' porphyroblasts that record has received popular support. However, the
consistent Si relating to some previously over- more radical interpretation that porphyroblasts
grown fabric (e.g. Fyson, 1980; Bell, 1985). do not rotate relative to fixed external co-ordi-
Steinhardt (1989), Johnson (1992) and Bell et nates during polyphase deformation, as
at. (1992) described examples in which S, of advanced by Bell (1985) and others, has not
porphyroblasts document near-constant orien- received the same level of support. On the basis
tation over large geographical areas despite of currently available data, the majority of
intense later deformation. This sort of informa- metamorphic geologists still favour the tradi-
tion has been used by the authors cited to indi- tional view that porphyroblasts do rotate rela-
cate non-rotation of porphyroblasts relative to tive to fixed external co-ordinates.
each other. Indeed, Bell (1985) went as far as to
suggest that constantly oriented S, indicates Millipede microstructures
that the porphyroblasts have not rotated rela- The term 'millipede microstructure' was intro-
tive to fixed geographical co-ordinates and that duced by Bell & Rubenach (1980) for certain
the preserved Si indicates the original orienta- syntectonic porphyroblasts which, at a given
tion of an earlier fabric. Passchier et at. (1992) margin at which Si passes into Sr, show S,
emphasised that in cases in which S, shows deflected in opposite directions (Plate 5 (e)).
constancy of orientation over a large area, Bell Millipede microstructures were interpreted by
and others would use this as evidence to advo- Bell & Rubenach (1980) as evidence for bulk
cate non-rotation of porphyroblasts. However, heterogeneous shortening that, at least locally,
when S, shows variable orientation across an is near coaxial. Critical discussions by
area, Bell and others have argued in terms of Passchier et al. (1992) and Johnson (1993b)
variable foliation orientation (due to folding) indicate that such microstructures can develop
prior to porphyroblastesis. With this sort of in various deformation regimes. Recent papers

Porphyroblast-foliation relationships

by Johnson & Bell (1996) and Johnson & morphism associated with uplift. Pseudo-
Moore (1996) discuss the usefulness of 'milli- morphing of one mineral by an aggregate of
pedes' and other types of oppositely concave another mineral (e.g. garnet pseudomorphed
microfolds (OCMs) seen adjacent to some by chlorite; see Plate 4(c) and Chapter 7)
porphyroblasts. The type 1 OCMs of Johnson is typically a late-tectonic or post-tectonic
& Bell (1996) are the cut-effect aD-type feature associated with retrogression during
(closed loop) inclusion trails (Section 9.1, Fig. uplift and waning P-T conditions. Since post-
9.1) described by Powell & Treagus (1970) and tectonic crystals have nucleated and grown
MacQueen & Powell (1977). They are seen in after the main phases of deformation, they will
thin sections cut parallel to the spiral axis of not show any strain effects. This means that
porphyroblasts with S-shaped inclusion trails. they will not exhibit features such as undulose
Porphyroblasts displaying such inclusion extinction or kinking, and will lack pressure
patterns are classically interpreted as forming shadows (Plate 6(a)).
during a single phase of growth in a regime of Elongate post-tectonic minerals usually
strongly non-coaxial deformation. Johnson show a lack of preferred orientation (Plate
(1993b) reports that such microstructures are 6(b)), in stark contrast to the pronounced
equally well explained by the porphyroblast alignment of comparable pre-foliation and
non-rotation model as they are by the porphy- syntectonic crystals. Post-tectonic overgrowth
roblast rotation model. The Type 2 OCMs of of crenulations commonly gives rise to gently
Johnson & Bell (1996) are the classic 'milli- curved to S-shaped internal fabrics in porphy-
pede' microstructures of Bell & Rubenach roblasts (Plates 6(c) & (f)). Superficially, these
(1980). As stated above, these are commonly may appear similar to the syntectonic
considered indicative of bulk coaxial shorten- 'spiralled' inclusion fabrics described earlier.
ing, but Johnson & Bell (1996) and Johnson & However, careful examination will reveal that
Moore (1996) have demostrated that they may the matrix does not deflect around post-
form by heterogeneous extension, and cannot tectonic porphyroblasts (whereas it does
be used to indicate specific deformation histo- around syntectonic crystals), and that the
ries (i.e. degree of non-coaxiality of the defor- crenulations being overgrown are present
mation). throughout the thin section. Where post-
tectonic crenulation overgrowth has occurred,
a range of different inclusion fabrics usually
9.2.3 Recognition and interpretation of
exists, as a function of different parts of the
post-tectonic crystals
crenulation being overgrown. It is not uncom-
Post-tectonic crystals are most easily distin- mon to see several crenulations overgrown by a
guished by their complete overgrowth of the single porphyroblast (e.g. Plate 6(f)).
main fabric (cleavage or schistosity) of the
rock. That is to say, the fabric does not
9.2.4 Complex porphyroblast inclusion trails
envelop the porphyroblasts, but abuts against
and multiple growth stages
the crystal, and in cases in which an inclusion
fabric is seen the fabric passes right through Late-stage metamorphism does not always
the porphyroblast (Plate 6(a)). Post-tectonic produce entirely new porphyroblasts, but in
crystals are characteristically associated with many examples post-tectonic rim growth on
contact metamorphism adjacent to igneous previously formed syntectonic porphyroblasts
intrusions, late thermal overprints during is recognised (Plates 6(d) & (e)). Where rocks
regional metamorphism and retrograde meta- (especially those of orogenic metamorphism)

Porphyroblast growth in relation to foliation

have experienced a late thermal overprint, it is (1989) are really deflection planes. According
common to find rim growth on pre-existing to Passchier et al. (1992), such deflection
porphyroblasts. In such cases a sharp break in planes can be explained in terms of a single
textural zonation is often seen. The core will phase of progressive deformation, and conse-
often show syntectonic relationships, while the quently the porphyroblasts illustrated by Bell
rim exhibits post-tectonic features and over- & Johnson (1989) could be interpreted in
grows the matrix foliation (e.g. Plate 6(d)). terms of far fewer deformation events and
Rim overgrowths can be complete (Plate 6(e)) overgrowth stages than they suggest.
or partial (Plate 6(d)). The post-tectonic rims From the above discussion, it is apparent
generally form at the margin in contact with that despite much research over many
mica-rich matrix, whereas at margins in decades, there remains considerable debate
contact with quartz-rich layers or pressure and controversy concerning the interpretation
shadows the rim is absent or poorly developed. of even the simplest types of inter-tectonic and
The most probable explanations for this are syntectonic porphyroblast inclusion trails. For
either in terms of locally unfavourable chem- porphyroblasts with multiple growth stages
istry of the quartz-rich areas, or easier nucle- and more complex inclusion trails, the prob-
ation and growth in the more highly strained lem is magnified, and it not surprising that the
areas out of the pressure shadows. interpretations vary greatly from one author
More complex porphyroblast inclusion to the next. There are clearly many secrets
patterns are also encountered, especially in locked away in inclusion trails of complex
rocks that have experienced a polyphase defor- syntectonic porphyroblasts, and despite their
mation and metamorphic history during great potential for unravelling deformation-
regional orogenesis (e.g. MacQueen & Powell, metamorphism interrelationships during oro-
1977; Bell & Johnson, 1989; Hayward, 1992). genesis, they remain some of the most difficult
These more complex inclusion patterns microstructures to interpret in a reliable and
commonly show distinct inflexions or breaks in objective manner.
inclusion trails, and are generally interpreted in
terms of multistage porphyroblast growth, and
overgrowth of more than one deformation
event (see also Chapter 12). Bell & Johnson Barker, A.j. (1994) Interpretation of porphyroblast
inclusion trails: limitations imposed by growth
(1989) illustrated porphyroblasts that they kinetics and strain rates. Journal of Metamorphic
interpret to have overgrown up to eight succes- Geology, 12, 681-694.
sive foliations relating to different deformation Bell, T.H. (1985) Deformation partitioning and
events. They suggested that given the right porphyroblast rotation in metamorphic rocks: a
radical reinterpretation. Journal of Metamorphic
garnet and a favourable situation, the full Geology, 3, 109-118.
history of orogenesis can be preserved in a Bell, T.H. & Johnson, S.E. (1989) Porphyroblast inclu-
single porphyroblast. They interpreted the sion trails: the key to orogenesis. Journal of
complex inclusion trails that they describe from Metamorphic Geology, 7,279-310.
Bell, I.H. & Rubenach, M.]. (1980) Crenulation cleav-
Vermont (USA), Pakistan and Nepal in terms age development - evidence for progressive, bulk
of non-rotation of porphyroblasts, and as indi- inhomogeneous shortening from 'millipede'
cating repeated cycles of uplift and collapse microstructures in the Robertson River Meta-
during orogenesis. Many researchers have seri- morphics. Tectonophysics, 68, T9-T15.
Bell, T.H. & Rubenach, M.]. (1983) Sequential
ous reservations about this, and Passchier et al. porphyroblast growth and crenulation cleavage
(1992) suggested that many of the so-called development during progressive deformation.
truncation planes reported by Bell & Johnson Tectonophysics, 92, 171-194.

Porphyroblast-foliation relationships

Bell, T.H., Johnson, S.E., Davis, B., Forde, A., Mawer, CK. (1987) Shear criteria in the Grenville
Hayward, N. & Wilkins, C (1992) Porphyroblast Province, Ontario, Canada. Journal of Structural
inclusion trail orientation data: eppure non son Geology, 9, 531-539.
girate! Journal of Metamorphic Geology, 10, Oleson, N.0. (1982) Heterogeneous strain of a phyllite
295-307. as revealed by porphyroblast-matrix relationships.
Fyson, W.K. (1975) Fabrics and deformation of Journal of Structural Geology, 4, 481-490.
Archaean metasedimentary rocks, Ross Lake - Pas schier, CW. (1987) Stable positions of rigid objects
Gordon Lake area, Slave Province, Northwest in non-coaxial flow: a study of vorticity analysis.
Territories. Canadian Journal of Earth Sciences, 12, Journal of Structural Geology, 9, 679-690.
765-776. Passchier, CW. & Sokoutis, D. (1993) Experimental
Fyson, W.K. (1980) Fold fabrics and emplacement of modelling of mantled porphyroclasts. Journal of
an Archaean granitoid pluton, Cleft Lake, Structural Geology, 15, 895-910.
Northwest Territories. Canadian Journal of Earth Passchier, CW., Trouw, R.A.J., Zwart, H.J. & Vissers,
Sciences, 17, 325-332. R.L.M. (1992) Porphyroblast rotation: eppur si
Hayward, N. (1992) Microstructural analysis of the muove? Journal of Metamorphic Geology, 10,
classical spiral garnet porphyroblasts of south-east 283-294.
Vermont: evidence for non-rotation. Journal of Powell, D. & Treagus, J.E. (1970) Rotational fabrics in
Metamorphic Geology, 10, 567-587. metamorphic minerals. Mineralogical Magazine, 37,
Johnson, S.E. (1992) Sequential porphyroblast growth 801-813.
during progressive deformation and low-P Prior, D. (1987) Syntectonic porphyroblast growth in
high-T (LPHT) metamorphism, Cooma Complex, phyllites: textures and processes. Journal of
Australia: the use of microstructural analysis in Metamorphic Geology,S, 27-39.
better understanding deformation and metamorphic Rosenfeld, J.L. (1970) Rotated garnets in metamorphic
histories. Tectonophysics, 214, 311-339. rocks. Geological Society of America Special Paper
Johnson, S.E. (1993a) Unravelling the spirals: a serial 129, 105 pp.
thin section study and three-dimensional computer- Schoneveld, Chr. (1977) A study of some typical inclu-
aided reconstruction of spiral-shaped inclusion trails sion patterns in strongly paracrystalline garnets.
in garnet porphyroblasts. Journal of Metamorphic Tectonophysics, 39, 453-47l.
Geology, 11,621-634. Spry, A. (1963) The origin and significance of snowball
Johnson, S.E. (1993b) Testing models for the develop- structure in garnet. Journal of Petrology, 4,
ment of spiral-shaped inclusion trails in garnet 211-222.
porphyroblasts: to rotate or not to rotate, that is Steinhardt, CK. (1989) Lack of porphyroblast rotation
the question. Journal of Metamorphic Geology, 11, in non-coaxially deformed schists from Petrel
635-659. Cove, South Australia, and its implications.
Johnson, S.E. & Bell, T.H. (1996) How useful are Tectonophysics, 158, 127-140.
'millipede' and other similar porphyroblast van den Driessche, J. & Brun, J.P. (1987) Rolling struc-
microstructures for determining synmetamorphic tures at large shear strain. Journal of Structural
deformation histories? Journal of Metamorphic Geology, 9, 691-704.
Geology, 14, 15-28. Vernon, R.H., Collins, W.J. & Paterson, S.R. (1993)
Johnson, S.E. & Moore, R.R. (1996) De-bugging the Pre-foliation metamorphism in low-pressure/high-
'millipede' porphyroblast microstructure: a serial temperature terrains. Tectonophysics, 219,
thin-section study and 3-D computer animation. 241-256.
Journal of Metamorphic Geology, 14,3-14. Wilson, M.R. (1971) On syntectonic porphyroblast
Macqueen, J.A. & Powell, D. (1977) Relationships growth. Tectonophysics, 11,239-260.
between deformation and garnet growth in Moine Zwart, H.J. (1962) On the determination of polymeta-
(Precambrian) rocks of western Scotland. Bulletin morphic associations and its application to the
of the Geological Society of America, 88, Bosost area (central Pyrenees). Geologische
235-240. Rundschau, 52, 38-65.

Chapter ten


10.1 Introduction assessment of the amount of shear, and could

give a false impression of shear sense. Because
It is common in a wide range of metamorphic
of this they should be avoided where possible.
rocks from a variety of metamorphic environ-
Another point to bear in mind is that fault
ments to see evidence for a component of non-
and shear zones that have been reactivated at
coaxial deformation. This chapter describes
various times during their evolution will typi-
some of the most commonly observed kine-
cally be dominated by textures and microstruc-
matic indicators in metamorphic rocks, and
tures developed in the last stage of movement,
discusses how they may be used to assess sense
or the most intense period of deformation.
of shear. In essence, there are seven main cate-
However, there may be domains within the
gories to consider, and these can be grouped
shear zone that preserve elements of earlier
under the following headings:
stages in the evolution, such that the shear
(i) sense of fold overturning and vein asym- sense may not always be consistently in the
metry; same direction. For example, it is quite
(ii) S-C fabrics, shear bands and mica-fish; common in fold and thrust belts to find early
(iii) differentiated crenulation cleavages; compression-related thrusts that become reacti-
(iv) spiralled inclusion fabrics; vated as extensional structures during late-
(v) mantled porphyroclasts; orogenic relaxation and collapse. As a result,
(vi) strain shadows; the shear-sense indicators in such zones will
(vii) grain-shape fabrics and crystallographic often show mixed polarity. It is therefore
preferred orientations. crucial to be fully aware of the deformation
phase to which a given shear-sense indicator
In attempting to determine the shear sense, and
relates, and not to mix shear-sense indicators
where possible the amount of shear, it is impor-
from various events.
tant to have an oriented sample. The thin
section should be cut parallel to the X-direction
of the finite strain ellipsoid (generally defined
10.2 Vein asymmetry and sense of fold
by stretching/extension lineation of the rock) overturning
and perpendicular to the X-Y plane (generally
defined by cleavage/schistosity). The slice The attitude of veins, and sense of vergence of
examined in thin section is therefore the X-Z asymmetrical folds, are features commonly used
plane, where X, Y and Z are the principal axes in the field to determine sense of shear (both on
of the finite strain ellipsoid (X ~ Y ~ Z). a local and on a regional scale). Such structures
Oblique cuts will certainly give erroneous exist on a variety of scales, and thin-section

Shear-sense indicators

studies often prove helpful in determining the establish that simple shear is primarily respon-
precise relationships between veining and sible for vein folding, remembering that folded
deformation. veins are also produced by pure shear flatten-
The history of vein development in a rock is ing of early veins oriented at a high angle to the
often complex, with some veins being present principal axis of compression.
pre-shearing, others forming during shearing In the case of veins formed during simple
and some forming post-shearing. By careful shear, they will tend to form as en echelon sets
study of cross-cutting relationships between in fractures developed perpendicular to the
different vein sets, and studies of relationships direction of maximum extension (Fig. 10.2(a)).
between veins and other structures of the rock These become rotated as shear progresses, and
(e.g. folds and foliations), it should be possible during rotation the vein develops a sigmoidal
to determine the full history of vein develop- trace (Figs 10.2(b) & 11.2) from which sense
ment. of shear can be determined ('S' -shape = sinis-
The superposition of a period of simple tral; 'Z'-shape = dextral). The sense of fold
shear on variably oriented early veins will lead overturning or vergence can also be used to
to folding or crenulation of some veins, while give a general indication of shear sense.
others become thinned and boudinaged. How a Initially gentle, slightly overturned fold struc-
particular vein behaves depends on its initial tures become increasingly overturned and
attitude relative to the superposed strain ellip- asymmetrical as shear progresses (Fig. 10.3).
soid (Fig. 10.1). Veins that lie in the shortening Such asymmetrical folds can range from
field of the strain ellipsoid (e.g. A) will become regional-scale structures down to mm-scale
folded, while those in the maximum elongation crenulations observable in thin section.
(extension) direction (e.g. B) will become Although commonly recorded, the use of fold
thinned and boudinaged. In the example illus- asymmetry as the sole line of evidence for shear
trated, increasing shear will rotate the veins in sense is not recommended. This is because
a clockwise sense, so that vein A may ulti- similar features can be produced by pure shear
mately move into the extensional field and flattening of features oblique to the principal
undergo unfolding and boudinage. compressional axes (Fig. 10.4), and in such
The foregoing discussion demonstrates that situations any shear-sense interpretation would
with careful observation early veins can give be entirely false.
important information relating to the sense of In brittle, sub-greenschist facies regimes,
any superimposed shear. In using this Sibson (1990) illustrates how vein and fracture
approach, care should always be taken to arrays and networks developed to accommo-


FIG. 10.1 A schematic illustration of veins deforming

during superimposed simple shear. The style of defor-
mation is a function of vein orientation relative to the
superimposed strain ellipsoid, and can be used to deter- FIG. 10.2 A schematic illustration showing the devel-
mine the shear sense. opment of sigmoidal tension gashes.

S-C fabrics, shear bands and mica-fish

FIG. 10.3 A schematic illustration of the development of asymmetrical folds during simple shear.

date local extension at dilational jog sites may

be used to evaluate shear sense (Fig. 10.5).
Although characteristically observed on the
exposure-scale, many such features can also be
seen on hand-specimen and thin-section scale.

10.3 S-C fabrics, shear bands and

In rocks such as mylonites and phyllonites that
have experienced high strain during intense
FIG. lOA A schematic illustration of asymmetrical non-coaxial ductile shear, it has been recog-
folded veins formed during pure shear.
nised that two fabrics commonly develop

FIG. 10.5 A schematic illustration of the various vein/fracture networks that are commonly observed at dila-
tional fault jog sites, and which can be used to evaluate shear sense. Such features are not scale-dependent, and
can be observed from map-scale down to thin-section scale: (a) ladder vein; (b) cymoid loop; (c) horsetail; (d, e)
alternate shear/extensional mesh models for dilational jogs (based on Fig. 4.16 of Sibson, 1990; courtesy of the
Mineralogical Society of Canada).

Shear-sense indicators

simultaneously in association with this shear- 10.7). At initiation the S fabric develops at an
ing. These are S-C fabrics (Fig. 10.6), the C angle (a) of about 45° to C, but as shearing
surfaces being parallel with the shear zone progresses the angle a diminishes as S rotates
margin while the S-surfaces are oblique to this towards C. Most typically a is between 15°and
(e.g. Berthe et aI., 1979; Lister & Snoke, 1984; 45°, but in extreme cases the two surfaces may
Passchier, 1991; Blenkinsop & Treloar, 1995). become sub-parallel. In general, as the angle a
The field appearance of such rocks is usually decreases, so the density of C surfaces
very distinctive, with the intersection of the Increases.
two cleavages giving rise to a texture which has With knowledge of the attitude of S fabrics,
commonly been described as 'fish-scale', and the shear-zone boundary orientation (which
'button schist' or 'oyster shell' texture (Fig. is parallel to C fabrics), it is thus possible to
ascertain the sense of shear of the zone in ques-
tion (Fig. 10.6). The detail of such fabrics is
often well exhibited in thin section, and when
present should corroborate evidence from D-type
porphyroclasts (Section 10.6). Unfortunately,
this simple picture of S-C fabrics is complicated
by secondary planar fabrics commonly devel-
oped at a later stage in such rocks. These are
usually antithetic to the orientation of S
s s s surfaces, and are often more pronounced and
more easily recognised than such surfaces. They
FIG. 10.6 A schematic illustration of SoC fabrics, and are termed C' fabrics (e.g. Ponce de Leon &
their relationship to the direction of shear. Choukroune, 1980) or extensional crenulation

FIG. 10.7 S-C and C' fabrics developed in phyllonite of the Abisko Nappe, Swedish Caledonides. The compass-
clinometer (lower left), for scale, is 8 em long.

S-C fabrics, shear bands and mica-fish

cleavages (e.g. Platt & Vissers, 1980; Platt,

1984; Dennis & Secor, 1987), and are especially -~~ ---- .......

-- -
~ :.---
well developed in phyllonites (Figs 10.7 &
10.8). In the brittle regime, synthetic Reidel - ---~---~-- --~:.:: ~

shears (e.g. RI of Logan et al., 1979; Chester et FIG. 10.10 A schematic illustration of mica-fish as
al., 1985; Arboleya & Engelder, 1995) have the shear indicators.
equivalent orientation to C' surfaces of ductile
shear zones.
In ductile thrust zones and 'slides', C'
surfaces characteristically dip towards the fore-
land, and may thus be used to indicate bulk
transport direction. They usually form an angle
of approximately 15-35° to C surfaces and the FIG. 10.11 Trails and stair-stepping between mica-

(b) ~

FIG. 10.8 A schematic illustration of C' fabrics in FIG. 10.12 Multiplication of mica-fish by break-up
relation to Sand C fabrics. and dispersion.

. -

FlG.10.9 Mica-(biotite-)fish in schist of the Hogtind nappe, Norwegian Caledonides. Scale = 1 mm (PPL).

Shear·sense indicators

shear-zone boundary. In thrusts of compres-

sional orogenic belts they are commonly associ-
ated with late-stage orogenic collapse. The
term shear band is used descriptively in many
publications to define discrete highly sheared
surfaces within mylonitic rocks. 'Shear bands'

are recognised in several orientations, and
while normally related to C' surfaces as
described above, C surfaces have also been
recognised as 'shear bands'. Since the term
'shear band' does not imply a specific orienta- FIG. 10.13 Crenulation cleavage asymmetry as a basis
for evaluation of local shear sense (based on Fig. 11 of
tion, it is recommended (in order to avoid Bell & Johnson, 1992).
confusion) that it only be used to describe the
nature of C or C' surfaces, and not to have any
general orientational implications.
Mica-fish, termed as such because of their
generally 'fish-like' lozenge-shaped appearance,
are useful kinematic indicators in many
mylonites, phyllonites and schists. The mica
crystals are large pre-existing grains (effectively
'porphyroclasts') or early porphyroblasts which
are deformed by a combination of brittle and
crystal-plastic processes. The asymmetry of the
mica 'fish' shape (Fig. 10.9), with 001 cleavage
typically either tilted sub-parallel to Sore
surfaces can be used to determine sense of FIG.10.14 Determination of the shear sense around a
shear (Fig. 10.10). Lister & Snoke (1984) fold on the basis of minor structure asymmetry (based
describe how mica-fish are usually linked by on Fig. 13 of Bell & Johnson, 1992).
thin trails of mica fragments, and when exten-
sively developed show a stair-stepping from
one mica-fish to the next (Fig. 10.11). Mica- try due to deflection when passing from
fish may multiply and disperse by a process domains of low strain to domains of high
typically involving slip on the basal plane, strain (Fig. 10.13). Deflection of old foliations
causing the break-up of early large micas: this in the manner illustrated in Fig. 10.13 is a
is illustrated in Fig. 10.12. common feature of many metamorphic rocks
of the greenschist facies, amphibolite facies and
in shear zones. They are especially pronounced
10.4 Differentiated crenulation cleavages where a strongly differentiated crenulation
Bell & Johnson (1992) describe how shear cleavage is developed, and since this a common
sense can be obtained from crenulation asym- feature, are potentially one of the most useful
metry against differentiated crenulation cleav- indicators of local shear sense. Bell & Johnson
ages. Indeed, they describe how any old (1992) illustrate how local shear sense on each
foliation being rotated during the development limb of a fold can be usefully determined using
of a new foliation can be used to determine differentiated crenulation cleavage as a shear-
local shear sense, as long as it shows asymme- sense indicator, and how this local sense of

Mantled porphyroclasts and 'rolling structures'

shear is comparable to the bulk sense of shear previously been considered (i.e. commonly < 1
in limbs between antiform and synform. It will Ma, and even < 0.1 Ma; Barker (1994) and
be noticed (Fig 10.14) that this shearing is Section 5.4 ). This means that under most
directed sub-parallel to the fold axial surface, circumstances porphyroblasts develop very
even though the area as a whole is experiencing rapidly with respect to ongoing deformation,
inhomogeneous bulk shortening perpendicular and that many S-shaped inclusion trails proba-
to the axial surface. In view of this, the impor- bly represent syntectonic crenulation over-
tance of gaining a complete understanding of growth. Clearly, this sheds still further doubt
structures on all scales should be apparent in on the use of S-shaped inclusion trails as kine-
order to evaluate local and regional shear sense matic indicators, and consequently their use is
and tectonometamorphic evolution of a given not recommended. More reliable are 'rolling
terrain. structures' and related features characteristic of
porphyroclast systems.
10.5 Spiralled inclusion trails
10.6 Mantled porphyroclasts and 'rolling
In Chapter 9 the traditional, and now contro-
versial interpretation of S-shaped inclusion
trails in syntectonic porphyroblasts as evidence Pas schier & Simpson (1986) review porphyro-
for porphyroblast rotation is discussed. The clast systems as kinematic indicators, with more
development of such fabrics in this manner, recent work undertaken by Passchier et al.
and their use as shear-sense indicators is illus- (1993) and Pas schier (1994). On geometric
trated in Fig. 9.5. More recently, it has been grounds, Passchier & Simpson (1986) subdivide
argued that porphyroblasts, and especially porphyroclasts into 0'- and 8-types (Fig. 10.15).
equidimensional porphyroblasts such as garnet, O'-type porphyroclasts have wedge-shaped tails
do not rotate (e.g. Bell, 1985). Instead, the and a 'stair-stepping' symmetry (Fig. 10.15(a)).
external fabric changes orientation with respect The median lines of the recrystallised tails (of
to the porphyroblast by way of progressive the same mineral as the porphyroclast) lie on
foliation transposition or shearing. When this each side of a central reference plane. The side
occurs synchronous with porphyroblast growth nearest the reference plane characteristically has
it too will give rise to an S-shaped inclusion a concave curvature, while that farthest away is
trail (Fig. 9.6), but with a completely opposite planar. O'-type porphyroclasts are themselves
sense of rotation to that deduced by assuming subdivided into 0'.- and O'b-types. The former
that the porphyroblast rotated (d. Fig. 9.5). represent isolated porphyroclasts (e.g. feldspar
Johnson (1993) reviews spiral-shaped inclusion or hornblende) in a homogeneous foliation of
trails of porphyroblasts in some depth, and uniform orientation, though locally deflected
concludes that most geometries are consistent around the porphyroclast (Fig. 10.16). In some
with both the rotation and the non-rotation cases porphyroclasts remain rigid, and pressure
models. Because of this, and in view of this shadows ('beards') of a different mineral phase
major difference in shear sense being a function (e.g. quartz) develop with geometry similar to
of the interpretation of the mechanism by O'.-type tails. O'b-type porphyroclasts are gener-
which the S fabric formed, any interpretation ally feldspars associated with S-C quartz
based on such fabrics should be made with feldspar mylonites (e.g. Berthe et aI., 1979).
extreme caution. Adding still further to the They have flat surfaces along the C planes, and
problem, it is now generally realised that tend to occur in clusters. Like 0'. -type clasts, the
porphyroblast growth is much faster than had recrystallised tails of O'b-type clasts can be of

Shear-sense indicators

different or mixed composition in comparison the tails crosses the central reference plane
to the clast itself. (Fig. 10.lS{b)). In B-type porphyroclasts the tails
&--type porphyroclast systems (also termed are thin, and tight embayments exist between
'rolling structures'), differ from a-type systems the tail and porphyroclast. Such tails only occur
by virtue of the fact that the median-line of around equidimensional or very slightly elongate
porphyroclasts, and usually extend for a consid-
erable distance from the porphyroclast (Figs
10.lS{b) & 10.17). On the basis of studies of
(e) ~
(8) granitoid mylonites from the Grenville Province,
CJ Mawer (1987) gave a clear illustration of how B-
type porphyroclasts can develop from a-type
porphyroclasts during progressive shearing (Fig.
(b) 10.lS{c)). In cases of prolonged or episodic
shearing, more complex examples are likely,
including folding of thin B-type tails (Fig. 10.18),
and complex a-B type porphyroclast relation-
ships (Fig. 10.19). It is important to note that Z-
shaped strain shadow and a-type porphyroclast
FIG. 10.15 Porphyroclast systems as kinematic indica- asymmetry indicates the same sense of shear as
tors. (a) A (J-type porphyroclast (modified after Fig. 2a S-shaped B-type porphyroclast ('rolling struc-
of Passchier & Simpson, 1986). (b) A c)-type porphyro-
clast (modified after Fig. 2e of Passchier & Simpson, ture') asymmetry (compare Figs 10.lS{a) & (b)).
1986). (c) The development of c)-type porphyroclasts Without this awareness and the ability to distin-
from (J-type porphyroclasts (modified from Mawer, guish B-type porphyroclasts, the sense of shear
1987). can easily be misinterpreted.

FIG. 10.16 A (Ja-type feldspar porphyroclast. Granitoid mylonite, South Armorican Shear Zone, France. Scale =
0.5 mm (PPL).

Strain shadows

FIG. 10.17 A O-type feldspar porphyroclast (after Fig. Sa (polished rock slab) of Mawer, 1987; courtesy of
Elsevier Science). Scale = 2 cm .

·overturned" systems
FIG. 10.18 A folded tail of I)-type porphyroclast (after
Fig. 3 of van den Driessche & Brun, 1987). complex (J • I'l systems

10.7 Strain shadows 'eeocncI~

Ramsay & Huber (1983), in their review of

pressure (strain) shadow characteristics distin-
guish two main types of pressure (strain) shad- FIG. 10.19 Overturned and complex (J-O porphyro-
clast systems (after Fig. 2h of Passchier & Simpson,
ows; namely, 'pyrite type' and 'crinoid type'. 1986).
The latter involves progressive fibre growth (of
the same mineral type as the rigid object)
between the object or its strain shadow and carbonates), but in general is rare in metamor-
the displaced matrix surface. This type of phic rocks. Crystallographic continuity is
strain shadow occurs around crinoid ossicles retained from object to fibre.
in deformed limestones (low-grade meta- 'Pyrite-type' strain shadows are very

Shear-sense indicators

common in deformed greenschist facies and the interface between the core (usually porphy-
lower-grade metamorphic rocks. The develop- roblast or porphyroclast) and the inner surface
ment of such fibrous shadows (or 'fringes') of its strain shadow. In some instances compos-
involves incremental fibre growth (of mineral ite pressure shadows develop involving fibres
species different from the rigid object 'core') at of several mineral species (Figs lO.20(a) & (b)).



Fig. 10.20 X-Z sections of three natural examples of pressure shadows. In all cases the scale is 1 mill. (a), (e)
XPL; (b) PPL. (a) Fine-grained cordierite schist/phyllite with complex deformable pressure shadows of calcite (a),
quartz (b) and Qtz + Ms (c) adjacent to the cordierite porphyroblasts. The cordierite appears very dark because it
contains extremely abundant and fine-grained inclusions of quartz and opaques. (b) Pyritic tuff with pressure
shadows of quartz (a) and chlorite (b) adjacent to euhedral pyrite crystals.

Strain shadows

Depending on the way in which they grow, along the displacement path, between pressure
fibres of 'pyrite-type' shadows can be subdi- shadow wall and towards the resistant object
vided into two categories. The first of these, (Fig. 10.21). The second type, 'face-controlled
'displacement-controlled fibres', show consis- fibres', exhibit fibre growth normal to the
tent geometry of progressive growth of fibres face(s) of the rigid object, irrespective of the


(c) A framboidal pyrite crystal in slate, with deformable pressure shadows of fibrous quartz (a) and calcite (b).

FIG. 10.21 Displacement-controlled fibres around pyrite (after Etchecopar & Malavieille, 1987). Scale = 1 mm
(XPL). (Reproduced with permission of Elsevier Science).

Shear-sense indicators

FIG. 10.22 Face-controlled quartz fibres developed In a pressure shadow adjacent to pyrite. Slate, south
Cornwall, England. Scale = 1 mm (XPL).

r tf:


FIG. 10.23 Geometric differences between rigid fibre (r) and deformable fibre (d) pressure shadows in situations
of coaxial (a) and non-coaxial (b) deformation (after Ramsay & Huber, 1983, Figs 14.15-14.16; courtesy of
Academic Press).

Strain shadows

displacement directions (Fig. 10.22). of the best criteria for determining sense of
Depending on the P-T conditions, the shearing in rocks. Such analysis clearly has
fibres/pressure shadows may be deformable or wide application in the study of metamorphic
rigid. At high temperatures and in more ductile terrains. Most obviously, pressure shadows
conditions, pressure shadows are generally around pyrite porphyroblasts from low-grade
deformable with strong recrystallisation while 'slate belts' can be studied, but pressure shad-
under lower-temperature conditions they are ows around porphyroblasts, such as garnet in
usually rigid and fibrous. The geometric higher-grade schists, can also be assessed in
contrasts between 'rigid fibre' (r) and order to determine the sense and amount of
'deformable fibre' (d) pressure shadows in the shear.
situation of (a) coaxial and (b) non-coaxial The types of pressure shadow fibres
deformation are visually summarised in Fig. predicted for successive amounts of shear (r =
10.23. 0-4; all other variables constant), for a model
Studies by Etchecopar & Malavieille (1987) based on 'deformable and face-controlled'
compare various computer-generated 'pyrite- fibres, are shown in Fig. 10.24. In contrast, the
type' pressure shadows with examples taken nature of pressure shadow fibres predicted for
from natural rocks. From this they are able to 'rigid and displacement controlled' fibres over
estimate the bulk strain and shear sense in the the same range of incremental shear strain is
deformed rocks. They consider pressure illustrated in Fig. 10.25. In Fig. 10.26, the
shadow asymmetry in X-Z sections to be one predicted developments in situations of pure



FIG. 10.24 A computer simulation of deformable, face controlled pressure shadow fibres developed after various
increments of shear strain (i.e. Y= 0-4) (modified after Fig. 5 of Etchecopar & Malaveille, 1987).

FIG. 10.25 A computer simulation of rigid, displacement-controlled pressure shadow fibres after various incre-
ments of shear strain (i.e. Y= 0-4) (modified after Fig. 6 of Etchecopar & Malavieille, 1987).

Shear·sense indicators

FIG. 10.26 A computer simulation of pressure shadow fibres expected in cases of pure shear flattening, assuming
the fibres to be rigid and displacement-controlled (after Fig. 7 of Etchecopar & Malavieille, 1987). (Reproduced
with permission of Elsevier Science) .

shear flattening are illustrated. By comparison Lister & Snoke, 1984; Schmid et ai., 1987;
with Fig. 10.24, it is apparent that the natural De Bresser, 1989; Shelley, 1995). Such fabrics
example of pressure shadow fibres illustrated are especially common in monomineralic
in Fig. 10.22 is face-controlled, and indicative greenschist facies mylonites, and usually show
of bulk shear of the order of r = 1. x-z an obliquity of 25-40' to the shear-zone
sections for some more natural examples of margin. Oblique shape fabrics are defined by
pressure shadows are shown in Fig. 10.20. By alignment of the long axes of dynamically
comparison with Figs 10.24-10.26, a rough recrystallised grains, especially of quartz or
visual estimate of the dominant fibre type and calcite, depending on the lithology. Since the
amount and sense of shear can be made for rock, and grains, may recrystallise more than
each (see the figure caption for details). once during the deformation and metamor-
phic history, the particular shape fabric will
10.8 Grain-shape fabrics and probably relate to the latest stages of defor-
crystallographic preferred orientations mation only. Nevertheless, this may be a
useful kinematic indicator, since the flattening
It is common, especially in quartzofeldspathic plane of the new grains will lie approximately
mylonites and highly sheared carbonates, to perpendicular to the direction of maximum
observe distinct grain-shape fabrics (Fig. compression during the final increment of
10.27) oblique to the shear-zone margins and deformation (Fig. 10.27). Lister & Snoke
C surfaces (e.g. Simpson & Schmid, 1983; (1984) illustrate how in mylonites, the atti-


Lister & Snoke, 1984). Similarly, the obliquity

of lattice preferred orientation in olivine with
respect to a given foliation is widely accepted
as an expression of rotational deformation and
may be used to determine sense of shear
(Etchecopar & Vasseur, 1987). Wenk et al.
(1987) give a detailed account of how calcite
FIG. 10.27 A schematic illustration of oblique grain
shape fabrics in relation to mica-fish and shear sense.
CPOs in marble can be used in distinguishing
pure shear from simple shear, assessing the
sense of shear and estimating finite strain.
Despite the undoubted value of CPOs as
tude of quartz grain fabrics agrees with the
shear-sense indicators, a degree of caution
sense of shear indicated by mica-fish. More
should always be exercised because, in some
recently, Schmid et al. (1987) give an account
cases, opposing senses of shear may be
of results from simple shear experiments on
recorded within different samples from the
calcite rocks. In particular, their study
same thrust zone (e.g. Bell & Johnson, 1992).
describes the nature of oblique calcite grain-
There are a number of reasons why this may
shape fabrics produced by dextral shearing of
occur. One reason is that the presence of
Carrara marble and Solenhofen limestone at
other phases in the matrix aggregate will
various temperatures.
affect the nature of the CPO recorded and,
As well as grain shape fabrics, many miner-
second, recrystallisation may cause weaken-
als (e.g. quartz, calcite and olivine) are known
ing of the CPO, such that the pattern is less
to have a crystallographic preferred orientation
distinct and shear-sense evaluation becomes
(CPO) in deformed rocks. Ultramylonites
less clear cut.
developed in quartz-rich rocks can superficially
resemble fine-grained quartzites (Fig. 8.13(c)).
However, the insertion of the 'sensitive tint References
plate' and rotation of the microscope stage will Arboleya , M.L. & Engelder, T. (1995) Concentrated
reveal how well aligned (crystallographically) slip zones with subsidiary shears: their development
the quartz grains are. Highly deformed on three scales in the Cerro Brass fault zone,
Appalachian valley and ridge. Journal of Structu ral
mylonitic rocks have a strong CPO such that in
Geology, 17, 519-532.
one position the bulk of grains in the field of Barker, A.j. (1994) Interpretation of porphyroblast
view should appear blue, whereas when rotated inclusion trails: limitations imposed by growth
through 90 they will be largely yellow-orange
0 kinetics and strain rates. Journal of Metamorphic
Geology, 12,681-694.
(Plates 7(a) & (b)).
Bell, T.H. (1985) Deformation partitioning and
Following early work by Nicholas et al. porphyroblast rotation in metamorphic rocks: a
(1971) on peridotites, many researchers have radical reinterpretation. Journal of Metamorphic
analysed CPOs when studying the kinematics Geology, 3, 109-118.
Bell, T.H. & Johnson, S.E. (1992) Shear sense: a new
of ductile deformation. The asymmetric nature
approach that resolves conflicts between criteria in
of quartz c-axes with respect to the principal metamorphic rocks. Journal of Metam orphic
foliation(s) of shear zones has been extensively Geology, 10, 99-124.
researched and discussed. It is now standard Berthe, D. Choukroune, P. & Jegouzo, P. (1979)
Orthogneiss, mylonite and non-coaxial deformation
practice to study such fabrics when interpreting of granites: the example of the south Armorican
the shear sense in mylonitic rocks (e.g. Lister & shear-zone. Journal of Structural Geology, 1,
Williams, 1979; Simpson & Schmid, 1983; 31-42.

Shear-sense indicators

Blenkinsop, T.G. & Treloar, P.]. (1995) Geometry, development of shear bands in rocks. Geologie en
classification and kinematics of S-C fabrics. journal Mijnbouw, 70,203-211.
of Structural Geology, 17, 397-408. Passchier, CW. (1994) Mixing in flow perturbations: a
Chester, F.M., Friedman, M. & Logan, J.M. (1985) model for development of mantled porphyroclasts
Foliated cataclasites. Tectonophysics, 111, in mylonites. journal of Structural Geology, 16,
139-146. 733-736.
De Bresser, J.H.P. (1989) Calcite c-axis textures along Passchier, CW. & Simpson, C (1986) Porphyroclast
the Gavarnie thrust zone, central Pyrenees. Geologie systems as kinematic indicators. journal of Structural
en Mijnbouw, 68, 367-376. Geology, 8, 831-844.
Dennis, A.J. & Secor, D.T. (1987) A model for the Passchier, CW., ten Brink, CE., Bons, P.D. &
development of crenulations in shear zones with Sokoutis, D. (1993) Delta-objects as a gauge
applications from the Southern Appalachians for stress sensitivity of strain rate in mylo-
Piemont. journal of Structural Geology, 9, nites. Earth and Planetary Science Letters, 120,
809-817. 239-245.
Etchecopar, A. & Malavieille, J. (1987) Computer Platt, J.P. (1984) Secondary cleavages in ductile shear
models of pressure shadows: a method for strain zones. journal of Structural Geology, 6, 439-442.
measurement and shear-sense determination. Platt, J.P. & Vissers, R.L.M. (1980) Extensional struc-
journal of Structural Geology, 9, 667-677. tures in anisotropic rocks. journal of Structural
Etchecopar, A. & Vasseur, G. (1987) A 3-D kinematic Geology, 2,387-410.
model of fabric development in polycrystalline Ponce de Leon, M.1. & Choukroune, P. (1980) Shear
aggregates: comparisons with experimental and zones in the Iberian Arc. journal of Structural
natural examples. journal of Structural Geology, 9, Geology, 2, 63-68.
705-717. Ramsay, ].G. & Huber, M.1. (1983) The techniques of
Johnson, S.E. (1993) Testing models for the develop- modern structural geology; Volume 1: Strain analy-
ment of spiral-shaped inclusion trails in garnet sis. Academic Press, London.
porphyroblasts: to rotate or not to rotate, that is the Schmid, S.M., Panozzo, R. & Bauer, S. (1987) Simple
question. journal of Metamorphic Geology, 11, shear experiments on calcite rocks: rheology and
635-659. microfabric. journal of Structural Geology, 9,
Lister, G.S. & Snoke, A.W. (1984) S-C mylonites. 747-778.
journal of Structural Geology, 6, 617-638. Shelley, D. (1995) Asymmetric shape preferred orienta-
Lister, G.S. & Williams, P.F. (1979) Fabric develop- tions as shear-sense indicators. journal of Structural
ment in shear zones: theoretical controls and Geology, 17, 509-517.
observed phenomena. journal of Structural Sibson, R.H. (1990) Faulting and fluid flow, in Fluids
Geology, 1, 283-297. in tectonically active regimes (ed. B.E. Nesbitt).
Logan, J.M., Friedman, M., Higgs, N., Dengo, C & Mineralogical Association of Canada, Short Course
Shimanto, T. (1979) Experimental studies of simu- No. 18,93-132.
lated gouge and their application to studies of Simpson, C & Schmid, S.M. (1983) An evaluation of
natural fault zones, in Analysis of actual fault zones criteria to deduce the sense of movement in sheared
in bedrock. u.S. Geological Survey, Open-file rocks. Bulletin of the Geological Society of America,
Report, 79- 1239, 305-343. 94, 1281-1288.
Mawer, CK. (1987) Shear criteria in the Grenville van den Driessche, J. & Brun, J.P. (1987) Rolling struc-
Province, Ontario, Canada. journal of Structural tures at large shear strain. journal of Structural
Geology, 9, 531-539. Geology, 9, 691-704.
Nicholas, A., Bouchez, ].L., Boudier, F. & Mercier, Wenk, H.-R., Takeshita, T., Bechler, E., Erskine, B.G.
J.C (1971) Textures, structures and fabrics due to & Matthies, S. (1987) Pure shear and simple shear
solid state flow in some European lherzolites. calcite textures. Comparison of experimental, theo-
Tectonophysics, 12,55-85. retical and natural data. journal of Structural
Passchier, CW. (1991) Geometric constraints on the Geology, 9, 731-745.

Chapter eleven

Veins and fluid


11.1 Controls on fluid migration and most veined rocks representing highest fluid
veining flux), or in terms of permeability differences. In
cases with little veining the permeability of the
Most prograde metamorphic reactions involve
rocks may have been sufficient to permit fluid
devolatilisation. The main volatile components
flow through the rock in a pervasive fashion,
produced are H 2 0 and CO 2 , although the rela-
maintaining P f below the value for fracturing.
tive proportions of these varies considerably
However, in intensely veined examples it is
from one rock to the next. Once produced,
likely that the permeability of the unfractured
volatiles tend to migrate upwards through the
rock was insufficient to allow fluid escape at
crust, the mechanism and rate of flow being
the rate at which it was being produced by
largely a function of temperature, fluid pres-
devolatilisation reactions. This would give rise
sure (Pf) gradients, and bulk rock permeability
to an increase in Pf until ultimately the rock
(e.g. Fyfe et al., 1978; Ferry, 1980; Etheridge et
strength would be exceeded and a series of
al., 1983; Wood & Walther, 1986; Yardley,
discrete fractures would develop to increase the
1986; Thompson, 1987).
rock's bulk permeability. Healing of fractures
Fluids become focused into zones of high
by precipitation of minerals out of solution
permeability and lower Pf' Penetratively foli-
competes with hydraulic fracturing, but
ated schists, fault zones and thrust zones are
progressive upward migration of fluids main-
particularly favoured, whereas massive
tains high P f and favours further micro crack
quartzites and marble (unless highly fractured)
propagation. Because Si02 solubility in aque-
have low permeabilities and experience much
ous solutions diminishes with decreasing P and
lower fluid flux. For a given rock, the degree of
T, Si0 2 saturated fluids will precipitate quartz
permeability relates to the effectiveness of fluid
as they ascend through the crust.
migration via grain boundaries, and discrete
fractures or cleavage. On the local scale,
segregations of new mineral growth are 11.2 Initial description and interpretation
commonly observed in strain shadows of of veins
porphyroblasts (e.g. Figs 10.20-10.22) and in
boudin necks on both the meso- and the When considering the characteristics of veins
micro scale (Fig. 11.1). and the nature of their formation, field studies
While veining is extensive in many metamor- provide key information concerning the
phic rocks and terrains, in other cases it is mineral assemblage, orientation and timing of
extremely scarce. This can be interpreted either each vein-set, both in relation to other veins
in terms of differences in bulk fluid flow (the and to specific deformation and metamorphic

Veins and fluid inclusions



FIG. 11.1 (a) Mesoscale vein quartz segregation in a boudin neck: a boudinaged psammite layer in an interbed-
ded psammite-pelite sequence. Snake Creek, Queensland, Australia. The lens cap is 50 mm in diameter. (b) A
microscale quartz segregation in a neck of fractured hornblende porphyroblast: Hbl-mica schist
('Garbenschiefer') from Troms, Norway. Scale bar = 1 mm (XPL).

events. The use of veins for shear-sense deter- identified on all scales. At moderate and high-
minations has already been covered in Section temperature metamorphism, most veins are
10.2. strongly recrystallised polygonal aggregates,
Since the majority of veins and segregations but at sub-greenschist facies, fibrous veins
have sharply defined margins, and are of differ- of quartz and calcite are common. In rela-
ent colour to the host rock, they are easily tion to certain high-temperature metamorphic

Initial description and interpretation of veins

processes such as charnockitic alteration of ation of one or more mineral phases at the vein
granulites (e.g. Newton, 1992) and develop- margin, whereas the central portion of the vein
ment of 'patch' and 'diktyonitic' migmatites often consists of one phase only (e.g. quartz).
(e.g. McLellan, 1988), diffuse veins and segre- In Fig. 11.3 actinolite is developed at the
gations may be encountered, with less well margin of a quartz vein passing through a
defined boundaries. metabasite. Although best observed in thin
Monomineralic veins such as quartz, calcite section, such features may be observable in the
and epidote are especially common, but veins field in coarse-grained veins.
that are bi-mineralic (e.g. qtz-cal and qtz-chl Although many features can be identified in
veins) or polymineralic (e.g. mineralised veins) the field, for a more complete understanding of
are also seen in metamorphic rocks. Some basic veins and mineral segregations, thin-section
observations about the degree of interaction studies are required. Depending on the mineral
with the wall-rock are important. For example, assemblage, it may be possible to gain a general
many quartz veins draw silica from the adja- understanding of the P-T conditions at the
cent wall-rock to leave darkened, quartz- time of vein formation. Clearly, quartz and
depleted margins. In slates and schists this is calcite are of little use here, since they are
often quite obvious in the field (Fig. 11.2). In stable over almost the full range of metamor-
other cases fluids infiltrating the wall-rock phic conditions. However, veins containing
adjacent to veins may cause metasomatic alter- minerals with more restricted stability fields
ation and new mineral growth in the wall-rock. (e.g. jadeite, andalusite and prehnite) can
The interaction of wall-rock and fluid passing provide important information. In recent
through a fracture commonly leads to nucle- decades, the use of cathodoluminescence

FIG. 11.2 Multiple quartz veins intruding slate. Bude, England. Scale = 5 cm. Note the darkened area of silica
depletion immediately adjacent to the main vein. Also note the fact that this 3 cm wide vein initiated as a zone of
weakness defined by an en echelon array of thin tension gashes. These have subsequently experienced a compo-
nent of right-lateral (dextral) shear, to give the sigmoidal traces now observed.

Veins and fluid inclusions

FIG. 11.3 Actinolite (arrowed) developed at the margin of a quartz vein (right) passing through a greenschist
facies metabasite (left). Pyrite Belt, Spain. Scale = 0.1 mm (XPL).

(Marshall, 1988) has proved highly revealing in the precipitating phase, and minerals growing
the study of vein mineralogy and genesis (Plates from the walls commonly develop large crystals
7(~) & (d)). In addition, it has been usefully with perfect form. In situations in which
employed for the recogmtlon of sealed increased fluid pressure (often due to local
microfractures through crystals, which are reaction) exceeds rock strength, hydraulic frac-
either unrecognisable or else poorly defined by turing will occur. Once such fracturing has
standard microscopy. Since the common initiated, fractures will propagate until such a
carbonate minerals luminesce very differently, time as fluid pressure falls below rock strength.
it has proved particularly useful for their study The propagating tip is characteristically
(Plates 7(c) & (d)), as well as for the study of tapered or splayed. At deeper levels in the crust
overgrowth textures on quartz. it is a widely held view that P f approximates to
With regard to the mechanism of vein devel- PI and open fractures only develop transiently,
opment, it is necessary to consider (a) whether and do not usually exist for any significant
a fracture already existed and fluid passed length of time. However, Ague (1995) presents
through and precipitated out of solution, or (b) convincing evidence for Ky + Grt + Qtz veins,
whether increased fluid pressure induced where the well-formed nature of the minerals
mechanical failure of the rock and promoted suggests growth into large aperture fractures at
fracture propagation, into which vein material P-T conditions corresponding to c. 30 km
precipitated. Open fissures can be maintained depth within continental crust. Analogue
in the upper 5-10 km of crust, since fluids modelling of single-stage fracture opening
passing through these fractures are generally at events and subsequent crystallisation and vein
pressures close to hydrostatic pressure and infill have been undertaken by Wilson (1994)
considerably less than lithostatic pressure. Such using ice and an aqueous solution. The
fractures are not always completely sealed by microstructures recorded have been used to

'crack-seal' mechanism of vein formation

evaluate the likely processes operating during M

single-stage fracture-opening and vein develop- (a)
ment in metamorphic rocks. Wilson (1994)
concludes that the presence of a deviatoric
stress plays a crucial role, influencing both
crystallisation and the resultant microstructure
of veins. Static growth in a fluid-filled cavity in
equilibrium with the local stress produces euhe-
dral prismatic crystals (e.g. 'drusy' or 'comb'
quartz cavity infill) by a mechanism of free-face
growth. In contrast, syntectonic growth from a
fluid-filled cavity in the presence of a deviatoric
stress produces veins comprising anhedral poly-
gonised aggregates, as a result of contact
growth and recrystallisation. Where vein
formation occurs in a solid grain aggregate in FIG. 11.4 Successive stages in the development of a
the presence of a deviatoric stress, Wilson vein by the crack-seal process (see text for details). M,
median line; I, inclusion bands.
(1994) shows that fibrous veins will develop by
incremental microcracking and sealing (the
'crack -seal mechanism').
open by 5-100 11m at a time. Precipitation of
solute species from the fluid leads to fracture
11.3 The 'crack-seal' mechanism of vein infill, until eventually the fracture is more or
formation less completely sealed. Once sealed, the perme-
The 'crack-seal' mechanism (Ramsay, 1980; ability of the rock becomes reduced once again,
and see Fig. 11.4) is an important process and assuming continual replenishment of fluid,
responsible for the formation of many veins Pf will rise once more until the critical value at
under low-grade conditions (e.g. greenschist which the rock fails again and further cracking
and sub-greenschist facies). It is often difficult is induced. The initial vein formation strongly
to assess the mechanism responsible for vein influences the location of subsequent veins due
formation under medium- and high-grade to the planar anisotropy set-up. Subsequent
conditions, since the veins are usually strongly fractures will commonly split earlier veins in
recrystallised (Fig. 11.5). The 'crack-seal' mech- half or form along one of the pre-existing
anism involves multistage crack opening and vein-wall contacts. The whole process of
vein infill in response to fluctuations in Pf. By 'crack-seal' and vein-widening continues by a
successive increments it can lead to veins of number of cycles such as described above. It is
considerable width (typically mm-cm-scale; the regular growth increments shown by
Plates 8(a)-(c)). Due to build-up of elastic 'crack-seal' veins that gives strong evidence for
strains in the rock in response to rising Pf' often hydraulic fracturing related to changing Pf as
induced by poor or limited rock permeability the vein-forming process. Characteristically,
(see above), a fracture is formed. This will prop- such veins record successive increments of
agate until such a time as Pf diminishes to a growth by thin 'inclusion bands' parallel to the
value less than rock strength. Judging by incre- vein walls (Plate 8(c)) and which generally
mental growth stages observed in veins (Plates represent detached screens of wall rock now
8(a)-(c)) it appears that such fractures typically incorporated in the vein. Even within veins that

Veins and fluid inclusions

FIG. 11.5 A recrystallised quartz vein let cutting amphibolite. Troms, Norway. Scale =0.5 mm (XPL).

FIG. 11.7 A schematic illustration of the sawtooth

form of the contacts between individual crystals in a
fibrous vein.

FIG. 11.6 A schematic illustration showing inclusion

trails developed perpendicular to the walls of a fibrous
veIn. rate optically parallel small inclusions of the
same mineral, and are usually entirely enclosed
within coarse fibrous crystals of the main vein
have recrystallised at higher grades of meta- phase or phases (typically quartz or calcite).
morphism, it may be possible to recognise these The contact between individual fibres is sub-
bands (Plate 8(d)) and thus understand how the perpendicular to the vein-wall, and commonly
vein formed. In addition to 'inclusion bands' has a stepped saw-toothed form (Fig. 11.7).
approximately parallel to vein walls, fibrous Planes of trapped fluid inclusions sub-parallel
veins may also show inclusions arranged in to the vein wall, and representing sealed
trails sub-perpendicular to the wall (Fig. 11.6). surfaces, are an additional feature of such
These inclusion trails normally comprise sepa- vems.

Interpretation of fibrous veins

11.4 Interpretation of fibrous veins of anchimetamorphic slate (Fig. 10.22) with

polygonal quartz pressure shadows in garnet-
Four main types of fibrous vein system can be
mica schist (Plate 5(c)). The various types of
distinguished (Fig. 11.8), and are common in
fibrous veins are now described and discussed.
rocks undergoing deformation at sub-green-
schist to mid-greenschist facies conditions. At
higher temperatures the fibrous form of the 11.4.1 Syntaxial fibre veins
crystal is thermodynamically unstable, and is These veins involve progressive fibre growth
superseded by equidimensional polygonal crys- from each wall of the fracture towards the
tals; compare fibrous quartz pressure shadows centre. The fibres thus fall into two halves
either side of a centrally located suture (Fig.
l1.8(a)). The crystals forming the fibres show a
A Syntaxlaf 8 Antltaxlaf close compositional link with the rock through
which the vein cuts (e.g. calcite veins in marble
or quartz veins in sandstone). In the case of
curved syntaxial fibres (Fig. l1.8(a)), the fibres
at the vein-wall contact are always perpendicu-
lar to the wall, whereas in the next type of
veins to be considered (antitaxial fibre veins)
the contact is oblique. Curved fibres reflect
changes in orientation of the line of maximum
incremental longitudinal strain during the
deformation history.

I 1

11.4.2 Antitaxial fibre veins

Antitaxial fibre veins involve growth towards
C Composite o 'Stretched' crystals the wall-rock rather than from it. The fibres
show crystallographic continuity from one wall
to the other and the mineral phase(s) involved
are often unrelated to the wall-rock mineral-
ogy, but are derived from some local source.
The centre of the vein is usually marked by a
screen of small wall-rock fragments to define
what is termed the 'median line'. Although
changes in orientation of the line of maximum
incremental longitudinal strain may give rise to
fA, nil.
curved antitaxial fibres, at the 'median line'
such fibres are always perpendicular to the
~, rock walls (Fig. l1.8(b)), since this coincides
'Zd with initial fracture opening. Some antitaxial
veins show several or many 'inclusion bands'
FIG. 11.8 (a) A syntaxial fibre vein. (b) An antitaxial parallel to the median line. Such trails indicate
fibre vein. (c) A composite fibre vein. (d) A stretched
successive vein openings by the 'crack-seal'
crystal fibre vein. (After Fig. 13.9 of Ramsay & Huber,
1983; courtesy of Academic Press.) m.l. is median line. mechanism (see above).

Veins and fluid inclusions

11.4.3 Composite fibre veins wall-rock (on to which they root) . In examples
from the Helvetic Alps, Ramsay & Huber
Such veins comprise a central zone (with median (1983) observe that the central portion of
line) of one mineral phase, bounded on each side composite veins has geometric characteristics
by marginal zones consisting of another crystal identical to that of an antitaxial vein, while the
species. The crystal species of the marginal zones marginal zones have characteristics of syntaxial
have compositional similarity to crystals of the veins (Fig. 11.8(c)).

FIG. 11.9 A stretched quartz fibre vein. Croyde Bay, Devon, England. Scale =2 cm.

FIG. 11.10 Detail of the serrated contacts between individual fibres of a stretched crystal fibre vein. Vein quartz,
Croyde Bay, Devon, England. Scale = 0.1 mm (XPL).

Fluid inclusions

11.4.4 'Stretched' (or 'ataxial') crystal fibre and Fe from the vein area (e.g. Kretz, 1994). If
veins the vein is forming by local differentiation, this
process will lead to depletion of biotite in the
This fourth category of fibrous veins involves vein and a build-up of biotite in a selvage
crystal fibres of the same or similar composi- around the vein or segregation. Similar diffu-
tion to those minerals found in the wall-rock. sive processes in amphibolite may produce
The fibres are typically perpendicular or at a irregular veins or segregations of PI + Qtz and
high angle to the vein margins (Fig. 11.9), and an area of surrounding host rock enriched in
always show crystallographic continuity from hornblende due to this localised subsolidus
one wall to the other (Fig. 11. 8 (d) ). In the case metamorphic differentiation, perhaps induced
of quartz, thin-section studies usually reveal by a localised pressure gradient (Kretz, 1994).
that contacts between adjacent fibres are During the development of stromatic
serrated (Fig. 11.10). In detail, the serrations migmatites by partial melting of specific layers
show some regularity suggestive of a crack-seal in a metasedimentary sequence, the biotite of
mechanism for vein formation. the protolith may do one of several things. It
The orientation of fibres in 'stretched' crys- may break down to form minerals such as K-
tal fibre veins is commonly used as an indica- feldspar, it may react out with plagioclase to
tion of the vein opening direction, but this is produce hornblende, or it may crystallise from
not strictly true, because fibrous crystals have a the partial melt into coarse biotite segregations
tendency to grow normal to the fracture within the leucosome, but more commonly as
surface (i.e. face-controlled growth rather than concentrated melanosome at the margins of the
displacement-controlled growth). Urai et at. leucosome (Johannes, 1983). Such biotite
(1991) suggest that displacement-controlled selvages a few millimetres thick, separating
fibres only develop under conditions with irreg- leucosome from mesosome (Fig. 4.12), are a
ular growth surfaces, small growth increments characteristic feature of stromatic migmatites
and isotropic crystal growth. (e.g. Mehnert, 1968; Johannes, 1983, 1988).
Unlike the subsolidus metamorphic differentia-
tion discussed above in relation to certain veins
11.5 Veins and melt segregations at high
and segregations of gneisses, the biotite
metamorphic grades
selvages at the margins of leucosome are
In migmatites, and upper amphibolite facies commonly interpreted to form by anatectic
schists and gneisses, quartzofeldspathic and differentiation (segregation from a melt) in the
granitoid veins are common. Some represent presence of H 2 0 (e.g. Johannes, 1988).
melt segregations, whereas others are veins and However, an alternative explanation presented
segregations that have precipitated from an by Maal0e (1992) is that contraction of melt-
aqueous fluid. In many cases, the veins and ing mesosome causes concentration of refrac-
segregations are demonstrably of local origin tory minerals such as biotite and hornblende in
(e.g. Vidale, 1974), and have developed due to a selvage at the margins of the leucosome.
metamorphic differentiation, involving diffu-
sion of elements in different directions along
concentration gradients. Commonly, it is found
11.6 Fluid inclusions
that veins are richer in K-feldspar and poorer Minute fluid inclusions are trapped in various
in biotite relative to surrounding gneisses, minerals from a range of metamorphic environ-
suggesting an exchange process involving trans- ments. The inclusions represent small quanti-
fer of K into the vein area, and removal of Mg ties of liquid or vapour trapped inside

Veins and fluid inclusions

individual crystals or healed microcracks at conditions. Second - and very i~portantly - it is

some time during the rock's evolution. generally considered that quartz shows no
Although the importance and usefulness of significant reaction (exchange of ions) with the
fluid inclusions was realised by Sorby (1858), trapped fluid, whereas other minerals such as
the general appreciation of such studies was calcite may well exchange cations with solute
slow to develop. However, in recent decades species.
fluid inclusion studies have received much In terms of their genesis there are two funda-
greater attention, and have been realised as a mental types of fluid inclusions, namely,
valuable tool for gaining insight into P- T PRIMARY and SECONDARY. Primary (P)
conditions and changing fluid chemistry during inclusions become trapped within minerals as
a rock's metamorphic history. For details of they crystallise within a fluid-rich environment.
methods and applications relating to fluid Secondary (S) inclusions represent small pockets
inclusion studies the publications of Roedder of fluids that have become trapped along post-
(1984), Shepherd et al. (1985) and De Vivo & crystallisation microcracks as they have sealed.
Frezzotti (1994) provide a comprehensive Planar surfaces of secondary inclusions are
insight. characteristic of healed cracks and are seen as
It is rare to observe fluid inclusions greater linear inclusion trails in thin section (Figs
than about 150 pm (0.15 mm) in diameter, and l1.11(a) & (b)). A third genetic class of fluid
more commonly they are less than 30 pm (0.03 inclusions, termed pseudo-secondary (PS) inclu-
mm) in size, with the majority being less than sions, develop in a similar way to S-inclusions,
15 pm. As a general approximation, the larger but differ by the fact that the fracture heals
the host crystal is, the larger is the inclusion before crystal growth is complete.
trapped. For this reason, vein minerals have During prograde metamorphism, general
received the greatest attention, and have recrystallisation and grain coarsening leads to
proved most fruitful for fluid inclusion studies. crystal lattice modifications, and the destruc-
This is also due to the fact that vein minerals tion of fluid inclusions trapped at lower
have formed in a very fluid-rich environment, temperatures. Overpressuring of early inclu-
and are thus more likely to trap more fluid. sions due to the thermal expansion of the
Useful studies can be made by standard trans- fluid being greater than the surrounding solid
mitted light microscopy, using the high-power commonly leads to 'decrepitation' (sudden
objective of the microscope and by adjusting the and abrupt leakage). In natural examples of
positions of the substage condenser and aperture decrepitated inclusions, the inclusions often
diaphragm to obtain optimum illumination. It is appear dark, may have several sharply
advisable to use lower magnifications initially in pointed corners, or exhibit a halo of daughter
order to locate inclusions, and to examine their inclusions. Experimental work with synthetic
distribution and morphological characteristics fluid inclusions (e.g. Sterner & Bodnar, 1989)
(Figs l1.11(a) & (b)). Fluid inclusions are has shown that identical features form by
reported in various minerals, but those in implosion, when a fluid inclusion is substan-
quartz, fluorite and calcite have been studied tially underpressured - that is, when the
most widely. Of these, it is inclusions in quartz confining pressure is substantially increased
which are of most general interest and use to the above internal fluid pressure. In natural
metamorphic petrologist. There are a number of samples of granulites, such decrepitated
reasons why this should be the case. First, quartz inclusions with a halo of microscopic satellite
is a common mineral, which often forms veins, inclusions are commonly encountered, and
and is stable over the full range of metamorphic are interpreted in terms of the rock experi-

Fluid inclusions

FIG. 11.11 (a) A typical example of a linear array of secondary fluid inclusions along healed microcracks in vein
quartz from Troms, Norway. Scale = 0.5 mm (PPL). (b) Detail of the inclusions shown in (a) . Scale = 100 ftm

. . .
encmg mcreasmg pressure accompanymg encountered. It IS fluid inclusions formed
cooling. synchronous with, or after, peak metamorphic
Because of the problems of recrystallisation temperatures that are most frequently preserved
and decrepitation associated with prograde meta- in veins and sealed microcracks. In consequence,
morphism, genuine primary inclusions related to fluid inclusion studies are most useful for assess-
the early history of the rock are infrequently ing conditions during the uplift trajectory of a

Veins and fluid inclusions

given rock. Careful petrographic study may

allow various generations of fluid inclusions to
be recognised. Non-planar clusters are most
1 o
likely to be the earliest inclusions, while linear
trails representing distinct inclusion planes relate
to later healed microcracks. In practice, it is
never easy to make definite identification of
primary inclusions; nor is it always easy to deter-
mine the relative ages of secondary inclusion
trails by petrographic observation alone.
Fluid inclusions show a variety of shapes,
which in part often relate to the structure of
the host mineral. Inclusions mirroring the
symmetry of the host and forming perfect
'negative crystals' are common in some miner-
als. Cubic inclusions in halite (Fig. 11.12( a))
are a good example of this. Inclusions in quartz
may also have a form relating to the symmetry
of the host (Fig. 11.12(b)), but invariably are o
not perfect, and have an irregular, or rounded
to oval shape (Fig. 11.12(c)). It is possible to
construct various elaborate schemes to classify
fluid inclusions in terms of the differing
proportions and types of solids, liquids and
vapour observed in inclusions at room temper-
ature. However, the vast majority of fluid
inclusions observed in metamorphic rocks fall
into one of five categories (Fig. 11.13):

(a) Monophase, liquid (usually aqueous) (Figs ..

11.13(a) & (f)).
FIG. 11.12 (a) A schematic example of fluid inclu-
(b) Two-phase, liquid-rich (i.e. liquid + vapour, sions in halite, forming perfect negative crystals. (b)
where L > V) (Figs 11.13(b) & (g)). Well-formed fluid inclusions in quartz, their shape
(c) Vapour-rich (vapour dominates, but in being controlled by the crystal structure of the host.
many cases a thin rim of liquid may still be Scale = 100 !1m (PPL). (c) An array of rounded to oval
fluid inclusions in quartz. Scale = 50 !1m (XPL).
present) (Figs 11.13(c) & (h)).
(d) CO 2-bearing (Figs 11.13(d) & (i)) . Such
inclusions usually consist of CO 2 + H 20 as (e) Solid-phase bearing - containing one or
immiscible liquids. If required, they can be more solid phases; cubes of NaCI or KCI being
subdivided according to which liquid domi- most common and having precipitated from
nates. At room temperature (e.g. Fig. 11.13(i)), the trapped fluid ('daughter minerals') (Figs
the inclusion often appears to be three-phase 11.13(e) & (j); Plate 8(f)). It is also possible
(COZ(Y) + CO 2 (l ) + H 2 0). The CO2-rich phase that solid phases suspended in the fluid (e.g.
may contain significant proportions of CH4 and mica) may become trapped in the inclusion
N z (Shepherd et ai., 1985). ('captive minerals') .

Fluid inclusions

FIG. 11.13 Schematic and natural examples of the five main types of fluid inclusions: (a, f) monophase; (b, g)
two-phase, liquid-rich; (c, h) two-phase, vapour-rich; (d, i) CO 2-bearing; (e, i) solid phase bearing. The natural
example shown in (i) is an aqueous fluid with 'daughter minerals' of halite (cubes). Scale bars are 20 "m.

FIG. 11.14 The necking-down of a fluid inclusion. Scale = 50 Jim (PPL).

Variable phase ratios in fluid inclusions can be physical flow characteristics of two-phase fluids
used as evidence for immiscible (phase-separated) are quite different from those of single fluids.
fluids at the time of trapping. Immiscible CO 2- When one fluid is significantly in the minority, it
H 2 0 and CH 4-H 2 0 fluids have been reported will either occur as isolated globules ('non-
from many metamorphic terrains (e.g. Yardley & wetting phase') suspended in the majority phase
Bottrell, 1988), and are significant in that the or, alternatively, will adhere to grain surfaces

Veins and fluid inclusions

, "
.' •



FIG. 11.15 Micro-inclusions ('bubbles') along grain boundaries (vein quartz, Trams, Norway): (a) scale bar = 50
flm, partial XPL; (b) scale bar = 125 flm, partial XPL.

('wetting phase'). The separation into two Having decided whether an inclusion or set
distinct phases generally leads to one phase flow- of inclusions is primary or secondary, and
ing more rapidly than the other and may also having classified in terms of the relative
lead to a reduction in effective permeability. proportions of solid, liquid and vapour, it is
These effects can have a profound influence on also necessary to determine whether or not the
reaction pathways and the progress of metamor- inclusion has leaked. The tell-tale signs of neck-
phism, especially in the case of carbonate rocks. ing-down and leakage are variable phase ratios

Fluid inclusions

in a group of inclusions, and inclusions

connected by a thin tube or with protruding
tails (Fig. 11.14). However, recent experimental
and TEM studies (e.g. Bakker & Jansen, 1994)
have established that even inclusions that show
none of the tell-tale signs of leakage when
examined during transmitted light microscopy
may in fact have experienced dislocation-
controlled diffusional leakage. This can have
important consequences when interpreting
fluid chemistry and trapping conditions,
because phase ratios can change during leak-
age. For example, Crawford & Hollister
(1986), and Bakker & Jansen (1991, 1994)
have shown that due to differences in wetting
characteristics, HzO preferentially leaks out of
HzO-CO z inclusions. Fluid inclusions are
particularly prone to leakage or else decrepita-
tion if they experience an underpressure or
overpressure in excess of 1 kbar relative to the
conditions under which they were trapped
(Bodnar et aI., 1989; Sterner & Bodnar, 1989), FIG.11.16 A TEM (dark field) image showing a sub-
or if there is extensive recrystallisation of the grain (in quartz) bordered by low-angle tilt boundaries.
mineral aggregate in which the fluid inclusions Adjacent sub-grains consist of two or more sets of
are contained. The effects of recovery and dislocations. Note the extrinsic dislocation loop
emanating from the fluid inclusion (FI) bubble project-
recrystallisation on fluid inclusions were ing into a sub-grain interior (top left, arrowed) (unpub-
considered in Section 8.2, but one of the most lished photograph courtesy of L. Hopkinson).
distinctive features is to see residual very fine
« 10 /lm) fluid inclusions decorating grain
boundaries (Figs 11.1S(a) & (b)), the fluid further, and obtain quantitative information
having been swept to these areas facilitated by about aspects such as the salinity of aqueous
dislocation climb and other intracrystalline fluids and the homogenisation temperatures of
processes (Kerrich, 1976; Wilkins & Barkas, two-phase inclusions. In metamorphic rocks
1978; O'Hara & Haak, 1992). From TEM salinities commonly range from 0 wt% NaCI
studies it has been confirmed that dislocations equivalent up to 30-40 wt% NaCl equivalent.
are often decorated with minute nanometre- Inclusions with >26 wt% NaCI characteristi-
scale fluid inclusions, termed 'bubbles', and cally show a daughter mineral (halite cube) at
that such dislocations lined with an array of room temperature (e.g. Fig. 11.13(j)). It is not
fluid bubbles commonly emanate from larger really possible to say too much about the envi-
fluid inclusions (Fig. 11.16), thus suggesting ronment or conditions of metamorphism
leakage (e.g. Bakker & Jansen, 1991, 1994; simply on the basis of fluid chemistry.
Reeder, 1992; Wang et aI., 1993 ). However, it is a common feature to find that
Having identified the different sets of inclu- peak metamorphic amphibolite facies fluid
sions present in a rock by optical studies, it is inclusions have a low-salinity HzO-CO z
possible by using a heating-freezing stage to go composition, while many greenschist facies

Veins and fluid inclusions

fluids associated with retrogression of minerals: methods and applications. Short Course
metapelites and metavolcanics have a highly of the Working Group (IMA) 'Inclusions in
Minerals' (Pontignano - Siena, 1-4 September
saline aqueous character, with one or more 1994) (eds B. De Vivo & M.L. Frezzotti). Virginia
daughter minerals (e.g. Bennett & Barker, Tech, USA, 25-44.
1992). Under granulite facies conditions, one Crawford, M.L. & Hollister, L.S. (1986) Metamorphic
of the most characteristic features is to find fluids: the evidence from fluid inclusions, in
Fluid-rock interactions during metamorphism (eds
CO2-rich fluid inclusions (e.g. Touret, 1971, J.V. Walther & B.]. Wood). Advances in Physical
1977). This agrees well with the theoretical Geochemistry, 5, 1-35. Springer-Verlag, New York.
considerations that suggests granulite facies De Vivo, B. & Frezzotti, M.L. (eds) (1994) Fluid inclu-
assemblages require low aH20to exist. sions in minerals: methods and applications. Short
Course of the working group (IMA) 'Inclusions in
More recently, the use of laser Raman spec- Minerals' (Pontignano - Siena, 1-4 September
troscopy (Burke, 1994) has enabled micro- 1994). Virginia Tech, USA, 376 pp.
analysis of daughter minerals and certain Etheridge, M.A., Wall, V.]. & Vernon, R.H. (1983)
species within individual fluid inclusions. This The role of the fluid phase during regional meta-
morphism and deformation. Journal of
technique involves laser beam excitation of Metamorphic Geology, 1, 205-226.
the various molecules and induces stretching Ferry, ].M. (1980) A case study of the amount and
and vibration of bonds between atoms, to distribution of heat and fluid during metamorphism.
produce Raman radiation. The Raman spectra Contributions to Mineralogy and Petrology, 71,
detected have enabled identification and Fyfe, W.S., Price, N.J. & Thompson, A.B. (1978)
quantification of various poly atomic species Fluids in the Earth's crust. Elsevier, Amsterdam,
(e.g. S04,CH4,C02'N2 ,NaCI, H 2 0, and so on) 383 pp.
present in the inclusions (either as solid, Johannes, W. (1983) On the origin of layered
migmatites, in Migmatites, melting and metamor-
vapour or fluid). phism (eds M.P. Atherton & C.D. Gribble). Shiva,
Nantwich, 234-248.
References Johannes, W. (1988) What controls partial melting in
migmatites? Journal of Metamorphic Geology, 6,
Ague, ].J. (1995) Deep crustal growth of quartz, kyan- 451-465.
ite and garnet into large-aperture, fluid-filled frac- Kerrich, R. (1976) Some effects of tectonic crystallisa-
tures, north-eastern Connecticut, USA. Journal of tion on fluid inclusions in vein quartz.
Metamorphic Geology, 13,299-314. Contributions to Mineralogy and Petrology, 59,
Bakker, R.]. & Jansen, J.B.H. (1991) Experimental 195-202.
post-entrapment water loss from synthetic Kretz, R. (1994) Metamorphic crystallization. John
CO 2-H2 0 inclusions in natural quartz. Geochimica Wiley, Chichester, 507 pp.
et Cosmochimica Acta, 55,2215-2230. McLellan, E.L. (1988) Migmatite structures in the
Bakker, R.]. & Jansen, J.B. (1994) A mechanism for Central Gneiss Complex, Boca de Quadra, Alaska.
preferential H 2 0 leakage from fluid inclusions in Journal of Metamorphic Geology, 6, 517-542.
quartz, based on TEM observations. Contributions Maal0e, S. (1992) Melting and diffusion processes in
to Mineralogy and Petrology, 116, 7-20. closed-system migmatization. Journal of
Bennett, D.G. & Barker, A.J. (1992) High salinity Metamorphic Geology, 10,503-516.
fluids: the result of retrograde metamorphism in Marshall, D.J. (1988) Cathodoluminescence of geologi-
thrust zones. Geochimica et Cosmochimica Acta, cal materials. Unwin Hyman, London, 172 pp.
56,81-95. Mehnert, K.R. (1968) Migmatites and the origin of
Bodnar, R.J., Binns, P.R. & Hall, D.L. (1989) granitic rocks. Elsevier, Amsterdam.
Synthetic fluid inclusions - VI. Quantitative evalua- Newton, R.C. (1992) Charnockitic alteration: evidence
tion of the decrepitation behaviour of fluid inclu- for CO 2 infiltration in granulite facies metamor-
sions in quartz at one atmosphere confining phism. Journal of Metamorphic Geology, 10,
pressure. Journal of Metamorphic Geology, 7, 383-400.
229-242. O'Hara, K. & Haak, A. (1992) A fluid inclusion study
Burke, E.A.J. (1994) Raman microspectrometry of fluid of fluid pressure and salinity variations in the
inclusions: the daily practice, in Fluid inclusions in footwall of the Rector Branch thrust, North


Carolina, USA. Journal of Structural Geology, 14, Urai, ].L., Williams, P.F. & van Roermund, H.L.M.
579-589. (1991) Kinematics of crystal growth in syntectonic
Ramsay, ].G. (1980) The crack seal mechanism of rock fibrous veins. Journal of Structural Geology, 13,
deformation. Nature, 284, 135-139. 823-836.
Ramsay, ].G. & Huber, M.l. (1983) The techniques of Vidale, R.]. (1974) Vein assemblages and metamor-
modern structural geology; Volume 1: Strain analy- phism in Dutchess County, New York. Bulletin of
sis. Academic Press, London, 307 pp. the Geological Society of America, 85,303-306.
Reeder, R.]. (1992) Carbonates: growth and alteration Wang, ].N., Boland, ].N., Ord, A. & Hobbs, B.E.
microstructures, in Minerals and reactions at the (1993) Microstructural and defect development in
atomic scale: Transmission electron microscopy (ed. heat treated Heavitree Quartzite, in Defects and
P.R. Buseck). Mineralogical Society of America, processes in the solid state: geoscience applications
Reviews in Mineralogy, No. 27, Ch. 10,381-424. (eds ].N. Boland & ].D. FitzGerald). Developments
Roedder, E. (1984) Fluid inclusions. Reviews in in Petrology No. 14 ('The McLaren Volume').
Mineralogy 12, Mineralogical Society of America. Elsevier, Amsterdam, 359-381.
Shepherd, T.]., Rankin, A.H. & Alderton, D.H.M. Wilkins, R.W.T. & Barkas, ].P. (1978) Fluid inclu-
(1985) A practical guide to fluid inclusion studies. sions, deformation and recrystallization in granite
Blackie, Glasgow, 239 pp. tectonites. Contributions to Mineralogy and
Sorby, H.C. (1858) On the microscopical structure of Petrology, 65,293-299.
crystals indicating the origin of rocks and minerals. Wilson, c.].L. (1994) Crystal growth during a single-
Quarterly Journal of the Geological Society, 14, stage opening event and its implications for syntec-
453-500. tonic veins. Journal of Structural Geology, 16,
Sterner, S.M. & Bodnar, R.J. (1989) Synthetic fluid 1283-1296.
inclusions - VII. Re-equilibration of fluid inclu- Wood, B.]. & Walther, ].V. (1986) Fluid flow during
sions in quartz during laboratory simulated metamorphism and its implications for fluid-rock
metamorphic burial and uplift. Journal of ratios, in Fluid-rock interactions during metamor-
Metamorphic Geology, 7, 243-260. phism (eds ].V. Walther & B.]. Wood). Advances in
Thompson, A.B. (1987) Some aspects of fluid motion Physical Geochemistry, 5, 89-108. Springer-Verlag,
during metamorphism. Journal of the Geological New York.
Society, 144, 309-312. Yardley, B.W.D. (1986) Fluid migration and veining in
Touret, ]. (1971) Les facies granulite en Norvege the Connemara Schists, Ireland, in Fluid-rock inter-
meridionale II: les inclusions fluides. Lithos, 4, actions during metamorphism (eds ].V. Walther &
423-436. B.]. Wood). Advances in Physical Geochemistry,S,
Touret,]. (1977) The significance of fluid inclusions in 109-131. Springer-Verlag, New York.
metamorphic rocks, in Thermodynamics in Yardley, B.W.D. & Bottrell, S.H. (1988) Immiscible
Geology (ed. D.G. Fraser). D. Reidel, Dordrecht, fluids in metamorphism: implications of two-phase
203-227. flow for reaction history. Geology, 16, 199-202.

Chapter twelve

and polymeta-
morphosed rocks

12.1 Polymetamorphism p- T conditions, are of little use in establishing

much about the rock's earlier history. However,
The majority of metamorphic rocks show in other cases, minerals such as chloritoid,
evidence of more than one phase of deforma- staurolite and glaucophane may be included.
tion, and often multiple stages of metamor- Because of their more restricted, and character-
phism. By careful study of the mineralogical istic, stability fields, they provide a useful indi-
and textural features of a rock, it should be cation of previous P-T conditions experienced
possible to build up a detailed understanding of by the rock. In certain cases it may be possible
the interrelationships between the various to establish with some confidence the nature of
mineral phases, and how they relate to specific some of the prograde reactions. In the case of
metamorphic and deformation events. garnet-staurolite schists for example, it is
Some porphyroblasts (especially garnets) common to observe chloritoid inclusions in the
may show chemical or else textural zonation garnet but no chloritoid in the matrix of the
resulting from more than one phase of growth rock (Plate 5(a)). This provides evidence for
(Plates 6(d) & (e)). The included minerals and chloritoid being part of the earlier assemblage
various growth stages can give an important of the rock, but having reacted-out to form
insight into precursor assemblages and the staurolite during prograde metamorphism by a
metamorphic conditions experienced by the reaction in the KFMA5H system, such as
rock. In many cases, they may provide a key to
Ctd ~ Grt + ChI + 5tt + H 20. (12.1)
some of the reactions that have occurred
during the early evolution of the rock. Unless Without the inclusion evidence, the nature of
there is extensive volume diffusion, included the staurolite-forming reaction would be much
phases in porphyroblasts become shielded from more speculative.
changing conditions in the matrix of the rock The inclusion trails common to many
as metamorphism proceeds. In many cases, the porphyroblasts can also provide valuable in-
inclusions are of minerals such as quartz formation relating to the timing and style of
which, being stable over such a wide range of previous deformation events. The detailed

Deciphering polydeformed and polymetamorphosed rocks

interpretations of porphyroblast-foliation rela- of these diagrams, the standard notation of S1>

tionships were described in Chapter 9, and S2 and S3 is used to denote planar fabrics asso-
consequently will not be repeated here. The ciated with successive deformation events 0 1,
conclusions regarding the timing of growth of O2 and 0 3. Ml and M2 denote specific meta-
various minerals with respect to specific defor- morphic events.
mation and metamorphic events can be usefully
summarised using 'mineral growth-deforma- 12.2 Local and regional complications
tion sequence' diagrams. The example shown
in Fig. 12.1 uses a bar of variable length to Across a given tectonometamorphic terrain
represent the times at which various minerals (especially one of orogenic metamorphism),
of the assemblage grew in relation to different diachroneity of metamorphism in relation to
tectonic fabrics (dashed lines represent less deformation may be observed. In orogenic belts
significant mineral growth or uncertainty). not only will the timing of specific deformation
Figure 12.2 is a slightly more refined version of phases be diachronous across the region (e.g.
Fig. 12.1, and illustrates the time(s) of maxi- as a given deformation event propagates from
mum nucleation and growth of particular hinterland to foreland), but the timing of
phases by bulges in the horizontal bar. In each these events with respect to metamorphism
may also show significant variation. If this is
the case, a systematic variation in porphyrob-
last-foliation relationships may be observed
S1 S2 8a on a regional scale. This grossly oversimpli-
Muscovite fied scenario is shown schematically in Fig.
12.3, to emphasise that deformation and
- ~
metamorphic events are not intrinsically
linked, and are likely to show both spatial and
temporal variation across a given terrain.
Notice in Fig. 12.3 that at time tl locality A is
FIG. 12.1 An example of a mineral growth-deforma- experiencing 0 1 deformation, whereas locality
tion sequence diagram. The solid lines represent the C does not experience this until tr Also notice
timing of the mineral growth with respect to the devel-
opment of different fabrics in the rock. The dashed that porphyroblast growth associated with MI
lines represent less significant mineral growth or uncer- metamorphism is inter-0 1-02 (SI-S2) at A, and
tainty. shows syn-S2 relationships at B, but post-S2
relationships at C.
When drawing conclusions from porphyrob-

last-foliation relationships, it is important to
S1 S2
realise that local variations in bulk rock chem-
-- - istry can have a profound effect on the
sequence of metamorphic reactions and precise
timing of porphyroblastesis. This means that
~ the observed relationships and textural features
-~-- -@>- may show significant variation between differ-
ent lithologies. For example, garnets will prob-
FIG. 12.2 An example of a mineral growth-deforma- ably nucleate more readily in pelites than in
tion sequence diagram, where the bars represent the
timing of the mineral growth. The acme of the growth semi-pelitic or psammitic rocks. Thus, even at
is shown by bulges in the bars. the same exposure, garnets from different


points should always be considered when

attempting to interpret regional patterns of
metamorphism. Additionally, if more than one
tectonic unit is involved (e.g. several nappes) it
is essential to treat each unit separately, and
determine the tectonometamorphic history of
each. Such an approach may reveal differences
B A that gIve grounds for defining separate
DISTANCE terranes.
Foreland - ......--..;,,;,,;;........- Hinterland

FIG. 12.3 A schematic diagram illustrating diachrone-

ity of deformation and metamorphism from the hinter- 12.3 P-T-t paths
land to the foreland of an orogenic belt.
12.3.1 Introduction
lithologies can show different growth relation- The evaluation of 'peak' metamorphic condi-
ships and zonation with respect to a given tions has traditionally been one of the main
tectonic fabric. Because of this, it is important objectives of the petrologist studying meta-
to compare samples from comparable litholo- morphic rocks. On the basis of empirical
gIes. observation, theoretical considerations and,
A further complication in polydeformed increasingly, by detailed experimental work, it
areas is that near to major fault or thrust zones is well established that certain minerals and
a more complex structural and metamorphic mineral assemblages characterise a given range
history will be experienced. Such differences of P- T conditions.
may be especially apparent in rocks in the foot- The metamorphic facies concept (Chapter 2)
wall and hangingwall to major thrusts during is based on equilibrium mineral assemblages
orogenic metamorphism. In such cases, the that characterise specific P- T conditions.
emplacement of a 'hot' slab on to a 'cold' foot- Assigning the equilibrium assemblage of a
wall gives rise to simultaneous heating and given rock to a particular metamorphic facies is
prograde reactions in the footwall, and cooling the first approach to quantifying P-T, and
with accompanying retrogression in the hang- should give a reasonable estimate for the
ingwall, promoted by infiltration of fluids magnitude of P-T at peak metamorphism (see
liberated during prograde reactions in the foot- Fig. 2.1 for relative positions of different facies
wall. This means that closely associated rocks in P- T space, and Appendix III for typical
each side of the thrust may show quite different assemblages of each metamorphic facies). A
P- T-t paths until the point at which movement 'metamorphic facies series' denotes the system-
on the thrust terminates, and they then have a atic change in peak metamorphic conditions
common, linked evolution. Documented exam- attained by a sequence of rocks across a given
ples of such histories include the work of metamorphic terrane. For example, the classic
Pecher (1989) in the Himalayas, and Anderson Barrovian sequence of Scotland (Barrow, 1893)
et al. (1992) in the north Scandinavian shows a typical medium-pressure facies series
Caledonides. Such differences between hang- from greenschist facies rising to upper amphi-
ingwall and footwall evolution next to major bolite facies. However, this should not be taken
thrusts mean that in porphyroblasts from these to represent a palaeogeothermal gradient; nor
areas inclusion fabrics and growth zones may should it be considered as the prograde P- T
in part reflect local features. These cautionary path taken by the highest-grade rocks. Spear et

Deciphering polydeformed and polymetamorphosed rocks

mineral inclusions in the porphyroblasts. With

such a rock it is necessary to establish which
phases formed the peak metamorphic assem-
blage, which minerals were present before
peak conditions (as determined by inclusions
in porphyroblasts) and which phases devel-
oped late (often as retrogressive phases). By
defining P-T at particular stages in the meta-
morphic evolution of a given rock, it should
be possible to mark on a petrogenetic grid the
approximate path that a particular rock has
taken through P-T space. On the basis of
TE MPERATURE ------ current knowledge, an inherent assumption of
most petrogenetic grids is that P total = P H20'
FIG. 12.4 A schematic P-T diagram (modified from Taking other fluids such as CO 2 into consider-
Spear et a/., 1984; courtesy of the Geological Society of ation, or simply lowering aH20, can cause
America) showing the different paths taken by three
rocks (1, 2 and 3) in a metamorphic sequence (in terms significant shifts in the position of univariant
of metamorphic grade, 1 > 2 > 3). Note (1) that the reactions and stability fields for specific
low-grade rock does not represent the earlier part of assemblages. This means that the construction
the high-grade rocks' metamorphic history, and (2) of a P-T path based solely on petrogenetic
that the change in metamorphic grade seen in the field
(the metamorphic field gradient) represents the locus of grids on which PtOtol = P H20 may have serious
the peak P-T conditions attained by each rock, and not limitations.
the palaeogeothermal gradient. The portions of the
P-T-t loop that can be constructed from different lines Geothermometry and geobarometry
of evidence are also indicated. A, Realm of entrapment
of mineral inclusions in porphyroblasts; B, 'peak' P-T In addition to the estimation of metamorphic
conditions from geothermometry/barometry; C, retro- conditions based on key mineral assemblages
grade path from reset mineral equilibria; D, realm of and petrogenetic grids, quantitative estimates
entrapment of fluid inclusions. of P- T conditions may also be obtained using
various experimentally, empirically or theo-
al. (1984) emphasise that the locus of 'peak' retically calibrated geothermometers and geo-
P-T conditions in samples across a given meta- barometers. Most of these rely on the
morphic terrain, termed the 'metamorphic field accurate measurement of the chemical
gradient', is rarely coincident with the rocks' compositions of mineral phases in the equi-
P-T path. It is more likely to be representative librium assemblage. For this reason, it is only
of the maximum temperature or maximum since the electron microprobe has become a
entropy state achieved by rocks along their standard tool of the metamorphic petrologist
individual paths (Fig. 12.4). that this approach to quantifying P-T condi-
It is highly erroneous to build up a P-T loop tions has been possible.
for an area based on assemblages in several Before attempting any geothermometry or
rocks from various grades across the particular geobarometry, it is of fundamental importance
metamorphic belt. Instead, it is necessary to to understand fully the mineralogical and
build up P-T trajectories for individual textural relationships preserved within the
samples. Some samples lend themselves to this rock. Of primary importance is to establish
more readily than others. Ideally suited are that the thermobarometry is based on an equi-
porphyroblastic samples with a range of librium assemblage. Thorough reviews by


Essene (1982, 1989) and Spear (1993, Chapter phases of the system, it is important to estab-
15) examine the various geothermometers and lish the nature of any chemical variations (e.g.
geobarometers used by metamorphic petrolo- zonation) within minerals, and to assess
gists. In these reviews, the positive points of whether this reflects prograde or retrograde
different approaches are identified, but the reactions. Chemical (X-ray) mapping and back-
assumptions and limitations of the different scattered electron imaging (Fig. 5.22) are often
systems and techniques are also critically evalu- employed to evaluate detailed mineralogical
ated. In addition to the use of petrogenetic relationships prior to microprobe analysis of
grids, the main approaches to geothermometry specific points for thermo barometry.
and geobarometry that have developed over the Having evaluated peak metamorphic condi-
past two decades fall under one of five main tions based on observed mineral assemblages
headings, namely: and, where possible, utilising one or more of
the available geothermometers and geobarome-
(a) exchange thermometry (e.g. Grt-Bt pairs or
ters, the position of this event on the P- T
Grt-Hbl pairs);
trajectory can be accurately defined. Spear et
(b) solvus thermometry (e.g. Cal-Dol pairs and
al. (1984) give a schematic summary diagram,
two-feldspar thermometry);
reproduced here as Fig. 12.4, showing the types
(c) solution models for multiple systems, or 'net-
of information that can be used to construct
transfer equilibria' (e.g. Grt-Rt-AlzSiO s
different parts of the P- T loop for individual
samples. Although temperatures can often be
(d) internally consistent thermodynamic data
evaluated quite well, accurate pressure esti-
set approach; and
mates are usually more difficult to obtain. A
(e) fluid inclusions.
study of porphyroblast-foliation relationships
In addition to these, illite crystallinity, vitrinite allows assessment of the timing of metamor-
reflectance and conodont colour index have all phic events with respect to given deformation
been used in low-grade rocks. It is outside the events. Combining all of the above information
scope of this book to give an in-depth analysis will give a P- T-t path, but without the
of the pros and cons of the various approaches absolute quantification of time.
to thermo barometry in metamorphic rocks, but
instead the reader is referred to the excellent Radiometric dating
reviews cited above. Essene (1989) gives a Dating of specific metamorphic phases associ-
particularly useful evaluation of optimal ther- ated with a given metamorphic event will
mobarometers for each metamorphic facies, provide a date for a specific part of the P-T-t
and gives an extensive list of references. Where loop. Similarly, radiometric dating of an intru-
possible, it is always advisable to base any eval- sion responsible for a certain metamorphic
uation of P- T conditions on more than one event, or constrained in the tectonometamor-
approach. If this is done with care, thermo- phic evolution by cross-cutting relationships
barometry in many metamorphic terrains may will place a specific time on part of the trajec-
be accurate to ± 50°C and ± 1 kbar (Essene, tory. Depending on the amount of information
1989). Of vital importance is the careful char- accrued, it may in some instances be possible to
acterisation of equilibrium mineral assemblages obtain an indication of uplift rates.
(including accessory phases), and a detailed There are various techniques that can be
evaluation of the microstructural interrelation- employed to obtain a whole-rock age, or a
ships of all phases in the rock. When undertak- metamorphic age based on specific mineral
ing chemical micro-analysis of individual separates. The various techniques, and the

Deciphering polydeformed and polymetamorphosed rocks

theory behind them, are dealt with in consider- TABLE 12.1 Estimated closure temperatures for vari-
able detail by Faure (1986), and will not be ous geochronological systems that are useful for
constraining metamorphic cooling histories (based on
dealt with here. The age equation, Spear, 1993; Zeitler, 1989).
1:1 1n (DOC
N" + 1 = t (12.2) Mineral Isotope system Tete}
Garnet U-Pb > 800
gives the basic relationship from which the age Zircon U-Pb > 700
of a sample can be determined. In this equa- Allanite U-Pb 600-750
Monazite U-Pb > 700
tion, D~- is the number of radiogenic daughter Sphene (titanite) U-Pb 500-600
atoms produced from the parent, N is the Garnet Sm-Nd '" 600
number of remaining parent atoms, and A is Hornblende 4oArj39Ar 470-550
the decay constant. Knowing A, and by measur- Muscovite 4oArl 39 Ar 350-425
Biotite 4oArl39 Ar 260-350
ing D'; and N, the age of the sample can be K-feldspar 4oAr/39 Ar 125-350
determined. Zircon Fission tracks '" 175-260
At high temperatures, a mineral will lose Sphene (titanite) Fission tracks '" 250
radiogenic daughter product, at medium Apatite Fission tracks '" 100-150
temperatures daughter product will start to
accumulate, and then at low temperatures
daughter product accumulation will take place minerals with different values of T" it is possi-
without loss. Eventually, the rate of increase of ble to obtain a series of dates corresponding to
daughter product with time becomes more or specific temperatures during the cooling history
less constant. The temperature at which this of the rock. This enables the cooling histories
rate of increase of daughter to parent becomes of individual rocks to be defined on an
constant is the closure temperature (TJ absolute rather than just a relative timescale. In
Closure temperatures of individual minerals Table 12.1 it is shown that U-Pb dating is most
provide one of the most useful approaches to useful for rocks that have attained high-
understanding the T -t evolution of a given temperature granulite facies conditions, and
rock, especially those that have cooled slowly. 40Ar/39Ar step-heating dates from micas and
By mathematical analysis, Dodson (1973) hornblende are most useful for rocks that have
devised an expression for Tc incorporating the experienced moderate- and high-temperature
various factors that have a direct influence; Barrovian-style orogenic metamorphism, while
namely, the cooling rate (dTldt), the chemical fission track dates are of greatest use at the
diffusivity (a function of activation energy for lowest temperatures in the final stages of meta-
chemical diffusion and temperature) and the morphic cooling, and in studies of basin evolu-
size of the diffusion domain of the crystal tion. Having established the detailed P- T-t
(which influences the length scale for diffu- evolution of a given rock, this can in turn be
sion). related to crustal-scale processes such as uplift.
Although T, is mostly dependent on the acti- This has proved increasingly useful towards
vation energy of diffusion for specific minerals, our understanding of crustal processes and the
it is also a function of cooling rate, and thus it tectono metamorphic evolution of individual
is not possible to define a unique closure rocks (e.g. Parrish et at., 1988; Mezger et aI.,
temperature for individual minerals. 1991; Anderson et aI., 1992). As with geother-
Nevertheless, closure temperature estimates mobarometry, it is crucial when undertaking
have now been established for various meta- thermochronology to be able to integrate it
morphic minerals (Table 12.1), and by using with petrological observations, and to be sure


of the precise mineralogical and microstruc- to rigid-block uplift and erosion, but can also
tural features of the rock being dated, and the result from extensional deformation along low-
particular metamorphic event to which the angle normal faults, and ductile thinning of the
dated material relates. crust (Hames et al., 1989; Haugerud & Zen,
In the following sections, examples are 1989). The high-grade rocks of the Tauern
discussed that represent the variety and Window, Eastern Alps provide a good example
complexity of P- T trajectories from rocks of of a situation in which the two differing inter-
different metamorphic environments. In partic- pretations of the decompression path have been
ular, those P- T-t paths that are most diagnos- presented. Until the work of Selverstone and
tic of a particular style of metamorphism are colleagues (e.g. Selverstone et al., 1984;
highlighted. In all cases, the emphasis is on the Selverstone & Spear, 1985; Selverstone, 1988),
mineralogical, microstructural and ther- the Tauern Window rocks had largely been
mochronological evidence used to establish a considered as a classic example of high-grade
particular P-T-t path. metamorphism as a result of overthrusting and
subsequent erosion-controlled exhumation.
However, Selverstone (1988) presented
12.3.2 Orogenic metamorphism
evidence to suggest that the western margin of
Orogenesis related to continental collision is the Tauern Window is a low-angle normal
associated with extensive crustal thickening. fault, and that a prolonged period of
Large areas of crust become deeply buried, and Oligocene-Miocene extension (after the main
the rocks involved experience an accompanying phase of nappe emplacement) involved ductile
pressure increase (England & Thompson, stretching and low-angle normal faulting to
1984). This buried crust then experiences a exhume the high-grade Tauern Window rocks.
period of heating as the perturbed crustal geot- Although the peak metamorphic and retro-
herm evolves towards a new equilibrium grade (exhumation) stage of a typical 'clock-
geotherm. Thrusting on various scales can give wise' P-T trajectory is usually well defined, the
rise to complex 'saw-toothed' geotherms (e.g.
Oxburgh & Turcotte, 1974; England &
Richardson, 1977), but with time the conduc- kbar GPa
tion of heat causes heating of the footwall and 16 1.6

cooling of the hangingwall, thus smoothing the

geotherm. As time progresses, erosion and/or 12 1.2
extensional thinning of the mountain belt I!!
causes unroofing (exhumation) of the deeply (f)

n. 8 0.8
buried and metamorphosed core of the orogen.
During this stage, the rocks in question experi-
ence both pressure and temperature decrease, 4 0.4

such that the most characteristic feature of

rocks displaying Barrovian-style orogenic meta- o -f--r--r........-I'-..,-"'r--1'''--r--i'--r-+ 0.0
morphism is one of a 'clockwise' P-T-t trajec- o 200 400 600 800 1000
Temperature fe)
tory (Fig. 12.5). The decompression stage of
such P-T-t trajectories is often referred to as
FIG. 12.5 A clockwise P-T trajectory of orogenic
'uplift', but terms such as 'unroofing' or metamorphism. Boxed areas shown on the diagram are
'exhumation' are a more accurate description, the metamorphic facies designated in Fig. 2.1. Al2 SiO s
since such decompression may not only be due triple-point and univariant curves are also shown.

Deciphering polydeformed and polymetamorphosed rocks

evidence for the early P- T evolution is usually morphism have been described for many
fairly limited. Inclusions of mineral phases in orogenic belts by numerous workers. For an
porphyroblasts provide some of the most useful area in the north Scandinavian Caledonides,
evidence for constraining early stages of the Barker (1989), Barker & Anderson (1989) and
evolution (Plate 5(a)), but in most cases there Anderson et al. (1992) used a combination of
are few diagnostic minerals. Alternatively (e.g. 4oAr/39Ar closure ages for hornblende and
Spear, 1986; Burton et ai., 1989), sections of muscovite, coupled with estimates of peak
the P- T trajectory can be modelled by use of metamorphism based on geothermobarometry,
chemical variations in zoned garnet porphyrob- to constrain the P- T-t evolution of individual
lasts and inclusions of biotite and plagioclase nappes. Included mineral phases were used as
contained within the individual porphyroblasts evidence for some of the prograde reactions,
(Fig. 12.6). Numerical modelling by England & and closure temperatures for muscovite,
Thompson (1984), Thompson & England coupled with fluid inclusion studies of retro-
(1984) and Ridley (1989) has also shown gression-related veins (Bennett & Barker,
that crustal thickening during orogenesis 1992), allowed the retrograde stage of the
produces Ky (± Sil) grade orogenic metamor- P- T-t evolution to be constrained to some
phism and characteristic 'clockwise' P-T-t degree. Two photographs of a garnet-mica
trajectories. schist from the Troms region of north Norway
On the basis of various lines of evidence, are shown in Fig. 12.7. The sample lies within
'clockwise' P-T-t paths associated with colli- one of the higher nappes of the Scandinavian
sional orogenesis and Barrovian-style meta- Caledonides, and has been metamorphosed


10 1·0


8 0·8 ~


6 0·6

440 460 540 560

Temp. 0 c

FIG. 12.6 A clockwise P-T trajectory based on porphyroblast chemical zonation (after Burton et aI., 1989; cour-
tesy of the Geological Society). The P-T paths were determined from modelling of the inclusion-free rims of garnets
in samples R60 (triangles) and R223 (circles) from the Skaiti Supergroup, Sulitjeima, Norway. The dashed portion
of the R60 curve was drawn using six P-T points (shown with associated error boxes) calculated from biotite and
plagioclase inclusions in the inclusion-rich core at varying distances from the centre. The modelling shows two
distinct periods of growth: cores growing during increasing P-T and the rims growing during decompression.


during the Scandian phase of the Caledonian although in some porphyroblasts (Fig. 12.7(b))
Orogeny. The garnets have two distinct the core exhibits a 'straight' inclusion fabric
growth stages; an early rather rounded core defined by small crystals of quartz. This is
region, followed by a later rim growth produc- interpreted as overgrowth of an early (5,)
ing subhedral to euhedral form. The early schistosity, the main fabric in the rock being
stage of growth is relatively inclusion-free, the regional (52) schistosity, defined by aligned


FIG. 12.7 A garnet-mica schist from a Caledonian nappe of the Troms region, Norway (see the text for details
of the polymetamorphic history). Scale =0.5 mm (PPL).

Deciphering polydeformed and polymetamorphosed rocks

muscovite and biotite (Fig. 12.7(a)). The AIm + Rt ~ 11m + Ky + Qtz.

boundary between the first and second growth
stages is marked by a distinct line, followed by For units north of the Main Karakorum
abundant quartz inclusions in the innermost Thrust, northern Pakistan (Himalayas), Allen
part of the second growth stage. This second & Chamberlain (1991) were able to infer
growth zone is only observed in a few other 'clockwise' P-T trajectories on the basis of
samples from the nappe, and there is no corre- reaction histories based on observed
lation with proximity to the basal or roof microstructural and textural relationships. For
thrusts. Because of this, possible frictional or example, in samples from the Braldu Nappe,
down-heating effects are ruled out as the cause they frequently recorded a corona of Qtz + Bt +
of the second growth stage. It is more likely Sil after kyanite, and replacing Grt + Ms.
that bulk rock chemistry is the main control, Additionally, they noted the common case of
and that the outer growth zone developed in kyanite incipiently replaced by sillimanite.
response to a specific discontinuous reaction Integrating these petrographic observations
within the rock, such as chloritoid breakdown. with the data obtained from geothermobarom-
The dense zone of quartz inclusions probably etry, they deduced the P-T path shown in Fig.
reflects rapid initial growth of the second 12.10. In a study of metamorphism in the
phase of garnet. This phase of growth Central Himalaya, Pecher (1989) discussed the
occurred synchronous with 52, but terminated P-T-t trajectories in units either side of the
before the 52 schistosity was completely devel- Main Central Thrust (MCT). Pecher (1989)
oped. In consequence, the porphyroblasts are concluded that the Midland Formation in the
wrapped by the schistosity (Fig. 12.7(a)). Plate footwall to the MCT experienced prograde
5(a) is from the same nappe, and shows inclu- metamorphism due to downward transfer of
sions of chloritoid in a garnet porphyroblast, heat during the 'hot' emplacement of the
whereas elsewhere in the matrix staurolite Tibetan Slab (also known as the 'High
occurs and chloritoid is absent. This provides Himalayan Crystalline Sequence') which forms
evidence that staurolite formed from a chlori- the hangingwall to the MCT. The Tibetan slab
toid breakdown reaction and, coupled with is interpreted to have experienced simultaneous
garnet 'core-rim' thermobarometry, helps to cooling as heat was transferred to the footwall.
constrain part of the P-T trajectory (Fig. This example usefully illustrates that while the
12.8). 'clockwise' P-T trajectories may have similar
In the Central Alps, Italy, Diella et al. (1992) forms for a range of rocks from different
studied the metamorphic evolution of South nappes in a given terrain, in detail the trajecto-
Alpine metamorphic basement and determined ries are more complex. When time is taken into
a 'clockwise' P-T trajectory on the basis of consideration there may be considerable
detailed microstructural and petrological obser- diachroneity between seemingly similar P-T
vations from metapelites. They established the trajectories. This point was also emphasised in
nature of key univariant reaction curves that the work of Anderson et al. (1992) and Barker
had been crossed during uplift. For example, in (1995) in a study of part of the Scandinavian
the Val Vedello basement (Fig. 12.9), they Caledonides.
showed how, during decompression, late kyan- The Taconian (Middle to Late Ordovician)
ite grew and rutile became destabilised, as indi- and Western Acadian (late Silurian to middle
cated by the development of ilmenite rims. This Devonian) metamorphism of New England,
is interpreted in terms of the reaction (Bohlen USA, also produced 'clockwise' P-T paths char-
et aI., 1983) acteristic of collisional orogenesis involving


20 ~--~--~----~--~----~--~--~----~--~

b ~ ,
... ,

:.... '


CU . ,
~ 10

, .


FIG. 12.8 A P-T trajectory for the Hogtind Nappe, Troms, Norway (based on Barker, 1989; Barker &
Anderson, 1989; Anderson et al., 1992; Bennett & Barker, 1992), superimposed on the calculated P-T grid of
Spear & Cheney (1989) and Spear (1993), for the KFMASH system. Only those assemblages containing Grt + Bt
are shown. The dashed contours show the Fe/(Fe + Mg) in garnet as a function of P and T in each assemblage.
Note that the Fe/(Fe + Mg) in garnet changes along the univariant curves.

Deciphering polydeformed and polymetamorphosed rocks

II 40
9 \: • • \1

\\'" .
8 30
,Q Ky .../
7 .::
v /

25 ~
::l ~ :. :'

'"'" .~~:./
Q. 6 .. II
20 ~

4 15

1. 5

300 400 500 600 700 800 900

Temperature (OC)

Fig. 12.9 A modified version of Fig. 9 of Oiella et al. (1992), courtesy of Blackwell Science, showing key univari-
ant reactions crossed during the uplift history of metapelites from the Val Vedello basement, Central Alps, Italy.
0\, 02 = P-T conditions for assemblages formed during 0\ and 0 2 structural events. AlzSiO s triple-points: H =
Holdaway (1971); RGB = Richardson et al. (1969). Initial melting of metapelite (wet and dry) after Thompson &
Tracy (1979). AIm + Rut = 11m + Ky + Qtz after Bohlen et al. (1983). Staurolite equilibria after Hoschek (1969),
Chatterjee (1972) and Rao & Johannes (1979). AIm + V = Fe-ChI + Qtz after Hsu (1968) and Naggar &
Atherton (1970). Sps + V = Mn-Chl + Qtz after Hsu (1968); 1 = Bt + Grt + AlzSiO s = ChI + Ms + Qtz from
Hirschberg & Winkler (1968). The dotted lines represent the minimum and maximum calculated In K for Grs +
Ky + Qtz = An (Ghent, 1976) and the Kd for Grt-Bt equilibrium (Perchuk & Lavrent'eva, 1983).

crustal thickening, followed by a period of observed P- T trajectory (constrained by

heating then rapid decompression during mineral equilibrium data and fluid inclusion
unroofing. This has been described by Hames data), Hames et al. (1989) concluded that the
et al. (1989, 1991) and Armstrong et al. rapid decompression following peak Acadian
(1992), by a combination of petrological work, metamorphism is best explained in terms of
microstructural relationships, geochronology rapid uplift coupled with erosion and tectonic
and thermal modelling (Fig. 12.11, path-A). denudation.
Comparing the theoretical models with the However, not all cases of orogenic (regional)


kbar GPa
metamorphism record 'clockwise' P-T-t trajec-
tories. In particular, some areas that have
9 __--"'2"'T":::------~7'"T'"0.9
experienced high- T, low-P (Buchan-type) meta-
8 0.8 morphism have been shown to exhibit an 'anti-
clockwise' P-T trajectory. For example, on the
7 0.7 basis of petrological, geochronological and
fluid inclusion evidence, various authors (e.g.
Schumacher et ai., 1989; Armstrong et ai.,
0.5 1992; Winslow et ai., 1994) have described an
'anticlockwise' path associated with Buchan-
4 0.4 type metamorphism in the Eastern Acadian
belt, Massachusetts, USA (Fig. 12.11, path-B).
3 0.3
The evidence for this path includes sillimanite
2 4---.,.---,...a...----,;---..... 0.2 pseudomorphs after andalusite, and late-stage
400 500 600 700 development of higher-pressure garnet-bearing
Temperature (OC) assemblages in place of low-P cordierite assem-
blages (Winslow et ai., 1994). Fluid inclusion
FIG. 12.10 The P-T path for the Braldu Nappe, studies also support an interpretation in terms
northern Pakistan, Himalayas (modified after Allen & of initiallow-P heating, and then thickening at
Chamberlain, 1991). The AI 2 SiO\ phase diagram is high temperature followed by isobaric cooling,
after Holdaway (1971) and the muscovite breakdown
curve is from Chatterjee & Johannes (1974). The before final decompressional unroofing. The
dashed line gives slopes calculated from entropy and early evolution is interpreted in terms of wide-
volume data for the reactions Ms + Grt ~ 2SillKy + spread melting in the lower crust of a back-arc
Qtz + Bt. The stippled area represents an error box of ± environment in response to a build-up of heat
2 kbar and ± 50°C for the average of the peak P-T
determinations for the Braldu Nappe. in basin-fill sediments enriched in radioactive
heat-producing elements. Back-arc extension
led to widespread emplacement of intermediate
kbar GPa and acidic plutons coupled with regional-scale,
18 1.8 low-P/high- T, Buchan-type metamorphism
(Armstrong et al., 1992; Winslow et al., 1994).
12 1.2
Subsequent east-west shortening during the
Acadian orogeny caused crustal thickening.
This caused pressure to increase while the
8 0.8
rocks were still at high temperature (but start-
ing to cool), and later unroofing completed the
0.4 'anticlockwise' P-T-t trajectory (Fig. 12.11,
path-B) .
..f-.......--"¥-....,....-+ 0.0
Temperature «,C) 12.3.3 Orogenic metamorphism with a
subsequent thermal overprint
FIG. 12.11 Path A is the Taconian clockwise P-T-t The previous section described the characteris-
trajectory, New England, USA; path B is the Eastern tic 'clockwise' P-T-t trajectories that typify
Acadian anticlockwise P-T-t trajectory from an adja-
cent metamorphic belt (based on Armstrong et ai.,
Barrovian-style orogenic metamorphism, but
1992). in cases in which there is more than one

Deciphering polydeformed and polymetamorphosed rocks

metamorphic event, it is quite common to kbar GPa

observe an earlier orogenic (regional) metamor- 16 1.6

phic event overprinted by a late-stage thermal

(or 'contact') metamorphic event, when grani- 12 1.2
toid melt generated in the lower crust ascends f
and comes to rest at upper crustal levels.
Modelling by De Yoreo et al. (1989) examined
i 8 0.8

the role of crustal anatexis and magma migra-

tion in relation to the thermal evolution of 4 0.4

thickened continental crust. When a single

granitoid sheet 2 km thick is emplaced into the o +-.--,..II....t'-"'T'"'..u,----t"--.-+"'""T"-+ 0.0
upper crust it causes a short-lived and localised o 200 400 600 800 1000
Temperatul'8 fC)
perturbation or 'spike' to sillimanite grade
conditions, the value for pressure depending on FIG. 12.12 A schematic illustration of a greenschist
the level at which the intrusion is emplaced. facies P- T trajectory with a late AndiSil overprint. This
However, with widespread melt generation and is the typical trajectory of low-grade metasediments
intruded by late-/post-orogenic granitoids (e.g.
melt migration to higher crustal levels, a Skiddaw Granite aureole, Lake District, England; see
regionally extensive high- Tilow-P metamorphic Fig. 12.13).
event may be initiated.
Inger & Harris (1992) described a 17-20
Ma, high-T, metamorphic overprint (M2) in 1 (c)). The distinctive chiastolite cross is still
metasediments of the upper part of the High clearly recognisable (Fig. 12.13(b)), despite the
Himalayan Crystalline Sequence in the pseudomorphing. The pseudomorphs show
Langtang Valley, northern Nepal, that had clear relationships to indicate that the original
previously experienced an M1 (pre 34 Ma) porphyroblasts statically overprinted a pre-
Barrovian (Ky-St grade) metamorphism. existing slightly crenulated slaty cleavage: that
Although this thermal overprint is recognised is, in both cases there is no deflection of the
elsewhere in the same sequence above the cleavage around the porphyroblasts, and there
MCT, there is still considerable debate concern- are no pressure shadows. In Fig. 12.13(a),
ing its cause. compositional layering trends from bottom left
A more straightforward and common exam- to top right, and has a continuous (or perva-
ple of a thermal overprint is the case in which sive) slaty cleavage (S1) sub-parallel to it. This
metasediments that have experienced sub- is overprinted by a close-spaced crenulation
greenschist or greenschist facies orogenic cleavage (S2) trending from bottom right to top
(regional) metamorphism in the upper crust left. In Fig. 12.13(b), the S2 cleavage is not
experience heating from a late-stage intrusion represented, but S1 (more highly magnified)
during their uplift trajectory (Fig. 12.12). The passes from top to bottom, cross-cutting the
chiastolite slates from the metamorphic aure- prominent compositional layering trending
ole of the Skiddaw Granite (Lake District, bottom left to top right. The earlier cleavages
England) document such a history. Figs and greenschist facies fine-grained matrix
12.13(a) & (b) show andalusite (var. chiasto- assemblages formed during low-grade regional
lite) formed at the time of granite intrusion metamorphism in the end-Silurian stage of the
(early Devonian), and subsequently pseudo- Caledonian Orogeny. The example in Fig.
morphed by an aggregate of sericite rimmed by 12.14(a) is a garnetiferous slate from the Isle
chlorite (for unaltered chiastolite, see Plate of Man, UK. It has a fine-grained matrix




FIG. 12.13 Chiastolite slates from the metamorphic aureole of the Skiddaw Granite, Lake District, England (see
the text for details of the polymetamorphic history). Scale = 1 mm in (a), 0.5 mm in (b) (PPL).

compnsmg Qtz-Ms-Chl-Bt-Ilm, which which are m turn overprinted by

defines a continuous slaty cleavage. In detail, euhedral spessartine-almandine porphyrob-
this is seen to have experienced a later fine- lasts (Gillott, 1955). This phase of porphrob-
scale crenulation (Fig. 12.14(b)). These lastesis occurred during the late-thermal
Caledonian structures are cross-cut by thin, overprint associated with a mid-Devonian
variably oriented chlorite-muscovite vems, granite intrusion.

Deciphering polydeformed and polymetamorphosed rocks



FIG. 12.14 Garnetiferous slate from the Isle of Man, England (see the text for details of the polymetamorphic
history). Scale =1 mm in (a), 0.5 mm in (b) (PPL).

12.3.4 Granulite facies P-T-t paths stage of near 'isothermal decompression'

(lTD), and paths that document a significant
Bohlen (1987), Harley (1989) and Spear period of near 'isobaric cooling' (IBC). The
(1993) have reviewed the characteristic P-T-t evidence for the two fundamentally different
paths associated with the evolution of gran- trajectories (Fig. 12.15) comes from the inter-
ulite facies rocks. There are two basic types of pretation of reaction features, geothermo-
path to consider, namely, paths showing a barometry and geochronology. The ability to


kbar GPa prograde stage of the tectonothermal evolution

16 1.6 is often poorly constrained.
Fig. 12.16, based on Spear (1993) and
12 1.2 Harley (1989), shows reaction microstruc-
tures (insets) that give evidence for two differ-
I 8 0.8
ent reaction types and two different P-T paths
for the individual rocks concerned. Path A,
based on reaction microstructure (corona) of
4 0.4 inset (a), suggests a granulite facies rock that
followed an IBC path. Sapphirine (centre of
o +---r---r..ll....-I1--r-..IIoo-_"--..--+-r--+O.O inset (a)) is separated from quartz, and thus
o 200 400 600 600 1000
interpreted to be out of equilibrium with
Temperatura «Ie)
quartz. Instead, an irregular corona of Opx +
FIG. 12.15 Typical isobaric cooling (IBC) and isother- Sil + Crd has formed. This indicates that the
mal decompression (lTD) paths recorded in granulites reaction Opx + Sil + Crd ~ Spr + Qtz has
based on reaction textures and geothermobarometry moved from right to left (i.e. down tempera-
(simplified from data in Harley, 1989). The boxed ture). Path B, based on the reaction
areas shown on the diagram are the metamorphic
facies designated in Fig. 2.1. Al2 SiO s triple-point and microstructure (symplectite) of inset (b), is
univariant curves are also shown. quite different from the previous case. In this
example, the reaction microstructure relates
to a continuous divariant reaction (Grt + Qtz
~ Crd + Opx) on an lTD path, with the
distinguish between the features indicative of position of the reaction curve in P-T space
an lTD path as opposed to features of an IBC varying as a function of composition. The
path is crucial, since the different paths imply sub-horizontal lines of Fig. 12.16, labelled 0.3
fundamentally different crustal processes. In to 0.9, represent the Fe/(Fe + Mg) composi-
practice, this distinction may not always be tion of garnet in the assemblage Grt + Crd +
straightforward, as Frost & Chacko (1989) Opx + Qtz at given P-T. The lines show how
and Spear (1993) point out. Initial tectonic the Fe : Mg ratio of garnet increases during
thickening of the crust and subsequent rapid decompression by the continuous reaction Grt
exhumation by extensional tectonism, + Qtz ~ Crd + Opx. Simultaneous with the
erosional processes or some combination of change in the Fe : Mg ratio, garnet is being
the two, is the most realistic interpretation of steadily consumed, as newly formed cordierite
an lTD path. and orthopyroxene grow to define an intricate
Bohlen (1987) and Ellis (1987) have symplectic intergrowth separating the
described 'anticlockwise' P-T paths with an corroded garnet core from the original coarse
IBC stage following an earlier high- T, low-P orthopyroxene and quartz (for further details,
thermal peak induced by magma accretion at see Harley, 1989; Perchuk, 1989). The same
the base of existing continental crust. However, reaction features were also used as evidence
other types of IBC path are also possible. One for isothermal decompression by Hisada &
of the problems when interpreting the cause of Miyano (1996), in a study of granulite facies
IBC paths is that while the retrograde IBC part rocks from the Botswanan Limpopo belt.
of the trajectory is often well defined on the There are many other well documented
basis of combined mineralogical and mICro- examples in which P-T paths of granulites
structural reaction features, the crucial have been constructed utilising information

Deciphering polydeformed and polymetamorphosed rocks

. GPa
16 1.6

14 1.4

12 1.2
10 1.0

a.. 8 0.8

4 0.4

2 0.2
900 1000 1100
Temperature (0 C)
FIG. 12.16 A composite diagram based on Harley (1989) and Spear (1993) to show examples of specific reac-
tion textures associated with IBC and ITO paths in granulites. The lines labelled 0.3 to 0.9 represent the Fe/(Fe +
Mg) composition of garnet in the assemblage Grt + Opx + Crd + Qtz. The garnet compositions become enriched
during decompression associated with the continuous reaction Grt + Qtz ~ Crd + Opx. The reaction lines
numbered 1 to 7 are as follows (right-hand side of equation represents higher-P assemblage): (1) Grt + Crd + Sil
~ Spr + Qtz; (2) Grt + Crd ~ Opx + Spr + Qtz; (3) Spl + Qtz ~ Grt + Spr + Sil; (4) Spl + Opx + Qtz ~
Grt + Spr; (5) Spl + Crd ~ Grt + Spr + Qtz; (6) Spl + Crd ~ Spr + Opx + Qtz; (7) Crd + Spl + Sil ~ Spr +
Qtz. Inset (a): a multi-corona reaction texture between Spr and Qtz in granulite from Enderby Land, Antarctica,
representing an IBC path (path A). High-density stipple = Spr; low-density stipple = Qtz; black = Opx; white =
Crd; cross-hatch =Sil. Scale =1 mm. Inset (b): the reaction texture of an ITO path (path B) for granulite from the
Sharyzhalgay Complex, Lake Baikal. 1 = Bt; 2 =11m; all other minerals labelled on inset. Scale = 1 mm.

from symplectite and corona reaction features Grt + Sil reaction being crossed during the
in conjunction with petrogenetic grids. For retrograde isobaric cooling stage of an anti-
example, Waters (1989), in a study of the clockwise P-T-t path. In contrast, Clarke &
Namaqualand granulites of southern Africa, Powell (1991), in a study of granulites from the
uses the observation of Sil + Grt coronas Musgrave Complex, central Australia, record
around spinel as evidence for the Spl + Qtz ~ coronas and symplectites defining the reaction


this deduced an isothermal decompression

retrograde stage of a clockwise P-T-t trajec-
tory (Fig. 12.17).
A study of silica-undersaturated granulites
from the Strangways Range, central Australia
(Goscombe, 1992) used coronitic and symplec-
IGn. OIzJ titic reaction textures in the FMAS system to
define the reaction sequence (1) Crd + Spl +

• Om] Crn ~ Spr + Sil, (2) Crd + Spl ~ Opx + Spr +
Sil, (3) Spr + Spl + Sil ~ Opx + Crn and (4) Spr
+ Sil ~ Crd + Opx + Crn, as evidence for an
'anticlockwise' P-T-t trajectory involving a
prograde heating stage followed by a retro-
grade near-isobaric cooling. In contrast, a study
of silica-deficient granulites from the Bamble
region, southern Norway (Kihle & Bucher-
Nurminen, 1992), defines a clockwise P-T-t
path, with evidence for a retrograde period of
isothermal decompression based on symplec-
tites and coronas that give evidence for the
discontinuous FMAS reaction Opx + Sil ~ Spr
+ Crd + Crn. In high-pressure garnet-
sapphirine granulites from the Central
Limpopo Mobile Belt, Zimbabwe, Droop
(1989) uses symplectitic and coronitic reaction
textures as evidence for the retrograde reac-
tions Krn ~ Spr + Crd and Grt ~ Ged + Crd +
Spr + Spl (Fig. 12.18). This in turn is used as
evidence for near-isothermal decompression
during a period of rapid uplift. A point to add
in relation to the use of reaction textures and
FIG. 12.17 An lTD retrograde reaction sequence, petrogenetic grids as an approach to defining a
determined for metapelitic granulites from In Ouzzal, path of near-isothermal decompression or
Algeria. Inset (a): the reaction Hy + Sil ~ Crd + Spr
near-isobaric cooling in granulite facies rocks
(sample C2). Scale = 0.1 mm. Inset (b): the reaction Grt
+ Qtz ~ Hy + Crd (sample C5). Scale = 0.1 mm is that, in the absence of good thermobaromet-
(modified after Bertrand et al., 1992). ric data, it is often difficult to define with any
accuracy the angle at which a given reaction
curve is crossed. This point was emphasised by
Grt + Sil ~ Crd + Spl, which they use as Vernon (1996), and while many authors show
evidence for retrogression during isothermal near-isothermal or near-isobaric paths, the
decompression. The detailed study by Bertrand truth of the matter is that although the direc-
et at. (1992) on Precambrian granulites from tion of the arrow is usually well-defined, the
the In Ouzzal craton, southern Algeria, also precise positions of the paths may not be
used corona and symplectite relationships to defined so accurately, and to some extent they
infer an array of retrograde reactions, and from are schematic.

Deciphering polydeformed and polymetamorphosed rocks

FIG. 12.18 (a) Symplectites and coronas associated with decompression-related reactions (see text for details) in
high-pressure Grt-Spr granulite from the Central Limpopo Mobile Belt, Zimbabwe (modified after Fig. 2a of
Droop, 1989).

. 12.3.5 Blueschist facies P-T-t paths

Ernst (1988), Perchuk (1989) and Spear (1993)
Ml have graphically summarised the typical style
~ M2 HARRIS , of P-T-t trajectories of rocks from subduction
(1"4) zone complexes (Fig. 12.19). Not suprisingly,
10 . . . . p-r'ATH the mineralogical and microstructural evidence
suggests that the typical path of a blueschist
facies rock involves an initial period of rapid
pressure increase at low temperatures, as the
rocks are subducted to deep levels. Following
this, the rocks heat up to some extent prior to
exhumation. In most cases, blueschist facies
assemblages display a greenschist or low-
• Spl·C,d
(17) amphibolite facies metamorphic overprint (e.g.
Holland & Ray, 1985; Kryza et at., 1990). This
is perhaps best explained in terms of rapid
exhumation and isothermal decompression
.00 700 eoo toO 1000
while still at temperatures in the range
r <"c) 300-500°C. Reactions to promote such a meta-
morphic overprint are greatly facilitated if
FIG. 12.18 contd (b) The interpreted P-T trajectory on fluids infiltrate the rocks during exhumation.
the basis of the petrographic evidence of (a) (Fig. 10 of Whatever the case, the most characteristic
Droop, 1989; courtesy of Blackwell Science).


Kb~ G~
12 1.2

8 0.8
a.. 4 0.4

o 0.0
o 600
Temperature (oC)
FIG. 12.19 Blueschist P-T-t trajectories from various classic areas (based on Ernst, 1988; Perchuk, 1989; Spear,
1993). The boxed areas shown on the diagram are the metamorphic facies designated in Fig. 2.1. Al 2 SiO s triple-
point and univariant curves are also shown. Arg ~ Cal transformation (after Johannes & Puhan, 1971;
Crawford & Hoersch, 1972).

FIG. 12.20 Lawsonite pseudomorphs in blueschist. Ile de Groix, Brittany, France. The coin is 24 mm in diame-

Deciphering polydeformed and polymetamorphosed rocks

situation of continental convergence, with

crustal thickening followed by rapid exhuma-
tion through a combination of erosion and
tectonic processes.
Some of the key reaction microstructures
that help to constrain the P-T evolution of the
'western Alpine type' trajectory include Czo +
Pg pseudomorphs after lawsonite (Fig. 12.20)
as evidence for increasing temperature, decreas-
ing pressure or a combination of both (e.g.
Forbes et ai., 1984). The development of reac-
tion rims or epitaxial overgrowth of actinolite
on glaucophane (or crossite) (Fig. 12.21) is
used as evidence of destabilisation of glauco-
phane during uplift, and transition into the
greenschist facies by the reaction GIn + Ep :t
CaC0 3 -7 Act + ChI + Ab (Kryza et ai., 1990).
FIG. 12.21 Actinolite rimming glaucophane: an exam- Holland & Ray (1985) describe blueschists
ple of a glaucophane-bearing assemblage being over- from the Tauern Window, Austria, where
printed by greenschist facies metamorphism from the
Kaczawa Complex, Sudetes, Poland (modified after petrographic studies have revealed that the
Kryza et al., 1990). Scale = 0.1 mm. Krs, Kaersutite (a assemblage Cros + Pg + Ep became unstable
brown amphibole with high Ti, typical of igneous during decompression and reacted to give the
amphiboles in alkali basic rocks); Gin, glaucophane; greenschist facies assemblage ChI + Ab + Act +
Act, actinolite; Spn, sphene; Chi, chlorite; Ab, albite;
Cal, calcite. Mag + TIc.
A second type of P-T trajectory displayed by
blueschist facies rocks is seen in the Franciscan
P- T-t trajectory of blueschist facies rocks has a complex of California, USA. In this case (the
'clockwise' sense. In the western Alps, the path 'Franciscan type' of Ernst, 1988) the retrograde
described above is particularly common (e.g. P-T path seems almost to retrace the same path
Chopin, 1984; Massone & Chopin, 1989), and as the prograde trajectory, and gives an overall
Ernst (1988) refers to this P-T-t trajectory as 'hairpin' trajectory (Fig. 12.19): that is, cooling
the 'western Alpine type'. Blueschists of the accompanies decompression, rather than rapid
Sanbagawa belt, Japan (e.g. Takasu, 1989; near-isothermal decompression as in the 'west-
Banno & Sakai, 1989; Otsuki & Banno, 1990), ern Alpine type' trajectory. The evidence for
the Seward Peninsula, Alaska (e.g. Forbes et significant cooling accompanying decompres-
ai., 1984), New Caledonia (e.g. Brothers, 1985; sion during uplift comes from the fact that
Brothers & Yokoyama, 1982), and many other many rocks still contain aragonite (high-pres-
blueschist terrains, also show this type of sure polymorph of CaC0 3 ). The fact that arag-
evolution. Ernst (1988) interprets the western onite is still present suggests that the Arg ~
Alpine type of trajectory with its near-isother- Cal transition is crossed at temperatures
mal decompression and greenschist or (perhaps < 100°C), at which the reaction kinet-
epidote-amphibolite facies overprint as being ics are sufficiently slow that aragonite does not
due to the nature of the convergence. In partic- react out (Carlson & Rosenfeld, 1981). There is
ular, he interprets the trajectory in terms of a still much debate regarding the tectonic
switch from subduction of oceanic crust to a process(es) by which the Franciscan blueschists


cate that the prograde path of an eclogite has

Grt been through the glaucophane stability field.
Pognante et al. (1987), studying part of the
Western Alps, similarly reported glaucophane
Gin inclusions within eclogitic garnets (Fig. 12.22).
It would be wrong to suggest that all eclogites
have followed this trajectory, but some at least
cBi° seem to have experienced rapid burial and a

clockwise P-T-t path comparable to western
Alpine type blueschists, but attaining higher
P-T at peak metamorphism. Other blueschist
FIG. 12.22 Schematic illustrations of glaucophane facies rocks that have passed up temperature
(Gin) and c1inozoisite (ezo) inclusions in eclogite facies into the low-temperature part of the eclogite
garnets (Grt) surrounded by omphacitic pyroxenes facies (see the review by Schielstedt, 1990),
(Omp); based on descriptions of Pognante et al. (1987),
for a suite of eclogites from the Western Alps.
include the high-pressure rocks on the Greek
islands of Sifnos and Syros (e.g. Ridley, 1984;
Schielstedt, 1986; Schielstedt & Matthews,
seemingly have an uplift trajectory that retraces 1987) and some of the high-P rocks from New
the prograde burial path. Whatever the case, the Caledonia (e.g. Brothers, 1985). As with many
preservation of aragonite seems to suggest that blueschist facies rocks, those rocks that have
compared to most other blueschist terrains, the entered the low-temperature eclogite facies
uplift path, although probably with a clockwise typically experience a greenschist facies over-
trajectory, is somewhat different. The presence print during exhumation.
of jadeiitic pyroxene in parts of the Franciscan Eclogites are known from various localities
complex, and in certain other blueschist facies throughout the Alps. Many of these eclogites
terrains indicates that the reaction Ab ~ Jd + have experienced an initial high-P (c. 8-30
Qtz has taken place and that the highest-pres- kbar), high-T (c. 400-450°C to 700-800°C)
sure part of the blueschist facies has been phase of eclogite facies metamorphism, subse-
attained. quently overprinted by blueschist facies assem-
blages, comprising abundant glaucophane (±
12.3.6 Eclogite facies P-T-t paths lawsonite), and in turn followed by a low-pres-
sure greenschist facies retrograde event.
Although the details of uplift trajectories for Although the prograde part of the trajectory is
eclogites are moderately well known due to not well constrained, Droop et al. (1990)
retrograde reaction assemblages and report that some rocks of the 'Eclogitic
microstructures, the prograde path of such Micaschist Complex' of the Sesia Zone
rocks is usually poorly constrained. Eclogites preserve as inclusions in eclogite facies miner-
such as those of the Najac-Carmaux thrust als, an earlier foliation defined by glaucophane
unit, Massif Central, France (Burg et aI., 1989), and epidote needles, indicating that the
have garnets that preserve inclusions of blue- prograde path passed through the glaucophane
green calcic clinoamphibole, zoned from glau- + epidote stability field. However, a 'clockwise'
cophane cores to barroisite rims. In some trajectory is not the only possibility for passing
instances, pure glaucophane inclusions occur in from early blueschist conditions to peak meta-
garnet cores (e.g. Barnicoat & Fry, 1989), but morphic eclogite facies conditions. For rocks
this is usually the only clear evidence to indi- from the Pie monte zone, western Alps,

Deciphering polydeformed and polymetamorphosed rocks

in the host mineral around the inclusion are

GPa believed to result from the stresses caused by
volume increase associated with the transfor-
mation coesite ---7 quartz. Apart from the fact
that ultra high-P conditions have been
obtained, the precise nature of the P- T-t
trajectory is not always well constrained, but
the evidence, based on included mineral
assemblages, geothermobarometry and retro-
grade assemblages, seems to suggest steep
slopes for both prograde and retrograde paths
(e.g. Zhang et ai., 1995).
There is no doubt that eclogite facies meta-
morphism occurs at deep crustal levels, and
Temperature (OC)
rocks preserving such metamorphic assem-
blages are generally considered to have been
FIG. 12.23 A schematic illustration of an eclogite rapidly exhumed. However, in most cases the
facies "hairpin' P-T-t trajectory (style of trajectory
based on Barnicoat & Fry, 1989).
available evidence is insufficient tightly to
constrain the full P-T evolution. Because of
this, it is usually difficult to say with any
certainty whether an individual eclogite has
Barnicoat & Fry (1989) presented evidence for experienced a 'clockwise' or 'anti-clockwise'
a tight 'hairpin' trajectory, that they considered trajectory. Indeed, it seems that a key feature of
to be 'anticlockwise' in character. many eclogites may be deep burial, followed by
Some eclogite terrains show evidence of rapid exhumation, to define what may best be
having experienced ultra high pressure meta- termed a 'hairpin' trajectory (Fig. 12.23).
morphism at moderate or high temperatures.
As well as theromobarometric studies, one of
12.4 Final comments
the key lines of evidence for ultra high-P
conditions (~ 20-25 kbar) comes from the The key to deciphering polydeformed and poly-
preservation of coesite (a high-P polymorph metamorphosed rocks lies in careful observa-
of Si0 2 ), or quartz pseudomorphs after tion, both in the field and during petrographic
coesite, as inclusions in garnet or jadeiitic work, whether by standard transmitted light
clinopyroxene. The preservation of coesite or microscopy or by more advanced techniques.
quartz pseudomorphs after coesite has now Throughout this chapter it has been emphasised
been recognised in a number of ultra high P that interpretation of the P-T-t evolution of a
eclogite facies terrains, including the Western metamorphic rock depends on the correct iden-
Alps (Chopin, 1984) and the Dabie Shan and tification of equilibrium assemblages and reac-
Su-Lu terrains, China (e.g. Okay, 1995; tion sequences, based on detailed observation.
Zhang et ai., 1995). The recognition of pre- The detailed petrogenetic grids that now exist
existing coesite is often revealed by radial allow P- T conditions to be constrained reason-
fractures emanating from an inclusion ably well if certain key minerals and assem-
comprising minute radiating quartz crystals blages have been identified. When coupled with
and perhaps a small core of remnant coesite other approaches to geothermometry and
(Fig. 7.2). The diagnostic radiating fractures geobarometry, as outlined in Section 12.3.1,

Final comments

and with additional geochronological informa- constrain. In most cases the only thing to go on
tion, it is often possible to constrain P-T-t is whatever minerals may be preserved as
paths with a fair degree of detail. When inte- included phases in the core of peak metamor-
grated with structural information, it is possi- phic minerals. In most cases, the minerals
ble to gain a more complete understanding of preserved may be phases such as quartz, which
tectonic processes and the tectono metamorphic are stable over a wide range of P-T conditions,
evolution of a particular portion of the Earth's and even if a more useful phase such as glauco-
crust at some time in the geological past. phane or chloritoid is preserved, the lack of a
Despite the potential for major insight into complete precursor assemblage makes it diffi-
crustal processes based on a firm understand- cult to constrain the P-T path in anything other
ing of metamorphic assemblages and their than general terms. Evidence for the exhuma-
microstructural interrelationships, there are a tion or retrograde P-T-t trajectory may be
few points that should be noted. First, the better preserved, due to incomplete reactions,
cause of a change in equilibrium assemblage fluid inclusion studies, geothermobarometry
may not necessarily be changing P-T condi- and metamorphic cooling ages.
tions, but could be changing fluid composition It is always important to give clear considera-
or some other factor. As previously stated, it tion to the errors on any part of the P- T-t
should be realised that many petrogenetic grids trajectory, and to consider other possible inter-
are constructed on the basis of the fluid being pretations that may exist. It is important to
pure HzO and that P f = PHzO . During the meta- avoid the temptation of simply joining up a
morphism of carbonate and calc-silicate rocks, continuous line between various constrained
it is well established that the fluid typically points in a rock's P-T evolution in order to
varies between pure HzO and pure CO 2, and obtain a P-T trajectory without giving proper
that the exact composition may vary as reac- consideration to the processes involved and the
tions proceed. For this reason, it is crucial to time interval between given points that have
appreciate the influence of different fluid been defined. Vernon (1996) emphasises this
compositions in controlling equilibrium assem- problem, and also discusses the problem of
blages at particular P-T conditions. For exam- inferring the direction of a P-T path from the
ple, Cartwright & Buick (1995) describe the crossing of a single reaction curve in P-T space.
formation of wollastonite-bearing layers in Although the particular reaction may have been
granulite facies marbles due to the infiltration identified with a high degree of confidence,
of water-rich fluids at peak metamorphic there is a 180° range of directions for the P-T
temperatures of around 700°C, rather than due path when crossing a constant slope (straight
to changing P-T conditions. The wollastonite- line) reaction curve, and when the reaction
forming reaction is Cal + Qtz ~ Wo + COz. curve is convex outwards in the direction in
The fact that adjacent marbles contain up to which the reaction is proceeding, there is an
11 % modal quartz, in equilibrium with calcite, even greater range of possible directions.
provides the evidence that some layers have Because of this range of possible interpretations
been infiltrated by fluid and others have not, when crossing individual reaction curves in P-T
and that it is influx of aqueous fluid that has space, it is always advisable (where possible) to
caused the reaction rather than changing P-T constrain the path to a higher degree of confi-
conditions. dence by identifying several reaction curves that
A second point to bear in mind is that the have been crossed and by using other
prograde path of a P-T-t trajectory for a partic- approaches to geothermobarometry to quantify
ular rock is generally the most difficult to particular metamorphic events. As Spear (1993)

Deciphering polydeformed and polymetamorphosed rocks

states, erroneous paths are worse than useless Bennett, D.G. & Barker, A.J. (1992) High salinity
because they are misleading and will certainly fluids: the result of retrograde metamorphism in
thrust zones. Geochimica et Cosmochimica Acta,
create more confusion than clarity, so be 56,81-95.
warned, and take care. Bertrand, P., Ouzegane, K.H. & Kienast, J.R. (1992)
P-T-X relationships in the Precambrian Al-Mg-rich
granulites from In Ouzzal, Hoggar, Algeria. Journal
of Metamorphic Geology, 10, 17-3l.
References Bohlen, S.R. (1987) Pressure-temperature-time paths
and a tectonic model for the evolution of granulites.
Allen, T. & Chamberlain, c.P. (1991) Metamorphic Journal of Geology, 95, 617-632.
evidence for an inverted crustal section, with Bohlen, S.R., Wall, V.J. & Boettcher, A.L. (1983)
constraints on the Main Karakorum Thrust, Experimental investigations and geological applica-
Baltistan, northern Pakistan. Journal of Meta- tions of equilibria in the system FeO-
morphic Geology, 9, 403-418. Ti02-AI20rSiOrH20. American Mineral-ogist,
Anderson, M.W., Barker, A.J., Bennett, D.G. & 68, 1049-1058.
Dallmeyer, R.D. (1992) A tectonic model for Brothers, R.N. (1985) Regional mid-Tertiary
Scandian terrane accretion in the northern blueschist-eclogite metamorphism in northern New
Scandinavian Caledonides. Journal of the Geo- Caledonia. Geologie France, 1, 37-44.
logical Society, 149, 727-74l. Brothers, R.N. & Yokoyama, K. (1982) Comparison
Armstrong, T.R., Tracy, R.J. & Hames, W.E. (1992) of the high-pressure schist belts of New Caledonia
Contrasting styles of Taconian, Eastern Acadian and Sanbagawa, Japan. Contributions to
and Western Acadian metamorphism, central and Mineralogy and Petrology, 79, 219-229.
western New England. Journal of Metamorphic Burg. J.P., Delor, GP., Leyreloup, A.F. & Romney, F.
Geology, 10,415-426. (1989) Inverted metamorphic zonation and Variscan
Banno, S. & Sakai, C. (1989) Geology and metamor- thrust tectonics in the Rouergue area (Massif Central,
phic evolution of the Sanbagawa metamorphic belt, France): P-T-t record from mineral to regional scale,
Japan, in Evolution of Metamorphic Belts (eds J.S. in Evolution of metamorphic belts (eds J.S. Daly,
Daly, R.A. Cliff & B.W.D. Yardley). Geological R.A. Cliff & B.W.D. Yardley). Geological Society
Society Special Publication No. 43, 519-532. Special Publication No. 43, 423-439.
Barker, A.J. (1989) Metamorphic evolution of the Burton, K.W., Boyle, A.P., Kirk, W.L. & Mason, R.
Caledonian nappes of north central Scandinavia, (1989) Pressure, temperature and structural evolu-
in The Caledonide Geology of Scandinavia (ed. tion of the Sulitjelma fold-nappe, central
R.A. Gayer). Graham & Trotman, London, Scandinavian Caledonides, in Evolution of meta-
193-204. morphic belts (eds J.S. Daly, R.A. Cliff & B.W.D.
Barker, A.J. (1995) Diachronous fluid release and vein- Yardley). Geological Society Special Publication No.
ing associated with out-of-sequence thrusts in the 43,391-41l.
north Scandinavian Caledonides. Australian Journal Carlson, W.D. & Rosenfeld, J.L. (1981) Optical deter-
of Earth Sciences, 42, 311-320. mination of topotactic aragonite-calcite growth
Barker, A.J. & Anderson, M.W. (1989) The kinetics: Metamorphic implications. Journal of
Caledonian structural-metamorphic evolution of Geology, 89, 615-638.
south Troms, Norway, in Evolution of Cartwright, 1. & Buick, LS. (1995) Formation of
Metamorphic Belts (eds J.S. Daly, R.A. Cliff & wollastonite-bearing marbles during late regional
B.W.D. Yardley). Geological Society Special metamorphic channelled fluid flow in the Upper
Publication No. 43, 385-390. Calcsilicate Unit of the Reynolds Range Group,
Barnicoat, A.C. & Fry, N. (1989) Eoalpine high-pres- central Australia. Journal of Metamorphic Geology,
sure metamorphism in the Piemonte zone of the 13,397-417.
Alps: south-west Switzerland and north-west Italy, Chatterjee, N.D. (1972) The upper stability limit of the
in Evolution of metamorphic belts (eds J.S. Daly, assemblage paragonite + quartz and its natural
R.A. Cliff & B.W.D. Yardley). Geological Society occurrence. Contributions to Mineralogy and
Special Publication No. 43, 539-544. Petrology, 34, 288-303.
Barrow, G. (1893) On an intrusion of Chatterjee, N.D. & Johannes, W. (1974) Thermal
muscovite-biotite gneiss in the southeastern stability and standard thermodynamic properties of
Highlands of Scotland, and its accompanying meta- synthetic 2M 1-muscovite, KAI 2 [AlSiPlo(OH)2].
morphism. Quarterly Journal of the Geological Contributions to Mineralogy and Petrology, 48,
Society of London, 49, 330-358. 89-114.


Chopin, e. (1984) Coesite and pure pyrope in high morphic belts (eds J.S. Daly, R.A. Cliff & B.W.D.
grade pelitic blueschists of the Western Alps. Yardley). Geological Society Special Publication No.
Journal of Petrology, 22, 628-650. 43,1-44.
Clarke, G.L. & Powell, R. (1991) Decompressional Faure, G. (1986) Principles of isotope geology. John
coronas and symplectites in granulites of the Wiley, New York, 464 pp.
Musgrave Complex, central Australia. Journal of Forbes, R.B., Evans, B.W. & Leyreloup, A.F. (1984)
Metamorphic Geology, 9, 441- 450. Regional progressive high-pressure metamorphism,
Crawford, W.A. & Hoersch, A.L. (1972) Calcite-arag- Seward Peninsula, Alaska. Journal of Metamorphic
onite equilibrium from 50 D -100 D e. American Geology, 2, 43-54.
Mineralogist, 57, 995-998. Frost, B.R. & Chacko, T. (1989) The granulite uncer-
De Yoreo, J.J., Lux, D.R. & Guidotti, C.V. (1989) The tainty principle: limitations on thermo barometry in
role of crustal anatexis and magma migration in the granulites. Journal of Geology, 97, 435-450.
thermal evolution of regions of thickened continen- Ghent, E.D. (1976) Plagioclase - garnet - Al2SiO s -
tal crust, in Evolution of metamorphic belts (eds J.S. quartz: a potential geobarometer- geothermometer.
Daly, R.A. Cliff & B.W.D. Yardley). Geological American Mineralogist, 61, 710-714.
Society Special Publication No. 43, 187-202. Gillott, J.E. (1955) Metamorphism of the Manx Slates.
Diella, V., Spalla, M.l. & Tunesi, A. (1992) Geological Magazine, 92,141-154.
Contrasting thermochemical evolutions in the Goscombe, B. (1992) Silica-undersaturated sapphirine,
Southalpine metamorphic basement of the Orobic spinel and kornerupine granulite facies rocks, NE
Alps (Central Alps, Italy). Journal of Metamorphic Strangways Range, Central Australia. Journal of
Geology, 10,203-219. Metamorphic Geology, 10, 181-20l.
Dodson, M.H. (1973) Closure temperatures in cooling Hames, W.E., Tracy, R.J., & Bodnar, R.J. (1989)
geochronological and petrological systems. Post-metamorphic unroofing history deduced
Contributions to Mineralogy and Petrology, 40, from petrology, fluid inclusions, thermochronom-
259-274. etry, and thermal modelling: an example from
Droop, G.T.R. (1989) Reaction history of southwestern New England. Geology, 17,
garnet-sapphirine granulites and conditions of 727-730.
Archaean high-pressure granulite-facies metamor- Hames, W.E., Tracy, R.J., Radcliffe, N.M. & Sutter,
phism in the Central Limpopo Mobile Belt, J.F. (1991) Petrologic, structural and geochrono-
Zimbabwe. Journal of Metamorphic Geology, 7, logic characteristics of the Acadian metamorphic
383-403. overprint of the Taconian zone in part of southwest-
Droop, G.T.R., Lombardo, B. & Pognante, U. (1990) ern New England. American Journal of Science,
Formation and distribution of eclogite facies rocks 291, 887-913.
in the Alps, in Eclogite facies rocks (ed. D.A. Harley, S.L. (1989) The origins of granulites: a meta-
Carswell). Blackie, Glasgow, 225-259. morphic perspective. Geological Magazine, 126,
Ellis, D.J. (1987) Origin and evolution of granulites in 215-247.
normal and thickened crusts. Geology, 15, 167-170. Haugerud, R.A. & Zen, E.-an (1989) Metamorphic
England, P.e. & Richardson, S.W. (1977) The influ- path studies - a critique and prospectus, in
ence of erosion upon the mineral facies of rocks Advances in Geochemistry (D.S. Korzhinskiy) (ed.
from different metamorphic environments. Journal L.L. Perchuk). Springer- Verlag, New York.
of the Geological Society, 134,201-213. Hirschberg, A. & Winkler, H.e.F. (1968) Stability
England, P.c. & Thompson, A.B. (1984) relations between cordierite, chlorite, and alman-
Pressure-tern perature-time paths of regional meta- dine during metamorphism. Contributions to
morphism, Part I: Heat transfer during the evolution Mineralogy and Petrology, 18, 17- 42.
of regions of thickened continental crust. Journal of Hisada, K. & Miyano, T. (1996) Petrology and
Petrology, 25, 894-928. microthermometry of aluminous rocks in the
Ernst, W.G. (1988) Tectonic history of subduction Botswanan Limpopo Central Zone: evidence for
zones inferred fron retrograde blueschist P-T paths. isothermal decompression and isobaric cooling.
Geology, 16, 1081-1084. Journal of Metamorphic Geology, 14, 183-197.
Essene, E.J. (1982) Geologic thermometry and barome- Holdaway, M.J. (1971) Stability of andalusite and the
try, in Characterization of metamorphism through aluminium silicate phase diagram. American Journal
mineral equilibria (ed. J.M. Ferry). Mineralogical of Science, 271, 97-131.
Society of America, Reviews in Mineralogy No. 10, Holland, T.J.B. & Ray, N.J. (1985) Glaucophane and
153-206. pyroxene breakdown reactions in the Pennine units
Essene, E.J. (1989) The current status of thermobarom- of the Eastern Alps. Journal of Metamorphic
etry in metamorphic rocks, in Evolution of meta- Geology, 3, 417-438.

Deciphering pol,deformed and pol,metamorphosed rocks

Hoschek, G. (1969) The stability of staurolite and the southern Omineca belt, British Columbia and
chloritoid and their significance in metamorphism Washington. Tectonics, 7, 181-212.
of pelitic rocks. Contributions to Mineralogy and Pecher, A. (1989) The metamorphism in the Central
Petrology, 22, 208-232. Himalaya. Journal of Metamorphic Geology, 7,
Hsu, L.e. (1968) Selected phase relationships in the 31-41.
system Al-Mn-Fe-Si-O-H: a model for garnet equi- Perchuk, L.L. (1989) P-T-fluid regimes of metamor-
libria. Journal of Petrology, 9, 40-83. phism and related magmatism with specific refer-
Inger, S. & Harris, N.B.W. (1992) Tectonothermal ence to the granulite facies Sharyzhalgay complex of
evolution of the High Himalayan Crystalline Lake Baikal, in Evolution of metamorphic belts (eds
Sequence, Langtang Valley, northern Nepal. Journal J.S. Daly, R.A. Cliff & B.W.D. Yardley). Geological
of Metamorphic Geology, 10, 439-452. Society Special Publication No. 43, 275-291.
Johannes, W. & Puhan, D. (1971) The calcite-arago- Perchuk, L.L. & Lavrent'eva, I.V. (1983) Experimental
nite transition reinvestigated. Contributions to investigation of exchange equilibria in the system
Mineralogy and Petrology, 31, 28-38. cordierite-garnet-biotite, in Kinetics and equilib-
Kihle, J. & Bucher-Nurminen, K. (1992) Ortho- rium in mineral reactions (ed. S.K. Saxena).
pyroxene-sillimanite-sapphirine granulites from the Springer-Verlag, New York, 199-239.
Bamble granulite terrane, southern Norway. Journal Pognante, U., Talirico, F., Rastelli, N. & Ferranti, N.
of Metamorphic Geology, 10, 671-683. (1987) High pressure metamorphism in the nappes
Kisch, H.J. (1987) Correlation between indicators of very of the Vall dell'Orco traverse (Western Alps colli-
low-grade metamorphism, in Low temperature meta- sional belt). Journal of Metamorphic Geology, 5,
morphism (ed. M. Frey). Blackie, Glasgow, 227-304. 397-414.
Kryza, R., Muszynski, A. & Vielzeuf, D. (1990) Rao, B.B. & Johannes, W. (1979) Further data on the
Glaucophane-bearing assemblage overprinted by stability of staurolite + quartz and related assem-
greenschist-facies metamorphism in the Variscan blages. Neues Jahrbuch fur Mineralogie, 1979,
Kaczawa complex, Sudetes, Poland. Journal of 437-447.
Metamorphic Geology, 8, 345-355. Richardson, S.W., Gilbert, M.e. & Bell, P.M.
Massonne, H.-J. & Chopin, e. (1989) P-T history of (1969) Experimental determination of the
the Gran Paradiso (western Alps) metagranites kyanite-andalusite and andalusite-sillimanite equi-
based on phengite barometry, in Evolution of meta- libria; the aluminium silicate triple point. American
morphic belts (eds ].S. Daly, R.A. Cliff & B.W.D. Journal of Science, 267, 259-272.
Yardley). Geological Society Special Publication No. Ridley, J. (1984) Evidence of a temperature-dependent
43,545-549. 'blueschist' to 'eclogite' transformation in high-pres-
Mezger, K., Rawnsley, e. Bohlen, S. & Hanson, G. sure metamorphism of metabasic rocks. Journal of
(1991) U-Pb garnet, sphene, monazite and rutile Petrology, 25, 852-870.
ages: Implications for the duration of high grade Ridley, J. (1989) Vertical movement in orogenic belts
metamorphism and cooling histories, Adirondack and the timing of metamorphism relative to defor-
Mts., New York. Journal of Geology, 99, 415-428. mation, in Evolution of metamorphic belts (eds ].S.
Naggar, M.H. & Atherton, M.P. (1970) The composi- Daly, R.A. Cliff & B.W.D. Yardley). Geological
tion and metamorphic history of some aluminium Society Special Publication No. 43, 103-115.
silicate-bearing rocks from the aureoles of the Schielstedt, M. (1986) Eclogite-blueschist relationships
Donegal granites. Journal of Petrology, 11, as evidenced by mineral equilibria in high-pressure
549-589. metabasic rocks of Sifnos (Cycladic Islands), Greece.
Okay, A.I. (1995) Paragonite eclogites from Dabie Journal of Petrology, 27,1437-1459.
Shan, China: re-equilibration during exhumation? Schielstedt, M. (1990) Occurrence and stability condi-
Journal of Metamorphic Geology, 13, 449-460. tions of low-temperature eclogites, in Eclogite facie~
Otsuki, M. & Banno, S. (1990) Prograde and retro- rocks (ed. D.A. Carswell). Blackie, Glasgow,
grade metamorphism of hematite-bearing basic 160-179.
schists in the Sanbagawa belt in central Shikoku. Schielstedt, M. & Matthews, A. (1987)
Journal of Metamorphic Geology, 8, 425-439. Transformation of blueschist facies rocks as a
Oxburgh, E.R. & Turcotte, D.L. (1974) Thermal consequence of fluid infiltration, Sifnos (Cyclades),
gradients and regional metamorphism in overthrust Greece. Contributions to Mineralogy and Petrology,
terrains with special reference to the eastern Alps. 97,237-250.
Schweiz Mineralogische und Petrographishe Schumacher, ].e., Schumacher, R. & Robinson, P.
Mitteilungen, 54, 641-662. (1989) Acadian metamorphism in central
Parrish, R.R., Carr, S.D. & Parkinson, D.L. (1988) Massachusetts and south-western New Hampshire:
Eocene extensional tectonics and geochronology of evidence for contrasting P-T trajectories, in


Evolution of metamorphic belts (eds j.S. Daly, R.A. morphic belt, Japan, in Evolution of metamorphic
Cliff & B.W.D. Yardley). Geological Society Special belts (eds j.S. Daly, R.A. Cliff & B.W.D. Yardley).
Publication No. 43, 453-460. Geological Society Special Publication No. 43,
Selverstone, j. (1988) Evidence for east-west crustal 533-538.
extension in the Eastern Alps: implications for the Thompson, A.B. & England, P.c. (1984)
unroofing history of the Tauern Window. Pressure-temperature-time paths of regional meta-
Tectonics, 7, 87-105. morphism II. Their inference and interpretation
Selverstone, j. & Spear, F.S. (1985) Metamorphic P-T using mineral assemblages in metamorphic rocks.
paths from pelitic schists and greenstones from the Journal of Petrology, 25,929-955.
Southwest Tauern Window, Eastern Alps. Journal Thompson, A.B. & Tracy, R.j. (1979) Model systems
of Metamorphic Geology, 3, 439-465. for anatexis of pelitic rocks. II. Facies series melting
Selverstone, ]., Spear, F.S., Franz, G. & Morteani, G. and reactions in the system CaO-KAl0 2-NaAl0 2-
(1984) High-pressure metamorphism in the AI 2 0 3 -Si0 2 -H zO. Contributions to Mineralogy
Southwest Tauern Window, Austria: P-T paths and Petrology, 70, 429-438.
from hornblende-kyanite-staurolite schists. Journal Vernon, R.H. (1996) Problems with inferring P-T-t
of Petrology, 25, 501-531. paths in low granulite facies rocks. Journal of
Spear, F.S. (1986) P-T PATH: a FORTRAN program Metamorphic Geology, 14, 143-153.
to calculate pressure temperature paths from zoned Waters, D.j. (1989) Metamorphic evidence for the
metamorphic garnets. Computers in Geoscience, 12, heating and cooling path of Namaqualand gran-
247-266. ulites, in Evolution of metamorphic belts (eds j.S.
Spear, F.S. (1993) Metamorphic phase equilibria Daly, R.A. Cliff & B.W.D. Yardley). Geological
and pressure-temperature-time paths. Mineral- Society Special Publication No. 43, 357-363.
ogical Society of America Monograph, Mineral- Winslow, D.M., Bodnar, R.j. & Tracy, R.j. (1994)
ogical Society of America, Washington, DC, 799 Fluid inclusion evidence for an anticlockwise meta-
pp. morphic P-T path in central Massachusetts. Journal
Spear, F.S. & Cheney, j.T. (1989) A petrogenetic of Metamorphic Geology, 12,361-371.
grid for pelitic schists in the system Zeitler, P.K. (1989) The geochronology of metamor-
SiOz-Alz03-FeO-MgO-KzO-HzO. Contributions phic processes, in Evolution of metamorphic belts
to Mineralogy and Petrology, 101, 149-164. (eds j.S. Daly, R.A. Cliff & B.W.D. Yardley).
Spear, F.S., Selverstone,]., Hickmott, D., Crowley, P. & Geological Society Special Publication No. 43,
Hodges, K.V. (1984) P-T paths from garnet zoning: 131-147.
a new technique for deciphering tectonic processes in Zhang, R-Y., Hirajima, T., Banno, S., Bolin Cong &
crystalline terranes. Geology, 12, 87-90. Liou, j.G. (1995) Petrology of ultrahigh-pressure
Takasu, A. (1989) P-T histories of peridotite and rocks from the southern Su-Lu region, eastern China.
amphibolite tectonic blocks in the Sanbagawa meta- Journal of Metamorphic Geology, 13, 659-675.

Appendix I:

Mineral abbreviations Di diopside

Dol dolomite
Mineral abbreviations used in the text and
En enstatite
diagrams are those of Kretz (1983), with the
Ep epidote
addition of those marked with ,~ (after Barker,
Fo forsterite
1990, and this edition), and those with + (after +Ged gedrite
Bucher & Frey, 1994).
GIn glaucophane
Ab albite Gr graphite
Act actinolite Grs grossularite
*Ads andesine Grt garnet
AIm almandine +Hc hercynite
An anorthite Hem haematite
And andalusite Hbl hornblende
Ank ankerite Hul heulandite
AnI analcite *Hy hypersthene
Ann annite III illite
Arg aragonite 11m ilmenite
Ath anthophyllite Jd jadeite
*Bar barroisite Kfs K-feldspar
+Brc brucite KIn kaolinite
Bt biotite +Krn kornerupine
Cal calcite +Krs kaersutite
ChI chlorite Ky kyanite
Cld chloritoid Lmt laumontite
+Cp Fe/Mg-carpholite Lws lawsonite
Cpx Ca -clinopyroxene Mag magnetite
Crd cordierite Mc micro cline
'Cros crossite Mgs magnesite
Crs cristo bali te Mnt montmorillonite
+Crn corundum Mrg margarite
Cs coesite Ms muscovite
Czo clinozoisite +Mul mullite
+Cum cummingtonite 01 olivine

Appendix I

"Olg oligoclase Additional abbreviations

Omp omphacite
Opx orthopyroxene CSD crystal size distribution
Or orthoclase CPO crystallographic preferred orientation
Osm osumilite ecc extensional crenulation cleavage
Pg paragonite EF elastico-frictional
+Phe phengite ((0 2 ) oxygen fugacity
PhI phlogopite P pressure
PI plagioclase PI lithostatic pressure
Pmp pumpellyite Pf fluid pressure
+Prh prehnite PPL plane-polarised light
Prl pyrophyllite QP quasi-plastic
Prp pyrope SEM scanning electron microscopy
Py pyrite T temperature
Pyx pyroxene t time
Qtz quartz TEM transmission electron microscopy
Rt rutile Th homogenisation temperature (used
Sa sanidine with respect to fluid inclusions)
+Scp scapolite T t trapping temperature (used with
+Ser sericite respect to fluid inclusions)
Sil sillimanite 11; chemical potential of i
*Smc smectite )lm micrometres (microns)
Spl spinel var. variety
Spn sphene XC0 2 proportion of CO2 in fluid in relation
+Spr sapphirine to H 2 0 (where XH2 0 + XC0 2 = 1.0)
Sps spessartine XH 2 0 proportion of H2o in fluid in relation
Srp serpentine to CO 2 (where XH2 0 + XC0 2 = 1.0)
St staurolite XRD X-ray diffraction
Stp stilpnomelane XPL cross-polarised light
Tlc talc 0"3 minimum compressive stress
Tr tremolite 0"1 maximum compressive stress
+Trd tridymite !J.(H enthalpy of formation
+Ves vesuvianite (idocrase) !J.T change in temperature
Wo wollastonite > greater than
Wr wairakiite ~ greater than or equal to
Zeo zeolites < less than
Zo zoisite ~ less than or equal to
Zm ZIrcon '# not equal to

Appendix II:

Absorption Amphibolite
The process by which additional units are A metamorphosed basic igneous rock with a
chemically bonded into the structure of a grow- mineral assemblage comprised largely of
ing crystal to form an integral part of it (modi- amphibole and plagioclase, usually with quartz
fied after Spry, 1969). and epidote (Fig. 4.S(a)).

The process of partial melting of high-grade
A term used to describe needle-like crystals
metamorphic rocks in the presence of H 20.
(Figs S.12(a) & (b); Plate 8(e)).
(This process produces granitoid melts and
typically operates in the middle to lower crust
Activation energy during orogeny.)
That energy required before a given reaction or
diffusive process can proceed. Anchimetamorphism
Sub-greenschist facies, very low grade meta-
Activity morphism. (The 'limits' of anchimetamorphic
(Symbol = a). One of several ways to describe conditions have been variably defined based on
the behaviour of given component in solution illite crystallinity.)
(solid or fluid). The relationship between the
activity of a given component (a j ), and the
concentration of that component (Xj ) is given A term used to describe metamorphic crystals
(especially porphyroblasts) with irregular form
by a j =a j IXj, where a j is the activity coefficient.
For gases a j = (If;°, where f; is the fugacity (or
thermodynamic pressure) of gas i, and f;0 is the Annealing
fugacity of i in a standard state. For an ideal A recovery process in deformed rocks while
solution, f; = a j = Xi' However, in real solu- still at high temperature. It involves static
tions, ai (and hence f;) may differ greatly from recrystallisation, and the formation of new
Xj (after Vernon, 1976). strain-free grains.

Adsorption Antiperthite
The loose attachment (or bonding) of matrix or A feldspar intergrowth comprising K-feldspar
fluid phases to the surface of a growing crystal. inclusions enclosed within plagioclase (the
(Fig. 6.3). converse of perthite).


Antitaxial vein Burial metamorphism

Vein with filling material grown from vem Low-grade metamorphism in response to ambi-
centre towards the vein walls (Ramsay & ent geothermal gradient generated in thick
Huber, 1983) (Fig. 11.8(b)). basinal sequences.

Calc-silicate rock
Atoll structure
A rock with a chemistry dominated by calcium
A structure common to granulites consisting of
and silica, consisting of hydrous or anhydrous
a core of one mineral entirely surrounded by a
calc-silicate minerals such as tremolite, diop-
rim of another mineral (e.g. garnet forming a
side and grossular. (Carbonate minerals are
core entirely surrounded by plagioclase) (Plate
also commonly present.)
2(a) ).

Augen gneiss A deformation mechanism in which crystal
A gneissose rock with abundant augen structure remains undistorted, but grains or
(German: eyes) represented by porphyroclasts groups of grains become cracked and the frag-
(typically K-feldspar) enveloped by the folia- ments may exhibit frictional sliding with
tion (Fig. 4.3). respect to one another (Rutter, 1986) (Fig.
A term used for mylonites with extensive Cataclasite
mineral growth synchronous with shearing or A cohesive largely unfoliated fault-rock
else widespread static recrystallisation immedi- containing angular clasts of variable size in a
ately after deformation. fine-grained matrix of similar composition.

Chemical potential
One of several expressions for the behaviour of
A metamorphosed mafic rock indicative of
a component in solution. The chemical poten-
high-P/low- T subduction-related metamor-
tial of a component i in an ideal solution is
phism. It contains large quantities of sodic
given by )1i = GO i + RTlnXi where GO i is a
(blue) amphibole (glaucophanelcrossite), and
constant (the partial molar Gibbs free energy of
has a pronounced schistosity.
pure i in the standard state), R is the gas
constant, Xi is the molar concentration of i in
Bow-tie structure solution and T is temperature (based on
A term used to describe aggregates of elongate Vernon, 1976).
prismatic, and acicular crystals that are arranged
to give the appearance of a 'bow-tie'. It is Chemical softening
commonly exhibited by amphiboles in garben- Enhanced ductility in a deforming rock due to
schiefer that have grown in the foliation plane changes in the trace element content of a
under low differential stress (Fig. 5.13(b)). mineral (e.g. water weakening of quartz)
(modified from White et at., 1980).
Burgers vector
A vector defining the amount and direction of Chemical zoning
lattice displacement associated with an Regular or abrupt changes in mineral chemistry
intracrystalline dislocation (Fig. 8.5). from mineral core to rim.


Chlorite-mica stack of concentric zones of varying metamorphic

Interlayered chlorite and mica (typically phen- grade based on mineral assemblages.
gite or muscovite) to form a lens or barrel-
shaped phyllosilicate stack, often as a Continuous reaction
replacement of detrital biotite in sub-green- See Sliding reaction.
schist facies metasedimentary rocks (especially
slates) (Figs 7.7(a)-(c)). Core-and-mantle microstructure
A characteristic feature of many mylonitic and
Cleavage protomylonitic rocks, also referred to as
A sub-parallel set of closely spaced approxi- 'mortar structure'. It comprises a porphyroclast
mately planar surfaces produced during rock of a single crystal, mantled by a fine-grained
deformation. The rock preferentially splits recrystallised aggregate of the same mineral
along such surfaces due to the alignment of phase (Fig. 8.2( d)).
platy or elongate grains (usually phyllosilicate
minerals) within the rock (Fig. 4.8). Corona
A monomineralic or polymineralic rim totally
Cleavage dome surrounding a core of another mineral phase. It
An unusual feature involving arching or represents an arrested reaction between the
doming of matrix cleavage in areas immedi- core phase and other components, often during
ately adjacent to faces of certain euhedral retrogression of the core phase (Figs 12.16(a)
porphyroblasts. These domes (often comprising & 12.17(b); Plates 2(a) & (b)).
muscovite and graphite), are interpreted in
terms of physical displacement of insoluble Crack-seal vein
matrix phases ahead of a growing porphyrob- A vein built up by successive development of
last in a bulk hydrostatic stress field. They microcracks, followed by successive periods of
represent one of the few lines of evidence used cementation (Ramsay & Huber, 1983) (Fig.
as an indicator of porphyroblast growth by 11.4; Plates 8(a)-(d)).
displacement rather than replacement.
Closure temperature Continuous, usually slow deformation of a rock
Symbol = Te. The temperature at which (for a or individual crystal resulting from relatively
given mineral) the rate of increase of radiogenic low stress acting over a long period of time.
daughter product relative to parent becomes
constant for a given isotopic system (Table 12.1). Crenulation
Microfolding on a em-scale, best developed in
Component quartz-mica-dominated metasedimentary rocks
The fundamental chemical constituents of a with well developed primary lamination or
rock system (e.g. Si0 2 , Al 2 0 3 or K2 0) are earlier continuous tectonic foliation.
described as components.
Cross-hatched twinning
Contact aureole 'Cross-hatched twinning', 'tartan pattern' or
The zone of rocks surrounding a plutonic 'gridiron twinning' is one of the most charac-
intrusion which are thermally metamorphosed teristic features of microcline (Fig. 5.19(b)). It
in response to the intrusion. The contact results from the high-angle intersection of
aureole can usually be subdivided into a series multiple albite and pericline twins.


Crystalloblastic series plagioclase and pyroxene. Deformation twins

Based on original work by Harker (1939), the terminating within the crystal have tapered ends
crystalloblastic series (Table 5.1) is a miner- (Figs 5.18(b) & (c); Plates 3(a) & (b)).
alogical sequence, with those minerals at the
top having the greatest tendency towards euhe- Dehydration reaction
dral form. The sequence reflects decreasing A reaction liberating H 2 0 from the reactant(s).
surface energy, and minerals higher in the
sequence always tend to form euhedral faces Diagenesis
against minerals lower in the sequence. All physical, chemical and biological changes
experienced by sediments during lithification
Decarbonation reaction prior to metamorphism. The boundary
A reaction resulting m the liberation of CO2 between diagenesis and low-grade metamor-
from the reactant( s). phism is transitional and poorly defined.

Decussate structure Differential stress

A term used to describe interlocking, randomly Usually defined as the difference between maxi-
oriented, elongate, prismatic or sub-idioblastic mum ((II) and minimum ((13) compressive
crystals, generally of one species (modified stresses in a non-hydrostatic (lithostatic) stress
after Spry, 1969) (Fig. 5.16). regIme.

Deformation bands Diffusion

Distinct bands of deformation in crystals. The process by which atoms, molecules or ions
Optically they are more sharply defined than move from one position to another within a
undulose extinction, and represent higher fluid or solid, under the influence of a chemical
strain. They can terminate either at grain potential gradient. In solids this means the peri-
boundaries or inside grains (Fig. 8.13(a)). odic jumping of atoms from one site in the
structure to another (modified after Jensen,
Deformation lamellae 1965; Vernon, 1976).
Sets of very narrow planar or tapered lenticular
structures in quartz and many other silicates. Dihedral angle
They terminate inside grain boundaries, and The interfacial angle formed at the triple-junc-
show a variety of orientations, although they tion of a particular mineral phase, A, in contact
are frequently perpendicular to bands of undu- with two adjacent crystals of a different
lose extinction (after Vernon, 1976) (Fig. 8.10). mineral phase, B.

Deformation mechanism map Discontinuous reaction

A plot of stress versus homologous temperature A reaction involving the complete breakdown
defining fields over which particular deforma- of one mineral or minerals and formation of
tion mechanisms dominate during deformation another mineral or minerals in response to
of a particular material (e.g. quartz) of speci- changing P, T and or fluidlchemical conditions.
fied grain size (Fig. 8.7).
Deformation twins An incompatible association of mineral phases
Twins developed in crystals in response to defor- and/or a combination of textures and struc-
mation. They are especially common in calcite, tures incompatible with prevailing conditions.


Dislocations Epitaxial growth

Line defects in crystals, produced during The oriented growth of one mineral phase in
growth or deformation, and which are thermo- optical continuity with another due to struc-
dynamically unstable (Fig. 11.16). tural similarities between the two phases (e.g.
growth of secondary amphibole rims on earlier
Dislocation tangle amphiboles of different composition) (Fig.
During intracrystalline deformation, the inter- 12.21; Plate 4(a)).
section of different slip systems leads to entan-
glement of migrating dislocations. Such Equilibrium
dislocation tangles make further deformation That state of a rock system in which the phases
of the crystal increasingly difficult, and greatly present are in the most stable, low-energy
contribute to overall strain (work) hardening. arrangement, and in which all phases are
Dissolution compatible with the given P, T and fluid condi-
The process by which minerals are chemically
corroded and the constituent elements pass into
solution. Euhedral
A term used to describe metamorphic crystals
Dynamic recrystallisation (especially porphyroblasts) with well developed
Deformation-induced reworking of the grain crystal faces and form (synonymous with
sizes, shapes or orientations with little or no idioblastic) (Fig. S.9(a)).
chemical change (Poirier & Guillope, 1979).
Exothermic reaction
Eclogite A reaction that gives off heat as it proceeds. It
A dense, high-grade metamorphic rock of is the converse of endothermic reactions, which
mafic composition with an essential assemblage consume heat.
of clinopyroxene (omphacite) and garnet.
Edge dislocation
The process whereby an initially homogeneous
A type of dislocation whereby the crystal has
solid solution separates into two (or possibly
an additional half lattice plane (Fig. 8.S(a)).
more) distinct crystalline phases (typically
Endothermic reaction during cooling) without the addition or
A reaction that consumes energy, such that heat removal of material (modified after Bates &
energy must be added to the system for the Jackson, 1980).
reaction to proceed (e.g. prograde dehydration
reactions in pelitic schists). Fabric
The geometric and spatial relationships
Enthalpy between the crystal components making up a
The energy that is associated with heat (e.g. the rock. The fabric can relate to the preferred
heat evolved when substances react is the orientations of grain shapes, to grain sizes, and
enthalpy of reaction) (Powell, 1978). to crystallographic orientations of the compo-
nents. Preferred orientations of platy- and
Entropy needle-shaped crystals can give rise to planar
A property that reflects the degree of disorder (S-fabrics), linear (L-fabrics) or to mixed linear-
in a system: the more disordered the system is, planar (L-S fabrics) (Ramsay & Huber, 1983)
the higher is its entropy (Powell, 1978). (Fig. 4.4).


Faserkiesel Garbenschiefer
See Fibrolite. Regionally metamorphosed impure calcareous
sediments and meta-tuffs commonly give rise to
Fault breccia garbenschiefer. These are rocks with abundant
A cataclastic rock associated with fault zones. amphibole prisms lying within the foliation
It consists of coarse angular fragments of vari- plane in 'bow-tie' and radiating aggregates.
able size in a fine-grained, often silicified, Spry (1969) defines the porphyroblasts as being
matrix. 'the size of caraway seeds', but in practice the
same term is also used for rocks with much
Fault gouge larger amphiboles (Fig.S .13 (b)).
A soft and incohesive fine-grained fault-rock. It
is usually rich in clay minerals resulting from Geometric softening
the chemical breakdown of adjacent wall Enhanced ductility in deforming rocks caused
rocks. by lattice reorientation of deforming grains. It
is most pronounced in materials with limited
Fibrolite slip systems such as granitoid mylonites, and
A fine fibrous or hair-like variety of sillimanite . generally involves reorientation of grains so
common in amphibolite facies schists and that their slip directions approach parallelism
gneisses. It often occurs in matted aggregates or with the shear direction.
knots, termed faserkiesel.
Fluid inclusion A curve expressing the thermal gradient
A term for microscopic and sub-microscopic throughout the lithosphere. The nature of the
(and rarely macroscopic) inclusions of fluid geotherm varies from place to place, and with
trapped in minerals during primary crystallisa- time.
tion or fracture healing. They are typically < 50
J.Lm in size (Section 11.6; Figs 11.11-11.13; Geothermal gradient
Plate 8(f)). The rate of change of temperature with depth
in the lithosphere. Many factors influence the
Foliation geothermal gradient, such that it varies greatly
A set of closely spaced planar surfaces from one place to another and at different
produced in a rock as a result of deformation depths in the lithosphere.
(e.g. schistosity or cleavage) (modified after
Park, 1983) (Figs 4.4-4.6). Gibbs free energy
The Gibbs free energy for a closed system is
Fugacity given by G = E + PV - TS = H - TS, where E is
An expression for the behaviour of a gas in the internal energy, V is the volume, T is the
a solid or fluid medium. The fugacity or absolute temperature, S is the entropy and H is
thermodynamic pressure of a gas i (f;) the enthalpy of reaction. The driving force of
is related to activity by a j = f;l~ where f~ metamorphic reactions is the change in Gibbs
is the fugacity of i in a standard state. free energy (~G). For equilibrium, the Gibbs free
For ideal solutions, f; = a j = X j , However, in energy of a system would be at its minimum so
real solutions, a j (and hence f;) may differ for reactions to proceed the reaction should give
greatly from Xj (Xj = mole fraction of i in rise to a lowering of G for the system (i.e . ..1G is
solution). negative) (based on Vernon, 1976).


Gneiss implication regarding metamorphic facies, and

A coarsely banded high-grade metamorphic the term granulite should strictly be reserved
rock consisting of alternating, mineralogically for a granulite facies rock with an essential
distinct (usually felsic and mafic) layers. assemblage of pyroxene (typically hypersthene)
and anorthitic plagioclase (Plate 2(a)). It is
Grain-boundary migration commonly granoblastic.
To help minimise the energy of the rock system
in response to changing P-T conditions, the Greenschist
atoms forming the contacts between individual A low-grade mafic rock with schistose texture
grains are rearranged to a more stable configu- and a mineral assemblage consisting largely of
ration. To a large extent this is achieved by the actinolite, chlorite, epidote, albite, quartz and
process of grain-boundary migration. This accessory sphene.
involves movement at a high angle to the plane
of the grain boundary. During prograde meta- Growth twin
morphism, such a process produces regular Primary (or growth) twins represent twins
interfaces and a polygonal aggregate of grains. present in a given crystal that formed at the
(The aggregate shown in Fig. 5.14(b) probably time of crystal growth (Fig. 5.18(a)).
developed from an aggregate such as Fig.
5.14(a) by this process.) Helicitic structure
Strictly, this refers to an organised, seml-
Grain-boundary sliding symmetrical, S-shaped arrangement of inclu-
Movement within the plane of the grain sions within poikiloblastic crystals (e.g. garnet
boundary. It can be envisaged in terms of the (Plate 5(c)) or staurolite).
physical movement of individual grains past
Heterogeneous nucleation
each other under an applied shear stress.
Non-random nucleation on some pre-existing
substrate, such as new crystals preferentially
Granoblastic structure
nucleating at pre-existing grain boundaries.
An aggregate consisting of equidimensional
crystals of approximately equal sizes. In many Homologous temperature
cases the crystals are rounded to anhedral, but The expression of temperature (in Kelvin) for a
granoblastic-polygonal aggregates are equally particular mineral (e.g. a mineral or rock) as a
common. Granoblastic structure is especially proportion of the melting temperature for that
characteristic of granulites, eclogites and many particular material (i.e. Tffm ). For example, an
hornfelses (Figs 5.14(b) & 5.15). homologous temperature of 0.6 means 60% of
the melting temperature (Tm) for the material
Granofels (mineral or rock) in question.
Non-foliated medium- to coarse-grained
granoblastic metamorphic rock. Hornfels
A hard, fine- to medium-grained granoblastic
Granulite rock produced by high-grade contact metamor-
The term granulite has been used in a number phism.
of ways, and some confusion has arisen
because of different usage. In strict terms, Hour-glass structure
granofels is the best term to use when describ- A structure common in chloritoid porphyrob-
ing an even-grained granular rock with no lasts and some other minerals, consisting of a


dense mass of fine-grained (usually opaque) Interfacial energy

inclusions arranged in the form of an 'hour- See Surface energy.
glass' (Plate l(b)). It forms due to the influence
of the host crystal structure. Intergranular slip
A deformation process involving slip between
Hydrostatic pressure individual crystals of the assemblage.
The pressure exerted by a vertical column of
water, and acting equally in all directions. Interstitial
Hydrostatic pressure increases at approxi- A point defect represented by an extra atom or
mately 0.1 kbar (10 MPa) per kilometre molecule in the crystal lattice (Fig. 8.4).
Intracrystalline plasticity
Idioblastic A deformation mechanism in which grains
See Euhedral. become internally distorted through dislocation
motion and deformation twinning (Rutter,
Idiotopic 1986).
A structural term used to describe rocks
comprised almost entirely of idioblastic crys-
Inversion twins
Twins formed due to a mineral changing its
crystal structure in response to changing P-T
Illite crystallinity
conditions. (Compare with deformation twins,
The degree of ordering of the structure of illite.
which form in response to superimposed stress)
Illite becomes increasingly 'crystalline' from
(Figs S.19(a) & (b)).
anchimetamorphic to greenschist facies condi-
tions. This change shows a regular relationship
with temperature and has been calibrated for Ionic reaction
use as a geothermometer. A reaction involving ionic exchange between
several different reaction sites within the
system in order to facilitate a particular trans-
A solid or fluid phase totally enclosed within a formation (Fig. 1.5).
mineral (e.g. fluid inclusions in vein minerals,
or inclusions of matrix phases in porphyrob- Isobaric cooling (IBC)
lasts) (Fig. 6.1; Plate l(a)). A temperature decrease at near-constant pres-
sure. It is a feature common to certain granulite
Inclusion trail facies retrograde P-T paths.
A regular shaped arrangement of inclusions
(fluid or solid) in a crystal (commonly a Isobaric reaction
porphyroblast) to define a distinct fabric (Fig. A reaction occurring at constant pressure, irre-
11.11; Plates Sial-if)). spective of temperature.

Index mineral Isochemical system

A mineral characterising a particular stage or A system maintaining constant bulk composi-
zone of a progressive metamorphic sequence tion during metamorphism; a closed system
(e.g. sillimanite is an index mineral for upper with no introduction or loss of chemical
amphibolite facies metamorphism). components.


Isograd weight of the column of overlying rock (after

A surface joining points of equal metamorphic Bates & Jackson, 1980).
grade in a unit of metamorphic rocks. It is
usually marked by the first appearance of a L-tectonite
particular index mineral in a progressive meta- A deformed rock contammg a recognisable
morphic sequence. linear structure (Fig. 4.4).

Isothermal decompression (lTD) L-S tectonite

A pressure decrease at near-constant tempera- A deformed rock containing recognisable linear
ture. It is a feature common to the early stages and planar structural components (Fig. 4.4).
of many granulite facies uplift trajectories.
Isothermal reaction A metamorphic rock comprised largely of
A reaction occurring at constant temperature, calcite (Plate 3(a)).
irrespective of pressure.
Kelyphitic rim The dark-coloured (typically biotite-rich)
A symplectic intergrowth mantling a core of an component of migmatites, often seen as
earlier, typically high-temperature mineral selvages around layers and lenses of leuco-
phase. The term kelyphitic rim is often used to somes (Fig. 4.12).
refer to intergrowths of pyroxene and spinel
developed around olivines in gabbros subjected Mesoperthite
to slow cooling at deep crustal levels. The term given to exsolved feldspars with sub-
equal volumes of intergrown K-feldspar and
Latticed-preferred orientation (LPO) plagioclase.
An anisotropic spatial arrangement of the crys-
tal lattices in a population of mineral grains Mesosome
(after Shelley, 1993). A portion of migmatites that is intermediate in
colour between the leucosome (light) and
Leucosome melanosome (dark) components. The mesosome
The light-coloured (leucocratic) component of is often considered to represent the residual
migmatites. It is usually represented by melt- unmelted fraction of high-grade metamorphic
derived granitoid material. rock, but this need not necessarily be the case.

Lineation Metamorphic facies

A recognisable linear component in rocks, A metamorphic facies is a subdivision of meta-
usually formed in response to deformation. morphic conditions in P-T space on the basis
Lineations are often contained within a folia- of diagnostic mineral assemblages which have
tion, and in metamorphic rocks are usually been shown by experimental and field observa-
defined by alignment of elongate minerals such tions to characterise a specific range of P- T
as amphiboles and micas (Fig. 4.4). conditions for a particular compositional group
of rocks (in association with H 20 fluid). This
Lithostatic pressure concept was introduced by Eskola (1915,
The vertical pressure at a point in the Earth's 1939), and is especially useful for the study of
crust, equal to the pressure caused by the me tap elites and metabasites (Fig. 2.1).


Metamorphic facies series Migmatite

An observed sequence of progressively chang- A coarse-grained heterogeneous rock type char-
ing metamorphic facies representing increasing acteristically with irregular and discontinuous
metamorphic grade across a given terrane (e.g. interleaving of leucocratic granitoid material
the change from greenschist to epidote amphi- (leucosome) and residual high-grade metamor-
bolite and finally amphibolite facies is a meta- phic material (restite). Migmatites are often
morphic facies series). intensely folded and heavily veined. General
opinion considers most migmatites to have
Metamorphic grade developed by in situ anatexis (Figs 4.11 &
A general term used to describe the relative 4.12).
rank of metamorphism of rocks in a given area,
without specifying a particular metamorphic Millipede microstructure
facies (e.g. rocks with slaty appearance would A term introduced by Bell & Rubenach
be described as low-grade, while those with (1980) for particular syntectonic usually
migmatitic appearance would be described as elongate porphyroblasts, that for a given
high-grade). porphyroblast margin where the external foli-
ation (S.) passes into the internal foliation
Metamorphism (S), shows Se deflected in opposite directions.
The mineralogical, chemical and structural The pattern is repeated approximately
adjustment of solid rocks to physical and symmetrically across the porphyroblast to
chemical conditions which have generally been give an appearance likened to the arrange-
imposed at depth below the surface zones of ment of legs on a millipede. Originally taken
weathering and cementation, and which differ to be indicative of bulk coaxial shortening,
from the conditions under which the rocks in Johnson & Moore (1996) have suggested that
question originated (Bates & Jackson, 1980). they can form during various types of defor-
mation (Plate Se).
Metamorphism involving modification of the Mimetic growth
bulk rock chemistry by influx or removal of Crystal growth that is influenced by and repro-
chemical components via a fluid phase (e.g. duces a pre-existing rock or crystal structure.
widespread potassium metasomatism associated (e.g. (1) micas preferentially growing along pre-
with granite intrusions is a common occurrence). existing foliation; or (2) pseudomorphing,
where the nucleating and growing mineral has
Mica-fish its form influenced by the structure of the old
'Fish' - or 'lozenge' -shaped mica porphyro- phase it is replacing rather than taking on its
clasts/porphyroblasts aligned within a finer- normal habit).
grained schistose matrix. They are
characteristic of phyllonites and highly Mortar structure
deformed schists, and can be used to determine A term used to describe the structure especially
shear sense (Figs 10.9-10.12). common in highly deformed, dynamically
recrystallised quartz-rich rocks, in which large
Microstructure porphyroclasts of quartz have abundant small
The geometric arrangement and interrelation- sub-grains and new grains developed around
ships between grains and internal features of their margins (Fig. 8.6(d)) (see also Core-and-
grains. mantle microstructure).


Mylonite parent grain and, unlike 'sub-grains', have

A cohesive, foliated and usually lineated rock sharply defined boundaries (Fig. 8.6(c)).
produced by tectonic grain-size reduction via
crystal-plastic processes in narrow zones of Omphacite
intense deformation. It contains abundant A sodium-rich variety of the clinopyroxene
porphyroclasts (10-50% of the rock), which augite: an essential and characteristic mineral
characteristically are of similar composition to of eclogites.
the matrix minerals. The term mylonite is not
restricted to a specific compositional range of Overgrowth
rocks. It is thus possible to have granitoid The nucleation and growth of the same or a
mylonites, carbonate mylonites, amphibolitic different mineral on the outer surface of
mylonites, and so on, depending on the another crystal (Figs 12.7(a, b) & 12.21; Plates
observed mineralogy. Wise et al. (1984) use the 4(a) & 6(d, e)).
term mylonite for all rocks in the spectrum
protomylonite to ultramylonite and suggest the Paragenesis
term orthomylonite for rocks in the middle of In metamorphic petrology, a term used synony-
the range. The definition preferred here still mously for the evolution of the mineral assem-
reserves the term for those rocks in the middle blage characterising a given rock.
of the range containing 10-50% porphyro-
clasts (Figs 8.2 & 8.6(c)-(e) and 8.13(b)). Pelite
A rock of argillaceous composItIOn with a
Myrmekite mineral assemblage dominated by phyllosilicate
A symplectic intergrowth of vermicular quartz minerals. Original sedimentary rocks are
and plagioclase resulting from the retrograde mudstones and siltstones, which when meta-
replacement of K-feldspar (Fig 6.13). morphosed become slates, phyllites, schists,
and so on (Fig. 4.6(a)).
N abarro-Herring creep
The dominant grain-scale rock deformation Peristerites
process at high temperatures and low shear Sodic plagioclases conslstmg of microscopic
stress. Nabarro-Herring creep is a type of and sub-microscopic intergrowths of albite
diffusion creep involving a combination of and oligoclase. Such feldspars characterise the
grain-boundary sliding and diffusional trans- greenschist facies-amphibolite facies transi-
port of matter through the crystal lattice and tion.
along grain boundaries.
Neoblasts A feldspar intergrowth with plagioclase inclu-
Crystals that are more newly formed compared sions enclosed within K-feldspar. It is common
to others in the rock. in high-grade metamorphic rocks and plutonic
igneous rocks (Figs 6.7 & 6.8).
New grains
New grains develop during recrystallisation of Petrogenetic grid
highly deformed rocks and are typically located A diagram the co-ordinates of which are inten-
at grain boundaries and the margins of porphy- sive parameters characterising the rock-forming
roclasts. They are stable, strain-free areas that environment (e.g. pressure and temperature) on
have significant optical discontinuity with the which may be plotted equilibrium curves


delimiting the stability fields of specific miner- Polygonal structure

als and mineral assemblages (Bates & Jackson, A term for rocks containing crystals (usually
1980) (e.g. Fig. 12.8). equigranular) with polygonal shapes
(commonly five- or six-sided) and dominantly
Phase straight boundaries meeting at triple-points
A real chemical entity composed of one or more (modified from Spry, 1969) (Figs S.14(b) &
components. The various minerals and fluids 5.15).
present in a given system are described as phases.
Phyllite One of two or more crystallographic forms of
A well cleaved pelitic rock characterised by a the same chemical substance. For example,
distinctive sheen on the planar surface. It is there are three common polymorphs of
generally of intermediate grain size and of AI 2SiO s , namely, andalusite, kyanite and silli-
metamorphic grade between slate and schist. manite (Fig. 7.9).

Phyllonite Porphyroblast
An intensely sheared phyllosilicate-rich rock A metamorphic mineral that has grown to a
with synchronously developed 'S-C fabrics', much larger size than minerals of the surround-
the intersection of which gives rise to so-called ing matrix (Fig. 5.4; Plate 1).
oyster shell or button-schist appearance in
outcrop. Phyllonites are characteristically asso- Porphyroblastesis
ciated with thrust zones, shear zones and The growth of porphyroblasts.
tectonic slides, and are often developed over
broad areas (Fig. 10.7).
A term used to describe a metamorphic rock
with large crystals (porphyroblasts) grown
The retrogression, especially of cordierite, to an
within a finer-grained matrix (Fig. 5.4).
ultra fine grained green or yellow felty mixture
of muscovite and chlorite (modified from Deer
et a/., 1966). Porphyroclast
A large relict crystal, or crystal fragment, in a
Pleochroism fine-grained matrix of a deformed rock (Figs
The ability of an anisotropic crystal differen- 8.3, 8.13(a, b), 9.3(a, b) & 10.15-10.19).
tially to absorb various wavelengths of trans-
mitted light in various crystallographic Porphyroclastic
directions, and thus show different colours in A term used to describe rocks with abundant
different directions (i.e. as the microscope stage porphyroclasts (e.g. mylonites) (Figs 8.3,
is rotated in plane-polarised light the mineral 8.13(a, b) & 9.3(a, b)).
shows a colour change, and is said to be
pleochroic) (modified after Bates & Jackson, Post-tectonic growth
1980) (Plate 2(d)). The growth of minerals or parts of minerals
(e.g. porphyroblast rims) after deformation of
Poikiloblastic the rock has ceased. It is often deduced from
A term used to describe porphyroblasts with porphyroblast-foliation relationships (Plate
abundant mineral inclusions (Plate l(a)). 6).


Pressure shadow Pseudotachylite

See Strain shadow. A glassy rock produced by frictional melting in a
fault or thrust zone (Mason, 1978) (Fig. 8.12).
Pressure solution
A deformation process whereby material under Pure shear
stress goes into solution at a localised point in An irrotational strain where the area dilation is
a material. This material is transported by flow zero (Ramsay & Huber, 1983).
or diffusion and is usually deposited at some
other locality in the rock system; a process Reactant
termed solution-transfer (Ramsay & Huber, A phase being consumed in a given reaction.
Reaction rim
Pre-tectonic growth A monomineralic or polymineralic rim
Mineral growth before deformation. (commonly of hydrous phases) surrounding the
core of another phase in the process of retro-
Primary twins gression (Plate 4(b)).
Twins developed during crystal growth (Plates
2(c)-(f)). Reaction softening
Enhanced ductility in a deforming rock due to
Product a metamorphic reaction (based on White et aI.,
A phase produced as a consequence of a given 1980).
Prograde metamorphism Any of the processes through which the
Metamorphic changes resulting from increasing number of grain dislocations (i.e. strain energy)
temperature and/or pressure conditions. produced during rock deformation can be
reduced (e.g. sub-grain and new grain develop-
Protomylonite ment) (after Bates & Jackson, 1980) (Figs
A mylonitic rock with porphyroclasts compris- 8.6(b) & 8.10).
ing more than 50% of the rock (Fig. 8.13(a)).
Psammite Solid state textural modification in rocks, involv-
A metamorphosed impure sandstone with a ing intra- and inter-crystalline rearrangement to
mineral assemblage comprised largely of quartz produce a more stable lower-energy system. This
with lesser amounts of feldspar and mica. typically involves grain-boundary migration and
the production of more regular crystal faces (e.g.
Pseudomorph polygonization of quartz). During the prograde
A mineral or aggregate of minerals having the metamorphism of quartz-rich rocks, abundant
form of another mineral phase being replaced. smaller quartz grains generally amalgamate to
A pseudomorph is described as being 'after' the produce fewer but larger crystals (Figs 8.6(c) &
mineral the outward form of which it has (e.g. (d), 8.11 & 5.14(b)).
chlorite after garnet). Pseudomorphing is a
gradual process such that at the arrested stage Regional (orogenic) metamorphism
seen in thin section the pseudomorph may be Metamorphism affecting large areas of the
'partial' or 'complete' (Plates 4(c)-(e)). Earth's crust and commonly associated with


collisional orogeny. Regional metamorphic epidote group minerals and sericite (± calcite)
rocks commonly exhibit complex interrelation- (Section 7.1.3; Plates 4(f) & (g)).
ships between mineral growth and deforma-
tion. S-C fabrics
Mylonites and phyllonites commonly exhibit
Resorption two fabrics (S-C fabrics) that simultaneously
A process involving partial or complete chemi- developed during the intense shearing that
cal modification of earlier formed crystals (e.g. formed such rocks. First described by Berthe et
porphyroblasts) that are no longer in equilib- al. (1979), the C-surfaces are parallel to the
rium. This typically involves diffusion at the shear zone margin, while S-surfaces are oblique
crystal margins penetrating to varying depths to this (typically by about 30°). In phyllonites
within the crystals. Distinct jumps in rim chem- the intersection of these two fabrics gives rise
istry of zoned porphyroblasts (e.g. Mn-rich to a texture that has been described as fish-
rims in garnet) are evidence of resorption (Fig. scale, button schist or oyster shell because of its
5.23(b)). distinctive appearance (Figs 10.6-10.8).

Retrograde metamorphism Schist

Metamorphic changes in response to decreas- Metamorphic rock commonly of pelitic compo-
ing pressure and/or temperature conditions sition, with a well developed schistosity (e.g.
(Plates 5(b)-(h)). Plate 5(c)) (modified after Mason, 1978).

Retrogression Schistosity
Modification of the primary mineral assem- A planar structure defined by the alignment of
blage due to waning P- T conditions and/or inequant minerals such as micas and amphi-
changing fluid chemistry. This process typically boles and where individual minerals are
involves partial or complete replacement of discernable in hand specimen. Rocks showing
high-grade largely anhydrous phases by lower- this structure are termed schists. Such a struc-
grade hydrous phases (Plates 4(b)-(h)). ture is common in regional (orogenic) meta-
morphic and blueschist facies metapelites and
Rheology metabasites (Plate 5(c); Fig. 4.5(a)).
A branch of physics that gives a phenomeno-
logical account of mechanical behaviour of Schlieren
matter, which involves its material properties Streaks or elongate segregations of non-leuco-
(after Poirier, 1985). some (usually biotite-rich) material within the
leucosome component of migmatites (Fig.
Ribbon quartz 4.11). The schlieren represent entrained restite
With intense ductile deformation associated that has not been entirely separated from the
with high strains (especially at high T), individ- melt.
ual quartz crystals often become exceptionally
elongate within the mylonitic fabric. This is Screw dislocation
known as ribbon quartz (Fig. 8.6(e)). A type of dislocation in which part of the crys-
tal is displaced by a lattice unit, giving a
Saussuritisation twisted lattice at the line of dislocation, but
The retrogressive replacement of anorthitic elsewhere the lattice planes line up (Fig.
plagioclase by a fine-grained aggregate of 8.5(b)).


Secondary twins crystals (particularly quartz) to form a skeletal

Twins that have formed subsequent to crystal or mesh-like network of thin interconnected
growth. This includes inversion twins formed strands (Fig. 5.11).
due to change in crystal habit as a result of
instability of the initial structure with changing Sliding reaction
P-T, and deformation twins formed in response A continuous, or sliding, reaction is one not
to deformation of the crystal lattice (Figs 5.18 involving the production of new minerals but
& 5.19; Plates 3(a) & (b)). involving a gradual change in chemistry of a
particular phase or phases in response to
Sericitisation changing metamorphic conditions (e.g. biotites
The alteration of a mineral or minerals to an of pelitic rocks typically show an increasing
aggregate of fine-grained white mica, known as Mg:Fe ratio with increasing temperature).
sericite (Fig. 7.6; Plate 4(g)).
Snowball structure
Serpentinite A term used to describe spiralled or S-shaped
A retrogressed ultramafic rock with a mineral inclusion fabrics in syntectonic porphyroblasts
assemblage comprised largely of serpentine (e.g. garnet). It was originally considered to
minerals (Fig. 7.3). form due to physical rotation of the porphy-
roblast during growth, but it is now largely
Serpentinisation accepted as resulting from differential rotation
A process involving the conversion of high- of the fabric with respect to the porphyroblast
temperature minerals (especially olivine) to an during growth, and/or crenulation overgrowth
aggregate of serpentine. This is common in (Fig. 9.4; Plate 5(d)).
retrogressed ultramafic rocks, and takes place
Solid solution
at temperatures below 500°C and often
A process involving the substitution of one or
dOO°C in the presence of aqueous fluids.
more elements between two (or more) end-
member phases to produce a complete range of
mineral composltlons between the end-
Synonymous with Poikiloblastic (Plate 1(a)).
members. For example, olivine forms a solid-
solution series between pure forsterite
Simple shear (Mg zSi0 4 ) and pure fayalite (Fe 2 Si04 ).
Rotational, constant volume, plane strain
deformation (essentially two-dimensional). Spherulitic
A term describing a sub-spherical mass of acic-
Skarn ular crystals radiating from a common point
A rock formed during contact metamor- (Fig.5.12(c)).
phism/metasomatism by reaction between
carbonate rocks and fluids rich in elements Stacking fault
such as iron and silica. Skarns have an assem- The surface defining the zone of mismatch
blage dominated by calc-silicate minerals, often between partial dislocations and the adjacent
in association with magnetite. ordered crystal lattice.

Skeletal crystals Static recrystallisation

Individual crystals that have nucleated and The modification of the grain structure (size,
grown between th