Nepheline Syenite Pegmatites in Langesundsfjord

JOURNAL OF PETROLOGY VOLUME 51 NUMBER 11 PAGES 2303^2325 2010 doi:10.
1093/petrology/egq058
Petrology of Nepheline Syenite Pegmatites in the

Oslo Rift, Norway: Zirconium Silicate Mineral
Assemblages as Indicators of Alkalinity and
Volatile Fugacity in Mildly Agpaitic Magma
Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

T. ANDERSEN1*, M. ERAMBERT1, A. O. LARSEN2 AND
R. S. SELBEKK3
1
DEPARTMENT OF GEOSCIENCES, UNIVERSITY OF OSLO, PO BOX 1047 BLINDERN, N-0316 OSLO, NORWAY
2
STATOIL ASA, HYDROVEIEN 67, N-3908 PORSGRUNN, NORWAY
3
NATURAL HISTORY MUSEUM, UNIVERSITY OF OSLO, SARS GATE 1, N-0562 OSLO, NORWAY
RECEIVED APRIL 7, 2010; ACCEPTED SEPTEMBER 7, 2010
Agpaitic nepheline syenites contain complex zirconium silicate min- eudialyte-dominated liquidus assemblages (and possibly further to
erals as primary magmatic phases; agpaitic liquidus mineral assem- hyperagpaitic residual liquids); (2) the increasing water trend, lead-
blages are controlled by compositional parameters such as alkalinity ing to catapleiite crystallization; (3) the increasing fluorine trend,
and concentrations of water and halogens in the magma. The giving transitionally agpaitic liquidus assemblages with wo« hlerite,
Larvik Plutonic Complex in the late Palaeozoic Oslo Rift, Norway lvenite and hiortdahlite, with or without zircon, but eventually
consists of hypersolvus monzonite (larvikite) and different varieties with magmatic fluorite; this type of evolution is applicable to
of nepheline syenite. The nepheline syenite members have miaskitic weakly agpaitic systems such as the Oslo Rift pegmatites.The evolu-
mineralogy (i.e. zircon stable), except for a suite of late, mildly tion of such magmas will most probably terminate at liquidus bound-
agpaitic pegmatites, which contain zirconium silicate minerals of aries also involving eudialyte and/or rosenbuschite, and they are
the wo« hlerite, rosenbuschite and eudialyte groups, in some cases to- unlikely to evolve to more strongly agpaitic compositions.
gether with zircon.These minerals form part of the liquidus mineral
assemblage of the nepheline syenite magma together with fluorite
and the rock-forming minerals alkali feldspar, aegirine(âugite), KEY WORDS: agpaitic rocks; Oslo Rift; nepheline syenite; zirconium
alkali amphibole, biotite, nepheline and sodalite. Low-variance min- silicate minerals
eral assemblages with two or three zirconium silicate minerals, fluor-
ite and the major rock-forming minerals can be modelled in terms of
phase equilibria in a multi-component system involving the sodium I N T RO D U C T I O N
disilicate component of the melt and the volatile components water, Zircon is an abundant accessory mineral, and in most
fluorine and chlorine. A semiquantitative isothermal and isobaric igneous rocks it is the main host mineral for zirconium.
petrogenetic grid in log activity space constructed from the observed Factors that control the crystallization of zircon from a sili-
mineral assemblages and the compositions of the coexisting minerals cate melt include the concentration of zirconium, the
indicate that there are at least three trends of evolution leading from major element composition and volatile content of the
a miaskitic to an agpaitic crystallization regime in peralkaline neph- magma, as well as temperature (e.g. Watson & Harrison,
eline syenites: (1) the increasing alkalis trend, typically leading to 1983; Baker et al., 2002). However, in some highly evolved
The Author 2010. Published by Oxford University Press. All

rights reserved. For Permissions, please e-mail: journals.permissions@
*Corresponding author. E-mail: tom.andersen@geo.uio.no oxfordjournals.org
JOURNAL OF PETROLOGY VOLUME 51 NUMBER 11 NOVEMBER 2010
Table 1: Mineral formulas for Zr-silicates and other (e.g. Motzfeldt and Gardiner complexes, Greenland;
McGerrigle and Mont Saint-Hilaire, Canada; Oslo Rift,
important minerals from nepheline syenite pegmatites in Norway; Srensen, 1997, and references therein), and
the Oslo Rift agpaitic mineral assemblages have formed by late- to
post-magmatic, fluid-induced processes in miaskitic neph-
Eudialyte-group eline syenite in Tamazeght, Morocco (Schilling et al., 2009,
Eudialyte Na15Ca6(Fe2þ,Mn2þ)3Zr3(Si25O73)(O,OH,H2O)3 and references therein).
(OH,Cl)2 The factors controlling the transition from a miaskitic
Ferrokentbrooksite Na15Ca6(Fe,Mn)3Zr3NbSi25O73(O,OH,H2O)3Cl2 (i.e. zircon present) to an agpaitic liquidus mineralogy

Zirsilite-(Ce) (Na11–12)(Ce,Na)3Ca6Mn3Zr3Nb(Si3O9)2(Si9O27)2 with complex zirconium silicate minerals in a peralkaline
(SiO)(OH)3CO3·H2O
nepheline syenite magma is an important problem in the
Wöhlerite group
petrology of alkaline rocks. An elevated sodium concentra-
Wöhlerite Na2Ca4ZrNb(Si2O7)2O3F
tion in the magma is clearly needed to stabilize minerals
Hiortdahlite Na2Ca4Zr(Nb,Ti)(Si2O7)2(F,O)4
such as sodium catapleiite (Na2ZrSi3O9·2H2O) in nepheline
Lafi venite Na2(Na,Ca)2(Fe2þ,Mn,Ti)2(Nb,Zr)2(Si2O7)2(O,F)4
syenite or elpidite (Na2ZrSi6O15·3H2O) in alkali granite,
and strong enrichment of alkali elements and other incom-
Rosenbuschite group1
patible components by extensive fractional crystallization
Rosenbuschite (Ca,Na)3(Zr,Ti)Si2O7FO
of an alkaline parent magma is generally thought to be ne-
Hainite Na2Ca5(Ti,Zr,Mn,œ)(Si2O7)2F2(OH)2
cessary for the formation of agpaitic magmas (e.g. Larsen
Kochite (Na,Ca)3Ca2(Mn,Ca)ZrTi(Si2O7)2(F,O)4
& Srensen, 1987). However, the zirconium silicate min-
Grenmarite (Zr,Mn)2(Zr,Ti)(Mn,Na)(Na,Ca)4(Si2O7)2(O,F)4
erals listed in Table 1 also contain essential water, fluorine
Rinkite group
or chlorine, and the activities of volatile species in
Mosandrite2 Ti(œ,Ca,Na)3Ca4(Si2O7)2[H2O,OH,F]4 · 1H2O
the magma will influence the mineralogy of zirconium-
Rinkite (‘Johnstrupite’3) Ti(Na,Ca)3(Ca,Ce)4(Si2O7)OF3
bearing phases in such rocks (e.g. Kogarko, 1974;
Catapleiite Na2ZrSi3O9 · 2(H2O)
Srensen, 1997). The effect of halogens to stabilize complex
Tritomite-Ce (Ce,Ca,La,)5(Si,B)3(O,OH)12(OH)
Na^Ca^Zr silicates was, in fact, pointed out already by
Pyrochlore (Na,Ca)2Nb2(O,OH,F)7
Brgger (1890) in his classical study of the minerals of
nepheline syenite pegmatites in the Oslo Rift, Norway.
Formulas after Strunz & Nickel (2001) with additional
data from the www.mindat.org database (http://www One convenient way to evaluate the relative effects of
.mindat.org/strunz.php). alkalinity and volatile activities for the transition from
œ: Vacancy. miaskitic to agpaitic conditions in nepheline syenite
1
Christiansen et al. (2003). 2Bellezza et al. (2009). 3Brøgger magmas is to examine zirconium silicate mineral assem-
(1890).
blages in groups of cogenetic rocks with both miaskitic
and agpaitic members. The nepheline syenite pegmatites
felsic rocks, zircon is absent despite a high concentration of in the Oslo Rift (Larsen, 2010) provide one such example.
zirconium. In such rocks, Zr is hosted by complex zirco- These pegmatites range from zircon-bearing to zircon-free
nium silicate minerals, some of which are listed in Table 1 types, and have provided the type specimens for cata-
(e.g. Srensen, 1974, 2006; Khomyakov, 1995; Salvi & pleiite, wo«hlerite, hiortdahlite, lvenite and other minerals.
Williams-Jones, 1995; Schilling et al., 2009). Based on stu- Based on the mineral assemblages of these pegmatites,
dies of nepheline syenite intrusions in the Gardar Rift in Semenov (in Srensen, 1974) defined the Langesundsfjord
Greenland, Ussing (1912) introduced the term agpaite for type as a separate class of nepheline syenite, intermediate
nepheline syenite with (Na þ K)/Al 1·2. In current petro- between the main miaskitic and truly agpaitic types, and
graphic terminology, agpaitic rocks are peralkaline Khomyakov (1995) assigned them to the low agpaitic
nepheline syenites with complex Na^Ca^Ti^Zr-silicate rocks. Despite nearly two centuries of active mineralogical
minerals (Le Maitre, 2003). In contrast, less alkaline neph- research (Larsen, 2010, and references therein), no studies
eline syenites in which zircon, titanite and ilmenite are have been made on the petrology of these pegmatites
characteristic minerals are called miaskitic, from a rock since that by Brgger (1890), and their coexisting zirco-
name originally coined for a non-peralkaline nepheline nium silicate mineral assemblages have not been properly
syenite (Le Maitre, 2003). Agpaitic nepheline syenites characterized.
are major constituents of large, highly alkaline intrusive Here we report the results of a systematic study of the
complexes such as the Il|¤ maussaq intrusion in southern zirconium silicate mineral assemblages and mineral
Greenland and the Khibina and Lovozero intrusions in compositions in selected nepheline syenite pegmatites
the Kola Peninsula, but they also occur as minor intrusive from the Larvik Plutonic Complex in the Oslo Rift. These
members in dominantly miaskitic alkaline rock complexes data provide the basis for a chemographic analysis in
2304
ANDERSEN et al. ZIRCONIUM SILICATE MINERAL ASSEMBLAGES

Fig. 1. (a) Simplified geological map of the Langesundsfjord archipelago and surrounding area, SW Oslo Rift, Norway. Localities: 1, Barkevik
area (including Barkevik, Barkevikskjr, Skutesundskjr, Kjeya); 2, Stokksund; 3, Langodden; 4, Aryskjrene; 5, Risya; 6, Trompetholmen;
7, Lven; 8, Vesle Arya. (b) Simplified map showing the 10 successive ring segments making up the Larvik Plutonic Complex, after Petersen
(1978). Location of the map area in (a) is indicated.
multi-component space, which is used to derive an empir- rocks range in composition from alkaline basalt to granite
ical, semiquantitative petrogenetic grid relating the zirco- and nepheline syenite, with a large proportion of inter-
nium silicate liquidus mineralogy to alkalinity and mediate, monzonitic intrusive rocks (larvikite) and com-
volatile activities. This, in turn, provides constraints for positionally equivalent latitic lavas known as rhomb
the relative importance of the parameters controlling the porphyries (Neumann et al., 2004, and references therein).
transition from a miaskitic to an agpaitic crystallization The Larvik Plutonic Complex (Fig. 1b) is a large, compos-
regime in peralkaline nepheline syenite magma. ite intrusion making up the southernmost onshore part of
the Oslo Rift (Larsen et al., 2008; Dahlgren, 2010). The
complex is built up of 10 arc-shaped or circular intrusions
GEOLOGIC A L S ET T I NG of hypersolvus monzonite (larvikite) and nepheline syenite
The Oslo Rift in southeastern Norway is part of a system (Fig. 1b; Petersen, 1978). U^Pb dating of magmatic zircon
of late Carboniferous to Permian continental rifts in north- suggests a relatively short period of emplacement and crys-
ern Europe (Neumann et al., 2004). Rift-related igneous tallization (299^292 Ma, Dahlgren, 2010), overlapping in
2305
time with the main rifting episode and widespread rhomb pegmatites. This stage is usually characterized by extensive
porphyry volcanism (Larsen et al., 2008). Trace element zeolitization and alteration of the magmatic minerals, and
distribution patterns and radiogenic isotopic signatures of by crystallization of low-temperature hydroxides and
larvikite and rhomb porphyry point towards an origin hydrous silicates. Many of the rare REE- and beryllium-
from the lithospheric mantle (Neumann et al., 2004, and bearing minerals belong to this stage of pegmatite forma-
references therein). The rocks of the Larvik Plutonic tion. A few minerals (e.g. goethite, hisingerite, neotocite)
Complex range from mildly quartz normative in the have formed as the result of supergene processes (Larsen,
early, eastern ring intrusions, to strongly nepheline norma- 2010). The present study is concerned only with the first,
tive in the younger intrusions in the west and magmatic stage of pegmatite crystallization. The host lar-

north (Petersen, 1978). This range in composition is explic- vikite crystallized at shallow-crustal pressures (P 1kbar),
able by polybaric fractionation of a mildly alkaline, and reached its solidus at 850^8608C (Neumann, 1976,
mafic parent magma in deep- to middle-crustal magma 1980), which defines the maximum pressure and tempera-
chambers, combined with a density filtering mechanism ture for emplacement of the pegmatites. From the compos-
(Neumann, 1980). itions of coexisting Fe^Ti oxides, Neumann (1976)
Syenite and nepheline syenite pegmatites are relatively determined a consistent T^fO2 cooling trend for the
common in the Larvik Plutonic Complex (Dahlgren, Larvik Plutonic Complex parallel to and less than one log
2010), and can be grouped in three main suites. fO2 unit below the quartz^fayalite^magnetite (QFM)
buffer curve; oxygen fugacity was buffered by equilibria
(1) Miaskitic pegmatites with zircon, titanite, pyrochlore involving silicate melt, Fe^Ti oxides and mafic silicate
and zirconolite (‘polymignite’), known as the eastern minerals during crystallization of the pluton.
type, as they are mainly found in the eastern part of
the complex.
(2) Mildly agpaitic nepheline syenite pegmatites contain-
ing complex, volatile-bearing zirconium silicate min- P E T RO G R A P H Y
erals (the Langesundsfjord type). These pegmatites In addition to samples collected for this study, c. 180 sam-
are relatively abundant in the southwestern part of ples of nepheline syenite pegmatite in the mineral collec-
the Larvik Plutonic Complex and its border zone in tions of the Natural History Museum, University of Oslo,
the Langesundsfjord archipelago and the adjoining and the National Museum of Natural History (NRM),
mainland area (Fig. 1). Stockholm, Sweden were examined under the binocular
(3) The rare, agpaitic Bratthagen type pegmatites microscope, and visual mineral identifications were
with parakeldyshite (Na2ZrSiO7), lorenzenite checked by X-ray diffraction (XRD) on selected speci-
(Na2Ti2Si2O9) and dalyite (K2ZrSi6O15) are asso- mens (Electronic Appendix 1, available at http://www
ciated with foyaite rather than larvikite, and re- .petrology.oxfordjournals.org). The Stockholm collection
stricted to a small area in the northern part of the is of special interest, as it contains many of the samples
complex (Larsen, 2010, and references therein). first studied by Brgger (1890). Detailed scanning elec-
tron microscopy and electron microprobe work was
The Langesundsfjord type pegmatites are intrusive in done on samples from selected pegmatites indicated
nepheline-normative larvikite belonging to ring segment in Fig. 1. Descriptions of the samples analysed by elec-
VI, and in contact-metamorphosed, Permo-Carboniferous tron microprobe in the present study are given
basalt (Fig. 1). Pegmatites in this area occur as more or in Electronic Appendix 2, at http://www.petrology.
less irregular veins, often not particularly coarse-grained. oxfordjournals.org.
The main minerals are white or greyish microcline, Some of the samples examined in this study are pegmati-
nepheline (often more or less altered to ‘spreustein’, i.e. tic in the classical sense; that is, ultra-coarse-grained
aggregates of pink, fibrous natrolite), albite sodalite rocks with well-developed crystals of the rock-forming
aegirine ferro-edenite (barkevikite) magnetite minerals. Others, however, are medium- to coarse-grained
biotite. Fluorite is a common mineral in these pegmatites, intrusive magmatic rocks, many of which are heteroge-
pyrophanite (MnTiO3) is also relatively common, ilmenite neous in terms of grain size and mineralogy on a
and rutile much less so (Larsen, 2010). In addition, a large hand-specimen scale.
variety of accessory minerals may be present, and the Fluorite has been observed in most samples examined in
abundance of Zr-, Ti-, Nb-, rare earth element- (REE-) this study, occurring in two textural settings (Fig. 2,
and Be-minerals is conspicuous (Brgger, 1890; Larsen, Table 2). In most samples it occurs as part of the early,
1996, 2010). The main rock-forming minerals and the zirco- magmatic mineral assemblage, forming intergrowths with
nium silicates formed during the magmatic crystallization and primary inclusions in the Zr-silicate minerals (Fig. 2h
stage in the pegmatites. A subsequent, hydrothermal stage and k). However, fluorite also occurs in younger, crosscut-
of evolution is discernible in most nepheline syenite ting veins (Fig. 2l) and in symplectites replacing zirconium
2306
silicate minerals (Fig. 2b). Zircon and the other zirconium lvenite has been replaced by symplectites consisting of
silicate minerals commonly occur as euhedral to subhedral fluorite, magnetite and zircon (Fig. 2k, arrows); according
crystals with well-developed crystal face terminations to Brgger (1890), this type of alteration is common for
against matrix feldspar, nepheline and sodalite, or lvenite in the area.
they are included in or intergrown with biotite and/or Rosenbuschite typically forms pale, acicular crystals
pyroxene. intergrown with hiortdahlite (Fig. 2f) and eudialyte
(Fig. 2g); in sample TRO, it replaces hiortdahlite (Fig. 2h).
Rosenbuschite is closely associated with fluorite, which is
Occurrence and textural relationships a general observation in these pegmatites (Brgger, 1890;
of the zirconium silicate minerals

Larsen, 2010). The type specimen (TYROS, NRM no.
Hiortdahlite forms plate-shaped crystals with pale yellow 531136) shows distinct, patchy variations in BSE brightness,
to greyish or brownish yellow colour (Fig. 2a). It is com- suggesting compositional heterogeneity (Fig. 2d).
monly in contact with or intergrown with fluorite and the Eudialyte in these pegmatites typically has a brown or
groundmass minerals microcline, nepheline and aegirine; orange^brown colour, and occurs as anhedral masses in
in samples TRO, RIS and ARS also with rosenbuschite contact with or enclosing rosenbuschite, hiortdahlite and
(Fig. 2f). Sample 24102 from the Langodden pegmatite con- lvenite (Fig. 2g). As pointed out by Brgger (1890), zircon
tains euhedral plates of hiortdahlite in contact with is not found in primary coexistence with eudialyte, but it
zircon (Fig. 2a), and hiortdahlite in sample ARS contains does occur as an alteration product in eudialyte-bearing
primary inclusions of zircon. In hiortdahlite-bearing sam- samples.
ples also containing eudialyte, the two minerals occur in Catapleiite commonly belongs to a late magmatic
direct contact. In sample ARS, hiortdahlite also occurs as stage of evolution in the pegmatites (Larsen, 2010).
primary inclusions in pyroxene. Hiortdahlite contains pri- However, in the present samples it occurs both as primary
mary inclusions of nepheline, microcline, apatite, fluorite, inclusions in hiortdahlite (sample IGH from Stokksund),
wo«hlerite and tritomite-(Ce). Alteration products of hiort- and as an alteration product after hiortdahlite (Fig. 2c).
dahlite include zircon þ fluorite intergrowths along grain Samples from the Lven pegmatite and sample 30047
boundaries (Fig. 2b), catapleiite (Fig. 2c) and lvenite^ from Vesle Arya contain both lvenite and catapleiite,
rosenbuschite (Fig. 2h) intergrowths along crosscutting but the two minerals have not been found in direct grain
fractures. contact.
Wo«hlerite occurs as aggregates of plate-shaped, bright Other minerals occurring in intergrowths with the
yellow crystals, intergrown with biotite and enclosed by zirconium silicate minerals or as primary solid inclu-
feldspar and nepheline. Samples containing wo«hlerite
sions are apatite (Fig. 2g), tritomite-(Ce), monazite
commonly contain only small amounts of late or secondary
(Fig. 2i), Ca-free REE silicate minerals, and pyrochlore
fluorite, or are free of visible fluorite, unless the samples
(Fig. 2j and l).
also contain eudialyte sensu lato (s.l.), hiortdahlite and/or
rosenbuschite. Where fluorite is found spatially associated
with wo«hlerite, it either fills fractures associated with in-
cipient alteration (Fig. 2l), or it occurs in aggregates to-
gether with hiortdahlite and/or rosenbuschite replacing M I N E R A L C H E M I S T RY
wo«hlerite, or it is in a clearly interstitial textural position Analytical methods
with respect to wo«hlerite. Wo«hlerite occurs as primary Scanning electron microscopy and electron microprobe
inclusions in hiortdahlite (Fig. 2b) and lvenite (Fig. 2j), (EMP) analysis was carried out at the Department of
and contains primary inclusions of pyrochlore (Fig. 2l). Geosciences, University of Oslo, and XRD at the Natural
Direct contacts between wo«hlerite and eudialyte have History Museum, University of Oslo. Mineral fragments
been observed in sample TRO. for XRD and electron microprobe work were separated
Lvenite forms aggregates of orange^brown, subhedral from the pegmatite matrix under a binocular microscope;
to anhedral grains in a matrix consisting of alkali feldspar, samples for EMP analysis were mounted in epoxy and
nepheline, biotite and clinopyroxene. It contains inclusions polished. The samples were imaged using backscattered
of wo«hlerite and pyrochlore (Fig. 2j), and in sample LV10, electrons on a Cameca SX100 electron microprobe and a
it forms primary intergrowths with fluorite (Fig. 2k). In JEOL JSM 6460LV scanning electron microscope with an
sample LV15, a lvenite-like mineral has formed along the Oxford Instruments LINK INCA energy dispersive ana-
boundary between wo«hlerite and a cross-cutting fluorite lyser prior to and during the analytical sessions, the BSE
vein (Fig. 2l). In sample TRO, a BSE-bright lvenite has images acting as guides for the selection of points to be
formed by alteration of hiortdahlite (Fig. 2h), and in analysed.
sample ARS, intergrowths of fluorite and a lvenite-like Electron microprobe analysis was performed on a
mineral form patches in eudialyte (Fig. 2e). At Lven, Cameca SX100 fitted with five wavelength-dispersive
2307

Fig. 2. Microstructural relationships of zirconium silicate minerals in nepheline syenite pegmatites from the Larvik Plutonic Complex. The
terms ‘primary’ and ‘secondary’ inclusions are used in the sense commonly used in the fluid inclusion literature, referring to inclusions trapped
during crystal growth, and along healed fractures in the host mineral, respectively (e.g. Roedder, 1984). Numbered points on the images repre-
sent electron microprobe analyses listed in Electronic Appendix 3. (a) Polarized light photomicrograph of pegmatite sample 24102 from
Langodden (locality 3 in Fig. 1). Elongated plates of hiortdahlite (Hio) and stubby prisms of zircon (Zrc) in mutual contact embedded in a
matrix of alkali feldspar (anhedral grains with visible cleavage) and nepheline (Ne). (b) EMP-BSE image of a fragment of the hiortdahlite holo-
type (TYHIO from the Langodden pegmatite, locality 3 in Fig. 1), showing primary inclusions of microcline (Kf) and wo«hlerite (Woe) in hiort-
dahlite, and aggregate of zircon (Zrc) and fluorite (Flu) formed at the grain boundary of hiortdahlite by post-crystallization processes. (c)
Secondary catapleiite (Cat) along a healed fracture in hiortdahlite from the hiortdahlite holotype (TYHIO). Dark spots within the catapleiite
vein are cavities limited by crystal faces (i.e. fluid inclusions cross-cut by the section surface). (d) A grain of the rosenbuschite holotype analysed
in the present study (TYROS from Skutesundskjr, locality 1 in Fig. 1). The patchy variations in BSE intensity are due to minor variations in
Ti/(Ti þ Zr þ Hf) ratio of the rosenbuschite (Fig. 3c and Electronic Appendix 3). (e) An intergrowth of a BSE-bright, lvenite-like mineral
(‘low-Fe lvenite’) and BSE-dark fluorite (Flu) enclosed by eudialyte (Eud). Eudialyte shows lower BSE brightness next to the aggregate.
Sample ARS from Aryskjrene (locality 4 in Fig. 1). (f) A grain of hiortdahlite (Hio) from the Risya pegmatite (locality 5 in Fig. 1), with a
cross-section through an euhedral fibre of rosenbuschite (Ros) occurring as a primary inclusion in hiortdahlite. (g) An aggregate consisting of
eudialyte (Eud), rosenbuschite (Ros) and apatite (Ap) from the Risya pegmatite (locality 5 in Fig. 1). Minor zircon (Zrc) is fracture-bound
and probably formed by alteration of rosenbuschite. (h) Hiortdahlite (Hio) from the Trompetholmen pegmatite (locality 6 in Fig. 1) with pri-
mary inclusions of zircon (Zrc) and fluorite (Flu). Black inclusions are alkali feldspar. The inset detail shows alteration of hiortdahlite to a
lvenite-like mineral (‘low-iron lvenite’, analysis 24) and rosenbuschite (analysis 25) along a secondary healed fracture. Analysis 26 is a hiort-
dahlite with slightly higher Zr/(Ti þ Zr) than the BSE-darker hiortdahlite away from the fracture (Electronic Appendix 3). The BSE contrast
of the enlarged inset has been enhanced. (i) Wo«hlerite (Woe) from the Trompetholmen pegmatite (locality 6 in Fig. 1) with inclusions of micro-
cline (Kf), monazite and tritomite-(Ce), which appear to be primary in the wo«hlerite, whereas monazite and rosenbuschite (Ros) are located
at a healed fracture trail, and are due to secondary mineral^fluid interaction processes at the late- or post-magmatic stage. (j) Homogeneous
lvenite (Lav) in sample LV04 from the Lven pegmatite (locality 7 in Fig. 1) with primary inclusions of wo«hlerite (Woe) and pyrochlore
(Pcl). (k) Lvenite (Lav) intergrown with fluorite (Flu) in sample LV10 from the Lven pegmatite. Zircon (Zrc) is a secondary mineral in this
sample, probably belonging to a stage of hydrothermal recrystallization at late- or post-magmatic conditions. Near the zircon crystal, lvenite
is replaced by symplectites (arrows) consisting of zircon (BSE-bright) and fluorite (BSE-dark). (l) Wo«hlerite (Woe) with a primary inclusion
of pyrochlore (Pcl) in sample LV15 from the Lven pegmatite. The wo«hlerite is penetrated by a fluorite vein (Flu), against which wo«hlerite is
rounded and resorbed, and locally replaced by a lvenite-like mineral (‘low-Fe lvenite’, analyses 125 and 126).‘Low-iron lvenite’ and wo«hlerite
show very similar BSE brightness, and can be distinguished only from electron microprobe analyses (Electronic Appendix 3).
2308
Table 2: Summary of critical mineral assemblages in pegmatites from the Larvik Plutonic Complex
Locality Fig. 11 Sample Mineral assemblages
Magmatic stage Late or secondary stage
Langodden 3 TYHIO a1: Wöhlerite þ hiortdahlite

a2: Hiortdahlite þ zircon þ fluorite a3: Zircon þ fluorite catapleiite
b: Lvenite þ eudialyte/catapleiite2

Vesle Arøya 8 30047
Trompetholmen 6 TRO c1: Wöhlerite þ pyrochlore þ eudialyte zircon
c2: Hiortdahlite þ rosenbuschite þ eudialyte þ fluorite c3: Low-Fe lvenite þ fluorite
Risøya 5 RIS d1: Wöhlerite þ eudialyte d3: Zircon þ fluorite
d2: Hiortdahlite þ rosenbuschite þ eudialyte þ fluorite
Arøyskjærene 4 ARS e1: Zircon þ fluorite
e2: Hiortdahlite þ rosenbuschite þ eudialyte þ fluorite e3: Catapleiite þ zircon þ fluorite
Skutesundsjkjær 1 TYROS f: Rosenbuschite þ fluorite
1 30086 g1: Wöhlerite þ pyrochlore
g2: Hiortdahlite þ rosenbuschite þ fluorite
Stokksund 2 IGH h: Hiortdahlie þ catapleite þ fluorite
Lven 7 TYLA i: Lvenite
7 LV04 j1: Wöhlerite
j2: Lvenite þ fluorite þ pyrochlore
7 LV10 k1: Lvenite þ fluorite k2: Zircon þ fluorite
7 LV15 l1: Lvenite þ wöhlerite
l2: Lvenite þ fluorite
7 LV08 m: Catapleiite
1
Localities marked in Fig. 1.
2
The paragenetic relationship between catapleiite and the other minerals is unclear.
a1, a2, a3, etc.: successive mineral assemblages in the pegmatite.
(WDS) spectrometers. An accelerating voltage of 15 kV, Matrix corrections were made according to the PAP
beam current of 15 nA and a focused beam were generally procedure (Pouchou & Pichoir, 1984).
used. Beam-sensitive minerals (catapleiite and eudialyte) In XRD analysis, crystalline materials examined in this
were analyzed at 10 nA and a beam diameter of 10 mm. study were identified using a Siemens D5005 X-ray
Counting time was 10 s on peak (and 5 s on each back- powder diffractometer equipped with a ceramic X-ray
ground positions). Na, F, Cl and K were analyzed first. tube (2·2 kW) and a Cu-Ka-anode and Debye^Scherrer
Prior to analysis, full-range spectra were acquired on sev- configuration, combined with BRUKER software for
eral grains of each mineral species to identify all the elem- evaluation (EVA) with the latest PDF database. Several of
ents present and possible interferences between X-ray the specimens in the Oslo collection were originally identi-
lines. Background positions for each mineral type were fied by XRD using 9 cm Debye^Scherrer cameras and
then selected on detailed WDS scans acquired around the Fe-Ka radiation; the films are kept in the archives of the
X-ray lines of interest. Natural History Museum, University of Oslo.
Calibration standards and X-ray lines used were wollas-
tonite (Si Ka, Ca Ka), Al2O3 (Al Ka), pyrophanite (Ti
Ka, Mn Ka), Fe metal (Fe Ka), MgO (Mg Ka) orthoclase Mineral compositions
(K Ka), albite (Na Ka), fluorite (F Ka), synthetic alforsite Electron microprobe data are given in Electronic
(Cl Ka), Monastery Mine zircon (Zr La), Hf metal (Hf Appendix 3 (http://www.petrology.oxfordjournals.org).
Ma), SiÂl^Ca glass with 15% ThO2 (Th Ma), Nb2O5 Hiortdahlite, wo«hlerite, lvenite and rosenbuschite were
(Nb La), Ta2O5 (Ta Ma), scheelite (W Ma), synthetic recalculated to structural formulae assuming Si ¼ 4·00 cat-
orthophosphates (Jarosewich & Boatner, 1991) for REE ions. This is justified by the A6þxB2^x(Si2O7)2(F,O,OH)4
and Y (La La, Ce La, Nd Lb, Y La). stoichiometry of these minerals, but it should be noted
2309
that a minor error in the Si analysis may propagate to a

significant error in the total charge balance. The minerals
considered here all have Zr/Ti41, which distinguishes
them from related, Ti-dominated species such as rinkite,
hainite, kochite and mosandrite (Table 1). Ta, W and U con-
centrations are below the detection limit of the electron
microprobe. Cl is insignificant in wo«hlerite- and rosen-
buschite-group minerals, but O2^ (and possibly OH) sub-
stitutes for fluorine in amounts ranging from 0·5 a.p.f.u.

(atoms per formula unit) in hiortdahlite to 42·0 a.p.f.u. in
wo«hlerite, which accounts for a small deficiency in the ana-
lytical totals.
Wo«hlerite shows little compositional variation within
and between the samples. If two analyses of low-Nb^
high-Ti wo«hlerite inclusions in the hiortdahlite holotype
(TYHIO analyses 27 and 28, Fig. 3c, Electronic Appendix
3) are excluded, the composition is given by the formula
Na2·0Ca3·8(Fe,Mn)0·3Ti0·1Zr1·0Nb0·7(Si4O14)F1·8O2·2, which is
indistinguishable from published analyses of wo«hlerite
from the region (Brgger, 1890; Mariano & Roeder, 1989).
Wo«hlerite has lower fluorine (54·5 wt %), Na and
REE concentrations, but significantly higher Nb than the
other wo«hlerite-group minerals analyzed in this study
(Fig. 3c).
Hiortdahlite is the most calcium-rich (430 wt % CaO)
and sodium-poor member of the wo«hlerite-group species
analyzed here (Fig. 3a); its fluorine concentration (c. 8 wt
%) is significantly higher than that of wo«hlerite, and com-
parable with that of rosenbuschite. It also shows restricted
compositional variation within samples, but there is a
minor but probably significant difference in Na/Ca ratio
between the Stokksund (IGH) sample and the hiortdahlite
holotype on one hand, (Na1·9Ca4·4(Fe,Mn)0·3REE0·1
Ti0·1Zr1·1Nb0·1(Si4O14)F3·4O0·6), and the Risya, Aryskjr
and Trompetholmen hiortdahlite on the other
(Na1·7Ca4·6(Fe,Mn)0·2REE0·1Ti0·1Zr1·1Nb0·1(Si4O14)F3·5O0·5).
Lvenite in TYLA and 30047 is homogeneous, with
composition close to the ideal endmember at (Na3·1Ca1·1
(Fe,Mn)1·3REE50·1Ti0·3Zr1·8Nb0·3(Si4O14)F2·2O1·8). Lvenite
in samples LV04 and LV10 shows much more variable
composition ranging towards c. Na3Ca2(Fe,Mn)0·7
analysis 24 in Electronic Appendix 3. (a) Na^Nb^Ca/5 (Keller et al.,

1995). It should be noted that the ‘low-iron lvenite’ has Na/
(Ca þ Na) between the compositional ranges of lvenite and rosen-
buschite. (b) Mn^Fe^Ca/5. In this plot, wo«hlerite, rosenbuschite and
hiortdahlite cannot be distinguished, but lvenite stands out with sig-
nificantly higher Fe þ Mn. Noteworthy features are the range of Fe/
(Fe þ Mn) ratios in lvenite, and the intermediate position of the
‘low-iron lvenite’. (c) Quadrivalent and pentavalent ions, repre-
Fig. 3. Cation distribution in hiortdahlite (Hio), wo«hlerite (Woe), sented by Ti^Zr þ Hf^Nb. Wo«hlerite stands out with high Nb in this
lvenite (Lav) and rosenbuschite (Ros); data from Electronic diagram. (Note also the range in Ti/(Ti þ Zr þ Hf) in rosenbuschite,
Appendix 3. Numbers refer to analyses mentioned in the text. and that the rosenbuschite holotype (TYROS) show an internal com-
Analysis numbers should be combined with the appropriate sample positional range corresponding to the total variation in rosenbuschite
identifier, so that the point marked ‘#24’ in (b) corresponds to TRO observed in the present study.)
2310
Ti0·3Zr1·8Nb0·2(Si4O14)F3·5O0·5, and higher Fe/(Fe þ Mn) 3M(2)3M(3)M(4)Z3[Si24O72]O4X2, in which X ¼Cl, F, OH

ratio (Fig. 3b). The overall compositional variations of or CO3, Z ¼ Zr þTi, and N and M are cation sites incor-
lvenite in these samples are controlled by simple FeMn^1 porating Na, K, Ca, Nb, Fe, Mn (Johnsen et al., 2003). Si
and Ca^1(Mn,Fe) substitutions, and a coupled cation^ is also accommodated in the off-ring, tetrahedral position
anion substitution of the type CaF2M4þ^1O^2, where M4þ M4, and it competes with Nb and other high-valence ions
is Zr, Hf and Ti (Fig. 3b and c). in four- to six-coordinated M3 position. The compositions
The BSE-bright alteration product in hiortdahlite from of eudialyte s.l. in Electronic Appendix 3 have been
Trompetholmen (TRO analysis 24, Fig. 2h) is lvenite recalculated to a constant sum of Si þ Al þ
with higher Fe/Mn ratio than any of the other lvenites Nb þ Zr þ Hf þ Ti ¼ 29·000 (Johnson & Grice, 1999), and

analysed (Fig. 3b). The BSE-bright mineral in hiortdahlite cation distributions have been calcualted using the algo-
from Risya (RIS analysis 33), the alteration zone in wo«h- rithm of Pfaff et al. (2010), modified so that excess
lerite against fluorite in LV15 (LV15 analyses 124 and 125), Ca þ REE is allocated to the N sites irrespective of other
and the BSE-bright phase intergrown with fluorite in the site occupancies. The eudialyte s.l. analysed in this study
symplectite-like patches in Aryskjrene eudialyte (ARS has total Fe/Mn41 (Fig. 4a). Most analyses have SiM3/
analyses 34, 35, 39, Fig. 2e) all have compositions clustering NbM3 1, and incorporate minor Ti þ Zr in M3 (Fig. 4b).
around an idealized composition Na3Ca2·5(Fe,Mn)0·3 Z is filled by Zr þ Hf, and M2 is fully occupied by Fe and
Ti0·2Zr1·8Nb0·2(Si4O14)F3·5O0·5, which falls within the Mn, with excess Mn in M1, which is otherwise dominated
compositional gap between lvenite and rosenbuschite, by Ca. The N positions hold Na, K, Ca and REE, with
hiortdahlite and wo«hlerite (Fig. 3). The observed com- total N-site occupancy 14·6 cations. The distribution of
positions are compatible with Si2O7 type structures cations in the M3 and N sites indicates that the eudialyte
of rosenbuschite or wo«hlerite type; the low Ti suggests s.l. in these pegmatites can best be described as a fekliche-
that they belong in the wo«hlerite group rather than in vite^ferrokentbrooksite^zirsiliteêudialyte solid solution
the rosenbuschite group (Fig. 3c). This composition is with H3OþN and/or vacancies relative to 16 N-positions
most similar to lvenite, but has lower Fe þ Mn and (Fig. 4c and d; endmembers from Johnsen et al., 2003).
higher Ca than any lvenite reported in literature, includ- Eudialytes from other pegmatites in the Larvik Plutonic
ing the yellow lvenite from Vesla Arya analysed by Friis Complex have been classified as intermediates between fer-
et al. (2010). Below, it will be referred to as ‘low-iron rokentbrooksite and zirsilite (Larsen et al., 2005a, 2005b).
lvenite’. The name ‘eudialyte’ is, however, retained in the following
Rosenbuschite contains Ti as an essential component in discussion. Chlorine amounts to between 0·75 and
its structure, and the new analyses of the mineral show 1·40 wt %, but the combined amount of Cl and F will not
consistently higher Ti concentrations than in the coexisting fill the X positions, indicating additional substitution of
wo«hlerite group minerals. It is also more variable in com- OH or CO3 in these positions, in agreement with the low
position than wo«hlerite and hiortdahlite. The new analyses sums.
correspond to an approximate range in major compo- In eudialyte from the Aryskjrene pegmatite,
nents from Na2·2Ca3·8(Fe,Mn)0·2REE0·2Ti0·7Zr0·9Nb0·1 BSE-dark zones rimming aggregates of fluorite and
(Si4O14)F3·2O0·8 to Na2·6Ca3·3(Fe,Mn)0·1REE0·1Ti0·6Zr1·2 low-iron lvenite (Fig. 2e) show simultaneous reduction of
Nb0·1(Si4O14)F3·2O0·8. Simple (Zr,Hf)Ti^1 substitution Na and Ca, slightly increasing cation deficiency and
and a coupled Na2M4þCa^3 are the main controls on increasing [Na/(Na þ Ca)]N towards the aggregates (com-
th compositional variation. An overall variation in F pare points for ARS analyses 36, 13 and 14 from the inter-
from 2·5 to 3·5 a.p.f.u. is uncorrelated with Ca. The growth zone shown in Fig. 2e with those for ARS analyses
sample of the rosenbuschite holotype (TYROS) is compos- 11, 12, 33 and 32 from the central part of the same grain).
itionally heterogeneous, with an internal variation in The most likely explanation for the reduced BSE bright-
Zr þ Hf and (Zr þ Hf)/Ti ratio comparable with the ness in these zones is increasing substitution of water in
total compositional variation of the rosenbuschite sam- the eudialyte.
ples analysed in this study (Fig. 3c). The variation in Catapleiite is a water-bearing silicate whose structure is
the (Zr þ Hf)/Ti ratio is qualitatively correlated based on Si3O9 rings (Table 1). The present analyses
with patchy variations in BSE intensity shown in Fig. 2d. (Electronic Appendix 2) range from a pure sodium
The analysis of the same sample reported by Christiansen catapleiite (Na2ZrSi3O9·2H2O) to a composition close to
et al. (2003) falls in the middle of the range of the new NaCa0·5ZrSi3O9·2H2O), controlled by the vacancy substitu-
analyses. tion mechanism CaœNa^2. This range of composition
The minerals of the eudialyte group have complex struc- mimics that reported by Larsen (2001) for catapleiite
tures based on three- and nine-membered rings and single from the Larvik Plutonic Complex. Catapleiite in LV08
SiO4 groups (Johnsen et al., 2003).The general formula is distinctly cation deficient at a composition
can be written [N(I)N(2)N(3)N(4)N(5)]3[M(la)M(lb)] Na0·6Ca0·3ZrSi3O9·2H2O.
2311
M O D E L L I N G T H E S TA B I L I T Y O F
Z I RC O N I U M S I L I C AT E L I Q U I D U S
A S S E M B L AG E S I N P E R A L K A L I N E
N E P H E L I N E SY E N I T E M AG M A
Primary and secondary zirconium silicate
mineral assemblages
The minerals of the wo«hlerite, rosenbuschite and eudialyte
groups are part of the magmatic mineral assemblage of

the pegmatites, together with the main rock-forming min-
erals microcline, albite, annite-rich biotite, aegirine/aegir-
ineâugite, magnetite, fluorite and minor and accessory
minerals (pyrochlore, apatite, tritomite-(Ce) and others).
Based on the present observations, sequences of successive
mineral assemblages related to magmatic and post-
magmatic processes can be established (Table 2). An as-
semblage is identified by a lower-case letter representing
the sample, and a subscript number giving the relative
timing of the assemblage in the crystallization history of
that particular sample. For example, the Langodden peg-
matite shows three successive zirconium silicate mineral
assemblages: early crystallized wo«hlerite and hiortdahlite
(assemblage a1 inTable 2) was followed first by hiortdahlite
and zircon (a2), and later by secondary assemblages with
zircon, fluorite and catapleiite (a3).
Hiortdahlite, rosenbuschite, lvenite and zircon form
part of primary mineral assemblages with magmatic
fluorite (assemblages a2, c2, d2, f, g2, h, j, k1), but there is no
evidence for equilibrium coexistence of wo«hlerite with
fluorite. Wo«hlerite does, however, occur in magmatic as-
semblages with eudialyte at Risya and Trompetholmen
(assemblages c1 and d1), with hiortdahlite at Langodden
(a1), with pyrochlore at Skutesundskjr (g1) and lvenite
at Lven (l1). Inclusions of rosenbuschite observed in wo«h-
lerite appear to be trail-bound and secondary (Fig. 2i). In
the samples from Trompetholmen and Risya, wo«hlerite
(assemblages c1, d1) and hiortdahlite þ rosenbuschite (as-
semblages c2, d2) occur in separate domains, which may
record different stages of crystallization or small-scale
heterogeneity in composition. Replacement of wo«hlerite
by assemblages with fluorite and hiortdahlite and/or rosen-
buschite is a common feature in wo«hlerite-bearing pegmat-
ites (Table 2), suggesting that the wo«hlerite þ eudialyte
and hiortdahlite þ eudialyte þ fluorite rosenbuschite
assemblages reflect successive stages of magmatic
crystallization.
indicated by black dots have been taken from table 3 of Johnsen et al.
(2003). (a) Total Mn^Fe^Nb distribution, illustrating that Fe4Mn in
Fig. 4. Cation distributions in eudialyte s.l., from data given in eudialyte in these samples. (b) M3 cation distribution, showing
Electronic Appendix 3. Structural formulae have been calculated roughly equal Nb and Si occupancies in this position. (c) N-cation
assuming Si þ Al þTi þ Zr þ Hf þ Nb ¼ 29·000; cations have been distribution, illustrating the significant contribution of Ca þ REE in
assigned to structural positions as given by Johnsen et al. (2003) using eudialyte from these pegmatites. (d) Vacant and occupied N-pos-
an adaptation of the algorithm of Pfaff et al. (2010), which allows itions, shown in a NaN^(Ca þ REE)N^(H3Oþ þ œ)N diagram. The
excess Ca þ REE to be allocated to the N-positions in the structure sum of vacancies and H3Oþ ions in the N position has been estimated
without regard for other site occupancies. Endmember compositions assuming 16 N-positions.
2312
Zircon forms part of early mineral assemblages in the ferric iron from the system, we disregard the effect of
Langodden and Aryskjrene pegmatites (a2, e1). oxygen fugacity. In these pegmatites, fO2 is, however, in-
Secondary fluorite þ zircon catapleiite assemblages are ternally controlled by reactions involving titanomagnetite,
found in several of the samples studied (assemblages a3, microcline and biotite.
d3, e3, k2; Table 2). Other late magmatic or secondary as- Stoichiometric endmember compositions for the major
semblages are made up by fluorite þ low-iron lvenite (as- rock-forming minerals microcline, albite, nepheline and
semblage c3), replacing hiortdahlite in the Trompetholmen sodalite are used in the model calculations (Table 3).
pegmatite (Fig. 2h) and wo«hlerite in LV15 (assemblage k2, Micas in the pegmatites are close to annite in composition
Fig. 2l). Lvenite is itself commonly replaced by pseudo- (Larsen, 1998), and we include a pure hydroxy annite end-

morphs consisting of zircon, fluorite and magnetite, which member in our calculations, assuming the F^Fe avoidance
must have formed during a later, post-magmatic stage principle in biotite (Munoz, 1984) to be applicable.
of evolution (Fig. 2k). The textural status of the Clinopyroxene in the nepheline syenite pegmatites is in
symplectite-like aggregates of fluorite and the BSE-bright general aegirineâugite to aegirine; c. 10 mol % of the
zirconium silicate mineral in the Aryskjrene eudialyte hedenbergite (CaFeSi2O6) component has been reported
(Fig. 2e) is enigmatic, but the observation that eudialyte is from the Lven and Aryskjrene pegmatites (Larsen &
increasingly Na þ Ca deficient and hydrated towards the Raade, 1997; Murad, 2006). Because the NaFe3þSi2O6 com-
contact to the aggregates suggests that this feature is due ponent cannot be accounted for in the present model, we
to interaction with an aqueous fluid, and that the aggre- use the hedenbergite component in the calculations. This
gates are pseudomorphs after hiortdahlite or rosen- will invoke an activity term, which is, however, the same
buschite. Although catapleiite forms part of the liquidus for all reactions involving pyroxene.
assemblage in some pegmatites, it also occurs in late re- Niobium is essential in wo«hlerite, and occurs as a minor
placement assemblages together with zircon and fluorite component in rosenbuschite, hiortdahlite, lvenite and
in others (Table 2). eudialyte. The main niobium (and tantalum) host in the
pegmatites is, however, pyrochlore. Pyrochlore in the sam-
Chemographic analysis
ples studied is metamict and altered, and we have therefore
The magmatic mineral assemblages including zirconium
not carried out systematic analyses of this mineral. We use
silicates and the main rock-forming minerals of the peg-
a theoretical endmember composition (NaCaNb2O6F) for
matites are related to the silicate liquid by a series of min-
modelling purposes.
eral^melt reactions that are influenced by parameters
Titanium is present in minor amounts in all of the wo«h-
such as temperature, pressure, oxygen fugacity, activities
lerite and eudialyte group minerals, and is an essential
of volatile species (e.g. water and halogens), and the per-
structural constituent in the rosenbuschite group
alkalinity of the melt. In the present study, we assume crys-
tallization at constant pressure and temperature for the (Sokolova, 2006). Ti is buffered by titanomagnetite and
magmatic stage, and evaluate the relative influence on other Ti-bearing minerals, and is ultimately related to
mineral stability of the excess alkali content of the melt oxygen fugacity. Here, we assume that it is externally buf-
(represented by the activity of the sodium disilicate com- fered by titanium-bearing oxides (titanomagnetite, pyro-
ponent, Nds ¼ Na2Si2O5, e.g. Andersen & Srensen, 2005) phanite) and silicates (mosandrite, rinkite, astrophyllite).
and the activities of water, fluorine and chlorine (repre- We also assume external buffering for the REE, which
sented by HF and HCl). The composition-dependent vari- are hosted in minerals such as apatite, monazite, trito-
ables aNds, aHF and aH2O define a 3D chemical potential or mite-(Ce) and other REE silicate minerals (Larsen,
log activity space in which a petrogenetic grid can be con- 1996), some of which contain major components not
structed from observed mineral assemblages and compos- accounted for in the 13-component system (phosphorus,
itions using the Schreinemakers method (Zen, 1966; White boron).
et al., 2008, and references therein). Chlorine is of interest only for reactions involving eudia-
Twenty or more components are necessary to give a full lyte. In general, an elevated activity of chlorine will in-
chemographic description of the zirconium silicate- crease the stability field of eudialyte at the expense of the
bearing mineral assemblages of the pegmatites. This is wo«hlerite and rosenbuschite group minerals. Saturation
clearly not possible, and some simplifications must be with sodalite represents the maximum attainable chloride
made. The number of components can be reduced by elim- activity of a nepheline syenite magma, and the stability
ination of minor species by combining Hf with Zr, Y with field of eudialyte will therefore have a maximum size in a
the REE, and Mn and Mg with Fe. The pertinent mineral system crystallizing sodalite þ nepheline (reaction (2) in
reactions can then be modelled in terms of the simplified Table 4). The highest possible activity of fluorine (or HF)
11 oxide components NaO0·5, KO0·5, CaO, FeO, REEO1·5, is given by saturation with fluorite. The fluorite saturation
ZrO2, TiO2, NbO2·5, AlO1·5, SiO2, HO0·5 and two volatile reaction is divariant in the eight-component sub-system
exchange components: FO^0·5 and ClO^0·5. By omitting NaO0·5^KO0·5^CaO^FeOÂlO1·5^SiO2^HO0·5^FO^0·5
2313
Table 3: Composition (in moles per formula unit) of phases and species used in modelling
SiO2 AlO1·5 ZrO2 TiO2 NbO2·5 REEO1·5 FeO CaO NaO0·5 KO0·5 HO0·5 FO–0·5 ClO–0·5
Major minerals in pegmatite

Albite Ab 3·0 1·0 0·0 0·0 0·0 0·0 0·0 0·0 1·0 0·0 0·0 0·0 0·0
Nepheline Ne 4·0 4·0 0·0 0·0 0·0 0·0 0·0 0·0 3·0 1·0 0·0 0·0 0·0
Microcline Kf 3·0 1·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 1·0 0·0 0·0 0·0

Annite component in biotite Annbi 3·0 1·0 0·0 0·0 0·0 0·0 3·0 0·0 0·0 1·0 2·0 0·0 0·0
Hedenbergite component in clinopyroxene Hdcpx 2·0 0·0 0·0 0·0 0·0 0·0 1·0 1·0 0·0 0·0 0·0 0·0 0·0
Pyrochlore Pcl 0·0 0·0 0·0 0·0 2·0 0·0 0·0 1·0 1·0 0·0 0·0 1·0 0·0
Sodalite Sod 3·0 3·0 0·0 0·0 0·0 0·0 0·0 0·0 4·0 0·0 0·0 0·0 1·0
Fluorite Flu 0·0 0·0 0·0 0·0 0·0 0·0 0·0 1·0 0·0 0·0 0·0 2·0 0·0
Zirconosilicate minerals
Wöhlerite1 Woe 4·0 0·0 1·0 0·2 0·7 0·0 0·2 3·8 2·0 0·0 0·0 1·8 0·0
Rosenbuschite1 Ros 4·0 0·0 1·0 0·7 0·1 0·1 0·2 3·6 2·4 0·0 0·0 3·0 0·0
Hiortdahlite1 Hio 4·0 0·0 1·1 0·1 0·1 0·1 0·3 4·5 1·8 0·0 0·0 3·4 0·0
Lvenite1 Lav 4·0 0·0 1·6 0·3 0·2 0·0 0·8 1·8 3·0 0·0 0·0 2·6 0·0
Eudialyte1 Eud 25·4 0·1 3·1 0·1 0·4 0·6 3·6 6·9 10·5 0·2 2·0 0·2 1·1
Catapleiite1,2 Cat 3·0 0·0 1·0 0·0 0·0 0·0 0·0 0·3 1·4 0·0 4·0 0·0 0·0
Zircon Zrc 1·0 0·0 1·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0
Externally buffered minor oxide components
TiO2 0·0 0·0 0·0 1·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0
REE2O3 0·0 0·0 0·0 0·0 0·0 2·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0
Components in melt/fluid
Na2Si2O5 Nds 2·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 2·0 0·0 0·0 0·0 0·0
H2O 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 2·0 0·0 0·0
HF 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 0·0 1·0 1·0 0·0
1
Mineral compositions from Electronic Appendix 1.
2
Additional data from Larsen (2001).
(reaction (1) in Table 4), and thus independent of which þ pyrochlore þ two zirconium silicate minerals þ volatile-
zirconium silicate minerals are stable. bearing silicate liquid) define a formally divariant mineral
Whereas there is compelling evidence for the presence of assemblage in the chlorine-free sub-system, which can be
a free, aqueous fluid phase during the later, miarolitic and represented by a plane in 3D log activity space. For reac-
hydrothermal stages of pegmatite crystallization (Brgger, tions involving eudialyte, ClO^0·5 must be included as a
1890; Larsen, 2010), this is commonly not the case during component, and the number of phases present must be
initial, magmatic crystallization of agpaitic nepheline sy- increased to 10, adding sodalite. The divariant planes sep-
enite magma (Srensen, 1997), and we do not invoke the arate (trivariant) stability volumes for the single zirco-
presence of a separate non-silicate fluid in the following nium silicate minerals, pairs of planes intersect along
analysis of magmatic phase reactions. It should, however, univariant lines (10 phases in the chlorine-free subsystem)
be noted that Brgger (1890) observed abundant two-phase and three planes intersect in isobaricîsothermal invariant
aqueous(?) fluid inclusions in catapleiite. No systematic points represented by 11 (or 12, including sodalite) coexist-
search for fluid inclusions has been made in the present ing phases.
study, but observations of crystal-face limited microcavities A reaction between several phases and/or chemical
in catapleiite in some samples (Fig. 2c) support the hypoth- species in a multicomponent system can be written as
esis that at least some of the late catapleiite has crystallized X
in the presence of a free, aqueous fluid. ni X i ¼ 0
i
At constant temperature and pressure, with external buf-
fering of TiO2 and REE2O3, nine coexisting phases (e.g. where ni is the stoichiometric coefficient of the ith species
microcline þ albite þ nepheline þ biotite þ clinopyroxene involved in the reaction (Xi). By convention, ni is positive
2314
Table 4: Divariant mineral reactions between zirconium silicate minerals, major pegmatite minerals and melt
Fluorite saturation
(1) 0·25 Ne þ 0·08 Kf þ 1·00 Hdcpx þ 2·00 HF ¼ 1·00 Flu þ 0·75 Ab þ 0·33 Annbi þ 0·67 H2O
Sodalite–nepheline–HCl reaction
(2) 0·38 Ab þ 0·88 Ne þ 0·50 H2O þ 0·50 Nds þ 1·00 HCl ¼ 1·00 Sod þ 0·88 Kf
Hiortdahlite volume
(3) 0·89 Hio þ 2·60 Ab þ 1·25 Annbi þ 0·26 H2O ¼ 1·00 Zrc þ 0·87 Ne þ 0·38 Kf þ 3·96 Hdcpx þ 0·08 TiO2 þ 0·05 Pcl þ 0·04 REE2O3 þ 0·78 Nds þ

3·01 HF
(4) 0·89 Hio þ 2·60 Ab þ 1·15 Annbi þ 2·36 H2O ¼ 1·00 Cat þ 0·86 Ne þ 0·28 Kf þ 3·66 Hdcpx þ 0·081 TiO2 þ 0·05 Pcl þ 0·04 REE2O3 þ 0·08 Nds þ
3·00 HF
(5) 0·87 Hio þ 0·45 Ab þ 0·01 Kf þ 0·14 Annbi þ 0·08 TiO2 þ 0·29 Pcl þ 0·61 H2O þ 0·07 Nds ¼ 1·00 Woe þ 0·15 Ne þ 0·41 Hdcpx þ 0·03 REE2O3 þ
1·50 HF
(6) 0·92 Hio þ 0·29 Ab þ 0·17 Annbi þ 0·58 TiO2 þ 0·02 REE2O3 þ 0·36 Nds ¼ 1·00 Ros þ 0·10 Ne þ 0·08 Kf þ 0·58 Hdcpx þ 0·01 Pcl þ 0·13 HF þ
0·11 H2O
Wöhlerite volume
(7) 1·02 Woe þ 2·14 Ab þ 1·10 Bi ¼ 1·00 Zrc þ 0·71 Ne þ 0·39 Kf þ 3·54 Hdcpx þ 0·17 TiO2 þ 0·35 Pcl þ 0·01 REE2O3 þ 0·85 Nds þ 0·44 H2O þ 1·48 HF
(8) 1·02 Woe þ 2·14 Ab þ 1·00 Annbi þ 1·66 H2O ¼ 1·00 Cat þ 0·71 Ne þ 0·29 Kf þ 3·24 Hdcpx þ 0·17 TiO2 þ 0·35 Pcl þ 0·01 REE2O3 þ
0·15 Nds þ 1·48 HF
(9) 1·05 Woe þ 0·06 Ne þ 0·02 Annbi þ 0·49 TiO2 þ 0·05 REE2O3 þ 0·29 Nds þ 1·45 HF ¼ 1·00 Ros þ 0·19 Ab þ 0·09 Kf þ 0·15 Hdcpx þ 0·32 Pcl þ
0. 83 H2O
Zircon–catapleiite
(10) 1·00 Zrc þ 0·10 Kf þ 0·30 Hdcpx þ 2·10 H2O þ 0·70 Nds ¼ 1·00 Cat þ 0·10 Annbi
Eudialyte volume
(11) 0·32 Eud þ 0·63 Ne þ 0·06 Kf þ 1·33 Hdcpx þ 0·15 TiO2 þ 0·27 Pcl þ 1·45 HF ¼ 1·00 Woe þ 1·85 Ab þ 0·75 Annbi þ 0·08 REE2O3 þ 0·13 H2O þ
0·84 Nds þ 0·35 Sod
(12) 0·37 Eud þ 0·90 Ne þ 0·05 Kf þ 1·99 Hdcpx þ 0·07 TiO2 þ 3·39 HF ¼ 1·00 Hio þ 2·49 Ab þ 1·02 Annbi þ 0·02 Pcl þ 0·06 REE2O3 þ 0·84 H2O þ
1·04 Nds þ 0·40 Sod
(13) 0·34 Eud þ 0·73 Ne þ 1·25 Hdcpx þ 0·64 TiO2 þ 2·98 HF ¼ 1·00 Ros þ 2·14 Ab þ 0·03 Kf þ 0·77 Annbi þ 0·03 Pcl þ 0·04 REE2O3 þ 0·59 H2O þ
0·41 Nds þ 0·37 Sod
(14) 0·33 Eud þ 0·25 Ab þ 0·24 Annbi þ 1·61 H2O þ 0·01 HF ¼ 1·00 Cat þ 0·07 Ne þ 0·23 Kf þ 1·89 Hdcpx þ 0·07 Pcl þ 0·02 TiO2 þ 0·09 REE2O3 þ
1·00 Nds þ 0·35 Sod
(15) 0·33 Eud þ 0·38 Ab þ 0·24 Ne þ 0·35 Annbi ¼ 1·00 Zrc þ 0·66 Kf þ 2·19 Hdcpx þ 0·02 TiO2 þ 0·07 Pcl þ 0·10 REE2O3 þ 0·68 H2O þ
1·52 Nds þ 0·36 Sod
Lvenite volume
(16) 0·60 Lav þ 0·40 Ne þ 0·50 Kf þ 2·44 Hdcpx þ 0·27 Pcl þ 0·01 REE2O3 þ 0·91 H2O ¼ 1·00 Woe þ 1·20 Ab þ 0·89 Annbi þ 0·01 TiO2 þ 0·03 Nds þ
0·05 HF
(17) 0·61 Lav þ 0·91 Ab þ 0·12 Kf þ 0. 19 Annbi þ 0·58 H2O ¼ 1·00 Zrc þ 0·30 Ne þ 1·05 Hdcpx þ 0·18 TiO2 þ 0·07 Pcl þ 0·01 REE2O3 þ 0·88 Nds þ
1·53 HF
(18) 0·69 Lav þ 0·63 Ne þ 0·58 Kf þ 3·28 Hdcpx þ 0·04 REE2O3 þ 0·36 H2O þ 1·66 HF ¼ 1·00 Hio þ 1·90 Ab þ 1·19 Annbi þ 0·02 Pcl þ 0·11 TiO2 þ
0·12 Nds
(19) 1·86 Lav þ 2·02 Ab þ 137 Kf þ 3·49 Hdcpx þ 0·28 REE2O3 þ 3·26 H2O þ 2·51 Nds þ 1·09 HCl ¼ 1·00 Eud þ 0·71 Ne þ 0·46 Annbi þ 0·49 TiO2 þ
4·70 HF
Rosenbuschite–catapleiite
(20) 0·97 Ros þ 2·23 Ab þ 0·98 Annbi þ 2·46 H2O ¼ 1·00 Cat þ 0·77 Ne þ 0·21 Kf þ 3·10 Hdcpx þ 0·04 Pcl þ 0·64 TiO2 þ 0·06 REE2O3 þ 0·43 Nds þ
2·88 HF
Abbreviations used for idealized endmember minerals and dissolved components in the melt are given in Table 3.
2315
for products and negative for reactants. The equilibrium 3D activity space represent the phase equilibria appropri-
constant for the reaction (log K) is given by ate for the nepheline syenite pegmatites. Although both
X the fluorite-saturated and -undersaturated domains must
log K ¼ ni log ai be considered, it is most convenient to start with the distri-
i
bution of stability fields at fluorite saturation, as this
where ai is the activity of the ith species. reduces the problem to a case of 2D Schreinemakers ana-
In the present case, the reactants and products are: (1) lysis. Figure 5a shows five possible configurations of stabil-
solid phases of constant composition (endmember feldspar, ity fields in the fluorite-saturated, Cl-free subsystem
nepheline, sodalite, hydroxy-annite, zirconium silicate surrounding the hiortdahlite (Hio) field (shaded) in a pro-

minerals); (2) endmembers in multicomponent solid solu- jection onto the log aNds^log aH2O plane. Mineral assem-
tions (the CaFeSi2O6 component in aegirineâugite, blages that are forbidden in scenarios iii, iv and v have
NaCaNb2O6F in pyrochlore); (3) externally buffered, been found to exist in the pegmatites (Flu þ Zrc þ Cat,
minor oxides (TiO2, REE2O3); (4) the three species of Flu þ Zrc þ Hio, Flu þ Hio þ Cat, Flu þ Hio þ Ros, Lav þ
interest, Nds, H2O and HF. Assuming that the activities Flu), whereas observed mineral assemblages agree with
of the pure solids are unity, the equation can then be the predictions of scenarios i and ii (Fig. 5b). If wo«hlerite
rewritten does not coexist in equilibrium with fluorite, the
X four-phase assemblage Hio þ Cat þ Ros þ Flu is stable
log K ¼ ni log ai (scenario i in Fig. 5). If, on the other hand, wo«hlerite is
solid solutions
X stable with fluorite, this assemblage would be replaced by
þ nj log aj three four-phase assemblages involving fluorite with
externally buffered components
þnNds log aNds þ nH2 O log aH2 O þ nHF log aHF : wo«hlerite and two of the minerals hiortdahlite, catapleiite
and rosenbuschite (scenario ii in Fig. 5). None of the critical
Only the three last terms include the activities of four-phase assemblages that would confirm the validity of
Na2Si2O5, H2O and HF in the pegmatite magma. In the either scenario i or ii have been observed. Although the
absence of thermodynamic data on the participating min- present observations do not support the existence of any
erals and oxide species in the silicate liquid, G for the magmatic mineral assemblage involving stable Woe þ Flu,
reaction, and thereby log K, cannot be determined. there is no conclusive evidence against. We therefore chose
Furthermore, the activity terms for the pyrochlore and scenario ii as the preferred alternative, but with a
clinopyroxene solid solution, and for the minor oxide spe- Woe þ Flu field that is small, and/or restricted to combin-
cies are (at best) poorly constrained. Nevertheless, all ations of H2O and Nds activities that are not realized in
these terms can be regarded as constant with respect to the nepheline syenite pegmatites. It is possible that other
the three activity terms of interest, and the equation geometrical configurations will be stabilized by changes
simplifies to in temperature, pressure or other parameters not con-
sidered here. For example, decreasing fO2 would be likely
nNds log aNds þ nH2 O log aH2 O þ nHF log aHF ¼ constant:
to expand the stability field of lvenite with essential Fe2þ
This is the equation of a plane in an orthogonal, 3D co- at the expense of the low-iron zirconium silicate minerals,
ordinate system with log aNds, log aH2O and log aHF as eventually stabilizing a configuration like scenario iv in
axes. Although the right-hand side of the equation (and Fig. 5a.
hence the absolute coordinates of the plane) remains un-
constrained, its orientation in activity space is uniquely The 3D petrogenetic grid
given by the ratios of the stoichiometric coefficients; for ex- The 2D configuration of the fluorite saturation plane
ample, the slope of the intersection of a plane with a sec- (Fig. 5a) constrains the geometry of the fluorite-
tion parallel with the log aNds and log aHF axes is given by undersaturated part of the 3D petrogenetic grid. The
^nNds/nHF, and similar for the slopes in the projections divariant fluorite saturation surface (reaction (1) in
defined by the two other combinations of orthogonal axes. Table 4) is parallel to the log aNds axis, and has a positive
slope of 0·33 in the log aH2O^log aHF projection (Fig. 6).
An internally consistent petrogenetic grid The stability field boundaries in Fig. 5 are the intersections
Using idealized phase compositions listed in Table 3, 20 of this plane with divariant surfaces in the fluorite-
formally divariant, balanced mineral^melt reactions can undersaturated system, the features of which are conveni-
be defined (Table 4). ently illustrated in a projection of the fluorite saturation
surface to the log aNds^log aH2O plane (Fig. 7a) and in sec-
Fluorite-saturated case tions parallel to the log aNds^log aHF and log aH2O^log aHF
For the fluorite-saturated case comparison with observed planes (Fig. 7b^g). The grid is semiquantitative, in
mineral assemblages is necessary to determine which of the sense that it is internally consistent and gives a
several possible configurations of divariant reactions in true representation of the relative slopes of the divariant
2316
Fig. 5. (a) Possible configurations of stability fields of hiortdahlite (Hio, shaded), zircon (Zrc), catapleiite (Cat), rosenbuschite (Ros), lvenite
(Lav) and wo«hlerite (Woe) in the presence of fluorite, projected to the log aNds^log aH2O plane (Nds is the Na2Si2O5 component in the pegmatite
melt). Scenarios i^v show five mutually exclusive configurations of stability fields, which are all consistent with a 3D Schreinemakers analysis
of the divariant reactions in Table 4 in log aNds^log aH2O^log aHF space. (b) Two- and three-phase assemblages predicted from the geometrical
configurations in (a). For each matrix, the header row shows single phases expected to be stable (þ) or unstable (^) with fluorite. In the body
of the matrix, pairs of phases that are expected to occur stably with fluorite are indicated by (þ), and combinations that are not expected to
occur by (^). Grey shades represent predicted phase assemblages that have actually been observed in the pegmatites, and black fields ‘forbidden’
combinations that have been found in the samples. Observation of one or more ‘forbidden’ phase assemblages is evidence against that particular
stability field configuration.
2317

Fig. 6. A 3D view of the petrogenetic grid in log aNds^log aH2O^log aHF space showing the phase stability fields at, and immediately below, the
fluorite saturation surface. The 3D stability volume of wo«hlerite (Woe), shown in grey and white stripes, has for illustration been drawn to inter-
sect the fluorite saturation surface (i.e. scenario ii in Fig. 5), although evidence for the stability of a Woe þ Flu assemblage is lacking (see discus-
sion in the text). The hiortdahlite and lvenite fields have been left semi-transparent, to expose the fluorite-absent stability fields involving
these minerals. The orientations of divariant planes below the fluorite saturation surface are indicated by strike-and-dip symbols. At fluorite sat-
uration, the eudialyte field (dark stripes) will be separated from the zircon field by the lvenite, hiortdahlite and wo«hlerite fields, which
wedge out towards lower aHf; this will allow Eud þ Zrc assemblages at aHf well below fluorite saturation, but not a stable Eud þ Zrc þ Flu
assemblage.
planes and univariant lines in 3D log activity space, and of (Fig. 7b and c). The wedge-shaped stability volumes for
the relative positions of the stability volumes of the single wo«hlerite, lvenite and rosenbuschite occur at higher aHF
zirconium silicate minerals. However, it gives no informa- levels, and expand towards increasing aNds, but are trun-
tion on the absolute values of component activities, and cated by the eudialyte volume at elevated HCl activity.
the relative sizes of the divariant fields and the stability The stability volume of hiortdahlite is bounded by the vol-
volumes of single phases are arbitrary. Quantitative cali- umes of zircon (at low aNds), catapleiite (high aH2O), rosen-
bration of the grid would require experimental data on buschite (high aNds), wo«hlerite (low aHF), lvenite (low
critical mineral assemblages, and/or thermodynamic data aH2O) and eudialyte (high aNds and aHCl). The relative
on the zirconium silicate minerals and on components in slopes of the divariant planes limiting the hiortdahlite
the volatile-bearing silicate melt, none of which is volume are such that the volume can never extend to
available. great depths within the fluorite-undersaturated part of the
The chlorine-free subsystem has stability volumes for diagram. Hiortdahlite is therefore stable in nepheline syen-
catapleiite, zircon, wo«hlerite, hiortdahlite, rosenbuschite ite magmas only at or close to saturation with fluorite.
and lvenite. The boundary between the zircon and cata- The same can be said about rosenbuschite, but in low-Cl
pleiite volumes is subparallel to the log aHF axis, and systems, the rosenbuschite volume and the Ros þWoe plane
at very low aHF, these two, and the stability volume of will extend to increasingly greater depths below fluorite
eudialyte are the only ones that need to be considered saturation at increasing aNds.
2318
Fig. 7. Sections through the semiquantitative, 3D petrogenetic grid in the multicomponent system in log aNds^log aH2O^log aHF space, drawn
according to scenario ii in Fig. 5. Zrc, zircon; Cat, catapleiite; Ros, rosenbuschite; Lav, lvenite; Hio, hiortdahlite; Woe, wo«hlerite. Inferred crystal-
lization conditions for the pegmatites in this study are shown as circles and ellipses. Inferred evolution with time is indicated by arrows.
Continuous-line outlines and arrows reflect conditions during magmatic crystallization, dashed-line outlines and arrows post-magmatic,
fluid-induced reactions. The circled fields are linked to the corresponding mineral assemblages in Table 2 by lower-case letters. (a) The fluorite
saturation surface (Fig. 6), projected to the log aNds^log aH2O plane. Thin dashed lines identified by letters enclosed in diamonds give the pos-
itions of the sections shown in Fig. 7b^g. The extent of the eudialyte field is unconstrained; its maximum possible size is given by dash^dot line
E1, which touches at the Flu þ Zrc þ Hio þ Lav point. The shaded field limited by dash^dot line E2 is a more realistic field for Eud þ Flu in the
presence of Ne þ Sod, as it allows observed assemblages involving fluorite, hiortdahlite, rosenbuschite and wo«hlerite within a restricted range of
component activities, as illustrated by assemblages c2, d2, e2, f and g2. At sodalite-undersaturated conditions, the eudialyte field is shifted to
higher aNds (dash^dot line E3). (b, c) log aHF^log aH2O sections through the 3D petrogenetic grid along profiles indicated in (a). The bold line
represents the fluorite saturation surface (i.e. the maximum possible aHF in this system). The shaded field is the assumed extent of the eudialyte
field in the presence of nepheline and sodalite. (d^g) log aNds^log aHF sections, as indicated in (a).
2319
The HCl activity is not tied to the other components by hiortdahlite and catapleiite requires higher aHF than any
mixed volatile reactions. The maximum aHCl is limited by of the other hiortdahlite-bearing mineral assemblages
the simultaneous crystallization of nepheline and sodalite observed.
(reaction (2) in Table 4). In the Cl-bearing system, a The evolution from wo«hlerite-bearing assemblages
volume for eudialyte will appear at high aNds, and expand (assemblages c1, d1, g1) to hiortdahlite þ rosenbuschite
towards lower aNds with increasing aHCl. The non-existence eudialyte (c2, d2, g2) in the pegmatites from
of a stable Zrc þ Eud þ Flu assemblage indicates that the Trompetholmen, Risya and Skutesundskjr suggests a
eudialyte field does not extend as far into the diagram as trend at increasing activities of HF and Nds, ending up at
the Zrc þ Lav þ Hio þ Flu point; the dash^dot line E1 in the Hio þ Ros þ Eud þ Flu point, which may be the

Fig. 7a therefore represents an absolute maximum for the high-alkali endpoint for any nepheline syenite magma
stability field of eudialyte in these pegmatites. However, crystallizing hiortdahlite. The Aryskjrene pegmatite
the existence of assemblages with rosenbuschite þ fluorite also has the Hio þ Ros þ Eud þ Flu assemblage (e2), but
indicates that the real eudialyte stability field is signifi- lacks evidence of the early, crystallization stage at lower
cantly smaller, as suggested by lines E2 and E3. We tenta- aHF.
tively assume that E2 represents the real maximum extent In the Lven pegmatite, an early stage of wo«hlerite þ
of the eudialyte field (i.e. in a system with nephel- lvenite (l1) crystallization is followed by lvenite þ
ine þ sodalite), as it will allow the observed mineral assem- fluorite (l2), combined with alteration of wo«hlerite to
blages with wo«hlerite, rosenbuschite, hiortdahlite low-iron lvenite. Hiortdahlite has not been reported
eudialyte within a relatively small range of component from this pegmatite, so, although the lvenite þ fluorite as-
activities. The predicted four-phase assemblage semblage suggests crystallization at lower aH2O than in
any of the other pegmatites studied, the relationship to
Hio þ Ros þ Eud þ Flu has been observed in the present
the Hio þ Lav þ Flu line in Fig. 7 cannot be determined. A
samples (Table 2). On the other hand, none of the three-
trend towards low water activity could be caused by local
and four phase assemblages involving both lvenite and
crystallization of a hydroxy-annite or other water-bearing
hiortdahlite have been observed, indicating that the
major minerals. The Lven pegmatite has a complex min-
Lav þ Hio þ Flu univariant in Fig. 6 and 7a represents a
eralogy and an uneven distribution of mineral assemblages
minimum limit for the water activity at the magmatic
(Brgger, 1890; Liestl, 1956; Larsen, 2006, 2010). The pres-
stage of crystallization of the hiortdahlite-bearing pegmat-
ence of catapleiite-bearing mineral assemblages at Lven
ites. On the other hand, primary coexistence of lvenite
is very well documented, as the island is the type locality
and fluorite without hiortdahlite has been observed in the for catapleiite (Weibye, 1850). Together, these observations
samples from Lven island, suggesting that the water suggest that this pegmatite had a heterogeneous distribu-
activity was low at least in parts of that pegmatite. tion of volatile activities or a fluid regime that changed
with time, or a combination of the two.
Whereas these findings show diversity of volatile activity
DISCUSSION in the pegmatites, there is also a clear tendency of cluster-
ing around the intersection of the eudialyte field with the
Contrasting crystallization conditions in
hiortdahlite, rosenbuschite and wo«hlerite fields, and a
Oslo Rift nepheline syenite pegmatites change from fluorite-undersaturated liquidus assemblages
The different liquidus mineral assemblages recorded in the with wo«hlerite to fluorite-saturated assemblages (Fig. 7a
nepheline syenite pegmatites reflect differences in water and c). This is most probably due to enrichment of fluorine
and hydrogen fluoride activity and alkali enrichment of in the residual melt caused by growth of fluorine-free
the magma (Fig. 7). In the Langodden pegmatite, the minerals (feldspar, nepheline, clinopyroxene) from a fluor-
inclusion assemblage with wo«hlerite þ hiortdahlite (a1) ine-bearing, but initially fluorite-undersaturated melt.
indicates an early stage of crystallization at fluorite- Increasing residual fluorine led to the stabilization of
undersaturated conditions. The following stages of evolu- wo«hlerite-bearing assemblages, and finally to fluorite-
tion with coexisting hiortdahlite þ zircon þ fluorite (a2) saturated assemblages with hiortdahlite, rosenbuschite or
and zircon þ catapleiite þ fluorite (a3) reflect crystalliza- lvenite (Fig. 7c). The post-magmatic evolution in the peg-
tion at higher aHF and lower aNds, followed by an increase matites has in common a strong decrease in aNds, and
in aH2O. This relative decrease of alkalinity and increase increase in water.
of water has not been seen in other pegmatites examined
in this study. The effect of oxygen fugacity on zirconium
In the Stokksund pegmatite, the hiortdahlite þ fluor- silicate mineral assemblages
ite þ catapleiite mineralogy (assemblage h in Table 2 and Although oxygen fugacity has not been included in the
Fig. 7a) reflects magmatic crystallization at elevated aH2O model calculations, some inferences on the effect of changes
and aHF, as fluorite saturation in the presence of in fO2 can be made. Magnetite and biotite are common
2320
minerals in the pegmatites, and fO2 is likely to be con- The most interesting behaviour is seen within the transi-
trolled by the reaction tional zone within the zircon primary field at low aHF in
Fig. 7. Within this area, the liquidus zirconium silicate min-
Annbi þ 05O2 ¼ Mag þ Kf þ H2 O
eral assemblage of a nepheline syenite magma can be
where Mag is the Fe3O4 component in titanomagnetite. forced to change from miaskitic with stable zircon, to ‘low
This is not a formally univariant oxygen buffer curve, as agpaitic’ assemblages with different successions of wo«hler-
it depends on mineral compositions and water activity, ite-, hiortdahlite- and lvenite-bearing assemblages by a
but as long as biotite and titanomagnetite remain stable, it simple increase in aHF without simultaneous changes in
will prevent wide excursions in oxygen fugacity level. This either water or alkali activity.

suggests that the pegmatites have crystallized at mildly This suggests that there are at least three liquid lines of
reducing conditions around QFM ^ 1, corresponding to descent leading from a miaskitic to an agpaitic crystalliza-
those of the host larvikite (Neumann, 1976). tion regime in peralkaline nepheline syenite magmas: (1)
As long as the major liquidus mineralogy of the pegmat- the classical case of increasing peralkalinity, leading to
ite magma remains unchanged, the relative stability rela- eudialyte-dominated mineral assemblages, and from
tionships of the iron-poor Zr-silicate minerals are likely to there, possibly, to hyperagpaitic residual liquids (e.g.
be little affected. The exception may be lvenite, which Andersen & Srensen, 2005); (2) the increasing water
has essential Fe2þ and Mn2þ. At increasing oxygen fuga- path, leading to the stabilization of catapleiite-bearing
city, the lvenite stability field will decrease, and lvenite liquidus assemblages without eudialyte, and possibly, to
will be replaced by zircon þ magnetite fluorite symplec- saturation with an aqueous fluid phase; (3) the increasing
tites (Fig. 2k). Increasing oxygen fugacity would also be ex- fluorine trend, leading to hiortdahlite-, lvenite- and
pected to increase the stability fields of hiortdahlite and wo«hlerite-dominated mineral assemblages, and eventually
rosenbuschite to lower water activity, at the expense of to crystallization of magmatic fluorite. Increasing fluorine
lvenite. (Fig. 6). Oxidation of lvenite took place at a late concentration in the melt may cause the degree of poly-
magmatic to hydrothermal stage in the pegmatites, but at merization of the magma to decrease (Mysen & Virgo,
the magmatic stage, the opposite relationship is observed: 1985).
hiortdahlite and wo«hlerite were replaced by low-iron For transitionally agpaitic magmas forming the Oslo
lvenite, and eudialyte by low-iron lvenite-bearing Rift pegmatites, the combination of a relatively low aNds
symplectite. This suggests an overall decreasing trend in in the parent magma and the elevated fluorine content
oxygen fugacity towards the final stages of magmatic crys- leading to the stabilization of magmatic fluorite causes a
tallization in the pegmatites. barrier against evolution towards the more extreme com-
positions seen in highly agpaitic rocks such as the lujavrites
of the Il|¤ maussaq intrusion. The alkali enrichment of a
Petrological implications magma crystallizing a low-Na assemblage with
The factors controlling the change from a miaskitic crys- hiortdahlite þ fluorite will terminate at the Ros þ Hio þ
tallization regime with zircon to one with agpaitic zirco- Flu or Eud þ Hio þ Flu field boundaries, and the four-phase
nium silicate assemblages can be illustrated by a constant assemblage Ros þ Eud þ Hio þ Flu may represent the high-
aHF section through log aNds^log aH2O^log aHf space at est attainable aNds in a such a fluorite-saturated
low HF activity, below the point of coexisting wo«hlerite, system. Further evolution to the extreme alkali enrichment
catapleiite, zircon and eudialyte in Fig. 7b (Fig. 8). A per- permitting crystallization of mineral assemblages with
alkaline nepheline syenite magma crystallizing within the villiaumite, ussingite and other minerals typical of hyper-
primary field of zircon may change to a liquidus assem- agpaitic rocks (Khomyakov, 1995; Srensen, 1997) is impos-
blage with eudialyte at increasing aNds, in agreement with sible or at least highly unlikely in such a magma.
the traditional view of the miaskitic to agpaitic transition Separation and loss of an aqueous fluid would deplete
in peralkaline nepheline syenites (e.g. Larsen & Srensen, the system in soluble components, mainly Na, K and Cl,
1987). In contrast, the presence of primary catapleiite indi- but also to some extent F, as shown by the existence of
cates elevated water activity, rather than elevated sodium late, crosscutting fluorite veins in many of the pegmatites.
silicate activity. It should be noted, however, that the shift Loss of such a fluid can explain the formation of less alka-
from zircon- to catapleiite-dominated liquidus mineralo- line secondary assemblages in some of the pegmatites (e.g.
gies is possible only in highly alkaline systems; at lower assemblages d3 at Risya, k2 at Lven). On the other
alkali activities, water saturation may be reached before hand, interaction between an alkali-chloride enriched,
catapleiite can form. Within the white part of the zircon mobile fluid phase and less alkaline rocks could lead to
primary field in Fig. 8, a magma will remain miaskitic at the formation of metasomatic agpaitic mineral assem-
increasing aHF, and the distinctive mineral assemblages blages in otherwise miaskitic rocks, as has been observed
will be zircon followed by zircon þ fluorite. in, for example, Tamazeght (e.g. Schilling et al., 2009).
2321

Fig. 8. A section through 3D log aNds^log aH2O^log aHF space at constant aHF well below the point of Woe þ Cat þ Zrc þ Eud coexistence in Fig. 7
showing the ‘agpaitic’ (grey), ‘transitional’ (ruled) and ‘miaskitic’ (white) fields of crystallization in peralkaline nepheline syenite systems.
Lines E1 and E2 are ‘absolute’ and ‘realistic’ maximum limits for the eudialyte field, taken from Fig. 7a. In the miaskitic field, zircon remains
stable at any aHF. Eudialyte is stable only at elevated aNds, and an increase in the peralkalinity of the system is needed to stabilize it at the ex-
pense of zircon or catapleiite at any level of aHF. Within the transitional field, a magma crystallizing zircon at low aHF can change to transition-
ally agpaitic phase assemblages with hiortdahlite, wo«hlerite or lvenite by a simple increase in aHF without a concomitant increase in
peralkalinity. Likewise, within large parts of the miaskitic and ‘transitional’ fields at low aHF, the phase assemblage can change to a nominally
agpaitic assemblage with catapleiite as the characteristic zirconium mineral by an increase in water activity with or without increasing peralk-
alinity. An approximate limit for water saturation is drawn to allow a free, aqueous fluid phase to develop within the stability field of catapleiite,
agreeing with the observation of abundant fluid inclusions in this mineral. This boundary will intersect the Cat þ Zrc plane at low aNds, prevent-
ing the stabilization of catapleiite as a liquidus mineral in less peralkaline magmas.
Metasomatic eudialyte has not been observed in any of the and Khibina (Konnerup-Madsen, 2001; Ryabchikov &
present samples, but crosscutting catapleiite microveins in Kogarko, 2009, and references therein). The presence of
the TYHIO sample (Fig. 2c, i.e. assemblage a3) may have such fluids is a consequence of the low oxygen fugacity
a similar origin. and low solidus temperature of these magmatic systems
Hydrocarbon-bearing fluids have formed during crystal- (Ryabchikov & Kogargo, 2009), rather than a controlling
lization of large, agpaitic intrusions such as Il|¤ maussaq factor for the stability of zirconium silicate mineral
2322
stability. Whether or not organic fluids were present in the concentration-dependent parameters in these highly
nepheline syenite pegmatites of the Larvik Plutonic evolved magmas.
Complex is not known. There are at least three types of process that may change
the liquidus assemblage of a nepheline syenite magma
Wider implications from miaskitic with zircon to agpaitic with complex zirco-
In the absence of thermodynamic data for the species nium silicate minerals: (1) an increase in peralkalinity,
involved in phase equilibria, the method developed here is leading to crystallization of eudialyte; (2) increasing water
semiquantitative. Its main importance is that differences activity (and eventually, saturation with an aqueous fluid
and trends in peralkalinity and volatile activities in a crys- phase), leading to the stabilization of catapleiite-bearing

tallizing nepheline syenite magma can be detected from mineral assemblages; (3) increasing fluorine (or HF) activ-
relatively simple identification of the liquidus mineral as- ity, stabilizing liquidus assemblages with wo«hlerite, hiort-
semblages. The stability diagram in Figs 6 and 7 has been dahlite, lvenite fluorite.
developed for one specific suite of mildly agpaitic rocks, The findings of this study confirm that the ‘low agpaitic’
characterized by high Zr/Ti silicates of the wo«hlerite, or ‘transitional’ magmas forming the nepheline syenite
eudialyte and rosenbuschite groups, which has formed at pegmatites in the Oslo Rift represent a distinct style of
mildly reducing conditions controlled by magnetite^bio- magmatic evolution, in which increasing fluorine (or HF)
tite^potassium feldspar equilibria. Except for eudialyte activity is as important as an elevated (Na þ K)/Al ratio
s.l., the zirconium silicate minerals show restricted com- for the stabilization of complex Na^Ca^Zr silicate liquidus
positional variation. The results of this study may therefore minerals, and thus for the transition from a miaskitic to
be applied directly to other examples where these minerals an agpaitic crystallization regime. A peralkaline nepheline
occur in rocks with the same major mineralogy (i.e. micro- syenite magma crystallizing liquidus hiortdahlite or
cline þ albite þ nepheline þ aegirineâugite þ biotite þ lvenite together with fluorite may eventually also crystal-
titanomagnetite sodalite fluorite). The underlying lize eudialyte s.l. if chlorine or HCl activity is sufficiently
principle (i.e. chemographic analysis based on observed as- high, but it is likely to terminate at liquidus boundaries or
semblages of coexisting minerals and mineral compos- low-variance points involving these minerals and rosen-
itions determined by electron microprobe) is universally buschite, and not to evolve further towards hyperagpaitic
applicable. Within the field of agpaitic rocks, the method residual melt compositions.
may be used to identify factors controlling the relative sta-
bility of high-Ti assemblages with mosandrite, lorenzenite,
rinkite, astrophyllite and other minerals that are reason- AC K N O W L E D G E M E N T S
ably common in agpaitic rocks, and to evaluate the param- Thanks are due to Hans-Jrgen Berg for help with X-ray
eters controlling the transition into the hyperagpaitic diffraction work, and to Ulf Hlenius and Jo«rgen Langhof
realm; for example, in the lujavrites of Il|¤ maussaq. Other for access to the collections of the National Museum of
chemical variables may be introduced for rocks where Natural History, and for practical assistance with the
they are relevant (e.g. CaO for high-Ca agpaitic rocks), as study of the material in the Stockholm collection. Helpful
may oxygen fugacity, provided the analysis is properly suggestions from Henrik Friis, Per Aagaard and Marlina
supported by observed mineral assemblages. Elburg, and comments from editor Ron Frost and review-
ers Michael Marks, Roger Mitchell and Henning
Srensen are gratefully acknowledged. Gregor Markl and
CONC LUSIONS Katharina Pfaff kindly provided their Excel spreadsheet
In the nepheline syenite pegmatites associated with larvi- for eudialyte structural formula calculation. Special
kite and nepheline syenite in the Larvik Plutonic thanks go to Henning Srensen, who introduced the first
Complex in the Oslo Rift, Na^Ca^Zr^F silicate minerals author to the fascinating world of agpaitic rocks.
of the wo«hlerite and rosenbuschite groups form part of the
liquidus mineral assemblage of highly evolved, peralkaline
and volatile-enriched nepheline syenite magmas, together FU N DI NG
with the major rock-forming minerals (feldspar, alkali-rich This study was supported by grants from the ‘‘Smforsk’’
pyroxene and amphibole, biotite, nepheline, sodalite), program of the Department of Geosciences, University of
fluorite and eudialyte s.l. Because the structures of the Oslo to T.A. in 2008 and 2009.
wo«hlerite- and rosenbuschite-group minerals allow only
restricted compositional variation within single species,
minor changes in magma composition or volatile activities S U P P L E M E N TA RY DATA
may lead to stabilization of distinct zirconium silicate Supplementary data for this paper are available at Journal
mineral assemblages, which are sensitive indicators of of Petrology online.
2323
Larsen, A. O., —sheim, A. & Gault, R. A. (2005b). Ferrokentbrooksite

R EF ER ENC ES (Brgger’s xanthopite) from Skutesundskjr in the
Andersen, T. & Srensen, H. (2005). Stability of naujakasite in Langesundsfjord district, Norway. Bergverksmuseets Skriftserie 30,
hyperagpaitic melts and the petrology of naujakasite lujavrite in 63^67.
the Il|¤ maussaq alkaline complex, South Greenland. Mineralogical Larsen, B. T., Olaussen, S., Sundvoll, B. & Heeremans, M. (2008). The
Magazine 69, 125^136. Permo-Carboniferous Oslo Rift through six stages and 65 million
Baker, D. R., Conte, A. M., Freda, C. & Ottolini, L. (2002). The effect years. Episodes 31(1), 52^58.
of halogens on Zr diffusion and zircon dissolution in hydrous meta- Larsen, L. M. & Srensen, H. (1987). The Il|¤ maussaq intrusionçpro-
luminous granitic melts. Contributions to Mineralogy and Petrology 142, gressive crystallization and formation of layering in an agpaitic
666^678. magma. In: Fitton, J. G. & Upton, B. G. J. (eds) Alkaline Igneous
Bellezza, M., Merlino, S. & Perchiazzi, N. (2009). Mosandrite:

Rocks. Geological Society, London, Special Publications 30, 473^488.
Structural and crystal-chemical relationships with rinkite. Le Maitre, R. W. (ed.) (2003). Igneous Rocks. A Classification and Glossary
Canadian Mineralogist 47, 897^908. of Terms. Cambridge: Cambridge University Press, 236 p.
Brgger, W. C. (1890). Die Mineralien der Syenitpegmatitga«nge der Liestl, G. B. (1956). Noen petrografiske og mineralogiske
su«dnorwegischen Augit- und Nephelinsyenite. Zeitschrift fu«r underskelser omkring pegmatittgangene i Langesundsfjorden.
Krystallographie 16, 1^235 þ1^663. MSc thesis, Hovedoppgave i mineralogi med petrografi,
Christiansen, C. C., Johnsen, O. & Makovicky, E. (2003). Crystal chem- University of Oslo.
istry of the rosenbuschite group. Canadian Mineralogist 41,1203^1224. Mariano, A. N. & Roeder, P. L. (1989). Wo«hlerite: Chemical compos-
Dahlgren, S. (2010). The Larvik Plutonic Complex: The larvikite and
ition, cathodoluminescence and environment of crystallization.
nepheline syenite plutons and their pegmatites. In: Larsen, A. O.
Canadian Mineralogist 27, 709^720.
(ed.) The Langesundsfjord. History, Geology, Pegmatites, Minerals.
Munoz, J. L. (1984). FÔH and CIÔH exchange in micas with appli-
Salzhemmerdorf, Germany: Bode, pp. 26^37.
cation to hydrothermal ore deposits. In: Bailey, S. W. (ed.) Micas.
Friis, H., Balic¤-Z›unic¤, T., Williams, T. & Petersen, O. V. (2010).
Mineralogical Society of America, Reviews in Mineralogy 13, 469^491.
Minerals of the lvenite group from South Greenland and
Murad, E. (2006). Mineralogy of aegirine from Lven island,
Norway. Norsk Bergverksmuseum Skrift 43, 35^39.
Langesundsfjorden, southern Norway. Norwegian Journal of Geology
Jarosewich, E. & Boatner, L. A. (1991). Rare-earth element reference
86, 435^438.
samples for electron microprobe analysis. Geostandards Newsletter 15,
Mysen, B. O. & Virgo, D. (1985). Structure and properties of
397^399.
fluorine-bearing aluminosilicate melts: the system Na2OÂl2O3^
Johnsen, O. & Grice, J. D. (1999). The crystal chemistry of the eudia-
SiO2^F at 1atm. Contributions to Mineralogy and Petrology 91, 205^220.
lyte group. Canadian Mineralogist 37, 865^891.
Neumann, E.-R. (1976). Compositional relations among pyroxenes,
Johnsen, O., Ferraris, G., Gault, R. A., Grice, J. D., Kampf, A. R. &
amphiboles and other mafic phases in the Oslo Region plutonic
Pekov, I. V. (2003). The nomenclature of eudialyte-group minerals.
rocks. Lithos 9, 85^109.
Canadian Mineralogist 41, 785^794.
Neumann, E.-R. (1980). Petrogenesis of the Oslo Region larvikites and
Keller, J., Williams, C. T. & Koberski, U. (1995). Niocalite and wo«hler-
associated rocks. Journal of Petrology 21, 498^531.
ite from the alkaline and carbonatite rocks at Kaiserstuhl,
Neumann, E.-R., Wilson, M., Heeremans, M., Spencer, E. A.,
Germany. Mineralogical Magazine 59, 561^566.
Obst, K., Timmerman, M. J. & Kirstein, L. (2004).
Khomyakov, A. P. (1995). Mineralogy of Hyperagpaitic Alkaline Rocks.
Carboniferous^Permian rifting and magmatism in southern
Oxford: Oxford Scientific Publications, Clarendon Press, 222 p.
Scandinavia and northern Germany: a review. In: Wilson, M.,
Kogarko, L. N. (1974). Role of volatiles. In: Srensen, H. (ed.) The
Neumann, E.-R., Davies, G. R., Timmerman, M. J.,
Alkaline Rocks. London: John Wiley, pp. 474^487.
Konnerup-Madsen, J. (2001). A review of the composition and evolu- Heeremans, M. & Larsen, B. T. (eds) Permo-Carboniferous
tion of hydrocarbon gases during solidification of the Il|¤ maussaq Magmatism and Rifting in Europe. Geological Society, London, Special
alkaline complex, South Greenland. Geology of Greenland Survey Publications 223, 11^40.
Bulletin 190, 159^166. Petersen, J. S. (1978). Structure of the larvikite^lardalite complex,
Larsen, A. O. (1996). Rare earth minerals from the syenite pegmatites Oslo Region, Norway, and its evolution. Geologisches Rundschau 67,
in the Oslo Region, Norway. In: Jones, A. P., Wall, F. & 330^342.
Williams, C. T. (eds) Rare Earth Minerals. Chemistry, Origin and Ore Pfaff, K., Wenzel, T., Schilling, J., Marks, M. A. & Markl, G. (2010).
Deposits. London: Chapman & Hall, pp. 151^166. A fast and easy-to-use approach to cation site assignment for
Larsen, A. O. (1998). Identiteten til sorte glimmere i biotittserien fra eudialyte-group minerals. Neues Jarhbuch fu«r Mineralogie,
syenittpegmatitter i Langesundsfjordomrdet. Bergverksmuseets Abhandlungen 187, 69^81.
Skriftserie 14, 5^8. Pouchou, J. L. & Pichoir, F. (1984). A new model for quantitative
Larsen, A. O. (2001). Chemical composition of catapleiites from the X-ray microanalysisI. Application to the analysis of homogeneous
syenite pegmatites in the Larvik plutonic complex, Norway. samples. Recherche Ae¤ rospatiale 3, 13^38.
Bergverksmuseets Skriftserie 18, 5^9. Roedder, E. (1984). Fluid Inclusions. Mineralogical Society of America,
Larsen, A. O. (2006). Lven, et fredet naturminne. Bergverksmuseets Reviews in Mineralogy 12, 644 p.
Skriftserie 33, 37^40. Ryabchikov, I. D. & Kogarko, L. N. (2009). Oxygen fugacity in the
Larsen, A. O. (2010). The Langesundsfjord. History, Geology, Pegmatites, apatite-bearing intrusion of the Khibina Complex. Geochemistry
Minerals. Salzhemmerdorf, Germany: Bode, 239 p. International 47, 1157^1169.
Larsen, A. O. & Raade, G. (1997). Pyroksener fra Oslofeltets syenitt- Salvi, S. & Williams-Jones, A. E. (1995). Zirconosilicate phase rela-
pegmatitter. Bergverksmuseets Skriftserie 12, 16^21. tions in the Strange Lake (Lac Brisson) pluton, Quebec^
Larsen, A. O., —sheim, A. & Gault, R. A. (2005a). Minerals of the Labrador, Canada. American Mineralogist 80, 1031^1040.
eudialyte group from the Sagsen larvikite quarry, Porsgrunn, Schilling, J., Marks, M. A. W., Wenzel, T. & Markl, G. (2009).
Norway. Bergverksmuseets Skriftserie 30, 58^62. Reconstruction of magmatic to subsolidus processes in an agpaitic
2324
system using eudialyte textures and composition: a case study from Watson, E. B. & Harrison, M. T. (1983). Zircon saturation revisited:
Tamazeght, Morocco. Canadian Mineralogist 47, 351^365. temperature and composition effects in a variety of crustal magma
Sokolova, E. (2006). From structure to chemical composition. I. types. Earth and Planetary Science Letters 64, 295^304.
Structural hierarchy and stereochemistry in titanium disilicate Weibye, P. H. (1850). Neue Mineralien aus Norwegen, beschrieben
minerals. Canadian Mineralogist 44, 1273^1330. von P. H. Weibye; analysiert von N. J. Berlin, K. A. Sjo«gren und J.
Srensen, H. (ed.) (1974). The Alkaline Rocks. London: John Wiley, B. von Borck (Erster Theil). Annalen der Physik und Chemie 79,
622 p. 299^304.
Srensen, H. (1997). The agpaitic rocksçan overview. Mineralogical White, R. W., Powell, R. & Baldwin, J. A. (2008). Calculated phase
Magazine 61, 485^498. equilibria involving chemical potentials to investigate the textural
Srensen, H. (2006). The Il|¤ maussaq Alkaline Complex, South evolution of metamorphic rocks. Journal of Metamorphic Geology 26,
Greenlandçan overiview of 200 years of research and an outlook. 181^198.

Meddelelser om Grnland, Geoscience 45, 1^70. Zen, E-a. (1966). Construction of Pressure^Temperature Diagrams for
Strunz, H. & Nickel, E. H. (2001). Strunz Mineralogical Tables, 9th edn. Multicomponent Systems after the Method of SchreinemakersçA Geometric
Stuttgart: Schweizerbart, 870 p. Approach. US Geological Survey Bulletin 1225, 56 p.
Ussing, N. V. (1912). Geology of the country around Julianehaab,
Greenland. Meddelelser om Grnland 38, 1^376.
2325

Nepheline Syenite Pegmatites in Langesundsfjord

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Nepheline Syenite Pegmatites in Langesundsfjord

Hochgeladen von

Copyright:

Verfügbare Formate

JOURNAL OF PETROLOGY VOLUME 51 NUMBER 11 PAGES 2303^2325 2010 doi:10.

Petrology of Nepheline Syenite Pegmatites in the

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

RECEIVED APRIL 7, 2010; ACCEPTED SEPTEMBER 7, 2010

The Author 2010. Published by Oxford University Press. All

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Locality Fig. 11 Sample Mineral assemblages

Magmatic stage Late or secondary stage

Langodden 3 TYHIO a1: Wöhlerite þ hiortdahlite

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

that a minor error in the Si analysis may propagate to a

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

analysis 24 in Electronic Appendix 3. (a) Na^Nb^Ca/5 (Keller et al.,

Ti0·3Zr1·8Nb0·2(Si4O14)F3·5O0·5, and higher Fe/(Fe þ Mn) 3M(2)3M(3)M(4)Z3[Si24O72]O4X2, in which X ¼Cl, F, OH

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Major minerals in pegmatite

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Larsen, A. O., —sheim, A. & Gault, R. A. (2005b). Ferrokentbrooksite

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Downloaded from https://academic.oup.com/petrology/article-abstract/51/11/2303/1428765 by guest on 07 November 2018

Das könnte Ihnen auch gefallen