Contributions to

Contrib Mineral Petrol (1981) 76:177 1-95

Mineralogy and
Petrology
9 Springer-Verlag 1981

Trace Elements Behavior in Granite Genesis: A Case Study

The Calc-Alkaline Plutonic Association
from the Querigut Complex (Pyr6n6es, France)
Serge Fourcade and Claude J. Allegre
Laboratoire de G6ochimie et Cosmochimie (LA 196), Institut de Physique du Globe, Universit6 de Paris VI, 2 place Jussieu, F-75005 Paris,
France
D6partement des Sciences de la Terre, Universit~ de Paris VII, 2 place Jussieu, F-75005 Paris, France
Laboratoire "Pierre S~e", Centre d'Etudes Nucl6aires de Saclay, F-78 Saclay, France

Abstract. The different granitoids of the zoned Querigut complex chemical modelling of the formation and evolution of the conti-
(Hercynian Pyrenees) are associated with a series of basic to nental crust (Patterson and Tatsumoto 1964; Hurley et al. 1962;
intermediate rocks ranging from hornblende-bearing peridotites Armstrong 1968; Moorbath 1975; Allegre 1972; O'Nions et al.
to quartz-diorites. The whole complex appears as a calc-alkaline 1979; Allegre and Ben Othman 1980) using radiogenic isotopes
plutonic suite typical of orogenic zones. The distribution of lan- as tracers requires the knowledge of the behavior of trace ele-
thanides and other trace elements amongst coexisting minerals ments like Rb, Sr, Sin, Nd, U, Th, Pb, during granite genesis.
indicate they are essentially held by accessory phases, particularly It is even more critical for models based on trace elements only,
in granitoids. This restricts the use of those elements in the such as those developped by Taylor (1979). The present paper
calculation of petrogenetic models for acidic plutonic rocks. is a detailed investigation on one of the most abundant granite
Magmatic differentiation, mainly by hornblende+plagioclase type occurring in the European Hercynian (Variscan) orogenic
fractionation, can produce the basic series. This differentiation belt: the calc-alkaline granite from Querigut (Pyr6n6es). Indeed,
cannot directly produce the different granitoids, which require only a few studies on R.E.E. in Hercynian granitoids are still
a preponderant contribution of crustal melts. The sequence of available (Emmermann et al. 1975; De Albuquerque 1978).
different granitoids can be explained either by an heterogeneity
in the source region, or by magmatic differentiation. The most
plausible interpretation of the whole complex calls for the empla- Geological Setting and Petrology
cement of a mantle-derived magma into a wet, anatectic conti- of the Querigut Eruptive Complex
nental crust, with interactions between basic rocks and the so- If we set aside the posttectonic alkaline granites of minor extent,
produced acidic melts. the major part of the Hercynian plutonic magmatism is represented
by two dominant groups (Didier and Lameyre 1969): (1) 2-mica granit-
oids subdivided into leucogranites and aluminous, 2-mica calc-alkaline
granitoids. (2) Calc-alkaline granitoids (mainly granodiorites) which
are the dominant magmatism.
Introduction Even though generalization to all hercynian massifs may not be
valid because of lack of geochronological data, there is a clear relation-
The last decade has been marked by an extensive use of trace
ship between age and type of granitoids in the Pyrenees and the South-
elements in petrogenetic problems. With the complementary ad-
ern Massif Central (Vitrac-Michard and Allegre 1975; Hamet and
vances in experimental petrology, we can say that this approach Allegre 1976). Most two-mica granitic bodies are affected by major
is one of the major recent advances in petrology. However, tectonic events of the orogeny and are generally emplaced in a high-
such an approach has been restricted almost exclusively to basalts grade metamorphic environment. They are older (335 m.y.) than calc-
and related rocks with fewer studies on acidic and particularly alkaline plutons (280 m.y.) emplaced at higher crustal levels at the
granitic rocks. This situation has recently changed but the number end of the hercynian tectonism. Some of the most representative mas-
of studies related to granite genesis is still small and far behind sifs have been described by Didier (1973) on the basis of the nature
our knowledge of basaltic systems (see Hart and Allegre 1979). of their inclusions. This author has shown the constant association
of mafic inclusions, generally of dioritic composition, with all the
Granitoids are the major components of continental crust
calc-alkaline complexes. These dioritic rocks may appear as small-sized
and the second most abundant component of the surface of
inclusions (less than 1 m) or as mappable individual stocks. On the
the earth after basalts. Their origin has remained a great debate contrary, muscovite-bearing granitoids do not usually carry appreci-
of this century with arguments for a genesis by fractional crystal- able amounts of dioritic inclusions but contain remnants of metamor-
lization of a basaltic magma or for remelting of metamorphic phic rocks or hyperbiotitic inclusions. This has been an argument
series (Bowen 1928; Tuttle and Bowen 1958; Winkler 1965 for an anatectic origin of those granitoids.
anqong others) and supporters of exotic hypotheses like the solid The Querigut eruptive complex, situated in the Oriental Pyr6n6es
state diffusion theory or the metasomatic origin (Perrin and has attracted the interest of many geologists because of its spectacular
Roubault 1939; Ramberg 1944; Walton 1960; Orville 1962). associations between mafic and acidic rocks. The general geological
A more precise knowledge of the genetic processes of granitoids context of the complex is illustrated in Fig. 1. Its description and
the map of Fig. 2 are taken from Leterrier (1972) and Marre (1973)
and the resulting behavior of trace elements will have a major
comprehensive studies. Schematically, the complex is intermediate be-
impact in petrology but also in geodynamics. Indeed the geo- tween a concentric and a stratoid type. It is composed of four main
acidic units from edge to center: a granodioritic-tonalitic unit, a mon-
Contribution I.P.G. n ~ 481 zogranite, a porphyritic monzogranite, and a biotite calc-alkaline gran-

0010-7999/81/0076/0177/$02.80
178

0 20 Km
Precambrian Gabbro
Paleozo'ic
Gneisses ~ Diorite
~ ] Monzonitic Calc-alkaline
granite Biotite Granite ~ Faults
HornbI-Biotite C a l c - a l k a l i n e Granite with Fe-
Granodlorlte Biotite and miarolitic Granite
Charnockitic Muscovite-Biotite Granite
Granodiorite
!

o o o o o~
0 0 0 0 0
o~o Aston o
i OoO o o o
o o ~
4Z~O'

o Host:italet o o

c: n e3 C
Caran~a o
o o o o i

Saint La u r e~'fl'T

0 ~ O"

Fig. 1. Simplified geological map of the Oriental Pyr6n6es, showing the different massives of Hercynian granitoids and of Cambrian gneisses
(after Autran et al. 1970)

ite. Small stocks and veins o f aplitic, potassic, 2-mica granites are the formation of granodiorites and tonalites in the vicinity o f mafic
found in various parts of the complex, generally in the vicinity of stocks from initially granitic rocks. For the latter author, the various
fractured zones. Sparse but thick dykes of porphyritic microgranite mafic and acidic rocks could result from the differentiation of a single
cross-cut granodiorites in the southern part of the massif. Contacts type of basic magma, giving rise to the whole series. This interpretation
between the different units seem to be progressive but rapid except is basically similar to that proposed by Autran et al. (1970) and follows
for the central one which is clearly cross-cutting the others and there- the classical Bowen's theory. The spectacular trend displayed in the
fore, probably the last to be emplaced. A.F.M. diagram by the whole series may be a good argument for
"Basic" rocks (for convenience we will group under this term the latter interpretation (Fig. 3). On the other hand, Vitrac-Michard
all of the different mafic rocks ranging from amphibolites and gabbro- and Allegre (1975), using Sr isotopes and Rb-Sr as trace-elements
diorites to quartz-diorites) outcrop in three different settings: favor a mixing hypothesis between a differentiated mantelic component
Heterogeneous zones in which elongated xenoliths of dioritic rocks and a crustal one. Such a model is closer to Marre's interpretation.
are imbedded in a foliated granodioritic to tonalitic matrix (mainly
in the south-western part of the complex) ;
- Isolated non deformed microgranular quartz-dioritic inclusions in Experimental Results and Discussion
granitoids; of the Trace Element Data
- Large stocks of dioritic rocks, some of kilometer-size, included in
various granitoids but mainly in granodiorites and monzogranites. 1. Analytical Procedure and Accuracy of the Measurements
Their contacts with acidic rocks consist of a thin fringe of tonalite.
Some of these stocks themselves contain numerous inclusions of Trace elements concentrations have been analyzed by Neutron Activa-
Cortlandites (hornblende peridotites) and hornblende norites. The tion Technique following two different experimental procedures de-
whole complex intrudes paleozoic sediments, mostly limestones and scribed by Treuil et al. (1973), Chayla et al. (1973). R.E.E. with Cs,
calcareous schists in the North and West, and schists in the South. Rb, Sr, Ba are measured after chemical separation. Non destructive
The contacts dip towards the center of the massif and a 2 hm-thick instrumental analysis is used for the other trace elements. La, Eu,
zone of contact metamorphism can be observed. Leterrier (1972) and Tb, Rb, Sr, Cs, and Ba are measured by both procedures. Counting
Marre (1973) have formulated two different interpretations of the uncertainties depend upon the concentration in the sample, but general-
whole complex. In the former one, two independent magmas were ly, the precision is in the range 1-5% except for G d = 7 % and Tm
emplaced simultaneously, a high-At basaltic magma giving birth to 10%.
the mafic rocks, and an acidic magma. In this interpretation, fluid For the whole rock analyses, samples of 5 to 15 kg were collected
interaction between acid and basic formations could have lead to and crushed in an agate mortar. The purity of separated minerals
 179

ic.
~ Ordovician :
Pelites and
~Granodiorites,
I,< , [ t o n a l i t e s rich
.~ Porphyrit
monzogran ire • Undifferentiated k Faults
quartzites in m a f i c foliated granito'fds
inclu s i o n s
Goth and
. an , ~^C l u l ~"
C ane - Biotite Dykes of porphyritic
r"F"~Devonlan : Calc- ~ basic rocks ~Monzogranite granite ~ microgranite '
' = J Jschists and r,.-I ;~.~;__E~.~

Fig. 2. Simplified geologicai map of the Querigut Complex (after Leterrier 1972)

F
was checked under a polarizing microscope. Duplicated analyses of
a whole rock (sample Qt30) and mineral (hornblende Qt04) are given
in Tables 1 and 3. Four samples of the central granitic unit located
far from each other give very close resuits, checking the whole analyti-
cal process. They are given with BCR 1 and GSN standard analyses
in Table 1.

2. E x p e r i m e n t a l Results on W h o l e R o c k s

The whole rock data for the different subunits and rock types
are reported in Table 1. A brief description of the analyzed
samples is given in appendix. The major elements analyses by
Leterrier (1972) and some additional ones are given in Table 2.
They correspond to either the samples studied for trace elements
or other samples from the same subunit.
We will comment the results first using the normalization
procedure for lanthanides (Masuda 1962; Coryell et al. 1963),
then using simple two by two correlation. For convenience, we
will lump all mafic rocks in the "basic series" and all acidic
rocks in the "acidic series", without any implication at present. A
We will come back to this distinction later on. Fig. 3. A.F.M. diagram constructed for the main Querigut rock-types
(after Marre 1973). Black circles: Cortlandites, Hornblende Norites,
a) Lanthanide Distribution Amphibolites; black triangles: Gabbro-dioritic series; black squares:
Tonalites and Granodiorites; open squares: Monzogranite; open dia-
The lanthanides patterns exhibit some constant characteristics monds: Porphyritic Monzogranite; open circles: Central biotite gran-
through the whole complex (Fig. 4): (1) Enrichment of light ite; crosses: Muscovite granitoids
over heavy elements in all the rock types; two kinds of rocks
display a less marked enrichment: On the one hand, ultramafic heavy R.E.E. is weakly fractionated or not at all. (3) with two
and mafic rocks containing olivine (cortlandite and hornblende exceptions (central biotite granite and 2-mica granitoids), no
norite) and, on the other hand, two-mica subordinate granitoids, important Eu anomaly appears.
particularly the aplitic vein sample Qt85; (2) If one excepts the These three features have been observed in many other calc-
subordinate granitic varieties (2-mica granitoids), the group of alkaline complexes from the Hercynian Pyrenees (Mont-Louis,
180

Table 1. Trace - elements concentrations (ppm) in whole rock samples from the Querigut complex and Aston orthogneisses
Table l a

Rock Cortlan- Hornbl- Gabbro-diorites Quartz-diorites Dioritic Standards Chon-

dite Norite xenoliths drites

N~ Qt 46 Qt 104 Qt 09 Qt 44 Qt 08 Qt 10 Qt 04 Qt 02 Qt 101 Qt 13b BCR1 GSN

La 7.2 6.8 14.8 15.4 22.4 25.2 27.3 56.4 22.9 11.1 26.0 67 0.32
Ce 14.5 14.7 32.0 27.4 47.2 58.0 48.9 99 44.2 38.4 53 117 0.787
Nd 10.2 10.2 20.9 14.0 25.3 33.3 33.9 43 20 40 29 47 0.58
Sm 1.95 2.0 3.5 3.4 5.5 7.4 5.2 8.8 3.6 14.1 6.6 7.2 0.185
Eu 0.51 0.53 0.98 0.98 1.37 1.86 1.17 1.76 1.2 0.92 1.94 1.52 0.071
Gd / / / / / / / / 3.1 / / 4.0 0.256
Tb 0.3 0.3 0.8 0.5 0.85 1.1 0.7 0.9 0.5 3.1 1.0 0.56 0.05
Yb 1.1 1.1 1.7 1.7 2.6 3.3 2.0 2.2 1.0 12.0 3.36 1.43 0.186
Lu 0.18 0.21 0.27 0.29 0.4 0.53 0.36 0.37 0.24 1.85 0.55 0.23 0.034
U 0.7 / 0.9 2.7 1.1 1.0 1.6 / / / 1.7 7.74
Th 1.8 2.4 3.3 7.3 3.6 1.7 7.9 13.4 8.2 3.0 5.5 40.8
Hf 1.45 1.70 2.85 3.0 4.75 4.3 4.7 10.35 6.6 7.6 5.0 6.5
Ta 0.19 0.24 0.63 0.79 0.70 0.83 0.98 1.26 0.71 5.7 0.80 2.81
Cr 2020 1380 780 520 300 66 150 37 50 76 57 55
Co 98 80 49 34.5 30 25 24 19.5 14.8 17 37 65.3
Ni 840 290 280 122 87 14 59 9.5 13.4 12.7 10 34
Sc 15.3 28.3 25.4 21.7 25.5 26.6 23.5 22 12.8 47 33.5 7.33
Ba 90 125 255 240 360 505 590 755 725 310 680 1380
Sr 70 / 250 300 325 410 300 305 330 125 340 568
Rb 21 25 160 75 65 68 99 105 88 220 49 179
Cs 1.88 1.25 3.86 7.80 1.60 1.12 4.22 2.67 2.04 22.9 0.94 5.43
Th/U 3.9 / 3.7 2.7 4.3 1.7 4.9 / / /
La/Sm 3.7 3.4 4.2 4.5 4.1 3.4 5.2 6.4 6.4 0.8

Table l b

Rock Granodiorites associated Granodiorites and transition with Monzogranites

with mafic rocks monzogranites

N~ Qt 25 Qt 86 Qt 103 Qt 102 Qt 52 Qt 40 Qt 51 Qt 45 Qt 12 Qt 13 Qt 80 Qt 82 Qt 83

La 24,2 17.6 41.9 34.7 37.5 50.3 39.2 38,0 36.7 37.3 31.2 35.7 36.1
Ce 38.5 32.1 71 60 61 81 59.3 65 62 60 / / /
Nd 21.3 / 24.4 22.6 28.0 32.6 26.7 26.1 32 30 / / /
Sm 4.5 3.5 4.2 4.3 5.0 4.6 4.2 4.3 5.0 4.5 / / /
Eu 1.64 1.36 1.54 1.19 1.23 1.29 1.09 1.10 0.95 0.92 1.04 0.99 1.35
Tb 0.6 0.65 0.5 0.6 0.6 0.5 0.45 0.5 0.7 0.6 / / /
Yb 1.35 1.7 0.63 1.05 1,20 1.03 1.03 1.63 1.95 1.98 / / /
Lu 0.21 0.24 0.17 0.23 0,20 0.14 0.15 0.26 0.34 0.34 0.38 0.40 0.41
u / / / / 1.3 2.4 1.2 4,2 2.3 5.0 / / /
Th / 4.7 / 1 9.3 13.1 9.9 14.6 16.4 15.3 16.0 15.3 15.4
Hf / 5.4 / / 5.3 6.1 5.2 5.2 4.5 4.6 5.1 5.8 5.0
Ta / 0.96 / / 0.88 0.94 0.86 1.09 1.26 1.31 1.42 1.17 1.36
Cr / 30 / / 32 40 23 24 22 18 19.5 9.7 27.5
Co / 8.3 / / 8.9 9.3 7.5 6.2 5.3 4.9 5.3 5.4 6.6
Ni / 4.4 / / 4.4 5.0 3.7 2.8 3.1 2.5 3.1 / 3.2
Sc i 10.3 / / 9.8 8.8 8.0 7.3 7.1 5.9 6.4 7.5 7.3
Ba / 730 / / 810 545 820 790 620 600 640 630 660
Sr / 270 / / 255 245 170 155 165 155 185 / 195
Rb / 103 / / 95 107 120 130 155 160 150 145 150
Cs / 4.73 / / 4.69 3.01 5.00 4.01 5.26 5.64 4.00 4.25 8.9
Th/U / / / / 7.2 5.5 8.3 3.5 7.1 3.1 / 1 1
La/Sm / 5.0 10.0 8.1 7.5 10.9 9.3 8.8 7.3 8.3 / / /

Andorre, Maladetta, Cauterets-Panticosa, Costabonne) and distributions (Fig. 5, after u n p u b l i s h e d data). Therefore the three
f r o m the M o n t a g n e N o i r e (Folat, Sidobre) ( F o u r c a d e , u n p u b - features described above, if n o t characteristic, seem to be at
lished data). O n the contrary, o t h e r types o f variscan granitoids least well r e p r e s e n t e d in the V a r i s c a n calc-alkaline plutonic asso-
in F r a n c e (2-mica granitoids in large massives, subalkaline gran- ciation.
itoids a n d alkaline granitoids) display r a t h e r different l a n t h a n i d e O t h e r c o m m e n t s can be m a d e o n the w h o l e rock d a t a : in
 181

Table 1 e

Rock Porphyritic Central biotite granite Porphyritic Muscovite- Aston Orthogneisses

Monzogranite Monzogranite Granitoids

N~ Qtl8 Qt43 Qt87 Qtl5 Qt30 Qt30 Qt31 Qt32 Qt33 Qt71 Qt72 Qt42 Qt85 Astl Ast2 M15G

La 32.8 39.1 34.0 24.0 23.7 23.2 24.6 / 21.6 26.2 23.3 19.1 3.1 27.5 29.0 25.2
Ce 56.8 66 / 44 44 46 46 / 37.3 47 44 37.7 5.5 53 57 50
Nd 29.7 27.4 / 23.0 23 21 25.5 / 22.5 20.7 21.5 18 2.3 27 28.5 25.5
Sm 4.7 5.05 / 5.1 5.1 5.0 5.4 / 4.8 5.3 5.6 3.0 0.7 6.1 6.2 6.1
Eu 1.02 1.18 0.94 0.60 0.53 0.54 0.53 / 0.53 0.59 0.57 0.49 0.054 0.77 0.66 0.85
Od / / / / / / / / / 4.8 4.4 2.8 / 4.8 4.5 4.2
Tb 0.55 0.4 / 0,8 0.95 0.9 0.9 / 0.8 0,85 0.9 0.45 0.25 0.96 0.95 0.9
Yb 1.32 0.90 / 2.24 2.3 2.6 2.33 / 2.6 2.7 2.7 0.65 1.2 2.78 2.76 2.67
Ln 0.33 0.15 0.26 0.35 0.37 0.39 0.38 / 0.40 0.45 0.44 0.10 0.42 0.43 0.41
U 2.0 2.5 / 5.0 5.6 / / 4.2 / 4.6 / 3.5 5.2 5.2 5.1 3.3
Th 16.7 19.6 18.7 14.2 14.0 / / 14.8 / 14.5 / 11.0 3.2 11.5 10.0 10.0
Hf 4.8 4.8 4.9 3.2 4.2 / / 3.5 / 3.8 / 2.8 1.8 5.5 5.8 5.3
Ta 1.33 1.43 1.18 2:33 2.45 / / 2.22 / 2.50 / 2.31 6.21 1.92 1.87 0.93
Cr 20 19 18 11.4 / / / / / / / 5.4 / 31 29 23.6
Co 4.5 4.5 4.0 1.8 2.2 / / 1.7 / 1.9 / 1.2 0.3 5.8 5.2 4.8
Ni 3.4 4.6 2.6 5.4 / / / / / / / 1.2 / 8.4 9.8 7.9
Sc 5.9 5.7 5.5 4.6 5.I / / 4.6 / 4.7 / 2.5 3.0 7.8 8.3 7.2
Ba 660 620 600 320 320 / / 320 / 390 / 360 530 580 980
Sr 155 150 150 75 80 / / 100 / 82 / 60 55 40 75
Rb 170 160 150 217 238 / / 218 / 250 / 240 317 190 240 190
Cs 9.02 5.96 4.70 8.22 11.38 / / 6.69 / 16.8 / 8.7 9.24 9.36 11.48 9.76
Th/U 8.4 7.8 / 2.8 2.5 / / 3.5 / 3.2 / 3.1 0.6 2.2 2.0 3.0
La/Sm 7.0 7.8 / 4.7 4.6 / 4.6 / 4.5 4.9 4.2 6.4 4.4 4.5 4.7 4.1

Table 2. Major element analyses of the Querigut main rock types (after Leterrier 1972)

Rock Cort- Grano- Monzo- Porphy- Central Porphy- Musco- Aplite Aston
lan- diorites granite ritic granite ritic vite (mean of orthogneisses
dite (mean of (mean of monzo- (mean of micro- gran- 3 anal.)
25 anal.) 42 anal.) granite 46 anal.) granite itoids
(mean of (mean of (mean of
36 anal.) 15 anal.) 15 anal.)
N~ Qt 09 Qt 44 Qt 10 Qt 02

SiO2 42.10 51.71 55.53 51.03 53.72 64.90 68.10 68.60 72.30 73.32 73.10 75.50 71.80 71.00 71.20
A1203 7.25 13.03 14.38 19.33 19.04 16.85 15.65 15.35 14.15 14.48 14.60 12.40 13.70 14.60 14.60
FezO3 13.90 8.12 6.93 8.53 8.05 4.04 3.09 2.96 1.91 1.94 1.28 0.59
MnO 0.20 0.09 0.12 0.05 0.1 0.06 0.05 0.05 0.04 0.05 0.03 0.01 0.06 0.05 0.05
MgO 28.24 10.94 8.2 4.02 3.11 0.39 1.09 1.02 0.37 0.27 0.29 0.04 0.90 0.90 0.70
CaO 4.31 7.27 7.34 8.97 6.21 4.20 2.70 2.67 1.33 1.27 0.90 1.04 1.30 1.45 1.25
Na20 0.78 1.95 2.7 3.57 3.63 3.55 3.55 3.47 3.49 3.58 3.28 2.43 2.85 3.25 3.50
KzO 0.39 3.32 1.79 2.04 2.85 3.03 3.78 3.88 4.56 4.62 4.98 6.44 4.80 4.20 4.80
TiO2 0.31 0.91 0.86 1.53 1.63 0.54 0.38 0.40 0.18 0.15 0.16 0.07 0.39 0.43 0.30
P205 / 0.15 0.13 0.29 0.4 / / / / 0.02 / / 0.16 0.21 0.18
P.F. 2.09 1.47 1.68 1.13 1.69 1.03 1.39 1.29 1.39 0.59 0.96 0.80 0.60 0.85 0.85

99.77 98.96 99.66 100.49 100.43 99.59 99.78 99.69 99.72 100.29 99.58 99.32

the mafic series (with the exception of dioritic inclusion Qt13b, C o n s i d e r i n g acidic rocks, we can individualize several types
l a n t h a n i d e s a b u n d a n c e s increase as M g a n d C a c o n t e n t s de- o f l a n t h a n i d e s d i s t r i b u t i o n (Fig. 4).
crease. T h e increase o f l a n t h a n i d e c o n c e n t r a t i o n s in mafic sam-
ples b r o a d l y c o r r e s p o n d s to a vertical t r a n s l a t i o n o f the n o r m a l - Type 1. G r a n o d i o r i t e s a n d m o n z o g r a n i t e s display R.E.E. a b u n -
ized patterns, b u t a gap a p p e a r s b e t w e e n the m o s t mafic rocks dances very similar to the m e a n a b u n d a n c e s for granitic rocks
(cortlandite a n d h o r n b l e n d e norite) a n d gabbrodiorites. ( H e r r m a n n 1970), for the p o s t - A r c h e a n shales ( H a s k i n a n d Has-
A special m e n t i o n s h o u l d be m a d e o f the m i c r o g r a n u l a r inclu- kin 1966) ( H a s k i n et al. 1968) ( N a n c e a n d T a y l o r 1976).
sion Q t l 3 b . Its I a n t h a n i d e p a t t e r n strongly differs (Fig. 4) f r o m
the whole mafic series. A s we will see later, this p a t t e r n is close Type 2. G r a n o d i o r i t e s a n d tonalites ( m a p Fig. 2) differ f r o m
to t h a t o f h o r n b l e n d e because o f the m a x i m u m o b s e r v e d b e t w e e n the previous type by lower a b u n d a n c e s a n d the presence of a
Ce a n d Sm a n d by the large negative E u a n o m a l y . small positive E u a n o m a l y .
182

A Mafic rocks Hercynian Granitoids

.~ from France
10 2
10 2

10 ~ IZ

Y
~E101
U)

~10 2
~ . B aranodiorites and
6 '~. transition with

E La Ce Nd Sm Eu Gd Tb Yb Lu
Fig. 5. Typical chondrite - normalized R.E.E. distribution in various
granitoid types from Hercynian terranes in France: black circles: Alka-
it. line series, Fayalite-hastingsite granite (Tolla Cauro, Corsica) - open
10 2 "'11.
circles." subalkaline series, Syenogranite (Ploumanach', Brittany) -
C Gr anodiorites Black triangles: calc-alkaline 2-mica granite (Axles Thermes, Oriental
" ' ~ - . ~ associated with Pyr6n6es). Open Triangles.9 leucogranite (Eyrein, Limousin)

10 ~
ide distribution: very low abundances, large negative Eu anoma-
.......i ' ly and contents of heavy R.E.E. increasing with atomic number.
L'a Ce Net Sm Eu Gd Tb Y'b Lu The present data indicates that analysis of lanthanides in
only one or a few samples taken by granitic complex would
give misleading information about its composition. However,
lanthanide patterns seem to be homogeneous inside a given mag-
D Mon zogranites
and Dioritic matic unit and can be used to characterize each unit of a compos-
inclusions ite intrusion.
10 2
---., ~._._ . . . . . . . . . . . . b) Lile
Ba concentrations in the marie series range from 90 to 750 ppm.
inc(usion ....... 2'ak. \ /
Qt 101 -1=~[s In granitoids, Ba decreases with SiO2 increase and has a maxi-
~10 ~ mum value around 800 ppm in granodiorites and a minimum
- - -2zV~5~.=..~.;. = around 300 ppm in the central biotite granite. Sr contents of
21024 E Central Biotite
marie rocks are higher (250 to 400 ppm) than in granitoids (250
~ Granite and to 25 ppm) with the exception of Cortlandites (Sr=70 ppm).
0~
O.
Porphiritic Rb increases in the mafic series (20 to 105 ppm) while Mg and
E Ca decreases, and continues to increase with SiO2 in granitoids
u)10~ (from 110 ppm in granodiorites up to 300 ppm in 2-mica gran-
ites). Cs in generally more enriched in granitoids than in the
marie series but is irregularly distributed. One marie sample
(dioritic inclusion Qtl3b) has a surprisingly high Cs content
/ Muscov!t?
100 Granitoids (23 ppm).

c) Transition Elements and Sc

b L'a Ce N'd Sm
' EuG'dTb Yb L'u
All transition elements are much more concentrated in marie
Fig. 4a, b. Chondrite normalized R.E.E. distribution in whole rocks rocks than in granitoids. Maximum concentrations for Ni, Cr,
from the Querigut complex. A, black circles represeht Cortlandite and
Co are found in Cortlandite (Ni-~850 ppm). These elements
Hornblende norite ; black triangles, Gabbro-diorites and black squares,
Quartz-Diorites. D, Black squares: isolated dioritic inclusions, open broadly decrease in the marie series with decreasing Mg and
squares: Monzogranites, open diamonds." Porphyritic Monzogranites. Ca contents (Ni-~ 840 to 9.5 ppm, Cr-~ 2000 to 25 ppm, Co-~ 100
E, the dotted area represents the range of variation for the Central to 20 ppm) and with increasing SiO2 content in granitoids ( N i~- 5
Biotite granite. Open circles represent the porphyritic microgranites to 2.5 ppm with one exception, central biotite granite (5.4 ppm)
for which a single measurement is available, Cr_~40 to 10 ppm,
Co ~ 9.5 to 1.7 ppm with a minimum value in 2-mica granitoids:
Type 3. The central biotite granite and the porphyritic micro- 0.3 ppm).
granites have extremely similar distributions, characterized by Sc does not display important variations in the gabbrodioritic
a lower enrichment of light R.E.E. compared to type 1 and series (21.7 to 26.6 ppm) with one exception: isolated microdior-
by an important negative Eu anomaly. itic inclusion Qt 13 b (47 ppm) while it decreases with increasing
Si02 content in granitoids ( ~ 10 to 3 ppm).
Type 4. The 2-mica granitoids with tourmaline nodules have
R.E.E. abundances significantly lower than the average granitic d) Incompatible Elements
rocks (Haskin and Paster 1979). Sample Qt42 displays a fraction-
ation amongst heavy R.E.E. and a negative Eu anomaly while The behavior of Th and U is roughly paralM. Abundances in
sample Qt85 of the same rock type has a very particular lanthan- acidic rocks are in general higher than in mafic rocks but vary
 183

Sa Ba inside a given family of rocks, particularly for U (e.g., U in

- ppm )pm
granodiorites varies from 1.2 to 4.2 ppm).
T h / U varies between 1.7 and 4.9 in the mafic series, go t h r o u g h
a m a x i m u m value a r o u n d 8 in granodiorites a n d monzogranites
and decrease in the more siliceous granitoids (0.6 in 2-mica
\
500 9 500 \\ granites).
\
9 \ A n i m p o r t a n t result for the forthcoming discussion is the
decrease of Th with SiO2 from monzogranites to central biotite
granite (mean contents): 16.7 a n d 14.3 ppm). In the mafic series
9 Sr eZ)Ni H f concentrations increase from Cortlandites, t h r o u g h gabbro-
ppm ppn

200
00-/o / B~
~"p p m
5()0
diorites, to quartz diorites (1.45. to 10.35 ppm), a n d then decrease
irregularly with increasing SiO2 in granitoids (6.1 to 1.8 ppm).
Ta increase in the mafic series from Cortlandites (0.19 ppm)
to quartz-diorites (maximum 1.23 ppm) and continues to increase
9\ O

100
?< Y
///~,\ 500 -
i
/I/A
/A
/11

I
iI
in the granitoid series (0.86 to 6.2 ppm).

11/ 9 II OOC e) Interelement Relationships

~,. 9 I 9
/
/ z/ /
i // In order to visualize the trace elements results, we give in Fig. 6
CS lee1 I I Ta
ppm ./ ppm several 2 by 2 correlation diagrams for elements pertaining to
, t , t ~ t
Ce 1() Hf 1 2 the same a n d different chemical families.
pprn
Yb Noteworthy is the fact that in all cases granodiorites are
I0- / !
,' closer to the basic trends than other acidic rock types.
// II The major characteristic in the mafic series is the dispersion
111 iI of samples while in some diagrams they display a more or less
continuous evolution. This dispersion appears also in the acidic
series for which rocks pertaining to a given type plot as clusters

17~11
/ //. ,/ /
of points. However the different granitoids seem to define b r o a d
linear trends in which the succession of types is the same : grano-
Eu 1~ ~ I ~ Ta diorites, monzogranites and central biotite granite. Relative to
Yb ppm
I I the basic series the acidic series shows contrasting trends in
1 inter-element correlations (Fig. 6):
Fig. 6. Interelement relationships between rocks of the mafic and acidic (1) Some element couples show continuous variations
types. Same symbols as Fig. 4 t h r o u g h o u t the two series, e.g., R b - C s .

Table 3. Trace element concentrations (ppm) in mineral separates

Rock Quartz-diorite Qt 04 Monzogranite Qt 12

Mineral Plagio- Biotite Horn- Apatite Sphene Zircon K-Feld- Plagio- Biotite Allanite Quartz
clase blende spar clase

La 8.0 0.82 22.0 535 870 / 1.48 8.14 2.42 3.72% /

Ce 9.95 1.77 66.1 1360 3010 / 2.13 11.65 4.77 6.61% /
Nd 2.52 0.50 58.1 880 2650 / 0.91 3.41 3.53 2.35% /
Sm 0.39 0.142 14.70 187 685 / 0.22 0.76 2.23 3510 /
Eu 0.98 0.061 1.97 34.2 118.1 / 0.98 1.28 0.088 53.4 /
Gd 0.39 / 13.4 139 470 / / / / / /
Tb 0.04 0.027 2.20 21.9 85.1 / 0.034 0.075 0.51 213 /
Yb 0.095 0.067 6.20 69.8 262 / 0.12 0.304 2.06 186 /
Lu 0.02i 0.017 0.98 11.70 37.7 / 0.023 0.053 0.34 25.8 /
U 0.10 0.51 1.17 / / / 0.31 1.14 2.80 / 0.072
Th 0.175 0.105 5.7 / 396 430 0.65 2.9 1.63 2.87% 0.03
Hf 0.11 0.26 2.72 / 130 5910 1.06 1.62 3.38 n.d. 0.04
Ta 0.01 0.69 0.72 / 405 0.9 0.055 0.12 11.9 n.d. 0.01
Cr 1.65 380 420 / 635 / n.d. 1.32 195 2550 0.12
Co 0.70 67 50 / 60 6.3 0.25 0.33 50.8 148 0.02
Ni 1.60 165 107 / n.d. 7.1 n.d. 0.21 26.5 n.d. 0.09
Sc 0.32 26.85 90.7 / 115 50 0.205 0.36 52.4 112 0.015
Ba 205 1580 55 / 500 630 1625 200 1860 / 0.45
Rb 7.1 425 13.5 / n.d. 6 375 35 715 n.d. n.d.
Cs 0.203 17.4 0.7 / n.d. 0.35 3.44 0.89 47.3 n.d. 0.045
184

Table 3 (continued)

Rock Quartz-diorite Qt 02 Granodiorite Qt 52 Porphyritic monzogranite Qt 18 Central biotite

granite Qt 15

Mineral Horn- Horn- Sphene K-Feld- Plagio- Bio- Horn- K-Feld-Global Plagio- Bio- Apatite K-Feld-Plagio- Bio-
blende blende spar clase tire blende spar K-Feld- clase tite spar clase rite
(dupl.) pheno- spar
crist

La 36.7 34.3 790 1.59 6.74 1.76 59.7 7.90 5.5 13.5 41.9 1110 2.33 15.2 43.8
Ce 129 122 2450 2.17 8.75 6.46 191 10.9 8.2 21.1 73.5 2360 2.95 24.4 85.0
Nd 135.7 134.0 1785 1.11 2.41 5.93 203 3.50 3.0 8.1 31.8 1460 1.22 12.0 41
Sm 42.4 40.8 420 0.23 0.46 2.10 63.7 0.82 0.51 1.55 7.0 440 0.31 2.2 10.35
Eu 3.01 2.96 72.54 1.02 1.40 0.215 2.63 1.29 0.82 0.87 0.255 15.40 0.41 0.95 0.30
Gd 31.1 34.0 285 / 0.43 4.07 70 / / / / 460 / / /
Tb 5.82 5.85 49.2 0.04 0.045 0.47 11.1 0.142 0.043 0.155 0.93 69.4 0.04 0.22 1.87
Yb 11.13 12.19 140 0.156 0.143 1.36 22.5 0.27 0.132 0.48 2.38 242 0.15 0.70 6.22
Lu 1.76 1.63 19.5 0.025 0.021 0.225 3,67 0.055 0.026 0.35 37.0 0.022 0.136 0.98

Quartz Diorite Qt 04 (2) Others have a common field but the basic and acidic
series diverge, e.g., H f - T a , B a - T a .
(3) Still others show no apparent relationships between series,
10 3 e.g. B a - Sr.

Analytical Results on Separated Minerals

and Consequences
10 2

Hb
The Rock Budget
We have analyzed the T.E. composition of major mineral phases
'o
and when possible, accessory mineral phases from various rock
~101 types. Particular attention was paid to the purity of mineral
separates. The results are given in Table 3 and some of them
are illustrated in Fig. 7.
E
Considering the lanthanides, the concentrations in coexisting
r 10 o minerals cover a wide range: some very enriched minerals (allan-
ite, sphene, apatite), some moderately rich minerals (hornblende)
La Ce Nd Sm Eu Gd Tb Yb Lu and some poor minerals (plagioclase, biotite, K-feldspar and
quartz). Note that biotite in porphyritic monzogranite (sample
Qtl8) appears rather rich in lanthanides but this may be due
AI to inclusions of apatite needles which were not completely re-
10 3
~,.~(dlvided }
moved.
In order to discuss quantitatively the general characteristics
of each kind of minerals, we have analyzed more completely
the trace element contents in separated phases from two major
~10 2 rock-types: quartz diorite Qt04 and monzogranite Qtl2. Using
0 the modal analysis, we have tried in both rocks to compute
PI the budget for every element. The results are illustrated in Fig. 8.
Some additional details should be given on this calculation. First,
I0 l Bt as the modal abundance of accessory phases is very difficult
(Y)
to determine accurately, we have evaluated it by the deficiency
observed in the rock budget for some particular elements specific
of these accessory minerals. H f and Ta were used to calculate
100 the respective amounts of zircon and sphene in quartz diorite.
The same procedure was applied for allanite in monzogranite
by using Th. Then, it was possible to calculate the contribution
of accessory minerals in the rock budget for other elements.
Secondly, the Th concentration in allanite Qt12 was not obtained
very precisely because of the small size of the sample. So, the
Fig. 7a, b. Chondrite normalized R.E.E. distribution in separated contribution of this mineral to the light lanthanides budget is
minerals from two selected rock samples: Sph, Sphene; Ap, Apatite; probably underestimated.
A/, Allanite; Hb, Hornblende; Bt, Biotite; P/, Plagioclase; KF, K- Taking into account the uncertainties of analyses (both in
Feldspar modal and chemical analyses) our computation explains reason-
 185

QUARTZ DIORITE Qt 04
ably well the total budget for each rock except for some elements %
for which a considerable deficiency appears. This is due to the 100
:::.:.:. ----
presence of accessory minerals in the rocks which were in such
small amounts that they cannot have been separated for analysis.
In quartz diorite, allanite appears in very small amounts by
microscopic examination and account for the deficiency of light
lanthanides, Th and, to a lesser degree, U as suggested by its
composition in monzogranite and the compositions given for so
example by Smith et al. (1957). The major part of the lacking
heavy lanthanides could be explained by the presence of zircon
known for its high content in those elements (Nagasawa 1970).
Similarly, in monzogranite, zircon would be responsible of the
deficiency in heavy lanthanides, Hf, Zr and in part U, and allan-
La Ce Nd Sm Eu Tb Yb Lu Rb Cs Ba Fe Cr Co Ni Sc Ta Hf Th U
ite for the other part of U.
Considering the mineral separate data, two kinds of trace MONZOGRANITE Qt 12
elements can be distinguished in the plutonic rocks studied: %
a) Trace elements whose behavior is mainly governed by loo
major phases (feldspars, hornblende, biotite). Such elements in-
clude Rb, Cs, Ba, Cr, Co, Ni, Sc, Eu, plus heavy lanthanides
in quartz-diorite and Ta in monzogranite.
b) Trace elements whose behavior is mainly governed by
accessory minerals: such elements are light lanthanides, Hf, Zr, 5o
Th, U, plus heavy lanthanides in monzogranite and Ta in quartz-
diorite.

Consequencesfor Trace Elements Modelling

The present results imply that modelling of the behavior of La Ce Nd Srn Eu Tb Yb Lu Na Rb Cs Ba Fe Cr Co Ni Sc Ta Hf Th U
trace elements based on major minerals only, a common practice
~ Plagioclase ~Hornblende [r[iTrnSph e n e ~Allanlte
in the past, is not valid. Such a conclusion is particularly clear
for lanthanides in acidic plutonic rocks and reinforces the com- r::=:~ K Feldspar ~ Biotite ~Apatite ~Zi ....
ments of Mac Carthy and Kable (1978) and Shaw (1977). Many
Fig. 8a, b. Cumulative trace element content of separated minerals, ex-
models of the origin of granitoids, which have been published,
pressed in percent of the whole rock abundances. The modal propor-
should be carefully reevaluated for these reasons. Hanson (1978) tions are: Monzogranite: Qz: 0.26, K-Feldsp: 0.18, Plag: 0.46, Biot:
has already pointed out the role of accessory minerals like apatite 0.094, Allan: 0.00052 Quartz-diorite: Qz: 0.08, Plag: 0.53, Hornb:
in the behavior of lanthanides during petrogenetic processes and 0.20, Biot: 0.185, Spheric: 0.00175, Apat: 0.0025, Zircon: 0.00065
more recently, a considerable progress was made by Watson
(1979) in including zircon in the model. However, zircon is not
Therefore, we will use two different approaches of the frac-
the only accessory phase which can play a role. We emphasize
tional crystallization process in the Querigut complex: first, the
here the fundamental role of allanite and spheric as carriers
discussion of trace elements which appear to be governed by
of trace elements in plutonic rocks. A similar conclusion was
major minerals. Secondly, we will follow in the mafic series
reached by Gromet (1979).
the evolution of trace elements concentrations in a single mineral
But the problem is not only to take into account the accessory
species throughout the whole mafic series. Hornblende was choo-
phases in the modelling of petrogenetic processes : we have also
sen, because as previously shown, it contains the most of the
to change the formalism of the calculations because trace ele-
heavy lanthanides in mafic rocks and because its partition coeffi-
ments abundances are not necessarily low in accessory minerals.
cients for R.E.E. are well documented (Schnetzler and Philpotts
Therefore, the use of the dilute solution theory and of constant
1970; Nagasawa and Schnetzler 1971; Zielinski and Lipman
partition coefficients is certainly not a completely correct proce-
1976; Arth and Barker 1976; Irving 1978). The trace elements
dure. These general behaviors will be treated in detail elsewhere.
concentrations in hornblende for various mafic samples are given
They are not yet very useful since most of the parameters we
in table 4 and the chondrite normalized patterns in Fig. 9.
need for those calculation are not yet experimentally determined.
The second difficulty is the fact that, in most cases, the
Consequences About Crustal Anatexis
trace elements which are enriched in accessory minerals are those
and More Generally Recycling Processes
known for their "incompatible" behavior in more basic mag-
mas; for example, R.E.E. (except Eu), U, Th, Ta, Zr, HI'. This Following the ideas put forward by Patterson and Tatsumoto
just reflects that accessory minerals should cristallize as soon (1964), many radiogenic isotope studies have suggested that re-
as the solubity of some elements is attained in the course of cycling of the sedimentary continental through metamorphism
solidification process. Such a behavior is Welt demonstrated by and anatexis greatly contributes to granitoid genesis. Before eval-
Watson (1979) for zirconium. Then, if we want to modellize uating the crustal recycling models, the knowledge of the behav-
trace elements evolutions in a differentiation process, we have ior of accessory mineral during sedimentary, metamorphic and
to compute the concentrations of T.E. which constitute zircon, anatectic processes is needed. Let us consider, for example, the
allanite, apatite, sphene, etc .... and determine when the solubili- alteration of continental crust: zircon, monazite and allanite
ty is reached. Such a program cannot be realized because of (Hutton 1950) and, to a smaller degree, sphene are extremely
a lack of data. resistant during chemical and mechanical alteration. Biotite and
186

Table 4. Trace element concentrations (ppm) in various hornblende separates from the Querigut plutonites

Rock Cort- Gabbro-diorites Quartz-diorites Granodiorites Dioritic

landite inclusion

N~ Qt 46 Qt 07 Qt 08 Qt 04 Qt 25b Qt 10 Qt 55 Qt 56 Qt 02 Qt 45 Qt 52 Qt 13b

La 10.5 12.7 17.7 22.0 23.7 29.5 28.7 30.2 35.5 43.5 60 34.8
Ce 30 51 61 66 76 88 101 103 126 164 191 151
Nd 18.5 45 50 58 59.5 59 90 82.5 135 177 203 162
Sm 5.2 12.1 12.7 14.7 14.6 17.0 24.8 22.0 41.6 55.5 64 62.5
Eu 1.41 1.48 2.34 1.97 1.46 2.91 3.16 1.89 2.98 2.27 2.63 1.97
Gd 5.1 9.8 i2.3 13.4 12.5 17.1 23.5 17.6 32.5 64 70 /
Tb 0.93 1.8 1.9 2.2 2.1 2.5 3.7 3.3 5.8 12.7 11.1 13.5
Yb 3.2 7.1 6.2 6.2 6.2 7.5 9.2 10.5 11.7 46.1 22.5 53.5
Lu 0.50 0.90 0.78 0.98 0.65 1.10 0.77 1.68 1.7 6.82 3.67 8.8
Hf / 2.0 2.8 2.72 2.7 3.8 3.1 3.1 4.7 / 11.5 4.6
Ta / 0.68 0.75 0.72 0.76 1.02 1.11 1.14 0.98 / 2.7 4.4
Co / 51 45 50 49 43 45 43 39 / 38 34
Ni / 252 125 107 51 20 32 28 17 / 4.3 27
Sc / 60 73 90.7 87 90 128 99 220 / 295 210
Rb / 30 10 13.5 6 6 31 9 8 / 17 18
Cs / 0.59 0.33 0.7 0.14 0.03 0.04 0.22 0.26 / 0.92 1.03

B Hornblendes from preferentially incorporated into the melt? The lack of experimen-
Granodiorites and tal investigation has restrictive consequences on the modelling
Dioritic inclusion
of the continental crust evolution by use of N d and Pb isotopes
tracers and of course in the modelling of lanthanides distribution
in the course of crustal anatexis.
= 10 2 . . . . . . ~-~--= =-~"~'1

c-
O
c-
Discussion
O
A Hornblendes from ~ As already quoted in the geological discussion, the major prob-
~)
EL mafic rocks lem set by the Querigut complex consists in a possible derivation
E of granitoids by differentiation of the mafic series. For the clarity
of the discussion, we will treat separately the case of the mafic
series and that of granitoids. So we will discuss in a first part
the genesis of mafic rocks, in a second part of granitoids and
finally, if some kind of relationships links the 2 series (as we
will see, this point is critical for the origin of granodiorites
J , , , , J , and tonalites).
La Ce Nd Sm Eu Gd T b Yb Lu

Fig. 9. Chondrite-normalized R.E.E. distribution in Hornblendes. a The Genesis of Mafic Rocks

From the mafic series (same symbols as Fig. 4). b From granodiorites The increase of R.E.E. abundances in mafic rocks (Fig. 4) togeth-
(black squares) and isolated dioritic dioritic inclusion Qt 13 b (black
er with a decrease of Cr, Ni, Co (Table 1) suggests that some
triangles)
kind of differentiation by crystal fractionation has taken place
within the mafic series. The regular decreasing Ni and increasing
l-If, Ta, and Sc concentrations in hornblendes from gabbro-
plagioclase are quickly destroyed as probably apatite, but K- diorites to quartz-diorites reinforces this hypothesis (Table 4).
feldspar is more or less resistant. Now, in the U - P b system, However, the interpretation of trace element chemistry in those
we know (see for example Allegre 1972) that if detrital sediments rocks is rather complex because we do not know " a priori"
are submitted to metamorphism and anatexis, granitoids may if they represent liquids at different stages of evolution, or cumu-
be produced with isotopic ratios somewhat different from that lates, or possibly products of mixing between cumulates and
of parent crust because partial melting is not necessarily an some residual liquid. The possible participation of accessory
equilibrium process. The accessory minerals which contain the minerals with very high partition coefficients for trace elements
inherited radiogenic information such as zircon, monazite ... can to the crystal fractionation makes its modelization more com-
partly remain in the unmelted residue without being equilibrated plex.
with the liquid (Gulson and Krogh 1973). For the S m - N d Therefore, we have calculated the composition for R.E.E.
system, the behavior of the inherted radiogenic information of liquids in equilibrium with a reference mineral, i.e. horn-
is insufficiently known because of a lack of data concerning. blende. This has a double interest: in mafic rocks, it is both
the behavior of lanthanide-rich accessory minerals in detrital and the major carrier of R.E.E. and a major mineral phase. The
anatectic processes. Are they resistant to alteration? Do they evolution of R.E.E. abundances in these calculated liquids could
stay in the unmelted residue without equilibration or are they give informations about that in the residual magma if one as-
 187

1]-300
B Granodiorites
o
ffl
"F.
":'x Oe-
(9
:xx t00
O.

~~"~ x'xx,. [ r 101

":5 50

,'at; N'd S'~iuG'd ib vi, i.

Fig. 11. Chondrite-normalized R.E.E. distribution in calculated liquids
~200- :l:l ~, ~ ~:-.% Z~.~.v:-:7.:r.r
_.w/ corresponding to the differentiation in the mafic series, f represents
the fraction of liquid. The shaded area is the composition range mea-
sured for mafic rocks (except quartz-diorite Qt02)
-~ I00 10
y roo

\\ // because we do not know the exact partition coefficients govern-

ing our particular series and because the proportions of crystalliz-
ing phases can vary, but it explains reasonably well our measure-
\\
ments (Fig. 11) except the most evolved rocks (quartz diorites).
The latter rocks can have suffered contamination by acidic mate-
\ rials. Indeed, in the C e / Y b - E u / Y b diagram (Fig. 6), a process
of fractional crystallization with small solid/liquid partition coef-
lO ficients does not produce important variations, that is observed
for the major part of mafic samples. On the contrary, the points
corresponding to quartz diorites (Qt04, Qt02) are shifted towards
\\ the acidic rocks, indicating some contamination of mafic rocks
\ by the host acidic magma. Indeed, a mixing process would be
traduced by a linear correlation i n that kind of diagram. We
may infer from these calculations that no important Eu anomaly
L'a C'e Nd ' Srn G'd Tb E'u Y~) 1'U developps in the liquids. This could be surprising as ptagioclase
Fig. IO. Chondrite-normalizeddistribution of calculatedliquids in equi- is a dominant phase in the gabbrodioritic rocks. In fact, some
librium with hornblendesA) from the mafic series;B) from granodio- partition curves obtained on hornblende (e.g., Arth and Barker
rites. Same symbols as Fig. 4 1976) display a significant negative Eu anomaly. So, crystalliza-
tion of hornblende could counterbalance in some degree the
effect of plagioclase fractionation. Moreover, plagioclase occur-
sumes that the hornblende fraction contained in a rock sample ing in the gabbrodioritic series is very calcic (especially the core)
does not integrate too large an interval in the degree of crystal and it was shown by Schnetzler and Philpotts (1970) and by
fractionation. The normalized pattern of calculated liquids is Dudas et al. (1971) that the Eu discrimination by this mineral
given in Fig. 10. For calculations, we have used the set of parti- depends upon its anorthite content. This justifies our choice
tion coefficients in agreement with the major element chemistry of partition coefficients from andesites because the usual compo-
of each kind of rock: the average of Arth (1976) on basaltic sition of plagioclase in those volcanics is close to that in our
compositions for cortlandite, those of Arth and Barker (1976) basic series. Moreover, the oxygen fugacities admitted for the
on dacitic composition for the gabbrodioritic series, and those crystallization of andesites are rather high (see for example Os-
of Nagasawa and Schnetzler (1971) on rhyolitic composition born 1962; Fudali 1965 amongst many others). Such conditions
for granodiorites. are suggested by the presence of ilmenite and magnetite in the
gabbrodioritic series.
We want to point out that accessory minerals were not taken
a) The Gabbrodioritic Rocks from the Large Stocks
into account in modeling crystal fractionation. As shown in
Fractional Crystallization. The evolution of R.E.E. patterns in Fig. 4, the R.E.E. patterns of sphene is very similar to that
the gabbrodioritic series is marked by an enrichment of abun- of hornblende, but with higher abundances. So, a small amount
dances, no significant fractionation between light and heavy ele- of sphene crystallizing from the magma could produce the same
ments and no Eu fractionation (the very small Eu negative anom- effect as hornblende. In fact, sphene is a late phase in these
aly remains grossly constant). These facts are the constraints mafic rocks and seems to result from subsolidus recrystallization
in the modelling of crystal fractionation, together with the miner- involving hornblende and biotite (Fourcade and Villemant, un-
alogy of mafic rocks. They imply that bulk solid-liquid partition published data). The experiments of Prince (1943) on the system
coefficients are lower than unity and nearly equal for all the albite-anorthite-sphene, clearly demonstate that in magmas of
R.E.E. With a cumulate composed of Hb: 0.35, Bt: 0.15, Plag: basaltic to dioritic composition, sphene can appear only in the
0.50, which is reasonable considering the mineralogy of mafic very late crystallization stages. So, crystallization of sphene, if
rocks, we can explain the evolution of R.E.E. in the mafic series it has occured at the expense of liquid, would have taken place
by using the mean partition coefficients compiled in Arth (1976) after the "rock" was isolated from the whole magmatic system.
for andesitic compositions. This modelization is evidently crude Although no detailed experimental data is available at present,
188

the same conclusion would be applied to allanite in such basic Its R.E.E. abundances (Fig. 4) do not display significant dif-
rocks. ferences with those found in the gabbrodioritic series. This rock,
The situation of cortlandites and hornblende-norites within as the border zones of some large mafic masses in contact with
the suite may now be examined. Leterrier (1972) suggests that granitoids, exhibits a fine-grained subdoleritic texture. This fea-
cortlandites could represent an early cumulate of a tholeiitic ture is interpreted as resulting from the chilling of the gabbro-
basalt evolving towards a high-A1 basaltic magma giving rise, dioritic magma at the contact with the colder acidic one (Didier
after hydratation, to the gabbro-dioritic series. Our measure- and Lameyre 1969). This inclusion is thought to represent a
ments do not contradict such an interpretation: The very high piece of stoped-in chilled margin as commonly occurs in the
Ni content found in cortlandite Qt46:840 ppm (Table 1), is field.
consistant with the presence of cumulate olivine in that rock. The second sample (Qtl3b) is one of the numerous small
However, the hydrated character of the gabbrodioritic series (-~0.25 dm 3 in size) enclaves found in the monzonitic granites
(exclusive presence of hydroxyl-bearing minerals) can be inherit- far from any large mafic stock. It represents the typical basic
ed from an independant parent source and we cannot exclude microgranular inclusions of quartz dioritic composition system-
that cortlandites and gabbro-diorites have independant origin. atically encoutered in granitoids of granodioritic affinity (Didier
This is suggested by the field occnrence of cortlandites which 1973). Its R.E.E. pattern is drastically different from those of
are embedded in the core of large gabbrodioritic masses as angu- the gabbrodioritic series and is dominated by that of hornblende
lar blocks with sharp contacts. So, they could represent xenoliths (Figs. 4 and 9) which displays a strong negative Eu anomaly.
dragged by the" gabbrodioritic" magma during its ascent, which This particular pattern cannot represent that of a basic liquid
are not necessarily related to the early differentiation stages. because it has no analogy with those found in large mafic stocks.
It cannot also be the consequence of an equilibration between
Are Dioritic Rocks Representative of Solids or Liquids? In our the mafic inclusion and a liquid (for example, the granitic mag-
modelling of differentiation inside the mafic series, we have con- ma) because the measured Eu negative anomaly is not consistent
sidered dioritic samples as representing successive liquids, that with the presence of plagioclase in the inclusion. Thereby, we
needs justifications. A mutual equilibration of REE contents be- are inclined to interpret this pattern as the result of a solid-state
tween the various mineral phases is likely to have proceeded reequilibration between minerals from the mafic inclusion and
much after isolation of the "rock" from the parent magma, those of the enclosing granitoid. Indeed, our measurements of
as indicated by (1) uncommon REE patterns of liquids in equilib- R.E.E. concentrations in coexisting feldspars from granitoids
rium with hornblende (Fig. 10), compared with whole-rock pat- (Fig. 7) demonstrate that Eu is more strongly discriminated by
terns (Fig. 4) (2) correlation of the Eu negative anomaly in horn- K-feldspar than by plagioclase. Then, the equilibration of the
blende with plagioclase content of the sample. So, we must ad- mafic inclusion with the granitoid containing much sodic plagio-
mitt that a considerable amount of liquid is trapped with the clase and K-Feldspar would produce a depletion of Eu in the
crystallized phases and isolated from the whole magmatic system whole rock and in the hornblende. In our opinion, this reequili-
to produce a "rock" sample. This interpretation is strongly bration would have taken place by means of fluid exchange
supported by the work of Shimizu et al. (1978) who analyzed between the two rocks.
hornblende crystals inside one dioritic sample from the Querigut This is suggested by the very high Cs content of the mafic
complex (sample Qt04) using an ion probe. They found a zoning inclusion (Table 1), an element which is known to be very mobile
of R.E.E. inside hornblende and an Eu depletion correlated when a fluid phase is present. The small size of the inclusion
with a decrease of the Mg/Fe ratio. This likely reflects a crystalli- can be responsible for the efficiency of exchange, because the
zation in closed system mainly governed by plagioclase and horn- phenomenon is not observed for the largest previous inclusion.
blende, perhaps following some kind of Rayleight distillation Now, if this interpretation is correct, it is obvious that the R.E.E.
process. pattern of the inclusion cannot give informations on its origin.
These considerations support the idea that dioritic samples However, its subdoleritic fine-grained texture, suggests that it
are more likely representative of liquids rather than cumulates. can also represent fragments of the border zones of the mafic
stocks, disrupted and disseminated into the acidic magma.
Parent Material. Now, how can be identified the parent magma
of the gabbrodioritic series? If we compare the R.E.E. abun-
dances in the most "primitive" samples from the gabbrodioritic The Granitoids
series to those present in volcanic series, we come to the conclu- Several "a priori" genetic models can be considered for the
sion that they have similarities with some low-Si andesites (Tay- origin of granitoids (Fig. 12). If we consider the diagrams illus-
lor 1969), or with some high-A1 basalts, (e.g., Masuda 1968), trated in Fig. 6, granitoids do not plot at random but seem
or with continental basalts (Haskin et al. 1968; Brannon et al. to define a trend for many couples of elements. Here, we point
in Haskin and Paster 1979) although no definite answer can out an important difference between plutonic and volcanic com-
be given to this problem with the present data. However, the plexes: in volcanism, the frequent sampling of the magmatic
low Ta content in the most primitive gabbrodiorites could indi- reservoir by volcanic eruptions permits to follow a more or
cate the presence of some Ta-bearing minerals in the source re- less continous evolution in the chemistry of the liquid (see for
gion. Sphene cannot be this residual phase because it is not stable example, Schilling and Winchester 1967; Allegre et al. 1977).
under mantle conditions in hydrous mafic materials (Hellman On the contrary granitoids, in general, represent the end-prod-
and Green 1979). ucts of the chamber itself which may consist of various mixtures
of cumulative solids and crystallized liquid. In the Querigut com-
b) Small-Sized Mafic Inclusions in Granitoids
plex, the compositions of each kind of granitoid appear rather
The only two samples analyzed are representative of two different constant and plot as clusters of points in the diagrams of Fig. 6.
settings. Sample Qtl01 was sampled in a large (0.25 m 3 in size) In this case, more extensive sampling inside a given magmatic
isolated inclusion in monzonitic granite not far from large mafic unit do not bring more informations on an eventual relationship
stocks. between the different rock-types.
 189

contamination by acidic materials was likely in quartz diorites.

u
' ~ 'YM Simple mixing without superimposed differentiation is apparent-
i
' ~"6
-" ." ; o ly not adequate to describe the whole Querigut complex.
T
C G DT'Yg'Y,v, u
D
G
C
Hypothesis C." The Acidic and the Basic Series
Constitute Two Independent Series of Differentiation
cl

--
~ c

?/,5
~G
D
-(3OO
T ~MYc
Y5
It could appear surprising to consider the effects of crystal frac-
tionation in acidic magmas, as the calculations of Shaw (1965),
Bartlett (1969), and Ewart et al. (1975) converge to the same
conclusion: the high viscosity of acidic magmas and the general
occurrence of convection in natural magmatic chambers renders
4r C
the process of crystal settling very unlikely. However, crystal
e Basic magma
fractionation in acidic melts is still possible by other processes.
9 crustal
For example, Bateman and Chappel (1979), explain the differen-
tiation of the Tuolumne intrusive series by a centripetal accretion
T magma
Differentiatiln of crystalline materials to the margins of the magma chamber.

F Contamination by
crustal a c i d i c
magF:la$
Granting that crystal fractionation is a possible process for differ-
entiation in the Querigut complex we next examine our data
in light of partition coefficients compiled for major minerals
in rhyolites by Arth (1976).
Fig. 12. Schematized genetic processes considered for the whole Queri- The following trends are readily explained by major minerals
gut complex. C Cortlandite; G Gabbro-diorites; D Quartz-diorites;
fractionation:
T Tonalites; y6 Granodiorites; 7M Monzogranites; 7c Central biotite
granite Decrease of Ba content by K-Feldspar crystallization;
Decrease of Sr and Eu contents by crystallization of feldspars ;
-

- Moderate increase of Rb in residual liquids. Indeed, biotite

Thus, the "trends" observed for granitoids are poorly de- has a high partition coefficient for Rb, but is not in sufficient
fined. We can discern the group of granodiorites, a group com- amounts in our granitoids to produce bulk solid/liquid partition
posed of both monzogranites and the central biotite granite coefficients higher than unity.
which is quite distinct. All these evolutions are observed in the acidic series from
In the foregoing discussion of genetic models, we are faced granodiorites-tonalites to the central granite (Table 1 and Fig. 6).
to an alternative: In such a way, granodiorites-tonalites with positive Eu anomalies
- the "trends" in chemical evolution of the granitoids is fortu- could contain some cumulate plagioclase, and the central granitic
itous and has no genetic significance. body could be close to a residual liquid. However, crystal frac-
the ~ trends" are significant and reflect some common genetic tionation of major mineral phases cannot easely explain the
process. variations of all trace element contents in the different granitic
units. This can be due to the participation of accessory minerals
Hypothesis A . Mafic Rocks and Granitoids Are Issued which concentrate trace elements during the process of crystal
from the Differentiation of a Single Type fractionation. For example, let us consider the decrease of Th
of Basic Magma and light lanthanides from monzogranites to the central granite
(Table 1) : it would correspond to bulk solid/liquid partition coef-
A pure derivation of the whole complex by differentiation of
ficients higher than unity. Allanite which occurs in monzogran-
a single type of basic magma is unrealistic :
ites can account for this effect: using the relative trace elements
1) Diagrams of Fig. 6 do not show well defined continuums
concentrations in minerals from monzogranite Qtl 2, we can eval-
and clearly discriminate between the mafic and the basic series.
uate partition coefficients of allanite for Th and La, assuming
In this regard, we stress that the very smooth variation in major
it is in equilibrium with coexisting minerals. Now, if we separate
element composition, particularly the AFM plot of Fig. 3, cannot
a solid having, for example, the mineralogical composition of
be taken as evidence for simple differentiation. Such smooth
monzogranite Qtl2 from a liquid, bulk solid/liquid partition
variation result in part from normalization of the oxide sums
coefficients would exceed 1 for Th and La with a small amount
to 100% (Skala 1979).
of allanite (10 3).
2) Isotopic data (Sr, Nd, O) display important variations
between the different rock-types (Fourcade et al., in So, differentiation in granitic melts can be accompanied by
a decrease of light lanthanides contents. One can object that
preparation).
allanite is often found in pegmatitic veins and can be a very
late phase in the crystallization of granitoids as suggested above
Hypothesis B." The Whole Complex Results
for mafic rocks. In fact, its euhedral shape in monzogranites,
from Simple Mixing Between a Mafic Magma
its optical zoning and its enclosure in large feldspar crystals
and an Acidic Crustal Melt
indicate this is not the case here. The enclosure of accessory
Such a process should result in straight line correlations in the minerals inside growing crystals of higher size explains they can
diagrams of Fig. 6. Clearly, concentration maxima for some trace play a role in crystal fractionation, whatever the process of sepa-
elements such as Ba, Hf, La that corresponds to rocks of interme- ration between solid and liquid can be. Allanite would be a
diate major elements chemistry (quartz diorites, tonalites, grano- critical mineral in this respects during the crystallization of gran-
diorites), are incompatible with this. On the other hand, we itic magmas.
have seen in previous sections that the major phenomenon occur- However, the evolution of lanthanides, Hf, Ta, Th from
ring in the mafic series was magmatic differentiation even if granodiorites associated with mafic rocks to other granodiorites
190

and finally to both monzogranites cannot be easily explained concerning the behavior of accessory minerals in anatexis. How-
by any plausible combination of mineral species during crystal ever, the trace elements bring important constraints in the genesis
fractionation. Such a process is particularly inadequate to explain of those granitoids: the high Th contents (Th= 15.3 to 19.6 ppm)
why granodiorites spatially associated with mafic rocks differ as the high Sr initial ratios of 0.710 (Vitrac and Allegre 1975)
from other granodiorites by their lanthanide distribution (Fig. 4). suggests a crustal origin. The lack of important negative Eu
On the other hand, if the granitic sequency is the result of anomaly indicate that feldspars were not an important phase
magmatic differentiation, this process cannot have occurred "in in the unmelted residue. The flat heavy R.E.E. pattern suggests
situ" for at least two reasons: first, a differentiation by crystal that the parent material was not containing garnet as an impor-
settling is not compatible with the great number of mafic inclu- tant phase in the residue but more likely hornblende or (and)
sions contained in the granitoids. Secondly, an "in situ" differen- clinopyroxene. The source material could be a metagreywacke
tiation is not compatible with the homogeneity of the different in which both volcanogenic and old crustal materials are recy-
granitic units for major and trace elements (if we exclude grano- cled. This is in agreement with isotope data (Ben Othmann et al.
diorites-tonalites associated with mafic rocks). in preparation). Such a source material is also invoked by Vitrac-
Although magmatic differentiation cannot explain all the Michard et al. (1980) for the Maladetta Pyrenean complex.
data, it cannot be excluded as the main process generating the The origin of granodiorites and tonalites intimately inter-
acidic sequency. But in this case, we must admitt it has occurred mingled with mafic rocks is rather obscure. Indeed they present
in a magmatic chamber located deeper than the level of the small but significant positive Eu anomalies (Fig. 4). Now if a
intrusion. granodioritic magma has existed, its particular Eu content is
hardly conceivable in terms of partial melting of crustal sources.
The extraction of an Eu depleted liquid (possible represented
Hypothesis D: The Different Granitoids by the central granitic body) during a first melting event can
Have "Independant" Origins produce an Eu excess in the unmelted residue. Larger degree
In this case, the trends defined by granitoids in diagrams of of partial melting of this residue can produce liquids enriched
Fig. 6 are considered to be fortuitous. As shown by their Sr in Eu. However, this process must be ruled out because magma
initial ratio (Vitrac and Allegre 1975), the Querigut granitoids corresponding to the central granitic unit is the last to be em-
mainly derive from crustal materials. So, what can be the crustal placed and intrudes granodiorites and tonalites as dykes.
materials which have undergone partial melting and in what Non equilibrium partial melting in which feldspars are prefer-
degree? Fortunately, the Pyrenean tectonism has cut off the entially incorporated into the melt without being equilibrated
Hercynian basement into differently uplifted blocks. So, the with the residue may be another possibility for generating liquids
structure of a good part of the Hercynian crust from that area with positive Eu anomalies.
is rather well known. An interesting feature of the region is All these possibilities call for "ad hoc" assumptions about
the occurrence of large massifs of Cambrian granitoids orthog- the source region of the melting process and for that reason
neissified during the Hercynian orogeny. Particularly interesting are open to criticism. In our opinion, the particular lanthanides
are the Rietes gneisses occurring in the major part of the Aston- distribution in these rocks must be related to their particular
Hospitalet massif (Raguin 1977) because they appear as a large spatial distribution, i.e. association with mafic rocks. More pre-
uplifted anatectic dome in the prolongation of the Querigut com- cisely, such lanthanide distributions could result from an equili-
plex (Fig. 1). The chemical composition of Aston orthogneisses bration with mafic rocks. Indeed, the parameters governing the
is very similar to that of the central unit of the Querigut complex final concentration of a trace element in granodiorites during
(Table 2). The similarity is much more pronounced for lanthan- a process of equilibration are: - Initial concentrations in the
ide concentrations. However, this agreement does not exist for 2 kinds of rocks. Partition coefficients between mineral phases.
other trace-elements (Table 1). This is not surprising because - Mineralogical composition of both kinds of rocks. - Respective
these gneisses are strongly anatectic and have suffered muscoviti- amounts of mafic and acidic rocks during equilibration. If the
zation which can have disturbed the distribution of mobile ele- granitoids are the most abundant, the mafic rock would acquire
ments. Hence, a near total melting of the Aston gneisses to a lanthanide distribution dominated by its dominant mineral:
produce the central unit of the Querigut complex appears plausi- hornblende. Such a case was previously discussed for the isolated
ble. dioritic inclusion Qtl3b. On the contrary, if granodiorites are
The idea of total melting of crustal materials is not often less abundant compared to mafic rocks, they will acquire a lan-
invoked in generating granitic magmas because liquidus tempera- thanide distribution dominated by their most abundant mineral:
tures are too high to be found in regional metamorphism. Howev- feldspar. Figure 13 shows the lanthanide pattern calculated for
er, magmas from the basic series can bring the heat for a high a mineralogy similar to that of granodiorites-tonalites assuming
degree of melting. Moreover, if an important proportion of a they are equilibrated with dioritic rocks in excess. Some differ-
rock of granitic composition is melted, it is conceivable that ences in absolute abundances compared to measured values can
residual refractory material is carried upward with the anatectic reflect the uncertainty on partition coefficients used for calcula-
liquid, as crystal mush (Piwinskii 1968; Presnall and Bateman tions and of course the degree of equilibration: The abundance
1973; Wyllie 1977; White and Chappel 1977). of mafic and acidic materials equilibrated can vary for various
Now, what can be the origin of monzogranites? Their lan- samples. Such a process does not imply that equilibration has
thanide distributions are extremely similar to that of post Ar- obligatorily taken place between crystallized rocks. A granodior-
chean Shales (NASC) (Haskin et al. 1966; Nance and Taylor, itic liquid carrying an important amount of feldspar crystals
1976). Haskin and Paster (1979) explain the general NASC type would led to the same result.
lanthanides distributions by inheritance from igneous precursors. Hence, the original magma corresponding to these particular
Young granitoids like the Querigut complex may, in turn, have rocks could have been that of "normal" granodiorites or monzo-
inherited their lanthanide distribution from a crustal parent granites present in larger amounts in the Querigut complex.
source. As we have already said, the modelling of partial melting Hypothesis D seems compatible with many aspects of the
of metamorphic rocks is difficult because of the lack of data Querigut complex. It explains the spatial relationships of the
 191

where subscripts m, A, B refer to the mixture, component A

~ 10 2 and B respectively, x being the weight proportion of component
A.
O
If we plot Cm~- CB~versus C ~ - Cs,i a straight line would result
-X whose slope gives the mass proportions of the mixture. Two
~10 ~
forbidden domains exist in this diagram, delineated by the slopes
0 and 1. Figure 14 gives the result of this test for the Querigut
granitoids. Major elements taken from Leterrier (1972) obey
Ago ,~. Ai~ ,;i. the mixing relation, but some trace elements do not match the
Fig. 13. Lanthanide distribution (dotted line) calculated for a granodio- model, particularly Th and Ni. Simple mixing of two poles is
rite in equilibrium with mafic rocks in excess. Partition coefficients not very consistant with our results unless we admit a third com-
used for calculations are those of rhyolites (Arth I976). The composi-
ponent. Such a component can be represented by the mafic rocks.
tion ofmafic rock (full line) as been taken as sample Qt 04. Mineralogy
Therefore the mixing hypothesis cannot be entirely ruled out on
of the rocks are: granodiorite: Qz 0.16, K. Feldspar 0.110, Plag. 0.544,
Biot. 0.120, Hornbl. 0.066; Mafic rock: Qz 0.115, Plag. 0.54, Biot. the base of trace elements only.
0.18, Hornbl. 0.165. (after Leterrier 1972)
Genetic Link Between Acidic and Mafic Series
As discussed previously, the whole Querigut series cannot result
intrusive units. Magmas generated in different source regions from a single mixing process between an acidic and a basic
would be emplaced as individual, chemically homogeneous intru- pole. However, in diagrams given Fig. 6, granodiorites-tonalites
sions. which appear to be the granitoid type the closest to the mafic
series, plot either in continuation to the field defined by the
Hypothesis E. The Different Granitoids Are Generated mafic samples or distinctly outside. Then, is there a contribution
Independently but Mixing Has Occurred Between of the mafic series to the generation of granitoids, especially
the Acidic Magmas for granodiorites and tonalites ? A direct contamination of granit-
oids by mafic rocks by means of direct mixing should be ruled
The main granitic types define in diagrams (Fig. 6) " t r e n d s "
out because in diagrams of Fig. 6, trends obtained for acidic
in which the sequence is always the same: granodiorites-monzo-
and basic series are in most cases divergent. The Ni content
granites-central granites. These trends possibly represent mixing
of granitoids is too low, compared to that of mafic rocks (Ta-
lines and monzogranites (which are the most abundant rock-
ble 1) to admit a general contamination by mafic rocks. Howev-
types) can result from a mixing between magmas corresponding
er, the existence of selective exchanges between mafic rocks and
to granodiorites and to central biotite-granite. Indeed, the com-
enclosing granitoids has been emphasized by Leterrier and De-
position of monzogranites for major elements (Table 2) is inter-
bon (1978) in the Querigut and the other Pyrenean complex
mediate between that of granodiorites and central biotite-granite
of Cauterets. Such a process which has been previously discussed
and they constitute the median units in the complex. Now, if
can have played an important role in determining the chemical
mixing occurs, each element would be concerned by the process
characteristics of granitoids.
and we can write for each element:
So, if granitoids are not directly contaminated by mafic rocks,
C i~ - CB=x(CA--Cs)
i i i
is a very differentiated end product of the mafic series able

;!:, )i / to 9
/
: !:}];i ?ii i /
/
.i al;~:.,~j}~:i!7< :; :i ~i} }:~ /
:::]! {;:: ?;v;7:?::~ : 1 /
! :{!(~!:::";t !i,t ::~: O'Y

/
'::.,::tt::.:::c.;: . : ?:5r ::,: ?:;;tiq: f.: / 1- ~
; :::: ?: : : :;i ~!:h ~ / /7 / / j i /~
: -: .} ~ i / / /
i ?::; : ::)}'y /
..... . < O.~,,,-
: ;; 2:; : ' : /" /
:: / /. ~"
:;t:; ?:: / .1

/ i9 . / ~"

//~ //~'~;~ @ / z z :
/ i / . ;!{7;;;;::::)::;i7;!
z
Fig. 14. Mixing test between the main granitic
Si i /
types from the Querigut complex: Cya, CyM,
C~/c respectively represents the concentration of
element i in granodiorites, in monzogranites,
and in the central biotite granite. The shaded
area corresponds to the domains where the
mixing model is not verified
192

to produce the whole or a part of the acidic magmas? This controversed: Flynn and Burnham 1978; Wendland and Har-
hypothesis is contradicted by the lack of transition terms between rison 1979). Anyway, extraction of lanthanides from granitic
the mafic series and granodiorites for some trace elements such rocks by means of fluids appears possible (Alderton et al. 1980;
as Sc (Fig. 6), Moreover the SiO2 content of the most evolved Taylor and Fryer 1980). The aplitic vein Qt85 could represent
rock from the mafic series (Qt02) is considerably lower than a late differentiate of the central granitic body, strongly alterated
in granodiorites (53.7% against 64.9%) (Table 2). So an impor- by interaction with magmatic fluids.
tant amount of crystal fractionation is needed from quartz-dior-
ites to granodiorites. As shown in Fig. 10, a very strong crystal Conclusions
fractionation within the mafic series would produce Eu depletion
in the residual liquids. This depletion would be enhanced because Several general conclusions can be inferred from our study:
the plagioclase amount of the crystallizing phases would increase 1. The quantitative modelling of trace elements in granitoids
for high degrees of differentiation and because this plagioclase is much more complex than in basaltic volcanic suites. In addi-
would become more sodic (Dudas et a1.1971) than in our calcula- tion to major phases, accessory minerals can play a fundamental
tions based on partition coefficients in andesites. In fact, neither role.
granodiorites nor monzogranites display the characteristics of 2. The Variscan calc-alkaline granitoids from Pyr6n+es are
residual liquids. On the contrary, they have positive or no Eu not produced by the differentiation of a "basaltic magma".
anomalies. The same reasoning can be made for Hf and Th: Nevertheless, such a magma is suspected to play a role as a
the most evolved quartz-diorite in the mafic series (Qt02) con- heat source in generating the melting of continental crust.
tains more Hf and Th than granodiorites (Table 1), although 3. The granitoids themselves are mainly derived by melting of
an extension of magmatic differentiation after the stage of quartz continental crust (which possibly includes some old component
diorites would increase the amount of Hf and Th in the residual of mantelic origin). Secondary differentiation processes may also
liquid (Treuil and Varet 1973). Therefore, if a very differentiated be required. The scale at which magmatic differentiation occurs
end product of the mafic series plays a part in the generation is not yet well-fixed but can be quite large.
of granodiorites, we must admit it was swamped by larger 4. The depth at which the different granitic magmas are
amounts of acidic magmas of other provenance. This agrees generated (either in the hypothesis of magmatic differentiation
with the study of Leterrier (1972) who estimates that the chilled or in the hypothesis of different melting sources) is, in any case,
margins encountered in mafic rocks at the contact with granit- greater than the level of intrusion. Then, the concentric character
oids, imply that both acidic and basic magmas were present of the complex is a result of multiple intrusive episodes.
at the same time. Strictly speaking, our study applies to the Variscan calc-
alkaline plutonic suite from Pyr6n6es. However, the study of
Gromet (1979) on the Peninsular Ranges batholith suggests that
Subordinate Granitic Varieties some of our conclusions may be of more general application.
This study by itself cannot give a unique answer regarding the
The origin of porphyritic dykes of calc-alkaline granitoids cutting
geodynamics settings for the Variscan calc-alkaline plutonism.
in some places the granodiorites is clearly demonstrated by the
A detailed stable and radiogenic isotope study in progress will
analysis of trace elements. The texture of these rocks in a single
hopefully resolve some of the questions that have been raised.
dyke ranges from medium-grained-porphyritic to fine-grained-
porphyritic as thickness of the dyke decreases. Trace elements Acknowledgements. We are very gratefull to Dr. M. Treuil, J.L. Joron,
and lanthanide contents in a medium-grained and a fine-grained and H. Jaffrezic for the welcome in the "laboratoire Pierre Siie",
type are almost identical and strictly similar to those of the groupe des Sciencesde la Terre, Saclay, France, where they introduced
central biotite granite (Table 1). These dykes represent portions the senior author to the techniques of neutron activation analysis.
of the central unit magma intruded in granodiorites and rapidly We are particularly indebted to F. Albarede, A. Autran, J. Leterrier,
cooled because of their moderate thickness (~-200 m). This fact M. Semet, and N. Shimizu for stimulating discussions and to J. Leter-
confirms the field evidences: the central biotite-granite was em- rier for analyses of major elements.
placed after the other main granitic units.
Appendix. Brief mineralogical description of the studied samples
The second subordinate type of granitoids is represented by
muscovite-bearing granitoids with tourmaline nodules. Field re- Co rtlandite
lationships indicate they were emplaced after the central granitic - globular, fractured crystals of Mg olivine included
unit as stocks or veins. The decrease of lanthanides concentra- in pyroxene
tions in aplitic, tourmaline-bearing vein Qt85 compared to the - orthopyroxene (more or less euhedral),
central granite is somewhat abrupt and has been often reported partly altered.
in the most siliceous granitic varieties associated to larger granitic QT46 - corroded clinopyroxene
bodies (Koljonen and Rosenberg 1974; Emmermann et al. 1975; - abundant poikilitic hornblende with variable
Fourcade et al. 1977, Cocherie et al. 1977). Such a pattern could pleochroism (brown in core and light green to the
result of strong interactions with a fluid phase. Aplites are con- edge). It gradually passes to a tremolitic fibrous
sidered as related to the crystallization of a H20-saturated mag- amphibole at the edge
ma, the separation of a fluid phase leading to the formation - abundant mica with very variable pleochroism,
ranging from phlogopite to biotite
ofpegmatites. Expulsion of a fluid phase from the parent magma
- accessory minerals: rare sphene included in mafic
of aplitic vein Qt85 could produce a depletion of light R.E. minerals; picotite ; rr~gnetite and ilmenite in joints
abundances. However, after Flynn and Burnham (1978), Webster and cleavages of minerals
and Holloway (1980), distribution coefficients of lanthanides are
always in favor of the silicate melt in experiments conducted Hornblende norite
with C1 or F-rich aqueous vapor phase. But natural fluids can - Same mineralogy than Cortlandite but no clinopy-
contain other complexing agents such as B, SO4..., whose be- roxene, less abundant olivine and more poikilitic
havior on lanthanides is not documented (the role of COz is hornblende
 193

QT104 Plagioclase in large subeuhedral crystals (core An 85- Qt83 rite), apatite as acicular prisms (often included in
70, edge with recurrent zoning, An 55-38) biotite or in the edge of plagioclase crystals). Euhe-
dral zoned allanite often included in feldspars or
Gabbrodioritic series associated to biotite
(quartz gabbro-diorites and quartz diorites)
Qt07 - variable, but generally medium-grained texture Porphyritic monzogranite
Qt08 - Plagioclase in synneusis associations some crys- Qtl8 - Same mineralogy than monzogranite but the major
Qt09 tals with strong recurrent or patchy zoning - Ca- part of K-feldspar is represented by large mega-
Qt44 rich core often encountered (An 75) crists of microcline (4-5 cm in length). Compared
Qt09 - Hornblende in isolated subeuhedral crystals or in to monzogranite, allanite is very rare but apatite
Qt55 polycrystalline agregates with xenomorphic habit. more abundant
Qt56 Often green at the edge, brownish or light green
at the core Central biotite granite
Qt25b - Xenomorphic biotite often included in plagioclase Qtl5 - Plagioclase An 28-20
Qt04 and hornblende crystals. Sometimes it appears as Qt31 - Microcline
Qt02 micrographic intergrowths with quartz, associated Qt32 - Biotite, quartz
to corroded hornblende crystals Qt33 Very rare accessory minerals (apatite, zircon)
Microcline sometimes present as very xenomor-
Sample Qtl5 is slightly altered (quartz with ondula-
phic, interstitial crystals in the biotite-rich rocks
tory extinction and traces ofsericitization and chlo-
Quartzwithveryinterstitialhabitorascentimetric
ritization in plagioclase and biotite)
-

nodules
2-micas granitoids
Accessory minerals:
Qt42 - Subeuhedral phlagioclase (An 22-13) and K-feld-
- Spheric as very small crstals in the cleavages and
spar
at the edges of biotite crystals or as large isolated
Qt85 - Biotite, quartz
xenomorphic crystals
- Muscovite in large poikilitic crystals or in small
- Rare euhedral zircon often included in biotite.
crystals inside plagioclase
Very scarce anhedral crystals
- Tourmaline as skeletal agregates rounded by a
Euhedral prismatic apatite in small crystals gener-
quartzofeldspathic rim without biotite
ally included in feldspar and quartz
Very rare sphene, zircon and clinozoisite
- Allanite in rare, zoned and xenomorphic crystals
of reddish brown colour (non metamict) Porphyritic microgranite
Ilmenite and magnetite in small xenomorphic crys-
- Largephenocrists ofmicrocline (1-4 cms), subeuhe-
-

tals
dral biotite, and globular quartz
- Xenomorphic pyrite crystals in some samples
- Matrix more or less fine-grained (0.~1 mm with
Particular rocks in the gabbrodioritic series quartz, plagioclase (An 35-22), K-feldspar and bio-
tite
Qtl0 - Fine-grained subdoleritic texture with lath-shaped
- Xenomorphic muscovite
(Q-diorite) plagioclase crystals and poikilitic hornblende
Very rare accessory minerals (apatite," zircon; and
Qtl3b - Same texture and granulometry than Qtl0. Pre- sphene as very small crystals in biotite cleavages)
(Q-diorite sence of large (5 mm) nodules of quartz with inclu- Sample Qt71 has suffered a slight mylonitization and
in isolated sions of acicular apatite quartz has recrystallized with mozaic microstruc-
inclusion) ture s

Tonalites and granodiorites

Q86 - Zoned plagioclase crystals often associated in syn-
Qtl03 nensis. Some large crystals (5 8 ram) with strong R e f e r e n c e s

Qt25 recurrent or patchy zoning (An 43-27) and corrod-

ed calcic cores (An 57) Albuquerque CAR De (1978) Rare earth elements in " Y o u n g e r " grani-
- Biotite and minor greenish yellow hornblende gen- tes, Northern Portugal. Lithos 11:219-229
erally in subeuhedral habit Alderton DHM, Pearce JA, Ports PJ (1980) Rare earth element mobil-
- Microcline (rare but large xenomorphic crystals) ity during granite alteration: evidence from Southwest England.
- Quartz Earth Planet Sci Lett 49:149 165
- Accessory minerals." Xenomorphic sphene, prisma- Allegre CJ (1972) Gdochronologie isotopique et d~veloppement de
tic or acicular apatite often included in biotite, la cr6ute terrestre au cours des temps gdologiques. In: "Structure
zircon (often included in biotite). Allanite in zoned et dynamique de la lithosph+re", Paris, Herman
euhedral crystals Allegre CJ, Treuil M, Minster JF, Minster B, Albarede F (1977) System-
atic use of trace elements in igneous processes. Part 1 : Fractional
crystallization processes in volcanic suites. Contrib Mineral Petrol
Granodiorites and transition with monzogranite 60 : 57 75
Qt40 - Intermediate characteristics Allegre CJ, Ben Othman D (1980) Nd-Sr isotopic relationship in granit-
Qt45 oid rocks and continental crust development: a chemical approach
Qt51 to orogenesis. Nature 286 : 335 342
Qt52 Armstrong RL (1968) A model for the evolution o f strontium and
Qtl02 lead isotopes in a dynamic earth. Rev Geophys 6:175 199
Arth JG (1976) Behavior of trace elements during magmatic processes.
Monzogranite A summary of theoretical models and their applications. J Res
Qtl2 - Normally zoned plagioclase (An 40-28) USGS, 4 (1) : 41 47
Qt13 - Xenomorphic K-feldspar, generally microcline Arth JG, Barker F (1976) Rare earth partitioning between hornblende
QtS0 - Biotite, quartz and dacitic liquid and implications for the genesis of trondhjemitic-
Qt82 Accessory minerals." zircon (often included in bio- tonalitic magmas. Geology 4:534-536
194

Autran A, Fonteilles M, Guitard G (1970) Relation entre les intrusions Hurley PMH, Hughes G, Faure HW, Fairbairn WH, Pinson WH
de granitoides, l'anatexie et le m6tamorphisme r~gional consid6r~s (1962) Radiogenic 87Sr model of continent formation. J Geophys
principalement du point de vue du r61e de l'eau: cas de la cha]ne Res 67 : 5315-5334
Hercynienne des Pyr~n6es Orientales. Bull Soc G6ol Fr 7 (XII): Hutton CO (1950) Studies of heavy detrital minerals. Geol Soc Am
673 731 Bull 61:635
Bartlett RW (1969) Magma convection, temperature distribution and Irving AJ (1978) A review of experimental studies of crystal liquid
differentiation. Am J Sci 267 : 1067-1082 trace element partitioning. Geochim Cosmochim Acta 42,
Bateman PC, Chappel BW (1979) Crystallization, fractionation and n ~ 6A: 743-770
solidification of the Tuolumne intrusive series, Yosemite National Koljonen T, Rosenberg RJ (1974) Rare earth elements in granitic
Park, California. Geol Soc Am Bull 1 (90):465 482 rocks. Lithos 7 : 249-261
Bowen NL (1928) The evolution of igneous rocks. Princeton, New- Leterrier J (1972) Etude p6trographique et g6ochimique du massif
Jersey granitique de Qu6rigut (Ari~ge). Th~se Sci Nancy, Fr, Sci Terre
Chayla B, Jaffrezic H, Joron JL (1973) Analyse par activation darts Mere 23
les neutrons ~pithermiques. Application g la d+termination d'616- Leterrier J, Debon F (1978) Caract~res g~ochimiques compares des
ments en traces dans les roches. CR Acad Sci Paris, 277, set D : 273- granitoides et de leurs enclaves microgrenues. Implications g6n6ti-
275 ques. Bull Soc G~ol Fr, ser7, 20:3-10
Cocherie A (1977) Donn~es pr~liminaires sur la g6ochimie des terres Mac Carthy TS, Kable EJD (1978) On the behavior of rare earth
rares dans le massif leucogranitique du Manaslu (N6paI central). elements during partial melting of granitic rocks. Chem Geol
Int Coll CNRS 268, Paris:93 110 22 : 21-29
Coryei1 CD, Chase JW, Winchester JW (1963) A procedure for geo- Marre J (1973) Le complexe &uptif de Qu6rigut: p&rologie, structnro-
chemical interpretation of terrestrial rare earth abundance patterns. logie, cin6matique de mise en place. Th6se, Univ Toulouse, Fr
J Geophys Res 68:559 566 Masuda A (1962) Regularities in variation of relative abundances of
Didier J (1973) Granites and their enclaves. Dev. in Petrology. Amster- lanthanides elements and an attempt to analyse separation index
dam, Elsevier patterns of some minerals. J Earth Sci Nagoya Univ 10:173 187
Didier J, Lameyre J (1969) Les granites du Massif Central frangais: Masuda A (1968) Geochemistry of lanthanides in basalts of central
Etude compar~e des leucogranites et granodiorites. Contrib Mineral Japan. Earth Planet Sci Lett 4:284 292
Petrol 24:21%228 Michard-Vitrac A, Albarede F, Dupuis C, Taylor HP jr (1980) The
Dudas M J, Schmitt RA, Harward ME (1971) Trace element partition- genesis of Variscan (Hercynian) plutonic rocks: inferences from
ing between volcanic plagioclase and dacitic pyroclastic matrix. Sr, Pb, and O studies on the Maladetta igneous complex, Central
Earth Planet Sci Lett 11 : 440-446 Pyr6n6es (Spain). Contrib Mineral Petrol 72 (1):57-72
Emmermann R, Daieva L, Schneider J (1975) Petrologic significance Moorbath S (1975) Evolution of Precambrian crust from strontium
of rare earths in granites. Contrib Mineral Petrol 52:267 283 evidence. Nature 254 : 395-397
Ewart A, Hildreth W, Carmichael ISE (1975) Quaternary acid magma Nagasawa H (1970) Rare earth concentrations in zircons and apatites
in New Zealand. Contrib Mineral Petrol 51 : 1 27 and their host dacites and granites. Earth Planet Sci Lett
Flynn RT, Burnham CW (1978) An experimental determination of 9:359 364
rare earth partition coefficients between a chloride containing vapor Nagasawa H, Schnetzler CC (1971) Partitioning of rare earth, alkaline
phase and silicate melts. Geochim Cosmochim Acta 42:685 701 and alkaline-earth elements between phenocrists and acidic igneous
Fourcade S, Hamet J, Allegre CJ (1977) Donn6es de la m6thode rubid- magma. Geochim Cosmochim Acta 35 : 953-968
ium-strontium et d6termination des terres rares dans le leucogranite Nance WB, Taylor SR (1976) Rare earth element patterns and crustal
du Manaslu: implications pour l'orog6n6se himalayenne. CR Acad evolution. I. Australian post-Archean sedimentary rocks. Geochim
Sci Paris, 284, ser D:717 720 Cosmochim Acta 40:1539 1551
Fudali RF (1965) Oxygen fugacities and andesitic magmas. Geochim O'Nions RK, Evensen NM, Hamilton PJ (1979) Geochemical model-
Cosmochim Acta 29:1063 1075 ing of mantle differentiation and crustal growth. J Geophys Res
Gromet P (1979) Rare earths abundances and fractionation and their 84:6091-6101
implication for batholithic petrogenesis in Peninsular Ranges Orville PM (1962) Alkali metasomatism and feldspars. Nor Geol
Batholith, California, USA, and Baja California, Mexico. Ph D Tidsskr 42: 283-316
Thesis, Caltech Osborn EF (1962) Reaction series for subalkaline igneous rocks based
Gulson BL, Krogh TE (1973) Old lead component in the young Bergell on different oxygen pressure conditions. Am Mineral 47 : 211-226
massif, South East Swiss Alps Contrib Mineral Petrol 40:239-252 Patterson C, Tatsumoto M (1964) The significance of Lead isotopes
Hamet J, Allegre CJ (1976) Hercynian orogeny in the Montagne Noire in detrital feldspar with respect to chemical differentiation within
(France) : Application of 8VRb/SVSrsystematics. Geol Soc Am Bull the arth's mantle. Geochim Cosmochim Acta 28:1 22
87 : 1429-1442 Perrin R, Roubault M (1939) Le granite et les r6actions/t l'~tat solide.
Hanson GN (1978) The application of trace elements to the petrogene- Carte G6ol Alg~rie Bull, 5~me s~r, 4:1 168
sis of igneous rocks of granitic composition. Earth Planet Sci Lett Piwinskii AJ (1968) Experimental studies of igneous rock series, Cen-
38 : 26-43 tral Sierra Nevada batholith, California. J Geol 76:548 570
Hart SR, Allegre CJ (1979) Trace elements constraints on magma Presnall DC, Bateman PC (1973) Fusion relations in the system
genesis. I n " Physics of magmatic processes", Bowen Vol, Princeton NaA1SiaOs-CaA12Si2Os-KA1Si3Os-H20 and generation of granitic
Univ magmas in the Sierra Nevada batholith. Geol Soc Am Bull
Haskin LA, Haskin MA (1966) Rare earth in European shales: A 84: 3181-3202
redetermination. Science 154:507-509 Prince AT (1943) The system Albite-Anorthite-Sphene. J Geol 51 : 1 16
Haskin LA, Frey FA, Schmitt RA, Schmitt RH (1968) Meteroritic, Raguin E (1977) Le massif de l'Aston dans les Pyr6n+es de l'Ari~ge.
solar and terrestrial rare earth distribution. In: Ahrens (ed) "Phys- Bull BRGM, Sect 1 (2) : 8%119
ics and chemistry of the earth, Vol VII, pp 169 321 Ramberg H (1944) The thermodynamics of the earth's crust. 1. Prelimi-
Haskin LA, Paster TP (1979) Chemistry and mineralogy of the rare nary survey of principal forces and reactions in the earth's crust.
earths. In: Gschneider and Eyring (eds) "Handbook of the physics Nor Geol Tidsskr 24: 98-111
and chemistry of the rare earths", Ch 21 Schilling JG, Winchester JW (1967) Rare earth fractionation and mag-
Hellman PL, Green TH (1979) The role of sphene as an accessory matic processes. In: Runcorn (ed) "Mantle of the earth and terres-
phase in the high-pressure partial melting of hydrous mafic compo- trial planets" Interscience, New-York, pp 267-283
sitions. Earth Planet Sci Lett 42:191-201 Schnetzler CC, Philpotts JA (1970) Partition coefficients of rare earth
Herrmann AG (1970) Yttrium and lanthanides. In: Wedepohl (ed) elements between igneous matrix and rock forming phenocrists,
"Handbook of geochemistry", Springer Verlag, 39 : 57 71 II. Geochim Cosmochim Acta 34:331-340
 195

Shaw HR (1965) Comments on viscosity, crystal settling and convec- Tuttle OF, Bowen NL (1958) Origin of granite in the light of experi-
tion in granitic magmas. Am J Sci 263:120-152 mental studies in the system NaA1Si308 - KAISi3Os - SiO2 - H20.
Shaw DM (1977) Trace element behavior during anatexis. In: Dick Geol Soc Am Bull 74:1 153
(ed) "Magma genesis", Proc Am Geophys Un, Bull 96:189-213 Vitrac-Michard A, Allegre CJ (1975) A study of the formation and
Shimizu N, Semet MP, Allegre CJ (1978) Geochemical applications history of a piece of continental crust by the STRb/STSr method:
of quantitative ion-microprobe analysis. Geochim Cosmochim Acta The case of the french oriental Pyr6n6es. Contrib Mineral Petrol
42:1321 1334 50:257-285
Skala W (1979) Some effects of the constant sum problem in geo- Walton M (1960) Granite problems. Science 131:635-645
chemistry. Chem Geol 27 : 1-9 Watson B (1979) Zircon saturation in felsic liquids: Experimental re-
Smith WL, Franck ML, Sherwood AN (1957) Uranium and Thorium sults and application to trace element geochemistry. Contrib Min-
in the accesory Allanite of igneous rocks. Am Mineral 42:367 eral Petrol 70:407 419
Taylor SR (1969) Trace element geochemistry of Andesites and asso- Webster EA, Holloway JR (1980) The partitioning of REE's, Sc, Rb
ciated calc-alkaline rocks. Proc Andesite Conf, Oregon, Dept Geol and Cs between a silicic melt and a C1 fluid. Eos 6i (46):1152
Mineral, Publ 65 Wendland RF, Harrison WJ (1979) Rare earth partitioning between
Taylor SR (1979) Chemical composition and evolution of the continen- carbonate and silicate liquid and CO2 vapor: results and implica-
tal crust. In: Mc Elhinny (ed) "The earth, its origin, structure tion for the formation of light rare earth-enriched rocks. Contrib
and evolution", Academic, New-York Mineral Petrol 69:409-419
Taylor RP, Fryer BJ (1980) Multiple-stage hydrothermal alteration White AJR, Chappel BW (1977) Ultrametamorphism and granitoid
in porphyry copper systems in northern Turkey: the temporal inter- genesis. Tectonophysics 43:7 22
play of potassic, propylitic and phyllic fluids. Can J Earth Sci Winkler HGF (1965) Petrogenesis of metamorphic rocks. Springer
17 : 901-926 Verlag, Berlin
Treuil M, Jaffrezic H, Deschamps N, Derre C, Guichard F, Joron Wyllie PJ (1977) Crustal anatexis: an experimental revew. Tectono-
JL, Pelletier B, Novotny S, Courtois C (1973) Analyse des lantha- physics 43:41-71
nides, Hf, Cr, Mn, Co, Cu, Zn dans les min6raux et les roches Zielinski RA, Lipman PW (1976) Trace element variations at Summer
par activation neutronique. J Radioanal Chem 18:55-68 Coon Volcano, San Juan Mountains, Colorado, and the origin
Treuil M, Varet J (1973) Crit6res volcanologiques, p6trologiques et of continental interior andesite. Geol Soc Am Bull 87:1477-1485
g6ochimiques de la g6nSse et de la differenciation des magmas
basaltiques : Exemple de l'Afar. Bull Soc G$ol Fr 7 (XV): 506-540 Received September 15, 1980; Accepted February 5, 1981