JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 PAGES 299^324 2012 doi:10.

1093/petrology/egr063

Detecting Peritectic Garnet in the Peraluminous

Cardigan Pluton, New Hampshire

MICHAEL J. DORAIS1* AND MICHAEL TUBRETT2

1
DEPARTMENT OF GEOLOGICAL SCIENCES, BRIGHAM YOUNG UNIVERSITY, PROVO, UT 84602, USA

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

2
CREAIT NETWORK, MICROANALYSIS FACILITY, INCO INNOVATION CENTRE, (MAF-IIC), MEMORIAL UNIVERSITY
OF NEWFOUNDLAND, ST. JOHN’S, NEWFOUNDLAND, A1B 3X5, CANADA

RECEIVED FEBRUARY 8, 2011; ACCEPTED NOVEMBER 8, 2011

ADVANCE ACCESS PUBLICATION DECEMBER 19, 2011

The peraluminous Cardigan Pluton of New Hampshire contains res- Chappell (1996) noted that many peraluminous plutons
titic garnetites with abundant peritectic garnet that serves as a show negative correlations between bulk-rock normative
basis for comparison with garnet in the host-rock. The peritectic corundum and silica. For example, the Cardigan Pluton
garnet is inclusion-rich and contains high concentrations of heavy of New Hampshire, USA, has a wider range in silica con-
rare earth elements (HREE), Sc, and Zr and low concentrations of tents than most peraluminous plutons, with some portions
Zn, P, and U. The high HREE, Zr, and Sc result from biotite and of the pluton having silica contents as low as 55 wt %
zircon melting, buffering liquid compositions. Because the HREE (Dorais et al., 2009a). These more mafic rocks have higher
and Sc are compatible in garnet, the peritectic garnet is rich in these normative corundum contents than the more evolved
elements. Some garnet crystals in the host-rocks have cores with iden- rocks, reaching 8% (Fig. 1). Simple magma mixing
tical inclusions and trace element compositions to the peritectic cannot produce these compositions because the mafic
garnet in the garnetites. We infer that these are also peritectic garnet end-member would have to have been strongly peralu-
grains dispersed in the host magma. These cores are mantled by minous and no basaltic liquids with such high Al/
inclusion-poor garnet. These mantles, along with other garnet crys- (Na þ K þCa) exist. Additionally, Stevens et al. (2007)
tals that are inclusion-poor throughout, show rimward trends of have argued that experimental melts of metasediments
decreasing HREE, Sc, and Zr, and increasing Zn, P, U, and have ferromagnesian contents that are considerably lower
Eu/Eu*.We interpret the trace element variations as records of frac- than those of most of the peraluminous granites of the
tionation as the melt decreased in HREE, Zr, and Sc and was en- Cape Granite Suite of South Africa. They suggested that
riched Zn, P, and U. The rimward increase in Eu/Eu* records the the compositions of the more mafic granitoids require an
effect of feldspar fractionation. Thus the peritectic garnet preserves addition of an Mg- and Fe-rich component to the magma.
evidence of crystallization from buffered melts whereas the pheno-
The addition of garnet satisfies the high normative corun-
crystic garnet shows variable trace element abundances indicative of
dum, Mg and Fe contents of the more mafic granitoids
crystallizing from fractionating melts. These differences provide cri-
for both the Cardigan Pluton and the Cape Granite
teria for the recognition of peritectic garnet in the host pluton.
Suite (Stevens et al., 2007; Dorais et al., 2009a; Villaros
et al., 2009a).
Two explanations have been proposed to explain the
KEY WORDS: granite; restite; peritectic garnet
excess garnet in the mafic rocks. The first is that the mafic
rocks contain accumulations of garnet phenocrysts. In this
scenario, the mafic rocks represent cumulates, accounting
I N T RO D U C T I O N for their non-liquid compositions. The other possibility is
A fundamental question regarding the petrogenesis of that the mafic rocks contain selectively entrained peritectic
peraluminous plutons is what process controls their com- garnet, resulting from biotite dehydration melting in
positional variations. White & Chappell (1988) and the source that produced the granitic melt. Mobilization

ß The Author 2011. Published by Oxford University Press. All

rights reserved. For Permissions, please e-mail: journals.permissions@
*Corresponding author. E-mail: [email protected] oup.com
 JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 FEBRUARY 2012

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 1. Bulk-rock SiO2 vs normative corundum diagram for garnetites (filled circles) and plutonic host-rocks (open circles) of the Cardigan
pluton (from Dorais et al., 2009a). The pluton defines a trend that projects back to garnetite samples. The high normative corundum contents
of the most mafic host-rock samples negates the role of simple magma mixing in the generation of this compositional trend, instead indicating
the presence of an aluminum-rich, garnetite component to the Cardigan Pluton. The gray field represents the composition of garnet in the
garnetites.

of mixtures of peritectic garnet and melt produces mafic 2009a). The pluton was syntectonically emplaced with
granitoids. The compositional spectrum of the Cape nappes during the Acadian Orogeny (Eusden & Lyons,
Granite Suite and other peraluminous plutons is thus inter- 1993; Dorais, 2003) where rapid magma ascent carried the
preted to result from varying proportions of peritectic high-density garnetites to the emplacement level. As res-
products entrained in the melt (Stevens et al., 2007; tite, the garnetites provide the opportunity to compare
Villaros et al., 2009a, 2009b; Kotkova¤ & Harley, 2010; the major and trace element compositions and inclusion
Taylor & Stevens, 2010). relations of peritectic garnet with those of garnet in the
These two models are fundamentally different. One em- host pluton.
phasizes the role of fractional crystallization and the con- Under equilibrium melting conditions, peritectic garnet
comitant accumulation of phenocryst phases, primarily should have trace element abundances that are consistent
garnet, to generate the more mafic granites. The other, with its formation from buffered liquids that were rich in
involving selective entrainment of peritectic phases, is a elements contributed from the melting phases [e.g. heavy
modification of the initial restite unmixing model of rare earth elements (HREE) from zircon, Sc from biotite].
White & Chappell (1977). The inferred restitic material These compositions should be distinguishable from those
need not consist of minerals in the source rocks that were of phenocrystic garnet because the phenocrysts should
not involved in the melting reaction or remnants of min- show rimward decreasing concentrations of compatible
erals partially consumed by anatexis, but instead is repre- elements and increasing concentrations of incompatible
sented by the solid products of incongruent melting elements as garnet crystallized from fractionating melts.
(Taylor & Stevens, 2010). These trace element differences, along with the inclusion-
In the case of plutons with abundant garnet in the mafic rich nature of the peritectic garnet in the garnetites, allow
granitoids it is imperative to determine whether the distinction between peritectic and phenocrystic garnet in
garnet is phenocrystic or peritectic. Without confident the host pluton. With this distinction, we can then proceed
identification of the origin of the garnet, the interpretation to evaluate the nature of the compositional zoning of the
of the chemical variation of a pluton remains elusive. The Cardigan Pluton.
410 Ma Cardigan Pluton of the New Hampshire Plutonic
Suite has a bearing on the problem because it contains dis-
tinctive pods of garnetite that have up to 70 modal % GEOLOGIC A L S ET T I NG
garnet with lesser amounts of quartz, plagioclase, silliman- The Cardigan Pluton is by far the largest pluton of the
ite, and biotite. The garnetites have been interpreted to syn-tectonic Acadian plutons in New Hampshire (Fig. 2).
represent restite, with the garnet representing a peritectic It measures c. 20 km  90 km at its maximum width and
product of biotite dehydration melting (Dorais et al., length (Lyons et al., 1997), trending roughly north^south

300
 DORAIS & TUBRETT PERITECTIC GARNET, CARDIGAN PLUTON

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 2. Simplified geological map of the Kinsman and Bethlehem members of the New Hampshire Plutonic Suite in New Hamsphire
(after Clark & Lyons, 1986). The Kinsman member intrudes the metasediments of the Central Maine Trough.

across southwestern New Hampshire (Fig. 2). According to quartz, K-feldspar, biotite, variable amounts of garnet
gravity studies, the pluton is tabular in shape, with a max- and some secondary muscovite. Accessory phases include
imum thickness of 2·5 km (Nielson et al., 1976). It is a monazite, zircon, ilmenite and rare sillimanite and graph-
medium- to coarse-grained, massive to rarely strongly ite. The appearance in the field is distinctive, with
foliated, granitoid body consisting of oligoclase^andesine, K-feldspar crystals up to 15 cm in length. The pluton is

301
 JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 FEBRUARY 2012

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 3. Photomicrograph of biotite in a garnetite sample breaking down from dehydration melting to produce peritectic garnet. Sillimanite
mats are present in the lower corners of the photomicrograph. Width of view is 4 mm.

predominantly granodiorite in composition, but a con- use of the term peritectic for the inclusion-rich garnet
tinuum exists between minor amounts of quartz diorite grains in the garnetites.
and relatively abundant granite (Clark & Lyons, 1986; The pluton contains rare but distinctive meter-sized
Dorais et al., 2009a). pods consisting of 50^70 modal % garnet (þ silliman-
Garnet-rich pods (garnetites) that are generally up to ite þ biotite þ plagioclase þ quartz; see Dorais et al.,
1m in size are present in the central portions of the 2009a, fig. 2). Garnet typically occurs as rounded subhe-
pluton. Larger bodies up to several tens of meters in size dral red^brown crystals; rare euhedral grains are also pre-
were mined in the past (Conant, 1935) but are no longer sent. Grain sizes range from 2 to 12 mm, with an average
accessible, as they either have been overgrown with vegeta- grain size of 8 mm. The garnet crystals contain inclusions
tion or are located on private property. Poor exposures of quartz, plagioclase, biotite, ilmenite, apatite, sillimanite,
make volumetric estimates of the garnetites difficult, but zircon, and monazite.
they represent a very minor (1%) amount of the The garnetites contain two generations of biotite. The
exposed area, less than 100 m2 in the 41000 km2 area of more abundant type is secondary biotite that occurs along
pluton. fractures and on the rims of garnet. A much rarer version
The Cardigan Pluton intruded a package of allochthon- occurs as large red^brown biotite grains with modal abun-
ous metasediments within the Central Maine Trough. dances of 5%. Some of these grains are rich in ilmenite
The thickness of the sedimentary sequence is estimated to inclusions that follow the biotite crystallographic orienta-
be from 3 to 10 km (Eusden & Lyons, 1993; Stewart et al., tions, forming a distinctive cross-hatching pattern, suggest-
1993; Thompson et al., 1993). ing that the mica was initially a Ti-rich, high-temperature
biotite. Some biotite grains have textures indicative of
EV I DENC E FOR PER I T ECT IC partial melting (Fig. 3). Muscovite is not present and silli-
manite occurs as fibrolitic mats associated with biotite.
GA R N ET I N T H E CA R DIGA N These fibrolitic mats commonly radiate from biotite
GA R N ET I T ES grains and are oriented at random angles to plagioclase
Evidence that the garnetites in the Cardigan Pluton repre- and quartz grains.
sent restite was presented by Dorais et al. (2009a). Here we The rocks immediately surrounding the garnetite pods
review the conclusions presented in that paper to provide are quartz diorite in composition and gradually grade
a basis for the results presented here, and to justify our into granodiorite within several meters from the pods.

302
 DORAIS & TUBRETT PERITECTIC GARNET, CARDIGAN PLUTON

In the vicinity of the garnetites, the host pluton contains up garnet formed as a peritectic product. It appears that the
to 25 modal % garnet. Fine-grained, muscovite- and peritectic garnet grew in the presence of apatite, zircon
biotite-rich (þ plagioclase þ quartz) metasedimentary and monazite, as these minerals form abundant inclusions
xenoliths are also present in the pluton. The xenoliths in the garnets. Dorais et al. (2009a) showed that although
range from centimeter- to meter-scale in length. many zircons in the garnetites formed at 413 Ma, a signifi-
Several mineralogical and geochemical characteristics cant number are inherited. Thus it appears that the local
of restites as defined by Clemens (2003) suggest that the magma was saturated in zircon and unable to dissolve all
garnetites are restitic, as follows. (1) Melt-depleted compos- the zircon grains at the site of melting. As peritectic
itions. The garnetites have low SiO2, K2O, Rb, Ba, and Sr garnet formed, it incorporated a large number of these
contents, as well as very high Fe2O3, MgO, TiO2, Sc, Cr, undissolved grains. In summary, the garnetites are melt-
V, Y, and HREE concentrations compared with any depleted, residual rocks that are in isotopic and mineral-

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Central Maine Trough metasediments reported in the lit- ogical equilibrium with their host granitic rocks. We
erature (Cullers et al., 1997; Dorais et al., 2009a, 2009b; see therefore refer to the garnet grains in the garnetites as
Dorais et al., 2009a, fig. 6). (2) Isotopic equilibrium with peritectic garnet throughout the remainder of this paper.
the host pluton. eNd(413 Ma) and 87Sr/86Sri data show that
the garnetites are in isotopic equilibrium with the host
A N A LY T I C A L M E T H O D S
granitic rocks (Dorais et al., 2009a, fig. 9). Both the garne-
tites and granitic hosts define a constrained field between Electron microprobe analyses and imaging were con-
ducted at Brigham Young University with a Cameca
eNd(413 Ma) of 3 and 5 and 87Sr/86Sri of 0·708 and 0·715.
SX50. Thick sections were used in order to conduct laser
(3) Mineralogical equilibrium with the host pluton.
ablation inductively coupled plasma mass spectrometry
Garnet, biotite, and plagioclase compositions are the
(LA-ICPMS) analyses on the same grains as analyzed by
same in both the garnetites and the host-rocks (Plank,
electron microprobe. Full section scans for Fe Ka, Mg
1987; Dorais et al., 2009a). (4) The garnetites are from
Ka, Mn Ka and Ca Ka were conducted at 100 nA to deter-
deeper sources and are not melt-depleted xenoliths of the
mine potential garnet zoning. Detailed traverses of these
wall-rocks. With the exception of the Rangeley Formation,
grains were conducted to determine the nature of the
the lowest formation in the Central Maine Trough, the iso-
zoning and to provide Si contents for LA-ICPMS data re-
topic compositions of the garnetites are distinct from all
duction. Representative analyses are presented in Table 1.
Central Maine Trough formations (Dorais et al., 2009a,
Trace element analyses were conducted at the Memorial
fig. 9). The vast majority of the metasediments have eNd(413 University of Newfoundland (MUN) using LA-ICPMS.
Ma) values between 8 and 10, lower than the 5 aver- Representative analyses are reported in Table 2, and the
age of the garnetites. The Rangeley Formation has similar full dataset is available in Electronic Appendix 1 (available
isotopic compositions to both the Cardigan Pluton and for downloading at http://www.petrology.oxfordjournals
the garnetites; however, a major difference between the .org/). The analytical system is a Finnigan ELEMENT
garnetites and the Rangeley Formation is that the garne- XR, a high-resolution double focusing magnetic sector
tites contain inherited zircons with ages of 600 Ma inductively coupled plasma mass spectrometry
(Dorais et al., 2009a, fig. 5). Ages of 600 Ma are indicative (HR-ICPMS) system coupled to a GEOLAS 193 nm ex-
of peri-Gondwanan sources. The Rangeley Formation has cimer laser system. A helium flow rate of 1·25 l min1 was
strong paleocurrent indicators of a Laurentian source used to carry ablated material to the ICP, with an add-
(Bradley & Hanson, 2002), a source that lacks 600 Ma zir- itional argon make-up gas added after the ablation cell.
cons. Another complication is that most analyses of the A laser spot size of 30 mm was used with an energy dens-
Rangeley Formation in the literature show the rocks are ity of c. 5 J cm2, with a laser repetition rate of 10 Hz.
rich in SiO2, generally greater than 70 wt %. If these ana- Time-resolved intensity data were acquired by peak-
lyses are representative of the Rangeley at depth, then pro- jumping in a combination of pulse-counting, analog and
ducing garnetites with SiO2 contents between 40 and Faraday modes, depending on signal strength, with one
50 wt % from this source would require extraordinarily point measured per peak.
high percentages of partial melting. Thus there seems to Concentrations were calibrated with NIST 612 glass.
be no known source in the Central Maine Trough that Silica (SiO2) was used as an internal standard to deal
could account for the composition of the melt-depleted with differences in ablation yields and matrix effects be-
garnetites. It appears that the garnetites are melt-depleted tween the garnet and the calibration material (NIST
rocks derived from sources below the Central Maine meta- glass). Approximately 30 s of gas background data were
sediments, consistent with the thermobarometric calcula- collected prior to each 60 s ablation of both standards and
tions of Dorais et al. (2009a). unknowns. The data acquisition method employed an ana-
Our interpretation of the garnetites is that they repre- lytical sequence of two analyses of the NIST 612 standard
sent residues from biotite dehydration melting, with and one of USGS BCR2G, followed by analyses of up to

303
 JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 FEBRUARY 2012

Table 1: Representative garnet electron microprobe analyses

CP-4F CP-4F CP-27 CP-27 CP-27 CP-27 CP-42 CP-42

Gar1 Gar1 Gar1 Gar1 Gar2 Gar2 Gar1 Gar1
Pe Pe Pe Pe Pe Pe Pe/Ph Pe/Ph

SiO2 38·093 37·876 37·376 37·242 37·561 37·714 37·841 38·388

TiO2 0·011 0·048 1·046 1·092 1·359 1·282 0·708 0·686
Al2O3 22·126 21·852 21·439 21·271 21·493 21·43 22·155 21·909
MgO 7·158 7·423 4·614 4·61 4·179 4·721 7·372 7·236

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

FeO 30·684 30·16 0·028 0·036 0·019 0·044 0·032 0·049
Fe2O3 0 0 0 0 0 0 0 0
MnO 1·14 1·05 33·632 33·997 34·43 33·915 30·632 30·795
CaO 1·443 1·449 1·286 1·337 1·301 1·359 1·343 1·387
Total 100·655 99·858 99·421 99·585 100·342 100·465 100·083 100·45
Si 5·93 5·934 5·982 5·967 5·981 5·981 5·915 5·975
Ti 0·001 0·006 0·142 0·148 0·183 0·172 0·094 0·09
Al 4·06 4·035 4·044 4·016 4·034 4·006 4·081 4·019
Mg 1·661 1·734 1·101 1·101 0·992 1·116 1·718 1·679
Fe2þ 3·995 3·952 0·003 0·004 0·002 0·005 0·004 0·006
Fe3þ 0 0 0 0 0 0 0 0
Mn 0·15 0·139 4·501 4·555 4·585 4·498 4·004 4·009
Ca 0·241 0·243 0·221 0·23 0·222 0·231 0·225 0·231
Total 16·038 16·043 15·994 16·021 15·999 16·009 16·041 16·009
Ad 0·02 0·084 0·05 0·064 0·034 0·078 0·056 0·086
Gr 3·949 3·87 3·615 3·7 3·654 3·71 3·632 3·709
Py 27·476 28·587 18·464 18·257 16·585 18·559 28·449 27·955
Sp 2·487 2·298 2·378 2·456 3·066 2·863 1·552 1·507
Al 66·069 65·161 75·493 75·523 76·662 74·789 66·311 66·744

CP-42 CP-42 CP-43 CP-43 CP-43 CP-43 K-1 K-1

Gar2 Gar2 Gar1 Gar1 Gar2 Gar2 Gar4e Gar4f
Pe/Ph Pe/Ph Ph Ph Ph Ph Pe/Ph Pe/Ph

SiO2 38·054 38·05 38·171 38·066 37·852 37·695 38·107 37·949

TiO2 0·686 0·69 0·674 0·679 0·677 0·642 0·766 0
Al2O3 22·068 21·842 21·753 21·832 21·902 22·276 21·513 21·621
MgO 7·154 6·966 7·17 7·307 7·007 7·048 7·781 7·675
FeO 0·022 0·047 0·033 0·027 0·027 0·07 0·051 0·041
Fe2O3 0 0 0 0 0 0 0·258 0
MnO 1·293 1·353 1·301 1·322 1·308 1·296 1·274 1·229
CaO 31·492 31·735 30·512 30·013 30·323 30·118 29·941 30·6
Total 100·769 100·683 99·614 99·246 99·096 99·145 99·691 99·115
Si 5·926 5·94 5·987 5·981 5·967 5·932 5·967 5·971
Ti 0·09 0·091 0·09 0·09 0·09 0·086 0·102 0
Al 4·05 4·019 4·021 4·043 4·069 4·132 3·97 4·009
Mg 1·661 1·621 1·676 1·711 1·647 1·653 1·816 1·8
Fe2þ 0·003 0·006 0·004 0·003 0·003 0·008 0·006 0·005
Fe3þ 0 0 0 0 0 0 0·03 0
Mn 4·101 4·143 4·002 3·944 3·998 3·964 3·921 4·027
Ca 0·216 0·226 0·219 0·223 0·221 0·219 0·214 0·207
Total 16·047 16·046 15·999 15·995 15·995 15·994 16·026 16·019
Ad 0·038 0·082 0·059 0·048 0·049 0·126 0·844 0·073
Gr 3·492 3·587 3·554 3·651 3·629 3·483 2·631 3·314
Py 27·374 26·67 28·013 28·686 27·658 27·946 30·028 29·851
Sp 1·492 1·501 1·497 1·514 1·518 1·447 1·68 0
Al 67·604 68·16 66·877 66·101 67·146 66·997 64·817 66·762

Pe, peritectic garnet; Ph, phenocrystic garnet; Pe/Ph, cores of grains are peritectic, mantled by phenocrystic garnet.

304
 DORAIS & TUBRETT PERITECTIC GARNET, CARDIGAN PLUTON

Table 2: Representative LA-ICPMS analyses of Cardigan garnets

CP-27 CP-27 CP-27 CP-27 CP-27 CP-27 CP-27 CP-27

Gar1A Gar1F Gar1G Gar1N Gar2B Gar2F Gar2I Gar2N
Core-Pe Pe Pe Rim-Pe Core-Pe Pe Pe Rim-Pe

P2O5 0·020 0·029 0·033 0·032 0·028 0·022 0·024 0·031

Y 434 581 616 71·6 559 531 520 230
Zr 12·7 17·4 15·4 11 8·61 11·7 12·5 6·05
Ce 0·020 0·020 0·026 0·021 0·007 0·014 0·008 50·008

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Pr 0·022 0·022 0·028 0·026 0·010 0·017 50·0086 0·015
Nd 0·585 0·72 0·567 0·502 0·258 0·459 0·0875 0·352
Sm 3·53 3·41 2·88 2·38 0·953 2·7 1·23 1·62
Eu 0·143 0·151 0·136 0·113 0·0754 0·136 0·0629 0·0825
Gd 21·8 21·8 18·4 12·1 5·24 19·2 15·1 9·75
Tb 7·7 8·11 7·36 3·55 3·14 6·96 6·68 3·78
Dy 72·7 86·6 85·4 22·1 54·6 76·1 78·5 36·7
Ho 17 22·9 25·9 2·51 22·4 21·5 22·1 8·23
Er 53·8 77·7 100 4·05 105 78·1 78 22·7
Tm 8·46 12·6 17·9 0·519 20·1 13·3 12·7 3·04
Yb 60·4 88·2 138 2·99 151 102 92·4 19·6
Lu 9·23 13·1 22·4 0·418 20·4 16·5 14·3 2·46
Hf 0·207 0·295 0·31 0·228 0·207 0·229 0·261 0·15
U 0·0059 0·0038 0·0078 0·0041 0·0023 0·0038 0·0042 0·0042
Sc 130 134 129 186 145 125 98·4 184
V 135 147 113 63·1 63·9 153 101 48·2
Cr 147 149 115 228 60·4 243 74·3 101
Co 21·7 21·1 20·1 15·6 20·4 23·8 22·4 16·9
Zn 79·2 79·3 69·7 48·5 62·1 99·1 90·8 51·8

CP-4F CP-4F CP-4F CP-4F CP-42 CP-42 CP42 CP-42

Gar1A Gar1C Gar1F Gar1J Gar1A Gar1D Gar1M Gar1N
Core-Pe Pe Pe Pe Core-Pe Pe Ph Rim-Ph

P2O5 0·036 0·044 0·033 0·034 0·023 0·024 0·034 0·034

Y 366 395 388 500 506 507 230 217
Zr 17·9 24·2 17·6 16·4 14·4 13·7 13·7 16·9
Ce 0·013 50·006 0·010 0·085 50·008 50·008 0·051 0·050
Pr 0·015 0·009 0·020 0·010 0·009 0·022 0·064 0·056
Nd 0·391 0·536 0·651 0·196 0·397 0·813 0·916 0·985
Sm 2·8 3·54 2·92 2·24 2·62 3·32 1·44 1·83
Eu 0·156 0·171 0·162 0·099 0·094 0·138 0·094 0·118
Gd 20 23·9 18·9 18·4 18·6 18·7 9·66 11·9
Tb 7·11 8·45 6·93 7·41 6·82 7·15 4 4·48
Dy 66·2 74·8 66·4 80·7 74·8 76·8 42·2 45·3
Ho 13·1 15·4 15·1 20·4 21·4 21·6 8·39 8·46
Er 34·8 42·3 42·6 60·5 74·8 76·6 15·4 12·9
Tm 4·63 5·82 6·14 8·38 11·9 12·8 0·907 0·711
Yb 30·1 37·5 41 52·8 83·5 92·8 2·9 2·47
Lu 4·08 5·16 5·31 6·94 12·8 13·7 0·255 0·274
Hf 0·328 0·663 0·373 0·354 0·278 0·245 0·266 0·321
U 0·0067 0·0045 0·0081 0·0248 0·0026 0·0022 0·0156 0·0184
Sc 167 143 142 135 140 130 89·6 91·9
V 125 124 102 108 108 134 148 138
Cr 175 77 125 96·6 194 191 92·3 97
Co 28·9 28·9 30·1 26·9 28·8 27·6 29·7 26·3
Zn 150 170 164 129 130 128 157 126

(continued)

305
 JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 FEBRUARY 2012

Table 2: Continued

CP-42 CP-42 CP-42 CP-42 K-1 K-1 K-1 K-1

Gar2A Gar2D Gar2K Gar2M Gar4A Gar4C Gar4L Gar4M
Core-Pe Pe Ph Rim-Ph Core-Pe Pe Ph Ph

P2O5 0·034 0·021 0·044 0·047 0·0515 0·0407 0·0357 0·0444

Y 163 479 44 40 346 385 142 60
Zr 18·7 8·2 11·0 15·9 15·8 22·0 10·4 11·4
Ce 0·034 0·02 0·051 0·076 50·00819 0·0393 0·0652 0·0663

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Pr 0·046 0·024 0·047 0·059 50·00965 0·0186 0·0446 0·061
Nd 1·1 0·701 0·89 1·191 0·432 0·684 0·737 0·735
Sm 3·927 2·29 2·615 3·621 3·13 4·27 1·91 2·37
Eu 0·165 0·109 0·113 0·145 0·133 0·178 0·0917 0·122
Gd 14·9 9·4 12·8 15·2 21·4 24·5 11·5 12·6
Tb 3·86 3·82 3·36 2·95 7·02 7·85 4·09 3·59
Dy 28·5 51·4 16·8 12·7 60·8 69·7 34·2 21·4
Ho 5·4 19·3 1·5 1·4 13·5 14·8 5·3 2·1
Er 15·3 94·6 2·3 3·0 40·5 42·2 7·0 2·9
Tm 2·35 20·57 0·28 0·39 5·96 6·22 0·42 0·30
Yb 16·61 177·41 1·69 2·36 41·00 41·90 1·73 1·86
Lu 2·41 30·70 0·25 0·32 5·89 5·77 0·25 0·26
Hf 0·328 0·174 0·246 0·261 0·395 0·501 0·177 0·278
U 0·0080 0·0100 0·0120 0·0060 0·0012 0·0063 0·0126 0·0103
Sc 88 102 60 67 114 125 88 66
V 167 108 114 147 110 166 108 121
Cr 212 78 115 176 172 176 89 122
Co 28·0 28·3 29·0 25·5 30·2 31·1 34·0 33·4
Zn 143·4 150·8 165·9 116·7 107·0 106·0 182·0 177·0

CP-43 CP-43 CP-43 CP-43 CP-43 CP-43 CP-43Gar2J CP-43

Gar1A Gar1E Gar1J Gar1L Gar2E Gar2D Gar2K
Core-Pe Ph Ph Rim-Ph Pe Pe Pe Ph

P2O5 0·038 0·046 0·052 0·046 0·020 0·043 0·038 0·042

Y 390 152 44 137 479 299 127 113
Zr 12·5 12·2 16·6 25·4 10·4 12·8 17 13
Ce 0·037 0·054 0·119 0·107 0·012 0·006 0·024 0·026
Pr 0·032 0·065 0·079 0·078 0·025 0·031 0·042 0·036
Nd 0·517 0·614 1·370 1·630 0·893 0·826 1·030 0·985
Sm 1·31 1·80 4·00 4·89 2·36 1·32 2·89 2·67
Eu 0·082 0·115 0·157 0·156 0·117 0·093 0·146 0·131
Gd 9·2 10·8 16·4 19·3 9·1 8·0 13·4 12·1
Tb 4·14 3·98 3·31 4·15 3·81 3·6 3·61 3·84
Dy 52·9 35·1 14·0 27·0 51·6 43·0 26·4 31·1
Ho 15·2 5·6 1·66 5·11 18·9 10·7 4·34 4·11
Er 44·8 7·23 3·39 13·9 87·1 38 8·48 4·84
Tm 4·67 0·463 0·423 2 17·4 7·73 1·04 0·361
Yb 17·5 2·08 2·44 14·3 137 71·8 6·88 1·9
Lu 1·16 0·234 0·356 2·03 20·6 12 1·03 0·252
Hf 0·31 0·276 0·326 0·427 0·198 0·294 0·298 0·226
U 0·0096 0·0179 0·0127 0·0201 0·0079 0·0042 0·0128 0·0147
Sc 150 116 51·2 78·1 134 111 113 99·5
V 65·5 70·5 156 191 96·1 134 134 138
Cr 82·3 96 176 253 86·9 79·2 153 85
Co 34·2 33·7 32·2 30 31·7 31·1 32·9 33·3
Zn 199 198 188 137 171 179 193 207

Pe, peritectic garnet; Ph, phenocrystic garnet.

306
 DORAIS & TUBRETT PERITECTIC GARNET, CARDIGAN PLUTON

Table 3: Statistical data for BCR2G analyses

Isotopic Unit Long-term This study LOD

mass average Average STD RSD (%) Study av./ (ppm)
(n ¼ 231) (n ¼ 18) MUN MPV (%)

P2O5 31 wt % 0·37* 0·27 0·03 11·5 27·0 0·0006

Y 89 ppm 31·3 30·08 0·94 3·1 3·9 0·0363
Zr 90 ppm 162 153·17 3·94 2·6 5·4 0·0062
Ce 140 ppm 51 48·64 1·01 2·1 4·6 0·0028

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Pr 141 ppm 6·4 5·97 0·16 2·7 6·7 0·0027
Nd 146 ppm 27·5 25·83 0·75 2·9 6·1 0·0133
Sm 147 ppm 6·2 5·83 0·29 5·0 6·0 0·0098
Eu 153 ppm 1·85 1·73 0·05 2·8 6·2 0·0027
Gd 157 ppm 6·03 5·81 0·17 2·9 3·7 0·0127
Tb 159 ppm 0·93 0·89 0·03 3·6 4·1 0·0017
Dy 163 ppm 5·92 5·73 0·22 3·9 3·2 0·0067
Ho 165 ppm 1·19 1·15 0·04 3·8 3·1 0·0020
Er 166 ppm 3·3 3·22 0·16 5·0 2·5 0·0053
Tm 169 ppm 0·48 0·48 0·03 5·8 0·6 0·0019
Yb 172 ppm 3·23 3·14 0·17 5·4 2·6 0·0078
Lu 175 ppm 0·48 0·47 0·03 6·5 3·0 0·0018
Hf 178 ppm 4·57 4·14 0·21 5·1 9·5 0·0041
U 238 ppm 1·58 1·63 0·06 3·8 3·4 0·0004
Sc 45 ppm 34·8 32·13 0·64 2·0 7·7 0·0742
V 51 ppm 400 422·03 6·13 1·5 5·5 0·0267
Cr 53 ppm 16·9 16·14 1·04 6·5 4·5 0·8776
Co 59 ppm 35·4 37·22 0·72 1·9 5·1 0·0331
Zn 64 ppm 169 144·24 4·10 2·8 14·6 0·1310

*GEOREM value.

14 unknowns, closing with a repetition of the same stand- the procedures described by Longerich et al. (1996).
ards in reverse order. The USGS BCR2G standard was LAMTRACE allows selection of representative signal
treated as an unknown and data were acquired to monitor intervals, background subtraction, and internal standard
the accuracy and precision of the dataset. The results are correction for ablation yield differences and instrument
given inTable 3. The error for this method when measuring sensitivity drift during the analytical session, and performs
homogeneous materials is estimated to be better than calculations converting count rates into concentrations by
4^7% relative based on the reproducibility of results for reference to the standards.
various reference materials measured from day to day Twelve garnet grains were chosen for study: five grains
over several months in the MUN laboratory. Elements in from garnetites and seven from the host pluton. We
the table showing high RSD values can be directly attribu- present X-ray dot maps, electron microprobe traverses,
ted to heterogeneities in BCR2G for particular elements, and chondrite-normalized REE patterns for eight repre-
especially at low concentrations and for elements prone to sentative grains. Only a few analyses were made of the re-
fractionation. Limits of detection (LOD) are given in maining four grains and these are similar to the data for
Table 3. These were calculated based on the calibration the eight garnets presented. Based on X-ray dot maps and
standards backgrounds and signals. For each element re- electron microprobe profiles, locations for garnet radius
ported in Table 3 and accompanying figures, all analytical traverses by LA-ICPMS were selected. Attempts were
data are above the LOD. made to avoid locations at the outermost portions of rims
Data were reduced using MUN’s in-house CONVERT and adjacent to biotite inclusions, to avoid potential diffu-
and LAMTRACE spreadsheet programs, which employ sion of trace elements from retrograde exchange and net

307
 JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 FEBRUARY 2012

transfer reactions. As a check, the garnets were subse- CP-43. The inclusion-rich cores have identical REE pat-
quently re-analyzed by electron microprobe at spots adja- terns with the same flat HREE profiles (Figs 9c, 10c, 11c
cent to the LA-ICPMS craters to monitor MnO as an and 12c) as the peritectic garnet in the garnetites.
indicator of any retrograde net transfer reactions (Kohn, However, the inclusion-poor mantles show a progressive
2003); any LA-ICPMS analyses in locations with anomal- decrease in HREE concentrations, with the rim analyses
ous MnO were not included in the database. showing the lowest HREE values, mimicking the compos-
itional profile of garnet 1 in CP-43.
Figure 13 illustrates chondrite-normalized Gd/Lu ratios
R E S U LT S
vs Sc and Zr concentrations. The (Gd/Lu)N values show
X-ray dot maps and electron microprobe the relative slope of the REE patterns, and are used in
analyses this and subsequent figures as a differentiation index.

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

The Ca Ka X-ray maps (Fig. 4) reveal that the peritectic Figure 13a shows that the peritectic garnet in the garnetites
garnet in the garnetite contains abundant apatite inclu- has high Sc concentrations compared with the majority of
sions (as well as zircon and monazite, not shown). Three the inclusion-poor mantles and garnet 1 of CP-43 in the
garnet grains (CP-27 garnet 1 and 2, CP-4F garnet 1) plutonic host-rocks. Whereas the analyses of the
show circular patterns for the apatite inclusions, suggestive inclusion-rich cores of the host-rock garnet have similar
of growth zones. Three of the five studied garnet grains in Sc and chondrite-normalized (Gd/Lu)N ratios to the peri-
the host-rocks show similar inclusion patterns. Samples tectic garnet, the inclusion-poor mantles and garnet 1 of
CP-42 garnet 1 and 2, and K-1 garnet 4 all show abundant CP-43 define overall rimward trends to lower Sc abun-
apatite inclusions in the interior portions of the grains, dances with the steepening of the HREE patterns. The
but are mantled by inclusion-free zones. Garnet 2 of core of garnet 1 (CP-43) has similar Sc concentrations to
CP-43 also contains apatite inclusions, but it has a larger, the peritectic garnet, but the rim has the lowest Sc content
inclusion-free mantle. The analytical traverses for all these of any garnet measured in this study.
grains extended from the inclusion-rich cores though the Similar trends are shown by Zr vs (Gd/Lu)N (Fig. 13b).
inclusion-free mantles. Garnet 1 from CP-43 is distinctly The inclusion-rich cores in the host-rock garnet grains
different from the other host-rock garnet crystals. It is
and the peritectic garnet in the garnetites are similar,
inclusion-poor with only one apatite inclusion in the inter-
showing considerable overlap, and range from 37 to
ior of the grain (Fig. 4h). In that sense, it is similar to the
7 ppm Zr at low (Gd/Lu)N values. The inclusion-poor por-
inclusion-free mantles of the other host-rock garnet grains.
tions of the host-rock garnet have low Zr concentrations,
Mg Ka X-ray maps for three peritectic garnet grains in
generally less than 15 ppm. The inclusion-poor garnet 1 of
the garnetite and five host-rock garnet grains are presented
CP-43 has a wide range in compositions with core values
in Figs 5a^12a with corresponding electron microprobe
approaching that of the peritectic garnet. (Gd/Lu)N for
traverses in Figs 5b^12b. Each garnet, regardless of
the core is as low as 1^2, with corresponding Zr as high as
whether it came from the garnetites or the host plutonic
25 ppm.
rocks, shows flat almandine, pyrope, grossular and spes-
In contrast to the depletion trends of Sc and Zr, the con-
sartine profiles across the majority of the grains. Only ad-
jacent to biotite inclusions and at the rims do the grains centrations of Zn, P2O5, and U increase towards the rims
show increases in almandine and spessartine with corres- as the HREE patterns steepen (Fig. 14). Peritectic garnet
ponding decreases in pyrope. in the garnetites shows low concentrations in these elem-
ents and there is considerable overlap with the
LA-ICPMS analyses inclusion-rich core analyses of the host-rock garnet grains.
The locations of LA-ICPMS analyses along the garnet tra- The inclusion-free mantles and rims of these grains are
verse are shown by the white dots in Figs 5a^12a. richer in Zn, P2O5, and U. The core of garnet 1, CP-43, is
Chondrite-normalized REE patterns for these analyses again similar to the peritectic garnet, having slightly
are plotted in Figs 5c^12c. Each analysis shows a steep higher (Gd/Lu)N, but from the core to the rim, the concen-
LREE pattern, increasing from Ce to Sm with a prominent tration of these elements increases in the same manner
negative Eu anomaly. The peritectic garnet grains from as the inclusion-poor mantles in the host-rock garnet.
the garnetites have fairly flat patterns from Gd to Lu with Figure 15 illustrates the size of the Eu anomaly (Eu/Eu*)
only a few rim analyses having slightly declining patterns versus the slope of the HREE patterns. The peritectic
(Figs 5c, 6c and 7c). Garnet 1 of CP-43 is distinctly differ- garnet in the garnetites exhibits a range in Eu/Eu*, but
ent. Not only does this grain lack abundant inclusions, its on average has smaller negative Eu anomalies compared
chondrite-normalized HREE patterns show a strong, rim- with the host-rock garnet. These garnet grains show a gen-
ward depletion in HREE (Fig. 8c). Other garnet grains in eral trend of increasing Eu/Eu* towards the rims, with
the host-rocks show characteristics of both the peritectic the inclusion-free mantles and garnet 1 of CP-43 having
garnet in the garnetites and the inclusion-poor garnet 1 of the largest Eu/Eu* values.

308
 DORAIS & TUBRETT PERITECTIC GARNET, CARDIGAN PLUTON

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 4. Ca Ka X-ray maps for garnets discussed in the text. Bright grains are apatite inclusions in garnet. The peritectic garnets (a, b and c)
from garnetites have inclusions that define circular patterns. Likewise, the garnets in the host-rock also have cores with abundant apatite inclu-
sions (d, e, f and g) that are mantled by inclusion-poor regions. Garnet 1 in CP-43 (h) is inclusion poor throughout the entire grain.

309
 JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 FEBRUARY 2012

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 5. (a) Mg Ka X-ray dot map for garnet 1 in sample CP-27. The white dots represent locations of LA-ICPMS analysis in this and subsequent
figures. (b) Electron microprobe results plotting almandine, pyrope, grossular and spessartine values. The bulk of the grain is homogeneous
with the rims showing retrograde compositions. (c) Chondrite-normalized REE patterns. (Note the flat patterns from Gd to Lu.)

310
 DORAIS & TUBRETT PERITECTIC GARNET, CARDIGAN PLUTON

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 6. (a) Mg Ka X-ray dot map for garnet 2 in sample CP-27. (b) Electron microprobe results plotting almandine, pyrope, grossular and
spessartine values. The bulk of the grain is homogeneous with the rims showing retrograde compositions. (c) Chondrite-normalized REE pat-
terns. Note the flat patterns from Gd to Lu.

311
 JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 FEBRUARY 2012

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 7. (a) Mg Ka X-ray dot map for garnet 1 in sample CP-4F. (b) Electron microprobe results plotting almandine, pyrope, grossular and spes-
sartine values. The bulk of the grain is homogeneous with the rims showing retrograde compositions and portions adjacent to biotite inclusions
showing retrograde exchange compositions. (c) Chondrite-normalized REE patterns. (Note the flat patterns from Gd to Lu.)

312
 DORAIS & TUBRETT PERITECTIC GARNET, CARDIGAN PLUTON

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 8. (a) Mg Ka X-ray dot map for garnet 1 in sample CP-43. (Note the inclusion-poor nature of this grain.) (b) Electron microprobe results
for a half-grain traverse plotting almandine, pyrope, grossular and spessartine values. The bulk of the grain is homogeneous with the rims show-
ing retrograde compositions (left portion of traverse). (c) Chondrite-normalized REE patterns. (Note the strong depletion in HREE for all
the analyses.) HREE concentrations decrease from core (C) to rim (R).

313
 JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 FEBRUARY 2012

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 9. (a) Mg Ka X-ray dot map for garnet 2 in sample CP-43. (b) Electron microprobe results for a half-grain traverse plotting almandine,
pyrope, grossular and spessartine values. The bulk of the grain is homogeneous with the rims showing retrograde compositions (right portion
of traverse). (c) Chondrite-normalized REE patterns. Analyses from the interior, inclusion-rich portion of the grain have flat patterns from
Gd to Lu. The grain shows core (C) to rim (R) depletion in HREE.

314
 DORAIS & TUBRETT PERITECTIC GARNET, CARDIGAN PLUTON

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 10. (a) Mg Ka X-ray dot map for garnet 1 in sample CP-42. (b) Electron microprobe results plotting almandine, pyrope, grossular and
spessartine values. The bulk of the grain is homogeneous with the rims showing retrograde compositions. (c) Chondrite-normalized REE
patterns. Analyses from the interior, inclusion-rich portion of the grain have flat patterns from Gd to Lu. The grain shows core (C) to rim (R)
depletion in HREE.

315
 JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 FEBRUARY 2012

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 11. (a) Mg Ka X-ray dot map for garnet 2 in sample CP-42. (b) Electron microprobe results plotting almandine, pyrope, grossular
and spessartine values. The bulk of the grain is homogeneous with the rims showing retrograde compositions. (c) Chondrite-normalized
REE patterns. Analyses from the interior portions of the grain have flat patterns from Gd to Lu. Mantle portions show rimward depletion
in HREE.

316
 DORAIS & TUBRETT PERITECTIC GARNET, CARDIGAN PLUTON

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 12. (a) Mg Ka X-ray dot map for garnet 4 in sample K-1. (b) Electron microprobe results plotting almandine, pyrope, grossular and spes-
sartine values. The bulk of the grain is homogeneous with the rims showing retrograde compositions. (c) Chondrite-normalized REE patterns.
Analyses from the interior, inclusion-rich portions of the grain have flat patterns from Gd to Lu. Mantle portions show rimward depletion
in HREE.

317
 JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 FEBRUARY 2012

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 13. Sc and Zr concentrations (ppm) vs (Gd/Lu)N. Black-filled circles represent analyses of peritectic garnet. Open circles represent ana-
lyses of inclusion-rich cores in host-rock garnet with their inclusion-free mantles plotted as light gray circles. The dark gray circles represent
inclusion-poor garnet 1 (CP-43) of the host-rock. The mantles and garnet 1 (CP-43) show a rimward decrease in Sc and Zr as the (Gd/Lu)N
value increases. Garnet 1 shows a decrease in Sc and Zr from its core (C) to the rim (R), with the core values approaching those of the peritectic
grains. Detection limits given in Table 2. (See text for discussion.)

profiles in phenocrystic garnets would require very

DISCUSSION long times to homogenize, of the order of tens of millions
Diffusion coefficients for selected trace of years. Hence we consider the trace element profiles
elements in garnet observed in the Cardigan Pluton garnet to represent
Before any interpretations of trace element zoning in the growth zoning profiles.
Cardigan garnets can be presented, a discussion of REE
diffusion rates in garnet is in order. Van Orman et al. Trace element zoning and inclusion
(2002) showed that the diffusion rates for Sm and Yb are relationships
two to three orders of magnitude slower than those of The presence of garnetites in the Cardigan Pluton allows
major elements. Several studies (Spear & Kohn, 1996; Van the textural and chemical characterization of the peritectic
Orman et al., 2002; Vielzeuf et al., 2005; Villaros et al., garnet. These characteristics serve as a basis for compari-
2009a; Lackey et al., 2011a) concluded that because the dif- son with the garnets in the plutonic host-rocks and are
fusion coefficients are also small for REE, Zr, Y, O, Ca, used to determine if the host-rock garnets are peritectic or
and other elements analyzed in this study, growth zoning phenocrystic.

318
 DORAIS & TUBRETT PERITECTIC GARNET, CARDIGAN PLUTON

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 14. Zn (ppm), P2O5 (wt %), and U (ppm) concentrations vs (Gd/Lu)N. Symbols as in Fig. 13. The inclusion-free mantles and garnet 1 of
CP-43 show rimward increases in Zn, P2O5, and U concentrations. Garnet 1 (CP-43) has core compositions that approach those of the peritectic
garnet. Detection limits given in Table 2.

319
 JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 FEBRUARY 2012

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

Fig. 15. Eu/Eu* vs (Gd/Lu)N. The mantles and garnet 1 (CP-43) show an increase in the size of the Eu anomaly with steepening of the HREE
pattern. Symbols as in Fig. 13.

As discussed above, the HREE concentrations in the melting should have formed from liquids that were
peritectic garnet in the garnetites are high and define flat HREE-rich and hence should have high HREE concentra-
HREE chondrite-normalized patterns (Figs 5c^7c). tions. This is precisely what is shown in Figs 5c^7c. The
Provided that the peritectic garnet, dissolving accessory relatively flat chondrite-normalized REE patterns indicate
phases, and the melt were in equilibrium during partial that the peritectic garnet in the garnetites grew from li-
melting (discussed below), then dissolution of accessory quids that maintained high HREE concentrations; no rim-
minerals should have imparted high concentrations of ward depletion of the HREE is present in these garnet
elements that were compatible in the accessory phases to grains, suggesting that there was no fractionation of
the melt. If these elements were also compatible in garnet, HREE in the liquid.
then the peritectic garnet should have formed from melts The inclusion-rich cores in the host-rock garnets not only
that are at their highest concentration of these compatible show the same type of inclusions and circular inclusion
elements; that is, equilibrium with accessory phases patterns as the peritectic garnets, but they also have the
should buffer the concentrations of these elements in the same HREE concentrations (Figs 9c^11c). We interpret
melt. For example, it has been well documented that Ycon- these inclusion-rich cores to be peritectic grains that were
centrations in garnet are strongly affected by the presence dispersed in the Cardigan magma. Subsequent fraction-
of xenotime (Spear & Kohn, 1996; Pyle & Spear, 1999, ation of zircon, as well as garnet itself, would have
2000). Xenotime dissolution liberates Y, producing high Y depleted the residual liquid in HREE, producing the
concentrations in garnet because garnet has high partition observed growth zoning of the mantles and rims of the
coefficients for Y. Once xenotime is consumed, host-rock garnet (Figs 8c^12c). Hence we interpret
Y concentrations in subsequently grown garnet diminish the inclusion-poor mantles and rims of these host-rock
considerably, producing step-like profiles in Y zoning. No peritectic garnet grains to be phenocrystic. The entire
xenotime has been observed in the Cardigan garnetites, garnet 1 in sample CP-43 is inclusion-poor (Fig. 4h). The
but similar partitioning between compatible elements REE patterns of this garnet are depleted in the HREE
from dissolving accessory phases and garnet is expected. (Fig. 8c) with decreasing HREE concentrations towards
The abundance of inherited zircon in the garnetites the rim. It appears that the entire grain grew under the in-
(Dorais et al., 2009a) indicates that the melt was saturated fluence of a fractionating liquid. Hence this grain is a
in zircon and unable to dissolve all the zircon in the phenocryst.
source rock. Because the HREE are highly compatible in The peritectic garnet in the garnetites also has the high-
both zircon and garnet (Watson, 1980; Fujimake, 1986; est concentrations of trace elements that are compatible in
Sisson & Bacon, 1992; Thomas et al., 2002), peritectic both biotite and garnet. Any element liberated by biotite
garnet growing in an environment where zircon was breakdown should have relatively high concentrations in

320
 DORAIS & TUBRETT PERITECTIC GARNET, CARDIGAN PLUTON

the melt and be available for partitioning into peritectic approaching those of the peritectic garnets. The core
garnet. The Sc biotite-rhyolitic melt partition coefficient appears to have crystallized from a liquid that was very
ranges from 4·9 to 20 (Nash & Crecraft, 1985). Biotite similar in composition to the anatectic melt, prior to sig-
breakdown should supply Sc for incorporation in peritectic nificant fractionation. The core could represent an early
garnet, especially as the Sc partition coefficient between phenocryst that essentially was in equilibrium with the
garnet and rhyolitic melt ranges from 13·6 to 63 (Sisson & restite.
Bacon, 1992). Peritectic garnet should have crystallized at
the melt’s maximum Sc concentrations; that is, before phe- Equilibrium among peritectic garnet,
nocrystic biotite and garnet crystallization diminished the accessory phases and melt
abundance of Sc in the fractionating melt. The Sc concen- Several studies have indicated that equilibrium conditions
trations of the Cardigan Pluton garnets are consistent between accessory minerals and anatectic melt should not

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

with this prediction (Fig. 13a). Peritectic garnet grains be assumed (Bea, 1996; Bea & Montero, 1999). In disequi-
have high Sc concentrations, ranging from 100 to 170 ppm. librium melting, peritectic garnet would have lower con-
Inclusion-rich cores of host-rock garnet overlap in Sc abun- centrations of elements that are shared between accessory
dances and the grains show rimward depletion in Sc in minerals and garnet (e.g. HREE) than those formed
the inclusion-poor mantles as the HREE patterns steepen, under equilibrium conditions when the melt is buffered.
preserving a record of magmatic fractionation in the man- Peritectic garnet in melanosomes of the Mikhondo Valley
tles. The lower Zr concentrations in the inclusion-poor, Metamorphic Suite of Swaziland provides evidence for
host-rock garnet, especially the decrease in Zr towards such disequilibrium melting (Taylor & Stevens, 2010).
the rims of these garnet grains (Fig. 13b), suggest that The HREE abundances in the peritectic garnet are con-
zircon was also crystallizing, which would, like garnet, siderably lower than those in garnet in the associated
contribute to the increasing (Gd/Lu)N seen from garnet granitic sheets, almost by an order on magnitude in
chondrite-normalized REE diagrams (Taylor & Stevens,
cores to rims.
2010, fig. 8). Apparently the accessory phases were not
Elements that were incompatible in peritectic garnet or
melting in their modal proportions; perhaps they were
other early crystallizing minerals should have increased in
shielded by minerals that were not melting, yielding low
the melt with fractionation. Thus peritectic garnet should
HREE concentrations to the anatectic melt.
have relatively low concentrations of these elements,
The peritectic garnet of the Cardigan Pluton is unlike
whereas garnet phenocrysts crystallizing from fractionated
the disequilibrium-influenced garnet of the Mikhondo
liquids should show corresponding enrichment in these
Valley Metamorphic Suite in that the Cardigan peritectic
elements. Figure 14 illustrates the low abundances of Zn, P
garnet is the most HREE-rich of our analyzed grains. We
and U in the peritetic garnet of the garnetites. The
infer that a buffered melt led to their high HREE values.
inclusion-rich portions of grains in several host-rock gar-
nets are identical to the peritectic garnet, again suggesting Garnet dissolution^reprecipitation
that the cores are also peritectic. However, the mantles of A complication in determining the phenocrystic or peritec-
these grains show rimward enrichments in these elements tic origin of garnet has been presented by Villaros et al.
with progressive depletion in HREE throughout the (2009a) and Taylor & Stevens (2010). They suggested that
inclusion-poor mantles, the same zoning as in grain 1 in garnet of any origin in peraluminous plutons, whether phe-
CP-43. All these mantles and the entire grain in CP-43 nocrystic, peritectic or xenocrystic, undergoes dissolution
are phenocrystic. and reprecipitation that keeps the garnet in constant equi-
Because feldspar would have been an early crystallizing librium with the changing magma compositions. Those
phase in peraluminous melts, Eu depletion should have workers proposed that a 1cm diameter garnet can be dis-
occurred in the residual melts. Hence the size of the Eu solved within a timescale of days, and concluded that the
anomaly in the phenocrystic garnet should increase as debate on the origin of garnet in peraluminous granites is
garnet crystallized from progressively more fractionated si- meaningless because dissolution^reprecipitation complete-
liceous magmas. This is portrayed in Fig. 15, where the ly masks the inheritance of garnet in such magmas.
size of the negative Eu anomaly increases from peritectic The combination of the same types of inclusions and
garnet to the rims of phenocrystic garnet. The negative the identical trace element abundances in the cores of
Eu anomalies of the peritectic garnet could reflect plagio- host-rock garnets to those of the peritectic garnets in the
clase in the restite, but we infer that the deepening of the garnetites allows recognition of peritectic garnet in the
anomaly in the phenocrysts is a reflection of feldspar frac- host pluton. Because these garnet cores have the same
tionation in an evolving melt and is indicative of pheno- type of inclusions and circular inclusion patterns as the
crystic versus peritectic garnet. garnets in the garnetites, we infer that these garnet cores
Garnet 1 of sample CP-43 shows a wide range in were not subject to dissolution^reprecipitation; otherwise,
(Gd/Lu)N and other trace elements with the core values they should have been purged of inclusions, as well as

321
 JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 FEBRUARY 2012

having different trace element abundances from the peri- step in our study of the Cardigan Pluton is to conduct an
tectic garnet. But what of the inclusion-poor mantles that extensive petrographic and analytical survey of garnet.
crystallized on these peritectic cores? They could be por- Determination of the amount of peritectic garnet will
tions of peritectic grains that dissolved and reprecipitated, permit evaluation of the cause of high normative corun-
thereby liberating inclusions and producing different trace dum in the more mafic rocks (Fig. 1). If this high-alumina
element profiles. phase is peritectic rather than phenocrystic garnet, then
It seems reasonable that if dissolution^reprecipitation the selective peritectic mineral entrainment model is a
could remake a 1cm diameter garnet in a matter of days valid explanation for the compositional variation of the
(Villaros et al., 2009a), then the entire garnet should main- Cardigan Pluton.
tain equilibrium with the magma, producing homoge- Whether this same approach will be fruitful for other
neous major and trace element abundances. If garnet peraluminous plutons remains to be determined. The

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

dissolves and reprecipitates faster than HREE can diffuse Cardigan Pluton may be an exceptional case because the
into the crystal, then the crystal could be zoned, but this fortuitous presence of garnetites provides a basis for deter-
is essentially due to fractional crystallization. We propose mining the microstructures and compositions of peritectic
that the trace element zoning in the mantles of the garnet; garnetites are generally absent in other peralumin-
host-rock garnet and in garnet 1 of CP-43 is consistent ous plutons. Additionally, partial melting under disequilib-
with zoning produced by fractional crystallization. rium conditions may shield accessory phases from
Phenocrysts that have not dissolved and reprecipitated melting, giving rise to nonsystematic trace element concen-
should be homogeneous in major element profiles because trations in peritectic garnet in melts that are not buffered.
of relatively rapid major element diffusion at magmatic However, we anticipate that phenocrystic garnet will
temperatures; however, trace elements with small diffusion show trace element zoning profiles that are suggestive of
coefficients should preserve growth zoning profiles as has differentiation and that are different from those of peritec-
been observed in this and other studies (Spear & Kohn, tic garnet, regardless of whether that garnet formed under
1996; Vielzeuf et al. 2005; Dutch & Hand, 2009). We there- equilibrium melting conditions or not.
fore conclude that garnet grains with trace element profiles
that are consistent with crystal fractionation are pheno-
crysts, a conclusion shared by Spear & Kohn (1996),
CONC LUSIONS
Vielzeuf et al. (2005) and Dutch & Hand (2009).
A recent study by Lackey et al. (2011b) demonstrated that Peritectic garnet of the Cardigan Pluton is rich in elements
garnet samples from the Hallowell and Togus plutons of that were compatible in both garnet and the melting
Maine contain garnet grains that are zoned in d18O. Some phases. HREE contents are at the maximum observed
hand samples contain several populations of garnet values of any analyzed garnet in this study owing to
grains, each with distinct d18O zoning profiles. The preser- zircon melting contributing large supplies of HREE to the
vation of these profiles also suggests that garnet did not dis- anatectic melt. Subsequent fractionation of both garnet
solve and reprecipitate in these plutons. and zircon depleted the residual melt in HREE, a deple-
tion recorded in rimward traverses of phenocrystic garnet
Selective entrainment of peritectic garnet grains. Likewise, Sc concentrations are at maximum
in peraluminous plutons values in peritectic garnet owing to the high supply of Sc
Stevens et al. (2007) determined that the Cape Granite from biotite dehydration melting. Subsequent fractionation
Suite of South Africa has major element trends that are of garnet and biotite depleted the melt in Sc, the record of
controlled by variable amounts of garnet retention. They which is preserved in phenocrystic garnet zoning.
proposed that the garnet is peritectic rather than pheno- Incompatible elements were enriched in the fractionated
crystic because accumulation of phenocrystic garnet in melts and hence phenocrystic garnet shows rimward en-
the more mafic portions of the pluton would have required richments in these elements. Garnet that crystallized from
significant amounts of fractionation; no corresponding vol- fractionated melts produced crystals that were no longer
umes of very silicic rocks are present that would counter- in trace element equilibrium with restite, instead forming
balance the large portions of relatively mafic rocks in the zoned phenocrysts.
pluton. They therefore concluded that the selective entrain-
ment of peritectic garnet controls the compositional vari-
ation of the pluton. LA-ICPMS studies of garnet can AC K N O W L E D G E M E N T S
complement the bulk-rock approach of Stevens et al. We are grateful to Mike Williams, Becky Jamieson and
(2007). The distinction of peritectic from phenocrystic John Schumacher for their constructive reviews of this
garnet in the Cardigan Pluton opens the possibility of paper, and to Ron Frost for his valuable editorial as-
determining the modal proportions of peritectic garnet sistance. M.J.D. thanks Carol Dorais for assistance in
across the compositional spectrum of a pluton. The next the field.

322
 DORAIS & TUBRETT PERITECTIC GARNET, CARDIGAN PLUTON

Kohn, M. J. (2003). Geochemical zoning in metamorphic minerals.

FUNDING In: Rudnick, R.L. (ed.) Treatise on Geochemistry, Vol. 3. Amsterdam:
This research was supported in part by funds from the Elsevier, pp. 229^261.
College of Physical and Mathematical Sciences, Brigham Kotkova¤, J. & Harley, S. L. (2010). High-pressure crustal metamorph-
Young University. ism: Evidence from garnet^whole-rock REE relationships and
zircon^rutile Ti^Zr thermometry in leucogranulites from the
Bohemian Massif. Journal of Petrology 51, 1967^2001.
Lackey, J. S., Erdmann, S., Hark, J. S., Nowak, R. M., Murray, K. E.,
S U P P L E M E N TA RY DATA Clarke, D. B. & Valley, J. W. (2011a). Tracing garnet origins in
Supplementary data for this paper are available at Journal granitoid rocks by oxygen isotope analysis: examples from the
South Mountain batholiths, Nova Scotia. Canadian Mineralogist 49,
of Petrology online.
767^789.
Lackey, J. S., Romero, G. A., Bouvier, A.-S. & Valley, J. W.

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

(2011b). Dynamic growth of garnet in granitic magmas. Geology
R EF ER ENC ES (in press).
Longerich, H. P., Jackson, S. E. & Gunther, D. (1996). Laser ablation
Bea, F. (1996). Residence of REE, Y, Th, and U in granites and crustal
inductively coupled plasma mass spectrometric transient signal
protoliths: implications for the chemistry of crustal melts. Journal data acquisition and analyte concentration calculation. Journal of
of Petrology 37, 521^552. Analytical Atomic Spectrometry 11, 899^904.
Bea, F. & Montero, P. (1999). A LA-ICP-MS evaluation of Zr reser- Lyons, J. B., Bothner, W. A., Moench, R. H. & Thompson, J. B. (1997).
voirs in common crustal rocks: implications for Zr and Hf geo- Bedrock Geologic Map of New Hampshire. Denver, CO: US
chemistry, and zircon-forming processes. Canadian Mineralogist 44, Geological Survey.
693^714. Nash, W. P. & Crecraft, H. R. (1985). Partition coefficients for trace
Bradley, D. C. & Hanson, L. S. (2002). Paleocurrent analysis of a de- elements in silicic magmas. Geochimica et Cosmochimica Acta 49,
formed Devonian foreland basin in the northern Appalachians, 2309^2322.
Maine, USA. Sedimentary Geology 148, 425^447. Nielson, D. L., Clark, R. G., Lyons, J. B., Englund, E. J. & Borns, D.
Chappell, B. W. (1996). Magma mixing and the production of compos- L. (1976). Gravity models and mode of emplacement of the
itional variation within granite suites: Evidence from the granites New Hampshire plutonic series. In: Lyons, J. B. & Brownlow, A.
of southeastern Australia. Journal of Petrology 37, 449^470. H. (eds) Studies in New England Geology. Geological Society of America,
Clark, R. G. & Lyons, J. B. (1986). Petrogenesis of the Kinsman Memoirs 146, 301^318.
Intrusive Suite: peraluminous granitoids of western New Plank, T. (1987). Magmatic garnets from the Cardigan pluton and the
Hampshire. Journal of Petrology 27, 1365^1393. Acadian thermal event in southwest New Hampshire. American
Clemens, J. D. (2003). S-type granitic magmasçpetrogenetic issues, Mineralogist 72, 681^688.
models and evidence. Earth-Science Reviews 61, 1^18. Pyle, J. M. & Spear, F. S. (1999). Yttrium zoning in garnet: Coupling
Conant, L. C. (1935). The New Hampshire garnet deposits. Economic of major and accessory phases during metamorphic reactions.
Geology 30, 387^399. Geological Materials Research 1, 1^49.
Cullers, R. L., Bock, B. & Guidotti, C. (1997). Sedimentary geochem- Pyle, J. M. & Spear, F. S. (2000). An empirical garnet (YAG)^xenotime
istry: environments and processes dominated by involvement of an thermometer. Contributions to Mineralogy and Petrology 138, 51^58.
aqueous phase. Philosophical Transactions of the Royal Society of London, Sisson, T. W. & Bacon, C. R. (1992). Garnet high-silica rhyolite
Series A 286, 353^372. trace-element partition-coefficients measured by ion microprobe.
Dorais, M. J. (2003). The petrogenesis and emplacement of the New Geochimica et Cosmochimica Acta 56(5), 2133^2136.
Hampshire plutonic suite. AmericanJournal of Science 303, 447^487. Spear, F. S. & Kohn, M. J. (1996). Trace element zoning in garnet as a
Dorais, M. J., Pett, T. K. & Tubrett, M. (2009a). Garnetites of the monitor of crustal melting. Geology 24, 1099^1102.
Cardigan Pluton, New Hampshire: Evidence for peritectic garnet Stevens, G., Villaros, A. & Moyen, J.-F. (2007). Selective peritectic
and implications for source rock compositions. Journal of Petrology garnet entrainment as the origin of geochemical diversity in
50, 1993^2016. S-type granites. Geology 35, 9^12.
Dorais, M. J., Wintsch, R. P., Nelson, W. R. & Tubrett, M. (2009b). Stewart, D. B., Wright, B. E., Unger, J. D., Phillips, J. D.,
Insights into the Acadian Orogeny, New England Appalachians: a Hutchinson, D. R., Luetgert, J. H., Bothner, W. A., Klitgord, K.
provenance study of the Carrabassett and Kittery formations, D., Liberty, L. M. & Spencer, C. P. (1993). Global Geoscience
Maine. Atlantic Geology 45, 50^71. Transect 8; Quebec^Maine^Gulf of Maine Transect, southeastern
Dutch, R. & Hand, M. (2009). Retention of Sm^Nd isotopic ages in Canada, northeastern United States of America. Investigations
garnets subjected to high-grade thermal reworking: Implications Series, US Geological Survey, Report I-2329, 17p.
for diffusion rates of major and rare earth elements and the Thomas, J. B., Bodnar, R. J., Shimizu, N. & Sinha, A. K. (2002).
Sm^Nd closure temperature in garnet. Contributions to Mineralogy Determination of zircon/melt trace element partition coefficients
and Petrology 159, 93^112. from SIMS analysis of melt inclusions in zircon. Geochimica et
Eusden, J. D., Jr & Lyons, J. B. (1993). The sequence of Acadian Cosmochimica Acta 66, 2887^2901.
deformations in central New Hampshire. In: Roy, D. C. & Thompson, J. B., Bothner, W. A., Robinson, P., Isachsen, Y. M. &
Skehan, J. W. (eds) The Acadian Orogeny: Recent Studies in New Klitgord, K.D. (1993). A guide to continental^ocean transect E-1.
England, Maritime Canada, and the Autochthonous Foreland. Adirondacks to Georges Bank. Centennial continent^ocean transect #17.
Geological Society of America, Special Papers 275, 51^66. Boulder, CO: Geological Society of America, 55 p.
Fujimake, H. (1986). Partition coefficients of Hf, Zr, and REE between Taylor, J. & Stevens, G. (2010). Selective entrainment of peritectic
zircon, apatite, and liquid. Contributions to Mineralogy and Petrology garnet into S-type granitic magmas: Evidence from Archaean
94, 42^445. mid-crustal anatectites. Lithos 120, 277^292.

323
 JOURNAL OF PETROLOGY VOLUME 53 NUMBER 2 FEBRUARY 2012

Van Orman, J. A., Grove, T. L., Shimizu, N. & Layne, G. D. (2002). disequilibrium melting and accessory phase entrainment in the
Rare earth element diffusion in a natural pyrope single source. Contributions to Mineralogy and Petrology 158, 543^561.
crystal at 2·8 GPa. Contributions to Mineralogy and Petrology 142, Watson, E. B. (1980). Some experimentally determined zircon/liquid
416^424. partition coefficients for the rare earth elements. Geochimica et
Vielzeuf, D., Veschambre, M. & Brunet, F. (2005). Oxygen isotope Cosmochimica Acta 44, 895^897.
heterogeneities and diffusion profile in composite metamorphic^ White, A. J. R. & Chappell, B. W. (1977). Ultrametamorphism and
magmatic garnets from the Pyrenees. American Mineralogist 90, granitoid genesis. Tectonophysics 43, 7^22.
463^472. White, A. J. R. & Chappell, B. W. (1988). Some supracrustal (S-type)
Villaros, A., Stevens, G., Moyen, J. F. & Buick, I. S. (2009a). Tracking grantites of the Lachlan fold belt. Transactions of the Royal Society of
S-type granite from source to emplacement: clues from garnet in Edinburgh: Earth Sciences 79, 169^181.
the Cape Granite Suite. Lithos 112, 217^235.
Villaros, A., Stevens, G., Moyen, J. F. & Buick, I. S. (2009b). The trace
element compositions of S-type granites: evidence for

Downloaded from https://academic.oup.com/petrology/article/53/2/299/1535894 by guest on 08 December 2020

324