3 Article

Geochemistry
G
Volume 3, Number 10
Geophysics 25 October 2002
1061, doi:10.1029/2002GC000355
Geosystems
AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES ISSN: 1525-2027
Published by AGU and the Geochemical Society
Correction published 11 February 2003

Volcanic evolution in the Galápagos:

The dissected shield of Volcan Ecuador
Dennis Geist
Department of Geological Sciences, University of Idaho-3022, Moscow, Idaho 83844, USA (dgeist@uidaho.edu)

William M. White
Department of Geological Sciences, Snee Hall, Cornell University, Ithaca, New York 14853, USA
(white@geology.geo.cornell.edu)

Francis Albarede
Laboratoire des Sciences de la Terre, Ecole Normale Supérieure de Lyon, 46 Allee d’Italie, 69364, Lyon Cedex 7, France
(albarede@ens-lyon.fr)

Karen Harpp
Department of Geology, Colgate University, 13 Oak Drive, Hamilton, New York 13346, USA
(kharpp@mail.colgate.edu)

Robert Reynolds
Central Oregon Community College, 2600 NW College Way, Bend, Oregon 97701, USA (breynolds@metolius.cocc.edu)

Janne Blichert-Toft
Laboratoire des Sciences de la Terre, Ecole Normale Supérieure de Lyon, 46 Allee d’Italie, 69364, Lyon Cedex 7, France
( jblicher@ens-lyon.fr)

Mark D. Kurz
Department of Chemistry, Woods Hole Oceanographic Institution, MS #25, Clark Laboratory, Woods Hole,
Massachusetts 02543, USA (mkurz@cygnus.whoi.edu)

[1] Volcan Ecuador, a young, active shield volcano at the northwest tip of Isabela Island, Galápagos,
experienced at least one sector collapse event that removed its western half. The compositions of lavas
exposed in the old caldera wall and fault scarps dissecting the outer shield together with young post-
collapse lavas provide insights into the history of this volcano. 40Ar/39Ar and cosmogenic 3He dating
reveal that sector collapse occurred at <100 Ka. Prior to sector collapse most magma erupted from
circumferential dikes that fed the summit carapace and other dikes that fed vents on the caldera floor.
Almost all of the erupted magmas cooled and underwent fractional crystallization in the lower crust or
upper mantle, where clinopyroxene crystallization dominated. Magmas then rose to a small (
0.1 km3),
steady state, subcaldera magma chamber where plagioclase, along with lesser amounts of olivine and
clinopyroxene, crystallized. Sector collapse perturbed the plumbing system in a way that now allows
some magmas, particularly those erupted on the East Rift, to bypass the magma chamber. Sector collapse
also opened new pathways to the surface that enabled magma to erupt at low elevation, effectively cutting
the magma supply to the summit. The paucity of summit eruptions over the past 100,000 years accounts
for the old, eroded appearance of the shield. Subtle variations in isotopic and trace element composition
occurred over the past 100,000 years, with the net change being toward slightly more depleted
compositions. These variations most likely reflect small-scale (<5 km) heterogeneity superimposed on
more regular regional patterns of mantle composition beneath the Galápagos. Overall, however, the

Copyright 2002 by the American Geophysical Union 1 of 32

 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

compositional variation at Volcan Ecuador has been minimal, suggesting that, unlike Hawaii, the
compositions of magmas do not substantially change while the volcano is in its shield-building phase.

Components: 12,170 words, 18 figures, 4 tables.

Keywords: Galápagos; ocean; island; mantle plume; igneous; petrology; caldera.
Index Terms: 1749 History of Geophysics: Volcanology, geochemistry, and petrology; 8499 Volcanology: General or
miscellaneous; 3640 Mineralogy and Petrology: Igneous petrology.
Received 4 April 2002; Revised 19 July 2002; Accepted 25 July 2002; Published 25 October 2002.

Geist, D., W. M. White, F. Albarede, K. Harpp, R. Reynolds, J. Blichert-Toft, and M. D. Kurz, Volcanic evolution in the
Galápagos: The dissected shield of Volcan Ecuador, Geochem. Geophys. Geosyst., 3(10), 1061, doi:10.1029/2002GC000355,
2002.

1. Introduction: the Importance of Albarede et al. [1997]; Kerguelen: Weis et al.

Compositional Evolution
[2] The evolution of ocean island volcanoes is
important not only because the regular changes
constitute one of their most fundamental attributes,
but these changes have also led to major break-
throughs in the development of modern ideas on
magma generation, and, in particular, the dynam-
ics of deeply-rooted mantle plumes. From the time
that they emerge from the deep seafloor until they
go extinct, most ocean island volcanoes experi-
ence changes in eruptive styles and rates, varia-
tions in magma composition, and shifts in the
isotopic and trace element compositions of the
lavas.
[3] Nowhere is the evolution of ocean island vol-
canoes better known than Hawaii, where the pio-
neering work of Stearns [1946] showed that the
volcanoes progress through a sequence of stages,
including emergence from the submarine phase,
then tholeiitic shield-building, alkaline shield-mod-
ifying, and finally posterosional activity. Accom-
panying the change in eruptive style, a systematic
compositional evolution has also been demonstra-
ted at timescales ranging from days [Garcia et al.,
1989] to centuries [Kurz and Kammer, 1991; Pie-
truzska and Garcia, 1999a] to hundreds of thou-
sands of years [McDonald and Katsura, 1964]. As
other archipelagos are investigated in detail, sim-
ilarly consistent changes in the eruptive styles and
compositions are being discovered (e.g., Reunion:
[1998]; Canaries: Hoernle and Schmincke [1993]).
[4] In the Galápagos archipelago, such a picture of
consistent volcanic, petrologic, and geochemical
evolution is not nearly so clear. Variations in ages,
volcano morphology, petrology, and geochemistry
do correlate with the direction of plate motion. The
western volcanoes are younger, higher in altitude,
steeper, and have calderas. In contrast, the volca-
noes to the east are generally older, lower in
elevation, have more subdued shields, and lack
calderas [McBirney and Williams, 1969]. Likewise,
the volcanoes in the western part of the archipelago
generally erupt compositionally-uniform tholeiitic
lavas, whereas those to the east erupt lavas of more
diverse composition, ranging from tholeiitic to
alkaline [e.g., Geist et al., 1988; White et al.,
1993; Harpp and White, 2001].
[5] The west-to-east differences in magmatic com-
positions are not necessarily caused by different
stages of evolution, however. According to this
hypothesis, western-like shields underlie the
younger rocks that cap Santiago, Santa Cruz,
Floreana, and San Cristobal volcanoes (Figure 1).
Another possibility, though, is that each of the
volcanoes in the Galápagos is fundamentally dis-
tinct throughout its life. In this case, the differences
in eruptive style and composition are due to differ-
ences in lithospheric structure, variable interaction
with the migrating Galápagos Spreading Center,
and transient plume processes. Erosion in the arid
2 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

3˚
The Galapagos Islands N

Wo
lf-
Da
Darwin I.

rw
in
Wolf I.

Lin
GSC

eam
Volcan Pinta

en
t
Ecuador Marchena Genovesa
Volcan
Wolf
Santiago
0˚

Fernandina Santa Cruz

San Cristobal

Isabela
Tortuga 1˚S
Floreana m
Espa–ola 2000

92˚W 88˚W
30 00 m

Figure 1. Regional setting of Volcan Ecuador. GSC is the Galápagos Spreading Center. Arrow at bottom shows
motion of the Nazca plate in the hot spot reference frame.

Galápagos is minimal, thus there is no access to the
older lavas on the eastern shields, so this hypoth-
esis is difficult to test. Compositional evolution has
been addressed at several of the western shields,
however. Geologic mapping and petrologic study
of Alcedo [Geist et al., 1994], Sierra Negra [Rey-
nolds et al., 1995], Cerro Azul [Naumann and
Geist, 2000], and Wolf [Geist et al., 2000] have
established a reasonably-constrained stratigraphy at
each volcano. No regular change in the composi-
tions of the lavas with time has been recognized at
the latter three volcanoes. Alcedo is exceptional,
because rhyolite has erupted in the younger episode
of activity, which has been attributed to it being in a
late stage of evolution [Geist et al., 1995]. Like-
wise, a comprehensive study of the historic lavas
from Fernandina volcano reveal remarkably little
compositional variation [Allan and Simkin, 2000].
[6] Volcan Ecuador, at the northwest tip of Isabela
Island (Figure 1), provides a unique opportunity to
explore volcanic and petrologic evolution, because
it has an unusually accessible caldera wall, and the
caldera floor is dissected. Further, the volcano is
small and at the leading edge of the northern part of
the archipelago, ‘‘upstream’’ of the much more
active Wolf and Santiago volcanoes. Therefore,
stratigraphic sampling combined with detailed geo-
logic mapping and geochemical analyses should
reveal how Galápagos volcanoes evolve, particu-
larly in their early shield-building stages. Another
attribute of Volcan Ecuador that makes it a target in
a search for compositional evolution is that several
reconnaissance samples indicated that its lavas
have relatively enriched isotopic signatures [White
et al., 1993]. This contrasts with neighboring Wolf
volcano, where lavas are the most depleted of all of
the western Galápagos volcanoes [Geist et al.,
2000]. Thus, one might predict that Ecuador’s lavas
might become increasingly depleted with time, as it
is carried toward a Wolf-like mantle domain.
2. Tectonic Setting
[7] The Galápagos Islands are a hot spot-related
chain that lies just to the south of the Galápagos
Spreading Center, a divergent plate boundary that
separates the Cocos plate to the north from the
Nazca plate to the south (Figure 1). The absolute
motion of the Nazca plate is 37 km/my at 91

[Gripp and Gordon, 1990], thus the youngest vol-
3 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Outer Flank ER
Section 3
NA
ER
Wolf
A IC
Volcano
Outer Flank VS CR
Section 2
CG
Caldera Wall
Section
IC
CF
CR

VS ER
PVR
Units
ER = East Rift A'
NA = North Apron
IC = Intracaldera 5 km
VS = Palagonite Tuff and Basaltic tephra and spatter
CF = Caldera Floor
CR = Caldera Rim

Figure 2. (Top) Geologic map of Volcan Ecuador. 100-m contours are from DEM produced by Topsar mission
[Mouginis-Mark et al., 1996], although some are interpolated from field data and aerial photo inspection where
Topsar data are lacking. A-A0 shows the line of section in Figure 3. CG = Cerro Grande and PVR = Punta Vincente
Roca. (Bottom) Aerial photograph of Volcan Ecuador taken in 1960.

canoes are found at the western end of the archipe-
lago. Extrapolation of magnetic stripes suggests that
Volcan Ecuador lies on lithosphere that is approx-
imately 8 million years old [Wilson and Hey, 1995].
Gravity and plate-flexure models suggest that the
crust supporting Volcan Ecuador is about 11 km
thick [Feighner and Richards, 1994]. Volcan Ecua-
dor abuts Volcan Wolf, which is very active, mor-
phologically youthful, and the highest volcano in
the archipelago. Volcan Ecuador, in contrast, has a
remarkably eroded appearance (Figure 2), which led
McBirney and Williams [1969] to conclude that it
4 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

a)
A IC A'
VS
ER
CF CR

Sector Caldera Circumferential

Collapse Fault Dike Swarm

b)

Figure 3. (a) Cross section of Volcan Ecuador. Vertical exaggeration is 2.5X. (b) View into the former caldera of
Volcan Ecuador. Block that has slumped from the caldera rim can be seen perched 2/3 of the way up the cliff. Dark
lavas of the younger intracaldera series are visible in the foreground. The tuff cone Cerro Grande is to the left. Sector
collapse occurred along a fault that coincides with the west coast of the volcano. Note the asymmetry (west is lower)
of the caldera rim; this is the primary dip direction of most of the caldera wall lavas.

was older than its neighbors. Nevertheless, many
young flows occur in the former caldera, the north-
ern apron, and the peninsula that connects it to
Wolf. Because of this, the low elevation of the
volcano (
800 meters), and the absence of volca-
noes to the west, we believe that Volcan Ecuador is
probably younger than its neighbors and in a
juvenile, shield-building stage.
3. Geology of Volcan Ecuador
[8] The most striking aspect of Volcan Ecuador is
that its western half is missing; the remaining
eastern half has most of the characteristics of the
Galápagos shield volcanoes, including unusually
steep upper slopes, circumferential dikes, and a
proportionately-large caldera (Figures 2, 3, and 4).
3.1. Geologic Units
[9] We have divided Volcan Ecuador into 6 geo-
logic units on the basis of age, geographic location,
and characteristic features observed in the field
(Figures 2 and 3). They are discussed in order of
decreasing age.
3.1.1. Caldera Rim Lavas
[10] The oldest lavas of Volcan Ecuador are best
exposed in the cliffs of the prominent west-facing
amphitheater, which we interpret as a former cal-
dera wall, and on the outer steep slope of the
volcano, where mass wasting has exposed them
(Figures 2, 3, and 4). Most of these lavas are 0.5 to
2.5 m thick pahoehoe flows, although about 20%
are a’a. Owing to the severe terrain of the caldera
walls, we were not able to measure the lateral
extent of individual lavas, but most are at least
tens of meters wide. We estimate that roughly 150
(±50) individual flow units are exposed in the
easternmost part of the amphitheater. Within the
amphitheater wall, the lavas are subhorizontal, with
dips <4
. The lavas exposed on the outer flanks
have outward primary dips ranging up to about 15
.
5 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Pre-Sector Collapse (ca. 100 Ka)

CF CR
CR

Sector Collapse

Modern Volcano

Figure 4. Structural development of Volcan Ecuador, looking north (approximately the line of section of Figure 2).
(Top) The volcano originally had the morphology of a typical Galápagos shield, with a large caldera. The west rim
was lower than the east rim, as indicated by the skewness of the dip of rim-facies lavas currently exposed in the
caldera walls. (Middle) Half of the volcano slides 3000 m down the western margin of the Galápagos platform,
bisecting the volcano. (Bottom) Recent eruptions mostly occur from the old caldera floor (unit IC), the northern apron
(not included in this section), and the newly-formed East Rift (ER).

The lavas exposed in the amphitheater and those of
the outer slopes do not dip perfectly radially; there
is a skewing of the primary dips to the east,
suggesting that the western part of the volcano
has been topographically lower for most of the
emergent history of the volcano. The thicknesses,
attitudes, and vesicle distributions in these lavas
resemble those ‘‘rim-facies’’ lavas that have been
identified in the caldera walls at Cerro Azul [Nau-
mann and Geist, 2000] and Wolf [Geist et al.,
2000] volcanoes. They are similar to lavas that
are erupted from circumferential fissures cutting the
summit platforms of the active shields, which tend
to be thin and dominated by pahoehoe. Rim facies
lavas flow from these eruptive fissures both out-
wards and inward down the caldera wall, then are
6 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Table 1. Isotope Ratios of Volcan Ecuador Lavasa

MC-ICP-MS TIMS
Sample
87 86 143 144 176 177 206 204 207 204 208 204 206 204 207
Number Sr/ Sr Nd/ Nd Hf/ Hf Pb/ Pb Pb/ Pb Pb/ Pb Pb/ Pb Pb/204Pb 208
Pb/204Pb

Caldera Rim
E97-50 0.703063 0.512965 0.283081 19.165 15.575 38.832 19.184 15.604 38.914
E97-52 0.703167 0.512947 0.283039 19.248 15.582 38.938
E97-58 0.703150 0.512945 0.283068 19.245 15.587 38.945 19.230 15.581 38.910
E97-67 0.703117 0.512957 0.283072 19.188 15.577 38.840 19.273 15.598 38.813
E97-70 0.703133 0.512949 0.283078 19.200 15.593 38.875 19.170 15.558 38.769
E97-76 0.703105 0.512960 0.283053 19.210 15.596 38.895
E97-79 0.703115 0.512962 0.283053 19.197 15.564 38.888 19.209 15.590 39.008
E97-88 0.703083 0.512965 0.283088 19.160 15.575 38.801 19.150 15.564 38.765
E97-99 0.703049 0.512975 0.283062 19.171 15.566 38.791
E97-101 0.703082 0.512975 0.283057 19.155 15.579 38.805
E97-49 0.703036 0.512975 0.283068 19.082 15.573 38.709
E97-08 0.703100 0.512963 0.283083 19.182 15.574 38.824
E97-09 0.703103 0.512951 19.204 15.584 38.865
E97-11 0.703274 0.512972 19.098 15.553 38.691
E97-14 0.703066 0.512960 19.186 15.577 38.826 19.185 15.575 38.822
E97-15 0.703047 0.512962 0.283056 19.184 15.578 38.832
E97-17 0.703079 0.512946
E97-127 0.703172 0.512941 0.283026 19.293 15.593 39.031
E97-130 0.703143 0.512958 0.283044 19.225 15.584 38.910
E97-131 0.703058 0.512972 0.283068 19.145 15.581 38.806 19.135 15.571 38.776
E97-134 0.703049 0.512956 0.283069 19.209 15.592 38.888

Caldera Floor
E97-108 0.703145 0.512959 0.283068 19.194 15.587 38.870
E97-110 0.703100 0.512961 0.283063 19.243 15.594 38.928

Postcollapse Intracaldera
E97-04b 0.703221 0.5129433 0.283055 19.253 15.594 38.986
E97-05 0.703139 0.5129487 19.159 15.583 38.857
E97-20 0.703256 0.51295585 0.283052 19.248 15.582 38.980

North Apron
E97-24b 0.703200 0.512964 0.283071 19.250 15.594 38.971 19.255 15.599 38.987
E97-48b 0.703193 0.512953 0.283065 19.238 15.592 38.953 19.229 15.583 38.925

East Rift
E95-02b 0.703271 0.512934 0.283052 19.289 15.601 39.037 19.288 15.605 39.040
E95-03 19.239 15.588 38.912
E95-10 0.703141 0.512940 19.309 15.586 39.011
E95-13 0.702724 0.513030 18.957 15.528 38.392
E95-18 0.703120 0.512949 19.253 15.608 39.021
E95-22b 0.703168 0.512935 0.283045 19.330 15.600 39.050 19.299 15.586 39.008
a
Analyzed values of standards are 176Hf/177Hf = 0.282160 for JMC 475, 143Nd/144Nd = 0.511855 for La Jolla, 87Sr/86Sr = 0.710245 for NBS
987, 206Pb/204Pb = 16.908, 207Pb/204Pb = 15.453, and 208Pb/204Pb = 36.591 for NBS 981 standard. Pb isotope ratios were corrected for mass
fractionation by applying a correction of 0.8 % per dalton. Errors (2s) estimated from replicate analysis of standards are 0.000013 for 87Sr/86Sr,
0.000012 for 143Nd/144Nd, 0.000005 for 176Hf/177Hf, 0.009 for 206Pb/204Pb, 0.012 for 207Pb/204Pb, and 0.040 for 208Pb/204Pb.
b
Data previously reported in Blichert-Toft and White [2001] and included here for convenience.

cut by younger caldera collapse and rock avalanch- wasting of the caldera wall removed circumferen-
ing. Although circumferential vents are notably tial vents and exposed the dikes.
absent from the summit of Volcan Ecuador (Figure
2), numerous dikes strike subparallel to the caldera [11] The caldera rim lavas are stratigraphically the
wall. Our interpretation is that late-stage mass oldest exposed on Volcan Ecuador. One of the
7 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems
uppermost flows from the easternmost part of the
wall has an 39 Ar/ 40Ar age of 127 ± 32 Ka.
Unfortunately, repeated attempts to determine the
ages of several of the lowermost flows in this
exposure yielded no meaningful results. Rim facies
flows near the bottom of two cliff sections cut in
the northwestern flank of the shield yield ages of 91
± 20 Ka and 100 ± 9 Ka (Figure 2). This cluster of
ages from widely-separated lavas suggests that the
Volcan Ecuador shield began growing about 100
Ka and grew rapidly. The rapid increase in sub-
aerial volume may have led to gravitational insta-
bility and the sector collapse that followed.
[12] Cosmogenic 3He exposure ages [Kurz, 1986]
were determined on two samples from the exposed
interiors of caldera rim lavas, one from midway up
the caldera wall (Section 1 in Figure 2) and one
from the base of a steep cliff in the outer part of the
shield (Section 3 in Figure 2). Two splits from the
sample from the outward-facing cliffs yield ages of
3400 ± 800 and 800 ± 300 years (Table 1). The
sample from the caldera wall yields an age of 2800
± 400 y. Note that the stated uncertainties reflect
only measurement uncertainties, and that geologi-
cal factors (such as erosion) and the production-rate
scaling contribute significantly to the uncertainties.
In particular, the caldera wall ages are minimum
estimates of the solidification because they are
obtained on cliffs exposed by caldera faulting and
mass wasting. The exposure ages are far younger
than the 39Ar/40Ar age measured on samples from
the same stratigraphic units, and are consistent with
recent mass wasting from the cliffs. The lack of
extensive talus piles beneath the outward-facing
cliffs suggests that the talus is quickly buried by
lava flows, and thus that the outer apron flows are
less than several millenia old.
[13] Dozens of dikes cut the caldera wall section
and are also well exposed on the cliffs of the south
shore, east of Punta Vincente Roca. The ones we
inspected range in thickness from 0.2 to 2.4 m. The
dikes strike subparallel to the caldera wall, indicat-
ing that they fed circumferential fissures that char-
acterize all of the Galápagos shields. An interesting
aspect of these dikes is that not one of the dozens
we observed intrudes a fault: nowhere are the
contacts between flows offset across a dike. This
´ PAGOS
provides strong support for the contention of Chad-
wick and Howard [1991] and Reynolds et al.
[1995] that the dikes do not follow caldera-forming
faults but instead create their own fractures and are
guided by stresses set up by the caldera walls and
steep outer slopes.
[14] Volcan Ecuador has a smooth, slightly east-
dipping summit platform. A younger cone and flow
lie on top of the platform, and the platform is cut by
four large pit craters that align with the East Rift.
The pit craters are not eruptive vents but are collapse
features. They expose thin, outward-dipping flows
(B. Nolf, personal communication, 2001).
[15] A large (2 km-long by 400 m wide) block is
perched part way up the southeast caldera wall
(‘‘Slump Block’’ in Figure 3b). This block is
fractured and is rotated so it now dips inward, and
the lavas are rim-facies and not caldera-floor facies.
Thus, the block has slumped from the summit.
[16] The steep outer flanks on the eastern side of
the shield are formed by unusual triangular flatirons
separated by V-shaped, upward-widening arroyos.
The flatirons are the remnants of formerly contin-
uous summit lava flows that drape the steep outer
flanks. Headward erosion of the arroyos has dis-
sected the flows into steeply-dipping triangular
wedges. Each arroyo has a fan shaped debris fan
at its base. Lava flows of adjacent flatirons can be
traced continuously to the break in slope that marks
the beginning of the summit platform. The lavas are
thin (<0.5 m) pahoehoe flows. Lava drips on the
underside of some flatiron flows are vertical indi-
cating that the flows maintain their primary dip and
have not been appreciably deformed since solid-
ification. These features indicate that erosion of the
outer flanks is top down rather than by basal wave
action [c.f. Rowland et al., 1994]. Although these
unusual geomorphic forms are not abundant else-
where in the Galápagos, upward-widening arroyos
are common on the flanks of large tuff cones and
thus are typical of steeply-dipping strata exposed to
arid climate.
3.1.2. Caldera Floor Lavas
[17] The caldera floor sequence is exposed on cliffs
below a flat bench that covers the upper half of the
8 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems
amphitheater floor (Figures 2 and 3). There has
been some difference of opinion as to the nature of
this bench and whether a caldera existed in Volcan
Ecuador prior to sector collapse [McBirney and
Williams, 1969; Simkin, 1984; Chadwick and
Howard, 1991; Rowland et al., 1994]. The flows
that crop out on this escarpment are thick (2 to 6
m), have flat primary dips, and have undergone
almost complete vesicle segregation from massive
interiors. All of these are characteristics of thick
flows that ponded in the flat floors of other Gal-
ápagos calderas and were cut by renewed caldera
faulting. Similar lavas that crop out in the caldera
walls of Alcedo [Geist et al., 1994], Cerro Azul
[Naumann and Geist, 2000], Fernandina [Rowland
and Munro, 1992], and Wolf [Geist et al., 2000]
volcanoes and have been coined ‘‘caldera-filling
facies’’. Hence, we interpret this bench as the
remains of a caldera floor and the cliff face bound-
ing it to the east as the old caldera wall.
[18] A specimen from the lowermost of these lavas
yields a 39Ar/40Ar plateau age of 50 ± 11 Ka.
Another lava from this sequence yields a 3He
exposure age of 96 ± 2 Ka. As mentioned earlier,
the stated uncertainty in the exposure age is ana-
lytical only; the time-integrated shielding from
cosmic ray flux due to the steep caldera floor
terrain is likely to be a major source of uncertainty
in this case (shielding is assumed to lead to 50%
reduction in production rate; Table 1). Neverthe-
less, the exposure age is in reasonable agreement
with caldera rim flow 39Ar/40Ar ages (90-130 Ka).
[19] The origin of the west-facing escarpment that
exposes the caldera floor flows is not known. It
could be the fault plane bounding the headwall of
the sector collapse. Alternatively, it could have
been the face of a bench in the caldera floor, such
as those that occur on other Galápagos shields and
caused by multiple filling and caldera collapse
events [Rowland and Munro, 1992; Naumann
and Geist, 2000].
3.1.3. Postsector Collapse Intracaldera
Lavas
[20] Over a dozen flows have erupted from small
cones and fissures from inside the former caldera
since it was cut by sector collapse (Unit IC of
´ PAGOS
Figure 2). Two groups of flows characterize this
unit. First, several small cones and fissures erupted
along an arc parallel to the caldera wall, on the
upper level of the bench formed by the old caldera
floor. Second, a set of fissures cutting the former
caldera floor has poured out massive a’a lava flows
that form the lowest bench and most of the west
coast of the volcano. The ages of these lavas are
unknown. They are certainly much younger than
the oldest caldera floor flows (50 to 100 Ka), and
the youngest ones which are almost completely
unweathered, thus unlikely to be more than a few
millennia old.
3.1.4. Palagonite Tuff and Basaltic Tephra
[21] Vents near the coast have produced two prom-
inent tuff cones: Punta Vincenta Roca and Cerro
Grande (Figure 3b). Cerro Grande is cut by two
east-striking faults located on the south slope.
Punta Vincenta Roca, located on the southwest
tip of Volcan Ecuador, is extensively eroded and
presumably the older of the two cones.
3.1.5. North Apron Lavas
[22] To the north of the main edifice lies a broad
coastal apron, extending about 2 km from the old
volcano to the coast. These lavas are dominantly
a’a, but tube-fed pahoehoe also occurs. Most vents
abut the old shield, and the apron lavas uncon-
formably overlie the flows of the old shield. The
youngest lavas are probably no more than a few
centuries old: they are virtually unweathered and
unvegetated.
3.1.6. East Rift Lavas
[23] Volcan Ecuador has one of the best-developed
rift zones in the entire Galápagos archipelago. This
rift forms a 6 km-long, 4 km wide isthmus that
connects Volcan Ecuador to the flanks of Volcan
Wolf. The ridge crest includes three sets of eruptive
fissures, each about 3 km long and which diverge
from one another down slope to the east. East Rift
lavas are a mix of a’a and pahoehoe that flowed
perpendicular to the fissures and form thin, narrow
flows that are 1 to 2 km long; most reach the coast.
[24] Rift flows cover older flows from the base of
the shield yet deflect around the debris fans asso-
9 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems
ciated with the arroyos, indicating that the subaerial
erosion of the outer shield was occurring prior to
the development of the East Rift. The East Rift
lavas are almost completely unweathered and are
morphologically the youngest on the volcano
(although possibly contemporaneous with the north
apron and intracaldera lavas). A single 3He expo-
sure age of an East Rift pahoehoe lava yields a
maximum age of 800 years (because cosmogenic
helium is below detection; Kurz and Geist [2000]).
Geologic observations, and comparison with his-
torical lava flows, suggest that most of the East Rift
lavas are no more than several centuries old.
3.2. Structural and Morphological
3.2.1. Development of Volcan Ecuador
[25] Four stages of growth are apparent in the
volcanic evolution of Volcan Ecuador (Figure 4).
3.2.1.1. Stage 1: Submarine Growth
[26] Little is known about the submarine growth of
Volcan Ecuador, but it had to have grown >3000 m
from the abyssal depths to the west, plus any
amount the volcano has subsided. The recent
DRIFT04 cruise has provided detailed bathymetric
and side-scan sonar images of the submarine flanks
of the volcano, and those results are forthcoming
[Fornari et al., 2001].
3.2.1.2. Stage 2: Shield Building
[27] As Volcan Ecuador emerged above the sea, it
took on the classic form of a Galápagos shield. It
did this by erupting low-volume flows of pahoehoe
from a circumferential fissure system established
along the caldera rim. A caldera probably has
existed through all of the emergent growth of the
volcano. The gravitational stresses imposed by the
caldera wall and steep outer slopes result in persis-
tent circumferential eruptive fissures. These vents
are not exposed, but the dikes that fed them crop
out in the caldera wall. Small-volume eruptions
from these fissures form steep upper slopes on the
flanks of the volcano [Simkin, 1972]. Eruptions
from within the caldera produced lavas that ponded
within the caldera, occasionally partly filling it. At
about 100 Ka, Volcan Ecuador had the morphology
´ PAGOS
of an asymmetric Galápagos shield (lower to the
west), with outer slopes of about 25
, a flat summit
rim, and a 600 m deep caldera. Radial lower flank
fissures, comparable to those of the other western
Galápagos shields, are notably absent from Volcan
Ecuador. However, the lower flanks of the volcano
are either covered by the young flows of the East
Rift and North Apron, or are under water.
3.2.1.3. Stage 3: Sector Collapse
[28] Because it grew on the edge of a steep, 3000
m high submarine escarpment, Volcan Ecuador
was highly susceptible to large-scale mass wasting,
similar to the prodigious landslides that occur
around the Hawaiian islands [Moore et al.,
1989], Tenerife [Marti et al., 1997], and Reunion
[Duffield et al., 1982]. Volcan Ecuador experi-
enced one or more catastrophic sector collapse
events, the net effect of which was to remove the
entire western half of the volcano. The age of the
sector collapse is not known, but it probably took
place near the time that the shield-building phase
of the volcano waned, at about 100 Ka. The extent
of the downdrop and run-out is unknown, but slip
had to have been at least 800 m and runout over a
kilometer, as evidenced by the absence of any slide
remnants along the west coast. In fact, it is likely
that the entire half of the volcano slid to abyssal
depths of over 3000 m. The DRIFT04 cruise
[Fornari et al., 2001] imaged the seafloor to the
west of Volcan Ecuador with side-scan sonar, and
some hummocky terrain is apparent to a distance
of about 20 km from the west coast, but most of
the submarine slope is obscured by young lava
flows.
[29] Whether sector collapse occurred in a single or
multiple events is difficult to judge. Volcanic activ-
ity on the caldera rim and from the caldera floor
effectively ceased by about 100 Ka. The most
likely reason for this cessation is sector collapse:
with loss of the western half of the volcano,
magmas no longer rose to the summit, as they
found easier outlets at low elevation.
[30] Poland [2001] has recently suggested that the
circumferential diking pattern of Galápagos-type
shields is due to the stresses imposed by loading
10 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems
of dense rocks beneath the caldera floor. In con-
trast, others have contended circumferential dikes
are due to radially-developed tension set up by the
caldera walls and steep outer slopes [Simkin, 1972;
Chadwick and Howard, 1991; Reynolds et al.,
1995; Naumann and Geist, 2000]. The coincidence
of sector collapse with the cessation of eruption
from circumferential vents points to the second
hypothesis.
[31] Sector collapse is known elsewhere in the
Galápagos, for example on the southwest flank of
Cerro Azul volcano [Naumann and Geist, 2000],
the west face of Pinta [Cullen and McBirney,
1987], and the west face of Santa Fe [Geist et al.,
1985]. Although sector collapse in the Galápagos is
not as common or voluminous as in Hawaii, Volcan
Ecuador is a particularly dramatic example, and
catastrophic collapse events appear to be a common
feature of volcanic evolution in the Galápagos, and
perhaps in the evolution of oceanic volcanoes
generally.
3.2.1.4. Stage 4: Rejuvenation and Rift Zone
Development
[32] Following sector collapse, volcanic activity
from the caldera rim dropped dramatically. Lower
elevation vents on the old caldera floor were
established, a focused rift zone was constructed to
the east, and an apron built to the north.
[33] The East Rift zone results from stresses created
by the loads of Volcan Ecuador and nearest-neigh-
bor Volcan Wolf [Chadwick and Dieterich, 1995].
Once a ridge was built, magmas were focused into
the rift by the weight of the ridge itself. While it is
impossible to say when the East Rift began to
develop, its lavas clearly overlie those of the
dissected shield, and the pit craters related to East
Rift fissuring cut the youngest lavas of the shield;
we suspect that the shift in vent locations to the
East Rift coincided with sector collapse, although a
long hiatus could have occurred. Before the rift
zone was constructed, Volcan Ecuador was an
isolated island, separate from Isabela. This ridge
now serves as a land-bridge for the biologic colo-
nization of Volcan Ecuador: although it never (so
far as is known) has been populated by tortoises, it
´ PAGOS
is home for terrestrial iguanas apparently of Wolf
stock (H. Snell, personal communication, 2001).
[34] Rowland et al. [1994] proposed that the steep
outer flanks of Volcan Ecuador resulted from
marine erosion during a prolonged period of inac-
tivity (20,000–30,000 years). While wave cutting
may well have accentuated primary steep slopes,
particularly on the south side of the volcano, we
find no evidence (wave-cut terraces, beach depos-
its) for such action. We suggest instead that the
eroded appearance of Volcan Ecuador, results not
from a hiatus of volcanic activity and marine
erosion, but from a shift in volcanic activity from
the summit to the flanks and a ‘‘top down’’ ero-
sional style accelerated by periodic El Niño cli-
matic events. The steep outer slopes of Volcan
Ecuador’s former shield are due principally to
construction, with highest accumulation rates at
the summit, a process originally proposed by Sim-
kin [1972].
4. Petrology and Geochemistry
4.1. Analytical Techniques
[35] Bulk rock analyses (124) of major elements
and Ni, Cr, Sc, V, Ba, Rb, Sr, Zr, Y, Nb, Ga, Cu,
Zn, Pb, and Th were performed by XRF in the
Geoanalytical Laboratory at Washington State Uni-
versity (Table 2 [Johnson et al., 1999]). Blind
triplicate analyses yield reproducibility at levels
of <1% RSD for the major analytes and <5% for
the trace elements, except Rb, which has a stand-
ard deviation of 12% at these low concentrations.
After all of the analyses were completed, it was
discovered that the Li-tetraborate used to prepare
some of the samples had been contaminated with
Cr and Ni (but no other analyte in measurable
quantities). Cr and Ni analyses of the affected
samples were therefore discarded. The contami-
nated material was not used for isotopic or ICP-
MS analyses.
[36] Additional trace element analyses were carried
out using the Hewlett-Packard HP4500 ICP-MS at
Colgate University (Table 3). All chemical proce-
dures were conducted in HEPA-filtered Plexiglas
clean boxes using prepurified reagents and ultra-
11 of 32
 Table 2. ICP-MS Analyses of Trace Elements in Volcan Ecuador Lavas [The full Table 2 is available at http://www.g-cubed.org]
Sample Number Rb Sr Y Zr Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U

Caldera Rim
E97-50a 379.9 29.28 172.6 123.2 17.43 40.2 5.46 24.71 6.05 1.97 6.24 1.01 5.43 1.04 2.76 0.39 2.35 0.36 4.1 1.24 1.53 1.67 0.52
E97-52a 8.97 321.6 26.69 142.8 98.1 12.54 29.2 4.11 18.59 4.74 1.61 5.12 0.85 4.65 0.92 2.4 0.34 2.03 0.31 3.5 0.97 1.27 0.39
E97-58a 337 32.56 180.4 130.3 17.66 40.8 5.47 25.07 6.35 2.05 6.65 1.08 5.91 1.14 2.97 0.42 2.49 0.38 4.47 1.46 1.41 0.45
E97-67a 30.22 164.5 128.9 17.66 40.2 5.68 25.39 6.23 2.03 6.44 1.05 5.7 1.1 2.89 0.41 2.46 0.37 3.93 1.23 1.38 1.66 0.50
Geophysics
Geosystems

E97-70a 418 30.23 170.1 121 16.06 37.5 5.12 23.02 5.66 1.9 5.93 0.97 5.2 1 2.64 0.38 2.23 0.34 4.23 1.34 1.3 1.92 0.61
Geochemistry

E97-76a 9.64 448.2 31.96 193.8 132.3 17.56 40.6 5.64 24.84 5.95 1.98 6.15 0.96 5.28 1.04 2.66 0.37 2.28 0.34
G

E97-79a 166.3 22.07 51.1 7.07 31.34 7.47 2.41 7.47 1.21 6.51 1.26 3.27 0.46 2.76 0.42
3

E97-88a 434 26.52 155.2 124.5 18.01 42.3 5.93 26.8 6.58 2.15 6.82 1.09 5.92 1.17 3.04 0.43 2.58 0.4 3.78 1.21 1.31 1.39 0.44
E97-99a 406.6 32.01 194.8 149.5 21.15 49 6.55 29.7 7.31 2.35 7.61 1.26 6.6 1.28 3.38 0.48 2.83 0.43 4.72 1.46 1.74 1.88 0.63
E997-101 415.2 44.57 267.4 190.2 27.94 65.2 8.85 38.04 8.32 2.72 8.79 1.5 8.06 1.55 3.98 0.63 3.5 0.54 6.47 1.79 2.1 2.53 0.76
E97-49a 158.6 23.02 54.2 7.52 33.06 7.74 2.49 8.11 1.3 7.06 1.38 3.6 0.51 3.06 0.47
E97-08 10.57 413.7 38.62 248.1 176.4 24.59 58.4 7.76 35.19 8.18 2.48 7.87 1.29 7.27 1.44 3.91 0.59 3.57 0.52 5.85 1.82 1.51 2.04 0.59
E97-09 10.2 477.8 35.68 223.1 164.4 21.15 50.5 6.68 29.66 6.83 2.2 6.96 1.23 6.99 1.4 3.57 0.53 3.14 5.76
E97-11 480.2 34.06 198.4 142.7 20.2 46.9 6.33 27.36 5.98 2.06 6.39 1.09 5.92 1.17 2.93 0.45 2.58 0.39 4.67 1.45 1.94 1.83 0.58
E97-14 7.95 441.3 34.39 219.1 145.5 21 49.1 6.75 28.85 6.88 1.9 6.74 1.17 6.69 1.3 3.35 0.52 3.03 0.45 5.1 1.52 1.44
E97-15 11.16 411.5 42.1 263.6 169.4 24.89 59.6 8.06 35.66 8.61 2.48 8.2 1.38 7.97 1.57 4.08 0.61 3.67 0.56 6.52 1.76 1.85 2.32 0.66
E97-17 27.52 165.7 115.4 15.75 38 5.17 22.64 5.01 1.82 5.36 0.91 5.04 1 2.52 0.38 2.17 0.33 3.88 1.32
E97-127 17.1 428 40.85 267.5 208.3 25.93 59 8.04 34.11 8.12 2.38 7.62 1.26 7.44 1.48 3.96 0.58 3.44 0.52 2.14 1.94 2.81 0.79
E97-130a 16.34 466.1 39.27 249.4 220.9 23.97 53.5 7.22 31.54 7.52 2.45 7.84 1.24 6.67 1.29 3.42 0.49 2.99 0.45
E97-131a 10.75 447.3 31.78 202.4 133.7 17.65 41.4 5.78 25.72 6.15 2.01 6.24 1.06 5.51 1.07 2.84 0.41 2.4 0.37 4.53 1.43 1.43 1.91 0.57
E97-134a 377.9 17.05 109.6 15.24 35.7 4.95 22.01 5.31 1.77 5.46 0.86 4.7 0.92 2.41 0.34 2.04 0.32

Caldera Floor
E97-108 165.9 21.95 51.5 6.94 31.52 7.83 2.54 8.27 1.38 7.34 1.45 3.81 0.55 3.22 0.49
E97-110a 5.96 468.3 20.79 137.8 122.6 15 33.5 4.64 20.15 4.8 1.67 4.98 0.78 4.28 0.84 2.16 0.3 1.82 0.27

Postcollapse Intracaldera
E97-04a 13.25 392 35.72 222.4 175.1 22.04 50.4 6.95 30.73 7.47 2.4 7.72 1.24 6.61 1.28 3.32 0.48 2.82 0.43 5.57 1.58 2.21 0.61
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA

E97-20a 14.31 394.9 35.84 230 189.8 23.47 53.5 7.4 32.81 8.04 2.58 8.43 1.31 7.15 1.38 3.57 0.5 3.05 0.46 5.8 1.91 1.83 2.01 0.59

North Apron
´ PAGOS

E97-24a 385.8 200.8 154.518.8 43.3 5.97 26.36 6.44 2.1 6.77 1.08 5.79 1.14 2.91 0.42 2.46 0.38 4.77 1.62 1.68 1.95 0.59
E97-40a 23.02 54.2 7.52 33.06 7.74 2.49 8.11 1.3 7.06 1.38 3.6 0.51 3.06 0.47
E97-48a 30.52 171.2 162.3 19.7 45 6.22 27.87 6.92 2.24 7.28 1.16 6.34 1.23 3.25 0.46 2.77 0.41 1.39 1.35 1.59 0.47

East Rift
E95-02a 96.9 12.37 28.8 4.04 18 4.57 1.51 4.94 0.81 4.44 0.88 2.28 0.33 1.99 0.3
E95-03 317.3 22.68 126.6 104.7 13.04 30.2 4.16 18.21 4.12 1.46 4.62 0.79 4.33 0.87 2.15 0.31 1.86 0.28 3.2 0.98 1.18 1.26 0.39
E95-05 358.5 34.85 203.9 145.2 20.16 46.6 6.38 27.71 6.14 2.1 6.69 1.14 6.23 1.23 3.12 0.48 2.76 0.41 4.75 1.52 1.59 1.85 0.57
E95-10 332.1 24.62 145.2 115.3 13.88 31.8 4.36 18.81 4.18 1.46 4.58 0.78 4.33 0.86 2.2 0.34 1.97 0.3 3.4 1.18 1.31 1.44 0.44
a
Ba and rare Earths analyzed at Colgate University. All other analyses were performed at Cornell University.
10.1029/2002GC000355

12 of 32
 Table 3. XRF Analyses in Volcan Ecuador Lavas [The full Table 3 is available at http://www.g-cubed.org]
SiO2 Al2O3 TiO2 FeO MnO CaO MgO K2O Na2O P2O5 Total Ni Cr Sc V Ba Rb Sr Zr Y Nb Ga

E9501 48.28 13.51 2.38 10.93 0.18 9.85 13.05 0.40 2.42 0.24 101.24 28 287 119 7 274 125 23 18.4 17
E9502 47.71 13.84 2.47 11.04 0.18 10.05 12.29 0.40 2.44 0.25 100.66 29 294 130 7 276 128 22 18.8 19
E9503 48.18 15.13 2.46 10.53 0.17 10.85 9.55 0.50 2.65 0.28 100.30 30 288 151 7 361 149 25 18.7 20
E9504 48.83 15.41 3.36 11.16 0.19 10.60 6.19 0.75 3.41 0.42 100.32 33 364 235 12 402 211 34 28.5 23
E9505 50.35 15.06 2.80 10.42 0.18 11.03 6.41 0.66 3.15 0.36 100.41 31 306 189 11 368 183 29 24.5 24
Geophysics
Geosystems
Geochemistry

E9506 48.87 14.77 3.35 11.48 0.19 10.71 5.83 0.76 3.64 0.43 100.04 37 371 207 10 409 226 35 30.1 21
E9507 48.76 17.37 2.59 9.99 0.16 11.33 6.13 0.58 3.09 0.32 100.32 31 282 167 9 443 179 28 23.0 24
G

E9508 48.28 15.20 2.68 10.76 0.17 11.23 8.02 0.56 2.72 0.30 99.92 30 302 155 9 370 161 27 20.7 22
3

E9509 48.16 15.61 2.55 10.66 0.17 11.10 8.24 0.54 2.76 0.30 100.08 34 294 168 8 383 154 26 19.0 18
E9510 47.41 15.67 2.62 10.79 0.18 10.13 9.36 0.63 2.99 0.34 100.12 34 290 176 9 374 166 27 23.3 23
E9511 47.68 15.73 2.61 10.69 0.18 10.18 9.39 0.61 2.99 0.34 100.40 32 283 177 9 373 165 26 22.4 20
E9512 49.63 15.09 2.87 10.91 0.18 11.01 6.52 0.64 3.15 0.36 100.36 30 335 186 10 368 177 29 23.0 22
E9513 49.36 17.23 2.57 9.01 0.16 12.32 6.02 0.42 3.05 0.31 100.45 36 301 90 4 413 184 28 13.7 19
E9514 47.24 15.38 2.62 11.19 0.18 9.80 9.61 0.64 2.94 0.33 99.92 26 276 179 12 373 167 26 23.1 22
E9515 49.48 15.01 2.95 10.90 0.19 10.94 6.29 0.69 3.24 0.36 100.05 30 343 204 11 387 186 32 26.3 24
E9516 49.20 16.91 2.71 10.06 0.17 11.36 5.79 0.61 3.33 0.35 100.49 32 309 171 8 431 186 30 24.6 23
E9517 49.59 15.08 2.98 10.62 0.19 10.89 6.29 0.67 3.19 0.36 99.85 29 336 200 10 385 185 30 25.3 22
E9518 48.50 15.35 3.35 10.93 0.19 10.61 6.16 0.77 3.45 0.42 99.73 34 366 231 10 404 212 35 28.2 24
E9519 47.82 15.05 2.44 11.11 0.17 9.96 9.89 0.55 2.87 0.30 100.16 26 244 151 8 356 155 26 19.7 18
E9520 47.80 15.00 2.20 10.47 0.17 10.55 10.33 0.47 2.62 0.27 99.88 35 268 148 8 348 144 23 18.9 23
E9521 48.67 15.31 2.36 10.56 0.17 10.84 8.95 0.53 2.79 0.29 100.47 34 281 148 9 354 155 25 20.6 19
E9522 48.33 15.46 2.57 10.74 0.17 9.91 9.59 0.63 3.00 0.33 100.73 28 257 172 10 376 165 26 23.7 20
E9523 49.20 15.03 2.97 11.09 0.18 10.98 6.45 0.68 3.19 0.36 100.13 33 330 195 11 380 185 29 24.2 23
E9524 49.11 14.50 3.35 12.00 0.20 10.43 5.67 0.68 3.65 0.42 100.01 37 370 186 9 357 220 37 26.2 23
E9525 48.70 16.71 2.84 10.63 0.17 11.31 5.81 0.63 3.29 0.35 100.44 32 333 177 8 422 192 30 25.0 21
E9526 50.15 14.80 3.31 11.69 0.19 10.55 5.94 0.77 3.43 0.42 101.25 31 364 235 13 397 211 33 30.7 25
E9527 48.03 15.24 3.53 11.45 0.19 10.16 6.33 0.74 3.56 0.49 99.71 33 368 217 12 423 231 37 30.9 28
E971 48.92 15.12 3.34 11.44 0.18 10.49 6.35 0.76 3.24 0.43 100.27 57 144 35 364 257 12 395 215 34 29.1 24
E972 49.22 16.05 3.00 11.32 0.18 10.59 7.15 0.56 3.14 0.36 101.56 73 190 38 347 222 7 369 184 30 22.0 23
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA

E973 49.50 15.27 2.93 11.57 0.18 10.86 6.41 0.66 2.94 0.36 100.67 55 113 38 326 240 9 374 185 30 24.0 24
E974 51.87 15.62 3.31 11.72 0.19 10.54 6.18 0.74 3.20 0.43 103.79 44 74 33 364 259 11 384 210 33 28.5 26
E975 49.68 15.73 3.23 11.63 0.18 10.55 6.55 0.62 3.18 0.40 101.75 54 172 38 343 201 8 382 206 34 25.4 25
´ PAGOS

E976 49.03 14.84 2.96 11.36 0.18 10.66 6.34 0.68 3.10 0.37 99.52 51 93 36 338 244 10 377 190 31 26.1 19
E977 49.25 15.90 2.86 10.91 0.17 10.83 6.89 0.56 3.02 0.35 100.74 29 325 208 8 376 182 30 22.7 25
E978 50.06 15.02 3.22 11.61 0.18 10.62 6.13 0.74 3.55 0.43 101.55 38 113 35 392 246 9 402 222 34 27.3 23
E979 49.98 17.87 2.56 9.91 0.15 11.40 5.34 0.57 2.95 0.34 101.07 43 160 34 292 194 8 456 178 28 22.3 24
E9710 49.41 16.83 2.81 9.56 0.16 11.47 5.73 0.67 3.53 0.39 100.55 64 144 31 322 189 8 445 197 30 23.4 20
E9711 50.21 16.73 2.82 10.29 0.16 10.90 6.04 0.62 3.04 0.38 101.19 78 178 33 321 227 10 479 197 30 23.3 26
E9712 50.08 14.85 3.77 11.22 0.18 9.76 5.53 0.68 3.90 0.50 100.46 43 80 34 382 241 10 374 246 42 27.8 24
E9713 46.64 11.67 1.88 12.08 0.18 7.48 19.87 0.41 2.07 0.26 101.54 594 1338 25 222 137 5 287 133 20 15.5 18
E9714 49.55 15.79 3.15 11.34 0.18 10.75 6.06 0.69 3.56 0.41 101.48 66 188 34 365 226 9 398 214 33 25.8 24
a
RSD is relative standard deviation, in percent, based on triplicate analyses.
10.1029/2002GC000355

13 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Table 4. 3He Exposure Agesa

4He 3He/4He 4He 3He/4He
Altitude, Crush, Crush, melt, melt, 3Hec, Age,
Sample m cc STP/g R/Ra ± cc STP/g R/Ra ± atoms ± Ka ± Location
b
GI92-19 24 3.04E-08 9.9 0.1 2.18E-08 9.9 0.1 <6.0E + 04 <.8 East Rift
E97-59c 597 1.29E-10 4.7 1.9 3.30E-10 18.6 0.9 1.70E + 05 1.29E + 04 2.8 0.2 Caldera Rim
E97-110c 161 6.00E-10 7.3 0.5 9.86E-10 113.8 1.2 3.90E + 06 6.43E + 04 95.8 2.2 Old Caldera Floor
E97-135Pc 55 1.64E-09 8.5 0.3 3.41E-10 18.4 2.3 1.26E + 05 2.93E + 04 3.4 0.8 Caldera Rim
E97-135c 55 1.37E-09 8.5 0.1 4.17E-10 10.2 0.8 2.78E + 04 1.21E + 04 0.8 0.3 Caldera Rim
a
All measurements on olivine mineral separates. 3He sea level production rate is assumed to be 72 atoms/g/yr [Kurz and Geist, 1999].
Calculations use altitude scaling of Lal [1991]. Uncertainties are 1 sigma analytical error only (see text).
b
Denotes sample from Kurz and Geist [2000].
c
Denotes caldera wall samples, which are assumed to be perfectly shielded over half the horizon (half the production rate).

pure water. Approximately 250 mg of powdered
material was digested in a closed PFA Teflon
container with 15 mL HNO3 and 3 mL HF for a
minimum of 24 hours [e.g., Harpp and White,
2001; Longerich, 1993]. The solution was evapo-
rated to dryness, and the solid residue dissolved in
2.5 mL HNO3, transferred to 250 mL polyethylene
bottles, and diluted with water. The final solution is
a thousand-fold dilution of the original sample in a

1% HNO3 solution. Between 5 and 8 replicate
analyses were carried out on each sample. Meas-
urements were made using a nonlinear drift correc-
tion procedure and external calibration against
matrix-matched standards as described by Cheat-
ham et al. [1993]. Additional signal noise was
corrected using an on-line internal standard correc-
tion (1:20 dilution of a 1 ppm 133Cs solution) [e.g.,
Eggins et al., 1997]. Contributions from polybaric
oxide and doubly-charged interferences were con-
sistently below 1%.
[37] Sr and Nd isotope ratios were analyzed at
Cornell University using a FISONS Sector-54
thermal ionization-mass spectrometer (Table 4).
Analytical and separation methods are the same
as reported in White et al. [1993]. All samples were
leached in HCl before separation to remove sea salt
and other contaminants. Hf isotope ratios were
determined by multicollector inductively coupled
plasma mass spectrometry (MC-ICP-MS) using a
VG Plasma 54 at ENS Lyon [Blichert-Toft and
Albarede, 1997]. Pb isotope ratios were analyzed
using two methods. The first was conventional
thermal ionization mass spectrometry using the
Cornell Sector 54 instrument and the same methods
described in White et al. [1993]. The second was
MC-ICP-MS using the VG Plasma 54 at ENS
Lyon. The method used, and a comparison between
MC-ICP-MS and thermal ionization results, are
reported in detail in White et al. [2000]. To avoid
contamination during grinding, only rock chips
were used in Pb analysis.
4.2. Major Element Compositions
[38] Ecuador basalts are transitional between alka-
line and tholeiitic, lying close to the Macdonald-
Katsura line on an alkali-silica diagram with
small amounts of normative nepheline or hyper-
sthene (Figure 5). Compared to lavas from nearby
Volcan Wolf [Geist et al., 2000], they are richer
in K and Na and all of the incompatible trace
elements.
[39] Volcan Ecuador basalts have a wide range in
MgO, from 19.0 to 4.8%, in contrast to Fernandina,
Wolf, and Sierra Negra volcanoes, each of which
has a very narrow range of MgO [Allan and Simkin,
2000; Geist et al., 2000; Reynolds and Geist,
1995]. Probably the most outstanding geochemical
attribute of Ecuador lavas is that they are very rich
in Al2O3 (Figure 6). All lavas with Al2O3 > 16%
have abundant plagioclase phenocrysts and a strong
correlation between Al2O3 and Sr/Y, best explained
by accumulation of plagioclase into basaltic liquids
with 14% < Al2O3 < 16%. Because many Ecuador
lavas have been strongly affected by plagioclase
accumulation, the molar MgO/(MgO + FeO*) ratio
(Mg#; expressed in percent) is used as a differ-
entiation index. Also, several of the diagrams (e.g.,
14 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

5 I 0.8 ● ●
.5 kb ●
Na2O + K2O I ●●
● ● Ca/Al
●● I● ● ●
4.5 L ●
▲● 0.75 ●❍
L
L ●
I ● ● ● ● ●■■● I I ■
L▲▲ ●I ▲ ●▼ ● ●●●●●▲ ▲■●▲ I IILILI ■■
2 kb LLL III▲ ▲■
▲
●
▲
●●
▲ ●
▲ ●▲▼▲ ▼ LI ■
L
●▲▲ ● LILI
LI
LI
LILI
LI
LI
LI
LILILI
LILI
LI
I
L
L I❍ ▼▲ ▼ I ■
4 LL I
L ❍●
I ▼ ▲
●●●
● ▼● 0.7 ❍ ●▼ ▼●
▲●
●
●▲ I
▲
I ■ ■
● ■● ●● ●●▲ ▲●I
▲ ▼
■I
▲
●
L I
I● ▲● ■▼▼ ▲● ▼ ▼
L
LL I I❍●▼ ▼▲
■ ▼▲● ▼▲I ▲
● I●I▼▲
▲
▲
●▼ ▲ ▲
▼ ●
■■ ■ LL I▲ I
I● ▼ ▲ ■
■ ●
L ● ● ●● ▼ ▼
3.5 I 0.65 I▲ ●● ❍ ■ ■
▲●●●
L
■ I I ● II
L ● ● ■■ ●
● ■ III■■
IL ● I
I
I
ILI
LI
LI
IL
L
L
I
L
I ●
L ●
■■
IL
IL
IL
IL L
IL ● LL ▲
3 0.6 I .5 kb
LL ▲
■■ ● Old Shield I LL
I LL
❍ Caldera Floor LL
2.5 ■ E Rift 0.55 LL 2 kb
● LL
▲ Apron
▼ Intracaldera
2 0.5
45 46 47 48 49 50 51 52 40 45 50 55 60 65 70 75
SiO2 Mg#
1 3.8 ● ●
I
.5 kb LI
I LI
I ●● ● 3.6 ❍LI●
K2O LI
● I●
L● ▲▲ ▲ TiO2
0.9 I 3.4 I LI▲ ▼●●▲ ❍
● ●L▲ I●▲
▲▲ ▼
■
●
L
IL● ● ● ▼▼LI▲▼
▲ ▲
▲
▲ ●● ●
▼ I
IL ▲ 3.2 I ●
❍ ▼▼
L
●LI▼I
IL ●
● ▲ ▼ ● ●●●
0.8 ●● ●L I
❍ILIL ▼▲ ▲▲▲▼ I ●●●●
●▲▼ ■ ●▲
3 L●I
▲■
▲ ▲ ▼■●L ▲
IL●▼
● IL ▼●▲ ● .5 kb ●▲▼ ▲ I
IL▲
●●
●●▼● ❍ 2.8
■L▼I
0.7 ●LLI
❍I●
LI▲■
● ●● ▲ ● ▲
L■
LI▼ ● ● ▲ L II■
●● ▼ ILI■● 2.6 ● ■ ■ ■
● ●▲ ● L ■
■■ ●●
●ILILI
L■
ILI
LI
LI L●
LILI ▲■
0.6 ▼ ▼ ILIL ■ ●
I
LI
LI
LI
LI
LILI
LI
■▲
LI■
LI
LI
L ■
● ● I
LLI 2.4 ● ■ ■
●
●
● ● ▲▼
● ● ▲ I●■ ●
II■ ■■
LILI
LILI
LI
LI
LILI 2.2 ● ■
●
LI
LI
LI
LILIL▲
LI ■
0.5 ● I
LILI
LI
LI
L
■ ▲ ● 2
1.8 ●
0.4 ■■ ●
40 45 50 55 60 65 70 75 40 45 50 55 60 65 70 75

Mg# Mg#
Figure 5. Major element variation of Ecuador lavas; units are in weight %. Lavas that are thought to have large
amounts of accumulated plagioclase (Al2O3 > 16%) are excluded. Modeled liquid lines of descent at those pressures
of 2 kb and 0.5 kb are labeled and were calculated using the Melts computer program [Ghiorso and Sack, 1995].

Figure 5) have been filtered to remove lavas with
Al2O3 > 16%, so as to assess processes other than
plagioclase accumulation involved in the petrogen-
esis of Ecuador magmas.
[40] The majority of Ecuador’s lavas exhibit a
limited range in Mg#, between 42 and 54 (Figure
5); fewer than 20% are more primitive than this.
Following Albarede et al.’s [1997] classification
scheme for Reunion lavas, we refer to lavas with 42
< Mg# < 54 as the ‘‘steady state’’ series and those
with Mg# > 54 as the ‘‘primitive’’ series.
4.3. Petrography and Mineral Chemistry
[41] Most Ecuador basalts are plagioclase-phyric,
with up to 20% plagioclase phenocrysts up to 4 mm
in length. Olivine phenocrysts, many with chro-
mium spinel inclusions, occur in the basalts with
high MgO, but most lavas have sparse or no olivine
15 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Figure 6. Many of Ecuador’s lavas are much richer in Al2O3 than could be generated by fractionation (liquid line of
descent models are shown at 0.5 and 2 kb); the aluminum-rich lavas contain up to 18% plagioclase phenocrysts. The
chemical effects of accumulation of 10% and 15% plagioclase are shown by the arrows.

phenocrysts. Many samples contain small augite
phenocrysts.
[42] Little relation exists between the compositions
of the cores of plagioclase phenocrysts and the
whole rock compositions (Figure 7). The most
common type of zoning profile is one of sharp
irregularities superimposed on a relatively constant
core composition (Figure 8, top) or broad normal
zoning (Figure 8, bottom). The irregularities tend to
be several tens of microns thick and with magni-
tudes >5 mole% An, which are likely due to
changes in the growth environment (e.g., different
compositions of liquids) rather than boundary layer
diffusional effects [Pearce and Kolisnik, 1990].
[43] All analyzed olivine grains are normally zoned
(Figure 9). In lavas with Mg#  60, olivine cores
have compositions that are in equilibrium with
liquids whose compositions are the same as that
of the bulk rock. The more primitive lavas contain
olivine that is richer in iron than would be in
equilibrium with liquids with the bulk rock com-
positions.
4.4. Isotopic and Trace Element
Geochemistry
[44] Isotope ratios of Volcan Ecuador basalts
show significant variation, well beyond analytical
error (Table 4; Figure 10). Nevertheless, the range
for all isotopes is much less than that for the
archipelago as a whole. One sample from the East
Rift, E95-13, is anomalous, plotting well outside
fields defined by all other Ecuador samples and
within the fields of lavas from Volcan Wolf. This
is an example of an ‘‘illegitimate magma’’ pro-
duced by the intrusion of Wolf magma into the
Volcan Ecuador system [Geist et al., 1999], and it
excluded from the discussion. With this excep-
tion, Ecuador basalts define isotopic arrays cen-
16 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Figure 7. Plagioclase composition (expressed as XAn) compared to bulk-rock compositions, which are indicated by
Mg# and Ca# (molar CaO/CaO + Na2O). The compositions of the phenocrysts do not bear any meaningful
relationship to the major element compositions of the host lavas.

Figure 8. Microprobe traverses of two representative Ecuador plagioclase phenocrysts (core is to the left, rim to the
right). Each phenocryst is approximately 3 mm long and 1 mm wide.
17 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Figure 9. Forsterite concentrations of olivine phenocrysts against Mg# of the bulk rock. Equilibrium composition
assumes liquid-olivine Kd (MgO/FeO) of 0.3 and FeO = 0.9*FeO*.

tered within a triangular region with apices
defined by Wolf, Sierra Negra, and Pinta isotopic
compositions.
[45] Small differences in isotopic composition
relate to the stratigraphic units defined earlier: the
older basalts, those from the caldera rim series and
the old caldera floor tend to have more depleted
isotopic signatures and plot closer to the Wolf and
MORB arrays than do the recent basalts (East Rift,
northeast apron, and postsector collapse intracal-
dera series). There appears to be little, if any,
difference in the isotopic compositions of contem-
poraneous groups (e.g., East Rift and northeast
apron) (Figure 10).
[46] All lavas are light rare Earth-enriched and
have nearly parallel chondrite normalized rare
Earth patterns (Figure 11); absolute concentrations
of the rare Earth elements range about two-fold.
Variations in incompatible element concentrations
are also limited, with most having about a two-fold
variation. Incompatible element ratios are typical of
oceanic basalts, e.g., Nb/Ta averages 17.27 ± 7%,
Zr/Hf averages 41.5 ± 3%, Nb/U averages 44.7 ±
10%, Pb/Ce averages 0.034 ± 11%, Th/U averages
3.30 ± 5%, Ba/Nb averages 5.98 ± 10%, and Nb/Zr
averages 0.127 ± 12%. The East Rift lavas tend to
be slightly more light rare Earth enriched (Figure
11) and have higher Nb/Zr at a given Nb concen-
tration (Figure 12).
4.5. Long-Term Compositional Evolution
[47] The long-term (
100,000 yr) evolution of
Volcan Ecuador can be evaluated by comparing
the compositions of the major stratigraphic units
(Figures 5, 6, 10, and 12). The few consistent
differences between the different stratigraphic
units may be summarized as follows: (a) a. Most
18 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Figure 10. Correlation between isotopic ratios of Volcan Ecuador lavas. Also shown are fields of Wolf, Floreana,
Sierra Negra, and Pinta isotopic values, which are close to the extremes of Galápagos compositions [Harpp and
White, 2001; Blichert-Toft and White, 2001].

Figure 11. Chondrite-normalized rare Earth patterns of Volcan Ecuador lavas. Shaded area shows the range of
patterns in 11 lavas from the caldera-rim series.
19 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Figure 12. Trace element variation of Ecuador lavas, grouped by eruptive unit. Lavas that are thought to have
accumulated plagioclase are not included.

of the ‘‘primitive’’ lavas (Mg# > 55) are from the
East Rift and north apron; these also tend to have
lower Al2O3 (Figure 6). The vast majority of the
caldera wall, caldera floor, and younger intracal-
dera lavas are part of the steady state series and
have exceptionally high Al2O3 (Figure 5). (b) The
incompatible elements (e.g., Ba, Zr, Nb, Y, Ti,
REE, and K) form predictably negative correla-
tions with Mg# (Figures 5 and 12). Seven of the
caldera wall samples have unusually low Ti and K
contents for their Mg# (Figure 5). (c) Among the
younger lavas, there is a systematic difference in
Nb/Zr ratios, with the East Rift lavas having a
higher value than the north apron and intracaldera
lavas (Figure 12). Nb/Zr ratios of the older caldera
wall and caldera floor lavas span the entire range
of the younger ones. (d) The East Rift lavas
(along with a few samples from the other units)
20 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems
display strong positive correlation between Sc/Y
(a measure of augite control) and Mg#. Most of
the lavas from the presector collapse edifice have
a large range in Sc/Y with no significant relation-
ship to Mg# (Figure 12). (e) On average, the lavas
of the older edifice have isotopic ratios that
suggest their sources were richer in a Wolf
(depleted) component (Figure 10) than postcol-
lapse lavas. Those of the East Rift are richest in
the Sierra Negra component, even though the East
Rift bounds Volcan Wolf, and at least one Wolf
dike has intruded it [Geist et al., 1999].
4.6. Short-Term Compositional Evolution
[48] The bulk of the volume of Volcan Ecuador is
made up of the old shield, and this unit is well
exposed by both the caldera wall and outward-
facing cliffs. As discussed above, this sequence
erupted over a limited time span, perhaps on the
order of 50,000 years, with no observed discon-
formities. The variations of several compositional
parameters with stratigraphic height are indicated
in Figures 13–17.
[49] The most notable feature of the caldera wall
section is the remarkably small range in Mg# (from
43 to 55); without exception, the lavas are of the
‘‘steady state’’ variety (Figure 13). The two sec-
tions from the eroded outer flanks contain some
higher-Mg# lavas, suggesting that different reser-
voirs fed at least some of these flows.
[50] K and Nb concentrations increase broadly with
stratigraphic height in the caldera wall section, but
the broad trend is superimposed on erratic flow-by-
flow variations that are greater than two-fold (Fig-
ures 14 and 15). K and Nb from the first outer flank
section overlap with the middle and upper parts of
the caldera wall section. In contrast, K and Nb
concentrations of the second outer flank section
start at the high end of the caldera wall section and
extend to higher values. The broad increase in
incompatible elements in the caldera wall section
is not accompanied by a systematic change in trace
element ratios, such as Nb/Zr and La/Sm (Figure
14), which have erratic, high frequency fluctua-
tions. In contrast, the Sc/Y ratio decreases with
stratigraphic height (Figure 16), and, in fact, there
´ PAGOS
is strong correlation between Sc/Y and [Nb] in the
caldera wall section (Figure 16).
[51] Isotopic variations in the stratigraphic sections
are also irregular. With the exception of the bottom-
most sample, the caldera wall section shows an
apparent decrease in 87Sr/86Sr with time, opposite
the net long-term trend discussed above (Figure 17).
The second outer flank section shows an apparent
increase in 87Sr/86Sr. If this section is younger than
the caldera wall section, then the isotopic evolution
is not unidirectional, but cyclic. The other isotopic
systems show similar changes, but their magnitude
is smaller, only slightly beyond analytical precision.
5. Discussion
[52] The parallel nature of the rare Earth patterns,
consistency of mass balance models to the observed
compositional variations (Figure 5), and the absence
of significant correlations between elemental con-
centrations and isotopic ratios suggests most of the
compositional variation of Volcan Ecuador magmas
is due to fractional crystallization. In addition to
simple removal of crystals from a cooling magma,
there is much evidence for crystal accumulation in
previously-evolved magmas.
5.1. Petrologic Evolution During Shield
Growth
5.1.1. Plagioclase Accumulation
[53] The most important petrologic process at Vol-
can Ecuador is the accumulation of plagioclase,
especially in lavas from the presector collapse
shield. Evidence for plagioclase accumulation
includes: (a) Many of the lavas have abundant
plagioclase phenocrysts, up to 18.5% by volume.
(b) Many of the shield lavas are much richer in
Al2O3 than is likely to be produced by crystallization
at reasonable pressures (Figure 6). (c) The compo-
sitions of Al-rich lavas are effectively modeled by
the addition of up to 18% plagioclase to liquids
having the compositions of the lavas with Al2O3 <
16% (Figure 6). (d) In lavas with Al2O3 > 16%, there
is strong correlation between Al and Sr/Y. Because
Sr is compatible in plagioclase but y is not, the
correlation can be produced by the accumulation of
21 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Figure 13. Variation of Mg# with stratigraphic sequence from sections collected in the caldera wall and two cliffs in
the outer flank. Dashed lines show limit of ‘‘steady state’’ compositions.

plagioclase but is unlikely to be produced by frac-
tional crystallization (many of the high-Al and high
Sr/Y lavas have relatively low Mg#). (e) There is no
relation between the compositions of plagioclase
cores and the bulk-rock composition (Figure 7).
All plagioclase-phyric lavas contain plagioclase
with a range in compositions and erratic zoning
patterns (Figure 8).
[54] These observations are all consistent with the
idea that plagioclase contained in many of the lavas
grew from discrete batches of magma, which had
different temperatures and compositions. These
plagioclase crystals then accumulated by flotation
(or at least neutral buoyancy) in the shallowest part
of the magmatic plumbing system, likely a sub-
caldera magma chamber. The density of an anhy-
drous parental Ecuador magma (Mg# = 61; T =
1163
C; P = 100 b) is calculated to be 2.69 gm/
cm3 [Lange and Carmichael, 1990], and the den-
sity of An85 at the same temperature is 2.65 gm/
cm3. After cooling 20
and the consequent crystal-
lization, the liquid’s density increases to 2.70 gm/
cm3, enhancing flotation. The concentration of
water in Galápagos magmas is not known, but a
reasonable estimate would be 0.5 to 1%. This
amount of water would lower the liquid’s density
by 0.03 to 0.06 gm/cm3 [Ochs and Lange, 1999].
Thus, the difference in density between plagioclase
and liquid in these magmas is minute.
5.1.2. Olivine Accumulation
[55] Because the olivine phenocrysts are poorer in
forsterite than would be in equilibrium with bulk-
rock compositions, lavas with Mg# > 60 are likely to
be accumulative, with Fo83 – 85 entrained in liquids
with Mg# of 60–62. The linear trend between Ni and
22 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Figure 14. Variation of K2O with stratigraphic sequence; vertical line indicates K2O = 0.75% for reference. Note
the broad increase in the caldera wall section.

MgO (Figure 12) is also evidence that the lavas with
Mg# > 62 formed by olivine accumulation [Hart and
Davis, 1978]. Most of the olivine-accumulative
lavas are from the East Rift, but two are from the
outer flanks of the presector collapse shield.
5.1.3. Thermal Regulation
[56] So long as olivine and pyroxene are the only
mafic minerals crystallizing, Mg# decreases with
decreasing temperature. The remarkable uniformity
of the Mg# of the lavas exposed in the caldera wall
(between 42 < Mg# < 55) corresponds to a temper-
ature variation of only 45
C (1160
to 1115
C), as
calculated using the Melts algorithm [Ghiorso and
Sack, 1995]. The small amount of temperature
variation and lack of systematic compositional
trends suggest that most Ecuador magmas have
passed through a well-regulated thermal environ-
ment prior to eruption, a process that has already
been noted at Sierra Negra [Reynolds and Geist,
1995], Fernandina [Allan and Simkin, 2000], and
Wolf volcanoes [Geist et al., 2000]. We suggest
that the subcaldera magma chambers at these
volcanoes are in a thermal and chemical steady
state, whereby heat loss and material transfer by
crystallization and eruption are balanced by input.
5.1.4. Dominant Control is by Augite
Fractionation
[57] The general increase in incompatible element
contents (Figures 14 and 15), the decrease in Sc/Y
(Figure 16), and correspondence between Sc/Y and
incompatible elements in the caldera wall section
(Figure 16) are best accounted for by clinopyrox-
23 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Figure 15. Variation of incompatible trace element Nb and Nb/Zr ratio with stratigraphic height. Dashed line is for
reference only.

ene fractionation that generally increased with time.
Trace element modeling suggests further that the
shield-forming magmas were not likely to be multi-
ply-saturated during the major episode of crystal-
lization: the covariation of Sc/Y and incompatible
elements requires a clinopyroxene-rich assemblage
to have fractionated (Figure 16) and precludes low
pressure gabbroic fractionation. A calculated P-T
phase diagram (using thermodynamic calculations
in Melts [Ghiorso and Sack, 1995]) of a relatively
primitive aphyric East Rift lava (Mg# = 61) shows
that plagiocase is the liquidus phase at low pressure
(Figure 18). Clinopyroxene replaces plagioclase as
the liquidus phase at about 0.2 GPa and progres-
sively dominates the crystallizing assemblage with
increasing pressure.
[58] There are two possibilities to explain the
increasing clinopyroxene fractionation with time:
either the magmas simply cooled and crystallized to
24 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Figure 16. (Top) Sc/Y ratios decrease with stratigraphic height. (Bottom) Sc/Y displays an excellent negative
correlation with Nb in lavas from the caldera wall. Two model fractionation paths are displayed, one for an
assemblage of pure clinopyroxene and the other for one with 30% clinopyroxene, 60% plagioclase, and 10% olivine.
Distribution coefficients are from Hart and Dunn [1993].

greater extents as the volcano evolved, or the
magmas crystallized at progressively greater depths
with time, with a consequent increase in the stability
of augite (Figure 18). We prefer the second explan-
ation, because the caldera wall lavas show a poor
correspondence between Mg# and Sc/Y (Figure
12), and there is no apparent stratigraphic trend to
lower temperature as judged from Mg# (Figure 13).
These considerations and the phase diagram suggest
that the level at which Ecuador magmas cooled and
fractionated increased from about 6 km to greater
depths as the shield grew to its maximum height.
Estimates of changing phase equilibria with depth
led Geist et al. [1998] to conclude that the East Rift
magmas underwent fractional crystallization at 0.3
GPa.
[59] The low Mg# and Ni contents of all of the
nonaccumulative lavas indicate that substantial
olivine fractionation has also taken place since
25 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

Figure 17. Variation of 87Sr/86Sr with stratigraphic height through the caldera wall section and the interpreted
stratigraphic positions of the two outer flank sections.

derivation in the mantle. It is not possible to say
whether fractionation occurred during the higher-
pressure, augite-dominated crystallization regime
or the lower-pressure plagioclase-dominated sys-
tem, or both. Given the uncertainty of the possible
compositions of the primary magmas, it is impos-
sible to estimate the amount of olivine crystalliza-
tion that might have occurred.
5.2. Residence Time and Magma
Chamber Size
[60] The high frequency fluctuations of incompat-
ible trace element ratios such as Nb/Zr (Figure 15)
with stratigraphic height provide evidence for the
size of the magma body beneath Volcan Ecuador’s
caldera. The volume of a magma reservoir can be
estimated by the fluctuations in the magmas’ com-
positions and eruptive rates through the use of
residence time analysis [Albarede, 1993; Pietruszka
and Garcia, 1999b]. The residence time is esti-
mated by the formula for isotopic or incompatible
trace element ratios [Albarede, 1993]:
C2in Rin  Rres
tres ¼ ; ð1Þ
C2res dR=dt

26 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems
Figure 18. Synthetic P-T phase diagram constructed
using Melts [Ghiorso and Sack, 1995]. Composition
used is an East Rift specimen with Mg# = 61.
where tres is the magma residence time in the
magma chamber, C2 is the concentration of the
element in the denominator of the elemental ratio,
superscripts res and in refer to resident and input
magma, R is the isotopic or trace element ratio, and
dR/dt is the rate of change of that ratio. Our
calculations use the Nb/Zr ratio, and note that some
of the largest fluctuations are between adjacent
samples (Figure 15). If we assume that these
fluctuations are due to mixing between the most
extreme compositions at Volcan Ecuador, the left
hand ratio is roughly 2, dt becomes the interval
between samples, and the numerator and dR cancel.
Our best estimate for the interval between samples
is about 1000 years (50,000 yr/50 sampled flows).
Thus the residence time is on the order of 2000
years. This is a maximum estimate; less extreme
contrasts between the intruding and resident
magmas would lower this estimate to 1000 years.
[61] The volume of the magma reservoir can be
calculated as the product of the residence time and
the supply rate (assumed to be equivalent to the
eruption rate). The volume of the presector collapse
shield is about 25 km3, thus the eruption rate on the
order of 5  105 m3/yr. This estimate is well within
the range of those of the other western shields that
have been studied in detail (as compiled by Nau-
´ PAGOS
mann and Geist [2000]). With a liberal estimate of
the residence time of magma in the chamber of
2000 yr, the chamber size is on order of 0.1 km3. If
the magma reservoir is a sill whose diameter was
that of the presector collapse shield caldera floor
[Amelung et al., 2000], then it could be no more
than 120 m thick. This thickness is strikingly
similar to the magma ‘‘lens’’ thickness at 9
N on
the East Pacific Rise estimated by seismic imaging
[Forsyth et al., 1998].
[62] The characteristic cooling time for a hot body
is:
t ¼ l2=k ; ð2Þ
where l is the length scale and k is the specific
thermal conductivity of rock (taken to be 106 m2/s
here). For a 100 m thick sill, the cooling time is on
order of 300 years. This appears to be discrepant by
over 100-fold with the estimated duration of the
shield building phase of Volcan Ecuador. It is also
at odds with the observation that the Mg# (a strong
function of temperature) in these magmas was
buffered during shield growth. In other words,
although Volcan Ecuador’s magma body appears to
be in a thermal steady state, it is in a constantly
fluctuating chemical one.
[63] One possible resolution of the discrepancy is
that the 100 m thick sill overlies a much larger
magmatic system, so the length scale is much larger
than 100 m. It was documented above that the lavas
of the old shield evolved by extensive olivine and
clinopyroxene crystallization in the deep crust and
upper mantle before undergoing plagioclase-domi-
nated crystallization in the shallow sill. Because we
have no idea of the composition of the primary
magmas, the extent of the deep cumulate layer is
difficult to estimate. For the cooling time to be
increased 100-fold, the thickness of the magma
body must be 10 times greater, and >90% crystal-
lization from the primary magma to create the
erupted basalts is surely excessive.
[64] Another (but not exclusive) resolution is that
in all of the calculations above, we assume that all
magma is erupted, and surely there must be a
considerable intrusive complex of unerupted
magma in addition to the cumulates that comple-
27 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems
ment the erupted lavas. One possibility is that there
was a plexus of sills beneath the caldera that were
not in chemical communication but whose heat
flow buffered the interior of the volcano to roughly
1150
. In any event, these calculations suggest that
only about 10% of all magma is erupted, a finding
consistent with crustal thickening on order of 4 to 8
km over the entire Galápagos Platform [Feighner
and Richards, 1994].
5.3. Mantle Sources
[65] Isotopic variations such as those depicted in
Figure 10 are almost certainly due to mixing either
within the mantle or between melts produced from
different parts of the mantle. In the Galápagos,
isotopic variation has been attributed to mixing
between a heterogeneous mantle plume and the
asthenosphere entrained by it [Geist et al., 1988;
White et al., 1993; Kurz and Geist, 2000; Harpp
and White, 2001; Blichert-Toft and White, 2001]. If
this interpretation is correct, the Volcan Ecuador
source is a mixture of 30% depleted asthenosphere
(e.g., the Wolf source) and 70% mantle plume (e.g.,
Sierra Negra source), with no appreciable differ-
ences in ratios between Sr, Nd, Pb, and Hf in the
two sources. These proportions have changed only
slightly through the 100,000 years of Volcan Ecua-
dor history that we have been able to sample.
[66] Nevertheless, subtle variations in isotopic
ratios over this period do occur. Figure 17 shows
that the Sr isotopic ratios of the caldera wall lavas
decline steadily, punctuated by sharp increases at
the base and top of the section. Hf, Nd, and Pb
isotope ratios show sympathetic, but more erratic
variations. After the sector collapse, 87Sr/86Sr
remained high, and 143Nd/144Nd and 176Hf/177Hf
remain low. Consistent with this, many of the
younger lavas are richer in incompatible elements
(e.g., Ti, K, Zr, Nb; Figures 5 and 12) than the older
shield lavas.
[67] In the past 100,000 years, plate motion has
carried Volcan Ecuador roughly 4 km to the east
relative to the hot spot, that is, toward the mantle
source supplying Wolf Volcano. On the basis of the
geographic pattern of isotopic variation in the
Galápagos [Geist et al., 1988; White et al., 1993]
´ PAGOS
and the more incompatible element-depleted char-
acter of Wolf lavas, we would have predicted that
the source of Volcan Ecuador lavas would become
more isotopically- and incompatible element-
depleted with time. However, the net change is to
a more incompatible element- and isotopically-
enriched source. There are a number of possibilities
for this inconsistency. The simplest one is that that
small-scale heterogeneity is superimposed on the
more regular large-scale geographic pattern of iso-
topic variation. If this is correct, the temporal
isotopic variations provide evidence of small-scale
(<3 km) spatial heterogeneity in the mantle beneath
the Galápagos. If mixing occurs between melts
rather than sources, small variations in simple
physical factors controlling melt migration might
cause random variations in the proportions of each
end-member in a mixed melt. For example, different
magma ascent paths could produce short-term tem-
poral variations by varying the extent of reaction
between an ‘‘enriched’’ magma and ‘‘depleted’’
wallrock (or visa versa) [Reiners, 1998]. Conceiv-
ably, the variation might be cyclical, with the trend
of the cycles starting with relatively enriched com-
positions and steadily decaying to a more depleted
one. This is consistent with initial melting of
enriched material with a slightly lower solidus that
is steadily diluted with melt from more depleted
rock with a slightly higher solidus [e.g., Morgan,
1999]. At present, there is simply not enough
evidence to decide between these possibilities.
5.4. Volcanologic and Petrologic Effects of
Sector Collapse
[68] One might expect that the catastrophe of half
of the volcano sliding away might have an effect on
the volcanic plumbing system. The most obvious
one is that it changed the locus of eruptive vents.
Prior to sector collapse, most eruptions were from
circumferential fissures along the caldera rim and
from the caldera floor. After sector collapse, most
eruptions were from the former caldera floor and
the north apron, and the East Rift formed. Only a
few eruptions took place from high on the old
shield after the sector collapse.
[69] A fundamental difference between the shield
lavas and the younger lavas is that there is a
28 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems
disproportionate number of younger lavas with
high Mg#, whereas the shield lavas are all of the
steady state variety (Figure 5). Our interpretation is
that prior to sector collapse, almost all of the
magmas passed through a long-lived, persistent
magma chamber that was in a thermal and compo-
sitional steady state, and any accumulated or early-
crystallized olivine settled out during late-stage
residence in the shallow magma body. The sector
collapse either perturbed that magma chamber, or it
altered the stress regime such that some magmas
began to bypass the reservoir, and they erupted
carrying olivine that crystallized at higher temper-
ature, or olivine that was picked up from the thick
cumulate pile.
[70] Another aspect of these younger lavas is that
some of them do not lie on the trends formed by the
shield building lavas on diagrams of Ca/Al and Sc/
Yversus. Mg# (Figures 5 and 12): they are dis-
placed to higher Mg#. The most likely explanation
for this is that the older caldera wall and younger
East Rift magmas may take different P-T paths as
they ascend through the lithosphere. The caldera
wall magmas cooled and fractionated at progres-
sively greater depths, with increasing amounts of
clinopyroxene removal. In contrast, some of the
younger lavas cooled and crystallized at shallow
levels, with olivine + plagioclase fractionation,
while others cooled and crystallized at deeper
levels where augite played an important role.
[71] The subtle changes in the lavas’ compositions
between the presector collapse shield and the
younger lavas are best developed in the East Rift
lavas. These have the most enriched isotopic com-
positions, highest ratios of Nb/Zr and La/Sm, and
the highest Mg#. Also, the East Rift lavas are not as
strongly affected by plagioclase accumulation as
the older shield lavas and the younger intracaldera
and north apron lavas, but most of the olivine-
accumulative are from East Rift (Figure 5). Most of
the north apron and younger intracaldera lavas are
not substantially different from the lavas of the old
shield. Thus, instead of simply destroying the
magmatic plumbing system, the sector collapse
may have simply changed the magmas’ litho-
spheric pathways. Prior to sector collapse, all
magmas passed through the subcaldera magma
´ PAGOS
reservoir, where they attained the steady state
magma composition and were strongly affected
by plagioclase accumulation. More recently, some
of the magmas, particularly those that are generated
in a slightly enriched part of the mantle to the east
of the old caldera, ascend to the surface without
extensive cooling, crystallizing, and homogenizing
through mixing en route.
6. Conclusions
[72] The older phase of volcanism at Volcan Ecua-
dor built a Galápagos-type shield volcano, with
steep flanks and a large caldera. The magmas
making up the shield underwent two stages of
cooling and crystallization. In the lower crust and
upper mantle, clinopyroxene fractionation was
dominant. The magmas then resided in a shallow
subcaldera magma chamber, which is likely to be a
sill on the order of 100 m thick. In the subcaldera
sill, plagioclase and olivine crystallized along with
clinopyroxene, and many of the lavas have accu-
mulated plagioclase owing to its buoyancy in the
shallow magma chamber.
[73] A catastrophic sector collapse bisected the
volcano, which disrupted the magmatic plumbing
system and changed the focus of the eruptions from
the summit to the former caldera floor, a northern
apron, and the East Rift. The lavas of the younger
phase are compositionally more diverse than the
older ones, reflecting less efficient thermal and
chemical buffering in a shallow magma chamber.
[74] Subtle variations in isotopic composition pro-
vide evidence for variation in source composition
over last
100,000 years. The net change has been
toward a more incompatible element-depleted
source, which is opposite to the change that would
be predicted from plate motion and regional iso-
topic patterns. As shorter term isotopic variations
are nearly as great as the total range of variation,
this change is most likely due to small-scale (4
km) mantle heterogeneity, or to variation in the
proportion of melt derived from plume and asthe-
nospheric components.
[75] The most important implication of this study is
that the extent of systematic compositional evolu-
29 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems
tion at Volcan Ecuador is small, and the question of
whether any of the Galápagos volcanoes evolves in
a style similar to the Hawaiian example remains
doubtful. Detailed study of most of the volcanoes is
now complete and has yet to reveal such evolution,
thus suggesting that the compositions of magmas,
and hence the nature of the source and the thermal
evolution of the magma during ascent, do not
substantially change as the volcano grows and
eventually goes extinct.
Acknowledgments
[76] This work was supported by NSF grant EAR9612110 to
D. Geist and W. M. White and CHE-9996141 and CHE-
9996136 to K. Harpp. J. Blichert-Toft and F. Albarede
acknowledge financial support from the French Institut
National des Sciences de l’Univers. M. Kurz’s Galápagos
work is supported by NSF grant EAR 0126097. Field work
in the Galápagos would be virtually impossible if it were not
for the support and permission of the Charles Darwin Research
Station and the Galápagos National Park Service, whom we
thank. We thank Christian Germanaz and Jacques Bertola for
their mountaineering skills in obtaining samples from section
3, and them and Eduardo Villema and Darwin Villagomez for
help in the field and valued camaraderie. Bruce Nelson
reviewed an early draft, and we thank him for his constructive
comments. Thanks to Bill Chadwick and Jamie Allan for
thorough journal reviews.
References
Albarede, F., Residence time analysis of geochemical fluctua-
tions in volcanic series, Geochim. Cosmochim. Acta, 57,
615 – 621, 1993.
Albarede, F., B. Luais, G. Fitton, M. Semet, E. Kaminski, G. J.
Upton, P. Bachelery, and J. L. Cheinee, The geochemical
regimes of Piton de la Fournaise volcano (Réunion) during
the last 530,000 years, J. Petrol., 38, 171 – 201, 1997.
Allan, J. F., and T. Simkin, Fernandina Volcano’s evolved,
well-mixed basalts: Mineralogical and petrological con-
straints on the nature of the Galápagos plume, J. Geophys.
Res., 105, 6017 – 6031, 2000.
Amelung, F., S. Jonsson, H. Zebker, and P. Segall, Widespread
uplift and ‘‘trapdoor’’ faulting on Galápagos volcanoes ob-
served with radar interferometry, Nature, 407, 993 – 998,
2000.
Blichert-Toft, J., and F. Albarede, The Lu-Hf isotope geochem-
istry of chondrites and the evolution of the mantle-crust
system, Earth Planet. Sci. Lett., 148, 243 – 258, 1997.
Blichert-Toft, J., and W. M. White, Hf isotope geochemistry of
the Galápagos Islands, Geochem. Geophys. Geosyst., 2, Pa-
per number 2000GC000138, 2001.
Chadwick, W. W., and J. H. Dieterich, Mechanical model-
ing of circumferential and radial dike intrusion on Galá-
´ PAGOS
pagos volcanoes, J. Volcanol. Geotherm. Res., 66, 37 – 52,
1995.
Chadwick, W. W., and K. A. Howard, The pattern of circum-
ferential and radial eruptive fissures on the volcanoes of
Fernandina and Isabela islands, Galápagos, Bull. Volcanol.,
53, 259 – 275, 1991.
Cheatham, M. M., W. F. Sangrey, and W. M. White, Sources of
error in external calibration ICP-MS analysis of geological
samples and an improved non-linear drift correction, Spec-
trochim Acta (Electronica), 48B, E487 – E506, 1993.
Cullen, A. B., and A. R. McBirney, The volcanic geology and
petrology of Isla Pinta, Galápagos archipelago, Bull. Geol.
Soc. Am., 98, 294 – 301, 1987.
Duffield, W. A., L. Stieltjes, and J. Varet, Huge landslide
blocks in the growth of Piton de la Fournaise, La Reunion,
and Kilauea Volcano, Hawaii, J. Volcanol. Geotherm. Res.,
12, 147 – 160, 1982.
Eggins, S. M., J. D. Woodhead, P. J. Kinsley, G. E. Mortimer,
P. Sylvester, M. T. McCulloch, J. M. Hergt, and M. R.
Handler, A simple method for the precise determination of
>40 trace elements in geological samples by ICPMS using
enriched isotope internal standardisation, Chem. Geol., 134,
311 – 326, 1997.
Feighner, M. A., and M. A. Richards, Lithospheric structures
and compensation mechanisms of the Galápagos Archipela-
go, J. Geophys. Res., 99, 6711 – 6729, 1994.
Fornari, D. J., M. D. Kurz, D. J. Geist, P. D. Johnson, U. G.
Peckman, and D. Scheirer, New perspectives on the structure
and morphology of the submarine flanks of Galápagos vol-
canoes: Fernandina and Isabela, Eos Trans. AGU, 82, Fall
Meet. Suppl., Abstract T41D-06, 2001.
Forsyth, D. W., et al., Imaging the deep seismic structure be-
neath amid-ocean ridge: The MELT experiment, Science,
280, 1215 – 1218, 1998.
Garcia, M. O., R. A. Ho, J. M. Rhodes, and E. W. Wolfe,
Petrologic constraints on rift-zone processes: Results from
episode 1 of the Puu Oo eruption of Kilauea Volcano, Ha-
waii, Bull. Volcanol., 52, 81 – 96, 1989.
Geist, D. J., A. R. McBirney, and R. A. Duncan, Geology of
Santa Fe Island: The oldest Galápagos volcano, J. Volcanol.
Geotherm. Res., 52, 65 – 82, 1985.
Geist, D. J., W. M. White, and A. R. McBirney, Plume asthe-
nosphere mixing beneath the Galápagos Archipeligo, Nat-
ure, 333, 657 – 660, 1988.
Geist, D. J., K. A. Howard, A. M. Jellinek, and S. Rayder,
Volcanic history of Volcán Alcedo, Galápagos Archipelago:
A case study of rhyolitic oceanic volcanism, Bull. Volcanol.,
56, 243 – 260, 1994.
Geist, D. J., K. A. Howard, and P. Larson, The generation of
oceanic rhyolite by crystal fractionation: The basalt-rhyolite
association at Volcan Alcedo, Galápagos Archipelago,
J. Petrol., 36, 965 – 982, 1995.
Geist, D., T. R. Naumann, and P. L. Larson, Evolution of
Galápagos magmas: Mantle and crustal fractionation without
assimilation, J. Petrol., 39, 953 – 971, 1998.
Geist, D., W. White, T. Naumann, and R. Reynolds, Illegiti-
mate magmas of the Galápagos: Insights into mantle mixing
and magma transport, Geology, 27, 1103 – 1106, 1999.
30 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems
Geist, D. J., T. Naumann, J. Standish, M. Kurz, and K. Harpp,
Volcan Wolf’s (Galápagos) caldera and magma chamber,
IAVCEI Abstracts and Addresses, 82, 2000.
Ghiorso, M. S., and R. O. Sack, Chemical mass transfer in
magmatic processes, IV, A revised and internally consistent
thermodynamic model for the interpolation and extrapolation
of liquid-solid equilibria in magmatic systems at elevated
temperatures and pressures, Contrib. Mineral. Petrol., 119,
197 – 212, 1995.
Gripp, A. E., and R. G. Gordon, Current plate velocities rela-
tive to the hotspots incorporating the NUVEL-1 global plate
motion model, Geophys. Res. Lett., 17, 1109 – 1112, 1990.
Harpp, K. S., and W. M. White, Tracing a mantle plume:
Isotopic and trace element variations of Galápagos sea-
mounts, Geochem. Geophys. Geosyst., 2, Paper number
2000GC000137, 2001.
Hart, S. R., and K. E. Davis, Nickel partitioning between
olivine and silicate melt, Earth Planet. Sci. Lett., 40, 203 –
219, 1978.
Hart, S. R., and T. Dunn, Experimental cpx/melt partitioning of
24 trace elements, Contrib. Mineral.Petrol., 113, 1 – 8, 1993.
Hoernle, K., and H.-U. Schmincke, The role of partial melting
in the 15-Ma geochemical evolution of Gran Canaria: A blob
model for the Canary hotspot, J. Petrol., 34, 599 – 626, 1993.
Johnson, D. M., P. R. Hooper, and R. M. Conrey, XRF Ana-
lysis of Rocks and Minerals for Major and Trace Elements
on a Single Low Dilution Li-tetraborate Fused Bead, Adv.
X-ray Anal., 41, 843 – 867, 1999.
Kurz, M. D., In-situ production of cosmogenic terrestrial he-
lium and some applications to geochronology, Geochim.
Cosmochim. Acta, 50, 2855 – 2862, 1986.
Kurz, M. D., and D. Geist, Dynamics of the Galápagos hotspot
from helium isotope geochemistry, Geochim. Cosmochim.
Acta, 63, 4139 – 4156, 1999.
Kurz, M. D., and D. P. Kammer, Isotopic evolution of Mauna
Loa Volcano, Earth Planet. Sci. Lett., 103, 257 – 269, 1991.
Lal, D., Cosmic ray labeling of erosion surfaces: In situ nuclide
production rates and erosion models, Earth Planet Sci. Lett.,
104, 424 – 439, 1991.
Lange, R. A., and I. S. E Carmichael, Thermodynamic proper-
ties of silicate liquids with an emphasis on density, thermal
expansion, and compressibility, in Modern Methods in
Igneous Petrology, vol. 24, Mineralogical Society of
America Reviews in Mineralogy, edited by J. Nichols and
K. Russell, M25-64, Miner. Soc. of Am., Washington, D.C.,
1990.
Longerich, H. P., Oxychlorine ions in inductively coupled
plasma-mass spectrometry: Effect of chlorine speciation as
Cl- and ClO4-, J. Analy. At. Spectrom., 8, 439 – 444, 1993.
Macdonald, G. A., and T. Katsura, Chemical composition of
Hawaiian lavas, J. Petrol., 5, 82 – 133, 1964.
Marti, J., M. A. Hurlimann, G. J. Ablay, and A. Gudmundsson,
Vertical and lateral collapses on Tenerife (Canary Islands)
and other volcanic ocean islands, Geology, 25, 879 – 882,
1997.
McBirney, A. R., and H. Williams, Geology and petrology of
the Galápagos Islands, Geol. Soc. Am. Mem., 118, 197, 1969.
Moore, J. E., D. A. Clague, R. T. Holcomb, P. W. Lipman, W.
´ PAGOS
R. Normark, and M. E. Torresan, Prodigious submarine land-
slides on the Hawaiian Ridge, J. Geophys. Res., 94, 17,465 –
17,484, 1989.
Morgan, J.P., Isotope topology of individual hotspot basalt
arrays: Mixing curves or melt extraction trajectories?,
Geochem., Geophys., Geosysts., 1, Paper number
1999GC000004, 1999.
Mouginis-Mark, P. J., S. K. Rowland, and H. Garbeil, Slopes
of western Galápagos volcanoes from airborne interfero-
metric radar, Geophys. Res. Lett., 23, 3767 – 3770, 1996.
Naumann, T., and D. Geist, Physical volcanology and structur-
al development of Cerro Azul volcano, Isabela island, Galá-
pagos: Implications for the development of Galápagos-type
shield volcanoes, Bull. Volcanol., 61, 497 – 514, 2000.
Ochs, F. A., and R. A. Lange, The density of hydrous mag-
matic liquids, Science, 283, 1314 – 1317, 1999.
Pearce, T. H., and A. M. Kolisnik, Observations of plagioclase
zoning using interference imaging, Earth Sci. Rev., 29, 9 –
26, 1990.
Pietruzska, A. J., and M. O. Garcia, A rapid fluctuation in the
mantle source and melting history of Kilauea volcano in-
ferred from the geochemistry of its historical summit lavas
(1790 – 1982), J. Petrol., 40, 1321 – 1342, 1999a.
Pietruzska, A. J., and M. O. Garcia, The size and shape of
Kilauea volcano’s summit magma storage reservoir: A geo-
chemical probe, Earth Planet. Sci. Lett., 167, 311 – 320,
1999b.
Poland, M. P., Determining the stress regime within a volcanic
edifice from igneous intrusions and deformation measure-
ments, Ph.D. dissertation, Arizona State Univ., Tempe, Ariz.,
2001.
Reiners, L. P., Reactive melt transport in the mantle and geo-
chemical signatures of mantle-derived magmas, J. Petrol.,
39, 1039 – 1061, 1998.
Reynolds, R. W., and D. J. Geist, Petrology of lavas from
Sierra Negra volcano, Isabela Island Galápagos Archipelago,
J. Geophys. Res., 100, 24,537 – 24,553, 1995.
Reynolds, R., D. Geist, and M. Kurz, Physical volcanology
and structural development of Sierra Negra volcano, Galá-
pagos Archipelago, Geol. Soc. Am. Bull., 107, 1398 – 1410,
1995.
Rowland, S. K., and D. C. Munro, The caldera of Volcan
Fernandina: A remote sensing study of its structure and re-
cent activity, Bull. Volcanol., 55, 97 – 109, 1992.
Rowland, S. K., D. C. Munro, and P. Perez-Oviedo, Volcán
Ecuador, Galápagos Islands: Erosion as a possible mechan-
ism for the generation of steep-sided basaltic volcanoes, Bull.
Volcanol., 56, 271 – 283, 1994.
Simkin, T., Origin of some flat topped volcanoes and guyots,
Geol. Soc. Am. Mem., 132, 183 – 193, 1972.
Simkin, T., Geology of the Galápagos Islands, in Galápagos
(Key Environments), edited by R. Perry, pp. 5 – 41, Perga-
mon, New York, 1984.
Stearns, H. T., Geology of the Hawaiian Islands, Hawaii Divi-
sion of Hydrography, Bulletin, 8, 112 pp., 1946.
Weis, D., F. A. Frey, A. Giret, and J. M. Cantagrel, Geochem-
ical characteristics of the youngest volcano (Mount Ross) in
the Kerguelen Archipelago: Inferences for magma flux,
31 of 32
 Geochemistry 3
G 10.1029/2002GC000355
Geophysics ´ PAGOS
GEIST ET AL.: EVOLUTION OF VOLCAN ECUADOR, GALA
Geosystems

lithosphere assimilation and composition of the Kerguelen
Plume, J. Petrol., 39, 973 – 994, 1998.
White, W. M., A. R. McBirney, and R. A. Duncan, Petrology
and geochemistry of the Galápagos Islands: Portrait of a
pathological mantle plume, J. Geophys. Res., 98, 19,533 –
19,564, 1993.
White, W. M., F. Albarede, and P. Telouk, High-precision
analysis of Pb isotope ratios by multi-collector ICP-MS,
Chem. Geol., 167, 257 – 270, 2000.
Wilson, D. S., and R. N. Hey, History of rift propagation and
magnetization intensity for the Cocos-Nazca spreading cen-
ter, J. Geophys. Res., 100, 10,041 – 10,056, 1995.

32 of 32