Chiaradia Et. Al., 2011

JOURNAL OF PETROLOGY VOLUME 52 NUMBER 6 PAGES 1107^1141 2011 doi:10.
1093/petrology/egr020
Enriched Basaltic Andesites from Mid-crustal

Fractional Crystallization, Recharge,
and Assimilation (Pilavo Volcano,
Western Cordillera of Ecuador)
MASSIMO CHIARADIA1, OTHMAR MU«NTENER2 AND

BERNARDO BEATE3
1
SECTION OF EARTH AND ENVIRONMENTAL SCIENCES, DEPARTMENT OF MINERALOGY, UNIVERSITY OF GENEVA,
RUE DES MARAI“CHERS 13, 1205 GENEVA, SWITZERLAND
2
INSTITUTE OF MINERALOGY AND GEOCHEMISTRY, UNIVERSITY OF LAUSANNE, ANTHROPOLE, CH-1015 LAUSANNE,
Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

SWITZERLAND
3
DEPARTMENT OF GEOLOGY, ESCUELA POLITECNICA NACIO¤NAL, QUITO, ECUADOR
RECEIVED OCTOBER 19, 2010; ACCEPTED APRIL 19, 2011

ADVANCE ACCESS PUBLICATION JUNE 8, 2011
The origin of andesite is an important issue in petrology because Petrographic, mineral chemistry, bulk-rock geochemical and isotopic
andesite is the main eruptive product at convergent margins, data indicate that the Pilavo magmatic rocks have evolved through
corresponds to the average crustal composition and is often associated three main stages: (1) generation of a basaltic magma in the mantle
with major CuÂu mineralization. In this study we present petro- wedge region by flux melting induced by slab-derived fluids (aque-
graphic, mineralogical, geochemical and isotopic data for basaltic ous, supercritical or melts); (2) high-pressure differentiation of the
andesites of the latest Pleistocene Pilavo volcano, one of the most basaltic melt (at the mantle^crust boundary or at lower crustal
frontal volcanoes of the Ecuadorian Quaternary arc, situated upon levels) through sustained fractionation of olivine and clinopyroxene,
thick (30^50 km) mafic crust composed of accreted Cretaceous leading to hydrous, high-alumina basaltic andesite melts with a
oceanic plateau rocks and overlying mafic to intermediate Late tholeiitic affinity, enriched in incompatible elements and strongly
Cretaceous^Late Tertiary magmatic arcs. The Pilavo rocks are bas- impoverished in Ni and Cr; (3) establishment of one or more
altic andesites (54^57·5 wt % SiO2) with a tholeiitic affinity as mid-crustal magma storage reservoirs in which the magmas evolved
opposed to the typical calc-alkaline high-silica andesites and dacites through dominant amphibole and clinopyroxene (but no plagioclase)
(SiO2 59^66 wt %) of other frontal arc volcanoes of Ecuador fractionation accompanied by assimilation of the modified plutonic
(e.g. Pichincha, Pululahua). They have much higher incompatible roots of the arc and recharge by incoming batches of more primitive
element contents (e.g. Sr 650^1350 ppm, Ba 650^1800 ppm, Zr magma from depth. The latter process has resulted in strongly
100^225 ppm, Th 5^25 ppm, La 15^65 ppm) and Th/La ratios increasing incompatible element concentrations in the Pilavo basaltic
(0·28^0·36) than Pichincha and Pululahua, and more primitive andesites, coupled with slightly increasing crustal isotopic signatures
Sr (87Sr/86Sr 0·7038^0·7039) and Nd (eNd þ5·5 to þ6·1) and a shift towards a more calc-alkaline affinity. Our data show
isotopic signatures. Pilavo andesites have geochemical affinities that, although ultimately originating from the slab, incompatible
with modern and recent high-MgO andesites (e.g. low-silica ada- element abundances in arc andesites with primitive isotopic signa-
kites, Setouchi sanukites) and, especially, with Archean sanukitoids, tures can be significantly enhanced by intra-crustal processes within
for both of which incompatible element enrichments are believed a thick juvenile mafic crust, thus providing an additional process
to result from interactions of slab melts with peridotitic mantle. for the generation of enriched andesites.
The Author 2011. Published by Oxford University Press. All

*Corresponding author.Telephone: þ4122 379 66 34. Fax: þ4122 379 32 rights reserved. For Permissions, please e-mail: journals.permissions@
10. E-mail: [email protected] oup.com
JOURNAL OF PETROLOGY VOLUME 52 NUMBER 6 JUNE 2011
KEY WORDS: andesite; assimilation; fractional crystallization; the possibility that its incompatible element enrichment is
recharge; Ecuador acquired from assimilation of mature continental crust.
The enriched geochemistry of Pilavo rocks resembles that
of modern low-silica adakites (e.g. Setouchi sanukite)
I N T RO D U C T I O N and, especially, that of Archean sanukitoids, from which
Establishing the origin of andesitic rocks is an important Pilavo differs, none the less, in having variably lower
goal of petrology because andesite is the main eruptive MgO and significantly lower Ni and Cr contents.
product at convergent margins, it corresponds to the Investigation of the Pilavo basaltic andesites thus has the
average crustal composition (e.g. Gill, 1981; Rudnick & potential to shed further light on the origin of enriched an-
Gao, 2004) and it is closely associated with major desitic rocks and their significance with respect to contin-
porphyry-related CuÂu mineralization (e.g. Richards, ental crust formation processes, as well as on the coupled
2003). Various petrogenetic models have been proposed for occurrence of tholeiitic and calc-alkaline rocks within the
the genesis of andesite, including, among others, partial same arc and volcanic edifice (e.g. Hora et al., 2009).
melting of peridotite under hydrous conditions (e.g.
Kushiro, 1969; Hirose, 1997; Wood & Turner, 2009), partial
melting of the subducting oceanic crust (e.g. Taylor et al., GEODY NA M IC S ET T I NG A N D
1969), crystal fractionation of mantle-derived basalts (e.g.
Crawford et al., 1987; Mu«ntener et al., 2001; Grove et al., A RC M A G M AT I S M
2003), mixing between mantle-derived basalt and Present-day arc volcanism in the Ecuadorian Andes is the
crust-derived felsic melts (e.g. Tatsumi & Kogiso, 2003) or result of subduction of the 12^20 Myr old Nazca plate and

assimilation of the plutonic roots of arcs (Reubi & Blundy, overlying aseismic Carnegie Ridge (Fig. 1). The active to
2008). In all these models, the geochemical characteristics recent volcanic arc of Ecuador extends over more than
of arc andesites imply a contribution from the subducted 100 km across the Andean chain (Fig. 1) starting in the
slab (sediments and oceanic crust) in the form of aqueous Western Cordillera, passing through the Interandean
fluids, supercritical liquids or hydrous silicate melts acting Valley and the Eastern Cordillera, and ending in the
as the metasomatic agent responsible for the flux melting Andean foreland (Oriente) where both alkaline (Sumaco,
of the mantle wedge, with supercritical fluids and hydrous Pan de Azucar, Puyo) and calc-alkaline (Reventador) vol-
silicate melts being the most likely fluxing agents (e.g. canoes occur. Such a broad volcanic arc (4100 km) con-
Kessel et al., 2005; Hermann et al., 2006). In particular, trasts with the narrower volcanic arc immediately to the
high-MgO andesites [or low-silica adakites of Martin north in Colombia, and with the volcanic gap south of
et al. (2005)] with high concentrations of large ion litho- 28S owing to the Peruvian flat slab segment (Barazangi &
phile elements (LILE) and light rare earth elements Isacks, 1976). The subduction of the aseismic Carnegie
(LREE), high Sr/Y and La/Yb, and heavy REE (HREE) Ridge in coincidence with the Ecuadorian volcanic arc
depletion, like those of the Miocene Setouchi belt of Japan suggests a direct link with such a broad arc. It has been
and of the western part of the Aleutian arc, are considered proposed that subduction of the Carnegie Ridge beneath
to derive from melting of peridotite previously metasoma- the continental crust of Ecuador has caused an increased
tized by slab melts or from the incomplete reaction of slab coupling between the subducting and overriding plates
melts with the peridotitic mantle wedge (Rapp et al., 1999; (e.g. Ego et al., 1996; Gutscher et al., 1999; Graindorge
Kelemen et al., 2003; Martin et al., 2005; Tatsumi, 2006). et al., 2004). Gutscher et al. (1999) suggested that flat sub-
Because such rocks are similar to the average crustal com- duction was related to the subduction of the Carnegie
position and bear strong similarities to Archean sanuki- Ridge flanked by subduction of two steeper segments
toids (Shirey & Hanson, 1984), their petrogenesis is north and south of the Carnegie Ridge. In contrast,
important for the understanding of continental crust for- Guillier et al. (2001) concluded that subduction of the
mation processes (e.g. Tatsumi, 2006; Rapp et al., 2010). Carnegie Ridge segment is not flat, but regularly dipping
In the present study we investigate andesitic rocks from at an angle of 258 (see also Lonsdale, 1978; Pennington,
Pilavo volcano, one of the most frontal volcanoes of the 1981; Pre¤vot et al., 1996) although immediately to the
Ecuadorian Quaternary arc. The petrogenesis of these north of it the dip increases slightly to 308.
rocks is of particular interest because Pilavo lavas are The estimated crustal thickness beneath the Western
incompatible element-rich basaltic andesites to low-SiO2 Cordillera, where Pilavo is situated, ranges between
andesites with a tholeiitic affinity as opposed to the typ- 25^30 km (Feininger & Seguin, 1983) and 40^50 km
ical calc-alkaline high-SiO2 andesites and dacites of the (Guillier et al., 2001). Gravimetric (Feininger & Seguin,
other frontal arc volcanoes of Ecuador (e.g. Pichincha, 1983) and isotopic (Chiaradia & Fontbote¤, 2001;
Pululahua). Like all frontal arc volcanoes of the Western Chiaradia, 2009) data indicate the presence of mafic crust
Cordillera of Ecuador (e.g. Feininger & Seguin, 1983), beneath the Western Cordillera of Ecuador (Fig. 1), with
Pilavo sits above Cretaceous mafic oceanic crust, excluding oceanic plateau-like geochemical characteristics (e.g. Kerr
1108
CHIARADIA et al. ENRICHED BASALTIC ANDESITE ORIGIN
Water depth (m)

2°N 3000-4000
2500-3000 COLOMBIA
2000-2500
1500-2000
1000-1500 Pilavo
<1000
Pululahua Figure 2a
0° Pichincha
CGR

0 100 km
2°S
Continental basement
Oceanic basement
GFZ Volcano with historical activity
Volcano active since <10 ka
4°S Plio-Quaternary volcano
PERU Crustal fault/suture
84°W 82°W 80°W 78°W

Fig. 1. Map of Ecuador showing the main topographic features of the subducting plates (CGR, Carnegie Ridge; GFZ, Grijalva Fracture Zone)
and the localization of recent and active volcanoes in Ecuador (modified after Litherland et al., 1994; Corredor, 2003).
et al., 2002; Mamberti et al., 2003). The oceanic plateau of slab fluids or melts towards the back-arc (the K^h rela-
crust represents an exotic terrane (e.g. Feininger, 1987; tionship: Dickinson, 1975).
Jaillard et al., 1990, 1997; Spikings et al., 2005), derived
from the large Caribbean^Colombian Oceanic Plateau
(CCOP), which was accreted to the continental margin of LOC A L G EOLO GY
Ecuador during Late Cretaceous times (Vallejo et al., Pilavo volcano (0831’N, 78822’W) is located about 20 km
2006, 2009). trenchward of the main axis of the active volcanic front
Various studies have highlighted systematic compos- (Figs 1 and 2a), which follows the crest of the Western
itional changes in the lavas of the Ecuadorian arc from Cordillera of the Ecuadorian Andes. It is thus one of
the frontal to the back-arc region (e.g. Barragan et al., the closest volcanoes to the trench of the Northern Andes
1998; Bourdon et al., 2003; Bryant et al., 2006; Chiaradia subduction system, located about 230 km to the west of
et al., 2009). These variations have been interpreted as the Pilavo (Fig. 1). The deep basement of Pilavo, like that of
result of progressively smaller degrees of partial melting of several other frontal volcanoes of the Western Cordillera
the mantle wedge metasomatized by decreasing amounts of Ecuador, consists of oceanic plateau mafic crust
1109
775000mE 80 85 90 95 0 5 10 15 20 25 30
(a) Figure 2b
Pilavo 60
Yanaurcu
Parulo Huanguillaro
50
Pulumbura
Urcuqui
Ibarra 40
Imbabura

Cotacachi- 30
Cuicocha
Otavalo Cubilche
002000mN
0 10 km
Pre-Cretaceous metamorphic rocks Avalanche deposits
Cretaceous to Eocene undifferentiated Undifferentiated

volcanic and sedimentary rocks Pleistocene-Holocene
tuffs and alluvial
Miocene sandstones (Chota Fm.) deposits
Mio-Pliocene volcanic rocks Pleistocene-Holocene
volcanic edifices
Plio-Pleistocene volcanic rocks
Caldera
Mid-Miocene intrusions
(granodiorite to quartz-diorite) Faults
Pleistocene-Holocene
rhyodacitic domes
Fig. 2. (a) Simplified geological map of the volcanic complexes around Pilavo (modified after BGS & CODIGEM, 1999); (b) topographic
map of Pilavo showing the locations of the samples investigated in this study. Lava flows are contoured. The earlier part of the volcanic edifice
is indicated by the horizontally striped area. UTM coordinates (relative to zone 17N) are reported on both maps.
1110

Fig. 2. Continued.
accreted to the continental margin during the Late xenoliths in Pilavo lava flows, and are recrystallized to a
Cretaceous, and of overlying Late Cretaceous to Tertiary fine-grained mosaic of epidote and amphibole (5200 mm
(Eocene to Miocene) magmatic arcs (Vallejo et al., 2006, crystal size).
2009; Chiaradia, 2009). The exposed basement of Pilavo Pilavo volcano consists of some 30 lava flows, reaching
volcano (Fig. 2a) consists of Cretaceous marine sequences lengths up to 8 km (Fig. 2b), with a thickness of more than
(Pilato¤n Unit) of volcanic sand- and siltstones as well 40 m in many cases. Its morphology is closer to that of a
as breccias derived from erosion of the Cretaceous Rio shield than to that of a strato-volcano (Fig. 2b), with a
Cala basaltic arc (Chiaradia, 2009). These rock types are total average basal diameter of about 8 km and a height of
occasionally found as microcrystalline, millimeter-sized 650 m. Pyroclastic deposits associated with Pilavo have
1111
not been observed (either they never formed, or they were if Parulo is earlier or later than Pilavo, based on field
scarce or, if related to early activity, they were buried by observations alone.
later lava flows). NNE^SSW- and NW^SE-striking struc-
tures appear to have controlled the location of eruptive
vents at Pilavo volcano, as well as at the neighbouring A N A LY T I C A L M E T H O D S
early Pleistocene andesitic Pulumbura volcano, the mid- Estimates of modal mineral abundances (Table 1) were ob-
to late Pleistocene Yanaurcu (andesite and dacite) and tained by image analysis carried out on thin sections with
Huanguillaro (andesite to rhyodacite) composite strato- an estimated uncertainty (1s) of 10%. Thicknesses of
volcanoes (Fig. 2a). Normal, NNE^SSW-striking faults are amphibole phenocryst reaction rims were measured under
also present in the area, affecting the CretaceousÊocene the microscope on 1^3 polished thin sections per sample,
basement (Silante Fm.), but their relation to volcanic activ- using both transmitted and reflected light. Table 1 also re-
ity is unknown. ports the average thickness and corresponding standard
Pilavo blocky lava flows consist mainly of fresh, gray deviation of the reaction rims measured around the entire
coloured, mostly non-vesicular, hornblende-, pyroxene-, amphibole population of each sample. Because we have
plagioclase-phyric basaltic andesites to andesites, which measured reaction rim thickness on all the amphibole
are covered by 1^2 m of soil with intercalated distal phenocrysts present in each thin section we assume that
the effects of oblique sections through amphibole rims are
Holocene pumice tephras (the youngest tephra, with an
similar for all samples. Although these measurements are
age of 58 ka, is from the Huanguillaro volcanic complex:
subject to some uncertainty (preventing a correct estimate
Aguilera, 1998). The morphology of lava flows is fairly
of ascent rates, which we therefore did not attempt to cal-

well preserved, as central flow waves, lateral levees, toes
culate), sample-to-sample variations are considered to be
and tongues, can be recognized on topographic maps significant.
(Fig. 2b) and on aerial photographs. Flows have been af- Microprobe analyses of pyroxene, amphibole, plagio-
fected slightly by glacial activity (at 10^12 ka) at higher clase, Fe^Ti-oxides and volcanic glass (Electronic
elevations, and the summit area shows a 1km diameter de- Appendixes 1^5, available for downloading at http://www
pression open to the NNW, which could have been .petrology.oxfordjournals.org) were carried out at the
formed by glacial erosion, focused in the central vent area Institute of Mineralogy and Geochemistry, University of
(summit). In contrast, Pilavo was not affected by the in- Lausanne (Switzerland), using a JEOL 8200 microprobe
tense glacial erosion of the pleniglacial period (at 20^40 equipped with five wavelength-dispersive spectrometers.
ka) like the adjacent Yanaurcu and Huanguillaro volca- Analytical conditions were: sample current 15 or 20^nA,
noes. Based on these observations, Pilavo is older than acceleration voltage 15 kV, measuring time 10^30 s on
8 ka (the age of the youngest tephra from the peak, and half the time on the respective backgrounds de-
Huanguillaro complex covering Pilavo) and probably pending on the element, and spot size of 1^5 mm depend-
younger than 20^40 ka. ing on the element and mineral. Both natural and
Although Pilavo lava flows have similar petrographic synthetic silicate, oxide, and sulfate standards were used
characteristics (hornblende-bearing basaltic andesites), for external calibration.
the morphology of the edifice suggests a three-stage devel- In situ trace element and REE analyses of volcanic glass,
opment over a relatively short period of time, as follows. amphibole and clinopyroxene (Electronic Appendixes
5^7) were carried out on polished thin sections by laser
(1) Eruption of several early lava flows mostly towards ablation inductively coupled plasma mass spectrometry
the west along a north^south- to NNE^SSW-oriented (LA-ICP-MS) using a Perkin Elmer ELAN 6100 DRC
fissure located 4 km west of the summit (Fig. 2b: e.g. system equipped with a 193 nm excimer laser (Geolas) at
samples E05150, E05151). the University of Lausanne. Operating conditions of the
(2) Eruption of abundant radiating lava flows (most of laser were: 8 Hz frequency, 100 mJ energy, 60 mm spot size.
the collected samples) from the glacially eroded cen- CaO contents determined by electron microprobe in
tral vent located in the summit area (Fig. 2b). These the area of subsequent ablation with the laser were used
lavas flowed around earlier flows mainly to the for internal standardization by reference to an SRM612
north, west and south, and surrounded the area of NIST external standard. Raw data were reduced off-line
earlier flows towards the north and south of the fis- using the LAMTRACE software (Jackson et al., 1992;
sure, constructing a lava flow-dominated edifice. This Longerich et al., 1996; Jackson, 2008). The reproducibility
is the main building phase of Pilavo volcano. (1s) of the measured trace elements in the SRM612 stand-
(3) Eruption of the small dome^flow complex of Parulo, ard was 510% for Ni and 55% for all other elements
of similar petrography and immediately west of the (Electronic Appendix 6).
Pin‹an river, to form Laguna Donoso by blocking the Thirteen whole-rock samples (Table 2 and Fig. 2b) were
drainage with a lava flow (Fig. 2b). It is not clear yet analyzed for major and trace elements by X-ray
1112
Table 1: Mineral modal abundances in volcanic rocks of Pilavo calculated from image analysis and average thickness of
reaction rims around amphibole phenocrysts with its corresponding standard deviation
Sample: E05064 E05065 E05066 E05067 E05148 E05149 E05150 E05151 E05152 E05153 E05154 E05155 E05156
Rock: Andesite Andesite Andesite Andesite Andesite Andesite Basaltic Andesite Basaltic Andesite Andesite Andesite Andesite
andesite andesite andesite
amph (modal %) 4 3 13 12 2 10 5 11 17 3 4 2 17
cpx (modal %) 20 13 5 8 27 6 20 1 15 14 27 20 5
plag (modal %)* 39 80 25 15 55 22 39 45 15 15 27 23 18
mt (modal %) 5 2 4 3 3 4 5 2 5 5 5 5 2
gm (modal %)y 32 3 53 62 14 59 31 41 48 63 38 50 58
Average thickness of reaction
rim around 130 183 44 28 80 75 183 96 72 90 68 90 18
amphibole (mm)
n 31 12 39 73 21 44 26 168 248 22 69 65 43
1s of thickness (mm) 103 78 29 7 33 53 39 55 38 40 44 42 3

*Includes microcrysts in the groundmass but not microlites (5150 mm).
yIncludes glass and plagioclase-clinopyroxene microlites (5150 mm).
Estimated 1s uncertainty for modal abundances is 10%. n, number of amphibole phenocrysts around which reaction rims
were measured. amph, amphibole; cpx, clinopyroxene; plag, plagioclase; mt, magnetite; gm, groundmass.
fluorescence (XRF) at the University of Lausanne. The again. Additionally we also analyzed Sr and Nd isotopes
BHVO-1 basaltic standard was used for quality control. on a microdrilled amphibole phenocryst and its host
The 1s uncertainties based on repeated measurements of groundmass (sample E05156) to check for possible disequi-
this standard are 50·5% for all major elements (except librium between phenocrysts and groundmass. The ana-
MgO, 1·5%, and Na2O, 3·3%; Table 2) and 51% for trace lyzed material (c. 8 mg in both cases) was obtained by
elements. REE and additional trace elements (e.g. Th, U, drilling a cylinder of c. 3 mm diameter out of a 350 mm
Ta, Cs, Hf) were measured by LA-ICP-MS on the thick thin section using a diamond microdrill.
fused glass beads used for major element XRF analysis Sr, Nd and Pb separation was carried out using cascade
(Table 2). Operating conditions of the laser were: 10 Hz fre- columns with Sr-spec, TRU-spec and Ln-spec resins fol-
quency, 140 mJ energy, 120 mm spot size. CaO contents pre- lowing a modified method after Pin et al. (1994). Pb was
viously determined by XRF were used for internal further purified with an AG-MP1-M anion exchange resin
standardization by reference to an SRM612 NISTexternal in a hydrobromic medium. Pb, Sr and Nd isotope ratios
standard. Raw data were reduced off-line using the were measured on a Thermo TRITON mass spectrometer
LAMTRACE software. For each sample 3^4 points were on Faraday cups in static mode. Pb was loaded onto Re
measured and results were then averaged. Uncertainties filaments using the silica gel technique and all samples
(1s) between results of repeat points of the same sample (and standards) were measured at a pyrometer-controlled
were510% for REE and55% for the other trace elements temperature of 12208C. Pb isotope ratios were corrected
analyzed, whereas the 1s uncertainties of repeated meas- for instrumental fractionation by a factor of 0·07% per
urements of the SRM612 NIST glass were 55% for all a.m.u. based on more than 90 measurements of the
elements (Table 2). SRM981 standard and using the standard values of Todt
All 13 whole-rocks were analyzed for their Sr, Nd and Pb et al. (1996). External reproducibility of the standard ratios
isotopic compositions at the Department of Mineralogy, is 0·11% for 206Pb/204Pb, 0·12% for 207Pb/204Pb and
University of Geneva (Table 2). About 130 mg of powdered 0·20% for 208Pb/204Pb.
rock (570 mm) were dissolved in closed Teflon vials for Sr was loaded on single Re filaments with a Ta oxide so-
7 days on a hot plate at 1408C with a mixture of 4 ml lution and measured at a pyrometer-controlled tempera-
conc. HF and 1ml HNO3 15 M. The sample was then ture of 14808C in static mode using the virtual amplifier
dried on a hot plate, and re-dissolved in 3 ml of 14 M design to cancel out biases in gain calibration among
HNO3 in closed Teflon vials at 1408C and dried down amplifiers. 87Sr/86Sr values were internally corrected for
1113
Table 2: Geochemical and isotopic data for the Pilavo lavas investigated in this study
Sample: E05064 E05065 E05066 E05067 E05148 E05149 E05150 E05151

Rock: Andesite Andesite Andesite Andesite Andesite Andesite Bas. andesite Andesite
SiO2 57·57 56·07 56·07 56·42 56·04 57·37 54·17 56·54

TiO2 0·73 0·71 0·76 0·81 0·76 0·75 0·69 0·66
Al2O3 17·66 17·85 17·43 17·22 17·22 16·81 17·1 18·17
Fe2O3(tot) 8·29 9·13 8·93 8·07 8·85 7·82 9·33 8·63
MnO 0·15 0·16 0·15 0·13 0·16 0·13 0·17 0·15
MgO 3·25 3·73 3·58 3·3 3·52 3·1 3·8 3·45
CaO 7·41 7·97 7·7 7·23 7·68 7·05 8·23 7·65
Na2O 3·16 3·01 2·95 3·15 3·05 3·26 2·9 3·04
K2O 1·77 1·13 1·86 2·36 1·67 2·36 1·38 0·95
P2O5 0·28 0·23 0·28 0·32 0·27 0·33 0·25 0·22
LOI 0·02 0·15 0·28 0·45 0·00 0·08 0·21 0·35
Total 100·3 99·81 99·99 99·55 99·25 99·12 98·26 99·78
Rb 32·3 17·1 29·8 42·9 30·9 44·1 22·2 14·2
Sr 907 666 1135 1298 904 1226 901 700

Y 16·5 16·2 18·5 16·6 18·3 15·9 18·4 15·5
Zr 133 80 156 222 131 206 108 85
Nb 4·0 3·2 4·1 5·4 4·0 5·2 3·4 3·1
Cs 0·6 0·4 0·8 1·1 0·4 1·1 0·6 0·6
Be 2·2 1·7 2·2 3·3 0·8 1·4 1·3 1·1
Ba 1119 734 1412 1465 1019 1439 968 648
Hf 2·8 2·0 3·2 4·8 2·6 3·7 1·9 1·8
Ta 0·3 0·3 0·3 0·4 0·2 0·2 b.d.l. 0·1
Pb 8·6 7·6 9·5 12·3 7·0 12·0 8·0 6·0
Th 8·4 4·5 15·2 14·6 7·4 11·9 8·0 4·1
U 2·3 1·2 3·6 3·5 2·1 3·3 2·4 1·2
Sc 21·7 23·8 22·8 22·1 21·3 18·2 23·1 21·0
V 138 161 168 159 241 174 203 183
Cr 4 13 6 17 6 14 11 12
Ni 4 6 6 10 8 12 9 10
Cu 56 48 37 37 63 57 44 33
Co 27 25 19 19 22 18 25 28
Zn 60 68 77 76 90 86 89 87
La 27·7 14·6 43·9 50·1 24·2 42·6 25·8 13·7
Ce 50·6 26·7 72·5 85·9 44·6 78·3 46·6 25·8
Pr 5·9 3·4 8·3 10·2 5·3 9·2 5·4 3·3
Nd 23·8 14·9 32·8 41·5 20·9 34·6 20·4 13·4
Sm 4·6 3·1 5·7 7·2 4·1 6·1 3·8 2·9
Eu 1·3 1·0 1·5 1·8 1·3 1·7 1·0 0·8
Gd 3·5 3·2 4·2 5·7 3·2 4·6 3·2 2·7
Tb 0·54 0·47 0·58 0·66 0·43 0·52 0·45 0·37
Dy 3·3 2·9 3·7 3·8 2·9 3·1 2·7 2·5
Ho 0·64 0·61 0·71 0·72 0·57 0·52 0·54 0·48
Er 1·78 1·85 1·98 1·86 1·67 1·37 1·48 1·41
Tm 0·26 0·26 0·31 0·25 0·21 0·2 0·22 0·19
Yb 1·83 1·75 1·94 1·48 1·29 1·38 1·44 1·3
Lu 0·26 0·26 0·30 0·24 0·24 0·19 0·22 0·20
Mg-no. 0·44 0·45 0·44 0·45 0·44 0·44 0·45 0·44
87
Sr/86Sr 0·703867 0·703765 0·703874 0·703898 0·703848 0·703884 0·703823 0·703859
143
Nd/144Nd 0·512934 0·512952 0·512933 0·512931 0·512923 0·512927 0·512939 0·51295
eNd (CHUR) 5·8 6·1 5·8 5·7 5·6 5·6 5·9 6·1
206
Pb/204Pb 19·024 19·059 19·065 19·027 19·021 19·038 19·050 19·028
207
Pb/204Pb 15·579 15·600 15·589 15·594 15·579 15·598 15·586 15·572
208
Pb/204Pb 38·631 38·750 38·760 38·724 38·692 38·742 38·738 38·667
(continued)
1114
Table 2: Continued
Sample: E05152 E05153 E05154 E05155 E05156 E05156 E05156

Rock: Bas. andesite Andesite Andesite Andesite Andesite Amphibole Groundmass
SiO2 53·74 56·94 56·79 56·59 57·47

TiO2 0·75 0·77 0·76 0·77 0·77
Al2O3 16·63 17·13 17·01 17·02 16·6
Fe2O3(tot) 8·92 8·48 8·28 8·53 7·85
MnO 0·16 0·14 0·14 0·15 0·13
MgO 3·56 3·36 3·25 3·42 3·2
CaO 7·48 7·43 7·27 7·37 7·11
Na2O 2·87 3·16 3·1 3·06 3·14
K2O 1·86 1·89 2·1 1·89 2·37
P2O5 0·28 0·29 0·31 0·29 0·33
LOI 2·76 0 0·32 0·31 0·36
Total 99·06 99·62 99·35 99·42 99·41
Rb 28·4 34·2 40·5 35·3 40·48
Sr 1344 983 1048 951 1188

Y 19·9 16·7 16·3 17·0 15·0
Zr 217 148 167 145 138
Nb 4·9 4·2 4·5 4·3 4·6
Cs 1·0 0·7 1·0 0·9 1·0
Be 2·4 1·3 2·5 1·7 4·4
Ba 1777 1170 1288 1136 1196
Hf 3·9 2·9 3·2 2·9 3·8
Ta 0·2 0·2 0·3 0·2 0·2
Pb 14·0 6·0 10·0 8·0 10·8
Th 23·9 9·1 10·6 9·3 11·8
U 6·0 2·5 2·9 2·5 3·2
Sc 22·0 19·8 19·1 20·4 18·1
V 203 191 185 190 181
Cr 11 5 4 6 19
Ni 10 9 9 9 14
Cu 29 59 45 64 45
Co 24 22 21 22 23
Zn 99 91 90 90 86
La 66·7 30·7 35·6 30·2 42·2
Ce 116·4 56·2 65·0 56·3 79·6
Pr 12·2 6·6 7·5 6·6 9·4
Nd 45·3 26·4 30·1 26·0 37·4
Sm 7·3 4·9 5·3 4·4 6·3
Eu 2·0 1·4 1·4 1·3 1·7
Gd 5·0 4·1 4·2 4·0 4·3
Tb 0·64 0·47 0·51 0·5 0·56
Dy 3·6 3·0 2·9 3·0 2·9
Ho 0·65 0·57 0·57 0·56 0·54
Er 1·73 1·58 1·42 1·55 1·39
Tm 0·25 0·22 0·21 0·22 0·18
Yb 1·68 1·34 1·37 1·35 1·22
Lu 0·25 0·20 0·20 0·22 0·19
Mg-no. 0·44 0·44 0·44 0·44 0·45
87
Sr/86Sr 0·703924 0·703873 0·703881 0·703864 0·703893 0·703889 0·703888
143
Nd/144Nd 0·512926 0·512926 0·512925 0·512929 0·512930 0·512948 0·512954
eNd (CHUR) 5·6 5·6 5·6 5·7 5·7 6·0 6·2
206
Pb/204Pb 19·137 19·023 19·025 19·072 19·023
207
Pb/204Pb 15·598 15·577 15·578 15·618 15·583
208
Pb/204Pb 38·850 38·694 38·708 38·804 38·694
(continued)
1115
Table 2: Continued
Sample: BHVO-1 SRM612
Rock: Basalt standard (average, n ¼ 8) 1s Glass standard (average, n ¼ 28) 1s
SiO2 49·60 0·08

TiO2 2·71 0·00
Al2O3 13·63 0·04
Fe2O3(tot) 12·22 0·01
MnO 0·17 0·00
MgO 7·21 0·11
CaO 11·39 0·03
Na2O 2·40 0·08
K2O 0·55 0·00
P2O5 0·28 0·00
LOI 0
Total 100·21 0·13

Rb 31·57 0·69
Sr 76·67 1·23
Y 38·51 0·75
Zr 36·23 0·72
Nb 38·32 0·73
Cs 41·92 0·84
Be 37·99 1·53
Ba 38·00 0·67
Hf 35·00 0·71
Ta 40·04 0·85
Pb 39·22 0·59
Th 37·48 0·73
U 37·40 0·77
Sc 41·33 0·77
V 39·49 0·45
Cr 299·20 4·76
Ni 130·20 1·10
Cu 36·96 0·61
Co 35·50 0·58
Zn 106·00 1·00
La 36·01 0·65
Ce 38·61 0·68
Pr 37·41 0·75
Nd 35·48 0·77
Sm 36·97 0·71
Eu 34·67 0·67
Gd 37·20 0·63
Tb 36·16 0·64
Dy 36·21 0·68
Ho 38·13 0·80
Er 37·68 0·78
Tm 37·80 0·81
Yb 40·22 0·88
Lu 37·96 0·81
Also reported are the averages and standard deviations of analyses obtained on the BHVO-1 basalt standard
(XRF analyses) and of the SRM612 glass standard (LA-ICP-MS analyses) under the same analytical conditions as used
for the analysis of whole-rocks (see text). Major elements, Cr, Ni and Zn were analyzed by XRF; all other elements by
LA-ICP-MS. b.d.l., below detection limit.
1116
fractionation using a 88Sr/86Sr value of 8·375209. Raw glass coronas of quartz xenocrysts, which are diopsidic
values were further corrected for external fractionation by (Electronic Appendix 1). Excluding sector-zoned clinopyr-
a value of þ0·03ø, determined by repeated measurements oxene phenocrysts, whose compositional zoning may be
of the SRM987 standard (87Sr/86Sr ¼ 0·710250). External the result of rapid growth, and the clinopyroxenes that
reproducibility of the 87Sr/86Sr ratio for the SRM987 are related to meltâmphibole and melt^quartz reactions,
standard is 7 ppm. Nd was loaded onto double Re fila- relic clinopyroxene phenocrysts surrounded by a corona of
ments with 1M HNO3 and measured in static mode with amphibole display the highest Na and Alvi contents,
the virtual amplifier design. 143Nd/144Nd values were in- followed by unzoned clinopyroxene of the most primi-
ternally corrected for fractionation using a 146Nd/144Nd tive rock (E05065), by clinopyroxene associated with
value of 0·7219 and the 144Sm interference on 144Nd was amphibole and finally by clinopyroxene in clots (Fig. 4a).
monitored on the mass 147Sm and corrected using a In particular, relic clinopyroxene phenocrysts have system-
144
Sm/147Sm value of 0·206700. External reproducibility of atically higher Na and Alvi contents than all other clino-
the JNdi-1 standard (Tanaka et al., 2000) is 55 ppm. pyroxene types within each sample, suggesting that
they were formed under higher pressure conditions.
Concentrically zoned clinopyroxene, both in clots and in
P E T RO G R A P H Y A N D M I N E R A L isolated grains, usually displays normal zoning with more
C H E M I S T RY magnesian cores with respect to rims. The various textural
Pilavo andesitic lavas are porphyritic with phenocrysts of types of clinopyroxene defined above may coexist in the
amphibole, clinopyroxene and, especially in the least same sample.

evolved rocks, plagioclase. Sizes of amphibole and clino- REE patterns of clinopyroxene are hump-shaped with
pyroxene phenocrysts range from several hundreds of an increasing slope from La to Nd and then a slightly
micrometers to almost 1cm in some cases, whereas plagio- decreasing slope from Sm to Lu, with no or very minor
clase phenocrysts range from a few hundreds of microm- Eu negative anomalies (Fig. 4b). Clinopyroxene trace
eters to a few millimeters. The groundmass consists of element compositions (Electronic Appendix 6) display sys-
plagioclase and clinopyroxene microlites (5150 mm) and tematic changes with host-rock composition. For example,
fresh glass. The abundance of amphibole phenocrysts sys- Sr, Sr/Y, Dy/Yb, and La/Yb of clinopyroxene increase
tematically increases and that of plagioclase systematically with the chemical evolution of the host-rock (e.g. increas-
decreases with increasing content of incompatible elements ing K2O of whole-rock: Fig. 4c and d).
in the whole-rock (see below). The amphibole-rich (plagio-
clase-poor) samples also contain quartz xenocrysts sur- Amphibole
rounded by glass and acicular clinopyroxene coronas (see We have distinguished three textural types of amphibole
below). In some samples (E05150) quenched mafic enclaves phenocryst: (1) amphibole with two or three concentric
in the basaltic andesite lava display a fluidal texture con- growth zones separated by a layer of fine-grained
sisting of plagioclase and pyroxene microlites (5150 mm). Fe-oxides (Fig. 3f); (2) amphibole similar to the type
above (two or three concentric growth zones separated by
Pyroxene a layer of fine-grained Fe-oxides) overgrown around a
Clinopyroxene is virtually the only pyroxene present, with single large relic clinopyroxene phenocryst or a cluster of
orthopyroxene found only in two cases (as a core replaced several clinopyroxene phenocrysts (Fig. 3a); (3) optically
by clinopyroxene in a corona around a quartz xenocryst homogeneous amphibole isolated (Fig. 3g) or in associ-
and as a reaction rim assemblage around an amphibole ation with clinopyroxene (Fig. 3d). Amphibole types (1)
phenocryst). We have distinguished five textural types of and (2) occur only in the most evolved rocks, whereas
clinopyroxene: (1) large (millimeter- to centimeter-sized) type (3) occurs in all the rocks of Pilavo.
relic phenocrysts surrounded by a corona of amphibole The three textural types of amphibole phenocrys are
(Fig. 3a); (2) isolated phenocrysts (Fig. 3b) with either con- surrounded, when in contact with the groundmass, by
centric or sector zoning (hundreds of micrometers to sev- variably thick reaction rims of two types (Fig. 3eî): (1)
eral millimeters) and idiomorphic to sub-idiomorphic opacite rims (Fig. 3g) and/or (2) gabbroic (plagioclase^
shapes; (3) clots of several clinopyroxene crystals (hun- pyroxene^Fe oxide) rims (Fig. 3e, f and h). Gabbroic rims
dreds of micrometers to several millimeters; Fig. 3c) with are usually considered the result of amphibole^melt reac-
either sector or concentric zoning; (4) clots of several tion during decompression (e.g. Rutherford & Hill, 1993).
clinopyroxene crystals in association with amphibole The thickness of the gabbroic rims, as well as the size and
plagioclase (hundreds of micrometers to several milli- shape of the crystals in these rims, varies systematically
meters; Fig. 3d); (5) products of the reactions of melt with from sample to sample (from a few tens to a few hundreds
amphibole (Fig. 3f) or quartz xenocrysts (Fig. 3j^l). of micrometers; Fig. 3e^h: see below) and even for different
All these textural types have augitic compositions, amphiboles within the same thin section. Crystals in the
except for the acicular clinopyroxenes surrounding the reaction rims may be more or less elongated (to acicular)
1117

Fig. 3. Photomicrographs of Pilavo lava textures. (a) Amphibole megacryst with a poikilitic core of high Al2O3-clinopyroxene and Fe-oxides
(sample E05156: parallel Nicols). (b) Sector-zoned clinopyroxene phenocryst (sample E05156: crossed Nicols). (c) Clinopyroxene clot (sample
E05150: crossed Nicols). (d) Amphibole and clinopyroxene (cpx) clot (sample E05066: parallel Nicols). (e) Rounded amphibole surrounded by
a fine-grained gabbroic rim and a corona of coarse-grained clinopyroxene (sample E05151: parallel Nicols). (f) Optically multi-zoned amphibole
with gabbroic rim (sample E05153: parallel Nicols). (g) Amphibole with opacite rim (sample E05067: parallel Nicols). (h) Amphibole with
thick gabbroic rim (sample E05064: parallel Nicols). (i) Amphibole completely replaced by pyroxene^plagioclase oxides (sample E05065:
crossed Nicols). (j) Corona texture around quartz xenocryst (sample E05067: binocular microscope). From core to rim: white-transparent
(quartz, qz), blackish (glass, gl), light grey (clinopyroxene, cpx). (k) Clinopyroxene (cpx) corona texture around quartz (qz) xenocryst
(sample E05067: binocular microscope). (l) Acicular clinopyroxene in corona texture around quartz xenocryst (accidentally removed during
thin section preparation; parallel Nicols: sample E05067).
and fairly large (up to a few tens of micrometers). In some Pilavo amphiboles can be classified as ferri-tschermakite
cases no residual amphibole is left (Fig. 3i). In a few cases to magnesio-hastingsite (Electronic Appendix 2) accord-
amphibole phenocrysts are surrounded by a corona of ing to the scheme of Leake et al. (1997). They are character-
equant clinopyroxene crystals (Fig. 3e), which in turn sur- ized by roughly constant Alvi with variable Aliv,
round the rim of plagioclase^pyroxeneôxide aggregate. suggesting the lack of pressure-sensitive Al-Tschermaks
1118
0.04
(a) (b)
rock/chondrite
100
0.03
Na
10
0.02
1
0.01
0.02 0.03 0.04 0.05 0.06 0.07 0.08 La Pr Pm Eu Tb Ho Tm Lu
Ce Nd Sm Gd Dy Er Yb
Alvi
7 (d)
140 (c)
6
La/Yb
Sr (ppm)
5
100
4

3
60
2
1
20
0.5 1.0 1.5 2.0 2.5 0.5 1.0 1.5 2.0 2.5
K2OWR (wt.%) K2OWR (wt.%)
Symbols for Figure 4A Symbols for Figure 4B-D

relic cpx with amph corona E05151 E05153
cpx associated with amph E05150 E05154
cpx clot E05064 E05156
cpx of sample E05065 E05066 E05067
Fig. 4. Clinopyroxene chemistry: (a) Na vs Alvi plot for clinopyroxenes in Pilavo lavas showing that clinopyroxene relict crystals surrounded by
amphibole coronas and clinopyroxene associated with amphiboles have generally the highest Na and Alvi contents; (b) REE patterns for
Pilavo clinopyroxenes; (c, d) Sr and La/Yb variations of Pilavo clinopyroxenes vs K2O contents of the corresponding host whole-rocks.
exchange. In contrast, they display a broadly positive cor- amphibole displays systematic changes with the chemistry
relation between (Na þ K)A and Aliv suggesting a T-de- of the host-rock; for example, K2O, Sr and Rb increase
pendent edenite exchange. Therefore, the variable Al2O3 with increasing whole-rock K2O (Fig. 5b).
contents of the Pilavo amphiboles reflect T rather than P Amphibole with optically distinct growth zones displays
changes. None the less, the overall high-Al2O3 ( 12 wt Mg-number (TiO2, K2O) increase and Na2O decrease
%) Pilavo amphiboles are similar in composition to the (Fig. 5c) from the core to the adjacent outer zone. Rims
high-Al2O3 amphiboles of Mount St. Helens, which are can be either more or (slightly) less magnesian than the
considered to have crystallized at depths 415 km preceding zone, but are generally more magnesian than
(Thornber et al., 2008). cores (Fig. 5c). Optically homogeneous amphiboles are
REE patterns of amphiboles are similar to those of generally richer in Mg than cores of zoned amphiboles
clinopyroxene, with slope increasing from La to Nd and and often as Mg-rich as their rims (compare E05153 and
slightly decreasing from Sm to Lu and no significant Eu E05066: Fig. 5c and d). They are characterized by a slight
negative anomalies (Fig. 5a; Electronic Appendix 7). decrease of Mg-number and K2O and by an increase of
REE concentrations of amphibole are comparable with to TiO2 from core to rim (Fig. 5d). Amphiboles with variably
slightly lower than those of clinopyroxene (Figs 4b and Mg-rich cores may also coexist in the same sample. Pilavo
5 a; Electronic Appendixes 6^7). Like clinopyroxene, amphiboles, especially the rims of crystals from the most
1119
Fig. 5. (a) REE patterns of Pilavo amphiboles. (b) Sr variations of Pilavo amphiboles vs K2O contents of the corresponding host whole-rocks.
(c, d) Mg-number, TiO2, K2O and Na2O concentration profiles in optically zoned and unzoned amphiboles from Pilavo lavas. Reported dis-
tances (mm) are always from the core (0 mm). (e, f) Correlations between average reaction rim thickness around amphiboles and their average
Sr and Ba contents (error bars represent the associated standard deviations).
1120
evolved rocks, can have high F contents, ranging from 0·1 and hercynite components are highest in the magnetite
to 2·7 wt % (fluor-magnesio-hastingsite). hosted by amphibole, in which MgO and Al2O3 can
The average thickness of the reaction rims of the amphi- reach concentrations up to 11wt % (Fig. 7c; Electronic
bole population in each thin section (Table 2) shows a Appendix 4). Also, hemoilmenite has high contents of
broad inverse correlation with average incompatible elem- MgO (up to 8 wt %) and Al2O3 (up to 3·5 wt %;
ent content of the amphibole (Fig. 5e and f). Electronic Appendix 4).
Application of the magnetiteîlmenite thermometer and
Plagioclase oxygen barometer [using the ILMAT software of Lepage
We have distinguished three texturally different types of (2003)] for the magnetite^hemoilmenite pairs of the
plagioclase, as follows. amphibole phenocrysts [which have all passed the equilib-
rium test of Bacon & Hirschmann (1988)] yields tempera-
(1) Plagioclase phenocrysts (up to a few millimeters in
tures ranging between 850 and 9008C, with one pair
size) consisting almost entirely of a heavily
yielding a significantly higher temperature of 9758C
sieve-textured core, which is corroded and overgrown
(Fig. 7d; Electronic Appendix 4). Oxygen fugacities calcu-
by a thin clear rim (Fig. 6a). The cores of these plagio-
lated from these pairs are high to very high, ranging
clase phenocrysts are anorthite-rich (An480 and up
from NNO þ1·5 to NNO þ 3·3, where NNO is the nickel^
to An92) with the sieve-filling plagioclase being
nickel oxide buffer.
An35^45 (Fig. 6a and b; Electronic Appendix 3).
The clear outer rim, with concentric zoning, is usually
An-poorer than the core (Fig. 6a; Electronic Quartz xenocrysts

Appendix 3). A typical textural feature of only the most evolved Pilavo
(2) Oscillatory zoned phenocrysts (300^600 mm) with rocks is the occurrence of oval zoned globules (Fig. 3j^l),
poorly to non-sieved cores characterized by high sub-millimeter to several millimeters in size, consisting of
anorthite contents (An80^85) and a thin rim with (1) an outer zone of coarse and equant crystals of clinopyr-
lower anorthite content. oxene followed inward by (2) a zone of acicular clinopyrox-
(3) Microcrysts (150^300 mm) in the matrix without ene crystals growing radially inwards in a glass
sieved cores, characterized by a strong anorthite de- groundmass passing to (3) pure glass and finally to (4)
crease from core to rim (Fig. 6b and c) and usually quartz. Some globules consist entirely of plagioclase and
less anorthitic than phenocrysts. clinopyroxene crystals with interstitial glass. In other
cases the globule is completely occluded by acicular clino-
Plagioclase displays systematic chemical changes with the
pyroxene crystals growing radially from the rim towards
composition of the host-rock. In particular, the maximum
the core of the globule. These textures are identical to the
Sr contents of plagioclase are positively correlated with
corona textures around quartz xenocrysts in basaltic lavas
increasing evolution of the host-rocks (e.g. increasing
described by Sato (1975). Acicular clinopyroxene in the
K2O, Sr and Th: Fig. 6d), despite the fact that Sr lacks
globules has a diopsidic composition (Electronic
any correlation with plagioclase anorthite content in our
Appendix 1) whereas equant clinopyroxene on the rim of
samples.
the globules is augitic like the clinopyroxene in the lava
and may surround a corroded core of orthopyroxene.
Fe^Ti oxides The glass of coronas from samples E05066 and E05155
Titanomagnetite occurs within amphibole (Fig. 7a), clino- (Electronic Appendix 5) has high SiO2 (73^80 wt %),
pyroxene (Fig. 7b) and, more rarely, plagioclase, and, to Al2O3 (8·6^9·3 wt %), K2O (4·6^5·7 wt %) and Na2O
the greatest extent, as single pheno- and microcrysts in (1·9^2·8 wt %), and low FeO (52 wt %), MgO (50·1wt
the groundmass. Titanomagnetite in amphibole pheno- %) and CaO (50·3 wt %), similar to glass compositions
crysts is often associated (in contact) with hemoilmenite in other coronas around quartz xenocrysts (e.g. Sato,
(Fig. 7a), whereas ilmenite has never been found within 1975). All major elements (except the most compatible
clinopyroxene and plagioclase phenocrysts or in the ones such as MgO, FeO plus TiO2) display negative correl-
groundmass, except as exsolution lamellae. ations with SiO2. REE and incompatible element spectra
Compositional variations of groundmass magnetite of the glass of the globules are similar to those of the
result from variations in the ulvo«spinel and combined whole-rocks (Fig. 8a and b), although they are poorer in
magnetite^magnesioferrite components, whereas the com- most incompatible elements. The only elements with con-
bined spinel^hercynite component remains low (3%). centrations in the glass higher than in the whole-rock
In contrast, magnetite included in amphibole and clino- are Rb, Cs, Pb and K (see also Sato, 1975). Glasses
pyroxene phenocrysts is characterized by an increasing in the corona textures contain significant amounts
spinel^hercynite component at the expense of ulvo«spinel of B (7^23 ppm) and F (1000^2000 ppm) (Electronic
and magnetite^magnesioferrite components. The spinel Appendix 5).
1121
(a) (b)
pl 1
pl 2
90 90
80
80
An (mol%)
An (mol%)
70
pl 1
60 70
50
60 pl 2
40

30 50
0 100 200 300 400 0 100 200 300 400
Distance (µm) Distance (µm)
5000
(c) (d)
Srplagioclase (ppm)
4000
3000
2000
1000
0
90 600 800 1000 1200 1400
SrWR (ppm)
80
An (mol%)
70 E05151 E05064 E05154

E05065 E05066 E05156
60 E05150 E05153 E05067
50
100 200 300 400
Distance (µm)
Fig. 6. Anorthite zoning of plagioclase phenocrysts for large phenocrysts with intensely sieved cores (a) characterized by albite-rich fillings of
the sieves (An30^40) (sample E05065), phenocrysts with weakly sieved cores (b) (sample E05067), and microcrysts in the groundmass (b and
c) (sample E05067); (d) variations of Sr in plagioclase with whole-rock Sr content.
1122

Fig. 7. Reflected light photomicrographs of Fe-oxide textures in Pilavo lavas. (a) Magnetite (mt)^hemoilmenite (ilm) association within
amphibole (amph) phenocryst (sample E05156); (b) magnetite (mt) hosted by clinopyroxene (cpx) (sample E05066); (c) Al2O3 vs TiO2 vari-
ations for Pilavo magnetites; (d) fO2^T plot for the magnetite^hemoilmenite pairs within amphibole and for the magnetiteîlmenite
pair of the groundmass (sample E05153) obtained using the ILMAT software of Lepage (2003) and ulvo«spinel and ilmenite molar proportions
calculated according to Lindsley & Spencer (1982). The two points yielding low T (7508C) correspond to non-equilibrium magnetite with la-
mellar ilmenite exsolution in the groundmass. All others are equilibrium pairs hosted by amphibole.
element concentrations display a wide range of variability

G E O C H E M I S T RY
(Rb ¼12^45 ppm, Sr ¼ 650^1350 ppm, Ba ¼ 600^
Major, trace and rare earth elements 1800 ppm, Zr ¼ 80^225 ppm, Th ¼ 4^24 ppm, Pb ¼
The Pilavo lavas are medium- to high-K basaltic andesites 6^14 ppm, U ¼1^6 ppm, La ¼10^70 ppm, Ce ¼ 20^
and low-SiO2 andesites (Fig. 9a) with only a small range 120 ppm) and are significantly higher than at Pichincha
in major element composition (SiO2 ¼54^57·5 wt %, and Pululahua (Fig. 9g and h), whereas compatible elem-
MgO ¼ 3·1^3·8 wt %, Fe2O3 ¼7·8^9·3 wt %; Table 2) in ents (Ni ¼ 4^14 ppm, Cr ¼ 4^19 ppm) are significantly
contrast with the medium-K high-SiO2 andesites and da- lower (Fig. 9i and j). Incompatible elements are strongly
cites of the other frontal volcanoes of Ecuador (e.g. correlated with each other (Fig. 9k).
Pululahua, Pichincha; Fig. 9a). Additionally, the Pilavo In primitive mantle-normalized trace element diagrams
rocks fall almost entirely within the tholeiitic field as (Fig. 8a), Pilavo rocks display typical subduction-related
opposed to those from Pululahua and Pichincha (Fig. 9b). geochemical signatures with LILE enrichment and strong
MgO displays positive correlations with Fe2O3 and CaO negative Nb and Ta anomalies. Compared with Pululahua
and negative correlations with Na2O and K2O (Fig. 9c^f). and Pichincha, they show deeper P and Ti negative anoma-
In contrast to major elements, incompatible trace lies (Fig. 8a), a smaller negative Th anomaly and a general
1123
1000
Rock/Primitive Mantle Pilavo (a) (b)
Pululahua
100
Pichincha
Rock/Chondrites
100
corona glass
(E05155)
10
10
1
1
Cs Ba U Ta La Pb Sr Nd Sm Ti Y Lu La Pr Pm Eu Tb Ho Tm Lu
Rb Th Nb K Ce Pr P Zr Eu Dy Yb Ce Nd Sm Gd Dy Er Yb
100 1000
(c) (d)

Rock/Chondrites
10
Rock/MORB
100
Setouchi sanukites 10
0.1 Sanukitoids
Pilavo
Continental crust
0.01 1
Rb Ta Ce Sr Zr Ti Yb Cr La Pr Sm Gd Dy Yb
Ba K Pb P Sm Dy Lu Ce Nd Eu Tb Er Lu
Nb La Pr Nd Eu Y Ni
Fig. 8. (a) Primitive mantle-normalized trace element variation diagrams for the Pilavo rocks and of the corona texture glass around quartz
xenocrysts (sample E05155) compared with those of Pichincha and Pululahua. The corona texture glass is the average of 12 LA-ICP-MS spot
measurements. Normalizing values are from Sun & McDonough (1989). (b) Chondrite-normalized REE patterns of the Pilavo lavas and of
the corona texture glass around quartz xenocrysts (sample E05155) compared with those of Pichincha and Pululahua. The corona texture
glass is the average of 12 LA-ICP-MS spot measurements. Normalizing values are from Sun & McDonough (1989). (c) MORB-normalized
trace element patterns of the Pilavo lavas compared with those of the average continental crust (Rudnick & Gao, 2004), Setouchi sanukites
(Shimoda et al., 1998; Tatsumi et al., 2003), and sanukitoids (Martin et al., 2005). The fields of Setouchi sanukites and sanukitoids represent 1
standard deviations around the averages of element concentrations. It should be noted that compared with (a) some elements have been
removed because they are not available in the sanukitoid and sanukite databases and Cr and Ni have been added. MORB-normalizing values
are from Kelemen et al. (2004). (d) Chondrite-normalized REE patterns of the Pilavo rocks and of the corona texture glass around quartz xeno-
crysts (sample E05155) compared with those of the average continental crust (Rudnick & Gao, 2004), Setouchi sanukites (Shimoda et al., 1998;
Tatsumi et al., 2003), and sanukitoids (Martin et al., 2005). The fields of Setouchi sanukites and sanukitoids represent 1 standard deviations
around the averages of element concentrations. Compared with (c) Pr and Ho have been removed because they are not available in the
sanukitoid database (see Martin et al., 2005). Normalizing values are from Sun & McDonough (1989).
enrichment in all incompatible elements. In a mid-ocean are characterized by a moderately steep LREE to middle
ridge basalt (MORB)-normalized diagram including com- REE (MREE) slope, a flat MREE to HREE slope and
patible elements (Fig. 8c) Pilavo lavas display some affi- the absence of negative Eu anomalies (Fig. 8b), and,
nities with high-MgO andesites such as Setouchi sanukites except for the two least evolved rocks, are shifted to
(and average continental crust), from which they differ higher concentrations (especially LREE) compared with
markedly in having higher Sr and P and significantly Pululahua and Pichincha. In the context of their REE pat-
lower Cr and Ni contents. Pilavo rocks, especially the terns the Pilavo rocks show significant affinity both with
most enriched ones, display a remarkable affinity with Archean sanukitoids (the most enriched rocks) and with
Archean sanukitoids, from which they differ only in terms Setouchi sanukites (the least enriched rocks) (Fig. 8d).
of their lower Ni and Cr contents (Fig. 8c). REE patterns The affinity with Setouchi sanukites and Archean
1124
10 4
(a) (b) TH
Rhyolite
Trachy- 29 Okmok
Trachy- dacite CA
Na2O+K2O (wt.%)
FeOTOT/MgO
Basalt. andesite 3
trachy- Parinacota
andesite 25
Dacite
6 Trachy-
basalt 2
15
Andesite
4 andesite
Basaltic
Basalt Pilavo 1
2 Pululahua Lassen
Pichincha Setouchi Piip Peak
0 0
45 50 55 60 65 70 75 48 52 56 60 64 68
SiO2 (wt.%) SiO2 (wt.%)
10 9
(c) 30 28 26 plag (d) mt
24 22
9 cpx 24 22 28 26
Fe2O3 (wt.%)
8 30
CaO (wt.%)
32

8 amph
amph plag
7 7
6 cpx
6
5 mt
4 5
5 3.0
(e) (f)
Na2O (wt.%)
2.5
K2O (wt.%)
mt
4
2.0
cpx mt
plag 1.5 cpx plag
3 30 26
amph 30
28 26 1.0 amph
24 22 22
2 0.5
1 2 3 4 5 1 2 3 4 5
MgO (wt.%) MgO (wt.%)
Fig. 9. (a) Total alkalis^silica diagram (Le Bas et al., 1986) for Pilavo lavas compared with Pichincha and Pululahua. (b) FeOtot/MgO^SiO2
diagram (Miyashiro, 1974) for Pilavo lavas compared with other arc volcanoes (TH, tholeiitic; CA, calc-alkaline). Sources of data: Piip,
Yogodzinski et al. (1994); Okmok, Finney et al. (2008); Lassen Peak, Clynne et al. (2008); Parinacota, Hora et al. (2009); Pululahua
and Pichincha, Chiaradia et al. (2009). (c^j) MgO versus selected major and trace elements for Pilavo lavas compared with Pululahua and
Pichincha. Modeled fractional crystallization trends of an average continental arc basalt (Kelemen et al., 2004) fractionating 33% olivine,
64% clinopyroxene, and accessory amounts of Cr-rich spinel (1%) and magnetite (2%) are indicated by continuous lines. Tick marks
and corresponding numbers represent the crystallized fraction (see Tables 4^6 for further details). Modeled fractional crystallization
trends for amphibole, clinopyroxene, magnetite and plagioclase starting from the composition of a melt corresponding to 25% fractionation of
the continental arc basalt according to the above model are indicated by the dashed lines (see Tables 4^6 for further details). (k) Modeling of
repeated recharge and assimilation^fractional crystallization (AFC) processes in Ba^La space (see Table 7 for details). (l) Sr/Y vs K2O for
Pilavo lavas.
1125
25
(g) 1300 (h)
20
Sr (ppm)
Th (ppm)
1100
mt
15
900 cpx
10 700 24 22
cpx, amph amph 30 28 26 plag
5 mt plag 500
24
32 26 22
0 300
70
60
(i) 160 (j)
50
Cr (ppm)
Ni (ppm)
120
40
30 22 80 22

20 cpx, amph, 24 40 24
10 mt plag cpx mt plag
0 0 amph
1 2 3 4 5 1 2 3 4 5
MgO (wt.%) MgO (wt.%)
80
(k) 5 (l)
80
60
La (ppm)
4
Sr/Y
40 60
3
2
20 1 40
0
500 1000 1500 2000 0.5 1.0 1.5 2.0 2.5
Ba (ppm) K2O (wt.%)
Fig. 9. Continued.
sanukitoids extends also to major elements (Table 3), with et al., 2005). Adakitic indices (e.g. Sr/Y, but also La/Yb
the exception of Fe2O3 (higher at Pilavo) and MgO and Dy/Yb), increase regularly with the degree of differen-
(lower at Pilavo, but only with respect to Setouchi tiation (e.g. K2O; Fig. 9l).
sanukites).
Some of the Pilavo rocks can be classified as adakite-like Sr, Nd and Pb isotopes
in terms of their Sr/Y (40^80) and La/Yb (10^40) values, The Sr, Nd and Pb isotopic compositions of the Pilavo
but compared with Pichincha and Pululahua they rocks (Table 2) show little variability (87Sr/86Sr ¼ 0·70377^
have higher concentrations of Y (15^20 ppm) and Yb 0·70390, 143Nd/144Nd ¼ 0·51292^0·51295, eNd ¼ þ5·5 to
(1·2^2·0 ppm), which, for some samples, are beyond the þ6·1, 206Pb/204Pb ¼19·02^19·08, 207Pb/204Pb ¼15·57^15·62,
208
threshold of adakites (Y 18 ppm; Yb 1·8 ppm; Martin Pb/204Pb ¼ 38·63^38·85) and are always more depleted
1126
Table 3: Comparison of average major element compositions of Pilavo lavas, Archean sanukitoids and Setouchi sanukites
SiO2 TiO2 Al2O3 Fe2O3 MgO CaO Na2O K2O P2O5 FeO/MgO
Sanukitoid average
(n ¼ 31) 58·76 0·74 15·8 5·87 3·9 5·57 4·42 2·78 0·39 1·35
1s 2·90 0·30 0·90 1·50 1·30 1·50 0·70 0·80 0·10 0·42
Setouchi sanukite
average (n ¼ 30) 57·38 0·6 15·87 6·23 7·22 6·59 3·15 1·74 0·15 0·78
1s 1·85 0·11 0·78 0·7 1·34 0·54 0·39 0·36 0·02 0·22
Pilavo average
(n ¼ 13) 56·29 0·75 17·22 8·55 3·42 7·51 3·07 1·81 0·28 2·25
1s 1·16 0·04 0·46 0·48 0·21 0·34 0·11 0·45 0·03 0·08
Residual difference
Pilavo – sanukitoids (%) ¼ ¼ þ0·3 þ8·2 ¼ þ1·3 –17·6 ¼ ¼ þ17·8
Residual difference
Pilavo – sanukite (%) ¼ ¼ þ0·6 þ13·3 65·8 0·5 ¼ ¼ 28·6 52·0

Residual differences (%) represent the minimum per cent difference between the averages taking into account the
1s errors (¼ means that the averages overlap within 1s) and are calculated with respect to Pilavo average value.
Data for Archean sanukitoids are from Martin et al. (2005) and those for Setouchi sanukites are from Shimoda et al.
(1998) and Tatsumi et al. (2003).
with respect to Sr and Nd but more radiogenic for Pb than such enrichment might have occurred. Finally, we discuss
rocks of companion frontal volcanoes (i.e. Pululahua and the tholeiitic affinity and the incompatible element enrich-
Pichincha; Fig. 10). In both Sr^Nd (Fig. 10a) and Sr^Pb ment of Pilavo basaltic andesites in a more general context.
(Fig. 10b) isotope spaces the Pilavo compositions are slight-
ly shifted towards the Galapagos Islands field compared Mantle source processes
with those of companion frontal arc volcanoes, but, like In agreement with the geodynamic setting of the volcano,
the latter, plot fully within the field of granulites and am- the available geochemical data indicate that Pilavo
phibolites of the Western Cordillera of Ecuador, which rep- magmas are subduction-related (strong negative Nb and
resent the metamorphic equivalents of the oceanic plateau Ta anomalies; LILE enrichment). Therefore, following es-
rocks accreted to the continent in the Late Cretaceous tablished models (e.g. Tatsumi, 1989) and in the absence of
(Amo¤rtegui et al., 2005; Chiaradia et al., 2009). The Sr and primitive magmatic rocks, the most reasonable parent
Nd isotopic compositions of the microdrilled amphibole magma of the Pilavo basaltic andesites is a hydrous basalt
phenocryst and groundmass of sample E05156 are identical derived from partial melting of the mantle wedge metaso-
within error and are also similar to the whole-rock com- matized by a slab component. The most primitive samples
position (Table 2). from Pilavo have Ba/La (50) and Pb/Ce (0·25^0·30),
Despite being characterized by a small range of isotopic ratios similar to those of the most primitive magmas of
variability the Pilavo lavas show significant correlations the adjacent frontal arc volcanoes in Ecuador (Pululahua
of Sr and Nd (not Pb) isotopes with incompatible elem- and Pichincha; Fig. 12a), indicating that the aqueous(?)
ents and their ratios (Fig. 11). In general, 87Sr/86Sr values (Ba and Pb are aqueous fluid-mobile elements; e.g. Kessel
increase whereas 143Nd/144Nd values decrease with increas- et al., 2005) slab component contribution was probably
ing incompatible element concentrations and their ratios similar for all three volcanoes. None the less, the more
(Fig. 11). primitive isotopic compositions and significant enrichment
in aqueous fluid-immobile incompatible elements (e.g. Th,
La, Zr, Nb) as well as their ratios (e.g. Th/La, Th/Nb) in
DISCUSSION the Pilavo lavas, compared with those of Pichincha and
The main focus of this study is the petrogenesis of the Pululahua (Fig. 12b and c), might suggest a fundamental
Pilavo basaltic andesites and, in particular, the processes difference in the slab component or in the mantle source.
responsible for their enrichment in incompatible elements. High Th/La values (0·28^0·36) and Th concentrations (4^
Below, we discuss the steps, from magma source to 24 ppm), like those of the Pilavo andesites, are usually con-
magma differentiation at various crustal levels, at which sidered to reflect a subducted sediment (melt) component
1127
0.5134
DMM Ecuadorian oceanic terranes (a)
EPR (Western Cordillera and coastal plain)
0.5132
GSC
0.5130
143Nd/144Nd GIs Early Tertiary
HIMU Macuchi island arc
0.5128
Western Cordillera BSE

0.5126
granulites and amphibolites
0.5124
EMI
0.5122
0.702 0.703 0.704 0.705 0.706
87Sr/86Sr

0.706
Early Tertiary (b)
Pilavo
Macuchi island arc
Pichincha
0.705 Pululahua
87Sr/86Sr
Western Cordillera
granulites and
0.704 amphibolites
GIs
0.703
GSC
Ecuadorian oceanic terranes
EPR (Western Cordillera and coastal plain)
0.702
18 19 20
206Pb/204Pb
Fig. 10. Nd^Sr (a) and Sr^Pb (b) isotope variation in the lavas of the studied volcanoes in comparison with fields of various potential source
reservoirs and rocks (data from Zindler & Hart, 1986; White et al., 1993; Litherland et al., 1994; Mamberti et al., 2003; Amo¤rtegui et al., 2005;
M. Chiaradia, unpublished data). DMM, Depleted MORB Mantle; EPR, East Pacific Rise; GSC, Galapagos Spreading Center; HIMU,
high m; GIs, Galapagos Islands; BSE, Bulk Silicate Earth; EMI, Enriched Mantle I.
in the mantle source (e.g. Plank, 2005) and therefore might and Carnegie Ridge lavas (Harpp et al., 2005) have low
suggest a higher subducted sediment melt contribution at Th/La values (0·021^0·15 and 0·034^0·092, respectively),
Pilavo than at Pichincha and Pululahua. However, the implying that the high Th/La values of the Pilavo rocks
high Th/La values and Th concentrations of Pilavo lavas are not the result of sediment and/or oceanic crust melting
are inconsistent with their isotopic signatures, which are and might, instead, be the result of intracrustal magmatic
more depleted than those of Pululahua and Pichincha processes (see also Plank, 2005). This, however, does not
(Figs 10a, b and 12 c). Furthermore, subducting sediments exclude the possibility that the metasomatizing agent re-
at adjacent trench locations (Colombia and Peru, Plank sponsible for mantle wedge flux melting in the subduction
& Langmuir, 1998; no geochemical data are available for zone beneath Pilavo may be a hydrous melt, as suggested
sediments of the Ecuadorian trench to our knowledge) for subduction zones in general by thermal modeling of
1128
0.70395 0.70395
(a) (b)
0.70390 0.70390
87Sr/86Sr
87Sr/86Sr
0.70385 0.70385
0.70380 0.70380
0.70375 0.70375
500 1000 1500 2000 600 800 1000 1200 1400
Ba (ppm) Sr (ppm)
0.70395 0.70395
(c) (d)
0.70390 0.70390
87Sr/86Sr
87Sr/86Sr
0.70385 0.70385

0.70380 0.70380
0.70375 0.70375
50 100 150 200 250 5 10 15 20 25
Zr (ppm) Th (ppm)
0.70395 0.70395
(e) (f)
0.70390 0.70390
87Sr/86Sr
87Sr/86Sr
0.70385 0.70385
0.70380 0.70380
0.70375 0.70375
35 40 45 50 55 60 65 70 75 80 85 25 30 35 40 45 50 55
Sr/Y Ba/La
0.51296 0.51296
(g) (h)
0.51295 0.51295
143Nd/144Nd
143Nd/144Nd
0.51294 0.51294
0.51293 0.51293
0.51292 0.51292
1.0 1.5 2.0 2.5 50 100 150 200 250
K2O (wt.%) Zr (ppm)
Fig. 11. Correlations between whole-rock radiogenic isotope compositions and various incompatible trace elements and ratios (for discussion
see text).
1129
0.5 (i.e. MgO-, Fe2O3-, CaO-richer) Pilavo rocks, which are

(a) much lower than the andesitic^dacitic companion frontal
0.4 volcanoes situated more to the south (Fig. 9), suggest exten-
sive fractional crystallization of olivine and clinopyroxene.
Pb/Ce
High Al2O3 (18 wt %) and Sr (700 ppm) contents in

0.3 the most primitive basaltic andesites of Pilavo indicate
the lack of significant plagioclase fractionation. A liquidus
0.2 assemblage of olivine and clinopyroxene (without plagio-
clase), resulting in depletion of MgO and Ni and enrich-
ment of Al2O3 but no significant enrichment of SiO2, is
0.1 typical of high-pressure (e.g. mantle^crust interface) frac-
20 40 60 80 100 tionation of mantle-derived basaltic melts and generates
Ba/La high-Al2O3 hydrous basalts and basaltic andesites (see
also Crawford et al., 1987; Foden & Green, 1992; Mu«ntener
5
(b) et al., 2001). We have modeled fractional crystallization of a
primitive magma with the composition of an average
4
continental arc basalt (Kelemen et al., 2004), involving a
clinopyroxene- and olivine-dominated high-pressure as-
3
Th/Nb
semblage (continuous black lines in Fig. 9). Modeling can

reproduce consistently the major and trace element geo-
2
chemistry of the least evolved rocks of Pilavo for 25%
fractional crystallization of a bulk assemblage consisting
1
of olivine (33%), clinopyroxene (64%), plus accessory
amounts of magnetite (2%) and magnesiochromite (1%)
0
0.0 0.1 0.2 0.3 0.4 (see Tables 4^6 for further details). Although based on a
primitive melt composition that does not necessarily cor-
Th/La respond to that of the Pilavo parental magmas and on pro-
0.7043 portions of the fractionating assemblage that cannot be
(c) Pichincha and Pululahua verified, this modeling exercise shows that, as a first ap-
0.7042 most primitive rocks proximation, high-pressure fractionation of spinel-bearing
olivine-clinopyroxenite from a primitive arc basalt is a
87Sr/86Sr
0.7041 Pilavo most primitive rock

suitable process to explain the strong compatible element
0.7040 depletion of the Pilavo basaltic andesites. Importantly, our
modeling also shows that such a fractionating assemblage
0.7039 will drive the primitive basaltic melt into the tholeiitic
field (Fig. 9b), reproducing the tholeiitic compositions of
0.7038
the least evolved Pilavo rocks for 25% crystallization,
0.7037 consistent with our modeling of the other major and trace
0 5 10 15 20 25 elements (Fig. 9c^j). However, if the least evolved rocks
Th (ppm) of Pilavo can be derived by high-pressure olivine-
clinopyroxenite fractionation of a primitive arc basalt, the
Fig. 12. Pb/Ce vs Ba/La (a), Th/Nb vs Th/La (b) and 87Sr/86Sr vs Th geochemical trends of the Pilavo rocks (Fig. 9b^j) require
(c) plots for Pilavo, Pululahua and Pichincha volcanoes (for discus-
sion see text). in contrast either a different fractionating assemblage or al-
ternative magmatic processes.
subducted slabs (van Keken et al., 2002; Kincaid & Griffiths, Recharge and mixing in crustal magma
2004; Plank et al., 2009), experimental studies on element chamber(s)
partitioning (e.g. Kessel et al., 2005), and trace element data
Clinopyroxene, amphibole, plagioclase and magnetite are
on high-pressure rocks (Hermann et al., 2006).
the dominant phenocryst phases in the Pilavo basaltic an-
desites as well as in the Pululahua and Pichincha andesites
High-pressure olivine and clinopyroxene or dacites. Figure 9 shows model fractionation trends
fractionation resulting from the removal of each of these mineral
The very low compatible element concentrations at phases. The variations in major and trace element concen-
low SiO2 contents (e.g. Ni ¼ 6^14 ppm, Cr ¼ 4^19 ppm, trations of Pululahua lavas (notably K2O, Th and Sr:
Mg-number 0·44^0·45) even of the least evolved Fig. 9f, i and j) can be explained almost entirely by
1130
Table 4: Compositions of primitive melts used for the modeling curves reported in Figure 9
Primitive melt* Model SiO2 TiO2 Al2O3 FeOtot MnO MgO CaO Na2O K2O P2O5 Cr Ni Th Sr Total Ref.
Continental arc 1 51·33 0·98 15·7 8·72 0·17 9·48 9·93 2·61 0·88 0·22 398 159 2·03 426 100·10 1
E05065 2 56·07 0·71 17·85 8·22 0·16 3·73 7·97 3·01 1·13 0·23 13 6 4·5 666 99·08 2
*Oxides in wt %, trace elements in ppm.

Model 1 refers to the modeling of the high-pressure stage at the mantle–crust interface or within the lower crust. Model 2
refers to the modeling of the Pilavo series rocks at mid-crustal levels (see text for further details). References: 1, Kelemen
et al. (2004); 2, this study. Because Kelemen et al. (2004) provided an average composition from n ¼ 427 analyses without
standard deviations we have allowed a variability of 0·5 wt % for CaO and Na2O and of 75 ppm for Sr in our model (for
which the average continental arc composition would be too high in CaO and Na2O and too low in Sr to reproduce
the primitive composition of sample E05065 for these elements). The allowed uncertainties are within the natural variability
of arc basalts.
Table 5: Compositions of mineral phases

Mineral Sample SiO2 TiO2 Al2O3 Cr2O3 FeOtot MnO MgO CaO Na2O NiO K2O Total Ref.
Model 1
cpx Run 10 49·35 0·34 7·49 0·03 5·65 0·16 15·57 21·24 0·32 0·01 100·16 1
olivine E05024-8 39·59 0·03 0·05 11·11 0·24 48·64 0·08 0·01 0·27 100·01 2
spinel Run 10 0 0·3 42·13 13·85 24·83 0 14·08 0·21 0·03 0·13 0 95·56 1
Magnetite E05151_1_2 0·07 5·58 3·49 0·02 77·79 0·50 7·79 0·03 95·27 3
Model 2
cpx E156a_6-2 49·71 0·590 4·690 0·050 9·75 0·430 13·38 20·89 0·460 0·010 0·020 99·98 3
amphibole E05067_2_1 41·66 1·74 12·70 0·01 13·09 0·12 13·44 11·27 2·00 0·02 1·26 97·31 3
magnetite E05151_1_2 0·07 5·58 3·49 0·02 77·79 0·50 7·79 0·03 95·27 3
plagioclase E05067 47·09 0·01 33·15 0·64 0·05 16·43 1·78 0·09 100·01 3
References: 1, Kägi (2001), mineral compositions are from experimental runs on calc-alkaline basalt at 1 GPa and
1170–12008C; 2, unpublished data (M. Chiaradia) from olivine of a basaltic andesite of Chacana (Eastern Cordillera of
Ecuador); 3, this study.
Table 6: Proportions of fractionating minerals in Model 1 variable proportions of clinopyroxene, amphibole^magnet-

ite plagioclase could explain the variations in Fe2O3,
CaO, Na2O (Figs 9cê), and possibly Cr and Ni with
Olivine Clinopyroxene Magnesiochromite Magnetite
MgO (Figs i^j), but clearly cannot explain the strong
enrichments in K2O, Th, Sr (Figs 9f, g^h) and all other in-
wt % 33 63·6 1·4 2
compatible elements (not shown). This is consistent with
the petrography, mineral chemistry and correlations of
Sr and Nd isotopes with incompatible elements, which
fractionation of a clinopyroxeneâmphibole^magnet- indicate that the Pilavo magmas have evolved in an open
ite plagioclase assemblage, consistent with the minimal crustal magma storage system through either combined
isotopic variations observed at this volcano (Chiaradia assimilation^fractional crystallization (AFC) or recharge
et al., 2009). At Pichincha, fractionation processes cannot accompanied by assimilation^fractional crystallization
entirely explain the Th and especially the K2O enrich- (RAFC).
ment: assimilation of crustal material is required, in A simple AFC process is incompatible with petrographic
agreement with the correlations observed between radio- evidence of magma mixing such as the occurrence of min-
genic isotopes and fractionation indices at this volcano gling textures, of texturally different phenocrysts within
(Chiaradia et al., 2009). At Pilavo fractionation of the same sample, and of magmatic quartz xenocrysts
1131
(see above). A dynamic process of RAFC (e.g. Bohrson & have produced the typical zoned corona textures with
Spera, 2003) with continuous homogenization of the evol- glass around the quartz xenocryst and clinopyroxene on
ving magmas, their partial solidification to form uncon- the outer part through reaction with the recharging mafic
solidated crystal mushes or proto-plutons and their magma. It is unlikely that they represent xenocrysts
assimilation by recharges of more mafic magma is more derived from mature continental crust because the
appropriate for the petrogenesis of the Pilavo lavas, as Western Cordillera of Ecuador is floored by mafic oceanic
indicated by the systematically changing petrographic, plateau crust (e.g. Feininger & Seguin, 1983; Vallejo et al.,
geochemical and isotopic compositions of whole-rocks and 2006, 2009; Chiaradia, 2009) and assimilation of a
minerals. quartz-rich mature continental crust should result in a sig-
Petrographic evidence for magma mixing (or mush or nificant shift of radiogenic isotope compositions, which
proto-pluton incorporation) in the Pilavo lavas is provided are instead primitive for all rocks (see above).
by the coexistence of texturally and compositionally differ- Amphibole and clinopyroxene phenocrysts in the Pilavo
ent pheno- or xenocrysts in the same thin section: (1) lavas display systematic correlations in their incompatible
unzoned to normally zoned Mg-rich amphibole pheno- element contents with whole-rock chemistry (Figs 4^6).
crysts and inversely multi-zoned large amphibole; (2) The intra-sample variability (e.g. standard deviation) of
amphiboles with or without equant clinopyroxene coronas, these elements is similar for all samples and somewhat
perhaps indicative of heating (Rutherford & Devine, less than the inter-sample variability. Because each sample
2003); (3) clinopyroxene associated with amphibole in investigated represents a different lava flow (Fig. 2b), this
clots (perhaps residues of magma mushes), as isolated suggests that amphiboles and clinopyroxenes have crystal-

large phenocrysts with sector zoning (probably crystallized lized from differently evolved and relatively (though not
rapidly near the surface; Brophy et al., 1999) or surrounded completely) homogenized magmas, before being erupted.
by amphibole coronas (relic clinopyroxenes); (4) plagio- The identical Sr and Nd isotopic compositions of a large
clases with heavily sieved, slightly sieved, and non-sieved zoned amphibole phenocryst and the surrounding ground-
cores. Additional evidence of magma mixing or mush mass (sample E05156) support such a conclusion.
incorporation comes from plagioclase compositions. Whole-rock incompatible element contents (e.g. Sr)
Variably sieved to non-sieved cores of plagioclase pheno- correlate with modal mineralogy (Fig. 13a, c and e). In par-
crysts display high anorthite contents (An80^92) indicating ticular, they increase with increasing modal amphibole
crystallization from H2O-rich magmas and/or from melts (Fig. 13c) and (in a semi-quantitative way; see above) with
with high CaO/Na2O ratios and Al2O3 contents (e.g. the proportion of quartz xenocrysts. Because in the more
Sisson & Grove, 1993; Panjasawatwong et al., 1995). Feig evolved Pilavo rocks a portion of the amphibole population
et al. (2006) have shown experimentally that plagioclase consists of multiply zoned amphibole crystals (see above),
with anorthite contents between 80 and 85 mol % can which reflect growth episodes under different magmatic
crystallize only for water contents of about 5 wt % at conditions and thus probably represent xenocrysts, and
P ¼ 0·2 GPa. Anorthite-rich cores of plagioclase pheno- because quartz crystals are magmatic xenocrysts, we
crysts in the Pilavo rocks have both low and high Sr con- suggest that the incompatible element enrichment was con-
tents: the least evolved rocks contain only plagioclase with comitant with a steadily growing incorporation of evolving
Sr-poor cores, whereas progressively more evolved rocks magmas, crystal mushes or proto-plutons.
contain plagioclase with both Sr-poor and increasingly In agreement with the above observations, geochemical
Sr-rich cores (compare analyses pl67_3-1, pl67_3-8 and modeling (Fig. 9k) shows that the enrichment of incompat-
pl67_7-2 in Electronic Appendix 3). Because Sr contents in ible elements of Pilavo rocks can be successfully repro-
plagioclase decrease with increasing anorthite content duced by a process in which the same primitive magma
(Blundy & Wood, 1991) and increase with Sr concentration (most primitive rock of the Pilavo series) recharges, frac-
in the melt, the occurrence of two types of similarly tionates, and assimilates a few times increasingly evolved
anorthite-rich plagioclase with different Sr contents in the magmas, mushes or proto-plutons (represented by samples
more evolved rocks indicates mixing of these plagioclases of the Pilavo series). The modeling implies a high r value
during or after their crystallization from melts with differ- of 0·9 (r ¼ mass of assimilant/mass of crystallized
ent Sr contents. magma), which is typical for a thermally mature assimi-
Evidence for recharge of felsic magmas, crystal mushes lant, such as crystal mush, unconsolidated proto-pluton or
or unconsolidated proto-plutons by more mafic magmas crystal-rich magma in the case of Pilavo. For each RAFC
comes from the presence of quartz xenocrysts only in the step the resulting amounts of fractionation and
more evolved rocks. Also, following previous interpret- assimilation are low (5^12% fractionation and 4·5^11%
ations (e.g. Sato, 1975; Stimac & Pearce, 1992; Mashima, assimilation; Fig. 9k, Table 7) and thus consistent with the
2004), these are most probably phenocrysts of a felsic small variability of major elements for a concomitant
magma (or mush or unconsolidated proto-pluton), which strong enrichment in incompatible elements.
1132
1400 0.70395
(a) (b)
1200 0.70390
87Sr/86Sr
Sr (ppm)
1000 0.70385
800 0.70380
600 0.70375
10 30 50 70 90 10 30 50 70 90
plagioclase (modal %) plagioclase (modal %)
1400 0.70395
(c) (d)
1200 0.70390
87Sr/86Sr
Sr (ppm)
1000 0.70385

800 0.70380
600 0.70375
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70
groundmass (modal %) groundmass (modal %)
1400 0.70395
(e) (f)
1200 0.70390
Sr (ppm)
87Sr/86Sr
1000 0.70385
800 0.70380
E05151
600 0.70375
0 4 8 12 16 20 0 4 8 12 16 20
amphibole (modal %) amphibole (modal %)
25 0.70395
(g) (h)
amphibole (modal %)
20
0.70390
87Sr/86Sr
15
0.70385
10
0.70380
5
0 0.70375
0 40 80 120 160 200 0 40 80 120 160 200
average reaction rim thickness (µm) average reaction rim thickness (µm)
Fig. 13. (a^f) Correlations between modal mineralogy and geochemical and isotopic composition of the whole-rocks. A noteworthy feature is
the high amphibole modal content (11%) in sample E05151, which is ‘anomalous’ for being characterized by a low Sr content (700 ppm) but a
radiogenic Sr isotope signature (0·703859). (g, h) Correlations between the average thickness of reaction rims around amphibole and corres-
ponding Sr isotopic composition and amphibole modal abundance of the whole-rocks.
1133
Table 7: Parameters used for modelling repeated recharge and assimilation^fractional crystallization processes at Pilavo
(refer to Fig. 9k)
Step DBa DLa Tick marks of Fractionation Assimilation Scheme of recharge

F in Fig. 9 (wt %) (wt %)
1 0·3 0 0·05 5 4·5 E05065 ! E05065 to yield E05148

2 0·3 0 0·05 5 3·6 E05065 ! E05148 to yield E05064
3 0·35 0 0·035 7 6·3 E05065 ! E05064 to yield E05154
4 0·35 0 0·04 8 7·2 E05065 ! E05154 to yield E05149
5 0·45 0 0·04 12 10·8 E05065 ! E05149 to yield E05152
Compositions of samples used for modelling are reported in Table 2.
Crustal assimilation and melting (see Fig. 10 and Chiaradia et al., 2009) and thus consistent
Petrographic and geochemical evidence that plagioclase with the small isotopic shifts observed within the Pilavo

crystallization postdates that of amphibole in the Pilavo series. Partial melting of mafic mid- to lower crustal
basaltic andesites suggests magmatic evolution at relatively amphibole-bearing lithologies (e.g. Davidson et al., 2007)
high pressures (0·4 GPa). Experimental studies have can produce K-rich granitic melts (Sisson et al., 2005) and
shown that in H2O-rich oxidized basaltic systems amphi- such melts could also be enriched in incompatible elem-
bole crystallization precedes plagioclase at pressures ents. Experimental work by Storkey et al. (2005) has
0·4 GPa (e.g. Grove et al., 2003; Feig et al., 2006). shown that incompatible element concentrations in partial
Petrographic evidence of late plagioclase crystallization melts of lower crustal metabasites are several hundred to
derives from the fact that the great majority of the plagio- several thousand times higher than primitive mantle con-
clase occurs as microcrysts and microlites in the ground- centrations. Partial melting of tonalite at mid-crustal
mass. Late crystallization of plagioclase (largely levels also produces high-K dacitic melts that are likely to
post-amphibole and post-clinopyroxene) is also supported be enriched in incompatible elements (e.g. Singh &
by steadily increasing Sr contents (and Sr/Y ratios) with Johannes, 1996; Reubi & Blundy, 2008). In concomitance
magma evolution not only in whole-rocks but also in with the RAFC processes described above, incorporation
amphibole and clinopyroxene phenocrysts of all lava sam- of these enriched partial melts into crystal mushes and
ples (Figs 4c and 5 b). proto-plutons recycled by recharging magmas might be
Assimilation and eventually melting of the crust at such an additional process to explain the strong and steady en-
mid-crustal levels is favored by a high rate of mafic richment of incompatible elements in concomitance with
magma supply and by magmas residing at these levels for subtle radiogenic isotope changes in the Pilavo magmas
enough time to cause thermal maturation of the host (Fig. 11).
rocks (e.g. Petford & Gallagher, 2001; Dufek & Bergantz,
2005; Annen et al., 2006; Lee et al., 2006). At Pilavo assimi- Mixing in the conduit during magma
lation of crustal material by magmas that were undergoing ascent
the RAFC processes described in the previous section is Analysis of amphibole reaction rim thickness provides evi-
supported by the systematic correlations of Sr and Nd iso- dence that further mixing occurred, at least in the more
topes with magma differentiation indices (Fig. 11) and min- primitive rocks, in magmatic conduits at pressures below
eral modal abundances (Fig. 13b, d and f). The basement those of amphibole stability. Hydrous silicate minerals
in this part of the Western Cordillera of Ecuador is repre- that have crystallized from an H2O-rich magma at depth
sented by a Cretaceous oceanic plateau, which at mid- to will become unstable as the solubility of H2O in the de-
lower crustal levels may have been modified by arc compressing silicate liquids decreases (e.g. Shaw, 1974).
magmas continuously moving through the crust since Destabilization through such a process results in the forma-
the Late CretaceousÊarly Tertiary (Chiaradia, 2009; tion of reaction rims around amphibole crystals in contact
Chiaradia et al., 2009). Such long-lived refinement pro- with melt, consisting of an assemblage of pyroxene, plagio-
cesses of the Cretaceous oceanic plateau by arc-related clase and Fe^Ti oxides (e.g. Rutherford & Hill, 1993). The
magmas might have created a hybrid crust with an isotopic onset of amphibole destabilization depends on the magma
composition close to that of the Quaternary arc magmas composition and temperature but typically occurs at
1134
pressures below 150^100 MPa (e.g. Rutherford, 2008). The remain high and vary between NNO þ 0·5 and about
thickness of the reaction rims developed around amphibole NNO þ 2.
phenocrysts, as well as the size and shape of the replacing The high fO2 of the Pilavo lavas might indicate oxidiz-
mineral phases (plagioclase, pyroxene, Fe-oxides), depends ing conditions in the mantle source, although further oxi-
on the magma ascent rate (e.g. Rutherford & Hill, 1993; dation may have occurred within the crust. In fact, in
Browne & Gardner, 2005). Thicker reaction rims indicate H2O-rich magmas that are crystallizing amphibole, like
longer exposure times (slower ascent rates) at disequilib- those of Pilavo, fO2 increases owing to liberation of free
rium between amphibole and the surrounding melt and O2 during amphibole crystallization (Frost & Lindsley,
are usually accompanied by equant and larger sized repla- 1991). The inverse MgO zoning in the Pilavo amphiboles
cing mineral phases (Browne & Gardner, 2005). In con- and associated opacitization of the rims at each subsequent
trast, thinner rims with a smaller grain size of the growth zone would be consistent with increasing fO2
replacing minerals and acicular shapes are typical of during magmatic evolution (e.g. Mason, 1978).
faster and continuous ascent rates (Browne & Gardner,
2005). Reaction rims around amphibole phenocrysts there- General model
fore provide qualitative information about the path of the Mantle-derived basaltic magmas initially underwent
magma between the depth of initial amphibole breakdown high-pressure fractionation of dominant olivine and clino-
and the effusion at the vent. pyroxene with lesser amounts of Cr-rich spinel and mag-
We note that the average thickness of reaction rims on netite at the mantle^crust interface and/or at lower
the amphibole population of each sample correlates broad- crustal levels (Fig. 14). This fractionating assemblage drove
derivative liquids into the tholeiitic field (Fig. 9b) and to-

ly with amphibole geochemical indices (Fig. 5e and f) as
well as with the modal amount of amphibole, whole-rock wards high H2O and Al2O3 contents, and low Ni and Cr
geochemical indices, and whole-rock 87Sr/86Sr values (Fig. contents. These hydrous, high-Al2O3 basalt to basaltic an-
13g and h). Less evolved rocks are thus characterized by desite melts migrated upwards, where they evolved at
less abundant amphibole with low incompatible element mid-crustal levels through dominant amphibole and clino-
contents (Fig. 13e and f) and variably thick reaction rims pyroxene plagioclase fractionation under high pH2O.
(Fig. 13h), whereas more evolved rocks contain more abun- Although no age determinations are available for the
dant amphibole enriched in incompatible elements and various lava flows investigated here, volcanic stratigraphy
with consistently thinner rims. This suggests that the less suggests that the more evolved and crustally contaminated
evolved magmas ascended slowly and mixed with each lava flows belong to the later (main) phase of build-up of
other in the conduit at a pH2O below that of amphibole the volcanic edifice (Fig. 2b). From this point of view, the
stability to account for the coexistence of amphiboles with correlations between modal abundance of amphibole,
variably thick reaction rims in the same thin section, mineral chemistry, radiogenic isotopes and geochemical
whereas more evolved magmas have ascended more rapid- indices of whole-rock differentiation point to the following
ly from the magma storage chamber with no significant model (Fig. 14). Initially (Stage a in Fig. 14) the
mixing in the conduit. This is qualitatively in line with high-Al2O3 basalt to basaltic andesite magmas ascended
the negative correlation of modal plagioclase with relatively rapidly, crystallizing little amphibole and
whole-rock geochemical differentiation (Fig. 13a and b). allowing for abundant decompression crystallization of
plagioclase and clinopyroxene only at shallow crustal
Because the vast majority of plagioclase in the Pilavo
levels (mostly outside the amphibole stability field) when
rocks consists of microcrysts (a likely result of decompres-
magmas started to exsolve water and consequently to slow
sion crystallization during slow ascent), the above correl-
down until stalling in the conduit or in shallow magma
ation suggests that the highest amount of decompression
chambers. At this stage a significant mid-crustal reservoir
crystallization occurred in the slowly ascending, least
was not yet established, as indicated by the paucity of
evolved magmas.
amphibole phenocrysts, the lack of zoned amphiboles and
the homogeneously low Sr contents of plagioclase in these
Oxygen fugacity of the magmas magmas. Mixing probably occurred in the conduit at
Oxygen fugacities estimated from magnetite^-ilmenite depths shallower than those of amphibole stability between
pairs in the Pilavo lavas are high (from NNO þ1·5 to magmas stalled owing to the onset of abundant decompres-
NNO þ3·3). However, Scaillet & Evans (1999), among sion crystallization and batches of magma rising from
others, have highlighted that fO2 values 4NNO þ1·5 are depth. This scenario is supported by the coexistence of
outside the calibrated range of Fe^Ti-oxide barometry amphiboles with variably thick reaction rims only in the
and, by measuring experimentally fO2 for Pinatubo, they early stage, more primitive rocks of Pilavo.
determined an fO2 that is about 1 log unit below that es- The subsequent establishment of one or several
tablished by Fe^Ti-oxide barometry. Assuming a similar mid-crustal magma reservoirs (Stages b and c in Fig. 14),
correction for the Pilavo lavas, fO2 values none the less where magmas evolved through amphibole and
1135
increasing ascent rate (=decreasing amphibole reaction rim thickness)

increasing modal amphibole and decreasing crystallinity
increasing incompatible elements in minerals and rocks
increasing crustal contamination
0
approximate
depth of
amphibole
destabilization magmatic chambers
10
cpx and amph

(±plag) fractionation,
20 recharge, partial melting of
modified mafic crust and
assimilation of magmas/
crystal mushes/proto-plutons
in mid-crustal magmatic
chambers
30
high-alumina basalts
40 km
with tholeiitic affinity
produced by
olivine-cpx-spinel
fractionation
Stage a Stage b Stage c
Fig. 14. Conceptual model of the magmatic plumbing system at Pilavo volcano, inspired by the deep hot zone model of Annen et al. (2006) (for
discussion see text). Lighter colors indicate progressively enriched magmas. Stages a, b, c do not necessarily correspond to systematically
younger stages of volcano evolution. At Pilavo, the volcanic stratigraphy seems to suggest that Stage b postdates Stage a, but there is no con-
straint on the temporal relationships between Stage b and Stage c. The model may also be valid for Stage a postdating Stage b as a result of a
changing plumbing system in which magmas short-circuit (Stage a) or not (Stages b and c) the mid-crustal chamber(s).
clinopyroxene fractionation under high pH2O conditions in progressively evolving magmas, as indicated by their
(P 0·4 GPa), accompanied by assimilation of felsic correlations with radiogenic isotopes (Fig. 11) and amphi-
magmas, crystal mushes or proto-plutons as well as bole abundance (Fig. 13), and as supported by our model-
mixing with incoming batches of mafic magma (RAFC ing (Fig. 9k). Magmas were driven towards the
processes), is marked by the appearance of multiply zoned calc-alkaline field as a result of such processes (Fig. 9b).
amphibole in lavas of the main volcanic stage, increasingly Correlations of average amphibole reaction rim thickness
Sr-rich plagioclase (coexisting with Sr-poor plagioclase) with amphibole mineral chemistry, whole-rock geochem-
and magmatic quartz xenocrysts. During these RAFC istry and isotopic compositions suggest that increasingly
processes incompatible elements became strongly enriched enriched magmas were amalgamated in the mid-crustal
1136
reservoir and ascended to the surface with steadily faster fractures affecting the extent of partial melting of the
ascent rates above the depths of amphibole stability, prob- mantle wedge; Miller et al., 1992), or to changes in struc-
ably owing to their increasingly H2O-rich nature. The tural features of the overriding plate affecting its response
above model is also valid for the case in which Stages a, b to subduction-induced stress (e.g. Singer & Myers, 1992).
and c of Fig. 14 represent plumbing systems at different,
but not sequential, times, in which magmas short-circuit Implications for continental crust
(Stage a) or not (Stages b and c) the mid-crustal cham- formation processes
ber(s) (see also Hora et al., 2009). Enriched high-MgO andesites (e.g. Kelemen et al., 2003)
and intrusive equivalents of Archean age (sanukitoids)
have geochemical features remarkably similar to those of
CONC LU DI NG R E M A R K S the average continental crust. Therefore, studying their
The main focus of this work has been to understand the petrogenesis helps to understand the processes responsible
formation processes of the enriched basaltic andesites of for the formation of the continental crust. In a
Pilavo. We have shown that Pilavo is distinct from adjacent subduction-related context the genesis of this type of rocks
frontal arc volcanoes in Ecuador in having a tholeiitic af- is usually explained by either partial melting of mantle
finity and a strong enrichment in incompatible elements. peridotite metasomatized by slab melts (e.g. Shirey &
Below we briefly discuss these features in a broader context Hanson, 1984; Martin et al., 2005; Tatsumi, 2006) or incom-
through comparisons of Pilavo with other tholeiitic and plete reaction of slab melts with mantle peridotite
andesitic arc volcanoes and with other types of enriched (Kelemen et al., 2003; Rapp et al., 2010). In both cases in-
andesites (e.g. Setouchi sanukites and sanukitoids).

compatible element enrichment is attributed to slab melt
interaction with mantle peridotite.
Tholeiitic vs calc-alkaline trends at Pilavo Enriched Pilavo basaltic andesites are geochemically
Figure 9b shows Pilavo lavas plotted in a FeOtot/MgO vs similar to Archean sanukitoids and, to a lesser extent, to
SiO2 diagram (Miyashiro, 1974) together with various arc Setouchi sanukites, except for the variably lower MgO
volcanoes displaying either tholeiitic or calc-alkaline affin- and significantly lower Ni and Cr contents resulting
ity. The Pilavo lavas plot almost entirely within the tholei- from high-pressure fractionation of spinel-bearing olivine-
itic field, although they do not define a typical tholeiitic clinopyroxenite. The processes of incompatible element en-
trend such as that of Okmok volcano. Instead, they define richment occurring at Pilavo may therefore be of relevance
a trend similar to those of calc-alkaline volcanoes for the debate on continental crust formation. Having
(Pichincha, Pululahua, Lassen Peak, Piip, Setouchi) with shown that incompatible element enrichment of Pilavo bas-
an almost invariant FeOtot/MgO for changing SiO2. altic andesites has resulted from intra-crustal RAFC pro-
However, Pilavo is clearly distinct from the calc-alkaline cesses within a juvenile mafic crust, we suggest that,
volcanoes because its ‘calc-alkaline trend’ occurs at higher although ultimately sourced in the slab component, incom-
FeOtot/MgO values (i.e. within the tholeiitic field). In patible elements in isotopically primitive andesites can be
detail, calc-alkaline volcanoes often define sub-horizontal enriched not only via slab melt^mantle wedge interactions,
trends at various FeOtot/MgO values, with Parinacota but also at mid-crustal levels through refining processes
being the closest to Pilavo. We concur with Hora et al. within a juvenile mafic crust. The main requirement for
(2009) that, in an arc setting, such sub-horizontal trends this to occur, as suggested by the high r value of our model-
at variably high FeOtot/MgO ratios reflect magmas that ing, is probably a sufficient thermal maturation of the
have evolved at lower crustal levels with little further evo- lower to mid-crust achieved by a focused and sustained
lution at shallower levels for the tholeiitic end of the trend transit of magmas through the same crustal volume. The
and magmas that underwent shallower level contamin- transit of arc magmas through the juvenile oceanic plateau
ation and fractionation processes for the calc-alkaline end crust accreted to the continent in the Late Cretaceous
of the trend. If so, then the occurrence of sub-horizontal seems thus a viable process for the formation of isotopically
trends at variable FeO/MgO values would reflect different primitive but geochemically enriched, continental
degrees of clinopyroxene- and olivine-dominated fraction- crust-type rocks in the Western Cordillera of Ecuador (see
ation at deep levels, with Pilavo being an extreme case of also Chiaradia, 2009).
olivine-clinopyroxenite fractionation.
Contrasting tholeiitic and calc-alkaline evolutionary
trends in adjacent volcanoes such as Pilavo, Pichincha and AC K N O W L E D G E M E N T S
Pululahua, are not uncommon in magmatic arcs (e.g. Thoughtful reviews by Andy Saunders and two anonymous
Singer & Myers, 1990). This has been attributed either to reviewers helped to improve the quality of this work.
changes in the magmatic plumbing system of the volcano Editorial handling by John Gamble and Marjorie Wilson
in time or space (e.g. Hora et al., 2009), to spatial changes is greatly appreciated. We thank A. Ulianov for precious
in the characteristics of the subducted oceanic crust (e.g. technical support during LA-ICP-MS analyses, C.
1137
Ginibre, K. Kouzmanov and J. Allibon for help with the volcanic rocks from the Andean Northern Volcanic Zone, Ecuador.
microprobe analyses, and F. Capponi for whole-rock XRF Journal of Petrology 47, 1147^1175.
analyses. Chiaradia, M. (2009). Adakite-like magmas from fractional crystal-
lization and meltingâssimilation of mafic lower crust (Eocene
Macuchi arc, Western Cordillera, Ecuador). Chemical Geology 265,
468^487.
F U N DI NG Chiaradia, M. & Fontbote¤, L. (2001). Radiogenic lead signatures in
This study was funded by the Swiss National Science Au-rich VHMS ores and associated volcanic rocks of the Early
Foundation (projects N. 200021-109636 and 200020-117616). Tertiary Macuchi island arc (Western Cordillera of Ecuador).
Economic Geology 96, 1361^1378.
Chiaradia, M., Mu«ntener, O., Beate, B. & Fontignie, D. (2009).
Adakite-like volcanism of Ecuador: lower crust magmatic evolu-
S U P P L E M E N TA RY DATA tion and recycling. Contributions to Mineralogy and Petrology 158,
Supplementary data for this paper are available at Journal 563^588.
of Petrology online. Clynne, M. A., Muffler, L. J. P., Siems, D. F., Taggart, J. E., Jr &
Bruggman, P. (2008). Major and EDXRF trace element chemical
analyses of volcanic rocks from Lassen Volcanic National Park
and vicinity, California. US Geological Survey Open-File Report
R EF ER ENC ES 2008^1091, 1^10.
Aguilera, E. (1998). The Chalupas and Chachimbiro geothermal fields Corredor, F. (2003). Seismic strain rates and distributed continental
in Ecuador. Geothermal Resources Council Transactions 22, 247^251. deformation in the northern Andes and three-dimensional seismo-
Amo¤rtegui, A., Lapierre, H., Jaillard, E., Martelat, J.-E., Bosch, D., tectonics of the northwestern South America. Tectonophysics 372,

Bussy, F., Demant, A. & Brunet, P. (2005). Accreted oceanic frag- 147^166.
ments below the Western Cordillera of Ecuador. 6th International Crawford, A. J., Falloon, T. J. & Eggins, S. (1987). The origin of island
Symposium on Andean Geodynamics (ISAG 2005, Barcelona), Extended arc high-alumina basalts. Contributions to Mineralogy and Petrology 97,
Abstracts 42^45. 417^430.
Annen, C., Blundy, J. D. & Sparks, R. S. J. (2006). The genesis of Davidson, J., Turner, S., Handley, H., Macpherson, C. & Dosseto, A.
calc-alkaline intermediate and silicic magmas in deep crustal hot (2007). Amphibole ‘sponge’ in arc crust? Geology 35, 787^790.
zones. Journal of Petrology 47, 505^539. Dickinson, W. R. (1975). Potash^depth (K^h) relations in continental
Bacon, C. R. & Hirschmann, M. M. (1988). Mg/Mn partitioning as a margin and intra-oceanic magmatic arcs. Geology 3, 53^56.
test for equilibrium between coexisting Fe^Ti oxides. American Dufek, J. & Bergantz, G. W. (2005). Lower crustal magma genesis
Mineralogist 73, 57^61. and preservation: a stochastic framework for the evaluation of
Barazangi, M. & Isacks, B. (1976). Spatial distribution of earthquakes basalt^crust interaction. Journal of Petrology 46, 2167^2195.
and subduction of the Nazca plate beneath South America. Geology Ego, F., Se¤brier, M., Lavenu, A., Yepes, H. & Egues, A. (1996).
4, 686^692. Quaternary state of stress in the Northern Andes and the restrain-
Barragan, R., Geist, D., Hall, M., Larsen, P. & Kurz, M. (1998). ing bend model for the Ecuadorian Andes. Tectonophysics 259,
Subduction controls on the compositions of lavas from the 101^116.
Ecuadorian Andes. Earth and Planetary Science Letters 154, 153^166. Feig, S. T., Koepke, J. & Snow, J. E. (2006). Effect of water on tholeiitic
BGS & CODIGEM. (1999). Mapa geologico de la Cordillera Occidental del basalt phase equilibria: an experimental study under oxidizing
Ecuador entre 08^18 N. Quito: Mision Britanica, CODIGEM. conditions. Contributions to Mineralogy and Petrology 152, 611^638.
Blundy, J. D. & Wood, B. J. (1991). Crystal-chemical controls on the Feininger, T. (1987). Allochthonous terranes in the Andes of Ecuador
partitioning of Sr and Ba between plagioclase feldspar, silicate and northwestern Peru. Canadian Journal of Earth Sciences 24,
melts, and hydrothermal solutions. Geochimica et Cosmochimica Acta 266^278.
55, 193^209. Feininger, T. & Seguin, M. K. (1983). Bouguer gravity anomaly field
Bohrson, W. A. & Spera, F. J. (2003). Energy-constrained open-system and inferred crustal structure of continental Ecuador. Geology 11,
magmatic processes IV: Geochemical, thermal and mass conse- 40^44.
quences of energy-constrained recharge, assimilation and fraction- Finney, B., Turner, S., Hawkesworth, C., Larsen, J., Nye, C.,
al crystallization (EC-RAFC). Geochemistry, Geophysics, Geosystems Rhiannon, G., Bindeman, I. & Eichelberger, J. (2008). Magmatic
4(2), 8002, doi:10.1029/2002GC000316. differentiation at an island-arc caldera: Okmok Volcano, Aleutian
Bourdon, E., Eissen, J. P., Gutscher, M. A., Monzier, M., Hall, M. L. Islands, Alaska. Journal of Petrology 49, 857^884.
& Cotten, J. (2003). Magmatic response to early aseismic ridge sub- Foden, J. D. & Green, D. H. (1992). Possible role of amphibole in the
duction: the Ecuadorian margin case (South America). Earth and origin of andesite: some experimental and natural evidence.
Planetary Science Letters 205, 123^138. Contributions to Mineralogy and Petrology 109, 479^493.
Brophy, J. G., Whittington, C. S. & Park, Y.-R. (1999). Sector-zoned Frost, B. R. & Lindsley, D. H. (1991). Occurrence of iron^titanium
augite megacrysts in Aleutian high alumina basalts: implications oxides in igneous rocks. In: Lindsley, D. H. (ed.) Oxide Minerals.
for the conditions of basalt crystallization and the generation of Mineralogical Society of America, Reviews in Mineralogy 25, 433^468.
calc-alkaline series magmas. Contributions to Mineralogy and Petrology Gill, J. B. (1981). Orogenic Andesites and Plate Tectonics. New York:
135, 277^290. Springer.
Browne, B. L. & Gardner, J. E. (2005). The influence of magma ascent Graindorge, D., Calahorrano, A., Charvis, P., Collot, J. Y. &
path on the texture, mineralogy, and formation of hornblende reac- Bethoux, N. (2004). Deep structure of the Ecuador convergent
tion rims. Earth and Planetary Science Letters 246, 161^176. margin and the Carnegie Ridge, possible consequences on great
Bryant, J. A., Yogodzinski, G. M., Hall, M. L., Lewicki, J. L. & earthquakes recurrence interval. Geophysical Research Letters 31,
Bailey, D. G. (2006). Geochemical constraints on the origin of doi:10.1029/2003GL018803.
1138
Grove, T. L., Elkins-Tanton, L. T., Parman, S. W., Chatterjee, N., Kessel, R., Schmidt, M. W., Ulmer, P. & Pettke, T. (2005). Trace elem-
Mu«ntener, O. & Gaetani, G. A. (2003). Fractional crystallization ent signature of subduction-zone fluids, melts and supercritical
and mantle-melting controls on calc-alkaline differentiation liquids at 120^180 km depth. Nature 437, 724^727.
trends. Contributions to Mineralogy and Petrology 145, 515^533. Kincaid, C. & Griffiths, R. W. (2004). Variability in flow and tem-
Guillier, B., Chatelain, J. L., Jaillard, E., Yepes, H., Poupinet, G. & peratures within mantle subduction zones. Geochemistry, Geophysics,
Fels, J. F. (2001). Seismological evidence on the geometry of the oro- Geosystems 5, Q06002, doi:10.1029/2003GC000666.
genic system in central^northern Ecuador (South America). Kushiro, I. (1969). The system forsterite^diopside^silica with and with-
Geophysical Research Letters 28, 3749^3752. out water at high pressures. American Journal of Science 267A,
Gutscher, M. A., Malavieille, J., Lallemand, S. & Collot, J. Y. (1999). 269^294.
Tectonic segmentation of the North Andean margin: impact of the Leake, B. E., Woolley, A. R., Arps, C. E. S. et al. (1997). Nomenclature
Carnegie Ridge collision. Earth and Planetary Science Letters 168, of amphiboles; report of the Subcommittee on Amphiboles of
255^270. the International Mineralogical Association Commission on
Harpp, K., Wanless, V. D., Otto, R. H., Hoernle, K. & Werner, R. New Minerals and Mineral Names. European Journal of Mineralogy
(2005). The Cocos and Carnegie aseismic ridges: a trace element 9, 623^651.
record of long-term plume^spreading center interaction. Journal of Le Bas, M. J., Le Maitre, R. W., Streckeisen, A. & Zanettin, B. (1986).
Petrology 46, 109^133. A chemical classification of volcanic rocks based on the total
Hermann, J., Spandler, C., Hack, A. & Korsakov, A. V. (2006). alkali^silica diagram. Journal of Petrology 27, 745^750.
Aqueous fluids and hydrous melts in high-pressure and ultra-high Lee, C.-T. A., Cheng, X. & Horodyskyj, U. (2006). The development
pressure rocks: Implications for element transfer in subduction and refinement of continental arcs by primary basaltic magma-
zones. Lithos 92, 399^417. tism, garnet pyroxenite accumulation, basaltic recharge and de-
Hirose, K. (1997). Melting experiments of lherzolite KLB-1 under hy- lamination: insights from the Sierra Nevada. Contributions to
drous conditions and generation of high-magnesian andesitic Mineralogy and Petrology 151, 222^242.

melts. Geology 25, 42^44. Lepage, L. (2003). ILMAT: an Excel worksheet for ilmenite^magnet-
Hora, J. M., Singer, B. S., Wo«rner, G., Beard, B. L., Jicha, B. R. & ite geothermometry and geobarometry. Computers and Geosciences
Johnson, C. M. (2009). Shallow and deep crustal control on differ- 29, 673^678.
entiation of calc-alkaline and tholeiitic magma. Earth and Planetary Lindsley, D. H. & Spencer, K. J. (1982). Fe^Ti oxide geothermometry:
Science Letters 285, 75^86. Reducing analyses of coexisting Ti-magnetite (Mt) and ilmenite
Jackson, S. (2008). LAMTRACE data reduction software for (Ilm). Abstracts, AGU 1982 Spring Meeting. EOS Transactions,
LA-ICP-MS. In: Sylvester, P. (ed.) Laser Ablation ICP-MS in the American Geophysical Union 63(18), 471.
Earth Sciences: current practices and outstanding issues. Litherland, M., Aspden, J. A. & Jemielita, R. A. (1994). The
Mineralogical Association of Canada Short Course. Mineralogical Metamorphic Belts of Ecuador. Keyworth: British Geological Survey.
Association of Canada Short Course Series 40, 305^307. Longerich, H. P., Jackson, S. E. & Gu«nther, D. (1996). Laser ablation^
Jackson, S. E., Longerich, H. P., Dunning, G. R. & Fryer, B. J. (1992). inductively coupled plasma^mass spectrometric transient signal
The application of laser ablation microprobeînductively coupled data acquisition and analyte concentration calculation. Journal of
plasma^mass spectrometry (LAMÎCP^MS) to in situ trace Analytical Atomic Spectrometry 11, 899^904.
element determinations in minerals. Canadian Mineralogist 30, Lonsdale, P. (1978). Ecuadorian subduction system. AAPG Bulletin
1049^1064. 62(12), 2454^2477.
Jaillard, E., Soler, P., Carlier, G. & Mourier, T. (1990). Geodynamic Mamberti, M., Lapierre, H., Bosch, D., Jaillard, E., Ethien, R.,
evolution of the northern and central Andes during early-to-middle Hernandez, J. & Polve¤, M. (2003). Accreted fragments of the Late
Mesozoic times: a Tethyan model. Journal of the Geological Society, Cretaceous Caribbean^Colombian Plateau in Ecuador. Lithos 66,
London 147, 1009^1022. 173^199.
Jaillard, E., Ben|¤ tez, S. & Mascle, G. H. (1997). Les de¤formations Martin, H., Smithies, R. H., Rapp, R., Moyen, J.-F. & Champion, D.
pale¤ogenes de la zone d’avant-arc sud-e¤quatorienne en relation (2005). An overview of adakite, tonalite^trondhjemite^granodiorite
avec l’e¤volution ge¤odynamique. Bulletin de la Socie¤ te¤ Ge¤ ologique de (TTG), and sanukitoid: relationships and some implications for
France 168, 403^412. crustal evolution. Lithos 79, 1^24.
Ka«gi, R. (2001). The liquid line of descent of hydrous, primary, Mashima, H. (2004). Time scale of magma mixing between basalt and
calc-alkaline magmas under elevated pressure. An experimental dacite estimated for the Saga^Futagoyama volcanic rocks in north-
approach, PhD thesis, ETH Zuu«rich, 115 pp. west Kyushu, southwest Japan. Journal of Volcanology and Geothermal
Kelemen, P. B., Yogodzinski, G. M. & Scholl, D. M. (2003). Research 131, 333^349.
Along-strike variation in the Aleutian Island Arc: Genesis of high Mason, D. R. (1978). Compositional variations in ferromagnesian
Mg# andesite and implications for continental crust. In: Eiler, J. minerals from porphyry copper-generating and barren intrusions
(ed.) Inside the Subduction Factory. Geophysical Monograph, American of the Western Highlands, Papua New Guinea. Economic Geology 73,
Geophysical Union 138, 223^276. 878^890.
Kelemen, P. B., Hanghoj, K. & Greene, A. R. (2004). One view of the Miller, D. M., Langmuir, C. H., Goldstein, S. L. & Franks, A. L.
geochemistry of subduction-related magmatic arcs, with an em- (1992). The importance of parental magma composition to
phasis on primitive andesite and lower crust. In: Holland, H. D. & calc-alkaline and tholeiitic evolution: evidence from Umnak
Turekian, K. K. (eds) Treatise on Geochemistry. Amsterdam: Elsevier, Island in the Aleutians. Journal of Geophysical Research 97(B1),
pp. pp, pp. 593^659. 321^343.
Kerr, A. C., Aspden, J. A., Tarney, J. & Pilatasig, L. F. (2002). The Miyashiro, A. (1974). Volcanic rock series in island arcs and active con-
nature and provenance of accreted oceanic blocks in western tinental margins. AmericanJournal of Science 274, 321^355.
Ecuador: geochemical and tectonic constraints. Journal of the Mu«ntener, O., Kelemen, P. B. & Grove, T. L. (2001). The role of H2O
Geological Society, London 159, 577^594. during crystallization of primitive arc magmas under uppermost
1139
mantle conditions and genesis of igneous pyroxenites: an experi- Scaillet, B. & Evans, B. W. (1999). The 15 June 1991 Eruption of
mental study. Contributions to Mineralogy and Petrology 141, 643^658. Mount Pinatubo. I. Phase equilibria and pre-eruption P^T^fO2^
Panjasawatwong, Y., Danyushevsky, L. V., Crawford, A. J. & fH2O conditions of the dacite magma. Journal of Petrology 40,
Harris, K. L. (1995). An experimental study of the effects of melt 381^411.
composition on plagioclase^melt equilibria at 5 and 10 kbar: Shaw, H. R. (1974). Diffusion of H2O in granitic liquids; Part 1,
implications for the origin of magmatic high-An plagioclase. Experimental data; Part 2, Mass transfer in magma chambers. In:
Contributions to Mineralogy and Petrology 118, 420^432. Hofmann, W., Giletti, B. J., Yoder Jr., H.S. & Yund, R.A. (eds)
Pennington, W. D. (1981). Subduction of the eastern Panama basin Geochemical Transport and Kinetics. Washington: Carnegie
and seismotectonics of northwestern South America. Journal of Institution of Washington, pp. 139^170.
Geophysical Research 86(B11), 10753^10770. Shimoda, G., Tatsumi, Y., Nohda, S., Ishizaka, K. & Jahn, B. M.
Petford, N. & Gallagher, K. (2001). Partial melting of mafic (amphi- (1998). Setouchi high-Mg andesites revisited: geochemical evidence
bolitic) lower crust by periodic influx of basaltic magma. Earth for melting of subducting sediments. Earth and Planetary Science
and Planetary Science Letters 193, 483^499. Letters 160, 479^492.
Pin, C., Briot, D., Bassin, C. & Poitrasson, F. (1994). Concomitant Shirey, S. B. & Hanson, G. N. (1984). Mantle-derived Archaean
separation of strontium and samarium^neodymium for isotopic monzodiorites and trachyandesites. Nature 310, 222^224.
analysis in silicate samples, based on specific extraction chromatog- Singer, B. S. & Myers, J. D. (1990). Intra-arc extension and magmatic
raphy. Analytica Chimica Acta 298, 209^217. evolution in the central Aleutian arc, Alaska. Geology 18, 1050^1053.
Plank, T. (2005). Constraints from thorium/lanthanum on sediment Singh, J. & Johannes, W. (1996). Dehydration melting of tonalites. 2.
recycling at subduction zones and the evolution of the continents. Composition of melts and solids. Contributions to Mineralogy and
Journal of Petrology 46, 921^944. Petrology 125, 26^44.
Plank, T. & Langmuir, C. H. (1998). The chemical composition of sub- Sisson, T. W. & Grove, T. L. (1993). Experimental investigations of
ducting sediment and its consequences for the crust and mantle. the role of H2O in calc-alkaline differentiation and subduction

Chemical Geology 145, 325^394. zone magmatism. Contributions to Mineralogy and Petrology 113,
Plank, T., Cooper, L. B. & Manning, C. E. (2009). Emerging 143^166.
geothermometers for estimating slab surface temperatures. Nature Sisson, T. W., Ratajeski, K., Hankins, W. B. & Glazner, A. F. (2005).
Geoscience 2, 611^615. Voluminous granitic magmas from common basaltic sources.
Pre¤vot, R., Chatelain, J.-L., Guillier, B. & Yepes, H. (1996). Contributions to Mineralogy and Petrology 148, 635^661.
Tomographie des Andes Equatoriennes: e¤vidence d’une continuite¤ Spikings, R. A., Winkler, W., Hughes, R. A. & Handler, R. (2005).
des Andes Centrales. Comptes Rendus de l’Acade¤ mie des Sciences, Se¤ rie II Thermochronology of allochthonous terranes in Ecuador:
323, 833^840. Unravelling the accretionary and post-accretionary history of the
Rapp, R. P., Shimizu, N., Norman, M. D. & Applegate, G. S. (1999). Northern Andes. Tectonophysics 399, 195^220.
Reaction between slab-derived melts and peridotite in the mantle Stimac, J. A. & Pearce, T. H. (1992). Textural evidence of mafic^felsic
wedge: experimental constraints at 3·8 GPa. Chemical Geology 160, magma interaction in dacite lavas, Clear Lake. American
335^356. Mineralogist 77, 795^809.
Rapp, R. P., Norman, M. D., Laporte, D., Yaxley, G. M., Martin, H. Storkey, A. C., Hermann, J., Hand, M. & Buick, I. S. (2005). Using in
& Foley, S. F. (2010). Continent formation in the Archean and situ trace-element determinations to monitor partial-melting pro-
chemical evolution of the cratonic lithosphere: melt^rock reaction cesses in metabasites. Journal of Petrology 46, 1283^1308.
experiments at 3^4 GPa and petrogenesis of Archean Mg-diorites Sun, S. S. & McDonough, W. F. (1989). Chemical and isotopic system-
(sanukitoids). Journal of Petrology 51, 1237^1266. atics of oceanic basalts: implications for mantle composition and
Reubi, O. & Blundy, J. (2008). Assimilation of plutonic roots, forma- processes. In: Saunders, A. D. & Norry, M. J. (eds) Magmatism in
tion of high-K ‘exotic’ melt inclusions and genesis of andesitic the Ocean Basins. Geological Society, London, Special Publications 42,
magmas at Volca¤n de Colima, Mexico. Journal of Petrology 49, 313^346.
2221^2243. Tanaka, T., Togashi, S., Kamioka, H., Amakawa, H., Kagami, H.,
Richards, J. P. (2003). Tectono-magmatic precursors for porphyry Hamamoto, T., Yuhara, M., Orihashi, Y., Yoneda, S., Shimizu, H.,
Cu^(MoÂu) deposit formation. Economic Geology 98, 1515^1533. Kunimaru, T., Takahashi, K., Yanagi, Y., Nakano, T., Fujimaki, H.,
Rudnick, R. L. & Gao, S. (2004). Composition of the continental Shinjo, R., Asahara, Y., Tanimizu, M. & Dragusanu, C. (2000).
crust. In: Holland, H. D. & Turekian, K. K. (eds) Treatise on JNdi-1: a neodymium isotopic reference in consistency with La
Geochemistry. Amsterdam: Elsevier, pp. 1^64. Jolla neodymium. Chemical Geology 168, 279^281.
Rutherford, M. J. (2008). Magma ascent rates. In: Putirka, K. D. & Tatsumi, Y. (1989). Migration of fluid phases and genesis of basalt
Tepley, F. J., III (eds) Minerals, Inclusions and Volcanic Processes. magmas in subduction zones. Journal of Geophysical Research 94(B4),
Mineralogical Society of America and Geochemical Society Reviews, in 4697^4707.
Mineralogy and Geochemistry 69, 241^271. Tatsumi, Y. (2006). High-Mg andesites in the Setouchi volcanic belt,
Rutherford, M. J. & Devine, J. D. (2003). Magmatic conditions and southwestern Japan: analogy to Archean magmatism and contin-
magma ascent as indicated by hornblende phase equilibria and ental crust formation? Annual Review of Earth and Planetary Sciences
reactions in the 1995^2002 Soufrie're Hills magma. Journal of 34, 467^499.
Petrology 44, 1433^1454. Tatsumi, Y. & Kogiso, T. (2003). The subduction factory: its role in the
Rutherford, M. J. & Hill, P. M. (1993). Magma ascent rates from evolution of the Earth’s crust and mantle. In: Larter, R. D. &
amphibole breakdown: An experimental study applied to the Leat, P. T. (eds) Intra-oceanic Subduction Systems. Geological Society,
1980^1986 Mount St Helens eruptions. Journal of Geophysical London, Special Publications 219, 55^80.
Research 98(B11), 19667^19685. Tatsumi, Y., Shukuno, H., Sato, K., Shibata, T. & Yoshikawa, M.
Sato, H. (1975). Diffusion coronas around quartz xenocrysts in andes- (2003). The petrology and geochemistry of high-magnesium andes-
ite and basalt from Tertiary volcanic region in Northeastern ites at the western tip of the Setouchi Volcanic Belt, SW Japan.
Shikoku, Japan. Contributions to Mineralogy and Petrology 50, 49^64. Journal of Petrology 44, 1561^1578.
1140
Taylor, S. R., Capp, A. C., Graham, A. L. & Blake, D. H. (1969). Trace van Keken, P. E., Kiefer, B. & Peacock, S. M. (2002). High-resolution
element abundances in andesites. Contributions to Mineralogy and models of subduction zones: implications for mineral dehydration
Petrology 23, 1^26. reactions and the transport of water into the deep mantle. In:
Thornber, C. R., Pallister, J. S., Lowers, H. A., Rowe, M. C., Kay, S.M., Ramos, V.A. & Dickinson, W.R. (eds) Backbone of the
Mandeville, C. W. & Meeker, G. P. (2008). Chemistry, mineralogy, Americas: Shallow Subduction, Plateau Uplift, and Ridge and
and petrology of amphibole in Mount St. Helens 2004^2006 Terrane Collision. Geological Society of America Memoir 204, 197^216.
dacite. In: Sherrod, D. R., Scott, W. E. & Stauffer, P. H. (eds) A White, W. M., McBirney, A. R. & Duncan, R. A. (1993). Petrology
Volcano Rekindled; the Renewed Eruption of Mount St. Helens, 2004^ and geochemistry of the Galapagos Islands: portrait of a
2006. US Geological Survey Professional Papers 1750, 727^753. pathological mantle plume. Journal of Geophysical Research 98,
Todt, W., Cliff, R. A., Hanser, A. & Hofmann, A. W. (1996). 19533^19563.
Evaluation of a 202Pb^205Pb double spike for high-precision lead Wood, B. J. & Turner, S. P. (2009). Origin of primitive high-Mg andes-
isotope analysis. In: Basu, A. & Hart, S. (eds) Earth Processes: ite: Constraints from natural examples and experiments. Earth and
Reading the Isotopic Code. Geophysical Monograph, American Geophysical Planetary Science Letters 283, 59^66.
Union 95, 429^437. Yogodzinski, G. M., Volynets, M. P., Koloskov, A. V., Seliverstov, N. I.
Vallejo, C., Spikings, R. A., Luzieux, L., Winkler, W., Chew, D. & & Matvenkov, V. V. (1994). Magnesian andesites and the subduction
Page, L. (2006). The early interaction between the Caribbean component in a strongly calc-alkaline series at Piip volcano, far
Plateau and the NW South American Plate. Terra Nova 18, 264^269. western Aleutians. Journal of Petrology 35, 163^204.
Vallejo, C., Winkler, W., Spikings, R. A., Luzieux, L., Heller, F. & Zindler, A. & Hart, S. (1986). Chemical geodynamics. Annual Review of
Bussy, F. (2009). Mode and timing of terrane accretion in the fore- Earth and Planetary Sciences 14, 493^571.
arc of the Andes in Ecuador. In: Geological Society of America,
Memoirs 204, 197^216.
1141

Chiaradia Et. Al., 2011

Uploaded by

Copyright:

Available Formats

Chiaradia Et. Al., 2011

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chiaradia Et. Al., 2011

Uploaded by

Copyright:

Available Formats

JOURNAL OF PETROLOGY VOLUME 52 NUMBER 6 PAGES 1107^1141 2011 doi:10.

Enriched Basaltic Andesites from Mid-crustal

MASSIMO CHIARADIA1, OTHMAR MU«NTENER2 AND

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

RECEIVED OCTOBER 19, 2010; ACCEPTED APRIL 19, 2011

The Author 2011. Published by Oxford University Press. All

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Water depth (m)

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

84°W 82°W 80°W 78°W

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Pre-Cretaceous metamorphic rocks Avalanche deposits

Cretaceous to Eocene undifferentiated Undifferentiated

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Sample: E05064 E05065 E05066 E05067 E05148 E05149 E05150 E05151

SiO2 57·57 56·07 56·07 56·42 56·04 57·37 54·17 56·54

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Sample: E05152 E05153 E05154 E05155 E05156 E05156 E05156

SiO2 53·74 56·94 56·79 56·59 57·47

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Sample: BHVO-1 SRM612

Rock: Basalt standard (average, n ¼ 8) 1s Glass standard (average, n ¼ 28) 1s

SiO2 49·60 0·08

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Symbols for Figure 4A Symbols for Figure 4B-D

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

70 E05151 E05064 E05154

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

element concentrations display a wide range of variability

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Western Cordillera BSE

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

0.5 (i.e. MgO-, Fe2O3-, CaO-richer) Pilavo rocks, which are

High Al2O3 (18 wt %) and Sr (700 ppm) contents in

semblage (continuous black lines in Fig. 9). Modeling can

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

0.7041 Pilavo most primitive rock

*Oxides in wt %, trace elements in ppm.

Table 5: Compositions of mineral phases

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Table 6: Proportions of fractionating minerals in Model 1 variable proportions of clinopyroxene, amphibole^magnet-

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Step DBa DLa Tick marks of Fractionation Assimilation Scheme of recharge

1 0·3 0 0·05 5 4·5 E05065 ! E05065 to yield E05148

Compositions of samples used for modelling are reported in Table 2.

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011

Downloaded from petrology.oxfordjournals.org at UniversitÃ© de GenÃ¨ve on July 11, 2011