JOURNAL OF
PETROLOGY
Journal of Petrology, 2015, Vol. 56, No. 2, 389–408
doi: 10.1093/petrology/egv004
Advance Access Publication Date: 24 February 2015
Original Article
Thermodynamic Modeling of Crustal Melting
Using Xenolith Analogs from the Cortlandt
Complex, New York, USA
Kristin M. Dorfler*, Mark J. Caddick and Robert J. Tracy
Geosciences, Virginia Polytech Institute, Blacksburg, VA 24061, USA
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
*Corresponding author: Department of Geology, Oberlin College, 52 W. Lorain St., Oberlin, OH 44074
USA. E-mail: [email protected]
Received July 13, 2014; Accepted January 20, 2015
ABSTRACT
Emplacement of gabbroic magmas of the Cortlandt Complex, New York, induced rapid (less than 5
years) heating of pelitic schist protoliths (up to 1200 C at 09 GPa). Xenoliths entrained within
the mafic melt experienced significant melting and melt-segregation, now represented as a series
of complex fabrics and structures comprising Si-rich veins and an Al-rich, Si-poor residuum (a
typical assemblage is spinel–magnetite–ilmeno-hematite–sillimanite 6 sapphirine 6 corundum).
Subtle microscopic textures in the residuum include corundum–magnetite symplectites, which are
interpreted to be a result of oxidative breakdown of the hercynite component in spinel during cooling. Aluminous orthopyroxene selvages in veins have typically grown along the contact between
the residuum and quartzofeldspathic ‘melt’. Hybrid monzonorite and monzodiorite crop out near
the xenoliths and are interpreted to represent assimilation by the mafic magma of some of the partial melt produced from the pelitic xenoliths. Equilibrium-melting and batch-melting thermodynamic models track the evolution of the pelitic schist, its partial melt upon heating, and the
residuum from melting and melt extraction. We introduce a ‘filter-pressing’ cooling calculation to
simulate the crystallization of the quartzofeldspathic veins. Modeling results yield the following: (1)
an initial partial melt that, when mixed with the estimated composition of the mafic melt, produces
a hybrid igneous rock consistent with the monzonorite found near the xenolith; (2) a high-T melt
that upon ‘filter-pressing’ crystallization produces a mineral assemblage that texturally and compositionally corresponds to the quartzofeldspathic veinlets retained in the samples within xenolith
interiors; (3) a residual material that, when oxidized, resembles the aluminous assemblages in the
residuum. Modeling of crystallization of the high-T melt predicts early orthopyroxene formation,
with the Al content of orthopyroxene consistent with that of analyzed selvage pyroxene. We
propose that this pyroxene reflects a primary melt crystallization phase rather than reaction-rim
margins of the veins against residual matrix.
Key words: emery formation; migmatization; partial melting; phase equilibria; UHT metamorphism
INTRODUCTION
Detailed studies of country rock–magma interactions
are vital to developing our understanding of processes
such as crustal contamination of mantle-derived mafic
magmas, partial melting in the crust, and melt loss
during high-grade metamorphism. Specifically,
understanding interactions between felsic crustal
xenoliths and mafic magma in plutons can yield
valuable insights into (1) larger-scale processes in
mid- to deep-crustal settings by shedding light on material and heat transfer, (2) the production of granitoid
melts through partial melting of aluminous protoliths,
and (3) the generation of dense silica-deficient residua
that may be an important constituent of the lower
crust (Preston et al., 1999; Markl, 2005; Johnson et al.,
2010).
C The Author 2015. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: [email protected]
V
389
390
that extensive xenolith melting produced a silicic melt,
which was then extracted and mixed with the mafic
magma, locally forming a hybrid monzonorite marginal zone and leaving behind an aluminum-, iron- and
titanium-enriched, silica- and alkali-depleted residuum
within the xenolith. The residuum consists of a mineral
assemblage of spinel–magnetite–ilmeno-hematite–sillimanite 6 sapphirine 6corundum that can vary in mineral proportion and is transected by a network of veins
and veinlets of fine-grained quartzofeldspathic material, interpreted to be remnants of some stage of partial
melt produced within the xenolith. Production of granitoid melt upon heating of Al-rich sediment is well
documented and often linked with magma contamination (Theriault et al., 1997; Preston et al., 1999; Markl,
2005; Johnson et al., 2010; Diaz-Alvarado et al., 2011).
However, clear field evidence of a granitic melt, a contaminated or hybridized magma and an unmodified
parental magma from a single occurrence has yet to
be documented, as far as we know. This study on the
Salt Hill xenolith and surrounding igneous rocks yields
petrological evidence of the pre-intrusion xenolith protolith, estimated parental host magma, aluminous
xenolith residuum, felsic melt produced during heating, and a hybrid igneous rock formed through mixing
of the locally derived partial melt with an apparently
primitive, mantle-derived mafic magma. This follows
from the detailed petrographic work of Tracy &
McLellan (1985) and classic work of Bowen (1928),
which recognized the presence of monzonorite around
inclusions of aluminous material in the Cortlandt
Complex and proposed that they indicate assimilation
of pelite in basaltic magma.
The strategy for this study was to use careful textural
and mineralogical observations and stacked chemical
compositional maps (‘chemical petrography’) for samples from the Cortlandt Complex to constrain and test
quantitative models for isobaric melting of the Salt Hill
metapelitic xenolith during magmatic heating, with the
ultimate goal of understanding the mechanisms
involved in the production of a siliceous melt and aluminum-rich residuum during contact metamorphism.
These fractionated products can ultimately be linked to
more widespread crustal melting and fractionation at
mid- to deep-crustal levels. Equilibrium-melting and
batch-melting models (at varying melt-removal intervals) were tested to simulate the range of possible melting processes and allow discrimination between them
based on observed petrological relationships. Modeling
results produced (1) an initial partial melt that when
mixed with the estimated composition of the mafic host
melt produces a hybrid igneous rock congruous with
the monzonorite found near the xenolith, (2) a high-T
melt that, upon crystallization, produces a mineral assemblage that texturally and compositionally corresponds to the quartzofeldspathic veinlets retained in the
samples from the xenolith interior, and (3) a residual
material that, when oxidized, resembles the assemblages in the residuum.
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
Laboratory-produced samples and experimental
studies (e.g. Hensen & Green, 1970, 1971, 1972;
Newton, 1972; Huang & Wyllie, 1973; Ackermand et al.,
1975; Newton & Wood, 1979; Peterson & Newton,
1990; Patiño Douce & Beard, 1995) are extremely useful in understanding how ultrahigh-temperature (UHT)
products (melts and residuum) form, as both the reactants and products are carefully controlled and the
processes are closely monitored. However, experimental studies cannot fully simulate the evolution of
natural rocks owing in part to obvious scaling, time,
and mass constraints, as well as to the fact that natural
materials may represent a complex interplay between
coeval heating, melt addition or loss, fractionation
processes, wall-rock interaction, and deformation (e.g.
Preston et al., 1999). To some extent this information is
recoverable through careful spatial characterization of
the mineral compositions and textures of natural igneous and metamorphic samples. On the other hand, in
most cases one or more of the protoliths or products
of melting and crystallization in natural environments
is absent or has been subsequently modified.
Although natural samples may not have been produced in an easily interpreted environment, if the products and reactants of UHT processes such as partial
melting are available and the melt is preserved, careful
thermodynamic and mass-balance modeling experiments can replicate the processes and products, allowing for quantitative insight into how the residual
xenolithic material and partial melt evolved, and how
they possibly modified the primary enclosing magma
(e.g. magma mixing or contamination).
Several recent studies have focused on interactions
between felsic xenoliths and mafic magma and resulting implications for crustal contamination and the
overall production of granitic melts, including localities
such as the Moine Supergroup (Preston et al., 1999),
French Massif Central (Maury & Bizouard, 1974),
Bushveld Complex (Johnson et al., 2010, 2011) and
Skaergaard intrusion (Markl, 2005). In addition to
these, another occurrence that shows country rock–magma interactions is the Cortlandt Complex in southeastern New York, a small (60 km2) Ordovician mafic
complex that contains many aluminum-rich xenoliths
that experienced extensive chemical interaction with
initially Mg-rich and presumably mantle-derived basaltic melts during the emplacement of the mafic
pluton.
This study focuses on a single xenolith (roughly
30 m across), which is representative of others in a
xenolith swarm at Salt Hill in the southeastern quadrant of the Cortlandt Complex, 6 km SE of Peekskill,
New York. A clear physical contact between the metapelitic xenolith and the mafic host cannot be identified
owing to quarrying and extensive vegetation, as well
as the possibility that the contact itself may be diffuse
or transitional. Mineral assemblages and textures from
samples collected in a single traverse across the xenolith and into the surrounding igneous rocks indicate
Journal of Petrology, 2015, Vol. 56, No. 2
Journal of Petrology, 2015, Vol. 56, No. 2
PETROLOGY OF THE CORTLANDT COMPLEX,
SALT HILL XENOLITHS AND HYBRID
IGNEOUS ROCKS
Cortlandt Complex and Manhattan Schist
the primary melt was likely to have been similar in composition to McBirney’s Karoo picritic dolerite.
There is no evidence of any systematic variation in
mineral assemblage within the study samples collected
from the core to the rim of the 30 m diameter xenolith
that would indicate a temperature gradient. Therefore it
is assumed that the xenolith and surrounding igneous
rocks achieved thermal equilibrium. In some samples,
spinel is rimmed by orthopyroxene or garnet,
secondary silicates that are probably reaction products
of residuum spinel and silica from adjacent quartzofeldspathic veins. Thermometry using THERMOCALC
(Powell & Holland, 1988, 1994, 2008; Powell et al., 1998)
for equilibria involving spinel, orthopyroxene, garnet,
sillimanite and quartz in the xenolith implies that the
secondary silicates equilibrated at temperatures of
1100 C during cooling, indicating that the peak temperature of the residuum in thermal equilibrium with
the magma must have been greater than 1100 C. The
composition of the evolved fractionate from the original
parental magma composition used in the MELTS simulation at 1200 C (effectively a mildly alkalic gabbro composition) is thus used as an analog for the external
magma surrounding the xenolith at the time of its contact heating and resulting dehydration reactions, including partial melting.
The country rocks that surround the intrusive complex include regionally metamorphosed and complexly
folded Neoproterozoic to Ordovician gneisses, schists,
marbles, and amphibolites. Both the Inwood Marble
and Manhattan Schist immediately surround the complex, the latter abutting the southeastern border of the
complex closest to Salt Hill. The Manhattan Schist as a
map unit contains diverse lithologies including pelitic
schists of varying composition, amphibolites and calcsilicates. It is likely that the protoliths for the Salt Hill
xenolith swarm were similar to the Manhattan Schist
units just outside the contact aureole near Salt Hill.
These metapelites are quartz-rich plagioclase–sillimanite–muscovite–biotite–garnet schists with variable
amounts of Fe–Ti oxides and rare staurolite.
Determining an exact protolith composition for the
xenolith under study is not feasible considering the
variable bulk composition of local schist lithologies and
the extensive metasomatic effects that altered the bulk
composition of the xenolith during melting and melt extraction. Rogers (1911) obtained and reported several
wet chemical analyses of Manhattan Schist; the four
schist analyses he obtained (analyses 24–27; see
Table 1) varied in the range of 4016–6298 wt % SiO2,
1688–295 wt % Al2O3 and 248–1966 wt % Fe2O3.
Additional information on the petrology and structure of the Cortlandt Complex and the surrounding
rocks, including the various xenoliths, has been given
by Friedman (1956), Barker (1964), Caporuscio & Morse
(1978), Domenick & Basu (1982), Ratcliffe et al. (1982),
Waldron (1986) and Johnson & Tracy (1999). The earliest known published work on the Cortlandt Complex is
that of Dana (1881).
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
The Cortlandt Complex is an 60 km2 composite intrusion of six mafic to ultramafic plutons near Peekskill,
New York, about 40 km north of New York City (Fig. 1).
The six plutons are numbered in order of emplacement
as distinguished by crosscutting relations, internal
structure, post-emplacement deformation, mineralogy
and chemistry (Ratcliffe et al., 1982). The complex
formed at c. 446 Ma (Ratcliffe et al., 2012), postdating regional metamorphism and deformation during the
Taconic Orogeny by c. 20–30 Myr. Its contacts cross-cut
steeply dipping Taconic Barrovian regional metamorphic isograds that range from biotite and garnet on
the lower-grade western end near the Hudson River to
sillimanite–muscovite at the eastern end. Given the relatively short west-to-east dimension of the complex
(<8 km), this probably represents a compressed metamorphic gradient that is thought to be unrelated to the
subsequent emplacement of the mafic melts studied
here.
The relatively short time interval between peak regional metamorphism and pluton emplacement suggests that the regional country rocks at the eastern end
of the complex were still at elevated temperatures
(480–550 C) at the time of emplacement (assuming an
average cooling rate of 3–5 C Ma–1 from a peak T of
c. 630 C over 25–30 Myr; Dorfler et al., 2014).
Thermobarometric characterization of a nearby contact
aureole around a contemporaneous small pluton to the
NNE of the Cortlandt Complex indicates that the pressure at the current level of erosion was 09 GPa at the
time of pluton emplacement (Dorfler et al., 2014). In an
unpublished thesis, Waldron (1986) estimated peak conditions for regional metamorphism on the eastern side
of the Cortlandt Complex outside the contact aureole
(sillimanite þ muscovite grade) to be 615–645 C and
064–072 GPa, using thermobarometers available at the
time.
The youngest pluton of the series (pluton VI) is host
to many metapelite and marble xenoliths and contains
the Salt Hill xenolith swarm, which is composed of numerous screens and curtains of aluminous xenoliths
located in the southeasternmost end of the pluton.
Unequivocal parental magma for the pluton has not
been identified anywhere in the complex owing to intense fractional crystallization, geochemical interactions,
and contamination or assimilation (Bender et al., 1984).
However, Tracy (2009) conducted a series of fractional
crystallization calculations at 08 GPa using MELTS
(Ghiorso & Sack, 1995; Asimow & Ghiorso, 1998) with
several potential compositions for a parental mafic
magma. Using a Karoo picritic dolerite (a mildly alkalic
Mg-rich basalt) as a parental magma (McBirney, 2007),
Tracy (2009) was able to produce a close match to the
compositions and proportions of minerals in the cumulate olivine-pyroxenite layers of pluton VI, implying that
391
392
Journal of Petrology, 2015, Vol. 56, No. 2
73º 55’
73º 45’
0
N
73º 35’
Peach Lake
Pluton
Croton Falls
Pluton
10
km
County
Dutchess
ds
nty
u
o
ster C Hudson Highlan
Westche
ky
sil
+
s
mu
Inwood Marble
Cortlandt Complex
Manhattan
Schist
e
et st +
rn
ga
41º 15’
me
ro
n’s
Lin
tite
bio
Ca
NY CT
New York
bi
ot
ite
Study Area
ga
et
rn
st +
ky
Peekskill
granite
Peekskill
2.5 km
Emery
Hill
Va
I
VI
III
IV
Vb
II
sil +
mus
Hybrid Igneous
Rock
Salt Hill
Manhattan Schist
Inwood Marble
Croton Falls Pluton
Hybrid Igneous Rock
Peekskill Granite
Fordham Gneiss
Mafic Igneous Plutons
Fig. 1. Simplified geological maps of the Cortlandt Complex and surrounding lithology. The Croton Falls Pluton studied by Dorfler
et al. (2014) is colored red in the upper map. The plutons in the Cortlandt Complex are outlined by bold black lines and Roman numerals indicate the relative ages of the plutons from oldest (I) to youngest (VI). Fine lines within each pluton indicate cumulate
layering. Dashed lines are regional metamorphic isograds (after Ratcliffe et al., 1983) cross-cut by the Cortlandt Complex.
Salt Hill xenoliths and monzonorite
General petrogenetic relationships within the
Cortlandt xenoliths, specifically from Salt Hill, have
previously been investigated in field and laboratory
studies, with emphasis on the petrogenesis of emery,
a colloquial term for corundum-bearing abrasive material (Friedman, 1956; Barker, 1964; Caporuscio &
Morse, 1978; Bender et al., 1984). Tracy & McLellan
(1985) were the first to provide a detailed petrographic
interpretation of the Salt Hill samples using textural
and mineral compositional data to determine the kinetic controls on the metamorphic processes, especially
reaction kinetics and the occurrence of localized equilibrium. They concluded that heating of the xenolith
probably produced a significant amount of silica-rich
peraluminous melt that eventually escaped and mixed
with the surrounding mafic magma to form a hybrid igneous rock zone surrounding the original schist
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
r
ve
Ri
on
ds
Hu
Fordham
Gneiss
Journal of Petrology, 2015, Vol. 56, No. 2
393
Table 1: Manhattan Schist analyses (in wt %) from Rogers
(1911)
Analysis no.:
24
25
26
27
5794
2170
157
590
249
058
174
468
217
029
Trace
101
Trace
—
—
019
10026
6298
1688
248
500
158
Trace
302
745
—
—
—
—
Trace
008
—
—
9947
6157
1953
544
261
190
Trace
348
214
—
—
—
153
Trace
085
—
—
9905
5512
2432
613
499
Trace
Trace
271
283
—
—
—
246
Trace
123
—
—
9979
4016
2950
1966
580
Trace
085
146
136
—
—
—
—
Trace
082
—
—
9961
Analysis 23 is a composite analysis of five specimens.
Analyses 24–27 are single analyses from schist samples near
the Cortlandt Complex.
xenolith. Crosscutting quartzofeldspathic veins and
veinlets (from millimeters to tens of millimeters in
thickness) are present within the xenolith samples, and
were interpreted by Tracy & McLellan (1985) to represent a remnant fraction of the highly siliceous, locally
derived melt. In this study we test this premise thermodynamically, showing that although the earlier study
captured the gross behavior of Salt Hill xenolith heating and resulting reactions, a more complex and subtle
suite of processes is required to explain all of the preserved lithologies and textures.
The Salt Hill xenolith samples examined in this study
consist of two chemically and physically segregated
materials: an alumina-rich, silica-deficient residuum
and SiO2-rich quartzofeldspathic veinlets (Fig. 2). The
residuum consists of varying mineral assemblages
including spinel (spl, 20–50 vol. %), complexly exsolved
ilmeno-hematite (ilm-hem, 10–20 vol. %), magnetite
(mt, 5–20 vol. %), sillimanite (sil, 10–50 vol. %), corundum (crn, 0–5 vol. %), sapphirine (spr, 0–40 vol. %), garnet (grt, 0–5 vol. %), and anorthite (an, 0–1 vol. %). In
general, the residuum can be either sapphirine-bearing
with Mg-rich spinel or sapphirine-absent, in which case
modal spinel is generally higher and the spinel has noticeably lower Mg/(Mg þ Fe) (Fig. 3). Corundum is rare
and can occur in either the sapphirine-bearing or sapphirine-absent residuum; it is present as either larger
prismatic crystals surrounded by magnetite, spinel, or
ilmeno-hematite solid solution, or as much finergrained corundum–magnetite symplectites along the
edges of spinel, apparently formed by oxidative breakdown of the hercynite component in spinel.
The quartz-grain-supported veinlets contain >90 wt
% SiO2 (based on modal analysis from image processing and average phase compositions) and consist of
roughly 80 modal % of 1–2 mm diameter rounded
quartz crystals with optically continuous microperthitic
ternary feldspar infilling the triangular interstices
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
SiO2
Al2O3
Fe2O3
FeO
MgO
CaO
Na2O
K2O
H2Oþ
H2O–
CO2
TiO2
P2O5
S
Cr2O3
MnO
Sum
23
between the crystals (average integrated feldspar composition of Or34Ab51An15) and minor prismatic sillimanite blades and hercynite-rich spinel (Figs 3 and 4). The
feldspar composition, plotted on a Ab–An–Or ternary
with solvus isotherms at 08 GPa from Fuhrman &
Lindsley (1988), indicates a minimum T of crystallization
of a homogeneous single feldspar at 1050 C.
Sapphirine, garnet and orthopyroxene reaction rims
surround spinel or Fe–Ti oxides where there is extensive physical (and chemical) interaction between the
veinlets and adjacent residuum. Sapphirine and orthopyroxene appear to form in more magnesian centimeter- to meter-scale lithological units, whereas the
less magnesian units contain more secondary garnet
and no sapphirine or orthopyroxene.
Uniformly thick orthopyroxene selvages lie along
the veinlet–residuum boundary in zones where there is
less physical interaction between veinlet and residuum
(Fig. 4). Major element analyses of these selvages
were performed using a CAMECA SX-50 electron
microprobe at Virginia Tech. Quantitative analysis of
major and minor elements used a 15 kV accelerating
potential with a beam current of 20 nA and a combination of natural and synthetic standards, with ZAF-type
matrix corrections performed using the PAP protocol
(Pouchou & Pichoir, 1984). Compositional maps in
Figures 3–5 are energy-dispersive spectrometry (EDS)
element maps collected using the Bruker SEM at
Virginia Tech with a beam current of 40 nA. Figures 3
and 5 involved an additional digital post-processing
method, overlaying two or more compositional maps
of different elements to emphasize different mineral
phases and textures. The orthopyroxene composition
is constant parallel to the vein–residuum boundary,
with Fe/(Fe þ Mg) ¼ 037. However, a strong compositional gradient in Al content of orthopyroxene occurs
normal to the boundary, with a higher average Al concentration (in cations per six oxygen formula basis) of
046 per formula unit (p.f.u.) on the residuum side of
the selvage and a lower Al concentration of 022 p.f.u.
adjacent to the vein (Table 2 and Fig. 5; complete
microprobe data for traverses 1–4 in Fig. 5 are provided in Supplementary Data Table 1; supplementary
data are available for downloading at http://www.petrology.oxfordjournals.org). We initially interpreted the
selvage as a reaction-rim product that formed during
silication of the spinel in the residuum by the reaction
spl þ melt ¼ opx þ sil. However, orthopyroxene is present regardless of the substrate phase in the residuum
(spl, ilm-hem, mt or even crn), which indicates that the
orthopyroxene more probably is a primary product of
crystallization from a melt, rather than a reaction product, and that it nucleated from melt along the irregular
interface, regardless of substrate mineral. This theory
is tested below.
A hybrid monzonorite near the margin of the xenolith
swarm in pluton VI contains intermediate plagioclase,
mesoperthitic alkali feldspar, ortho- and clinopyroxene,
and texturally late hornblende and biotite (this last
394
Journal of Petrology, 2015, Vol. 56, No. 2
Fig. 2. (a) Hand sample from the xenolith highlighting the two chemically and physically segregated materials. The white outlined
sections are the quartzofeldspathic veinlets. The dark material is the Al-rich residuum. Rock hammer handle for scale. (b) Close-up
of a hand sample showing the same segregation of the two materials. Paperclip for scale.
THERMODYNAMIC MODELING METHODS
To understand progressive reaction and melting processes during heating of the xenolith to well over
1100 C, we forward-modeled the isobaric heating of a
Manhattan Schist protolith (analysis 23 in Table 1) at
09 GPa in the system TiO2–Na2O–CaO–FeO–Fe2O3–
K2O–MgO–Al2O3–SiO2–H2O. Calculations utilized the
ds5.5 update to the Holland & Powell (1998) thermodynamic dataset, further modified for sapphirine
end-members following Kelsey et al. (2004). Complex
phases were modeled with activity–composition models for melt (White et al., 2001, 2007), feldspar (Holland
& Powell, 2003), sapphirine (Taylor-Jones & Powell,
2010), white mica (Coggon & Holland, 2002), ilmenohematite (White et al., 2000), spinel–magnetite (White
et al., 2002), garnet (White et al., 2005), orthopyroxene,
biotite (both Powell & Holland, 1999), and staurolite
(Holland & Powell, 1998). All calculations used Perple_X
6.6.8 (Connolly, 2005), with modifications made as
described below.
Having a well-constrained protolith bulk-rock composition is crucial for a forward-modeling study.
Previous xenolith–magma studies (i.e. Markl, 2005)
inferred the protolith composition from its altered
product; however, numerous analyses of the regionally metamorphosed Manhattan Schist are available.
For example, Tracy & McLellan (1985) suggested that
there were at least two protolith Manhattan Schist lithologies in one xenolith, based on the bimodal Fe/
(Fe þ Mg) composition of matrix spinels in the residuum and the presence or absence of sapphirine and
cordierite in the residuum. For the calculations shown
here, we chose to use one average Manhattan Schist
composition based on a composite of five separate
schist analyses reported by Rogers (1911). The specific
lithologies that produced either type of residuum (sapphirine-bearing and sapphirine-absent) are not known
and cannot be reconstructed with any confidence
owing to extensive melting. We acknowledge that use
of this ‘average Manhattan Schist’ composition introduces some uncertainty, but will demonstrate that the
results using this estimated protolith composition are
consistent with all melt and reaction products
observed.
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
mineral typically occurs in sprays radiating from Fe–Ti
oxide grains). X-ray fluorescence (XRF) whole-rock analyses of Salt Hill monzonorite samples are provided in
Table 3. The presence of a ‘mixed’ rock (e.g. norite, diorite
or tonalite), which typically represents contamination
(Bowen, 1928; Duchesne et al., 1989; Theriault et al., 1997;
Diaz-Alvarado et al., 2011; Alasino et al., 2014), indicates
that chemical and physical mixing of xenolith-derived
partial melt and surrounding mafic magma occurred during xenolith heating and melting, following xenolith incorporation in the magma. In thin section, replacement of
pyroxene along grain margins by amphibole is common,
as is the above-noted growth of radiating sprays of biotite
on oxide grains. Two notable additional features of this
hybrid rock are excessive modal amounts of apatite and
severely deformed plagioclase crystals (some are bent
into horseshoe shapes). Bender et al. (1984) interpreted
excess apatite as a product of contamination by country
rocks, and the deformation of plagioclase is consistent
with the hybrid lithology being a largely crystalline mush
during final pluton emplacement.
One of the goals of this study was to determine if the
monzonorite could have been formed via simple mixing
of a felsic melt produced during heating of the xenolith
(at temperatures well above its solidus) with an estimated fractionated composition of the external magma.
Mixing of felsic and mafic magmas requires compatibility of their physical properties (viscosity) at the temperature of equilibria (Sparks & Marshall, 1986). We
aim, therefore, to understand the formation of the hybrid igneous rock, the residuum, the quartzofeldspathic
veinlets, and their orthopyroxene selvages by constructing thermodynamic heating, crystallization, and
oxidation models, and by examining the viscosity contrasts of each of the melt phases to identify whether
hybridization is feasible.
Journal of Petrology, 2015, Vol. 56, No. 2
395
(a)
(b)
SH31
1 mm
SH20
SH29
(c)
feldspar
sillimanite
Edge of
Figure 4
quartz
hercynite
300 microns
Fig. 3. Examples of various types of residuum and texture of the quartz veinlets. (a) Sapphirine-free, spinel-rich residuum. Spinel is
green, rhombohedral oxides are black, large white crystals are corundum, white fibrous crystals are sillimanite. (b) Sapphirine-rich
residuum. Blue and tan crystals are sapphirine, black crystals are rhombohedral oxides, large green crystals are spinel, colorless
crystals are sillimanite. (c) False-color SEM-EDS image overlay of Si and Al maps showing the quartzofeldspathic vein. Yellow crystals are quartz, lavender blue material is interstitial ternary feldspar. Sillimanite (light blue) and hercynite (dark blue) are sometimes
present within rounded quartz crystals. The edge of Fig. 4 is shown for reference.
We estimate that the Manhattan Schist xenolith
cooled from 625 C to 500 C in the interval between
the Taconic-age sillimanite-grade regional metamorphic peak at c. 465 Ma and contact heating upon
incorporation into the magma at c. 446 Ma. It must then
have been rapidly heated to magmatic temperatures
(up to an estimated 1200 C) once it was fully immersed
in the mafic melt. One-dimensional heat conduction
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
1 mm
396
Journal of Petrology, 2015, Vol. 56, No. 2
Veinlet
Residuum
ic
mat
pris il
s
Fe-Ti
Oxide
opx
spl
Al
Fe
crn
mt
sil
0.25 mm
Relative Element Abundance
low
high
Fig. 4. Part of a thin section photomicrograph showing
the orthopyroxene rimming texture between residuum and
quartzofeldspathic veinlet. EDS chemical images show the
relative abundance of elements in the outlined area of the
micrograph.
modeling, assuming a 30 m diameter spherical xenolith
and an infinite reservoir of magma at a temperature of
1200 C (i.e. assuming that the heating of the xenolith
did not significantly cool the melt), predicts that the
core of the xenolith would reach 1150 C in 3 years
(Fig. 6). It is likely that prograde metamorphic dehydration reactions at lower temperatures, which did not involve melting, were significantly overstepped until melt
was present to increase intergranular diffusion rates.
Because the outermost 85% (by volume, assuming a
spherical xenolith) would have reached 900 C less than
1 year after initial immersion in the mafic melt, it is
likely that the rate of melt production significantly overstepped the rate of melt removal from the xenolith at
stages of its heating. Therefore, we propose that a melting model without removal of melt from the system
(equilibrium melting) best represents processes during
the initial heating of the Manhattan Schist. We note,
however, that calculated low-temperature equilibria
may not accurately reflect the earliest stages of xenolith
formation.
At temperatures much higher than that of the solidus, silica-rich melt loss from the xenolith ocurred,
yielding a silica-deficient residuum. This requires a different modeling strategy, whereby the calculated bulk
composition is progressively altered upon melt loss.
Both the solid phase and melt products of this calculation must be shown to evolve to form the observed mineralogy and measured phase compositions of the
aluminum-rich residuum and silica-rich veinlets.
RESULTS
Initial heating and melt generation
In an equilibrium thermodynamic model for heating the
average Manhattan Schist at 09 GPa, melt is first produced at 695 C from a complex melting reaction involving the breakdown of quartz, muscovite, biotite, and
feldspar (Fig. 8a). Muscovite begins to become markedly depleted at 750 C and disappears from the system
by 757 C. Over the temperature range from 750 to
763 C, biotite and quartz content decreases significantly
whereas total feldspar content increases (the calculated
K2O content of that feldspar increases to 10 wt % at
760 C, corresponding to a composition of Or57Ab40
An2). Kyanite transforms to sillimanite at 763 C. Modal
garnet content doubles from 7 to 14 vol. % over the temperature range from 760 to 815 C, and at 815 C the last
biotite disappears from the system. Quartz and plagioclase continue to dissolve into the melt above 815 C,
until quartz is finally lost at 912 C. Upon continued temperature increase, plagioclase is the only phase that
continues to be depleted markedly, before its final disappearance at 970 C. At this point the calculated residuum consists of garnet, sillimanite and ilmenite.
Once temperature reaches 1007 C, sapphirine forms at
the expense of sillimanite and garnet, which are lost at
1034 C and 1059 C, respectively. Sapphirine peaks in
abundance at 1059 C (67 vol. %) but dissolves into the
melt with further heating, completely melting out at
1187 C and leaving only an Mg-bearing ilmenite in the
residuum.
This simple model predicts that quartz and feldspar
should be absent from the high-T residuum assemblage
(matching observations), but the calculated mineral assemblages above 1100 C are significantly different
from those observed in the xenolith samples (the only
stable phases are sapphirine and ilmenite) and at no
point does the model calculate the observed
assemblage
spl þ mt þ ilm-hem þ sil 6 spr 6 crn 6 an.
Although the modeled melt composition produced at
1200 C in Fig. 8a is silica rich (60 wt %), its modeled
crystallization does not reproduce the observed mineralogy of the quartzofeldspathic veinlets. In fact, no
single melt composition from this model is appropriate
to form the quartzofeldspathic veinlets and we do not
believe that the modeled melts illustrated in Fig. 8a
were retained within the xenolith as veinlets.
Accordingly, we examined a possible scenario in which
the melt left its xenolith parent. We mixed the melt
derived at 900 C (where 60 vol. % of the xenolith is
melted; Fig. 8a and Table 4, column 1), with the fractionate of the assumed parental mafic melt at 1200 C
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
Si
Figure 7 is a flowchart that shows the sequential calculations that we undertook to model formation of the
various products, starting with initial heating of a
Manhattan Schist protolith to 900 C. Each of these steps
required different approaches and assumptions, and is
described in detail in the following sections.
Journal of Petrology, 2015, Vol. 56, No. 2
397
1
(a)
Al
2.5
Opx
Qz
Spl
1.5
Fe
0.5
Mg
opx
0
quartz
8
16 24 32 40
2 Mg
1
spinel
Fe
0.6
Al
0.2
0
30 microns
3
1.0
4
8
12
Opx
20
Qz
Mg
Fe
sil
3
4
mt
quartz
4
0
12
24
36
48
Mg
1.0
Fe
crn
spinel
0.6
Al
0
0.1 mm
2.0
1.5
1.0
0.5
0.2
6
18 24 30 36
Distance (microns)
Fig. 5. Electron microprobe (EMPA) traverses across an orthopyroxene selvage. (a, b) False-color SEM-EDS image overlays of Si
(yellow), Al (blue), Fe (violet), and Ti (green). Locations of traverses 1 and 2 are shown in (a); traverses 3 and 4 are shown in (b). The
orthopyroxene composition is constant parallel to the vein–residuum boundary (traverses 2 and 4). Al content shows a steady decrease in traverses (1 and 3) toward the quartzofeldspathic side. Dashed line in traverse 4 corresponds to a crack in the orthopyroxene. Grey chevrons highlight representative analyses in Table 2. (See Supplementary Data Table 1 for complete microprobe
analyses for each traverse.)
Table 2: Representative electron microprobe analyses from orthopyroxene selvage
Traverse 1
Spl
Opx next
to Spl
12
Opx next
to Qz
34
Traverse 2
Traverse 3
Opx
Opx next
to Crn
10
Opx next Qz
Opx
44
24
10
Traverse 4
Distance (mm):
4
SiO2
TiO2
Al2O3
FeO
MnO
MgO
CaO
043
002
5875
2894
048
974
002
4773
012
1009
2144
100
1997
009
5134
008
527
2091
130
2204
012
4703
015
918
2153
108
1965
015
4648
011
904
2133
102
1919
011
5062
007
562
2077
111
2155
009
4861
013
801
2151
109
2022
011
Total
Si
Ti
Al
Fe
Mn
Mg
Ca
9839
002
000
254
089
001
053
000
10044
176
000
044
066
003
110
000
10107
188
000
023
064
004
120
000
9877
177
000
041
068
003
110
001
9727
178
000
041
068
003
109
000
9983
188
000
025
064
003
119
000
9967
181
000
035
067
003
112
000
Sum
400
400
400
400
400
400
400
Formula calculated on the basis of four cations per six oxygens. Distances correspond to grey arrows in Fig. 5.
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
Al
(b)
16
4 Cations/6 Oxygen
2
ilm–hem
398
Journal of Petrology, 2015, Vol. 56, No. 2
Table 3: XRF analyses of two Salt Hill monzonorite samples,
and the averaged analysis of the two samples used in the melt
mixing calculations (see Table 4)
Ferric/ferrous ratio based on titration. LOI, loss on ignition.
(Table 4, column 2). A best-fit ratio of approximately
40% ‘felsic’ melt (i.e. melt produced by heating the pelitic schist to 900 C) and 60% ‘mafic’ melt yields a hybrid
melt that best matches the average composition of two
Cortlandt monzonorite samples in terms of all 10 oxides
considered (albeit with lower MgO content presumably
reflecting a small difference between estimated and actual mafic melt compositions; Table 4, column 3).
We thus propose that the monzonorite formed via mixing of an initial partial melt (which began leaving the
pelitic xenolith at 900 C as it continued to heat to
higher temperatures) with the host mafic magma in
which the xenolith resided, and below we explore the
feasibility of this process in terms of melt viscosities.
Continued melting from 900 C and veinlet
formation
The second stage of our model simulates continued
heating of the xenolith from 900 C, beginning with the
same bulk composition as used in the initial calculation
(Fig. 8a). Once melt begins to escape from the xenolith,
it is likely to continue to do so (Marschall et al., 2013) either continuously or cyclically depending on the applied
differential stress and rate of melt pressure buildup
(Brown, 2001). We therefore modified our strategy to
Temperature, ˚C
5459
162
1411
1069
017
679
643
232
322
037
10028
149
851
123
107
12
36
247
45
111
417
497
2395
16
358
1835
74
277
2315
685
1105
38
9685
1.50
900
1.25
12
11
10 –
9 ti m
8
e
7 6 (months) ––––
–––––|
5 4
3
2
1
800
700
600
500
0
2
4
6
8
10
12
radial distance from xenolith core, m
14
Fig. 6. One-dimensional (1-D) heat conduction modeling
assuming a 30 m diameter spherical xenolith with an initial T of
500 C and an infinite, non-cooling reservoir of magma at
1200 C. Calculated using equation (6.18) of Crank (1975), modified for temperature and plotted as described for figure 3 of
Caddick et al. (2010). Model predicts that the core of the xenolith would reach magmatic temperatures in about 3 years.
calculate phase equilibria in 50 C intervals, removing
any melt produced in each step, to simulate small-step
batch melting. Smaller temperature increments yielded
similar results, as discussed below. Figure 8b shows
the result at each stage by plotting the residuum and
melt at each temperature interval, with the ‘bulk composition’ recast after each step in which melt was produced and extracted. At 900 C, the phase proportions
are exactly the same as they were in the previous model
(qz þ sil þ grt þ ilm þ plg coexisting with 60% melt that
is extracted before the next calculation step). Removal
of this hydrous melt leaves a completely dry residuum
that is unable to melt further until dry melting reactions
begin above 1100 C, so there is little change in the calculated proportions of the phases between 950 and
1100 C. At 1150 C, however, the system reaches the
solidus for its new (depleted) composition and begins
to produce a new silica-rich melt, leaving behind an
even further SiO2-depleted residuum (spr þ grt þ ilm).
The calculated melt at 1150 C is silica- and aluminarich (66 wt % SiO2, 19 wt % Al2O3). Assuming that this
melt may have formed the veinlets shown in Figs 3c
and 4, we calculated its crystallization products upon
cooling using the same thermodynamic data as in the
heating calculations (Fig. 9). This new calculation
begins with a bulk composition fixed as the silica-rich
melt composition produced at 1150 C (Fig. 8b), but permits modification of the bulk composition by accounting for loss of partial melt during cooling. Phase
equilibria were calculated in 1 C increments, permitting
25% of the calculated melt fraction at each step to be
removed. This essentially simulates a filter-pressing
type of crystallization. We acknowledge that a more
sophisticated approach would involve calculation of the
melt fraction and viscosity at each temperature before
calculating the amount of melt that should be extracted
(which should also be cooling-rate dependent).
However, we consider the simple strategy of removing
25% of the melt present in each 1 C increment over the
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
5592
189
1382
1170
017
609
532
241
263
033
10028
173
894
176
1124
11
36
441
45
90
644
359
300
16
550
178
62
284
249
79
132
39
454
–
5326
135
1439
967
016
748
754
222
380
040
10027
124
807
070
1015
13
<05
52
45
132
190
635
179
16
167
189
86
270
214
58
89
37
1483
2.00
1000 1.75
––
Average
|– –
T-80-68
3.00
1100 2.50
time (years)
SiO2
TiO2
Al2O3
Fe2O3 (total)
MnO
MgO
CaO
Na2O
K2O
P2O5
Total
LOI
FeO
Fe2O3
Rb
Pb
U
Th
La
Ce
Nb
Sr
Y
Sc
Zr
V
Ni
Cr
Ga
Cu
Zn
Co
Ba
T-82-68
1200 4.00
Journal of Petrology, 2015, Vol. 56, No. 2
heating and
Manhattan melt retention
900˚C
Schist
see Fig 8a
399
Incremental heating and
residuum
batch melt removal to 1150˚C
see Fig 8b
residuum
melt escapes
hybridize with
mafic gabbro to
form monzonorite
see Table 2
some melt retained
crystallisation to
Opx + Sill + Qtz + Fsp
veinlets by 1100˚C
see Fig 3 and Fig 9
cooling + oxidation to
final residuum
approximate residuum
by 970˚C
see Fig 3 and Fig 10
Fig. 7. Flow chart showing progression of models calculated to produce the hybrid monzonorite, quartzofeldspathic veinlets and
aluminous residuum. Figures and tables that correspond to the appropriate calculation for each stage are indicated in red. Figures
illustrating the material that corresponds to the end results (veinlets or residuum) are indicated in blue.
Evolution of the residuum
At 1150 C, the batch-melting model produces a silicarich melt that crystallizes to a quartzofeldspathic rock
consistent with most observations of the natural veinlets, but the calculated residuum assemblage
(spr þ grt þ ilm) does not resemble the natural residuum. Based on coexistence of ilmeno-hematite with
almost pure magnetite in the natural samples, we estimate that the xenolith experienced increasingly
oxidizing conditions, almost to the hematite–magnetite
buffer at some point during residuum formation or evolution. (It should be noted that the bulk composition of
exsolved ilmenite-rich grains with hematite-rich lamellae in terms of pure FeTiO3 and Fe2O3 is 30% ilmenite,
70% hematite, as shown to the left of Fig. 5a and in
Supplementary Data Fig. 1.) We thus explored postmelting oxidation and cooling of the residuum that was
calculated in the batch-melting model at 1150 C
(Fig. 8b) with a T–Fe2O3/FeO diagram (Fig. 10) in which
the horizontal axis represents increasing ferric/ferrous
ratio; the initial ratio suggested for the residuum from
the calculations for Fig. 8b is on the left, and Fe2O3/FeO
increases progressively to the right. As noted above,
the natural residuum assemblage is variable but is predominantly composed of spl þ mt þ ilm-hem þ sil 6 spr
6 crn with <2 vol. % each of garnet and almost pure
anorthite. Simply increasing the ferric/ferrous iron ratio
at 1150 C predicts the formation of a residuum assemblage (spl þ ilm þ spr þ sil þ an þ crn) with phase proportions moderately close to the range observed in the
natural sample (Fig. 10). Furthermore, oxidizing the residuum to an Fe2O3/FeO of 035–058 and cooling to
temperatures between 960 and 1000 C produces the
assemblage spl þ ilm þ spr þ sil 6 crn 6 grt 6 an with
phase proportions that match the observed range in the
natural samples (see areas identified in Fig. 10).
DISCUSSION
Alternative heating models
The steps outlined above describe our preferred model
for the heating and formation of the Salt Hill xenoliths,
yielding the composition and phase proportions consistent with the monzonorite, quartzofeldspathic veinlets, and residuum from the Salt Hill samples. This
model has obvious assumptions such as the starting
composition of the pelite and the suitability of the
thermodynamic database; however, the results imply
that the problem, albeit complex, is solvable and can
yield a possible and plausible solution. To test
alternative possible methods of formation we compared the initial equilibrium-heating stage (Fig. 8a)
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
narrow crystallization interval to be a reasonable simplification for these dry melts. A minor modification of
Perple_X was required to permit this partial fractionation calculation, and the modified code is available for
downloading from http://www.metamorphism.geos.vt.
edu/Resources.html.
Crystallization illustrated in Fig. 9 begins with 100%
felsic melt at 1150 C and ends with complete crystallization to qz þ sil þ fsp þ opx by 1114 C, thus agreeing well
with the observed phase assemblage in the veinlets.
The small crystallization interval reflects the anhydrous
nature of the melt. Quartz is the first solid phase to form
in this model at 1148 C and is followed by sillimanite,
ternary feldspar (Or50Ab45An4) and an aluminous orthopyroxene at 1133, 1132, and 1130 C, respectively. The
orthopyroxene begins crystallizing with an Al content of
034 (per six oxygens) and falls to 026 by 900 C. The
Fe/(Fe þ Mg) ratio in the orthopyroxene is initially 048
at 1130 C, increases to 060 at 1115 C, and remains constant during further cooling. The final calculated rock is
composed of 83 vol. % quartz, 12 vol. % ternary feldspar, 4 vol. % sillimanite and 1 vol. % orthopyroxene,
with an overall bulk composition containing 92 wt %
SiO2. The modeled assemblage, phase proportions and
change in composition of the orthopyroxene as it
continues to crystallize are almost identical to our observations of the quartzofeldspathic veinlets and
orthopyroxene selvages in the natural samples
(e.g. Figs 3–5). Furthermore, mineral texture observations of the silica-rich veinlets, such as large, rounded
quartz crystals within interstitial ternary feldspar, are
consistent with the crystallization sequence modeled
here.
400
Journal of Petrology, 2015, Vol. 56, No. 2
100
(a)
90
Quartz
Volume Percent
Pre
me
elltt composition
elt
itio
d
ccte
ted
ed melt
me
co
co
om
mpossititio
on
on
ed
diict
Predicted
C(wt%):
(w
(wt
at 900
at
90
9
00
0 ˚C
wt%
t%
%)):
%):
900˚C
SiO
69..8
69
82
2
S O2 = 69.82
Al
15.01
1
Al2O3 = 15
1
5.
5.0
01
FeO
0.93
FeO
0.9
93
Fe
.9
3
O=0
MgO
0.25
gO
O = 0.2
0 25
M
Mg
5
Na
2.38
Na2O = 2.3
2
.3
38
8
CaO
C
0.
0.6
Ca
aO
O = 0.66
0 66
6
K2O = 6.35
6.3
6.3
35
4.6
60
H2O = 4.60
Melt
80
70
60
Muscovite
50
40
30
Biotite
Feldspar
10
Kyanite
Garnet
Sillimanite
Ilm
men
me
en
nite
Ilmenite
Sap
Sap
apphi
phiri
hirin
ine
ne
e
Sapphirine
0
600 650 700 750 800 850 900 950 1000 1050 1100 1150 1200 1250 1300
Temperature (˚C)
100
1150-1200 ˚C = approximate
peak T reached by xenolith
(b)
90
Volume Percent
80
70
60
50
40
30
20
10
0
900
950
1000
1050
1100
1150
1200
1250
1300
Temperature (˚C)
Spinel
Orthopyroxene
Sapphirine
Quartz
Sillimanite
Garnet
Ilmenite
Plagioclase
Melt
Fig. 8. (a) Phase abundances calculated from a 1-D (isobaric at 09 GPa) pseudosection. Gibbs free energy minimizations were in
1 C intervals, and a constant bulk composition of the Manhattan Schist (Table 1, analysis 23) was assumed. Activity–composition
models used are listed in the text. It should be noted that ‘feldspar’ combines plagioclase and alkali feldspar. (b) Phase abundances
calculated in 50 C intervals, with recasting of the bulk composition to simulate extraction of melt produced at any step. It should be
noted that the 900 C calculation is identical to the result at 900 C in (a). Diagonally shaded area to the right of the dashed line represents the part of the calculation not applicable to this study (see text for details). Composition of melt and residuum at 1150 C is
used in cooling (Fig. 9) and T–Fe2O3/FeO (Fig. 10) calculations.
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
20
Journal of Petrology, 2015, Vol. 56, No. 2
401
Monzonorite formation
Oxide Calculated
Calculated
Average
39% silica(wt %) silica-rich
picritic dolerite Salt Hill
rich melt,
melt at 900 C melt at 1200 C monzonorite 61% picrite
The calculation used to mix the ‘felsic’ melt produced at
900 C in the initial heating model (Fig. 8a) and the
‘mafic’ melt derived by fractionation from the inferred
parental picritic magma yielded results that match the
composition of the monzonorite samples found near
Salt Hill. In general, although mafic magmas should solidify on contact with cooler felsic melts, thereby
decreasing the efficiency of mixing, higher proportions
(>50%) of mafic melt intermingled with lower proportions of felsic melt can allow mixing of both compositions (Sparks & Marshall, 1986). The felsic melt will
have a higher viscosity than the mafic melt at a given
temperature, but if a sufficiently large proportion of the
hot mafic magma is present to continue heating of the
silica-rich melt, both melts could mix to form a hybrid
(Sparks & Marshall, 1986). The 60:40 mafic to felsic melt
ratio used to simulate the production of the hybrid igneous rock above and the estimated maximum temperature reached by the felsic melt (1200 C) imply that
formation of monzonorite from magma mixing is
plausible.
Melt mixing is physically possible only if the viscosities of the two melts are similar. In cases where the viscosity contrasts are low and slow cooling is combined
with strong convection, it is possible to generate quasihomogeneous hybrid igneous rocks (Gerdes et al.,
2000). The viscosities of the felsic and mafic melts were
therefore calculated with a composition-dependent
silicate melt viscosity calculator (available at http://
www.eos.ubc.ca/krussell/VISCOSITY/grdViscosity.html;
Giordano et al., 2008). This algorithm is calibrated to
calculate viscosity over a wide range of compositions
and temperatures at 1 bar pressure and we thus assume
a minimal pressure effect on calculated viscosity. This
assumption is clearly important, but should lead to less
error for the relatively low dissolved-H2O contents that
we model here than it would for very H2O-rich melts
(with the solubility of depolymerizing H2O in melt being
strongly P-dependent; e.g. Goranson, 1936, 1938;
Boettcher & Wyllie, 1969; Burnham & Davis, 1971;
Hamilton & Oxtoby, 1986; Mysen & Virgo, 1986). For
simplicity, the viscosity calculations assume that each
melt composition is fixed over the temperature range.
At 900 C the calculated viscosity of the felsic melt is
1042 Pa s, decreasing to 1024 Pa s as temperatures
reach 1200 C (Fig. 11). The calculated viscosity of the
mafic melt is 1028 Pa s at 1050 C (at which point the picritic dolerite is calculated to contain >17% melt, with
80% of crystals having been fractionally removed). This
decreases to 1015 Pa s at 1200 C (at which point the
melt fraction is >045). At temperatures greater than
1050 C, the viscosities for ‘felsic’ melt originally produced at 900 C and the mafic magma differ by less than
half an order of magnitude, implying that physical and
chemical mixing of the felsic melt and mafic magma
was unlikely to have been impeded by viscosity contrasts following extraction of the felsic melt from the
xenolith. We note also that the viscosity of the high-T
SiO2
TiO2
Al2O3
Fe2O3
FeO
MnO
MgO
CaO
Na2O
K2O
6982
000
1500
000
093
000
025
066
238
635
4454
268
1571
205
1193
017
658
914
288
131
5459
162
1411
123
851
017
679
643
232
322
5449
163
1543
124
760
010
409
580
268
329
Column 1 is the melt composition calculated at 900 C
(Fig. 8), with a melt model that does not permit Fe2O3. Column
2 is the composition of the assumed host magma calculated
with MELTS at 1200 C (see text). Column 3 is the averaged
composition of the Salt Hill monzonorite samples (Table 3).
Column 4 is a combination of 39% of column 1 and 61% of
column 2 and is comparable with the Salt Hill monzonorite
(column 3).
with results of ‘ideal’ batch-melting models in which
100% of the melt generated at any step was
removed before the next calculation step. These models
utilized phase equilibria calculated in 1, 20, 30, 40 and
50 C increments and used the same thermodynamic
database, protolith composition, total pressure,
and heating interval as in Fig. 8a (see Supplementary
Data Fig. 2).
The 1 C increment batch-melting model produced a
residuum (at T > 1200 C) containing 107 vol. % quartz.
Because the natural residuum is devoid of primary
quartz, this model was discarded. In contrast, the 20, 30,
40 and 50 C increment batch-melting models predict
that a trace amount of quartz is stable at temperatures
above 1140 C (the 30 C increment model predicts that
8 vol. % quartz is stable at 1200 C, but diminishes to
<1 vol. % at 1230 C, reaching 0 vol. % by 1290 C). In
fact, all of these models predict that the mineral assemblage of the residuum will be composed of spr, ilm and
spl (with minor amounts of grt, opx and qtz) at temperatures above 1140 C (see Supplementary Data Fig. 2).
These results are similar to the calculated residuum assemblage at 1150 C in our preferred model. However,
when each calculated melt (at 1140, 1140, 1160 and
1150 C from the 20, 30, 40, and 50 C interval batch-melt
models, respectively) was used in the ‘filter-pressing’
crystallization model described above, the calculated
crystallization products yielded inappropriate phase
proportions (e.g. 56–69 vol. % ternary feldspar), which
do not match observations from the quartzofeldspathic
veinlets in the Salt Hill samples. In addition, the xenolith-derived melt required to form the hybrid monzonorite by mixing is not produced in these alternative
calculations, leading us to believe that retention of melt
up to a melt fraction of 60% during rapid heating to
900 C, followed by progressive melt loss at higher
temperature (and lower melt viscosity; see below), is
more appropriate.
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
Table 4: Melt mixing results
402
Journal of Petrology, 2015, Vol. 56, No. 2
100%
Orthopyroxene
Ternary Feldspar
90%
Sillimanite
80%
Volume Percent
70%
60%
50%
Melt
Quartz
40%
30%
20%
0%
1150
1140
1130
1120
1110
1100
1090
1080
1070
1060
1050
Temperature (˚C)
Fig. 9. Phase abundances calculated from a 1-D (isobaric at 09 GPa) pseudosection. Gibbs free energy minimizations were in 1 C
intervals, and a starting bulk composition of the melt produced at 1150 C in Fig. 8b was used. At each step, 25% of the calculated
volume of melt was extracted (see text for details). All solid phases included in other calculations were considered in these calculations, but only four were found to be stable.
silica-rich melt generated at 1150 C (Fig. 8b) is substantially higher than those for either of the other two calculated melt compositions, its limited mobility being
consistent with the observation of retention of some
proportion of this melt within the xenolith during crystallization to form veinlets.
The exact composition of the melts that mixed to
produce the hybrid igneous rock and the temperature
at which melt began to escape the xenolith cannot be
definitively determined. However, our models show
that melt produced during initial rapid heating of the
pelitic schist can compositionally and physically mix
with the mafic melt to yield a hybrid igneous rock
similar to the monzonorite upon its escape from the
xenolith. Stokes’ Law calculations suggest an almost
immediate settling of the dense, melt-depleted xenolith to the base of the magma chamber. Even small
(1 m diameter) xenolith fragments (for which Stokes’
Law is more appropriate) are calculated to settle rapidly owing to the 800 kg m–3 density contrast between them and the host melt. Therefore it is possible
that the segregation of less-dense silica- and alkalirich partial melt into veinlets or lenses contributed
to the collapse, agglomeration and settling of the
iron–magnesium–titanium-rich residuum. Some residuum samples contain triple-junction grain boundary textures of spinel and Fe–Ti oxides that are
characteristic of annealing, which may have taken
place during melt accumulation. Additionally, our
modeling supports previous studies that reported contamination of mafic magma by granitic partial melt
derived from metapelites and production of intermediate compositions (e.g. Bowen, 1928; Duchesne et al.,
1989; Theriault et al., 1997).
Crystallization of orthopyroxene selvages
We initially interpreted the orthopyroxene selvage texture (e.g. Figs 4 and 5) to be the product of successive
reaction of the residuum (principally spinel) with silica
from the quartzofeldspathic veins (silication), initially
proposed by Friedman (1954) in a paper on the spinel–
silica reaction succession. However, there are two observations that conflict with that simple interpretation:
(1) the selvages have a generally uniform thickness of
about 30 mm along veinlet margins and adjacent to the
residuum, irrespective of the substrate mineral in the
residuum, whether spinel, corundum, ilmeno-hematite, magnetite or sillimanite; (2) the composition of
the orthopyroxene shows the same variation in Al content and Fe/(Fe þ Mg) irrespective of the adjacent residuum and veinlet phases. Microprobe analytical
traverses along the length of the selvage reveal little
variation in Al content or Fe/(Fe þ Mg) ratio (see Fig. 5)
immediately adjacent to the residuum. Traverses that
transect the selvage have the same Fe/(Fe þ Mg) ratio
across the orthopyroxene but decreasing Al content
toward the veinlet. Based on these observations, a
simple localized reaction-margin origin of the orthopyroxene seems less credible, and instead it seems
probable that the orthopyroxene crystallized from the
veinlet-forming melt and nucleated uniformly at the irregular veinlet contact, on the surface of the
residuum.
The ‘filter-pressing’ cooling model developed for this
study, where 25% of the melt is removed at each 1 C increment, best represents crystallization of the calculated silica-rich melt at 1150 C. If the partial melt is
segregated into veinlets and melt pockets (as observed
in natural samples) by the agglomeration of the dense
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
10%
Journal of Petrology, 2015, Vol. 56, No. 2
403
1150
spl
ilm
grt
spl ilm spr
spr
sill an cor
spl
sill
ilm grt
an
spr sill
Temperature (˚C)
1100
1050
1000
900
850
spl ilm grt cor
750
ilm spr
sill an
cor
0%
ilm
grt
spr
sill
cor
spl ilm grt
spr cor
opx
ilm
grt
spr
cor
opx spl
ilm grt cor
35%
Garnet
spl
ilm
grt
spl ilm spr
spr
sill an cor
spl
sill
ilm grt
an
spr sill
spl ilm grt
spr sill cor
950
800
Corundum
ilm spr
sill an
cor
spl ilm grt
spr sill cor
spl ilm grt cor
0%
ilm
grt
spr
sill
cor
spl ilm grt
spr cor
opx
ilm
grt
spr
cor
opx spl
ilm grt cor
25%
1150
Temperature (˚C)
1000
900
850
750
spl ilm grt cor
0%
ilm
grt
spr
sill
cor
spl ilm grt
spr cor
opx
ilm
grt
spr
cor
opx spl
ilm grt cor
28%
1150
spl
ilm
grt
spl ilm spr
spr
sill an cor
spl
sill
ilm grt
an
spr sill
1100
Temperature (˚C)
1050
1000
900
850
750
spl ilm grt cor
0%
Sillimanite
spl ilm grt cor
0%
0.19
15%
0.39
0.70
1.50
Fe2O3/FeO
ilm
grt
spr
sill
cor
spl ilm grt
spr cor
opx
ilm
grt
spr
cor
opx spl
ilm grt cor
28%
spl
ilm
grt
spl ilm spr
spr
sill an cor
spl
sill
ilm grt
an
spr sill
ilm spr
sill an
cor
ilm
grt
spr
sill
cor
spl ilm grt
spr cor
opx
ilm
grt
spr
cor
opx spl
ilm grt cor
ilm spr
sill an
cor
spl ilm grt
spr sill cor
spl ilm grt
spr sill cor
950
800
ilm spr
sill an
cor
spl ilm grt
spr sill cor
950
Ilmenite
spl
ilm
grt
spl ilm spr
spr
sill an cor
spl
sill
ilm grt
an
spr sill
Sapphirine
ilm spr
sill an
cor
spl ilm grt
spr sill cor
spl ilm grt cor
0%
4.71 0.19
ilm
grt
spr
sill
cor
spl ilm grt
spr cor
opx
ilm
grt
spr
cor
opx spl
ilm grt cor
70%
0.39
0.70
1.50
4.71
Fe2O3/FeO
Fig. 10. T–Fe2O3/FeO diagrams with phase abundances contoured for corundum, garnet, spinel, ilmenite, sillimanite and sapphirine. Dotted oval denotes area where phase abundances most closely match the residuum samples. Color bar shows the observed
abundance range of each mineral.
residuum, then phases crystallizing from the high-T
melt would remain in the fractures as melt continued to
be compressively ‘squeezed’ or pumped. The filterpressing crystallization model predicts that 1 vol. %
orthopyroxene forms from cooling the calculated melt
produced at 1150 C. This orthopyroxene is predicted to
begin to crystallize only 18 C after the initial crystallization of quartz, the first phase to crystallize at 1148 C,
and 3 C after sillimanite, the second phase to crystallize
at 1133 C. This cooling model was the only calculation
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
1050
800
Spinel
spl
ilm
grt
spl ilm spr
spr
sill an cor
spl
sill
ilm grt
an
spr sill
1100
404
Journal of Petrology, 2015, Vol. 56, No. 2
7
Silica-rich melt
produced at 1150˚C
(see Fig 8)
6
logη (Pa s)
5
Felsic melt produced
at 900˚C (see Fig 8)
4
3
T above which
dolerite contains
> 17 % melt
2
1
0
900
950
T above which
dolerite contains
< 70 % crystals
1000
1050
1100
Temperature (˚C)
1150
1200
that produced results consistent with the quartzofeldspathic veinlet texture and chemistry (see, for example, sillimanite–orthopyroxene–quartz relationships
in Fig. 4).
The model predicts that orthopyroxene crystallized
at 1130 C contains 034 p.f.u. Al, decreasing to
026 p.f.u. by 900 C. These values differ slightly from
measurements of the orthopyroxene selvages (which,
on average, decrease from 046 p.f.u. on the residuum
side of the selvage to about 022 p.f.u. adjacent to the
vein; see Fig. 5), but the trend of decreasing aluminum
content from the inferred first-crystallizing orthopyroxene on the veinlet wall to the last orthopyroxene is
consistent with our observations. The calculated Fe/
(Fe þ Mg) ratio of the orthopyroxene is initially
048 p.f.u. at 1130 C and increases to 060 p.f.u. at
1115 C, subsequently remaining constant for the duration of cooling. This differs from the Fe/(Fe þ Mg) ratio
in the samples, which remains a consistent 037 p.f.u. at
any distance across the selvage. This discrepancy may
be due to the differences in bulk composition of the protolith, or to the possibility that there was diffusional mobility of Fe and Mg within the selvage (in contrast to
more limited Al mobility) and thus diffusional re-equilibration. Tracy & McLellan (1985; compare Preston et al.,
1999; Johnson et al., 2010) suggested that the local Fe/
(Fe þ Mg) of the residuum spinel, and the occurrence of
discrete secondary silicate assemblages (e.g. spl–grt–opx versus spl–spr–opx–crd that could be correlated
with this variation in spinel chemistry) were functions
of variations in the original Fe/(Fe þ Mg) of the
Manhattan Schist protolith within the xenolith.
Oxidation of the residuum and symplectite
formation
The calculated T–Fe2O3/FeO diagram (Fig. 10) illustrates
a potential residuum oxidation process in which oxygen
content increases (indicated by the change in ferrous/
ferric iron ratio) during cooling from 1150 to 700 C. In
CONCLUSIONS
The results described in detail above show that, on a
bulk scale, our model for partial melting, melt mixing
and crystallization produces lithologies and assemblages consistent with observations from each of the
main rock types studied. Although the multi-phase
melting model is specific for the Salt Hill xenolith samples, the products of the processes described here resemble prevalent mid- to lower-crustal materials and
we suggest that certain aspects of these processes may
be more widespread than currently thought. The general model may be applicable to understanding (1) the
partial melting of the lower crust and production of hybrid igneous rocks (e.g. in or near under-plated mafic
magmas) and (2) attainment of UHT crustal conditions
without recourse to enrichment in radiogenic heat-producing elements or unusual thermal relaxation
scenarios.
The generation of intermediate arc magmas is
widely attributed to either partial melting of crustal
rocks (e.g. Smith & Leeman, 1987; Chappell & White,
2001) or differentiation of primary magmas by crystallization (e.g. Rogers & Hawkesworth, 1989; Müntener
et al., 2001). Several previous models have focused on
the interaction of primary mafic melts with the silicic
crustal melts that they can generate, reasoning that
more secondary felsic melt can be produced deeper in
the crust, where the ambient temperatures are higher
(Annen & Sparks, 2002; Annen et al. 2006; Lyubetskaya
& Ague, 2010). The fertility of the crust that is intruded
is also important (e.g. Petford & Gallagher, 2001), and
models reveal that 50 km deep mafic intrusions into pelitic crust are likely to produce melt fractions reaching or
exceeding 03 (Lyubetskaya & Ague, 2010). Our study
shows the results of such melt production and demonstrates that, at least on local scales, crustal partial melt
mixes with the mafic melt to generate hybrid magmas.
The very large melt fraction that our model extracts
from the pelitic protolith results from (1) its fertile
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
Fig. 11. Calculated viscosities for equilibrium composition of
the mafic melt (picritic dolerite) at 1200 C and felsic melts generated at 900 C (see Fig. 8a) and at 1150 C (Fig. 8b). For simplicity, viscosity calculations assume that each melt composition
is fixed.
general, the model generates a reasonable match for
the observed range of phase proportions for most of
the phases observed in the various types of the natural
residuum. Although there are textural and mineralogical observations in the samples that are not
explicitly explained by the T–Fe2O3/FeO diagram (specifically, ilmenite–hematite ratio in the rhombohedral
oxide phase and the breakdown of hercynite to form
magnetite–corundum symplectite), the underlying principle that the model demonstrates is that the residuum
must have been oxidized, probably at a late stage in its
evolution, to produce the residuum mineral assemblages observed. Further research on understanding in
detail how and when oxidation occurred, including
micro-scale observations and textural analysis on the
exsolved ilm-hem grains and the crn–mt symplectites,
is in progress and will be presented in a later
contribution.
Journal of Petrology, 2015, Vol. 56, No. 2
metapelite minerals (especially spinel) throughout the
region north of the Cortlandt Complex led Brock &
Brock (1998) to suggest a late Proterozoic granulite-facies regional metamorphism in this terrane, followed
by pervasive but incomplete retrogressive reactions,
which would also be consistent with the more widespread presence of currently unexposed mafic units. In
this case, we expect characteristically brief durations
near peak metamorphic conditions of 900 C or higher
(e.g. Galli et al., 2011; Clark et al., 2014b). This rapid,
melt-enhanced heating contrasts with UHT belts that
experienced longer heating and cooling durations (tens
of million years) and have been interpreted as requiring
high concentrations of heat-producing elements
coupled with low erosion rates (e.g. Clark et al., 2014a).
This study successfully models heating and partial
melting of a pelitic schist to produce the residual aluminous xenolith assemblage, hybrid igneous rock compositions, and quartzofeldspathic veinlet compositions
and textures observed in the natural samples. These results enhance the quantitative understanding of the
rapid production of siliceous melts upon contact metamorphism and illustrate the multi-stage and complex
nature of melting and hybridization processes that are
likely to characterize the deep crust. This provides inspiration for further investigation of the details of the
heating and cooling processes such as those that
occurred at Salt Hill, and demonstrates the utility of
quantitative thermodynamic modeling for understanding rocks undergoing complex, multi-stage melt
production.
ACKNOWLEDGEMENTS
We thank Michael Brown for stimulating conversation
concerning modeling techniques and initial results, and
James Beard for careful editorial handling of this contribution. Michael Williams, Jodie Hayob and Jay Ague
gave extremely detailed, considered and helpful reviews,
for which we are grateful. KMD gratefully acknowledges
a research stipend from the Department of Geosciences
at Virginia Tech, which supported fieldwork. Many
thanks to the patient and trusting field assistants H. Gall
and A. Frayne, and to C. Visweswariah, P. Buchanan,
G. Pacchiana and Thalle Industries, Inc. for granting
access to the Salt Hill quarry.
SUPPLEMENTARY DATA
Supplementary data for this paper are available at
Journal of Petrology online.
REFERENCES
Ackermand, D., Seifert, F. & Schreyer, W. (1975). Instability of
sapphirine at high pressures. Contributions to Mineralogy
and Petrology 50, 79–92.
Ague, J. J. & Baxter, E. F. (2007). Brief thermal pulses during
mountain building recorded by Sr diffusion in apatite and
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
nature, with a high H2O content prior to contact metamorphism, and (2) the geometry of the case study, with
immersion of a 30 m diameter xenolith in a much larger
body of mafic magma permitting the pelitic rocks to effectively reach mafic melt temperatures, irrespective of
the ambient geotherm. Wall-rocks adjacent to the intrusion are subject to different constraints, experiencing
‘energy-limited’ conditions owing to conductive heat
loss and thus reaching a maximum temperature that is
governed by the initial temperature. However, it is important to note, when considering the potential for
more widespread genesis of intermediate melts, that
our models generate a substantial proportion of the silica-rich melt end-member at temperatures of <900 C.
This demonstrates that wholesale heating of the crust
up to gabbroic emplacement temperatures is not necessarily required to produce a significant volume of hybrid melt, in agreement with previous studies (e.g.
Annen & Sparks, 2002). Although the specific geometry
of sedimentary xenoliths fully immersed in a mafic melt
might be relatively rare, we nevertheless anticipate that
country rocks on the outer margins of the mafic intrusions also generate silicate melt appropriate for the hybridization described here.
Dorfler et al. (2014) showed that prior to contact
metamorphism, metamorphic rocks now forming the
aureole of a small intrusion geologically similar to and
near the Cortlandt Complex had reached regional metamorphic temperatures of 600 C. Those contact metamorphosed rocks now record equilibration at 800 C,
and the xenolith studied here preserves assemblages
that formed at over 1000 C. We pose the question here
of how these contact aureole rocks might have been interpreted in the absence of field relations suggesting
contact metamorphism: is it possible that the high-temperature metamorphism would be interpreted as representing the culmination of regional metamorphism if
the mafic bodies were not exposed (e.g. if they had
been removed by or not yet exposed by erosion)?
The characteristic timescale of the contact event
described here is probably far shorter than that of the
regional metamorphism that it followed, and this brevity may be diagnostic of the role of magma emplacement. This is similar to recent findings from Dalradian
metasediments of the Barrovian sequence, which suggest that garnet-grade samples experienced long
duration metamorphism (>10 Myr) but that sillimanitegrade rocks experienced brief (<05 Myr) pulses of
crystallization (e.g. Ague & Baxter, 2007). This leads to
the conclusion that the highest metamorphic temperatures may have been reached owing to thermal conduction from mafic magmas (Ague & Baxter, 2007; Viete
et al., 2011; Vorhies & Ague, 2011). We suggest that
emplacement of such mafic melts, currently exposed or
inferred, may be particularly important or even necessary in attaining UHT metamorphic conditions, particularly in cases where the occurrence of UHT mineral
assemblages is only patchy or localized. For example,
observation of the cryptic occurrence of high-T
405
406
Crank, J. (1975). The Mathematics of Diffusion. Clarendon
Press.
Dana, J. D. (1881). The geological relations of the limestone
belts of Westchester County, New York. American Journal
of Science (1880) 22, 103–119.
Diaz-Alvarado, J., Castro, A., Fernandez, C. & Ignacio MorenoVentas, I. (2011). Assessing bulk assimilation in cordieritebearing granitoids from the Central System Batholith,
Spain; experimental, geochemical and geochronological
constraints. Journal of Petrology 52, 223–256.
Domenick, M. A. & Basu, A. R. (1982). Age and origin of
the Cortlandt Complex, New York: Implications from
Sm–Nd data. Contributions to Mineralogy and Petrology 79,
290–294.
Dorfler, K. M., Tracy, R. J. & Caddick, M. J. (2014). Late-stage
orogenic loading revealed by contact metamorphism in the
northern Appalachians, New York. Journal of Metamorphic
Geology 32, 113–132.
Duchesne, J. C., Wilmart, E., Demaiffe, D. & Hertogen, J. (1989).
Monzonorites from Rogaland (Southwest Norway): a series
of rocks coeval but not comagmatic with massif-type anorthosites. Precambrian Research 45, 111–128.
Friedman, G. M. (1954). The spinel–silica reaction succession: A
study of incompatible mineral phases. Journal of Geology
62, 366–374.
Friedman, G. M. (1956). The origin of spinel–emery deposits,
with particular reference to those of the Cortlandt Complex,
New York. University of the State of New York.
Fuhrman, M. L. & Lindsley, D. H. (1988). Ternary feldspar
modeling and thermometry. American Mineralogist 73,
201–215.
Galli, A., Le Bayon, B., Schmidt, M. W., Burg, J.-P., Caddick, M.
J. & Reusser, E. (2011). Granulites and charnockites of the
Gruf Complex: evidence for Permian ultra-high temperature
metamorphism in the Central Alps. Lithos 124, 17–45.
Gerdes, A., Woerner, G. & Finger, F. (2000). Hybrids, magma
mixing and enriched mantle melts in post-collisional
Variscan granitoids: the Rastenberg Pluton, Austria. In:
Franke, W., Haak, V., Oncken, O. & Tanner, D. (eds) Orogenic
Processes: Quantification and Modelling of the Variscan
Belt. Geological Society, London, Special Publications 179,
415–43.
Ghiorso, M. S. & Sack, R. O. (1995). Chemical mass transfer in
magmatic processes. IV. A revised and internally consistent
thermodynamic model for the interpolation and extrapolation of liquid–solid equilibria in magmatic systems at elevated temperatures and pressures. Contributions to
Mineralogy and Petrology 119, 197–212.
Giordano, D., Russell, J. K. & Dingwell, D. B. (2008). Viscosity of
magmatic liquids: A model. Earth and Planetary Science
Letters 271, 123–134.
Goranson, R. W. (1936). Silicate–water systems: The solubility
of water in albite–melt. Transactions of the American
Geophysical Union 17, 257–259.
Goranson, R. W. (1938). Silicate–water systems: Phase equilibria in the NaAlSi3O8–H2O and KAlSi3O8–H2O systems at
high temperatures and pressures. American Journal of
Science 35A, 71–91.
Hamilton, D. L. & Oxtoby, S. (1986). Solubility of water in albite–melt determined by the weight-loss method. Journal of
Geology 94, 626–630.
Hensen, B. J. & Green, D. H. (1970). Experimental data on coexisting cordierite and garnet under high grade metamorphic
conditions. Physics of the Earth and Planetary Interiors 3,
431–440.
Hensen, B. J. & Green, D. H. (1971). Experimental study of the
stability of cordierite and garnet in pelitic compositions at
high pressures and temperatures: I, compositions with
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
multicomponent diffusion in garnet. Earth and Planetary
Science Letters 261, 500–516.
Alasino, P. H., Casquet, C., Larrovere, M. A., Pankhurst, R. J.,
Galindo, C., Dahlquist, J. A., Baldo, E. G. & Rapela, C. W.
(2014). The evolution of a mid-crustal thermal aureole at
Cerro Toro, Sierra de Famatina, NW Argentina. Lithos
190–191, 154–172.
Annen, C. & Sparks, R. S. J. (2002). Effects of repetitive emplacement of basaltic intrusions on thermal evolution and
melt generation in the crust. Earth and Planetary Science
Letters 203, 937–955.
Annen, C., Blundy, J. D. & Sparks, R. S. J. (2006). The genesis of
intermediate and silicic magmas in deep crustal hot zones.
Journal of Petrology 47, 505–539.
Asimow, P. D. & Ghiorso, M. S. (1998). Algorithmic modifications extending MELTS to calculate subsolidus phase relations. American Mineralogist 83, 1127–1131.
Barker, F. (1964). Reaction between mafic magmas and pelitic
schist, Cortlandt, New York. American Journal of Science
(1880) 262, 614–634.
Bender, J. F., Hanson, G. N. & Bence, A. E. (1984). Cortlandt
Complex; differentiation and contamination in plutons of alkali basalt affinity. American Journal of Science (1880) 284,
1–57.
Boettcher, A. L. & Wyllie, P. J. (1969). Phase relationships in the
system NaAlSiO4–SiO2–H2O to 35 kilobars pressure.
American Journal of Science 267, 875–909.
Bowen, N. L. (1928). The Evolution of the Igneous Rocks.
Princeton, NJ: Princeton University Press.
Brock, P. C. & Brock, P. W. G. (1998). Taconian-aged granulite
facies metamorphism in the Manhattan Prong, NY and CT.
Abstracts: Geology of Long Island and Metropolitan New
York. Stony Brook, NY: Long Island Geologists, SUNY Stony
Brook, pp. 28–30.
Brown, M. (2001). Orogeny, migmatites and leucogranites: A review. Journal of Earth System Science 110(4), 313–336.
Burnham, C. W. & Davis, N. F. (1971). The role of H2O in silicate
melts: I. P–V–T relations in the system NaAlSi3O8–H2O to 1
kilobar and 1000 C. American Journal of Science 270,
54–79.
Caddick, M. J., Konopasek, J. & Thompson, A. B. (2010).
Preservation of garnet growth zoning and the duration of
prograde metamorphism. Journal of Petrology 51,
2327–2347.
Caporuscio, F. A. & Morse, S. A. (1978). Occurrence of sapphirine plus quartz at Peekskill, New York. American Journal of
Science (1880) 278, 1334–1342.
Chappell, B. W. & White, J. R. (2001). Two contrasting granite
types: 25 years later. Australian Journal of Earth Sciences
48, 489–499.
Clark, C., Healy, D., Johnson, T., Collins, A. S., Taylor, R. J.,
Santosh, M. & Timms, N. E. (2014a). Hot orogens and supercontinent amalgamation: A Gondwanan example from
southern
India.
Gondwana Research doi:10.1016/
j.gr.2014.11.005.
Clark, C., Kirkland, C. L., Spaggiari, C. V., Oorschot, C., Wingate,
M. T. D. & Taylor, R. J. (2014b). Proterozoic granulite formation driven by mafic magmatism: An example from the
Fraser
Range
Metamorphics,
Western
Australia.
Precambrian Research 240, 1–21.
Coggon, R. & Holland, T. J. B. (2002). Mixing properties of muscovite–celadonite–ferroceladonite–paragonite micas and
revised garnet–phengite thermobarometers. Journal of
Metamorphic Geology 20, 683–696.
Connolly, J. A. D. (2005). Computation of phase equilibria by linear programming: a tool for geodynamic modeling and its
application to subduction zone decarbonation. Earth and
Planetary Science Letters 236, 524–541.
Journal of Petrology, 2015, Vol. 56, No. 2
Journal of Petrology, 2015, Vol. 56, No. 2
Patiño Douce, A. E. & Beard, J. S. (1995). Dehydration-melting
of biotite gneiss and quartz amphibolite from 3 to 15 kbar.
Journal of Petrology 36, 707–738.
Peterson, J. W. & Newton, R. C. (1990). Experimental biotite–quartz melting in the KMASH–CO2 system and the role of
CO2 in the petrogenesis of granites and related rocks.
American Mineralogist 75, 1029–1042.
Petford, N. & Gallagher, K. (2001). Partial melting of
mafic (amphibolitic) lower crust by periodic influx of
basaltic magma. Earth and Planetary Science Letters 193,
483–489.
Pouchou, J.-L. & Pichoir, F. (1984). A new model for quantitative
X-ray microanalysis; Part I: application to the analysis of
homogeneous samples. Recherche Aerospatiale 3, 13–38.
Powell, R. & Holland, T. J. B. (1988). An internally consistent
dataset with uncertainties and correlations: 3. Applications
to geobarometry, worked examples and a computer program. Journal of Metamorphic Geology 6, 173–204.
Powell, R. & Holland, T. J. B. (1994). Optimal geothermometry
and geobarometry. American Mineralogist 79, 120–130.
Powell, R. & Holland, T. J. B. (1999). Relating formulations of
the thermodynamics of mineral solid solutions: Activity
modeling of pyroxenes, amphiboles, and micas. American
Mineralogist 84, 1–14.
Powell, R. & Holland, T. J. B. (2008). On thermobarometry.
Journal of Metamorphic Geology 26, 155–179.
Powell, R., Holland, T. J. B. & Worley, B. (1998). Calculating
phase diagrams with THERMOCALC: methods and examples. Journal of Metamorphic Geology 16, 577–588.
Preston, R. J., Dempster, T. J., Bell, B. R. & Rogers, G. (1999).
The petrology of mullite-bearing peraluminous xenoliths:
Implications for contamination processes in basaltic magmas. Journal of Petrology 40, 549–573.
Ratcliffe, N. M., Armstrong, R. L., Mose, D. G., Seneschal, R.,
Williams, N. & Baiamonte, M. J. (1982). Emplacement history and tectonic significance of the Cortlandt Complex,
related plutons, and dike swarms in the Taconide Zone of
southeastern New York based on K–Ar and Rb–Sr investigations. American Journal of Science (1880) 282, 358–390.
Ratcliffe, N. M., Bender, J. F. & Tracy, R. J. (1983). Tectonic
setting, chemical petrology, and petrogenesis of the
Cortlandt Complex and related igneous rocks of southeastern New York State. Geological Society of American
Northeastern Section Meeting, 2–26 March 1983, Concord
Hotel, Kiamesha Lake, NY, Guidebook for Field Trip 1, pp.
1–101.
Ratcliffe, N. M., Tucker, R. D., Aleinikoff, J. N., Amelin, Y.,
Merguerian, C. M. & Panish, P. T. (2012). U–Pb zircon and
titanite ages of late- to post-tectonic intrusions of the
Cortlandt–Beemerville magmatic belt, CN, NY, and NJ: relation to Iapetan closure in the Taconian orogeny. Geological
Society of America, Abstracts with Programs 44, 73.
Rogers, G. & Hawkesworth, C. J. (1989). A geochemical traverse
across the North Chilean Andes—evidence for crust generation from the mantle wedge. Earth and Planetary Science
Letters 91, 271–285.
Rogers, G. S. (1911). Geology of the Cortlandt Series and its
emery deposits. Annals of the New York Academy of
Science 21, 11–86.
Smith, D. R. & Leeman, W. P. (1987). Petrogenesis of Mount St.
Helens dacitic magmas. Journal of Geophysical Research
92, 10313–10334.
Sparks, R. S. J. & Marshall, L. A. (1986). Thermal and mechanical constraints on mixing between mafic and silicic magmas. Journal of Volcanology and Geothermal Research 29,
99–124.
Taylor-Jones, K. & Powell, R. (2010). The stability of
sapphirine plus quartz: calculated phase equilibria in
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
excess alumino-silicate. Contributions to Mineralogy and
Petrology 33, 309–330.
Hensen, B. J. & Green, D. H. (1972). Experimental study of the
stability of cordierite and garnet in pelitic compositions at
high pressures and temperatures; II, compositions without
excess alumino-silicate. Contributions to Mineralogy and
Petrology 35, 331–354.
Holland, T. J. B. & Powell, R. (1998). An internally consistent
thermodynamic data set for phases of petrological interest.
Journal of Metamorphic Geology 16, 309–343.
Holland, T. J. B. & Powell, R. (2003). Activity–composition relations for phases in petrological calculations: an asymmetric
multicomponent formulation. Contributions to Mineralogy
and Petrology 145, 492–501.
Huang, W. L. & Wyllie, P. J. (1973). Melting relations of muscovite–granite to 35 kbar as a model for fusion of metamorphosed subducted oceanic sediments. Contributions to
Mineralogy and Petrology 42, 1–14.
Johnson, A. M. & Tracy, R. J. (1999). Textural and chemical relations among spinel-sapphirine-garnet-orthopyroxene, Salt
Hill emery mine, Cortlandt Complex, NY. Geological Society
of America Abstracts with Programs 31, 40.
Johnson, T. E., Brown, M. & White, R. W. (2010). Petrogenetic
modelling of strongly residual metapelitic xenoliths within
the southern Platreef, Bushveld Complex, South Africa.
Journal of Metamorphic Geology 28, 269–291.
Johnson, T. E., White, R. W. & Brown, M. (2011). A year in the
life of an aluminous metapelite xenolith—the role of heating
rates, reaction overstop, H2O retention and melt loss. Lithos
124, 132–143.
Kelsey, D. E., White, R. W., Holland, T. J. B. & Powell, R. (2004).
Calculated phase equilibria in K2O–FeO–MgO–Al2O3–SiO2–
H2O for sapphirine–quartz-bearing mineral assemblages.
Journal of Metamorphic Geology 22, 559–578.
Lyubetskaya, T. & Ague, J. J. (2010). Modeling metamorphism
in collisional orogens intruded by magmas: II. Fluid flow and
implications for Barrovian and Buchan metamorphism,
Scotland. American Journal of Science 310, 450–491.
Markl,
G.
(2005).
Mullite–corundum–spinel–cordierite–
plagioclase xenoliths in the Skaergaard Marginal Border
Group: multi-stage interaction between metasediments and
basaltic magma. Contributions to Mineralogy and Petrology
149, 196–215.
Marschall, H. R., Dohman, R. & Ludwig, T. (2013). Diffusioninduced fractionation of niobium and tantalum during continental crust formation. Earth and Planetary Science Letters
375, 361–371.
Maury, R. C. & Bizouard, H. (1974). Melting of acid xenoliths
into a basanite: an approach to the possible mechanisms of
crustal contamination. Contributions to Mineralogy and
Petrology 48, 275–286.
McBirney, A. R. (2007). Igneous Petrology. Jones and Bartlett.
Müntener, O., Kelemen, P. B. & Grove, T. L. (2001). The role of
H2O during crystallisation of primitive arc magmas under
uppermost mantle conditions and genesis of igneous pyroxenites: an experimental study. Contributions to Mineralogy
and Petrology 141, 643–658.
Mysen, B. O. & Virgo, D. (1986). Volatiles in silicate melts at
high pressure and temperature. 1. Interaction between OH
groups and Si4þ, Al3þ, Ca2þ, Naþ and Hþ. Chemical Geology
57, 303–331.
Newton, R. C. (1972). An experimental determination of the
high-pressure stability limits of magnesian cordierite under
wet and dry conditions. Journal of Geology 80, 398–420.
Newton, R. C. & Wood, B. J. (1979). Thermodynamics of water
in cordierite and some petrologic consequences of cordierite as a hydrous phase. Contributions to Mineralogy and
Petrology 68, 391–405.
407
408
Waldron, K. A. (1986). High pressure contact metamorphism in
the aureole of the Cortlandt Complex, southeastern New
York. PhD thesis, Yale University, New Haven, CT.
White, R. W., Powell, R., Holland, T. J. B. & Worley, B. (2000).
The effect of TiO2 and Fe2O3 on metapelitic assemblages at
greenschist and amphibolite facies conditions: mineral equilibria calculations in the system K2O–FeO–MgO–Al2O3–
SiO2–H2O–TiO2–Fe2O3. Journal of Metamorphic Geology 18,
497–511.
White, R. W., Powell, R. & Holland, T. J. B. (2001).
Calculation of partial melting equilibria in the system
Na2O–CaO–K2O–FeO–MgO–Al2O3–SiO2–H2O (NCKFMASH).
Journal of Metamorphic Geology 19, 139–153.
White, R. W., Powell, R. & Clarke, G. L. (2002). The interpretation
of reaction textures in Fe-rich metapelitic granulites of the
Musgrave Block, central Australia: constraints from mineral
equilibria calculations in the system K2O–FeO–MgO–
Al2O3–SiO2–H2O–TiO2–Fe2O3. Journal of Metamorphic
Geology 20, 41–55.
White, R. W., Pomroy, N. E. & Powell, R. (2005). An in situ
metatexite–diatexite transition in upper amphibolite facies
rocks from Broken Hill, Australia. Journal of Metamorphic
Geology 23, 579–602.
White, R. W., Powell, R. & Holland, T. J. B. (2007). Progress relating to calculation of partial melting equilibria for metapelites. Journal of Metamorphic Geology 25, 511–527.
Downloaded from http://petrology.oxfordjournals.org/ at Oberlin College on March 24, 2015
FeO–MgO–Al2O3–SiO2–TiO2–O. Journal of Metamorphic
Geology 28, 615–633.
Theriault, R. D., Barnes, S.-J. & Severson, M. J. (1997).
The influence of country-rock assimilation and silicate to
sulfide ratios (R factor) on the genesis of the Dunka
Road Cu–Ni–platinum-group element deposit, Duluth
Complex, Minnesota. Canadian Journal of Earth Sciences
34, 375–389.
Tracy, R. J. (2009). High pressure crystallization of mafic
magma: field observations, compositional measurements
and computer modeling. EOS Transactions, American
Geophysical Union 90, Abstract #V11A-04.
Tracy, R. J. & McLellan, E. L. (1985). A natural example of the
kinetic controls of compositional and textural equilibration.
In: Thompson, A. B. & Rubie, D. C. (eds) Metamorphic
Reactions; Kinetics, Textures, and Deformation. Advances in
Physical Geochemistry 4, 118–13.
Viete, D. R., Hermann, J., Lister, G. S. & Stenhouse, I. R. (2011).
The nature and origin of the Barrovian metamorphism,
Scotland: diffusion length scales in garnet and inferred thermal time scales. Journal of the Geological Society, London
168, 115–13.
Vorhies, S. H. & Ague, J. J. (2011). Pressure–temperature
evolution and thermal regimes in the Barrovian zones,
Scotland. Journal of the Geological Society, London 168,
1147–1166.
Journal of Petrology, 2015, Vol. 56, No. 2