Myron G. Best. Igneous and metamorphic 2003 Blackwell Science

Подождите немного. Документ загружается.

Migmatite is a composite, heterogeneous rock (the

Greek migma means mixture) on the outcrop scale that

consists of maﬁc metamorphic rock mingled with felsic

rock in the form of planar to folded and contorted

layers, criss-crossing veins, and irregular pods (Fig-

ures 11.24, 15.12, 15.13a, and 15.14). There may be

no sharp distinction between quartzo-feldspathic gneiss

and migmatite; hence labels such as migmatitic gneiss

and gneissic migmatite are frequently used. The richly

illustrated Mehnert (1968) shows the wide variety of

migmatites. The leucocratic component is sharply

bounded to gradational with the more maﬁc host schist,

gneiss, or amphibolite. In many migmatites, the leuco-

cratic leucosome is separated from the mesosome

(a descriptive term for intermediate rock) or paleo-

some (a genetic term for the “average” protolith), by a

more maﬁc selvage, or melanosome (Figures 15.13a

and 15.15).

The origin of migmatites has been the subject of

heated controversy since their ﬁrst description by the

Finnish geologist J. J. Sederholm in 1907 (Ashworth,

1985). A bewildering terminology, partly descriptive

and partly genetic (such as mesosome versus paleo-

some), has grown during this debate. The origin of

migmatites seems to be linked in some way with their

typical occurrence in high-T sectors of regional meta-

morphic terranes where mineral assemblages deﬁne the

upper amphibolite facies (Figure 14.33). Under these

conditions, decomposition of muscovite (Figure 14.31)

and perhaps biotite liberates water that may be essen-

tial in creating local water-saturated conditions for

minimum-T partial melting just above the granite solidus

(Figures 5.24–5.26; see also Section 11.6.1 and Figure

11.21, top). Thus, the felsic leucosome in migmatites

can be generated by in situ partial melting and segre-

gation from the paleosome, the melanosome being a

maﬁc residue. Mass-balance investigations disclosing

complementary compositions support this closed sys-

tem origin, as does the presence of euhedral, slightly

zoned plagioclases of magmatic character in some leu-

cosomes. Nonetheless, a complementary compositional

relation does not always exist and in such cases it is

argued that leucosomes could have been intruded as

magmas from a more distant source, despite the neces-

sarily pervasiveness of the injections. Another proposed

origin consistent with the pervasive heterogeneity of

migmatites and the more or less complementary com-

position of leucosome and melanosome with the paleo-

some is metamorphic differentiation; maﬁc and felsic

components segregate via diffusion. Exactly how this

ordering and decrease in entropy of the system can

occur spontaneously is a puzzle. Yet another origin

involves metasomatic transformation of the parent

rock by introduction of felsic components via pene-

trating ﬂuids. The origin of migmatites has yet to be

resolved. They may, in fact, be polygenetic, originating

Petrography of Metamorphic Rocks: Fabric, Composition, and Classiﬁcation

457

quartz grains may be still evident in typically low-grade

slates (Figure 14.25b). Weakly foliated metagraywackes

(Figure 14.17b) are commonly interlayered with slates,

as are various metavolcanic rocks such as greenstones

and metatuffs that have relict magmatic fabrics. Slates

are purple or red (oxidized, with abundant hematite),

green (chlorite) and, most commonly, gray to black

(reduced, with graphite). Some of the latter have

porphyroblasts of euhedral pyrite. Low-temperature

recrystallization of slate in contact aureoles creates

poikiloblasts of cordierite, andalusite, or mica, forming

spotted slates (Figures 14.10 and 14.25c); locally,

altered cordierites preferentially weather out, leaving

pits.

Phyllite is also aphanitic, but because of a slightly

coarser grain size than slate, has a lustrous or silky

sheen on foliation surfaces. Phyllites are transitional

between slates and schists and share associations and

properties of both.

Widespread schists are phaneritic, commonly

porphyroblastic, and have a weak to well developed

segregation layering of felsic and maﬁc material that

usually parallels and augments the already strong lepido-

blastic fabric (Figure 14.17d). Lineation is commonly

expressed by elongate segregations of contrasting min-

erals (Figure 15.9) or by folds and wrinkles in the foli-

ation. Relative to slates and phyllites, coarser-grained

schists have better developed imposed metamorphic

fabrics that have modiﬁed and usually erased relict

magmatic and sedimentary features. Compositional

layering in schists may have originated by metamorphic

differentiation in addition to or in lieu of inherited

bedding.

All chemical classes are represented in schists,

but pelitic and quartzo-feldspathic compositions with

abundant micas and chlorites are most common.

Diagnostic minerals indicate low- to intermediate-

grade conditions of the subgreenschist, greenschist,

blueschist, and amphibolite facies.

15.2.3 Weakly Foliated Rocks

The typically higher grade of metamorphism of a

phaneritic gneiss is evident in the low modal propor-

tion, or absence, of micas and chlorite. Grains of major

rock-forming minerals, mostly quartz and feldspar, are

more equant. Foliation ranges from barely perceptible to

conspicuous and is expressed by sparse or imperfectly

aligned platy minerals, parallelism of augen in augen

gneiss (Figures 14.38a and 15.7), ﬂattened mineral

aggregates (Figures 15.6 and 15.13b), or mechanically

passive compositional layering (Figure 15.5; see also

Figure 19.24a). Maﬁc layers may be more strongly foli-

ated where there are lepidoblastic biotites or decussate

amphiboles, whereas alternating felsic layers tend to be

granoblastic. Abundant ﬁeld photographs in Passchier

et al. (1990) illustrate the character of gneisses.

458 Igneous and Metamorphic Petrology

(a) (b)

in different ways in different places at different times,

some being magmatic and others metamorphic.

Deep crustal migmatites have long been regarded

as spawning grounds for granitoid plutons. However,

no direct genetic connection has been unequivoc-

ally demonstrated. One tentative conclusion is that

migmatites represent the culmination of prograde

metamorphism and are not embryonic granitic plutons.

Mylonite is a sheet-like, laminated body of ﬁne-

grained rock developed in high-strain-rate ductile

shear zones that are the deep-seated counterparts of

brittle cataclastic fault rocks in the shallower crust

(Figure 8.9). The microcrystalline, streaked matrix

consists of intensely ﬂattened, “smeared out” grains

surrounding larger, less (but still visibly) deformed

porphyroclasts deﬁning a porphyroclastic or ﬂaser

fabric (Figure 14.16c, d). Though usually visible, the

foliation is generally passive and does not control how

the rock breaks. Elongate mineral aggregates may

express a stretching lineation. A protomylonite is

15.12 Layered hornblende–plagioclase schist. The slabby character of the outcrop and ease of breaking into slabs qualiﬁes the rock as a schist.

The well developed felsic segregations, which would qualify the outcrop as a gneiss for some geologists, possibly represent migmatitic

partial melts. Camera lens cap for scale. Proterozoic rocks in Black Canyon of the Gunnison River, Colorado.

(a) (b)

15.13 Migmatitic quartzo-feldspathic (felsic) gneiss. Note massive character of rock that does not weather or break along its subtle mechanic-

ally insigniﬁcant foliation. The rock may have been derived from a granitoid protolith; if true, it could be called an orthogneiss.

Precambrian terrane in the Colorado Front Range. (a) Irregular more felsic leucosome layers probably formed by partial melting and

segregation from the host rock. Note thin melanosome borders to left of and above and to right of and below hammer. Compare

Figure 15.15. (b) Close-up view, rotated 90° from (a) orientation, of more homogeneous mesosome (paleosome) gneiss between seg-

regated melanosome–leucosome layers. Weak foliation is deﬁned by small lenticular maﬁc aggregates.

less deformed and has 50% porphyroclasts (Figure

14.16b), whereas the more deformed ultramylonite

(Figure 14.16d) has 10% porphyroclasts and matrix

grains are so minute (0.01 mm) that the hand sample

may resemble chalcedony or chert (Figure 15.16) or

even appear glassy, not unlike pseudotachylite (Special

Interest Box 11.1). Phyllonites are mylonites that con-

tain abundant phyllosilicates and have an S–C fabric

(see Section 17.3.5).

Mylonites can originate in any type of rock, but

seem best preserved in quartzo-feldspathic protoliths.

Study of mylonites by optical and electron micro-

scopy, with complementary laboratory experiments on

Petrography of Metamorphic Rocks: Fabric, Composition, and Classiﬁcation

459

(a) (b)

15.14 Migmatite, Black Canyon of the Gunnison River, Colorado. Pocketknife is 8 cm long. (a) Irregular, contorted veins and pods of coarse

leucocratic felsic rock lie in ﬁner-grained amphibolite. Note that the major leucocratic area at the top encloses material more maﬁc than

the host rock, as if the two resulted from segregation, or differentiation, of the original intermediate rock (a leucosome–melanosome–

paleosome relationship; compare Figure 15.13). (b) Similar to (a) except that a late granitic dike cuts across the migmatitic segregations.

Mesosome

(paleosome)

Melanosome

Leucosome

Melanosome

Mesosome

(paleosome)

15.15 Schematic segregation of felsic and maﬁc (dashes) minerals in

a migmatite resulting from in situ partial melting. Mesosome

is a descriptive term for rock intermediate in composition be-

tween maﬁc melanosome and leucosome, whereas paleosome

is a genetic term for the protolith.

15.16 Sheet of aphanitic ultramylonite in granite, Kern Mountains,

eastern Nevada. Note slight discordance between faint trace of

foliation in mylonite and its contact with the host granite.

Pocketknife is 8 cm long.

ductile deformation, has demonstrated that the typ-

ical, and once controversial, very ﬁne grain size results

from dynamic recrystallization (Figure 15.17) rather

than by brittle pulverization. At high strain rates under

elevated P–T metamorphic conditions, drastically

straining grains simultaneously recrystallize into a

mosaic of tiny, strain-free grains.

15.2.4 Nonfoliated Maﬁc Rocks

Metamorphosed rocks of basaltic composition called

metabasites are widespread in regional metamorphic

terranes of orogenic belts. In Archean–early Protero-

zoic cratons, belts of low-grade metabasites lie between

more voluminous metamorphosed granitoids that are

now felsic orthogneisses (Figure 9.14). Metabasites

constitute much of the modern-day oceanic crust, frag-

ments of which are exposed in accreted terranes of

ophiolite on land in subduction zones.

Despite widespread nonhydrostatic states of stress,

granoblastic or only subtly foliated greenstones and

amphibolites appear to dominate over mineralogically

equivalent, strongly foliated schists. Inhomogeneous

deformation and uneven strain partitioning may allow

parts of metabasite bodies to escape development of

anisotropic fabric while adjacent parts or neighboring

rocks that have lesser ductile strength experience

ductile ﬂow and development of tectonite fabric.

Greenstone is an aphanitic, maﬁc, low-grade meta-

morphic rock composed of a haphazard isotropic array

of combinations of epidote-group minerals, prehnite,

pumpellyite, chlorite, actinolite, albite, titanite, and

locally a carbonate mineral. This low-grade assem-

blage of minerals partially to completely replaces prim-

ary magmatic minerals but relict magmatic fabrics,

including pillows and amygdaloidal, porphyritic, and

ophitic texture (Figure 14.6a, b), are commonly pre-

served. With increase in metamorphic grade, complex

reactions over a range of T consume the lower grade

assemblage of albite, epidote, chlorite, and actinolite,

forming generally coarser grains of oligoclase–andesine

and aluminous hornblende that erase any vestiges of

protolith fabric in amphibolite.

Amphibolite is a widespread phaneritic rock type

made of plagioclase and hornblende (Figure 14.6c).

Depending on grade and bulk chemical composition,

almandine-rich garnet and pyroxene may also occur

in garnet amphibolites (Figure 14.8); subordinated

amounts of biotite, titanite, Fe–Ti oxides, and quartz

may also be locally present. Some petrologists use the

label amphibolite for any plagioclase–hornblende rock,

regardless of fabric. Although the rock type amphi-

bolite is the namesake for the amphibolite facies, the

common name, unfortunately, causes confusion. Not

all amphibolites belong to this facies nor is the entire

facies made up of amphibolite.

Eclogite is a very rare but geologically signiﬁcant

rock that was the focus of considerable attention in the

late twentieth century (Smith, 1988; Carsell, 1990). The

460 Igneous and Metamorphic Petrology

15.17 Anorthosite mylonite from a shear zone in the Jotun nappe complex, Jotunheimen, Norway (Ji and Mainprice, 1990). The estimated T

during ductile deformation and concurrent dynamic recrystallization was about 600°C. Photomicrographs under cross-polarized light

courtesy of Shaocheng Ji. (a) Lenticular porphyroclasts of strained plagioclase (An

) in a granoblastic matrix of small polygonal

plagioclases resulting from dynamic recrystallization (see Section 17.2.4). Higher relief grain top center of view is clinopyroxene.

(b) Enlarged view of porphyroclasts. Strain is manifest in undulatory extinction and lenticular tapered deformation twins. Small,

unstrained polygonal grains along boundaries of porphyroclasts have completely replaced some narrow high-strain bands.

(a) (b)

oceanic lithosphere before being exhumed by tec-

tonic and erosion processes.

2. As local layers, lenses, and boudins that are pos-

sibly relict dikes in amphibolite-facies migmatitic

felsic gneisses in the Caledonide orogen of western

Norway (Figure 15.19). More recent discoveries are

in continental sutures in the western Alps and east-

central China.

Coleman et al. (1965) classiﬁed eclogites into three

groups corresponding to these three petrotectonic

associations but emphasizing contrasts in the chemical

composition of the constituent clinopyroxenes and

garnets. Subsequent work has not entirely supported

the mineralogical distinctions. Because of their wide

range in equilibration T, Carswell (1990) adapted an

earlier proposal by S. Banno that subdivides eclogites

into three thermal categories:

1. High-T (900°C). Includes most mantle-derived

xenoliths in maﬁc and ultra-maﬁc magmatic host

rocks.

2. Medium-T (900–550°C). Includes some xenoliths

as well as eclogite intercalated with felsic gneisses.

3. Low-T (550°C). Associated with blueschist-facies

rocks in near-trench settings.

15.2.5 Nonfoliated High-Grade Felsic Rocks

Phaneritic granoblastic rocks are widespread in meta-

morphic terranes. Those composed mostly of calcite

are called marble, of quartz quartzite, and of plagio-

clase  amphibole amphibolite. Those made mostly of

quartz and feldspar are called granofels. Biotites, if

present, are randomly oriented; other high-grade maﬁc

minerals such as amphiboles, pyroxenes, and garnets

do not contribute to a planar fabric. Granofelses are

typical of the upper amphibolite and higher grade

granulite facies. Protoliths include granitoids and

Petrography of Metamorphic Rocks: Fabric, Composition, and Classiﬁcation

461

0.1 mm

Quartz

Coesite

Garnet

15.18 Partially replaced coesite grains enclosed within garnet in an

ultrahigh-P phengite-bearing eclogite, Guanjialing, Dabie

Mountains, central China. In the photomicrograph under

plane polarized light the ragged relics of high-relief coesite

(monoclinic, with a feldspar-like atomic structure) are rimmed

by low-relief, polycrystalline, replacement quartz. Note the

radial tensile cracks in the host garnet that were produced

as the larger molar volume (less dense) quartz replaced the

smaller molar volume (more dense) coesite. In the absence of

coesite, such cracks surrounding polygranular quartz provide

evidence for the former presence of the denser coesite poly-

morph. Sample CD41 described in Carswell et al. (1997).

Photomicrograph and caption courtesy of Tony Carswell.

exact deﬁnition, distinction from allied rocks, and con-

ditions of formation remain somewhat controversial.

As originally deﬁned by R. J. Haüy in 1822, eclogite is

essentially a bimineralic rock composed of a red-brown

pyrope-rich garnet solid solution and a green clino-

pyroxene solid solution called omphacite whose end-

members are essentially jadeite and diopside (Appendix

A). Other common accessory constituents include quartz,

kyanite, orthopyroxene, rutile, and, rarely, coesite (Fig-

ure 15.18). Plagioclase is absent. These mineralogical

features as well as its much greater density (3.4–3.5

g/cm

) compared to its basaltic protoliths (Figure 8.15)

bespeak its equilibration under high pressures.

The occurrence of eclogite as mantle-derived

xenoliths in highly alkaline basaltic rocks, kimberlite,

orangeite, and lamproite has already been described

(Sections 11.2.1 and 13.12.2). Controversial possible

sources for the xenoliths are bodies of crystallized

basalt magma trapped in the mantle and transformed

basalt crust in ancient subducted oceanic lithosphere.

As noted by P. Eskola in 1921, eclogite also occurs in

two metamorphic terranes in orogens:

1. With blueschist-facies rocks in forearc settings near

the oceanic trench (Figure 14.24). These eclogites

are believed to be basalt crust that was metamor-

phosed while deeply buried in the subducted

15.19 Pod of coesite-bearing eclogite under hammer surrounded by

felsic paragneiss on the northeast shore of Nordpollen about

1 km northwest of Vetrhus, Norway (Carswell et al., 1999).

Photograph courtesy of Tony Carswell.

thoroughly reconstituted psammitic and pelitic meta-

sedimentary rocks. In the past, granofelses were called

granulites but this designation is not recommended

because of confusion with the facies of the same name.

Not all granulites of older geologic writings belong to

the granulite facies and other rock types occur in it.

Charnockite is a granofels consisting mostly of

perthitic alkali feldspar and quartz plus minor Fe-rich

orthopyroxene and plagioclase (Figure 15.20). Feldspars

are commonly dark green, brown, and red, and even

black. Antiperthitic plagioclases have small amounts

of exsolved alkali feldspar and alkali feldspars are

intergrowths of subequal amounts of K- and Na-

endmembers (mesoperthites). These two perthites

indicate high temperatures of equilibration under

granulite-facies conditions, followed by slow cooling.

Low water fugacities stabilize anhydrous maﬁc min-

erals in lieu of biotite and hornblende. Some quartz

grains in hand sample show an unusual blue opales-

cence due to the presence of minute needles of rutile

apparently formed by exsolution of TiO

originally dis-

solved in the quartz at high temperatures. The name

charnockite was coined by T. H. Holland in 1900 from

the tombstone of Job Charnock at St John’s Church,

Calcutta, India!

Charnockite is one rock type within the charnockitic

rock suite that includes other, more plagioclase-rich

felsic phaneritic rocks in which the dominant maﬁc

mineral is orthopyroxene. (In lieu of the detailed classi-

ﬁcation of the charnockitic rock suite in Le Maitre

(1989, p. 13), they can simply be referred to as ortho-

pyroxene monzonite, orthopyroxene granodiorite, and

so on depending on their modal composition and prob-

able magmatic protolith.) Garnet and clinopyroxene

may also occur in charnockitic rocks that are wide-

spread in Proterozoic associations with anorthosite

and related gabbroic rocks (see section 19.5.2). Some

petrologists question whether charnockitic rocks are

metamorphic, originating by recrystallization in the

solid state from a granitoid protolith. Alternatively,

they may be magmatic, crystallizing directly from dry

hot magmas (Kirkpatrick and Ellis, 1992) under

granulite-facies conditions, in contrast to the biotite–

hornblende-bearing magmas of widespread Mesozoic–

Cenozoic continental arcs. Magmatic charnockitic

rocks may have suffered overprinting subsolidus grain

boundary adjustment during very slow cooling—a sort

of autometamorphism. This would confuse their origin

with rocks produced by a distinct episode of metamor-

phism independent of, and perhaps tens to hundreds

of millions of years after, the origin of the protolith.

15.2.6 Other Nonfoliated Metamorphic Rocks

Quartzite forms by recrystallization of relatively pure

quartz sandstones. In such protoliths, rounded detrital

quartz grains that are typically outlined by adhering

clay or other foreign particles have recognizable optic-

ally continuous overgrowths of cementing quartz. In

metamorphic quartzites, recrystallization has essen-

tially or completely erased this clastic texture and pro-

duced a granoblastic aggregate. Quartz grains in some

quartzites may be somewhat strained and others may

have ragged, sutured margins resulting from deforma-

tion and dynamic recrystallization. Massive quartzites,

possibly with relict bedding, grade into laminated or

slabby quartz schists in which thin folia of mica altern-

ate with thicker lenses and layers of quartz; protoliths

were thinly bedded shaly sandstones. Marbles, schists,

phyllites, and slates representing metamorphosed lime-

stones and shales are commonly interlayered with

quartzites in metasedimentary sequences.

Some quartzites are coarse-grained metacherts

derived from deep marine deposits that contain Mn-

silicates, such as piemontite (an epidote-like mineral)

and spessartine garnet.

Metamorphism of relatively pure limestone and

dolomite yields generally phaneritic marble (Figure

14.4b) and dolomite marble, or dolomarble, respect-

ively. Impure carbonate protoliths that contain vari-

able amounts of quartz sand grains and shaly material

may have, in addition to calcite and dolomite, quartz,

diopside, tremolite, talc, phlogopite, wollastonite, calcic

plagioclase, vesuvianite, forsterite, and grossular–

andradite. Relict bedding may be preserved. Gray

coloration, often uneven and streaked, in otherwise

pure marbles is produced by minute specks of graphite

or other minerals. But graphite and carbonaceous

462 Igneous and Metamorphic Petrology

Perthite

Quartz

Orthopyroxene

0mm1

15.20 Charnockite from Madrasteria, St Thomas Mountain, Salem

District, India. The Fe-rich orthopyroxene is strongly

pleochroic green to pink. Sample courtesy of A. P. Subramanian.

material can react with water under metamorphic

conditions, liberating CO

and CH

that leaves the

rock snow-white (see Special Interest Box 16.2).

Hornfels results from thermal metamorphism within

contact aureoles surrounding magmatic intrusions.

Granoblastic fabric is typical and grain size ranges

from ﬁne phaneritic to aphanitic; conchoidal fracturing

occurs in very ﬁne-grained hornfelses. Relict bedding

may be present. Protoliths range widely in composition

in aureoles. High-grade derivatives of maﬁc rocks are

combinations of plagioclase, hornblende, pyroxene,

and Fe–Ti oxides. Metamorphism of impure carbonate

rocks consist of Ca–Al–Mg–Fe silicates, such as tremo-

lite, diopside, wollastonite, forsterite, and so on. Pelitic

hornfelses commonly contain poikiloblastic cordierite

and/or andalusite (Figure 14.25c, d).

15.2.7 Serpentinite

This generally massive, nonfoliated rock type consti-

tutes a large proportion of ophiolites emplaced in near-

trench sectors of contractional orogens and, therefore,

merits special attention.

Serpentinites are ultramaﬁc, typically aphanitic

rocks composed mostly of serpentine, together with

accessory magnetite, brucite, other Mg-silicates, and

Ca–Al-silicates (O’Hanley, 1996). Most protoliths are

ultramaﬁc rocks that contain abundant Mg-rich oli-

vine and/or pyroxene; that is, peridotite, dunite, and

pyroxenite. However, rare metasomatized gabbro and

dolomite may also be protoliths. Serpentinites have a

bewildering array of fabrics on all scales of observation

as well as colors, but most are shades of yellow and

green ranging to black. Their color and mottled

scaly appearance is the basis of the name from the

Latin serpentinus, meaning serpent rock. Weathered

surfaces can be also be shades of orange, red, and

brown.

Three low- to medium-grade serpentine minerals—

lizardite, antigorite, and chrysotile—are all approxim-

ately Mg

(OH)

but limited amounts of Fe

2

3

, and Al can substitute for Mg and Si. Because of

subtle chemical contrasts, as well as differences in

atomic structure, these three minerals are not poly-

morphs with separate P–T stability ﬁelds. Both anti-

gorite (stable up to 550°C) and lizardite (300°C) are

platy, whereas chrysotile (300°C) consists of scrolled

layers manifest in ﬁbrous form as asbestos. All three are

found in subgreenschist-facies rocks, lizardite and

antigorite in blueschist-facies rocks, and antigorite in

greenschist- and amphibolite-facies rocks.

Serpentinization. A wide range of ﬂuid compositions

are apparently involved in the open-system serpentin-

ization of Mg-rich anhydrous protoliths, but ﬂuids must

be essentially aqueous in order to create serpentine

minerals that contain 12–14 wt.% water. Some serpen-

Petrography of Metamorphic Rocks: Fabric, Composition, and Classiﬁcation

463

Residual

Cr-Fe spinel

Mesh-textured

serpentine

olivine grainof orthopyroxene

0.50mm

15.21 Partially serpentinized peridotite (harzburgite). Serpentine

replacement of olivine progressed inward from internal frac-

tures as water moved in, creating mesh texture. Serpentine

has mostly replaced primary orthopyroxene, creating bastite

texture. Primary Cr–Fe spinel is unaltered.

tinites form by interaction between ocean water and

hot ultramaﬁc rocks at spreading ridges, whereas other

protoliths are serpentinized by lower

O meteoric

waters (Figure 2.24) on continents. Locally unusually

low fugacities of oxygen during serpentinization are

indicated by the occurrence of secondary native iron

(Fe

) and Fe

–Ni–Co alloys. Ferrous iron in magmatic

olivines and pyroxenes apparently reacts with oxygen

in inﬁltrating water to form magnetite (Fe

2



while H liberated from the dissociated water reduces

other Fe, Ni, and Co to their metallic state (Frost,

1985). Some protoliths serpentinized at ocean ridges

can be subsequently overprinted by thermal, tectonic,

and ﬂuid-induced metamorphism as they are emplaced

into accretionary prisms as ophiolite (Section 13.6).

Serpentine replacement of olivine grains begins by

hydration reactions along cracks, yielding after com-

plete serpentinization a mesh texture (Figure 15.21).

In contrast, replacement of chain silicates, such as

orthopyroxene, creates a so-called bastite texture. Relict

fabrics can be obliterated by penetrative deformation

under nonhydrostatic stresses. Bodies of serpentinite

are mechanically weak and fault movement is com-

monly focused in them. In sheared serpentinite, a

multitude of differently oriented, polished slip surfaces

(slickensides) can be visible in hand sample.

Simple hydration of peridotite to serpentine miner-

als without any change in proportions of other chem-

ical constituents theoretically results in a decrease in

density of 20% from 3.3 to 2.65 g/cm

. But whether

serpentinization actually proceeds in this manner and

results in a volume increase or proceeds without a

change in volume as some original components are

removed has been controversial for many decades. Two

model reactions involving olivine exemplify these two

end-member processes:

15.1 2Mg

SiO

3H

O Mg

(OH)

Mg(OH)

forsterite serpentine brucite

88 cm

111 cm

25 cm

V  48 cm

15.2 6.15Mg

SiO

 6H

O  2Mg

(OH)

forsterite serpentine

271 cm

221 cm

865 g 90 g 554 g

 2Mg(OH)

 4.3MgO  2.15SiO

brucite removed in solution

49 cm

116 g 173 g 129 g

V  0

Resolution of this controversy has far-reaching geo-

logic implications that go beyond just the nature of the

serpentinization itself. For example, if serpentinization

involves a volume increase in the thin oceanic litho-

sphere at spreading ridges some of the uplift might be

related to inﬁltration of seawater into hot peridotite in

this extensional rift environment. The global budget

of Mg and Si in the oceans could depend on which

reaction prevails in the ridge metasomatic system.

Advocates of a constant volume replacement (reac-

tion 15.2) cite preservation of relict magmatic textures

in serpentinized peridotite; primary grain outlines on

the scale of millimeters do not seem to show expansion.

Yet voluminous Mg- and Si-metasomatism in neigh-

boring rocks is not always apparent. On a larger out-

crop scale of tens of meters, it is impossible to evaluate

volume changes because of nearly ubiquitous shear

zones that cut through the serpentinite body, confusing

the issue of whether the deformation is related to ex-

pansion or to tectonic emplacement. O’Hanley (1992)

points out that with the predicted volumetric expan-

sion during serpentinization the change in dimensions

of relict olivine and pyroxene grains is only on the

order of hundredths of a millimeter, which is too small

to be reliably discerned. However, the cumulative

effect of volumetric expansion might be more apparent

on the scale of tens of centimeters to perhaps a meter.

Indeed, he documents instances where unsheared

outcrops consist of distinct, fracture bounded blocks,

or “kernels,” of this dimension of peridotite that are

surrounded by serpentinite grading inwardly into less

altered rock. Serpentine-ﬁlled fractures radiate from

the interior of the kernel and become wider outward,

consistent with their development during progressive

inward serpentinization and volumetric expansion of

the block of peridotite (Figure 15.22). Some petrolo-

gists have documented a constant ratio between SiO

and (MgO  CaO  MnO  NiO  total Fe oxide)

in variably serpentinized peridotites, ruling out reac-

tion 15.2 and supporting 15.1.

Ca in serpentinizing pyroxenes, especially Ca-rich

clinopyroxenes, in ultramaﬁc protoliths cannot be

sequestered in serpentine minerals or several common

accessory phases. It consequently must exit the open

system, and is presumed to migrate via advecting ﬂuids

into adjacent rock where it produces metasomatic

rodingite. A wide variety of protoliths adjacent to ser-

pentinizing rock can be so metasomatized into Ca-rich,

silica-poor assemblages of epidote group minerals, Ca-

rich garnet, diopside, prehnite, vesuvianite, and many

other Ca–Al minerals. Na purged from the metasoma-

tizing rocks can in turn form nearby metasomatic

464 Igneous and Metamorphic Petrology

Peridotite

Serpentinite

~0.5 m

Initial

15.22 Kernel pattern and cross-fractures in serpentinized peridotite.

Redrawn from O’Hanley (1992). As serpentinization penetrates

into the core of peridotite, the accompanying volume increase

produces tensile fractures in the surrounding serpentinite jacket.

Progressive inward serpentinization creates outwardly wider

chrysotile-ﬁlled fractures, as shown at successive steps in time

, t

, and t

. Actual expansion and fracturing is more contin-

uous. Fractures are oriented radially around the core, essentially

perpendicular to the sharp core–rim interface, and, in a plane

parallel to it, form a grid-pattern visible in top of the block.

albitites or, at higher pressures in a subduction setting,

jadeitic rocks.

Intimate mixtures of green serpentinite and white

carbonate rock, seen in polished ﬂoor tiles and build-

ing façades, is called ophicalcite, from the Greek

ophites, meaning like a serpent. The carbonate is

commonly introduced into openings in cataclastic

serpentinite formed in fault zones.

15.2.8 Metasomatic Rock Types

Generally nonfoliated rocks originating by replacement

of pre-existing minerals are described in the following

paragraphs. Serpentinite could have been included here.

Skarn is a calc-silicate rock produced by metaso-

matic replacement of predominantly carbonate rock

(limestone and dolostone) by Ca–Mg–Fe–Al silicates

(Einaudi et al., 1981). Replacement is accomplished by

advecting hydrothermal solutions usually derived from

a nearby magmatic intrusion and on a smaller scale by

interdiffusion of components at boundaries between

layers of silicate and carbonate rock in contact and

regional settings. A wide variety of silicate minerals

are formed in contact aureoles, including diopside–

hedenbergite, vesuvianite, grossular–andradite, scapolite,

and epidote minerals. Economically valuable deposits

of magnetite as well as sulﬁdes and oxides of Cu, Mn,

Zn, and Pb also occur locally. Most of the world’s

tungsten (W) is obtained from scheelite, CaWO

, and

wolframite, (Fe,Mn)WO

, in skarns.

Jasperoid is produced in much the same manner as

skarn by advecting silica-laden hydrothermal solutions,

usually derived from nearby magmatic intrusions.

Limestone and dolostone are replaced to varying de-

grees by microcrystalline quartz so that the protolith

appears unchanged, even preserving a dark gray color.

Jasperoid is locally host to economic deposits of Au,

Ag, and other metals.

Greisen is a metasomatized granitic rock, commonly

of the S-type, produced by interaction with hydrother-

mal solutions having high concentrations of F, B, Li,

W, and Sn. Very large time-integrated ﬂuid/rock ratios

create a greisen composed chieﬂy of quartz and Li-

bearing mica (muscovite, zinnwaldite, or lepidolite)

plus accessory topaz, tourmaline, ﬂuorite, rutile, cassi-

terite, and wolframite. Some greisens are economic

resources of the listed elements.

Fenite is a syenitic product of alkali metasomatism

of felsic country rock surrounding intrusions of car-

bonatite and some highly alkaline magmas (Figures

13.39 and 13.40). The fenitizing Na–K-rich solutions

desilicate the protolith by replacing any unstable

quartz by Na–Fe-rich amphibole and/or pyroxene

(Figure 14.21), whereas alkali feldspar remains stable.

Apatite and titanite are accessories.

Rodingite is a massive light-colored rock composed

of Ca-rich minerals produced by inﬁltration of Ca-

bearing solutions into rocks surrounding serpentiniz-

ing ultramaﬁc protoliths.

Spilite is a metasomatized basalt produced by hydro-

thermal activity at ocean ridges and more locally by

burial metamorphism. Ca has been removed (chieﬂy

from primary calcic plagioclase), along with transition

metals such as Cu, Ni, Co, and Mn, while Na, H

and perhaps Mg have been added. Ti and Al remain

constant. The metasomatic assemblage is of low-grade

albite–oligoclase  chlorite  tremolite–actinolite and

locally includes epidote, pumpellyite, prehnite, and

quartz. A light gray (in thin section) aggregate of cryp-

tocrystalline titanite (“leucoxene”) commonly replaces

primary Fe–Ti oxides. Magmatic clinopyroxenes and

protolith textures may only partially be replaced. Wide-

spread amygdules are of chalcedony, opal, chlorite,

calcite, and hydrous Ca–Al silicates. First recognized

over 150 years ago as pillow lavas within ophiolitic

sequences, spilites were believed to have have crystal-

lized directly from a unique spilitic magma.

The protolith of rare anthophyllite–cordierite rocks

has been a long-standing controversy, but they are now

believed to originate by recrystallization at moderate-T

and moderate- to low-P of highly metasomatized,

Ti–Ca–Na-poor, ocean ridge basalts.

15.2.9 Misﬁts

Inevitably, there will be rocks that cannot be appro-

priately matched to any of the foregoing deﬁned

rock types. Common misﬁts are weakly or nonfoliated

rocks of variable grain size and commonly composed

of a partially retrograded mineral assemblage. These

may simply be called “rock” with an appropriate com-

positional qualiﬁer, such as feldspar–quartz–chlorite

rock.

15.2.10 High-Strain-Rate Rocks in Fault and

Shear Zones

Brittle to ductile rock-forming processes (Sections 8.2

and 14.1.3) operate where bodies of rock slip rapidly

past one another. During shallow crustal low-T faulting,

brittle rock is mechanically broken, crushed, and pul-

verized to form a structureless (isotropic) mass of angu-

lar fragments possessing cataclastic fabric (Figure 8.7).

Coarser fragments make up tectonic or fault breccia,

whereas more thoroughly and intensely pulverized

sandy to clay-like material is fault gouge. Subsequent

cementation of these cohesionless porous materials

may occur where dissolved minerals precipitate from

inﬁltrating ground water, forming a solid cataclastic

rock. This term is also used for a nonfoliated cohesive

rock formed by processes transitional between brittle

and ductile. Very local melting can create pseudo-

tachylite (Special Interest Box 11.1).

Ductile deformation in generally deep crustal shear

zones creates ﬁne grained, cohesive mylonite.

Petrography of Metamorphic Rocks: Fabric, Composition, and Classiﬁcation

465

15.2.11 Veins

Veins are a widespread feature of metamorphic terranes,

regardless of their grade. In contrast to magmatic dikes,

veins are typically a result of advecting hydrothermal

solutions that precipitate minerals into openings in

the rock mass. Their size, form, composition, source

of ﬂuids, and chronology relative to deformation and

recrystallization in the wall rock can be complex, as

discussed by Fry (1984) and Barker (1990).

Veins range from microscopic dimensions to meters

in thickness and tens to hundreds of meters in length.

Some are planar sheets of uniform thickness, whereas

lenticular (Figure 15.8) and irregularly shaped ones

(Figure 15.23) seem more common. Like dikes, they

can be discordant to wall rock structure or be hosted in

isotropic rock, but many are more or less concordant to

foliation in the host rock. Many vary from concordant

to discordant along their extent.

Some veins are monomineralic, consisting of calcite

or epidote (in granitoids such as Figure 14.27). But

the most common vein is of white quartz; locally

these may carry economic concentrations of gold, as in

the “Mother Lode” of the Sierra Nevada Foothills,

California. Fewer veins consist of two or more con-

spicuous minerals.

Vein-forming ﬂuids are high-T solutions of H

and usually lesser CO

in which various chemical

species, especially silica, are dissolved. Although some

may be expelled from nearby crystallizing bodies of

volatile oversaturated magma, most ﬂuids are derived

from within the metamorphic rock, either as ground-

water heated by a nearby magma (Figure 4.12), or

liberated from rocks undergoing prograde devolatiliza-

tion reactions during metamorphism. Fluid migration

is governed by gradients in ﬂuid pressure, P

, and by

bulk rock permeability (Section 8.5). Devolatilizing

carbonate-bearing rocks subjected to prograde meta-

morphism tend to develop grain-scale permeabilities

so that ﬂuids migrate by porous ﬂow along grain

boundaries; consequently, veins are scarce. On the

other hand, in silicate rocks permeabilities tend to be

less and as P

increases during prograde reactions it

can exceed the tensile strength of the rock, creating

hydraulic fractures (Section 8.2.1) that can either cut

across host rock foliation or insinuate along it. The

ﬂuid immediately moves into the crack where minerals

may precipitate as the solution is chemically or ther-

mally perturbed by the wall rock. This crack-seal

mechanism can be repeated many times in the his-

tory of a metamorphic terrane, producing multistage

vein growth. Vein fabric and mineral assemblages can

potentially provide insights into prograde mineral reac-

tions; vein minerals having restricted P–T stabilities

(such as andalusite, kyanite, jadeite) are especially

useful in recording the character of advecting ﬂuids

associated with metamorphism.

15.3 GRAPHICAL REPRESENTATION OF

MINERAL ASSEMBLAGES IN

COMPOSITION DIAGRAMS

A composition, or chemographic, diagram that can

represent the correspondence between the stable

mineral assemblages in rocks and their bulk chemical

composition is of paramount importance and utility in

metamorphic petrology. Such diagrams permit easy visu-

alization of what minerals can stably coexist together

under restricted conditions of P, T, and ﬂuid composi-

tion in a rock of particular chemical composition, as

well as contrasts in mineral assemblage depending on

contrasts in rock compositions. Comparisons can be

made, for example, among assemblages in different

zones in pelitic rocks in a regional Barrovian terrane by

using several diagrams to represent the differing con-

ditions of their metamorphism. Although the assump-

tions and simpliﬁcations made in the construction of

composition diagrams (e.g. Thompson, 1982; Winter,

2001, pp. 487–93) compromise thermodynamic rigor,

which means that they cannot represent exact states

of equilibrium in all cases, their advantages outweigh

shortcomings.

15.3.1 Fundamentals

The basis of a chemographic diagram is rooted in the

phase rule (Section 5.1.1). In divariant mineral assem-

blages where both P and T can be freely and inde-

pendently varied without any perturbation in the state

of equilibrium, compositional variables (X) are also

usually signiﬁcant, increasing the degrees of freedom,

so that F  2. The common recurrence of certain min-

eral assemblages in metamorphic terranes around the

world is consistent with at least divariant equilibria

where P, T, and X were independent variables, such as

within a particular mineralogical zone in a regional

466 Igneous and Metamorphic Petrology

15.23 Veins of white quartz in schist.