MINERALOGICAL MAGAZINE

VOLUME 60 NUMBER 401 AUGUST 1996

The conditions of metamorphism of a grossular-

wollastonite vesuvianite skarn from the Omey
Granite, Connemara, western Ireland, with
special reference to the chemistry of vesuvianite

Y. AHMED-SAID*AND BERNARDE. LEAKE

Department of Geology and Applied Geology, University of Glasgow, Glasgow G12 8QQ, UK

Abstract
The Fountain Hill skarn, which was produced by thermal metamorphism of impure limestone in the Omey
Granite aureole, consists of wollastonite, calcite, grossular-andradite garnet, diopside, B-free vesuvianite and
small amounts of albite, K-feldspar and quartz. Zoning in vesuvianite is, overall, independent of birefringence
but oscillatory concentric zoning in the mineral is controlled mainly by Ti. In addition to extensive within-site
substitutions (e.g. F for OH in the OH-sites, Fe2+ for Mg, Fe3+ for A1, Ti for AI etc, in the Y-sites), there are
significant cross-site combined substitutions involving Y - X and Y - Z but not X - Z sites so that ideal solution
models for this mineral are not applicable. The thermodynamic mole fraction of Hoisch (1985) is modified to
account for the excess of the Y.Y cations in the Y sites and can be applied to both B-bearing and B-free
vesuvianites. Using the thermodynamic dataset of Holland and Powell (1990) and taking into account the
existence of andalusite, sillimanite and corundum in associated pelites, leads to the conclusion that the
metamorphic conditions were about 640-t-20~ and 3.3 __+0.3 kb at 0.15__+0.05 Xco2-

KEYWORDS:activity, grossular, hornfels, thermodynamic mole fraction, vesuvianite, wollastonite.

Introduction morphic aureole. The earlier regional metamorph-

isms in the Dalradian rocks to the SE of the granite
TRE circular 6.5 km diameter Omey Granite in produced quartz-plagioclase-biotite-muscovite-
western Connemara, Co. Galway, Ireland, which was garnet-staurolite with rare kyanite followed by
described by Townend (1966) and dated at about 415 fibrolite sillimanite in the pelites. The Omey
Ma by Leggo et al. (1966), intrudes the Dalradian aureole developed andalusite, cordierite, new
Connemara schists with development of a meta- biotite, and microcline in pelites up to 1400 m from
the granite with prismatic sillimanite and corundum
in the inner aureole (< 250 m from the contact) and in
* Previous address: Oum-Toub (Skikda), P.O. Box 56, marbles garnet-wollastonite-calcite up to 470 m from
21450, Algeria the granite (Ferguson and Harvey, 1979; Ferguson
Mineralogical Magazine, August 1996, Vol. 60, pp. 541-550
9 Copyright the Mineralogical Society
542 Y. AHMED SAID AND B. E. LEAKE

TABLE 1. Mineral chemistry

Gamet Diopside Plagioclase Wollastonite Vesuvianite

Oxides maximum a ~ s
rim core

SiOz 38.55 39.56 52.59 0.28 67.39 0.65 51.41 0.23 35.83 35.65
TiOE 0.76 0.15 0.06 0.01 0.04 0.016 2.76 0.46
A1203 16.30 19.75 0.70 0.18 20.73 0.56 0.06 0.06 16.92 18.97
Fe203 8.59 3.81 1.15 0.07 0.28 0.08 1.72 1.51
FeO 1.70 0.41 6.27 0.39 T 2.02 1.77
MgO 0.05 0.01 13.23 0.34 1.89 1.83
MnO 0.30 0.24 0.48 0.32 0.13 0.05 0.06 0.02
Cat 34.88 36.50 24.83 0.04 0.07 0.02 45.5 0.26 36.15 35.12
NaEO 0.17 0.03 10.84 0.78 0.16 0.09
KEO 0 0.06 0.02 F 1.56 2.06
Total 101.13 100.47 99.48 99.09 97.42 99.53 97.83

Cations on 24 oxygens for garnet, 6 for diopside, 32 for feldspars, and 18 for woUastonite and 50 cations for
vesuvianite

Si 5.934 6.001 1.973 0.004 11.846 0.132 5.999 0.013 17.438 17.536
AITM 0.066 - 0.027 0.005 4.295 0.105 0.014 0.013 0.562 0.440
AIvl 2.842 3.530 9.143 10.57
Ti 0.088 0.170 0.0012 0.0002 0.003 0.001 1.010 0.170
Fe 3§ 0.996 0.434 0.033 0.005 0.025 0.007 0.630 0.560
Fe 2§ 0.218 0.052 0.197 0.015 0.822 0.729
Mg 0.012 0.002 0.740 0.021 1.371 1.344
Mn 0.040 0.03 0.015 0.010 0.013 0.005 0.025 01008
Ca 5.758 5.936 0.998 0.004 0.013 0.004 5.940 0.009 18.848 18.532
Na 0.012 0.002 3.691 0.252 0.0012 0.150 0.086
K 0.011 0.005
F - 2.40 3.20
0 2 - 1.0 1.0
O H - 6.60 5.62

The minerals were analysed in the Department of Geology and Applied Geology, University of Glasgow using a
Camebax SX 50 microprobe. All analyses were performed with an accelerating voltage of 15 kv (30 kv for trace
elements), sample current of 5 ~tA, counting time of 1 0 - 3 0 s using a beam of 5 txm in diameter. ZAF corrections
were applied to all apparent concentrations. ~ is the average of four minerals with four analysed spots per mineral,
Fe 2§ and Fe 3§ are calculated on charge balance basis except for vesuvianites where Fe 2+ = 0.54Fe 3+ (see text for
more details), T = trace. Vesuvianites contain B 0.00, 0.00; SO2 0.01, 0.13; ZnO 0.00, 0.00; C1 0.00, 0.02.

and A1-Ameen, 1985). The hornfelsing of the pelites good exposures occur at Fountain Hill beach on the
has been studied in some detail by the latter authors south side of the Omey Granite about 1 4 0 - 1 5 0 m
who deduced conditions of 6 1 5 + 2 5 ~ at 2.5-1-0.25 south of the granite contact. The marble bands in the
kbar and with localized flow of an HzO-rich fluid Lakes Marble Formation are generally calcite
away from the granite but the skams and their marbles following the regional metamorphism of
minerals have not been previously described in the Dalradian rocks as distinct from the green
detail. New understanding of the role of boron in ophicalcitic 'Connemara Marble' used ornamentally,
vesuvianite (Groat et al., 1992a) also makes the study which represents a metamorphosed siliceous dolts-
of assemblages including this mineral more profit- tone and has been described in detail by Leake et al.
able than previously. (1975). The Lakes Marbles regionally contain quartz,
The skams occur where marble in the Lakes phlogopite, titanite, diopside and K-feldspar with
Marble Formation (Leake and Tanner, 1994) is voluminous calcite and were presumably derived
within 300 m from the granite margin and especially from original sedimentary quartz, clay minerals
 WOLLASTONITE-VESUVIANITE SKARN 543

(especially illite) and Mg-bearing calcite in a Y-group total (Y.Y) significantly higher than the ideal
limestone and the "impurities" reacted with calcite value of 13 and hence they recalculated the mineral
to give the skarn silicates during partial decarbona- formula on the basis of 18 Si instead. Four of this
tion of the calcite. "small number of exceptional vesuvianite crystals"
were further refined by these authors and their results
showed the Z-sites to be exclusively occupied by Si
Mineralogy
which lead them to suggest an additional tetra-
Near the contact, the rocks consist of wollastonite, hedrally coordinated site (T1). In B-free vesuvianites,
calcite, diopside, garnet, vesuvianite, and small these T(1) sites are either vacant or partly occupied
amounts of albite and K-feldspar with accessory by the excess (A1, Fe 3§ which lead to a new revised
zircon. Garnet occurs as both poikiloblastic grains formula X19Y13T(1)sZ18068WI o where T(1) = A1, Fe 3+
(usually 5 - 1 0 mm in diameter but up to 30 mm) in excess of the 13 cations in the Y sites (see below).
containing calcite, diopside, wollastonite and quartz We have renormalized the four "exceptional
inclusions or less commonly as inclusion-free vesuvianites" studied by Groat et al. (1994a) on the
euhedral crystals 5 mm in diameter. Like garnet, basis of 50 cations and plotted them with many B-
vesuvianite, which is brown in hand specimen and free vesuvianites in Fig. 1. There are a number of
colourless to slightly pinkish in thin section, is also interesting features: (1) the four "exceptional
found as euhedral and anhedral grains; the latter are vesuvianites" have Si = 17.92+0.13 a.p.f.u. (atoms
most common and usually replace calcite and garnet, per formula unit) which is statistically the same to the
and these two minerals together with wollastonite ideal value of 18 a.p.f.u. The same applies to the
and diopside are also found as inclusions within the vesuvianites of Hoisch (1985) and Groat et al.
vesuvianite. The euhedral vesuvianites, 6 mm in (1992a, 1994a, 1995). (2) There is no correlation
diameter, are inclusion-free and replaced plagioclase. between A1 and Si (R = 0.00) for vesuvianites
Diopside (<1 mm) occurs as isolated grains or as published by these authors either as separate groups
inclusions in anhedral garnet and vesuvianite. or all together, thus excluding any A1 for Si
Wollastonite, which sometimes replaced garnet, substitutions. (3) EY-cations>13 a.p.f.u, for all
vesuvianite, diopside and calcite, is found as radial vesuvianites shown in Fig. 1 except four samples
aggregates with big elongated crystals displaying published by Groat et al. (1995).
simple twinning. Albite is scarce but calcite is The Connemara vesuvianites are strikingly
ubiquitous. Chemical analyses of the minerals are different; they are generally more aluminous
given in Table 1.
Both garnet types exhibit the same zoning pattern,
having Ti- and Fe-rich cores without significant 12
variations of Mn, Mg and Ca. The maximum core [+ Hoisch (1985)
[• Groat et al. (1992)
and rim compositions are Gross87.3oAndll.68 ]9 Groat etal. (1994)
]A Groatetal.(1995) [
Spo.5oAlmo.a9Pyro.o3 and Gross68.2oAnd2s.55Alm2.41 /O Frimmel and van ]
Spo.66PyroA8 respectively. Plagioclase is unzoned 11
albite containing variable amounts of Sr (0-2198 o0~ 1 7 6 A
ppm) with only little Ba (Ba = 0 - 3 0 0 ppm).
Wollastonite is also unzoned, does not form solid
solutions with pyroxene, and is an almost pure m
~ e ~ A
calcium silicate with only small substitutions of Fe 10
and Mn for Ca. Sr is detected in all analysed
o
wollastonites and averages 761--+307 ppm (range
507-1353 ppm). Diopside is u n i f o r m ++~+ ++
Ca49.9oFe13.1Mg37.oo with 676 ppm Sr and 0.11 []
wt.% ZnO.
9 []
Recent structural refinements of vesuvianite
(Hoisch, 1985; Valley et al., 1985; Groat et al.,
1992a, b, 1994a, b , 1995) suggest that the formula has 8
the basis of 50 cations. This normalization scheme 17.2 1 7 . 4 17.6 1 7 . 8 1 8 . 0 1 8 . 2 18.4
agrees well with the number of available sites in the I o n s Si
ideal structure of B-free vesuvianites based on the
formula X19Y13ZlsO68Wlowhere X = Ca, Na, R E E , FIG. 1. A1 vs. Si showing AI for Si substitution in the
Pb 2§ Sb3+; Y= AI, Mg, Fe 3+, Fe 2+, Ti 4+, Mn, Cu, Zn; Connemara vesuvianites (open circles). Note also the
Z = Si and W = OH, F, O. However, Groat et al. broad negative correlation between the two ions for the
(1994a) noted that a few B-free vesuvianites have a analyses as a whole.
544 Y. AHMED SAID AND B. E. LEAKE

20 Kerrick et al., 1973; Gibson and Wallmach, 1995).

Traverses through the mineral allowed direct
18
substitutional studies (Fig. 2; Table 2a; see also
16 Table 2b for a summary of statistics of the analysed

14 .1= ol mineral). Table 2a shows very important facts:

(1) EX- and EZ-cations do not correlate (R=0.05)
12 excluding any interactions a m o n g the cations
occupying these sites
Z
10 (2) EY- and EZ-cations e x h i b i t a n e g a t i v e
8 correlation (R --- - 0 . 1 3 ) which is principally due to
2.5 A1 and to a lesser extent Ti substitutions for Si
O (R = - 0 . 1 1 and - 0 . 0 9 respectively).
2.0 (3) EY- and EX-cations display a strong anti-
pathetic correlation ( R = - 0 . 6 7 ) indicating strong
1.5 interactions among the various X and Y sites,
probably through combined substitutions.
1.0 It is, however, relevant to establish substitutions in
each site first. Thus, Table 2a shows that there is no
0.5 significant one to one Na substitution for Ca
0.0 ( R = - 0 . 0 5 ) within the X sites but combined substitu-
tions involving the unanalysed REE, which preferably
occupy the X sites, are possible (Groat et al., 1992a,b).
FIG. 2. A traverse through a 2.5 mm zoned vesuvianite There are a large number of substitutional possibilities
(ions based on 50 cations). Stipple areas are exaggerated among the Y cations but these are dominated by Ti for
high-birefringence zones. A1 replacements. These include: Fe 2+ = Mg2+; Fe 3§ =
A13+; Fe 2+ +Tin+ = 2A13+; Mg 2+ +Tin+ = 2AI 3+ and
possibly (Groat et al., 1992); H § + AI 3§ = Tin§ Mg 2§ +
(A1 = 10.24___0.50 a.p.f.u.), less siliceous (Si = O H - = A13§ + 0 2 - ; Fe 2§ + O H - = A13§ + 0 2 - ; Tin+ +
17.58___0.07 a.p.f.u.) and more importantly, A1 and 0 2 - = A13§ + OH-. Substitutions involving M n are
Si correlate antipathetically, suggesting continuous less clear due to its low concentrations in the mineral
AI for Si substitutions. Particularly relevant is that, (Mn = 0.02-t-0.01 a.p.f.u.) but positive correlation
with the exception of the four vesuvianites mentioned between Mn and Mg suggests that both elements
above, all vesuvianites shown in Fig. 1 have XY follow similar trends rather than substitute for one
cations in excess of 13 which make it a common another. In the Z sites: Tin+ = Sin+ and A13+ + H § =
rather than an exceptional feature. The Connemara Sin+. Many possibilities exist for X - Y cation group
vesuvianites were checked for accuracy and the substitutions (Table 2) but only those which are
existence of Si-deficient vesuvianites from other believed realistic are given: Na § Tin+ = Ca 2§ + A13§ ;
localities (Fig. 1) supports the fact that our data are Na § Tin+ = Mg 2§ + AI3+; Ca 2+ + Mg 2+ = Na + + A13+;
not exceptional. Ca 2+ + Mg 2+ = 2Fe2+; Na + + A13+ = Tin+; Na + + A13+ =
Some of the Connemara vesuvianites display 2Fe2+; Na + + A13+ = Fe 2+ + Ca2+; Ca 2+ + Na + = A13+;
oscillatory concentric birefringence-zoning with Ca 2§ + Mg 2§ = Fe3§ Na+. It should be emphasized
alternating low and high birefringence zones. There that the strong negative correlations in the equations
is a general decrease in A1 and Fe and increase in Ti involving Ti and A1 on each side are strongly
and Mg from core to rim in the low birefringence governed by Ti for A1 substitution.
zones (Fig. 2) or in the optically unzoned grains. The There is a decrease of the XY cations in passing
high birefringence zones are very narrow and from the vesuvianite core to rim which can be reliably
systematically characterized by low A1 and high Ti attributed mainly to A1 and to a lesser extent Fe (Figs
contents (Fig. 2). Although Fe tends to be higher and 2 and 3) implying that the excess of the Y cations over
Mg lower in the high birefringence zones, the trends 13 a.p.f.u, is related to these cations. Groat et al.
are not always maintained especially toward the rims (1994a) showed that the Y group excesses, which they
(Fig. 2). Aluminium can also be eliminated because also attributed to A1 and Fe in B-free vesuvianites,
zones of high and low birefringence can have similar occupy a T(1) site which is a tetrahedrally coordinated
AI concentrations (compare high birefringence zones site linked to anions occupying 0(7) and O(11). Based
in the rims with low birefringence zones in the cores on local bond valence of their studied vesuvianites,
in Fig. 2). It is therefore concluded that the high these authors suggested two main substitutions:
birefringence zones are related to high Ti concentra- T(1)(AI,Fe) + H(1)~4 = r 0 ) ~ 2 + H(1)H4 + x(3)Ca
tions in agreement with previous findings (e.g.
 WOLLASTONITE-VESUVIANITE SKARN 545

o5'ooo* 5'5'~ ~

- o oooo ooooo
Mid d d " ' ' o " d d d d d d
I t I I I I I I
0

,5 ,:5
I I I I
0

d d d d d d d d d d d d d ,.-.,
.~ I I I I I I I I I

r..)
5, o5, oooooooo--
I I
e.-,

+ d d d e d d d d d d d--
Z I I I 1 I I I I

0
~+mNN~EN-nN
~ d d d d d d ~ d d ,.-,
~ I I I I I I I 0 0 ~ 0 0
I
0

Z I

I I

o d o o d ,-., II
I I
o d d d o d
II
d d d d d ~
+ I 1 +
,.~'N

N ~ d d d ~

d ~ d d d ~
II
546 Y. AHMED SAID AND B. E. LEAKE

14.0 Ca, YA1 and Mg+Fe 2§ but also with ZAl (i.e. A1
substituting for Si in the Z sites) confirming the
a b o v e s u b s t i t u t i o n s but also i n d i c a t i n g t h a t
13.8 r(1)(A1,Fe3+) + ZA1 = (m)Hz + Zsi is an important
form of substitution in the Connemara vesuvianites
compared with those published by Groat et al.
13.6 (1994a), probably because the Z sites of their
published vesuvianites are fully occupied by Si.
The Fe2+/Fe 3§ ratio is computed in the following
13.4 way. Iron correlates negatively with Mg and A1
(R = - 0 . 2 1 and - 0 . t 8 respectively) indicative of
Fe2+>Fe 3+ . Thus, Fe2+/Fe 3§ = 0.54 was taken; the
Z 13.2 assumption is that all Fe 3§ and Fe 2§ correlate with
A13§ and Mg respectively. The average Connemara
vesuvianites have the composition:
13.0 (Cals.60Nao.15)(A19.81Tio.63Fe(III)o.T1Fe(II)o.83Mno.02-
0 o ZY
~Y-A1T Mgl.24)(Alo.42SilT.58),
12.8 9 Y.T(1)=(AI + Fe3+)
indicating that the T(1) sites are occupied by 0.24
0.5 cations. If the Fe3+/A13+ ratios in the Y- and T(1)-sites
0.4 are the same then the Connemara vesuvianites would
have the formula:
0.3
(Ca 18.6oNao.15)(A19.59Tio.63Fe(IIl)o.69Fe(II)o.83Mno.o2-
0.2
Mgl.24)(Alo.z2Fe(III)o.og(Alo.42Si~7.58)O68(OH6.36F2.640).
0.1
The SO2 does not exhibit any zoning patterns but
0.0 correlates negatively with both M g O and T i t 2
(R = - 0 . 2 2 ) indicative of the substitutions 2H § +
F~G. 3. Plot showing the variation of Y.Y, Y-Y-A1T and S 4§ = Mg 2§ + Ti 4+ or possibly S4§ = Tin+; S 4§ =
ZT1 ions across the 2.5 mm vesuvianite (see text for 2Mg 2§ The C1 and F substitute for each other; ZnO,
details). which is always between 0.02 and 0.09 wt.%, does
not show any preferred core to rim distributions and
correlates negatively with Mg suggesting
and Zn 2+ = Mg 2§
~l)(A1,Fe3§ + H(1)~2 + r(Mg,Fe2§ =
7"(1)<> + nO)H2 +r(A1,Fe3§ ) Phase equilibria and conditions of metamorphism
where <> denotes a vacancy. The internally consistent thermodynamic dataset of
The estimated T(1) cations in the Connemara Holland and Powell (1990; see also Holland and
vesuvianites (see below), correlate negatively with Powell 1985; Powell and Holland 1985, 1988, 1994)

TABLE 3. Mineral activities and abbreviations used in the text

Garnet cores Garnet rim

Gr Aim Di Hed Cats Gr Aim

a 0.768 5.3e -5 0.738 0.196 0.0136 0.503 5.86e -4

sd(ln a) 0.0064 0.8337 0.0206 0.1863 0.7353 0.0377 0.7557
The activity of all other minerals is taken as unity except muscovite (amu = 0.92).

Mineral abbreviations
Ab = Albite; An = anorthite; Ksp = K-feldspar; Cats = Ca-Tschermak's molecule; Di = Diopside;
H e d = Hedenburgite; A l m = Almandine; Gr = Grossular; Wo = Wollastonite; Cc = Calcite; Qz = Quartz;
Mu = Muscovite; Cor = Corundum; Sill = Sillimanite; And = Andalusite; Ky = Kyanite; Vsv = Vesuvianite;
Zo = Zoisite.
 WOLLASTONITE-VESUVIANITE SKARN 547
700 9 t . , . + . , , t +
temperatures exceed 700~ around felsic intrusions
such as the Omey Granite. The pressure is restricted
i,+ : ~ tt_.__HP+ to 2<P<4 kbar because reactions in the pelites (see
65O below) took place within the stability field of
andalusite extending to the sillimanite field only in
3 Gt-cot, 12 & 13
the innermost parts of the aureole. The values of
600 1 ~.7IGt.care I 11 ($3) . 0.0<Xco <0.4 are taken because the upper limits are
defined by the intersections between the pelite and
t calc-silicate equilibria under the optimum pressure
2 (see below), and bringing in zoisite following the
550
reaction: (7) 3Qz + 5Cc + 2Zo = 3Gr + 5CO2 + HEO,
for agr= ff768-0.503 indicate Xco2 t>0.012. Similar
minimum Xco2 values are obtained using reactions
500 -
0,05 010 0.15 0.20 0.25 0.30 0.35 0.40
involving Zo-Gr-An.
0.00
XC02
In the calc-silicates, Xr was probably <0.2
because; (a) the occurence of euhedral-subhedral
albite, spatially associated with garnet, suggests that
FIG. 4. Isobaric (3.3 kbar) T-X plot in the CASCH system
Ca in plagioclase was consumed following the
for the calc-silicates (the intersections of equilibria 1-3
reactions; (8) Cc + Wo + (Ca+Na)plag = Gr + CO2
are shown for P = 3.3 ___ 0.3 kbar for garnet core (open
(+albite); (9) 2Cc + Wo + (Ca+Na)plag = Gr + 2CO2
circle) and rim (filled circle)). Also shown are reaction 7
(+albite); (10) 2Wo + (Ca+Na)plag = Gr + Qz
for both garnet core and rim and reactions 11-14 in the
(+albite)
pelites. HP = Holland and Powell (1990), $3 = prismatic
Grossular produced through these equilibria is
sillimanite of Salje (1986) and HM = Holdaway and
stable only at Xcoz<0.3 (Deer et al., 1982), and (b)
Mukhopadhyay (1993). Dark stipple = likely T-Xco2 in
the coesistence of Mg-vesuvianite, wollastonite and
the Omey Granite aureole.
quartz, although the considerable Fe substitution for
Mg and F for OH in the present vesuvianites, requires
XH,O>0.8 (Valley et al., 1985).
through the program THERMOCALC vl.2 (1990) Within the pelites, sillimanite coexists with
has been used throughout the study, instead of the andalusite in the innermost parts of the aureole and
elemental partitions between coexisting minerals; the muscovite is broken down to K-feldspar and both
advantages are discussed by Bucker-Nurminen andalusite (sillimanite) and corundum suggesting the
(1990). following CO2 insensitive equilibria: (11) Sill -- And;
The studied sample (BEL 4917) consists of (12) Qz + Mu = And + Ksp + H20; (13) Qz + Mu =
vesuvianite, grossular, diopside, calcite, wollastonite sill + Ksp + H20; and (14) Mu = C o r + Ksp + H20.
and quartz, defining divariant equilibria in the system When plotted on the T-Xco~ diagram the curves of
CMASCH. As the activity of vesuvianite is unknown reactions 12 and 13, which are believed to be in
(discussed below), the calc silicates phase equilibria equilibrium in the present aureole, coincide with each
will be restricted to the remaining phases taking the other at 0<Xco2<0.4 and 3.6<P<3.0 kbar for the
mineral activities given in Table 3. likely T = 640+_20~ (Fig. 4). Under these P-T
Of the six possible equilibria (Mineral abbrevia- conditions, reaction 14, on the other hand, would
tions are given in Table 3) (1) Cc + Qz = Wo + CO2; only be stable if 0.28<Xco2<0.60 for the same
(2) Cats + 2Qz + 2Cc = Gr + 2CO2; (3) Cats + Wo = temperature range (see below).
Gr; (4) A l m + 4Qz + 4Cc = 3 Hed + Cats + 4CO2; (5) The physical conditions of metamorphism around
2Gr + A l m = 3Hed + 3 Cats; and (6) A l m + 6Qz + Cc the Omey Granite are best estimated by plotting
= Gr + 3Hed + CO2; only the first three reactions are equilibria 1 - 3 and 11-13 on a T-Xco2 plot at 2<P<4
usable as the aalm is far too low, resulting in large kbar (Fig. 4). Then, a P-T projection of their
standard errors on activities (Table 3) and hence intersections will trace P-T lines, contoured by
excluding equilibria 4 - 6 as potential petrogenetic Xco2, along which the pelites and calc-silicates
indicators. Although equilibria 2 and 3 are less useful could have been in equilibrium. However, since
than reaction 1 due to the large errors in the reactions garnet is zoned and despite the uncertainty of the
involving Cats ( _ 100~ T.J.B. Holland, pers. reactions involving Cats, it is more relevant to
comm.), the first three reactions are considered in the include the intersections of reactions 1 - 3 rather
poT-X calculations but with special emphasis on the than the individual curve of reaction l (Fig. 4) so that
best experimentally determined reaction 1. the isobaric increase of temperature with garnet
On the T-Xco2 plot (Fig. 4), the temperature is growth is emphasized especially as X~gincreases
chosen as <700~ because it is unlikety that contact from core to rim. Figures 4 and 5 indicate that the
548 Y. AHMED SAID AND B. E. LEAKE
5
country rock for more than 5 0 - 1 0 0 m for high-
~ Granite solidus level intrusions such as the Omey Granite. Also, it is
not at all certain that anorthite exists in the skam
4
assemblages and its absence invalidates the use of
Wo-Gr-An equilibria. Further, there is no convincing
evidence that the calc-silicates were in perfect
equilibrium with the pelites especially as the
CO2/H20 ratios could differ even at the same locality
2 depending on the rock composition (e.g. pelites and
marbles), porosity, permeability, grain size and
diffusivity. The restriction of corundum to within
1 the muscovite plates (Ferguson and A1-Ameen,
1985), suggests that the H20 released from the
decomposing muscovite according to reaction 14
0 may have locally increased the C O J H 2 0 ratios of the
500 550 600 650 700 750 800
Temperature(ol2) circulating fluids thereby forming corundum at
relatively lower temperatures. Some localized
convective H20-rich fluid might have been involved.
Fro. 5. P-T projection diagram showing P-T lines along
which the pelites and calc-silicates could have been in
The vesuvianite mole fraction
equilibrium. Dark stipple = likely P-T around the Omey
Granite. Arrow indicates increasing temperature with The P - T - X around the Omey Granite are now
garnet growth. RGB = Richardson et al. (1969). Other constrained and hence varying the avsv for the
symbols as in Fig. 2. reaction(s): (15) 2Vsv + 5Wo + CO2 = 11 Gr +
6Cc Di + 9H20 and/or, (16) 2Vsv + 6Qz = Wo + 11
Gr + 4Di + 9H20, until an agreement with the above-
determined P - T - X would enable the activity of
physical conditions of metamorphism around the vesuvianite (avsv) to be estimated. Taking the
Omey Granite were P = 3.3-t-0.3 kbar, T = relevant mineral activities as given in Table 3,
640+20~ Xc% = 0.15-t-0.05. These values are in avsv = 0.01. More importantly, the activity coefficient
excellent agreement with those deduced from (Tvsv) could now be computed from the relation
crystallization of the Omey granite where the Qz- avsv = XvsvYvsvwhere Xvsv = mole fraction assuming
Ab-Or minima follow the XH~O = 0.8--0.9 curves of an ideal solution.
Ebadi and Johannes (1991). However, while the Based on the formula CaI9Mg2AlllSi18069(OH)9,
values determined above agree with those of assuming ideal mixing on sites and all the Z, X and
Ferguson and A1-Ameen (1985), the pressure is Y2 sites are occupied by 18Si, 19Ca and 4A1
about 0.7 kbar higher. The low-pressure terms respectively, Hoisch (1985) proposed the thermo-
obtained by these authors is probably due to the dynamic mole fraction:
mineral activities of grossular, diopside and musco-
X ~vsv Y1 Y3 Y3 7
g = 20.4(X~g)(X~g)(X)a H 7
) (X~oH) ,
vite which were not accounted for but also to the fact
that their calc-silicate equilibria involve pure where Y1- and Y3-sites refer to B- and A1Fe-sites in
anorthite, which is absent from the present calc- the terminology of Hoisch (1985). This mole fraction
silicate and they also used the andalusite-sillimanite needs to be modified to account for A1 for Si and F
equilibrium of Holdaway (1971). for OH substitutions and the existence of Tl-sites.
Ferguson and A1-Ameen (1985) argued that the Thus, the new formula, which is also applicable to B-
growth of corundum at the expense of muscovite bearing vesuvianites, becomes:
according to reaction 13 above, which occurred about X~Vg = 20.4(XXa)'9(XVMlg)(X~3)(X~3)7(X~I)5(xZi)'8(X~o~) 7
100 m further away from the granite contact than
their Wo + Gr + An assemblage (i.e. 250 m from the where:
granite contact), would require Xco~ of 0.5-0.6.
(Xr~1) = 1 - Fe(II) m (if Fe 2§ 3§ ratio is not
They concluded that H20-rich magmatic fluids
available then Fe(II) rl = 0.5 Fetot),
derived from the main granite body must have
percolated through the country rock. Although our vs ( M.~A1 ,~vs
results also suggest higher Xc% (0.28<Xc%<0.6) for (X~/g,A1) = ~,Mg+Al+a~+vo ++vo3++Mn,+/ ,
this reaction, magmatic fluid was probably insignif-
Mg rs = Mgtot - MgY1; AI ~ = Altot - A1v2 _
icant because Ferry (1994) showed that it is very
AIr(AI ~ = 4 a.p.f.u.),
unlikely that such fluids percolate through the
 WOLLASTONITE-VESUVIANITE SKARN 549

= AITI+Fea++B ) Ferguson, C.C. and AI-Ameen, S.I. (1985) Muscovite

breakdown and corundum growth at anomalously
A1T1 + Fe 3+ = Y.Y cations - 13 low fH20: A study of contact metamorphism and
convective fluid movement around the Omey
OH Granite, Connemara, Ireland. Mineral. Mag., 49,
(X~oH) : (()H+F+"CI+O)'
505-14.
power of seven because 2 - 3 H § ions occur as Ferguson, C.C. and Harvey, P.K. (1979) Thermally
bonding cations with O(10)s, Groat et al., 1995), overprinted Dalradian rocks near Cleggan,
( XC xa ) , (xZi) = (kCn+Na+REE/
c~ .~x and [ sl ~z Connemara, western Ireland. Proc. Geologists'
~i+~-u 9 Assoc., 90, 43-50.
If the EY cations = 13 and the X and Z sites are Ferry, J.M. (1994) Role of fluid flow in the contact
fully occupied by Ca and Si respectively and metamorphism of siliceous dolomitic limestones.
vesuvianite is F- and B-free, then the above Amer. Mineral., 79, 719-36.
formula reduces to Hoisch's(1985). Frimmel, H.E. and van Achterbergh, E. (1995)
Using the formula given above, the average Metamorphism of calc-silicate and associated rocks
Connemara vesuvianites yield X ~ = 0.001 implying in the Pan-African Kaaimans Group, Saldania Belt,
that the 7vsv = 10 confirming the non-ideal solution of South Africa. Mineral. Petrol., 53, 75-102.
vesuvianites. The importance of these activity Gibson, R.L. and Wallmach, T. (1995) Complex zoning
models and the petrogenetic significance of vesuvia- in vesuvianites from the Canigou Massif, Pyr6nEes,
nite with metamorphic grade will be published France. Canad. Mineral., 33, 77-84.
elsewhere. Groat, L.A., Hawthorne, F.C. and Ercit, T.S. (1992a)
The chemistry of vesuvianite. Canad. Mineral., 30,
19-48.
Conclusions
Groat, L.A., Hawthorne, F.C. and Ercit, T.S. (1992b)
The Connemara vesuvianites have Y-cations in The role of fluorine in vesuvianite: a crystal-
excess of 13 a.p.f.u, consistent with an additional structure study. Canad. Mineral., 30, 1065-75.
tetrahedrally coordinated site (T1) occupied by (A1, Groat, L.A., Hawthorne, F.C. and Ercit, T.S. (1994a)
Fe 3+, B), Groat et al. (1994a). The zoned Excess Y-group cations in the crystal structure of
vesuvianites show the importance of within-site, vesuvianite. Canad. Mineral., 32, 497-504.
mainly among the Y cations, and cross-site, Groat, L.A., Hawthorne, F.C. and Ercit, T.S. (1994b)
dominated by X-Y and Y-Z interactions, combined The incorporation of boron into vesuvianite
substitutions. The high birefringence zones in structure. Canad. Mineral., 32, 505--23.
vesuvianites are related to high Ti concentrations. Groat, L.A., Hawthorne, F.C., Rossman, G.R. and Ercit,
The application of a modified thermodynamic mole T.S. (1995) The infrared spectroscopy of vesuvianite
fraction to the Connemara vesuvianites confirms a in the OH region. Canad. Mineral., 33, 609-26.
non-ideal solution model for the mineral. The studied Hoisch, T.D. (1985) The solid solution chemistry of
skarn is consistent with crystallization at 640+20~ vesuvianite. Contrib. Mineral. Petrol., 89, 205-14.
at 3.3-t-0.3 kbar and with Xco2 at 0.15__+0.05 (i.e. Holdaway, M.J. and Mukhopadhyay (1993) A reevalua-
XHO = 0.85_+0:05). tion of the stability relations of andalusite:
Thermochemical data and phase diagram for
aluminium silicates. Amer. MineraL, 78, 298-315.
Acknowledgements
Holhmd, T.J.B. and Powell, R. (1985) An internally
We thank Mr R. Macdonald and Dr C. Farrow for consistent thermodynamic dataset with uncertainties
assistance with the use and calibration of the Cameca and correlations. II Data and results. J. Meta. Geol.,
SX 50 electron microprobe. Dr T.J.B. Holland is 3, 343-70.
particularly thanked for significant scientific sugges- Holland, T.J.B. and Powell, R. (1990) An internally
tions to improve the first draft. consistent thermodynamic dataset with uncertainties
and correlations: The system Na20-K20-CaO-MgO-
MnO-FeO-Fe203-AI203-SiO2-TiOz-C-H2-O2. J.
References
Meta. Geol., 8, 89-124.
Bucher-Nurminen, K. (1990) Geological phase diagram Kerrick, D.M., Crawford, K.E. and Randazzo, A.F.
software. Terra, 4, 401-10. (1973) Metamorphism of calcarous rocks in three
Deer, W.A., Howie, R.A. and Zussman, J. (1982) Rock- roof pendants in the Sierra Nevada, California. J.
forming minerals, volume 1A, Longman, London. Petrol., 14, 303-25.
Ebadi, A. and Johannes, W. (1991) Beginning of melting Leake, B.E. and Tanner, P.W.G. (1994) The geology of
and composition of first melts in the system Qz-Ab- the Dalradian and associated rocks of Connemara,
Or-H20-CO2. Contrib. Mineral. Petrol., 106, 286-95. Western Ireland. Royal Irish Academy, Dublin 96pp.
550 Y. AHMED SAID AND B. E. LEAKE

Leake, B.E., Tanner, P.W.G., and Senior A. (1975) The Richardson, S.W., Gilbert, M.C. and Bell, P.M. (1969)
composition and origin of the Connemara dolomitic Experimental determination of kyanite-andalusite
marbles and ophicalcites. J. Petrol., 16, 237-57. and andalusite-sillimanite equilibrium; the alumi-
Leggo, P., Compston, W. and Leake, B.E. (1966) The nium silicate triple point. Amer. J. Sci., 267, 259-72.
geology of the Connemara granites and its bearing Salje, E. (1986) Heat capacities and entropies of
on the antiquity of the Dalradian series. Q. J. Geol. andalusite and sillimanite: the influence of fibroliti-
Soc., 122, 91-116. sation on the phase diagram of the A12SiO5
Powell, R. and Holland, T.J.B. (1985) An internally polymorphs. Amer. Mineral., 71, 1366-71.
consistent thermodynamic dataset with uncertainties Townend, R. (1966) The geology of some granite
and correlations. 1 Methods and worked example. J. plutons from western Connemara, Co, Galway. Proc.
Meta. Geol., 3, 327-42. Roy. Ir. Acad., 65(B), 157-202.
Powell, R. and Holland, T.J.B. (1988) An internally Valley, J.W., Peacor, D.R., Bowman, J.R., Essene, E.J.
consistent thermodynamic dataset with uncertainties and Allard, M.J. (1985) Crystal chemistry of Mg-
and correlations. III Application methods, worked vesuvianite and implications of phase equilibria in
examples and a computer program. J. Meta. Geol., 6, the system CaO-MgO-A1203-SiO2-HzO-CO2. J.
173-204. Meta. Geol., 13, 1-12.
Powell, R. and Holland, T.J.B. (1994) Optimal
g e o t h e r m o m e t r y and geobarometry. Amer. [Manuscript received 28 February 1995:
Mineral., 79, 120--33. revised 1 November 1995]