holness-saywer-2008-jpet

8/2/2019 holness-saywer-2008-jpet

1/21

On the Pseudomorphing of Melt-filled PoresDuring the Crystallization of Migmatites

MARIAN B. HOLNESS1* AND EDWARD W. SAWYER2

1DEPARTMENT OF EARTH SCIENCES, UNIVERSITY OF CAMBRIDGE, DOWNING STREET, CAMBRIDGE CB2 3EQ, UK

2SCIENCES DE LA TERRE, DE PARTEMENT DES SCIENCES APPLIQUE ES, UNIVERSITE DU QUE BEC A' CHICOUTIMI,

CHICOUTIMI, QUE BEC G7H 2B1, CANADA

RECEIVED SEPTEMBER 13, 2007; ACCEPTED MAY 1, 2008ADVANCE ACCESS PUBLICATION MAY 30, 2008

Pseudomorphs of melt-filled pores, recognized by their generally cus-pate shape, are used as diagnostic for the former presence of partial

melt. They are commonly observed in migmatites from the mid- to

deep crust although they occur in the smaller pores in migmatites

from shallower levels (1^2 kbar). The pseudomorphing of melt-

filled pores is controlled by the kinetics of nucleation and is a conse-

quence of the greater supersaturation required for nucleation in a

small pore compared with a larger one.We examine three migmatites

in detail: a contact metamorphosed cherty band from an iron forma-

tion; an Archaean regional granulite from an accretionary prism;

and an amphibolite-facies sample from the roots of an Archaean

mountain chain.The greater undercooling required for nucleation in

progressively smaller pores is recorded by the composition of plagio-

clase pseudomorphs. A study of dihedral angles at the corners of pseudomorphed pores demonstrates that melt^solid textural equilib-

rium was probably attained only in the contact aureole.The regional

granulite preserves an almost unmodified reaction-controlled

melt distribution, with little evidence for either melt^solid textural

equilibration or solid^solid re-equilibration, whereas the reaction-

controlled melt distribution in the regional amphibolite-facies exam-

ple has been modified by a partial approach to solid^solid textural

equilibrium. It is not clear whether the differences in dihedral angle

population are due to differences in uplift and exhumation rates or

due to the presence of H2O on grain boundaries.

KEY WORDS: migmatite; microstructure; dihedral angle; crystallisation

I N T R O D U C T I O N

Whereas it is relatively easy to map melt distribution

in migmatites once the melt has amalgamated into

centimetre-scale pockets or leucosomes (e.g. Brown, 2007),it is not always straightforward to infer the former presence

of melt in rocks that have either undergone very small

degrees of partial melting or in which the melt has been

almost completely expelled. The grain-scale liquid distri-

bution in melt-poor crustal rocks has commonly been

inferred from the presence of highly cuspate grains with

low dihedral angles (Fig. 1a). Their shape and the low dihe-

dral angles are reminiscent of melt-filled pores in experi-

mental charges (e.g. Jurewicz & Watson, 1985) and it has

been suggested that the cuspate grains nucleated within,

and infilled, melt-filled porosity (Platten, 1982, 1983;

Pattison & Harte,1988; Harte et al., 1991; Riller et al., 1996;

Holness & Clemens, 1999; Sawyer,1999, 2001; Rosenberg &Riller, 2000; Marchildon & Brown, 2002). Such grains have

been termed interserts (Rosenberg & Riller, 2000), pseudo-

morphs after melt (Marchildon & Brown, 2002) or melt

pseudomorphs (e.g. Harte et al., 1993; Clemens & Holness,

2000; Brown, 2001; Walte et al., 2005). The process of pseu-

domorphing of melt-filled porosity and the conditions

under which it occurs are not well understood. Why do

some interstitial grains form perfect pseudomorphs of

a melt-filled pore (clinopyroxene in the gabbroic example

shown in Fig. 1b and c), with no overgrowth of the pore

walls, whereas other examples show the (multiply satu-

rated) liquid being replaced by a poly-phase intergrowth

(clinopyroxene and plagioclase in the gabbroic exampleshown in Fig. 1d)?

The mineral forming the pseudomorph is commonly dif-

ferent from that of the pore walls (Fig. 1a^c). In quartzo-

feldspathic crustal migmatites, it is feldspar that forms the

pseudomorphs in quartz-dominated volumes, whereas in

*Corresponding author. E-mail: [email protected]

The Author 2008. Published by Oxford University Press. All

rights reserved. For Permissions, please e-mail: journals.permissions@

oxfordjournals.org

JOURNAL OF PETROLOGY VOLUME 49 NUMBER 7 PAGES1343^1363 2008 doi:10.1093/petrology/egn028


2/21

feldspar-dominated volumes it is quartz (e.g. Harte et al.,

1991; Riller et al., 1996; Rosenberg & Riller, 2000). Petro-

graphic studies of migmatites most commonly describe

quartz, alkali and plagioclase feldspar pseudomorphs (e.g.

Sawyer, 2001; Marchildon & Brown, 2002), although other

minerals (such as cordierite and biotite) may also form

pseudomorphs. Uncommonly the pseudomorph is formed

of the same mineral as one or both of the walls, although

it may have a different composition (e.g. plagioclase,

Sawyer, 2001).

The composition of single pseudomorphs is unlikely to

represent the composition of the last liquid in that pore:

even if other components of the liquid solidified on

the walls of the pore itself, some of the liquid must have

solidified elsewhere, forming pseudomorphs themselves or

overgrowths on existing grains. The formation of pseudo-

morphs therefore necessitates mass transport through the

pore network on at least the grain scale.

If averaged over a centimetre length scale, the composi-

tion of all pseudomorphs approximates that of the final

liquid. Table 1 gives four examples of melt compositions

estimated from point counting melt pseudomorphs within

single thin-sections of meta-sedimentary migmatites; these

are all of broadly granitic composition. Overgrowth of res-

tite grains depletes the bulk composition of the pseudo-

morphs in the phase that forms the dominant restite

phase. The data from the two granulite-facies metagrey-

wackes from Ashuanipi in Table 1 show a low modal

proportion of plagioclase, probably because plagioclase is

the major residual phase in both these rocks: it is easy to

underestimate the amount of overgrowth on restitic plagio-

clase, especially if the overgrowths are thin or in opticalcontinuity with their substrate.

Developing an understanding of the pseudomorphing of

melt-filled porosity necessitates the pseudomorphs being

set into the context of the microstructural development of

migmatites. Migmatite microstructures themselves pose a

considerable interpretative problem, requiring not only an

understanding of the controls on the grain-scale dis-

tribution of melt, both during and after reaction, and an

Fig. 1. (a) Photomicrograph under crossed polars of a migmatite from the Opatica Subprovince, Quebec. Note the elongate cuspate grains ofmicrocline on grain boundaries. Scale bar represents 200 mm. (b) Gabbro nodule from Iceland in which glass-filled pores (black) betweenplagioclase grains are partially filled with clinopyroxene. (Note how the clinopyroxene inherits the angle at the pore corners.) Scale bar repre-sents 100 mm. (c) Gabbro nodule from Iceland in which thick grain boundary melt films (now glass) are replaced by clinopyroxene. Note theapparent absence of any plagioclase overgrowth on the pore walls in both this and in (b). Scale bar represents 200 mm. (d) Gabbro nodule fromIceland, in which large melt-filled pores formed between plagioclase grains are partially filled by plumose intergrowths of clinopyroxene andplagioclase. Scale bar represents 1mm.

Table 1: Composition of pseudomorphed melt pockets,

determined by point counting 1000 points per section

Melt fraction

(%)

Quartz

(%)

Plagioclase

(%)

K-feldspar

(%)

D ul uth a ure ol e p el it e 64 39 35 26

B1-384-20 251 45 24 31

Ashuanipi metagreywacke

BL86-091A

54 37 22 41

Ashuanipi metagreywacke

BL86-044A

42 37 26 37

Source: Sawyer (2001).

JOURNAL OF PETROLOGY VOLUME 49 NUMBER 7 JULY 2008

1344


3/21

appreciation of how and why textures formed during

solidification may vary, but also an understanding of how

these solidification textures may be modified in the sub-

solidus.We thus first set the scene by providing an overview

of melt distribution in partially melted rocks, together with

a discussion of the role textural equilibration may play

in modifying these. A review of published observations isthen used to demonstrate the joint controls on crystalliza-

tion textures of cooling rate and pore size: we will show

that pseudomorphs form in the smallest pores in the slow-

est cooled rocks.

We then present detailedanalysis of partially melted rocks

from a contact aureole and two regionally metamorphosed

migmatite terranes using petrographic observations, con-

ventional geochemical analysis and cathodoluminescence

(CL) imaging to reveal successive stages of micro-

structural evolution during melting and solidification.

We demonstrate how the details of pseudomorph geom-

etry can provide information on the balance of reaction

and textural equilibration during anatexis, and qualita-tive constraints on the rates of cooling after the metamor-

phic peak.

A N A LY T I C A L T E C H N I Q U E S

A N D C H O I C E O F S A M P L E S

Plagioclase compositions and the titanium content of

quartz were determined using wavelength-dispersive elec-

tron microprobe spectrometry at the Department of Earth

Sciences, University of Cambridge, using a Cameca SX100

electron probe microanalyser. For the plagioclase analyses,

operating conditions were an accelerating potential of

15 keV, with a nominal beam current of 10 nA for Na, Al,Si and Ca, and 100 nA for Mg, Sr, Ti, Mn and Fe, and

beam diameter of 5 mm. The quartz analyses used an accel-

erating voltage of 25 keV, with a nominal beam current of

300 nA and a beam diameter of 5 mm. The detection limit

was 7 ppmTi. The PAP corrections procedure was used for

data processing and was run using Cameca Peak Sight

software.

Recent work coupling trace element concentration with

CL in quartz has demonstrated a strong link between con-

centration of Ti and brightness (Wark & Spear, 2005),

which can further be exploited using the TitaniQ thermo-

meter (Wark & Watson, 2006; Wiebe et al., 2007). CL and

back-scatter electron (BSE) images were obtained using a Jeol JSM-820 scanning microscope fitted with a Gatan

MonoCL. Polished thin-sections were scanned at a large

working distance (37 mm) with an accelerating voltage of

15 kV, and a beam current of 10 nA. The BSE aperture was

centred while the CL detector was off-centrethis per-

mitted BSE and CL images to be obtained over a wide

area (3 mm 3 mm) with minimal disruption. The pres-

ence of the CL detector created some astigmatism of the

beam but the effect on the final images at low magnifica-

tion was not significant.

CL spectra were obtained from four quartz generations

discernible by variations in brightness, using a beam cur-

rent of 30nA and a dwell time of 2 s. Spectra wererecorded with the Gatan MonoCL spectrograph, with a

paraboloidal mirror for light collection. The spectra are

shown in Fig. 2, with the detector background omitted for

clarity. The darkest, most weakly luminescent, quartz has

four clearly defined peaks at 320, 430, 580 and 660 nm.

All four peaks are present in more strongly luminescent

quartz but the 430 nm peak is dominant. The intensity of

the 430 nm peak, and thus the brightness of the quartz

under CL, correlates with Ti concentration, consistent

with the conclusions of Wark & Spear (2005) and Wiebe

et al. (2007).

In a texturally equilibrated poly-mineralic rock in which

grains have no preferred crystallographic orientation,dihedral angles form a narrow population with a spread

about the mean determined by the extent of anisotropy of

interfacial energies (Herring, 1951; Laporte & Provost,

2000; Holness, 2006). When the microstructure is out of

equilibrium, the dihedral angle population may no longer

be unimodal, and the standard deviation is likely to be

large (Holness et al., 2005b). Although it is possible to con-

strain the extent of textural disequilibrium from an analy-

sis of apparent angles measured in two dimensions using a

conventional microscope stage (Jurewicz & Jurewicz, 1986;

Elliott et al., 1997), a direct picture of the population of

true, three-dimensional (3-D), dihedral angles can be

gained using a universal stage mounted on an opticalmicroscope (Vernon, 1997). As the details of the population

of true angles are essential to interpreting disequilibrium

microstructures, we measured dihedral angles using a uni-

versal stage mounted on a J. Swift monocular optical

microscope at a magnification of320. The error on each

measurement is of the order of 28, and populations of

90^162 angles were measured for each sample. The univer-

sal stage was also used to identify and obtain a qualitative

Fig. 2. CL spectra obtained for quartz from the Biwabik iron forma-tion sample ELY-6B. The locations of the spots from which thesespectra were obtained are shown in Fig. 10.

HOLNESS & SAWYER PSEUDOMORPHING OF MELT-FILLED PORES

1345


4/21

picture of the distribution of grain boundary melt films,

which become difficult to see if they lie at a low angle to

the plane of a conventional microscope stage.

M E LT D I S T R I B U T I O N O N

T H E G R A I N S C A L EBecause of the long time scales involved it is impossible to

find a regionally metamorphosed anatectic rock preserving

unmodified microstructures dating from the early stages

of melting: primary textures are invariably overprinted.

Access to the earliest stages of reaction, during which the

melt distribution and geometry are totally controlled by

the kinetics of reaction, is generally possible only via

laboratory experiments and by examination of natural

examples of contact metamorphism in which time scales

of the metamorphic event were short. These show that

melting results in the development of progressively thicker

melt films separating reactant grains (Fig. 3a; Butler, 1961;

Mehnert et al., 1973; Bu sch et al., 1974; Paquet & Francois,1980; Platten, 1982, 1983; Maddock, 1986; Holness,

1999; Cesare, 2000; Holness & Watt, 2002; Holness et al.,

2005a; Acosta-Vigil et al ., 2006), which may contain

euhedral grains of peritectic phases (Fig. 3b; Kenah

& Hollister, 1983; Vernon & Collins, 1988; Grant &

Frost, 1990; Brown, 2001; Marchildon & Brown, 2002;

Vernon, 2004).

The volume increase associated with some fluid-absent

melting reactions may generate over-pressure, which can

result in hydrofracture (Fig. 3c and d; Clemens & Mawer,

1992; Connolly et al., 1997; Rushmer, 2001; Holness & Watt,2002; Holness et al ., 2005b), although micro-cracking

appears to be confined to experimental charges and pyro-

metamorphic environments. This is perhaps because only

in such extreme environments is the overstep sufficient to

overcome the natural tendency of systems to buffer along

reactions in P^Tspace.

If the melt production rate is sufficiently slow (with

reaction and deformation rates commensurate with that

of grain-boundary mass transport) melt-bearing systems

can attain textural equilibrium. Melt geometry becomes

a function of porosity or melt fraction (f), the equili-

brium melt^solid dihedral angle () and the extent of

anisotropy of interfacial energies. Equilibrium dihedralangles are controlled by the balancing of the two inter-

facial energies involved: the grain boundary energy

between the two grains of the same phase, gb, and

the energy of the interface between the two different

Fig. 3. (a, b) Brown glassy melt rims in partially melted psammitic rocks from the pyrometamorphic aureole of the Glenmore gabbro plug,Ardnamurchan, Scotland, imaged in plane-polarized light. Feldspar is partially melted, forming a sieve-like texture. Biotite is entirely replacedby spinel, new biotite and mullite. Scale bars represent 200 mm. Note the parallelism of the grain boundary melt films, indicating no relativemovement of liquid and restite (a), and in (b) the acicular orthopyroxene grains, which grew consequent to biotite breakdown. (c) CL image ofpsammite collected 44 cm from the contact of the Traigh Bhan na Sgurra s ill, Isle of Mull. The melt present at the peak of metamorphism wasderived from reaction between quartz and feldspar. Each feldspar grain in this image is surrounded by a brightly luminescing granophyric rim,from which inferred melt-filled cracks emanate, forming a completely interconnected network. Scale bar represents 500 mm. (d) Reacted mus-covite grains in a psammite collected 80 cm from the contact with the Traigh Bhan na Sg urra sill, Isle of Mull, imaged in plane-polarized light.The muscovite is now replaced by mullite, biotite, K-feldspar and spinel, with some prismatic grains of cordierite (now entirely replaced byyellow sheet silicates). The sites of the reacted muscovite grains are linked by melt-filled fractures, which are filled with brown granophyre. Mostof the fractures in this image contain melt but it is only apparent in BSE or CL. Scale bar represents 200 mm.


1346


5/21

phases, gi (Fig. 4a). For isotropic materials satisfies the

relation

gb 2gi cos

2

(Smith, 1948), but because most materials of interest to

geologists have significant anisotropy of interfacial energy(e.g. Kretz, 1966; Vernon, 1968; Laporte et al., 1997), the

equation above is modified to incorporate torque forces,

which act to rotate interfaces to lower energy orientations

(Herring, 1951; Hoffman & Cahn, 1972; Cahn & Hoffman,

1974). This stabilizes planar interfaces at pore corners (e.g.

Kretz, 1966; Vernon, 1968; Laporte & Watson, 1995; Cm|ral

et al., 1997; Lupulescu & Watson,1999; Laporte & Provost,

2000), and a range of possible equilibrium angles, depend-

ing on the relative orientation of the solid grains (Herring,

1951; Laporte & Provost, 2000).

For geologically relevant systems the equilibrium melt^

solid dihedral angle is less than 608 [see Holness (2006)

for a recent review] with a standard deviation of $128

(Laporte & Provost, 2000; Holness, 2006) so the equilib-

rium distribution is one of interconnected channels along

three-grain junctions (Fig. 4b and c; Smith, 1964; Beere ,

1975; von Bargen & Waff, 1986). For isotropic materials

these channels remain open and interconnected even for

vanishingly low porosities but, because all minerals areanisotropic to some extent, there is a threshold level for

interconnectivity of f less than a few volume per cent

(Wark & Watson, 1998; Maumus et al., 2004; Price et al.,

2006; Yoshino et al., 2006).

In a system in which liquid can move freely, complete

equilibration will entail melt flow (e.g. Jurewicz &

Watson, 1985) until it reaches the minimum energy poros-

ity, which ranges from 0 vol. % at 4608 to $23 vol. %,

or the porosity required for complete disaggregation, for

08 (Park & Yoon, 1985; Cheadle, 1989). For the likely

range of equilibrium melt^solid dihedral angles in geologi-

cal systems (20^408, Holness, 2006), the minimum energy

Fig. 4. (a) Rounded pigeonite glomerocrysts in textural equilibrium with the andesitic fine-grained matrix, Mull, Scotland (from Holness,2006). The direction of the force exerted by the energy of the grain boundary is labelled gb, and that of the interface between melt and solid isshown as gi. The dihedral angle,, is a function of the balance between these two energies. Scale bar represents 500 mm. (b) Schematic illustra-tion showing the difference in equilibrium melt topology for dihedral angles above and below the critical value of 60 8 for melt interconnectivity.After Watson & Brenan (1987). (c) Scanning electron microscope image of an experimental charge in which quartz was equilibrated with brineat 94 kbar and 8008C: the dihedral angle under these conditions is less than 608 (Holness,1992). (Note the channels on all three-grain junctions.)The small acicular crystals scattered on the quartz surface formed during the quench. Scale bar represents 20 mm.


1347


6/21

porosity is 10^18 vol. % (Cheadle, 1989). The corollary of

this is that textural equilibration in a melt-present system

in which the liquid is free to move will result in the infiltra-

tion of melt along previously dry three-grain junctions

(Watson, 1982; Daines & Kohlstedt, 1993; Hammouda &

Laporte, 2000).

The extent of textural equilibration depends on the rateand mechanism of deformation during anatexis. Although

diffusion creep does not change the melt distribution, dis-

location creep can result in a significantly different melt

distribution compared with the static case (AveLallemant

& Carter, 1970; Urai, 1983; Jin et al ., 1994; Hirth &

Kohlstedt,1995; Kohlstedt & Zimmerman, 1996; Daines &

Kohlstedt, 1997). Similarly, if the rates of reaction, whether

solidification in igneous rocks or melting in migmatites,

are faster than that of textural equilibration, melt geome-

tries during reaction will be controlled by reaction

kinetics.

Much of our understanding of the process of textural

equilibration has been obtained from studies of plutonicigneous rocks (Hunter, 1987; Holness et al., 2005b; Higgins,

2006). These show that textural equilibration, or matura-

tion, takes place in a series of stages, the earliest of which

is the rotation of grain boundaries in the vicinity of three-

grain junctions to achieve the equilibrium dihedral angle

(Holness et al., 2005b). As a consequence of angle change,

the grain boundaries in the vicinity of the pore corner

develop a step-change in curvature. Because changes in

mean curvature increase the energy of the interface

(Beere , 1975), the pore wall must move to attain constant

mean curvature, at least in isotropic systems (e.g. Holness

& Siklos, 2000). There is thus a complex interplay between

the gradually changing angle at the melt^solid^solid junc-tion and the propagation of a change in grain boundary

orientation and curvature further from the junction.

Because continuous films of liquid on grain boundaries

are stable only for equilibrium melt^solid dihedral angles

of 08 (Smith, 1964), and equilibrium dihedral angles are

always 408 for silicate systems, reaction-generated grain

boundary melt films will neck down to form strings of

melt-filled lenses of a shape controlled by the dihedral

angle. Melt-filled, intra-grain cracks formed by hydrofrac-

ture will heal, leading to arrays of melt-filled inclusions

with a shape controlled by the interfacial energies of the

host crystal.

The details of establishment of the equilibrium dihedralangle at pore corners depends on the geometry of the

initial reaction-controlled melt geometry and on whether

the equilibration takes place in the super- or sub-solidus

(or both). Thin grain boundary melt films, generated

either by reaction between adjacent phases or by inter-

grain fracturing, in a rock with an initially well-

equilibrated solid-state microstructure will initially gener-

ate dihedral angles of$1208 (Fig. 5). As the walls melt back

and the melt rim becomes thicker, the dihedral angles will

tend towards 1808, with a low standard deviation. Melt

films propagating down two-grain junctions are akin to

propagating fractures and likely to have very low dihedral

angles at their tips. Melt^solid^solid dihedral angles in a

melting rock in which melt distribution is dominated

by grain boundary films will therefore form as many asthree separate peaks (Fig. 5c). The rate at which pore junc-

tions will progress towards melt-present equilibrium will

depend on their geometry. The tips of propagating films

are close to melt solid solid equilibrium and will not

change significantly. The junctions of grain boundaries

with thick melt films are farthest from equilibrium, will

have the greatest driving force for microstructural change,

and will therefore undergo rapid adjustment of the pore

walls. The peak at $1208 will migrate relatively slowly

towards lower angles.

M I C R O S T R U C T U R E S F O R M E D

D U R I N G C R Y S T A L L I Z A T I O N

Cooling rate exerts a primary control on the microstruc-

tures formed during crystallization. At one extreme we

find supercooled melt (glass), whereas more slowly cooled

melt crystallizes as increasingly coarse polycrystalline

aggregates as the cooling rate is decreased. This can be

illustrated by a consideration of quartzo-feldspathic rocks

from a series of contact aureoles.

Pyrometamorphism at 120bar around the 50 m dia-

meter gabbro plug at Glenmore, Ardnamurchan, resulted

in glassy melt rims separating quartz and feldspar (Fig. 3a

and b; Butler, 1961; Holness et al., 2005a). Solidification ofmelt rims in the 150 bar aureole of the gabbro plug in

Kinloch, Isle of Rum, and in the 600 bar aureole of the

Traigh Bhan na Sgurra sill (Holness & Watt, 2002)

resulted in a fine granophyric intergrowth (Fig. 6a).

Solidification of melt rims in amphibolitic gneiss metamor-

phosed by the Rum Igneous Complex at 100^200 bar

formed a coarser granophyric intergrowth in which

quartz paramorphs after tridymite are visible (Fig. 6b and

c; Holness & Isherwood, 2003).

A further reduction in cooling rate results in melt rims

on quartz^feldspar grain boundaries being replaced

by either an aggregate of equigranular, equant grains of

feldspar and quartz (termed the string of beads texture;

Holness & Isherwood, 2003) with randomly oriented

grains, or by a highly irregular grain boundary (Fig. 6e

and f). The irregularity of the latter is due to the formation

of extremely coarse granophyric intergrowths, with each

phase having nucleated and grown on its respective side-

wall of the melt rim. Which of these two textures forms is

probably a consequence of the ease of nucleation during

solidification.


1348


7/21

As the cooling rate decreases further, for larger or

deeper contact aureoles and for regional metamorphism,

the extended time scale permits the onset of melt migra-

tion, resulting in the grain-scale draining of melt from its

localized source, the destruction of thick, parallel-sided,

melt films such as those depicted in Fig. 6b, and segrega-

tion of melt into large pockets. This permits the identi-fication of the second major control of crystallization

microstructures: the size of the melt pocket.

In the deeper parts of the contact aureole associated

with the Rum Layered Suite (Holness & Isherwood,

2003), and in that associated with the 3 kbar, 10 km dia-

meter, Ballachulish Igneous Complex (Pattison & Harte,

1988), the largest melt pockets are filled with granophyre

(Fig. 7a) or by oikocrysts with abundant inclusions

(Fig. 7b; Grant & Frost, 1990; Harte et al., 1991). Thick melt

films on quartz^feldspar grain boundaries are replaced by

highly cuspate intergrowths or a string of beads (Fig. 6e

and f), whereas the smallest pores (the thinnest melt rims,

and melt pockets bounded by 3^4 grains) are generally

pseudomorphed by single crystals of the volumetrically

minor phase (Fig. 7c).

For migmatites in the mid- to deep crust, the same

pattern of solidification microstructures is observed.

Smaller melt pockets (typically those bounded by fewer

than four grains) are partially or wholly pseudomorphed

by single grains. In contrast, large melt pockets, or leuco-

somes, crystallize as poly-mineralic aggregates (although

granophyric intergrowths are apparently confined to

shallow to mid-crustal environments). We suggest that this

microstructural progression, from poly-mineralic aggre-

gates in the larger pores to monocrystalline pseudomorphs

of the smaller pores, which can be observed within a single

rock (and therefore at a constant cooling rate), reflects anincreasing barrier to nucleation as the pore size becomes

smaller.

The role of pore size in nucleationBecause an atom on the surface of a small crystal sus-

pended in a solution is in a highly energetic condition com-

pared with an atom of the same material on a planar

surface (i.e. on the surface of an infinitely large, spherical,

crystal) in contact with the same solution, it has a stronger

Fig. 5. (a) Image of Ashuanipi granulite-facies migmatite in whichthe melt phase has been pseudomorphed by K-feldspar. The sketch in(b) shows grain outlines, with melt as white, and the quartz and pla-gioclase as light and dark grey, respectively. Field of view is 0 9 mmwide. The different types of pore^solid^solid junctions should benoted: those corresponding to the tips of grain boundary melt films

have dihedral angles $208, whereas those at the junction of

grain boundaries intersecting melt films at a high angle havedihedral angles of 1808. Those formed at erstwhile solid^solid junc-tions of 1208 have an initial dihedral angle of $1208. (c) Duringmelt-present textural equilibration, the low angles at film tips do notneed to increase much to attain equilibrium dihedral angles of$308,whereas those at high-angle, grain boundary intersections need todecrease by $1508 the driving force is greatest for these highangles. (d) During sub-solidus textural equilibration, the endpointfor dihedral angles is now $1208, and it is the low angles that nowhave the greatest departure from, and hence driving force towards,textural equilibrium.


1349


8/21

tendency to leave the crystal (i.e. to dissolve). To maintain

chemical equilibrium, to permit a correspondingly higher

rate of movement of atoms from the solution onto the crys-

tal, the small crystal needs to be in contact with a stronger

solution than does a larger crystal. The concentration of

solute, C, in equilibrium with a spherical crystal of radius

r is given by

2g

r

RT

Vln

C

C0

(the Gibbs^Thompson, or Ostwald^Freundlich equa-

tion), where g is the energy of the solid liquid

interface, R the gas constant, T the temperature, V

the molar volume of the solid phase, and C0 is the solubi-

lity of a liquid in equilibrium with an infinitely large

spherical crystal or a planar interface (Cahn, 1980;

Adamson, 1990).

This means that the thermodynamics of crystallization

in pores is different from that in a free fluid: most impor-

tantly, a confined fluid can become more supersaturated

before crystallization begins compared with an unconfined

fluid of the same composition (e.g. Bigg, 1953; Melia &

Moffitt,1964; Cahn, 1980; Scherer, 1999; Putnis & Mauthe,

2001). For liquid-bearing rock with a range of pore sizes the

point at which crystals (be they interstitial grains in a plu-

tonic igneous rock or cement in a sedimentary rock) nucle-ate will be determined by pore size. The first observation of

this effect in rocks was reported by Putnis & Mauthe

(2001), who showed that small pores in sandstone remain

empty whereas nearby larger pores are filled with halite

cement. A similar effect is probably implicated in the pre-

servation of the thick grain boundary films of melt in

almost completely solidified crystalline mafic nodules

described by Holness et al. (2007).

Fig. 6. A progression of solidification microstructures developed in grain boundary melt films in quartzo-feldspathic rocks as the cooling ratedecreases. (a) Fine-grained granophyre replacing a feldspar grain in the 600 bar aureole of the Traigh Bhan na Sgurra sill, Isle of Mull,Scotland. Scale bar represents 200 mm. (b) Continuous and parallel-sided granophyric rims separating quartz and feldspar in gneiss fromthe 100^200 bar aureole of the Rum Igneous Complex, Scotland. Scale bar represents 1mm. (c) Close-up view of the granophyric rim from(b), showing the elongate quartz paramorphs after tridymite. Scale bar represents 200 mm. (d) The string-of-beads texture developed in moreslowly cooled, deeper parts of the Rum aureole. Quartz paramorphs after tridymite are no longer present. Scale bar represents 200mm.(e) Highly cuspate quartz^feldspar grain boundaries are the manifestation of extremely coarse-grained intergrowths of quartz and feldspar,in which the two phases nucleate on the walls of the melt rim. Gneiss from the deeper parts of the Rum aureole. Scale bar represents 1mm.(f) Close-up of (e) showing the highly cuspate grain boundaries. Scale bar represents 200 mm.


1350


9/21

T E X T U R A L M O D I F I C A T I O N I N

T H E S U B - S O L I D U SFurther microstructural adjustment towards lower internal

energies may occur in the sub-solidus (Fig. 5d). This pro-cess is much slower than the melt-present case because

mass transfer is achieved by grain boundary diffusion

rather than mass transport through a liquid. Furthermore,

whereas sub-solidus textural equilibration of microstruc-

tures in monomineralic domains is rapid, as it necessitates

only mass movement across grain boundaries (Hunter,

1987), equilibration in polymineralic domains requires

mass transport along grain boundaries and is therefore

sluggish. It is rare to find completely equilibrated solid-

state microstructures in polymineralic rocks.

Typical median angles for solid-state textural equilib-

rium fall in the range 100^1408 (e.g. Kretz, 1966; Vernon,

1968, 1970), with standard deviations of 10^208 (Vernon,

1968, 1970). Therefore it is the pseudomorphed propagating

film tips that are furthest from equilibrium and likely tomigrate the fastest (Fig. 5d). The 1808 angles on thick melt

films will migrate to lower angles relatively slowly.

C O N T A C T - M E T A M O R P H O S E D

Q U A R T Z O - F E L D S P A T H I C

R O C K S AT 2 k b a r

The Lower Proterozoic Biwabik Iron Formation in

Minnesota comprises a cherty and slaty banded iron for-

mation (Morey, 1992) and is part of the Animike Group,

which was metamorphosed at $2 kbar by the Duluth

Igneous Complex, an arcuate mass of troctolitic and

anorthositic intrusions of Middle Proterozoic age locatedin the mid-continent rift of North America (Van Schmus

& Hinze, 1985). Samples of cherty horizons of the Biwabik

Formation were obtained from a drill core through the

intrusion and into the footwall.

The cherty horizons are dominated by coarse-grained

quartz, feldspar, clinopyroxene (with well-developed exso-

lution lamellae of Ca-poor pyroxene), magnetite and ilmen-

ite (Fig. 8a). Coarse-grained granophyric intergrowths are

common and mark centimetre-sized pockets of melt

(Fig. 8b). Quartz-rich parts of the rock are dominated by

rounded quartz grains separated by a network of thin

single crystals of plagioclase (with minor clinopyroxene),

which outline the grain boundaries and form cuspategrains at three-grain junctions. Each plagioclase grain is

approximately one, or at most two, quartz grains in

length. This network occupies $10 vol. % of the rock.

A similar pattern is observed in relatively quartz-poor

regions, in which the Fe-oxides and clinopyroxene form

the network of interserts (Fig. 8a).

A hundred randomly chosen plagioclase^quartz^quartz

dihedral angles form a unimodal population with a

median of 688, and a standard deviation of 1738 (Fig. 9b).

The median is significantly higher, and the standard devia-

tion slightly higher, than expected for melt^quartz equilib-

rium ($208 and $108, respectively; Holness, 2006). For

comparison with solid-state equilibrium, plagioclase^quartz^quartz angles were measured in texturally well-

equilibrated regions of an upper amphibolite-facies rock

(sample EL180, which is described at length below). The

median and standard deviation of a population of 100

angles are 1128 and 1608, respectively (Fig. 9a), comparable

with a mean angle of 1058 (and standard deviation of

1248) measured in granulites from Broken Hill (Vernon,

1968), and also with a median of 1178 measured by

Fig. 7. (a) Granophyric intergrowths of quartz and feldspar, nucleat-ing on euhedral crystals of plagioclase in Lewisian gneiss from theaureole of the Rum Igneous Complex. Scale bar represents 1 mm;crossed polars. (b) Melt pool defined by oikocrystic quartz enclosing

euhedral plagioclase crystals that grew directly from the liquid.Contact metamorphosed Lewisian gneiss, Isle of Rum, Scotland.Scale bar represents 200 mm; crossed polars. (c) Contact metamor-phosed arkose from the aureole of the Ballachulish Complex,Scottish Highlands, under crossed polars. This sample was collected1m from the contact and contains rounded quartz grains with highlycuspate feldspar grains with low feldspar^quartz quartz dihedralangles.The feldspar grain shapes are suggestive of melt pseudomorph-ing. Scale bar represents 200 mm.


1351


10/21


11/21

R E G I O N A L Q UA R T Z O -

F E L D S PAT H I C G R A N U L I T E S

Ashuanipi Subprovince, QuebecThe Ashuanipi Subprovince, lying NE of the Opatica

Subprovince, forms the eastern end of a 2000 km long

meta-sedimentary belt that crosses the entire width of

the Superior Province (Card & Ciesielski, 1986; Percival

et al., 1992) and is believed to be the remains of a single,

late Archaean, accretionary prism (Percival, 1989).

Sample DL96-1006A [previously described by Sawyer

(2001)] is a meta-greywacke that experienced a prolonged

Fig. 9. Populations of true, 3-D dihedral angles, measured with the universal stage. The number of measurements for each population is givenby n. (a) Equilibrium populations, showing the spread of angles as a result of anisotropy of interfacial energies.The melt^plag^plag populationis taken from Holness (2006), whereas the plag^quartz^quartz angles were measured on sample EL180 from regions in which melt-derivedtextures were absent. (b) Plag^plag^quartz angles from the Biwabik iron formation. (c) All K-feldspar angles from the Ashuanipi granulite-facies migmatite. The bimodal population comprises a peak at low angles corresponding to the tips of melt films on (predominantly)plagioclase^quartz grain boundaries, whereas the higher peak corresponds to that on quartz^quartz, or plagioclase^plagioclase, boundaries

intersecting melt films at high angles. The K-feldspar grain is outlined for clarity. (d) All microcline di hedral angles from the upper amphibo-lite-facies m igmatite from the Opatica Subprovince.


1353


12/21

(40^50 Myr, Guernina & Sawyer, 2003) metamorphic

event that peaked in the granulite facies at about 850 8C

and 6^65 kbar (Guernina & Sawyer, 2003; Percival, 1991).

It is a foliated, medium-grained, quartzo-feldspathic

rock containing abundant plagioclase, biotite and ortho-

pyroxene, and with a melt-depleted bulk composition

(Guernina & Sawyer, 2003).

Melt, pseudomorphed predominantly by non-twinned

K-feldspar (which appears only as pseudomorphs), with

subordinate amounts of quartz and plagioclase, is concen-

trated in films and pools around reactant biotite grains(Fig. 12a). Quartz and plagioclase pseudomorphs tend

to be more compact than the K-feldspar, with high dihe-

dral angles against the adjacent quartz and plagioclase

grains. In contrast, K-feldspar pseudomorphs form

extended, cuspate films several grain diameters in length.

Melt films occur mainly on biotite^quartz, biotite^

plagioclase, and plagioclase^quartz grain boundaries

and are associated with rounded, corroded grains of

biotite, plagioclase and/or quartz (Fig. 12a and b), suggest-

ing the reaction

biotite plagioclase quartz orthopyroxene

melt ilmenite

(Sawyer, 2001). Melt films on plagioclase^plagioclase and

quartz^quartz boundaries are uncommon, are generally

in close proximity to biotite grains, and are always a con-

tinuation of a thick melt film on a plagioclase^quartz or

biotite^quartz grain boundary. The thinnest films termi-

nate along an inter-phase boundary (Fig. 12c and d) andisolated thin lenses may be present. Slightly thicker films

are usually only a single grain width long and terminate

at a three-grain junction. The thickest films are several

grain diameters in length, with several melt-free grain

junctions terminating at the film wall (Fig. 5a and b).

Quartz is generally featureless under CL, apart

from a slightly brighter grain core compared with

the rim, and the presence of numerous, late-stage,

Fig. 10. Main image shows a composite of CL images of the Biwabik sample ELY-6B from the region shown in Fig. 6c, with single quartz grainsoutlined in white (the quartz component of the granophyre is outlined only where it forms an overgrowth on restitic grains). The four distinctquartz generations, distinguishable by variations in luminescence, should be noted. Profiles of Ti concentration measured along the four tra-verses shown as white lines in the CL image were converted to minimum temperature estimates using the thermometer of Wark & Watson(2006) and are shown as plots to the right of the CL image, each colour-coded for the four generations of quartz. Spot analyses of plagioclaseare shown, calculated to molar per cent albite, on the CL image. The positions of the CL spectra shown in Fig. 1 are shown as spots labelled

x, y and z, located both on the CL image and in the temperature traverses (a and b).


1354


13/21

dark fractures (Fig. 13). The absence of the kind of detailobserved in the Biwabik samples is probably because Ti

variations are able to decay by diffusion during extended

high-temperature episodes.

Dihedral angles at the junctions of the K-feldspar melt

pseudomorphs with either plagioclase or quartz were mea-

sured in sample DL96-1006A, which is typical of the ter-

rane (Sawyer, 2001). The angle population is strongly

bimodal (Fig. 9c). The bimodality is controlled by the

geometry of the melt film. The terminations of the thinnest

melt films have low dihedral angles, forming a well-defined

peak at about 208 (examples are labelled a in Fig. 14). The

slightly thicker melt films, which terminate at three-grain

junctions, have angles generally below about 608, whereas

the junctions of melt-free grain boundaries, which termi-

nate at the thickest melt films, form a well-defined peakat $1008 (examples of which are labelled b in Fig. 14).

Although this does correlate with mineralogy, as low-

angle, terminating melt films almost invariably occur on

plagioclase^quartz grain boundaries whereas it is predom-

inantly quartz^quartz grain boundaries that terminate at

the thickest melt films (Figs 12 and 14), we do not think

that these variations are caused by interphase variations

in interfacial energy. This is because equilibrium dihedral

angles for melt^quartz compared with melt^plagioclase,

and quartz plagioclase^plagioclase compared with

plagioclase^quartz^quartz systems are very similar

(Vernon, 1968; Holness, 2006).

Opatica Subprovince, QuebecThe Opatica Subprovince is a 200 km wide belt of amphi-

bolite-facies plutonic gneisses from the southeastern part of

the Superior Province, Quebec. It is interpreted as the

deeply eroded interior of an Archaean mountain chain

and represents a section of reworked Archaean crust with-

out major tectonic disruption (Sawyer & Benn, 1993). The

highest-grade, and later exhumed, migmatites are in the

centre of the Subprovince and record maximum tempera-

tures of 760^8108C at a pressure of 6^7 kbar (Sawyer,1998).

Sample EL180, collected from approximately 508330N

and 778330W, is a typical residual grey gneiss containing

plagioclase, green hornblende and biotite, with accessory

apatite and zircon [see Sawyer (1998) for a bulk composi-tion analysis]. Titanite is abundant and forms strings of

small grains within aggregates of biotite and amphibole.

Orthopyroxene is absent. Biotite grains are commonly par-

tially chloritized, and are deformed by the adjacent growth

of lenses of untwinned K-feldspar, many of which are par-

tially albitized. Both chloritization and lens growth are

indicators of low-temperature migration of hydrous fluids

(Holness, 2003), which therefore significantly postdate for-

mation of the migmatites.

Solidified melt is pseudomorphed by microcline, plagio-

clase and quartz. Extended, fine-grained, poly-mineralic

aggregates commonly occur along boundaries between

larger grains (Fig. 15a), and we interpret these as solidifiedmelt rims. The quartz and plagioclase grains tend to be

equant, with 1208 dihedral angles against adjacent quartz

and plagioclase grains (Fig. 15b). Only the microcline

grains are cuspate and elongate, with dihedral angles

51208, reminiscent of the interserts described elsewhere

(Fig. 15b and c), but even the microcline pseudomorphs

are relatively rounded and equant compared with those in

the Ashuanipi granulite (Fig. 12).

Fig. 11. (a) Plagioclase composition as a function of width of theplagioclase pocket in sample ELY-6B from the Biwabik Formationin the aureole of the Duluth Igneous Complex. (b) Detail of (a) forpockets smaller than 200 mm. The grey area shows the range of com-positions observed in the larger pockets with the average (An508)shown by the horizontal line. (c) BSE image showing the position ofeach analysis together with the measured plagioclase composition.The most albitic plagioclase is found in the thinnest part of the filmsand at the film tips.


1355


14/21

Melt is not concentrated adjacent to biotite or amphibole

but is evenly distributed, forming thick grain boundary

films on plagioclase^plagioclase, plagioclase^quartz,

plagioclase^biotite and quartz^biotite grain boundaries,

together with compact, triangular pools at quartz or feld-

spar three-grain junctions: these are rounded with blunt

apices. Thin melt films are uncommon, and intra-grain

cracks are rare (only one example was observed).

The melting reaction is not as clear-cut as in the

Ashuanipi rocks. However, the dearth of melt around bio-

tite grains, the absence of peritectic phases, and the con-

centration of melt rims on plagioclase^quartz grain

boundaries, suggest that it may have been

Quartz Plagioclase K-feldspar H2O Melt:

The confinement of K-feldspar to melt pseudomorphs sug-

gests not only that all of the original K-feldspar grains have

been melted, but also that significant melt loss has

occurred (see Sawyer, 1998). The source of the hydrous

fluid that drove melting is uncertain.

In a similar fashion to the Ashuanipi samples, the

Opatica rocks do not reveal any detailed information in

CL images. Once again, the study concentrated on

dihedral angles formed at the corners of melt pores pseu-

domorphed by K-feldspar. The dihedral angle population

is bimodal with peaks at $708 and $1008 (Fig. 9d). Angles

below 508 are rare. In a similar manner to DL96-1006A,

the lower peak corresponds to the tips of grain boundary

melt films and the corners of pores at three-grain junc-

tions, whereas the higher peak corresponds to grain

boundaries truncated by melt films.

T H E D E V E L O P M E N T O F P O R E

P S E U D O M O R P H SThe best information about the development of pore pseu-

domorphs is preserved in the cherty bands in the Biwabik

Formation. The four generations of quartz shown in Fig. 10

record metamorphism, anatexis and solidification of the

protolith. The last episode of quartz growth, to form the

granophyre, occurred at a temperature consistent with

melt solid chemical equilibrium in the Qtz^Ab^H2O

system at 2 kbar (Johannes, 1978; although he presented

data only for 5 kbar). The earlier growth of highly lumi-

nescent quartz may have occurred during a period of

Fig. 12. Microstructures in Ashuanipi granulite-facies migmatite, DL96-1006A. (a) Back-scattered electron image, showing the distribution ofK-feldspar in thin grain boundary films associated with biotite (elongate pale grey grain). Note the termination of these films part-way alongplagioclase^quartz grain boundaries, and the rounded and resorbed nature of plagioclase and quartz within K-feldspar pools. Scale bar repre-sents 50 mm. (b) Photomicrograph under crossed polars. The K-feldspar-filled melt pools surround corroded biotite grains, with much of themelt forming elongate pools parallel to the margins of the biotite grains. Scale bar represents 200 mm. (c) Back-scattered electron image demon-strating the general absence of melt films along plagioclase^plagioclase and quartz^quartz grain boundaries. Scale bar represents 50 mm.(d) Photomicrograph under crossed polars. Note the tapering melt films ending part-way along plagioclase^quartz grain boundaries and thecorroded nature of the quartz and plagioclase. Scale bar represents 200 mm.


1356


15/21

lower H2O activity (and higher melt temperatures);

the details of this remain unclear.

The quartz overgrowths on the restitic grains bounding

the largest pores show that at the metamorphic peak the

melt was saturated only in quartz. This melt occupied

$20 vol. % of the quartz-rich regions (of which about

half now comprises feldspar) and also formed much

larger, scattered pockets (Fig. 10). During cooling, over-

growths formed on the quartz restite until the liquid

moved onto the quartz^feldspar cotectic. At this point feld-

spar nucleated, and quartz and feldspar grew simulta-neously to form granophyric intergrowths in the largest

pores. CL images of the largest granophyre-filled pores

demonstrate that undulose quartz^feldspar boundaries sig-

nify simultaneous crystallization of quartz and feldspar. In

contrast, the smooth quartz^feldspar grain boundaries in

the smaller pores denote sequential growth, with feldspar

(or clinopyroxene or Fe-oxides in other parts of the rock;

Fig. 8a) crystallizing alone, following quartz overgrowth

on the pore walls. The critical pore width at which this

change occurs is $250 mm (Fig. 8c). It should be noted

that in some pores quartz continued crystallizing until the

pore was completely filled (Fig. 10).

There is no kinetic barrier to overgrowth of pore walls,

but crystallization of the phases not forming the pore

walls requires homogeneous nucleation. Although thisbecomes easier as overgrowth on the walls proceeds

(which increases the concentration of the other phases in

the remaining liquid), the smaller the pore, the greater

the supersaturation required for nucleation. Nucleation

of the minor phase therefore cannot occur until the tem-

perature has decreased to create sufficient undercooling to

overcome the barrier to nucleation. Once nucleation does

occur, the small pores are pseudomorphed by a few grains,

which spread out in the porosity over 1^2 restitic grain dia-

meters. In the case of the Biwabik migmatite, this plagio-

clase growth must have been accompanied by quartz

overgrowth elsewhere, as the rejected quartz component

migrated through the remaining porosity away from thegrowing feldspar.

Pore-size-controlled nucleation delay of the plagioclase

in the Biwabik sample, ELY-6B, is consistent with the gen-

eral increase in Na content of plagioclase in progressively

smaller pores (Fig.11). Within any single pore, the variation

in Na content shows that growth occurred first in the

widest region and subsequently propagated into the smal-

lest extremities (Fig. 11c). When all the data are taken as a

Fig. 13. (a) Back-scattered electron image of Ashuanipi granulite-facies migmatite DL96-1006A with corresponding region imagedunder CL (b). Note the absence of fine detail, the generally moreluminescent cores to the quartz grains and the fine tracery of non-luminescing fractures associated with late-stage hydrous fluids.

Scale bar represents 50 mm.

Fig. 14. Photomicrographs of Ashuanipi migm atite, DL96-1006Ashowing examples of low dihedral angles preserved at the tips ofgrain boundary melt films (examples are marked a in both images),whereas high angle junctions of grain boundaries between non-reactants (i.e. intra-phase boundaries such as quartz^quartz) withmelt films have high dihedral angles of 120^1808 (examples aremarkedbin both images). Scale bars in both imagesrepresent 200mm.


1357


16/21

whole it appears that significantly delayed nucleation

occurs in melt films below about 40mm thickness. For

ELY-6B, reduced solidification rates associated with

delayed feldspar nucleation in the small pores permitted

attainment of smoothly curved pore walls and (perhaps)

also the equilibrium melt^quartz^quartz dihedral angle

($208, Holness, 2006). The change from Ab50 to Ab90corresponds to a solidus temperature decrease of 308C in

the Qtz^Ab^An^H2O system at 5 kbar (Johannes, 1978).

Although the cooling rate of the aureole of the Duluth

Complex is not known, it is likely that a decrease intemperature of several tens of degrees would take longer

than the few hundreds of years thought to be required to

equilibrate a melt-bearing rock with a 1mm grain size

(Cheadle, 1989; Holness & Siklos, 2000).

A critical issue to address is the preservation of melt^

solid dihedral angles in pseudomorphed pores. In the

Biwabik rock, nucleation delay appears to have promoted

melt^solid textural equilibration, and it is these equilib-

rium angles that are retained by the plagioclase pore-

filling. However, in the Ashuanipi granulite, and perhaps

also in the Opatica migmatite, dihedral angles recorded

by the K-feldspar pseudomorphs are far from textural

equilibrium. This implies both that any nucleation delaywas insufficient to promote melt solid textural equi-

libration before the melt was replaced by K-feldspar, and

also that overgrowth on the walls of K-feldspar-filled

pores in the vicinity of the pore corners was minimal.

To interpret the preserved angle population in terms of

melt distribution and subsequent cooling history it is

essential to be certain that this population accurately

reflects melt^solid angles, rather than being an artefact

of overgrowth on the pore walls. Although a mechanism

for pseudomorphing based on the kinetics of nucleation

in small pores can be used to explain the observed micro-

structure in the Biwabik migmatites, there is no corre-

sponding information on the progressive stages ofsolidification of either the Ashuanipi or the Opatica

migmatites.

The Opatica sample described here contains a pseudo-

morph population that can plausibly account for a granitic

liquid (i.e. quartzplagioclasemicrocline). All three

phases crystallized together in larger pools and the thicker

melt rims, whereas isolated single grains infill the smaller

pores and thinner melt rims. Restitic plagioclase grains

commonly have either overgrowths or elongate extensions

into adjacent inferred melt pools, discernible by a variation

in extinction position (Fig. 15a). In contrast, the Ashuanipi

granulite is dominated by K-feldspar pseudomorphs.

The absence of preserved CL information means that wecannot be sure about the quartz pore walls, but plagioclase

pore walls bounding K-feldspar pseudomorphs appear

optically to be compositionally uniform, suggesting that

at least the plagioclase has not been overgrown. It is

not clear why some melt-filled pores in the Ashuanipi

migmatite were apparently perfectly pseudomorphed by

K-feldspar, with the plagioclase and quartz components of

the melt crystallizing elsewhere.

Fig. 15. Photomicrographs of Opatica amphibolite-facies migmatiteEL180. (a) An annealed and recrystallized string-of-beads texturedeveloped between reacting grains of plagioclase and quartz. Thelarger grains are interpreted as restite and are marked qtz and plag,whereas the crystals solidified from thick grain boundary melt filmsare marked m (microcline), q (quartz) and p (plagioclase). Note howthe microcline within the thinner melt film (bottom left-hand corner ofimage) is elongate and cuspate, in contrast to the quartz and plagio-clase grains within the thicker melt film. The plagioclase restite grainat the top centre of the image has an extension with a slightly differentextinction position separating the two white quartz restite grains.Thisextension nucleated on the plagioclase restite grain during solidifica-tion of the thick melt film. Scale bar represents 200 mm. (b) String ofquartz grains forming a poly-crystalline pseudomorph of a grainboundary melt film. The asterisk marks the spot where four quartzgrains meet, two of which are inferred to have crystallized from melt.The two three-grain junctions have 1208 angles, demonstrating sub-solidus textural equilibration. The microcline grains are cuspate.

Scale bar represents 200mm. (c) Microcline pseudomorphs after meltfilms. Note the cuspate shape denoting limited sub-solidus texturalequilibration, although the dihedral angles at the corners haveincreased somewhat in the sub-solidus. Scale bar represents 200 mm.


1358


17/21

I N T E R P R E T A T I O N O F

P S E U D O M O R P H E D P O R E S H A P E S

Melt^solid textural equilibrium is most likely to have been

attained in the Biwabik sample, ELY-6B, which contained

about 20 vol. % melt within the quartz-rich regions

(Fig. 10). This is close to both the disaggregation porosity

of 23% and the minimum energy porosity of 18 5% for a

dihedral angle of 208 (Cheadle, 1989). Distinction between

the possibilities that the quartz-rich parts of the rock

formed a crystal slurry at the metamorphic peak, or that

their microstructure records complete textural equilib-

rium, with excess melt expelled to form large pockets

(such as that shown in Fig. 10; Jurewicz & Watson,1985) is

difficult, as the melt percentage is variable on the milli-

metre scale.

If textural equilibrium were achieved in rocks con-

taining $5 vol. % melt, inferred for the Ashuanipi and

Opatica samples, melt solid dihedral angles of $208

would lead to most of the melt residing in tubular channels

on three-grain junctions (Smith, 1964; Beere , 1975). A ran-

domly oriented population of such junctions will be predo-

minantly sectioned at high angles to their length, so

tubular melt pockets would generally appear as cuspate

pockets at three-grain junctions. The observed dominance

of elongate pockets on two-grain junctions in both EL180

and DL96-1006A cannot be accounted for by random

cross-sectioning (see Wark et al., 2003) but must reflect a

dominance of grain-boundary melt films, suggesting tex-

tural disequilibrium on the grain scale (even if equilibrium

dihedral angles are established on a smaller scale). Melt

films are most common on grain boundaries between reac-

tant phases, suggesting that the grain-scale melt distribu-

tion is controlled predominantly by reaction kinetics.Reported examples of dihedral angles associated with

pore pseudomorphs (Holness & Clemens,1999; Rosenberg

& Riller, 2000; Marchildon & Brown, 2002) only rarely

form populations consistent with either melt^solid (e.g.

Rosenberg & Riller, 2000) or solid^solid textural equilib-

rium. This is in agreement with the results presented

here for the contact metamorphosed rock from the

Biwabik Formation, in which the unimodal plagioclase^

quartz^quartz dihedral angle population (Fig. 9b) is inter-

mediate between melt solid and solid solid textural

equilibrium (Fig. 9a). A comparison with partially

re-equilibrated angle populations in olivine cumulates

(Holness et al., 2005b), in which a similar, intermediate,

unimodal population occurs in clinopyroxene oikocrysts,

suggests that the present dihedral angle population in the

Biwabik Iron Formation sample represents the partial

approach to solid-state textural equilibrium of an inherited

fully equilibrated population of melt^solid dihedral angles,

potentially providing a measure of the cooling rate in the

sub-solidus.

On the assumption that the angles preserved in the

Ashuanipi granulite reflect those inherited from the melt

with no modification by pore-wall overgrowth, their

strongly bimodal population is also clearly out of any kind

of textural equilibrium, supportive of the wider disequilib-

rium recorded by the dominance of grain boundary reac-

tion rims. As the films on plagioclase^quartz boundariesare commonly adjacent to reacting biotite grains, these

films probably represent a propagating reaction front. The

angle at propagating fronts is likely to be low and so melt^

solid equilibration, at least in the immediate vicinity of the

tip, would require relatively little adjustment. The absence

of isolated K-feldspar lenses on plagioclase^quartz grain

boundaries (see Fig. 1b) demonstrates that equilibration of

the films, either melt^solid or solid^solid, did not occur.

The 908 peak, which represents the evolution of

reaction-controlled geometries with an initial 1808 angle,

is below that expected for solid^solid textural equilibrium

(with a median of $1208), suggesting that these angles

were adjusting towards melt^solid equilibrium. The anglepopulation comprises a significant proportion of angles

between the two main peaks and these generally corre-

spond to melt film geometries that appear to follow origi-

nal grain boundaries, where the initial angle was likely to

be in the range 1208551808 (Fig. 5).

Melt-filled pores pseudomorphed by K-feldspar in the

Ashuanipi granulite thus appear to record a snapshot of

melt geometries preserved at or close to the metamorphic

peak, with only minor modifications in the sub-solidus.

Melting occurred on grain boundaries between reacting

phases, concentrated near biotite grains and propagating

out along plagioclase^quartz grain boundaries. Reaction

occurred sufficiently fast that equilibration of the micro-structures was not possible apart from perhaps establish-

ment of the equilibrium dihedral angle in the immediate

vicinity of pore corners that already had low angles.

The geometry of the pseudomorphed melt in the

Opatica amphibolite-facies migmatite is very different

from that in both the Biwabik chert and the Ashuanipi

granulite. Although thin cuspate interserts are common,

much of the melt has been replaced by poly-mineralic

aggregates on grain boundaries (Fig. 15a and b). These are

highly reminiscent of the string-of-beads texture shown in

Fig. 6d, and we suggest that their coarser grain size and

more equilibrated texture (i.e. straight grain boundaries

and 1208

triple junctions) are due to annealing and equili-bration in the sub-solidus. The melt films in the Opatica

migmatite that solidified as poly-crystalline aggregates are

significantly thicker than those replaced by microcline

alone (Fig. 15a and b), reflecting the greater ease of nuclea-

tion within the larger pores.

The population of angles preserved at the corners of

microcline-filled melt pores in the Opatica amphibolite-

facies migmatite shows the same peak at 908 seen in the


1359


18/21

Ashuanipi sample, but the lower peak has a median of 608,

instead of 208. The large driving force for grain boundary

migration near the tips of pseudomorphed reaction-

generated melt films permitted substantial approach to

solid^solid textural equilibrium, whereas the inherited

angles at the terminations of intra-phase boundaries at

the walls of thick melt films did not evolve as much,because of their generally higher angles.

The relatively uncommon, isolated, pseudomorphs filled

by quartz and plagioclase have very different shapes from

those filled by microcline, with angles close to, or in sub-

solidus textural equilibrium (Fig. 15b). This is probably

because the adjacent grains are quartz and plagioclase

and so microstructural adjustment towards lower energy

configurations requires only mass movement across the

grain boundary (Hunter, 1987). For the microcline (and

the K-feldspar in the Ashuanipi granulite) there is no cor-

responding K-feldspar in the pore walls and so sub-solidus

textural equilibration is much slower. The important corol-

lary of this is that recrystallization and sub-solidus equilib-ration of inherited melt microstructures is dependent on

the minerals forming the pseudomorphs. They are likely

to retain a shape that is recognizable as a pseudomorph of

a melt-filled pore only if they are formed of the minor

component.

The three samples thus show a range of histories

recorded by the geometry of the melt-filled pores. The

Biwabik rock is likely to have attained complete melt^

solid textural equilibrium, not only with equilibrated

melt solid dihedral angles but also possibly with an

equilibrated pore geometry on the grain scale. This was

followed by a partial approach to sub-solidus textural equi-

librium during cooling. The Ashuanipi granulite recordsan almost unmodified reaction texture in which only par-

tial approach to melt^solid equilibrium occurred, and the

Opatica amphibolite records a reaction texture that has

been significantly modified in the sub-solidus.

It is somewhat surprising that textural equilibrium

was not attained, even on the scale of the pore corners,

while melt was present in the granulite-facies regional

(Ashuanipi) migmatite, whereas the Biwabik sample

demonstrates the attainment of melt^solid textural equilib-

rium during contact metamorphism. The contrast may be

due to the slightly higher temperatures in the Biwabik

sample (49008C compared with 820^9008C) or perhaps

to the relative mineralogical simplicity of the Biwabik mig-matite. Additionally, the quartz grains in the Biwabik rock

record a complex thermal history and this may have con-

tributed to a prolonged period of (fairly) constant high

temperature during which reaction was slowed and melt^

solid equilibration was facilitated.

The two regional terranes studied here record very

different extents of sub-solidus textural equilibrium. This

difference may have been a function of thermal history,

with the Ashuanipi granulite cooling much more rapidly

than the Opatica rocks as a consequence of their very dif-

ferent tectonic settings. This is consistent with the presence

of microcline in the latter, and untwinned K-feldspar in

the former. The Opatica Subprovince is thought to be the

deep roots of a mountain chain whereas the Ashuanipi

Subprovince is an ancient accretionary prism. It is highlylikely that the Ashuanipi samples experienced rapid exhu-

mation as a result of tectonism and active faulting in a

terrane that remained at, or close to, a cratonic margin,

whereas the Opatica rocks could have been exhumed only

by erosion.

However, a further possibility is that the difference in

sub-solidus textural equilibration was caused by the differ-

ence in melting reactions (with the Ashuanipi terrane

undergoing dehydration melting whereas melting of the

Opatica rocks was probably fluxed by externally derived

H2O). The presence of H2O is known to facilitate textural

equilibration in the sub-solidus (Holness et al., 1991) and it

is possible that sufficient free H2O was present after melt-ing in the Opatica rocks, whereas the grain boundaries in

the Ashaunipi rocks remained dry. A much slower rate of

sub-solidus textural maturation for the Ashuanipi rocks

would avoid the necessity for an abrupt decrease in tem-

perature coinciding with the almost complete depletion of

K-feldspar that is required to preserve reaction textures in

a residual gneiss. Given the potential of microstructural

analysis for the decoding of migmatite history it is impor-

tant that this question is addressed in future work.

S U M M A R Y A N D C O N C L U S I O N S

Comparison of reported microstructures in migmatitesfrom a variety of crustal environments shows that the

time scale of metamorphism exerts a crucial control on

melt distribution and on the microstructures formed

during solidification. Pseudomorphing of pores to form

cuspate grains and pockets on grain boundaries generally

necessitates temperature^time histories sufficient to permit

melt migration. This is because pore pseudomorphing is

dependent on there being a barrier to nucleationsuch a

barrier occurs in a small pore.

Observation of the distribution and shape of pseudo-

morphed pores in migmatites from a range of crustal

environments demonstrates that the balance between tex-

tural equilibration, melting and deformation is only rarelyin favour of textural equilibrium in melt-bearing rocks,

even in regionally metamorphosed granulites. Observed

low dihedral angles only rarely form populations consis-

tent with melt^solid equilibrium (e.g. Rosenberg & Riller,

2000), and even where melt^solid equilibrium is achieved

at pore corners, the grain-scale distribution may still be

dominated by grain boundary melt films rather than chan-

nels on three-grain junctions.


1360


19/21

Because the smallest pores are those that require the

greatest supersaturation, these will be the last to solidify.

The minerals filling the smallest pores will thus have

more evolved compositions than grains of the same

mineral in larger pores. A further consequence of the

later in-filling of the smallest pores is that these will show

the lowest solid-state dihedral angles. This is because theywill have had a shorter time in which to undergo solid-state

textural maturation compared with the earlier-formed

interstitial grains in the larger pores (see Holness et al.,

2005b). The control of pore size on solidification kinetics

means that, given the same rate of cooling, a migmatite in

which the melt is distributed in many small pores will take

longer to solidify than one in which the same amount of

melt is contained in fewer, larger, pores. This will affect

the rheological response to cooling.

A C K N O W L E D G E M E N T S

We are grateful to Stephen Reed for assistance with theCL and for suggesting that we examine Ti in quartz.Chris Hayward is thanked for help with microprobe anal-yses. Helpful comments from Jon Blundy, Jean-LouisVigneresse, Richard White and Gary Stevens improvedand clarified the manuscript.

R E F E R E N C E SAcosta-Vigil, A., London, D. & Morgan, G. B., VI (2006).

Experiments on the kinetics of partial melting of a leucogranite at

200 MPa H2O and 690^8008C: compositional variability of melts

during the onset of H2O-saturated crustal a natexis. Contributions to

Mineralogy and Petrology 151, 539^557.

Adamson, A. W. (1990). Physical Chemistry of Surfaces, 5th edn. New York:John Wiley.

AveLallemant, H. G. & Carter, N. L. (1970). Syntectonic recrystallisa-

tion of olivine and modes of flow in the upper mantle. Geological

Society of America Bulletin 81, 2203^2220.

Beere , W. (1975). A unif ying theory of the stability of penetrating liquid

phases and sintering pores. Acta Metallurgica 23, 131^138.

Bigg, E. K. (1953). The supercooling of water. Proceedings of the Physical

Society (London) 66B, 688^694.

Brown, M. (2001). Orogeny, migmatites and leucogranites: a review.

Proceedings of the Indian Academy of Science (Earth and Planetary

Science) 110, 313^336.

Brown, M. (2007). Crustal melting and melt extraction, ascent and

emplacement in orogens: mechanisms and consequences. Journal of

the Geological Society, London 164, 709^730.

Bu sch, W., Schneider, G. & Mehnert, K. R. (1974). Initial melting atgrain boundaries. Part II: melting in rocks of granodioritic, quartz-

dioritic and tonalitic composition. Neues Jahrbuch fur Mineralogie,

Abhandlungen 1974, 345^370.

Butler, B. C. M. (1961). Metamorphism and metasomatism of rocks of

the Moine Series by a dolerite plug in Glenmore, Ardnamurchan.

Mineralogical Magazine 32, 866^897.

Cahn, J. W. (1980). Surface stress and the chemical equilibrium of

small crystals: I. The case of the isotropic surface. Acta Metallurgica

28, 1333^1338.

Cahn, J. W. & Hoffman, D. W. (1974). A vector thermodynamics for

anisotropic surfaces, II. Application to curved surfaces. Acta

Metallurgica 22, 1205^1214.

Card, K. D. & Ciesielski, A. (1986). DNAG#1. Subdivisions of

the Superior Province of the Canadian Shield. Geoscience Canada

13, 5^13.

Cesare, B. (2000). Incongruent melting of biotite to spinel in a quartz-

free restite at El Joyazo (SE Spain): textures and reaction character-isation. Contributions to Mineralogy and Petrology 139, 273^284.

Cheadle, M. J. (1989). Properties of texturally equilibrated two-phase

aggregates. Ph.D. thesis, University of Cambridge, 166 pp.

Clemens, J. D. & Holness, M. B. (2000). Textural evolution and partial

melting of arkose in a contact aureole: a case study and implica-

tions. Electronic Geosciences 5, 4.

Clemens, J. D. & Mawer, C. K. (1992). Granitic magma transport by

fracture propagation. Tectonophysics 204, 339^360.

Cm|ral, M., Fitz Gerald, J. D., Faul, U. H. & Green, D. H. (1997).

A close look at dihedral angles and melt geometry in olivine-

basalt aggregates; a TEM study. Contributions to Mineralogy and

Petrology 130, 336^345.

Connolly, J., Holness, M., Rubie, D. & Rushmer, T. (1997). Reaction-

induced microcracking: an experimental investigation of a

mechanism for enhancing anatectic melt extraction. Geology 25,

591^594.

Daines, M. J. & Kohlstedt, D. L. (1993). A laboratory study of melt

migration. Philosophical Transactions of the Royal Society of London,

Series A 342, 43^52.

Daines, M. J. & Kohlstedt, D. L. (1997). Influence of deformation on

melt topology in peridotites. Journal of Geophysical Research 102,

10257^10272.

Elliott, M. T., Cheadle, M. J. & Jerram, D. A. (1997). On the identifi-

cation of textural equilibrium in rocks using dihedral angle mea-

surements. Geology 25, 355^358.

Grant, J. A. & Frost, B. R. (1990). Contact metamorphism and partial

melting of pelitic rocks in the aureole of the Laramie anorthosite

complex, Morton Pass, Wyoming. American Journal of Science 290,

425^472.

Guernina, S. & Sawyer, E. W. (2003). Large-scale melt-depletion in

granulite terranes: an example from the Archean AshuanipiSubprovince of Quebec. Journal of Metamorphic Geology 21, 181^201.

Hammouda, T. & Laporte, D. (2000). Ultrafast mantle impregnation

by carbonatite melts. Geology 28, 283^285.

Harte, B., Hunter, R. H. & Kinny, P. D. (1993). Melt geometry, move-

ment and crystallisation, in relation to mantle dykes, veins and

metasomatism. Philosophical Transactions of the Royal Society of London,

Series A 342, 1^21.

Harte, B., Pattison, D. R. M. & Linklater, C. M. (1991). Field

relations and petrography of partially melted pelitic and

semi-pelitic rocks. In: Voll, G., To pel, J., Pattison, D. R. M. &

Seifert, F. (eds) Equilibrium and Kinetics in Contact Metamorphism: the

Ballachulish Igneous Complex and its Aureole. Berlin: Springer,

pp. 181^210.

Herring, C. (1951). Surface tension as a motivation for sintering.

In: Kingston, W. E. (ed.) Physics of Powder Metallurgy. New York:McGraw^Hill, pp. 143^179.

Higgins, M. D. (2006). Quantitative Textural Measurements in Igneous

and Metamorphic Petrology. Cambridge: Cambridge University Press,

265 pp.

Hiraga,T., Nishikawa, O., Nagase, T., Ak izuki, M. & Kohlstedt, D. L.

(2002). Interfacial energies for quartz and albite in pelitic schist.

Contributions to Mineralogy and Petrology 143, 664^672.

Hirth, J. G. & Kohlstedt, D. L. (1995). Experimental constraints on

the dynamics of the partially molten upper mangle: deformation


1361


20/21

in the diffusion creep regime. Journal of Geophysical Research 100,

1981^2001.

Hoffman, D. W. & Cahn, J. W. (1972). A vector thermodynamics for

anisotropic surfaces, I. Fundamentals and applications to plane sur-

face junctions. Surface Science 31, 368^388.

Holness, M. B. (1992). Equilibrium dihedral angles in the system

quartz^CO2^H2O^NaCl at 8008C and 1^15 kbar: the effects of

pressure and fluid composition on the permeability of quartzites.Earth and Planetary Science Letters 114, 171^184.

Holness, M. B. (1999). Contact metamorphism and anatexis of

Torridonian arkose by minor intrusions of the Rum Igneous

Complex, Inner Hebrides, Scotland. Geological Magazine 136,

527^542.

Holness, M. B. (2003). Growth and albitisation of K-feldspar in crys-

talline rocks in the shallow crust: a tracer for fluid circulation?

Geofluids 3, 89^102.

Holness, M. B. (2006). Melt^solid dihedral angles of common minerals

in natural rocks. Journal of Petrology 47, 791^800.

Holness, M. B. & Clemens, J. D. (1999). Partial melting of the Appin

Quartzite driven by fracture-controlled H2O infiltration in the

aureole of the Ballachulish Igneous Complex, Scottish Highlands.


Holness, M. B. & Isherwood, C. (2003). The aureole of the RumTertiary Igneous Complex, Inner Hebrides, Scotland. Journal of the

Geological Society, London 160, 15^27.

Holness, M. B. & Siklos, S. T. C. (2000). Rates of textural equilibration

in fluid-bearing systems: kinetic limitations to surface-energy con-

trolled permeability. Chemical Geology 162, 137^153.

Holness, M. B. & Watt, G. R. (2002). The aureole of the Traigh Bhan

na Sgurra sill, Isle of Mull: reaction-driven micro-cracking during

pyrometamorphism. Journal of Petrology 43, 511^534.

Holness, M. B., Bickle, M. J. & Graham, C. M. (1991). On the kinetics

of textural equilibration in forsterite marbles. Contributions to

Mineralogy and Petrology 108, 356^367.

Holness, M. B., Dane, K., Sides, R., Richardson, C. & Caddick, M.

(2005a). Melting and melt segregation in the aureole of the

Glenmore Plug, Ardnmurchan. Journal of Metamorphic Geology 23,

29^43.Holness, M. B., Cheadle, M. J. & McKenzie, D. (2005 b). On the uses

of changes in dihedral angle to decode late-stage textural evolution

in cumulates. Journal of Petrology 46, 1565^1583.

Holness, M. B., Anderson, A. T., Martin, V. M., Maclennan, J.,

Passmore, E. & Schwindinger, K. (2007). Textures in partially

solidified crystalline nodules: a window into the pore struc-

ture of slowly cooled mafic intrusions. Journal of Petrology 48,

1243^1264.

Hunter, R. H. (1987). Textural equilibrium in layered ig neous rocks.

In: Parsons, I. (ed.) Origins of Igneous Layering. Dordrecht: D.

Reidel, pp. 473^503.

Jin, Z., Green, H. W. & Zhou, Y. (1994). Melt topology in partially

molten mantle peridotite during ductile deformation. Nature 372,

164^167.

Johannes, W. (1978). Melting of plagioclase in the system AbÂn^H2O

and QzÂbÂn H2O at PH2O 5 kbars, an equilibrium problem.


Jurewicz, S. R. & Jurewicz, A. J. G. (1986). Distribution of

apparent angles on random sections with emphasis on

dihedral a ngle measurements. Journal of Geophysical Research 91,

9277^9282.

Jurewicz, S. R. & Watson, E. B. (1985). The distribution of partial melt

in a granitic system: the application of liquid phase sintering

theory. Geochimica et Cosmochimica Acta 49, 1109^1121.

Kenah, P. & Hollister, L. S. (1983). Anatexis in the Central Gneiss

complex, British Columbia. In: Atherton, M. P. & Gribble, C. D.

(eds) Migmatites, Melting and Metamorphism. Nantwich: Shiva,

pp. 142^162.

Kohlstedt, D. L. & Zimmerman, M. E. (1996). Rheology of partially

molten mantle rocks. Annual Review of Earth and Planetary Sciences

24, 41^62.

Kretz, R. H. (1966). Interpretation of the shape of mineral grains inmetamorphic rocks. Journal of Petrology 7, 68^94.

Laporte, D. & Provost, A. (2000). Equilibrium geometry of a fluid

phase in a polycrystalline aggregate with anisotropic surface ener-

gies: Dry grain boundaries. Journal of Geophysical Research 105,

25937^25953.

Laporte, D. & Watson, E. B. (1995). Exper imental and theoretica l con-

straints on melt distribution in crustal sources: the effect of crystal-

line anisotropy on melt interconnectivity. Chemical Geology 124,

161^184.

Laporte, D., Rapaille, C. & Provost, A. (1997). Wetting angles, equili-

brium melt geometry, and the permeability threshold of partially

molten crustal protoliths. In: Bouchez, J.-L., Hutton, D. H. &

Stephens, W. E. (eds) Granite: From Segregation of Melt to Emplacement

Fabrics. Norwell, MA: Kluwer Academic, pp. 31^54.

Lupulescu, A. & Watson, E. B. (1999). Low melt fraction con-nectivity of granitic and tonalitic melts in a mafic crustal rock

at 8008C and 1 GPa. Contributions to Mineralogy and Petrology 134,

202^216.

Maddock, R. H. (1986). Partial melting of lithic porphyroclasts in

fault-generated pseudotachylytes. Neues Jahrbuch fur Mineralogie,

Abhandlungen 155, 1^14.

Marchildon, N. & Brown, M. (2002). Grain-scale melt distribution

in two contact aureole rocks: implications for controls on melt

localisation and deformation. Journal of Metamorphic Geology 20,

381^396.

Maumus, J., Laporte, D. & Schiano, P. (2004). Dihedral angle

measurements and infiltration property of SiO2-rich melts in

mantle peridotite assemblages. Contributions to Mineralogy and

Petrology 148, 1^12.

Mehnert, K. R., Bu sch, W. & Schneider, G. (1973). Initial melting atgrain boundaries of quartz and feldspar in gneisses and granulites.

NeuesJahrbuch fur Mineralogie, Monatshefte 4, 165^183.

Melia, T. P. & Moffitt, W. P. (1964). Crystalli sation fro m aqueous solu-

tion. Journal of Colloid Science 19, 433^447.

Morey, G. B. (1992). Chemical composition of the Eastern Biwabik

Iron formation (Early Proterozoic), Mesabi Range, Minnesota.

Economic Geology 87, 1649^1658.

Paquet, J. & Francois, P. (1980). Experimental deformation of partially

melted granitic rocks at 600^9008C and 250 MPa confining pres-

sure. Tectonophysics 68, 131^146.

Park, H. H. & Yoon, D. N. (1985). Effect of dihedral angle on the

morphology of grains in a matrix phase. Metallurgical Tranaction s

16, 923^928.

Pattison, D. R. M. & Harte, B. (1988). Evo

Documents

holness-saywer-2008-jpet