JOURNAL OF PETROLOGY
VOLUME 48
NUMBER 5
PAGES 829^852
2007
doi:10.1093/petrology/egm003
Diverse Origins of Xenoliths from Seamounts
at the Continental Margin, Offshore Central
California
A. S. DAVIS*, D. A. CLAGUE AND J. B. PADUAN
MONTEREY BAY AQUARIUM RESEARCH INSTITUTE, 7700 SANDHOLDT ROAD, MOSS LANDING, CA 95039-9644, USA
RECEIVED AUGUST 10, 2005; ACCEPTED JANUARY 31, 2007
ADVANCE ACCESS PUBLICATION MARCH 14, 2007
Xenoliths are samples of the mantle lithosphere underlying
a volcano and/or the crust the host magma traversed. Their
compositions provide information about the temperatures
and pressures at which they originated or last equilibrated.
They might record metasomatic processes that modify
the lower lithosphere during rock^melt interactions.
Mantle xenoliths have been described from various
tectonic settings, including continental rifts (e.g. Frey &
Prinz, 1978; Kempton, 1987; McGuire, 1988), island arcs
(e.g. Takahashi, 1980), and ocean islands such as Hawaii
(e.g. Sen & Presnall, 1986; Sen, 1988; Sen et al., 2005), the
Canary Islands (e.g. Neumann, 1991; Neumann et al.,
2000, and references therein), and the Society Islands (Qi
et al., 1994). Some ocean island lavas containing mantle
xenoliths also include fragments of old ocean crust (e.g.
Clague & Chen, 1986; Fodor & Vandermeyden, 1988;
Schmincke et al., 1998; Neumann et al., 2000). Other
xenoliths are cumulates of ocean island (e.g. Sen &
Presnall, 1986; Clague, 1987; Bohrson & Clague, 1988;
Fodor & Moore, 1994; Fodor & Galar, 1997) or mid-ocean
ridge magma chambers (e.g. Hekinian et al., 1985; Dixon
et al., 1986; Davis & Clague, 1990). If the xenoliths ascend
rapidly, there might be minimal interaction with their host
magma. In contrast, if they are in prolonged contact with
the melt, the xenoliths might be mineralogically and
chemically modified.
This study describes the petrography and mineral
and host lava compositions of a diverse suite of xenoliths
from Davidson and Pioneer seamounts, offshore central
California. Unlike most intra-plate ocean island volcanoes,
the seamounts are built on top of spreading center segments that were abandoned at the continental margin
when the tectonic regime changed from subduction to a
transform margin. The host lavas erupted millions of
years after mid-ocean ridge basalt (MORB) volcanism
ended (Davis et al., 1995, 2002). The xenoliths provide a
window into the upper mantle and lower crust in this
unusual environment. We use the mineral chemistry to
identify and distinguish mantle and ocean crust cumulates
*Corresponding author. Telephone: 831-775-1857.
E-mail: [email protected]
ß The Author 2007. Published by Oxford University Press. All
rights
reserved.
For
Permissions,
please
e-mail:
[email protected]
A diverse assemblage of small mafic and ultramafic xenoliths occurs
in alkalic lava from Davidson and Pioneer seamounts located at the
continental margin of central California. Based on mineral compositions and textures, they form three groups: (1) mantle xenoliths of
lherzolite, pyroxenite, and dunite with olivine of 4Fo90; (2) ocean
crust xenoliths of dunite with olivine 5Fo90, troctolite, pyroxenegabbro, and anorthosite with low-K2O plagioclase; (3) cumulates
of seamount magmas of alkalic gabbro with primary amphibole and
biotite and anorthosites with high-K2O plagioclase. The alkalic
cumulates are genetically related to, but more evolved than, their host
lavas and probably crystallized at the margins of magma reservoirs.
Modeling and comparison with experimentally derived phases
suggest an origin at moderate pressures (05^09 GPa). The high
volatile contents of the alkalic host lavas may have pressurized the
magma chambers and helped to propel the xenolith-bearing lavas
directly from deep storage at the base of the lithosphere to the eruption
site on the ocean floor, entraining fragments of the upper mantle and
ocean crust cumulates from the underlyingabandoned spreading center.
KEY WORDS: basaltic magmatism; continental margin seamounts;
geothermobarometry; mineral chemistry; xenoliths
I N T RO D U C T I O N
JOURNAL OF PETROLOGY
VOLUME 48
from xenoliths related to the alkalic volcanism that built
the seamounts. We estimate the depth of origin based on
temperatures and pressures recorded by mineral equilibria
in the xenoliths and draw inferences concerning magma
generation and transport processes.
GEOLOGIC A L S ET T I NG
Davidson and Pioneer are two seamounts of a group of
four located at the continental margin, offshore central
California (Fig. 1). Morphologically similar seamounts are
more abundant offshore southern and Baja California.
Unlike typical ocean island volcanoes or near-ridge
seamounts, all of the seamounts are complex NE^SWtrending ridges that reflect the ridge-parallel structure of
the underlying oceanic crust (Davis et al., 2002). Davidson
Seamount is built on a fossil spreading center based on
symmetric magnetic anomalies of Chron 6, indicating an
ocean crust age of about 20 Ma (Lonsdale, 1991). Mapped
Chron 6C magnetic anomalies near Pioneer Seamount are
not symmetrical about the seamount but suggest an age of
about 24 Ma for the underlying ocean crust. Published
39
Ar/40Ar laser fusion ages indicate volcanism at about
12 Ma on Davidson and at 11Ma on Pioneer (Davis et al.,
2002), younger by 8^13 Myr than the underlying oceanic
crust. New Ar^Ar incremental heating results for some
Davidson samples expand the age of volcanism at Davidson
from 17 to 10 Ma (D. A. Clague, unpublished data),
indicating that episodes of volcanism occurred on 3 to 10
Myr old ocean crust. Such prolonged volcanic activity to
form the seamounts suggests very low magma supply
rates and long hiatuses between eruptions, as suggested
based solely on seamount morphology by Davis et al. (2002).
Some whole-rock and glass chemistry data from
Davidson and Pioneer seamounts were given by Davis
et al. (2002), who presented petrography, Ar^Ar ages, and
trace element and isotope compositions for lavas from the
four seamounts offshore central California and for one
located farther south. Volcanic rocks are predominantly
alkalic basalt, hawaiite, and mugearite, but also include
some tholeiitic basalt and rare trachyte. Radiogenic isotopes indicate a variably enriched MORB source (Davis
et al., 2002; P. Castillo, personal communication).
S A M P L I N G A N D A N A LY T I C A L
M ET HODS
The xenoliths occur in volcanic rocks that were collected by
dredging on several cruises of the US Geological Survey
(USGS) in 1976,1978, and 1979 and on dives of the remotely
operated vehicle (ROV) Tiburon on three cruises of the
Monterey Bay Aquarium Research Institute’s R.V. Western
Flyer in 2000 and 2002. The xenoliths studied were selected
to include the widest variety of minerals and textures, but
they represent only a small fraction of the inclusions present
NUMBER 5
MAY 2007
in the lavas. Whole-rock lava samples were analyzed by Xray fluorescence (XRF) at the GeoAnalytical Laboratory
of Washington State University, and the standards
used, precision and accuracy are available at their
web site (http://www.wsu.edu/geology/geolab/note.html).
Minerals of xenoliths and glass of pillow rinds and of volcanic breccias were analyzed with aJEOL 8900 Superprobe at
the USGS, Menlo Park using natural and synthetic glass
and mineral standards (Davis et al.,1994). Glass and plagioclase were analyzed with a defocused beam (10 mm) and
20 nA and 15 nA specimen current, respectively. A focused
beam and 25 nAwere used for pyroxene, amphibole, biotite,
and apatite; for olivine and oxides the current was increased
to 30 nA. Back-scattered electron images of textural
features and compositional zoning were also determined
with the same microprobe rasterizing with a focused beam
and 40 nA current over a variable-sized area, according
to the area of interest. The complete analytical dataset is
available at http://petrology.oxfordjournals.org (Electronic
Appendices 1^3).
H O S T L AVA
The host lavas containing the xenoliths are alkalic basalt,
hawaiite and mugearite (Table 1, Fig. 2). The rare tholeiitic
basalt and trachyte do not contain xenoliths. One calcalkaline andesite that was also recovered is not included
here because it is inferred to be an erratic. Exotic rocks,
including granitic, sedimentary, and metamorphic rocks,
occur on all of these seamounts (Davis et al., 2002;
Paduan et al., 2004). Several of the xenoliths occur in fresh
or altered glass of volcanic breccia. The xenolith-bearing,
alkalic lavas are moderately to highly vesicular and
are typically porphyritic with variable proportions of clinopyroxene, plagioclase, olivine. Broken fragments of
large plagioclase crystals in a number of samples may be
pieces of xenoliths. The rims of these crystals are more
calcic than the cores and overlap with the compositions
of microphenocrysts and microlites (4An50, K2O
03^07 wt%). The clinopyroxene is pinkish brown, typically complexly zoned, with high TiO2 (to 7 wt%) and
Al2O3 (to 13 wt%) diagnostic of alkalic basalts (LeBas,
1962). The olivine is often replaced by clays and iron
oxides but when unaltered is Fo78^87 with 02^03 wt%
CaO and 003^025 wt% NiO.
Xenoliths are found in different flows at numerous locations on the seamounts (Fig. 1). A wide range of xenoliths,
including dunite, lherzolite, pyroxenite, amphibole-gabbro,
anorthosite, and amphibole and titanomagnetite megacrysts has been found; these are especially abundant in
one mugearite sample (L2-79NC-D1-R35) recovered in a
dredge from the northern flank of Davidson Seamount.
Representative major element compositions of 54 wholerock samples and 93 glasses are listed in Table 1 and shown
in Fig. 2. Whole-rock compositions range from tholeiitic
830
DAVIS et al.
XENOLITHS FROM SEAMOUNTS
(a)
(b)
T627
T119
T603
T142
S4-78-D6
L2-79-D1
T139
S5-79-D13
T429
T146
T147
T426
T427
T430
T141
T144
T140
124°
38°
Pioneer
120°
San Francisco
Monterey
T145
36°
Davidson
Los Angeles
34°
100 km
contour interval 500 m
32°
Fig. 1. Illuminated Simrad EM300 bathymetric images of (a) Pioneer and (b) Davidson seamounts showing ROV Tiburon dive tracks and
USGS dredge locations. Black lines indicate tracks where xenoliths were recovered. Contour interval is 500 m. Inset map shows location of
seamounts relative to the continental margin.
831
JOURNAL OF PETROLOGY
VOLUME 48
NUMBER 5
MAY 2007
Table 1: Representative major element compositions of whole rocks and glasses
Major element compositions
Davidson Seamount
Pioneer Seamount
Sample:
T429R24
79D1-37*
T427R2
T429R4
T140R16
T145R10
T603R6
T627R1
76D4-4
Rock type:
Trachyte
Mug
Haw
Haw
AB
TranB
Haw
AB
ThB
SiO2
604
509
470
486
486
502
517
474
502
Al2O3
191
177
166
176
159
146
176
170
145
TiO2
086
242
351
271
201
FeO
372
679
861
767
895
MnO
008
019
014
045
022
013
013
015
CaO
242
750
884
894
962
911
866
936
200
274
112
302
784
223
107
109
015
112
MgO
076
380
628
498
839
725
364
556
553
K2O
505
292
259
227
178
125
208
164
045
Na2O
606
472
388
352
316
314
432
323
366
P2O5
042
146
085
138
077
032
083
091
Total
988
984
983
982
993
992
995
037
989
992
Glass compositions
Davidson Seamount
Pioneer Seamount
Sample:
T429R17
T430R14
T139R6
T426R5
T140R10
T145R5
T119R15
T119R5
S4-78D4-7*
Rock type:
Ben
Mug
Haw
Haw
AB
ThB
Mug
Haw
Haw
SiO2
586
476
508
472
503
498
493
468
472
Al2O3
191
182
182
176
174
163
154
178
164
TiO2
125
322
250
331
240
FeO
450
903
774
963
869
MnO
013
015
015
016
CaO
231
893
879
966
MgO
127
376
448
439
K2O
392
292
263
240
016
209
104
322
126
354
103
366
113
015
019
017
020
961
625
982
901
550
657
304
433
363
199
081
282
248
251
101
Na2O
466
442
336
437
342
336
453
360
434
P2O5
062
090
078
077
067
031
078
093
088
Total
964
992
996
995
1007
996
983
999
993
AB, alkalic basalt; ThB, tholeiitic basalt, TranB, transitional basalt; Haw, hawaiite; Mug, mugearite. Ben, benmoreite.
*Full cruise identification is L2-79NC and S4-78NC.
and mildly alkalic basalt to trachyte but hawaiite compositions are most abundant (Fig. 2). Glass rims of lava are
typically more evolved than their corresponding wholerock compositions, but trends parallel those of the wholerock samples. The dense, aphyric trachyte has no glass
rind. Similar to other ocean island suites (e.g. fig. 1 of
Nevkasil et al., 2004, and references therein), the greatest
variability, especially in alkalis, occurs over a narrow
range of silica contents (48^50 wt%, Fig. 2). More
evolved compositions show better developed trends,
but they also represent a limited number of samples.
MgO, CaO, FeO, and TiO2 decrease with increasing SiO2
whereas Na2O and K2O, and to a lesser extent Al2O3,
increase. P2O5 has the greatest scatter (figures not shown).
DESC R I PT ION OF X ENOL I T H S
A N D M E G AC RY S T S
Petrographic descriptions are summarized in Table 2,
images of representative thin sections are shown in Fig. 3,
832
DAVIS et al.
XENOLITHS FROM SEAMOUNTS
12
sa
n
it e
10
8
CA seamounts
(Davis et al., 2002)
Ba
Na2O + K2 O (wt.%)
the other three samples have Fo87^88 and the small crystals
forming the mosaic at the margin of D3 is Fo861. Except
for these small crystals with 026 wt% CaO, all are typically low in CaO (001^005 wt%) relative to olivine in
the host lava (Fig. 5a). NiO content is 4030 wt% for samples with 4Fo90 (Table 3, Fig. 5b) but lower Fo olivine has
correspondingly lower NiO (020^030 wt%). The spinel,
occurring only in the three samples with lower Fo olivine,
has TiO2504 wt% and Al2O3 contents of about 30 wt%
or less (Table 4).
itic
Trachyte
ol e
Benmoreite
on hrit
h
P ep
T
Mugearite
Trachyandesite
Haw
Dacite
6
Andesite
Bas.
And.
4
Alkalic
Basalt
2
44
Thol.
Basalt
48
52
Whole rock
Glass
56
60
64
Lherzolite and pyroxenite
68
SiO2 (wt%)
Fig. 2. Whole-rock and glass compositions of lavas recovered from
Davidson Seamount on an alkali vs silica plot that shows the range
from tholeiitic and mildly alkalic basalt to trachyte. Hawaiite compositions are most abundant. Field of lava compositions from five seamounts offshore central California (Davis et al., 2002) includes some
from Davidson and Pioneer Seamount. Data are normalized to 100%.
Analytical errors are indicated. Classification is that of Cox et al. (1979).
Haw, hawaiite; Bas. And., basaltic andesite;Thol., tholeiitic.
and back-scattered electron (BSE) images of selected areas
are shown in Fig. 4. Xenoliths include dunite, lherzolite,
pyroxenite, troctolite, anorthosite, and gabbro. The
gabbro xenoliths can be divided into three groups based
on the presence or absence of amphibole and whether the
amphibole is primary or secondary. Megacrysts are
plagioclase, amphibole, clinopyroxene, and titanomagnetite (Table 2, Fig. 3). Except for two 10 cm gabbro samples
from Davidson Seamount, xenoliths are small (505 to
5 cm) inclusions in crystalline or glassy basaltic lava or volcaniclastic breccia. Megacrysts are single large crystals
ranging in size from 4 to 9 mm that were identified as
xenocrysts based primarily on mineral compositions,
discussed below.
Dunite
The five analyzed dunite xenoliths (Table 2) are small
(5 to 8 mm) and angular with some planar surfaces
and/or rounded corners (Fig. 3a). All have fractures and
joints in two or more directions; some are severely sheared.
Brown iron oxide alteration lines many of the fractures
and replaces some olivine, which is commonly strained.
Textures are predominantly fine- to medium-grained,
allotriomorphic^granular (Pike & Schwarzman, 1977),
except for sample D3 (Fig. 4a), which has a narrow layer
of a mosaic of small, anhedral olivine crystals at the
margin. Other than this margin, contacts with the host
lavas are typically sharp and mostly without reaction
rims. Three samples contain subhedral to rounded Cr-rich
spinel crystals to 05 mm in size.
Only two dunites (D3, D5) have olivine 4Fo90 (Table 3,
Fig. 5, and Electronic Appendix 1). Olivine compositions in
Four lherzolite xenoliths range from triangular to blocky in
shape (Table 2, Fig. 3b). There are few if any reaction rims,
although abundant blebs and stringers of lava have penetrated into some of the xenoliths along fractures (Fig. 4b).
Fractures or joints in at least two directions are typically
present. Only sample L1 (Fig. 3b) is deformed, although it
does not have well-developed foliation. This sample is the
only one to which the textural term porphyroclastic (Pike
& Schwarzman, 1977) is applicable. Olivine crystals are
strained and may be partially or completely replaced by
iddingsite and/or iron oxide. Orthopyroxene crystals typically have undeformed clinopyroxene exsolution lamellae
(Fig. 4b). Spinel is present in only two samples (L1, L2) in
the form of minute, anhedral crystals along pyroxene crystal boundaries.
The olivine compositions are comparable with those in
the dunites (Table 3, Fig. 5, and Electronic Appendix 1).
One sample has Fo87^88 and the other three have 4Fo90.
CaO ranges from 001 to 006 wt% and NiO from 030 to
045 wt%. Clinopyroxene is calcic Cr-diposide, and orthopyroxene is enstatite (Fig. 6a). Both pyroxenes have high
Mg-numbers ranging from 87 to 94 (Fig. 6b and c). TiO2
in clinopyroxene is low (000^04 wt%) and Cr2O3 is high,
up to 29 wt% (Table 5, Electronic Appendix 2), relative to
that of the host lavas. Al2O3 in clinopyroxene is also typically low, ranging from 08 to 3 wt%. Only small subhedral clinopyroxene neoblasts along the margins and in the
fractures of the deformed lherzolite (L1) are more aluminous (4^7 wt%). Despite the somewhat higher Al2O3,
they are uniformly low in TiO2 (003 wt%, Fig. 6c) and
have some of the highest Cr2O3 contents (to 26 wt%).
Spinel in sample L3 is more aluminous than in the dunites
(37 wt%) but lower in TiO2 (007 wt%). Rare spinel in
the deformed lherzolite (L1) is the most Cr2O3-rich
(475 wt%, Table 4, Fig. 7).
The two pyroxenites are small (56 mm), angular slivers
of mostly orthopyroxene with clinopyroxene occurring as
small crystals and as undeformed exsolution lamellae.
One sample (P1) has euhedral spinel inclusions up to
1mm in size (Fig. 3c). The pyroxenes have Mg-numbers
ranging from 86 to 94 (Table 5, Electronic Appendix 2).
The clinopyroxene is low in TiO2 (to 017 wt%) and
Al2O3 (54 wt%, Fig. 6c). Spinel compositions are similar
833
JOURNAL OF PETROLOGY
VOLUME 48
NUMBER 5
MAY 2007
Table 2: Summary of xenoliths and xenocrysts
Sample
Full sample identification
Size (cm)/shape/dominant texture
Dominant minerals
Remarks
Dunite
D1
L2-79NC-D1-31
05/angular/allotrio
Oli, sp
No reaction rim, fractured
D2
L2-79NC-D1-53G
07/angular/allotrio
Oli, sp
Sharp contact, fractured
Mosaic margin of small olivine
D3
T147-R28x2
08/triangular
Oli
D4
T426-R5.2
07/triangular
Oli, sp
Highly fractured
D5
L2-79NC-13D-2H
07/triangular
Oli
Sharp contact, fractured
Lherzolite
L1
L2-79NC-D1-1
12/triangular/porph
Opx4oli4cpx
Deformed; vesicular host glass
L2
L2-79NC-D1-2
30/blocky/hypid
Opx4cpx4oli
Olivine replaced by iddingsite
L3
T141-R13a
14/angular/allotrio
Opx, oli, cpx
No reaction rim
L4
T147-R28x1
15/angular/allotrio
Opx4cpx4oli
Exsolution lamellae
P1
L2-79NC-D1-35I.1
04/triangular
Opx, cpx, spinel, oli
Euhedral spinel (mm-sized)
P2
L2-79NC-D1-35J
06/angular
Cpx4opx
Severe reaction rims; thin opx lamellae
L2-79NC-D1-53H
15/angular/allotrio
Oli, plag, spinel
Tiny, anhedral to subhedral spinel
Olivine replaced by iddingsite
Pyroxenite
Troctolite
T1
Gabbro/no amphibole
Gn1
L2-79NC-D1-42
20/rectangular/hypid
Cpx plag4oli
Gn2
T147-R2a
100/partly disaggregated
Plag4cpx4oli
Severe reaction rims
Gn3
T147-R2b
100/partly disaggregated
Cpx4plag4oli
Severe reaction rims
Gabbro/secondary amphibole
Gs1
L2-79NC-D1-3
30/blocky/allotrio
Plag4cpx4oli, amph
Olivine to iddingsite/Fe-oxide
Gs2
L2-79NC-D1-15
20/angular/allotrio
Plag4cpx4oli, amph
Reaction rims, Fe–Ti oxide lamellae
Gs3
L2-79NC-D1-35I.2
13/angular to rounded
Plag4cpx4opx, amph
Severe reaction rims
Embayed margins
Gabbro/primary amphibole
Gp1
L2-79NC-D1-26A
11/angular/allotrio
Plag4amph
Gp2
L2-79NC-D1-27A
00/rounded, hypid
Plag4amph, il
Gp3
L2-79NC-D1-27K
10/triangular/poikilitic
Amph4plag
Amph encloses euhedral plagioclase
Gp4
L2-79NC-D1-35F
12/blocky
Plag4bio4amph
Euhedral apatite in plagioclase
Gp5
L2-79NC-D1-35H
12/blocky/allotrio
Amph cpx4plag
Amphibole replacing cpx
Gp6
L2-79NC-D1-35S
13/rounded/poikilitic
Amph4plag
Euhedral plagioclase
Gp7
L2-79NC-D1-52
20/disaggregated
Amph4plag, ox, ap
Plagioclase has reaction rims
Gp8
T141-R13b
14/blocky/allotrio
Plag4amph, ox, ap
Ilmenite þ magnetite
Anorthosite
A1
L2-79NC-D1-27C
10/angular to rounded
Plag (An43–53)
No reaction rim
A2
L2-79NC-D1-35W
22/blocky
Plag (An13–14), zircon
Brecciated, extensive reaction rims
A3
L2-79NC-D1-53C
11/blocky to rounded
Plag (An44–48)
Apatite inclusions
A4
T142-R15b
15/elongate
Plag (An58–61)
Embayed along cleavages
A5
T144-R11x
19/ovoid
Plag (An58–65)
No reaction rim
A6
T146-R14
23/rounded
Plag (An19–26)
Severe reaction rim
A7
T147-R1x
21/elongate to rounded
Plag, ox, ap
Vesicular host glass
A8
T147-R3
09/elongated
Plag (An25–31), ap
Resorbed margins
A9
T426-R5.1
16/rounded to blocky
Plag (An26–41), ox, ap
Dark brown glass intrusion
A10
T427-R2
17/tabular
Plag (An28–32)
Devitrified inclusions, fractures
A11
S4-78NC-6D-16
37/elongated
Plag (An38–43), ox, ap
In vesicular host glass
(continued)
834
DAVIS et al.
XENOLITHS FROM SEAMOUNTS
Table 2: Continued
Sample
Full sample identification
Size (cm)/shape/dominant texture
Dominant minerals
Remarks
M1
L2-79NC-D1-26B
05/ovoid
Amphibole
No reaction rim
M2
L2-79NC-D1-35C
09/ovoid
Amphibole
Embayed margin, no reaction rim
M3
L2-79NC-D1-27I
08/angular
Amphibole
Broken crystal, no reaction rim
M4
L2-79NC-D1-35I.3
05/rounded
Amphibole
No reaction rim
M5
L2-79NC-D1-35L
11/amoeboid
Titanomagnetite
Embayed margins
M6
S5-79NC-13D-3
06/rounded
Augite
Compositionally zoned
Megacrysts
Because of small sample size, modal mineralogy (i.e. 1000 point counts) was not determined. Size is maximum dimension.
opx, orthopyroxene; cpx, clinopyroxene; oli, olivine; plag, plagioclase; amph, amphibole; ox, Fe–Ti oxide; il, ilmenite; ap,
apatite; bio, biotite; allotrio, allotriomorphic–granular; hypid, hypidiomorphic–granular; porph, porphyroclastic.
to those in dunite and lherzolite, having low TiO2
(5010 wt%) and Al2O3 contents of 32 wt% (Table 4).
Troctolite
One xenolith consists of plagioclase and olivine, which
encloses small (501mm), subhedral spinel crystals. The
angular, 15 cm fragment has an equigranular texture
(Fig. 3d). There is no reaction rim and virtually no alteration of either olivine or plagioclase crystals. The plagioclase is labradorite (An58^65) with a low K2O
(010 wt%) content (Table 6, Fig. 8). The olivine is Fo85
with a high NiO (to 03 wt%) and low CaO (5007 wt%)
content. The small Cr-rich spinel crystals are highly aluminous (375^435 wt%) and have low TiO2 (003 wt%)
contents.
Gabbro without amphibole
Three gabbro xenoliths do not contain amphibole. One of
these (Gn1) is composed of plagioclase and clinopyroxene
with traces of olivine replaced by iddingsite and Fe-oxide.
It is medium to coarse grained (5 mm). Contact with the
host lava is sharp and without reaction rims. Only one
fracture extends through this xenolith and into the host
lava (Fig. 3e). As no lava has penetrated into the fracture,
fracturing must have occurred post-emplacement, possibly
during sample collection or preparation. The plagioclase is
low-K2O (020 wt%) labradorite, and the clinopyroxene
is low in TiO2 (05^15 wt%) and Al2O3 (37^6 wt%)
relative to pyroxene in the host lava (Fig. 6b). The
Mg-numbers of clinopyroxene in Gn1 range from 77 to
80 (Table 5, Electronic Appendix 2).
The two other gabbro samples (Gn2, Gn3), the largest of
all the xenoliths (10 cm), are composed of clinopyroxene,
plagioclase, and minor olivine. They have basically the
same lithology except that one has a larger proportion
of pyroxene relative to plagioclase than the other.
Both samples are almost disaggregated by their host
lava (Fig. 3f). The plagioclase is highly anhedral with
embayed margins and zones of sieve-texture, with some
crystals having a narrow rim of a more calcic overgrowth.
Compositions are higher in K2O than for Gn1, ranging
from An54 and 04 wt% K2O (Table 6) for cores to An70
and 03 wt% K2O (Fig. 8) for rims. The clinopyroxene
crystals are also anhedral with severely embayed margins
and devitrified glass inclusions. Most crystals are optically
zoned, and Mg numbers range from 71 to 767. The TiO2
(14^42 wt%) and Al2O3 (62^112 wt%) contents of clinopyroxene are significantly higher than those of Gn1
(Fig. 6c), indicating that they crystallized from an alkalic
melt. The unaltered olivine is Fo76^80 and NiO ranges
from 002 to 024 wt% and CaO from 014 to 025 wt%
(Table 3, Fig. 5). The compositions of olivine cores in these
two xenoliths are lower in Fo and CaO, but similar in NiO,
to those of the rims, which overlap with the olivine composition in the surrounding lava (Fig. 5a).
Gabbro with secondary amphibole
Three gabbro xenoliths contain dark brown, dusty-looking
amphibole that we interpret to be secondary because
it occurs as discontinuous, anhedral inclusions within clinopyroxene, or as a replacement along crystal margins and
in fractures (Fig. 3g). All three samples are medium- to
coarse-grained, allotriomorphic^granular in texture and
have extensively reacted with the host melt. Two (Gs1,
Gs2) are composed primarily of plagioclase, clinopyroxene, and minor olivine, replaced by iddingsite. One
sample (Gs2) has pyroxene with crisscrossing Fe^Ti oxide
lamellae (Fig. 4c) that may have replaced orthopyroxene.
A third sample (Gs3) has relict orthopyroxene lamellae
that are too narrow to analyze. The feldspar in Gs1 and
Gs2 is labradorite (An52^65) with a low K2O content
(014^022 wt%), whereas Gs3 has some core compositions
of labradorite of An450, K2O 03 wt% but rims are
835
JOURNAL OF PETROLOGY
VOLUME 48
(a)
NUMBER 5
MAY 2007
(b)
(c)
Spinel
(d)
Plag
Oli
(e)
Plag
(f)
Cpx
Cpx
Plag
Fig. 3. Photomicrographs of thin sections of xenoliths and megacrysts: (a) dunite (D5) with4Fo90 olivine; (b) porphyroclastic lherzolite (L1) in
vesicular glass; (c) pyroxenite (P1) with large spinel crystal; (d) troctolite (T1) in vesicular lava; (e) amphibole-free gabbro (Gn1) with low-K2O
plagioclase; (f) amphibole-free gabbro (Gn2) with high-K2O plagioclase; (g) gabbro with secondary amphibole replacing clinopyroxene;
(h) gabbro (Gp6) with euhedral plagioclase poikilitically enclosed in amphibole; (i) amphibole gabbro (Gp7) partially disaggregated by the
host lava; (j) rounded anorthosite (A5) of labradorite without reaction rim; (k) angular andesine anorthosite, embayed along cleavages;
(l) ovoid amphibole megacryst (M1); (m) amoeboid Fe^Ti oxide megacryst (M5); (n) compositionally zoned clinopyroxene (M6) in crosspolarized light. Scale bar represents 5 mm.
836
DAVIS et al.
XENOLITHS FROM SEAMOUNTS
(h)
(g)
Plag
Cpx
Amph
(j)
Pl
ag
(i)
Amph
(m)
(l)
(k)
(n)
Fig. 3. Continued.
837
JOURNAL OF PETROLOGY
VOLUME 48
NUMBER 5
MAY 2007
more sodic and have K2O up to 1wt% (Fig. 8). Rim compositions of plagioclase in the xenoliths are comparable
with those in the host lava and probably reacted with the
host melt. The unaltered clinopyroxene cores of Gs1 and
Gs2 have Mg-numbers ranging from 72 to 91. TiO2 and
Al2O3 range from 06 to 16 wt% and 28 to 55 wt%,
respectively. The clinopyroxene in Gs3 is too altered to
analyze. The amphiboles in all three samples have SiO2 of
41 1wt%, with TiO2 ranging from 37 to 54 wt% and
K2O from 09 to 134 wt% (Table 7). The lower K2O and
TiO2 compositions are limited to round or anhedral inclusions in the clinopyroxene. Only these three gabbro xenoliths show extensive mineral replacement indicative of
modal metasomatism (e.g. Kempton, 1987).
(a)
Ol
Ol
1 mm
Gabbro with primary amphibole
(b)
Ol
Cpx
Glass
Opx
Glass
0.5 mm
(c)
Fe-Ti oxide
Cpx
Amph
1mm
Fig. 4. Back-scattered electron (BSE) images: (a) the mosaic formed
by small iron-rich olivine crystals along margins of a dunite xenolith
(D3); (b) clinopyroxene occurring as small crystals and as exsolution
lamellae in orthopyroxene of lherzolite L1; (c) amphibole inclusions
and Fe^Ti oxide lamellae cross-cutting clinopyroxene with orthopyroxene lamellae in gabbro xenolith Gs2.
Eight gabbro xenoliths consist of medium- to coarsegrained, hypidiomorphic^granular aggregates of predominantly plagioclase and brown, strongly pleochroic
amphibole; only one of these (Gp5) also contains clinopyroxene. We interpret the amphibole to be magmatic in
origin because it occurs as optically homogeneous, large
crystals (to 09 cm), bounded in part by crystal faces, and
because it poikilitically encloses plagioclase. Four samples
are mostly plagioclase with minor amphibole inclusions or
anhedral amphibole attached at the margins, whereas two
others (Gp3, Gp6) consists mostly of amphibole poikilitically enclosing small, euhedral plagioclase crystals
(Fig. 3h). The amphibole of Gp5 forms coronas around
the clinopyroxene. Traces of anhedral biotite are present
along the margins of amphibole crystals in three samples
and euhedral apatite and Fe^Ti oxide are included in plagioclase of several samples (Table 2). Rare iron sulfide
(pyrrhotite) inclusions occur in some amphiboles. One of
the gabbro samples (Gp1) is almost disaggregated by the
host lava (Fig. 3i). The amphibole, spanning a narrow compositional range (Table 7), is kaersutite [classification of
Leake (1978)] with SiO2 ranging from 38 to 397 wt%,
TiO2 from 45% to 8 wt%, and K2O from 11 to 14 wt%
(Fig. 9). The biotite has 8^9 wt% K2O and 4^8% wt TiO2.
The clinopyroxene in sample Gp5 is moderately high in
TiO2 (19 wt%) and Al2O3 (75 wt%), comparable with
some in the host lavas. The feldspar compositions are predominantly andesine but range from labradorite to oligoclase with K2O from 021 to 10 wt%. Reverse zoning is
ubiquitous, and the edges of feldspar crystals in contact
with the host lava have an overgrowth of labradorite, comparable with the compositions of small crystals in the host
lava (Fig. 8). Oxide present is mostly titanomagnetite, but
one sample has both titanomagnetite and ilmenite (Gp8),
and a third has only ilmenite (Gp2, Table 8). Euhedral apatite crystals are enclosed in feldspar and have F contents to
14 wt% and Cl to 05 wt% (Table 9).
838
DAVIS et al.
XENOLITHS FROM SEAMOUNTS
Table 3: Representative olivine compositions
Dunite
Sample:
Lherzolite/pyroxenite
D1
D2
D3
D3*
D4
D5
401
400
403
411
1327
112
458
475
SiO2
406
402
FeO
121
118
MgO
475
471
920
491
D5
886
494
410
908
495
L1
L3
411
835
497
405
L4
L4
P1
412
408
406
119
861
472
860
488
491
754
502
CaO
003
004
004
029
001
001
002
005
003
001
003
007
MnO
019
016
011
018
023
014
014
011
019
010
009
012
NiO
033
024
041
020
027
044
036
036
031
045
038
Total
Fo
036
1007
995
993
998
995
999
1001
996
1001
992
990
990
876
877
905
857
884
917
906
914
875
910
911
916
Troctolite
Gabbro
Host lava
Sample:
T1
T1
T1
Gn2
Gn2
Gn3
S579-13D-2H
S579-13D-2H
T147-R2A
T147-R2B
T147-R28X2
T426-R5
SiO2
398
401
399
383
386
381
396
402
385
388
394
392
FeO
121
143
145
199
194
214
161
143
189
181
144
148
MgO
449
448
447
407
414
393
439
453
416
420
449
450
CaO
007
005
006
014
017
014
026
021
021
023
026
026
MnO
022
022
022
033
027
037
024
019
029
022
025
023
NiO
025
025
029
008
002
006
015
020
010
007
019
018
Total
999
997
996
994
999
993
1003
1004
996
993
994
996
Fo
845
848
846
766
792
766
829
850
797
806
847
845
*Small crystals at margin.
Anorthosite
Eleven xenoliths consist primarily of feldspar. Inclusions of
euhedral apatite and subhedral titanomagnetite are present in or attached to the feldspar of six samples. In one
sample (A2), feldspar encloses a large (08 mm) zircon
crystal (Table 2). The anorthosite xenoliths range in size
from about 1 to 4 cm and occur in two basic shapes:
nearly elliptical with rounded corners (Fig. 3j) or
elongated^tabular with deeply embayed margins that are
aligned along cleavage planes (Fig. 3k). Fractures in two
or more directions are often present. Devitrified glass
inclusions in feldspar are highly abundant in some samples.
Reaction rims are virtually absent in some samples,
whereas others have embayed and sieve-textured margins
and alteration along fractures and cleavage planes.
Compositionally, the feldspar is predominantly andesine
but ranges from An65 to An13 with correspondingly
increasing K2O (030^185 wt%, Fig. 8). Reaction rims
are more pronounced in samples with more sodic feldspar
and are most severe for the oligoclase (An13) of the
zircon-bearing xenolith (A2). All analyzed crystals are
reversely zoned, including labradorite crystals that show
no evidence of resorption. As observed for the primary
amphibole-gabbros, titanomagnetite and apatite inclusions
occur only in the more sodic plagioclase and have compositions (Tables 8 and 9) comparable with those in the gabbros with primary amphibole. Coexisting ilmenite and
titanomagnetite were found in one sample (A9).
Megacrysts
Large, single crystals of amphibole, titanomagnetite, feldspar, and clinopyroxene present in some lava samples
appear to be xenocrysts, based on compositions.
Distinctive, large (to 09 cm) rounded amphibole crystals
(Fig. 3l) are present in hawaiite and mugearite lava samples (Table 2) that also contain amphibole-gabbro xenoliths. Similar in size and pleochroism, and with
comparable kaersutite compositions (TiO2 55 to 7 wt%
and K2O 12 wt%, Fig. 9), they appear to be disaggregated xenoliths. Their shape is largely ovoid and some
have embayed margins. One centimetre-size, amoeboid
titanomagnetite crystal (M5, Fig. 3m), present in the most
xenolith-rich mugearite, may also be a xenocryst because
it is higher in Al2O3 and TiO2 than oxide crystals in
the host lavas or anorthosite xenoliths, suggesting that it
crystallized from a more evolved magma composition.
839
JOURNAL OF PETROLOGY
0.3
VOLUME 48
0.2
Gorda Ridge
Xenoliths
CaO (wt.%)
Gn2&3
cores rims
Cayman Rise
0.1
Hess Deep
Dunite
Lherzolite
Pyroxenite
Troctolite
Gabbro/no
amphibole
0.0
(b)
0.4
DISCUSSION
Origin of xenoliths
NiO (wt.%)
0.3
0.2
0.1
Gn2&3
xenoliths
cores rims
0.0
76
78
Gorda Ridge
Xenoliths
80
MAY 2007
One centimetre-size, clear, compositionally zoned clinopyroxene megacryst (M6, Fig. 3n) has a core low in TiO2
and Al2O3 (075 wt% and 55 wt%, respectively, Fig. 6b
and c) comparable with the pyroxene in the amphibolefree gabbros. From core to rim, the crystal becomes progressively higher in Al2O3 over a narrow range in TiO2.
Other complexly zoned, lavender-coloured clinopyroxene
crystals and plagioclase are abundant in most lava samples
and may be disaggregated from gabbro xenoliths like Gn2
and Gn3. Because their compositions overlap with those of
the host lava (Fig. 6b and c), we do not consider them separately but have included their compositions in the field for
host lavas.
(a)
Host lava
NUMBER 5
82
84
86
88
90
92
Fo (mol.%)
Fig. 5. Variation of olivine composition (mole% Fo) vs (a) CaO and
(b) NiO content for magnesian olivine of dunite, lherzolite, and
pyroxenite xenoliths relative to those of ocean crust or alkalic gabbro
(Gn2 and Gn3). Compositions of olivine from the Hess Deep (Allan
& Dick, 1996), in xenoliths from Cayman Rise (Elthon, 1987) and
Gorda Ridge (Davis & Clague, 1990), as well as in the host lavas are
shown as fields for comparison.
Despite its volumetric significance, direct knowledge of the
composition of the lower ocean crust and underlying
mantle comes from relatively few studies of rocks exposed
in fracture zones and a few Ocean Drilling Program
(ODP) drill sites. As our interpretation of the origin of
the xenoliths relies primarily on mineral compositions
and rock textures, we summarize pertinent data from the
Hess Deep (Hekinian et al., 1993; Allan & Dick, 1996) and
Mid-Cayman Rise (Elthon, 1987) because these studies
provide detailed mineral chemistry. Fields for these minerals are shown for comparison with our xenolith and xenocryst compositions in Figs 5^8. The Hess Deep drill site
895 provides a view into the lower crust and upper mantle
under a fast-spreading center, whereas the Cayman Rise
study provides a view into magma chambers beneath
a slow-spreading and dying ridge segment, presumably
analogous to the abandoned spreading center under
Table 4: Spinel compositions
Lherzolite
Sample:
TiO2
Al2O3
L1
000
196
Pyroxenite
L2
006
L2
007
P1
011
Troctolite
P1
007
T1
Dunite
T1
003
003
T1
003
D1
018
D2
032
D2
031
D4
031
D4
029
315
369
313
324
435
394
375
232
296
291
307
300
109
104
118
120
133
148
153
157
153
154
168
172
FeO
913
Fe2O3
542
364
317
285
288
650
588
611
620
547
543
569
MnO
016
017
015
020
021
020
021
031
028
030
026
031
MgO
165
167
178
163
163
166
149
144
122
141
139
133
519
026
130
Cr2O3
475
361
314
374
365
201
239
260
403
359
361
330
351
Total
983
990
1000
999
1003
1001
991
996
980
1008
1006
1001
1009
Mg-no.
763
731
753
711
709
690
643
627
581
622
618
586
573
Cr-no.
619
435
364
446
430
236
289
318
549
449
435
434
447
Mg-number ¼ 100Mg/(Mg þ Fe2þ); Cr-number ¼ 100Cr/(Cr þ Al).
840
DAVIS et al.
XENOLITHS FROM SEAMOUNTS
Davidson Seamount. At Hess Deep, the stratigraphic section extends from MORB, through diabase and isotropic
gabbro, into gabbroic cumulates and mantle rocks of
lherzolite and harzburgite (e.g. Hekinian et al. 1993;
(a)
Di
Hd
host lava
Lherzolite
Pyroxenite
Gabbro/
no amphibole
Gabbro/ primary
amphibole
Gabbro/ second.
amphibole
Megacryst
En
6
Fs
(b)
Dunite
Lherzolite
Pyroxenite
Gabbro/
no amphibole
Gabbro/ second.
amphibole
Megacryst
host lava
5
4
3
2
Cayman Rise
Gorda Ridge
TiO2 (wt.%)
1
2σ
0
64
4
72
80
88
96
Mg#
Mantle xenoliths
(c)
host lava
3
2
Cayman Rise
Gorda Ridge
1
2σ
0
0
2
4
6
Allan & Dick, 1996; Dick & Natland, 1996). Diagnostic
mineral compositions of the mantle rocks are highly
magnesian olivine (4Fo90) with high NiO and low CaO
contents (Fig. 5), clino- and orthopyroxene and spinel with
high Mg-numbers. Both spinel and clinopyroxene have low
TiO2 and relatively high Cr2O3 contents (Figs 6 and 7).
Although some of the Hess Deep ultramafic and mafic
rocks have undergone complex wall-rock^melt interaction,
olivine with 4Fo90, high NiO and low CaO contents in
dunite is undoubtedly of mantle origin. Olivine in gabbro
cumulates extends to lower Fo compositions but overlaps
with mantle olivine in the higher Fo range. However,
CaO in these olivines is consistently lower than in phenocrysts in ocean floor basalt, reflecting slow cooling and/or
greater pressure. Spinel compositions in cumulate gabbros
are distinctly lower in Mg-number and have several times
greater TiO2 contents than those of mantle rocks.
No mantle rocks were recovered at the Mid-Cayman
Rise (Elthon, 1987) but the suite of gabbros, troctolite, and
anorthosites described show the diversity of magma cumulates present under a slow-spreading center. The olivine in
these rocks spans a large range in Fo (88^73) and NiO contents (026^011wt%) but all have low CaO contents
(003^008 wt%), reflecting slow cooling of deep-seated
rocks. Likewise, clinopyroxenes have Mg-numbers ranging
from 88 to 62, with TiO2 contents higher than for any of
the mantle rocks but typical for tholeiitic compositions.
Spinel compositions have an enormous range (Mg-numbers 60^10) presumably as a result of re-equilibration at
lower temperatures (Elthon, 1987). Plagioclase compositions ranges from 4An70 to An35, but all are low in K2O
(002^022 wt%), typical of normal MORB (N-MORB).
8
10
12
Al2O3 (wt.%)
Fig. 6. Pyroxene compositions shown in (a) Ca^Mg^Fe ternary,
(b) TiO2 vs Mg-number and (c) TiO2 vs Al2O3 plots. Mantle xenoliths are highly magnesian and exceedingly low inTiO2 over a narrow
range of Al2O3, whereas xenoliths of alkalic affinity have high concentrations of these two oxides. Ocean crust xenoliths are intermediate in
composition. It should be noted that the salitic pyroxene compositions
of alkalic samples extend beyond the Di^Hd boundary because of
their high CaO content (Basaltic Volcanism Study Project, 1981).
The mineral compositions of the lherzolite, pyroxenite,
and dunite with 4Fo90 olivine indicate a mantle origin for
these xenoliths. The high Mg-numbers for both clino- and
orthopyroxene with low TiO2 and Al2O3 contents (Fig. 6),
as well as the low CaO and high NiO contents of olivine
(Fig. 5), are comparable with those in harzburgite and
spinel lherzolite recovered from beneath the ocean crust
in Hess Deep (e.g. Hekinian et al., 1993; Allan & Dick,
1996). Compositions of spinel enclosed in olivine in two of
the lherzolites (L1, L2) and in clinopyroxene in one of the
pyroxenite (P1) samples are low in TiO2, similar to those
from Hess Deep (Fig. 7), although the spinel in the porphyroclastic sample (L1) has a somewhat higher Crnumber. Except for L1, all plot within the field for abyssal
peridotites (Dick & Bullen, 1984). The olivine is only
slightly strained and has no kink bands except for that in
the porphyroclastic lherzolite (L1). The lamellae in the
orthopyroxene are undeformed, indicating that no deformation occurred after exsolution. The mantle xenoliths in
the seamount lavas appear less deformed than the Hess
Deep mantle rocks and, unlike them, show no evidence
841
JOURNAL OF PETROLOGY
VOLUME 48
NUMBER 5
MAY 2007
Table 5: Representative pyroxene compositions
Host lava
Sample:
T426-R5
SiO2
421
TiO2
Al2O3
Lherzolite
79-13DH2
454
444
320
125
109
Pyroxenite
L1
L1
L2
L2
L3
L3
L4
P1
P2
514
569
528
565
541
562
528
534
540
019
00
002
003
002
001
007
017
007
422
156
274
282
247
227
287
343
166
FeO
740
817
275
556
179
519
256
672
213
208
182
Cr2O3
001
000
263
053
093
064
115
060
098
067
034
MnO
014
013
011
MgO
107
119
175
CaO
212
194
198
Na2O
066
096
992
1001
992
Wo
504
456
En
357
393
Fs
138
Mg-no.
721
SiO2
520
237
069
005
988
425
07
529
911
151
46
722
919
007
162
225
000
084
986
999
484
14
482
892
82
34
918
934
015
318
086
006
008
008
007
182
165
177
224
223
236
013
096
009
986
997
995
993
473
17
451
470
471
477
863
515
488
495
95
50
121
34
42
34
904
905
877
939
922
935
Gabbro with secondary amph
Gabbro with primary amphibole
Megacrysts
Gn2
Gn3
Gs1
Gs1
Gs2
Gp5
Gp5
M6
M6
490
444
477
507
533
494
482
472
509
492
086
148
Al2O3
368
633
FeO
782
740
Cr2O3
007
007
022
014
326
Gn1
TiO2
MnO
005
005
168
999
Gabbro without amphibole
Gn1
040
060
Total
Sample:
014
348
018
387
196
072
036
156
154
189
075
128
719
495
403
545
654
748
528
674
625
804
653
281
700
616
711
743
627
010
005
003
055
001
029
020
055
054
109
013
018
016
002
024
017
017
019
011
MgO
167
145
115
129
134
174
129
139
130
177
149
CaO
186
191
209
203
221
206
221
209
207
156
194
Na2O
Total
045
132
090
074
069
063
058
069
085
058
066
1003
993
990
990
993
997
992
985
986
990
990
Wo
386
422
498
454
480
436
484
462
466
351
429
En
487
450
385
405
409
517
396
431
409
526
464
Fs
128
129
117
141
112
47
121
107
126
124
109
Mg-no.
792
778
767
741
786
917
767
801
765
810
809
Mg-number ¼ atomic 100 Mg/(Mg þ Fe2þ), where all iron is FeO.
for retrograde metamorphism in the form of serpentine
and greenschist minerals.
The textures and the exceedingly low TiO2 content of
clinopyroxene and spinel suggest that these xenoliths are
of depleted upper mantle from which N-MORB has been
extracted. Glass penetrating into L1 is that of the host
magma and not interstitial glass produced by partial melting. Similar mantle xenoliths have been found in alkalic
lavas from all over the world, including continental rifts
(e.g. Ellis, 1976; McGuire, 1988) and ocean islands such as
Hawaii (Sen, 1988), Tahiti (Qi et al., 1994), and the Canary
Islands (e.g. Neumann, 1991; Neumann et al., 2000, and
references therein). These xenoliths are fragments of upper
mantle entrained in ascending alkalic basalt magma.
Ocean crust cumulates
The low-K2O in plagioclase of the troctolite and several
gabbro samples (Gn1, Gs1, Gs2, Gs3), as well as the low
842
DAVIS et al.
60
(a)
XENOLITHS FROM SEAMOUNTS
Hess Deep
harzburgite
100Cr/(Cr+Al)
50
40
Hess Deep
gabbro
30
2σ
20
Gorda
xenoliths
Abyssal
peridotite
dunite
lherzolite
pyroxenite
troctolite
1.0
(b)
Gorda xenoliths
TiO2(wt.%)
0.8
0.6
Hess Deep
gabbro
0.4
0.2
2σ
Hess Deep
harzburgite
0
30
40
50
60
70
80
100Mg/(Mg+Fe2+)
Fig. 7. Compositions of spinel in dunite, lherzolite, and troctolite
xenoliths in terms of (a) Cr-number and (b) TiO2 vs Mg-number.
The low-Ti spinel of the mantle xenoliths suggests depletion
comparable with that of the Hess Deep harzburgite. The ocean crust
spinel is somewhat lower inTi than in the Hess Deep gabbro or Gorda
Ridge xenoliths. Except for the higher Cr-spinel of lherzolite L1,
all plot within or near the field for abyssal peridotite spinel of Dick
& Bullen (1984).
TiO2 and Al2O3 in clinopyroxene in these gabbros, indicates crystallization from tholeiitic melts (LeBas, 1962;
Basaltic Volcanism Study Project, 1981). These gabbros,
together with the dunite with olivine of 5Fo88, are similar
to dunite and gabbro (Figs 5^7) overlying the more
depleted mantle peridotite at Hess Deep (e.g. Hekinian
et al., 1993; Allan & Dick, 1996), and also closely resemble
ocean crust gabbros from the Mid-Cayman (Elthon, 1987)
and East Pacific Rise (Hekinian et al., 1985). They are
probably ocean crust cumulates from the underlying,
abandoned spreading center.
The medium- to coarse-grained, allotriomorphic^ to
hypidiomorphic^granular textures of the troctolite and
gabbro samples are typical of magmatic textures, as are
the twinning and zoning of clinopyroxene and the poikilitically enclosed spinel in the olivines of the troctolite and
dunites. These spinel compositions are lower in Mgnumber and higher in TiO2 than in the lherzolite and pyroxenite (Fig. 7) and suggest an ocean crustal origin.
Although the cores of plagioclase with K2O5020 wt%
are labradorite with 4An55, rims of some crystals at the
margins of the xenoliths have lower An and higher K2O
overgrowth, probably as a result of reaction with the host
lava (Fig. 8). One anorthosite xenolith (A10) has plagioclase that plots along the low-K2O trend of the ocean
crust xenoliths (Fig. 8), but the An20^32 compositions suggest that it crystallized from a melt more evolved than
typical ocean-ridge basalt (e.g. Basaltic Volcanism Study
Project, 1981; Bryan, 1983; Stakes et al., 1984; Davis &
Clague, 1987, 1990). Similar, only slightly less sodic plagioclase from inclusions in Mid-Cayman Rise basalts was proposed to have crystallized from the last melt remnants at a
dying spreading center (Elthon, 1987). Because Davidson
Seamount is built on top of an extinct spreading center
(Lonsdale, 1991), this xenolith could be related to the final
stage of spreading. Similar gabbro, dunite, and troctolite
inclusions found in lavas from ocean islands (e.g. Hawaii,
Clague & Chen, 1986; Schmincke et al., 1998; Canary
Islands, Neumann et al., 2000) have also been interpreted
as ocean crust cumulates.
The compositionally zoned clinopyroxene megacryst
(M6, Table 2, Fig. 3n) is probably also an ocean crust fragment. Its low TiO2 and Al2O3 composition indicates crystallization from a tholeiitic parent magma (e.g. LeBas,
1962; Basaltic Volcanism Study Project, 1981; Davis &
Clague, 1990). Alternatively, it could be a cumulate from a
tholeiitic magma related to seamount formation. Tholeiitic
basalt is rare among the seamount lava samples, but one
sample was recovered from Pioneer Seamount, and several
samples of tholeiitic to transitional basalt were recovered
on the deepest dive (T145) at the southernmost end of
Davidson Seamount. An Ar^Ar age of 10 Ma (D. A.
Clague, unpublished data) for one of these basalts suggests
that it is not from an early shield-building stage but is one
of the youngest eruptions, postdating the eruption of the
xenolith-rich lavas.
Alkalic cumulates related to the host magmatic system
With the exception of anorthosite (A10) mentioned above,
the feldspar of the anorthosite xenoliths and of the primary
amphibole-gabbros have K2O contents (Fig. 8) consistently
higher than those found in ocean crust gabbros, indicating
that they crystallized from alkalic magmas. The high K2O
and TiO2 contents of the amphibole are also distinct from
that of amphibole in MORB gabbros (Fig. 9). Kaersutite
and biotite in the gabbros and as megacrysts are diagnostic
of a volatile-rich, alkalic melt. Salitic clinopyroxene,
high in TiO2 and Al2O3, occurs in three gabbro samples
843
JOURNAL OF PETROLOGY
VOLUME 48
NUMBER 5
MAY 2007
Table 6: Representative feldspar compositions
Troctolite
Sample:
T1
Gabbro without amphibole
T1
Gn1
Gn1
Gabbro with secondary amphibole
Gn2
Gn2
Gn3
Gn3
Gs1
Gs1
Gs2
Gs2
Gs3
SiO2
525
529
507
526
526
539
522
550
514
530
513
545
544
Al2O3
307
305
310
305
299
283
301
277
309
298
310
286
280
FeO
009
001
014
014
035
029
039
032
026
021
010
031
MgO
002
001
007
000
005
008
008
007
000
000
002
002
CaO
130
130
136
128
125
110
128
106
133
121
133
018
000
107
110
K2O
008
018
016
013
043
054
035
067
016
020
018
031
031
Na2O
420
409
358
423
400
486
385
507
387
459
375
521
506
Total
1006
1007
993
1004
998
990
998
994
999
999
997
997
991
An
627
629
671
621
617
538
634
515
648
584
654
520
536
Ab
368
360
320
372
358
430
345
447
342
404
335
461
446
Or
05
11
09
07
25
32
21
39
09
12
11
18
18
Gabbro with primary amphibole
Sample:
Gp1
Anorthosite
Gp3
Gp3
Gp4
Gp4
Gp5
Gp6
Gp7
Gp8
Gp8
A2
A2
A4
SiO2
554
527
519
589
578
574
563
557
577
561
642
641
528
Al2O3
284
295
301
253
266
269
276
279
263
271
217
219
291
FeO
039
025
030
032
029
025
032
032
027
031
019
020
MgO
004
002
004
001
002
003
002
002
003
004
001
001
687
808
887
940
977
814
921
268
285
CaO
102
117
127
037
008
120
K2O
053
035
029
084
068
067
054
055
074
056
185
185
045
Na2O
541
451
386
691
637
617
604
559
619
566
864
852
419
Total
1004
990
992
992
998
1003
1002
999
994
990
993
994
990
An
494
578
634
337
396
425
448
475
402
457
131
139
597
Ab
475
402
349
614
565
537
522
493
555
510
762
753
377
Or
30
20
17
49
40
38
31
32
43
33
107
108
27
Anorthosite
Sample:
A4
A5
SiO2
534
532
Al2O3
283
291
A5
A6
A6
522
583
627
300
260
227
A7
A8
A9
A9
559
604
578
588
279
244
266
260
A10
A10
A10
A11
A11
633
600
610
572
587
231
250
246
274
263
FeO
058
033
033
022
022
031
029
021
017
009
007
013
051
051
MgO
008
009
010
003
001
003
002
003
002
001
001
000
003
005
779
396
989
591
838
767
428
664
576
889
791
066
149
049
119
060
065
044
036
040
059
073
CaO
K2O
Na2O
113
049
452
118
041
432
133
035
370
640
813
552
719
621
680
903
749
801
617
677
Total
987
993
1000
994
992
1000
994
998
1001
1003
996
999
1008
1008
An
563
586
651
386
193
483
290
412
369
202
322
278
428
375
Ab
408
390
329
575
720
489
641
554
594
773
658
700
539
583
Or
29
24
21
39
86
29
69
35
37
25
21
23
34
42
844
DAVIS et al.
XENOLITHS FROM SEAMOUNTS
features suggest that these xenoliths are genetically related
to the alkalic volcanism that built the main edifices
of these seamounts and that they probably crystallized
at the margins of the magma reservoirs or along
transport paths.
1.8
Troctolite
Gabbro/no
amphibole
Gabbro/second.
amphibole
Gabbro/primary
amphibole
Anorthosite
1.6
1.4
K2O (wt.%)
1.2
Depth of xenolith origin
Mantle xenoliths
1.0
host lava
0.8
0.6
Cayman Rise
0.4
A10
0.2
0.0
Gorda Ridge
Xenoliths
20
30
40
50
60
70
An (mol.%)
Fig. 8. Plagioclase compositions show a trend of higher K2O with
decreasing An content for the xenoliths of alkalic affinity. The ocean
crust gabbros have plagioclase compositions that overlap with those
from Cayman Rise and Gorda Ridge but most are less calcic than
the N-MORB xenoliths from the Gorda Ridge. Plagioclase compositions from Hess Deep are off scale (4An70).
(Gn2, Gn3, Gp5) and is typical of alkalic lava, in contrast
to the Cr-diopside of the ocean crust and mantle samples
(LeBas, 1962; Basaltic Volcanism Study Project, 1981).
Olivine, present only in the two largest amphibole-free
gabbro xenoliths (Gn2, Gn3), is reversely zoned with cores
lower in Fo and in NiO than the rims or in the host lava.
This zoning suggests that they crystallized from a melt
more evolved than the host lava. All of these samples exhibit textures that are clearly magmatic, such as large crystals (to 09 cm) bounded in part by crystal faces;
amphibole poikilitically enclosing euhedral plagioclase,
plagioclase enclosing euhedral apatite and/or subhedral
magnetite; and twinning and zoning in the clinopyroxene.
Although compositional zoning is common in the pyroxene and ubiquitous in the feldspar, the kaersutite is relatively homogeneous for a given sample. All of the
kaersutite analyzed, whether in gabbro or occurring as
megacrysts, shows a narrower compositional range
(Fig. 9) than for amphibole in xenoliths of similar
alkalic lavas from the Canary Islands (Neumann et al.,
2000) suggesting crystallization from similar alkalic melt
compositions. The more sodic, higher K2O cores of plagioclase, and the presence of apatite, titanomagnetite, and
especially of zircon inclusions in plagioclase indicate
that the xenoliths formed from alkalic melts similar to,
but more evolved than, the host lavas. All of these
The presence of mantle xenoliths unequivocally indicates
that the host magmas originated in the upper mantle.
Dense mantle xenoliths, albeit small ones, preclude prolonged storage in shallow reservoirs, where they would
have settled out (Clague, 1987). Spinel lherzolite xenoliths
demonstrate that the host magma rose from a depth of at
least 25 km, below the plagioclase stability field (08^
1GPa, e.g. Gasparik, 1984; Sen, 1985). Despite the many
thermometers published for spinel lherzolite (e.g. Wood &
Banno, 1973; Wells, 1977; Mercier, 1980; Lindsley 1983;
Lindsley & Andersen, 1983; Sen, 1985; Brey & Ko«hler,
1990), there are no reliable geobarometers available.
Within and between the various thermometers, there are
large uncertainties, but collectively calculated temperatures range from 800 to 11008C, with most from 9008 to
10008C. Because the two-pyroxene thermometer of Brey &
Ko«hler (1990) is based on the largest dataset, and they
provided a comparative assessment of the various other
two-pyroxene geothermometers, we used it to calculate
temperatures ranging from 11188 to 8198C for the samples
with coexisting clino- and orthopyroxene (Table 10). These
temperatures are within the range of published values as
well as those determined experimentally for the spinel
peridotite stability field (900^11008C, 08^16 GPa, e.g.
Gasparik, 1984; Sen, 1988). Because exsolution lamellae
are present, the range we calculated probably represents
subsolidus temperatures. Similar temperatures were calculated for spinel lherzolite xenoliths from Hawaii (Sen et al.,
2005), the Society Islands (Qi et al., 1994) and the Canary
Islands (Neumann, 1991); for the latter two estimated pressures were 12^16 GPa.
The clinopyroxene structural geobarometer of Nimis
(1999) can be applied to a range of pressures and compositions and does not require knowledge of the equilibrium
liquid compositions (Nimis & Ulmer, 1998; Nimis, 1999).
Using the major element composition of clinopyroxene
alone, pressures can be calculated for anhydrous (BA) or
hydrous basalt (BH) and for more evolved liquid compositions along the tholeiitic (BT) and mildly alkaline (MA)
trends. Except for the BA model, pressures are temperature dependent and small changes in temperature can
result in large uncertainties in pressure (02 GPa). Using
the temperatures determined with the Brey & Ko«hler
thermometer, we obtained pressures ranging from 09
to 17 GPa (Table 10) with the BT model for the
845
JOURNAL OF PETROLOGY
VOLUME 48
mantle xenoliths. As higher temperatures yield lower
pressures, these are maxima.
Ocean crust xenoliths
The xenoliths inferred to be ocean crust cumulates are
similar to gabbros from the Mid-Cayman Rise (Elthon,
1987) and to gabbros and some dunites overlying the harzburgite drilled at ODP site 895 near the Hess Deep (Allan
& Dick, 1996, Figs 5^8). Elthon (1987) proposed that the
Mid-Cayman Rise gabbros did not crystallize from typical
low-pressure (1atm to 02 GPa) MORB magmas but probably formed at moderate pressure (05^10 GPa) within
deep-seated magma chambers under this slow-spreading
NUMBER 5
MAY 2007
center, after spreading ceased. The ocean crust xenoliths
from Davidson, like those from the Cayman Trough,
formed within ocean crust layer 3. Pressures, however,
must be less than 10 GPa because the xenoliths contain
abundant plagioclase. Pressures estimated using geobarometers based on mineral compositions are relatively
unconstrained. Pressures calculated arbitrarily at 10008
and 11008C for clinopyroxene in gabbro and the clinopyroxene megacryst range widely from 068 to 13 GPa
(Table 10), and are greater than our physical estimates presented below. Pyroxene compositions in samples Gs1^Gs3
may not be suitable for these calculations because of
secondary amphibole alteration.
Table 7: Amphibole and biotite compositions
Gabbro with secondary amphibole
Gabbro with primary amphibole
Sample:
Gs1
Gs1
Gs2
Gs2
Gs2*
Gp1
Gp2
Gp2
Gp3
Gp3
Gp5
Gp5
SiO2
408
414
396
402
401
389
387
397
397
384
391
375
TiO2
373
383
544
450
495
692
721
586
577
710
514
438
Al2O3
140
140
141
112
150
148
151
141
141
151
151
163
FeOT
118
116
111
112
108
108
101
108
110
100
114
127
MnO
011
010
019
015
016
015
013
018
016
012
020
MgO
121
123
127
126
120
122
125
128
129
123
120
CaO
119
116
120
118
120
118
116
111
112
117
114
010
163
001
Na2O
237
258
209
217
231
234
223
247
233
239
230
058
K2O
134
124
117
134
087
117
116
118
119
115
141
924
F
020
021
021
022
022
014
018
031
031
010
023
041
Cl
006
008
003
003
000
002
001
002
001
002
002
Total
984
989
986
984
984
992
989
985
986
Gabbro with primary amphibole
Gp4
Gp4
Gp4
Gp4
Gp7
SiO2
364
365
392
394
349
722
699
577
004
983
976
Megacrysts
Sample:
TiO2
984
582
827
Gp7
399
711
Gp7
Gp8
Gp8
M1
M2
M4
392
391
398
392
394
393
632
600
574
616
664
621
Al2O3
150
150
138
137
151
139
145
143
137
143
144
146
FeOT
151
152
119
122
152
114
112
118
120
115
107
111
MnO
MgO
016
140
012
140
017
022
121
119
112
115
015
139
014
021
022
016
017
018
127
120
122
123
124
124
124
117
116
112
110
113
116
116
CaO
001
002
Na2O
114
118
238
239
119
247
235
244
269
233
224
233
K2O
841
848
117
120
814
115
128
124
115
122
123
120
F
070
090
013
007
010
010
014
014
016
017
018
014
Cl
004
004
003
002
005
003
001
003
003
001
001
Total
982
984
979
984
003
006
970
1005
*Inclusion in clinopyroxene.
FeOT, total iron as FeO.
846
987
987
988
988
989
001
991
DAVIS et al.
XENOLITHS FROM SEAMOUNTS
These xenoliths presumably crystallized at normal layer
3 crustal depths but were re-equilibrated to slightly greater
pressure as the seamount grew and depressed the crust.
We can therefore estimate the depth of equilibration of the
ocean crust xenoliths by determining the depth to the
crust^mantle boundary beneath the seamounts. Miller
et al. (1992) have shown that complex tectonics resulted in
local thickening of underplated oceanic crust at the central
California margin but farther offshore, near Davidson
Seamount, they show crustal thickness of 7 km. Adding
about 2 km for the height of the seamounts to the crustal
Canary Islands
1.6
K2O (wt.%)
1.2
inclusions in cpx
0.8
kaersutite of Nevkasil et al. (2004)
at 0.93 GPa
Gabbro/ second.
amphibole
Gabbro/ primary
MORB gabbro
amphibole
Megacryst
0.4
0
0
2
4
TiO2 (wt.%)
6
8
Fig. 9. Amphibole compositions of megacrysts, alkalic gabbros, and
as secondary replacement in ocean crust gabbros on a K2O vs TiO2
(wt%) plot. The largest amount of variation is found in the inclusions
partially replacing clinopyroxene in ocean crust xenoliths (Gs2).
However, compared with amphibole compositions of xenoliths from
the Canary Islands (Neumann et al., 2000), they cluster within
a narrow compositional range. Compositions of amphibole in
MORB gabbro (Cannat et al., 1997; Gaggero & Cortesogno, 1997) are
significantly lower in TiO2 and K2O.
thickness and an additional 2^5 km estimated for isostatic
depression as a result of the seamount load, we infer that
the crust^mantle boundary was about 11^14 km deep
(equivalent to 505 GPa) when the seamounts formed.
Alkalic cumulates
All of the mineral compositions of the xenoliths and megacrysts of alkalic affinity indicate that they crystallized from
a melt similar to but more evolved than their hawaiite and
mugearite host lavas. However, the parental alkalic melts
may have crystallized at depths greater or less than the
depths of origin of the mantle and crustal xenoliths. We
have used a variety of techniques to estimate the temperature and pressure at which the alkalic cumulate xenoliths
could have crystallized from the seamount magmas.
Kaersutitic amphibole can crystallize over a considerable pressure range (e.g. Best, 1970; Dawson & Smith,
1982, and references therein) and is not diagnostic of a specific depth. It is unstable and tends to oxidize at magmatic
temperatures in shallow reservoirs and is absent at pressures above the amphibole stability field (425 GPa, e.g.
Niida & Green, 1999). Similar kaersutite inclusions and
megacrysts from continental settings in Arizona (Best,
1975), Australia (Ellis, 1976), and the Tertiary volcanic
rocks of Germany (Vinx & Jung, 1977) were interpreted as
having formed in the upper mantle and/or lower crust at
41GPa. The amphibole compositions in the gabbros and
of megacrysts indicate magmatic temperatures of 411008C
at 03 GPa, using the semi-empirical thermometer of Otten
(1984). At 05 GPa the temperatures are about 80^908 lower
and are in agreement with the ilmenite^magnetite temperature calculated for sample Gp8 (Table 10). Only one
sample (Gp5) contains clinopyroxene, which apparently
crystallized from a mildly alkalic melt, based on the Ti
and Al contents. Using the temperature obtained for coexisting magnetite^ilmenite (Andersen et al., 1983), pressures
calculated with the Nimis clinopyroxene barometer range
Table 8: Fe^Ti oxide compositions
Gabbro/primary amphibole
Anorthosite
Megacryst
Sample:
Gp2
Gp6
Gp7
Gp8
Gp8
A4
A4
A9
A9
A11
A11
M5
M5
TiO2
460
133
140
497
156
493
478
445
118
147
165
152
188
Al2O3
FeO*
034
404
805
668
599
722
096
382
589
699
105
395
148
397
049
497
537
754
547
728
603
667
813
669
635
637
MnO
074
024
062
047
069
034
027
096
076
078
049
045
MgO
748
710
377
943
431
863
905
334
353
259
576
580
681
Cr2O3
004
002
000
001
000
004
005
000
001
00
000
003
002
V2O3
Total
038
955
035
960
036
971
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
988
965
989
984
990
969
FeO*, total iron as FeO. n.d., not determined.
847
025
967
037
960
047
970
044
063
970
JOURNAL OF PETROLOGY
VOLUME 48
NUMBER 5
MAY 2007
Table 9: Apatite compositions
Anorthosite
Sample:
A11
Gabbro with primary amphibole
A11
A1
A1
Gp2
Gp4
Gp6
Lava
Gp6
Gp7
Gp7
DA1-35L*
FeO
078
071
065
052
084
045
098
081
075
075
049
MnO
019
010
016
009
014
007
009
010
010
012
018
CaO
SrO
P2O5
540
544
027
420
026
421
540
016
420
543
539
019
423
538
019
419
537
016
018
423
421
540
019
421
542
016
418
542
017
017
131
120
129
142
131
130
140
117
115
140
116
Cl
043
042
049
044
053
046
054
051
048
053
064
990
991
988
992
988
985
989
990
986
008
553
004
422
F
Total
013
542
421
992
79D13-H2
427
171
012
991
1001
*Inclusion in feldspar phenocryst.
Table 10: Temperature and pressure estimates based on mineral compositions
Sample
Type
Temperature (8C)
Opx–Cpx1
Pressure (kbar)
Ti in amphibole2
Magnetite/ilmenite3
Cpx4
(BA)
Al/hbl5
(BT)
(MA)
Mantle
L1
Lherzolite
1118
67
89
105
L3
Lherzolite
830
41
146
181
P1
Pyroxenite
819
59
172
211
Ocean crust
Gn1
Gabbro
71
130
157
Gs1
Gabbro
28
81
107
Gs2
Gabbro
15
68
91
M6
Clinopyroxene
90
129
170
M6
Clinopyroxene
74
99
123
Alkalic cumulate
Gn2
Gabbro
84
107
129
Gn3
Gabbro
47
70
90
Gp2
Gabbro
1087
Gp5
Gabbro
1090
49
68
81
89
Gp5
Gabbro
1090
55
76
90
91
Gp8
Gabbro
1020
A9
Anorthosite
92
1053
84
878
1
Brey & Köhler (1990); 2Otten (1984) at 05 GPa; 3Andersen et al. (1993); 4Nimis (1999); 5Schmidt (1992) at 05 GPa.
BA, anhydrous; TH, tholeiitic; MA, mildly alkaline model. (See text for details.)
from 049 to 055 GPa for the anhydrous (BA) and from
08 to 09 GPa for the alkaline (MA) model (Table 10).
The MA model always yields pressures significantly
higher than the BA model. Considering that the minerals
in sample Gp8 span a narrow compositional range and
that the magnetite^ilmenite thermometer is probably one
of the most reliable ones, this calculation may provide the
best constraint on pressure. Similar pressures (Table 10)
were calculated with the Al-in-hornblende barometer of
Schmidt (1992), although it may not be applicable to these
848
DAVIS et al.
0.8
XENOLITHS FROM SEAMOUNTS
(a)
CaO/ Al2O3
0.6
1 kb
0.4
3 kb
5 kb
0.2
whole rock
glass
0
65
(b)
1 kb
SiO2 (wt.%)
60
55
3 kb
50
5 kb
45
40
0
2.5
5
7.5
10
MgO (wt.%)
Fig. 10. Liquid lines of descent for Davidson Seamount lavas calculated with the MELTS program show (a) CaO/Al2O3 vs MgO
strongly decreasing, indicating significant clinopyroxene fractionation. (b) SiO2 vs MgO shows that silica content decreases at higher
pressure, making it difficult to produce trachyte at greater depths.
compositions as it was calibrated for amphibole crystallized from tonalite.
The MELTS program (Ghiorso & Sack, 1995) has been
widely used to evaluate liquid lines of descent for various
magmas. Davidson lava compositions are not well replicated at any single pressure from 01 to 09 GPa (Fig. 10)
with a range of water contents (05^1%) and oxygen fugacity at the quartz^fayalite^magnetite buffer. We used the
high- and low-SiO2 end-members in our modeling. The
observed range of lava and glass compositions clearly
requires a range of starting parental magma compositions.
The prominent decrease in CaO/Al2O3 observed in the
lava and glass compositions (Fig. 10a) requires significant
clinopyroxene fractionation. Clinopyroxene becomes the
dominant phase with increasing pressure, resulting in a
large increase in alkalis without changing the silica content
significantly in the basalt to hawaiite range, as observed for
the Davidson lavas. However, at pressures of 07^09 GPa
at low MgO contents, amphibole never appears as a crystallizing phase and the SiO2 enrichment observed in trachyte cannot be attained (Fig. 10b). The required SiO2
enrichment is possible at low pressure (01^03 GPa),
but amphibole is again absent from the crystallizing
assemblage and the Al2O3 enrichment observed in the
trachyte is not attained. We could not find a combination
of pressure and water content that produced both the SiO2
and Al2O3 enrichment observed in trachyte, and none of
the runs produced amphibole, clearly an important phase
in the evolution of these lavas. Crystallization of kaersutite
at the expense of plagioclase and garnet (not observed)
would result in higher SiO2 and Al2O3 in the melt and
match the observed trachyte compositions. In summary,
the MELTS program does not match the observed mineralogy of the xenoliths or the sequence of observed resultant
liquid compositions at any pressure or combination of
pressures. Our evaluation of its utility in modeling hydrous
alkalic basalt compositions suggests that the recognized
problem in modeling intermediate to silicic calc-alkaline
compositions involving fractionation of hornblende and
biotite (Ghiorso & Sack, 1995) extends to more alkaline
compositions as well.
Experimental studies of alkalic lavas similar to the host
lavas may provide better constraints for their evolution.
Nevkasil et al. (2004) showed that an increase in alkalis
relative to nearly constant silica, as seen in Davidson lavas
and commonly observed in alkalic lava suites, is due to the
dominance of clinopyroxene and suppression of plagioclase
in early fractionating assemblages at elevated pressures
(09 GPa, 11008C). At the same pressure, but at intermediate temperatures (1090^9408C), they found that kaersutite became a dominant phase under hydrous (05% bulk
water) conditions but was replaced by a Ti-rich biotite
under less hydrous conditions and at the lower end of the
temperature range. The higher temperature (410008C)
calculated for the amphibole-gabbro (Table 10) is close to
that in the high-pressure experiments of Nevkasil et al.
(2004) that yielded kaersutite, apatite, and ilmenite. The
compositions of these experimentally derived phases,
including plagioclase, are also similar to those in
Davidson xenoliths (Fig. 9). In agreement with earlier
experimental studies (Mahood & Baker, 1986), Nevkasil
et al. (2004) confirmed that plagioclase tends to be more
sodic at elevated pressures. Experimental studies of
MORB at high pressures (Bender et al., 1978; Green et al.,
1979) also showed a decrease of An contents in plagioclase
with increasing pressure, suggesting that this effect may be
independent of lava composition. The similarity of the
Davidson xenoliths to these experimental results suggests
that the Davidson magmas crystallized the alkalic cumulates at pressures 509 GPa, or a depth of about 24 km,
at the base of the lithosphere (Zhang & Lay, 1999).
Nevkasil et al. (2004) further suggested that water content could play a more important role than pressure,
especially as a variable in suppressing early feldspar
crystallization. Bulk water contents measured for
some hawaiite glasses from Davidson are 07 wt%
849
JOURNAL OF PETROLOGY
VOLUME 48
(Davis & Clague, 2003), but must have been much greater
at the margins of the magma reservoirs where the amphibole crystallized. The high fluorine and chlorine contents
of the amphibole and apatite (14 and 09 wt%, respectively) indicate high halogen contents in the magmas, but
their effects on crystallization have not been well investigated experimentally and so cannot be evaluated.
Transport of xenoliths to the surface
We have presented evidence that these alkalic magmas
fractionated at 07^09 GPa, at or near the base of the
lithosphere, before migrating to the surface and eruption.
During their ascent they entrained and transported some
of the partly crystallized wall rocks (alkalic cumulates
and megacrysts) of their deep magma reservoirs and crystalline xenoliths of upper mantle (509 GPa) and ocean
crustal rocks (505 GPa). These magmas probably rose to
the surface rapidly to maintain the dense xenoliths in suspension, and because of the high volatile contents we infer
for the parent magmas and the explosive character of
many of the eruptions (Davis & Clague, 2003).
The kaersutite, biotite, and apatite, all with high fluorine and chlorine contents, testify to high contents of water
and halogens in the magma. We think these volatiles accumulated and became enriched at the upper margins of the
alkalic melt pockets and in veins extending into the mantle
country rock and crystallized hydrous phases such as kaersutite and biotite. With continued fractionation, the bulk
water contents increased, thereby lowering the melt density and viscosity, and the increase in volatiles pressurized
the system and eventually propelled the lava to the eruption site on the sea floor. Many of the xenoliths and megacrysts are contained in highly vesicular hyaloclastite
breccias, demonstrating the explosivity of the eruptions
(Davis & Clague, 2003).
The rapid ascent rate of the host magma is supported by
the lack of diffusion of CaO in unaltered olivine (Klu«gel,
1998), by thin or absent lava selvages on mantle and some
ocean crust xenoliths, and by decompression that fractured
the xenoliths and also increased the temperature of the
host lava. The increased temperatures caused extensive
resorption of sodic plagioclase and the formation of thin
overgrowths of more calcic plagioclase. The rounded
shapes of amphibole and embayed margins of titanomagnetite megacrysts suggest dissolution and ablation of these
lower temperature phases. Many of the fractures along
congugate joint surfaces that were filled by host lava probably originated at this time as a result of rapid
decompression.
Individual eruptions might have been initiated by, or at
least were aided by, regional extensional tectonics.
Synchronous volcanism occurred at geographically widely
separated places on- and offshore along the continental
margin during the Miocene that might have been related
to movement along major faults as the tectonic regime
NUMBER 5
MAY 2007
changed from a convergent to a transform margin
(e.g. Davis et al., 1995, 2002; Dickinson, 1997). Bailey (1970,
1972) suggested that lithospheric structures focused areas of
alkalic volcanism in continental rift settings. The alignment of volcanic cones at Davidson Seamount parallel to
the ocean crust fabric suggests a similar pattern with
ascending melts channeled along existing zones of
weakness.
S U M M A RY A N D C O N C L U S I O N S
The three types of xenoliths included in the seamount
lavas provide information on the characteristics of the
mantle and crust underlying the volcanoes. The mantle
xenoliths indicate that the initial melt rose from mantle
depths below the plagioclase stability field (1GPa). The
ocean crust xenoliths are cumulates from the final stages
of spreading centers that were abandoned when the tectonic regime changed from a divergent to a transform
margin. The alkalic gabbros, anorthosites, and kaersutite
and titanomagnetite megacrysts are cumulates formed at
moderate pressure (05^09 GPa) at the base of the lithosphere from melts that are genetically related to, but more
evolved than, the host lava. Increasing water and other
volatile constituents decreased magma density and pressurized the system, leading to rapid ascent and eruption
on the ocean floor.
AC K N O W L E D G E M E N T S
Most of the xenolith samples were collected in dredges carried out by Gary Greene and Brent Dalrymple on several
USGS cruises in the late 1970s. Four xenoliths are from
dives T141 and T142 by Peter Lonsdale and Pat Castillo
and three additional xenoliths were collected by Andrew
DeVogeleare on dives T426 and T427, funded by NOAA’s
Ocean Exploration Program. These principal investigators
kindly made the additional samples available to us,
thereby increasing the number of samples available. We
thank the ROV Tiburon pilots and the captain and crew of
the Western Flyer for their skill in recovering the dive samples. Robert Oscarson assisted with microprobe analysis.
We thank Gautam Sen and Paolo Nimis for sharing their
EXCEL spreadsheets for geothermobarometry. Reviews by
M. Coombs, R. Fodor, S. Keshav, and especially Editor W.
Bohrson of an earlier version greatly improved the manuscript. The support of the David and Lucile Packard
Foundation through a grant to MBARI is gratefully
acknowledged.
S U P P L E M E N TA RY DATA
Supplementary data for this paper are available at Journal
of Petrology online.
850
DAVIS et al.
XENOLITHS FROM SEAMOUNTS
R EF ER ENC ES
Allan, J. F. & Dick, H. J. (1996). Cr-rich spinel as a tracer for melt
migration and melt^wall rock interaction in the mantle: Hess
Deep, leg 147. In: Mevel, C., Gillis, K. M., Allan, J. F. & Meyer,
P. S. (eds) Proceedings of the Ocean Drilling Program, Scientific Results,
147. College Station, TX: Ocean Drilling Program, pp. 157^172..
Andersen, D. J., Lindsley, D. H. & Davidson, P. M. (1993). QUILF: a
Pascal program to assess equilibria among Fe^Mg^Mn^Ti oxides,
pyroxenes, olivine, and quartz. Computers and Geosciences 19(9),
1333^1350.
Bailey, D. K. (1970). Volatile flux, heat focusing and the generation
of magma. GeologischesJahrbuch, Special Issue 2, 177^186.
Bailey, D. K. (1972). Uplift, rifting and magmatism in continental
plates. Journal of Earth Sciences 8, 225^239.
Basaltic Volcanism Study Project (1981). Basaltic Volcanism on the
Terrestrial Planets.New York: Pergamon Press, 1288 pp.
Bender, J. F., Hodges, F. N. & Bence, A. E. (1978). Petrogenesis of
basalts from the project FAMOUS area: experimental studies
from 0 to 15 kilobars. Earth and Planetary Science Letters 41, 277^302.
Best, M. G. (1975). Mantle-derived amphibole within inclusions and
kindred megacrysts in basanitic lavas, Grand Canyon, Arizona.
Contributions to Mineralogy and Petrology 27, 25^45.
Bohrson, W. A. & Clague, D. A. (1988). Origin of ultramafic xenoliths
containing exsolved pyroxenes from Hualalai volcano, Hawaii.
Contributions to Mineralogy and Petrology 100, 139^155.
Brey, G. P. & Ko«hler, T. (1990). Geothermobarometry in fourphase lherzolites II. New thermobarometers and practical assessment of existing thermobarometers. Journal of Petrology 31,
1353^1378.
Bryan, W. B. (1983). Systematics of modal phenocryst assemblages
in submarine basalts: petrologic implications. Contributions to
Mineralogy and Petrology 83, 62^74.
Cannat, M., Chatin, F., Whitechurch, H. & Ceuleneer, G. (1997).
Gabbroic rocks trapped in the upper mantle at the Mid-Atlantic
Ridge. In: Karson, J. A., Cannat, M., Miller, D. J. & Elthon, D.
(eds) Proceedings of the Ocean Drilling Program, Scientific Results, 153.
College Station, TX: Ocean Drilling Program, pp. 243^264.
Clague, D. A. (1987). Hawaiian xenolith populations, magma supply
rates and development of magma chambers. Bulletin of Volcanology
49(4), 577^587.
Clague, D. A. & Chen, C.-H. (1986). Ocean crust xenoliths from
Hualalai volcano, Hawaii. Geological Society of America, Abstracts 18,
565.
Cox, K. G., Bell, J. D. & Pankhurst, R. J. (1979). The Interpretation of
Igneous Rocks.London: Allen and Unwin, 450 pp.
Davis, A. S. & Clague, D. A. (1987). Geochemistry, mineralogy, and
petrogenesis of basalt from the Gorda Ridge. Journal of Geophysical
Research 92, 10467^10483.
Davis, A. S. & Clague, D. A. (1990). Gabbroic xenoliths from the
northern Gorda Ridge: implications for magma chamber processes
under slow-spreading centers. Journal of Geophysical Research 95,
10885^10990.
Davis, A. S. & Clague, D. A. (2003). Hyaloclastite from Miocene seamounts offshore central California: compositions, eruption styles,
and depositional processes. In: White, J. D. L., Smellie J. L., &
Clague, D. A. (eds) Explosive Subaqueous Volcanism. Geophysical
Monograph, American Geophysical Union 140, 129^142.
Davis, A. S., Clague, D. A. & Friesen, W. B. (1994). Petrology and
mineral chemistry of basalt from Escanaba Trough, southern
Gorda Ridge. US Geological Survey Bulletin 2022, 153^170.
Davis, A. S., Gunn, S. H., Bohrson, W. A., Gray, L. B. & Hein, J. R.
(1995). Chemically diverse sporadic volcanism at seamounts
offshore southern and Baja California. Geological Society of America
Bulletin 107, 554^570.
Davis, A. S., Clague, D. A., Bohrson, W. A., Dalrymple, G. B. &
Greene, H. G. (2002). Seamounts at the continental margin of
California: a different kind of oceanic volcanism. Geological Society
of America Bulletin 114, 316^333.
Dawson, J. B. & Smith, J. V. (1982). Upper-mantle amphiboles: a
review. Mineralogical Magazine 45, 35^36.
Dick, H. J. B. & Bullen, T. (1984). Chromian spinel as a petrogenetic
indicator in abyssal and alpine-type peridotites and spatially associated lavas. Contributions to Mineralogy and Petrology 86, 54^76.
Dick, H. J. B. & Natland, J. H. (1996). Late-stage melt evolution and
transport in the shallow mantle beneath the East Pacific Rise.
In: Mevel, C., Gillis, K. M., Allan, J. F. & Meyer, P. S. (eds)
Proceedings of the Ocean Drilling Program, Scientific Results, 147. College
Station, TX: Ocean Drilling Program, pp. 103^134.
Dickinson, W. R. (1997). Tectonic implications of Cenozoic volcanism
in coastal California. Geological Society of America Bulletin 109,
936^954.
Dixon, J. E., Clague, D. A. & Eissen, J. P. (1986). Gabbroic xenoliths
and host ferrobasalt from the southern Juan de Fuca Ridge. Journal
of Geophysical Research 91, 3795^3820.
Ellis, D. J. (1976). High pressure cognate inclusions in the Newer
Volcanics of Victoria. Contributions to Mineralogy and Petrology 58,
149^180.
Elthon, D. (1987). Petrology of gabbroic rocks from the mid-Cayman
rise spreading center. Journal of Geophysical Research 92, 658^682.
Fodor, R. V. & Galar, P. (1997). A view into the subsurface of Mauna
Kea volcano, Hawaii: crystallization processes interpreted through
the petrology and petrography of gabbroic and ultramafic xenoliths. Journal of Petrology 38(5), 581^624.
Fodor, R. V. & Moore, R. B. (1994). Petrology of gabbroic xenoliths in
1960 Kilauea basalt: crystalline remnants of prior (1955) magmatism. Bulletin of Volcanology 56, 62^74.
Fodor, R. V. & Vandermeyden, H. J. (1988). Petrology of gabbroic
xenoliths from Mauna Kea volcano, Hawaii. Journal of Geophysical
Research 93(B5), 4435^4452.
Frey, F. A. & Prinz, M. (1978). Ultramafic inclusions from San Carlos,
Arizona: petrologic and geochemical data bearing on their petrogenesis. Earth and Planetary Science Letters 38, 129^176.
Gaggero, L. & Cortesogno, L. (1997). Data report: metamorphic
mineralogy of Leg 153 gabbro. In: Karson, J. A., Cannat, M.,
Miller, D. J. & Elthon, D. (eds) Proceedings of the Ocean Drilling
Program, Scientific Results, 153. College Station, TX: Ocean Drilling
Program, pp. 531^546.
Gasparik, T. (1984). Two-pyroxene thermobarometry with new experimental data in the system CaO^MgO^Al2O3^SiO2. Contributions to
Mineralogy and Petrology 87, 87^97.
Ghiorso, M. S. & Sack, R. O. (1995). Chemical masstransfer 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. Contributions to Mineralogy and Petrology 119,197^212.
Green, D. H., Hibberson, W. O. & Jaques, A. L. (1979). Petrogenesis
of mid-ocean ridge basalts. In: McElhinny, M. W. (ed.) The Earth:
its Origin, Structure and Evolution. London: Academic Press,
pp. 265^299.
Hekinian, R., Hebert, R., Maury, R. C. & Berger, E. T. (1985).
Orthopyroxene-bearing xenoliths in basalts from the East Pacific
Rise axis near 128500 N. Bulletin de Mine¤ ralogie 108, 691^698.
Hekinian, R., Bideau, D., Francheteau, J., Cheminee, J. L.,
Armijo, R., Lonsdale, P. & Blum, N. (1993). Petrology of the East
851
JOURNAL OF PETROLOGY
VOLUME 48
Pacific Rise and upper mantle exposed in Hess Deep (Eastern
equatorial Pacific). Journal of Geophysical Research 98, 8069^8094.
Kempton, P. D. (1987). Mineralogical and geochemical evidence for
differing styles of metasomatism in spinel lherzolite xenoliths:
enriched mantle source regions of basalts? In: Menzies, M. A. &
Hawkesworth, C. J. (eds) Mantle Metasomatism.London: Academic
Press, pp. 45^87.
Klu«gel, A. (1998). Reactions between mantle xenoliths and host
magma beneath La Palma (Canary Islands): constraints on
magma ascent rates and crustal reservoirs. Contributions to
Mineralogy and Petrology 131, 237^257.
Leake, B. E. (1978). Nomenclature of amphiboles. Canadian Mineralogist
16, 501^520.
LeBas, M. J. (1962). The role of aluminum in igneous clinopyroxene
with relation to their parentage. American Journal of Science 260,
267^288.
Lindsley, D. H. (1983). Pyroxene thermometry. American Mineralogist
68, 477^493.
Lindsley, D. H. & Andersen, D. J. (1983). A two-pyroxene thermometer. In: Proceedings of the 13th Lunar and Planetary Conference. Journal
of Geophysical Research, Supplement 88, A887^A906.
Lonsdale, P. (1991). Structural patterns of the Pacific floor offshore of
peninsular California. In: Dauphin, J. P., Ness, G. & Simoneit, B.
R. T. (eds) The Gulf and Peninsular Province of the Californias. American
Association of Petroleum Geologists, Memoirs 47, 87^125.
Mahood, G. A. & Baker, D. R. (1986). Experimental constraints on
depths of fractionation of mildly alkalic basalts and associated
felsic rocks: Pantelleria, Strait of Sicily. Contributions to Mineralogy
and Petrology 93(2), 251^264.
McGuire, A. V. (1988). Petrology of mantle xenoliths from Harrat al
Kishb: the mantle beneath western Saudi Arabia. Journal of Petrology
29, 73^92.
Mercier, J.-C. C. (1980). Single-pyroxene thermobarometry.
Tectonophysics 70, 1^37.
Miller, K. C., Howie, J. M. & Ruppert, S. D. (1992). Shortening
within underplated oceanic crust beneath the central California
margin. Journal of Geophysical Research 97, 19961^19980.
Nekvasil, H., Dondolini, A., Horn, J., Filiberto, J., Long, H. &
Lindsley, D. H. (2004). The origin and evolution of silica-saturated
alkalic suites: an experimental study. Journal of Petrology 45(4),
693^721.
Neumann, E.-R. (1991). Ultramafic and mafic xenoliths from Hierro,
Canary Islands: evidence for melt infiltration in the upper mantle.
Contributions to Mineralogy and Petrology 106, 236^252.
Neuman, E.-R., Sorensen,V. B., Simonsen, S. L. & Johnson, K. (2000).
Gabbroic xenoliths from La Palma, Tenerife and Lanzarote,
Canary Islands: evidence for reactions between mafic alkaline
Canary Islands melts and old oceanic crust. Journal of Volcanology
and Geothermal Research 103, 313^342.
Niida, K. & Green, D. H. (1999). Stability and chemical composition
of pargasitic amphibole in MORB pyrolite under upper mantle
conditions. Contributions to Mineralogy and Petrology 135, 18^40.
Nimis, P. (1999). Clinopyroxene geobarometry of magmatic rocks,
Part 2: Structural geobarometers for basic to acid, tholeiitic and
mildly alkaline magmatic systems. Contributions to Mineralogy and
Petrology 135(1), 62^74.
NUMBER 5
MAY 2007
Nimis, P. & Ulmer, P. (1998). Clinopyroxene geobarometry of magmatic rocks, Part 1: An expanded structural geobarometer for
anhydrous and hydrous, basic and ultrabasic systems. Contributions
to Mineralogy and Petrology 133(1^2), 122^135.
Otten, M. T. (1984). The origin of brown hornblende in the Artfja«llet
gabbro and dolerites. Contributions to Mineralogy and Petrology 86,
189^199.
Paduan, J. B., Clague, D. A. & Davis, A. S. (2004). Evidence that three
seamounts off Southern California were ancient islands. EOS
Transactions, American Geophysical Union 85(47), 1463.
Pike, J. E. N. & Schwarzman, E. C. (1977). Classification of textures in
ultramafic xenoliths. Journal of Geology 85, 49^61.
Qi, Q., Beard, B. L., Jin, Y. & Taylor, L. A. (1994). Petrology and geochemistry of Al-augite and Cr-diopside group mantle xenoliths
from Tahiti, Society Islands. International Geology Review 36, 152^178.
Schmidt, M. W. (1992). Amphibole composition in tonalite as a
function of pressure: an experimental calibration of the Al-inhornblende barometer. Contributions to Mineralogy and Petrology 110,
304^310.
Schmincke, H. U., Kluegel, A., Hansteen, T. H., Hoernle, K. & van
den Bogaard, P. (1998). Samples from the ocean crust beneath Gran
Canaria, La Palma and Lanzarote (Canary Islands). Earth and
Planetary Science Letters 163, 342^360.
Sen, G. (1985). Experimental determination of pyroxene compositions
in the system CaO^MgO^Al2O3^SiO2 at 900^12008C and
10^15 kbar using PbO and H2O fluxes. American Mineralogist 70,
678^695.
Sen, G. (1988). Petrogenesis of spinel lherzolite and pyroxenite suite
xenoliths from the Koolau shield, Oahu, Hawaii: implications for
petrology of the post-eruptive lithosphere beneath Oahu.
Contributions to Mineralogy and Petrology 100, 61^91.
Sen, G. & Presnall, D. C. (1986). Petrogenesis of dunite xenoliths from
Koolau Volcano, Hawaii: implications for Hawaiian volcanism.
Journal of Petrology 27, 197^217.
Sen, G., Keshav, S. & Bizimis, M. (2005). Hawaiian xenoliths and
magmas: composition and thermal character of the lithosphere.
American Mineralogist 90, 871^887.
Stakes, D. S., Shervais, J. W. & Hopson, C. A. (1984). The volcanic^
tectonic cycle of the FAMOUS and AMAR Valleys, Mid-Atlantic
Ridge (368470 N): evidence from basalt glass and phenocryst compositional variations for a steady state magma chamber beneath
the Valley Midsections, AMAR-3. Journal of Geophysical Research
89(NB8), 6995^7028.
Takahashi, E. (1980). Thermal history of lherzolite xenolithsçI.
Petrology of lherzolite xenoliths from the Ichinomegata crater,
Oga peninsula, northeast Japan. Geochimica et Cosmochimica Acta 44,
1643^1658.
Vinx, R. & Jung, D. (1977). Pargasitic^kaersutitic amphibole from a
basanitic diatreme at the Rosenberg, North Kassel (North
Germany). Contributions to Mineralogy and Petrology 65, 135^142.
Wells, R. A. (1977). Pyroxene thermometry in simple and complex systems. Contributions to Mineralogy and Petrology 62, 129^139.
Wood, B. J. & Banno, S. (1973). Garnet^orthopyroxene^clinopyroxene
relationships in simple and complex systems. Contributions to
Mineralogy and Petrology 42, 109^124.
Zhang, Y.-S. & Lay, T. (1999). Evolution of oceanic upper mantle
structure. Physics of the Earth and Planetary Interiors 114, 71^80.
852